Download The PIN-domain ribonucleases and the prokaryotic VapBC toxin

Protein Engineering, Design & Selection vol. 24 no. 1 –2 pp. 33–40, 2011
Published online November 29, 2010 doi:10.1093/protein/gzq081
REVIEW
The PIN-domain ribonucleases and the prokaryotic
VapBC toxin – antitoxin array
Vickery L.Arcus 1,3, Joanna L.McKenzie 1,
Jennifer Robson 2 and Gregory M.Cook 2
1
Department of Biological Sciences, University of Waikato, Private Bag
3105, Hamilton 3240, New Zealand and 2Department of Microbiology and
Immunology, Otago School of Medical Sciences, University of Otago,
P.O. Box 56, Dunedin 9054, New Zealand
3
To whom correspondence should be addressed.
E-mail: [email protected]
Received September 21, 2010; revised September 21, 2010;
accepted September 23, 2010
Edited by Daniel Otzen
The PIN-domains are small proteins of ∼130 amino acids
that are found in bacteria, archaea and eukaryotes and are
defined by a group of three strictly conserved acidic amino
acids. The conserved three-dimensional structures of the
PIN-domains cluster these acidic residues in an enzymatic
active site. PIN-domains cleave single-stranded RNA in a
sequence-specific, Mg21- or Mn21-dependent manner.
These ribonucleases are toxic to the cells which express
them and to offset this toxicity, they are co-expressed with
tight binding protein inhibitors. The genes encoding these
two proteins are adjacent in the genome of all prokaryotic
organisms where they are found. This sequential arrangement of inhibitor-RNAse genes conforms to that of the socalled toxin–antitoxin (TA) modules and the PIN-domain
TAs have been named VapBC TAs (virulence associated
proteins, VapB is the inhibitor which contains a transcription factor domain and VapC is the PIN-domain ribonuclease). The presence of large numbers of vapBC loci in
disparate prokaryotes has motivated many researchers to
investigate their biochemical and biological functions. For
example, the devastating human pathogen Mycobacterium
tuberculosis has 45 vapBC loci encoded in its genome
whereas its non-pathogenic relative, Mycobacterium smegmatis has just one vapBC operon. On another branch of the
prokaryotic tree, the nitrogen-fixing symbiont of legumes,
Sinorhizobium meliloti has 21 vapBC loci and at least one of
these loci have been implicated in the regulation of growth
in the plant nodule. A range of biological functions has been
suggested for these operons and this review sets out to
survey the PIN-domains and summarise the current knowledge about the vapBC TA systems and their roles in diverse
bacteria.
Keywords: growth regulation/mRNA interferases/
mycobacteria/RNAse/VapBC
Preface
My research career began as a PhD student in Alan Fersht’s
lab in Cambridge in 1992. It was here that I joined the
efforts of the lab to unravel the folding pathway of the small,
single-domain proteins barnase and CI2. I studied the early
events on the folding pathway of barnase and this was my
first foray into molecular biology having come from a chemistry undergraduate degree. I enjoyed my time in Alan’s lab
immensely and I remain deeply indebted to Alan for all that
he taught me during those formative PhD years. I returned to
New Zealand in 1996 and, via a circuitous research route,
my group now works on bacterial PIN-domain ribonucleases
that have some uncanny similarities to barnase! Like the
barnase/barstar interaction, the PIN-domain ribonucleases are
co-expressed with cognate inhibitors that form a protein –
protein complex. Thus, I thought it fitting to write a review
of these domains for this special issue. Upon reflection, my
peripatetic research path over the last 18 years describes a
certain unintended symmetry between Cambridge and New
Zealand despite being on opposite sides of the world.
Vic Arcus, September 2010.
Introduction
The PIN-domains are a large protein family with representatives in all three major branches of life. They were originally
named for their sequence similarity to the N-terminal
domain of an annotated PilT protein (PilT N-terminal
domain), although this historical annotation stems from a
domain fusion between a PIN-domain and a PilT ATPase
domain. A functional link that connects the PIN-domains
with type IV pili (PilT), has not been demonstrated.
PIN-domains are small protein domains of 130 amino
acids that are identified by the presence of three strictly conserved acidic residues. Apart from these three residues, there
is poor sequence conservation across the family, however,
this is offset by the conservation of three-dimensional structure seen in 11 PIN-domain structures in the Protein Data
Bank. The three-dimensional structures cluster the conserved
acidic residues together in a putative active site.
The Pfam database (Finn et al., 2010) lists 3457 proteins
belonging to the PIN-domain family (PF01850) from 490
different species including eukaryotes, eubacteria and
archaea. PIN-domains are found in nearly half of all
sequenced prokaryotes including many important pathogens,
for example, Neisseria gonorrhoeae (Hopper et al., 2000;
Mattison et al., 2006) and Mycobacterium tuberculosis
(Arcus et al., 2005; Ramage et al., 2009), and yet their biological functions in these bacteria are not well understood or
studied. Biochemically, the PIN-domains are ribonucleases
and in eukaryotes, are associated with nonsense-mediated
decay of RNA (Takeshita et al., 2007) and processing of
pre-18S ribosomal RNA (Lamanna and Karbstein, 2009). In
prokaryotes, the vast majority of PIN-domain proteins are the
toxic components of so-called toxin– antitoxin (TA) systems.
# The Author 2010. Published by Oxford University Press. All rights reserved.
For Permissions, please e-mail: [email protected]
33
V.L.Arcus et al.
