Download University of Groningen Lactococcus lactis systems biology Eckhardt

University of Groningen
Lactococcus lactis systems biology
Eckhardt, Thomas Hendrik
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to
cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2013
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Eckhardt, T. H. (2013). Lactococcus lactis systems biology: a characterization at different growth rates
Groningen: s.n.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.
Download date: 09-08-2017
Lactococcus lactis Systems Biology:
A characterization at different growth rates
Paranimfen
Ana Solopova
Martijn Herber
Printing
Printed by Koninklijke Wöhrmann B.V., Zutphen, the Netherlands
Cover
Artist impression of DNA transcription (Elisabeth Eckhardt)
ISBN: 978-94-6203-350-4
The work described in this thesis was carried out in the Molecular Genetics Group of the Groningen
Biomolecular and Biotechnology Institute (Faculty of Mathematics and Natural Sciences, University of
Groningen, The Netherlands)
Printing of this thesis was financially supported by the Faculty of Mathematics and Natural Sciences,
University of Groningen
This research is supported by the Dutch Technology Foundation STW, which is part of the Netherlands
Organisation for Scientific Research (NWO) and partly funded by the Ministry of Economic Affairs,
Agriculture and Innovation (08080)
Lactococcus lactis Systems Biology
A characterization at different growth rates
Proefschrift
ter verkrijging van het doctoraat in de
Wiskunde en Natuurwetenschappen
aan de Rijksuniversiteit Groningen
op gezag van de
Rector Magnificus, dr. E. Sterken,
in het openbaar te verdedigen op
vrijdag 31 mei 2013
om 14.30 uur
door
Thomas Hendrik Eckhardt
geboren op 27 september 1984
te Sliedrecht
Promotores: Prof. dr. O.P. Kuipers
Prof. dr. J. Kok
Beoordelingscommissie: Prof. dr. R. Siezen
Prof. dr. M. Heinemann
Prof. dr. G.J.W. Euverink
Voor Elize
Contents
Chapter 1
Introduction
Page 9
Chapter 2
Comparative transcriptome analysis of Lactococcus lactis MG1363
propagated at varying growth rates using chemostats
Page 41
Chapter 3
Metabolic regulation governs the metabolic shift from mixed-acid to
homolactic fermentation of Lactococcus lactis: A multi-level study
Page 67
Chapter 4
Transcriptional regulation of fatty acid biosynthesis in Lactococcus lactis
Page 105
Chapter 5
The role of YfiA in ribosomal stalling in Lactococcus lactis
Page 137
Chapter 6
Summary and conclusion
Page 163
Chapter 7
Addendum
Page 173
Nederlandse samenvatting voor niet-ingewijden
Page 203
Dankwoord
Page 209
Chapter 1
Introduction
9
Lactococcus lactis as an industrial work horse
The lactic acid bacteria (LAB) form a collection of Gram-positive, acid-tolerant, nonsporulating rod-shaped or coccoid bacteria with low G+C content grouped for a common characteristic: the production of lactic acid, an important metabolic end-product
of glycolysis for industrial use. The best-studied members of the LAB are Lactococcus, Lactobacillus, Leuconostoc, and Streptococcus. Other LAB family members
are also of industrial interest, since some of these species are involved in fermented
products or in food spoilage. The amount of research on LAB species like Aerococcus, Carnobacterium, Enterococcus, Oenococcus, Sporolactobacillus, Pediococcus,
Tetragenococcus, Vagococcus, and Weisella has recently increased.
Several members of the LAB have a GRAS status (Generally Recognized As Safe),
meaning that the bacteria are considered to be safe for consumption. In fact, some
LAB members are beneficial in the production, conservation and shelf life of food
products, like cheese (Lactococcus lactis) 1 and yoghurt (Streptococcus thermophilus
and Lactobacillus delbrueckii ssp bulgaricus) 2.
In general, LAB have complex nutritional requirements, due to their low biosynthetic capacities 3. For that reason these bacteria are usually found in environments
where such requirements are met, such as soil, the oral cavity, on plants or in dairy
products. LAB are used in the production of cheese, yoghurt, a variety of fermented
vegetables, sausages, beer, wine and bread. About 500 varieties of cheese are produced worldwide, and the present annual production is 107 tonnes 4. The main role
of starter cultures, deliberately added mixtures of strains of LAB, is to produce lactic
acid from the available sugar in milk, leading to a drop in the pH of the final product. Cheese starter cultures consist of cheese-specific defined or undefined mixtures
of L. lactis and often strains of Leuconostoc. The bacteria produce lactic acid from
the milk sugar lactose, which promotes the curdling of the milk. Curd and whey are
separated, after which the former is pressed and salted. In the case of cheddar cheese,
the salt is added before the pressing step. During cheese ripening, L. lactis starter
cultures contribute to flavor formation, since several of their enzymes are involved
in glycolysis, proteolysis, lipolysis and the conversion of peptides and amino acids
into flavor compounds 5. The species L. lactis encompasses the subspecies (ssp.),
lactis, cremoris, and hordniae; only the first two are used in dairy fermentations.
10
The use of microorganisms in industrial production is quite challenging. The industrial
cultures should be robust and perform optimally under the sometimes rapidly changing and challenging conditions. Cultures themselves need to be non-contaminated to
ensure consistency in the fermentation process and to guarantee a safe product with a
convenient shelf-life and with attractive flavors for consumption 6–8. During industrial
production bacteria are under a constant bacteriophage threat, they may suffer from
loss of industrial traits by failing to keep their plasmids, and are fighting acidification, starvation or temperature and/or salt shock 9,10. The principal mechanisms of
stress resistance have been extensively studied in L. lactis 11–13. These studies have
gained insight in the how L. lactis attempts to achieve homeostasis, the stable intracellular conditions that bacteria aim to maintain when facing environmental changes.
Our understanding of the regulation and adaptive decisions that L. lactis makes when
facing particular stress conditions has expanded13. Such knowledge is of industrial
importance to be able to predict starter culture performance and robustness in harsh
fermentation conditions. In order to be able to use L. lactis as a “factory”, more precise
cellular models are needed to predict fermentative behavior. Omics technologies (transcriptomics, proteomics, metabolomics) enable the study of growth and fermentation
in an integrated way. Understanding the genomic build-up, transcriptional regulation,
protein abundance and activity, as well as the relationships between the cellular components and processes they are involved in is very useful for the dairy industry which
employs L. lactis as a fermentation work horse. More and more, this also holds for
the pharmaceutical industry, which is exploring the capacity of L. lactis to serve as a
vehicle for drug delivery and vaccine production 14.
When the consumption of living microorganisms results in health benefits for the host
organism, the bacteria are classified as probiotics. Not all LAB are beneficial for their
hosts. Important clinical infections caused by Streptococcus and Enterococcus species
comprise sepsis, meningitis, pneumonia, urinary tract and surgical wound infections
15,16
. An increasing extent of antibiotic resistance is found in infections caused by variants of Streptococcus 17 and Enterococcus 18. This is without a doubt a worrying threat
for human beings.
11
The strength is to combine the –omics…
By investigating enzymes and genetic pathways of the cell, more and more has
become known about a number of cellular systems. In L. lactis, sugar utilization,
citrate fermentation, phage resistance and proteolysis among others have been studied
by biochemical and genetic techniques. For instance, very thorough analyses of the
industrially important proteolytic system of L. lactis were performed over the last
decades 19,20. The main components of the protein degradation machinery were characterized at the biochemical and genetic level, describing its functions and activities,
the interactions between the enzymes and the regulation of the genes encoding the
major players in proteolysis in L. lactis. The L. lactis protease PrtP splits extracellular
proteins 21, the oligopeptides are transported by the oligopeptide permease Opp and
di-/tripeptide transporters DtpT and DtpP 22–26. The oligopeptides are further broken
down into smaller peptides and amino acids by a large number of peptidases. Peptidases with different specificities have been described 27,28. The obtained knowledge
was essential for the engineering of flavor formation in the production of cheese 29,30.
The most important proteins in proteolysis are regulated at the transcriptional level, by
the pleiotropic transcriptional regulator CodY. Via CodY the cell senses the abundance
of the branched-chain amino acids leucine, isoleucine and/or valine. CodY represses
a number of important genes for proteins involved in proteolysis; the branched-chain
amino acids hereby act as a co-repressor of CodY 31,32. The different approaches and
techniques, and the results obtained in this way, have been used to develop insights
that allow a complete description of, in this case, the proteolytic system of L. lactis. A
similar detailed analysis has been performed of the other important pleiotropic regulator, CcpA, a transcriptional regulator involved in sugar metabolism of L. lactis 33.
By describing systems on the basis of their specific function and interactions of their
components, a full understanding of how those systems function is sought. Ultimately,
by combining all systems that operate in L. lactis, a complete understanding of the
functioning of this important industrial bacterium might be achieved. It is exactly this
integration of components that forms the strength of systems biology.
12
Delivering data for systems biology
Understanding the systems that underlie and define cellular physiology is a recurrent
theme in biology 34. Since the beginning of this millennium, thanks to the great technological progress in high-throughput analyses methods of biomolecules 35such as whole
transcriptome 36, proteome and metabolome techniques and more recently, next generation sequencing (NGS) 37, the exciting possibility of describing a cell in its entirety
is now a real option. Combining data from different–omics techniques is extremely
powerful, since a more complete overview of cellular arrangements is obtained.
A systems biology approach can give insights in the networks of different gene interactions, biochemical pathways and metabolic routes. The type of network reconstruction depends on which (interdisciplinary) experiments are used. The technological
platforms to map the changes in the most informative biomolecules, i.e. DNA, RNA,
proteins and metabolites are called genetics, transcriptomics, proteomics and metabolomics respectively 38,39.
From genes to genetics
Genomics is the study of the structure, function, plasticity, evolution and mapping of
genomes. An important achievement in the field of genomics is the discovery of DNA
as the carrier of genetic information and the decryption of its structure and code 40.
Decoding, correctly assembling and annotating the entire genomic code of organisms is another important aspect of genomics. The first genome to be fully sequenced
was the bacteriophage phi-x174 in 1977 41; the first completely sequenced bacterial
genome was that of Haemophilus influenzae in 1995 42. And in 2001 the first “complete” human genome was presented 43,44. Today, more than 1600 different publically
available prokaryotic and over 40 different eukaryotic genomes have been sequenced
to completion, and many more are in progress 45.
The first complete genome sequence of L. lactis IL1403 was published in 2001 46, after
which in 2005 strain L. lactis MG1363 was added to the genome sequence library.
By automated annotation and manual curing, the accuracy of the lactococcal genome
sequence has improved over the years. The availability of a correct annotation and
gene (cluster) categorization has been imperative for both transcriptomics and proteomics as will be discussed below.
13
From mRNA to transcriptomics
Transcriptomics is a technology used in molecular biology research to uncover expression profiles of all genes in a specific cell type or in (sub)populations of cells at a
given moment in time. The gene and promoter activity is expressed as the number
of transcripts or messenger RNAs (mRNAs). With transcripts all different types of
RNA are represented, but with transcriptomics we mean to study only messenger RNA
(mRNA) changes. The size and amount of a specific mRNA can be quantitatively
estimated by Northern blotting.
RT-qPCR (reverse transcriptase combined with quantitative polymerase chain reaction) measures the formation of double stranded DNA from mRNA templates using a
fluorescent probe. The quantity of mRNA is calculated on the basis of amplification
time. Comparing these times to a standard curve, relative gene expression levels can
be obtained 47.
With the advent of the non-radioactive DNA microarray technique, it became possible
to measure changes in expression levels of thousands of genes in a single experiment.
Together with decreasing costs and improved data analysis methods, this explains the
popularity of using DNA microarrays in current gene transcription research.
The procedure for DNA microarray analysis at the department of Molecular Genetics
is described here. In short it consists of slide preparation, RNA isolation from a bacterial culture, reverse transcription to synthesize copyDNA (cDNA), labeling of cDNA,
hybridization of labeled cDNA onto the slide and data analysis. The slides used are
either homemade, or are obtained from commercial suppliers and contain 70-meric
nucleotides of all annotated genes.
In a typical DNA microarray experiment, a comparative analysis of the genome
expression between two different conditions is performed. For example, a comparison is made between a wild-type strain and a mutant of the same strain, or between
an exponentially growing culture and a culture in stationary phase or between two
different growth conditions. After harvesting the cultures to be compared, cells are
snap frozen in liquid nitrogen to avoid stress responses and RNA degradation. Total
RNA is obtained via phenol-chloroform extraction and purification by a commercially
available RNA isolation kit. From the total RNA, cDNA is synthesized by reverse
transcription with the incorporation of amino-allyl nucleotides to allow subsequent
fluorescent labeling. The cDNA of each strain/condition is labeled differentially by a
14
fluorescent dye (like Cy3 and Cy5 or more recently by DyLight fluor probes). Comparable amounts of cDNAs are mixed and allowed to compete for binding to the oligonucleotides on the slides. This hybridization step is performed at a precise and controlled
temperature and humidity. To correct for possible labeling differences between the
dyes used, a dye-swap is a commonly employed duplication. For statistical reasons, a
biological replicate is always performed.
The data analysis begins with scanning of the microarray slides for “spot intensity”:
the amount of label is a measure for the amount of cDNA bound to the oligonucleotide
on that spot. Thus, that spot intensity is a representation of the amount of mRNA
of the gene of that spot. As the fluorescence on the spot is measured in two channels, one for each dye, a measure is obtained for the amount of gene transcript under
both conditions. Individual signals are quantified by software (for example ArrayPro
4.5 from Media Cybernetics Inc.) after subtraction of the local slide backgrounds,
and normalized to total spot intensity for all of the spots on each microarray. With
LOWESS (locally weighted scatterplot smoothing methods) normalization the signals
are compensated for non-linear dye bias, caused by e.g. variation in spotting of the
oligonucleotides. Dye-swap and replicate data are pooled and the ratios between the
average normalized intensity signals are calculated with their confidence level. We use
the p values of the t-test from a software package called Cyber-T 48. DNA microarrays
allow high-throughput and are reasonably priced. They mostly examine expression
of known annotated genes only. A more advanced microarray type is the tiling array
49
. By including oligonucleotides representing intergenic regions to slides, information about non-coding and small RNAs can also be obtained. This technique can give
information on the transcriptional structure of genes, and is more widely used to study
eukaryotic transcription 50. More recently, the development of a new generation highthroughput DNA sequencing (NGS) methods have been used for both mapping and
quantifying transcriptomes. This method, termed RNA-sequencing (RNA Seq), has
the advantage of counting transcripts rather than estimating their ratios 51. Detailed
information on the complete RNA content is acquired by NGS.
From proteins to proteomics
The study of the total protein content of a cell and their structure and function is
called proteomics. The amount of a specific mRNA does not always correlate with the
15
amount of the encoded protein 52. In addition, post-translational modifications such
as phosphorylation and methylation can influence the activity of proteins 53. In proteomics two techniques are distinguished: gel-based proteomics and non-gel-based
proteomics. In gel-based proteomics 2D-electrophoresis is used to separate proteins on
the basis of their isoelectric points and their masses. After separation of the proteins,
gels can be used for visualization, quantification or identification of proteins using e.g.
specific antibodies 54,55. Classical 1D-, 2D- or blue native gel electrophoresis methods
are not sensitive enough to examine all proteins in a cell at a given moment. To more
thoroughly investigate the hundreds of proteins in a proteome, high-throughput nongel-based techniques are increasingly used and have paved the way for a proteomics
analysis that describes all proteins of a proteome.
The most common technique is mass spectrometry (MS), a very sensitive method to
determine masses of ionized particles, that allows us to identify peptides and even
determine their sequence 56. The ionized molecules and their specific trajectory of the
fragments in a vacuum system yields an indication of the mass-to-charge ratio. The
number of ions specifies how many fragments with a specific mass-to-charge ratio
passed the detector. The combination of these values is plotted in a mass spectrum.
Another technique to analyze proteins is MALDI-TOF (matrix-assisted laser desorption/ionization – time of flight). Hereby, the sample is mixed with a matrix, after which
this matrix-peptide mixture is irradiated with a laser. This causes the combination of
peptide and matrix to be desorbed and ionized. The time of flight gives a direct measure for the mass of the peptide.
Proteins can be covalently labeled with isobaric reagents specific for certain amine
groups. In iTRAQ (isobaric tags for relative and absolute quantitation) all proteins
in different proteome samples are labeled. By mixing up to eight of such differently
labeled samples, all proteins can be measured simultaneously. Based on the different
masses of the reporters the proteins of the various experiments can still be distinguished 57. The strength of this technique lies not only in the absolute quantification
of the peptides, but also in the fact that relative amounts of proteins can be compared
between experiments.
Many different strategies exist for peptide identification and quantification, because
of the different separation (2D-electrophoresis, HPLC, affinity chromatography,
ion exchange chromatography or a combination like LC-LC) and analysis methods
(MALDI-TOF, MS, MS-MS) 58,59. Like many roads lead to Rome, all approaches pro16
vide a huge amount of data of the proteome that need to be analyzed. With the help
of computer algorithms 60 and database search programs like SEQUEST, MASCOT,
PEAKS and many others, proteins can be identified from the proteomic datasets 61.
However, knowing which proteins are present in a cell and in what abundance is not
sufficient to understand an organism in its entirety. Like stated above, phosphorylation
of proteins may influence the activity of proteins. Today, a lot of effort is put in the
study of the phosphoproteome 62.
Phosphoproteomics
Determining all the protein abundances and annotating their functions does not solve
how metabolites are shuttled through the cell. Although gene expression and protein
synthesis rates determine the enzyme abundance, the in vivo catalytic flux of an enzyme
does not solely depend on the protein count per cell. Many types of post-translational
modifications that affect enzyme activity are studied, the most common of which,
phosphorylation, is discussed below.
Post-translational modification of enzymes alters their function and activity, and can
be used by cells to modulate enzyme efficiency rapidly in response to environmental
signals 63. A very important reversible post-translational modification in bacteria is
phosphorylation 64, which can be detected by Eastern blotting and advanced mass
spectrometry phosphoproteomics 65. Prokaryotic phosphorylation typically results in
a conformational change of the protein by coupling of a phosphate group to a serine,
threonine or tyrosine residue 66,67. Even phosphorylated residues have shown to be biologically relevant. Recently, arginine phosphorylation was shown to play a functional
role in cells 68. Phosphorylation plays a dual role in the control of carbon catabolite repression 69. In order to function as a repressor, CcpA needs to bind a Ser-46
phosphorylated HPr. But before this HPr-CcpA complex is formed the histidine phosphorylation on His-15 first needs to be removed 70. Another type of phosphorylation
is that of lysine residues in for example histones, which has only been characterized
in eukaryotes, where this has been shown to be important in cell-cycle regulation 71.
In prokaryotes, lysines are not known to be phosphorylated. However, the lysines of
glycolytic enzymes in E. coli were either succinylated 72, and/or acetylated 73, showing
yet two other mechanisms of post-translational modification.
17
Metabolites to metabolomics
Metabolomics techniques aim to yield a systematic measure of all the metabolites
being present in a cell at a given moment. Metabolites are the chemical intermediates
and end-products of metabolism. With separation methods such as chromatography
and capillary electrophoresis, metabolites are sorted in fractions according to their
chemical properties. Subsequently, sorted fractions containing metabolites are subjected to a detection method. The two most common techniques for metabolic detection are MS and Nuclear Magnetic Resonance (NMR). MS identifies and quantifies
metabolites as described above for proteins. NMR is spectroscopy based on the interaction of radiomagnetic radiation with atomic nuclei that possess a magnetic moment.
The energy difference between states is very small, yielding transitions in the radio/
microwave region of the spectrum, i.e. very small energies are involved. Which makes
it safe to use in living tissues or organisms in a non-invasive, non-destructive way 74.
With NMR, metabolite- and protein structures can be solved 75,76. In metabolomics,
NMR can be used to detect and quantify a whole range of small molecules. NMR is a
non-invasive technique. An in vivo NMR setup, whereby living cells are subjected to
NMR, provides valuable insights in the different metabolic pathways in L. lactis, most
importantly glycolysis 77,78.
In a flux balance analysis (FBA) is a mathematical method to estimate the metabolic
fluxes in a steady state condition. In FBA, the cellular objective is defined; in most
cases this is optimal biomass. As a function, the metabolites of the media are given
to the metabolic network to define the input of nutrients. By setting constraints to the
flux rates, the reaction capacity is defined of the uptake system or enzyme. Finally,
with linear programming, a calculation is made that yields a prediction of the cellular
objective and potential flux values for each reaction.
Many experimental biologists use and create models these days. Visualization tools
are hereby essential to provide a useful and interpretable model for biologists 79.
Bioinformatics for data analyzes and integration
Bioinformatics has played an important role in retrieving, analyzing and classifying
biological data. For instance, automatic gene finding and annotation was enabled after
the development of a software system 42.
18
Performing the –omics techniques as described in this chapter usually generates huge
lists of transcripts and protein numbers, enzyme activities, metabolites, etc. To compare and integrate the data from these different –omics studies, generally two types of
analyses are used: correlation analysis and functional categorization.
The most common correlation analysis is the Pearson’s correlation coefficient 80. This
correlation analysis provides a measure for linear dependences between two datasets
based on relative values 80. The Pearson’s correlation coefficient is calculated from
the covariance of two different datasets; e.g. those describing the transcriptional and
proteomic fold changes of an experiment, divided by the product of their standard
deviations 80. For ranked instead of relative value data, Spearman’s rank correlation
coefficient is used to test linear dependency 81.
Functional categorization provides a data structure in which genes, enzymes, metabolic pathways, regulons etc. are grouped and then classified. Assigning different
genes, enzymes, metabolic pathways or regulons into a group based on similarity, is
called clustering. For the functional categorization, a classification database for the
investigated bacterium is provided by bioinformatics databases. The most common
functional class databases are Gene Ontologies 82, Metabolic Pathways by KEGG 83,
cluster of orthologous groups (COG) 84, protein families (PFAM) 85 and regulons. For
L. lactis all functional class databases are available and accessible on the MolGen
server (server.molgenrug.nl) 89.
With functional categorization, matching genes, enzymes, metabolic pathways and
regulons provide biological insight into the data. For instance, similarities in expression profiles of genes can be noticed upon functional categorization, when comparing
multiple strains 86 or various growth conditions 87,88. The results of these experiments
lead to a list of significantly altered genes. In order to check for significantly altered
groups in the dataset software can calculate the overrepresentation of clustering classes
in a dataset. The functional categorization algorithm of the MolGen department makes
use of overrepresentation. A functional class is considered to be overrepresented when
more ‘hits’ are scored on a hypergeometric distribution when compared against a random distribution 89. Obtaining the functional categories for an experiment can be helpful in understanding the cellular rearrangements taking place during the experiment.
19
Modeling of L. lactis
In even a relatively simple biological system like a biological cell, the interactions
and interdependencies of the constituting modules are numerous and very complex.
An abstraction in the form of a model can provide more comprehensive insights with
biological meaning. In systems biology two research approaches are commonly used.
The first one is top-down, starting with sets of high-dimensional data, i.e. measurements on many variables. By perturbing the environment of a bacterium and subsequently measuring the effects on the transcriptome, proteome and/or metabolome,
etc. the consequences of perturbations on the organism are identified. This allows
uncovering of new molecular insights. By combining different perturbations, the
general pathways a bacterium uses to regain homeostasis can be determined. Via an
iterative process of experimentation, read out and statistical analysis, the overview
becomes more and more detailed until finally a descriptive and comprehensive model
is obtained that describes the specific behavior of the investigated bacterium.
On the other hand, the bottom-up approach starts with knowledge about thoroughly
investigated and well-characterized molecular parts. The knowledge is used to synthesize mathematical models of an organism or parts of an organism. For example,
with detailed kinetic data of a certain reaction cascade in a bacterium, a first model
is made explaining the physiological behavior of the bacterium as a consequence of
the enzyme kinetics. By adding additional data, such as coming from flux balance
data, the model becomes more and more precise in its description. By perturbing the
reaction cascade, a further insight in the flexibility of the sub-system can be obtained,
but more importantly this also tests the generality of the model. Once comprehended
at a satisfying level, the sub-systems are embedded, and the elements of these submodules are linked to other sub-modules. Ultimately the aims of the bottom-up and
top-down strategies do not differ, namely a full understanding of the physiology of the
organism, but while a top-down approach uses induction to yield insights, bottom-up
modeling tries to deduce the functioning of the entire organism by giving an in-depth
description of the modules 90. In practice, a combination of both methods is often used,
and the well-defined sub-modules are integrated in the whole system models to yield
comprehensive insights into physiology 91.
In this thesis the latter modeling approach is used, i.e. a combination of reaction stoichiometry with the transcriptional activities and protein abundances of L. lactis.
20
Systems biology of L. lactis
System biology is concerned with the integration of large datasets. For one of the best
studied Gram-positive bacteria, L. lactis, large datasets combined with transcriptomics
are available 92–93. The fact that we have a lot of these data makes L. lactis a very useful
microorganism to study with a systems biology approach.
A bottom up approach to model bacterial growth was attempted by Molenaar et al. This
self-replicator model was used to explain the regulation of growth rate dependent protein allocation to different biosynthetic modules, like protein synthesis and membrane
synthesis machinery and transporters, as well as the switching from efficient to inefficient metabolic pathways at increasing growth rates, also sometimes called overflow
metabolism. The model assumes a condition of a steady state or ‘balanced growth’,
where all modules of the cell grow at the same rate. The second assumption is that
protein allocation is regulated in a manner that maximizes the growth rate. The model
is designed in such a way that, when the substrate concentration changes, the relative
amounts of the different modules can be easily computed and compared. For example,
at a higher growth rate, the model predicts that more protein synthesis is needed, resulting in a higher number of ribosomes. It also predicts that the relative amount of substrate transporters as well as relative membrane surface decreases with growth rate. As
it is known that at low nutrient concentrations cells have larger numbers of transporters
per cell compared to growth in rich media, 95,96, a decrease in total number of transporters is also expected with increasing growth rate. In this thesis a growth-rate dependent
experiment is detailed on L. lactis cells, growing at steady state in order to test the model.
Growth rate-dependent regulation of metabolism in LAB has drawn special attention
over the years. The growth rate determines whether an efficient glycolytic fermentative pathway (leading to mixed acid formation) is used or a rather inefficient pyruvate
conversion (homolactic fermentation) 97. Shifting from an efficient to an inefficient
glycolytic pathway, thereby producing side metabolites such as acetate is known as
overflow metabolism, and is not exclusive for LAB 98,99. It has been described for
Escherichia coli and Saccharomyces cerevisiae as the Crabtree effect, and is comparable to the Warburg effect in cancer cell lines 100.
L. lactis is the model organism for LAB, and a good choice to study growth-rate
dependent overflow metabolism, since its glycolytic pathways are well-studied 78.
21
Figure 1. Schematic overview of glycolysis in L. lactis. Glucose is converted into glucose-6phosphate, and subsequently metabolized into the intersection molecule pyruvate. Pyruvate
can be dehydrogenated into lactate (orange), yielding NAD+ (purple arrows). On the other
hand, via the mixed acid branch acetate, ethanol, formate, 2,3-butanediol and CO2 (blue)
are produced with a more efficient energy production in the form of ATP (green arrow) and/
or NAD+. Before energy is generated, first energy needs to be invested in the form of ATP
hydrolysis (red arrow) or NAD+ reduction (orange arrow). All enzyme names are abbreviated:
GLK, glucokinase; PGI, phosphoglucose isomerase; PFK, phosphofructokinase; ALD, fructose
bisphosphate aldolase; TPI, triosephosphate isomerase; GAPDH, glyceraldehyde phosphate
dehydrogenase; PGK, phosphoglycerate kinase; PGM, phosphoglycerate mutase; ENO,
enolase; PYK, pyruvate kinase; LDH, lactate dehydrogenase; ALS, acetolactate synthase;
ALD, acetolactate dehydrogenase; BUTA butanol dehydrogenase/ acetoin reductase; PDH,
pyruvate dehydrogenase; PTA, phosphotransacetylase; ACK, acetylkinase; ADH, alcohol
dehydrogenase; PFL, pyruvate formate lyase.
22
Glucose as a limiting substrate can accurately serve as the determining factor for L.
lactis growth rate.
Transcriptional regulation in Lactococcus lactis
Regulation of transcription in bacteria depends on extrinsic variables like the nutrition levels, salt concentration, pH, temperature and on more intrinsic factors such as
the metabolic state of the organism, concentration of regulators and metabolite. The
transcription regulation mechanisms of only those sub-modules that are relevant for
this thesis are described below.
Glycolysis
Glycolysis is the metabolic pathway that converts monosaccharides, often glucose,
into pyruvate with concomitant side- and end products like ATP and lactate. Not only
glucose can be used for glycolysis, however glucose as a carbon source is preferred
by L. lactis 101,102. Via phosphotransferase systems (PTS), and non-PTS permeases
mono- and disaccharides are taken up by the cell, where the intermediate carbohydrate
can be converted upon entry in glycolysis 78. Glucose is taken up and immediately
phosphorylated into glucose-6-phosphate (G6P) (Fig. 1). This G6P is converted into
fructose-bisphosphate (FBP) by phosphoglucose isomerase and phosphofructokinase.
For the last reaction one molecule of ATP is used. FBP is broken down by FBP aldolase
into glyceraldehyde-3-phosphate (GAP) and dihydroxyacetone phosphate (DHAP).
DHAP can be converted into GAP by triphosphate isomerase. Phosphoglyceric acid
(3-PGA) is produced from the GAP molecules by the GAP dehydrogenase and phosphate glucokinase yielding NADH and ATP. Subsequently, 3-PGA is converted into
2-PGA, then into phosphoenol-pyruvate (PEP) and finally into pyruvate, an important
intermediate molecule of the glycolytic pathway (Fig. 1).
Pyruvate can be converted by pyruvate formate lyase, yielding acetyl-CoA and formate. Pyruvate dehydrogenase also produces acetyl-CoA, and generates NADH and
CO2 when oxygen is present. Acetyl-CoA can either be converted into ethanol via
alcohol dehydrogenase, thereby generating two molecules of NAD+, or into acetate
with the help of phosphotransacetylase and acetylkinase with a net gain of one ATP
molecule. A less frequently used pathway in L. lactis is the conversion of pyruvate into
2,3-butanediol. Per 2,3-butanediol molecule formed, one molecule of ATP and two
23
molecules of CO2 are produced. Even though this pathway is not generally used under
normal conditions, improved production of 2,3-butanediol can be easily achieved by
metabolic and genetic engineering strategies
103
. The production of 2,3-butanediol,
acetate, ethanol and formate is a process that is called mixed-acid fermentation, mainly taking place under sub-optimal conditions of growth 97. L. lactis generally grows
homofermentatively, meaning that most of the pyruvate formed in glycolysis is converted into lactate with recovery of NAD+. Thus, homolactic fermentation yields less
energy per pyruvate under optimal growing conditions than mixed-acid fermentation.
How L. lactis exactly controls its carbohydrate metabolism is already for decades a
matter for scientific research. An important role is played by carbon catabolite repression (CCR). When extracellular glucose is present, the uptake and metabolic pathways
of less favored sugars are downregulated 33. CCR is caused by a complex of the proteins CcpA and HPr phosphorylated at its serine on position 46 104. This complex binds
to CcpA binding motifs (cre-sites) that lie upstream of many genes, the proteins of
which are often involved in central metabolism 88. The phosphorylation state of HPr
depends on an elegant system of phosphorylation transfer between different enzymes
(EI, EIIA, EIIB and HPr), a cascade called the phosphotransferase system (PTS).
Glycolytic intermediate PEP donates its phosphoryl group to EI, transfers it to HPr,
which finally phosphorylates EIIA and EIIB. EIIB phosphorylates imported glucose,
forming G6P. At low extracellular glucose concentrations the cell accumulates phosphorylated EII and HPr His-15 105. At high glucose concentrations, HPr His-15P gets
dephosphorylated. Via the ATP-dependent kinase/phosphatase called HprK/P, HPr can
then be phosphorylated at Ser-46 106. HPr Ser46-P can form a complex with CcpA that
results in CCR. In addition to acting as a repressor, CcpA-HPr Ser46-P is also capable
of binding to cre-sites upstream of the las-operon and activating the genes of three
important glycolytic enzymes, namely pfk, pyk and ldh 107.
Arginine metabolism
Arginine is synthesized by proteins of which the genes are clustered in three operons:
argCJDBF, gltS-argE and argGH 108. Two L. lactis membrane proteins are responsible
for the import of arginine. GltS is annotated as an arginine/ornithine ABC transporter
substrate-binding protein 3, and the ArcD1/D2 complex is a well described antiporter
that exchanges arginine for ornithine 109. The genes of the arginine deiminase (ADI)
pathway enzymes involved in arginine catabolism are all located in the arcAB24
D1C1C2TD2 cluster. Arginine metabolism via the ADI pathway leads to carbamoyl-P,
which can be used for de novo synthesis of pyrimidine. Carbamoyl-P can also be
converted into ammonia and CO2 with generation of one molecule of ATP. Ammonia
production from arginine through ADI is a way to fight acid stress
110
and has been
used in the past to discriminate between L. lactis ssp. lactis (ADI+) and L. lactis ssp.
cremoris (ADI-) 111.
Regulation of arginine metabolic gene clusters takes place by different regulators.
When arginine is limiting, the regulator ArgR forms a hexamer that functions as a
repressor of the arc catabolic cluster 112. When arginine is abundant, ArgR and AhrC
are presumed to form a heterohexamer that binds to the upstream regions of the arginine synthesis and transport clusters argCJDBF, gltS-argE and argGH, thereby blocking transcription 113. In addition, the activity of the ADI pathway is subjected to CCR.
A cre site is present in the upstream region of the arc cluster in L. lactis (unpublished
results). It has been shown in several organisms that, when glucose is present, CcpA
binds to the cre-site of the arc cluster, which leads to repression of arginine catabolism
114,115
. Another regulator that might play a role in arginine metabolism is CodY, which
is mainly involved in nitrogen metabolism (see above). Upon deletion of this global
regulator a derepression of parts of the arc regulon was observed 32. The exact role of
CodY in arginine metabolism has not been clearly identified yet.
Regulation of ribosome synthesis
Ribosomes are the protein synthesis machines of the cell. Bacterial ribosomes
consist of a large 50S subunit and a small 30S subunit. The L. lactis 50S subunit
contains a 5S and a 23S rRNA and at least 32 proteins. The lactococcal 30S subunit is made up of a 16S rRNA and 21 proteins. As is the case in most organisms,
multiple loci for the genes encoding rRNAs are present in the genome of L. lactis
MG1363. Six operons of rRNA are asymmetrically located in the proximity of the
origin of replication and contain genes coding for transfer RNA (tRNA) 3 (Fig. 2).
tRNA are RNA molecules that upon charging with an amino acid serve as a carrier to deliver the amino acid in the ribosome. Delivery of specific amino acids is
a task directed by the complementary nature of the anticodon of the tRNA116.
rRNA transcription is best described in, and thought to be close to optimal in E.
coli. By site-directed mutations the rrn promoter regions in E. coli were characterized, but no further optimization of transcription could be achieved 117. The rrn
25
upstream regions of E. coli contain a P1 and a P2 promoter that increase RNA polymerase binding 118. Both P1 and P2 have a promoter upstream element (UP) that
enhances transcription, as was shown by assaying different upstream regions with
LacZ-fusions
. L. lactis possesses at least one rrn promoter, with a canonical -35
119
sequence of TTGACA, a spacing of 17 bp and a nearly consensus -10 sequence
TAGAAT. This promoter has already been described nearly 25 years ago as promoter
P59 120. The +1 position (transcription start site) is a G, and the initiating NTP for
transcription from B. subtilis rRNA promoters is GTP 121, and is an indication that
rrn transcription is under the regulation of signal nucleotide (p)ppGpp levels 122.
Other stimulating elements in the transcription of rRNA are the Fis (factor for inversion
stimulation) elements 123. The E.coli Fis protein recognizes three sequences upstream
of the P1 promoter and at the same time interacts with the RNA polymerase, thereby
stimulating rrn transcription 124. L. lactis possesses a protein (Llmg1742) with a Fistype helix-turn-helix domain that possibly functions as a transcriptional enhancer, but
no specific data is available for the protein. Transcription of the rRNA/tRNA operons
leads to a polycistronic pre-messenger that is processed by RNAse III into the several components 125. This process leads to a molar balance between the 23S, 16S and
5S RNAs, and after proper folding to a stable ratio between ribosomal subunits and
tRNAs.
Single genes encode ribosomal proteins (rProtein) of L. lactis. These genes are mostly
scattered throughout the genome. Some rProtein genes are present in clusters like
llmg2353-llmg2357, llmg2362-llmg2367 and llmg2370-llmg2390 3. As the genes and
operons coding for ribosomal components are not all combined, the biosynthesis of
rProtein is probably uncoupled from rRNA and tRNA synthesis at the transcriptional
level. When an E. coli rProtein is synthesized it binds to an available binding location
in the ribosome. Free rProtein immediately binds to its own mRNA target as shown by
mutating a rProtein binding place 126,127. Upon less favorable conditions rRNA synthesis can be reduced, so less binding places for rProtein become available, and this feedback loop prevents undesirable production of rProteins. The regulation of ribosomal
components is described below.
The most stringent response for rRNA and tRNA synthesis in E. coli is caused by
the high energy molecule (p)ppGpp. When aminoacyl-tRNA molecules cannot be
recharged due to a shortage in amino acids, RelA is activated. The gene product of
relA is an enzyme that phosphorylates GTP to form (p)ppGpp. A second gene in L.
26
Figure 2. L. lactis has six copies of the 16S, 23S and 5S operon. All of these contain a gene
encoding tRNA for alanine (ala). Many genes that encode other tRNAs can be found in operon
1, 2, 5 and 6 and are abbreviated according to the three-letter standard. The upstream regions
of all versions of 16S hold a -10 and -35 region (red) and a spacing that is identical to the
consensus. Transcription of the 16S rRNA starts with a G (red).
lactis is spoT, the product of which is also capable of pyrophosphoryl transfer to form
(p)ppGpp. How relA and spoT are precisely regulated and how (p)ppGpp functions
on the rrn operons is still under debate 128. One option is the disruption of the stability
of the open complexes between unstable rrn DNA and RNAP by ppGpp. A second
probable option is that ppGpp binds RNAP directly, forming a closed complex and
preventing any transcription. When amino acid concentrations are low, ppGpp most
likely functions at both levels, ultimately reducing the transcription and formation of
rRNA and tRNA. L. lactis possesses both relA and spoT in the genome 3, but their role
in regulation of rRNA and tRNA remains to be elucidated.
When cells are at high growth rate, a high ribosomal activity is demanded, because
more proteins are produced, due to an increase in cell division. Many reports have
tried to give a description of the growth-rate dependence of ribosome synthesis 119,129–
131
. There is definitely a relationship between growth rate and the total amount of
ribosomes, but the molecular mechanisms behind this observation remain relatively
27
unknown. Stalling ribosomes, when less are needed, is a way for a cell to economically handle the number of ribosomes without degrading them. More and more becomes
known about this interesting phenomenon of ribosomal stalling 132, and is also a topic
of investigation in this thesis.
Fatty acid biosynthesis regulation
The biosynthesis of fatty acids is a vital part of cell maintenance. When growth rate
increases, the necessity for membrane elements like phospholipids rises accordingly.
For most bacteria, and likewise for L. lactis, the genes for fatty acid biosynthesis
(FAB) are organized in one regulon. Biosynthesis of saturated fatty acids (SFA) in
bacteria is performed by multiple conserved enzymes in a multistep process, starting
with the production of precursor malonyl-CoA by the enzymes AccABCD and followed by a series of elongation rounds to couple the fatty acid components together by
the enzymes FabDHGZIF. In the best-studied bacterium, E. coli, the regulation of FAB
is well-coordinated. The bifunctional FadR protein transcriptionally represses the fad
regulon by binding downstream of the promoter region 133.These fad gene products are
enzymes that oxidize fatty acids. At the same time, FadR can function as activator of
the FAB regulon by binding upstream of fabA 134 and fabB 135. The repressor of the FAB
genes is FabR 136. Interplay between these two regulatory proteins balances the production- and degradation rates of acyl chains, necessary for membrane phospholipids.
In L. lactis no fatty acid oxidation enzymes are found yet, so regulation of fatty acid
degradation is unknown. The genes for FAB are organized in two gene clusters, fabZ1/I
and fabT/H/acpA/fabD/G1/F/accB/fabZ2/accC/D/A, and one open reading frame,
annotated as fabG2 3. The fabT encodes a MarR-family regulator, and as shown in this
thesis, seems to be responsible for repressing the fab regulon. In the bacteria S. pneumoniae and E. faecalis, where a very similar organization of the fab genes is present, a
comparable repressor of the FAB system is reported 137,138. Generally, for the members
of Streptococcaceae, a single helix-turn-helix regulator (FabT) seems to be responsible for transcriptionally repressing the production of FAB enzymes. Most probably
FabT is regulated via a negative feedback loop. For S. pneumoniae, when sufficient
end products in the form of acyl-ACPs with an appropriate length are produced, they
stimulate the repressive activity of FabT 139. Because of the similarity in sequence and
genetic organization, it is likely that this system is universal for the Streptococcaceae.
28
Outline of the thesis
The work presented in this thesis was done as part of an STW (Stichting Toegepaste
Wetenschappen) project, called “Optimizing growth rate, biomass and product formation of Lactococcus lactis by a Systems Biology approach”. The research objectives of
this project were to: (1) elucidate the relationship between growth rate and fermentative capacity of L. lactis, (2) determine how the relationship between growth and product formation is affected by long-term cultivation (3), integrate the knowledge gained
into a cellular model that can predict fermentative behavior. Genomics technologies
(transcriptomics, proteomics, metabolomics) are used to define the core model. To
combine these different –omics approaches, a collaboration was initiated between the
groups Molecular Cell Physiology (metabolomics, modeling) at the Vrije Universiteit
in Amsterdam, Membrane Enzymology (proteomics, protein characterization) and
Molecular Genetics (transcriptomics, genetics, bioinformatics), both at the Rijksuniversiteit Groningen. This team designed a glucose-limited chemostat setup in which
the growth rate of L. lactis was varied, and determined transcriptome, proteome, part
of metabolome and fluxome at these different growth rates. The assigned tasks for the
Molecular Genetics group within the project were to perform genetic engineering and
to investigate the transcriptomics (Chapter 2 and 3), regulation of protein synthesis,
ribosome (Chapter 5) and phospholipid biosynthesis (Chapter 4). The results of that
work are described in this thesis.
In Chapter 2, we describe the transcriptome analysis of L. lactis growing at four
different growth rates. We used the common reference design in which direct and
indirect comparisons were used to increase statistical significance. Our extensive transcriptome analysis was used as part of the multi-omics study as described in Chapter
3. In this chapter, we provide a high-quality dataset of the transcriptome, proteome
and fluxome of L. lactis growing at varying growth rates. With increasing growth rate,
the focus of study was mostly on the metabolic shift from mixed-acid to homo-lactic
fermentation and the increase of ribosomes and ribosomal activity. Being an important
part of the lactococcal cell, and being part of the envisaged model, an overview of
fatty acid biosynthesis and its regulation is given in Chapter 4. The phenomenon
of ribosomal dimerization and the role that YfiA plays in this process is detailed in
Chapter 5. An overview and further discussion of the work presented in this thesis, is
given in Chapter 6.
29
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
30
Chapot-Chartier, M.-P., Deniel, C., Rousseau, M., Vassal, L. & Gripon, J.-C.
Autolysis of two strains of Lactococcus lactis during cheese ripening. Int. Dairy
J. 4, 251–269 (1994).
Mater, D. D. G. et al. Streptococcus thermophilus and Lactobacillus delbrueckii
subsp. bulgaricus survive gastrointestinal transit of healthy volunteers consuming yogurt. FEMS Microbiol. Lett 250, 185–187 (2005).
Wegmann, U. et al. Complete genome sequence of the prototype lactic acid bacterium Lactococcus lactis subsp. cremoris MG1363. J. Bacteriol. 189, 3256–3270
(2007).
Fox, P. F. Cheese: General aspects. (Academic Press, 2004).
Fox, P. F. et al. Acceleration of cheese ripening. Antonie van Leeuwenhoek 70,
271–297 (1996).
Ray, B. & Bhunia, A. Fundamental Food Microbiology, Fourth Edition. (CRC
Press, 2007).
Smit, G., Smit, B. A. & Engels, W. J. M. Flavour formation by lactic acid bacteria
and biochemical flavour profiling of cheese products. FEMS Microbiol. Rev. 29,
591–610 (2005).
Nouaille, S. et al. Transcriptomic response of Lactococcus lactis in mixed culture
with Staphylococcus aureus. Appl. Environ. Microbiol. AEM.02653–08 (2009).
Even, S., Lindley, N. D., Loubière, P. & Cocaign-Bousquet, M. Dynamic response
of catabolic pathways to autoacidification in Lactococcus lactis: Transcript profiling and stability in relation to metabolic and energetic constraints. Mol. Microbiol 45, 1143–1152 (2002).
Wouters, J. A. et al. Changes in glycolytic activity of Lactococcus lactis induced
by low temperature. Appl. Environ. Microbiol. 66, 3686–3691 (2000).
Kilstrup, M., Jacobsen, S., Hammer, K. & Vogensen, F. K. Induction of heat
shock proteins DnaK, GroEL, and GroES by salt stress in Lactococcus lactis.
Appl. Environ. Microbiol. 63, 1826–1837 (1997).
Rallu, F., Gruss, A., Ehrlich, S. D. & Maguin, E. Acid- and multistress-resistant
mutants of Lactococcus lactis: identification of intracellular stress signals. Mol.
Microbiol. 35, 517–528 (2000).
Van de Guchte, M. et al. Stress responses in lactic acid bacteria. Antonie Van
Leeuwenhoek 82, 187–216 (2002).
Villatoro-Hernández, J., Kuipers, O. P., Saucedo-Cárdenas, O. & Montes-deOca-Luna, R. Heterologous protein expression by Lactococcus lactis. Methods
Mol. Biol. 824, 155–165 (2012).
Bogaert, D., De Groot, R. & Hermans, P. Streptococcus pneumoniae colonisation: the key to pneumococcal disease. Lancet Infect. Dis. 4, 144–154 (2004).
Fisher, K. & Phillips, C. The ecology, epidemiology and virulence of Enterococcus. Microbiology 155, 1749–1757 (2009).
Yao, K.-H. et al. Surveillance of antibiotic resistance of Streptococcus pneu-
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
moniae isolated from hospitalized patients with pneumonia in four children’s
hospitals in China. Zhongguo Dang Dai Er Ke Za Zhi 10, 275–279 (2008).
Sun, J., Sundsfjord, A. & Song, X. Enterococcus faecalis from patients with
chronic periodontitis: virulence and antimicrobial resistance traits and determinants. Eur. J. Clin. Microbiol. Infect. Dis. (2011).
Tjwan Tan, P. S., Poolman, B. & Konings, W. N. Proteolytic enzymes of Lactococcus lactis. J. Dairy Res. 60, 269–286 (1993).
Savijoki, K., Ingmer, H. & Varmanen, P. Proteolytic systems of lactic acid bacteria. Appl. Microbiol. Biotech. 71, 394–406 (2006).
Kok, J., Van Dijl, J. M., Van Der Vossen, J. M. & Venema, G. Cloning and expression of a Streptococcus cremoris proteinase in Bacillus subtilis and Streptococcus lactis. Appl. Environ. Microbiol. 50, 94–101 (1985).
Tynkkynen, S. et al. Genetic and biochemical characterization of the oligopeptide transport system of Lactococcus lactis. J. Bacteriol. 175, 7523–7532 (1993).
Kunji, E. R., Smid, E. J., Plapp, R., Poolman, B. & Konings, W. N. Di-tripeptides
and oligopeptides are taken up via distinct transport mechanisms in Lactococcus
lactis. J. Bacteriol. 175, 2052–2059 (1993).
Hagting, A., Kunji, E. R., Leenhouts, K. J., Poolman, B. & Konings, W. N. The
di- and tripeptide transport protein of Lactococcus lactis. A new type of bacterial
peptide transporter. J. Biol. Chem. 269, 11391–11399 (1994).
Kunji, E. R. et al. Transport of beta-casein-derived peptides by the oligopeptide
transport system is a crucial step in the proteolytic pathway of Lactococcus lactis. J. Biol. Chem. 270, 1569–1574 (1995).
Foucaud, C. et al. Specificity of peptide transport systems in Lactococcus lactis:
evidence for a third system which transports hydrophobic di- and tripeptides. J.
Bacteriol. 177, 4652–4657 (1995).
Geis, A., Bockelmann, W. & Teuber, M. Simultaneous extraction and purification
of a cell wall-associated peptidase and β-casein specific proteas from Streptococcus cremoris. Appl. Microbiol. Biotechnol. 23, 79–84 (1985).
Kiefer-Partsch, B., Bockelmann, W., Geis, A. & Teuber, M. Purification of an
X-prolyl-dipeptidyl aminopeptidase from the cell wall proteolytic system of Lactococcus lactis subsp. cremoris. Appl. Microbiol. Biotechnol. 31, 75–78 (1989).
Law, J., Fitzgerald, G. F., Uniacke-Lowe, T., Daly, C. & Fox, P. F. The contribution of lactococcal starter proteinases to proteolysis in cheddar cheese. J. of
Dairy Sci. 76, 2455–2467 (1993).
Lee, K. D., Lo, C. G. & Warthesen, J. J. Removal of bitterness from the bitter
peptides extracted from cheddar cheese with peptidases from Lactococcus lactis
ssp. cremoris SK11. J. of Dairy Sci. 79, 1521–1528 (1996).
Guedon, E., Sperandio, B., Pons, N., Ehrlich, S. D. & Renault, P. Overall control
of nitrogen metabolism in Lactococcus lactis by CodY, and possible models for
CodY regulation in Firmicutes. Microbiol. 151, 3895–3909 (2005).
Hengst, C. D. den et al. The Lactococcus lactis CodY Regulon identification of a
31
conserved cis-regulatory element. J. Biol. Chem. 280, 34332–34342 (2005).
33. Warner, J. B. & Lolkema, J. S. CcpA-dependent carbon catabolite repression in
bacteria. Microbiol. Molec. Biol. Rev. 67, 475 –490 (2003).
34. Heinemann, M. & Sauer, U. Systems biology of microbial metabolism. Curr.
Opin. Microbiol. 13, 337–343 (2010).
35. Karas, M. & Bahr, U. Laser desorption ionization mass spectrometry of large
biomolecules. TrAC Trends in Anal. Chem. 9, 321–325.
36. Schena, M., Shalon, D., Davis, R. W. & Brown, P. O. Quantitative monitoring of
gene expression patterns with a complementary DNA microarray. Science 270,
467 –470 (1995).
37. Mukhopadhyay, R. DNA sequencers: the next generation. Anal. Chem. 81, 1736–
1740 (2009).
38. Dove, A. Proteomics: translating genomics into products? Nat. Biotechnol. 17,
233–236 (1999).
39. Romero, R. et al. The use of high dimensional biology (genomics, transcriptomics, proteomics, and metabolomics) to understand the preterm parturition
syndrome. BJOG: Intern. J. Obstet. Gyn. 113, 118–135 (2006).
40. Watson, J. D. & Crick, F. H. Molecular structure of nucleic acids; a structure for
deoxyribose nucleic acid. Nature 171, 737–738 (1953).
41. Sanger, F. et al. Nucleotide sequence of bacteriophage phi X174 DNA. Nature
265, 687–695 (1977).
42. Fleischmann, R. et al. Whole-genome random sequencing and assembly of Haemophilus influenzae Rd. Science 269, 496 –512 (1995).
43. Venter, J. C. et al. The sequence of the human genome. Science 291, 1304 –1351
(2001).
44. Lander, E.S. et al. Initial sequencing and analysis of the human genome. Nature
409, 860–921 (2001).
45. Kim Pruitt, K. C. BioProject Help. (2011)
46. Bolotin, A. et al. The complete genome sequence of the lactic acid bacterium
Lactococcus lactis ssp. lactis IL1403. Genome Res. 11, 731–753 (2001).
47. Logan, J. M. J., Edwards, K. J. & Saunders, N. A. Real-Time PCR: Current technology and applications. (2009).
48. Baldi, P. & Long, A. D. A Bayesian framework for the analysis of microarray
expression data: regularized t -test and statistical inferences of gene changes.
Bioinf. 17, 509 –519 (2001).
49. Yazaki, J., Gregory, B. D. & Ecker, J. R. Mapping the genome landscape using
tiling array technology. Curr. Opin. Plant Biol. 10, 534–542 (2007).
50. Mockler, T. C. et al. Applications of DNA tiling arrays for whole-genome analysis. Genomics 85, 1–15 (2005).
51. Wang, Z., Gerstein, M. & Snyder, M. RNA-Seq: a revolutionary tool for transcriptomics. Nat. Rev. Genet. 10, 57–63 (2009).
52. Gygi, S. P., Rochon, Y., Franza, B. R. & Aebersold, R. Correlation between Pro32
tein and mRNA abundance in yeast. Mol. Cell. Biol. 19, 1720–1730 (1999).
53. Walsh, G. & Jefferis, R. Post-translational modifications in the context of therapeutic proteins. Nat. Biotech. 24, 1241–1252 (2006).
54. Renart, J., Reiser, J. & Stark, G. R. Transfer of proteins from gels to diazobenzyloxymethyl-paper and detection with antisera: a method for studying antibody
specificity and antigen structure. Proc. Natl. Acad. Sci. 76, 3116 –3120 (1979).
55. Gygi, S. P., Corthals, G. L., Zhang, Y., Rochon, Y. & Aebersold, R. Evaluation of
two-dimensional gel electrophoresis-based proteome analysis technology. Proc.
Natl. Acad. Sci. 97, 9390 –9395 (2000).
56. Monteoliva, L. & Albar, J. P. Differential proteomics: An overview of gel and
non-gel based approaches. Brief. Func. Genomics & Proteomics 3, 220 –239
(2004).
57. Ross, P. L. et al. Multiplexed protein quantitation in Saccharomyces cerevisiae
using amine-reactive isobaric tagging reagents. Mol. Cell Proteomics 3, 1154–
1169 (2004).
58. Raska, C. S. et al. Direct MALDI-MS/MS of Phosphopeptides affinity-bound to
immobilized metal ion affinity chromatography beads. Anal. Chem. 74, 3429–
3433 (2002).
59. Peng, J., Elias, J. E., Thoreen, C. C., Licklider, L. J. & Gygi, S. P. evaluation of
multidimensional chromatography coupled with tandem mass spectrometry (LC/
LC−MS/MS) for large-scale protein analysis: The yeast proteome. J. Proteome
Res. 2, 43–50 (2003).
60. Zhang, W. & Chait, B. T. ProFound: An expert system for protein identification using mass spectrometric peptide mapping information. Anal. Chem. 72,
2482–2489 (2000).
61. Xu, C. & Ma, B. Software for computational peptide identification from MS-MS
data. Drug Discov. Today 11, 595–600 (2006).
62. Mijakovic, I. & Macek, B. Impact of phosphoproteomics on studies of bacterial
physiology. FEMS Microbiol. Rev. 36, 877–892 (2012).
63. Stock, J. B., Ninfa, A. J. & Stock, A. M. Protein phosphorylation and regulation
of adaptive responses in bacteria. Microbiol. Rev. 53, 450–490 (1989).
64. Stülke, J. More than just activity control: phosphorylation may control all aspects
of a protein’s properties. Mol. Microbiol. 77, 273–275 (2010).
65. Bodenmiller, B. et al. Phosphoproteomic analysis reveals interconnected systemwide responses to perturbations of kinases and phosphatases in yeast. Sci. Signal
3, rs4 (2010).
66. Macek, B. et al. The serine/threonine/tyrosine phosphoproteome of the model
bacterium Bacillus subtilis. Mol. Cell Proteomics 6, 697–707 (2007).
67. Macek, B. et al. Phosphoproteome analysis of E. coli reveals evolutionary conservation of bacterial Ser/Thr/Tyr phosphorylation. Mol. Cell Proteomics 7,
299–307 (2008).
68. Elsholz, A. K. W. et al. Global impact of protein arginine phosphorylation on the
33
physiology of Bacillus subtilis. Proc. Natl. Acad. Sci. 109, 7451–7456 (2012).
69. Görke, B. & Stülke, J. Carbon catabolite repression in bacteria: many ways to
make the most out of nutrients. Nat. Rev. Microbiol. 6, 613–624 (2008).
70. Deutscher, J., Küster, E., Bergstedt, U., Charrier, V. & Hillen, W. Protein kinasedependent HPr/CcpA interaction links glycolytic activity to carbon catabolite
repression in gram-positive bacteria. Mol. Microbiol. 15, 1049–1053 (1995).
71. Nowak, S. J. & Corces, V. G. Phosphorylation of histone H3: a balancing act
between chromosome condensation and transcriptional activation. Trends Genet.
20, 214–220 (2004).
72. Zhang, Z. et al. Identification of lysine succinylation as a new post-translational
modification. Nat. Chem. Biol. 7, 58–63 (2011).
73. Zhang, J. et al. Lysine acetylation is a highly abundant and evolutionarily conserved modification in Escherichia coli. Mol. Cell Proteomics 8, 215–225 (2009).
74. Gadian, D. G. NMR and its applications to living systems. (1996).
75. Pérez-Flores, J. C. et al. Insight into the channel ion distribution and influence on
the lithium insertion properties of hexatitanates A(2)Ti(6)O(13) (A = Na, Li, H)
as candidates for anode materials in lithium-ion batteries. Dalton Trans (2012).
76. Hellmich, U. A., Duchardt-Ferner, E., Glaubitz, C. & Wöhnert, J. Backbone
NMR resonance assignments of the nucleotide binding domain of the ABC multidrug transporter LmrA from Lactococcus lactis in its ADP-bound state. Biomol
NMR Assign. 6, 69–73 (2012).
77. Ramos, A., Neves, A. R. & Santos, H. Metabolism of lactic acid bacteria studied
by nuclear magnetic resonance. Antonie Van Leeuwenhoek 82, 249–261 (2002).
78. Neves, A. R., Pool, W. A., Kok, J., Kuipers, O. P. & Santos, H. Overview on
sugar metabolism and its control in Lactococcus lactis - the input from in vivo
NMR. FEMS Microbiol. Rev. 29, 531–554 (2005).
79. Boele, J., Olivier, B. G. & Teusink, B. FAME, the flux analysis and modeling
environment. BMC Syst. Biol. 6, 8 (2012).
80. De Jong, A. Development and use of genomics mining and visualization tools for
prokaryotes. (2012).
81. Myers, J. L. & Well, A. D. Research Design & Statistical Analysis. (2002).
82. Ashburner, M. et al. Gene ontology: tool for the unification of biology. The gene
ontology consortium. Nat. Genet. 25, 25–29 (2000).
83. Ogata, H. et al. KEGG: Kyoto encyclopedia of genes and genomes. Nucleic
Acids Res. 27, 29–34 (1999).
84. Tatusov, R. L. et al. The COG database: an updated version includes eukaryotes.
BMC Bioinf. 4, 41 (2003).
85. Bateman, A. et al. The Pfam protein families database. Nucleic Acids Res. 32,
138–141 (2004).
86. Eppinger, M., Baar, C., Raddatz, G., Huson, D. H. & Schuster, S. C. Comparative
analysis of four Campylobacterales. Nat. Rev. Microbiol. 2, 872–885 (2004).
87. Deng, Y., Yao, J., Wang, X., Guo, H. & Duan, D. Transcriptome sequencing and
34
comparative analysis of Saccharina japonica (Laminariales, Phaeophyceae)
under blue light induction. PLoS ONE 7, e39704 (2012).
88. Zomer, A. L., Buist, G., Larsen, R., Kok, J. & Kuipers, O. P. Time-resolved determination of the CcpA Regulon of Lactococcus lactis subsp. cremoris MG1363.
J. Bacteriol. 189, 1366–1381 (2007).
89. De Jong, A. Genome2D MolGen Bioinformatics server. Functional analysis
(2012).
90. Bruggeman, F. J. & Westerhoff, H. V. The nature of systems biology. Trends
Microbiol. 15, 45–50 (2007).
91. Wilson, I. Top-down versus bottom-up—rediscovering physiology via systems
biology? Mol. Syst. Biol. 3, (2007).
92. Oliveira, A. P., Nielsen, J. & Förster, J. Modeling Lactococcus lactis using a
genome-scale flux model. BMC Microbiol. 5, 39 (2005).
93. Pinto, J. P. C., Kuipers, O. P., Marreddy, R. K. R., Poolman, B. & Kok, J. Efficient overproduction of membrane proteins in Lactococcus lactis requires the
cell envelope stress sensor/regulator couple CesSR. PLoS ONE 6, e21873 (2011).
94. Michelsen, O., Hansen, F. G., Albrechtsen, B. & Jensen, P. R. The MG1363 and
IL1403 laboratory strains of Lactococcus lactis and several dairy strains are diploid. J. Bacteriol. 192, 1058–1065 (2010).
95. Molenaar, D., Van Berlo, R., De Ridder, D. & Teusink, B. Shifts in growth strategies reflect tradeoffs in cellular economics. Mol. Syst. Biol. 5, (2009).
96. Teixeira de Mattos, M. & Neijssel, O. Bioenergetic consequences of microbial
adaptation to low-nutrient environments. J. Biotechnol. 59, 117–126 (1997).
97. Thomas, T. D., Ellwood, D. C. & Longyear, V. M. C. Change from homo- to heterolactic fermentation by Streptococcus lactis resulting from glucose limitation
in anaerobic chemostat cultures. J. Bacteriol. 138, 109–117 (1979).
98. Dijken, J. P., Weusthuis, R. A. & Pronk, J. T. Kinetics of growth and sugar consumption in yeasts. Antonie van Leeuwenhoek 63, 343–352 (1993).
99. Vemuri, G. N., Altman, E., Sangurdekar, D. P., Khodursky, A. B. & Eiteman, M.
A. Overflow metabolism in Escherichia coli during steady-state growth: transcriptional regulation and effect of the redox ratio. Appl. Environ. Microbiol. 72,
3653–3661 (2006).
100.Diaz-Ruiz, R., Rigoulet, M. & Devin, A. The Warburg and Crabtree effects: On
the origin of cancer cell energy metabolism and of yeast glucose repression. Biochim. Biophys. Acta 1807, 568–576 (2011).
101.Thompson, J. In vivo regulation of glycolysis and characterization of sugar:
phosphotransferase systems in Streptococcus lactis. J. Bacteriol. 136, 465–476
(1978).
102.Kowalczyk, M. & Bardowski, J. Regulation of sugar catabolism in Lactococcus
lactis. Crit. Rev. Microbiol. 33, 1–13 (2007).
103.Gaspar, P., Neves, A. R., Gasson, M. J., Shearman, C. A. & Santos, H. high
yields of 2,3-butanediol and mannitol in Lactococcus lactis through engineering
35
of NAD+ cofactor recycling. Appl. Environ. Microbiol. 77, 6826–6835 (2011).
104.Gunnewijk, M. G. & Poolman, B. Phosphorylation state of HPr determines the
level of expression and the extent of phosphorylation of the lactose transport
protein of Streptococcus thermophilus. J. Biol. Chem. 275, 34073–34079 (2000).
105.Deutscher, J. The mechanisms of carbon catabolite repression in bacteria. Curr.
Opin. Microbiol. 11, 87–93 (2008).
106.Reizer, J. et al. A novel protein kinase that controls carbon catabolite repression
in bacteria. Molec. Microbiol. 27, 1157–1169 (1998).
107.Luesink, E. J., Van Herpen, R. E. M. A., Grossiord, B. P., Kuipers, O. P. & De
Vos, W. M. Transcriptional activation of the glycolytic las operon and catabolite
repression of the gal operon in Lactococcus lactis are mediated by the catabolite
control protein CcpA. Molec. Microbiol. 30, 789–798 (1998).
108.Larsen, R., Buist, G., Kuipers, O. P. & Kok, J. ArgR and AhrC are both required
for regulation of arginine metabolism in Lactococcus lactis. J. Bacteriol. 186,
1147–1157 (2004).
