Microbes at their best: first Mol Micro Meeting Würzburg

Authors


E-mail alexander.boehm@uni-wuerzburg.de; joerg.vogel@uni-wuerzburg.de; Tel. (+49) 931 3182575; Fax (+49) 931 3182578.

Summary

Founded on ground-breaking discoveries such as the operon model by Jacob and Monod more than 50 years ago, molecular microbiology is now one of the most vibrant disciplines of the life sciences. The first Mol Micro Meeting Würzburg (‘M3W’) hosted more than 160 scientists from 14 countries to exchange their latest ideas in this field of research. Divided into the four main sessions Gene Regulation, Pathogenesis, Microbial Cell Biology and Signalling, the conference provided insight into current advances and future goals and challenges.

Introduction

The conference took place at the Institute for Molecular Infection Biology of the Julius-Maximilians-University Würzburg, Germany in early May 2011. The concept of the meeting is to bring together scientists working in different areas of microbiology, but all living up to a certain style of research with a strong focus on deciphering molecular mechanisms underlying basic physiological processes. After an opening keynote lecture, work related to 16 selected papers, all published between September 2009 and August 2010 in Molecular Microbiology, was presented by the senior authors of these papers in 25 min talks. Young researchers were given exposure through seven short talks selected from submitted abstracts as well as the well-received evening poster session. The poster prize, which was sponsored by Molecular Microbiology, went to Jessica Hankins from Stephen Trent's lab at the University of Texas at Austin. Her presentation on lipopolysaccharide (LPS) modification in Vibrio cholerae was selected from almost 80 posters. The second M3W meeting, which will celebrate the 25th anniversary of Molecular Microbiology, will take place at the same venue 25–27 April, 2012 (see http://www.m-3-w.de for information).

Keynote lecture

The keynote speaker was John D. Helmann, chair of the department of microbiology at Cornell University, Ithaca (USA) and Editor-in-chief of Molecular Microbiology. In his latter position, he began with an historical overview of the journal and a review of recent and planned changes in journal scope and content. He then presented an excellent example of how transcriptomics and molecular genetic approaches have been used to reveal adaptational responses to cell envelope stress in Bacillus subtilis. At least four of the seven known extracytoplasmic function (ECF) σ factors orchestrate transcriptional reprogramming when B. subtilis is challenged with substances affecting cell envelope integrity or biogenesis (Jordan et al., 2008). Many of the genes induced by the ECF σ factors encode defence factors mediating stress tolerance. While tolerance against certain antimicrobial compounds depends on specific ECF σ factors, other compounds appear to activate several ECF σ pathways simultaneously. For example, resistance is conferred by σX to cationic antimicrobial peptides, σW to the peptidoglycan synthesis inhibitor fosfomycin, σM to the transglycosylase inhibitor moenomycin, and σV to lysozyme. In other cases, multiple ECF σ factors appear to play redundant roles in antibiotic resistance (Luo et al., 2010). This is due, in part, to overlapping promoter recognition by different σ factors, which leads to cross protection against certain cell envelope stressors.

An interesting member of the σW regulon is the fabHa-fabF operon which encodes for one of two variants of the initiating enzyme (FabHa) and the sole elongation enzyme (FabF) for fatty acid biosynthesis. In response to detergent stress, σW binds to a promoter within the coding region of the proximal fabHa gene and thereby causes repression of fabHa and induction of fabF. As a result, fatty acid biosynthesis initiates more frequently from the action of the other FabH paralogue (FabHb) which increases the proportion of straight rather than branched acyl chains. In addition, upregulation of FabF leads to an increased chain length. Together, these two changes in fatty acid composition decrease membrane fluidity which thereby counters the fluidity stress caused by detergents (Kingston et al., 2011). In a conceptually similar study, the key role of σM in mediating cephalosporin resistance was dissected genetically and found to involve the recently discovered secondary messenger cyclic-di-AMP: The cyclic-di-AMP synthase DisA is a member of the σM regulon and mutations in yybT, encoding a cyclic-di-AMP-specific phosphodiesterase (Rao et al., 2010), can restore cephalosporin resistance to sigM mutants.

Gene regulation

Following John Helmann's introduction to the world of ECF σ factors, Thorsten Mascher from LMU Munich (Germany) presented a combination of bioinformatics analysis and molecular biology to show that these alternative sigma factors represent one of the most prevalent signalling and gene regulatory pathways in the bacterial kingdom. A comprehensive biocomputational search revealed thousands of ECF σ factors from 369 available microbial genome sequences. More than half of these can be assigned to one of 43 distinct phylogenetic classes (Staron et al., 2009). One of these classes, ECF41, has roughly 300 representatives from 10 different bacterial phyla. Members of ECF41 might be involved in responses to oxidative stress and seem to be controlled by a novel regulatory mechanism. In contrast to canonical ECF σ factors that are kept inactive by interaction with a separately encoded anti-σ factor, members of ECF41 appear to harbour an anti-σ domain at the C-terminus of the same polypeptide. How input signals might unleash ECF41 activity remains to be seen.