Fig. 1. A visual depiction of the hidden Markov model that defines the PIN-domain family of proteins. The height of each letter is proportional to the amount
of ‘information’ it provides about the respective position in the PIN-domain family. The width of each column is also an indication of the importance of this
position in defining the family. Sequence regions that contain insertions (and therefore almost no information about the family) are shown as dark and light
pink columns. The width of these columns (dark þ light pink) represents the expected length of the insertion and the width of the dark pink column is the
probability that at least one amino acid is inserted at this point in the sequence. Taken from PFam (Finn et al., 2010) using HMM Logo (Schuster-Böckler
et al., 2004).
In this case, the ‘toxicity’ arises by virtue of their ribonuclease activity. To offset the toxicity, the prokaryotic
PIN-domains are co-expressed with an inhibitor [in a manner
similar to barnase/barstar expression (Hartley, 1988)]. Where
the PIN-domains differ from barnase/barstar is in the fact
that the PIN-domain/inhibitor complex is not secreted, but
maintained inside the cell where proteolytic degradation of
the inhibitor leads to activation of the PIN-domain ribonuclease. The biological consequences of this activation are the
subject of a great deal of current research. For example, the
role of the TAs in the pathogenesis of M. tuberculosis, one
of the most devastating infectious microorganisms known in
humans, is of great interest (Gupta, 2009; Ramage et al.,
2009).
The PIN-domain TA systems are now called VapBC TAs
(virulence associated proteins), where VapB is the inhibitor
and VapC, the PIN-domain ribonuclease toxin. The proteins
are encoded by a vapBC operon, often with overlapping open
reading frames, and these TA systems are very widespread in
prokaryotes. Indeed, the vapBC genes are the largest family
of the nine TA families that have been classified (Gerdes
et al., 2005; Van Melderen and Saavedra De Bast, 2009).
Despite this, they remain the least well characterised.
Intriguingly, the vapBC TAs are expanded in number in the
genomes of several unrelated prokaryotes. For example, M.
tuberculosis has 45 vapBC operons encoded in its genome,
whereas the related environmental organism Mycobacterium
smegmatis has just one (Arcus et al., 2005; Ramage et al.,
2009). This phenomenon of expanded numbers in various
organisms spans the prokaryotic tree: the archaeal sulphurmetabolising organism Archaeoglobus fulgidus has 22
vapBC operons and the ammonia-oxidising soil bacterium
Nitrosomonas europaea has 13. This begs the question: what
are the biological roles of the vapBC operons in these
diverse bacteria? This review will outline the defining
34
features of PIN-domains and summarise the evidence for
their biochemical and biological functions. It will then
discuss the VapBC TA systems and collect the disparate evidence for their roles as response operons to stressors and the
changing environments encountered by the microbes in
whose genomes they reside.
PIN-domain bioinformatics, sequence and structure
Ten years ago, Clissold and Ponting (2000) used bioinformatics to predict that the PIN-domain proteins were
Mg2þ-dependent RNAses, and suggested that the conserved,
active site residues were similar in architecture to phage T4
RNAse H and the Flap endonucleases. They used
PSI-BLAST to detect remote sequence homologues and the
PIN-domain sequence signature is now more formally
encoded in a Hidden Markov Model (found at PFam) that
can be visualised using the tool HMM Logo
(Schuster-Böckler et al., 2004; Fig. 1). The low overall
sequence conservation for this large family is evident with
small stacks of letters at a majority of positions signifying
accommodation of a range of amino acids along most of the
sequence (Fig. 1). The three well-conserved acidic residues
are clear at positions 4, 40 and 93. A fourth acidic residue is
less well conserved at position 112. Another sequence
feature for the family is the presence of a polar residue (Thr,
Ser or Asn) at position i þ 1 or i þ 2 following the first conserved aspartic acid ( positions 5 and 6 in Fig. 1). Patches of
hydrophobic residues (e.g. positions 94 – 100) are indicative
of the hydrophobic cores of the folded PIN-domain—positions 94 –100 constitute a buried b-strand.
Subsequent bioinformatics analysis of the genomic context
for many PIN-domains in prokaryotes showed that in very
many cases the gene preceding each PIN-domain had features that resembled known transcription factors and this
The PIN-domain ribonucleases and the prokaryotic
Fig. 2. The structure of the PIN-domains. (A) A cartoon representation of a PIN-domain structure coloured blue-to-red from the N-terminus to the C-terminus.
The 5-stranded parallel b-sheet points out of the plane and has strand order 32145 from left to right. The third b-strand is yellow and partially obscured by
a-helices 3 and 4. (B) The same structure as a Ca trace showing the four conserved acidic residues, the conserved polar residue and the active site Mn2þ ion
(in black). (C) A PIN-domain dimer as a partially transparent surface, showing the active site channel and two Mn2þ ions (in black). Conserved residues are
visible beneath the surface. (D) A PIN-domain tetramer shown as an electrostatic surface. A tunnel is visible in the centre of the structure with protruding,
positively charged Lys residues inside the tunnel. Coordinates were taken from 2FE1 (Bunker et al., 2008) and 1V8P (Arcus et al., 2004) and images were
drawn with Pymol (http://www.pymol.org/).
gene formed an operon with the PIN-domain. This led to the
prediction that these operons were TA loci (Anantharaman
and Aravind, 2003; Arcus et al., 2005; Pandey and Gerdes,
2005).
Domain fusions are indicative of functional linkages
between domains (Marcotte et al., 1999) and in 5% of
cases where PIN-domains are found, they are fused with
TRAM domains, KH domains and AAAþ ATPase domains.
