Download Transient responses and adaptation to steady state: Gene regulation

Transient responses and adaptation to steady state:
Gene regulation from a population perspective
Erez Braun and Naama Brenner
Depts. Of Physics and Chemical Engineering
Technion, Israel Institute of Technology,
Haifa, 32000, ISRAEL
Genetic regulatory networks in yeast are currently at the center of interest. Most recent studies
focus on structural, static or short-term properties of these networks from a cellular genomewide perspective. However, genetic regulatory networks are dynamical systems, and
studying their long-term dynamics under controlled conditions is necessary for their complete
characterization. Since gene expression is a relatively slow process, its dynamics displays
behavior with time scales comparable to and longer than the cell cycle. This suggests the
relevant “systems” level for such problems as that of the cell population. Our research focuses
on the dynamics of regulatory networks in yeast, emphasizing the perspective of clonal
populations. This perspective entails following the dynamics for many generations under
controlled condition; carefully distinguishing between transient responses and steady state;
measuring gene expression at single-cell resolution to account for cell variability; taking into
account population factors such as protein inheritance; and more.
Using our specially designed experimental system, we have demonstrated novel adaptive
dynamics in the well-known GAL system in yeast. This is a classic model for a eukaryotic
genetic switch, induced by galactose and repressed by glucose. We followed the expression of
a reporter gfp under a GAL promoter at single-cell resolution in large populations of yeast
cells. Experiments were conducted for long time scales, several generations, while controlling
the environment in continuous culture. This combination enabled us, for the first time, to
distinguish between transient responses and steady state. We have found that both galactose
induction and glucose repression are only transient responses. Over several generations, the
system converges to a single robust steady state, independent of external conditions; this
adaptation is physiological and does not involve random genetic mutations. Thus, at steady
state the GAL network loses its hallmark functionality as a sensitive carbon source rheostat.
This result suggests that, while short-term dynamics are determined by specific modular
responses, over long time scales inter-modular interactions take over and shape a robust
steady-state response of the regulatory system. [1]
To shed light on the functional significance of this adaptation, we next placed an essential
gene (HIS3) under GAL regulation. This construction mimics the process of gene recruitment,
where one gene is placed under a foreign regulation system, which is a major driving force in
evolution. Growing the cells in glucose and using a competitive inhibitor, we were able to
load the GAL system and apply variable degrees of selection pressure on the population. The
results show that the population utilizes the adaptation from glucose repression to overcome
the pressure and survives otherwise lethal conditions. This experimental technique enables to
study the adaptive capabilities of the regulatory system in a yeast population. [Unpublished].
[1] E. Braun and N. Brenner, Phys. Biol. 1, 67 (2004).
Mass Isotope Ratio Analysis of Intracellular Metabolites by LC-MS/MS: New Roads
towards Accurate Steady State and Dynamic Fluxome Analysis
W.A. van Winden, L. Wu, M.R. Mashego, W.M. van Gulik, J.C. van Dam, C. Ras,
A.M. Pröll, J.L. Vinke, J.J. Heijnen*
*Dept. Of Biotechnology, Delft University of Technology, Julianalaan 67, 2628 BC, Delft, The
Netherlands, tel. :+31-15-2785307, fax : +31-15-2782355. ([email protected])
Rational metabolic engineering of micro-organisms consists of an iterative sequence of three
stages: culturing and quantitative analysis of micro-organisms, redesign of the microorganism and, finally, synthesis of a new genotype that has improved phenotypical features.
The quantitative data cover four levels of the phenotype: fluxome, metabolome, proteome,
and transcriptome. In case of highly interconnected metabolic and regulatory networks that
are to be redesigned, quantitative analysis will have to include a mathematical modelling step.
Our group focuses on the development and application of tools for the quantitative analysis
and modelling of the intracellular metabolome (the size of the metabolite pools) and fluxome
(the rates of the metabolic reactions) in the central carbon metabolism of Saccharomyces
cerevisiae.
Liquid chromatography coupled to tandem mass spectrometry forms the heart of our quantitative analysis. We can routinely analyse most intermediates of the glycolysis and TCA cycle,
plus some intermediates of the pentose phosphate pathway. Recently, we developed a method
for the analysis of the mono, di and triphosphate forms of four nucleotides. Despite the use of
optimized protocols and equipment for rapid sampling, quenching and extraction of the
biomass [1], quantitative recovery of the metabolites remains challenging due to the low
concentrations and fast intracellular turnover rates of the metabolites and due to the multiple
processing steps of the sample.