Kelly Hughes from the University of Fribourg (Switzerland) described how Salmonella enterica assembles its flagellum. This fascinating nanomachine consists of three major parts, the rotary motor, a propeller and a flexible joint, called the ‘hook’, that connects motor and propeller. Together the three substructures are built from more than 25 different proteins (with between a few and several thousand copies of each) that have to self-assemble in a coordinated fashion. Biogenesis of the extracytoplasmic parts of the flagellum is driven by export of the different subunits through a flagellum-specific secretion apparatus that is related to type III secretion systems. One of the major questions about flagellar biogenesis is how the length of the hook is controlled. Several models have been proposed in the past that involve molecular cups, clocks and rulers (Chevance and Hughes, 2008). In his presentation Hughes proposed a refined model for hook length control. According to his ‘infrequent molecular ruler’ model, the hook length control protein FliK and the hook structural protein FlgE are alternately secreted through the hollow structure of the flagellar type III secretion apparatus. Only when the accurate hook length is reached (when the correct numbers of hook subunits are assembled) can the N- and C-termini of the ruler protein FliK interact for an extended period of time with the distal tip of the growing hook and the FlhB protease at the proximal substructure of the type III secretion apparatus respectively. This interaction triggers the so-called secretion-substrate specificity switch, which prevents further secretion of FlgE and thus terminates hook growth (Erhardt et al., 2010; 2011) (Fig. 1).

Figure 1.

The ‘infrequent ruler model’ for flagellar hook length control. From left to right: While the hook is growing due to the secretion and assembly of FlgE subunits, hook length is probed by intermittent secretion of the ruler protein FliK. In too short hooks, FliK rapidly leaves the secretion channel of the nascent hook, possibly because folding of the FliK N-terminus ‘pulls’ FliK out of the channel. When the hook has reached its target length of roughly 55 nm FliK secretion is delayed, allowing for an extended interaction time of the FliK C-terminus with the FlhB protease. This interaction triggers the secretion-specificity switch, which in turn prevents secretion of further hook subunits and allows secretion of proteins needed for the biogenesis of the hook-distal flagellar structures, i.e. the filament and cap. OM, outer membrane; PG, peptidoglycan; IM, inner membrane. Figure provided by Marc Erhardt and Kelly Hughes.

Charles Dorman from Trinity College Dublin (Ireland) introduced the audience to the differences in transcriptional regulation due to changes in DNA topology between the two related enterobacterial species Escherichia coli and S. enterica. While in both organisms many transcription factors that mediate gene expression function similarly, profound differences exist at the level of global regulation mediated by DNA supercoiling and the nucleoid-associated proteins FIS and H-NS (Dorman, 2010). An example is the FIS-regulated ssrA gene, a virulence master regulator of Salmonella Pathogenicity Island 2. The expression of ssrA is induced by the DNA relaxation that accompanies a reduction in metabolic flux (O Croinin et al., 2006). This effect can be mimicked experimentally by inhibiting some of the ATP-dependent DNA supercoiling activity of DNA gyrase with low concentrations of the drug novobiocin or by modulating the osmolarity of the growth medium. Control by FIS contributes to the influences of novobiocin and osmolarity on DNA topology (and hence transcription). Intriguingly, when transferred to E. coli, the Salmonella ssrA promoter displays a sophisticated, partially FIS-dependent regulation by novobiocin or salt treatment, which however differs fundamentally from the pattern observed in Salmonella. This has consequences for horizontally acquired genes that may display strongly deviant expression profiles even in closely related species such as S. enterica and E. coli (Cameron et al., 2011). This work shows that fine-tuning of gene regulatory circuits is likely to be a prerequisite for successful regulatory integration of horizontally acquired genes.