109.Poolman, B., Driessen, A. J. & Konings, W. N. Regulation of arginine-ornithine
exchange and the arginine deiminase pathway in Streptococcus lactis. J. Bacteriol. 169, 5597–5604 (1987).
110. Cotter, P. D. & Hill, C. Surviving the acid test: responses of gram-positive bacteria to low pH. Microbiol. Mol. Biol. Rev. 67, 429–453 (2003).
111. Schleifer, K. H. et al. Transfer of Streptococcus lactis and related Streptococci to
the genus Lactococcus gen. Syst. Appl. Microbiol. 6, 183–195 (1985).
112. Larsen, R., Van Hijum, S. A. F. T., Martinussen, J., Kuipers, O. P. & Kok, J.
Transcriptome analysis of the Lactococcus lactis ArgR and AhrC regulons. Appl.
Environ. Microbiol. 74, 4768–4771 (2008).
113. Larsen, R., Kok, J. & Kuipers, O. P. Interaction between ArgR and AhrC controls regulation of arginine metabolism in Lactococcus lactis. J. Biol. Chem. 280,
19319–19330 (2005).
114. Dong, Y., Chen, Y.-Y. M., Snyder, J. A. & Burne, R. A. Isolation and molecular
analysis of the gene cluster for the arginine deiminase system from Streptococcus
gordonii DL1. Appl. Environ. Microbiol. 68, 5549–5553 (2002).
115. Griswold, A., Chen, Y.-Y. M., Snyder, J. A. & Burne, R. A. Characterization
of the arginine deiminase operon of Streptococcus rattus FA-1. Appl. Environ.
Microbiol. 70, 1321–1327 (2004).
116. Crick, F. H. C. The origin of the genetic code. J. Molec. Biol. 38, 367–379 (1968).
117.Gaal, T. et al. Saturation mutagenesis of an Escherichia coli rRNA promoter
and initial characterization of promoter variants. J .Bacteriol. 171, 4852–4861
(1989).
118. Gourse, R. L., Gaal, T., Bartlett, M. S., Appleman, J. A. & Ross, W. rRNA transcription and growth rate–dependent regulation of ribosome synthesis in Escherichia coli. Annu. Rev. Microbiol. 50, 645–677 (1996).
119. Gourse, R. L., De Boer, H. A. & Nomura, M. DNA determinants of rRNA syn36
thesis in E. coli: growth rate dependent regulation, feedback inhibition, upstream
activation, antitermination. Cell 44, 197–205 (1986).
120.Van der Vossen, J. M., Van der Lelie, D. & Venema, G. Isolation and characterization of Streptococcus cremoris Wg2-specific promoters. Appl. Environ. Microbiol. 53, 2452–2457 (1987).
121.Lew, C. M. & Gralla, J. D. Mechanism of stimulation of ribosomal promoters by
binding of the +1 and +2 nucleotides. J. Biol. Chem. 279, 19481–19485 (2004).
122.Krásný, L. & Gourse, R. L. An alternative strategy for bacterial ribosome synthesis: Bacillus subtilis rRNA transcription regulation. EMBO J. 23, 4473–4483
(2004).
123.Johnson, R. C. & Simon, M. I. Hin-mediated site-specific recombination requires
two 26 by recombination sites and a 60 by recombinational enhancer. Cell 41,
781–791 (1985).
124.Gourse, R. L., Gaal, T., Bartlett, M. S., Appleman, J. A. & Ross, W. rRNA transcription and growth rate–dependent regulation of ribosome synthesis in Escherichia coli. Annu. Rev. Microbiol. 50, 645–677 (1996).
125.Young, R. A. & Steitz, J. A. Complementary sequences 1700 nucleotides apart
form a ribonuclease III cleavage site in Escherichia coli ribosomal precursor
RNA. Proc. Natl. Acad. Sci. 75, 3593–3597 (1978).
126.Babitzke, P., Baker, C. S. & Romeo, T. Regulation of translation initiation by
RNA binding proteins. Annu. Rev. Microbiol. 63, 27–44 (2009).
127.Cole, J. R. & Nomura, M. Translational regulation is responsible for growth-ratedependent and stringent control of the synthesis of ribosomal proteins L11 and
L1 in Escherichia coli. Proc. Natl. Acad. Sci. 83, 4129–4133 (1986).
128.Potrykus, K. & Cashel, M. (p)ppGpp: Still magical? Annu. Rev. Microbiol. 62,
35–51 (2008).
129.Maaløe, O. & Kjeldgaard, N. O. Control of macromolecular synthesis: a study of
DNA, RNA, and protein synthesis in bacteria. (1966).
130.Koch, A. L. Overall controls on the biosynthesis of ribosomes in growing bacteria. J. Theor. Biol. 28, 201–231 (1970).
131.Nomura, M., Gourse, R. & Baughman, G. Regulation of the synthesis of ribosomes and ribosomal components. Annu. Rev. Biochem. 53, 75–117 (1984).
132.Ueta, M. et al. Role of HPF (Hibernation Promoting Factor) in translational
activity in Escherichia coli. J. Biochem. 143, 425 –433 (2008).
133.Cronan, J. E. & Subrahmanyam, S. FadR, transcriptional co-ordination of metabolic expediency. Mol. Microbiol 29, 937–943 (1998).
134.Henry, M. F. & Cronan, J. E. A new mechanism of transcriptional regulation:
release of an activator triggered by small molecule binding. Cell 70, 671–679
(1992).
135.Campbell, J. W. & Cronan, J. E., Jr Escherichia coli FadR positively regulates
transcription of the fabB fatty acid biosynthetic gene. J. Bacteriol. 183, 5982–
5990 (2001).
37
136.Zhang, Y.-M., Marrakchi, H. & Rock, C. O. The FabR (YijC) transcription factor
regulates unsaturated fatty acid biosynthesis in Escherichia coli. J. Biol. Chem.
277, 15558–15565 (2002).
137.Lu, Y.-J. & Rock, C. O. Transcriptional regulation of fatty acid biosynthesis in
Streptococcus pneumoniae. Mol. Microbiol. 59, 551–566 (2006).
138.Lu, Y.-J., White, S. W. & Rock, C. O. Domain swapping between Enterococcus
faecalis FabN and FabZ proteins localizes the structural determinants for isomerase activity. J. Biol. Chem. 280, 30342 –30348 (2005).
139.Jerga, A. & Rock, C. O. Acyl-acyl carrier protein regulates transcription of fatty
acid biosynthetic genes via the FabT repressor in Streptococcus pneumoniae. J.
Biol. Chem. 284, 15364–15368 (2009).
38
39
40
Chapter 2
Comparative transcriptome analysis of
Lactococcus lactis MG1363 propagated at
varying growth rates using chemostats
Thomas H. Eckhardt, Anne de Jong, Jan Kok and Oscar P. Kuipers
41
Abstract
In this study we provide a high-quality dataset of the transcriptional activity of Lactococcus lactis MG1363 at different growth rates. We compared four different growth
rates using a glucose-limited chemostat setup, with a common reference method that
greatly improved the numbers of significant changes found. By analyzing growth ratedependent changes across the complete set of lactococcal mRNA, we show that for
most cellular processes transcription exhibits a quite stable pattern. Transcriptional
regulation plays a modest role in the adaptation of L. lactis to different growth rates.
Introduction
The model organism for studying lactic acid bacteria (LAB) is Lactococcus lactis. L.
lactis strains are used as starter cultures for the production of many dairy products, such
as a large variety of cheeses. During milk fermentation the bacteria are confronted with
rapidly changing conditions due to their conversion of lactose into lactic acid, leading
to a fast lowering of the pH, and because of various technical insults like temperature
changes, salt addition and more. L. lactis is able to quickly adapt to such a changing
environment. Its changeable behavior makes it difficult to assess the robustness of L.
lactis when it is used in a starter culture. From an industrial perspective this is unwanted
since it can lead to a significant decrease of production efficiency 1,2. The unpredictabil-
42
ity of L. lactis is mainly due to a lack of understanding with regard to its fermentative
behavior. For instance, L. lactis can use either of two different metabolic pathways for
the degradation of glucose; an energetically efficient mixed-acid fermentation at low
growth rates, and an energetically inefficient homolactic fermentation at high growth
rates 3. Variables such as enzyme concentration, all kind of effectors involved in allosteric regulation, or small molecules like (p)ppGpp and small regulatory RNAs can
play a role in regulation of glycolysis and more general in the physiology of bacteria 4–8.
In this study we focus on other factors that are of influence for metabolic activity, namely the transcriptional activity. We characterized the transcriptional activity of the annotated genes in L. lactis MG1363 9 as a function of increasing growth
rate. The aim was to provide a high-quality dataset yielding information on the
growth rate-dependency of the transcription of all genes of L. lactis MG1363.
In order to study the transcriptome of an organism, the most popular method for the
last decade has been the DNA-microarray technology 10. The simple whole-genome
technology using DNA microarrays is cheap, accurate, and fast and generates data
that is easy to interpret. For these reasons we chose the DNA-microarray technology
for our in vitro gene expression studies. When comparing two different strains or two
different culture conditions, a direct comparison method is most often sufficient to
observe differences at the transcriptional level. For more complex experiments that
require differential expression analysis, the design of the analysis depends strongly
on the setup of these experiments. In the current study we aimed to identify the transcriptionally altered genes in relation to a changing growth rate by comparing gene
transcription in L. lactis cultures maintained in separate chemostats. The chemostats
were operated at different dilution rates, such that four stable cultures were established
that grew with growth rates of µ=0.15 h−1; µ= 0.3 h−1; µ= 0.5 h−1; µ= 0.6 h−1 (Fig. 1A).
The comparisons of (quasi) steady-state cultures of L. lactis result in small and often
insignificant effects on the transcriptome 11. For statistical reasons, we combined both
the direct and indirect comparisons of the transcriptomes of the various cultures (Fig.
1A). We present this method as a useful technique for transcriptional data analysis
when comparing strains and/or conditions with small transcriptional effects.
Material and Methods
Strain and growth conditions
L. lactis MG1363 12 cells were precultured in a chemically defined medium for prolonged
cultivations (CDMPC) 13 and subsequently grown in chemostats. CDMPC contains 25
mM glucose and is composed of: (A) buffer (in g · liter−1), KH2PO4, 2.75; Na2HPO4,
2.85; NaCl, 2.9; (B) vitamins (in mg · liter−1), dl-6,8-thioctic acid, 2; d-pantothenic acid
hemicalcium salt, 0.5; biotin, 0.1; nicotinic acid, 1; pyridoxal hydrochloride, 1; pyridoxine hydrochloride, 1; thiamine hydrochloride, 1; (C) metals (in mg · liter−1), ammonium molybdate tetrahydrate, 0.3; calcium chloride dihydrate, 3; cobalt(II) sulfate heptahydrate, 0.3; copper(II) sulfate pentahydrate, 0.3; iron(II) chloride tetrahydrate, 4;
magnesium chloride hexahydrate, 200; manganese chloride tetrahydrate, 4; zinc sulfate
heptahydrate, 0.3; and (D) amino acids (in mg · liter−1), l-alanine, 130; l-arginine, 244;
l-asparagine, 80; l-aspartic acid, 137; l-cysteine hydrochloride monohydrate, 61; l-glutamic acid, 97; l-glutamine, 96; glycine, 29; l-histidine, 24; l-isoleucine, 82; l-leucine,
117; l-lysine monohydrochloride, 187; l-methionine, 38; l-phenylalanine, 64; l-pro43
line, 412; l-serine, 172; l-threonine, 68; l-tryptophan, 36; l-tyrosine, 50; l-valine, 86.
Glucose-limited chemostat cultures were grown in 2 L bioreactors with a working volume of 1.2 L at 30°C, under continuous mild stirring. The headspace was flushed at 5
headspace volume changes per hour, with a gas mixture of 95% N2 (99.998% pure) and
5% CO2 (99.7% pure) with oxygen impurity less than 34 vpm. A pH of 6.5±0.05 was
maintained by automatic titration with 5 M NaOH. Chemostats were inoculated with
4% (v/v) of standardized precultures consisting of 45 mL of CDMPC inoculated with
300 µL of a glycerol stock of L. lactis MG 1363 and incubated for 16 h at 30°C. After
batch growth until an optical density at 600 nm (OD600) of around 1.8 was reached, the
specific growth rate was adjusted by the in- and outflux of medium, and set at µ= 0.15
h−1, µ= 0.3 h−1 µ= 0.5 h−1 and µ= 0.6 h−1. Chemostats, with independent precultures,
were run in triplicate. From every chemostat, the cells from two samples of 30 ml were
harvested by centrifugation (6,000 g for 5 min). The pellets were immediately frozen
in liquid nitrogen prior to storage at -80°C.
DNA microarray analysis
For RNA isolation the frozen cells were thawed on ice. Subsequent cell disruption,
RNA purification, reverse transcription and Cy3/Cy5 labeling were done as described
previously 14. Labeled cDNAs were hybridized to full-genome DNA microarray slides
of L. lactis MG1363 (GEO accession number GLP16042) 15. All reagents and glassware for RNA work were treated with diethylpyrocarbonate (DEPC) (Sigma-Aldrich,
St Louis, MO). RNA and cDNA quantity, quality, and the incorporation of the cyaninelabels were examined by NanoDrop (ThermoFisher Scientific Inc. Waltham, MA) at
260 nm for RNA and cDNA, 550 nm for Cy3, and 650 nm for Cy5. As shown in Figure
1A, samples of every chemostat were compared with those of every other chemostat,
with a dye-swap for each comparison. The three biological replicates (A,B,C) were
compared independently, so that only the cDNA from the culture grown at µ= 0.15(A)
was compared with that of µ= 0.3(A), and that of µ= 0.15(B) only with that of µ=
0.3(B) etc. DNA microarray slide images were analyzed using ArrayPro 4.5 (Media
Cybernetics Inc., Silver Spring, MD). The Limma R package 16 was used to analyze
the DNA microarray data using that of the culture at µ= 0.15 as the common reference.
Fold changes are considered to be significantly altered when the Benjamini-Hochberg
adjusted p-value was ≤ 0.05.
44
Bioinformatics analysis
An in-depth analysis of the transcriptome data was performed with a variety of bioinformatics tools. In order to do so, we used 2log-transformed gene expression ratios
of comparisons of the different growth rates as input. The COG category 17 and the L.
lactis regulon classification 18 were used to determine significant changes in functionally related genes. The software package MeV 19 was used to cluster the transcriptional data for all genes, using k-means clustering based on the Pearson’s correlation.
A distance matrix of all experiments was created using the Pearson’s correlation of
the growth rate comparisons 19. Genome visualization was obtained by plotting the
data points of the direct comparisons of µ= 0.3, µ= 0.5, or µ= 0.6 each, versus µ= 0.15
against the gene location data in the Microbial Genome Viewer 20.
45
Results
Global transcriptional analysis of L. lactis growing at four different dilution rates
In this study, an in-depth analysis of the genome-wide transcriptional alterations as a
function of the growth-rate of L. lactis was performed. To this end, the transcriptomes
of triplicates of four chemostat cultures with different growth rates were determined
using DNA microarrays. The differences between the various cultures were obtained
by comparing the transcriptional profiles of the cells in two chemostats directly.
The analyses of the transcriptomes were performed using the chemostat operated
at a dilution rate of D= 0.15 h-1 as a common reference, allowing indirect comparisons to potentially provide additional information (Fig. 1A). Indeed, when adding
the results of the indirect comparisons into the common-reference equation, as supported by the Limma-package 16, the sensitivity of every direct comparison increases.
This is indicated by the 1.6- to 7-times increase in the number of all genes that were
found to be significantly up- or downregulated (Fig. 1B). Especially the detection of
genes of which the transcriptional activity changed only slightly as a consequence
of the changing growth rate was improved considerably (Addendum, Table A1).
Whether or not the different growth rate comparisons correlate with each other
was calculated by making a Pearson’s correlation distance matrix (Fig. 1C). The
correlations between the comparisons of the samples taken from the 0.6 h−1 and
0.15 h−1 cultures, the 0.6 h−1 and 0.3 h−1 cultures and 0.6 h−1 and 0.5 h−1 are very
high. On the other hand, the comparisons between values from the samples of
the 0.15 h−1 and 0.3 h−1 cultures shows a very low correlation only with values from the samples of the 0.3 h−1 and 0.5 h−1 or the 0.5 h−1 and 0.6 h−1 cultures.
This shows the reproducability of the setup and the high quality of the dataset.
As a next step in the global analysis of the data, transcriptional activities of all genes
as a function of the growth rate were plotted on a chromosomal map (Fig. 2). The
results show that hotspots of transcriptionally active areas are scattered all over the
genome, without there being specific chromosomal regions with exceptionally altered
transcription. When B. subtilis is growing with a high growth rate, the expression
of genes around the oriC is higher 21. Such a location effect seems to be absent in L.
46
lactis, as no local increase of transcriptional activity is observed around oriC (Fig. 2).
A higher growth rate in bacteria is reflected in the transcriptional activity of more spe-
A
B
C
Figure 1. (A) Experimental design used for comparing the transcriptomes of L. lactis, grown
at the indicated growth rates. Double-headed arrows indicate that dye-swaps were performed
for each comparison. The cultures growing with µ= 0.15 was used as the common reference.
All direct and indirect comparisons e.g., µ= 0.15 compared to µ= 0.5 compared to µ= 0.3
were analyzed. An example of the (in) direct comparisons is presented in transparent colors.
The indirect comparisons, as shown in red, contain information about the direct comparison
(in blue) and are therefore pooled with the direct comparison. This scheme was used three
times for three biological replicates. (B) The amount of significant genes increases (number of
genes with significantly changed transcription levels) when adding indirect comparisons. (C)
Pearson correlation distance matrix of the different growth rate as compared against the growth
rate of µ= 0.15 h-1. The distance of expression patterns, between two samples is calculated
using a Pearson distance metric (TMeV). By comparing all sample-to-sample distances, an
insight is provided in the extent of cluster similarity. A high similarity is visualized as a dark
green square, lower similarities are lighter.
cific parts of the bacterial chromosome, namely those regions that contain for instance
the genes encoding RNA polymerase, rRNAs and tRNAs 22. In B. subtilis increased
transcriptional activity at high growth rates was observed for ribosomal genes. The
first base of these transcripts is a G nucleotide 23. Transcripts of L. lactis rRNA also
start with a G (Chapter 1, Fig. 2). We did not discern a specific base preference for the
+1 position of L. lactis genes transcriptionally activated at high growth rates.
47
Figure 2. Genome of L. lactis MG1363 indicating the position of the origin of replication
(oriC) and termination (terC) and distances in mega basepairs. The second ring shows all
annotated genes, in red are the genes on the +strand, yellow is indicative for the genes on the
–strand. The inner circle depicts the transcriptional fold change of the respective genes with
an increasing growth rate (the effect of 0.3 is turquoise, of 0.5 is purple and of 0.6 is blue).
Classes of genes that respond to changes in growth rate
To look for genes of which transcription changed at increasing growth rates, a selection was made of all genes that were either significantly upregulated or significantly
downregulated for at least 2 out of the 6 comparisons (supplementary material, Table
S1). A total of 516 genes were significantly upregulated while 431 genes were significantly downregulated. Within the latter group, 288 genes fit a COG-category e.g., lipid
transport and metabolism (10 out of 37 genes belonging to this category), carbohydrate transport and metabolism (34 out of 163), cell motility (2 out of 9) and defense
mechanisms (8 out of 36) (supplementary material, Table S1). Of the genes that are
upregulated at increasing growth rate, 430 fit in a COG-category; a high proportion of
these genes are involved in the biological processes of signal transduction mechanisms
(14 out of 41), coenzyme transport and metabolism (28 out of 66), nucleotide transport and metabolism (37 out of 78). Also, COG-categories like cell wall/membrane/
48
envelope biogenesis (28 out of 98), cell cycle control, cell division, chromosome partitioning (6 out of 17), translation, ribosomal structure and biogenesis (63 out of 145)
are overrepresented (supplementary material, Table S1). When looking at the core
genome as defined for L. lactis 24, 672 out of the 1216 core genes do not respond to an
increase in growth rate, indicating that many genes in the core genome (55%) are not
growth-rate adaptive. Many of these genes belong to the COG-categories posttranslational modification, protein turnover, chaperones, cell motility, inorganic ion transport
and metabolism, and energy production and conversion. Of the other core genes, 353
are upregulated and 191 are downregulated (supplementary material, Table S1). The
only COG-category in the core genome that is downregulated as a consequence of the
increase in the growth rate is lipid transport and metabolism. The significantly upregulated genes in the core genome belong to the COG-categories nucleotide transport and
metabolism, coenzyme transport and metabolism, and translation, ribosomal structure
and biogenesis. Gene expression of all genes as a function of the growth rate was stud-
fold change (log) Effect (h-­‐1) 0.3 0.5-­‐0.3 0.5 0.6-­‐0.5 0.6-­‐0.3 0.6 2 1 0 -­‐1 2 1 0 -­‐1 fold change (log) -­‐2 -­‐2 0.3 0.5-­‐0.3 0.5 0.6-­‐0.5 0.6-­‐0.3 0.6 Effect (h-­‐1) Figure 3. Cluster analysis of the complete growth rate dependent experiment. Fuzzy k-means
clustering was performed on the 1,762 mRNAs with k=4, to group similar expression patterns.
The 2log fold change is plotted against the transcriptional effect of each growth rate compared
with µ= 0.15 h-1. Clockwise, starting in the left upper corner; (1) unchanged till 0.5 h-1,
repression at highest growth rate; (2) transcripts mainly positively correlated with the growth
rate; (3) unchanged; (4) slightly activated at 0.15 h-1, unchanged at other growth rates. Every
grey line signifies a single transcription dataset, its group average is symbolized by a pink line.
49
ied using the multiple experiment viewer (MeV 19) . With k-means clustering we aimed
to group genes that have a similar expression profile over the various growth rate comparisons. To achieve this goal, four different groups of genes with comparable transcriptional profiles were defined (Fig. 3). Group 1 contains genes of which transcript
ratios show a slight increase from 0.15 h-1 to 0.5 h-1, but which are repressed at 0.6 h-1.
Genes of which transcription is positively correlated with the increasing growth rate
over all four growth rates examined are collected in group 2; these will be described
in other sections below. Group 3 contains genes of which the transcripts remain
unchanged throughout all experimental samples. Finally, group 4 contains genes that
are transcriptionally activated exclusively at a growth rate of 0.15 h-1. Most of the genes
that do not or hardly respond to an increase in the growth rate (Fig. 3, group 3) are not
annotated or have been annotated as coding for proteins with an unknown function.
Among group 1 are genes that are part of a large cluster of genes, ranging from llmg0206llmg0217. This region contains genes coding for glycosyltransferases rgpABCEF and
carbohydrate isomerases rmlACBD. Homologs of these genes are also located in a
long operon in the genome of L. lactis IL1403 that is prone to phage absorption 25. Part
of this large operon has been shown to be involved in the formation of a sugar pellicle
around cells of L. lactis 26. Also the mannitol specific PTS system genes mtlAF and their
regulator mtlR behave alike with respect to their transcriptional activity. These genes
belong to the CcpA regulon 27. Indeed, the top five of genes showing the highest drop
in transcription at 0.6 h-1 all belong to the CcpA regulon and will be discussed below.
Some genes belonging to group 4, the transcription of which is highest at a growth rate
of 0.15 h-1 and subsequently does not change at higher growth rates, are involved in
central metabolism. The genes encoding glyceraldehyde 3-phosphate dehydrogenase
gapB, phosphoglucosidase ascB and fructose 1-phosphate kinase fruC are examples
of such genes. Other genes, not related to carbohydrate metabolism but with a similar trend in the fold-changes between 0.15 h−1 and the other growth rates are ispA
(encoding geranyltranstransferase), llmg1612 (for a transposase helper protein), ponA
(putative penicillin-binding protein), llmg1468 and llmg1552 (coding for two putative
ABC type transport system proteins), osmC (osmotically inducible protein) C, cadA
(cation-transporting ATPase), llmg0276 (oxidoreductase), uspA and llmg2023 (encoding two universal shock proteins).
50
Table 1. Number of genes significantly changed with increasing growth rates compared with
μ: 0.15 h-1 and classified in their functional COG categories. Sorting is based on the number
of total significantly effected genes per category. Low numbers of genes per experiment
are indicated by a lighter background color, separated between upregulated (up, green) and
downregulated (down, red).
Functional categories
Total
Total
Total
up
down
up
down
up
down
up
down
up
down
up
down
genes
up
Down
[S] Function unknown
23
32
38
44
33
39
32
26
39
42
20
35
403
185
218
[J] Translation, ribosomal structure and biogenesis
34
8
51
18
61
23
30
10
56
27
46
26
390
278
112
[R] General function prediction only
24
32
29
41
30
32
30
27
34
37
28
35
379
175
204
[E] Amino acid transport and metabolism
42
17
48
25
39
25
29
22
31
34
19
31
362
208
154
[G] Carbohydrate transport and metabolism
25
31
36
35
24
32
29
16
27
26
18
34
333
159
174
[M] Cell wall/membrane/envelope biogenesis
27
17
32
25
28
16
17
16
16
19
15
16
244
135
109
[K] Transcription
24
17
30
23
20
17
18
14
23
17
18
19
240
133
107
[L] Replication, recombination and repair
18
17
28
25
21
16
21
16
20
16
13
16
227
121
106
[F] Nucleotide transport and metabolism
23
2
33
8
36
3
14
10
25
6
19
8
187
150
37
[P] Inorganic ion transport and metabolism
14
12
34
10
19
11
16
6
15
17
8
18
180
106
74
[H] Coenzyme transport and metabolism
11
3
28
6
23
2
11
6
21
7
15
6
139
109
30
[O] Posttranslational modification, protein turnover, chaperones
12
4
20
5
12
5
20
5
13
10
7
14
127
84
43
[C] Energy production and conversion
10
7
17
10
10
12
11
7
8
11
6
11
120
62
58
[T] Signal transduction mechanisms
9
5
13
6
13
8
10
4
5
9
6
8
96
56
40
[I] Lipid transport and metabolism
4
8
7
5
13
5
6
7
11
9
9
5
89
50
39
[V] Defense mechanisms
9
6
9
4
4
7
6
1
3
8
1
7
65
32
33
[D] Cell cycle control, cell division, chromosome partitioning
2
2
6
5
6
4
5
4
6
1
7
0
48
32
16
[U] Intracellular trafficking, secretion, and vesicular transport
3
1
6
4
5
4
3
2
5
2
4
3
42
26
16
[Q] Sec. metabolites biosynthesis, transport and catabolism
2
4
3
6
2
5
3
1
2
4
4
2
38
16
22
[N] Cell motility
2
1
3
4
0
1
2
2
0
0
1
2
18
8
10
318
226
471
309
399
267
313
202
360
302
264
296
3727
2125
1602
Total number of genes
0.3
0.5
0.6
0.5-0.3
0.6-0.3
0.6-0.5
Effect of growth rate on nucleotide metabolism
Nucleotides are the building blocks for the biosynthesis of DNA and RNA. L. lactis
is auxotrophic and needs a precursor of purine or pyrimidine in the medium in order
to be able to grow 28,29. In the experiments performed here, all amino acids, of which
some serve as precursors in nucleotide metabolism, were provided at sufficient levels
13
. The bacterium can partly engage in nucleotide recycling by purine and pyrimidine salvage pathways. With increasing growth rate, most transcripts of the purine
(pur) regulon are more abundant (Addendum, Table A1). However, not all pur-genes
respond, such as e.g., purAFMHDE. A similar observation was made for the transcripts of pyrimidine (pyr) de novo synthesis genes. Transcription of the pyrPDbFEC
genes and carA is increased with an increasing growth rate and with comparable fold
changes, but transcription of other members of the pyr regulon, such as pyrR1BKDa
and carB remains unchanged.
51
Growth rate-dependent components of cell division and cell biogenesis
Among the most responsive categories to an increase in the growth rate are those
that encompass genes involved in cell growth and biogenesis (Table 1, categories J,M,D). These include for instance the genes of the cell division (Category
D) proteins FtsK (llmg0766), Llmg0016 and Llmg0769, as well as the genes rodA
(llmg1643) and gidA (llmg2035), which are involved in B. subtilis cell shape
determination 30 and translation 31, respectively. Not all genes involved in cell
division and biogenesis are upregulated with increasing growth rate, however.
Transcription of the gene encoding the essential cell division protein FtsZ hardly
changes, while e.g., the transcriptional activity of the gene coding for chromosomal condensing protein SMC even decreases at the highest growth rate (Table 1).
With respect to the cellular translation machinery (Category J), fusA, encoding translation elongation factor EF-G, is reduced at higher growth rates, while
for the other annotated elongation factor genes efp (llmg1880), tsf (llmg2429) and
llmg0562 mRNA levels remain more or less unchanged. Initiation factor 1 and
3 genes, infA and infC respectively, show a further increase in transcriptional
activity with increasing growth rate, while the amount of transcripts for the initiation factor 2 gene infB appears to go down, but only at the highest growth rate.
The ribosome, as a major player in translation is an important cell component responding to growth rate differences (Table 1) 32. The general trend is a small increase of transcripts for the ribosomal proteins, with the exception of rplJKL and rpsAB, for which a
relatively large increase was seen of at least 2.2-fold between the samples taken from
the 0.15 h−1 and the 0.6 h−1 cultures. L. lactis possesses two neighboring clusters of
multiple genes for ribosomal proteins (llmg2362-llmg2367; llmg2370-llmg2384). For
most of the genes in these clusters a gradual increase in transcription is observed with
increasing growth rate. The only genes that show a decrease in transcriptional activity are llmg2370 to llmg2376 and llmg2380. The amount of transcripts of ribosomal
52
proteins encoded by llmg2377, llmg2378 and llmg2379 was not significantly changed.
The transcriptional response of the aminoacyl-tRNA synthetase genes can
be divided in three categories. Transcripts of thrS, lysS, glySQ, leuS, tyrS,
ileS, hisS and argS increase in abundance with increasing growth rates,
while transcription of trpS, aspS and proS is decreased. Transcripts of
valS, alaS, ansS and cysS decrease with growth rates of 0.5 h−1 and 0.6 h−1.
The proteins (category M) encoded by the mur genes play a role in the formation of the pepti-
doglycan in the cell wall. Especially murB and murF are highly upregulated with an increasing growth rate; their products function as UDP-N-acetylenolpyruvoylglucosamine reductase and UDP-N-acetylmuramoyl-tripeptide-D-alanyl-D-alanine ligase, respectively.
Finally, no coherent increase is observed in the transcription of fatty acid biosynthesis
genes. Only llmg0538, encoding FabZ1, the putative dehydratase and isomerase of
acyl-ACP intermediates, shows a consistent increase in transcriptional activity that
appears to be related to the growth rate. The transcripts of fabD and acpA are only
upregulated from 0.3 h−1 to 0.6 h−1, while the transcripts for the acetyl coenzyme A carboxylase (ACC) complex encoded by accABCD are downregulated with an increasing
growth rate.
The influence of growth rate on the CodY and CcpA regulon
The L. lactis MG1363 genome contains at least two global regulators of metabolism, namely CodY and CcpA. The proteolytic system in L. lactis is regulated by
the transcriptional repressor CodY 33,34. First, the growth rate-dependent transcription
of genes that are part of the CodY regulon 33 are described, followed by a description of the response of other amino acid metabolism and amino acid transport genes.
The highest expression of members of the CodY regulon is observed when cells growth
at a rate of 0.3 h−1 or 0.5 h−1. At a growth rate of 0.6 h−1 the arc-genes of the arginine
deiminase pathway (ADI), are highly downregulated. A similar trend is observed for
the genes involved in the biosynthesis of isoleucine ilvACBD, of glutamate gltBPQ, of
serine and glycine serB and for the gene encoding endopeptidase PepO, a hydrolase
for oligopeptides 35. Since no (oligo)-peptides were supplied to the medium, transcriptional activity of the genes encoding the uptake machinery for these macromolecules
(oppDFABC and dtpT) is under these chemostat conditions at lowest growth rate not
changed. Additionally, at the highest growth rate strong repression of the transcription of these genes takes place, but not for dtpT. Repression of genes responsible for
the biosynthesis of histidine takes place at growth rates of 0.5 h−1 and 0.6 h−1, even
though these genes are not functional 36. While most expression levels for CodY regulon genes are lowest at the highest growth rate, the ammonium transporter gene amtB
and lysine biosynthesis gene lysA are upregulated at growth rates 0.5 h−1 and 0.6 h−1.
Other genes that play a role in amino acid metabolism and/or transport, and which
are regulated independent from CodY, follow either no specific trend with the
growth rate or a trend that is positively correlated with the growth rate. Amino acid
53
permease genes llmg0375, llmg0376, llmg1591, ctrA, araT and glycine/proline
transporter genes busAA and busAB display this positively correlated trend. The
di-/tripeptide transporter gene dtpT and the DAHP synthase gene aroF are only
upregulated at the highest growth rate. The genes of the arg cluster are mildly downregulated with an increasing growth rate, as are the aspC, cysD and glnA genes, all
involved in amino acid anabolism. Other genes involved in amino acid metabolism and/or transport do not follow a specific trend with increasing growth rate.
Under the glucose-limiting conditions used here, a response might be expected from
the CcpA regulon. Indeed, with increasing growth rates a substantial increase is seen
of transcription of ccpA and ptsH, the two main regulator genes involved in carbon
catabolite response (CCR) 37. We examined the response of the entire CcpA regulon, as
defined for L. lactis previously 27, on the variation in the growth rate of L. lactis. As can
be seen in AddendumTable A1 (or for the regulon sorted data, Chapter 3, Table S1),
expression of most genes of the CcpA regulon was significantly reduced at increasing
growth rate. Most changes (up and down) in the CcpA regulon take place between 0.5
h-1 and 0.6 h-1, when the increase in transcription of ccpA and ptsH was highest.
Discussion
A high-quality dataset provides insights in the growth rate dependence of
transcription in L. lactis
In the work presented here the transcriptional activities were compared of all genes of
the genome of L. lactis MG1363 growing at four different growth rates in chemostats.
By controlling the growth of L. lactis in chemostats at a stable temperature, pH and
nutritional environment, an industrially relevant and complete dataset of the transcriptional responses was obtained. In addition, by using long-term cultivation conditions
we attempted to mimic common industrial conditions that L. lactis encounters when
it is used as a production strain. Our experimental setup does not allow a strict comparison with a batch-culture experiment, but during a batch culture experiment a range
54
of growth rates are observed. Therefore we assume that during at a certain point in
transition phase in such a batch culture compares to the lowest growth rate used here.
Subsequently we assume that exponential batch growth rate is comparable to the highest growth rate employed here. The growth rate in our experiment was adjusted by
changing the medium influx rate until stable growth was obtained. The reproducibility
of the setup and the subsequent measurements is high, because we did chemostats
experiments at same growth rate of which the transcriptomics is in high accordance.
We obtained very high numbers of genes of which the expression level change met
the significance threshold of p-value ≤ 0.05 (Fig. 1B). This is not only due to culture
stability; also performing the analysis with a common reference increased the number
of significantly differentially expressed genes substantially. Remarkably, all direct
comparisons with a growth rate of 0.15 h-1 are less affected by the choice for Limma
analysis than the indirect comparison, e.g. between growth rates of 0.6 h-1- 0.3 h-1. This
might be because of the choice of the growth rate of 0.15 h-1 as the common reference.
An indication that the common reference analysis is performed accordingly is seen in
the differences in the Pearson’s correlation distances (PCD). The Pearson’s correlation
is relatively high when the comparing effects are overlapping. A clear trend, however,
is not visible for all genes, as some subsets of genes behave different from others (Fig.
1C), and induce no local variation of transcriptional activity in the genome (Fig. 2).
Growth rate dependence of genes involved in cell biogenesis is variable
A large fraction of the L. lactis genes does not respond to a change in the growth
rate. Part of this fraction includes genes belonging to COG-categories that are without an identified function for L. lactis. Our results suggest that these genes with an
unknown function possess a specific role that is not related to growth and growth rate.
The genes that show a positive correlation between transcriptional activity and the
growth rate are mainly involved in cell division and protein synthesis, and to a lesser
extent in stress responses and energy metabolism. This observation is in line with the
transcriptome studies performed in yeast 38,39 and with time-series transcriptomics data
of L. lactis growing in batch culture in milk 40 or GM17 (Brouwer et al., unpublished,
data not shown). In all of these studies, it was shown that the increase of transcription
of cell biogenesis genes is not necessarily positively correlated with the growth rate.
The mRNA levels of only certain genes of the transcription- and translation machineries are increased, as are those of cell division genes. By measuring transcripts of
cellular processes genes in a population of L. lactis cells, we could not observe a clear
coordination. A possible explanation might be that L. lactis is not able to synchronize the synthesis of all components for cell division, under the examined conditions,
as has been shown for the metabolic cycle in yeast 41. In that study a very specific
starvation regime had to be employed in order to allow observing a coherent tran55
scriptional response for metabolism. There are indications that cell growth and cell
division in an individual cell are linked processes 42 and that with an increase in the
growth rate, the genes for cell division in L. lactis could be transcriptionally activated.
Even though cell division is expected to be a very controlled and coordinated process
43
, our transcriptome data hints at a more strict coordination taking place at a differ-
ent level. For example, the mRNA levels of the genes of some of the cell division
proteins, like FtsK, Llmg0016 and Llmg0769 increase with increasing growth rate.
In contrast, the gene of the essential cell division protein FtsZ hardly shows a transcriptional response to an increase in the growth rate. This was also the case for B.
subtilis FtsZ, for which mRNA levels could not be linked to a changing growth rate 44.
We observe a remarkable variation in the mRNA levels of ribosomal protein genes. Most
of these are upregulated with increasing growth rate. The same subset of genes shows a
distinct transcriptional pattern in the time-series data 40 (Brouwer et al., unpublished): all
of these genes are mostly active around the mid-logarithmic phase of growth during batch
culturing (high growth rate). Also ribosomal protein genes llmg2370-llmg2380 which
were downregulated at the highest growth rates in the batch experiments, have the lowest
transcript levels at logarithmic phase of growth. All together, these results show that the
genes coding for ribosomal proteins are indeed growth rate responsive, both in batchculture as well as under artificially imposing increased growth rates using chemostats.
A different trend in transcription is observed for the growth rate dependence of genes
for initiation factors and other ribosomal modulation proteins. L. lactis has three genes
coding for an initiation factor: infABC. Transcription of infA and infC is highest at the
56
highest growth rate, yet infB is repressed at that growth rate. All three initiation factor
genes show a similar trend in transcriptional activity in an L. lactis batch culture in
milk, decreasing at late stationary phase 40. This is in contrast with results from the time
series data for L. lactis growth in rich GM17 medium (Brouwer et al., unpublished),
where infA and infC are repressed upon entry of stationary phase, but mRNA levels
of infB are unchanged throughout the log- and stationary phases. L. lactis translation
initiation factor IF-2 seems to behave differently from IF-1 and IF-3, since transcription of infB is more activated at lower growth rates. IF-2 stabilizes 70S ribosomes 45
and could protecting ribosomes against ribonuclease activity at lower growth rates.
Another ribosome protection mechanism is established with protein YfiA. In E.coli,
this protein is known to prevent ribosomal dimerization 46,47. In our studies presented
elsewhere (Chapter 5) we show that L. lactis YfiA is responsible for 70S ribosome
protection by dimerizing lactococcal ribosomes at low growth rates. Transcription of
yfiA is strongly repressed at the highest growth rate.
Carbohydrate metabolism is not transcriptionally controlled by growth rate
In this study the growth rate of L. lactis is preset by controlling and fixing the influx
rate of the medium. Under the conditions employed here, we attempted to set the
amount of glucose as the limiting factor for growth. Thus, transcriptional response
was supposed to be regarded as a direct or indirect response to a varying glucose
concentration. However, at highest growth rate, residual glucose was detected in the
effluent (Chapter 3). By maintaining high extracellular levels of glucose, it was not
limiting anymore at 0.6 h-1, while in batch growth conditions we achieved a maximal
growth rate of 0.7 h-1. We have not been able to identify the limiting factor for growth
in a chemostat, since we have not determined all intermediate metabolites and did not
define the possible toxic effect of the accumulated lactate levels at 0.6 h-1 for L. lactis
48
.
This residual glucose had a very small effect on transcription of the enzymes of the
glycolytic pathway, as we see a rather modest transcriptional response. From metabolic flux data obtained from the same chemostat cultures used here (Chapter 3) we know
that at the growth rates of 0.15 h-1 and 0.3 h-1 L. lactis ferments pyruvate more into
the end products of the mixed-acid branch, while at higher growth rates of 0.5 h-1 and
0.6 h-1 mainly lactate is produced. The metabolic shift from one type of fermentation
into the other is not readily discernable in the transcriptional activity of the glycolytic
enzymes. This is in line with observations in yeast, where it was shown that a different
carbon supply does have an effect on the in vivo glycolytic flux, but that this could
not be seen in the transcript ratio changes of the genes for the glycolytic enzymes 49.
With increasing glucose levels, an effect in CCR might have been expected 50. On
the transcriptional level this was shown to be the case. First, the transcripts of regulator genes ccpA and ptsH were highly expressed at increasing growth rate. However,
mRNA levels of these genes can only hint at CCR; they do not provide proof, particularly since HPr needs to be phosphorylated on its Ser46 residue in order to function as
a co-repressor for CcpA 51. Nevertheless, with an increasing growth rate a particular
alteration of transcriptional activity in the CcpA regulon is observed. The differences
in transcriptional activity in the CcpA regulon is not linear with growth rate increase.
CCR predominantly takes place between 0.5 h-1 and 0.6 h-1. Thus, the occurrence of
57
CCR is related to the decreasing glucose limitation and is uncoupled from growth rate.
The ratios of most transcripts of the genes that are part of the CcpA regulon decrease.
Proteins that are encoded by those genes are mainly non-glycolytic proteins/enzymes
like maltose and mannitol transporters. Besides repression, the complex of CcpA-HPr
His46-P is capable of activation of the las-operon, comprising the genes pfk, pyk and
ldh 52. However, no coordinated activation of the las-operon is observed under the
conditions employed. Only at 0.6 h-1, the ratio of pyk transcription is very slightly
increased (Addendum, Table A1). While the transcript ratios of most genes of the CcpA
regulon point at a CCR response at increasing growth rate, the lack of transcriptional
activation of the las-operon shows that the CCR response is not entirely coherent.
Genes of nucleotide metabolism are derepressed at growth rates of 0.5 h-1 and 0.6 h-1
Other parts of metabolism mostly show transcriptional activation with an increasing growth rate, as for example the nucleotide synthesis pur- and pyr genes.
Stress- and time-series experiments have indicated that those genes for the synthesis of nucleotides are attenuated at lower growth rates 40,53,54 (Brouwer et al.,
unpublished). The relationship between growth rate and mRNA abundance of pur
and pyr genes has been shown several times, and depends strongly on intermediate
molecules that are self-inhibitors. The purine genes are activated upon increasing
phosphoribosyl pyrophosphate (PRPP) levels 55, PyrR termination is promoted by
uridine monophosphate (UMP) 56 and pyrG transcription is blocked when cytidine
triphosphate (CTP) levels are low 57. So, at high growth rates, when precursors of
nucleotides are readily available, the attenuation on nucleotide synthesis genes is lost.
Growth rate limitation by glucose does not affect the CodY regulon of L. lactis
In our experimental setup, none of the amino acids in the growth medium used were
supposed to be limiting at any growth rate. Nevertheless, many transcripts of genes
involved in amino acid synthesis and transport were altered (Table 1). For this group
of genes, the global regulator CodY, involved in nitrogen metabolism, plays a central role. Most of the CodY repression takes place when nutrients are abundant, and
CodY repression is relieved upon entrance of the cells into the stationary phase (e.g.
low growth rate). In B. subtilis, CodY also represses the expression of genes that
are not directly related to nitrogen metabolism but are involved in processes such as
competence development, chemotaxis and antibiotic synthesis 58–60. L. lactis CodY
58
represses expression of a range of peptidases and peptide uptake systems, and is activated by sensing intracellular branched-chain amino acids
. In our experiments,
34,61
growth rate is limited by the amount of glucose and not dependent on the levels of
intracellular branched-chain amino acids. The effect of CodY repression was therefore
expected to be relatively low. Indeed no significant changes in codY regulon transcripts were observed. A few members of the CodY regulon (opp, pep and a range
of amino acid biosynthetic genes) are transcriptionally repressed at high growth
rates, despite the absence of complex peptides. At growth rates of 0.15 h−1 and 0.3
h−1 these genes are not repressed. Moreover, at a growth rate of 0.6 h-1 we observed
a slight downregulation of transcription of codY itself. Therefore, we estimate that
the role of CodY as a transcriptional repressor is very small under these conditions.
59
Conclusion
We hereby provide a complete and comprehensive dataset for the transcriptional
feedback of L. lactis at different growth rates. Overall, the transcriptional response
to variations in the growth rate is diverse. Each imposed growth rate in the chemostat
yields a different transcriptomic state, which is adjusted to the specific growth rate
condition. Transcriptional (de)activation is a rather fast response, and our experiments
were executed on homeostatic cultures. Nevertheless, the transcriptional responses
of many genes and gene clusters show certain trends with a growth rate increase.
For physiological interpretations, the transcriptome dataset requires data from both
the proteome and metabolome of the same experimental samples (Chapter 3). As our
experiments show, transcriptional regulation plays an important role in many cellular
components like cell division and biogenesis. For many other components in L. lactis,
transcriptional regulation at a varying growth rate results very often in a rather low
or modest alteration in transcription of the genes. For a more detailed comprehension
of the cellular rearrangements at different growth rates, levels of metabolites, and
proteins (Chapter 3), and the post-translational modifications of proteins also need to
be taken into account.
60
References
1. Pedersen, M. B., Iversen, S. L., Sørensen, K. I. & Johansen, E. The long and
winding road from the research laboratory to industrial applications of lactic acid
bacteria. FEMS Microbiol. Rev. 29, 611–624 (2005).
2. Deveau, H., Labrie, S. J., Chopin, M.-C. & Moineau, S. Biodiversity and classification of lactococcal phages. Appl. Environ. Microbiol. 72, 4338–4346 (2006).
3. Thomas, T. D., Ellwood, D. C. & Longyear, V. M. C. Change from homo- to heterolactic fermentation by Streptococcus lactis resulting from glucose limitation in
anaerobic chemostat cultures. J. Bacteriol. 138, 109–117 (1979).
4. Melchiorsen, C. R., Jokumsen, K. V., Villadsen, J., Israelsen, H. & Arnau, J. The
level of pyruvate-formate lyase controls the shift from homolactic to mixed-acid
product formation in Lactococcus lactis. Appl. Microbiol. Biotechnol. 58, 338–344
(2002).
5. Magnusson, L. U., Farewell, A. & Nyström, T. ppGpp: a global regulator in Escherichia coli. Trends Microbiol. 13, 236–242 (2005).
6. Mercade, M., Cocaign-Bousquet, M. & Loubière, P. Glyceraldehyde-3-phosphate
dehydrogenase regulation in Lactococcus lactis ssp. cremoris MG1363 or relA
mutant at low pH. J. Appl. Microbiol. 100, 1364–1372 (2006).
7. Papenfort, K. & Vogel, J. Multiple target regulation by small noncoding RNAs
rewires gene expression at the post-transcriptional level. Res. Microbiol. 160,
278–287 (2009).
8. Romby, P. & Charpentier, E. An overview of RNAs with regulatory functions in
gram-positive bacteria. Cell. Molec. Life Sci. 67, 217–237 (2009).
9. Wegmann, U. et al. Complete genome sequence of the prototype lactic acid bacterium Lactococcus lactis subsp. cremoris MG1363. J. Bacteriol. 189, 3256–3270
(2007).
10.Naidu, C. N. & Suneetha, Y. Current knowledge on microarray technology - An
overview. Trop. J. Pharm. Res. 11, 153–164 (2012).
11. Lahtvee, P.-J. et al. Multi-omics approach to study the growth efficiency and amino acid metabolism in Lactococcus lactis at various specific growth rates. Microb.
Cell Fact. 10, 12 (2011).
12.Gasson, M. J. Plasmid complements of Streptococcus lactis NCDO 712 and other
lactic streptococci after protoplast-induced curing. J. Bacteriol. 154, 1–9 (1983).
13.Goel, A., Santos, F., Vos, W. M. de, Teusink, B. & Molenaar, D. Standardized
assay medium to measure Lactococcus lactis enzyme activities while mimicking
intracellular conditions. Appl. Environ. Microbiol. 78, 134–143 (2012).
14.Larsen, R., Kloosterman, T. G., Kok, J. & Kuipers, O. P. GlnR-mediated regulation of nitrogen metabolism in Lactococcus lactis. J. Bacteriol. 188, 4978–4982
(2006).
15.Kuipers, O. P. et al. Transcriptome analysis and related databases of Lactococcus
lactis. Antonie Van Leeuwenhoek 82, 113–122 (2002).
16.Smyth, G. Bioinformatics and computational biology solutions using R and bio61
conductor (Gentleman, R., Carey, V. J., Huber, W., Irizarry, R. A. & Dudoit, S.)
397–420 (2005).
17.Tatusov, R. L. et al. The COG database: an updated version includes eukaryotes.
BMC Bioinf. 4, 41 (2003).
18.De Jong, A. Genome2D MolGen Bioinformatics server. Functional analysis
(2012).
19. Saeed, A. I. et al. TM4 microarray software suite. Meth. Enzymol. 411, 134–193
(2006).
20.Kerkhoven, R., Enckevort, F. H. J. van, Boekhorst, J., Molenaar, D. & Siezen, R. J.
Visualization for genomics: the microbial genome viewer. Bioinf. 20, 1812–1814
(2004).
21.Okumura, H. et al. Regulation of chromosomal replication initiation by oriCproximal DnaA-box clusters in Bacillus subtilis. Nucleic Acids Res. 40, 220–234
(2012).
22.Couturier, E. & Rocha, E. P. C. Replication-associated gene dosage effects shape
the genomes of fast-growing bacteria but only for transcription and translation
genes. Mol. Microbiol. 59, 1506–1518 (2006).
23.Krásný, L. & Gourse, R. L. An alternative strategy for bacterial ribosome synthesis: Bacillus subtilis rRNA transcription regulation. EMBO J. 23, 4473–4483
(2004).
24.Siezen, R. J. et al. Genome-scale diversity and niche adaptation analysis of Lactococcus lactis by comparative genome hybridization using multi-strain arrays.
Microbiol. Biotechnol. 4, 383–402 (2011).
25.Dupont, K., Janzen, T., Vogensen, F. K., Josephsen, J. & Stuer-Lauridsen, B. Identification of Lactococcus lactis genes required for bacteriophage adsorption. Appl.
Environ. Microbiol. 70, 5825–5832 (2004).
26.Chapot-Chartier, M.-P. et al. Cell Surface of Lactococcus lactis is covered by a
protective polysaccharide pellicle. J. Biol. Chem. 285, 10464–10471 (2010).
27.Zomer, A. L., Buist, G., Larsen, R., Kok, J. & Kuipers, O. P. Time-resolved determination of the CcpA regulon of Lactococcus lactis subsp. cremoris MG1363. J.
Bacteriol. 189, 1366–1381 (2007).
28.Dickely, F., Nilsson, D., Hansen, E. B. & Johansen, E. Isolation of Lactococcus
lactis nonsense suppressors and construction of a food-grade cloning vector. Mol.
Microbiol. 15, 839–847 (1995).
29.Kilstrup, M., Hammer, K., Ruhdal Jensen, P. & Martinussen, J. Nucleotide metabolism and its control in lactic acid bacteria. FEMS Microbiol. Rev. 29, 555–590
(2005).
30.Ikeda, M. et al. Structural similarity among Escherichia coli FtsW and RodA
proteins and Bacillus subtilis SpoVE protein, which function in cell division,
cell elongation, and spore formation, respectively. J. Bacteriol. 171, 6375–6378
(1989).
31.Osawa, T. et al. Conserved cysteine residues of GidA are essential for biogen62
esis of 5-carboxymethylaminomethyluridine at tRNA anticodon. Structure 17,
713–724 (2009).
32.Gausing, K. Regulation of ribosome production in Escherichia coli: Synthesis and
stability of ribosomal RNA and of ribosomal protein messenger RNA at different
growth rates. J. Molec. Biol. 115, 335–354 (1977).
33.Hengst, C. D. den et al. The Lactococcus lactis CodY regulon identification of a
conserved cis-regulatory element. J. Biol. Chem. 280, 34332–34342 (2005).
34.Guedon, E., Sperandio, B., Pons, N., Ehrlich, S. D. & Renault, P. Overall control
of nitrogen metabolism in Lactococcus lactis by CodY, and possible models for
CodY regulation in Firmicutes. Microbiol. 151, 3895–3909 (2005).
35.Tuler, T. R., Callanan, M. J. & Klaenhammer, T. R. Overexpression of peptidases
in Lactococcus and evaluation of their release from leaky cells. J. Dairy Sci. 85,
2438–2450 (2002).
36.Delorme, C., Godon, J. J., Ehrlich, S. D. & Renault, P. Gene inactivation in Lactococcus lactis: histidine biosynthesis. J. Bacteriol. 175, 4391–4399 (1993).
37.Deutscher, J., Küster, E., Bergstedt, U., Charrier, V. & Hillen, W. Protein kinasedependent HPr/CcpA interaction links glycolytic activity to carbon catabolite
repression in gram-positive bacteria. Mol. Microbiol. 15, 1049–1053 (1995).
38.Brauer, M. J. et al. Coordination of growth rate, cell cycle, stress response, and
metabolic activity in yeast. Mol. Biol. Cell 19, 352–367 (2008).
39.Castrillo, J. I. et al. Growth control of the eukaryote cell: a systems biology study
in yeast. J. Biol. 6, 4 (2007).
40.De Jong, A., Ejby, M., Kuipers, O. P., Kilstrup, M. & Kok, J. Detection of environmental changes by Lactococcus lactis MG1363 during fermentation of milk. PLoS
ONE (2013).
41.Tu, B. P., Kudlicki, A., Rowicka, M. & McKnight, S. L. Logic of the yeast metabolic cycle: temporal compartmentalization of cellular processes. Science 310,
1152–1158 (2005).
42.Qu, Z., Weiss, J. N. & MacLellan, W. R. Coordination of cell growth and cell division: a mathematical modeling study. J. Cell. Sci. 117, 4199–4207 (2004).
43.Wu, L. J. & Errington, J. Coordination of cell division and chromosome segregation by a nucleoid occlusion protein in Bacillus subtilis. Cell 117, 915–925 (2004).
44.Weart, R. B. & Levin, P. A. Growth rate-dependent regulation of medial FtsZ ring
formation. J. Bacteriol. 185, 2826–2834 (2003).
45.Tsai, A. et al. Heterogeneous pathways and timing of factor departure during
translation initiation. Nature 487, 390–393 (2012).
46.Maki, Y., Yoshida, H. & Wada, A. Two proteins, YfiA and YhbH, associated with
resting ribosomes in stationary phase Escherichia coli. Genes Cells 5, 965–974
(2000).
47. Ueta, M. et al. Ribosome binding proteins YhbH and YfiA have opposite functions
during 100S formation in the stationary phase of Escherichia coli. Genes Cells 10,
1103–1112 (2005).
63
48.Magni, C. et al. Mechanism of citrate metabolism in Lactococcus lactis: Resistance against lactate toxicity at low pH. J. Bacteriol.181, 1451-1457 (1999).
49.Daran-Lapujade, P. et al. Role of transcriptional regulation in controlling fluxes
in central carbon metabolism of Saccharomyces cerevisiae A chemostat culture
study. J. Biol. Chem. 279, 9125–9138 (2004).
50.Warner, J. B. & Lolkema, J. S. CcpA-dependent carbon catabolite repression in
bacteria. Microbiol. Molec. Biol. Rev. 67, 475 –490 (2003).
51.Gunnewijk, M. G. & Poolman, B. Phosphorylation state of HPr determines the
level of expression and the extent of phosphorylation of the lactose transport protein of Streptococcus thermophilus. J. Biol. Chem. 275, 34073–34079 (2000).
52.Luesink, E. J., Van Herpen, R. E. M. A., Grossiord, B. P., Kuipers, O. P. & De
Vos, W. M. Transcriptional activation of the glycolytic las operon and catabolite
repression of the gal operon in Lactococcus lactis are mediated by the catabolite
control protein CcpA. Molec. Microbiol. 30, 789–798 (1998).
53.Marreddy, R. K. R. et al. The response of Lactococcus lactis to membrane protein
production. PLoS ONE 6, e24060 (2011).
54. Nouaille, S. et al. Transcriptomic response of Lactococcus lactis in mixed culture
with Staphylococcus aureus. Appl. Environ. Microbiol. 02653–08 (2009).
55.Kilstrup, M. & Martinussen, J. A transcriptional activator, homologous to the
Bacillus subtilis PurR repressor, is required for expression of purine biosynthetic
genes in Lactococcus lactis. J. Bacteriol. 180, 3907–3916 (1998).
56.Lu, Y. & Switzer, R. L. Transcriptional attenuation of the Bacillus subtilis pyr
operon by the PyrR regulatory protein and uridine nucleotides in vitro. J. Bacteriol. 178, 7206–7211 (1996).
57.Jensen-MacAllister, I. E., Meng, Q. & Switzer, R. L. Regulation of pyrG expression in Bacillus subtilis: CTP-regulated antitermination and reiterative transcription with pyrG templates in vitro. Mol. Microbiol. 63, 1440–1452 (2007).
58.Lazazzera, B. A., Kurtser, I. G., McQuade, R. S. & Grossman, A. D. An autoregulatory circuit affecting peptide signaling in Bacillus subtilis. J. Bacteriol. 181,
5193–5200 (1999).
59.Mirel, D. B. et al. Environmental regulation of Bacillus subtilis sigma(D)-dependent gene expression. J. Bacteriol. 182, 3055–3062 (2000).
60.Molle, V. et al. Additional targets of the Bacillus subtilis global regulator CodY
identified by chromatin immunoprecipitation and genome-wide transcript analysis. J. Bacteriol. 185, 1911–1922 (2003).
61.Guedon, E., Renault, P., Ehrlich, S. D. & Delorme, C. Transcriptional pattern
of genes coding for the proteolytic system of Lactococcus lactis and evidence
for coordinated regulation of key enzymes by peptide supply. J. Bacteriol. 183,
3614–3622 (2001).
64
Supplementary material Table Table S1.
Overview of all genes significantly upregulated or significantly downregulated for at
least 2 out of the 6 comparisons, and functional categorized. The tables can be found
on http://www.molgenrug.nl/supplementary_data/thesisEckhardt/Chapter2_TableS1.
xlsx
65
66
Chapter 3
Metabolic regulation governs the metabolic
shift from mixed-acid to homolactic fermentation of Lactococcus lactis: A multi-level
omics study
Thomas H. Eckhardt, Anisha Goel, Pranav Puri, Filipe Santos, Douwe Molenaar,
Willem M. de Vos, Jan Kok, Bert Poolman, Anne de Jong, Fabrizia Fusetti, Bas
Teusink and Oscar P. Kuipers
67
Abstract
Using a simple model bacterial system, we test the hypothesis that investment in protein synthesis influences the metabolic strategy employed. We characterized the physiological responses of the bacterium Lactococcus lactis growing in glucose-limited
chemostats at various growth rates as relative transcription and protein ratios, enzyme
activities and fluxes. A drastic shift in carbon flux from 10 to 75% lactate formation
was accompanied by very limited changes in transcription, protein ratios or even enzyme activities, in glycolytic and fermentative pathways. These minimal changes were
reflected even in the ribosomal proteins, a major investment of cellular machinery.
Thus, contrary to the original hypothesis, L. lactis displays a strategy in which its central metabolism appears always prepared for high growth rate and in which it primarily employs regulation of enzyme activity rather than alteration of gene expression.
Only at the highest growth rate and during batch growth we observed down-regulated
stress proteins and up-regulated glycolytic proteins - triggered by glucose excess. We
conclude that for glucose, transcription and protein expression largely follow a binary
feast / famine logic.
Introduction
Microorganisms synthesize various metabolites depending on their growth conditions.
Under certain conditions, they also undergo metabolic shifts. The occurrence of metabolic shifts in a wide variety of (micro)organisms has been investigated extensively
1–8
; for a recent review see 9. Among the multitude of theories and explanations about
metabolic shifts, several consider one or more aspects of microbial physiology important, e.g. metabolism, gene expression, or competitive advantage, to name only a
few 3,10–12. Others suggest the importance of biochemical constraints 13, spatial structure 14 and limited intracellular and membrane space 15. In light of the many studies
emphasizing the role of protein cost in overall cellular behaviour 16–21, one cannot
help to wonder about a possible relationship with the metabolic shift. The theories of
limited intracellular and membrane space translate to protein investment influencing
the metabolic shift. Similarly, protein investment and metabolism were linked and
the suggestion is that evolutionary optimization of resource allocation underlies the
68
metabolic shift 22. So, a self-replicating system integrating several cellular subsystems
was proposed. The predictions of this self-replicator model lead to the hypothesis
that a trade-off between protein investment and metabolic yield ultimately governs
the metabolic strategy in a growth-optimized microbial system. Depending on the
proteins involved in the different metabolic pathway branches, investment of proteins
(enzymes) varies with varying substrate availability and consequently growth rate,
altering the metabolic profile of the microorganism. To test this hypothesis, a good
model system exhibiting a metabolic shift and the ability to quantify protein investment would be necessary. Surprisingly, only few experimental studies investigate the
metabolic shift at multiple cellular levels. We chose the model lactic acid bacterium
Lactococcus lactis because of its classic mixed-acid-to-lactic-acid metabolic shift, its
simple metabolism without respiration (in the absence of exogenously supplied hemin
23
), its relatively small and well-characterized genome and its industrial importance.
It exhibits the shift between mixed-acid and homolactic fermentation upon changing
growth rate under the same steady state environmental conditions 3. This is inherently
different from the other well-studied microbial model organisms, yeast, E. coli and B.
subtilis, with respect to the fact that L. lactis does not show a diauxic shift upon using the overflow metabolites. L. lactis also lacks an electron transport chain and the
capacity to respire under normal conditions, and hence the ATP gain in the metabolic
shift from mixed acids to lactic acid is over 10 times smaller. A well designed multilevel-omics study can provide sufficient data on protein investment as well as insights
into its regulation 24. Here, we present an inter-laboratory, standardized, multi-omics
study of the metabolic shift in L. lactis, carried out in the light of the predictions of the
self-replicator model. First, we investigate the metabolic shift at multiple cellular levels, and second, we examine growth rate-related regulation in L. lactis. We study the
metabolic shift by focussing on the energy generating pathways of L. lactis: glycolysis
and arginine metabolism and also investigate the ribosomal protein investment.
69
Results
The metabolic shift and bioenergetics
In order to study the classic metabolic shift from mixed acid fermentation (production
of formate, acetate and ethanol) to homolactic fermentation, i.e. production of lactic
acid in L. lactis in glucose-limited chemostat cultures at varying growth rates, we set
the dilution rates (D, henceforth referred to as growth rate) accordingly to 0.15 h-1; 0.3
h-1, 0.5 h-1 and 0.6 h-1, but measured the actual dilution rate on the day of harvesting.
This resulted in a slightly different dilution rate for some chemostat cultures. The largest deviation was found for one chemostat, which ran at a D of 0.45 h-1 while being
set at a D of 0.5 h-1 (Table 1). For transcriptomic and proteomic analysis, the deviation
did not result in an outlier; the samples derived from this chemostat still contribute to
the triplicate measurement. A growth rate of 0.6 h-1 is close to the maximal growth rate
in this medium, and a growth rate of 0.15 h-1 is expected to mainly yield products of
mixed acid fermentation. Indeed, L. lactis displayed a rather steep metabolic shift at
increasing growth rates (Fig. 1A). The fraction of lactate flux (normalized to the total
carbon flux) increased at higher growth rates, from about 10% to 75%. The decreasing
mixed acid branch flux and increasing lactate flux overlapped between growth rates
0.3 and 0.5 h-1. Energetically, homolactic fermentation generates 2 ATP per glucose,
while mixed-acid fermentation generates 3 ATP per glucose. Yet, a 65% increase in
lactate flux corresponding to a 22% decrease in ATP production was not accompanied
by a drastic decrease in biomass concentration, which decreased only 14% at D = 0.6
Fraction of total carbon flux
A
100%
Homolactic
B
Mixed acid
80%
60%
40%
20%
0%
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
-1
Dilution rate (h )
Figure 1. (A) Fraction of total carbon flux towards homolactic and mixed-acid branches.