The TRAM and KH domains have general annotations as
RNA-binding domains and the AAAþ ATPase domains are
known to form hexameric ring structures that act as
ATP-driven molecular motors. However, in 95% of cases,
the PIN-domains are single domain proteins preceded in the
genome by a transcription factor.
The PIN-domain fold is described in SCOP (Andreeva
et al., 2008) as a 3-layer a/b/a sandwich containing a
5-stranded parallel b-sheet in the centre of the structure
(Fig. 2A). The topology of the protein traces repeating units
of b-strand, helix, helix, in a series of right-handed turns.
The stand order of the parallel b-sheet is 32145 and, in combination with the right-handed spiral topology, places helices
1 –4 above the b-sheet and helices 5 – 7 below it. This fold
brings together three or four conserved acidic residues to
form an active site which binds Mg2þ or Mn2þ ions
(Fig. 2B; Arcus et al., 2004; Bunker et al., 2008). The polar
residue at i þ 1 or i þ 2 following the first acidic residue
appears to play a structural role, often forming a hydrogen
bond with its neighbouring acidic residue and presumably
positioning it appropriately for catalysis (Fig. 2B).
All of the prokaryotic PIN-domain structures to date are
dimers and dimerisation configures the active sites in a
groove along the long-axis of the structure (Fig. 2C). In one
35
V.L.Arcus et al.
case, PAE2754 from Pyrobaculum aerophilum, two dimers
form a tetramer such that the four active sites are inside a
tunnel that is large enough to accommodate single-stranded
RNA with positively charged lysine residues projecting into
the tunnel. These would facilitate binding to the phosphate
backbone of the RNA (Fig. 2D; Arcus et al., 2004; Bunker
et al., 2008).
It is potentially dangerous for an organism to express a
cytosolic RNAse and so VapC proteins are nearly always
co-expressed with a cognate protein inhibitor, VapB, forming
an inert VapBC protein complex. VapC may then be activated by the degradation of its inhibitor, VapB. The vapBC
genes belong to a larger group of operons called the TA
operons.
TA operons
TA protein pairs were discovered more than 20 years ago as
factors which protect low copy number plasmids in bacteria
from segregational loss (Jaffe et al., 1985; Gerdes et al.,
1986). One protein of the pair is toxic to the cell and stable,
while its cognate antitoxin is unstable and requires continuous transcription to inhibit the toxin. Thus, if the plasmid is
lost during cell division, the daughter cell is left with the
proteins—one stable and toxic, the other, an easily degraded
inhibitor. With time, the antitoxin is degraded, leaving the
toxin to kill the plasmid-free cell. Here, the TA system operates as a suicide mechanism for those cells that do not carry
the plasmid. This has been variously described as ‘postsegregational killing’, ‘death upon curing’ (DOC), an ‘addiction’ system and plasmid stability elements (Van Melderen
and De Bast, 2009).
More recently, chromosomally encoded TA operons have
been identified by a range of bioinformatics techniques and
are present in a very wide range of bacteria and archaea. The
biological roles for some TA systems have been intensively
studied as a result of their association with antibiotic resistance, virulence factors and pathogenicity islands in pathogenic bacteria (Gerdes, 2000; Hayes, 2003). In these cases it
has been found that the TAs potentially play biological roles
associated with growth arrest under conditions of nutritional
stress (Gerdes, 2000), challenge by antibiotics (Sat et al.,
2001), DNA damage or genomic rearrangement (Hazan
et al., 2004).
There are nine families of TA operons. The toxic components exert their effects in different ways: The CcdB and
ParE toxins are DNA gyrase inhibitors; the RelE, Doc and
HigB toxins directly or indirectly cleave ribosome-associated
mRNAs; HipA is a protein kinase which targets EF-Tu; and
the VapC and MazF families are RNases which cleave
mRNA transcripts independently of the ribosome. The VapC
and MazF toxins have been collectively described as mRNA
interferases (Nariya and Inouye, 2008; Zhu et al., 2008)
whose biochemical activity results in the inhibition of translation (Robson et al., 2009). The VapBC family is the largest
TA family and is the only toxin that contains a PIN-domain
(Arcus et al., 2005; Sevin and Barloy-Hubler, 2007; Van
Melderen and De Bast, 2009).
The organisation and regulation of vapBC operons is
shown in Fig. 3. Two overlapping genes form the operon
and encode a transcription factor-like antitoxin (VapB)
and a PIN-domain ribonuclease, respectively (VapC).
36
Fig. 3. Organisation of a typical vapBC TA system. The vapB and vapC
genes overlap in an operon (shown as cyan arrows). These encode two
proteins that form a tight complex. This complex binds to inverted repeats
(IR) in the promoter region leading to auto-regulation of operon
transcription. Activation of VapC RNase activity is probably a result of
proteolytic degradation of the more labile VapB antitoxin.
These proteins form a benign protein – protein complex that
binds to the promoter DNA via the ribbon helix helix (RHH)
domain of VapB. Promoter-binding autoregulates vapBC
transcription. The mechanism of VapC activation is
unknown, but is thought to be via proteolytic degradation of
VapB. Once activated, VapC cleaves cohorts of mRNA transcripts presumably via sequence-specific RNAse activity analogous to MazF (Zhang et al., 2003; Zhu et al., 2008).