Complete recovery is not required when measuring the mass isotopomer fractions of
metabolites in biomass that is grown on partially 13C-labeled substrates, since the
differentially 13C-labeled metabolites are co-processed, co-eluted in the LC and co-ionised in
the MS-inlet and only separated in the final stage of the analysis. This application of LCMS/MS in our laboratory adds to the established 13C-NMR and GC-MS methods for 13Clabeling analysis, which are increasingly used for steady state fluxome analysis [2].
We also used the above advantage of differentially 13C-labeled metabolites in a new method
for metabolome quantification following a substrate pulse to the fermentor. To do so, we
cultured 100% uniformly 13C-labeled biomass and processed the biomass to obtain a mix of
fully 13C-labeled intracellular metabolites. Fixed amounts of this mix are added to samples of
an unlabeled culture of which the metabolome is to be analysed. The 13C-labeled metabolite
extract then serves as a mix of ideal internal standards for LC-MS/MS analysis [3]. The mix
has been show to contain 13C-labeled equivalents of all the components we currently analyse.
An additional advantage of the use of the approach is that the 13C-labeled intermediates serve
as an identity check of compounds that are not commercially available: they should co-elute
with their unlabeled equivalent and should have the correct number of carbon atoms in the
mother molecule and daughter fragments in the MS/MS.
References
[1] Lange, H.C. et al. Biotechnol. Bioeng., 75, 406 – 415, 2001
[2] Sauer, U. Curr. Opinion Biotechnol., 15, 1 - 6, 2004
[3] Mashego, M.R. et al. Biotechnol. Bioeng., 85, 620 – 628, 2004
Keywords: yeast, LC-MS/MS, mass isotope distribution, fluxome, metabolome
Session: New tools/ technology for physiological studies
Dynamics of the yeast cell cycle interactome - an integrative systems
biology approach
Ramus Wernersson, Ulrik de Lichtenberg, Lars Juhl Jensen, Søren Brunak & Peer
Bork
Center for Biological Sequence Analysis, BioCentrum-DTU, The Technical University
of Denmark, Building 208, DK-2800 Lyngby, Denmark
In recent years, a large number of high-throughput experimental techniques have
been developed and applied in particular to the yeast Saccharomycs cerevisiae.
These include DNA microarrays, protein-protein interaction screens, subcellular
localization experiments and chromatin IP investigations of protein-DNA interactions.
This has made yeast the most widely used platform for large-scale analysis such as
studies of biological networks. Most of these analysis, however, have dealt with just
one data type and mainly focussed on global proporties of the transcriptome or
interactome. Typical examples include the identification of subsets of genes
regulated in response to particular conditions or static topological proporties of
interactions networks.
In this work, we construct an interaction network for the yeast cell cycle and analyze
the dynamics of this network by integrating protein-protein interaction data with
information on gene expression, protein subcellular localization and post-translational
modifications.
The resulting time-dependent interaction network not only reveals novel components
and modules, it also shows that only two thirds of the subunits of the central cell
cycle machineries are periodically transcribed. Yet, all modules but one are subject to
transcriptional control through at least one periodic component. This appears to be a
general design principle which provides temporal control of complex assembly and
function. Our work also reveals that the proteins that are periodically expressed are
often also subject to additional regulation, through phosphorylation by the cyclin
dependent kinase Cdc28p (Cdk1).
Increasing fermentative capacity: an integrative approach
Sergio Rossell*, Coen C. van der Weijden, Alexander Lindenbergh, Arthur L. Kruckeberg,
Barbara M. Bakker & Hans V. Westerhoff.
BioCentrum Amsterdam, Faculty of Earth and Life Sciences, Department of Molecular Cell
Physiology. De Boelelaan 1085 NL-1081 HV Amsterdam. The Netherlands.