(B) Total ATP formation rates calculated by substrate-level phosphorylation (Vtotal, SLP), by
the genome-scale stoichiometric network model (Vtotal, GS), and the maximum possible ATP
production rate (VATP).
70
Table 1. Biomass concentration, apparent catabolic carbon balance and total carbon balance
with standard deviations, in glucose limited chemostat cultures of L. lactis MG1363.
Dilution rate
Biomass
Catabolic
(h-1)
0.15
0.15
0.15
0.3
0.3
0.3
0.5
0.45
0.5
0.6
0.61
0.613
(gDW.L-1)
0.803 ± 0.068
0.797 ± 0.116
0.826 ± 0.017
0.842 ± 0.097
0.806 ± 0.105
0.840 ± 0.029
0.762 ± 0.023
0.790 ± 0.074
0.722 ± 0.022
0.734 ± 0.005
0.719 ± 0.002
0.641 ± 0.008
C balance %a,b
81.02 ± 8.24
84.41 ± 14.1
86.45 ± 3.82
83.48 ± 10.9
79.14 ± 11.6
77.39 ± 3.58
84.02 ± 4.28
85.96 ± 9.97
79.29 ± 4.13
72.84 ± 2.74
85.98 ± 3.61
81.85 ± 5.03
C balance %a,c
100.3 ± 9.84
103.5 ± 16.7
106.2 ± 4.54
103.7 ± 13.1
98.56 ± 13.9
97.53 ± 4.34
102.3 ± 4.77
104.9 ± 11.5
96.59 ± 4.60
90.44 ± 2.95
107.85 ± 4.05
102.51 ± 5.95
a % C-Balance = % (qC-out / qC-in); C-moles: glucose=6, lactate=3, pyruvate=3, ethanol=2, acetate=2, succinate=4,
biomass=27.8 gDW/C-mole 26
b Excluding biomass
c Including biomass
h-1 (homolactic metabolism) compared with D = 0.15 h-1 (mixed-acid metabolism)
(Table 1). The catabolic and total carbon balances were closed with less than 17%
standard deviation. We calculated the ATP formation rates by substrate-level phosphorylation (Vtotal, SLP) and also via the genome-scale stoichiometric network model
(Flahaut, et al., manuscript submitted), (Vtotal, GS) and the maximum possible ATP production rate (vATP) (Fig. 1B) 25. The curve Vtotal, SLP plateaus above D = 0.5 h-1. Thus, at
D = 0.6 h-1 cells grow faster than at D = 0.5 h-1, but at the same rate of ATP formed per
unit biomass. The curve Vtotal, GS is steeper, but also plateaus. The maximal ATP formation rate increases linearly to D=0.5 h-1 but remains stable at the highest growth rate
of D = 0.6 h-1.
Fluxes do not relate to Vmax, protein- or transcript abundance
We obtained fluxes by averaging the flux ranges predicted by flux variability analysis
(FVA) on the genome-scale model of L. lactis MG1363 (Fig. 2). The glycolytic flux
linearly increased with growth rate. The flux through the homolactic-branch enzyme
71
mRNA
Glucose
Protein
2
ATP
GLK
ADP
2
1
0
1
●
●
●
●
●●
●
−1
−2
NADP+
G6PDH
G6P
2
NADPH −1
Pentose −2
phosphate 2
pathway 1
PGI
0
F6P
PFK
ADP
●
●
●●
ALD
●
●
●
●
●
DHAP
Pi
TPI
NAD+
NADH
PGK
2
1
0
●
●
●
●
●
●●
●
Pyruvate
CoA
0
1200
1
900
0
2
●
●
●
●
●
●
PFL
Pi
Acetyl CoA
NADH
ALDH
PTA
0
1
0
0
20
●
●
●
●
●●
●
0
●
●
●●
●
●
●
●●
Ethanol
0
●
●
●
●
●
●
●●
−1
−2
2
●
●
●
●
●
●
●
●
1
0
●
●
●
−1
●●
●
2
●
●
●
●
●
●
1
0
−2
2
●
●
●
●
●
●
●
2
1
10
50
40
30
20
10
0
120
90
60
0
250
200
150
100
50
0
●
●
0.2
0.1
100
50
●
●
●
●
●
● ●
●
● ●
●
●●
●
●
●
●
●
0.0
●
●
●
●
●
●
●
●
● ●
20
●
●
●
15
●
●
●
10
5
●
●
●
●
●
●
●
●
●
●
●
●
15
●
●
●
10
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
15
●
●
●
●
●
●
●●
60
40
20
●
●
●
●
● ●
●●
●
● ●
●●
●
●
●
●
20
●
15
●
●
●
●
200
100
●
●
●
●
10
5
●
●
●
5
● ●
●
0
●
●
●
15
10
●
●
10
●
●
●
●
●●
●
20
5
●
●
● ●
●
0
●
●
●
●
●
20
5
20
10
●
●
0
●
●
●
●
●
●
0
●
●
40
●
●
●
● ●
●
●
●
●
30
●●
●
10
●
●
●
20
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
40
●
●
●
30
●
●
●
20
10
3
2
1
●
●
0
●
30
20
10
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
0
●
●
●
●
●
●
●
●
●
●
●
40
●
●
●
30
10
●
●
●
20
●
2
1
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
40
●
●
●
30
●
●
●
20
10
4
●
●
2
●
●
●
●
●
●
●
●
●
0
●
●
●
●
●
●
●
●
●
15
5
●
●
●
●
●
●
●
●
●●
●
● ●
●
●●
●
●
●
20
●
●
10
●
●
●
●
0
●
●
●
20
●
●
●
●
●
●
●
●
●
●
20
10
●
● ●
●
●●
●
0
30
−2
●
●
−2
●
300
200
100
●
●
●
●
●
10
0
● ●
●
●
●●
●
●
●
1
●
●
●
●
●
●
●
0
−1
−1
−2
−2
●
●
●
●
●
●
15
2
1
1
●
●
●
●
●●
●
0
0
2
10
●
●
●
●
●
●●
●
●
●
●
●●
●
0
−1
−1
−2
−2
1
1
2
●
●
●
200
●
●
●
●
●
2
●
●
●
●
●
●
0
−1
−1
−2
−2
0.2 0.3 0.4 0.5 0.6
●
●
●
●
●
●
●
●
●
●
5
300
1
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
0.2 0.3 0.4 0.5 0.6
●
●
●
4
4
●
●
●
●
2
●
●
●
●
0.2 0.3 0.4 0.5 0.6
0
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
2000
1000
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
6
●
●
0
60
40
20
●
●
●
0
●
●
●
●
2
2
0
20
●
4
6
●
100
10
6
0
0
2
1
0
●●
0
ACK
ADH
●
●
15
−2
−2
NAD+
●
●
●
1
●
●
●
●
0
20
2
●
●
●
●
100
5
−1
0
●
●●
−1
−1
0
●
1
0
10
200
●
2
●
●
●
●
●
●
0
−2
●
●
●
1
0
0
●
300
−1
1
0
●●
●
1
●
●
●
1
0
−1
●
2
1
0
●
●
−2
1
0.3
●
●
300
2
−1
Acetate
600
−2
−2
NADH
●
●
●
1
−1
Acetaldehyde
●
●
−2
NAD+ Acetyl-P
ADP
ATP
●
●
−1
2
CoA
0
−1
2
Lactate
Formate
200
−2
−2
NAD+ LDH
●
●●
1
−1
NADH
●
−2
2
PYK
●
●
●
100
−2
ATP
0
●
●
●
300
1
−1
−1
ADP
40
−1
2
PEP
●
●
0
−2
H2O
●
●
−2
−1
ENO
0
●
0
60
●
●●
Vmax/flux
●
●
●
●
50
−2
2
2PGA
●●
●
20
−2
PGM
●
−1
−1
3PGA
●
●
●
●
●
2
−1
2
ATP
0
Flux
●
4
●●
●
2
−2
ADP
●
●
0
1
−1
2,3BPG
6
●
●
100
−2
●
●
●
●
1
1
●
●
●
4
−2
−2
2
GAPDH
0
−1
●
●
6
0
2
●
●
2
GA3P
●●
−2
−1
2
FBP
●
●
2
−1
0
●
●
●
●
1
●
●
2
ATP
0
−1
2
1
0
Vmax
10
5
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
0.2 0.3 0.4 0.5 0.6
0.2 0.3 0.4 0.5 0.6
Figure 2. 2Log regulation ratios of mRNA and proteins, Vmax’s and fluxes in mmol.gdw-1.h-1 and
Vmax/flux ratios of glycolytic and fermentative pathways at various dilution rates. We estimated
all metabolic fluxes by averaging the flux ranges predicted by flux variability analysis (FVA)
on the genome-scale model of L. lactis MG1363, constrained by all measurements of nutrient
consumption and product formation rates. Gray areas represent standard error. (s)AckA1 (l)
AckA2, PGM transcript shows llmg0355 (gpmA), ADH transcript shows adhE.
72
lactate dehydrogenase (LDH) increased nonlinearly, and fluxes through the mixedacid-branch enzymes pyruvate formate lyase (PFL), acetate kinase (ACK) and alcohol dehydrogenase (ADH) increased at D = 0.3 h-1 and then decreased further on. In
contrast to these changes in fluxes, the Vmax’s, protein- and transcript ratios altered
very little. (Fig. 2, supplementary material, Table S1 and S2).
The Vmax’s and protein ratios of all glycolytic enzymes remained more or less constant
up to D = 0.5 h-1 except for PFL, of which the protein ratios decreased linearly with a
growth rate above 0.15 h-1. Between 0.5 h-1 and 0.6 h-1, the enzymes encoded by the
las operon, phosphofructokinase (PFK), pyruvate kinase (PYK) and LDH showed increases in Vmax‘s and protein ratios. Apart from these enzymes, phosphoglucose isomerase (PGI), fructose bisphosphate aldolase (ALD), triosephosphate isomerase (TPI),
phosphoglycerate kinase (PGK) and phosphoglycerate mutase (PGM) all showed a
rise in Vmax value at D = 0.6 h-1. Of these, ALD and TPI also showed a rise in protein
ratios. Glucokinase (GLK) Vmax showed a gradual decrease overall, while the GLK
protein ratio increased at D = 0.6 h-1. The enzymes involved in the mixed-acid fermentation pathway showed a decreasing trend. Phosphotransacetylase (PTA) protein
ratios, and ADH and ACK Vmax and protein ratios decreased at D = 0.6 h-1. The two
copies of acetate kinase showed antagonistic behaviour. At higher growth rate, the
ackA1 transcript increased while the protein ratio was constant, whereas the ackA2
AA
B
B
●
1
●
PFK
●
● ●
●●
●●●
● ● ●
●● ●●●
●
●
●
●
●●
●
●
●
●● ●
●
●
●●●
●
●●
●
●●
●●
●
log2Vmax ratio
0
Enzyme
●
●
−1
Dilution
0.15 - 0.3 0.15 - 0.5 0.15 - 0.6 0.3 - 0.5 0.3 - 0.6 0.5 - 0.6
rate
PGI
●
●
●
GLK
ACK
ADH
ALD
TPI
GAPDH
PGK
Dilution rate PGM
●
−2
0.3
0.5
0.6
ENO
PYK
>1
1
0.8
0.6
0.4
0.2
LDH
−3
PTA
r=0.589
−1
0
ACK
ADH
0
<0
log2Protein ratio
Figure 3. (A) Correlation between Vmax and protein 2log regulation ratios of all glycolytic
and fermentative enzymes at various growth rates. Enzyme activities for which the regulation
ratio deviates by at least 1 in one of the measurements are coloured. (B) Metabolic regulation
coefficients for different growth rate pairs.
73
transcript was constant while the relative AckA2 protein ratio decreased to a third
compared with that at 0.3 h-1. The correlation between enzyme activities and their
respective protein ratios as measured in the proteome studies are shown in Fig. 3A.
Except for GLK, ACK and ADH, most enzymes lie on the perfect correlation line with
a slope of 1 denoting excellent correspondence between regulation of protein ratios
and regulation of enzyme activities, but overall, the changes in Vmax and protein ratios
do not correlate proportionally with the changes in fluxes.
At the transcript level, pgiA (PGI), fbaA (ALD) and gpmA (PGM) were linearly activated with growth rate (Fig. 2). Glycolysis in L. lactis is regulated at different levels
by a variety of mechanisms. One of the predominant ones is carbon catabolite repression (CCR) orchestrated by carbon catabolite control protein CcpA 27,28. The number
of significantly regulated transcripts of genes of the CcpA regulon increased, as well
as their fold-changes, with higher growth rate (supplementary material, Table S1).
Genes displaying CCR have an upstream binding motif (cre-site) to which the CcpAHPr Ser46-P complex can bind 27,29. An important part of the CcpA regulon is the las
operon. The transcriptional activity of ldh, pyk and pfk is enhanced upon binding of the
CcpA complex in the las promoter region 30. Even though at high growth rates CCR is
functional both at the transcriptional and protein level, no coherent influence of CcpA
on the transcription of glycolytic genes is seen. For instance, at a growth rate of 0.6 h-1
only the pyk gene showed a significant transcriptional increase, while transcription of
ldh and pfk genes did not change. Thus, overall we did not observe many significant
changes in transcription of genes encoding glycolytic enzymes, except for a few (butAB and pgiA) at the highest growth rate of 0.6 h-1 (supplementary material, Table S1).
The metabolic shift is predominantly regulated at the enzyme activity level
To investigate the growth rate-related flux regulation in L. lactis, we used regulation
analysis 31 for a quantitative analysis of the control of glycolytic and fermentative
fluxes with increasing growth rate. The hierarchical regulation coefficient (ρh) represents the extent of flux regulation through gene expression and via changes in enzyme
concentration. It can be defined as:
74
ρh 
 lnVi d ln ei d ln ei


 ln ei d ln J
d ln J
for a pathway flux J, with concentration ei of enzyme i which carries a flux at a rate
Vi. The coefficient ρh for a set of two dilution rates was thus calculated as the ratio of
the difference in the logarithm of the fluxes at both dilution rates to the difference in
the logarithm of the enzyme activity. The metabolic regulation coefficient (ρm) repre-
sents the extent of flux regulation as a result of metabolic regulation of enzyme activity, defined as
 lnV  ln x
ρm  

 ln x  ln J
ρh  ρm  1
x
At steady state, the sum of the regulation coefficients ρh and ρm is one. In our data,
hierarchical regulation coefficients of the glycolytic and fermentative pathways are
close to zero (Fig. 3B) because the flux increases substantially, while the Vmax remains
more or less unchanged except for the change in D from 0.5 h-1 to 0.6 h-1, where the
Vmax values increase. In other words, the metabolic coefficients are close to 1, representing the constant maximal enzyme activity over the change in glycolytic flux and
the metabolic shift. All ρm’s between D = 0.15 h-1 and 0.5 h-1 are above 0.8. Between
D = 0.3 h-1 and 0.6 h-1, ρm’s are lower, indicating partial hierarchical regulation, and
from 0.5 h-1 to 0.6 h-1, except for glyceraldehyde phosphate dehydrogenase (GAPDH),
PTA and ACK, the ρm’s are zero, indicating complete hierarchical regulation. ADH is
hierarchically regulated at all growth rate-pairs except for 0.15 h-1 to 0.5 h-1.
Ribosome investment
One of the predictions of the self-replicator model 22 is the proportional increase in
investment in ribosomal protein (rProtein) with increasing growth rate. The largest
fraction of lactococcal RNA consists of ribosomal RNA (rRNA). In E. coli, rRNA is
approximated to amount to 85% of the total RNA (totRNA) in the cell 32. The total
amount of RNA divided by the total amount of protein (totProt) is an accepted method
to calculate the ribosomal content of a cell. We therefore quantified totRNA and totProt
for a fixed cell density for all chemostat samples. The totRNA/totProt ratio increased
linearly with, but not proportionally to the growth rate, and its increase relative to the
lowest growth rate levels off at the highest growth rate (Fig. 4). For the transcript and
protein ratios of rProteins, a very small increase is seen relative to that of the lowest
growth rate (Fig. 4). The rProtein ratios show a somewhat steeper increase at increasing growth rate than the mRNA ratios of rProteins (Fig. 4). The largest increase is
observed in the ratio of totRNA/totProt, an indication for the amount of rRNA (Fig. 4).
75
BB
●
2
●
●
0.5
●
1
●
0.0
0
●
●
0.3
0.4
Dilution rate (h-1)
0.5
0.6
D
D
2
1
log2 protein ratio
log2 Ribosomal protein ratio
C
●
●
●
1.0
0.2
C
●
●
log2 mRNA ratio
log2[total RNA / total Protein] ratio
AA 1.5
0
0.3
0.4
0.5
Dilution rate (h−1)
0.6
0.3
0.4
0.5
0.6
0.2
0.3
0.4
0.5
Dilution rate (h−1)
2
1
0
Dilution rate (h−1)
Figure 4. Change in totRNA/totProt ratio (A), mRNA ratios of rProteins (B) and rProtein ratios relative to the lowest growth rate 0.15 h-1, for L. lactis (C) (this study) and relative rProtein
ratios from the proteome data of E. coli (D) 33.
Amino acid consumption
To better understand the overall metabolism of L. lactis, we also quantified the amino
acid consumption rates (supplementary material, Fig. S1). The consumption rates of
most amino acids steadily increased with the increase in the growth rate. Aspartate and
glutamate measurements could not reliably be determined as the concentrations were
too close to the detection limit. The production of ornithine, citrulline and ammonia
steadily increased with concomitant consumption of arginine (Fig. 6, supplementary
material, Fig. S1), which changed in a bow-like fashion with growth rate. At 0.15 h-1,
arginine was consumed at a rate of 0.28 mmol∙gdw-1∙h-1; this consumption rate gradually increased up to 0.8 mmol∙gdw-1∙h-1 at a growth rate of 0.5 h-1. Subsequently, however, at growth rate 0.6 h-1, the consumption rate dropped to a value lower than that at
growth rate 0.15 h-1. The same pattern was observed at the transcript and protein ratios
of the responsible proteins ArcABC1C2D1D2. An initial increase in the expression
76
ratios of the genes arcAC1C2D1 was followed by a significant reduction of the same
transcripts plus arcB when comparing dilution rates 0.5 h-1 and 0.6 h-1. At the protein
level, ArcABC2 showed a similar trend. The ArcA protein ratio did not change between 0.15 h-1 and 0.5 h-1, a decrease was only seen at 0.6 h-1.
Fatty acid biosynthesis
Another module in the self-replicator model 22 is that of lipid biosynthesis as an essential part of membrane biogenesis. The composition of acyl chains in the phospholipids
of lactococcal membranes was investigated at varying growth rates. Most of the fatty
acid biosynthesis genes are organised in one large operon in L. lactis, and are regulated by FabT (this thesis, Chapter 4). Transcription of these genes does not change
coherently with increasing growth rate (supplementary material, Table S1). Only at
the highest growth rate transcription of fabZ1 (llmg0538) is upregulated significantly.
The dehydratase FabZ1 in Enterococcus faecalis is known to function as an isomerase
that decreases the acyl-chain saturation ratios 34. Thus, we determined the length and
degree of saturation of the acyl chains of the lactococcal cell membrane at different
growth rates (Fig. 5). At the lowest growth rate, the cell membranes contain more
saturated acyl chain (66%) than at the other growth rates. Also, the short C14 and C16
Figure 5. Acyl chain composition analysis with increasing growth rate. Saturated acyl chains
are shown in blue, while unsaturated acyl chains are shown in red. Lighter colours indicate
longer acyl chains. Shown are the averages of three biological experiments.
77
A
B
Figure 6. (A) Schematic overview of the arginine metabolic pathway in relation to the relative
transcription and protein ratios and the metabolic fluxes per enzyme. With the available omics
data fluxes were generated; empty graphs indicate that no significant data was available. (B)
Representation of the different repression mechanisms that balance arginine metabolism. At
low glucose concentrations, ArgR (homohexamer) blocks the catabolic arc-operon at low
arginine levels. At high arginine levels, ArgR and AhrC form a heterohexameric complex that
blocks the anabolic arg-operon 35. At high glucose concentrations, repression of the arc genes
at high arginine levels is taken over by carbon catabolite repression through CcpA/HPr-Ser-P.
78
Table 2. Stress related proteins from GO annotation, as determined by EBI genome
reviews.
Llmg0080
Llmg0093
Llmg0638
Llmg0986
Llmg1080
Llmg1088
Llmg1350, Lmg1351, Llmg1352
Llmg1574
Llmg1575
Llmg1576
Llmg1662, Llmg2023
Llmg2047
Llmg2302
osmotically inducible protein C
hypothetical protein
ATP-dependent Clp protease
ATP-dependent Clp protease
universal stress protein A2
glutathione peroxidase
putative tellurium resistance protein
molecular chaperone DnaK
heat shock protein GrpE
heat-inducible transcription repressor
universal stress protein A
universal stress protein E
non-heme iron-binding ferritin
acyl chains are more abundant at the lowest growth rate (69%). Both the saturation
rate and the contribution of short acyl chain lengths decrease significantly when the
growth rate increases. At the highest growth rate the degree of saturation is 54% and
the percentage of small acyl chains is 55%.
Growth rate-related stress
Proteins and their transcripts related to stress were obtained from the GO-database
(Table 2) 36. Most stress-related transcript and protein ratios remain unchanged between the growth rates of 0.3 h-1 and 0.5 h-1, while most changes, if any, take place between growth rates of 0.5 h-1 and 0.6 h-1 (Fig. 7A). The correlation between transcripts
and proteins for this class is relatively high with a r of 0.632 (Fig. 7B).
List of putative isoenzymes
The proteome of L. lactis, as obtained by the annotation of the genome sequence contains a number of isoenzymes that could effectively perform the same function. This
feature is not unique to L. lactis, since it is prevalent in a multitude of organisms. The
possible reason for having isoenzymes is that it offers flexibility of regulation under
a wide variety of environmental conditions 37. We classified all the observed proteins
of our dataset according to their annotated enzyme functions to compile a “putative”
79
A
Stress-related ProteinsStress-related Transcripts
log2mRNA ratio
B
●●
●
●●
●●
●
●
●
●
●
●●
●
●●
●
●
●●● ●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●
●
●●
●●
●
●● ●
●
●
●●
●
● ●●●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●●
●●
●
●
●
● ●● ●
●
●
●●●
●●●
●
● ● ●●
● ●● ●
●
●
●
●
●●
●
●
●
0
−1
●
●
−2
Dilution
rate (h−1)
● 0.3
● 0.5
● 0.6
r=0.632
−1
0
log2Protein ratio
1
Figure 7. (A) Stress-related proteins and their transcript ratios at increasing growth rate. Gray
bars indicate the standard deviation. (B) Correlation between stress-related transcript and
protein ratios.
isoenzyme list of 115 enzyme groups (supplementary material, Table S3). In about
half of the enzyme groups, proteins were not detected in our samples.Next to that,
there was considerable variation in the spectral counts of the detected peptides. This
variation might indicate which proteins are more prominently involved in carrying out
the respective reactions under glucose-limited chemostat growth of L. lactis. When a
protein is undetected by our experimental setup this does not confirm its absence, but
it does indicate that such protein may be present at very low quantities. This kind of
information can be useful for (iso)enzyme studies with respect to engineering metabolic pathways under specific environmental conditions.
80
Discussion
Experimental design
In this study we characterized the metabolic shift from mixed-acid to homolactic fermentation in the model lactic acid bacterium L. lactis. We designed our experimental
setup, and based it on the predictions of the self-replicator model 22, to test the hypothesis that shifts in metabolic strategies are outcomes of evolutionary optimization of
resource allocation. The proteome data show that a major fraction of total protein is
invested in glycolysis and ribosome synthesis (supplementary material, Table S2). To
quantify the protein investment in L. lactis, we measured the maximal enzyme activities and relative protein levels. To investigate possible regulation in protein investment, genome-wide responses in transcription were examined. We characterized the
metabolic shift by also measuring extracellular metabolites. Intracellular fluxes were
inferred by variability analysis on the genome-scale model, taking the average values
to depict each flux.
There are no studies characterising the metabolic shift of L. lactis at the multi-level
scale and only a few studies investigated the response of growth rate on changing
environmental conditions. Of these, one focussed on amino acid metabolism in L.
lactis strain IL1403 by growing chemostat cultures under isoleucine limitation 38. This
experimental design resulted in glucose excess in all chemostats. Since glucose is a
preferred substrate, glucose repression was constantly on in this study and the metabolic shift was not observed. Another study did look at glucose limitation by growing the same strain in accelerostats, but the metabolic shift to mixed-acid fermentation was not observed at the lower growth rates 39. L. lactis strain ML3, an ancestor
from MG1363 40, is known to exhibit the metabolic shift at different growth rates in
glucose-limited chemostats 3. Therefore, we characterized the glucose-limited, growth
rate-mediated metabolic shift in L. lactis and employed inter-laboratory standardized
protocols. A graphical representation of the experimental setup is shown in Fig. 8.
Lack of correlation between transcriptomics and proteomics data
The correlation between the data of the transcriptome and the proteome experiments
was generally low with r = 0.262. When looking at individual functional categories,
for instance the CcpA regulon, a linear dependence is seen for all significantly altered
81
transcripts and proteins with r = 0.536 (supplementary material, Table S4). Similar
results were obtained for several other pathways represented in the KEGG database,
with the highest correlation coefficients being 0.61 for starch and sucrose metabolism
and 0.56 for glycine, serine and threonine metabolism. All other pathways had lower
linear dependencies or contained less than 10 genes (supplementary material, Table
S4). Altogether the change in protein ratios is not proportional to the change in transcript ratios. In a recent chemostat study, in which the growth rate of one culture of
L. lactis was gradually increased, the correlation between data of transcriptome and
proteome was up to 0.69 39. Using a similar experimental setup, comparable correlations were seen between the transcriptome and proteome of E. coli 33. In independent
chemostat cultures of Saccharomyces cerevisiae, for instance, transcriptional changes
do not largely contribute to glycolytic behaviour 11,41. It is more likely that the glycolytic fluxes are regulated at the post-transcriptional level, partly explaining the poor
correlation between gene transcription ratios and protein expression ratios. Other studies in yeast show that genes with a high correlation between their output at the transcriptomic and proteomic level are often adaptive. Examples are genes involved in
stress-response and usually not genes encoding proteins involved in core metabolic
functions like glycolysis 42,43. The fact that a correlation between mRNA and protein
ratios is lacking for glycolytic enzymes suggests that an mRNA-buffer exists for glycolysis genes in L. lactis. Indeed, protein turnover and mRNA lifetimes are important
posttranscriptional aspects playing key regulatory roles 44,45.
82
Correlation of fluxes with proteomics: overcapacity of enzymes over flux
The glycolytic flux increases proportionally with the growth rate. However, this is
not the case for transcripts, proteins and enzyme activities. In fact, the ratio of Vmax/
flux shows that the enzyme activities are much higher than the actual flux inside the
cell at all growth rates for all enzymes except for PGM, GAPDH and enolase (ENO)
(Fig. 2). The overcapacity of enzymes suggested by the Vmax/flux ratios can explain
the unchanged enzyme levels in spite of a four-fold increase in the flux. Therefore, it
is not surprising to see metabolic regulation coefficients with a value of 1 because the
increase in flux might simply be because of a higher glucose concentration resulting in
enzymes with higher flux. What is most surprising is that the enzymes of the mixedacid branch and the lactic-acid branch also show an overcapacity. We cannot, however,
ignore the possibility that enzyme activities were measured in vitro and might over-
estimate in vivo Vmax’s. GAPDH, PGM and ENO are exceptions because of technical
issues; as detailed in the Materials and Methods section, GAPDH and PGM activi-
ties were determined by old sub-optimal enzyme assays since the in vivo-like assay
medium 46 specifically designed for this study had not been finalized yet. The activity
of ENO is sensitive to ammonium sulphate, present in the buffer because of adding
coupling enzymes that are suspended in ammonium sulphate solution 46. PFL activity
could not be measured, as the enzyme is prone to oxidation 47. It was shown before that
protein ratios of PFL is a determining factor for the metabolic shift from mixed-acid 48.
Indeed we see PFL protein ratios go down with increasing growth rate (Fig. 2).
Correlation of Vmax with protein ratios
One would normally expect a good correlation between the amount of an enzyme
and the activity of that enzyme because the Vmax is a product of total enzyme concentration and the catalytic turnover number (kcat). Of course, this is only true, when
post-translational modification affecting the enzyme activity does not occur. When
looking at the correlation of Vmax and the respective protein ratio for all measured
enzymes (Fig. 3A), we found that GLK, ACK and ADH do not lie on the perfect correlation line (slope = 1). This might indicate that these enzymes undergo post-translational modification resulting in differential regulation of enzyme activity without
much change in the protein level. Post-translational modification of proteins is abundant in eukaryotes 49 as well as prokaryotes 50. A recent study revealed phosphorylation of proteins at the amino acid residues serine, threonine and tyrosine in a strain
of L. lactis 51. The importance of the post-translational modification as a regulatory
mechanism has generated increased attention and a database of phosphorylated proteins (PHOSIDA) has recently been developed 52.
It is interesting to note that of the enzymes that do not perfectly correlate with their
protein ratios, one is involved in phosphorylating glucose and two (ACK and ADH)
are involved in the metabolic shift. Although surprising at first, it seems logical that
the enzymes of the mixed-acid branch (ACK, ADH) show indications of post-translational modification, because they are important in determining the metabolic flux in
either direction of the metabolic shift in L. lactis. It is unclear why GLK appears in
this database of phosphorylated proteins. L. lactis predominantly uses PTS systems
for glucose transport, while GLK is only useful when a glucose permease (GlkU) is
used. There could be two explanations for this phenomenon. One, GLK might un83
dergo post-translational modification as its need will most likely rise at fast growth.
Second, we might be missing an activator for GLK during Vmax assay measurements
with resultant underestimated GLK Vmax’s.
The increase of ribosomes is not proportional to the increase in growth rate
In order to define the relationship between growth rate and ribosomal content in L.
lactis, we characterized the amount of both ribosomal RNA and total protein for each
growth rate. The ratio of totRNA over totProt is proportional to growth rate can be
equated with the proportionality of ribosome abundance with growth rate. Since, an
increase in growth rate demands a higher mRNA translation capacity of the cell. Gausing reported that a rise in the growth rate of E. coli is concomitant with an increase
in rProtein synthesis. From a specific growth rate of 0.6 h-1 to 2.2 h-1 this increase is
even proportional 53. In other words, doubling the growth rate of E. coli also doubles
the amount of ribosomes. In our experiments we observed a gradual, albeit non-proportional increase in ribosomal content with increasing growth rate. This was for a
major part caused by a steep rise in the amount of rRNA accompanied by only a slight
increase in the rProtein pool with increasing growth rates. If we take into account that
a rather stable level of mRNA of most of the related rProtein yields more synthesized
protein at higher growth rates, our data indicate a restrained yet efficient strategy for
total ribosomal content of a lactococcal cell. At all growth rates, a certain amount
of mRNA for rProteins exists that can be used by the cell to synthesize the required
amount of rProteins, as demanded by the growth rate.
The results presented here support the idea that the major limitation for an increase in
the ribosomal content is the synthesis of rRNA, as shown by the upwards linear trend
of the totRNA/totProt ratio (Fig. 4A). Even when the rProteins are limiting at high
growth rates, this does not necessarily mean that ribosome activity is affected. First of
all, most rProteins are located on the outside of the ribosome, and do not play a direct
role in protein synthesis 54. Secondly, part of the rProteins can be removed without a
loss in ribosome activity 55,56, which then points toward the idea that rRNA as such is
indeed functional as a ribozyme without the requirement of rProteins 57.
The relationship between ribosome abundance and ribosome synthesis rate is not obvious. In the assembly of a ribosome, the early-assembly rProteins are thought to
structure the rRNA in such a way that it functions as a ribosome 58. In our dataset the
amount of early-assembly rProteins shows a gradual increase with increasing growth
84
rate (supplementary material, Fig. S2), but it is not proportional to the rRNA synthesis
rate as derived from the totRNA/totProt ratio (Fig. 4A). So even the minimally required subset of rProteins does not follow the same trend as the rRNA production rate
with an increasing growth rate. In the proteomic dataset of E. coli, growing at different
growth rates 33, the relative ratios of rProteins show trends comparable to those in our
study for the rProtein synthesis rate (Fig. 4C and D). Since not many available datasets
provide this type of rProtein abundances from cells growing at varying growth rate, it
is too early to speculate on a general phenomenon for bacteria. However, we seem to
have found that at increasing growth rates ribosomes are less decorated with rProteins
than at lower growth rates.
L. lactis strategy to avoid degradation of ribosomes at low growth rates
The fact that at a low growth rate the ratio of rRNA/rProt is much lower than at high
growth rates strengthens the idea that when the growth rate decreases excess rRNA
is degraded until the moment where the rProteins prevent total breakdown of the ribosome. In E. coli 59 and Staphylococcus aureus 60 it has been shown that ribosomes
tend to dimerize when the cells reach stationary phase. Because detached ribosomal
subunits are the actual targets of ribonuclease activity 61, dimerized ribosomes have an
increased protection against RNA degrading enzymes. The ribosome dimerizing factors in E. coli possess a mutual homology with the YfiA protein in L. lactis. The level
of yfiA transcripts is downregulated when growth rate increases. We have shown recently that L. lactis YfiA is involved in ribosome dimerization when the cells enter the
stationary phase (this thesis, Chapter 5). Dimerization of ribosomes by YfiA is seen as
a possible strategy for L. lactis when, at lowest growth rates, part of the ribosome pool
is protected against degradation. Since the cells conserves their ribosomes at lower
growth rate, they allow a quick recovery of translation, and thus of growth, as soon as
the cells encounter new resources in the environment.
Arginine metabolism as a function of the growth rate
Arginine serves as a source of carbon and nitrogen in L. lactis. This strain can catabolize arginine to obtain an additional ATP via the arginine deiminase (ADI) pathway
62,63
. The concomitant production of ammonia leads to less acidification of the environment 64. Arginine degradation via ArcA produces citrulline, which can be further
catabolized by the ornithine carbamoyltransferase ArcB either to carbamoyl-P or or85
nithine. ArcC then degrades carbamoyl-P into ammonia and carbon dioxide, producing 1 mole of ATP per mole of arginine. Ornithine is exchanged for arginine by the
arginine-ornithine antiporter ArcD1/2 62,65. The transcriptomic- and proteomic analyses reveal that up to a growth rate of 0.5 h-1 arginine catabolism increases, after which
a steep decrease occurred at a growth rate of 0.6 h-1, leading to a strongly reduced
amount of arginine catabolic enzymes and their transcripts. This was confirmed by the
consumption rates of arginine and the production rates of ornithine, citrulline and ammonia. The upstream region of the arginine catabolic gene cluster arcABD1C1C2TD2
contains 6 ARG boxes for binding of ArgR6 66, a putative CodY operator site and a
cre-site for CcpA binding 27. At high levels of arginine, the two arginine regulators
ArgR and AhrC derepress the ADI pathway 67. With increasing growth rates, arginine
degradation could function as a glycolysis-independent system for ATP generation.
However, at a growth rate of 0.6 h-1 arginine might be required for biomass production,
which would, reduce the intracellular arginine level and repress the arginine catabolic
arc operon (Fig. 5). If the arginine concentration were indeed low at a high growth
rate, a transcriptional response of the arginine biosynthetic genes argCJBF, argGH
and gltSargE might be expected. However, no significant upregulation of arginine
biosynthesis is seen. It is therefore more likely that CcpA repression causes the steep
decrease in ADI activity as at near-maximal growth rates the amount of residual glucose quickly increases (Fig. 5B), which would lead to carbon catabolite repression of
the arc operon.
Behaviour at near-maximal growth rate
The limits of growth were approached in the condition with D = 0.6 h-1. We suppose
that L. lactis MG1363 cannot grow much faster under these conditions, and with the
growth rate employed here cells were approaching “a washout”. The biomass went
down at highest growth rates (Table 1), and the concentration of residual glucose in
the chemostats changed quite drastically from undetectable in the chemostats at D =
0.15 h-1 up to 0.5 h-1 to a few mM at the highest dilution rate, 0.6 h-1. The medium was
supplemented with all components up to concentrations that only glucose was the
limiting factor. The amounts of some stress proteins went down at the highest growth
rate. This is reminiscent to the feast / famine kind of behaviour seen in B. subtilis 68
and E. coli 69, which comprises cells being prepared for all kinds of stresses at low
growth rate and then suddenly investing much less in stress machineries when they
86
encounter high glucose concentrations, at higher growth rates. The transcript ratios
of the CcpA and HPr encoding genes increased significantly with a rising growth rate
(supplementary material, Table S1). Importantly, the phosphorylation state of HPr, i.e.
HPr-Ser-P, increases with growth rate and this may directly impact the glycolytic and
other metabolic rates 70,71. Both the transcriptome and proteome are significantly, but
not drastically, reorganized upon the growth rate at 0.6 h-1 (supplementary material,
Table S1 and S2). Altogether, the impact of growing at a growth rate close to maximal
as compared to lower growth rates yield mild differences in the proteome and the respective transcriptome of L. lactis.
Why does L. lactis behave differently, compared to other microorganisms?
As has been detailed above, L. lactis maintains an almost identical proteome throughout a wide range of growth rates, which is, to say the least, non-optimal in view of
protein costs. An assumption of the self-replicator model is that there is a choice between a metabolically (energetically) efficient (high-yield) and kinetically efficient
(high-rate) pathway. In L. lactis the mixed-acid route is analogous with the former
and the lactate route with the latter. From our data it seems that L. lactis fermentation
type is largely determined by effectors that change in response to external factors.
When L. lactis is growing fast, the level of determining factors like PFL protein ratios is reduced (Fig. 2), thereby reducing the inhibition of the lactate flux 48. Another
candidate for regulation of LDH activity is fructose-bis-phosphate (FBP) 72. The flux
of FBP formation increases linearly with growth rate (Fig. 2) 73. Both the decrease of
PFL protein levels and the increase for FBP formation flux confirm the observations
from earlier experiments. When the growth rate increases, a shift towards lactate fermentation takes place under the influence of both the reduced protein level (PFL) and
an increase of FBP at high growth rates 73. Despite our multi-omics approach we can
however not exclude that other glycolytic intermediates and/or unknown factors play
an important role in the metabolic shift from mixed-acid to homo-lactic fermentation.
Another inherent assumption of the self-replicator model is that organisms are evolutionarily optimized, and in this optimal state, differential protein allocation results in
a metabolic shift. If this assumption of evolutionary optimization were true, then the
actual test of the hypothesis would be via long-term evolution experiments. In such
a case one explanation for the non-conformation of L. lactis to the protein economy
hypothesis is that it might not be evolutionarily optimized for the conditions tested in
87
the laboratory. Native to a rich environment of milk, this microorganism might have
been selected for growth on high sugar concentrations, always facing enough substrate
to support heavy investments in protein. The cost of using resources scantily and not
being able to use an unexpected abundance of nutrients might in fact be penalizing in
a rich substrate environment in the presence of many competitors.
Materials and methods
Strain and growth medium
Lactococcus lactis ssp. cremoris MG 1363 74 was grown on chemically defined medium for prolonged cultivation (CDMPC) as described by Santos et al., (manuscript in
preparation) with 25 mM glucose as the limiting nutrient and the medium composition
as detailed before 46.
Culture conditions
Glucose-limited chemostat cultures were grown in 2 L bioreactors with a working
volume of 1.2 L at 30°C, under continuous stirring. The headspace was flushed at 5
headspace volume changes per hour, with a gas mixture of 95% N2 (99.998% pure)
and 5% CO2 (99.7% pure) with oxygen impurity less than 34 vpm. A pH of 6.5±0.05
was maintained by automatic titration with 5 M NaOH. Fermenters were inoculated
with 4% (v/v) of standardized precultures consisting of 45 mL of CDMPC inoculated
with 300 µL of a glycerol stock of L. lactis MG 1363 and incubated for 16 h at 30°C.
After batch growth until an optical density at 600 nm (OD600) of around 1.8, medium
was pumped at the appropriate dilution rate (0.15, 0.3, 0.45, 0.5, 0.6 h-1).
Harvesting of cells from chemostats
The chemostats were harvested assuming a steady state at 10 working volume changes
75
. At harvest, the medium inflow was stopped and the entire culture in the chemostat
was pumped out at a high flow rate into sampling tubes placed on ice; the whole
procedure taking less than 90 s. Samples were collected for cell density, extracellular
metabolite analysis, DNA microarray analysis, enzyme activity assays and finally for
proteomic and fatty acid composition analysis.
88
Cell density
Cell density was measured spectrophotometrically at 600 nm and calibrated against
cell dry weight measurements performed in triplicate for each sample as follows. 4 mL
of culture was filtered through a pre-dried, pre-weighed 0.2 µm cellulose nitrate filter
(Whatman GmbH, Dassel, Germany), washed twice with deionized water and dried
to a constant weight. For one unit change of optical density, the change in dry weight
was determined to be 0.31±0.02 g.L-1.OD600-1.
Fermentation end-product, ammonia and amino acid analysis
Supernatant samples from medium bottles and chemostat fermentations were prepared by filtering through a 0.20 mm polyethersulfone (PES) filter (VWR international B.V., Amsterdam, the Netherlands) and storing the flow-through at -20ºC until further analysis. Extracellular concentrations of lactate, acetate, ethanol, formate,
and glucose were determined by High Performance Liquid Chromatography (HPLC)
on a Shimadzu (Tokyo, Japan) LC-10AT liquid chromatograph equipped with a Shimadzu (Tokyo Japan) RID-10A refractive index detector for ethanol and glucose,
and a Shimadzu (Tokyo, Japan) SPD-10AVP UV-Vis absorbance detector set at 210
nm for the remaining metabolites. Separation was carried out on a Bio-Rad Aminex
Ion exclusion HPX-87H column equilibrated at 55ºC with an isocratic flow of 5 mM
H2SO4 set to 0.5 mL · min-1. The injection volume used was 50 mL and concentrations
were estimated in triplo by comparison of peak areas to a calibration curve obtained
with standards analyzed under the same conditions. Residual glucose concentrations
were determined by enzymatic coupling with NADP+ in an assay containing 100 mM
HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid)-KOH, 5 mM MgSO4,
2 mM ATP, 4.5 mM NADP+, 1.5 U∙mL-1 hexokinase, 1 U∙mL-1 glucose-6-phosphate
dehydrogenase (G6PDH) and sample or standard. Ammonia concentrations were
measured using a commercially available ammonia assay kit (catalogue no. AA0100,
Sigma-Aldrich, St. Louis, MO). Amino acid vials were prepared in 950 μL volumes by
adding 25 μL each, of 0.1 M Borate, 1 mM of Norvaline as an internal standard, and
culture supernatant samples, and the rest, milli-Q water. Amino acid concentrations
were determined with an AminoQuant 1090 High Performance Liquid Chromatography (Shimadzu, Kyoto, Japan). After precipitation of proteins at 4°C by addition
of four volumes of methanol, samples were chemically modified (derivatization in
presence of 3-mercaptopropionic acid by ortho-phthalaldehyde and 9-fluorenylmethyl
89
chloroformate for primary and secondary amino acids, respectively), separated with
a C18 column and detected by spectrophotometry at 338 and 262 nm. Fluxes qi (in
mmol∙gdw-1∙h-1) were calculated as: qi = D x (Ci,supernatant - Ci,medium)/Xbiomass, where C
is the concentration of compound i (mmol∙L-1), Xbiomass is the biomass concentration
(gDW∙L-1), and D is the dilution rate (h-1).
Enzyme activities: sampling, cell extract preparation and assay conditions
An amount of cell culture containing 100 mg dry weight was centrifuged (4°C, 5
min, 8,000 rpm), washed once and resuspended in 3 to 6 mL 50 mM HEPES-KOH at
pH 7.5, containing 15% glycerol supplemented with Halt Protease Inhibitor single-use
cocktail, EDTA-free (Thermo Fischer Scientific, Rockford, IL). This suspension was
divided into 0.5 mL aliquots added to 0.5 mg glass beads with 100 µm diameter (BioSpec Products, Bartlesville, OK) in screw capped tubes, snap-frozen in liquid nitrogen
and stored at -20°C until further analysis. Frozen samples were thawed on ice and
MgCl2 was added to a final concentration of 2 mM. Cells were disrupted in a FastPrep
FP120 homogenizer (BIO 101, Vista, CA) at a speed setting of 6, in 3 bursts of 20 s,
with 120 s intermittent cooling. After centrifugation (4°C, 10 min, 10,000 g), the supernatant was collected and a series of dilutions were prepared, which were used immediately for enzyme assays. Protein concentrations of cell extracts were determined
on the same day by the bicinchoninic acid (BCA) method 76 with a BCA Protein Assay
Kit (Pierce, Thermo Fisher Scientific) using bovine serum albumin (BSA, 2 mg·mL−1
stock solution; Pierce), containing 2 mM MgCl2 and Halt Protease Inhibitor cocktail,
as the standard. Enzyme activities were assayed at 30°C at pH 7.5 in freshly prepared
cell extracts within 2 weeks of harvesting the chemostats. The enzymes GLK, G6PDH, PGI, PFK, ALD, TPI, GAPDH, PGK, PGM, ENO, PYK, LDH, ACK, PTA, ADH
and aldehyde dehydrogenase (ALDH) were assayed with the in vivo-like assay medium (version 1) as described 46 with the following differences: the coupling enzymes
were not desalted, GAPDH was assayed with 5 mM arsenate 77 and PGM was assayed
in the absence of activator 2,3-bisphosphoglycerate. ALDH activity was not detected.
All assays were checked for linearity and proportionality with increasing cell extract,
with at least 4 technical replicates. The values obtained from the assays yield the total
activity of all isoenzymes in the cell extract and are expressed as the rate of substrate
converted, relative to total protein in the extract. Obtained activities in μmol∙ min-1∙
mg protein-1 converted to fluxes (in mmol∙gdw-1∙h-1) by multiplying activities with the
90
ratio of total protein content per dry weight estimated for each chemostat culture.
Total cell protein and total RNA
Cells from each chemostat were taken and lysed with 2% SDS and incubated at 96°C
for 2 h. The total amount of protein in the obtained cell lysates was determined by using a BCA Protein Assay Kit (Pierce, Thermo Fisher Scientific) using BSA (2 mg·mL−1
stock solution; Pierce) as the standard 76. Total protein data was obtainedfrom technical triplicates and biological triplicates. TotRNA was acquired from cell lysate after
cell disruption (Qiagen Tissue Lyser, 15 Hz, 2 cycles, 5 min each), extracted by phenol/ chloroform/ isoamylalcohol (25:24:1 v/v), and extracted again with chloroform/
isoamylalcohol (24:1 v/v). The totRNA was precipitated by the addition of isopropanol and by adding KAc to a final concentration of 150 mM, supplemented with
Diethyl Phosphorocyanidate (DEPC). By vacuum-centrifugation the solvents were
removed from the RNA. Finally, samples were mixed in MilliQ-DEPC until completely dissolved and measured at 260 nm by NanoDrop (Thermo Fisher Scientific
Inc.). Data reported is an average of technical duplicates for each biological sample.
DNA microarray analysis
L. lactis cells (2 x 30 mL) were harvested by centrifugation (5 min, 4500 g); pellets
were immediately frozen in liquid nitrogen and stored at -80°C. For RNA isolation the
frozen cells were thawed on ice. Subsequent cell disruption, RNA purification, reverse
transcription and Cy3/Cy5 labeling were done as described previously 78. Labeled cDNAs were hybridized to full-genome DNA microarray slides of L. lactis MG1363 79,
with the addition of probes for rProteins. All reagents and glassware for RNA work
were treated with DEPC. RNA, cDNA quantity and quality, and the incorporation of
the cyanine-labels were examined by NanoDrop (ThermoFisher Scientific Inc.) at 260
nm for RNA and cDNA, 550 nm for Cy3, and 650 nm for Cy5. The four chemostats
with increasing growth rate were run as biological triplicates. Thus, three times the
samples of an increasing growth rate were compared directly with each other in combination with a dye-swap (Fig. 8). DNA microarray slide images were analyzed using
ArrayPro 4.5 (Media Cybernetics Inc., Silver Spring, MD). Filtering of bad- and lowintensity spots and signals, data parsing, automated grid-based Lowess normalization,
scaling, data visualization and outlier detection were performed using the Limmapackage 80. We used the common reference design in which direct and indirect com91
Ef°Bl
x-Rl
Xbiomass
BSo
FWlbiomassWlBSi
CC
BB
l
l
QRdl4WlRdlBO4
FWlbiomassWlBSo
4Rll
BUMPBl
Mixed acid
Waste
Culture density
Proteomics
f-4lμm
filter
q°B
ql
f-4lμm
PES
filter
Rl
xfll
ll
Sampling
l4l
Homolactic
f-xRlllllllf-Ellllllf-qRlllf-Rllf-8
Ul°h(xM
l8-Rllf-fR
Metabolites
HPLB
lll
AA
4l
(4f°B
4Rfl
mL
PllllBl°xfflμl-mL(xM
WashedWlllxfll
ll°RfllWllT-RWl
xRdlM
DNA microarray
(Af°B
q°B
(Af°B
4l
Efl
mL
4l
Bl
RN/l
l
ByE)ByRlll
RN/
l4l
Enzyme activities
l4l
l
4ffl
mL
BlWlEQll
xfflll
Ultracentrifuged
WashedWlllE(8lll
Rflll°lT-RMl
PlxRdll
Pll
l
fraction
l
fraction
l4l
f-E/
f-E:
f-EB
f-R/
f-RB
f-8/
f-8:
f-8B
f-RB
f-8/
f-8:
lE
f-E:
Bl
l4l
f-E/
/nalysis
/lq-RWl(l
q°B
lll
in vivo(ll
f-R/
Ulylf-8l(x
f-xRB f-xRB
lx
Bl
Ulylf-El(x
f-E:
ll
lllllA(l/l
f-E/
Ulylf-Rl
(4f°B
(x
f-xR: f-xR: f-xR:
Ulylf-xRl
(x
f-xR/ f-xR/ f-xR/
f-Rl
l+
Figure 8. Experimental setup. (A) Glucose limited chemostats of 1.5 L volume were
fed with chemically defined medium for prolonged cultivation (CDMPC) with 25
mM glucose at a flow rate F mL·h-1. CSi is the concentration of nutrients in the inflow,
CSo, that in the chemostat and outflow, Xbiomass, the biomass concentration inside the
chemostat and outflow. (B) Various dilution rates were chosen to span the metabolic
shift of L. lactis. (C) Chemostats were harvested after 10 volume changes for samples
to determine culture density, mRNA and protein levels, and enzyme activities.
92
parisons were used to increase statistical significance. Fold changes are considered to
be significantly altered when the p value ≤ 0.05.
Proteomic analysis
For protein expression profiling 2 x 250 mL of culture from each chemostat was collected by directly pouring it in pre-chilled centrifuge bottles containing chloramphenicol at a final concentration of 10 μg·mL-1 (2.5 mL stock solution, 10 mg·mL-1). The
cells were harvested by centrifugation (4°C, 5 min, 8,000 rpm). Supernatant was discarded and the pellet was washed with 50 mL of wash buffer (50 mM HEPES-NaOH
pH 7.5, 15% glycerol) and centrifuged. The washed cell pellets were resuspended in
10 mL wash buffer, frozen in liquid nitrogen and stored at -80°C. Cells corresponding
to OD600 of 50 in a total volume of 6 mL with 1 mM MgCl2 were disrupted at 39 kPsi
with a Constant Systems cell disrupter. The crude cell lysates were centrifuged (4°C,
15 min, 12,000 g); the supernatant was carefully recovered and subsequently centrifuged (4°C, 15 min, 267,000 g). The supernatant, containing the soluble fraction was
removed and stored at -80°C. The residual membrane fraction was washed once and
finally resuspended in 500 μL of wash buffer and stored at -80°C. Protein concentrations for both soluble and membrane fractions were determined with BCA kit (Pierce).
For Trypsin digestion 50 mg of protein was resuspended in 50 mL of 500 mM TEAB,
2% acetonitrile and 0.08% SDS. The disulfide bonds were reduced with 3 mM Tris
(2-carboxymethyl) phosphine hydrochloride, and the cysteine residues were modified
with 4 mM iodoacetamide. The 8-plex iTRAQ labeling was performed three times
(Fig. 8), according to the manufacturer’s protocol with few modifications as described
81
. The peptide mixture was subjected to chromatography and spectrometric analysis.
The pre-fractionation of peptides was performed on a silica based polysulfoethyl aspartamide strong cation exchange (SCX) column (catalogue number: 202SE0502 Poly
LC inc., Columbia).
93
Proteomic data analysis and statistics
Raw proteome for each sample data consisting of four sets of 8-plex iTRAQ signal
strengths annotated with a peptide and protein identifiers. Two data sets each originated from membrane and soluble protein fractions. Membrane and soluble protein
fraction were analyzed separately. Peptide identifiers could only be compared within
and not between an 8-plex iTRAQ data set. Individual samples within an 8-plex dataset were signal normalized by LOESS regression on an M-A transformation of the
signals, as is common in microarray analysis. Using the assumption that the bulk of
log-transformed signal ratios between different samples or between replicates will be
ideally located symmetrically around 0 (no regulation) independently of the signal
strength underlies this normalization technique. Since this technique is used originally
when comparing only two samples, an adaptation for 8 samples was made. LOESS
normalization was performed for each of the 28 unique pairs of samples within an
8-plex set, and these normalizations were reconciled by linear modeling. The normalized data were used to fit the logarithmically transformed ratios of protein amounts
at the different growth rates (relative to growth rate 0.15 h-1) taking into account the
additional effects of peptide and iTRAQ 8-plex set.
Fatty acid composition analysis
Samples from L. lactis chemostats, were pelleted and washed as described for proteomic analysis. All samples were transmethylated and analyzed on a gas chromatograph for acyl chain composition according to the methods described 82. The data
presented is an average from the biological triplicates.
Constraint-based modelling: flux balance analysis and flux variability analysis
The genome scale metabolic model was based on that of L. lactis MG1363 (Flahaut,
et al., manuscript submitted) with modifications in growth and maintenance energy
parameters which were estimated as described earlier 25. The network was constrained
with all measured experimental fluxes with the objective of maximising ATP dissipation to estimate the maintenance coefficient as the maximum ATP dissipation rate, and
the ATP requirement for precursor biosynthesis was estimated by the reduced cost
of biomass flux for ATP dissipation. This exercise was repeated to calculate the ATP
parameters for each dilution rate resulting in 12 models. Flux variability analysis at a
fixed growth rate was carried out for all models and the flux distribution was obtained
94
by calculating the average of the flux range for each individual flux. All analyses were
carried out using the web-based modelling tool: Flux Analysis and Modelling Environment (FAME) 83.
Acknowledgements
This work is supported by the Dutch Technology Foundation STW (grant 08080)
which is part of the Netherlands Organisation for Scientific Research (NWO) and
partly funded by the Ministry of Economic Affairs, Agriculture and Innovation, the
Kluyver Centre for Genomics of Industrial Fermentation and the Netherlands Consortium for Systems Biology (NCSB), within the framework of the Netherlands Genomics Initiative (NGI) / NWO.
95
References
1. Wolfe, A. J. The acetate switch. Microbiol. Mol. Biol. Rev. 69, 12–50 (2005).
2. Sauer, U. & Eikmanns, B. J. The PEP-pyruvate-oxaloacetate node as the switch
point for carbon flux distribution in bacteria. FEMS Microbiol. Rev. 29, 765–794
(2005).
3. Thomas, T. D., Ellwood, D. C. & Longyear, V. M. Change from homo- to heterolactic fermentation by Streptococcus lactis resulting from glucose limitation in
anaerobic chemostat cultures. J. Bacteriol. 138, 109–117 (1979).
4. Neves, A. R., Pool, W. A., Kok, J., Kuipers, O. P. & Santos, H. Overview on
sugar metabolism and its control in Lactococcus lactis - the input from in vivo
NMR. FEMS Microbiol. Rev. 29, 531–554 (2005).
5. Postma, E., Verduyn, C., Scheffers, W. A. & Van Dijken, J. P. Enzymic analysis
of the Crabtree effect in glucose-limited chemostat cultures of Saccharomyces
cerevisiae. Appl. Environ. Microbiol. 55, 468–477 (1989).
6. Huberts, D. H. E. W., Niebel, B. & Heinemann, M. A flux-sensing mechanism
could regulate the switch between respiration and fermentation. FEMS Yeast Res.
12, 118–128 (2012).
7. Warburg, O. On the origin of cancer cells. Science 123, 309–314 (1956).
8. Gatenby, R. A., Gillies, R. J. & Brown, J. S. Evolutionary dynamics of cancer
prevention. Nat. Rev. Cancer 10, 526–527 (2010).
9. Goel, A., Wortel, M. T., Molenaar, D. & Teusink, B. Metabolic shifts: a fitness
perspective for microbial cell factories. Biotechnol. Lett. (2012).
10.Vemuri, G. N., Altman, E., Sangurdekar, D. P., Khodursky, A. B. & Eiteman, M.
A. Overflow metabolism in Escherichia coli during steady-state growth: transcriptional regulation and effect of the redox ratio. Appl. Environ. Microbiol. 72,
3653–3661 (2006).
11. Daran-Lapujade, P. et al. Role of transcriptional regulation in controlling fluxes
in central carbon metabolism of Saccharomyces cerevisiae A chemostat culture
study. J Biol. Chem. 279, 9125–9138 (2004).
12.Piskur, J., Rozpedowska, E., Polakova, S., Merico, A. & Compagno, C. How did
Saccharomyces evolve to become a good brewer? Trends Genet. 22, 183–186
(2006).
13.Heinrich, R., Schuster, S. & Holzhütter, H. G. Mathematical analysis of enzymic reaction systems using optimization principles. Eur. J. Biochem. 201, 1–21
(1991).
14.Kreft, J.-U. Biofilms promote altruism. Microbiol. 150, 2751–2760 (2004).
15.Zhuang, K., Vemuri, G. N. & Mahadevan, R. Economics of membrane occupancy and respiro-fermentation. Mol. Syst. Biol. 7, 500 (2011).
16.Dean, A. M., Dykhuizen, D. E. & Hartl, D. L. Fitness as a function of beta-galactosidase activity in Escherichia coli. Genet. Res. 48, 1–8 (1986).
17.Dong, H., Nilsson, L. & Kurland, C. G. Gratuitous overexpression of genes in
Escherichia coli leads to growth inhibition and ribosome destruction. J. Bacte96
riol. 177, 1497–1504 (1995).
18.Snoep, J. L., Yomano, L. P., Westerhoff, H. V. & Ingram, L. O. Protein burden in
Zymomonas mobilis: negative flux and growth control due to overproduction of
glycolytic enzymes. Microbiol. 141, 2329–2337 (1995).
19.Dekel, E. & Alon, U. Optimality and evolutionary tuning of the expression level
of a protein. Nature 436, 588–592 (2005).
20.Stoebel, D. M., Dean, A. M. & Dykhuizen, D. E. The cost of expression of Escherichia coli lac operon proteins is in the process, not in the products. Genetics
178, 1653–1660 (2008).
21.Shachrai, I., Zaslaver, A., Alon, U. & Dekel, E. Cost of unneeded proteins in E.
coli is reduced after several generations in exponential growth. Mol. Cell 38,
758–767 (2010).
22.Molenaar, D., Van Berlo, R., De Ridder, D. & Teusink, B. Shifts in growth strategies reflect tradeoffs in cellular economics. Mol. Syst. Biol. 5, 323 (2009).
23.Duwat, P. et al. Respiration capacity of the fermenting bacterium Lactococcus
lactis and its positive effects on growth and survival. J. Bacteriol. 183, 4509–
4516 (2001).
24.Wessely, F. et al. Optimal regulatory strategies for metabolic pathways in Escherichia coli depending on protein costs. Mol. Syst. Biol. 7, 515 (2011).
25.Teusink, B. et al. Analysis of growth of Lactobacillus plantarum WCFS1 on a
complex medium using a genome-scale metabolic model. J. Biol. Chem. 281,
40041–40048 (2006).
26.Oliveira, A. P., Nielsen, J. & Förster, J. Modeling Lactococcus lactis using a
genome-scale flux model. BMC Microbiol. 5, 39 (2005).
27.Zomer, A. L., Buist, G., Larsen, R., Kok, J. & Kuipers, O. P. Time-resolved determination of the CcpA regulon of Lactococcus lactis subsp. cremoris MG1363.
J. Bacteriol. 189, 1366–1381 (2007).
28.Warner, J. B. & Lolkema, J. S. CcpA-dependent carbon catabolite repression in
bacteria. Microbiol. Mol. Biol. Rev. 67, 475 –490 (2003).
29.Schumacher, M. A., Sprehe, M., Bartholomae, M., Hillen, W. & Brennan, R. G.
Structures of carbon catabolite protein A–(HPr-Ser46-P) bound to diverse catabolite response element sites reveal the basis for high-affinity binding to degenerate DNA operators. Nucleic Acids Res. 39, 2931–2942 (2011).
30.Luesink, E. J., Van Herpen, R. E., Grossiord, B. P., Kuipers, O. P. & De Vos, W.
M. Transcriptional activation of the glycolytic las operon and catabolite repression of the gal operon in Lactococcus lactis are mediated by the catabolite control protein CcpA. Mol. Microbiol. 30, 789–798 (1998).
31.Ter Kuile, B. H. & Westerhoff, H. V. Transcriptome meets metabolome: hierarchical and metabolic regulation of the glycolytic pathway. FEBS Lett. 500,
169–171 (2001).
32.Bremer, H. & Dennis, P. P.in Escherichia coli and Salmonella 1553–1569
(1996).
97
33.Valgepea, K. et al. Systems biology approach reveals that overflow metabolism
of acetate in Escherichia coli is triggered by carbon catabolite repression of acetyl-CoA synthetase. BMC Syst. Biol. 4, 166 (2010).
34.Lu, Y.-J., White, S. W. & Rock, C. O. Domain swapping between Enterococcus
faecalis FabN and FabZ proteins localizes the structural determinants for isomerase activity. J. Biol. Chem. 280, 30342 –30348 (2005).
35.Larsen, R. Transcriptional regulation of central acid metabolism in Lactococcus
lactis. (2005).
36.Ashburner, M. et al. Gene ontology: Tool for the unification of biology. The gene
ontology consortium. Nat. Genet. 25, 25–29 (2000).
37.Postmus, J. et al. Isoenzyme expression changes in response to high temperature
determine the metabolic regulation of increased glycolytic flux in yeast. FEMS
Yeast Res. 12, 571–581 (2012).
38.Dressaire, C. et al. Growth rate regulated genes and their wide involvement in
the Lactococcus lactis stress responses. BMC Genomics 9, 343 (2008).
39.Lahtvee, P.-J. et al. Multi-omics approach to study the growth efficiency and
amino acid metabolism in Lactococcus lactis at various specific growth rates.
Microb. Cell Fact. 10, 12 (2011).
40.Le Bourgeois, P., Lautier, M., Van den Berghe, L., Gasson, M. J. & Ritzenthaler,
P. Physical and genetic map of the Lactococcus lactis subsp. cremoris MG1363
chromosome: comparison with that of Lactococcus lactis subsp. lactis IL 1403
reveals a large genome inversion. J. Bacteriol. 177, 2840–2850 (1995).
41.De Groot, M. J. L. et al. Quantitative proteomics and transcriptomics of anaerobic and aerobic yeast cultures reveals post-transcriptional regulation of key cellular processes. Microbiol. 153, 3864–3878 (2007).
42.Brauer, M. J. et al. Coordination of growth rate, cell cycle, stress response, and
metabolic activity in yeast. Mol. Biol. Cell. 19, 352–367 (2008).
43.Castrillo, J. I. et al. Growth control of the eukaryote cell: a systems biology
study in yeast. J. Biol. 6, 4 (2007).