VapBC demographics in prokaryotes
The distribution of vapBC operons across prokaryotes raises
some intriguing evolutionary questions. What evolutionary
processes have led to their expansion in number in several
unrelated prokaryotes? What are the common factors among
the seemingly disparate organisms that contain large
numbers of vapBC operons? Several surveys have been
undertaken to automatically identify TA operons in fully
sequenced genomes and these provide a starting point for
hypotheses
about
their
distribution
(Sevin
and
Barloy-Hubler, 2007; Makarova et al., 2009). A list of 10
organisms whose genomes have been sequenced and contain
large numbers of vapBC operons is initially confounding
(Table I). These organisms are unrelated and there appears to
be little or no common factors of lifestyle or environment.
There is good evidence that the TAs can behave as selfish
genetic elements and that vapBC operons conform to this.
Indeed, the first TAs were identified on plasmids and implicated in plasmid maintenance acting as suicide elements in
daughter cells that do not inherit the plasmid upon cell division (Jaffe et al., 1985; Gerdes et al., 1986). Other studies
have shown that TAs placed on plasmids outcompete the
same plasmid without this operon over 100s of generations
(Cooper and Heinemann, 2000). Thus, it is highly likely that
TAs move around among the horizontal gene pool and
confer fitness on the mobile genetic elements which carry
them. The evolutionary origin of chromosomal copies of
The PIN-domain ribonucleases and the prokaryotic
Table I. Organisms harbouring large numbers of vapBC operons
Organism
Description
vapBC loci
Mycobacterium tuberculosis H37Rv
Microcystis aeruginosa
Pyrococcus kodakaraensis
Arthrospira maxima
Sulfolobus tokodaii
Archaeoglobus fulgidus
Gloeobacter violaceus
Microcoleus chthonoplastes
Caulobacter sp. K31
Actinobacterium, aerobic, mesophilic
Cyanobacterium, aerobic, aquatic
Euryarchaeota, anaerobic, hyperthermophilic
Cyanobacterium, aerobic, aquatic
Chrenarchaea, aerobic, hyperthermophilic
euryarchaeota, anaerobic, hyperthermophillic
Cyanobacterium, terrestrial, mesophilic
Cyanobacterium, aquatic, halophilic
Proteobacteria, aerobic, aquatic, psychrophilic
45
34
28
26
24
22
19
19
18
Data taken from (Sevin and Barloy-Hubler, 2007; Makarova et al., 2009).
TAs, as mobile selfish elements, is also supported by their
association in genomes with large numbers of transposable
elements. For example, the organism with the most number
of TAs, Microcystis aeroginosa has 11.8% of its genome
composed of insertion sequences and inverted-repeat transposable elements (Kaneko et al., 2007). TAs in the genome of
M. tuberculosis have also been associated with mobile
elements (Arcus et al., 2005; Ramage et al., 2009). This is
further reinforced by the observation that the closest homologues of many of the M. tuberculosis TAs are found in unrelated soil organisms and do not conform to the phylogeny of
the mycobacteria (VLA, unpublished results).
Although it seems likely that the evolutionary origin of the
TAs is among the horizontal gene pool and that they are able
to invade plasmids and genomes and behave as selfish
elements, it is also likely that they have been co-opted over
time for functions which benefit the host organism. There are
several lines of evidence for this co-option hypothesis.
First, TA systems can be lost from a genome and so their
selfish behaviour does not permanently fix them in a parent
genome. In the cases of the obligate intracellular pathogens,
Mycobacterium leprae and Treponema pallidum, there are no
TAs and the TA loci have most probably been lost via
reductive evolution. For M. leprae there are pseudo TA loci
containing stop codons and this attests to the loss of
these elements over time. In contrast, their facultative intracellular relatives, M. tuberculosis, Mycobacterium bovis and
Treponema denticola, respectively, have large numbers of
TA loci. Genome decay along with concomitant loss of TAs
can also be seen when comparing the genomes of Rickettsia
conorii (15 TAs) and Rickettsia prowzekii (0 TAs). Genome
decay is thought to be an important evolutionary process in
the context of intracellular pathogens whose evolution has
driven these organisms into more and more specialised intracellular niches (Andersson and Andersson, 1999).
Secondly, a number of TA knockout strains have deleterious phenotypes. A hipBA deletion mutant showed a
reduction in Escherichia coli persister cells (a subpopulation
of cells with low metabolic rates) and by inference, a
reduction in antibiotic tolerance (Lewis, 2007; Schumacher
et al., 2009). A Myxococcus mazF deletion mutant was
unable to make the transition from vegetative growth to fruiting body formation and MazF was shown to be required for
programmed cell death during Myxococcus development
(Nariya and Inouye, 2008). These observations are tempered
by the fact that many TA knockout strains show no discernable phenotype under experimental conditions (Tsilibaris
et al., 2007; Robson et al., 2009).
Thus, it appears likely that the chromosomally encoded
TAs have their evolutionary origins among mobile genetic
elements, but have then been co-opted for contemporary
functional roles in many organisms. The VapBC TA systems
conform to this evolutionary scheme and have been studied
in a number of diverse organisms where their contemporary
functions are broadly associated with growth regulation in
response to changing environments.