* Corresponding author:
Tel: +31 20 4446966
Fax: +31 20 4447229
[email protected]
Fermentative capacity (FC) is defined as the specific rate of carbon dioxide production under
anaerobic conditions with excess of sugar. It is a well-defined physiological trait and a key
quality parameter in the industrial production of the yeast Saccharomyces cerevisiae. Nutrient
starvation, which is also an industrially relevant phenomenon, has been reported to result in a
decreased fermentative capacity. However, it has not been possible to correlate this decrease
with the activity of any single enzyme activity. Here we use an integrative approach to
address the problem of understanding and then perhaps engineering a physiological trait such
as FC. We have evaluated the activities of all the enzymes involved, the fluxes through the
pathway, and those that branch from it. In contrast to earlier attempts we have included the
determination of the glucose transport activity and of the branching fluxes. These
measurements were performed for three different conditions: unstarved (exponentially
growing), nitrogen-starved, and carbon-starved (deprived of a nitrogen or carbon source,
respectively, for 24 h); and for two different strains: a wild type (CENPK 113-7D) and a
hxk2∆ deletion mutant (KY116) that lacks hexokinase II.
In the wild type, both types of nutrient starvation led to a decrease in FC, which coincided
with a decrease in the glucose transport capacity. In the case of nitrogen starvation, glucose
transport activity quantitatively accounted for the reduction in glucose consumption. A
surplus ethanol production was observed, which could be explained by the degradation of
storage carbohydrates. In carbon-starved cells 40 % of the decrease of glucose consumption
could be explained by a decrease of glucose transport capacity, implying that the other 60 %
is regulated metabolically. Surprisingly, no other glycolytic enzyme activities were reduced
and no detectable amounts of storage carbohydrates were synthesized. We anticipate that
changes outside glycolysis (e.g. ATP utilization) contribute to the decreased FC. In the hxk2∆
mutant strain, nutrient starvation did not lead to a significant decrease of the FC, although it
did result in a decrease of some enzyme activities including glucose transport. Metabolic
modeling and regulation analysis are applied to understand the relative contribution of each of
the enzymes to the overall fermentative capacity.
Keywords: nutrient starvation, glucose transport, glycolysis, metabolic modeling
Impact of Cyclic AMP Signaling on Cell Cycle Progression
and Energy Metabolism in Saccharomyces cerevisiae
D. Müller, L. Aguilera-Vázquez, H. Diaz-Cuervo, E. Guerrero-Martín,
J. O. Marquetand, P.K. Murugan, A. Niebel, and M. Reuss
Institute of Biochemical Engineering, University of Stuttgart,
Allmandring 31,D-70569 Stuttgart, Germany
The pursuit of a systems-level understanding of biological processes constitutes one of
the central goals of systems biology. We have used a combination of experimental
techniques and mathematical modeling to analyze the impact of signal transduction via
cyclic AMP (cAMP) and protein kinase A (PKA) on the coordination of energy
metabolism and cell cycle progression in Saccharomyces cerevisiae. The aim is to
develop a single cell model which provides a combined description of metabolic
processes and signaling events in conjunction with a dynamic representation of the cell
cycle.
Experiments in synchronous cultures and in chemostat cultures have demonstrated
distinct dynamics of cAMP and the downstream targets of PKA in energy metabolism
during the cell cycle. However, the upstream signal responsible for the differential
activation of this signaling pathway under these conditions is elusive. We have obtained
evidence suggesting that intracellular nucleotide concentrations may play a role in
linking cAMP signaling and hence cell cycle progression to the energetic state of the
cell, at least under glucose-limited conditions.
A modular mathematical model has been developed, which aims at capturing not only
the dynamics of cAMP-PKA signal transduction, but also that of central carbon
metabolism and the cell cycle machinery itself. The signaling module comprises the
dynamics of cAMP synthesis and degradation as well as the resulting PKA activation.
Simulation studies demonstrate that the model is able to reproduce, e.g., the adaptive
response of the signaling pathway to a persistent stimulus of extracellular glucose
observed in experiments. The metabolic module is based on a previously established
model of glycolysis and the pentose phosphate pathway, which has been extended to
include the dynamics of the storage carbohydrates trehalose and glycogen along with
their associated regulation by PKA. Cell growth is described based on the output from
the metabolic module. A kinetic model of the yeast cell cycle machinery [1,2] constitutes
the basis of the cell cycle module, which has been amended to account for PKA
influence on cell cycle progression. Population distributions of cell cycle stages were
quantified by fluorescence microscopy and were subsequently employed for parameter
estimation.