44.Picard, F., Dressaire, C., Girbal, L. & Cocaign-Bousquet, M. Examination of
post-transcriptional regulations in prokaryotes by integrative biology. C. R. Biol.
332, 958–973 (2009).
45. Taniguchi, Y. et al. Quantifying E. coli proteome and transcriptome with singlemolecule sensitivity in single cells. Science 329, 533–538 (2010).
46.Goel, A., Santos, F., Vos, W. M. de, Teusink, B. & Molenaar, D. Standardized
assay medium to measure Lactococcus lactis enzyme activities while mimicking
intracellular conditions. Appl. Environ. Microbiol. 78, 134–143 (2012).
47.Melchiorsen, C. R. et al. Synthesis and posttranslational regulation of pyruvate
formate-lyase in Lactococcus lactis. J. Bacteriol. 182, 4783–4788 (2000).
48.Melchiorsen, C. R., Jokumsen, K. V., Villadsen, J., Israelsen, H. & Arnau, J. The
level of pyruvate-formate lyase controls the shift from homolactic to mixed-acid
product formation in Lactococcus lactis. Appl. Microbiol. Biotechnol. 58, 338–
98
344 (2002).
49.Gnad, F. et al. High-accuracy identification and bioinformatic analysis of in vivo
protein phosphorylation sites in yeast. Proteomics 9, 4642–4652 (2009).
50.Macek, B. et al. Phosphoproteome analysis of E. coli reveals evolutionary conservation of bacterial Ser/Thr/Tyr phosphorylation. Mol. Cell Proteomics 7,
299–307 (2008).
51.Soufi, B. et al. The Ser/Thr/Tyr phosphoproteome of Lactococcus lactis IL1403
reveals multiply phosphorylated proteins. Proteomics 8, 3486–3493 (2008).
52.Gnad, F., Gunawardena, J. & Mann, M. PHOSIDA 2011: The posttranslational
modification database. Nucleic Acids Res. 39, 253–260 (2011).
53.Gausing, K. Regulation of ribosome production in Escherichia coli: Synthesis
and stability of ribosomal RNA and of ribosomal protein messenger RNA at different growth rates. J. Mol. Biol. 115, 335–354 (1977).
54.Klein, D. J., Moore, P. B. & Steitz, T. A. The roles of ribosomal proteins in the
structure assembly, and evolution of the large ribosomal subunit. J. Mol. Biol.
340, 141–177 (2004).
55.Bubunenko, M. et al. 30S ribosomal subunits can be assembled in vivo without
primary binding ribosomal protein S15. RNA 12, 1229–1239 (2006).
56.Noller, H. F., Hoffarth, V. & Zimniak, L. Unusual resistance of peptidyl transferase to protein extraction procedures. Science 256, 1416–1419 (1992).
57.Steitz, T. A. & Moore, P. B. RNA, the first macromolecular catalyst: the ribosome is a ribozyme. Trends Biochem. Sci. 28, 411–418 (2003).
58.Nierhaus, K. H. The assembly of prokaryotic ribosomes. Biochimie 73, 739–755
(1991).
59.Ortiz, J. O. et al. Structure of hibernating ribosomes studied by cryoelectron tomography in vitro and in situ. J. Cell Biol. 190, 613 –621 (2010).
60.Ueta, M., Wada, C. & Wada, A. Formation of 100S ribosomes in Staphylococcus aureus by the hibernation promoting factor homolog SaHPF. Genes Cells 15,
43–58 (2010).
61.Zundel, M. A., Basturea, G. N. & Deutscher, M. P. Initiation of ribosome degradation during starvation in Escherichia coli. RNA 15, 977 –983 (2009).
62.Poolman, B., Driessen, A. J. & Konings, W. N. Regulation of arginine-ornithine
exchange and the arginine deiminase pathway in Streptococcus lactis. J. Bacteriol. 169, 5597–5604 (1987).
63.Budin-Verneuil, A., Maguin, E., Auffray, Y., Ehrlich, D. S. & Pichereau, V. Genetic structure and transcriptional analysis of the arginine deiminase (ADI) cluster in Lactococcus lactis MG1363. Can. J. Microbiol. 52, 617–622 (2006).
64.Lu, C.-D. Pathways and regulation of bacterial arginine metabolism and perspectives for obtaining arginine overproducing strains. Appl. Environ. Microbiol. 70,
261–272 (2006).
65.Driessen, A. J., Poolman, B., Kiewiet, R. & Konings, W. Arginine transport in
Streptococcus lactis is catalyzed by a cationic exchanger. Proc. Natl. Acad. Sci.
99
84, 6093–6097 (1987).
66.Larsen, R., Van Hijum, S. A. F. T., Martinussen, J., Kuipers, O. P. & Kok, J.
Transcriptome analysis of the Lactococcus lactis ArgR and AhrC regulons. Appl.
Environ. Microbiol. 74, 4768–4771 (2008).
67.Larsen, R., Buist, G., Kuipers, O. P. & Kok, J. ArgR and AhrC are both required
for regulation of arginine metabolism in Lactococcus lactis. J. Bacteriol. 186,
1147–1157 (2004).
68.Buescher, J. M. et al. Global network reorganization during dynamic adaptations
of Bacillus subtilis metabolism. Science 335, 1099–1103 (2012).
69.Ferenci, T. The spread of a beneficial mutation in experimental bacterial populations: the influence of the environment and genotype on the fixation of rpoS mutations. Heredity (Edinb) 100, 446–452 (2008).
70.Gunnewijk, M. G. & Poolman, B. Phosphorylation state of HPr determines the
level of expression and the extent of phosphorylation of the lactose transport protein of Streptococcus thermophilus. J. Biol. Chem. 275, 34073–34079 (2000).
71.Gunnewijk, M. G. et al. Hierarchical control versus autoregulation of carbohydrate utilization in bacteria. J. Mol. Microbiol. Biotechnol. 3, 401–413 (2001).
72.Wolin, M. J. Fructose-1,6-diphosphate requirement of Streptococcal lactic dehydrogenases. Science 146, 775–777 (1964).
73.Even, S., Lindley N. D. & Cocaign-Bousquet, M. Molecular physiology of sugar
catabolism in Lactococcus lactis IL1403. J. Bacteriol. 183, 3817-3824 (2001).
74.Gasson, M. J. Plasmid complements of Streptococcus lactis NCDO 712 and
other lactic streptococci after protoplast-induced curing. J. Bacteriol. 154, 1–9
(1983).
75.Even, S., Lindley, N. D. & Cocaign-Bousquet, M. Transcriptional, translational
and metabolic regulation of glycolysis in Lactococcus lactis subsp. cremoris MG
1363 grown in continuous acidic cultures. Microbiol. 149, 1935–1944 (2003).
76.Stoscheck, C. M. Quantitation of protein. Meth. Enzymol. 182, 50–68 (1990).
77.Garrigues, C., Loubiere, P., Lindley, N. D. & Cocaign-Bousquet, M. Control of
the shift from homolactic acid to mixed-acid fermentation in Lactococcus lactis: predominant role of the NADH/NAD+ ratio. J. Bacteriol. 179, 5282–5287
(1997).
78.Larsen, R., Kloosterman, T. G., Kok, J. & Kuipers, O. P. GlnR-mediated regulation of nitrogen metabolism in Lactococcus lactis. J. Bacteriol. 188, 4978–4982
(2006).
79.Kuipers, O. P. et al. Transcriptome analysis and related databases of Lactococcus
lactis. Antonie Van Leeuwenhoek 82, 113–122 (2002).
80.Smyth, G. Bioinformatics and computational biology solutions using R and bioconductor (Gentleman, R., Carey, V. J., Huber, W., Irizarry, R. A. & Dudoit, S.)
397–420 (2005).
81.Steen, A., Wiederhold, E., Gandhi, T., Breitling, R. & Slotboom, D. J. Physiological adaptation of the bacterium Lactococcus lactis in response to the production
100
of human CFTR. Molec. Cell. Proteomics 10, (2011).
82.Muskiet, F. A., Van Doormaal, J. J., Martini, I. A., Wolthers, B. G. & Van der
Slik, W. Capillary gas chromatographic profiling of total long-chain fatty acids
and cholesterol in biological materials. J. Chromatogr. 278, 231–244 (1983).
83.Boele, J., Olivier, B. G. & Teusink, B. FAME, the flux analysis and modeling
environment. BMC Syst. Biol. 6, 8 (2012).
Supplementary material Figures
Figure S1. Quality control for proteomics data: For all the chemostats at different growth the
cells were lysed and the soluble and membrane proteome was isolated as described in Materials
and Methods. For each growth rate the relative quantification for proteins was performed on
soluble and membrane fractions in independent experiments. This figure shows that the 2log
fold change regulation ratio of a protein in soluble or membrane fraction is correlated. Thus a
change in the regulation ratio for a protein is equally well represented in both the fractions. The
number of spectra obtained for a particular protein indicates its enrichment in either soluble
or membrane fraction. Proteomics on exclusive membrane fractions allows investigating the
minor changes in the abundance of membrane proteins with high confidence.
101
Figure S2. All amino acid consumption rates (q-rates) at various dilution rates.
Figure S3. rProtein ratios relative to the lowest growth rate 0.15 h-1, for L. lactis, separated in
early assembly, secondary assembly and late assembly 58.
102
Supplementary material Tables
Table S1. Transcriptome analysis of L. lactis MG1363 at varying growth rates. Tab1
(all) contains all genes sorted at accession number. Tab2 (regulon) contains the genes
sorted based on their regulon and Tab3 (COG-cat) based on their COG-category. The
table can be found on http://www.molgenrug.nl/supplementary_data/thesisEckhardt/
Chapter3_TableS1.xlsx
Table S2. Proteome analysis of L. lactis MG1363 at varying growth rates. All measured proteins are sorted on their accession number. Indicated are the number of
spectral reads and degrees of freedoms from the membrane (mem) and soluble (sol)
fractions of every measured protein. The table can be found on http://www.molgenrug.nl/supplementary_data/thesisEckhardt/Chapter3_TableS2.xlsx
Table S3. List of (iso)enzymes of L. lactis MG1363 obtained from the proteome
analysis. The table can be found on http://www.molgenrug.nl/supplementary_data/
thesisEckhardt/Chapter3_TableS3.xlsx
Table S4. Correlations between transcripts and proteins of KEGG-categories (Tab1)
and regulons (Tab2). The table can be found on http://www.molgenrug.nl/supplementary_data/thesisEckhardt/Chapter3_TableS4.xlsx
103
104
Chapter 4
Transcriptional regulation of fatty acid
biosynthesis in Lactococcus lactis
Thomas H. Eckhardt, Dorota Skotnicka, Jan Kok and Oscar P. Kuipers
105
Abstract
Here we study the influence of the putative fatty acid biosynthesis (FAB) regulator
FabT (originally called RmaG (Llmg1788)) on gene transcription in Lactococcus lactis MG1363. A strain with a knockout mutation of the putative regulator was constructed, and its transcriptome was compared to that of the wildtype strain. Almost
all FAB genes were significantly upregulated in the knockout. Using EMSAs and
DNAseI footprinting the binding motif of the regulator and the binding locations in
the genome were characterized. Fatty acid composition analysis revealed that a strain
lacking FabT contained significantly more saturated acyl chains in its phospholipids.
This observation demonstrates that the vital pathway of FAB in L. lactis is regulated
by the repressor FabT.
Introduction
Membrane phospholipids are essential for life; they contain a hydrophilic head group
and two hydrophobic tails esterified to a glycerol moiety. The hydrophobic tail is usually composed of a stretch of hydrocarbons denoted as acyl chains. Biosynthesis of
saturated fatty acids (SFA) in bacteria is performed by multiple conserved enzymes
in a multistep process. Based on the sequence similarity of genes in the fab regulon,
we describe here the most likely fatty acid biosynthesis (FAB) route in Lactococcus
lactis. The acetyl coenzyme A carboxylase (ACC) complex, consisting of AccABCD,
catalyzes an acetyl-CoA into malonyl-CoA conversion 1. The CoA is replaced by an
acyl-carrier protein (ACP) by FabD, a malonyl-CoA:ACP transacylase 2. Fatty acid
elongation rounds are initiated by FabH (β -ketoacyl-ACP synthase III) by condensing
an acetyl-CoA with malonyl-ACP 3. The first reductive step in the FAB elongation is
performed by β -ketoacyl-ACP reductase (FabG) producing a β -ketoacyl-ACP and
NADP- 4. This β-ketoacyl-ACP is dehydrated by FabZ (β-hydroxyacyl-ACP dehydratase), resulting in a trans-2-enoyl-ACP 5,6. The final step in lactococcal FAB elongation is a second reduction step executed by trans-2-enoyl-ACP reductase I FabI, giving an acyl-ACP 7,8. Further elongation rounds start by the condensation enzyme FabF
β -ketoacyl-ACP synthase II through the addition of an acyl group from malonyl-ACP
9,10
. The resulting β -ketoacyl-ACP can continue through the elongation cycle by re106
duction by FabG again. For L. lactis, no enzymes are identified that can process the
acyl-ACP into phospholipids. The only protein is PlsX, annotated in L. lactis as a putative acyltransferase. Investigations on the PlsX from Bacillus subtilis showed that the
enzyme is able to form acylphosphate from acyl-ACP 11,12.
FAB has been shown to be a coordinated process in the model organisms Escherichia
coli and B. subtilis, where FAB is under tight control of the transcriptional regulators FadR/FabR and FapR, respectively 13. The bifunctional E. coli FadR activates the
essential gene fabA 14. When sufficient amounts of long chain acyl-CoA have been
produced, some of these molecules bind to FadR, which results in de-repression of
the fatty acid degradation pathway (β-oxidation) specified by the fad-operon 15. FabR
is the transcriptional repressor of fabA and fabB, two genes that are required for the
synthesis of unsaturated fatty acids (UFA). FabR represses FAB and the first steps of
phospholipid synthesis in E. coli 16. The B. subtilis FapR regulator functions as a malonyl-CoA sensor, whereby complex-formation of FapR and its co-repressor malonylCoA results in the repression of the transcription of the FAB genes 17.
Regulation of FAB in Lactococcus lactis is poorly understood; it is important to understand this regulation, though, in view of the possible involvement of FAB in flavor formation pathways in this industrially relevant microorganism. Because of the
synteny of their fab-gene clusters, L. lactis, Enterococcus faecalis and S. pneumoniae
were grouped together 18. The regulator of FAB in S. pneumoniae and E. faecalis is
FabT. In S. pneumoniae this protein binds to the upstream region of fabK and to that of
fabT itself and is co-repressed by the acyl carrier protein (ACP) coupled to C16:0 and
C18:0 acyl chains 19. There seem to be only two binding sites for the regulator FabT
in the fab gene cluster. The L. lactis operon carries more and larger intergenic spaces
where a regulator could possibly bind. A similar situation occurs in E. faecalis, where
FabT is capable of binding to three regions upstream of the genes fabT, fabK and fabI/
fabF1. In L. lactis two fab genes possess a paralog outside the operon i.e. fabZ (outside: llmg0538/fabZ1; inside: llmg1781/fabZ2) and fabG (outside: llmg1760/fabG2;
inside: llmg1784/fabG1). Special attention should be given to fabZ1 since it shares an
upstream region with enoyl-acp reductase gene fabI, an essential part of FAB. In S.
pneumoniae and E. faecalis, fabI is present in the fab cluster where it is named fabK.
E. faecalis and L. lactis both contain fab genes on two locations on the chromosome.
The remainder of the cluster, the genes for the acetyl-CoA carboxylases accABCD, the
acyl carrier protein gene acp and the FAB genes fabDFGHZ are present in L. lactis,
107
E. faecalis and S. pneumoniae with similar synteny (Fig. 1A). In addition, all genes of
the fab cluster of L. lactis MG1363 share around 70% sequence similarity with those
of S. pneumoniae D39. Although the L. lactis, E. faecalis and S. pneumonia fab gene
clusters have a similar genetic organization 18, several differences remain. In this study
we establish the regulation of FA biosynthesis in L. lactis, and compare it to that of
E. faecalis and S. pneumoniae. We show that the regulator of FA biosynthesis in L.
lactis is a repressor that we renamed from RmaG in FabT. Moreover, we determined
its regulon and its DNA binding motif by EMSAs and DNAseI footprinting.
Material and Methods
Bacterial strains, plasmids and growth conditions
The strains and plasmids used in this study are listed in Table 1. E. coli was grown
aerobically at 37°C in TY medium (1% Bacto-Tryptone, 0.5% Bacto-yeast extract and 1% NaCl). L. lactis strains were grown as standing cultures in M17 medium (Difco Laboratories, Detroit, MA) with 0.5% (w/v) glucose (GM17) at 30°C.
Solid media contained 1.5% agar. Chloramphenicol (5 µg/ml) and erythromycin (120 µg/ml for E. coli and 2.5 µg/ml for L. lactis) was added when required.
General DNA techniques
General molecular biology techniques were performed essentially as described by
Sambrook 20. Plasmid DNA was isolated using a High Pure Plasmid Isolation Kit
and protocol (Roche Applied Science, Indianapolis, IN). Chromosomal DNA from L.
lactis was isolated according to the method described by Johansen and Kibenich 21.
Polymerase chain reactions (PCR) for (sub-) cloning were performed with Phusion
(Finnzymes, Espoo, Finland), colony PCR with the Taq Polymerase from Fermentas
(ThermoFisher Scientific Inc, Waltham, MA). Primers are listed in Table S1; they
were purchased from Biolegio BV (Nijmegen, the Netherlands). PCR products were
purified with a High Pure PCR Product Purification Kit (Roche Applied Science) according to the protocol of the supplier. DNA electrophoresis was performed in 1x TBE
buffer (89 mM Tris-HCl, 89 mM boric acid, 2 mM EDTA, pH 8.3) in 1% agarose
gels with 2 µg/ml ethidium bromide. Electrotransformation was performed using a
Bio-Rad Gene Pulser (Bio-Rad Laboratories, Richmond, CA). All DNA modification
108
enzymes were purchased from Fermentas and used according to the manufacturer’s
instructions. Sequencing reactions were done at ServiceXS (Leiden, the Netherlands).
Construction of an L. lactis fabT deletion mutant
Upstream and downstream regions of fabT, PCR-amplified using primer pairs
Pr3/Pr4 and Pr1/Pr2 (Table S1), respectively, were cloned in the integration
vector pCS1966 22 using the enzymes XbaI and XhoI. The resulting plasmid,
pCS1966ΔfabT, was obtained in E. coli and introduced in L. lactis to allow integration via single cross-over homologous recombination. An L. lactis integrant carrying the pCS1966 construct was selected on GM17 plates with chloramphenicol.
Screening for plasmid excision was done on plates containing 5’-fluoroorotic acid,
selecting against the oroP gene on pCS1966. A mutant carrying a clean knockout
of fabT was obtained and confirmed using PCR and nucleotide sequence analysis.
L. lactis FabT overexpression and protein purification
The fabT gene was amplified by PCR using L. lactis chromosomal DNA as the template and primers Pr6 and Pr61 (Table S1). The PCR product consisting of the fabT
gene, extended at the 5-end with the codons for the Strep-tag 23 was purified, digested
with NcoI and XbaI and ligated into pNZ8048 24 cut with the same enzymes. The
resulting plasmid (pNZstrepfabT), in which the Strep-fabT construct was under the
control of the nisin-inducible promoter PnisA, was obtained in E. coli and subsequently
introduced in L. lactis NZ9000 25. The nucleotide sequence was confirmed by sequencing using the primers pNZSeq and nis_fw_aldert (Table S1). Strep-tagged FabT was
overexpressed using the nisin-inducible system (NICE) 24. As a source of nisin, filtersterilized culture supernatant of the nisin-secreting strain L. lactis NZ9700 was used.
An overnight culture of L. lactis NZ9000 (pNZstrepfabT) was diluted 100-fold in
1 L of fresh GM17 medium with 5 μg/ml chloramphenicol and incubated at 30°C.
Nisin-containing supernatant (1:500) was added when the OD600 of the culture had
reached 0.5. After 2 h of further incubation cells were pelleted (6,000 g for 10 min),
resuspended in 10 ml lysis buffer (100 mM Tris-HCl, 100 mM NaCl, 1 mM EDTA,
5 mg/ml lysozyme, Roche complete mini protease inhibitor, pH 8.0), incubated for
40 min at 30°C and centrifuged at 9,000 g (20 min, 4°C). Subsequently, 0.1 g DNAseI powder was added, and the lysozyme treated cells were broken using a French
press (Aminco, Silver Springs, MD). Strep-tagged FabT was purified to homogene109
ity on a Streptactin Sepharose column according to the manufacturer’s instructions
(IBA-GmbH, Göttingen, Germany). Samples from each step in the purification were
analyzed by 12% sodium dodecylsulphate-polyacrylamidegel electrophoresis (SDSPAGE)
26
and Western hybridization using anti-Strep-tag antibodies (IBA-GmbH).
The concentration of purified protein was determined via spectroscopy (Nanodrop,
ThermoFisher Scientific Inc). Protein (100 μM) was kept at -80ºC in 10% glycerol, 100 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 2.5 mM desthiobiotin, pH 8.0.
DNA microarray analysis
L. lactis cells were grown to the mid-exponential growth phase (OD600 = 0.8) in
GM17 containing 5 µg/ml chloramphenicol for plasmid containing strains. For induction of cells containing pNZstrepfabT, 1000-fold dilutions of supernatant from
the nisin-producing strain L. lactis NZ9700 were added at OD600 ≈ 0.5. Cells were
harvested by centrifugation (6,000 g for 10 min); pellets were immediately frozen in liquid nitrogen and resuspended in 1.6 ml milliQ that was treated with diethylpyrocarbonate (DEPC) (Sigma-Aldrich, St Louis, MO) and divided in 4 portions prior to storage at -80°C. For RNA isolation the frozen cells were thawed on
ice. Subsequent cell disruption, RNA purification, reverse transcription and Cy3/
Cy5 labeling were done as described previously 27. Labeled cDNAs were hybridized to full-genome DNA microarray slides of L. lactis MG1363 28. All reagents and
glassware for RNA work were treated with DEPC. RNA, cDNA quantity and quality and the incorporation of the cyanine-labels were examined by NanoDrop (ThermoFisher Scientific Inc.) at 260 nm for RNA and cDNA, 550 nm for Cy3, and 650
nm for Cy5. Two biological replicates were used in combination with a dye-swap.
DNA microarray slide images were analyzed using ArrayPro 4.5 (Media Cybernetics Inc., Silver Spring, MD). Filtering of bad and low-intensity spots and signals, data parsing, automated grid-based Lowess normalization, scaling, data
visualization and outlier detection were performed using the Microprep software 29. Differential expression tests were done on expression ratios with Cyber-T on a local server implementation of a variant of the t-test 30. Fold-changes are considered to be significant when the p value ≤ 0.05. and the ratio ≥ |1.5|.
110
Fatty acid composition analysis
Samples from L. lactis cultures, either induced with nisin to modulate expression of Strep-FabT or not, were pelleted and washed three times in phosphate-buffered saline (PBS). All samples were transmethylated and analyzed on a gas chromatograph for acyl chain composition according to the
methods described by Muskiet
31
. Data presented are from biological duplicates.
SDS-PAGE and Western hybridization
Protein samples were mixed (1:1) with sample buffer (120 mM Tris pH 6.8, 50%
glycerol, 100 mM DTT, 2% (w/v) SDS and 0.02% (w/v) bromophenol blue), vortexed,
boiled at 100ºC for 5 min and separated on SDS 12%-PAA gels. PageRuler Prestained
Protein Ladder (ThermoFisher Scientific Inc.) was used as a marker. Gels were stained
with Coomassie Brilliant Blue (0.25% CoomassieBrilliantBlue R-250 dissolved in
25% isopropanol with 10% acetic acid) and destained in boiling demineralized water.
Western blot analysis was performed using a SNAP (Millipore Corp., Billerica, MA)
system as follows: the 12%-PAA gel was equilibrated in transfer buffer (25 mM Tris,
192 mM glycine, 10% methanol). Proteins were transferred to a PVDF (Roche Applied
Science) blotting membrane for 30 min at 20 V. The blot holders containing the blots
were placed in the SNAP system, after which blocking buffer in PBST (58 mM Na2HPO4, 17 mM NaH2PO4, 68 mM NaCl, pH 7.3, 0.1% v/v Tween-20) with 0.5% skim
milk was added and the vacuum was applied. Antibodies (anti Strep-tag conjugated to
peroxidase (IBA-GmbH)), diluted 1,000-fold in blocking buffer, were added to the blot
holders; incubation was for 10 min at room temperature. The vacuum was removed and
the blots were washed three times with PBST prior to visualization of immunoreactive
proteins using the ECL detection kit and protocol (GE Healthcare, Buckingham, UK).
Electrophoretic mobility shift assays (EMSAs)
DNA probes comprising the upstream sequences of all genes in the fab operon were
obtained using PCR. Probes of upstream regions of fabI/ fabZ1 from L. lactis and
of fabT from S. pneumoniae were also amplified. Purified PCR products (5 μl) were
end-labeled with polynucleotide kinase T4-PNK (ThermoFisher Scientific Inc.) for
2 h at 37°C by using 2 μl of 30 μCi [γ-33P]ATP (PerkinElmer, Waltham, MA), 2 μl
of One-For-All buffer (Roche Applied Science) and 1 μl of 10 units/μl T4-PNK in
a total volume of 20 μl. End-labelled DNA was purified with PCR Purification Kit
111
(Roche Applied Science) after which the counts per minute was determined using 1
μl of each DNA sample in 4 ml of scintillation liquid (Ultima Gold, PerkinElmer).
DNA-protein binding studies were carried out in 20 μl reaction volumes containing 40 mM Tris-HCl (pH 8.0), 17.4% (v/v) glycerol, 2 mM EDTA (pH 8.0), 10
mM MgCl2, 200 mM KCl, 1 mM dithiothreitol, labeled DNA fragment (5,000
cpm), and various amounts of purified Strep-tagged FabT protein (concentrations
0.05 μM - 1 μM). Bovine serum albumin (2 μg) and 0.05 mg/ml poly(dI-dC) were
added to the reaction mixtures in order to reduce nonspecific interactions. After incubation for 20 min at 30°C, 12 μl of the samples were loaded on a 4%-PAA gel.
Electrophoresis was performed in electrophoresis buffer (44.5 mM Tris, 44.5 mM
boric acid, 1 mM EDTA, pH 8.0) at 100 V for 90 min. Subsequently, the gel was
dried onto Whatman 3MM filter paper, and radiolabelled bands were visualized
by autoradiography, using a Cyclone phosphorimager (Packard, Meriden, CT).
DNAseI footprinting assay
DNAseI footprinting was performed using a protocol that was largely based on the
Sure Track Footprinting Kit (GE Healthcare), as described previously 32. The PfabT
region of L. lactis was amplified by PCR using a forward primer (Table S1) endlabelled with 30 μCi [γ-33P]ATP for 2 h at 37°C using T4-PNK. Binding reactions
using purified Strep-FabT were identical to those used in the EMSAs, in a total volume of 40 μl and in the presence of approximately 150,000 cpm of DNA probe.
Membrane permeabilization assay
Cells were grown at 30°C to an optical density of 1.0±0.1, after which 1 ml of cells
were pelleted, resuspended in 100 µl PBS and incubated with 1 µl of 1 µg/ml ethidium
bromide for 15 min at 30°C. Cells were pelleted, and washed two times with PBS. The
cells were resuspended in 100 µl PBS and subjected for FACS analysis (BD FACSCanto, BD Biosciences, Palo Alto, CA). All cells were grown as biological replicates
in triplicate.
112
Table 1. Bacterial strains and plasmids used in this study. Abbreviations: pepN; aminopeptidase;
nisRK; two component system which senses extracellular nisin to activate gene transcription
by PnisA; recA; recombinational repair enzyme; relA; ppGpp synthase; endA; endonuclease;
hsdR17; methylation endonuclease; emR; erythromycin resistance; oroP; orotate transporter;
cmR; chloramphenicol resistance; PnisA; nisin inducible promoter.
Strains
Characteristics
Reference
L. lactis NZ9000
MG1363 derivative; pepN::nisRK;
plasmid-free strain; NICE gene
expression host
24
L. lactis NZ9000 ∆fabT
NZ9000 derivative; chromosomal
deletion of fabT
This study
L. lactis NZ9700
Nisin producer strain
33
E. coli DH5α
Strain carrying deletions of recA, relA,
endA, hsdR17 host of recombinant
plasmids for L. lactis
34
pCS1966
EmR; oroP; non-replicating, integration
vector in L. lactis
35
pCS1966 ∆fabT
pCS1966 derivative, carrying flanking
regions of fabT.
This study
pNZ8048
CmR; nisin-inducible gene expression
vector carrying the PnisA.
pNZstrepfabT
pNZ8048 derivative, carrying LlfabT
with the coding region for Strep-tag at its
end under the control of PnisA.
Plasmids
24
This study
113
A
B
C
Figure 1. (A) Comparison of the fab clusters of E. faecalis, L. lactis and S. pneumoniae.
Arrows signify the fab genes that correspond to the open reading frames in the three organisms.
Upstream regions that can bind FabT are indicated with an asterisk. (B) Schematic overview of
the conserved fatty acid biosynthesis pathway, based on the annotated genes in L. lactis and the
model of B. subtilis 38. AccABCD initiate the formation of malonyl-ACP. Malonyl-CoA:ACP
transacylase FabD substitutes the CoA for an acyl carrier protein (ACP). Malonyl-ACP is
con-densed by FabH to acetyl-CoA after which the ketoacyl-ACP is formed. β-ketoacyl-ACP
reductase FabG starts with a reduction and is followed by a dehydration executed by FabZ. A
second dehydration by FabI produces the acyl-ACP. Once elongated sufficiently, acyl-ACPs
are processed by several unknown enzymes, the only known component being acyltransferase
PlsX and inserted into the membrane. In order to continue the elongation reaction acyl-ACPs
are dehydrated by FabF. (C) Adding or removing LlFabT has a small effect on the transcription
of the FAB genes. The first microarray analysis compares the deletion of regulator FabT against
the wildtype. In the second analysis, induction of fabT on a pNZ8048 plasmid was compared
against the induction of an empty pNZ8048. The bottom analysis shows the complementation
of FabT, meaning the same induction of fabT as in the second microarray, but in a fabT
deletion background. Black colored genes represent upregulated genes, whereas striped genes
indicate a downregulation under these conditions. Upstream regions that can bind FabT are
indicated with an asterisk.
114
Results
Two regions in the genome of Lactococcus lactis carry genes putatively involved
in fatty acid biosynthesis (FAB; Fig. 1A). The products of these genes show high
similarities with the enzymes involved in FAB in other prokaryotes, such as Streptococcus pneumoniae and Enterococcus faecalis. Not only are the proteins conserved,
the gene synteny is largely preserved (Fig. 1A). A large gene cluster extending from
llmg1777 to llmg1787 contains most of the fab genes, while the putative fabI gene and
a gene (llmg0538/fabZ1) of which the product has homology with FabZ2 (llmg1781)
are located in a head-to-head orientation elsewhere on the lactococcal chromosome.
The gene llmg1788 (rmaG) is annotated as encoding a regulator protein of the MarR
family 36. This family consists of a group of dimer-forming proteins, in which both
subunits possess a winged-helix DNA binding motif 37. The rmaG gene is located upstream of the large fab cluster in the L. lactis genome (Fig. 1A); its product shares 59%
similarity with SpFabT, the transcriptional regulator of FAB in S. pneumoniae. Thus,
Llmg1788 is hypothesized to perform an analogous function in the control of the L.
lactis fab genes. Based on this and on the results presented in this work, we propose to
rename Llmg1788 (RmaG) into L. lactis FabT.
FabT affects fab gene expression in L. lactis
The elimination of the fabT gene was achieved by complete removal of the open
reading frame, so that no partial fabT transcripts are produced in the fabT deletion strain. Removal of the fabT gene caused a slight growth delay of L. lactis on
GM17 (data not shown), suggesting that FabT is not essential for L. lactis, yet serves
a specific function under the conditions tested. In order to determine the effects of
the fabT mutation on gene expression in L. lactis, a transcriptome analysis was performed on exponential-phase cells growing in rich GM17 medium. Only a modest
decrease of the fabT transcript was observed in the fabT mutant relative to its parent
strain, suggesting that the quantity of fabT mRNA molecules is close to the background signals under these conditions in the wildtype strain. The expression of some
of the genes downstream of fabT as well as that of fabI and llmg0538/fabZ1 is upregulated several-fold in the knockout strain (2.1 to 6.5-fold) (Fig. 1C). The transcript abundances of acpA, fabHFZ2 and accACD were not significantly altered.3
A complementation strain was made which carries a copy of a gene expressing N115
terminally Strep-tagged FabT, strep-fabT, on a plasmid downstream of the nisininducible promoter PnisA. Both the wildtype strain and the fabT mutant strain were
complemented and compared. A 136-fold upregulation of fabT transcripts was seen
after nisin induction of strep-fabT (Fig. 1C). First, this shows that Strep-FabT is able
to produce transcripts of fabT after nisin induction. Under these circumstances a small
but significant decrease in the expression of most of the fab genes was observed. These
data support the hypothesis that FabT functions as a repressor of fab genes in L. lactis.
With the exception of fabH, all genes responsible for FAB are under the regulation of
FabT. In addition to the fab cluster, transcripts of proteins with a putative or unknown
function were affected. The gene llmg0538/fabZ1 is located next to and in opposite
orientation to fabI; both genes thus share the same upstream region. The former is
significantly downregulated (5.3-fold) when strep-fabT is overexpressed, while it is
6.5-fold overexpressed in the fabT knockout background) (Table S2).
FabT-DNA interactions
Electrophoretic mobility shift assays (EMSAs) were performed to examine whether
fab genes are under direct or indirect control of LlFabT. Strep-FabT was overexpressed
in L. lactis via the nisin-inducible system (NICE) 24 and purified. LlFabT runs at the
position of a protein of 18 kDa, just above the band of lysozyme (14 kDa) that was
used to break open the cells (Fig. S1). A band of a protein of twice the size of StrepFabT (36 kDa) most probably represents a dimer of the regulator.
Purified Strep-FabT was incubated with the upstream DNA regions of all fab genes
and subjected to EMSA (Fig. 2). The results show that Strep-FabT formed proteinDNA complexes with the upstream fragments of fabI/fabZ1, fabD, accC and fabT. In
a number of cases, two shifted bands can be observed. The fact that Strep-FabT also
forms a complex with a PCR fragment encompassing the LlfabT promoter suggests
that LlFabT regulates its own transcription. As the upstream region of LlfabT used
was quite large (460 bp), it was further investigated (Fig. 2). No binding was observed
to the 198-bp proximal fragment immediately upstream of the start of LlfabT. In contrast, significant in vitro binding was detected to the more distal PCR fragment, starting 167 bp upstream of the start codon in LlfabT (Fig. 2). As LlFabT and SpFabT are
highly similar, binding of the former to the promoter region of S. pneumoniae fabT
was examined. A shift was also observed when the regulator of L. lactis was mixed
with the upstream fragment of pneumococcal fabT 19, containing the SpFabT binding
116
Figure 2. In vitro binding studies of LlFabT. Electrophoretic mobility shifts (EMSA) show
the interaction between Strep-FabT and upstream regions of genes involved in FAB. The
upstream region of fabT was split into two fragments, one proximal fragment (-53 bp to 189
bp upstream of the ATG), and one distal fragment (152 bp to 407 bp upstream of the ATG).
SpfabT fragment is the upstream region of S. pneumoniae fabT, all others are lactococcal
fragments. EMSA reactions were carried out with increasing concentrations of Strep-FabT (0,
0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 1.0 μM) from left to right.
site (Fig. 2). Binding was not seen with upstream parts of other L. lactis fab genes and
with those of two genes of B. subtilis that served as negative controls (nasD, nitrate
reductase; htrA, membrane associated serine protease (data not shown)). Strep-FabT
binds to all the probes in a concentration-dependent manner and with different affinities for the various fragments. The lowest concentration of protein tested (0.05 µM)
was enough to cause a shift of the distal upstream region of LlfabT, while at least 0.8
µM of the regulator was required to cause a shift of the upstream region of fabD.
117
A
B
C
Figure 3. Purified Strep-FabT binds to its own promoter in vitro. (A.) DNaseI footprint
analysis of Strep-FabT binding to the upstream region of LlfabT. The first lane is the MaxamGilbert GA marker. The footprints are obtained by the addition of 0, 0.1 or 0.5 mM of protein
Strep-FabT as indicated. The black bars indicate the protected regions. (B.) The sequence of
the promoter region of LlfabT with the protected area is underlined and the hypersensitive
basepairs are shown in bold. The transcription start site (TSS) of LlfabT is indicated with
double underlining. The TSS of fabT is at a position where the Strep-FabT protein is able
to bind. (C.) The weight matrix and the alignment of the four binding sites that revealed the
general recognition site for FabT.
The DNA binding motif of lactococcal FabT
Using bioinformatic tools like MEME 39 and MDscan 40 no clear motif could be identified in the upstream regions bound by Strep-FabT. To pinpoint the L. lactis FabT
DNA binding site, DNaseI footprinting was performed on the distal fragment of the
upstream region of fabT. Fig. 3 shows that incubation of this 33P-labeled fragment with
purified Strep-FabT resulted in protection of an AT-rich region of 52 bp approximately
200 bp upstream of the fabT start-codon. Adding this information to the bioinformatics search allowed the identification of the presumed lactococcal FabT operator site
(TTTGAWAWAGAAA) (Fig. 3C). This motif occurs in the upstream regions of fabD,
118
fabT and fabZ1/fabI. The fabI and fabZ1 genes are not part of the fab gene cluster (Fig.
1). Nevertheless, Strep-FabT is able to form complexes with the fabI/fabZ1 intergenic
region (Fig. 2). The transcription start sites (TSS) of both fabI and fabZ1 were determined via 5’ RACE (Fig. 4B). The DNA binding motif overlaps the TSSs of both
fabZ1 and fabI; thus, FabT could block the transcription of both genes simultaneously.
Effect of FabT on L. lactis membrane phospholipid composition
L. lactis FabT controls the transcription of most fab genes, the encoded proteins of
which produce acyl chains that are incorporated into phospholipids. Consequently,
the effect of fabT deletion or overexpression on the fatty acid composition of the cytoplasmic membrane phospholipids was examined by gas chromatography. Overproduction of Strep-FabT in both the wild type strain of L. lactis and in the L. lactis fabT
mutant does not affect the length of the fatty acids or their degree of saturation (Fig.
5). Removal of FabT results in a shift from unsaturated (18:1n7) to saturated (16:0 and
18:0) fatty acids. Approximately 50% of the acyl chains in the fabT mutant are 16:0
molecules. The relative amount of 18:0 molecules is at least three times higher in the
A
B
Figure 4. Transcription start sites of fabT (A) and fabZ1 and fabI (B). With 5’RACE the
transcription start sites were determined (underlined). Nucleic acid bases in blue are the -10
and -35 sites. The PCR fragments used for EMSA are named FabT proximal fragment (purple)
and FabT distal fragment (green). The binding motif of Strep-FabT is indicated with a box.
Upon FabT binding, transcription of fabI and fabZ1 could be blocked simultaneously.
119
fabT knockout mutant than in the parent strain. The main decrease is observed in oleic
acid (18:1n9), an unsaturated fatty acid. This shift in UFA/SFA ratio had an effect on
the membrane permeability. The fabT mutant, possessing higher relative amounts of
saturated fatty acids, had a higher membrane permeability for the fluorescent DNAbinding probe ethidium bromide (Fig. 5B).
A
B
Figure 5. (A) Fatty acid composition of L. lactis strains. Deleting FabT (∆fabT) creates a
shift from unsaturated (18:1n7) to saturated (16:0 and 18:0) acyl chains. Induction of fabT
(pNZfabT) does not affect fatty acid composition. Cyc19:7 is a cyclic fatty acid, and saturated
acyl chains are shown in blue, while unsaturated acyl chains are shown in red. Shown are
the averages of two biological experiments. (B) Effect of the fabT mutant on membrane
permeability to ethidium bromide. The control (red line) is the NZ9000 strain without
exposure to fluorescent ethidium bromide, while NZ9000 (green line) and NZ9000 ∆fabT
(black line) both were incubated with 1% v/v of 1 µg/ml ethidium bromide for 15 minutes,
and subsequently washed to remove extracellular ethidium bromide. The graphs shown are
representative of two independent biological experiments performed with triplicates.
120
Discussion
Elucidating the regulation of the essential pathway of fatty acid biosynthesis in the
industrially relevant bacterium L. lactis is of interest from both a fundamental and
application point of view. Here, we characterized the regulation of FAB in L. lactis.
The gene for the MarR-type regulator RmaG, here renamed FabT, is located in a gene
cluster containing most of the L. lactis fab genes. The genetic make-up of the gene
cluster, the synteny and orientation of the fab genes are very similar to those of the fab
clusters of S. pneumoniae and E. faecalis 41,42. A transcriptome analysis of the effect
of deleting the fabT gene from the chromosome of L. lactis clearly revealed that FabT
is involved in the regulation of the fab gene cluster and in fabZ1/fabI expression. No
other genes were significantly regulated by FabT, indicating that it is a dedicated local
regulator of FAB. Upon overexpressing strep-fabT only fabF and fabD respond with
a significant downregulation of -1.7 and -1.9 fold, respectively. The repressive capacities of FabT are more apparent when its gene was removed from the genome. When
the fabT mutant was complemented by overexpression of Strep-FabT, almost the complete fab cluster responded (Fig. 1C). The only gene in the fab operon that does not
seem to respond to the deletion or overexpression of FabT is fabH. It was previously
reported that an L. lactis β-ketoacyl-ACP synthase III (FabH) mutant strain survives
43
. In that case, β-ketoacyl-ACP synthase I (FabF) took over the condensation function
of the deleted fabH. Similarly, S. pneumoniae fabM transcription does not respond to
the deletion of the streptococcal fab regulator FabT 41. Isomerase FabM was described
not to be essential in this organism 5. In Streptococcus mutans FabM is essential at low
pH 44. FabH of L. lactis seems to be non-essential as its transcription does not respond
to deletion of fabT nor to the overexpression of Strep-FabT. Altogether, it seems that
non-crucial fab genes do not directly respond to the presence or absence of the L. lactis
FAB regulator.
Under the conditions employed, exponential growth in a rich medium, it seems that
the repressive effect of L. lactis FabT on fab gene expression is mild, only delicately
repressing the genes involved in FAB appears to be sufficient to control regulation.
A mild repression of the FAB enzyme genes that can easily be relieved could therefore be sufficient for cell maintenance. The DNA microarrays report that significant
changes occur in transcription of the fab cluster. FabT does not regulate genes for
121
membrane proteins like PlsX and PlsY, which are involved in glycerolipid formation
in B. subtilis 11, or cyclic fatty acid synthase Cfa 45. Thus, L.lactis FabT seems to be a
dedicated repressor: it only regulates fatty acid biosynthesis.
In E. coli the regulator FadR has a dual function of stimulating biosynthesis and limiting the degradation pathway of FA 46. Such a dual function of FabT of L. lactis or S.
pneumoniae is difficult to examine at this point, as in these bacteria fatty acid degradation has not yet been characterized, nor have the genes been identified. The fact
that the regulon of L. lactis FabT seems to be strictly confined to the FA biosynthesis
genes would suggest that in this organism FA degradation is controlled by a different
mechanism. Investigating a range of different candidates for binding to S. pneumoniae
FabT shows that acyl-ACPs of appropriate length (C16:0/C18:0) are the best ligands
19
. It seems reasonable that for Streptococcaceae transcription of the FAB genes is low
but constitutive and is diminished even more when more unsaturated acyl-chains are
present than needed.
The proposed binding motif of S. pneumoniae FabT is GTTTTGATTGTAAAAGT 41,
while the consensus binding motif of FabT in E. faecalis was proposed as AGTTTGATAATCAAATT 42. Using these motifs did not automatically reveal the binding site
for L. lactis FabT. Ultimately, the consensus binding motif of L. lactis FabT was determined to be TTTGAWAWAGAAA. The lactococcal FabT binding motif displays a
lot of variation, with a bias towards adenines in the 3’-end of the motif, and a distinct
TTTGA in the 5’-end, as is also the case in the binding sites of the FabT regulators of
the other two bacterial species. The similarity of the FabT binding sites is high enough
for L. lactis FabT to recognize and bind to the S. pneumoniae FabT binding site (Fig.
2). It is therefore highly likely that the presence of saturated fatty acids as corepressors will affect he binding affinity, as has been shown for S. pneumoniae 19,41, but not
the nature of the recognition sequence. While a lot of MarR regulators prefer to bind
inverted repeats, not all members of the helix-turn-helix family do so 47,48. Also here,
a clear palindrome or repeat cannot be seen in the FabT binding motifs in L. lactis,
S. pneumoniae and E. faecalis, except for the outermost triple Ts and triple As. Apparently, the minimal common motif of FabT in Streptococcaceae does not require a
palindromic motif. Whereas FabT of S. pneumoniae only binds to the promoter region
of its own gene and to that of fabK, L. lactis FabT binds to the upstream regions of its
own gene fabT, and further downstream in fab, to the upstream regions of fabD, accC
and fabI (Fig. 1C).
122
In order to adapt to changing environments, membrane fluidity and permeability in
bacteria are changed by altering acyl-chain composition and head-group modification.
It was suggested that acyl-chain composition is mainly modified by de novo synthesis
49
. L. lactis possesses all enzymes necessary to produce saturated acyl chains. How-
ever, for the production of unsaturated acyl chains no equivalent of a desaturase like B.
subtilis DesK 50 is annotated in the genome of L. lactis. E. coli uses dehydratase FabA
and the condensation enzyme FabB 10 to produce cis double bonds, while two dehydratase FabZ variants exist in E. faecalis, of which one (FabN) could function as an
isomerase to create unsaturated fatty acids 42. The S. pneumoniae cis-trans isomerase
FabM functions as an interconverter in competition with enoyl-ACP reductase FabK:
the 10:1-trans-2 intermediate is either processed into saturated fatty acids or altered
via FabM into an 10:1-cis-3 intermediate that is further converted into other unsaturated fatty acids 41. If L. lactis would produce unsaturated fatty acids according to the
models of FAB of E. coli or S. pneumoniae the corresponding enzymes still need to be
identified. It seems more likely that L. lactis uses a mechanism similar to that operative in E. faecalis as its genome contains two genes, fabZ1 and fabZ2, of which the
encoded enzymes are similar (75% and 69%) to FabN and FabZ of E. faecalis V583,
respectively. Also, L. lactis fabZ1 is in close proximity to fabI, as is the case for E.
faecalis fabN/fabI (Fig.1A). Even though fabZ1/fabI are on a different location in the
chromosome, the L. lactis FabT binding motif is present in their intergenic region. The
5’RACE and microarray results further support the idea that fabZ1 is under the direct
regulation of FabT in L. lactis.
The repressor FabT has a direct effect on transcription of most of the FAB genes. To
determine the physiological effects of the fabT mutant, the acyl chain composition
was determined. Deletion of fabT in L. lactis leads to an increase in SFA from 51% to
66% (Fig. 5). Complementation of the mutation by overexpression of the Strep-FabT
repressor, even for as long as two hours (approximately two doubling times), does not
significantly change the UFA/SFA ratio. Apparently, two generations of repressive
circumstances on the genes in the fab operon do not deplete the already available FAB
enzymes to such an extent as to have an effect on this ratio. Alternatively, the high intracellular amounts of Strep-FabT, in the absence of co-repressor molecules due to the
observed repression of the acp gene, might result in such a disturbed FAB regulation
that the wildtype UFA/SFA ratio of 50% can no longer be obtained.
A change in membrane fluidity due to a change in UFA/SFA ratio can be observed
123
by an increased uptake of ethidium bromide 51. Higher relative amounts of saturated
fatty acids in the L. lactis fabT mutant result in a higher permeability for a compound
as ethidium bromide (Fig. 5B). This increased sensitivity for the uptake of such compounds could potentially be an interesting phenotype for industrial applications. Engineering FAB for industrial purposes has shown to be effective for E. coli, indicating
the robustness of this synthetic route 52.
In studies in which various stress conditions have been applied to L. lactis, changes in
expression of the genes for fatty acid biosynthesis are often seen 53–56. Further characterization of the fab cluster and the regulatory mechanism identified here will help to
understand the intricacies of membrane integrity and modulation in L. lactis.
Acknowledgements
We thank STW for financial support by funding project 08080.
We would like to thank Ingrid Martini and Frits Muskiet from University Hospital in
Groningen for analyzing the samples for their acyl chain composition.
124
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
James, E. S. & Cronan, J. E. Expression of two Escherichia coli acetyl-CoA
carboxylase subunits is autoregulated. J. Biol. Chem. 279, 2520–2527 (2004).
Verwoert, I. I., Verbree, E. C., Van der Linden, K. H., Nijkamp, H. J. & Stuitje,
A. R. Cloning, nucleotide sequence, and expression of the Escherichia coli
fabD gene, encoding malonyl coenzyme A-acyl carrier protein transacylase. J.
Bacteriol. 174, 2851–2857 (1992).
Choi, K.-H., Heath, R. J. & Rock, C. O. β-Ketoacyl-acyl carrier protein synthase III (FabH) is a determining factor in branched-chain fatty acid biosynthesis. J. Bacteriol. 182, 365–370 (2000).
Price, A. C., Zhang, Y. M., Rock, C. O. & White, S. W. Structure of beta-ketoacyl-[acyl carrier protein] reductase from Escherichia coli: negative cooperativity and its structural basis. Biochem. 40, 12772–12781 (2001).
Marrakchi, H., Choi, K.-H. & Rock, C. O. A new mechanism for anaerobic unsaturated fatty acid formation in Streptococcus pneumoniae. J. Biol. Chem. 277,
44809–44816 (2002).
Mohan, S., Kelly, T. M., Eveland, S. S., Raetz, C. R. & Anderson, M. S. An
Escherichia coli gene (FabZ) encoding (3R)-hydroxymyristoyl acyl carrier protein dehydrase. Relation to fabA and suppression of mutations in lipid A biosynthesis. J. Biol. Chem. 269, 32896–32903 (1994).
Heath, R. J. & Rock, C. O. Enoyl-acyl carrier protein reductase (fabI) plays a
determinant role in completing cycles of fatty acid elongation in Escherichia
coli. J. Biol. Chem. 270, 26538–26542 (1995).
Marrakchi, H. et al. Characterization of Streptococcus pneumoniae enoyl-(acylcarrier protein) reductase (FabK). Biochem. J. 370, 1055–1062 (2003).
Edwards, P., Sabo Nelsen, J., Metz, J. G. & Dehesh, K. Cloning of the fabF
gene in an expression vector and in vitro characterization of recombinant fabF
and fabB encoded enzymes from Escherichia coli. FEBS Lett. 402, 62–66
(1997).
Wang, H. & Cronan, J. E. Functional replacement of the FabA and FabB proteins of Escherichia coli fatty acid synthesis by Enterococcus faecalis FabZ and
FabF homologues. J. Biol. Chem. 279, 34489 –34495 (2004).
Paoletti, L., Lu, Y.-J., Schujman, G. E., De Mendoza, D. & Rock, C. O. Coupling of fatty acid and phospholipid synthesis in Bacillus subtilis. J. Bacteriol.
189, 5816–5824 (2007).
Yoshimura, M., Oshima, T. & Ogasawara, N. Involvement of the YneS/YgiH
and PlsX proteins in phospholipid biosynthesis in both Bacillus subtilis and
Escherichia coli. BMC Microbiol. 7, 69 (2007).
Zhang, Y.-M. & Rock, C. O. A rainbow coalition of lipid transcriptional regulators. Molec. Microbiol. 50, 115-119 (2010).
Henry, M. F. & Cronan, J. E. Escherichia coli transcription factor that both activates fatty acid synthesis and represses fatty acid degradation. J. Molec. Biol.
125
222, 843–849 (1991).
15. DiRusso, C. C., Tsvetnitsky, V., Højrup, P. & Knudsen, J. Fatty acyl-CoA binding domain of the transcription factor FadR. Characterization by deletion, affinity labeling, and isothermal titration calorimetry. J. Biol. Chem 273, 33652–
33659 (1998).
16. Feng, Y. & Cronan, J. E. Complex binding of the FabR repressor of bacterial
unsaturated fatty acid biosynthesis to its cognate promoters. Molec. Microbiol.
80, 195–218 (2011).
17. Martinez, M. A. et al. A novel role of malonyl-ACP in lipid homeostasis. Biochem. 49, 3161–3167 (2010).
18. Zhang, Y.-M. & Rock, C. O. Transcriptional regulation in bacterial membrane
lipid synthesis. J. Lipid Res. 50, 115–119 (2009).
19. Jerga, A. & Rock, C. O. Acyl-acyl carrier protein regulates transcription of fatty
acid biosynthetic genes via the FabT repressor in Streptococcus pneumoniae. J.
Biol. Chem. 284, 15364–15368 (2009).
20. Sambrook, J., Fritsch, E. F. & Maniatis, T. Molecular cloning: a laboratory
manual. (1989).
21. Johansen, E. & Kibenich, A. Isolation and characterization of IS1165, an insertion sequence of Leuconostoc mesenteroides subsp. cremoris and other lactic
acid bacteria. Plasmid 27, 200–206 (1992).
22. Defoor, E., Kryger, M.-B. & Martinussen, J. The orotate transporter encoded by
oroP from Lactococcus lactis is required for orotate utilization and has utility as
a food-grade selectable marker. Microbiol. 153, 3645–3659 (2007).
23. Skerra, A. & Schmidt, T. G. Use of the Strep-Tag and streptavidin for detection
and purification of recombinant proteins. Meth. Enzymol. 326, 271–304 (2000).
24. De Ruyter, P. G., Kuipers, O. P. & De Vos, W. M. Controlled gene expression
systems for Lactococcus lactis with the food-grade inducer nisin. Appl. Environ. Microbiol. 62, 3662–3667 (1996).
25. Kuipers, O. P., De Ruyter, P. G. G., Kleerebezem, M. & De Vos, W. M. Quorum
sensing-controlled gene expression in lactic acid bacteria. J. Biotechnol. 64,
15–21 (1998).
26. Laemmli, U. K. Cleavage of structural proteins during the assembly of the head
of bacteriophage T4. Nature 227, 680–685 (1970).
27. Larsen, R., Kloosterman, T. G., Kok, J. & Kuipers, O. P. GlnR-mediated regulation of nitrogen metabolism in Lactococcus lactis. J. Bacteriol. 188, 4978–4982
(2006).
28. Kuipers, O. P. et al. Transcriptome analysis and related databases of Lactococcus lactis. Antonie Van Leeuwenhoek 82, 113–122 (2002).
29. Van Hijum, S. A. F. T., García de la Nava, J., Trelles, O., Kok, J. & Kuipers, O.
P. MicroPreP: a cDNA microarray data pre-processing framework. Appl. Bioinf.
2, 241–244 (2003).
30. Long, A. D. et al. Improved statistical inference from DNA microarray data
126
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
using analysis of variance and a bayesian statistical framework. J. Biol. Chem.
276, 19937 –19944 (2001).
Muskiet, F. A. J., Van Doormaal, J. J., Martini, I. A., Wolthers, B. G. & Van der
Slik, W. Capillary gas chromatographic profiling of total long-chain fatty acids
cholesterol in biological materials. J. Chromatogr. 278, 231–244 (1983).
Hamoen, L. W., Van Werkhoven, A. F., Bijlsma, J. J. E., Dubnau, D. & Venema,
G. The competence transcription factor of Bacillus subtilis recognizes short
A/T-rich sequences arranged in a unique, flexible pattern along the DNA helix.
Genes Dev. 12, 1539–1550 (1998).
Rauch, P. J. & De Vos, W. M. Characterization of the novel nisin-sucrose conjugative transposon Tn5276 and its insertion in Lactococcus lactis. J. Bacteriol.
174, 1280 –1287 (1992).
Bethesda Research Laboratories BRLpUC host E. coli DH5αTM competent
cells. Bethesda Res. Lab. Focus 8, 9–12
Solem, C., Defoor, E., Jensen, P. R. & Martinussen, J. Plasmid pCS1966, a
new selection/counterselection tool for lactic acid bacterium strain construction
based on the oroP gene, encoding an orotate transporter from Lactococcus lactis. Appl. Environ. Microbiol. 74, 4772 –4775 (2008).
Wegmann, U. et al. Complete genome sequence of the prototype lactic acid
bacterium Lactococcus lactis subsp. cremoris MG1363. J. Bacteriol. 189,
3256–3270 (2007).
Alekshun, M. N., Levy, S. B., Mealy, T. R., Seaton, B. A. & Head, J. F. The
crystal structure of MarR, a regulator of multiple antibiotic resistance, at 2.3 Å
resolution. Nat. Struct. Mol. Biol. 8, 710–714 (2001).
Schujman, G. E., Paoletti, L., Grossman, A. D. & De Mendoza, D. FapR, a
bacterial transcription factor involved in global regulation of membrane lipid
biosynthesis. Dev. Cell 4, 663–672 (2003).
Bailey, T. L. et al. MEME SUITE: tools for motif discovery and searching.
Nucleic Acids Res. 37, W202–208 (2009).
Liu, X. S., Brutlag, D. L. & Liu, J. S. An algorithm for finding protein/DNA
binding sites with applications to chromatin-immunoprecipitation microarray
experiments. Nature Biotechnol. 20, 835–839 (2002).
Lu, Y.-J. & Rock, C. O. Transcriptional regulation of fatty acid biosynthesis in
Streptococcus pneumoniae. Molec. Microbiol. 59, 551–566 (2006).
Lu, Y.-J., White, S. W. & Rock, C. O. Domain swapping between Enterococcus
faecalis FabN and FabZ proteins localizes the structural determinants for isomerase activity. J. Biol. Chem. 280, 30342 –30348 (2005).
Morgan-Kiss, R. M. & Cronan, J. E. The Lactococcus lactis FabF fatty acid
synthetic enzyme can functionally replace both the FabB and FabF proteins of
Escherichia coli and the FabH protein of Lactococcus lactis. Arch. Microbiol.
190, 427–437 (2008).
Fozo, E. M. & Quivey, R. G., Jr The fabM gene product of Streptococcus mu127
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
128
tans is responsible for the synthesis of monounsaturated fatty acids and is necessary for survival at low pH. J. Bacteriol. 186, 4152–4158 (2004).
Budin-Verneuil, A., Maguin, E., Auffray, Y., Ehrlich, S. D. & Pichereau, V.
Transcriptional analysis of the cyclopropane fatty acid synthase gene of Lactococcus lactis MG1363 at low pH. FEMS Microbiol. Lett. 250, 189–194 (2005).
Cronan, J. E. & Subrahmanyam, S. FadR, transcriptional co-ordination of metabolic expediency. Mol. Microbiol. 29, 937–943 (1998).
Evans, K., Adewoye, L. & Poole, K. MexR repressor of the mexAB-oprM
multidrug efflux operon of Pseudomonas aeruginosa: Identification of MexR
binding sites in the mexA-mexR intergenic region. J. Bacteriol. 183, 807–812
(2001).
Martin, R. G., Jair, K. W., Wolf, R. E. & Rosner, J. L. Autoactivation of the
marRAB multiple antibiotic resistance operon by the MarA transcriptional activator in Escherichia coli. J. Bacteriol. 178, 2216–2223 (1996).
Zhang, Y.-M. & Rock, C. O. Membrane lipid homeostasis in bacteria. Nat. Rev.
Micro. 6, 222–233 (2008).
Cybulski, L. E. et al. Mechanism of membrane fluidity optimization: isothermal control of the Bacillus subtilis acyl-lipid desaturase. Molec. Microbiol. 45,
1379–1388 (2002).
Viegas, C. A. et al. Yeast adaptation to 2,4-dichlorophenoxyacetic acid involves
increased membrane fatty acid saturation degree and decreased OLE1 transcription. Biochem. Biophys. Res. Comm. 330, 271–278 (2005).
Lu, X., Vora, H. & Khosla, C. Overproduction of free fatty acids in E. coli: Implications for biodiesel production. Metab. Eng. 10, 333–339 (2008).
Kramer, N. E., Van Hijum, S. A. F. T., Knol, J., Kok, J. & Kuipers, O. P. Transcriptome analysis reveals mechanisms by which Lactococcus lactis acquires
nisin resistance. Antimicrob. Agents Chemother. 50, 1753–1761 (2006).
Marreddy, R. K. R. et al. The response of Lactococcus lactis to membrane protein production. PLoS ONE 6, e24060 (2011).
Pinto, J. P. C., Kuipers, O. P., Marreddy, R. K. R., Poolman, B. & Kok, J. Efficient overproduction of membrane proteins in Lactococcus lactis requires
the cell envelope stress sensor/regulator couple CesSR. PLoS ONE 6, e21873
(2011).
Zomer, A. Adaptation of Lactococcus lactis to its environment, a genomics approach. (2007).
Elution 6
Elution 5
Elution 4
Elution 3
Elution 2
Elution 1
Wash step
Flowthrough
A
Cell free extract
Marker
Supplementary material Figures
70kD
55kD
40kD
35kD
25kD
15kD
B
70kD
55kD
40kD
35kD
25kD
15kD
Figure S1. Purification of Strep-FabT under native conditions. (A) SDS-PAGE and Coomassie
Brilliant blue staining on purification steps of Strep-FabT. (B) Western blot of the same gel
decorated with anti-Streptag antibodies. Closed arrow: Strep-FabT, open arrow: Strep-FabT
dimer
129
Supplementary material Tables
Table S1. Primers used in this study. (A) Restriction sites are underlined, KO stands for
knockout. (B) F: forward primer, R: reverse primer.
A
Name of primers
used for cloning
Sequence (5’ - 3’)
Purpose
Pr1
GGCCGCTCGAGGTAGACAACTTACCAAAAGTAGGC
fabT downstream
region, for KO
Pr2
GGCCGGGATCCATGAAATAATGACTTTTGCG
fabT downstream
region, for KO
Pr3
GGCCGGGATCCTGTTTTCATCTGTCATCCTCTTC
fabT upstream region, for KO
Pr4
GGCCGTCTAGACGATTGATGTCTTGAATATTCCGGG
fabT upstream region, for KO
Pr6
GGCCGTCTAGATTATTTCATATTCTCCAAAG
To place fabT
under control of
PnisA
Pr61
GGCCGCCATGGCTTGGAGCCATCCACAATTTG
AAAAAGGTTCTAAAACAGATTTTGATAAGG
To add an N-terminal Strep-tag
to FabT
pNZSeq
ACGTGCTGTAATTTGTTTAATTGC
Sequencing
nis_fw_aldert
GGCTCTGATTAAATTCTGAAG
Sequencing
B
130
Name of primers
used for EMSA
Sequence (5’ - 3’)
Upstream fragment of
gene
F fabT1 up
CAGCCATTCATTTACCTTATC
Proximal upstream fabT
R fabT1 up
GCAATATAAAAACAAATTGAGTTC
Proximal upstream fabT
F fabT2 up
CTTTCGCGCAAGATGAACTC
Distal upstream fabT
R fabT2 up
GCCAGCAATCCAGCCTAATC
Distal upstream fabT
F fabD up
GTCATAATATATAGTAATCAACTA
fabD
R fabD up
AAATAATTCGGGATGGG
fabD
F accC up
TAAGTCCGAATTTACCATAAG
accC
R accC up
GCGGTTGTTGAAGCAGAAGC
accC
F fabI up
GCCCGAGTAATCAGGCGAAAAGAGGCGTCACTC
fabI
R fabI up
GCAACGCCCATGATAACAATTTTTTTACCTTC
fabI
F fabH up
CGTTTTCTGGCACATAATGTGCCGCTTGCG
fabH
Name of primers
used for EMSA
Sequence (5’ - 3’)
Upstream fragment of
gene
R fabH up
GTGACTCATATTATACTGACATCAAAAGGTGACGAG
fabH
F acpP up
CCGAGTTCGTCAACGATAATATCTTGTAC
acpP
R acpP up
GAGTCAAAAAGTTTAGCCTACTTTTGG
acpP
F fabG1 up
CCACGAGTTGAACCAGTTACAAAGACG
fabG1
R fabG1 up
GTCAGTAAAATTGGAGAAAGGAATTGAG
fabG1
F fabF up
CACCATAACCAGTGATAACAACTCTATTTGTC
fabF
R fabF up
GCAATGAAAGGGCAAATTCCAATGAAACG
fabF
F accB up
GGTTCATCAAATCTTTTACTTCTGAAATG
accB
R accB up
CCAGTAAATGCTGGAACAACTGAACTTGATGAAGG
accB
F fabZ2 up
GCCTCCATAATTTCAGTTACATTAATG
fabZ2
R fabZ2 up
GGAGAAGATATCGTTGAATTTGGTCAAG
fabZ2
F accD up
R accD up
GGCTGCTCGGCTTGCTTTTCAATAATGGAACGATTTGG
CTTACCAGAATATCAAAAAGAAGATTAATCAATAG
accD
accD
F accA up
CTTACAGCAATTTCGCCACGATTGGCAATC
accA
R accA up
CCATAAGATTTTTTACTTATGGTAAATTCGGAC
accA
F fabT sp D39 up
TTCAAAGAAGGAGTGCGAGC
fabT
R fabT sp D39 up
AAAACGACTACCTCTCAAGTTCAC
fabT
Table S2. List of differentially expressed genes found in the three transcriptomic comparisons
(Fig. 1). For this list of genes, the statistical significance (p value ≤ 0.05 and a ratio ≥ |1.5|)
in at least two of the three experiments was required, with the notable exception for the fab
genes. The genes are sorted by their COG-category.