Characterised VapBC TA systems
NtrPR from Sinorhizobium meliloti
The ntrPR operon of Sinorhizobium meliloti (a nitrogenfixing symbiont of legumes), is a member of the VapBC TA
family. It was originally discovered as a mutation in the
ntrPR operon that improved nitrogen fixation efficiency in
alfalfa nodules (Olah et al., 2001). The ntrP open reading
frame overlaps ntrR and the operon that is negatively regulated by the NtrPR protein complex by binding to promoter
DNA through NtrP (Bodogai et al., 2006). The ntrP gene
encodes a protein of 90 amino acids that includes a SpoVT/
AbrB DNA-binding domain and ntrR encodes a 134 amino
acid protein that belongs to the PIN-domain protein family
(Puskas et al., 2004). NtrP binds to the direct repeat
sequences 50 GGCATATACA-TTA-GGGATATACA30 and the
second repeat overlaps the transcriptional start site causing
auto-inhibition of transcription (Bodogai et al., 2006). Thus,
NtrPR satisfies all the elements of a VapBC TA as depicted
in Fig. 3.
Expression of NtrR in E. coli leads to a reduction in
colony formation and cell growth (Bodogai et al., 2006).
Although it was thought that NtrPR regulated nitrogen fixation genes in S. meliloti, a comparison between the transcriptional profile of wild type and DntrPR strains implicated
this TA module in the regulation of a range of metabolic
transcripts. Therefore it was hypothesised that NtrPR adjusts
metabolic processes under symbiosis and other stress conditions (Puskas et al., 2004). As NtrR belongs to the
PIN-domain family and the conserved acidic residues are
present, it is likely that it exerts its effect via RNase activity
that targets a cohort of transcripts for degradation.
Intriguingly, S. meliloti is one of the organisms with a
greatly expanded number of TA loci [42 TAs, (Makarova
et al., 2009)]. Of these, 21 belong to the VapBC family
(Sevin and Barloy-Hubler, 2007) and these are evenly spread
among the chromosome and two megaplasmids.
37
V.L.Arcus et al.
Fig. 4. The structure of FitAB bound to inverted repeats from its promoter DNA. (A) At the top of the figure FitA forms a dimer (two chains coloured pink
and purple). This FitA dimer binds to one half of the DNA inverted repeat. A FitA dimer is also seen at the bottom of the figure (two chains coloured blue and
light blue) binding to the second half of the DNA inverted repeat. Each FitA monomer binds to a FitB monomer—the two FitB dimers lie to the left and right
of the DNA (FitB chains are coloured clockwise from top right—light green, dark green, tan, yellow). Double-stranded DNA lies behind the hetero-octomeric
FitAB structure. (B) The C-terminal peptide of a FitA monomer is shown in grey and side chains are shown as sticks. Arg68 from FitA binds into the active
site of a FitB monomer, which is depicted as an electrostatic surface.
FitAB from Neisseria gonorrhoeae
The fitAB (fast intracellular trafficking) locus is a VapBC TA
system present on the chromosome of the sexually transmitted pathogen N. gonorrhoeae. In a similar manner to
ntrPR, the fitAB locus was first identified via mutations in
this locus that resulted in N. gonorrhoeae traversing epithelial cells at a faster rate than the wild type. The DfitAB
mutant also showed a faster intracellular replication rate
within epithelial cells (Wilbur et al., 2005). Neisseria gonorrhoeae infection can be asymptomatic, but remain culturable
and transmissible and thus, a human host may act as a carrier
for the disease. Carriers are the key to the spread of the
disease, however, the mechanism of the asymptomatic persistence of the organism is unknown. One hypothesis is that
the bacteria evade the immune response by residing within
the epithelial cells instead of crossing the cell into the subepithelial matrix. This spurred on the search for trafficking
mutants.
FitAB is hypothesised to slow intracellular trafficking and
replication of N. gonorrhoeae. The model proposed by
Mattison et al. (2006) is as follows: FitAB binds to its promoter
DNA when N. gonorrhoeae are in an extracellular environment. This sequesters the FitAB complex and represses transcription of the fitAB locus. Upon invasion into epithelial cells,
FitAB is released from the DNA and dissociates to slow N.
gonorrhoeae replication. This model is consistent with FitAB
behaving as a VapBC TA where VapC is released and slows
growth through disruption of translation via mRNA cleavage.
The fitAB locus has a one-base-pair overlap between the
two genes and the N-terminus of FitA is a RHH protein
domain that binds to inverted repeat sequences in the fitAB
promoter to regulate its own transcription (Wilbur et al.,
2005; Mattison et al., 2006). The DNA target for the FitAB
38
protein complex is an inverted repeat found in the promoter,
TGCTATCA-N12-TGATAGCA, which covers the putative
-10 promoter sequence for the locus. FitA binds to this
sequence with relatively weak affinity (Kd ¼ 178 nM) but the
binding affinity is improved 38-fold when FitA forms a
protein – protein complex with FitB (Kd ¼ 4.5 nM; Wilbur
et al., 2005).
The striking three-dimensional structure of FitAB bound
to a DNA fragment was solved by Mattison et al. (2006).
Four FitAB heterodimers associate to form a tetramer of
FitAB heterodimers bound to DNA. The FitA homodimer is
primarily responsible for recognition of the inverted repeat
sequences on the DNA and the tetramer of FitAB heterodimers associates in a circle that determines the separation of
the inverted repeat sequences (Fig. 4A).