Upon integration of the developed modules, the resulting single-cell model will yield a
dynamic description of the cAMP-dependent regulation of metabolism and cell cycle
progression during the different cell cycle phases. The chosen modular approach is
potentially applicable to systems of medical importance where the link between signal
transduction, energy metabolism, and the cell cycle is crucial, e.g. when modeling tumor
cell behavior. Moreover, the model can also serve as a basis for a segregated
description of heterogeneous cell populations, an issue of major importance in the
operation of large-scale bioprocesses.
References
[1] Chen KC, Csikász-Nagy A, Györffy B, Val J, Novák B, Tyson JJ (2000) Mol Cell Biol 11:369-391.
[2] Cross, FR (2003) Dev Cell 4(5):741-52.
Vertical Genomics: from gene expression to function, … and back
J. Bouwman, R.J.M. van Spanning, H.V. Westerhoff, and B.M. Bakker*
Molecular cell physiology, Molecular Cell Biology (IMC), Vrije Universiteit,
Amsterdam
Functional behavior of cells largely occurs at the level of 'fluxes', i.e. rates of
processes such as product formation, protein production and gene expression. How
these processes are interconnected has not been studied quantitatively. Thanks to
genomics, it should be possible to evaluate the implications of processes at the level
of transcription for functional fluxes. Glycolysis in yeast is a good model system to
test this relation, for it is one of the few pathways for which the kinetic properties of
the enzymes are known sufficiently to calculate the flux from the enzyme activities
and yeast can be brought under the well-defined steady-state and transient conditions.
To measure alterations in the expression of glycolytic genes a new method will be
used for quantitative and rapid monitoring mRNA levels. This method, MLPA
(Multiplex Ligation-dependent Probe Amplification) uses amplifiable probes of a
certain length. These are made by a ligation reaction dependent on the amount of
cDNA present in the sample. A mixture is used of different MLPA-probes that are
amplified with the same primers and in which each probe is specific for a certain
mRNA. In addition protein levels, metabolite levels and fluxes will be measured.
Ultimately these data and bioinformatics shall then deduce precisely how altered gene
expression leads to altered function.
Jildau Bouwman
Department of Molecular Cell Physiology
Molecular Cell Biology (IMC)
Faculty of Biology
Vrije Universiteit
De Boelelaan 1085
1081 HV Amsterdam (The Netherlands)
Tel: +31 20 4446966
Fax: +31 20 4447229
Email: [email protected]
Rob J.M. van Spanning
Email: [email protected]
Barbara M. Bakker*
Email: [email protected]
H.V. Westerhoff
Email: [email protected]
Discovering Activated Regulatory Networks in the DNA Damage Response Pathways of
Yeast
Christopher Workman, Scott A. McCuine, Craig Mak, Trey Ideker
UC San Diego, La Jolla, USA;
Jean-Bosco Tagne, Richard A. Young
Whitehead Institute, Cambridge, USA
To further elucidate DNA alkylation damage response in yeast, we have applied an
integrative approach to study the regulatory pathways responding to the DNA damaging agent
methyl methanesulfonate (MMS). Models were constructed using data derived from classical
genomic and microarray approaches, and were refined using computational systems biology
approaches. After a phenotypic screening was used to identify a set of transcription factors
important for DNA damage response, the implicated regulatory pathways were systematically
interrogated using chromatin immunoprecipitation (chIP-chip) and DNA microarrays to
monitor protein-DNA interactions and genome-wide expression patterns in single geneknockout strains. These data were integrated and modeled using tools for comparison of
networks across multiple conditions, for statistical identification of expression-activated
network regions (ActiveModules)1, and visualized using software we have developed for
operating on network models (Cytoscape)2. Using this systems approach, we have generated
new hypotheses revealing the complex web of interactions, cellular factors, and mechanisms
involved in DNA damage response.
1.
Ideker, T., Ozier, O., Schwikowski, B. & Siegel, A.F. Discovering regulatory and
signaling circuits in molecular interaction networks. Bioinformatics 18 Suppl 1,
S233-40 (2002).
2.
Shannon, P. et al. Cytoscape: a software environment for integrated models of
biomolecular interaction networks. Genome Res 13, 2498-504 (2003).