NZ9000 vs.
NZ9000 ΔfabT
NZ9000 pNZ8048 vs.
NZ9000 pNZstrepfabT
p
NZ9000 ΔfabT pNZ8048 vs.
NZ9000 ΔfabT pNZstrepfabT
p
p
(Proposed) function
Energy production and conversion
llmg_1726
llmg1726
2,2
2E-01
-1,8
1E-05
-2,1
4E-03
galactose-1-phosphate uridylyltransferase
llmg_1735
noxA
1,6
6E-01
-2,3
2E-03
-2,4
3E-04
NADH dehydrogenase
1E-05
-1,8
1E-02
-1,7
3E-03
branched chain amino acid ABC transporter carrier protein
Amino acid transport and metabolism
llmg_0650
brnQ
2,0
131
llmg_1185
lysA
1,3
2E-03
1,5
6E-03
2,5
6E-04
LysA protein
llmg_1700
choQ
1,2
5E-02
-3,0
1E-03
-3,5
9E-05
choline ABC transporter ATP binding
protein
llmg_1925
aroK
1,2
4E-03
-1,4
1E-02
-1,6
1E-03
AroK protein
llmg_1926
aroA
1,3
5E-04
-1,5
1E-03
-1,5
4E-02
3-phosphoshikimate 1-carboxyvinyltransferase
llmg_1938
aroB
1,1
2E-03
1,2
4E-01
1,2
2E-03
3-dehydroquinate synthase
llmg_2330
llmg2330
2,2
2E-02
2,0
4E-03
1,8
4E-03
amino-acid ABC transporter extracellular
binding protein
Nucleotide transport and metabolism
llmg_0467
pyrG
1,4
5E-01
2,0
7E-04
1,7
3E-03
CTP synthetase
llmg_1089
llmg1089
-1,1
7E-01
-1,5
3E-04
-3,0
3E-03
carbamoyl phosphate synthase large
subunit
llmg_1188
llmg1188
1,1
6E-01
1,3
3E-03
1,7
4E-04
hypothetical protein
llmg_1444
mutX
-1,0
1E+00
-1,5
2E-03
-1,3
2E-04
ADP-ribose pyrophosphatase
1E-01
-1,5
1E-03
-1,3
9E-04
N-acetylglucosamine catabolic protein
Carbohydrate transport and metabolism
llmg_1414
nagD
1,8
Coenzyme transport and metabolism
llmg_0544
dfpB
1,4
1E-03
1,0
9E-01
1,6
3E-03
phosphopantothenate--cysteine ligase
llmg_2162
birA2
2,0
2E-01
1,4
4E-03
1,2
2E-03
acetyl-CoA carboxylase ligase / biotin
operon repressor bifunctional protein
Lipid transport and metabolism
llmg_0538
fabZ1
6,5
4E-03
-1,5
7E-02
-5,3
8E-06
(3R)-hydroxymyristoyl-ACP dehydratase
llmg_0539
fabI
2,5
7E-04
-1,2
6E-03
-2,7
3E-03
enoyl-ACP reductase
llmg_1777
accA
3,3
7E-02
-1,4
2E-01
-3,2
2E-03
AccA protein
llmg_1778
accD
1,7
3E-01
-1,4
1E-01
-2,6
1E-04
acetyl-CoA carboxylase subunit beta
llmg_1779
accC
2,3
1E-01
-1,2
2E-01
-2,5
3E-03
acetyl-CoA carboxylase biotin carboxylase
subunit
llmg_1781
fabZ2
3,4
2E-02
-1,8
2E-01
-3,6
8E-04
(3R)-hydroxymyristoyl-ACP dehydratase
llmg_1782
accB
3,5
1E-04
-1,2
3E-02
-3,4
3E-04
acetyl-CoA carboxylase biotin carboxyl
carrier protein subunit
llmg_1783
fabF
2,7
2E-02
-1,7
4E-03
-2,7
4E-03
3-oxoacyl-ACP synthase
llmg_1784
fabG1
4,4
2E-03
-1,7
1E-01
-3,0
3E-05
3-ketoacyl-ACP reductase
llmg_1785
fabD
4,5
8E-04
-1,9
3E-03
-3,4
5E-03
malonyl CoA-acyl carrier protein transacylase
llmg_1786
acpA
1,8
1E-01
-1,4
7E-02
-2,9
5E-04
acyl carrier protein
llmg_1787
fabH
-1,1
3E-01
1,2
3E-01
1,2
5E-01
3-oxoacyl-ACP synthase
llmg_1788
fabT
-2,5
4E-02
-2,3
2E-02
135,9
2E-05
MarR family transcriptional regulator
Translation, ribosomal structure and biogenesis
rluE
1,1
8E-01
-1,7
1E-03
-1,9
2E-03
ribosomal large subunit pseudouridine
synthase
llmg_0557
prfA
1,1
7E-01
-1,7
4E-03
-1,4
5E-03
peptide chain release factor 1
llmg_1271
gidC
2,6
2E-01
-1,5
1E-03
-1,6
2E-03
tRNA (uracil-5-)-methyltransferase Gid
llmg_0384
132
Transcription
llmg_0439
llmg0439
-3,9
9E-04
1,8
2E-03
-1,0
8E-01
LacI family transcription regulator
llmg_0775
ccpA
-1,3
6E-01
-1,9
2E-03
-3,4
2E-04
catabolite control protein A
llmg_0947
cmhR
2,2
1E-03
1,4
4E-02
2,0
2E-04
HTH-type transcriptional regulator cmhR
llmg_1238
cspD2
1,9
4E-01
-3,6
3E-03
-3,7
4E-03
cold shock-like protein cspD2
llmg_1627
rmaH
1,1
2E-01
3,2
3E-04
3,4
2E-04
MarR family transcriptional regulator
llmg_2339
llmg2339
-1,7
5E-03
1,1
3E-01
1,7
3E-03
transcriptional regulator
llmg_2517
llmg2517
2,4
8E-02
-1,6
1E-03
-3,3
2E-04
hypothetical protein
Replication, recombination and repair
llmg_0359
recR
2,1
2E-01
-1,5
4E-04
-1,7
7E-04
recombination protein RecR
llmg_0606
recJ
1,1
6E-01
1,9
7E-04
1,7
3E-03
single strand DNA-specific exonuclease
llmg_1133
sbcC
1,3
1E-01
-1,4
4E-04
-1,4
3E-03
nuclease sbcCD subunit C
llmg_1270
llmg1270
-1,2
2E-02
2,0
2E-04
1,5
3E-03
site-specific tyrosine recombinase XerS
Cell wall/membrane/envelope biogenesis
llmg_0600
llmg0600
-1,7
1E-01
-1,9
8E-04
-2,0
1E-03
glycosyl transferase
llmg_1329
murB
2,4
6E-04
1,1
3E-01
-2,5
7E-04
UDP-N-acetylenolpyruvoylglucosamine
reductase
llmg_1516
glmS
-1,8
1E-01
1,9
2E-04
2,6
2E-03
glucosamine--fructose-6-phosphate
aminotransferase
llmg_1667
llmg1667
-1,5
3E-01
-2,3
2E-04
-2,4
1E-04
glycosyltransferase
llmg_1699
choS
-1,1
5E-01
-2,7
1E-05
-3,1
5E-03
choline ABC transporter permease and
substrate binding protein
llmg_1989
murE
1,4
2E-01
1,3
6E-04
1,4
2E-04
UDP-N-acetylmuramoylalanyl-D-glutamate--2, 6-diaminopimelate ligase
Posttranslational modification, protein turnover, chaperones
llmg_0411
groEL2
-1,2
7E-01
1,8
3E-03
2,3
2E-03
molecular chaperone GroEL
llmg_0986
clpB
-1,7
3E-02
1,5
6E-04
2,2
2E-03
ATP-dependent Clp protease
llmg_1088
gpo
-1,1
9E-01
-2,7
2E-03
-2,4
3E-04
gluthatione peroxidase
llmg_1588
trxB1
1,8
1E-01
-1,4
5E-03
-2,0
2E-03
TrxB1 protein
llmg_2419
htrA
1,2
1E-01
-1,9
4E-03
-2,3
3E-03
housekeeping protease
Inorganic ion transport and metabolism
llmg_0032
ps124
1,0
8E-01
15,4
2E-04
4,8
3E-03
hypothetical protein
llmg_1086
llmg1086
-1,2
5E-01
2,4
2E-03
1,8
1E-03
cation (calcium) transporting ATPase
llmg_1138
mtsA
-2,1
4E-01
-2,0
1E-03
-1,6
2E-04
manganese ABC transporter substrate
binding protein
llmg_1353
orf53
-1,2
1E-03
-1,6
7E-04
-1,6
4E-02
tellurite resistance protein
llmg_1490
mntH
1,7
5E-03
-1,1
9E-05
-1,3
1E-01
proton-dependent manganese transporter
group C beta
llmg_1770
noxC
-1,5
3E-01
2,2
1E-04
2,2
1E-03
NADH oxidase
llmg_1771
llmg1771
-1,2
5E-01
2,1
4E-03
1,9
2E-03
rhodanese-related sulfurtransferase
3E-03
-1,2
4E-02
hypothetical protein
Secondary metabolites biosynthesis, transport and catabolism
llmg_2543
llmg2543
2,0
4E-03
1,3
133
General function prediction only
llmg_0030
ps126
-1,3
2E-01
19,5
2E-04
8,8
9E-05
DNA primase
llmg_0604
llmg0604
-1,2
4E-01
2,3
2E-03
2,2
3E-03
ribonuclease Z
llmg_1135
llmg1135
-1,4
2E-01
-2,1
2E-04
-2,4
5E-04
hypothetical protein
llmg_1852
llmg1852
1,3
1E-02
-1,7
2E-03
-1,5
4E-03
hypothetical protein
Function unknown
llmg_0029
ps127
-1,0
9E-01
8,2
2E-04
4,0
8E-05
hypothetical protein
llmg_0034
ps122
-1,0
7E-01
5,4
1E-03
2,6
2E-03
hypothetical protein
-1,3
5E-01
10,4
4E-03
5,2
1E-03
DNA binding protein
llmg_0035
llmg_0043
ps113
-2,3
1E-01
-1,5
4E-04
-1,4
2E-03
hypothetical protein
llmg_0222
WefC
1,0
3E-01
-2,0
9E-04
-1,7
5E-03
hypothetical protein
llmg_0322
llmg0322
1,2
5E-01
1,9
5E-05
1,8
4E-03
cation transporter
llmg_0502
llmg0502
-1,2
5E-02
-1,3
2E-03
-3,7
3E-03
ABC transporter permease
llmg_0527
llmg0527
1,2
3E-01
1,5
2E-03
1,8
7E-04
hypothetical protein
llmg_0542
llmg0542
1,2
2E-01
1,7
3E-03
2,4
6E-05
hypothetical protein
llmg_0599
llmg0599
1,4
6E-01
-1,9
5E-03
-2,1
2E-03
hypothetical protein
llmg_0754
llmg0754
1,2
4E-01
-1,4
1E-04
-1,5
3E-03
Rgg family transcriptional regulator
llmg_0792
ps303
1,1
8E-01
-1,7
2E-03
-1,5
4E-04
hypothetical protein
llmg_1061
llmg1061
-1,1
2E-01
-1,6
4E-03
-2,0
5E-05
hypothetical protein
llmg_1186
llmg1186
1,4
3E-01
1,4
3E-03
1,6
8E-04
hypothetical protein
-2,3
2E-03
-3,0
1E-03
-2,2
3E-02
hypothetical protein
-3,0
6E-04
-1,9
1E-02
-1,7
2E-03
hypothetical protein
llmg_1300
llmg_1301
llmg1301
llmg_1302
-3,3
1E-03
-2,4
5E-04
-2,4
2E-02
hypothetical protein
llmg_1330
llmg1330
4,9
2E-04
-1,7
6E-04
-4,3
6E-03
hypothetical protein
llmg_1366
orf41
-1,2
3E-01
-1,6
2E-03
-1,5
7E-05
hypothetical protein
llmg_1499
llmg1499
1,0
1E+00
-1,5
4E-04
-1,5
2E-03
hypothetical protein
llmg_1666
llmg1666
-2,3
1E-01
-1,4
3E-03
-1,4
1E-03
hypothetical protein
llmg_1668
llmg1668
-1,6
5E-01
-1,7
5E-04
-2,0
9E-04
hypothetical protein
llmg_1760
llmg1760
1,6
4E-01
-1,4
3E-03
-1,4
4E-03
hypothetical protein
llmg_1822
llmg1822
-1,6
4E-03
-1,2
8E-05
-1,1
8E-02
hypothetical protein
1,4
2E-01
-1,9
5E-04
-1,5
1E-03
hypothetical protein
llmg_1962
llmg_2148
llmg2148
-1,3
2E-01
-1,2
9E-04
-1,2
5E-03
hypothetical protein
llmg_2194
llmg2194
2,2
3E-01
-1,7
6E-03
-2,9
4E-04
hypothetical protein
llmg_2251
ps518
-1,1
4E-01
27,1
4E-06
10,0
1E-04
phage protein by Glimmer/Critica
llmg_2256
ps513
1,0
1E+00
5,0
1E-03
2,7
9E-05
hypothetical protein
llmg_2282
llmg2282
-1,6
1E-02
1,8
2E-03
1,8
2E-03
hypothetical protein
1,9
2E-02
1,8
5E-04
2,3
4E-04
hypothetical protein
llmg_2527
ps603
-1,1
4E-01
36,3
1E-05
12,6
1E-05
hypothetical protein
llmg_2528
ps604
-1,2
9E-02
19,4
2E-04
6,6
2E-03
hypothetical protein
llmg_2529
-1,3
5E-01
24,2
7E-05
27,0
1E-03
hypothetical protein
llmg_2530
-1,2
5E-01
26,8
7E-07
32,1
8E-04
hypothetical protein
llmg_2531
2,3
4E-01
48,1
9E-06
49,6
7E-05
hypothetical protein
llmg_2495
134
llmg_2532
ps608
-1,2
3E-02
12,8
4E-05
4,2
5E-03
hypothetical protein
llmg_2533
ps609
-1,3
6E-02
3,9
1E-03
2,3
1E-03
hypothetical protein
llmg_2533
ps609
-1,4
2E-01
21,4
9E-05
6,1
4E-04
hypothetical protein
Signal transduction mechanisms
llmg_0908
llrA
-1,9
3E-01
-1,6
1E-03
-1,4
2E-04
two-component system regulator llrA
llmg_1559
flpA
1,9
3E-02
2,0
9E-04
1,9
1E-03
FNR like protein A
ypbC
1,2
1E-01
1,3
1E-03
2,0
2E-03
hypothetical protein
Defense mechanisms
llmg_1015
135
136
Chapter 5
The role of YfiA in ribosomal stalling in
Lactococcus lactis
Thomas H. Eckhardt, Pranav Puri, Linda E. Franken, Fabrizia Fusetti, Marc C. A. Stuart,
Bert Poolman, Jan Kok and Oscar P. Kuipers
137
Abstract
Dimerization of ribosomes, so-called ribosome stalling, inEscherichia coli is a twostep process. First, ribosome modulation factor RMF holds the large and small subunits of the ribosome together, after which hibernation promotion factor HPF binds to
complete the ribosome dimer state. Lactococcus lactis MG1363 has a paralog of HPF,
annotated as YfiA (YfiALl; Llmg0616), but none for RMF. In this study we show that
YfiALl is involved in the dimerization of ribosomes upon entry of lactococcal cells
into the stationary phase. The N-terminal amino acid sequence of YfiALl is homologous with that of HPF of E. coli. L. lactis ΔyfiA, in which the entire gene encoding
YfiA has been removed, does not have a strong growth effect, but failed to dimerize
ribosomes in the stationary growth phase. Introduction in L. lactis ΔyfiA of a variant
of YfiALl in which the C-terminal 59 amino acids were removed does not restore ribosome dimerization, while complementation with full-length YfiALl does. Thus, YfiALl
is a two-edged sword: the N-terminus of the protein contains the very well conserved
HPF domain, while the C-terminus is a functional homolog of the RMFEc protein.
Introduction
Dimerization of ribosomes in Escherichia coli is a stationary phase process that increases the viability of the cell 3. During the transition phase when one or more nutrients become scarce 4, or when facing stressful conditions 5, E. coli 70S ribosomes
can form 90S dimers upon binding with ribosome modulation factor RMF 1. These
intermediate ribosome dimers can subsequently bind a hibernation promotion factor
(HPF) molecule to form a mature 100S ribosomal particle in which the dimerization
interface is located between the two small subunits of the two participating ribosomes
2
. These ribosome dimers represent a hibernation state and are translationally inactive
6
. Ribosomal hibernation has been suggested to constitute a method for cells to prevent
ribosomes from being degraded by ribonucleases 7. A third protein that can bind to
ribosomes when E. coli cells enter the stationary phase is YfiA (RaiA). HPF and YfiA
are structurally similar and by X-ray diffraction and cryo-electron tomography studies
it was demonstrated that both proteins can bind to the catalytic A- and P-sites of the
ribosome 8,9. By creating heterologous complexes of Thermus thermophilus ribosomes
138
with E. coli YfiA, HPF and/or RMF, detailed information was obtained on the interaction of these proteins with the ribosome and how those lead to ribosome inactivation.
RMF blocks the ribosome-binding site of mRNAs, thus preventing interaction of the
messenger with the 16S rRNA. HPF and YfiA have nearly identical binding sites in
the ribosome, which overlap the sites where the mRNA, tRNA and initiation factors
would normally bind. Both HPF and YfiA are in the immediate vicinity of ribosomal
proteins S9 and S10 and, because their positively charged surfaces are in close vicinity
to conserved residues in the ribosomal peptidyl transferase center, protein translation
is disturbed 2.
Multiple alignment and phylogenetic analyses indicate that most bacteria have at least
one HPF homologue. These homologues can be classified into three types, long HPF,
short HPF and YfiA based on the presence of a conserved domain and additional homologous sequences 10. An open reading frame annotated as yfiA occurs in the genome
of the lactic acid bacterium Lactococcus lactis. The encoded protein, YfiALl, shares
32% amino acid sequence identity with YfiA of E. coli. YfiALl belongs to the long HPF
type: it shares 64% identity with the long HPF sequence of S. pyogenes 10. The L. lactis
genome does not contain other orthologs of rmf and hpf 11.
Not much is known about the transcriptional behavior of yfiA, rmf and hpf genes. It
was shown that transcription of rmf in E. coli is positively regulated by (p)ppGpp 3
and growth rate 5. L. lactis yfiA transcription was shown to be growth rate-dependent
(Chapter 2), and transcriptionally activated in the exponential growth phase 12 (Brouwer et al., unpublished).
Here, we studied ribosomal dimerization in L. lactis and the role of YfiA therein. Ribosome dimers in L. lactis were only seen when full-length lactococcal YfiA was present
in the cell. Next to that, we show that the C-terminal tail of YfiALl is essential for the
formation of ribosomal dimers. A yfiA mutant in L. lactis was not able to produce ribosome dimers. Finally, with RNA degradation assays we show that dimerization might
be beneficial for L. lactis by preventing rRNA degradation.
139
Material and Methods
Bacterial strains, plasmids and growth conditions
The strains and plasmids used in this study are listed in Table 1. E. coli was grown
aerobically at 37°C in TY medium (1% Bacto-Tryptone, 0.5% Bacto-yeast extract
and 1% NaCl). L. lactis strains were grown as standing cultures at 30°C in M17
medium (Difco Laboratories, Detroit, MA) supplemented with 0.5% (w/v) glucose
(GM17). Solid media contained 1.5% agar. Chloramphenicol (5 µg/ml) and erythromycin (120 µg/ml for E. coli and 2.5 µg/ml for L. lactis) were added when required.
General DNA techniques
General molecular biology techniques were performed essentially as described by
Sambrook et al. 13. Plasmid DNA was isolated using a High Pure Plasmid Isolation
Kit and protocol (Roche Applied Science, Indianapolis, IN). Chromosomal DNA from
L. lactis was isolated according to the method described by Johansen and Kibenich
14
. Polymerase chain reactions (PCR) were either performed with the Phusion enzyme
(Finnzymes, Espoo, Finland) or with a modified version of it, named PfuX7 15. This enzyme yields a uracil excision-ready PCR fragment that was subsequently ligated with
a mixture of uracil DNA glycosidase and DNA glycosylase-lyase endo VIII, commercially available as USER, using the protocol by the company (New England Biolabs,
Ipswich, MA). Colony PCR was performed with the Taq Polymerase (ThermoFisher
Scientific Inc, Waltham, MA). Primers listed in Table 2 were purchased from Biolegio
BV (Nijmegen, the Netherlands). PCR products were purified with a High Pure PCR
Product Purification Kit (Roche Applied Science) according to the protocol of the supplier. DNA electrophoresis was performed in 1x TBE buffer (89 mM Tris-HCl, 89 mM
boric acid, 2 mM EDTA, pH 8.3) in 1% agarose gels with 2 µg/ml ethidium bromide.
Electrotransformation was performed using a Bio-Rad Gene Pulser (Bio-Rad Laboratories, Richmond, CA). Sequencing reactions were done at MacroGen (Seoul, Korea).
L. lactis YfiA overexpression and protein purification
The yfiA gene was amplified by PCR using L. lactis MG1363 chromosomal DNA as
the template and primers Pr8 and Pr9 (Table 2). The PCR product, consisting of the
yfiA gene extended at the 5’-end with the codons for the Strep-tag 16, was purified, di140
gested with NcoI and XbaI and ligated into pNZ8048 17 cut with the same enzymes. The
same strategy was used to clone the yfiA gene without the coding region for a Strep-tag
(primers Pr7 and Pr9). The resulting plasmids (pNZstrepyfiA and pNZyfiA), in which
strepyfiA and the yfiA were under the control of the nisin-inducible promoter PnisA, were
obtained in E. coli and subsequently introduced in L. lactis NZ9000 18. Nucleotide se-
quences were confirmed by nucleotide sequence analysis. YfiA and Strep-tagged YfiA
(Strep-YfiA) were overexpressed using the nisin-inducible system (NICE) 17. Filtersterilized culture supernatant of the nisin-secreting strain L. lactis NZ9700 was used as
a source of nisin. Overnight cultures of L. lactis NZ9000 carrying either pNZstrepyfiA
or pNZyfiA were diluted 100-fold in 1 L of fresh GM17 medium with 5 μg/ml chloramphenicol and incubated at 30°C. Nisin-containing supernatant was added (1:500)
when the OD600 of the culture had reached 0.8. After 4 h of further incubation cells
were pelleted (7,000 g for 10 min), resuspended in either 10 ml phosphate-buffered saline (PBS) or in 10 ml PBS with 0.6% paraformaldehyde and incubated at 37°C for 20
min. Cells were centrifuged (7,000 g for 10 min at 4°C), washed once with wash buffer (150 mM NaCl, 50 mM Tris-HCl, pH 7.0), resuspended in 10 ml wash buffer and
stored at -80°C. Cells were thawed, and treated with 10 mg/ml lysozyme and Complete Mini Protease Inhibitor, according to the manufacturer’s instructions (Roche Applied Science) for 60 min at 30°C. Subsequently, 0.1 g DNAseI powder and 10 mM
MgSO4 were added and the lysozyme-treated cells were broken using a sonicator for
5 cycles of 45 sec at 65% and 15 sec of incubation on ice (Aminco, Silver Springs,
MD). The suspension was centrifuged at 9,000 g (15 min, 4°C). Strep-tagged YfiA
was purified to homogeneity using a Streptactin Sepharose column according to the
manufacturer’s instructions (IBA-GmbH, Göttingen, Germany). Samples from each
step in the purification protocol were analyzed by sodium dodecylsulphate 12%-polyacrylamidegel electrophoresis (SDS 12%-PAA) 19 and Western hybridization using
anti-Strep-tag antibodies (IBA-GmbH). The concentration of purified protein was determined via spectroscopy (Nanodrop, ThermoFisher Scientific Inc). Protein (100 μM)
was kept at -80ºC in 10% glycerol, 100 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA,
2.5 mM desthiobiotin, pH 8.0.
SDS-PAA and Western hybridization
Protein samples were mixed (1:1) with sample buffer (120 mM Tris-HCl, pH 6.8, 50%
glycerol, 100 mM DTT, 2% (w/v) SDS and 0.02% (w/v) bromophenol blue), vortexed,
141
boiled for 5 min and separated on SDS 12%-PAA gels. PageRuler Prestained Protein
Ladder (ThermoFisher Scientific Inc.) was used as a protein size marker. Gels were
stained with Coomassie Brilliant Blue (0.25% Coomassie Brilliant Blue R-250 dissolved in 25% isopropanol with 10% acetic acid) and destained in boiling demineralized water. Western blotting analysis was performed with the SNAP (Millipore Corp.,
Billerica, MA) system as follows: the PAA gel was equilibrated in transfer buffer (25
mM Tris, 192 mM glycine, 10% methanol). Proteins were transferred to a PVDF blotting membrane (Roche Applied Science) for 30 min at 20 V. The blot holders containing the blots were placed in the SNAP system, after which blocking buffer in PBST
(58 mM Na2HPO4, 17 mM NaH2PO4, 68 mM NaCl, pH 7.3, 0.1% v/v Tween-20) with
0.5% skim milk was added and the vacuum was applied. Anti-Strep-tag antibodies
conjugated to peroxidase (IBA-GmbH), diluted 1000-fold in blocking buffer, were
added to the blot holders; incubation was for 10 min at room temperature. The vacuum
was removed and the blots were washed three times with PBST prior to visualization
of immunoreactive proteins using the ECL detection kit and protocol (GE Healthcare,
Buckingham, UK).
Construction of an L. lactis yfiA double crossover mutant
Upstream and downstream regions of yfiA, PCR-amplified using primer pairs Pr3/Pr4
and Pr5/Pr6 (Table 2), respectively, were inserted in the integration vector pCS1966 20,
amplified with Pr1 and Pr2, and ligated with USER as described above. The resulting
plasmid, pCS1966ΔyfiA, was obtained in E. coli and introduced into L. lactis to allow integration via single crossover homologous recombination. An L. lactis integrant
carrying the pCS1966 construct was selected on GM17 plates with chloramphenicol.
Screening for subsequent plasmid excision was done on plates containing 5-fluoroorotic
acid by selecting against the oroP gene on pCS1966 20. A mutant carrying a clean knockout of yfiA was obtained and confirmed using PCR and nucleotide sequence analysis.
Complementation of ΔyfiA
Complementation of L. lactis ΔyfiA was performed with three different plasmids:
pIL253yfiALl, pIL253yfiALl1-126 and pIL253yfiAEc. The yfiA genes from L. lactis and E.
coli were amplified using primer pairs Pr10/Pr11 and Pr12/Pr13, respectively. A truncated version of L. lactis YfiA lacking the 59 amino acids at the C-terminus (YfiALl1-126)
was amplified using primer pair Pr10/Pr14. The amplified products were ligated in the
142
pIl253 vector 21 amplified with primer pair Pr15/Pr16 employing USER enzyme. Constructs were introduced in L. lactis ∆yfiA via electrotransformation 22.
Isolation of ribosomes
L. lactis cells were harvested by centrifugation at 7,000 g for 10 min at 4°C after 3
h or 7 h of growth in GM17 at 30°C. The cell pellet was resuspended in buffer I (20
mM Tris-HCl (pH 7.6), 15 mM magnesium acetate, 100 mM ammonium acetate, and
6 mM 2-mercaptoethanol) containing 1 mM phenylmethylsulfonyl fluoride (PMSF).
Subsequently, the cells were lysed by mixing them in two cycles with 0.2 mg glass
beads in an ice-cold Tissue-lyser (Qiagen, Venlo, the Netherlands). The homogenate
was centrifuged at 9,000 g for 15 min at 4°C. The supernatant was saved on ice and
the pellet was resuspended in buffer I with 1 mM PMSF. The suspension was centrifuged again under the same conditions. The combined supernatants (cell extracts)
were layered onto a 30% sucrose cushion in buffer I and centrifuged in a MLA 80 rotor
(Beckman, Fullerton, CA) at 206,000 g for 3 h at 4°C. By resuspending the pellet in
buffer I a crude preparation of ribosomes was obtained. To isolate ribosome monomers and dimers, this crude ribosome 400 µl preparation was layered onto a 10–40%
12 ml sucrose density gradient column in buffer I and centrifuged in a SW 32.1 Ti
rotor (Beckman) at 125,000 g for 80 min at 4°C 25. The gradient was fractionated by
collecting 400 µl fractions from the top, and the absorbance of each fraction at 260
nm was measured with a UV-1700 spectrometer CARY bio UV-Visible Spectrophotometer (Agilent Technologies, Palo Alto, CA). The fractions containing the ribosome
monomers and dimers were dialyzed at 4°C against buffer I and prepared for mass
spectrometry, RNA degradation assay and electron microscopy.
Mass spectrometry
For in-solution tryptic digestion, the 400 µl dialyzed fractions containing ribosome
monomers and dimers were precipitated overnight at -20ºC using 4 volumes of acetone. The precipitated proteins were first resuspended in 2 µl of 6 M urea and subsequently diluted to 20 µl with 100 mM triethylammonium bicarbonate (TEAB).
For reduction and alkylation of the cysteine residues, the samples were incubated
for 60 min at 55°C in the presence of 5 mM Tris (2-carboxymethyl) phosphine hydrochloride (TCEP) followed by the addition of 10 mM methyl methanethiosulfo143
nate (MMTS) and incubation at room temperature for 10 min. The protein mixture
was incubated overnight with 0.25 µg of trypsin (Trypsin Gold, mass spectrometry
grade, Promega 10 ng/µl in 25 mM NH4HCO3) at 37 ºC. Subsequently 10 µl of 25
mM NH4HCO3 was added to prevent drying, and the incubation was prolonged over-
night at 37°C. The tryptic peptides were recovered by three subsequent extractions
with 50 µl of 35%, 50%, and 70% acetonitrile in 0.1% TFA. The extracted peptides
were pooled and concentrated under vacuum. The digested peptides were analyzed
by LC/MSMS on an LTQ Obitrap XL (ThermoFisher) as described previously 26.
In-gel tryptic digestion for mass spectrometry-based protein identification was performed by excision of the gel bands and treated with 10 mM DTT followed by 55 mM
iodoacetamide in 50 mM NH4HCO3 to reduce and alkylate cysteine residues. Subsequently the gel slices were dehydrated by incubation for 5 min in 100% acetonitrile,
rehydrated in 10 µl of trypsin solution (Promega) and incubated for 2 h at 37°C. The
digested peptides were analyzed as described above.
RNAse degradation assay
Ribosome fractions from the sucrose gradient were incubated at 30 ºC for 30
min in the presence or absence of 1 µl of 10 mg/ml RNAse (Roche Applied Science), 2 or 3 µl freshly purified Strep-YfiA, and 2 or 3 µl RNAse-free buffer TEDEPC (Tris-HCl pH 8.0 plus diethylpyrocarbonate) in a total volume of 11 µl.
Reaction mixtures in RNA-loading buffer (10 mM Tris-HCl, 10 mM EDTA,
40% v/v glycerol, 0.1% v/v diethylpyrocarbonate) were loaded on an RNA gel 27.
Transfer electron microscopy and single particle analysis
Purified ribosome monomer and dimer samples were prepared for negative staining
with 2% uranyl acetate on glow-discharged carbon-coated copper grids. Electron microscopy was performed on a CM120 electron microscope (Philips, Eindhoven, the
Netherlands) equipped with a LaB6 cathode, operated at 120 kV. Images were recorded with a 4000 SP 4K slow-scan CCD camera (Gatan, Pleasanton, CA) at 80,000-fold
magnification with a pixel size of 0.375 nm at the specimen level after binning of the
images. GRACE software 28 was used for semi-automated data acquisition. Single
particles were analyzed with the Groningen Image Processing software using standard
procedures 29. After multiple alignment steps and multivariate statistical analysis, hierarchical classification based on group averages resulted in several classes, of which
144
four were visibly distinct and considered representative of the dimer side view. Images
were optimized by application of conditional summing with the correlation coefficient
of the final alignment step as a quality parameter to select the most homogeneous images in each class (500-2000 images per sum). The final sums were filtered to further
reduce noise and bring out smaller details. The surface area of the small and large
subunits of both monomer and dimer images were drawn over the final image sums.
Subsequently, they were overlaid on the electron micrographs.
Results
YfiA is conserved among Streptococcaceae
Of the proteins involved in ribosome dimerization in Escherichia coli 25, ribosome
modulation factor (RMF) and hibernation promotion factor (HPF) seem not to be encoded in the genomes of members of the Gram-positive family of Streptococcaceae.
Only the gene for a YfiA homolog is present in the annotated streptococcal genomes,
and the sequence of this protein is highly conserved (Fig. 1). A member of the Streptococcaceae, Lactococcus lactis, contains a single gene (llmg0616) coding for the YfiALl protein, which is very similar to Staphylococcus aureus HPF (SaHPF; Fig. 1). In
E. coli, RMF and HPF together dimerize ribosomes, while YfiAEc does not, neither
alone nor with any combination of RMF and HPF 1. In fact, YfiAEc prevents ribosome
dimerization 1. By contrast, SaHPF was shown to be solely responsible for ribosome
dimerization in S. aureus 32. The N-terminus of YfiALl has similarities with the E.
coli proteins YfiAEc (31%), HPF (32%) (Fig. 1). The similarity is confined to the Nterminus of YfiALl because the E. coli ribosome dimerization factors are shorter than
YfiALl and the other streptococcal YfiA proteins, and SaHPF. The latter have a wellconserved C-terminal domain that is not present in the E. coli proteins
Growth characteristics of the L. lactis ΔyfiA mutant
In order to determine the importance and functionality of yfiA for L. lactis a clean yfiA
knockout mutant was made. The fact that it was possible to make L. lactis ΔyfiA shows
that the gene is not essential. The yfiA mutant does not show a growth defect in rich
medium, nor is the viability of the culture reduced, as determined by colony counts.
The mutation also does not affect the length of the lag-phase after re-inoculation (Fig.
2).
145
Table 1. Strains and plasmids used in this study.
146
Strains
Characteristics
Reference
L. lactis NZ9000
MG1363 derivative; plasmid-free, pepN::nisRK;
host for nisin-induced protein expression
18
L. lactis NZ9000
∆yfiA
NZ9000 derivative; chromosomal deletion of
yfiA
This work
L. lactis NZ9700
Nisin producer strain
30
E.coli MC1061
High efficient transforming strain, host of recombinant plasmids for L. lactis
31
Plasmids
Characteristics
Reference
pCS1966
oroP; EmR; non-replicating, integration vector in
L. lactis
20
pCS1966 ∆yfiA
pCS1966 derivative, carrying flanking regions
of yfiA
This work
pNZ8048
Cm; nisin inducible expression vector carrying
PnisA; NcoI site used for translational fusions
17
pNZstrepyfiA
pNZ8048 derivative, with yfiA containing an
extra coding region for a N-terminal Strep-tag
inserted under the nisA promoter
This work
pIL253
EmR; High-copy number inducible expression
vector carrying Para and lac genes for selection
21
pIL253 yfiALl
pIL253 derivative, with yfiA of L. lactis
This work
pIL253 yfiALl1-126
pIL253 derivative, with yfiA of L. lactis, truncated 59 amino acids at the C-terminus
This work
pIL253 yfiAEc
pIL253 derivative, with yfiA of E. coli K12
This work
Table 2. Primers used in this study. Restriction sites are underlined. For Pr8, the nucleotides
encoding the Strep-tag are in bold and those representing the linker between the strep-tag and
yfiA are in italic.
Primer
name
Sequence (5’ - 3’)
Purpose
Pr1
ATCGTACCCUCGAGTGTTCCCTTTAGTGAGGGT
Pr2
ATACTAGTTCUAGAGCGGCCGCCAACAACC
Pr3
AGAACTAGTAUGACTAAATCTGAAAGCGACCG
Pr4
ATCCTTTGAUCATAAGAGTACCTCTTC
Pr5
ATCAAAGGAUCCACAGAATAAAAATTAAGG
Pr6
AGGGTACGAUCATGAATTCTTGGAAAGC
Pr7
GGCCGCCATGGTCAAATTTAATATCCGTGG
Pr8
GGCCGCCATGGCTTGGAGCCAT
CCCAATTTGAAAAAGGT
TCTAAAAGCATCAAATTTAATATCCG
Amplification of
pCS1966 vector
Amplification of
pCS1966 vector
Amplification of yfia upstream region
Amplification of yfia upstream region
Amplification of yfia
downstream region
Amplification of yfia
downstream region
Placing yfia downstream
of PnisA
Pr9
Pr10
Pr11
Pr12
Pr13
Pr14
Pr15
Pr16
Adding an N-terminal
Strep-tag to yfia
Placing yfia downstream
of PnisA
ATTTTGCAUGATCAAATTTAATATCCGTGGC- Amplification of L. lactis
GAA
yfiA gene
ACTTGAUTTATTATTCTGTTTCAATTAAGCAmplification of L. lactis
CATAACGACCATCTG
yfiA gene
Amplification of E. coli
ATTTTGCAUGACAATGAACATTACCAGCAAAC
yfiA gene
Amplification of E. coli
ACTTGAUCTACTCTTCTTCAACTTCTTCGAC
yfiA gene
Amplification of L. lactis
ACTTGAUTTATTAATCCTCAGCAACTTCATyfiA gene, removing 59
CAG
amino acids from Cterminus
Amplification of pIL253
ATCAAGUGTTCGCTTCGCTCTCACTG
vector
ATGCAAAAUTCCTCCGAATATTTTTTAmplification of pIL253
TACCTACC
vector
GGCCGTCTAGATAATTTTTATTCTGTTTC
147
Lactlaccre
Leucopseud
Leucocitre
Weissellak
Lactobacjo
Entcasseli
Tetragenoc
Entfaecali
Carnobacte
Lactobacca
Pediococcu
Lactobacva
Aerococcus
Lactlaclac
Lactgarvie
Streppneum
Strepmutan
Oenococcus
Staphyloco
EcoK12YfiA
EcoK12HPF
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
*
20
*
40
*
60
*
80
*
100
---------------MIKFNIRGENVEVTDAIRAYVEDKIGKLDKYFNDGHEVTAYVNLKVYTEK-RAKVEVTLPAKNVTLRAEDTSQDMYSSIDFVEEKLERQI
---------------MLDYNVRGENIEVTDAIRDYVEKRLTKLNRYIEE--SAKANVNLRTYRSDNSGKVEVTIVLPYVVLRAEDTNADMYAAVDNVSEKIERQI
---------------MLDYNVRGENIAVTDALRDYVEKRLTKLNRYIEE--KSSANVNLRTYKSDNSGKAEVTIVLPYVVLRAEDTNPDLYAAVDAVSEKLERQI
---------------MLEFIVRGENIEVTEAIKAYVIKRLTRLERYLNDDNKYVAHVNLRSYQER-TFKIEVTIQLPYLLLRAEDTQDDLYQSIDFVSDKLERQI
MCKKKRHKTLEGESTMLKYNVRGENVEVTDALREYVQKRLNKLEKYFEINSDVIAHINLKVYPDH-TAKVEVTIPLPYLVLRAEDTTDDMYKSIDFVSEKLERQI
---------------MFRYNVRGENIEVTEAIRDYVEKKVGKLERYFNDVPEATAHVNLKVYTEK-TAKVEVTIPLPFLVLRAEETSPDLYASIDLVVDKLERQI
---------------MFSYNVRGENIEVTQAIRDYVEKKVGKLERYFNDVPEATAHVNLKVYTEK-TAKAEVTIPLPYLVLRAEETSPDLYGSIDLVVDKLERQV
---------------MFRYNVRGENIEVTEAIRDYVEKKVGKLERYFSDSPEATVHVNLKVYTEK-TAKVEVTIPLPYLVLRAEETSPDLYASIDLVVDKLERQI
---------------MFKYNVRGENIEVTAAIRSYVEKKVGKVEKYFNDVPEATAHVNLKTYSDK-TAKVEVTIPLPYLVLRAEETSPDLYGSVDLVSDKLERQM
---------------MLTYNVRGENIEVTEAIRSYVEKRISKLNKFFGGSVTATAHVNLKVYSDK-TAKVEVTIPLSFLTLRAEETSPDLYASIDLVTDKLERQV
---------------MLDFNVRGENIEVTEAIRDYVEKRIGKIEKYFEQGANPKAHVNLKVYQDK-TAKVEVTIPLPYLVLRAEETSPDLYASVDLVTDKLERQL
------MLELKGEITMLKFNIRGENIEVTDSIRDYVVKRINKLQKFFEDNVEATAHVNLKVYPNR-TYKVEVTIPLPYLTLRAEETSNDMYGSVDLVTDKLERQI
---------------MFTYNVRGENIEITPAIRDYAENKISKIEKYFKDAPDTTVYVNAKVYQNG-EAKAEVTVPLPRLTLRAEETSQDLYGSIDLVVDKLERQV
---------------MIKFNIRGENVEVTDAIRAYVEDKIGKLDKYFNDGHEVTAYVNLKVYSEK-RAKVEVTLPAKNVTLRAEDTSQDMYSSIDFVEEKLERQI
---------------MIKYNIRGENFELTDSIRSYVEDKVGKLEKYFNDGHEVTAYVNLKVYSDK-RYKAEVTMPAKNVTLRAEDTAADMYAAIDFVEEKLERQI
---------------MIKYSIRGENLEVTEAIRDYVVSKLEKIEKYFQPEQELDARINLKVYREK-TAKVEVTIPLGSITLRAEDVSQDMYGSIDLVTDKIERQI
---------------MIKYSIRGENIEVTDAIRNYVESKLKKIEKYFNAEQELDARINLKVYREK-TAKVEVTIPLAPVTLRAEDVSQDMYGSIDLVVDKIERQI
---------------MIEYQIRGENMSTTDAINNYIKLRLEKLNNYIDQKNNPIAHINVRKYNEK-TFKIEVTIPLPYLTLRAEETQSDFYNAVDLVSAKLLRQI
---------------MIRFEIHGDNLTITDAIRNYIEEKIGKLERYFNDVPNAVAHVKVKTYSNS-ATKIEVTIPLKNVTLRAEERNDDLYAGIDLINNKLERQV
----------------MTMNITSKQMEITPAIRQHVADRLAKLEKWQTHLINPHIILSK----EPQGFVADATINTPNGVLVASGKHEDMYTAINELINKLERQL
----------------MQLNITGNNVEITEALREFVTAKFAKLEQYFDRINQVYVVLKVE----KVTHTSDATLHVNGGEIHASAEGQDMYAAIDGLIDKLARQL
: 89
: 88
: 88
: 89
: 104
: 89
: 89
: 89
: 89
: 89
: 89
: 98
: 89
: 89
: 89
: 89
: 89
: 89
: 89
: 85
: 85
HPF homology domain
Lactlaccre
Leucopseud
Leucocitre
Weissellak
Lactobacjo
Entcasseli
Tetragenoc
Entfaecali
Carnobacte
Lactobacca
Pediococcu
Lactobacva
Aerococcus
Lactlaclac
Lactgarvie
Streppneum
Strepmutan
Oenococcus
Staphyloco
EcoK12RMF
EcoK12YfiA
EcoK12HPF
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
*
120
*
140
*
160
*
180
*
200
*
RKYKTRMNRKPRNAVPTGQVFGDEFAP-----LDTTDEVAEDHVDIVRTKHVALKPMDAEEAVLQMDMLGHDFYVFTDADSNGTHVVYRRTDGRYGLIETE-----RKYKTKINRKSRETGFK---GID------SHDEIPETDVDDSALQIVRTKQVDLKPMSPEEAALQMDLLEHDFFIFKDAESNTDSVIYKRTDGKYGLLETAE----RKYKTKINRKSRETGFK---GID------SKNEIPATDSDDEALQIVRTKQVDLKPMSPEEAALQMDLLGHNFFIFRDADTNTDSVIYKRQDGKYGLLETVED---RKYKTKVNRKSREKGYK---GID-TFMNESLDEPQLEEEEDAQFEIVRTKHLSLKPMDVEEAILQMDLLGHNFFVFQDAESDLPAVVYKRKDGKYALIDTDL----RKYKTRINRKSREKGLKDFF----------YEDLEEEKKAPKEFDIVRNKHLDLKPMSAEEAVLQMDLLGHDFFVFEDADTNGTSIVYRRNDGRYGLIETNE----RKFKTKINRKSRETALVPPTDGDL-----FTTEANDEETNGSDLDIVRTKRLSLKPMDSEEAVLQMNMLGHNFFIFEDAETNGTSIVYRRKDGKYGLIETD-----RKFKTKINRKKRE-ASKPEANQEV-----FFEDEEPEENNGSDLDIVRTKRLSLKPMDSEEAVLQMNMLGHNFFIFEDADTTGTSIVYRRKDGKYGLIETD-----RKFKTKINRKSRETGRNNTKAAVF-----LVG--EETEETPSELDIVRTKRLSLKPMDSEEAVLQMNMLGHNFFIFEDAETNGTSIVYRRKDGKYGLIETD-----RKYKTKINRRTRGSNVVVPP----------VLPGEEQESHEDEVNIVRTKRLSLKPMDSEEAVLQMDMLGHNFFIFEDADTNGTSIVYRRKDGKYGLIETE-----RKFKTKINRKSREKGFGQIDIDAT-----APAEPKPAEDD-DNLTVVRTKRVSLKPMDSQEAILQMDMLGHNFFIFEDADTNGTSIVYKRRDGRYGLIETDE----RKYKTKINRKSRETGYKGVEIP--------SGDIVENEEEASQFDIVRTKQVSLKPMGSEEAILQMDMLGHNFFIYEDAESGSVDIVYRRRDGRYGLIESAND---RKYKTKVNRKSREKGFKSLEFVP----------TDDDAESQDDLKIVRTKQISLKPMDPEEAVLQMDMLGHDFFVFQDAETDGTSIVYRRKDGRYGLIEAE-----KKYKTRINRKSREKGISDVMFTEN-------NQEDSKDDNDSNIEIVRTKSIAVKPMSAEEAVLQMEMLGHSFFIYEDAESESVSLVYKRHNGKYGLIEIEKDIVNE
RKYKTRMNRKPRNAVPTGQAFGDEFAP-----LEKADEVAEDHVDIVRTKHVALKPMDAEEAVLQMDMLGHDFYVFTDADSNGTHVVYRRTDGRYGLIETE-----RKYKTRVNRKSKINVPTGQAFGDEFAP-----LDEAEEVNEDQVKIVRTKHVSLKPMDAEEAVLQMDMLGHDFYIFTDADTDSTNVVYRREDGNIGLIETK-----RKNKTKIERKNKNKVATGQLFTDALVE-----DSNIVQ-----SKVVRSKQIDLKPMDLEEAILQMDLLGHDFFIYVDVEDQTTNVIYRREDGEIGLLEVKES---RKNKTKIAKKHREKKPAAHVFTAEFEA-----EEMEEAPA---IKVVRTKNITLKPMDIEEARLQMDLLGHDFFIYTDANDNTTNVLYRREDGNLGLIEAK-----RKFKTRVNRKSRERGFKGIDFNEAIDP-VPSDTN-----EDKKIDVIRRKNLSLKPMDIEEAVLQMEMLDHDFFLFLNSDTNQLDIVYKRDDGKYGLIETENVESSK
RKYKTRINRKSRDRGDQEVFVAELQEMQET-QVDND-AYDDNEIEIIRSKEFSLKPMDSEEAVLQMNLLGHDFFVFTDRETDGTSIVYRRKDGKYGLIQTSEQ-----------------------MK-RQK—RDRLERAHQRGY-QAGI-AGRSKEMCPYQTLNQ--RSQW--LGG----WREAMADRVVMA-------------------NKLQHKG--------------------------------------EARRAATSVKDANFVEEVEEE----------------------------------------TKHKDKLKQH-------------------------------------------------------------------------------------------------
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
185
181
182
187
196
185
184
183
180
185
184
189
189
185
185
182
182
190
190
55
113
95
Extended C-terminal domain
Figure 1. Sequence alignment of YfiA of streptococcal members and compared with E. coli
YfiA HPF and RMF. The multiple sequences were aligned with ClustalW 31 and presented
using Boxshade 3.2.3 30. Black backgrounds indicate identical residues and grey backgrounds
indicate conserved residues. The HPF homolog conserved domain (blue box) and the extended C-terminal domain (yellow box) are adapted from Ueta et al., 2008 10. Leucopseud: Leuconostoc pseudomesenteroides, Leucocitre: Leuconostoc citreum KM20, Weissellak: Weissella
koreensis KACC15510, Lactobacjo: Lactobacillus johnsonii ATCC3320, Entcasseli: Enterococcus casseliflavus EC10, Tetragenoc: Tetragenococcus halophilus NBRC1, Entfaecali: Enterococcus faecalis V583,Carnobacte: Carnobacterium sp. AT7, Lactobacca: Lactobacillus
casei ATCC334, Pediococcu: Pediococcus pentosaceus ATCC2574, Lactobacva: Lactobacillus vaginalis ATCC4954, Aerococcus: Aerococcus urinae ACS-120-V-Col1, Lactlaccre:
Lactococcus lactis ssp. cremoris MG1363, Lactlaclac: Lactococcus lactis ssp. lactis Il1403,
Lactgarvie: Lactococcus garvieae ATCC49156, Streppneum: Streptococcus pneumoniae
TIGR4, Strepmutan: Streptococcus mutans UA159, Oenococcus: Oenococcus oeni PSU-1,
Staphyloco: Staphylococcus aureus ssp. aureus, EcoK12RMF: Escherichia coli K12RMF
EcoK12YfiA: Escherichia coli K12, EcoK12HPF: Escherichia coli K12HPF.
148
Ribosome dimerization in L. lactis
Subsequently, we examined whether L. lactis ribosomes dimerize during stationary
phase, and whether or not this was abrogated in L. lactis ΔyfiA. The dimerization
state of ribosomes in both strains was investigated using sucrose-gradient ultracentrifugation. Using this procedure, ribosomes sediment at a specific sucrose percentage
according to their sedimentation or Svedberg (S) coefficient. Cells of the wild-type
strain, harvested in the logarithmic phase of growth, showed a single major peak.
Below, we show that the fractions constituting this peak contain 70S ribosomes (Fig.
3A). Dimerization of ribosomes, indicated by the formation of particles sedimenting
at 100S, exclusively occurs in the wild-type cells that were harvested in the stationary
phase (Fig. 3B). Not all ribosomes dimerize or a fraction may dissociate in the sample
preparation, as a minor fraction of 70S ribosomes still remains. In cells lacking YfiALl,
100S ribosome particles are never observed, neither in exponentially growing (data
not shown) nor in stationary phase cells (Fig. 3C).
To verify the constitution and conformation of the ribosomes in the 70S and 100S
peaks, fractions were taken and subjected to electron microscopy. Indeed, 70S fractions contain monomeric ribosomes (Fig. 3GI), while in the 100S fractions the ribosome dimers are present (Fig. 3HJ). After aligning and filtering of the electron microscopy images, at least 500 homologous images per view were selected and summed to
obtain presented images (Fig. 3G-K). Ribosomes that dimerize are present in a specific conformation with the dimer interface located between the small subunits of two
intact ribosomes (Fig. 3H). As is clear from the superpositions presented in Fig. 3IJK,
the relative position of the small and large subunits of the ribosome in the dimerized
state has changed relative to that of the monomeric ribosome.
The C-terminus of YfiALl is essential for ribosome dimerization
As presented above, dimerization of ribosomes does not occur in L. lactis ΔyfiA (Fig.
3C). Ribosome dimerization was restored by complementing the mutation in this
strain with plasmid pILyfiALl, in which yfiA was (over)expressed by induction with
nisin of PnisA::yfiALl (Fig. 3D). A C-terminally truncated variant of the YfiALl protein lacking the 59 C-terminal amino acids, YfiALl1-126, did not restore the ribosome
dimerization defect in L. lactis ΔyfiA (Fig. 3E). Full-length YfiAEc is shorter than
YfiALl. It is homologous to the N-terminal domain of YfiALl but lacks the conserved
149
1E+11 1E+10 1E+09 1E+08 1E+07 1E+06 1E+05 1E+04 1E+03 1E+02 1E+01 1E+00 B 10.0 WT WT ∆yfiA ∆yfiA 1.0 OD600 (log) viability (CFU) A
0.1 0.0 5 9 32 54 (hours) 74 105 128 0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105 112 119 126 (hours) C 0.9 1 0.8 0.7 WT(9H) OD600 0.6 ∆yfiA(9H) 0.5 WT(32H) 0.4 WT(54H) ∆yfiA(32H) 0.3 ∆yfiA(54H) 0.2 ∆yfiA(74H) WT(74H) WT(105H) 0.1 ∆yfiA(105H) 8 9 10
11
12
13
14
15
16
17
18
19
20
21
7 6 5 4 3 2 0 -­‐0.1 1 0 (hours) a.er re-­‐inocula5on Figure 2. Viability of L. lactis wild-type (WT) and ∆yfiA strains. (A) Viability in GM17
medium. Samples were taken at the indicated time-points and used to re-inoculate GM17
(1:100) to measure the subsequent lag-phases (B) and plated to GM17 plates to determine cell
viability (C). The legenda of C indicates after how many hours re-inoculation was performed.
Results are presented as average of triplicate experiments.
C-terminal domain of YfiALl (see above and Fig. 1). Overexpressing E. coli yfiA in L.
lactis ΔyfiA does not lead to dimerization of lactococcal ribosomes (Fig. 3F).
YfiA is present in ribosome dimers of L. lactis
Ribosome monomers (70S fraction) and dimers (100S fraction) from the sucrose gradients (Fig. 3A-F) were subjected to mass spectrometric analysis. As expected, ribosomal proteins were detected in both samples. Ribosomal proteins were by far the
most abundant in the ribosome monomer fraction and other proteins detected were
very low in abundance and probably contaminants. Additional proteins that were exclusively present in the ribosome dimer preparations were the ribosomal protein RplL
(L7/L12) and the pyruvate dehydrogenase components PdhB and PdhC. The fourth
protein in the dimer fraction was YfiALl (Fig. 4A).
150
A
B
70S
70S
C
100S
D
70S
100S
E
F
70S
70S
G
H
I
J
K
Figure 3. Ribosome formation in L. lactis. Ribosome profiles of L. lactis after 5-25% sucrose
gradient ultracentrifugation were obtained (A-F). The x-axis shows the fraction numbers (top
to bottom), the peak heights are derived from the absorbance at 260 nm. Arrows indicate
the fractions used for mass spectrometry, RNA degradation assay and electron microscopy.
Cells were grown in GM17 at 30°C and were harvested in the logarithmic growth phase (log
phase) or in the stationary phase (stat phase). Peak-containing fractions number 9 and 13 were
subjected to electron microscopy, where the presence of 70S monomer (G and I) and 100S
dimer (H and J) was confirmed. By superpositioning the large (yellow) and small (purple
or pink) subunits, a tordation of the small subunits relative to the large subunit between the
monomer (I) and the dimer (J) can be observed (K).
151
Protein interactions of L. lactis YfiA
YfiA containing an N-terminal Strep-tag (Strep-YfiA; 22.8 kDa) was overexpressed in
L. lactis grown to stationary phase and purified by Strep-Tactin purification. We chose
to place the Strep-tag on the N-terminus, since the C-terminus appeared to play a major role in the dimerization, giving a risk of diminishing functionality when hooking it
up with a Strep-tag. A protein of the expected size (approximately 26 kDa) was visible
in a Coomassie-stained SDS 12%-PAA gel and on a Western blot after immunolabeling with anti-Strep antibodies (Fig. 4B). A second protein with a size of approximately
45 kDa also reacts with the anti-Strep antibodies. When the cells were treated with
4% paraformaldehyde (PFA) prior to harvesting and subsequent affinity purification,
a smear of proteins with sizes ranging from 50 kDa to 150 kDa was detected by Coomassie staining and Western hybridization (Fig. 4B). The lower amount of Strep-YfiA
observable after PFA-fixation and the appearance of a smear of bands at positions
of high-molecular weight proteins in the Western blot indicate that Strep-YfiA can
form complexes with other proteins. When the experiment was performed without
PFA-fixation no clear high-molecular weight products were observed (Fig. 4B). The
proteins from the fixated sample and the material in the equivalent gel area from the
non-fixated sample were eluted from equally sized gel-slices and subjected to MS/
MS. As a control, a similar area of an SDS 12%-PAA gel, containing the proteins from
stationary-phase cells from the L. lactis wild-type strain, was also examined (Table 3).
Table 3 shows a selection of proteins that were identified by MS/MS that appear to
have an interaction with Strep-YfiA. In the non-fixated sample, a number of proteins
appeared on the gel around the molecular weight of the specific protein plus approximately 22.8 kDa for Strep-YfiA (Table 3), suggesting that direct interactions might
exist between Strep-YfiA and these proteins. A number of these proteins are (possibly)
involved in protein synthesis (e.g. EF-Tu, RpsA, EF-G and IF-2). We also observed
co-purification of proteins involved in central metabolism, like PdhC and PdhD. These
proteins, part of the pyruvate dehydrogenase complex, are known in bacteria to copurify with rRNA modification enzymes as well as with 50S ribosome subunits 33,34.
Also, the PDH-complex is protected against trypsin digestion by membrane-bound
ribosomes 36. Other ribosomal proteins and glycolytic enzymes were not observed in
these high molecular weight fractions (Table 3). The only protein not belonging to either of these two groups is the cell division protein FtsZ. Cross-linking with PFA leads
to very large complexes. All of the aforementioned proteins that were pulled down
152
s
ell
xa
c
ted
N
[
[
Fi
ed
fix
on
3 hours
7 hours
0 524
PdhC (28)
PdhB (9)
YfiA (20)
RplL (12)
)
kD r
ke
tio mar
c
l fra
ro
ein
nt W
ot
o
r
C
M
P
(
ns
170+
170
100
70
60
50
40
35
30
25
18
12
rProteins
Figure 4. Identification of binding partners of YfiA. (A) After fixation with 4% PFA or
not, cells (L. lactis overexpressing Strep-tagged YfiA) grown for 7 h in GM17 at 30°C
were harvested and disrupted. The protein extracts were subjected to SDS-12% PAA and
Coomassie staining (panels 1,3 and 5, from the same gel) and Western hybridization using
anti-Strep-tag antibodies (panels 2 and 4, from one blot from a duplicate gel). After that,
gels were cut in fractions, based on their molecular weight as indicated by the numbers on
the right and proteins were identified by MS/MS. See Table 1 for the numbering system. (B)
Proteins identified in ribosome fractions after sucrose gradient ultra-centrifugation of L. lactis
wild-type cells after 3 h of growth in GM17 at 30°C (logarithmic growth phase) and 7 hours
of growth (stationary growth phase). 52 proteins were found with high significance in both
samples, and these are mainly ribosomal proteins. In the stationary phase fraction 4 unique
proteins were identified, namely YfiA, PdhC, PdhB and RplL (L7/L12), the numbers behind
the proteins indicate the spectral counts.
with Strep-YfiA in the absence of PFA appear in these complexes after cross-linking,
which can be up to and exceeding 170 kDa.
YfiA prevents ribosomal degradation
Previous studies have shown that stabilization of the 50S/30S conformation of intact
ribosomes with Mg2+ prevents their degradation by ribonuclease 7. We investigated
whether ribosomes from L. lactis are protected against ribonuclease activity when
purified Strep-YfiA is added to the reaction mixture. As can be seen in Fig. 5, elec153
Table 3. Proteins that co-purify with Strep-YfiALl. Nisin-induced L. lactis (pNZstrepyfiA),
grown to stationary phase, were treated with 4% PFA (fixated), or not (non-fixated). The
cells were disrupted and the proteins were subjected to Strep-tactin column chromatography.
Bound proteins were specifically eluted with desthiobiotin and separated by SDS 12%-PAGE
(Fig. 4B). The gel was sliced into 12 equally-sized slices containing proteins of up to 12
kDa and around 18, 25, 30, 35, 40, 50, 60, 70, 100, 170 or larger than 170 kDa. As a control,
proteins from L. lactis wild-type stationary phase cells were separated on the same SDS 12%PAA gel. Proteins were eluted and subjected to MS/MS analysis in which the proteomes of
the various samples were determined. Only co-purified proteins are shown here. A protein
was identified to co-purify with Strep-YfiALl when it was detected in a gel slice from the nonfixated cells that contained proteins at least appr 20 kDa larger than the native molecular mass
of that specific protein and not in the equivalent gel slice from the control cells.
Identified
protein
Annotation
Mass
(kDa)
YfiA
Putative sigma 54 protein
21
18
All gel slices
All gel slices
29
25; 30
40
100; 170; 170+
RpsB (S2) Ribosomal protein S2
154
Control
cells, occurrence in gel
slice
Non-fixated
cells, occurrence
in gel slice
Fixated cells,
occurrence in gel
slice
RpoA
RNA polymerase subunit α
34
35
50
170+
GapB
Glyceraldehyde 3-phosphate
dehydrogenase
36
35; 40
60
70; 100; 170;
170+
Pfk
6-phosphofructokinase
36
30
50
170; 170+
Tsf
Elongation factor Ts
37
35
50
170+
CcpA
Catabolite control protein
37
40
50
170; 170+
TufA
Elongation factor Tu
43
40
60
100; 170; 170+
FtsZ
Cell division protein
44
40
60
170+
RpsA (S1) Ribosomal protein S1
45
40
60
100; 170; 170+
Tig
Trigger factor
47
50
70
170+
PdhD
Dihydrolipoamide dehydrogenase
50
50
70
170+
PdhC
Pyruvate dehydrogenase
component
56
50; 60
100
170+
GroEL
Chaperonin
57
60
70
170+
DnaK
Chaperone protein
65
60
100
170+
GlmS
D-fructose-6-phosphate amidotransferase
66
50; 60
100
170+
TypA
GTP-binding protein
68
60
100
170+
FusA
Elongation factor G
78
70
100
170; 170+
PurL
Phosphoribosylformylglycinamidine synthase II
80
60; 70; 100
100
170+
GyrA
Gyrase subunit A
93
70
170
170+
AlaS
Alanyl-tRNA synthase
96
70
170
170+
ClpB
ATP-dependent protease
97
70
100; 170
170+
AdhE
Alcohol-acetaldehyde dehydrogenase
98
InfB
Translation initiation factor
IF-2
105
70; 100
170
170+
100
170
170+
trophoresis of the 70S ribosomal fraction leads to three bands on a standard RNA gel.
The upper fragment seems protected, but the smaller subunits 23S and 16S were not,
against degradation after incubation with Strep-YfiA. The same ribosomal fractions
without Strep-YfiA were completely degraded within 30 minutes of incubation with
RNAse (Fig. 5). Nuclease protection by Strep-YfiA is specific for ribosomal RNA,
since mRNA degradation could not be prevented (data not shown). It is tempting to
speculate that the upper band represents the 70S ribosome, representing an on-gel
disrupted ribosomal dimer and the lower bands the monomeric subunits, although this
should be verified by Northern blots.