FitB is a dimer and, like all other prokaryotic
PIN-domains, shows an extensive dimer interface formed by
interactions between a3 and a4 of each monomer. The four
conserved residues across PIN-domain proteins (Asp5,
Glu42, Asp104 and Asp122) are at the C-terminal end of the
b-sheet and form a negatively charged pocket near the centre
of the FitB molecule. No nucleic acid cleavage of ssDNA,
dsDNA, ssRNA or dsDNA by the FitAB complex was
observed (Mattison et al., 2006). This is because Arg68,
present at the C-terminus of FitA binds into the acidic
pocket that constitutes the active site of FitB, where it interacts with the carboxyl groups of Asp5, Glu42 and Asp104 of
FitB and forms strong electrostatic interactions that are not
easily displaced. Thus, the efficient inhibition of FitB by
FitA is clearly demonstrated in the structure (Fig. 4B).
The cellular RNA targets and the role of FitAB in N.
gonorrhoeae virulence are currently unknown. However, all
the features of FitAB are consistent with the VapBC TAs
The PIN-domain ribonucleases and the prokaryotic
and so it is a reasonable hypothesis that FitB cleaves cohorts
of mRNA transcripts which result in slow intracellular
growth of N. gonorrhoeae. Unlike S. meliloti, fitAB is the
sole vapBC TA loci in the genome of N. gonorrhoeae, which
harbours just five TA loci in total (two copies of mazEF and
single copies of vapBC, relBE and higAB).
VapBCs from mycobacteria
Tuberculosis (TB) is one of the most devastating infectious
diseases worldwide. Mycobacterium tuberculosis, the causative agent of this disease claims 1.7 m people per annum.
In 1993, the World Health Organisation (WHO) declared TB
a global emergency as an estimated one-third of the world’s
population carries the organism and around 9 – 10 m new
cases of TB are reported each year. One of the main problems in the treatment of TB infection is the capacity of M.
tuberculosis to enter a persistent stage that is less susceptible
to antibiotics, dictating long treatment regimes. The global
TB burden has recently been compounded by the dramatic
increase in multi-drug resistant and extensively drug resistant
strains of M. tuberculosis and the increased susceptibility of
HIV-infected individuals (Caminero et al., 2010).
Mycobacterium smegmatis has been used as a model
species for studying the Mycobacteria genus. It is fast
growing, non-pathogenic and provides a powerful molecular
model for studying mycobacterial physiology. Unlike
M. tuberculosis, M. smegmatis has a small number of TA
operons encoded in its genome: single copies of phd/DOC,
mazEF and vapBC. Mycobacterium smegmatis VapBC has
all the hallmarks of a bona fide TA system (Fig. 3): vapB
and vapC are polycistronic; formation of a benign VapBC
protein complex upon expression; binding of VapBC to
inverted repeat DNA sequences in the promoter region
leading to auto-regulation of transcription; inhibition of
translation and growth regulation via VapC RNAse activity;
and this effect is offset by co-expression of VapB with VapC
(Robson et al., 2009). In M. smegmatis, conditional
expression of VapC is bacteriostatic and not bactericidal
pointing towards a role in persistence and targets that arrest
growth.
Recently, Ramage et al. (2009) undertook a comprehensive
analysis of TA systems in M. tuberculosis, and identified 88
putative TA systems on the M. tuberculosis chromosome.
About 37% of these putative TA systems were located in
regions linked to horizontal gene transfer (HGT), adding
weight to the hypothesis that the evolutionary origin of TA
systems in mycobacteria is via HGT. Ramage et al. (2009)
were able to show that 20 of the 45 VapBC TAs in M. tuberculosis were toxic when just VapC was overexpressed
in M. smegmatis and that this toxicity was offset by coexpression of the cognate VapB antitoxins. Further, they
demonstrated RNase activity for two M. tuberculosis VapC
proteins in vitro and translation inhibition for four VapC proteins in vivo. They were able to rule out cross talk among
four related VapBC TAs and confirmed that in these cases,
each VapB antitoxin was a specific inhibitor of the cognate
VapC toxin (Ramage et al., 2009).
The structures for two VapBC TA complexes from M.
tuberculosis have been solved by Miallau et al. (2009) and
these elegantly illustrate the tight binding between VapB and
VapC and the tetramer of VapBC heterodimers in an
arrangement similar to the FitAB structure.
On the basis of their structures, Miallau et al. (2009) proposed a catalytic mechanism for M. tuberculosis VapC involving two metal ions similar to that of the endo- and
exo-nuclease FEN-1, a member of the FLAP nuclease superfamily. This is controversial as comparisons of substrate
interactions for the PIN and NYN (Nedd4-BP1, YacP nuclease) domains with the FLAP nuclease domains suggests that
PIN and NYN domains are more likely to coordinate a single
metal for catalysis (Anantharaman and Aravind, 2006).
Indeed, a crystal structure of VapC from the thermophilic
archaea P. aerophilum contains just a single Mn2þ ion in the
active site (Bunker et al., 2008).
The role of VapBC TA systems in M. tuberculosis is
largely unknown and is of significant interest to TB researchers. VapBC operons are often found associated with virulence factors, transposases and repetitive elements,
suggesting a role in the maintenance of virulence factors
(Arcus et al., 2005). It is also possible that different subsets
of TA systems are activated in response to different stressors.
This would enable the organism to adapt to many different
environmental conditions. It is also possible that the ribonuclease activity of VapC proteins is different from protein to
protein; some proteins may target a limited number of
mRNAs and therefore not significantly inhibit the bulk translation rate. Together, an array of VapBCs would enable the
organism to efficiently erase its transcriptional profile in
response to different stressors, thereby reprogramming the
proteome and promoting a rapid change in the metabolic
state of the cell during changing environmental conditions.