-
+
+
-
Strep-YfiA
-
-
+
+
RNAse
High molecular weight
RNA fragment
23S rRNA
16S rRNA
Figure 5. In vitro RNA degradation assay. Isolated ribosome fractions were loaded on an RNA
gel and incubated with Strep-YfiA or TE-DEPC or RNAse.
155
Discussion
Here, we identify the protein YfiALl as an essential factor in ribosome dimerization
in L. lactis. Dimerization of ribosomes in this organism takes place when the cells
enter into stationary phase. YfiALl is not essential as cells lacking the yfiA gene are
viable. Since the yfiA mutant does not show ribosome dimers, it can be concluded
that other proteins, such as putative functional homologs of the E. coli dimerization
proteins HPR and RMF, cannot take over that function in L. lactis. Heterologously
overexpressed YfiAEc from E. coli did not rescue the yfiA mutation in L. lactis. Also,
C-terminally truncated YfiALl1-126 does not allow ribosome dimer formation. Thus, important dimerization functions in YfiALl are located in the C-terminus of the protein.
L. lactis YfiA is an example of the so-called long HPF type 10. In that respect, it resembles the SaHPF protein from S. aureus. This protein has been shown recently to
be involved in ribosome dimerization 32. Apart from the domain that is conserved in
all HPF homologs, both proteins contain an extended, long and mutually similar Cterminal domain 10. YfiAEc contains only a short C-terminal extension to its conserved
HPF-domain (Fig. 1). In E. coli, this extension of YfiAEc prevents ribosome dimerization by interfering with RMF-dependent 90S formation. In the present study we show
that the extended C-terminal domain in YfiALl performs a completely opposite function. It is crucial for dimerization of ribosomes in L. lactis during stationary phase. The
C-terminus of YfiALl functionally resembles E. coli RMF, as suggested previously for
that of SaHPF 32. Indeed the C-terminal domains of SaHPF and YfiALl show limited
homology (Fig. 1).
In a follow-up experiment, L. lactis ΔyfiA will be complemented with YfiALl1-126 in
combination with a protein fragment encompassing the C-terminal domain of YfiALl
or E. coli RMF. This should prove whether the C-terminal domain of YfiALl indeed
is a functional homolog of RMFEc. The amino acid residues at the extreme end of the
YfiALl1-126 protein could interfere with RMFEc binding, as was postulated on the basis
of the model for YfiAEc binding to the ribosome.
The relative position of the large 50S and small 30S subunits in the lactococcal ribosome monomers has changed in the dimers compared to the monomers (Fig. 3G-K).
In E. coli it has been shown that RMFEc is involved in repositioning of the 30S subunit relative to the 50S subunit. The hypothesis is that this conformational change
156
facilitates the formation of 100S ribosome dimers 2. Similarly, we propose that the Cterminus of YfiALl performs the RMFEc-like function of inducing the conformational
change, allowing ribosome dimers to be formed in L. lactis.
L. lactis ΔyfiA does not have a strong phenotype. The growth-rates of the mutant and
the wild-type strain do not show differences. Viability and re-growth were also equal
for both strains. A minor phenotype was reported for E. coli yfiA::Km. The cells of
this strain may live somewhat longer in stationary phase 1. The enhanced viability has
been explained by increased protection of 70S ribosomes and ribosome dimers against
degradation by RNA hydrolases 7,9. Ribosome dimerization is thought to allow rapid
recovery of translation by tmRNA, an RNA molecule with tRNA- and mRNA-mimicking domains. tmRNA is responsible for trans-translation, which takes place when
ribosomes are stalled on, for instance, damaged mRNA 37. Interestingly, parts of the
tmRNA complex were found in this study to co-purify with Strep-YfiALl in stationary
phase cells. Of the proteins known to be part of the tmRNA complex, EF-Tu 38 and
ribosomal protein S1 39 were indeed detected. The third and essential protein factor for
trans-translation, SmpB 40,41 was not detected. This is most likely because this protein
does not directly interact with Strep-YfiALl, but we cannot exclude that trypsin-digested SmpB is too small to allow detection by our assay. Whether and how YfiALl plays
a role in trans-translation is not known, and needs further research.
Other factors known to interact with the ribosome and involved in various aspects of
translation and ribosome recycling were co-purified with Strep-YfiALl. We identified
initiation factor 2 (InfB), a protein that allows fMet-tRNA and 30S and 50S subunits
to form a 70S ribosome 42 and elongation factor Ts, catalyzing the release of guanosine diphosphate from EF-Tu. Also, protein-folding factors such as DnaK, GroEL and
trigger factor TG were co-purified. Co-purification with Strep-YfiALl of glycolytic
enzymes could be the result of unspecific binding since these proteins are very abundant in the cytosol 43. Also in S. aureus, glycolytic enzymes were shown to sediment
together with ribosomes after a modified 2D-PAGE separation 32. More specifically, in
that study and in ours, proteins of the pyruvate dehydrogenase (PDH) complex were
identified. This co-purification of (parts of) the PDH complex with YfiALl need not
be an artifact. It has been shown previously that the so-called S-complex in B. subtilis, or a membrane-bound ribosome protein complex in S. aureus, in fact consists of
a membrane-located complex of 4 pyruvate dehydrogenase proteins in tight contact
with ribosomes. In fact, the ribosomes were shown to protect the PDH complex from
being degraded by trypsin 36. It is an intriguing possibility that, conversely, the ribo157
some dimers are protected from RNAses by their location at or near the membranelocated PDH complex. The complex would thus serve as a “parking lot” for hibernated
ribosome dimers, a speculation that warrants further research.
158
References
1.
Ueta, M. et al. Ribosome binding proteins YhbH and YfiA have opposite
functions during 100S formation in the stationary phase of Escherichia coli. Genes
Cells 10, 1103–1112 (2005).
2.
Polikanov, Y. S., Blaha, G. M. & Steitz, T. A. How hibernation factors RMF,
HPF, and YfiA turn off protein synthesis. Science 336, 915–918 (2012).
3.
Yamagishi, M. et al. Regulation of the Escherichia coli rmf gene encoding
the ribosome modulation factor: growth phase- and growth rate-dependent control.
EMBO J. 12, 625–630 (1993).
4.
Roche, E. D. & Sauer, R. T. SsrA-mediated peptide tagging caused by rare
codons and tRNA scarcity. EMBO J. 18, 4579–4589 (1999).
5.
Izutsu, K. et al. Escherichia coli ribosome-associated protein SsrA, whose
copy number increases during stationary phase. J. Bacteriol. 183, 2765 –2773
(2001).
6.
Wada, A., Igarashi, K., Yoshimura, S., Aimoto, S. & Ishihama, A. Ribosome
modulation factor: stationary growth phase-specific inhibitor of ribosome functions
from Escherichia coli. Biochem. and Biophys. Res. Comm. 214, 410–417 (1995).
7.
Zundel, M. A., Basturea, G. N. & Deutscher, M. P. Initiation of ribosome
degradation during starvation in Escherichia coli. RNA 15, 977 –983 (2009).
8.
Vila-Sanjurjo, A., Schuwirth, B.-S., Hau, C. W. & Cate, J. H. D. Structural
basis for the control of translation initiation during stress. Nat. Struct. Mol. Biol. 11,
1054–1059 (2004).
9.
Ortiz, J. O. et al. Structure of hibernating ribosomes studied by cryoelectron
tomography in vitro and in situ. J. Cell Biol. 190, 613 –621 (2010).
10. Ueta, M. et al. Role of HPF (Hibernation Promoting Factor) in translational
activity in Escherichia coli. J. Biochem. 143, 425 –433 (2008).
11.
Wegmann, U. et al. Complete genome sequence of the prototype lactic acid
bacterium Lactococcus lactis subsp. cremoris MG1363. J. Bacteriol. 189, 3256–
3270 (2007).
12. De Jong, A., Ejby, M., Kuipers, O. P., Kilstrup, M. & Kok, J. Detection of
environmental changes by Lactococcus lactis MG1363 during fermentation of milk.
PLoS ONE (2013).
13. Sambrook, J., Fritsch, E. F. & Maniatis, T. Molecular cloning: a laboratory
manual. (1989).
14. Johansen, E. & Kibenich, A. Isolation and characterization of IS1165, an insertion sequence of Leuconostoc mesenteroides subsp. cremoris and other lactic acid
bacteria. Plasmid 27, 200–206 (1992).
15. Nørholm, M. H. A mutant Pfu DNA polymerase designed for advanced uracil-excision DNA engineering. BMC Biotechnol. 10, 21 (2010).
16. Skerra, A. & Schmidt, T. G. Use of the Strep-Tag and streptavidin for detection and purification of recombinant proteins. Meth. Enzymol. 326, 271–304 (2000).
17. De Ruyter, P. G., Kuipers, O. P. & De Vos, W. M. Controlled gene expression
159
systems for Lactococcus lactis with the food-grade inducer nisin. Appl. Environ. Microbiol. 62, 3662–3667 (1996).
18. Kuipers, O. P., De Ruyter, P. G. G., Kleerebezem, M. & De Vos, W. M. Quorum sensing-controlled gene expression in lactic acid bacteria. J. Biotechnol. 64,
15–21 (1998).
19. Laemmli, U. K. Cleavage of structural proteins during the assembly of the
head of bacteriophage T4. Nature 227, 680–685 (1970).
20. Solem, C., Defoor, E., Jensen, P. R. & Martinussen, J. Plasmid pCS1966,
a new selection/counterselection tool for lactic acid bacterium strain construction
based on the orop gene, encoding an orotate transporter from Lactococcus lactis.
Appl. Environ. Microbiol. 74, 4772 –4775 (2008).
21. Simon, D. & Chopin, A. Construction of a vector plasmid family and its use
for molecular cloning in Streptococcus lactis. Biochimie 70, 559–566 (1988).
22. Holo, H. & Nes, I. F. High-frequency transformation, by electroporation, of
Lactococcus lactis subsp. cremoris grown with glycine in osmotically stabilized media. Appl. Environ. Microbiol. 55, 3119–3123 (1989).
23. Hakenbeck, R., Chhatwal, S., Prudhomme, M. & Claverys, J.-P. Molec. Biol.
of Streptococci. (2007).
24. Pinto, J. P. C. et al. pSEUDO, a genetic integration standard for Lactococcus
lactis. Appl. Environ. Microbiol. 77, 6687–6690 (2011).
25. Maki, Y., Yoshida, H. & Wada, A. Two proteins, YfiA and YhbH, associated
with resting ribosomes in stationary phase Escherichia coli. Genes Cells 5, 965–974
(2000).
26. Drop, B. et al. Photosystem I of Chlamydomonas reinhardtii contains nine
light-harvesting complexes (Lhca) located on one side of the core. J. Biol. Chem.
286, 44878–44887 (2011).
27. Aranda, P. S., LaJoie, D. M. & Jorcyk, C. L. Bleach gel: A simple agarose gel
for analyzing RNA quality. Electrophoresis 33, 366–369 (2012).
28. Oostergetel G.T., Keegstra W. & Brisson A. Automation of specimen selection and data acquisition for protein electron crystallography. Ultramicroscopy 74,
47–59 (1998).
29. Boekema, E. J., Van Roon, H., Van Breemen, J. F. & Dekker, J. P. Supramolecular organization of photosystem II and its light-harvesting antenna in partially
solubilized photosystem II membranes. Eur. J. Biochem. 266, 444–452 (1999).
30. Rauch, P. J. & De Vos, W. M. Characterization of the novel nisin-sucrose
conjugative transposon Tn5276 and its insertion in Lactococcus lactis. J. Bacteriol.
174, 1280 –1287 (1992).
31. Casadaban, M. J. & Cohen, S. N. Analysis of gene control signals by DNA
fusion and cloning in Escherichia coli. J. Molec. Biol. 138, 179–207 (1980).
32. Ueta, M., Wada, C. & Wada, A. Formation of 100S ribosomes in Staphylococcus aureus by the hibernation promoting factor homolog SaHPF. Genes Cells 15,
43–58 (2010).
160
33. Jiang, M. et al. The Escherichia coli GTPase CgtAE is involved in late steps
of large ribosome assembly. J. Bacteriol. 188, 6757–6770 (2006).
34. Sergiev, P. V. et al. How much can we learn about the function of bacterial
rRNA modification by mining large-scale experimental datasets? Nucleic Acids Res.
(2012).
35. Hemilä, H., Palva, A., Paulin, L., Arvidson, S. & Palva, I. Secretory S complex of Bacillus subtilis: sequence analysis and identity to pyruvate dehydrogenase.
J. Bacteriol. 172, 5052–5063 (1990).
36. Adler, L. A. & Arvidson, S. Correlation between the rate of exoprotein synthesis and the amount of the multiprotein complex on membrane-bound ribosomes
(MBRP-complex) in Staphylococcus aureus. J. Gen. Microbiol. 133, 803–813
(1987).
37. Keiler, K. C. Biology of trans-translation. Annu. Rev. Microbiol. 62, 133–151
(2008).
38. Rudinger-Thirion, J., Giegé, R. & Felden, B. Aminoacylated tmRNA from
Escherichia coli interacts with prokaryotic elongation factor Tu. RNA 5, 989–992
(1999).
39. Ramrath, D. J. F. et al. The complex of tmRNA-SmpB and EF-G on translocating ribosomes. Nature 485, 526–529 (2012).
40. Karzai, A. W., Susskind, M. M. & Sauer, R. T. SmpB, a unique RNA-binding
protein essential for the peptide-tagging activity of SsrA (tmRNA). EMBO J. 18,
3793–3799 (1999).
41. Weis, F. et al. tmRNA-SmpB: a journey to the centre of the bacterial ribosome. EMBO J. 29, 3810–3818 (2010).
42. Myasnikov, A. G., Simonetti, A., Marzi, S. & Klaholz, B. P. Structure–function insights into prokaryotic and eukaryotic translation initiation. Curr. Opin. Struc.
Biol. 19, 300–309 (2009).
43. Steen, A., Wiederhold, E., Gandhi, T., Breitling, R. & Slotboom, D. J. Physiological adaptation of the bacterium Lactococcus lactis in response to the production
of human CFTR. Molec. Cell. Proteomics 10, (2011).
161
162
Chapter 6
Summary and conclusion
163
This thesis describes a systems biology approach to explain the fermentative behavior
of Lactococcus lactis, grown at specific varying growth rates. Different -omics technologies were accessible by collaboration between three research groups in The Netherlands. The coworkers of the Vrije Universiteit of Amsterdam provided the chemostat
setup and performed the metabolomics experiments. Our team members of Membrane
Enzymology at Rijksuniversiteit Groningen provided the expertise for proteomics,
and our group Molecular Genetics at Rijksuniversiteit Groningen was responsible
for performing genomics and transcriptomics studies and bioinformatics analyses.
Technology Foundation Stichting Toegepaste Wetenschappen (STW) funded this project. Its aim was to stimulate the transfer of knowledge between technical sciences and industry. The fundamental knowledge derived from this research is integrated into a cellular model of L. lactis that will be able to predict fermentative behavior of this industrially
very important microorganism. Our multi-omics data obtained from L. lactis growing at
varying preset growth rates provides the fundament for this model. From an industrial
perspective, such model is useful for further optimization of L. lactis as a microbial cell
factory and for choosing the best growth conditions, while retaining desired properties.
When cells of L. lactis were grown at varying growth rates in chemostats, under
glucose limiting conditions, the bacteria employ a metabolic shift from mixed-acid
to homolactic fermentation. L. lactis mainly produces formate, acetate and ethanol
(mixed-acid) at low growth rates, while at high growth rates the major end-product
of glycolysis is lactate (homolactic fermentation). Researchers have been trying to
understand for a long time why L. lactis employs this metabolic shift 1–7. The hypothesis for our study was that the shift from mixed-acid to homolactic fermentation
is an outcome of evolutionary optimization of resource allocation, and is based on
the predictions of the self-replicator model, which presumes that there is a tradeoff
between investments in enzyme synthesis and metabolic yields for alternative catabolic pathways 8. Most of the proteome in a lactococcal cell consists of glycolytic
and ribosomal proteins. How much a cell invests in the different cellular ‘modules’
is defined as the protein investment and is, for an important part, determined by the
transcriptional activity of its genes. A primary aim of our study was to characterize the
protein investments in the different cellular modules at different growth rates (Fig. 1).
In Chapter 3, a comprehensive and high-quality dataset of mRNA and protein ratios
and enzyme activities of most of the glycolytic enzymes is presented. The data from
our study shows that the transcription of the genes encoding glycolytic enzymes re164
mains equal with increasing growth rate. Also the protein levels and the kinetics of
the glycolytic pathway enzymes hardly change when the growth rate changes. Yet, the
metabolic flux data shows that under the growth rate-controlled conditions examined,
a metabolic shift from mixed-acid to homolactic fermentation is observed. Thus, the
theory of protein investment as a means to optimize the expressed proteins to the
growth rate, does not explain the metabolic shift between mixed-acid and homolactic
fermentation in L. lactis. The strain of L. lactis used in this study, MG1363, keeps a
large majority of the glycolytic enzymes at a basal level to be prepared for possibly
changing conditions.
The change in maximal catalytic capacity of a subset of glycolytic enzymes only
partly explains how L. lactis shifts from mixed-acid to homolactic fermentation. In
our experimental setup we determined enzyme kinetics in vitro. Even though the
assay buffer mimicked the in vivo composition of the lactococcal cytoplasm, the
kinetics of glycolytic enzymes should ideally be determined in vivo. Next to that,
in order to point out what causes allosteric regulation, data of all allosteric modifications onto the glycolytic enzymes should be collected, as is recently attempted
in the PHOSIDA database 9. In our study we did not determine effectors, known to
influence the metabolic shift, like the NADH/NAD+ ratios 2, fructose-bisphosphate
10
, or the phosphorylation state of HPr 11. To increase completeness of the systems
biology analysis, information of all intermediate metabolites should be determined.
In vivo NMR is a useful technique for tracking glycolytic intermediates 6. However, the relatively low sensitivity of in vivo NMR would require much thicker cellsuspensions than what we grew in chemostats. Alternatively, metabolites can be
measured in cell extracts obtained using fast sampling and quenching methods.
Transcriptional changes in L. lactis under varying growth rates
In this thesis, a main focus was on the determination of the transcript level changes in
cells of L. lactis when comparing four different growth rates. We present a method in
Chapter 2, using the Limma package of R 12, which combines both direct and indirect
comparisons of the transcriptomes at the different growth rates. This scheme improved
the significance of all comparisons, thus enabling to detect very small changes in the
transcriptomes. A thorough analysis of the transcriptome dataset revealed that mRNA
levels of genes coding for the glycolytic enzymes hardly alter of cells growing at different growth rates in chemostat cultures.
165
Chapter3
Chapter4
FABenzymes
Ribosome
Glycoly�c
enzymes
Metabolites
Energy
Chapter5
Endproducts
3´
5´
5´
3´
5´
3´
Transcriptome
Chapter2
Figure 1. Subject of study of the various chapters in this thesis (coloured squares) projected
onto the modular model of L. lactis. In Chapter 2, transcriptomics was performed, as part of
the multi-omics study of L. lactis growing at different growth rates presented in Chapter 3.
Regulation of fatty acid biosynthesis is detailed in Chapter 4. In Chapter 5, a characterization
is given of YfiA, a protein essential for dimerization of ribosomes.
The transcriptional activities of genes of other important modules of the model were
also characterized. Most of the genes encoding the ribosomal proteins show a very
small increase with increasing growth rate. A non-linear correlation with the growth
rate is seen for the transcription of the arc-genes, coding for the arginine catabolic
ADI-pathway. Between growth rates of 0.15 h-1 and 0.5 h-1, an increase in transcriptional activity of the arginine catabolism genes was followed by a steep decrease at
0.6 h-1. Other important modules that were expected to be changed in the protein investment of the cell through changes in the transcriptional activity of the involved
genes are fatty acid biosynthesis, membrane transport and cell division. We did not
discern any general trends in the transcription of genes of those modules in response
to an increase in the growth rate. All together we conclude that if the transcriptional
activity of genes changes in response to growth rate, these changes are most often
166
very modest. The concentrations of proteins in the core proteome (e.g. those involved
in metabolism, ribosomal proteins, etc.) are thereby less dependent on the change in
transcript abundance.
Ribosomal composition at different growth rates
Ribosomes, throughout all kingdoms of life, are composed of two very important molecules: RNA and protein. As our data shows, the composition of a ribosome changes
in response to the growth rate. Our experiments in which we determined the increase
of totRNA/totProt and of rProteins hint that at increasing growth rates rRNA is less
occupied by rProteins. Not all rProteins are functionally characterized, making it difficult to explain why some rProteins numbers per cell increase with the growth rate and
why some others do not. The small subunit rProteins S1, S2, S5, S9, S19, S20, S21 and
large subunit rProteins L7/L12, L10 and L12 do not respond to growth rate changes.
None of the genes coding for these rProteins are located in llmg2370-llmg2380, the
only cluster of genes that showed decreased transcription upon increasing growth rate.
Many of the rProteins are not functionally characterized; they are thought to operate as
a ‘ribosomal glue’13. It is tempting to speculate that, (1) a subset of rProteins is minimally required for a ribosome to function as a ribozyme 14. Another speculation may
be that, (2) each ribosome has a different set of rProteins to stabilize it. Alternatively,
(3) rRNA is synthesized to higher levels than needed at high growth rates. At lower
growth rates, excessive rRNA is degraded and used as an energy source. At least the
first hypothesis can be tested by isolating ribosomes from cells growing at different
growth rates and examining the rRNA concentration in combination with quantitative
MS/MS to reveal the identity and relative abundance of each ribosomal protein per
ribosome.
The growth rate-adaptivity of rRNA and rProteins shows that at lower growth rates
a minimal number of ribosomes per cell must be present. While at low growth rates,
or after entering stationary phase, E. coli ribosomes are more prone to degradation
by ribonucleases 15. In Chapter 5 we describe the functional properties of YfiA, a
protein that prevents ribosomes from being degraded by ribonucleases. It does so by
binding 70S ribosomes and dimerizing them. The research on L. lactis YfiA started by
observing a significant decrease in yfiA transcription at high growth rates (Chapter 3,
supplementary Table S1). The yfiA gene of L. lactis is annotated as a protein capable
of binding both protein and RNA 16. We concluded that, similar to YfiA of E. coli,
167
YfiAEc 17, L. lactis YfiALl might play a role in ribosome binding. The E. coli proteins
RMF and HPF stabilize the large and small ribosomal subunits and link two 30S subunits of two ribosomes together to form a ribosome dimer 18. We show here that YfiALl performs an opposite function in L. lactis: in contrast to YfiAEc, YfiALl stimulates
dimerization of ribosomes. YfiALl contains a long C-terminal tail that is not present
in YfiAEc. By performing a deletion analysis on YfiALl we prove that this C-terminal
extension is essential for ribosome dimer formation in L. lactis. All sequenced members of the family of Streptococcaceae contain YfiA homologs that are conserved in
both their N-terminal domains, which have strong homology to the HPF protein of E.
coli, and the extended C-terminal domains. The same organisms all lack proteins that
resemble HPF and/or RMF of E. coli. In Staphylococcus aureus, the protein SaHPF
is responsible for ribosome dimerization 20. The amino acid sequences of SaHPF and
YfiALl are very similar; both contain the extended C-terminus (Chapter 5, Fig. 1). The
results from our work and that of others suggests that bacteria can either dimerize ribosomes in two steps using HPF and RMF or in a single step employing a single protein
that contains a C-terminal extension as is present in YfiALl and SaHPF. In order to
prove the interchangeability of the C-terminal domain of YfiALl with RMF of E. coli,
a complementation of YfiALl1-126 with the latter protein should be made and tested for
ribosome dimerization.
Fatty acid biosynthesis and regulation
Regulation of fatty acid biosynthesis (FAB) in S. pneumoniae and E. feacalis has been
well studied and served as a basis for our studies 21–23. In Chapter 4, a reconstruction is
made of all enzymes required for FAB in L. lactis. Next to that, we show that FabT is a
dedicated regulator for FAB in L. lactis and characterized the location of binding sites
of FabT, both by bioinformatics techniques and with EMSA and DNAseI footprinting.
When cultures of L. lactis are grown at different growth rates, the acyl chain composition of phospholipid membranes changes accordingly (Chapter 3, Fig. 6). How L.
lactis alters the balance between saturated and unsaturated acyl chains in the membrane has not been solved yet. Dehydratase FabZ1 of L. lactis could function as an
isomerase, like FabN in E. faecalis 21. The substrate specificity of E. faecalis FabN is
less constrained than that of dehydratase FabZ of E. faecalis, since the positioning of
central helix α-3 is altered due to the specific placement of β3 and β4 sheets 21. The
crucial β3 and β4 sheets of L. lactis FabZ1 are mostly similar to those β-sheets of E.
168
faecalis FabN. By employing two FabZ dehydratase variants, one of which is capable
of isomerizing acyl chains, E. faecalis can regulate the saturation of acyl chains in the
phospholipid membrane 21. Based on the synteny of the genes involved and the structural similarities between L. lactis FabZ1 and E. faecalis FabN and L. lactis FabZ2 and
E. faecalis FabZ, it is likely that L. lactis regulates acyl chain saturation in the same
way.
169
References
1. Thomas, T. D., Ellwood, D. C. & Longyear, V. M. C. Change from homo- to heterolactic fermentation by Streptococcus lactis resulting from glucose limitation
in anaerobic chemostat cultures. J. Bacteriol. 138, 109–117 (1979).
2. Garrigues, C., Loubiere, P., Lindley, N. D. & Cocaign-Bousquet, M. Control of
the shift from homolactic acid to mixed-acid fermentation in Lactococcus lactis: predominant role of the NADH/NAD+ ratio. J. Bacteriol. 179, 5282–5287
(1997).
3. Neves, A. R. et al. Metabolic characterization of Lactococcus lactis deficient in
lactate dehydrogenase using in vivo 13C-NMR. Eur. J. Biochem. 267, 3859–3868
(2000).
4. Koebmann, B. J., Andersen, H. W., Solem, C. & Jensen, P. R. Experimental determination of control of glycolysis in Lactococcus lactis. Antonie Van Leeuwenhoek 82, 237–248 (2002).
5. Melchiorsen, C. R., Jokumsen, K. V., Villadsen, J., Israelsen, H. & Arnau, J. The
level of pyruvate-formate lyase controls the shift from homolactic to mixed-acid
product formation in Lactococcus lactis. Appl. Microbiol. Biotechnol. 58, 338–
344 (2002).
6. Neves, A. R., Pool, W. A., Kok, J., Kuipers, O. P. & Santos, H. Overview on
sugar metabolism and its control in Lactococcus lactis - the input from in vivo
NMR. FEMS Microbiol. Rev. 29, 531–554 (2005).
7. Voit, E. O. et al. Regulation of glycolysis in Lactococcus lactis: an unfinished
systems biological case study. Syst. Biol. 153, 286–298 (2006).
8. Molenaar, D., Van Berlo, R., De Ridder, D. & Teusink, B. Shifts in growth strategies reflect tradeoffs in cellular economics. Mol. Syst. Biol. 5, 323 (2009).
9. Gnad, F., Gunawardena, J. & Mann, M. PHOSIDA 2011: The posttranslational
modification database. Nucleic Acids Res. 39, D253–260 (2011).
10.Wolin, M. J. Fructose-1,6-diphosphate requirement of Streptococcal lactic dehydrogenases. Science 146, 775–777 (1964).
11. Gunnewijk, M. G. & Poolman, B. Phosphorylation state of HPr determines the
level of expression and the extent of phosphorylation of the lactose transport protein of Streptococcus thermophilus. J. Biol. Chem. 275, 34073–34079 (2000).
12.Smyth, G. Bioinformatics and computational biology solutions using R and bioconductor (Gentleman, R., Carey, V. J., Huber, W., Irizarry, R. A. & Dudoit, S.)
397–420 (2005).
13.Brodersen, D. E. & Nissen, P. The social life of ribosomal proteins. FEBS J. 272,
2098–2108 (2005).
14.Nierhaus, K. H. The assembly of prokaryotic ribosomes. Biochimie 73, 739–755
(1991).
15.Zundel, M. A., Basturea, G. N. & Deutscher, M. P. Initiation of ribosome degradation during starvation in Escherichia coli. RNA 15, 977 –983 (2009).
16.Wegmann, U. et al. Complete genome sequence of the prototype lactic acid
170
bacterium Lactococcus lactis subsp. cremoris MG1363. J. Bacteriol. 189, 3256–
3270 (2007).
17.Maki, Y., Yoshida, H. & Wada, A. Two proteins, YfiA and YhbH, associated with
resting ribosomes in stationary phase Escherichia coli. Genes Cells 5, 965–974
(2000).
18.Polikanov, Y. S., Blaha, G. M. & Steitz, T. A. How hibernation factors RMF,
HPF, and YfiA turn off protein synthesis. Science 336, 915–918 (2012).
19.Ueta, M. et al. Ribosome binding proteins YhbH and YfiA have opposite functions during 100S formation in the stationary phase of Escherichia coli. Genes
Cells 10, 1103–1112 (2005).
20.Ueta, M., Wada, C. & Wada, A. Formation of 100S ribosomes in Staphylococcus aureus by the hibernation promoting factor homolog SaHPF. Genes Cells 15,
43–58 (2010).
21.Lu, Y.-J., White, S. W. & Rock, C. O. Domain swapping between Enterococcus
faecalis FabN and FabZ proteins localizes the structural determinants for isomerase activity. J. Biol. Chem. 280, 30342 –30348 (2005).
22.Lu, Y.-J. & Rock, C. O. Transcriptional regulation of fatty acid biosynthesis in
Streptococcus pneumoniae. Mol. Microbiol. 59, 551–566 (2006).
23.Jerga, A. & Rock, C. O. Acyl-acyl carrier protein regulates transcription of fatty
acid biosynthetic genes via the FabT repressor in Streptococcus pneumoniae. J.
Biol. Chem. 284, 15364–15368 (2009).
171
172
Chapter 7
Addendum
173
Table A1. Transcriptome analysis of L. lactis MG1363 at varying growth rates belonging to
Chapter 2 and 3. All genes are sorted on the COG-category and within the COG-category
sorted on their accession number. Fold-changes are obtained by comparing with 0.15 h-1 as
common reference.
CLASS
0.3
P-value
0.5
P-value
3.7E-02
-1.9
7.6E-03
0.6
P-value
-2.7
-3.8
-1.5
2.1E-02
-1.4
1.6E-02
0.5-0.3
P-value
0.6-0.3
P-value
0.6-0.5
P-value
6.7E-05
-2.3
1.2E-03
-2.2
1.7E-03
1.2E-07
-2.8
3.7E-05
-2.4
3.2E-04
-1.4
2.8E-02
[C] Energy production and conversion
174
llmg_0071
pdhD
llmg_0072
pdhC
llmg_0195
llmg0195
-1.7
llmg_0245
llmg0245
llmg_0392
ldhB
-1.6
llmg_0447
nifJ
llmg_0475
llmg0475
llmg_0559
llmg0559
1.4
llmg_0568
llmg0568
llmg_0629
pfl
llmg_0634
pycA
2.6E-04
-1.4
2.9E-02
-1.9
2.3E-02
1.0E-02
1.5
2.9E-03
1.7
4.8E-04
-1.5
3.3E-02
-1.7
9.4E-03
1.5
7.7E-03
1.6
1.0E-02
1.4
4.0E-02
citB
llmg_0637
icd
-1.3
4.6E-02
1.8
2.5E-04
llmg_0730
fer
-1.3
2.7E-02
-1.5
5.5E-03
llmg_0763
pta
llmg_1028
llmg1028
llmg_1098
glpD
2.3
1.3E-03
llmg_1099
glpK
1.6
1.2E-02
llmg_1114
gpsA
llmg_1120
ldh
llmg_1429
ldhX
llmg_1441
frdC
1.6
5.4E-04
1.6
4.5E-04
llmg_1637
mleP
1.4
3.4E-03
1.3
1.5E-02
2.0
2.9E-03
1.8
2.3E-02
llmg_1638
mleS
llmg_1734
noxB
llmg_1735
noxA
llmg_1807
llmg1807
2.5
-1.4
qor
-1.4
1.5E-02
-1.5
5.4E-03
1.5
5.7E-03
1.7
8.3E-04
llmg_1863
cydB
-1.4
4.5E-02
-1.6
5.3E-03
llmg_1864
cydA
llmg1904
llmg_1916
llmg1916
llmg_1948
atpA
1.3
3.2E-02
1.0E-02
4.1
2.4
1.5
2.0E-02
-1.3
3.4E-02
2.1
1.6
-1.7
1.3E-03
1.5
4.2E-03
1.3E-02
-3.3
6.2E-05
-3.8
9.1E-07
1.2E-02
-1.4
2.4E-02
-1.6
1.7E-02
1.5
1.7E-02
1.5
3.2E-02
1.8
1.6E-02
2.5
2.8E-04
2.6
2.5E-04
3.2
1.3E-04
2.0
1.9E-03
2.1
6.2E-04
-1.5
1.7E-02
-1.5
5.7E-03
-3.5
1.5E-03
-2.9
8.0E-03
1.8
4.1E-03
1.5
2.6E-02
1.9
8.9E-05
1.6
5.6E-03
-1.8
3.3E-03
-1.6
4.8E-03
-1.9
9.3E-05
2.4E-02
-2.3
3.4E-02
-4.2
1.8E-04
atpF
1.6
1.0E-02
1.5
3.7E-02
atpB
1.5
3.1E-02
1.8
2.6E-03
llmg_1970
nifU
1.5
3.0E-04
1.4
3.8E-03
1.7
2.6E-03
1.4
5.0E-02
ppa
1.7
1.3E-03
llmg_2288
ackA2
1.7
7.5E-03
2.7E-02
1.5E-02
8.5E-05
-1.5
llmg2272
1.7
-1.9
4.8E-08
2.8E-04
llmg_2272
2.8E-02
4.3E-02
-1.8
llmg_1996
-1.5
-1.3
llmg_1950
1.1E-02
2.4E-02
1.8E-02
9.8E-03
llmg_1951
1.3
-1.5
-1.5
-1.4
1.0E-02
cydD
llmg1915
-1.6
-1.4
llmg_1861
llmg_1915
1.3E-03
4.2E-04
llmg_1850
llmg_1904
-2.6
4.7E-03
2.6E-02
4.5E-02
-1.6
llmg_0636
-1.4
-1.3
3.1E-02
-1.5
7.2E-03
1.6
4.2E-03
1.7
3.3E-02
1.4
2.1E-02
1.3
6.7E-03
1.4
2.2E-02
1.6
1.3E-02
llmg_2289
ackA1
2.0
1.5E-02
2.2
3.0E-03
6.6
1.3E-11
llmg_2432
adhE
2.3
1.8E-04
1.7
1.9E-03
-1.1
1.5E-02
6.6E-03
1.9
7.5E-03
1.4
2.9E-02
-2.3
1.8E-03
-1.3
4.2E-02
4.7
1.7E-08
4.3
4.3E-08
-4.5
2.5E-08
-2.6
3.1E-04
1.3
3.2E-02
[D] Cell cycle control, cell division, chromosome partitioning
llmg_0016
llmg0016
llmg_0019
mesJ
llmg_0284
llmg0284
llmg_0633
ftsW1
1.9
1.5
1.9E-02
1.4
3.5E-03
llmg_0766
ftsK
1.7
5.8E-03
2.6
6.6E-06
llmg_0769
llmg0769
1.7
4.8E-02
2.8
1.0E-03
1.8
2.8E-02
llmg_1409
orf3
-1.5
3.2E-03
-1.3
3.3E-02
1.6
1.3E-02
llmg_1545
ftsX
-1.7
9.9E-04
-1.6
9.6E-03
llmg_1546
ftsE
-1.4
2.5E-02
-1.4
4.1E-02
llmg_1643
rodA
1.6
1.9E-02
1.5
2.9E-02
llmg_1680
llmg1680
-1.5
8.3E-03
2.0
9.2E-04
-1.6
9.8E-03
2.0
1.0E-03
1.8
3.4E-03
3.0
4.6E-04
2.4
8.1E-03
1.5
1.5E-02
2.1
1.2E-05
1.4
4.2E-02
1.7
1.3E-02
2.2
3.3E-03
2.0
1.1E-03
-1.5
1.9E-02
llmg_1752
smc
-1.9
4.3E-04
-1.6
1.1E-02
-1.5
1.5E-02
llmg_2035
gidA
2.9
3.7E-05
3.3
1.0E-05
2.2
3.3E-03
2.4
1.2E-03
llmg_2060
ftsZ
1.4
1.7E-02
llmg_2428
ezrA
7.7E-03
-2.2
3.2E-03
-1.8
2.9E-02
1.7
5.1E-03
1.5
4.0E-02
1.7
6.5E-03
1.6
3.0E-02
1.7
6.6E-03
1.6
3.1E-02
-1.5
2.0E-02
-1.6
1.3E-03
-1.3
3.6E-02
1.7
7.8E-03
1.6
1.2E-02
-1.5
3.4E-02
2.2
1.3E-03
2.1
1.6E-03
3.3
1.6E-06
1.9
9.3E-03
2.0
3.3E-03
4.5
6.9E-07
4.2
3.3E-06
3.5
2.7E-05
-1.6
1.8E-02
-1.8
3.5E-04
-2.7
1.4E-05
-2.0
1.3E-03
-1.4
1.3E-02
-1.8
[E] Amino acid transport and metabolism
llmg_0066
araT
llmg_0091
cysD
llmg_0096
llmg0096
llmg_0118
ctrA
llmg_0125
aroF
llmg_0138
argG
-1.9
6.2E-04
-1.7
6.8E-03
-1.6
4.4E-03
-1.8
3.1E-04
1.5
2.4E-02
llmg_0139
argH
llmg_0171
aspC
-2.4
1.3E-04
llmg_0184
llmg0184
-1.4
3.8E-02
llmg_0238
llmg0238
-1.6
1.8E-02
-1.7
1.1E-02
-1.7
1.0E-02
llmg_0291
dapD
-3.2
3.0E-04
-3.6
4.0E-05
-2.3
3.1E-03
1.8
3.3E-02
1.9
1.2E-02
3.0
9.3E-05
2.5
8.5E-04
1.4
4.0E-02
llmg_0319
pepN
1.4
1.6E-02
llmg_0362
dppA
-1.3
4.2E-02
llmg_0366
dppD
llmg_0372
asnB
-1.3
2.7E-02
-1.3
6.0E-03
llmg_0375
llmg0375
2.8
8.4E-06
2.6
2.5E-05
llmg_0376
llmg0376
2.3
3.1E-03
2.7
2.2E-04
5.0
8.9E-09
llmg_0386
lysQ
1.5
1.0E-02
1.8
7.0E-05
1.8
1.9E-04
1.4
2.3E-02
llmg_0466
aspC
-1.7
7.3E-03
-2.5
1.5E-07
-1.8
5.3E-04
-1.7
1.6E-03
1.5
1.8E-02
llmg_0493
llmg0493
1.5
3.7E-02
1.4
4.8E-02
-1.7
6.6E-04
1.7
6.6E-04
-1.5
4.5E-02
llmg_0505
nifZ
llmg_0507
llmg0507
1.4
3.8E-02
-1.5
1.6
2.1E-03
llmg_0536
argE
-1.6
2.7E-02
1.3
3.4
4.8E-02
1.6E-07
9.8E-04
llmg_0563
glyA
llmg_0567
serB
1.5
2.1E-02
1.8
1.9E-03
2.6
llmg_0650
brnQ
1.8
7.0E-04
1.8
1.3E-03
llmg_0700
llmg_0701
llmg_0774
pepQ
3.0E-04
2.3
9.1E-04
1.3E-02
-1.5
3.3E-02
-1.5
1.2E-02
7.8E-03
oppC
-2.2
1.1E-03
-2.2
7.5E-04
-2.3
3.8E-04
oppA
-1.9
5.4E-03
-2.3
3.8E-04
-2.3
3.1E-04
2.3
4.9E-04
2.4
1.9E-04
5.2E-03
2.0
1.9E-02
2.1E-04
1.6
1.9
-1.4
2.7
-1.6
2.0E-03
175
llmg_0782
aroD
llmg_0871
asd
llmg_0874
dapA
llmg_0885
proB
llmg_0886
proA
1.9
llmg_0894
carA
2.0
llmg_0911
glnB
llmg_0940
dapB
llmg_0943
pepDB
llmg_0955
llmg0955
llmg_1030
trpE
llmg_1038
trpC
1.3
1.4
2.0
7.1E-03
2.6
3.5E-04
1.5
4.3E-03
1.5
3.7E-03
1.3
2.8E-02
4.7E-03
2.0
2.2E-03
1.7E-03
2.1
2.6E-04
3.8E-02
-1.3
2.8
1.5E-04
1.3
2.7E-02
1.3
2.4E-02
1.3
1.7E-02
1.3
1.0E-02
1.6
3.9E-03
2.9E-02
3.3E-02
1.9
trpB
1.5
2.2E-02
trpA
-1.5
3.1E-03
-1.4
llmg_1048
busAA
1.8
1.8E-03
2.2
2.4E-05
2.0
3.7E-04
llmg_1049
busAB
2.5
1.3E-04
3.3
6.0E-07
3.2
3.5E-06
1.6
1.8E-02
2.0
1.4E-02
llmg_1077
llmg1077
1.6
1.6E-02
llmg_1149
ocd
-1.3
1.8E-02
llmg_1151
llmg1151
1.5
9.7E-03
1.5
1.2E-02
llmg_1153
pabA
llmg_1154
pabB
1.4
3.8E-03
1.5
7.1E-04
llmg_1179
gadB
-1.3
3.2E-02
llmg_1183
gltB
llmg_1185
lysA
1.8
6.8E-03
llmg_1200
llmg1200
-1.4
6.5E-03
1.6
1.5
1.8E-02
1.7
1.3E-04
1.7
1.5E-03
1.5
4.0E-02
1.5
7.0E-03
1.8
5.1E-03
1.3
4.2E-02
1.7
5.9E-03
3.5
3.0E-07
3.0
6.0E-06
-1.3
4.4E-02
-1.4
3.1E-02
1.7E-03
1.5
6.5E-03
1.4
3.3E-02
1.6
7.8E-03
llmg_1279
ilvB
1.8
5.4E-04
2.0
3.4E-05
llmg_1284
leuC
1.5
1.0E-02
1.7
3.3E-03
llmg_1290
hisF
llmg_1291
hisA
2.0
1.4E-02
llmg_1292
hisH
1.6
1.1E-02
llmg_1294
hisB
1.7
1.4
1.2E-02
1.6
1.4E-02
-1.3
2.2E-02
1.6
-1.7
6.4E-04
2.3E-03
2.7
2.2E-05
1.1E-02
-1.5
1.2E-02
3.4E-02
-1.6
1.9E-02
-1.7
1.7E-02
-1.4
3.9E-02
-2.5
5.5E-04
-1.9
1.2E-02
-1.6
9.5E-03
-1.5
1.4E-02
-1.9
2.2E-03
-1.6
3.0E-02
-1.4
4.0E-02
llmg_1295
hisD
llmg_1296
hisG
-1.6
2.3E-02
llmg_1297
hisZ
-1.7
1.4E-02
llmg_1298
hisC
llmg_1309
als
-2.1
4.2E-03
-2.6
1.7E-04
-2.0
7.5E-03
llmg_1325
potD
1.8
1.2E-02
-1.6
4.5E-02
-2.4
1.8E-04
-1.7
2.2E-02
llmg_1326
potC
1.4
9.6E-03
-1.3
3.6E-02
-1.4
2.1E-02
1.6
8.8E-03
1.3
1.2E-02
1.6
6.0E-03
1.6
1.5E-02
1.6
3.4E-03
1.8
5.6E-04
1.5
2.0E-02
1.7
2.1E-02
1.4
4.0E-02
llmg_1327
potB
llmg_1328
potA
llmg_1331
thrB
1.4
1.5
2.8E-02
llmg_1042
ilvA
1.7E-02
3.1E-03
4.4E-02
llmg_1041
llmg_1276
-1.6
2.8E-02
1.5
-1.5
3.1E-03
1.5E-02
llmg_1227
176
1.6
2.2
3.5E-02
-2.8
5.0E-05
-2.1
-4.3
5.2E-04
2.0E-08
llmg_1452
llmg1452
1.7
1.8E-02
1.8
4.2E-03
llmg_1458
llmg1458
-1.3
2.4E-02
-1.5
1.2E-02
-1.7
5.8E-04
llmg_1461
llmg1461
-1.4
4.2E-02
1.5
3.9E-02
llmg_1591
llmg1591
llmg_1593
llmg1593
-1.3
2.2E-02
2.1
llmg_1609
llmg1609
-1.4
1.4E-02
-1.5
3.1E-03
-1.3
9.0E-04
3.3E-02
2.4
8.0E-05
2.1
7.4E-04
1.5
1.2E-03
1.3
3.6E-02
llmg_1642
butB
llmg_1700
choQ
-1.8
llmg_1706
pepV
llmg_1732
sdaA
1.6
8.5E-04
1.6
llmg_1743
prsA
-1.3
3.5E-02
-1.5
1.2E-03
llmg_1758
argC
-1.4
1.5E-02
-1.6
3.5E-03
1.7
1.4
1.3E-02
6.5E-04
-2.1
1.4
2.8E-02
llmg_1776
metC
llmg_1820
thrA
llmg_1849
metE2
-1.5
6.5E-03
llmg_1857
llmg1857
1.3
1.6E-02
llmg_1865
dtpT
llmg_1881
pepP
-1.4
4.7E-03
-1.4
llmg_1909
pepF
1.4
2.9E-02
llmg_1924
pheA
1.4
2.8E-02
-1.4
4.4E-02
1.3
4.0E-02
-1.4
4.6E-02
4.0E-03
3.3
3.1E-05
6.9E-03
-1.3
2.9E-02
-1.1
1.4E-02
-1.4
2.6E-02
1.6
8.8E-04
1.4
3.5E-02
-1.4
1.1E-02
-1.4
7.6E-03
2.5
1.4E-03
-1.4
1.8E-02
aroK
-1.5
5.5E-03
-1.6
1.3E-02
llmg_1927
tyrA
1.8
1.7E-04
1.7
1.7E-03
llmg_1934
aroC
1.4
3.1E-02
-1.4
5.1E-03
-1.4
1.7E-02
llmg_1938
aroB
1.9
8.7E-04
-1.3
4.1E-02
-1.6
1.4E-03
1.4
2.0E-02
2.9E-02
1.8
5.8E-03
llmg_1939
aroE
1.1
1.5E-02
-1.3
3.4E-02
llmg_1943
glnQ
1.3
2.1E-02
1.5
1.5E-02
1.5
1.8E-03
llmg_1972
llmg1972
1.4
2.0E-02
1.4
1.6E-02
1.6
6.8E-04
llmg_1978
gltQ
1.7
5.1E-03
llmg_1979
gltP
1.5
7.7E-04
llmg_1993
llmg1993
llmg_1994
pepT
llmg_2011
llmg2011
llmg_2024
oppA2
llmg_2026
oppB2
1.6
-1.7
9.5E-03
6.6E-04
-1.5
7.9E-04
1.4
2.2E-02
1.4
3.6E-02
1.6
2.1E-02
-1.5
4.3E-02
-1.3
4.0E-02
2.5
1.5
1.5
-1.5
1.7
1.3E-02
-1.6
2.5E-03
1.4
5.1E-03
1.6E-02
llmg_1925
1.3
-1.8
1.3E-03
-1.5
1.4
1.8E-03
3.3E-02
4.8E-03
2.1E-02
llmg_2046
prsB
2.0
4.1E-03
llmg_2048
nifS
1.5
3.8E-02
1.5
2.0E-02
-1.6
1.4E-02
1.9
8.2E-04
llmg_2064
llmg2064
1.0
3.5E-02
llmg_2069
pepC
2.1
2.4E-02
4.0
1.1E-05
llmg_2077
proC
1.8
4.9E-04
1.4
3.3E-02
1.4
3.4E-02
-1.5
7.4E-03
-1.3
4.3E-02
2.9
2.2E-04
-1.5
9.3E-03
-1.5
3.2E-03
1.7
4.9E-03
1.6
1.3E-02
-2.3
4.8E-05
-1.7
5.8E-03
9.3E-05
1.8
4.5E-02
-1.8
2.1E-02
2.0E-02
2.0
3.8E-04
3.5E-04
1.3
2.3E-02
3.8
2.3E-05
2.7
2.0E-03
-1.4
3.2E-02
llmg_2182
metA
1.4
2.3E-02
1.1
1.4E-02
-1.5
1.5E-03
llmg_2307
arcD2
1.5
3.2E-02
1.5
2.4E-02
1.5
4.5E-02
llmg_2308
arcT
-1.7
3.8E-02