VapBCs from Haemophilus influenzae, Leptospira
interrogans, enterobacteria and Sulfolobus solfataricus
Non-typeable Haemophilus influenzae (NTHi) is a pathogenic bacteria that enters a bacteriostatic state in the middle
ear during an otitis media infection (Daines et al., 2007).
The NTHi genome has two vapBC operons, VapBC-1 and
VapBC-2. VapBC-1 has been characterised by Daines et al.
(2007). The vapBC-1 promoter is more active during the lag
phase of growth and expression of the vapBC-1 operon displays an inverse relationship to culture density (Daines et al.,
2007). VapC-1 is a ribonuclease that cleaves free RNA
in vitro independently of the ribosome and does not degrade
single-stranded or double-stranded DNA (Daines et al.,
2007).
vapBC TA loci have also been investigated in the pathogen
Leptospira interrogans. Expression of VapC from L. interrogans in E. coli resulted in inhibition of growth that is counteracted by expression of its cognate antitoxin VapB. The
presence of vapBC on an unstable plasmid prevents plasmid
loss indicating that the VapBC proteins act as a typical
selfish TA system in the context of plasmids. VapB binds to
two 9 bp inverted repeat sequences in its promoter to autoregulate its transcription (Zhang et al., 2004). However, the
cellular targets of VapBC from L. interrogans are unknown.
In a similar vein the vapBC loci from Salmonella LT2
(vapBCLT2) and the Shigella plasmid pMYSH6000
(vapBCpMYSH) have also been confirmed as bona fide TA loci
(Winther and Gerdes, 2009). However, Winther and Gerdes
(2009) show that VapC induces mRNA cleavage at stop
codons, but this is dependent on the yefM-yoeB TA locus.
Induction of the Salmonella or Shigella vapC in E. coli
resulted in degradation of ompA, dksA and tmRNA mRNA,
39
V.L.Arcus et al.
by cleaving at stop codons between the second and third
codon bases via activation of YoeB.
The Sulfolobus solfataricus genome has at least 22 vapBC
loci (Cooper et al., 2009). Thermal stress triggered high transcription levels of the vapBC-22 operon. A null-mutation in
the VapC-22 toxin resulted in no obvious phenotype but
100 ORFs were differentially transcribed when the deletion
mutant was compared with the wild type strain (Cooper
et al., 2009). It was hypothesised that disruption of VapBC
loci leads to an increased susceptibility to thermal stress.
Conclusions
Since their discovery 25 years ago (Jaffe et al., 1985; Gerdes
et al., 1986), the TA operons have been the subject of a great
deal of research effort, and several possible biological roles
for these ubiquitous prokaryotic genes have been postulated
(Hayes, 2003; Gerdes et al., 2005; Kolodkin-Gal et al., 2007;
Szekeres et al., 2007). Indeed, the proposed biological roles
range from selfish genetic elements to programmed cell
death, to roles in growth regulation, persistence and antibiotic
resistance (Magnuson, 2007; Van Melderen and Saavedra De
Bast, 2009). The vapBC operons are by far the largest family
of TAs, and contain a PIN-domain ribonuclease as the toxic
component. Their expansion in number in the important
human pathogen M. tuberculosis and the nitrogen-fixing rhizobium S. meliloti are of great interest to human health and
plant physiology, respectively. Unravelling the biological
roles for the arrays of VapBC TA operons in the genomes of
diverse and important microbes is challenging, but if brought
to fruition, will have significance for microbiology from
pathogens to plant symbiosis.
Acknowledgements
We thank Kenn Gerdes for hosting J.L.M. as a visiting student in his laboratory, and Valerie Mizrahi for helpful discussions.
Funding
We gratefully acknowledge funding from the Health
Research Council of New Zealand that has supported
research in the Arcus and Cook laboratories. Funding for
J.L.M. and J.R. is acknowledged through a University of
Waikato PhD scholarship and a Tertiary Education
Commission Bright Futures PhD scholarship, respectively.
References
Anantharaman,V. and Aravind,L. (2003) Genome Biol., 4, R81.
Anantharaman,V. and Aravind,L. (2006) RNA Biol., 3, 18– 27.
Andersson,J.O. and Andersson,S.G.E. (1999) Curr. Opin. Genet. Dev., 9,
664–671.
Andreeva,A., Howorth,D., Chandonia,J.-M., Brenner,S.E., Hubbard,T.J.P.,
Chothia,C. and Murzin,A.G. (2008) Nucleic Acids Res., 36, D419– 425.
Arcus,V.L., Backbro,K., Roos,A., Daniel,E.L. and Baker,E.N. (2004) J. Biol.
Chem., 279, 16471–14478.
Arcus,V.L., Rainey,P.B. and Turner,S.J. (2005) Trends Microbiol., 13,
360–365.
Bodogai,M., Ferenczi,S., Bashtovyy,D., Miclea,P., Papp,P. and Dusha,I.
(2006) Mol. Plant-Microbe Interact., 19, 811–822.
Bunker,R.D., McKenzie,J.L., Baker,E.N. and Arcus,V.L. (2008) Proteins,
72, 510–518.
Caminero,JA., Sotgiu,G., Zumla,A. and Migliori,G.B. (2010) The Lancet
Infect. Dis., 10, 621– 9.
Clissold,P.M. and Ponting,C.P. (2000) Curr. Biol., 10, R888– R890.
40
Cooper,T.F. and Heinemann,J.A. (2000) Proc. Natl. Acad. Sci. USA, 97,
12643–12648.
Cooper,C.R., Daugherty,A.J., Tachdijan,S., Blum,P.H. and Kelly,R.M.