-4.1
1.3E-07
-3.5
1.6E-06
llmg_2309
arcC2
1.9
1.9E-02
-2.1
1.6E-02
-3.3
2.9E-05
llmg_2310
arcC1
1.8
2.3E-02
2.3
1.8E-03
-2.1
1.0E-02
-2.7
2.9E-04
llmg_2311
arcD1
1.8
2.1E-02
2.7
4.3E-05
-1.8
2.5E-02
-2.7
5.1E-05
llmg_2312
arcB
-7.0
6.7E-06
-9.7
2.9E-08
llmg_2313
arcA
5.8E-03
-2.7
7.2E-03
-5.2
1.1E-06
llmg_2330
llmg2330
1.6
6.5E-04
llmg_2345
llmg2345
llmg_2387
thrC
llmg_2477
llmg2477
llmg_2484
glnA
llmg_2506
asnH
3.5
1.4
3.7E-03
-1.3
-2.3
3.3E-04
1.2E-04
-4.7
6.5E-09
-2.4
4.2E-03
1.8
1.5E-02
-6.7
1.1E-05
-2.0
3.0E-02
1.4
1.7E-02
-1.4
2.1E-02
-1.4
2.7E-02
1.8
6.6E-05
-1.3
1.7E-02
1.4
1.3E-02
-1.8
6.6E-03
-1.8
8.6E-03
2.7
3.1E-02
-2.5
3.2E-05
1.6
1.0E-02
1.7
2.3E-02
-1.6
1.5E-02
177
[F] Nucleotide transport and metabolism
llmg_0020
hpt
1.5
3.0E-03
llmg_0191
dut
-1.4
8.8E-03
-1.4
1.2E-02
1.3
llmg_0192
llmg0192
1.6
7.0E-03
1.5
1.5E-02
-1.7
1.2E-03
llmg_0230
guaB
1.6
4.7E-04
9.2E-03
llmg_0232
llmg0232
1.6
1.7E-03
llmg_0281
llmg0281
1.4
9.7E-03
1.6
1.5E-03
1.6
5.2E-03
1.8
5.2E-04
llmg_0299
add
llmg_0316
cpdC
1.4
2.2E-02
1.4
2.4E-02
1.4
3.4E-02
1.8
1.2E-03
-1.5
llmg_0467
pyrG
llmg_0480
dukA
-1.3
4.3E-02
1.4
4.9E-02
1.3E-02
3.5E-02
1.5
1.7E-03
-1.4
2.6E-02
1.7
4.1E-03
1.6
6.8E-03
1.3
3.1E-02
1.7
1.0E-02
llmg_0499
adaA
llmg_0555
tdk
1.4
3.0E-02
1.5
9.1E-03
llmg_0607
apt
3.3
8.1E-05
3.6
1.8E-05
3.1
1.6E-04
llmg_0732
cmk
2.3
5.5E-04
2.5
1.9E-04
1.9
5.8E-03
llmg_0762
udk
-1.4
1.4E-02
llmg_0783
purB
1.7
3.3E-02
5.2
7.8E-10
1.7
4.7E-02
5.3
5.2E-10
4.1
6.5E-08
llmg_0872
llmg0872
2.3
5.7E-04
3.1
3.2E-06
5.4
8.4E-11
1.7
5.0E-02
3.2
4.8E-06
2.4
4.4E-04
llmg_0873
llmg0873
2.2
3.4E-06
2.4
2.4E-07
2.0
5.8E-05
1.7
2.4E-03
1.5
1.1E-02
-1.5
9.7E-03
llmg_0891
pyrP
1.7
2.2E-03
1.9
2.6E-04
llmg_0893
pyrB
2.1
7.5E-04
2.0
1.4E-03
llmg_0952
pyrDA
1.4
2.6E-02
llmg_0964
thyA
1.6
2.4E-03
1.4
2.9E-02
llmg_0973
purC
1.5
2.0E-03
1.4
5.8E-03
1.4
9.3E-03
1.9
2.9E-03
1.9
1.9E-03
3.3
5.7E-08
2.1
7.0E-04
2.0
7.0E-04
2.1
5.5E-03
2.2
2.6E-03
3.7
7.1E-07
2.2
3.5E-03
2.1
5.2E-03
1.8
1.1E-02
2.1
1.0E-03
1.6
3.0E-02
1.8
1.5E-02
2.0
2.4E-03
2.1
6.7E-03
2.3
5.2E-03
1.9
4.7E-03
1.9
7.9E-03
llmg_0974
178
1.4
llmg_0975
purQ
llmg_0976
purL
llmg_0977
purF
llmg_0988
purN
llmg_0993
hprT
llmg_0994
purH
-1.3
2.1E-02
1.2
3.6E-02
llmg_1000
purK
1.7
7.1E-03
1.6
1.7E-02
llmg_1008
guaA
2.1
1.1E-03
llmg_1062
deoC
1.7
1.6E-02
1.6
9.8E-03
llmg_1063
cdd
1.7
2.1E-02
llmg_1089
llmg1089
1.6
4.1E-03
1.7
4.1E-03
llmg_1106
pyrDB
2.1
1.8E-04
2.1
1.7E-04
2.5
1.9E-05
llmg_1107
pyrF
3.0
6.1E-06
2.9
1.0E-05
4.1
1.9E-08
llmg_1188
llmg1188
1.4
2.7E-02
1.5
8.0E-03
llmg_1333
dukB
1.3
3.6E-02
1.3
2.3E-02
llmg_1335
llmg1335
-1.4
3.3E-03
llmg_1345
pbuX
1.7
1.0E-02
3.7
3.9E-06
llmg_1346
xpt
2.3
1.0E-03
llmg_1412
guaC
llmg_1444
mutX
llmg_1488
llmg1488
llmg_1508
pyrC
llmg_1509
pyrE
llmg_1542
nrdI
1.5
2.3
3.1E-02
1.3E-04
2.3
1.8
6.5E-05
1.9E-03
-1.3
2.1E-02
1.8
4.5E-02
1.8
1.2E-02
2.0
2.0E-02
-1.3
3.8E-02
-1.4
3.9E-02
1.9
7.7E-03
2.7
1.8E-05
1.7
4.2E-03
1.6
6.4E-03
1.6
9.8E-03
1.7
2.5E-02
1.3
4.8E-02
4.1
6.8E-07
6.6
7.1E-10
2.6
2.1E-04
2.5
2.7E-04
-1.4
1.2E-02
1.5
2.3E-02
-1.6
1.2E-02
-1.5
4.3E-02
1.9
2.2E-03
1.7
8.4E-04
-1.5
3.3E-02
1.8
2.4E-03
1.6
1.1E-02
-1.5
4.6E-02
1.5
2.4E-02
1.6
2.0E-02
-1.6
2.6E-02
1.5
3.5E-02
llmg_1543
nrdE
1.6
4.0E-03
2.6
6.3E-05
2.9
1.7E-04
llmg_1544
nrdF
1.6
9.4E-03
1.6
9.1E-03
1.7
5.0E-03
1.7
-1.8
1.6E-02
-2.1
1.8E-03
-3.3
-1.4
1.0E-02
2.1
3.3E-03
2.7
2.4E-03
-1.6
1.0E-02
2.4E-03
1.0
6.5E-07
-2.2
2.9E-02
1.4
2.7E-02
1.0E-03
-1.9
6.4E-03
-1.5
8.3E-03
llmg_1595
fhs
llmg_1599
deoD
llmg_1720
udp
llmg_1843
llmg1843
1.7
5.4E-04
1.5
6.4E-03
llmg_1884
flaR
-1.7
1.5E-02
-1.5
4.7E-02
llmg_2073
pfs
1.7
1.9E-03
1.4
3.4E-02
-1.5
2.5E-02
llmg_2149
llmg2149
-1.7
9.3E-03
-1.6
2.0E-02
-1.6
2.0E-02
llmg_2155
gmk
2.8
2.1E-04
2.9
1.1E-04
2.3
1.5E-03
llmg_2170
llmg2170
-1.6
1.9E-02
llmg_2183
llmg2183
llmg_2201
purA
1.4
2.8E-02
llmg_2285
pyrH
-1.7
9.0E-03
-1.6
2.1E-02
-2.0
llmg_2359
adk
1.4
1.3E-02
1.6
2.6E-04
1.6
7.6E-04
llmg_2510
mutT
1.8
1.4E-03
1.8
4.7E-03
-3.1
1.4E-03
1.5
2.0E-02
1.5
7.5E-03
1.5
6.0E-03
1.8
2.3E-03
1.7
7.3E-03
-3.1
1.2E-03
-2.4
1.3E-02
5.3E-04
[G] Carbohydrate transport and metabolism
llmg_0022
mtlA
llmg_0024
mtlF
1.6
2.5E-03
1.6
2.0E-03
llmg_0112
llmg0112
1.6
5.4E-03
1.5
3.1E-02
-1.6
1.9E-02
-1.6
2.0E-02
llmg_0126
ptsH
2.2
2.3E-02
2.7
2.2E-03
9.1
8.3E-12
6.5
6.1E-09
5.7
4.1E-08
llmg_0127
ptsI
2.0
2.7E-03
2.2
6.4E-04
1.8
1.6E-02
llmg_0133
blt
1.3
4.3E-02
1.6
1.1E-02
llmg_0158
glgB
-1.4
3.1E-02
-1.7
1.3E-02
-1.6
3.6E-02
llmg_0190
bglS
llmg_0249
llmg0249
1.6
2.7E-03
1.5
3.0E-02
-1.4
8.8E-03
-1.4
8.7E-03
3.2E-02
llmg_0255
dhaK
llmg_0257
dhaM
llmg_0264
fbp
llmg_0266
llmg0266
1.3
4.3E-02
llmg_0293
xynD
-1.4
2.5E-02
1.3
4.9E-02
1.4
1.5E-02
llmg_0320
napC
-1.7
3.4E-04
-1.4
3.3E-02
-1.4
3.3E-02
llmg_0351
llmg0351
-1.3
1.8E-02
-1.3
3.1E-02
llmg_0355
pmg
3.0
1.1E-04
3.8
2.0E-06
3.4
8.0E-06
llmg_0383
ppnK
-1.3
-1.3
1.3E-02
-1.3
2.7E-02
llmg_0403
pepA
-1.5
1.2E-02
llmg_0437
ptcB
1.7
2.2E-03
-1.5
1.6E-02
llmg_0438
ptcA
-2.5
5.3E-03
llmg_0441
bglA
-1.4
2.7E-02
-1.4
3.0E-02
llmg_0446
msmK
2.4
3.4E-02
2.8
1.0E-02
llmg_0451
femD
llmg_0453
llmg0453
llmg_0454
llmg0454
llmg_0455
TrePP
-1.4
1.8
1.4E-02
-1.6
-1.3
-1.4
1.6
1.7
2.1E-02
llmg_0488
llmg0488
llmg_0489
llmg0489
1.5
3.8E-02
llmg_0490
llmg0490
1.4
4.5E-02
llmg_0494
nagZ
2.9E-04
-1.4
9.8E-03
-1.4
4.8E-02
3.7E-02
5.9E-03
3.0E-02
1.4E-02
1.6
-1.4
-2.3
3.9E-04
1.5
3.2E-02
1.5
1.1E-02
1.7E-02
-2.5
3.9E-02
-2.8
9.8E-03
-2.4
3.1E-04
-2.3
5.4E-04
2.6E-02
2.4
7.0E-05
1.7
7.0E-03
-1.9
5.7E-03
2.3
3.1E-04
2.0
5.5E-03
-2.2
1.0E-03
1.6
7.4E-03
-1.5
2.7E-02
1.8
3.2E-03
1.5
1.4E-02
1.6
1.1E-02
1.7
5.4E-03
2.1
2.8E-04
1.6
1.2E-02
1.8
2.9E-03
179
llmg_0516
suhB
llmg_0530
180
-1.8
5.2E-03
-1.8
1.1E-02
llmg_0586
gnd
llmg_0617
enoA
llmg_0631
pmrA
llmg_0649
icaB
-1.5
1.4E-02
llmg_0727
ptnD
-2.2
1.3E-02
llmg_0728
ptnC
llmg_0729
ptnAB
llmg_0737
malG
llmg_0738
llmg_0739
1.6
3.0E-04
1.6
3.7E-02
-2.6
1.8
2.1E-03
-2.2
1.2E-03
1.9
8.5E-03
1.9
9.1E-04
3.6
2.2E-08
1.7E-02
-2.7
1.5E-04
malF
-2.3
1.7E-03
malE
-2.0
3.2E-02
-1.6
2.9E-02
dexC
llmg_0741
dexA
llmg_0743
amyY
llmg_0744
agl
1.8
llmg_0745
mapA
llmg_0751
ascB
llmg_0785
rbsK
llmg_0786
rbsD
llmg_0787
rbsD
1.7
1.2E-02
3.0E-03
1.7
1.2E-02
1.6
2.0E-02
1.8
1.3E-02
-2.1
5.6E-04
-2.1
6.4E-04
1.6
2.3E-03
1.5
1.7E-02
2.8
1.1E-05
llmg_0788
rbsC
1.7
4.4E-02
llmg_0856
llmg0856
-1.3
2.3E-02
llmg_0858
uxuB
-1.4
4.9E-02
-1.7
6.8E-04
llmg_0859
uxuA
-1.5
4.6E-02
-1.5
3.2E-02
llmg_0860
uxuT
-1.3
3.3E-02
llmg_0862
uxaC
-1.4
9.5E-03
llmg_0866
llmg0866
llmg_0868
tkt
llmg_0957
rpe2
llmg_0958
rpiB
llmg_0959
llmg0959
llmg_0960
llmg0960
llmg_0961
llmg0961
llmg_0963
1.5
2.0E-02
7.5E-03
-3.2
1.4
-1.7
4.3E-02
1.9
7.5E-03
-1.3
4.5E-02
2.0
9.6E-03
1.7
1.2E-02
2.2
9.2E-03
2.6
2.6E-04
2.9
3.3E-06
2.9
4.8E-06
-2.6
3.4E-04
-2.3
1.1E-03
-2.1
6.4E-03
-2.0
8.7E-03
-1.9
9.2E-03
-1.8
1.4E-02
-1.9
1.7E-03
-2.0
3.4E-04
-2.0
9.3E-04
-1.8
3.1E-03
1.5E-02
4.9E-07
1.3E-02
-1.5
2.6E-02
-2.1
8.0E-04
-2.0
3.3E-03
1.5
4.9E-02
-2.2
1.1E-03
-1.9
2.9E-03
1.5
9.8E-03
-1.6
4.4E-02
-2.0
4.3E-03
1.4
1.2E-02
1.4
4.0E-02
-1.5
2.2E-02
2.0
2.7E-03
1.3
2.8E-02
-1.4
2.3E-02
1.4
2.1E-02
1.9E-03
-1.3
4.4E-02
-1.5
3.7E-03
2.0
1.6E-04
-1.0
3.4E-02
1.5
3.1E-02
1.5
6.9E-03
1.6
2.0E-03
llmg0963
-1.4
2.5E-02
1.5
1.8E-02
-1.4
2.4E-02
4.2E-03
llmg_0967
llmg0967
-1.6
5.8E-03
llmg_1002
xylA
-1.6
6.8E-03
-1.6
2.4E-02
llmg_1004
llmg1004
-1.6
6.4E-05
-1.7
3.1E-06
-1.5
7.2E-04
llmg_1006
scrK
-1.4
3.5E-02
llmg_1011
lplA
-1.4
3.2E-02
llmg_1046
bglH
1.4
2.3E-03
1.3
2.1E-02
llmg_1097
glpF2
1.8
2.7E-02
-2.9
7.8E-05
llmg_1104
llmg1104
1.3
1.6E-03
-1.6
-2.0
1.4
2.1E-02
2.1
3.1E-02
4.9E-02
1.6
1.5
1.6
1.5
2.0E-03
-1.9
llmg_0740
1.1E-02
-2.1
1.7
4.5E-02
-1.3
2.9E-02
1.4
4.8E-02
1.7
4.8E-02
1.2E-02
-1.3
4.4E-02
-1.4
4.0E-02
-1.7
2.8E-04
-1.4
4.3E-02
-2.4
2.5E-04
-2.6
5.4E-05
-1.3
3.0E-02
-4.0
1.3E-07
-3.0
4.1E-05
-1.4
5.3E-03
llmg_1117
nagA
-1.6
8.9E-03
-1.6
4.4E-03
llmg_1118
pfk
1.4
4.6E-02
1.5
8.2E-03
llmg_1119
pyk
2.3
1.8E-03
llmg_1163
llmg1163
llmg_1164
llmg1164
llmg_1165
llmg1165
-1.4
-1.4
1.4E-02
1.0E-02
-1.3
3.3E-02
-1.5
2.3E-02
-1.5
1.3E-03
2.2
9.3E-03
-1.4
1.8E-02
-1.4
2.1E-02
-1.7
4.4E-03
llmg_1166
llmg1166
llmg_1168
llmg1168
-1.4
2.6E-03
1.6
8.4E-03
llmg_1169
aguA
-1.4
1.3E-02
-1.5
1.3
4.8E-02
1.5
3.3E-02
1.5
1.7E-02
-1.3
2.9E-02
-3.7
3.4E-07
llmg_1244
llmg1244
-1.3
3.9E-02
llmg_1280
ilvD
1.5
2.7E-02
llmg_1320
llmg1320
-1.3
3.7E-02
llmg_1358
orf48
llmg_1414
nagD
llmg_1455
bglA2
llmg_1456
bglX
llmg_1569
fruC
-3.6
3.5E-07
llmg_1570
fruR
-1.5
6.6E-03
llmg_1579
gpmB
1.6
1.1E-02
llmg_1601
deoB
-1.5
3.8E-02
llmg_1608
llmg1608
-1.3
4.0E-02
llmg_1622
tagG
llmg_1636
llmg1636
llmg_1738
llmg1738
llmg_1767
rdrA
llmg_1789
pmi
llmg_1839
llmg1839
-1.5
-1.4
2.3E-03
-3.3
1.8E-06
1.8
7.2E-04
-1.4
1.9E-02
-1.4
3.4E-02
9.1E-03
1.6
1.7E-03
2.9
3.2E-05
1.5
1.3E-02
-1.5
1.0E-03
-1.4
8.5E-03
1.3
4.6E-02
1.4
2.5E-02
-1.4
4.2E-02
-1.6
3.1E-02
-1.5
2.1E-02
1.8
1.2E-04
1.5
3.4E-02
1.7
1.4E-02
-1.8
1.0E-02
1.6
3.8E-03
1.4
3.7E-02
-1.5
4.5E-03
2.5
3.4E-04
2.5
2.7E-04
1.3
1.5E-02
1.7
1.1E-02
-1.4
1.7E-02
-1.3
3.8E-02
2.0
2.1E-04
8.2E-03
1.5
-1.4
8.1E-03
1.2E-02
llmg_1869
apu
1.9
3.3E-03
1.7
9.3E-03
1.7
1.8E-02
llmg_1871
glgP
1.7
3.2E-04
1.5
4.4E-03
1.6
2.1E-02
llmg_1872
glgA
-1.5
1.9E-03
-1.4
6.4E-03
-1.6
3.9E-03
-2.0
3.5E-05
-1.4
9.1E-03
-1.4
2.3E-02
llmg_1873
glgD
-1.5
7.2E-03
llmg_1894
llmg1894
-1.3
4.6E-02
3.0E-02
llmg_1923
gpmC
1.5
llmg_2066
llmg2066
-1.3
3.0E-02
llmg_2157
nagB2
-1.4
4.5E-02
llmg_2167
fbaA
1.7
3.1E-03
2.7
3.7E-07
1.5
2.6E-02
2.5
4.0E-06
llmg_2185
lacX
1.5
1.5E-02
1.7
4.2E-03
1.5
1.7E-02
1.7
3.7E-03
llmg_2199
chiC
-2.2
6.9E-03
-1.9
1.6E-02
-2.4
2.6E-03
llmg_2235
galK
-1.6
2.9E-02
1.6
3.8E-02
llmg_2236
galM
3.4
8.2E-06
3.6
5.2E-06
llmg_2237
galP
-1.4
2.9E-02
llmg_2273
llmg2273
-1.4
2.6E-02
llmg_2327
glpF3
2.6
8.1E-04
llmg_2346
gmhA
llmg_2431
llmg2431
llmg_2448
pgiA
llmg_2467
gntP
llmg_2468
gntK
llmg_2469
gntZ
llmg_2499
llmg2499
llmg_2511
rpiA
llmg_2513
llmg2513
llmg_2539
gapB
llmg_2561
glcU
2.0
1.1E-02
2.0
-2.8
1.9E-07
5.3E-03
1.5
2.7E-02
1.4
4.0E-02
-2.7
3.1E-07
3.6
1.9
1.6E-03
-1.4
6.9E-03
-1.3
2.9E-02
-1.3
3.8E-02
2.1E-06
-9.1
5.5E-08
-9.1
6.2E-08
-7.3
8.3E-07
2.9
7.5E-06
3.0
1.4E-05
2.2
4.2E-04
-1.8
7.1E-04
-2.0
7.6E-05
-1.9
2.6E-04
-1.4
3.2E-02
1.5
8.7E-03
1.5
4.1E-02
-2.9
9.3E-08
1.7
8.4E-03
1.7
1.5E-03
1.5
2.7E-02
1.3
4.1E-02
1.4
4.7E-02
1.4
1.6E-02
1.6
1.7E-02
181
[H] Coenzyme transport and metabolism
llmg_0075
lplL
llmg_0181
llmg0181
1.5
llmg_0196
llmg0196
llmg_0197
menA
llmg_0236
nadD1
llmg_0237
llmg0237
llmg_0463
thiD2
llmg_0543
dfpA
1.6
llmg_0544
dfpB
1.7
llmg_0718
llmg0718
llmg_0749
dxsB
llmg_0753
ubiE
llmg_0759
ubiA
llmg_0934
hemH
llmg_1026
llmg1026
1.7
1.4
1.4
1.5
7.2E-03
coaA
1.4
7.7E-03
llmg1124
1.5
1.3E-02
llmg_1334
folC
1.5
3.1E-02
1.6
llmg_1336
folP
1.4
4.1E-02
1.6
1.3
llmg_1473
nadD2
9.9E-04
4.8E-03
2.7E-02
nadE
2.6E-05
1.6
1.4
llmg1313
pncB
2.8
1.6E-02
ilvC
llmg_1472
4.8E-02
1.5
llmg_1313
llmg_1470
1.7E-03
4.5E-02
llmg_1277
hemN
1.9
1.4
6.6E-03
llmg_1418
2.1E-02
1.2E-05
-1.5
dfrA
7.4E-03
2.9
ilvE
llmg_1342
1.7
1.5
3.8E-02
5.4E-03
1.7
1.4E-02
llmg_1529
llmg1529
1.5
2.9E-02
1.7
llmg_1530
ribA
2.0
7.0E-04
2.4
7.5E-06
3.0
1.3E-07
llmg_1531
ribB
1.5
7.4E-03
1.6
2.4E-03
llmg_1532
ribD
1.6
1.4E-02
2.2
1.7E-04
llmg_1583
llmg1583
-1.3
2.1E-02
-1.3
1.3E-02
-5.0
1.7E-08
-2.6
3.6E-04
llmg_1693
folD
1.4
1.2E-02
1.4
1.7E-02
-1.4
6.4E-03
llmg_1719
pnuC1
llmg_1828
menF
llmg_1829
menD
1.7
8.0E-03
llmg_1831
menB
llmg_1859
llmg1859
1.5
9.1E-03
llmg_2013
thiN
1.3
4.0E-02
llmg_2070
llmg2070
1.5
3.7E-02
1.8
5.9E-05
1.6
9.3E-03
llmg_2160
metK
llmg_2162
birA2
1.7E-02
2.7E-02
2.4
4.8E-04
2.3
5.2E-04
1.9
2.4E-03
2.6
9.9E-05
2.0
2.9E-03
1.9
1.2E-02
1.8
2.2E-02
1.4E-03
2.4
6.6E-04
-1.5
3.2E-03
-1.5
3.1E-03
1.4
3.0E-02
1.4
4.4E-02
1.4
1.7E-02
3.0
7.0E-08
1.4
1.7E-02
1.4
3.0E-02
1.7
1.4E-03
2.3E-03
1.4
4.2E-02
1.6
4.3E-03
1.6
3.3E-03
1.7
2.0E-03
-1.4
1.5E-02
-1.4
2.7E-02
1.8
4.9E-03
1.7
7.5E-03
2.2
2.1E-04
1.6
1.5E-02
-3.1
5.1E-05
1.3
3.8E-02
1.5
2.8E-02
1.6
2.4E-02
-1.5
4.0E-02
-3.5
1.3E-05
1.5
3.9E-02
1.4
2.5E-02
2.1E-02
2.6E-03
2.8E-03
1.2E-02
1.2E-02
3.2E-02
1.6
8.5E-03
2.2
1.6
1.5
-1.4
1.6
-1.3
1.6
1.4
-1.4
2.1
7.6E-04
1.6E-02
1.0E-02
1.1E-02
5.9E-06
1.4
-2.0
3.7E-02
-1.6
1.6E-03
2.5E-02
pnuC2
1.4E-02
-1.4
3.1
1.8E-02
ispA
1.6
1.5E-02
2.2
1.1E-02
llmg_1664
1.8E-03
4.4E-03
1.4
llmg_1689
-1.5
3.8E-02
1.6
2.1E-03
3.8E-02
1.8
1.5
1.1E-03
4.5E-02
1.6
1.4
6.4E-03
-1.4
2.0E-02
1.2E-02
6.7E-04
1.5E-02
1.2E-02
1.7
3.5E-03
1.7
7.4E-04
-1.4
4.0E-02
1.6
2.5E-02
1.4
4.6E-02
2.1E-02
2.2E-03
1.4
1.7
1.5
1.7
-1.4
1.5
1.7
-1.4
4.2E-03
2.0
9.6E-03
thiE
folE
2.7E-03
1.3E-02
1.5
llmg_1218
llmg_2158
182
1.7
llmg_1181
folB
8.4E-04
2.0E-02
llmg_1124
llmg_1338
2.1
-1.6
llmg_1058
llmg_1337
4.9E-02
1.8
5.2E-03
1.4
3.7E-02
2.5
6.4E-06
1.5
5.4E-03
-1.4
4.7E-02
-1.5
4.8E-03
2.3
2.2E-05
llmg_2205
llmg2205
1.5
3.2E-02
1.4
4.2E-02
llmg_2241
nadR
-1.4
6.6E-03
-1.3
2.9E-02
llmg_2242
llmg2242
-1.5
3.6E-02
1.4
8.7E-03
[I] Lipid transport and metabolism
llmg_0119
llmg0119
llmg_0397
llmg0397
llmg_0425
mvk
llmg_0426
mvaD
llmg_0427
llmg0427
llmg_0431
llmg0431
llmg_0445
llmg0445
llmg_0457
llmg0457
llmg_0538
llmg0538
llmg_0539
fabI
llmg_0627
fadD
llmg_0929
hmcM
llmg_0930
thiL
llmg_0931
mvaA
1.6
-1.4
-1.5
2.1E-02
1.6E-02
1.8
-1.4
1.2E-02
-1.4
3.3E-03
1.7E-02
1.5
-1.5
1.0E-03
1.4
4.5E-02
1.7
4.0E-03
1.9
2.6E-03
1.6
3.7E-02
3.7
2.7E-04
1.9E-02
1.5E-02
-1.4
1.4
3.0E-02
-1.6
8.6E-03
1.9
llmg0935
1.8
3.1E-02
llmg_1415
llmg1415
1.6
2.6E-03
2.0
3.4
2.4E-02
8.1E-03
1.9
9.1E-04
3.5
5.5E-04
1.7
7.2E-04
-1.5
9.3E-03
-1.3
1.5E-02
1.5
1.3E-02
1.6
2.3E-03
1.5
2.3E-02
1.7
9.8E-04
2.0
8.2E-03
2.5E-02
4.1E-06
llmg1517
1.5
1.8E-02
1.5
6.5E-03
1.5
2.1E-02
1.9
4.6E-03
2.2
1.2E-03
1.8
9.1E-03
llmg_1606
tagD2
1.5
1.6E-02
acpS
accA
llmg_1779
accC
-1.6
3.3E-04
-1.4
1.6
clsA
llmg_1705
3.5
1.2E-03
llmg_1517
llmg_1777
3.5E-03
4.4E-02
llmg_1558
-1.3
2.4E-02
-1.5
3.2E-02
7.6E-03
1.5
1.9
-1.5
-1.3
llmg_0935
1.9E-03
2.4
1.2E-03
-1.4
4.1E-02
2.1
2.3E-03
-1.5
3.4E-02
-1.5
1.3E-02
-1.6
1.5E-02
-1.5
1.6E-02
-1.5
4.9E-04
-1.8
7.2E-03
-1.6
2.4E-02
-1.8
1.3E-02
2.3
1.7E-03
2.3
1.4E-03
2.0E-02
1.1E-02
-1.6
9.3E-04
-1.7
2.7E-04
-1.5
1.0E-03
-1.4
3.4E-02
llmg_1781
fabZ2
llmg_1782
accB
-1.6
4.6E-02
llmg_1785
fabD
3.0
3.9E-05
llmg_1787
fabH
llmg_1919
llmg1919
-1.7
2.7E-02
-1.8
1.9E-02
llmg_1980
llmg1980
1.3
4.6E-02
1.3
4.3E-02
-1.5
9.7E-03
-1.5
2.8E-02
llmg_2224
pgsA
1.6
4.1E-02
1.5
4.5E-02
llmg_2391
llmg2391
-2.3
1.9E-04
-2.2
4.3E-04
-2.3
1.0E-04
llmg_2414
cdsA
2.2
4.7E-06
1.7
4.0E-03
1.5
9.3E-03
2.0
5.4E-05
1.7
2.4E-03
1.9
3.2E-04
1.8
9.5E-03
1.9
3.2E-03
1.3
3.9E-02
1.3
1.5E-02
1.5
4.7E-02
1.6
5.5E-03
1.6
7.1E-03
llmg_2415
llmg_2494
ywjF
-1.5
2.6E-03
-1.5
7.1E-03
-1.6
5.6E-03
1.6
2.4E-02
1.9
2.4E-03
-1.4
2.3E-03
[J] Translation, ribosomal structure and biogenesis
llmg_0007
llmg0007
llmg_0012
pth
llmg_0015
llmg0015
llmg_0079
trpS
llmg_0098
rpmG
llmg_0099
-1.3
6.3E-03
1.3
1.7
-1.4
1.9E-02
-1.8
4.4
3.5E-04
5.3E-04
2.7E-02
1.8E-02
-1.6
4.3E-03
1.4
2.3E-02
11.1
3.2E-09
-1.7
2.9
2.8E-03
2.0E-02
-1.4
4.6E-02
1.5
2.0E-02
1.4
3.2E-02
8.6
1.7E-07
5.0
1.3E-04
-1.7
3.3E-03
llmg_0110
prmA
-1.5
3.1E-02
-2.0
2.2E-04
llmg_0115
dtd
-1.4
4.0E-02
-1.4
3.7E-02
llmg_0164
tgt
1.6
1.7E-02
1.7
7.5E-03
1.6
1.4E-02
1.7
5.9E-03
183
llmg_0174
184
1.5
3.7E-02
llmg_0175
gatA
1.4
2.1E-02
1.5
3.9E-02
llmg_0176
gatB
1.4
2.8E-02
llmg_0177
llmg0177
-1.3
3.6E-02
llmg_0204
rpmB
1.7
2.9E-03
llmg_0296
rpsD
llmg_0369
rheA
1.4
2.3E-02
1.5
1.2E-02
llmg_0384
rluE
1.5
8.6E-03
1.3
4.5E-02
4.5
1.7E-05
5.0
4.0E-06
1.9
2.3E-03
1.7
4.5E-02
4.6E-02
-1.3
4.5E-02
1.2E-02
1.5
5.0E-03
-1.5
8.8E-03
-1.3
1.4
1.8
1.2E-03
1.6
4.0E-02
1.7
2.2E-02
1.9
4.1E-03
1.5
4.5E-02
1.5
3.1E-02
1.3
4.4E-02
1.5
4.3E-02
1.6
1.1E-02
2.4
6.3E-05
2.2
1.7E-04
llmg_0389
lysS
llmg_0401
tyrS
2.3
2.1E-04
3.8
3.2E-04
llmg_0460
llmg_0462
llmg0460
1.8
7.6E-03
2.3
9.2E-05
2.1
4.4E-04
truA
2.0
4.0E-03
2.3
6.3E-04
2.0
llmg_0532
llmg0532
4.0E-03
-1.4
4.8E-02
llmg_0557
prfA
llmg_0562
llmg0562
1.4
2.4E-02
llmg_0616
yfiA
llmg_0879
speG
1.4
3.7E-02
-7.3
1.7E-08
-6.7
7.8E-08
-6.6
5.2E-08
llmg_0932
rpsP
2.0
6.4E-03
2.6
5.1E-05
1.8
2.1E-02
llmg_0942
papL
1.6
2.2E-02
2.0
4.2E-04
-1.8
1.8E-02
1.9
1.1E-03
1.6
llmg_1040
llmg1040
-1.3
4.1E-02
2.3E-02
llmg_1100
rluC
llmg_1194
rluB
llmg_1207
rplJ
1.8
1.5E-02
2.3
llmg_1208
rplL
2.3
1.9E-03
2.9
llmg_1271
gidC
-1.3
5.0E-02
llmg_1293
llmg1293
llmg_1315
llmg1315
llmg_1343
spoU
llmg_1471
llmg1471
-1.4
4.3E-02
-1.5
9.0E-03
1.5
1.7E-02
-1.5
1.3
3.5E-02
1.7
6.9E-03
-1.4
1.4E-02
1.7
1.9E-02
1.7
1.4E-02
1.9
2.3E-03
1.5
3.8E-02
5.5E-04
4.8
1.2E-09
3.3
2.1E-06
2.8
4.7E-05
1.1E-04
5.2
3.3E-09
3.1
5.3E-05
2.6
6.4E-04
2.1
1.7E-03
1.8
2.9E-03
1.7
5.7E-03
-1.4
1.8E-02
-1.5
3.3E-03
3.5
2.7E-07
3.3
6.1E-07
3.1
2.1E-06
1.7
4.1E-03
1.6
6.2E-03
1.5
2.6E-02
1.2E-02
1.6
4.7E-03
-1.4
4.4E-03
-1.4
2.6E-02
llmg_1477
glyS
2.0
3.9E-03
2.2
6.6E-04
2.2
1.3E-03
1.6
4.6E-02
llmg_1478
glyQ
1.5
1.1E-02
1.8
2.1E-04
1.6
3.0E-03
1.5
2.1E-02
llmg_1491
rpmA
1.5
3.9E-02
2.0
2.6E-04
1.7
5.6E-03
llmg_1492
llmg1492
1.9
6.0E-03
1.7
3.2E-02
llmg_1493
rplU
1.7
2.7E-02
1.9
1.0E-02
llmg_1524
rluD
1.6
1.6E-02
1.6
1.3E-02
1.5
7.8E-03
llmg_1547
prfB
llmg_1550
llmg1550
1.7
4.2E-02
-1.4
-1.6
2.9E-02
2.5
2.6E-04
-1.5
3.0E-02
1.6
2.6E-02
2.0E-03
1.5
llmg_1645
llmg1645
llmg_1671
rplS
2.5
1.5E-05
2.5
2.0E-05
2.5
6.0E-06
llmg_1687
llmg1687
-1.7
1.0E-02
-1.6
1.4E-02
-1.8
3.4E-03
llmg_1724
rpsA
2.3
1.3E-02
4.6
1.6E-06
4.4
4.1E-06
2.7
2.1E-03
llmg_1725
trmU
1.7
1.9E-02
3.1
2.3E-06
2.8
1.5E-05
2.1
1.9E-03
llmg_1741
leuS
1.5
1.0E-02
1.3
3.5E-02
-2.0
2.0E-03
-2.0
2.8E-03
-2.3
2.2E-04
-2.0
9.6E-03
-2.5
1.0E-03
-1.5
5.8E-03
-2.0
1.2E-03
-1.7
8.6E-03
1.3
3.9E-02
1.4
8.1E-03
llmg_1766
llmg1766
llmg_1791
rbfA
llmg_1792
infB
llmg_1793
llmg1793
llmg_1815
rplI
llmg_1823
ansB
-1.6
2.9E-03
1.4
1.5
2.7E-02
1.6E-02
1.1E-02
1.7
1.3E-02
-1.3
4.4E-02
-1.8
6.3E-03
2.1
2.7E-02
llmg_1880
efp
llmg_1882
ksgA
llmg_1906
alaS
llmg_1921
rpsT
llmg_2017
asnS
llmg_2029
rplT
1.4
-1.7
1.4
2.7E-02
1.5
8.1E-03
llmg_2030
llmg_2031
infC
llmg_2040
cysS
llmg_2044
pnpA
llmg_2050
tufA
llmg_2053
ileS
1.3
1.7
4.6E-02
9.6E-04
6.2E-03
1.5
1.5E-02
2.2
1.9E-05
1.8
2.7E-03
2.0
6.4E-03
2.8
5.3E-05
3.4
2.3E-03
7.8
6.8E-08
1.3
4.6E-02
1.7
4.6E-02
1.9
1.8E-04
trmH
rpsO
1.9
6.3E-04
llmg_2081
sunL
2.0
8.9E-03
llmg_2147
fmt
1.6
4.5E-03
llmg_2169
thrS
2.1
1.6E-04
2.1
llmg_2195
pheT
1.7
1.8E-03
llmg_2209
llmg2209
1.4
3.2E-02
aspS
hisS
llmg_2277
rplK
llmg_2284
llmg_2314
5.9E-03
3.4E-02
llmg_2078
llmg_2217
1.5
1.4
llmg_2063
llmg_2215
-1.5
1.6
5.6E-03
-2.3
3.2E-03
1.7
3.3E-03
2.3
2.6E-04
2.6
3.1E-04
1.8E-04
1.8
5.1E-03
1.8
3.2E-03
1.9
8.9E-03
1.4
4.2E-02
7.7
3.8E-11
frr
-1.5
8.1E-03
argS
1.8
2.0E-03
-1.6
1.2E-02
1.7
2.9E-02
llmg_2332
gltX
llmg_2353
rplQ
llmg_2355
rpsK
llmg_2356
rpsM
-1.6
6.7E-03
llmg_2357
2.2
2.1E-03
-1.6
1.4E-05
6.3E-03
-1.7
1.1E-02
2.0
9.1E-04
2.3
3.4E-04
3.7
2.4E-07
2.9
3.1E-05
5.1
6.5E-10
llmg_2358
infA
llmg_2362
rplO
1.4
1.6E-02
2.0
3.2E-02
llmg_2363
rpmD
llmg_2364
rpsE
llmg_2365
rplR
2.0
8.6E-04
2.5
llmg_2366
rplF
1.6
1.9E-02
llmg_2367
rpsH
2.1
llmg_2370
rpsN
-1.6
llmg_2371
rplE
llmg_2372
rplX
llmg_2373
rplN
llmg_2374
rpsQ
-1.6
1.9
2.6E-02
1.8E-05
2.7
1.6E-02
-1.5
6.3E-03
1.1E-03
1.9
2.6E-03
2.0
5.9E-03
1.7
3.8E-02
6.8
5.9E-07
3.9
6.4E-04
-1.4
3.0E-02
-1.3
2.6E-02
-2.8
1.7E-04
-2.1
6.1E-03
2.1
5.2E-04
3.6E-02
2.3E-02
1.9
1.4E-02
2.5
5.0E-04
1.4
3.7E-02
-1.4
2.2E-02
1.6
2.6E-02
2.0
5.5E-03
1.7
3.7E-02
2.5
2.4E-03
6.0
4.8E-09
1.2E-02
2.0E-02
-1.6
1.5
1.8
3.2E-02
1.6
1.4E-02
1.5
1.7
2.2E-02
-1.5
3.0E-02
-1.5
1.3E-02
-1.5
3.8E-02
3.4
4.3E-05
1.0E-02
-1.5
8.9E-03
1.2E-03
2.0
9.9E-04
1.9
1.9
1.1E-02
1.9
8.3E-03
1.7
9.4E-03
1.7
1.6E-02
2.0
1.3E-03
3.0
1.3E-05
2.4
5.8E-04
3.1
1.1E-05
2.5
4.4E-04
2.0
1.5E-02
1.6
1.2E-02
7.3E-06
1.6
2.1E-02
2.1
2.5E-04
1.5
5.0E-02
2.2
9.6E-05
1.5
3.7E-02
2.1
2.9E-04
1.5
3.6E-02
8.0E-04
3.3
3.1E-07
1.8
9.6E-03
2.9
3.8E-06
1.9
6.0E-03
3.4E-02
-2.2
1.4E-04
-1.6
1.5E-02
-1.7
8.4E-03
-1.4
1.4E-02
-1.7
1.7E-02
2.5E-02
llmg_2375
-1.9
-1.5
2.7E-02
-1.5
2.6E-02
-1.7
3.8E-03
-1.5
3.1E-02
-1.9
9.2E-04
-1.7
2.1E-02
llmg_2376
rplP
llmg_2380
rplB
-1.6
4.4E-02
-1.7
4.9E-03
1.7
2.3E-03
llmg_2382
rplD
1.7
2.6E-03
1.9
2.5E-04
2.6
5.5E-07
llmg_2383
rplC
1.8
1.1E-03
2.5
2.0E-06
3.3
2.3E-09
llmg_2384
rpsJ
1.8
5.4E-03
2.4
8.2E-05
3.2
3.7E-07
-1.6
5.0E-02
llmg_2381
llmg_2390
2.2
1.3
-1.5
-1.4
3.6
2.6E-02
4.0E-02
-1.8
1.5
1.5
2.2E-03
1.7E-02
1.7E-02
2.1E-02
-1.9
6.7E-03
-1.5
9.2E-03
-1.5
3.4E-02
-1.5
2.5E-02
-1.9
4.3E-04
-1.6
7.4E-03
-1.8
1.5E-03
-1.5
2.6E-02
1.8
7.1E-04
1.5
9.6E-03
1.7
5.4E-03
1.5
1.1E-02
2.2
8.1E-05
1.6
1.5E-02
2.0
1.9E-03
1.6
4.6E-02
185
llmg_2396
rluA
llmg_2412
proS
llmg_2429
tsf
llmg_2430
rpsB
llmg_2455
valS
llmg_2464
llmg2464
1.5
5.1E-03
1.5
2.4E-02
1.5
1.7E-02
-1.7
9.4E-03
1.8
1.1E-03
1.5
2.6E-02
1.8
1.1E-03
-2.2
4.2E-04
-1.7
7.1E-03
-1.4
8.1E-03
-2.6
9.4E-06
-1.6
1.8E-02
6.0
8.9E-09
5.4
5.0E-08
4.7
4.2E-07
-3.7
3.9E-06
-2.6
1.5E-04
-2.7
1.2E-04
1.6
1.6E-02
1.7
8.0E-03
-1.5
3.7E-02
-1.5
4.3E-02
-1.6
1.7E-02
1.5
1.8E-02
llmg_2473
rpsR
llmg_2518
rsuA
llmg_2545
rpsI
1.3
3.3E-02
1.6
3.3E-04
1.6
1.3E-03
1.5
8.8E-04
llmg_2546
rplM
2.1
1.9E-03
4.0
4.3E-08
1.8
2.1E-02
3.3
2.6E-06
2.8
5.1E-05
llmg_2556
fusA
1.2
9.3E-03
-1.8
3.1E-03
-2.6
7.9E-06
-1.6
3.2E-02
-2.1
3.7E-04
-2.0
1.2E-03
llmg_2557
rpsG
1.4
4.7E-02
1.5
8.2E-03
llmg_2558
rpsL
1.6
2.3E-02
1.8
1.4E-03
3.3
7.2E-09
2.4
7.8E-06
2.1
9.2E-05
llmg_0013
mfd
1.4
1.0E-02
1.5
1.7E-02
llmg_0023
mtlR
-3.1
5.8E-06
-2.7
5.0E-05
llmg_0089
llmg0089
1.3
1.2E-02
llmg_0102
parA
llmg_0161
llmg0161
llmg_0172
codY
llmg_0180
cspE
llmg_0248
llmg0248
-1.5
llmg_0250
rmeC
llmg_0297
tex
llmg_0301
llmg0301
llmg_0323
llmg0323
-1.4
1.2E-02
llmg_0350
fhuR
-1.6
1.7E-02
llmg_0353
llmg0353
1.6
1.3E-02
llmg_0390
rlrG
llmg_0393
rlrD
llmg_0414
llrC
1.8
1.3E-03
llmg_0432
llmg0432
-1.5
2.6E-02
-1.8
4.0E-03
1.6
1.6E-02
[K] Transcription
186
llmg_0433
rheB
llmg_0435
hexR
2.0E-02
-2.3
8.9E-05
-2.5
2.3E-05
-2.0
9.2E-04
1.7
3.5E-02
2.5
1.4E-03
2.1
3.1E-03
1.8
1.7E-02
1.0E-03
-1.3
3.6E-02
-1.4
1.8E-02
-2.2
4.1E-02
1.6
2.0E-02
1.5
2.8E-02
-1.7
6.4E-03
-1.5
3.8E-02
-1.4
1.8E-02
-1.5
5.8E-03
-1.4
6.1E-03
1.3
4.5E-02
1.6
2.3E-02
-1.4
2.4E-02
-1.7
8.9E-03
-1.8
4.3E-03
1.7
6.1E-03
-1.6
8.6E-03
-1.7
2.6E-02
1.4
-1.6
3.4E-02
llmg0503
-1.3
4.2E-02
rpoD
-1.5
2.1E-02
llmg_0522
llmg0522
-1.3
3.3E-02
1.5
1.1E-02
llmg_0610
greA
1.3
1.7
-1.4
3.9E-02
1.5
2.9E-02
1.8E-02
llmg_0521
rpoE
3.8E-02
-1.8
llmg_0503
llmg_0608
1.3
3.8E-03
3.5E-04
llmg0439
llmg0572
3.0E-03
2.6E-02
1.5E-06
-1.6
-1.7
lytR
llmg0576
-1.4
-1.3
-3.3
3.2E-04
llmg_0461
llmg_0576
1.3E-02
-1.6
llmg_0439
llmg_0572
-1.8
1.6
1.3E-02
1.7
3.7E-03
-1.4
2.9E-02
1.5
1.2E-02
1.6
1.2E-02
3.5E-02
2.7E-03
1.4
2.9E-02
-1.5
3.5E-02
1.6
1.4E-02
1.3
2.7E-02
1.6
4.6E-03
-1.4
1.1E-02
-1.5
1.9E-03
1.8
1.3E-03
1.9
1.9E-03
1.6
4.9E-03
1.8
4.0E-03
1.5
2.5E-02
llmg_0614
ctsR
1.7
4.6E-03
2.0
llmg_0626
llmg0626
1.6
2.2E-02
1.7
9.9E-03
llmg_0709
llmg0709
1.7
3.2E-03
1.5
2.1E-02
llmg_0733
llmg0733
1.4
1.7E-02
llmg_0746
malR
1.6
5.0E-02
1.4
2.3E-02
1.6
-1.4
8.5E-03
2.5E-02
1.0E-03
1.3
3.9E-02
-2.0
5.3E-03
llmg_0747
llrF
1.5
2.0E-02
llmg_0775
ccpA
3.2
7.7E-03
llmg_0784
rbsR
-1.5
4.7E-03
llmg_0793
ps304
llmg_0797
ps308
1.5
2.0E-02
-2.2
6.2E-03
1.6
4.7E-03
2.2
6.2E-04
-1.5
2.4E-02
1.4
-2.0
12.8
1.7E-10
1.4E-02
3.2E-02
3.3E-02
llmg_0857
kdgR
llmg_0888
rarA
llmg_0947
cmhR
llmg_0966
rmaI
llmg_1001
xylR
llmg_1005
llmg1005
-1.4
7.8E-03
-1.3
7.3E-03
-1.3
4.5E-02
llmg_1054
llmg1054
1.8
3.3E-03
2.3
3.9E-05
2.2
1.3E-04
llmg_1065
rgrB
-1.5
9.1E-03
llmg_1068
llrB
-1.5
3.6E-02
llmg_1190
llmg1190
1.6
3.0
-1.4
1.7E-02
4.4E-02
2.0
-1.4
llmg_1193
scpB
1.5
3.2E-02
llmg_1204
hdiR
1.6
2.0E-02
llmg_1224
llmg1224
-1.3
5.0E-02
llmg_1238
cspD2
-1.3
2.6E-02
1.5
2.3
2.7E-02
4.8E-02
1.4
3.8E-02
1.6
1.2E-02
6.9E-03
llmg_1252
llmg_1255
cspC
llmg_1256
cspD
llmg_1304
vacB2
-1.4
3.0E-02
-1.4
2.0E-02
1.6
-1.4
4.0E-02
1.1E-02
llmg_1324
llmg1324
-1.4
2.9E-02
llmg_1359
orf47
-7.4
7.6E-09
llmg_1425
llmg1425
llmg_1462
llmg1462
-1.8
4.0E-04
llmg_1576
hrcA
2.3
4.0E-02
llmg_1586
vacB1
1.5
4.1E-03
1.9
1.9E-02
llmg_1602
llmg1602
llmg_1627
rmaH
llmg_1674
mleR
llmg_1686
ahrC
llmg_1731
copR
llmg_1740
llmg1740
-1.3
5.0E-02
-1.5
7.0E-03
1.6
-3.2
1.2E-03
1.4E-02
1.5
6.7E-03
1.4
-1.3
2.3E-02
-1.3
2.0E-02
1.9
1.8E-02
4.7
1.8E-08
-1.4
2.5E-02
-1.3
-1.7
-1.7
llmg_1753
rnc
-1.3
4.4E-02
llmg_1794
llmg1794
1.8
4.2E-03
1.8
3.7E-03
1.8
9.7E-03
2.0
1.1E-03
llmg_1796
nusA
llmg_1826
llmg1826
1.3
3.5E-02
llmg_1846
cspB
-1.6
2.9E-03
llmg_1847
cspA
llmg_1860
rmaB
6.3E-03
4.5E-02
2.9E-04
1.8E-03
-1.3
1.5
4.5E-02
-1.4
3.3E-02
3.5E-10
-1.4
3.4E-02
8.9E-04
3.7E-02
1.3
12.3
-1.6
5.6E-03
1.4
4.0E-02
1.4
3.1E-02
1.3
4.8E-02
1.4
2.2E-02
1.6
1.0E-02
2.4
3.6E-03
1.4
3.7E-02
1.9
6.1E-03
-1.7
6.9E-04
-8.4
8.9E-10
1.4
4.8E-02
1.5
2.6E-02
-5.1
3.5E-06
llmg_1868
llmg1868
llmg_1878
nusB
llmg_1981
rpoC
llmg_1982
rpoB
1.6
llmg_2021
dinG
1.5
llmg_2067
rlrB
llmg_2134
ps409
3.0E-07
1.4
1.0E-02
-1.7
1.5E-02
-1.7
1.1E-02
1.5
1.3E-02
-1.7
5.4E-03
1.9
2.8E-02
1.6
3.3E-02
1.3
1.7E-02
-7.7
1.9E-09
-4.4
1.5E-05
3.2
1.6E-05
-1.3
3.9E-02
3.4
7.6E-06
-1.5
6.0E-03
-1.7
5.9E-03
-1.6
1.3E-02
-1.4
5.0E-02
-1.6
1.4E-02
1.3
3.1E-02
1.6E-02
4.2E-02
1.4
1.3
7.8
1.5E-02
4.3E-02
1.7
3.6E-02
7.3E-03
1.7
6.5E-04
4.2E-02
1.8
2.4E-03
1.3
3.5E-02
1.8
8.9E-04
1.6
1.4E-02
1.5
-3.6
2.2E-06
-2.7
1.4E-04
-2.5
2.9E-04
-1.4
3.9E-02
1.7
5.8E-03
1.3
3.1E-02
2.5
9.7E-06
1.8
1.7
-1.4
2.4E-03
2.4E-03
3.5E-02
187
llmg_2154
rpoZ
1.5
1.2E-02
1.5
llmg_2198
padR
1.5
1.9E-02
1.4
2.0E-02
1.5
1.5E-02
llmg_2238
llmg2238
-1.6
7.0E-04
-1.4
1.5E-02
-1.4
1.7E-02
llmg_2262
ps507
1.4
1.4E-02
llmg_2315
argR
-1.4
1.1E-02
-1.3
4.9E-02
1.4
7.2E-03
llmg_2339
llmg2339
1.6
5.2E-03
1.7
2.0E-03
llmg_2354
rpoA
1.5
1.5E-02
llmg_2388
nusG
llmg_2453
llmg2453
-1.4
7.0E-03
llmg_2470
gntR
llmg_2485
glnR
llmg_2517
llmg2517
1.5
8.0E-03
llmg_2542
llmg2542
-1.3
4.8E-02
1.6
1.3E-02
2.8
1.3E-04
-1.3
1.3E-02
2.0
2.7E-04
3.7E-03
2.5
7.1E-04
1.8
1.6E-03
-1.4
3.4E-02
-1.4
2.6E-02
1.7
2.3E-03
1.5
3.9E-02
-1.3
4.0E-02
-1.4
3.7E-02
1.7
2.6E-03
-1.3
2.0E-02
2.4
1.2E-03
-1.4
6.8E-03
1.8
2.1E-03
-1.7
3.2E-03
-2.3
2.7E-05
[L] Replication, recombination and repair
188
llmg_0001
dnaA
llmg_0002
dnaN
1.6
3.1E-02
llmg_0004
rexA
1.3
1.0E-02
llmg_0055
ps101
llmg_0067
recO
1.5
-1.7
llmg_0103
cshA
llmg_0151
llmg0151
llmg_0244
ung
llmg_0359
recR
llmg_0374
recA
llmg_0409
ssbA
llmg_0416
holB
llmg_0444
ligA
llmg_0483
dnaE
llmg_0496
hllA
llmg_0520
dnaG
llmg_0534
uvrB
1.4
1.9
6.7E-04
1.8
3.5E-03
1.7
7.7E-03
1.6
3.7E-02
1.5
5.1E-04
1.3
3.6E-02
-1.4
2.3E-02
-1.4
2.0E-02
1.6E-02
1.6
3.9E-03
1.5
1.8E-02
2.4E-03
-1.6
8.0E-03
-1.5
3.4E-02
5.1E-04
-1.5
2.9E-03
1.5
1.4E-02
-2.2
2.2E-04
2.4
1.2
-1.4
4.3E-02
llmg_0594
ps201
recJ
-1.4
1.4E-02
3.9E-03
9.1E-04
3.5
9.1E-05
2.9
8.6E-04
-1.4
2.6E-02
-1.3
4.9E-02
1.4
1.7E-02
3.1
3.1E-04
1.5
5.6E-03
-1.6
3.4E-03
1.5
4.8E-03
1.3
2.1E-02
1.3E-03
1.3
4.2E-02
3.7E-02
4.4E-02
1.4
2.7E-02
-1.4
4.5E-03
3.2E-02
9.4E-06
-1.5
-1.3
1.3
4.0
2.0
llmg0655
sbcC
-1.7
1.7E-02
holA
llmg_1133
1.8E-02
2.4E-03
3.2E-02
llmg_0768
3.9E-02
4.8E-02
-1.4
1.7
1.3
llmg_0655
-1.4
6.9E-04
1.7
-1.3
sbcD
3.2E-03
1.5
3.0E-02
xerD
ps333
2.0
1.6E-02
5.0E-04
llmg_0618
llmg_0824
9.8E-05
1.6
1.3
3.3E-03
llmg_1132
2.5
2.2
1.8E-03
ps301
3.9E-02
4.5E-02
-1.5
llmg_0790
-1.3
1.5E-02
-1.5
5.5E-03
6.3E-03
3.1E-02
1.6
llmg0611
1.5
-1.7
-1.3
1.4
llmg0612
mutS
6.2E-03
3.8E-02
llmg_0611
llmg0771
1.4E-02
1.7
3.1E-04
-1.6
-1.6
-1.5
-1.7
-1.4
llmg_0612
llmg_0771
1.5E-02
-1.6
1.1E-02
llmg_0606
llmg_0778
-1.7
3.5E-02
llmg_1176
rnhB
1.7
6.2E-03
xerD2
-1.3
2.5E-02
llmg_1268
llmg1268
-1.3
3.3E-02
llmg_1270
llmg1270
1.4
2.0E-02
-1.3
1.5E-02
1.6
3.6E-03
-1.3
3.2E-02
2.2
2.3E-02
1.6
4.0E-03
1.3
4.4E-02
-1.3
3.4E-02
-1.4
2.1E-02
-1.3
3.5E-02
1.6
5.6E-03
1.4
1.7E-02
1.5
8.5E-03
1.7
1.0E-02
1.7
9.4E-03
1.7
1.3E-02
-1.3
4.5E-02
2.0E-02
llmg_1191
1.0E-03
llmg_1272
topA
-1.4
3.1E-02
llmg_1445
pcrA
1.7
9.0E-03
1.6
2.3E-02
1.8
llmg_1451
gyrA
1.4
4.9E-02
1.4
1.7E-02
llmg_1483
comFA
-1.5
4.5E-03
-1.4
1.4E-02
1.6
2.0E-02
2.1
4.8E-04
llmg_1502
dnaD
llmg_1515
radC
-1.5
4.7E-02
-1.4
1.8E-02
llmg_1539
parE
-1.4
llmg_1566
llmg1566
1.8
3.0E-03
llmg_1612
llmg1612
-2.2
4.2E-06
-3.0
3.4E-09
llmg_1631
gyrB
1.4
2.7E-02
1.8
3.0E-04
llmg_1654
tnp981
llmg_1657
xerD
llmg_1661
hslB
llmg_1691
xseB
-2.2
6.0E-06
-1.7
2.2E-02
1.7
3.8E-03
1.6
2.3E-02
2.4
xseA
mutY
1.5
9.2E-03
llmg_1718
uvrC
-1.4
4.2E-02
-1.4
1.2E-02
llmg_1765
exoA
-1.6
3.1E-02
llmg_1808
dnaI
-1.4
2.5E-02
-1.3
3.2E-02
llmg_1814
dnaC
1.6
4.6E-04
1.6
8.0E-04
llmg_1893
tnp904
1.4
3.3E-02
llmg_1922
recD
1.4
5.7E-02
1.7
2.6E-04
1.8
3.6E-04
llmg_1955
comEA
-1.3
4.0E-02
-1.4
2.4E-03
llmg_1986
llmg1986
recQ
llmg2007
-1.5
5.3E-03
-1.4
uvrA
llmg_2010
llmg2010
llmg_2075
llmg2075
1.4
2.7E-02
1.6
1.7E-03
llmg_2102
ps440
-1.2
4.9E-02
-1.3
1.6E-02
2.6E-02
1.5
8.4E-03
1.3
2.5E-02
llmg_2109
llmg2109
1.4
6.3E-03
-1.4
1.9E-02
llmg_2110
tnp712
-1.3
2.2E-02
-1.4
6.6E-03
llmg_2142
ps401
llmg_2153
priA
llmg_2268
ps501
1.5
1.1E-02
llmg_2290
llmg2290
-1.3
4.4E-02
llmg_2305
dinP
1.6
2.4E-03
1.5
1.2E-02
1.4
4.1E-02
-1.8
9.2E-04
snf
-1.4
4.8E-02
polC
1.4
2.7E-02
-1.8
6.8E-03
polA
-1.7
1.4E-02
ruvB
1.3
4.7E-02
-2.0
2.3E-03
1.8
3.8E-02
6.4E-03
1.6
2.8E-03
1.4
2.7E-02
-1.7
1.4E-02
-1.6
1.3E-02
-1.5
1.2E-02
1.8
6.9E-04
1.5
1.1E-02
1.8
2.3E-03
2.0
-1.5
2.1E-02
llmg_2488
ruvA
1.5
7.6E-03
llmg_2491
mutS
1.5
1.1E-02
1.5
1.5E-02
llmg_2493
tnp905
-1.7
2.8E-03
-1.7
7.7E-04
-1.9
4.5E-04
llmg_2523
recG
1.5
7.8E-03
1.5
1.2E-02
-1.4
3.6E-02
llmg_2525
ps601
1.6
9.9E-03
1.6
2.1E-02
1.7
2.4E-03
llmg_2549
rnhA
1.4
3.8E-02
4.0E-04
1.7
4.5E-03
1.6
3.7E-02
-1.4
1.9E-02
-1.6
2.2E-03
-1.3
4.4E-02
-1.0
2.1E-02
1.6
-1.8
4.3E-02
1.5
2.5E-02
1.4
4.2E-02
-1.4
3.5E-02
1.4
6.9E-03
-1.6
2.7E-03
1.3
4.8E-02
-1.3
4.2E-02
-1.4
1.8E-02
1.7
6.5E-04
1.4
5.0E-02
-1.1
1.4E-02
-1.4
1.9E-02
1.7
3.7E-02
-1.9
1.1E-03
-2.5
4.2E-05
-1.5
7.7E-03
-1.5
6.2E-03
1.4
4.1E-02
-1.4
2.2E-02
1.3
llmg_2409
llmg_2425
1.6
1.4
llmg_2319
llmg_2487
4.2E-02
2.4E-02
llmg_2008
1.3
1.4
1.3E-05
llmg_1717
llmg_2007
1.8E-03
1.8E-02
llmg_1692
llmg_1992
1.7
6.2E-03
3.4E-03
4.2E-04
2.5E-02
1.4
2.0E-02
1.5
9.4E-03
-1.3
3.9E-02
-1.3
2.7E-02
-1.4
4.7E-02
-1.5
4.9E-02
-1.4
4.8E-02
-1.6
4.9E-03
-1.6
9.8E-03
-1.4
2.4E-02
-1.3
5.0E-02
189
[M] Cell wall/membrane/envelope biogenesis
llmg_0104
llmg0104
llmg_0114
llmg0114
-1.4
3.6E-02
-1.4
2.7E-02
llmg_0134
llmg0134
-1.4
2.6E-02
llmg_0156
dltE
-1.4
2.4E-02
llmg_0162
llmg0162
1.9
5.4E-04
llmg_0206
rmlA
-1.7
4.1E-03
llmg_0207
rmlC
1.3
4.6E-02
llmg_0209
rmlB
1.5
1.5E-02
llmg_0210
rmlD
1.9
8.5E-04
llmg_0211
rgpA
2.1
llmg_0212
rgpB
1.3
2.7E-02
llmg_0213
rgpC
1.5
1.3E-02
1.9
3.0E-02
llmg_0216
rgpE
1.8
8.0E-03
1.8
3.9E-03
llmg_0217
rgpF
1.5
3.5E-02
1.5
4.1E-02
llmg_0221
llmg0221
1.5
2.7E-02
llmg_0229
llmg0229
llmg_0246
llmg0246
190
llmg_0326
murA1
1.3
2.9E-02
llmg_0358
pbp2B
1.4
3.9E-02
3.6E-02
5.4E-03
-1.8
1.5E-03
1.8
6.6E-05
-1.6
8.2E-04
1.5
8.3E-03
1.6
2.0E-03
-1.3
4.2E-02
1.4
6.8E-03
1.6
1.5E-03
1.4
4.1E-02
-2.2
7.7E-03
-1.6
3.1E-02
1.4
4.5E-02
-1.5
2.8E-03
1.5
8.7E-03
-2.0
1.4E-02
1.5
2.9E-02
llmg_0511
ponA
-1.8
1.2E-02
-2.0
2.3E-03
1.8
3.2E-03
-1.4
7.2E-04
ps356
2.2E-02
-1.3
1.7
1.5E-02
icaA
1.5
2.6E-02
1.7
llmg_0646
4.7E-02
3.0E-02
1.4
pbp1B
llmg_0851
-1.5
-1.5
-1.7
1.8
1.8E-03
-1.5
2.5E-03
2.0
-1.4
1.6
3.5E-02
1.4E-02
5.9E-03
1.9E-03
-2.6
9.5E-05
1.5
3.8E-02
1.2E-02
-1.5
2.9E-02
1.6
1.2E-02
-1.3
4.7E-02
-1.4
1.6E-02
2.3
2.6E-04
1.8
6.7E-03
-1.5
3.6E-02
-1.6
1.6E-02
-1.6
3.0E-02
1.5
3.7E-02
1.3
9.0E-03
4.3E-03
1.6
1.5E-02
-1.4
1.9E-02
-1.4
1.6E-02
-1.5
2.7E-02
-1.4
1.8
-1.4
2.4E-02
4.9E-07
llmg_0402
llmg0625
4.0E-02
9.5E-03
murF
llmg_0625
1.4
1.5
llmg0360
llmg0600
1.8E-02
3.6E-02
llmg_0360
murA2
1.5
1.4
llmg_0361
llmg_0517
3.6E-02
-1.4
llmg_0214
llmg_0600
-1.4
4.8E-02
3.5E-02
llmg_0913
murG
1.4
1.6E-02
1.4
llmg_0914
ftsQ
1.5
3.4E-03
-1.3
2.1E-02
llmg_0992
llmg0992
-1.7
5.4E-03
-1.5
4.0E-02
llmg_1109
gidB
1.5
3.9E-02
1.6
1.5E-02
1.7
3.8E-03
1.5
1.3E-02
llmg_1113
galU
1.7
8.5E-04
1.9
4.9E-05
1.4
4.5E-02
-1.5
1.1E-02
1.7
9.1E-03
2.6
8.2E-06
llmg_1180
llmg1180
llmg_1215
llmg1215
llmg_1220
dltB
llmg_1222
dltD
llmg_1240
llmg1240
llmg_1329
murB
llmg_1417
llmg1417
-1.8
-1.7
6.2E-03
1.3
4.4E-02
4.8E-05
1.9
9.5E-03
-1.4
4.6E-02
llmg_1422
pacA
llmg_1449
srtA
llmg_1463
lepA
-1.8
1.4E-02
1.0
6.8E-04
llmg_1465
murC2
-1.4
1.9E-02
-1.4
2.7E-02
llmg_1467
llmg1467
-1.3
4.2E-02
-1.5
2.2E-03
1.4
3.0E-02
2.0
5.1E-04
llmg_1486
dgkA
llmg_1516
glmS
-1.8
6.9E-05
4.0
4.6E-08
1.5
2.3E-02
1.6
2.2E-02
1.8
3.4E-03
1.6
2.2E-02
-1.6
8.2E-04
1.5
3.1E-02
-1.5
-1.6
4.2E-03
-1.4
2.6E-02
1.7
2.0
2.6
7.0E-06
1.2E-02
7.2E-04
2.9E-02
-1.7
3.3E-02
-1.8
1.4E-04
-1.5
2.5E-02
3.0
1.2E-05
2.6
9.5E-05
1.4
5.6E-03
1.6
1.3E-02
1.9
7.6E-04
2.0
2.7E-04
1.4
2.7E-02
1.4
2.5E-02
-1.6
2.9E-02
6.8E-03
7.3E-03
2.0
-1.6
llmg_1521
llmg1521
1.4
4.7E-02
llmg_1552
llmg1552
-1.6
7.5E-03
-1.7
3.4E-03
llmg_1578
dacB
1.7
1.3E-03
1.7
3.1E-03
llmg_1594
llmg1594
1.8
3.6E-03
-1.3
1.2E-02
llmg_1603
tagB
-1.5
5.4E-04
llmg_1604
tagF
-1.3
3.3E-02
llmg_1616
ugd
llmg_1621
llmg1621
-1.3
3.1E-02
llmg_1667
llmg1667
-1.4
2.7E-02
llmg_1676
llmg1676
2.5
1.5E-03
llmg_1678
mraY
llmg_1679
pbpX
llmg_1681
mraW
llmg_1699
choS
llmg_1704
alr
llmg_1708
llmg1708
llmg_1751
llmg1751
llmg_1799
lrgB
llmg_1801
srtC
llmg_1959
bar
llmg_1976
llmg1976
-1.4
2.1E-02
1.6
1.4E-03
llmg_1989
murE
llmg_2076
glmU
llmg_2082
ps461
1.7
4.5E-03
llmg_2161
cfa
1.6
6.4E-03
llmg_2233
galE
1.7
1.5E-03
llmg_2316
murC
llmg_2344
llmg2344
llmg_2350
llmg2350
llmg_2351
llmg2351
llmg_2385
mscL
llmg_2392
pbp2A
llmg_2420
llmg2420
llmg_2421
llmg2421
1.5
1.1E-02
-1.4
9.6E-03
-1.3
2.0E-02
3.6
4.2E-06
-1.7
1.1E-02
-1.4
1.3E-02
-1.3
2.6E-02
-1.5
4.1E-04
1.8
1.5E-04
-1.3
3.1E-02
1.7
3.0E-03
1.2
1.3E-02
2.2
2.4E-04
1.4
2.4E-02
-2.0
2.0E-04
-1.3
1.4E-02
-1.4
1.5E-02
-1.9
5.3E-04
2.0
2.2E-02
1.5
1.1E-02
-2.6
1.3E-03
1.5
6.3E-03
3.0
1.8E-05
1.8
1.1E-02
1.3
5.0E-02
1.6
3.6E-03
1.6
7.9E-03
-1.7
1.0E-02
-2.1
2.7E-03
1.3
4.6E-02
-1.5
4.4E-03
-1.4
1.5
2.6E-02
-1.5
8.6E-03
1.5
1.8E-02
-1.5
2.1E-02
1.5
2.6E-02
-1.4
1.9E-02
2.1
1.2E-02
-1.7
-1.4
-1.5
2.7E-03
-1.5
5.1E-04
1.6
5.0E-03
-1.3
1.8E-02
2.1
-1.4
7.7E-04
4.8E-03
2.3E-03
3.6E-02
-3.3
2.3E-05
1.4
2.7E-02
3.1
7.4E-06
-1.5
1.0E-02
1.9
4.8E-03
-1.8
1.1E-03
-1.6
3.6E-02
-1.3
4.2E-02
-1.4
3.0E-02
2.5E-02
-1.5
1.3E-02
1.6
1.9E-02
1.6
3.2E-02
1.6
6.7E-03
1.5
8.9E-03
1.6
1.6E-02
1.7
3.5E-03
2.2
1.5E-04
1.5
4.8E-02
1.5
1.0E-02
1.6
3.8E-03
1.6
6.9E-03
llmg_2508
llmg2508
llmg_2509
mreC
1.5
1.8E-02
2.3
1.6
6.3E-03
7.6E-04
1.9
1.1E-02
llmg_2547
llmg2547
-1.9
2.8E-03
llmg_2560
dacA
-1.6
3.9E-02
-2.4
6.8E-05
-1.8
9.1E-03
1.7
1.3E-02
1.9
1.9E-03
1.6
2.7E-02
-1.5
1.5E-02
-1.8
1.1E-03
-2.2
6.6E-03
1.5
3.5E-02
1.4
1.1E-03
-2.7
4.3E-04
2.5
1.6E-04
-1.4
4.8E-02
1.5
4.3E-02
-2.1
3.2E-04
-2.2
2.0E-05
[N] Cell motility
llmg_0280
acmA
llmg_0509
acmD
1.7
6.4E-03
2.5
7.7E-07
-1.4
2.2E-02
1.8
2.7E-04
llmg_0836
ps342
-1.6
2.6E-02
-1.5
3.9E-02
llmg_1977
llmg1977
1.5
2.5E-02
2.2
2.3E-03
2.3
7.0E-03
-2.1
2.3E-02
llmg_2165
acmB
-1.5
3.5E-02
-1.5
3.3E-02
-1.6
2.8E-04
1.3
2.9E-02
llmg_2304
comC
1.4
3.3E-02
llmg_2407
comGB
-1.5
7.2E-03
-1.6
2.5E-02
191
[O] Posttranslational modification, protein turnover, chaperones
llmg_0021
ftsH
1.8
3.2E-03
2.7
1.7E-05
3.4
1.7E-06
llmg_0080
osmC
-2.3
2.2E-04
-2.8
5.3E-06
-5.1
2.8E-11
llmg_0201
msrB
-1.6
1.5E-02
1.4
4.2E-02
llmg_0282
nrdG
1.7
2.0E-03
llmg_0306
llmg0306
llmg_0309
gcp
-1.5
5.0E-02
llmg_0318
tpx
1.6
1.4E-02
llmg_0356
ahpC
llmg_0357
ahpF
llmg_0410
groES
1.6
llmg_0411
groEL2
llmg_0519
tig
llmg_0528
clpE
llmg_0593
llmg0593
llmg_0615
clpC
llmg_0663
pcp
llmg_0702
pepO
llmg_0776
trxB2
llmg_0951
msrA
-1.8
llmg_0986
clpB
1.9
llmg_1088
gpo
llmg_1340
clpX
llmg_1400
orf9
2.1
5.7E-05
1.4E-02
1.5
1.1E-02
2.0
3.6E-03
-1.6
3.3E-02
6.7E-03
1.6
5.8E-03
-1.7
6.3E-03
2.3
3.0E-04
2.3
7.2E-04
-2.6
5.3E-05
1.6
1.6E-02
3.9E-05
4.1E-02
1.0E-05
2.8
4.9E-06
2.8E-02
-1.7
3.7E-02
-2.0
7.9E-03
7.8E-03
2.1
2.8E-03
-2.5
6.3E-03
1.6
6.7E-03
-1.3
2.9E-02
1.7
5.5E-03
3.0E-03
2.8
1.1E-04
2.3E-03
-1.9
3.5E-02
llmg_1575
grpE
1.7
7.5E-03
1.7
3.0E-03
llmg_1646
ppiB
1.7
3.2E-02
llmg_1902
smpB
1.4
2.9E-02
llmg_1907
pmpA
1.6
1.1E-02
1.6
4.6E-03
llmg_1969
sufB
1.9
5.2E-03
llmg_1973
llmg1973
1.7
1.4E-03
llmg_1974
llmg1974
llmg_1985
pepO2
1.7
3.3E-03
llmg_2204
hslO
1.7
3.1E-03
-1.6
1.8E-05
-1.3
2.6E-02
dnaJ
1.4
2.4E-05
2.3
htrA
1.9E-02
3.7E-02
2.2
llmg_2419
-1.5
1.6
7.6E-04
llmg_2502
4.4E-02
4.7E-05
2.5
llmg2243
-1.4
2.0
nrdH
llmg2244
9.4E-04
2.8
dnaK
llmg_2244
1.8
1.4E-03
llmg_1574
llmg_2243
4.2E-02
3.1E-02
3.0E-02
2.6E-03
2.6E-02
8.3E-05
1.5
1.4
1.7
llmg_1541
1.4
2.5E-03
-2.5
1.7
2.4
2.6E-02
2.4
3.9E-06
1.1E-03
1.9
1.3
4.0E-06
3.2E-06
4.8E-02
1.8E-04
3.8
-3.0
2.3
-1.4
2.1
1.8E-02
2.4
3.4
1.3
1.7
-1.3
4.1E-02
-1.6
2.7E-03
1.5
4.9E-02
1.4
2.0E-02
5.5E-03
3.0E-02
1.6
1.7E-02
1.5
3.2E-02
5.1E-03
1.4
2.6E-02
6.5E-05
-4.4
9.3E-07
-8.4
2.3E-11
-1.4
1.0E-02
1.2E-04
1.9
4.3E-03
-3.2
4.8E-05
-3.1
5.2E-05
1.3
3.3E-02
2.2E-02
9.7E-03
1.6
-2.2
3.2E-05
1.8
2.5E-02
3.7E-02
2.0E-03
3.1E-03
1.7
-1.8
2.5
1.6E-03
1.5
1.8E-03
-1.9
1.7
1.7E-02
2.1E-03
3.8E-02
-1.5
1.8
2.5
-1.9
-1.4
8.8E-03
1.8
1.5E-02
2.7E-04
6.9E-04
1.5
1.4
3.0
1.6
3.0E-04
2.1E-02
5.6E-03
-1.6
2.0
1.4
-1.9
2.4
1.0E-03
-3.3
1.1E-05
-3.3
8.4E-06
-1.6
1.2E-02
-1.8
1.7E-03
1.7
1.6E-03
1.6
1.2E-02
1.7
5.6E-03
-1.7
2.3E-03
-1.7
2.7E-02
1.7
4.6E-03
-1.6
6.6E-03
1.7
5.3E-03
-1.7
2.1E-03
-1.4
1.8E-03
1.3
1.9E-02
1.6
3.1E-02
-1.7
1.5E-02
-2.1
4.4E-02
-3.3
2.1E-04
-1.4
3.1E-02
-1.4
2.5E-02
1.7
5.7E-04
[P] Inorganic ion transport and metabolism
192
llmg_0100
cadA
llmg_0132
sugE
-3.3
llmg_0199
feoB
1.4
3.1E-02
llmg_0200
feoA
1.7
1.5E-03
1.5
1.2E-02
llmg_0289
cbiQ2
1.6
9.2E-04
1.6
9.5E-04
llmg_0312
phnD
1.5
4.2E-03
1.7
7.3E-04
3.2
1.6E-08
llmg_0313
phnC
2.1
1.4E-04
2.6
1.9E-06
2.6
2.4E-06
llmg_0314
phnB
1.9
4.7E-03
2.5
4.2E-05
2.1
1.5E-03
-4.3
7.5E-06
1.4
4.4E-02
-1.5
4.9E-03
-1.5
4.2E-03
3.4
5.3E-07
3.0
3.8E-06
llmg_0315
llmg0315
llmg_0322
llmg0322
1.5
3.5E-02
llmg_0335
plpA
1.0
3.1E-02
1.5
2.0E-02
1.5
1.7E-02
1.3
4.4E-02
llmg_0336
plpB
llmg_0338
plpC
-1.5
llmg_0342
llmg0342
llmg_0345
cbiQ
-1.3
2.4E-02
llmg_0346
fhuC
-1.4
2.8E-02
llmg_0349
fhuD
1.6
3.6E-02
llmg_0429
sodA
-1.9
8.5E-03
llmg_0547
ctpE
llmg_0588
kupB
2.4
7.3E-03
llmg_0628
llmg0628
-1.4
3.7E-02
-1.5
-2.0
1.4E-02
1.8E-06
llmg_0640
trmA
llmg_0643
pacL
1.6
4.8E-04
llmg_0661
llmg0661
1.8
4.0E-03
llmg_0666
oxlT
llmg_0690
llmg0690
-1.5
1.6E-02
-1.3
3.6E-02
llmg_0692
llmg0692
-1.5
8.2E-03
-1.4
2.2E-02
llmg_0699
oppB
-2.2
3.3E-04
llmg_0910
amtB
1.8
1.6E-03
2.3
3.6E-05
-1.5
4.7E-03
-1.4
1.9E-02
llmg_1016
llmg1016
-1.6
1.2E-03
llmg_1023
fur
-1.6
4.1E-02
llmg_1086
llmg1086
llmg_1116
telA
llmg_1130
llmg1130
llmg_1137
mtsC
llmg_1138
mtsA
llmg_1155
trmA
llmg_1223
mgtA
llmg_1248
arsA
llmg_1250
llmg1250
1.8
2.0
-1.5
3.4E-03
6.9E-04
4.0E-04
1.9
3.2E-03
2.0
2.8E-03
8.0E-04
3.3
6.9E-05
2.4
2.6E-03
1.2
1.7E-02
-1.7
1.8E-02
-1.7
1.1E-02
-1.5
3.3E-03
-2.1
1.5E-03
-1.8
9.1E-03
-1.4
2.8E-02
-1.6
4.0E-02
4.2E-03
1.4E-03
4.6E-05
-1.4
2.3E-02
1.9
2.6E-03
1.3
3.7E-02
1.4
1.8E-02
-1.4
2.6E-02
-1.5
1.9
1.9E-03
-1.8
9.0E-03
-1.7
1.9E-02
1.8
4.9E-05
-1.4
1.5E-02
1.3
9.1E-03
1.7
2.6E-03
1.6
1.2E-02
-1.4
2.8E-03
1.3
3.3E-02
1.3
3.9E-02
1.7E-04
llmg_1460
llmg1460
-1.3
2.0E-02
-1.3
4.3E-02
llmg_1490
mntH
2.1
llmg_1513
llmg1513
-1.4
1.1E-02
-1.4
llmg_1533
llmg1533
llmg_1561
llmg1561
-1.3
3.4E-02
1.4
3.0E-02
1.7
1.1E-02
1.4
2.6E-02
1.4
3.0E-02
1.5
1.6E-02
noxC
3.0
2.3
1.4E-03
1.7
llmg1768
3.5E-05
-2.0
2.1
2.2E-04
llmg_1768
3.8
1.8E-03
5.0E-05
2.7E-02
4.7E-02
1.8
2.3
-2.0
-1.3
2.2E-02
7.2E-03
5.1E-03
1.0E-02
1.7
-1.7
1.7
llmg_1770
3.4E-03
-1.6
1.5
6.6E-03
1.5E-02
orf53
llmg1703
-2.0
4.0E-02
1.6E-02
1.0
-1.4
llmg1318
copA
2.7E-05
2.7E-02
3.4E-02
llmg_1318
llmg_1729
3.5
1.5
-1.6
llmg_1353
llmg_1703
3.7E-02
1.5
-1.5
1.1E-02
-1.3
4.6
4.8E-02
1.5
1.6E-02
1.5
4.9E-03
1.5E-04
2.2
1.3E-04
5.3E-03
-1.3
2.5E-02