(2009) Biochem. Soc. Trans., 37, 123–126.
Daines,D.A., Wu,M.H. and Yuan,S.Y. (2007) J. Bact., 189, 5041– 5048.
Finn,R.D., Mistry,J., Tate,J., et al. (2010) Nucleic Acids Res., 38,
D211– 222.
Gerdes,K. (2000) J. Bact., 182, 561–572.
Gerdes,K., Rasmussen,P.B. and Molin,S. (1986) Proc. Natl. Aca. Sci. USA,
83, 3116–3120.
Gerdes,K., Christensen,S.K. and Lobner-Olesen,A. (2005) Nat. Revs
Microbiol., 3, 371–382.
Gupta,A. (2009) FEMS Microbiol. Lett., 290, 45–53.
Hartley,R.W. (1988) J. Mol. Biol., 202, 913–915.
Hayes,F. (2003) Science, 301, 1496–1499.
Hazan,R., Sat,B. and Engelberg-Kulka,H. (2004) J. Bact., 186, 3663–3669.
Hopper,S., Wilbur,J.S., Vasquez,B.L., Larson,J., Clary,S., Mehr,I.J.,
Seifert,H.S. and So,M. (2000) Infect. Immun., 68, 896– 905.
Jaffe,A., Ogura,T. and Hiraga,S. (1985) J. Bact., 163, 841– 849.
Kaneko,T., Nakajima,N., Okamoto,S., et al. (2007) DNA Res., 14, 247– 256.
Kolodkin-Gal,I., Hazan,R., Gaathon,A., Carmeli,S. and Engelberg-Kulka,H.
(2007) Science, 318, 652–655.
Lamanna,A.C. and Karbstein,K. (2009) Proc. Natl. Acad. Sci. USA, 106,
14259–14264.
Lewis,K. (2007) Nat. Rev. Microbiol., 5, 48– 56.
Magnuson,R.D. (2007) J. Bact., 189, 6089–6092.
Makarova,K.S., Wolf,Y.I. and Koonin,E.V. (2009) Biol. Direct, 4,19.
Marcotte,E.M., Pellegrini,M., Thompson,M.J., Yeates,T.O. and Eisenberg,D.
(1999) Nature, 402, 83–6.
Mattison,K., Wilbur,J.S., So,M. and Brennan,R.G. (2006) J. Biol. Chem.,
281, 37942– 37951.
Miallau,L., Faller,M., Chiang,J., Arbing,M., Guo,F., Cascio,D. and
Eisenberg,D. (2009) J. Biol. Chem., 284, 276–283.
Nariya,H. and Inouye,M. (2008) Cell, 132, 55– 66.
Olah,B., Kiss,E., Gyorgypal,Z., Borzi,J., Cinege,G., Casnadi,G., Batut,J.,
Kondorosi,A. and Dusha,I. (2001) Mol. Plant-Microbe Interact., 14,
887–894.
Pandey,D.P. and Gerdes,K. (2005) Nucleic Acids Res., 33, 966–976.
Puskas,L.G., Nagy,Z.B., Keleman,J.Z., Ruberg,S., Bodogai,M., Becker,A.
and Dusha,I. (2004) Mol. Gen. Genomics, 272, 275– 289.
Ramage,H.R., Connolly,L.E. and Cox,J.S. (2009) PLoS Genet., 5, 1 –14.
Robson,J., McKenzie,J.L., Cursons,R., Cook,G.M. and Arcus,V.L. (2009)
J. Mol. Biol., 390, 353– 367.
Sat,B., Hazan,R., Fisher,T., Khaner,H., Glaser,G. and Engelberg-Kulka,H.
(2001) J. Bact., 183, 2041–2045.
Schumacher,M.A., Piro,K.M., Xu,W., Hansen,S., Lewis,K. and
Brennan,R.G. (2009) Science, 323, 396– 401.
Schuster-Böckler,B., Schultz,J. and Rahmann,S. (2004) BMC Bioinfo., 5, 7.
Sevin,E.W. and Barloy-Hubler,F. (2007) Genome Biol., 8,R155.
Szekeres,S., Dauti,M., Wilde,C., Mazel,D. and Rowe-Magnus,D.A. (2007)
Mol. Micro., 63, 1588– 1605.
Takeshita,D., Zenno,S., Lee,W.C., Saigo,K. and Tanokura,M. (2007)
Proteins, 68, 980– 989.
Tsilibaris,V., Maenhaut-Michel,G., Mine,N. and Van Meldern,L. (2007)
J. Bact., 189, 6101– 6108.
Van Melderen,L. and Saaverdra De Bast,M. (2009) PLos Genet., 5, 1 –6.
Wilbur,J.S., Chivers,P.T., Mattison,K., Potter,L., Brennan,R.G. and So,M.
(2005) Biochemistry, 44, 12515– 12524.
Winther,K.S. and Gerdes,K. (2009) Mol. Microbiol., 72, 918–930.
Zhang,Y., Zhang,J., Hoeflich,K.P., Ikura,M., Qing,G. and Inouye,M. (2003)
Mol. Cell, 12, 913– 923.
Zhang,Y.X., Guo,X.K., Wu,C., Bi,B., Ren,S.X., Wu,C.F. and Guo Ping,Z.
(2004) Cell Res., 14, 208– 216.
Zhu,L., Phadtare,S., Nariya,H., Ouyang,M., Husson,R.N. and Inouye,M.
(2008) Mol. Microbiol., 69, 559– 569.