1.5
1.5E-02
llmg_1771
llmg1771
llmg_1896
pstA
1.9
1.8E-02
llmg_1898
pstC
1.6
1.6E-02
llmg_1899
pstD
llmg_1900
pstE
1.5
llmg_1960
llmg1960
1.7
1.5
3.9E-03
1.4
1.9E-03
1.5
9.7E-04
1.4
1.1E-02
1.9
8.2E-04
2.0
7.4E-04
1.5
1.2E-02
-1.5
2.1E-02
1.6
-1.4
1.9E-02
3.1E-02
1.7
1.0E-02
1.7
6.6E-03
-1.5
4.6E-03
-1.7
3.8E-03
-1.5
1.7E-02
-1.6
5.4E-03
-1.4
3.3E-02
7.7E-03
1.4
2.6E-02
1.4
1.5E-02
1.6
1.5E-03
1.4E-02
2.3
2.5E-04
1.6
4.8E-02
2.1
1.2E-03
-1.6
4.1E-02
-1.5
1.6E-02
-1.7
3.9E-02
-1.9
4.3E-04
1.6
3.0E-02
193
llmg_2171
mntA
-1.4
4.8E-02
llmg_2203
cadD
1.8
3.3E-02
2.4
3.0E-04
1.6
4.2E-02
2.5
1.7E-04
llmg_2269
cutC
2.0
3.3E-04
2.3
4.9E-05
1.6
1.2E-02
1.9
2.0E-03
llmg_2302
dpsA
-1.5
2.5E-03
-1.4
9.2E-03
llmg_2398
ZitP
1.4
1.7E-02
1.5
1.6E-02
-1.4
llmg_2399
ZitQ
llmg_2400
ZitS
llmg_2541
llmg2541
1.4
4.3E-03
1.4
2.6E-02
1.4
9.6E-03
1.7
3.1E-02
1.2E-02
-1.3
2.4E-02
[Q] Secondary metabolites biosynthesis, transport and catabolism
llmg_0277
llmg0277
llmg_0339
dar
llmg_0734
llmg0734
-1.3
4.3E-02
llmg_1219
dltA
1.5
8.7E-03
llmg_1275
aldB
llmg_1464
aldC
llmg_1538
llmg1538
llmg_1585
llmg1585
llmg_1742
llmg1742
llmg_1762
llmg1762
llmg_1832
menE
llmg_1834
llmg1834
llmg_2543
llmg2543
-1.8
7.3E-04
-1.5
2.2E-02
-2.3
6.2E-06
-1.5
2.4E-02
-1.8
2.0E-03
-1.3
3.2E-02
1.3
-1.3
2.6E-02
1.4
1.5E-02
1.7
1.8E-02
5.0E-02
-1.3
3.8E-02
1.5
1.2E-02
1.7
7.3E-03
-1.6
3.8E-02
2.0
4.1E-03
1.5
3.8E-03
1.5
1.9E-02
1.9
1.5E-02
-1.6
2.7E-03
-1.7
9.7E-04
1.7E-02
-1.4
2.6E-02
-1.7
4.3E-02
-1.5
3.8E-02
-1.7
3.8E-03
-1.7
1.3E-03
-2.0
2.8E-04
-1.4
4.0E-02
1.4
3.0E-02
1.4
7.1E-03
1.4
4.2E-03
1.4
2.6E-02
1.8
3.4E-03
1.6
1.9E-02
2.2
1.3E-03
2.9
6.0E-06
1.6
8.3E-03
1.4
4.8E-02
-1.7
8.5E-03
-1.5
4.4E-02
-1.8
6.2E-03
-1.3
3.6E-02
1.5
3.3E-02
1.4
3.9E-02
[R] General function prediction only
194
llmg_0084
llmg0084
llmg_0092
llmg0092
llmg_0106
llmg0106
llmg_0117
llmg0117
llmg_0130
llmg0130
llmg_0144
llmg0144
llmg_0146
llmg0146
llmg_0154
cbr
llmg_0167
llmg0167
llmg_0182
llmg0182
llmg_0185
llmg0185
-1.8
-1.5
1.5
-1.5
-2.1
-1.5
9.1E-03
1.1E-05
2.4E-03
1.6
-1.4
3.5E-02
-1.5
4.9E-02
-1.6
4.9E-04
llmg_0194
llmg0194
llmg0202
llmg_0225
rfbX
1.5
2.1E-02
llmg_0233
hflX
-1.4
1.4E-02
llmg_0254
hadL
2.8E-02
-1.8
7.9E-04
llmg_0265
llmg0265
llmg_0276
llmg0276
llmg_0292
hipO1
2.5
llmg_0302
llmg0302
-1.6
llmg_0307
llmg0307
llmg_0308
rimI2
llmg_0344
cbiO
llmg_0352
llmg0352
-1.4
2.5E-02
-1.5
7.2E-03
-1.6
-1.6
1.5
4.1E-03
1.5
3.2E-02
1.8
9.8E-05
1.6
2.6E-03
1.5
3.4E-02
1.5
4.7E-03
-1.4
3.6E-02
1.4
2.4E-02
1.6
1.9E-02
1.6
1.8E-02
1.7
4.1E-03
-1.4
4.1E-03
1.4
1.8E-02
2.1E-02
-1.5
7.1E-04
-1.3
3.4E-02
4.6E-05
-2.1
8.3E-05
2.7E-05
1.5
2.1E-02
2.6E-02
-2.4
6.2E-05
-1.5
2.5E-03
1.2E-02
9.7E-03
4.6E-04
-2.1
1.1
-2.0
2.0E-02
1.7E-03
llmg_0202
-1.4
2.9E-03
-1.4
1.7E-02
-1.4
8.6E-03
2.3
1.5E-04
1.5
1.3
3.3E-02
-1.6
3.3E-02
-1.0
6.3E-03
-1.2
1.9E-02
-1.4
4.4E-02
8.0E-03
-1.3
2.6E-02
1.6
1.1E-03
llmg_0367
dppF
-1.4
5.9E-03
1.1
2.8E-02
llmg_0387
llmg0387
-1.4
3.9E-02
-1.3
3.0E-02
llmg_0407
pheT
-1.7
4.2E-02
-2.3
1.3E-03
llmg_0408
noxE
1.5
2.5E-02
-1.7
2.3E-03
llmg_0412
vicX
llmg_0419
llmg0419
llmg_0456
pgmB
1.7
3.9E-02
2.8
4.6E-05
llmg_0459
llmg0459
-1.5
2.5E-02
-1.8
7.2E-04
llmg_0479
llmg0479
-1.6
2.0E-02
-1.5
4.5E-02
llmg_0491
hly
-1.4
6.0E-03
llmg_0510
llmg0510
1.4
5.9E-03
1.3
2.0E-02
llmg_0533
nudH
llmg_0546
xylH
llmg_0561
llmg0561
llmg_0584
llmg0584
llmg_0605
llmg0605
llmg_0609
pabC
llmg_0651
llmg0651
llmg_0664
amd
llmg_0698
oppF
llmg_0742
maa
llmg_0761
llmg0761
llmg_0765
llmg0765
llmg_0772
llmg0772
llmg_0777
llmg0777
llmg_0825
ps334
llmg_0876
llmg0876
llmg_0880
llmg0880
llmg_0904
llmg0904
llmg_0933
llmg0933
2.3
1.8
2.5E-03
2.0
1.3E-03
-1.6
4.1E-02
1.2E-02
4.6E-02
2.5E-02
1.9
1.1E-02
-1.6
1.3
-1.4
1.5
-1.5
2.9E-02
1.7
9.3E-03
1.7
4.6E-03
1.4
1.7E-02
1.4
3.5E-02
-1.4
2.0E-02
2.2
1.1E-03
1.8
7.5E-03
-1.4
2.3E-02
1.6E-02
-1.6
2.2E-03
-1.4
2.9E-02
-1.7
1.6E-02
1.7
5.1E-03
-1.5
2.7E-02
-1.6
1.0E-02
-2.4
6.7E-06
-1.8
1.2E-02
-2.0
-2.6
2.3E-06
-1.6
2.6E-02
-2.5
1.2E-05
-1.3
3.6E-02
-1.6
4.7E-04
-1.6
4.2E-02
-2.2
7.5E-04
-1.7
2.1E-03
1.7
3.0E-02
-1.4
4.9E-02
-1.4
2.2E-03
2.1
-1.5
-2.0
2.3E-03
llmg_0948
llmg0948
1.5
5.6E-03
llmg_0950
llmg0950
-1.6
4.0E-03
llmg_0968
llmg0968
-1.3
4.0E-02
2.4
8.4E-05
llmg_0995
llmg0995
1.7
2.6E-02
llmg_1019
llmg1019
1.5
2.1E-02
2.7E-02
7.7E-03
3.7E-03
2.3E-02
1.7E-04
1.5
1.4
6.5E-04
-1.3
-2.6
-1.4
2.0
1.0E-02
5.7E-04
5.5E-03
-1.5
5.7E-04
2.0
2.4E-03
2.4E-03
2.0
3.9E-03
1.5
-1.5
-1.5
5.9E-03
1.7
-1.4
1.2E-02
2.1E-04
2.0
3.2E-02
-1.4
-1.5
2.5E-03
1.8
4.9E-03
-1.4
3.0E-02
-1.7
4.4E-03
-1.6
8.8E-03
3.2E-03
-1.9
4.1E-03
2.1
1.8E-04
2.0
1.9E-04
-1.3
1.7E-02
-2.0
5.8E-04
-1.3
4.4E-02
-1.4
2.5E-02
-1.7
2.2E-03
1.7E-04
3.6E-03
1.8
5.1E-03
-1.5
8.8E-03
-1.3
4.3E-02
-1.7
8.5E-04
-1.8
7.2E-03
llmg_1052
llmg1052
-1.5
2.7E-02
llmg_1064
bmpA
1.4
3.6E-02
1.8
1.1E-03
1.4
3.5E-02
llmg_1101
llmg1101
-1.5
2.3E-02
-1.6
4.2E-03
-1.6
4.0E-03
llmg_1121
nupC
-1.9
1.1E-03
-1.8
3.1E-03
-1.5
2.7E-02
llmg_1122
nupB
-1.5
1.5E-02
llmg_1123
nupA
2.5
2.8E-03
3.0
1.5E-04
2.5
1.2E-03
llmg_1135
llmg1135
-2.9
1.8E-05
-2.1
2.5E-03
-3.4
1.8E-06
llmg_1150
llmg1150
-1.4
5.4E-03
-1.4
9.9E-03
1.7
2.8E-03
-1.5
4.1E-02
llmg_1175
llmg1175
llmg_1184
gltD
-1.3
2.2E-02
1.6
1.5E-02
llmg_1287
llmg1287
llmg_1288
hisK
llmg_1311
llmg1311
1.6
1.3E-02
llmg_1339
engB
1.8
2.2E-02
1.9
1.7
-1.5
2.8
2.3E-02
4.7E-05
4.1E-02
1.8E-03
1.5
1.7E-02
-1.8
1.5E-03
-1.5
2.5
7.7E-03
2.6E-04
2.3
1.2E-03
-1.9
1.1E-02
-1.6
6.9E-03
1.4
1.1E-02
-1.6
1.9E-02
-1.5
3.4E-02
1.8
2.1E-02
195
196
llmg_1420
llmg1420
llmg_1453
llmg1453
1.4
4.3E-02
llmg_1466
cobQ
1.7
8.5E-03
llmg_1484
comFC
1.4
3.5E-02
llmg_1496
llmg1496
llmg_1497
llmg1497
1.8
1.8E-03
llmg_1500
llmg1500
-1.8
9.5E-05
llmg_1512
llmg1512
2.1
3.1E-03
llmg_1514
rex
llmg_1548
llmg1548
-1.3
3.0E-02
llmg_1557
llmg1557
1.3
4.2E-02
llmg_1633
llmg1633
1.5
2.2E-02
llmg_1641
butA
llmg_1647
llmg1647
1.7
2.7E-02
1.5
3.7E-02
-1.3
1.7
3.1E-02
llmg1652
1.6
2.7E-02
llmg_1737
cpo
2.1
6.5E-03
llmg_1746
llmg1746
llmg_1759
llmg1759
llmg_1764
metS
llmg_1784
fabG1
llmg_1798
lrgA
llmg_1811
llmg1811
llmg_1817
7.0E-03
1.7
4.0E-02
-1.5
2.1E-02
-1.5
1.8E-02
-1.5
2.7E-02
1.3
2.8E-02
-1.3
1.5
2.7E-02
-1.6
2.0E-02
3.4E-02
-1.7
9.3E-04
1.4
4.7E-02
2.7
1.0E-04
2.2
2.4E-03
5.5
1.4E-08
4.9
7.5E-08
-1.4
1.1E-02
llmg_1652
-1.6
-1.3
2.5
4.7E-04
5.1
5.7E-08
4.7E-02
4.0E-02
-1.4
2.1E-02
-1.7
2.3E-03
-2.2
2.0E-05
-2.1
2.5E-05
-10.7
1.8E-08
-9.7
8.0E-08
-7.6
1.6E-06
1.5
2.6E-02
1.7
1.6E-02
3.8
4.7E-07
3.1
1.4E-05
2.6
1.6E-04
2.7
9.2E-06
2.6
1.8E-05
2.3
2.0E-04
-1.4
3.1E-02
1.5
1.8E-02
-1.5
7.5E-03
3.3
3.0E-06
3.1
1.1E-05
2.4
5.1E-04
2.1
5.2E-05
1.7
5.6E-03
1.6
8.5E-03
3.5
1.7E-06
3.2
5.1E-06
2.8
4.7E-05
llmg1817
1.5
4.7E-03
1.8
1.4E-04
1.5
1.0E-02
llmg_1825
llmg1825
-1.9
8.9E-04
-2.0
3.5E-04
-1.8
3.8E-03
llmg_1830
menX
-1.4
2.4E-02
-1.4
3.6E-02
-1.4
3.7E-02
1.3
4.7E-02
2.8
2.4E-06
-1.4
1.3E-02
llmg_1833
menC
llmg_1845
llmg1845
llmg_1888
ardA
-1.9
1.4
1.4E-04
1.5E-02
1.4
2.9E-02
1.4
1.5E-02
llmg_1908
llmg1908
llmg_1912
llmg1912
llmg_1913
pbuO
1.8
3.5E-03
llmg_1914
llmg1914
-1.3
3.7E-02
1.5
3.4E-03
llmg_1918
llmg1918
llmg_1953
estA
llmg_1957
llmg1957
1.1
2.8E-03
llmg_1964
bioY
-1.3
3.9E-02
1.4
2.1E-02
1.5
3.3E-02
llmg_1988
llmg1988
llmg_1991
adhA
llmg_1999
llmg1999
1.9
1.9E-02
1.3
3.1E-02
1.6
1.2E-02
1.6
4.0E-02
1.6
2.0E-03
-1.5
2.2E-02
1.6
1.5E-02
1.0
llmg2012
-1.7
8.3E-04
-1.4
4.2E-02
llmg_2034
llmg2034
-1.3
1.8E-02
-1.4
2.7E-03
llmg_2059
llmg2059
1.7
2.6E-02
llmg_2065
llmg2065
-1.4
1.7E-03
llmg_2079
pknB
1.5
3.5E-02
llmg_2100
ps442
3.1E-02
2.0
2.7
4.3E-03
1.3
1.8E-04
4.4E-02
1.6
5.3E-03
-1.0
1.6E-02
1.8
4.6E-03
-1.8
2.7E-03
4.0E-06
7.8E-03
llmg_2012
1.7
1.5
1.8
3.6E-03
-1.4
2.3E-02
1.8
7.8E-03
-1.5
7.8E-03
1.8
5.4E-03
2.1
4.8E-04
-1.4
2.2E-02
1.3
4.0E-02
1.7
1.2E-02
1.4
5.0E-02
-1.4
4.3E-02
3.1
5.8E-07
1.3
3.4E-02
-1.4
3.8E-02
-1.4
1.2E-02
-1.5
4.0E-03
-1.1
1.4E-02
2.1
7.6E-04
1.6
3.3E-02
1.7
8.2E-03
llmg_2172
llmg2172
2.5
3.5E-04
llmg_2184
llmg2184
1.4
1.2E-02
1.4
7.9E-03
1.3
4.9E-02
llmg_2186
cbf
-1.5
2.0E-02
-1.6
2.4E-03
-1.3
3.5E-02
llmg_2216
llmg2216
-1.5
1.5E-02
-1.4
3.7E-02
llmg_2232
llmg2232
llmg_2252
ps517
llmg_2325
llmg2325
llmg_2328
pepXP
llmg_2329
llmg2329
llmg_2336
llmg2336
llmg_2341
llmg2341
llmg_2348
llmg2348
1.4
1.4
7.2E-03
1.4
1.3E-02
-1.3
3.2E-02
2.2E-02
2.1
2.3E-02
-1.4
1.3E-02
-1.3
2.9E-02
1.1E-04
-1.5
2.8E-03
-1.8
1.2E-03
llmg2424
-1.3
3.5E-02
llmg_2426
llmg2426
-1.3
3.5E-02
llmg_2436
llmg2436
llmg_2440
llmg2440
-1.6
9.1E-04
-1.7
llmg_2449
llmg2449
-1.4
3.1E-02
-1.8
7.1E-04
llmg_2479
llmg2479
-1.3
2.5E-02
llmg_2500
llmg2500
llmg_2505
llmg2505
llmg_2521
trmE
llmg_2534
ps610
1.4
1.3
5.0E-02
1.8
1.7E-02
-1.4
4.8E-03
3.2E-03
-1.6
2.4E-04
-1.6
2.0E-02
2.0E-03
-1.4
5.2E-03
1.6
2.8E-02
-1.4
1.5E-02
-1.5
2.9E-03
1.8
1.3E-03
2.1
1.3E-04
-1.5
1.4E-02
-1.4
4.0E-02
-1.3
2.9E-02
-1.4
2.7E-02
1.3
2.9E-02
1.7
1.6E-03
-1.3
4.5E-02
3.5E-02
4.1E-02
-1.4
1.1E-02
-1.5
8.8E-05
-1.5
llmg_2424
-1.7
-1.4
-1.5
4.9E-04
-1.5
6.5E-03
-1.4
2.9E-02
-1.4
4.2E-02
-1.6
7.2E-03
1.5
3.2E-02
[S] Function unknown
llmg_0086
llmg0086
-1.6
1.9E-03
llmg_0101
llmg0101
1.4
8.1E-03
llmg_0116
llmg0116
-1.3
2.8E-02
-1.6
9.9E-04
llmg_0128
llmg0128
llmg_0150
llmg0150
llmg_0152
llmg0152
llmg_0153
llmg0153
-1.6
-1.7
8.1E-04
1.7E-02
2.0
-3.2
7.2E-04
-2.6
5.8E-05
llmg_0155
llmg0155
-1.5
3.5E-03
-1.7
1.4E-04
-1.7
2.0E-04
llmg_0165
llmg0165
-1.6
6.8E-03
1.7
3.1E-03
-1.7
4.3E-03
llmg_0193
llmg0193
1.4
3.7E-02
3.0
7.6E-06
-1.7
5.2E-03
llmg_0242
llmg0242
llmg_0298
llmg0298
llmg_0304
llmg0304
1.7
3.3E-02
3.1
3.3E-06
1.7
llmg_0333
llmg0333
-1.6
5.1E-04
-1.5
2.8E-03
llmg_0334
llmg0334
-1.3
2.6E-02
-1.4
llmg_0380
llmg0380
-1.4
2.6E-02
-1.7
6.0E-05
-1.4
-1.3
2.8E-02
-1.4
4.2E-03
1.9
4.3E-04
-1.8
5.7E-03
1.6
-1.9
llmg_0418
llmg0418
1.4
3.9E-02
llmg_0434
llmg0434
llmg_0448
llmg0448
llmg_0450
llmg0450
1.8
5.3E-04
llmg_0465
llmg0465
-2.5
6.0E-05
llmg_0476
llmg0476
3.3E-02
4.0E-06
1.7E-02
1.1E-03
1.7
2.3
1.7
-1.8
1.6E-06
3.5E-02
llmg0239
llmg0417
-2.4
-1.4
llmg_0239
llmg_0417
6.5E-05
2.1E-02
7.1E-04
llmg0391
2.0E-03
-2.5
-1.4
2.2
llmg_0391
1.2E-02
-2.9
1.0E-02
llmg0203
llmg0381
1.4
2.8E-03
-1.8
llmg0205
llmg0382
1.1E-02
2.5E-02
llmg_0205
llmg_0381
1.4
-2.6
1.7
llmg_0203
llmg_0382
2.9E-03
1.5
1.6E-02
1.5
9.5E-03
2.1
1.3E-03
2.9
1.5E-05
1.4
4.1E-02
-1.8
1.0E-03
-1.9
4.0E-04
3.0E-02
3.1
2.8E-06
2.2
1.4E-03
-1.3
4.9E-02
-1.3
2.9E-02
8.9E-03
-1.5
3.0E-03
-1.6
5.4E-03
2.0E-02
-1.3
3.4E-02
2.0
1.5E-04
1.6
5.6E-03
4.0E-02
1.5
4.2E-02
1.8
4.6E-03
1.6
1.9E-02
1.3E-03
-1.6
1.4E-02
-1.9
1.4E-03
-1.7
8.1E-04
-1.5
1.2E-02
-1.6
4.9E-02
-1.9
2.4E-02
1.7
2.9E-03
-1.6
1.1E-02
-1.8
1.4E-03
-1.9
2.9E-03
-2.0
3.1E-03
-2.5
6.9E-05
1.6
1.7E-02
197
llmg_0482
llmg0482
-2.0
1.3E-03
llmg_0495
llmg0495
1.6
1.8E-02
llmg_0513
llmg0513
llmg_0542
llmg0542
-2.5
4.1E-05
-1.3
4.7E-02
-1.5
2.2E-03
-1.7
2.4E-02
-1.3
5.0E-02
1.9
1.9E-02
-1.4
9.8E-03
-2.0
2.0E-03
-1.5
2.5E-02
llmg_0571
llmg0571
-1.6
2.7E-03
-1.5
2.3E-03
llmg_0590
llmg0590
1.8
3.9E-02
2.7
1.9E-04
llmg_0644
llmg0644
llmg_0645
llmg0645
-1.3
4.2E-02
llmg_0724
llmg0724
1.5
3.4E-02
llmg_0726
llmg0726
1.4
4.2E-02
llmg_0750
llmg_0752
pip
llmg_0867
llmg0867
llmg_0875
llmg0875
llmg_0938
llmg0938
llmg_0939
llmg0939
llmg_0969
llmg0969
llmg_0971
yphI
llmg_0972
llmg0972
-1.4
llmg1025
llmg_1029
llmg1029
llmg_1057
llmg1057
llmg_1079
llmg1079
9.3E-03
-2.2
1.9E-02
1.3
2.9E-02
1.5
1.7E-02
-2.2
2.9E-02
1.5
-1.4
4.0E-02
1.5
4.2E-03
1.5
2.3E-02
1.6
1.1E-02
1.6
1.8E-02
2.0
5.9E-04
1.5
7.2E-03
-1.3
2.7E-02
3.4E-02
1.7
4.3E-03
1.7
2.9E-03
1.5
2.6E-02
1.5
2.5
2.5E-04
2.3
2.0E-03
1.9
1.4E-02
1.8
2.1E-02
1.4
5.0E-02
1.5
2.0E-02
-1.3
3.3E-02
3.7
2.0E-05
4.0E-02
llmg_1081
llmg1081
llmg_1085
llmg1085
2.1
2.0E-02
4.4
1.3E-06
llmg_1103
llmg1103
1.6
3.7E-03
1.8
1.4E-03
llmg_1111
GerCA
1.5
2.1E-02
llmg_1112
llmg1112
1.3
1.4
4.6E-02
1.7
3.9E-03
1.4
4.0E-02
-1.3
2.7E-02
3.8
2.2E-05
-1.3
4.0E-02
2.7
1.0E-03
1.4
2.5E-02
-1.4
3.1E-02
2.7
1.3E-03
4.9E-02
llmg_1159
llmg1159
llmg_1186
llmg1186
2.7
1.6E-03
2.9
9.3E-05
llmg_1192
scpA
-1.3
2.7E-02
-1.5
3.2E-03
llmg_1195
llmg1195
-2.0
1.7E-02
-2.3
4.8E-03
llmg_1236
llmg1236
-1.3
3.6E-02
-1.3
1.6E-02
llmg_1257
llmg1257
-1.6
4.9E-02
-2.4
1.0E-05
3.5
4.3E-06
-1.3
4.8E-02
-1.7
3.5E-03
-5.6
2.6E-07
-6.1
3.5E-08
-2.2
2.3E-02
-3.1
2.8E-04
-3.7
1.8E-04
-5.1
9.8E-07
-2.1
3.3E-03
-2.3
7.4E-04
-2.2
3.1E-03
llmg_1263
llmg1263
-1.4
6.7E-03
-1.4
2.1E-03
llmg_1299
llmg1299
1.5
1.3E-02
1.6
8.7E-03
orf57
1.7
1.1E-02
llmg_1360
orf46
1.6
2.3E-02
1.8
2.1E-02
llmg_1413
llmg1413
1.6
5.1E-03
1.6
4.3E-03
llmg_1428
llmg1428
1.5
8.1E-03
1.3
3.6E-02
llmg_1442
llmg1442
-1.9
1.2E-02
-1.8
2.3E-02
llmg_1448
llmg1448
-1.4
3.1E-02
-1.3
2.2E-02
-2.1
2.7E-02
-4.7
6.5E-06
1.4
4.6E-03
1.4
3.3E-02
-3.0
2.7E-05
4.2E-02
1.0E-02
1.5E-03
3.0E-06
-1.5
-1.7
4.1E-02
-1.8
-4.8
5.1E-04
2.0E-02
-1.5
llmg1260
llmg_1349
6.8E-06
9.7E-03
2.7
-1.6
llmg1259
llmg_1302
3.4E-02
-2.6
2.5E-03
4.7E-02
llmg_1259
llmg1301
-1.5
2.4
1.4
1.3
llmg_1260
llmg_1300
198
2.0E-02
6.8E-03
llmg_1158
llmg_1301
-1.4
2.1E-02
-1.6
1.3
1.2E-02
3.0E-02
2.3E-02
llmg_1024
llmg_1025
1.7
-2.0
-1.6
2.2
1.5E-02
-1.3
4.6E-02
-1.9
1.7E-02
1.4
3.8E-02
llmg_1482
llmg1482
-1.5
7.4E-03
llmg_1485
llmg1485
1.4
3.4E-02
llmg_1495
llmg1495
-1.5
7.3E-03
-1.1
5.6E-03
-1.4
3.8E-02
llmg_1498
llmg1498
-2.0
3.5E-04
-2.0
2.3E-04
-2.3
8.5E-05
llmg_1499
llmg1499
-1.4
3.0E-02
-1.3
1.2E-02
-1.5
3.8E-02
llmg_1523
llmg1523
-1.4
6.3E-03
llmg_1540
llmg1540
3.1
4.5E-05
2.7
3.8E-04
llmg_1555
llmg1555
1.6
3.0E-02
1.7
2.9E-02
-2.0
1.6E-03
llmg_1556
llmg1556
llmg_1563
llmg1563
llmg_1572
mycA
llmg_1573
llmg1573
llmg_1597
llmg1597
1.4
4.8E-02
llmg_1644
llmg1644
-1.4
2.8E-02
-1.3
7.1E-03
llmg_1650
llmg1650
llmg_1659
llmg1659
llmg_1666
llmg1666
llmg_1670
llmg1670
llmg_1672
llmg1672
llmg_1683
llmg1683
llmg_1727
llmg1727
llmg_1749
llmg1749
llmg_1760
llmg1760
llmg_1806
llmg1806
-1.3
1.4
1.4
2.0E-02
-1.5
4.9E-02
-1.8
7.2E-03
-1.7
9.8E-03
-1.5
1.1E-03
-1.3
1.9E-02
-1.4
4.1E-03
-1.5
2.1E-02
2.2
7.8E-05
-1.3
4.9E-02
1.7
7.0E-04
1.9
5.6E-04
1.8
3.2E-03
1.4E-03
-1.5
2.1E-02
-1.4
4.2E-02
1.7
8.8E-04
1.6
2.4E-02
3.8
2.3E-05
1.4
2.3E-02
1.6
4.1E-03
1.8
7.0E-03
llmg1813
2.1
1.2E-03
llmg1822
1.8
1.5E-02
-1.6
5.3E-04
-1.6
4.2E-04
-3.2
4.1E-08
llmg_1851
llmg1851
-1.5
1.3E-02
llmg_1855
llmg1855
-1.7
1.0E-04
llmg_1917
llmg1917
-2.1
1.6E-04
-2.7
8.5E-07
llmg_1933
llmg1933
-1.3
4.4E-02
-1.3
3.8E-02
1.6
2.0E-02
llmg1936
llmg1937
llmg_1941
llmg1941
1.3
3.7E-02
llmg_1962
llmg_1971
llmg1971
llmg_2020
llmg2020
1.4
4.8E-03
llmg_2037
llmg2037
1.5
2.2E-02
1.5
4.3E-02
-1.6
3.5E-03
1.6
3.0E-02
-1.5
5.8E-03
1.4
3.0E-02
1.9
9.5E-05
-1.3
3.7E-02
1.9
1.6
1.1E-02
llmg2039
llmg_2041
llmg2041
-1.3
1.3E-02
llmg_2056
llmg2056
1.4
4.7E-02
1.8
1.2E-02
llmg2058
llmg_2062
llmg2062
llmg_2098
ps444
llmg_2127
llmg2127
llmg_2143
llmg2143
llmg_2144
1.4
-1.4
3.5E-03
6.9E-03
2.3E-02
-1.3
2.5E-02
2.3
2.6E-03
-1.8
6.4E-03
1.3
2.8E-02
-1.7
2.0E-03
-1.4
5.0E-02
2.2
4.3E-05
2.0
1.1E-03
-1.4
8.2E-03
-1.4
2.3E-02
-1.5
1.7E-02
1.4E-02
-1.5
7.5E-03
-1.6
2.4E-03
1.5
1.2E-02
1.4
2.9E-02
1.7
5.3E-03
-1.7
3.1E-03
-2.2
4.1E-05
-1.8
1.3E-03
3.4
1.2E-04
3.8
1.9E-05
1.7
1.3E-02
1.4
3.8E-02
-1.8
6.7E-03
-1.4
3.2E-02
1.4
3.9E-02
2.0
7.5E-04
1.6
9.4E-03
1.8
1.1E-02
-1.7
1.2E-02
-1.4
2.3E-02
5.0E-04
llmg_2039
llmg_2058
-1.7
6.1E-03
llmg_1813
llmg_1936
1.7
-1.5
5.0E-03
llmg_1822
llmg_1937
3.9E-02
1.6E-02
1.4
1.6
1.4
3.3E-02
1.5
1.4E-02
1.3
3.2E-02
-1.3
4.3E-02
-1.4
6.7E-03
1.6
5.5E-03
-1.4
5.4E-03
-1.3
4.4E-02
-1.4
5.2E-03
6.7E-03
1.7
2.3E-03
1.7
2.1E-02
2.2
7.4E-04
1.5
4.4E-03
1.3
4.7E-02
1.6
-1.3
4.8E-02
-1.4
2.3E-02
-4.5
9.5E-07
-3.8
1.7E-05
-3.3
7.8E-05
llmg2144
-2.7
2.1E-02
-11.4
2.3E-10
-10.3
1.4E-09
-7.5
1.3E-07
llmg_2146
llmg2146
-2.4
2.1E-02
-9.1
2.3E-10
-7.9
3.2E-09
-6.2
1.2E-07
llmg_2148
llmg2148
1.5
2.0E-02
-1.5
1.9E-02
-1.4
4.8E-02
llmg_2164
llmg2164
2.2
6.5E-03
2.0
3.1E-02
4.1
5.8E-07
3.4
3.0E-05
2.3
3.3E-03
199
llmg_2168
llmg2168
llmg_2187
llmg2187
1.5
2.1E-02
1.7
2.1E-03
-1.4
llmg_2194
llmg2194
1.5
2.8E-02
1.6
8.9E-03
1.4
1.9E-02
llmg_2212
llmg2212
llmg_2214
llmg2214
llmg_2225
llmg2225
llmg_2228
llmg2228
-1.4
1.8E-02
llmg_2230
pcaC
-1.3
3.1E-02
llmg_2261
ps508
llmg_2280
llmg2280
-1.7
6.0E-03
llmg_2282
llmg2282
llmg_2306
llmg2306
-1.3
3.6E-02
llmg_2324
llmg2324
llmg_2337
llmg2337
llmg_2338
llmg2338
llmg_2418
llmg2418
llmg_2422
llmg2422
llmg_2463
llmg2463
llmg_2496
llmg2496
llmg_2507
usp45
llmg_2516
llmg2516
1.6
1.0E-02
1.6
3.2E-02
2.1
1.4
1.7
9.8E-05
1.4
-1.7
2.0E-04
-1.9
2.7E-02
-1.5
2.9E-02
-1.5
2.4E-02
7.2E-03
2.1E-02
-1.4
3.5E-02
1.9
4.7E-03
-1.4
4.3E-03
-1.5
3.0E-02
3.7E-02
-1.5
4.2E-03
-1.5
2.2E-03
1.7
1.1E-02
1.8
6.3E-03
2.3E-02
2.2
3.5E-04
-1.5
1.3E-02
1.9
9.3E-04
1.5
5.4E-03
-1.7
1.0E-03
1.8
2.1E-02
-1.4
9.5E-03
-1.5
8.7E-03
1.6
6.2E-03
1.6
6.8E-03
-1.1
3.4E-03
1.9
6.1E-05
-1.1
1.9E-02
-1.4
1.5E-02
-1.8
1.0E-02
-1.6
2.3E-02
2.0
3.0E-04
2.0
2.3E-04
-1.4
4.9E-02
-1.6
1.8E-03
-1.8
7.5E-03
1.7
9.7E-04
-2.4
6.8E-04
[T] Signal transduction mechanisms
200
llmg_0093
llmg0093
-2.3
1.8E-03
llmg_0113
relA
1.7
1.6E-03
1.6
5.5E-03
1.8
8.0E-04
-1.5
3.4E-02
1.5
4.0E-02
llmg_0136
llmg0136
-1.5
4.9E-02
-1.6
2.4E-04
-1.5
6.1E-03
1.1
1.2E-02
-1.5
2.5E-02
1.5
1.5E-02
1.4
2.4E-02
1.4
3.4E-02
1.9
3.9E-04
1.7
1.6E-03
-1.4
1.8E-02
-1.3
2.9E-02
1.6
5.5E-03
1.6
2.6E-02
llmg_0273
luxS
llmg_0413
kinC
llmg_0430
cstA
llmg_0582
ptsK
llmg_0748
kinF
llmg_0831
ps337
-1.4
1.8
7.2E-03
3.0E-03
1.6
1.6
1.7
5.9E-03
1.6
3.0E-03
5.8E-03
6.5E-03
-1.6
1.5E-02
1.4
4.2E-02
llmg_0865
llmg0865
-1.3
2.6E-02
llmg_0902
arsC
-1.4
1.6E-02
llmg_0908
llrA
-2.7
5.7E-04
-2.5
1.6E-03
-2.7
4.7E-04
llmg_1080
uspA2
-3.8
1.5E-07
-3.9
8.6E-08
-3.8
7.5E-08
llmg_1235
llmg1235
-3.3
2.9E-06
-2.5
3.8E-04
-3.6
2.8E-07
-1.5
7.4E-03
-1.4
1.0E-02
llmg_1303
llmg1303
llmg_1350
telC
llmg_1351
telB
1.5
1.2E-02
1.3
4.5E-02
llmg_1352
telA
1.4
1.3E-02
1.5
3.1E-03
llmg_1518
kinE
1.6
2.9E-02
2.1
3.4E-04
llmg_1559
flpA
2.4
3.0E-03
1.5
1.8
2.9E-02
2.4
2.9E-03
1.2E-03
1.8
6.3E-03
1.3
3.6E-02
llmg_1582
llmg1582
2.1
1.4E-03
1.8
8.7E-03
2.0
1.3E-02
llmg_1649
kinD
1.6
5.1E-03
1.8
1.5E-03
1.8
5.7E-04
llmg_1662
uspA
-2.3
7.2E-04
-4.8
7.0E-10
llmg_1816
llmg1816
1.6
5.5E-03
1.5
3.5E-02
-2.0
1.3E-02
-4.4
9.7E-08
1.8
3.6E-06
2.3
5.5E-08
-2.5
1.6E-04
llmg_2023
llmg2023
-1.9
2.8E-02
llmg_2047
llmg2047
1.7
5.2E-03
llmg_2080
pppL
2.1
4.2E-07
1.6
-1.6
1.7
2.1E-03
-2.6
1.7E-04
-2.8
2.1E-05
-3.1
5.9E-05
-3.0
5.1E-05
1.6
1.7E-03
1.4E-02
4.3E-03
llmg_2163
llmg2163
1.9
3.1E-03
1.7
1.7E-02
llmg_2292
typA
-1.5
4.1E-03
-1.5
7.8E-03
llmg_2442
llmg2442
llmg_2480
llmg2480
1.3
4.8E-02
-1.4
1.7E-02
llmg_2486
ptpL
llmg_2512
rcfB
llmg_2514
llmg2514
2.5
1.4
5.5E-04
1.5
1.4E-02
-1.3
4.7E-02
2.6
2.0E-04
2.0
9.2E-04
2.2E-02
2.8
3.3E-04
2.7
5.6E-04
2.6
8.3E-04
2.6
3.4E-04
2.7
3.4E-04
2.1
4.6E-03
1.4
1.4E-02
2.0
1.2E-02
[U] Intracellular trafficking, secretion, and vesicular transport
llmg_0124
secA
llmg_0143
llmg0143
llmg_0540
oxaA2
2.0
7.2E-03
llmg_0638
clpP
1.5
6.0E-03
llmg_0878
ffh
llmg_1273
smf
-1.5
3.1E-02
llmg_1391
llmg1391
1.1
4.7E-03
1.3
2.8E-02
1.4
1.5E-02
llmg_1525
lspA
llmg_1587
secG
llmg_1744
ftsY
llmg_2097
ps445
1.3
3.7E-02
2.9
8.3E-05
-1.6
1.3E-02
2.3
2.4E-02
-2.2
3.9E-02
-2.0
2.3E-02
1.1
5.4E-03
-1.7
4.4E-02
1.4
1.8E-02
1.4
2.2E-02
1.5
1.8E-02
-1.3
2.4E-02
-1.3
3.4E-02
1.4
2.9E-02
llmg_2270
ecsB
1.3
1.5E-02
1.5
1.1E-03
1.4
2.3E-03
llmg_2361
secY
1.7
1.7E-02
2.0
2.3E-03
3.1
8.9E-07
2.1
1.3E-03
1.8
5.9E-03
-2.1
1.0E-02
-2.1
9.8E-03
-3.3
4.3E-05
-2.3
4.1E-03
-2.4
1.7E-03
-1.3
1.6E-02
-1.2
4.6E-02
-1.5
2.1E-02
1.4
3.0E-02
-1.5
1.1E-02
-1.5
4.8E-02
llmg_2389
llmg_2408
comGA
-1.2
3.0E-02
llmg_2416
yajC
-1.5
4.1E-02
1.8
2.7E-04
[V] Defense mechanisms
llmg_0178
dinF
1.5
3.3E-02
llmg_0324
lmrC
1.6
5.0E-03
llmg_0329
llmg0329
-1.7
4.5E-03
llmg_0501
llmg0501
1.5
1.2E-02
llmg_0624
llmg0624
1.5
1.0E-02
llmg_0658
hsdR
llmg_0659
hsdM
llmg_0660
hsdS
llmg_0852
ps357
llmg_0989
llmg_1202
1.5
3.5E-02
llmg0989
-1.7
4.2E-02
llmg1202
-1.4
4.7E-02
llmg_1203
llmg1203
-1.6
2.1E-02
llmg_1468
llmg1468
-1.8
4.8E-03
llmg_1553
llmg1553
llmg_1635
llmg1635
1.6
llmg_1675
llmg1675
1.5
2.5E-02
1.7
3.9E-02
1.7
1.2E-02
2.7
8.4E-06
3.0E-03
-3.3
1.1E-05
-1.5
2.5E-02
3.0E-03
1.5
4.7E-03
2.3
5.8E-04
2.8
2.8E-05
2.0
5.6E-05
llmg1853
lmrA
1.6
1.1E-02
llmg_2271
ecsA
1.3
3.8E-02
llmg_2279
bcrA
llmg_2322
llmg2322
1.1
6.7E-03
llmg_2386
llmg2386
1.7
2.3E-02
llmg_2548
llmg2548
-1.7
2.4E-03
5.5E-03
1.8E-02
-1.4
3.2E-02
3.1E-02
1.8E-02
2.3
1.6E-04
1.9
2.9E-03
-2.3
2.7E-03
-2.0
1.5E-02
2.1E-02
-1.7
7.1E-03
-1.4
4.3E-02
-3.1
1.5E-07
-2.0
9.8E-04
-2.4
3.3E-05
-1.8
6.8E-04
-1.5
9.3E-03
-1.4
4.5E-02
-1.4
2.8E-02
-1.5
1.7E-02
-1.9
8.2E-03
-1.6
1.2E-02
-1.4
-1.6
1.7
1.5
4.2E-02
llmg_1853
7.4E-03
-2.2
-1.4
llmg_1856
1.8
-1.5
-1.5
2.1E-02
-1.3
1.3E-02
1.5
2.7E-02
1.5
2.2E-02
1.6
9.6E-03
2.0
1.5E-03
1.1
4.7E-02
1.9
1.2E-03
1.4
4.3E-02
201
202
Samenvatting
Nederlandse samenvatting voor
niet-ingewijden
203
Dit proefschrift beschrijft een systeembiologische aanpak om het fermentatiegedrag
van Lactococcus lactis bij verschillende groeisnelheden nader te bestuderen.
Indien Lactococcus lactis groeit in de aanwezigheid van glucose, zal ze op lage
groeisnelheden voornamelijk formaat, acetaat en ethanol produceren, terwijl bij hoge
groeisnelheden hoofdzakelijk melkzuur wordt gevormd. Onze werkhypothese was dat
de omschakeling tussen de fermentatietypen wordt veroorzaakt door een investering
in andere enzymverhoudingen die uiteindelijk de verschillende eindproducten
maken. Hoofdstuk 2 en 3 bevatten naast een inzichtelijke en hoogwaardige dataset
van transcript- en eiwitratio’s ook de datapunten van enzymactiviteiten van vrijwel
alle glycolytische enzymen. We hebben de groeisnelheid in de bioreactor gevarieerd
door een limiet in te stellen van de enige suikerbron aanwezig in het groeimedium.
Suiker werd toegevoegd in de vorm van glucose. Uit deze dataset is af te leiden dat
met een verhoging van de groeisnelheid de transcriptie- en eiwithoeveelheden, alsook
de maximale enzymactiviteiten van de glycolytische enzymen, vrijwel onveranderd
blijven, terwijl de eindproducten wel degelijk een metabole verschuiving laten zien.
Hiermee is aangetoond dat voor Lactococcus. lactis de verschillen in investering in
eiwitexpressie niet verantwoordelijk zijn voor het wisselen van de fermentatietypen.
Met behulp van dezelfde dataset hebben we ook inzicht verkregen in de dynamiek van
de verschillende onderdelen van het ribosoom tijdens verschillende groeisnelheden.
Het grootste gedeelte van de totale eiwitproductie van de cel (het proteoom) bestaat uit
glycolytische enzymen en ribosomale eiwitten. Vanuit het oogpunt van regulatie door
investering in eiwitproductie was het dus belangrijk ook hier inzicht in te verkrijgen.
Het blijkt dat de ratio tussen ribosomaal RNA en ribosomaal eiwit verschilt bij
verschillende groeisnelheden. Bij lage groeisnelheden is er relatief veel ribosomaal
eiwit ten opzichte van het ribosomaal RNA. Bij hogere groeisnelheden is er een
toename van ribosomaal RNA ten opzichte van het ribosomale eiwit (hoewel beide
absoluut gezien toenemen). De toename van het ribosomaal eiwit wordt niet vergezeld
door een toename van transcripten voor diezelfde ribosomale eiwitten; dat wil zeggen,
de toename wordt veroorzaakt door een efficiëntere translatie.
Een ander belangrijk proces in de Lactococus lactis cel is het arginine metabolisme,
aangezien dit ook energie kan leveren, in de vorm van ATP. We hebben bepaald hoeveel
arginine wordt afgebroken bij welke groeisnelheid en het blijkt dat de hoeveelheden
van de betrokken eiwitten exponentieel toenemen met de groeisnelheid. Bij de laagste
groeisnelheid was er sprake van een lage activiteit in argininemetabolisme. Vanaf de
204
laagste groeisnelheid lopen transcriptie van de arginine-genen en de eiwithoeveelheden
steeds sneller op. Deze metingen worden bevestigd door de hoeveelheid arginine die
wordt aangetroffen in het medium na het oogsten van de cellen. Echter, bij de hoogste
groeisnelheid die we onder de geteste groeiomstandigheden konden behalen in onze
chemostaten was er vrijwel geen gebruik van arginine als energiebron. Wij vermoeden
dat dit wordt veroorzaakt door repressie gemedieerd door de eiwitten CcpA en HPr
Ser-46P (tezamen CCR genoemd). Op de hoogste groeisnelheid kan er sprake zijn
van repressie door CCR van de arginine metabolisme genen doordat er voldoende
glucose is in de omgeving van Lactococcus lactis. Overgebleven glucose en de
transcriptieniveaus van andere genen in het CCR regulon bevestigen deze waarneming
bij de hoogste groeisnelheid. We vermoeden dat doordat er genoeg glucose is voor
de cel bij de hoogste groeisnelheid, er geen noodzaak meer is tot het afbreken van
arginine.
In Hoofdstuk 4 is de regulatie van de vetzuursynthese in Lactococcus lactis in kaart
gebracht. Deze melkzuurbacterie produceert de onderdelen van vetzuren door acetylCoA om te zetten in malonyl-CoA, dit gebeurt door de enzymen AccA, AccB, AccC
en AccD. Het enzym FabD vervangt de CoA-groep met een chaperonne-eiwit (ACP)
en FabH bindt daar een acetyl-CoA aan om uiteindelijk een ketoacyl-ACP te vormen.
Dit molecuul wordt verlengd met behulp van de enzymen FabG, FabZ, FabI en FabF.
Vrijwel alle genen die coderen voor de betrokken enzymen bevinden zich in het fab
(fatty acid biosynthesis) operon. De uitzondering is fabZ die zich in twee varianten op
verschillende locaties bevindt; fabZ2 bevindt zich in het fab operon, terwijl fabZ1 is
gelokaliseerd naast fabI (Hoofdstuk 4, Fig. 1A). Al deze genen staan in Lactococcus
lactis onder regulatie van de repressor FabT. Door middel van zogenaamde EMSAs
(Electrophoretic Mobility Shift Assay) en DNAseI-footprinting technieken is zowel
de plek als de DNA sequentie (het bindingsmotief) waar FabT bindt in kaart gebracht.
Tijdens het analyseren van transcripten onder verschillende groeisnelheden viel op
dat er een transcript van het gen yfiA zeer actief was onder lage groeisnelheden maar
niet onder hogere groeisnelheid. Het eiwit afkomstig van dit gen bleek betrokken te
zijn bij het dimeriseren van ribosomen wanneer de cel zich in de stationaire groeifase
bevond. Dit werk wordt beschreven in Hoofdstuk 5. Het blijkt dat wanneer het eiwit
YfiA afwezig is door deletie van zijn gen yfiA, er geen ribosomale dimeren meer
kunnen worden gevormd in Lactococcus lactis. Bij een complementatie studie waarbij
het gen yfiA weer opnieuw wordt ingebracht en tot (over)expressie gebracht, en het
205
YfiA dus weer in de cel aanwezig is, komen ook de ribosomale dimeren terug. Dit
complementeren was echter niet succesvol bij het terugplaatsen van het yfiA gen van
de ongerelateerde, Gram-negatieve bacterie Escherichia coli. Dit wordt veroorzaakt
door het ontbreken van de staart van het eiwit in E. coli YfiA. Deze YfiA staart van
Lactococcus lactis is essentieel in het verbuigen van de ribosomale kleine subeenheid.
Door deze verbuiging komt een gedeelte vrij aan de onderkant van het ribosoom
waardoor een tweede ribosoom daar aan bindt en zo een ribosomaal dimeer-complex
kan vormen.
De in dit proefschrift beschreven eigenschappen van de melkzuurbacterie kunnen in
de zuivelindustrie zeer van pas komen; wanneer Lactococcus lactis wordt ingezet als
een microbiële fabriek kunnen de groeiomstandigheden met behulp van deze kennis
slim geoptimaliseerd worden.
206
207
208
Dankwoord
209
Je houdt het niet voor mogelijk wat er uiteindelijk uit gaat komen wanneer je begint aan
je aio-project. Ik wist als student dat Oscar Kuipers en Jan Kok goede wetenschappers
en prettige professoren zijn. MolGen als vakgroep kende ik als student wel. Microarray
had ik nog niet gedaan, maar leek me ‘wel leuk’ om te doen. Mijn studentenproject
was aan B. subtilis/ B. coagulans, maar L. lactis was toen ook al wel een interessante
bacterie, vooral omdat je er kaas mee kon maken. Als ik bekijk waar ik stond en hoe ik
dacht toentertijd, heb ik behoorlijk wat geleerd. En dat leren heb ik niet alleen gedaan.
Ik denk dat ik me persoonlijk ontwikkeld heb in een mooie periode van mijn leven. Ik
ben veel mensen dankbaar doordat ze mij in de gelegenheid hebben gesteld te worden
tot de persoon die ik nu ben. Graag grijp ik de gelegenheid aan om jullie op schrift
hier te bedanken.
Op de eerste plek komen mijn twee begeleidende professors, in willekeurige volgorde.
Prof. Dr. Kok, beste Jan, je bent binnen de melkzuurbacteriewereld een erkende naam
en terecht, want je kennis van onder andere L. lactis is enorm. Je bent door velen
geroemd om je scherpe pen, en ik kan me alleen maar bij de menigte scharen. Ook
wil ik graag kwijt dat ik heb genoten van onze gezamenlijke trip in India. Te midden
van allerlei planten, zaden, kruiden, insecten en bomen heb ik een rasbioloog gezien.
Bedankt voor de mooie samenwerking.
Prof. Dr. Kuipers, beste Oscar, ik ben nogal verwonderd om je. Zelfs na al die jaren.
Hoe kan het zijn dat iemand die zo succesvol is met publiceren, die zoveel mensen
begeleidt, die allerlei samenwerkingen aanknoopt, zoveel documenten verwerkt,
colleges geeft, nog zo aardig blijven? Je bent onvermoeibaar en hebt me voorzien
van veel kennis en daarmee gevormd, zowel binnen als buiten het wetenschappelijke.
Ik waardeer het feit dat je me altijd ruggensteun hebt gegeven, niet alleen tijdens de
contractperiode, maar ook vooral daarna. Je maakte mijn aio-tijd tot een fijne tijd.
Voor beide heren geldt, het is me een waar genoegen en een grote eer om met jullie
namen op de publicaties te staan. Ik hoop dat we elkaar nog vaak ontmoeten en wens
jullie nog veel successen toe in de komende jaren.
Dan zou ik graag de mensen bedanken uit ons consortium. Het werk uit dit proefschrift
moest vaak worden besproken en afgestemd. Ik keek altijd uit naar deze dagen.
Bedankt, Douwe Molenaar en Bas Teusink voor alle inhoud en discussies die ons
werk zoveel beter heeft gemaakt. But Amsterdam was not Amsterdam without Anisha
Goel. Thanks for your positive input and your tremendous working attitude. You have
a great personality and I am still very happy with some of your Indian recipes. If I was
in Amsterdam, part of the fun over there was also because of Filipe Santos and Herwig
Bachmann. Keep up the good work guys!
Wat dichter bij huis is het kantoor van Bert Poolman. Ondanks je drukke schema
vond je meestal wel tijd om onze data en experimenten te bespreken. Bedankt voor
de plezierige samenwerking. My closest collaborator was without any doubts Pranav
Puri. Both for the STW project and the YfiA story we collaborated really close. I
210
admire your skillful and precise handling of samples and data in the lab. Without
you, this thesis would have looked quite different. At the Membrane Enzymology
department I also would like to thank Fabrizia Fusetti. Tante grazie per aiutarmi e le
nostre discussioni vivaci.
Bij MolGen is het wat lastiger bedanken. In één keer de totale club bedanken doet geen
recht aan heel wat mensen. Ik doe een poging en begin bij mijn paranimfen. Dear Ana
Solopova, thanks for all your help cheering me up when I was down, your charm, your
comments to my work. But best of all, without a doubt you were my best audience
for all my lame jokes. Beste Martijn Herber, caro Martino. Wat een genoegen is jouw
aanwezigheid op het werk, maar ook buiten het lab. Zwemmen, koken, eten, politiek
bespreken, e naturalmente nostra passione per questa bellissima lingua. Zonder twijfel
ben jij een van de meest intellectuele mensen die ik ken, ik heb veel van je geleerd.
Andere kantoorgenoten, met de nadruk op genoten zijn Araz Zeyniyev, Rutger Brouwer,
Sjoerd van der Meulen (ouwe grillkoning). Dat jullie het hebben uitgehouden met me
is een prestatie op zich. Op het grote en drukke L. lactis lab was het dringen geblazen.
Maar de muziek van Auke, de grapjes van Dongdong, de grijns van Manolo, de hulp
van Andrius maakten het een prettig eind van mijn aio-tijd. Special thanks goes to
Ganesh, your wedding was fantastic and you gave me great stories to tell. The same
goes for João Pinto(san), thanks for the great adventures. Good luck with your life,
PhD-defense and most important your family. My mentor in the lab was definitely
Claire Price. Thanks for correcting some of my experimental setups, improving my
English language and for all the advices on babystuff and careerplanning. Omdat jouw
Nederlands zo goed is, wens ik je in mijn eigen taal veel succes bij DSM. Anne de
Jong, je bent zeer van waarde geweest bij de analyses en bij het meedenken voor
dit enorme microarray-experiment van hoofdstuk 2 en 3. Het mooiste aan je was dat
ik nooit enig leeftijdsverschil heb ervaren tussen ons. Wout, gypsy, veel relaxeter
dan jij lopen er niet rond op aarde. Sterkte met de afronding van jouw aio-project,
ik bewonder jouw doorzettingsvermogen. Over doorzetten gesproken, kom ik uit
bij onze Duitsers, want die twee gaan ook door tot de laatste minuut. Robin, alter
Pirat, bedankt voor het beachvolleyballen, het samen spelen bij Veracles H2 en je
onverzettelijke optimisme op het lab (Jaaaaah man!). Katrientje, multi-talent, bedankt
voor het logeren en al je hulp. Dank je wel Lieke, je bent heel inspirerend geweest
voor mijn ribosoomexperimenten met je creatieve gedachten. Laeti, merci pour tout
(y compris le vin) et au revoir. JW, professor in spé, fijn om af en toe te sparren met
je, veel plezier samen met Anne-Stephanie en jullie fotomodelletje. Robyn, degelijke
wetenschapper en vrolijke quizmaster, we zijn naar elkaar toegegroeid tot mijn plezier.
Julio, velkommen til Chr. Hansen, erg leuk weer collega’s te worden. Dank je wel
Jeroen, ol’ Flabberhand. Zelden zo een goedzak in zo een groot lichaam gezien.
Maarten, enorme sarcast met een goed gevoel voor humor. Ákos, mentor tijdens mijn
masterproject en prettige collega tijdens mijn aio-project, dank voor de mooie jaren en
211
succes als principal investigator. Tomas, Imke, Sulman, Evert-Jan het was een plezier
met jullie in Haren. Bogusia, je was een fijne collega voor me en je bruiloft was
geweldig. Bedankt nog voor de luxe behandeling. Iemand die me veel geholpen heeft
met schrijven is Anja Ridder. Dank je wel en succes met je carrière verder. Ruud en
Jelle, altijd vriendelijk en in voor een geintje, succes met jullie projecten. De pilaren
van MolGen zijn Peter, Mozes, Mirelle, Jannet, Siger, Harma en Anne Hesseling, jullie
werk is zeer waardevol voor het draaiend houden van de vakgroep wat mij betreft,.
Jullie worden dan ook hartelijk bedankt voor de afgelopen tijd. Andere MolGenners
die ik nu niet bij naam heb genoemd, dank jullie wel voor je bijdrage aan de prettige
werksfeer.
Aan mijn studenten heb ik veel plezier gehad, van jullie allemaal heb ik wat geleerd.
Ik hoop jullie ook van mij. Sommige projecten hebben geresulteerd in het bijspijkeren
van mijn kennis, zoals de bachelorprojecten van Matthijs Schokker en Joachim
Jungmann. Het project van Erik Dalenoord gaf me dankzij zijn experimenten een
inzicht in de transcriptie van yfiA. However, the output of Dorota was so great that
she (completely deserved) made it to the publication of Chapter 4. Good luck all
of you with your career, I hope some day you think again of our time at MolGen.
Niet alle experimenten staan in dit boekje. Ik heb ook projecten gedaan samen
met Da Liu, Antoine van Ooijen en Hermie Harmsen. Helaas is gebleken dat
het ribosoom een lastig onderdeel van de cel is om te bestuderen. Ik waardeer
jullie inzet en bedank jullie ook voor alle hulp. Ik vind het nog altijd jammer dat
onze samenwerking niet heeft geleid tot concrete resultaten in mijn proefschrift.
Linda Franken en Marc Stuart, dank voor een fijne samenwerking die leidde tot een
van de mooiste figuren uit het boekje wat mij betreft. Voor de vetzuuranalyses ging ik
graag richting UMCG om de monsters af te leveren bij Ingrid Martini uit de vakgroep
van Frits Muskiet. De mensen van mijn leescommissie, bedankt voor alle tijd en
energie die jullie in mijn proefschrift hebben gestopt. STW, de technologiestichting
en zijn partners die het consortium financieel hebben ondersteund, bedankt voor het
mogelijk maken hiervan. Vergaderen met jullie werknemers heb ik altijd erg plezierig
ervaren.
My new colleagues at Chr. Hansen, thanks for the help during the last few months.
I am proud of being a scientist at such a company. The work is so nice and I look
forward to the upcoming exiting times in LAB research.
Nu wil ik niet paternalistisch doen, maar toch een advies. Blijf sporten tijdens je aio-tijd.
Al is het nog zo druk, de fysieke uitdaging is nodig naast de geestelijke. Volleyballers
van Veracles, bedankt voor alles wat jullie voor me betekend hebben. Er zijn veel
Veraclessers die vrienden zijn geworden, dat is me een waar genoegen. En we blijven
groen-zwart voor het leven. Jullie allemaal opnoemen zou het dankwoord dikker
maken dan mijn hoofdstukken, al moet ik toch de Schneelite even noemen. Hetzelfde
geldt voor Bedum H2, jullie hebben me een mooi anderhalf seizoen gegeven. Geen
212
spijt van mijn transfer gehad.
Cara Elena, mille grazie per tutte le belle lezioni. Spero di non dimenticare l’italiano,
sopratutto che ora apprendo danese. Anche gli altri studenti, grazie per una bella
atmosfera.
Dan kom ik nu uit bij mijn vrienden. Jeroen en Jantien, Lies, Heijman, Michiel,
Willemijn, Niek en Sven. Bedankt voor jullie etentjes, bioscoop- en marktbezoekjes,
de goede wijnen of sterker, al jullie raad en daad. Herman Schonewille en Richard
Otten, zit ik toch weer in het buut’nland. Nu jullie weer! Markie W, Ernstje, Willem de
Vries, binnenkort maar eens in Denemarken gekke-dingen-eet-avondje doen, kan wel
uit hier. Ali Tfayli, mon grand, quel honneur de te connaître, notre voyage au Liban
m’a changé. Alle vrienden, dank jullie wel voor het luisterende oor bij al mijn sores
op het lab. Ik ben jullie erkentelijk voor jullie getoonde begrip. Voor veel van mijn
vrienden geldt dat ik ze lang ken, we hebben al heel wat meegemaakt samen. Ik hoop
dat er nog vele jaren volgen met jullie.
Lieve papa en mama, Rik en Lotte. Ik kom uit een warm gezin, en we zien elkaar
graag. Dank jullie wel voor het vormen van mijn persoonlijkheid. Met de fysieke
afstand tussen ons in is dit soms lastig, maar we spreken elkaar gelukkig vaak genoeg.
Shanna, je bent als een zusje, en ik waardeer je prikkelende opmerkingen. Opa van
Dalen, uw wijze woorden: ‘Al wentelend van trap tot trap bereik ik de hoogste
wetenschap’ zijn uitgedrukt in de voorkant van mijn boekje. Ik hoop dat u het mooi
vindt. Oma Eckhardt, wat had opa het mooi gevonden, zijn kleinzoon een doctorstitel.
Ik ben blij dat u het meemaakt en hoop nog vele jaren even te horen wat u nou toch
weer ondernomen heeft. Oma van Dalen en opa Eckhardt worden gemist. De rest
van de familie van Dalen en familie Eckhardt, ik waardeer jullie aanwezigheid.
We zien elkaar graag met feestjes, dus organiseer ik er maar weer eens een.
Mijn schoonfamilie, ik heb jullie niet uitgekozen maar ‘kreeg’ jullie erbij.
Bedankt ook voor jullie begrip van de afgelopen periode en van onze keuze
naar Denemarken te gaan. We maken veel mee samen, en ik vind het prachtig
hoe jullie zijn als tante of als grootouders. Elize viert met jullie Sinterklaas!
Helemaal op het eind kom ik bij mijn gezinnetje. Beatrix, het is nu toch echt klaar.
Gooi het proefschrift maar uit het raam als je dat dan zo graag wilt. Ik druk ik wel
een extra exemplaar af. Je bent een lieverd, en een geweldige mama. Dat je met me
mee bent gegaan naar Denemarken zal ik niet vergeten. Elize, als papa zich niet zo
goed voelde of als ik moe was sleepte je me er doorheen. Je hoefde maar te lachen en
ik vergat alle problemen. Alle levensfasen ervaar ik als hoofdstukken in het boek van
mijn leven. Dit proefschrift is het tastbare bewijs van één van die hoofdstukken en
daarom is dit boekje voor jou.
Thomas H. Eckhardt
213