Kenn Gerdes from Newcastle University (UK) presented exciting data on the role of toxin–antitoxin (TA) systems in bacterial persistence. It has been known for almost 70 years that it is often not possible to sterilize a bacterial culture with antibiotics. This is because many antibiotics require active growth to be effective, and a small fraction of dormant, so-called persister cells, will often survive antibiotic treatment. The switch to the dormant state occurs stochastically, and toxins from TA systems that target translation have been suspected to play a role in persister formation. This idea is based on the finding that toxin overproduction or partial inactivation of antitoxins can increase persister frequencies (Moyed and Bertrand, 1983). However, deletion of individual TA loci in E. coli, fails to impair persister formation (Lewis, 2010). Gerdes now showed that this is based on functional redundancy of the many E. coli TA systems. Successive deletion of up to 10 TA loci gradually decreased persister formation frequencies, with a ‘Δ10TA’ strain experiencing a two order of magnitude lower propensity to form persisters (Maisonneuve et al., 2011). Interestingly, all 11 E. coli antitoxins that in a normal cell keep their cognate toxins inactive are a substrate for the Lon protease. Thus, all antitoxins are predicted to display higher steady-state levels in a lon mutant and the probability for individual cells to switch into the persister mode should be decreased. Indeed, a Δlon strain shows an up to 400-fold reduced level of persister cells.

Eduardo Groisman from Yale University, New Haven (USA), presented his model on how the leader of an mRNA can sense two different metabolic signals to control gene expression (Park et al., 2010). The 264-nucleotide-long leader of the S. enterica Mg2+ transporter gene mgtA has been known to constitute a Mg2+-responsive riboswitch controlling transcription of the mgtA coding region via transcription attenuation (Cromie et al., 2006). New data show that this mRNA leader sequence also encodes a short proline-rich peptide. According to the presented model, the mgtA leader has the ability to form two alternative secondary structures. One of these structures prevents transcription elongation into the mgtA open reading frame when both Mg2+ and proline are high. But when either Mg2+ or proline (or both) fall below a threshold concentration, an alternative stem loop is formed that favours transcription elongation into the mgtA coding region. While magnesium is sensed by direct binding to the riboswitch structure, proline is sensed indirectly via translation of the 18 amino acids short proline-rich peptide that is encoded in the mgtA leader region. Proline shortage results in ribosomal pausing at the proline codons, which in turn allows formation of the productive (elongation permissive) stem loop. In contrast, complete translation of the leader peptide under high proline conditions prevents formation of this transcriptional elongation competent secondary structure and shuts off transcription of mgtA. The biological role of the combined Mg2+ and proline sensing converging at the mgtA leader is yet to be understood.

Transcription initiation of mgtA depends on the response regulator PhoP which is relevant for the response to low magnesium, low pH or challenges with anti-microbial peptides (Kato and Groisman, 2008). In a selected short talk Maude Guillier (IBPC, Paris) showed that E. coli phoP is controlled at the post-transcriptional level by two trans-encoded small RNAs, MicA and GcvB. MicA represses phoP in stationary phase or upon outer membrane stress (Coornaert et al., 2010), while GcvB acts as a second repressor of phoP, though under different conditions.

Closing the gene regulation session, Yvonne Göpel from the laboratory of Boris Görke, University of Göttingen, Germany, elaborated on the complex transcriptional control of GlmY and GlmZ, two conserved trans-acting sRNAs of enteric bacteria. Despite their conservation at the RNA level, transcriptional control of these regulators has evolved differentially in various enterobacteria, suggesting diversified activities of σ70 and σ54, as well as a contribution of the GlrR/GlrK two-component system and the globally acting IHF protein (Gopel et al., 2011).

Pathogenesis

The first speaker of the pathogenesis session, Anna Arnqvist from Umeå University (Sweden) presented her work on Helicobacter pylori adhesion to the gastric mucosa, as well as some interesting implications of outer membrane vesicles for induction of inflammatory responses of the host tissue (Olofsson et al., 2010). H. pylori cells employ two different adhesins, the BabA adhesin which recognizes the fucosylated ABO blood group antigens in healthy mucosa and the SabA adhesin which recognizes the sialyl-Lewis a/x antigen in the inflamed mucosa. Since irreversible adherence to inflamed tissue probably is disadvantageous for the pathogen, a dynamic cycling between adherence and non-adherence might represent a critical strategy for long-term host colonization. A model for the accurately balanced expression of the babA and sabA genes was presented. In addition, it was discussed how babA expression levels influence the outcome of Hpylori infections.

Peter Sebo from the Czech Academy of Science in Prague explained the molecular events required for Bordetella pertussis adenylate cyclase toxin (CyaA) translocation into host phagocytes. In contrast to the classical pertussis toxin, which interferes with host cell signalling by preventing inhibition of host adenylate cyclases, CyaA exhibits adenylate cyclase activity by itself. The unusually large (∼ 200 kDa) CyaA polypeptide is exported across the bacterial cell envelope by an associated type I secretion system and finds its phagocytic target cells by binding to αMβ2 integrins. Interestingly, two different conformers of the toxin can take different routes at this step. The first one can form a cation selective pore in the host cell membrane that leads to K+ leakage and by this contributes to host cell damage (Osickova et al., 2010). The second conformer can form a Ca2+-permeable intermediate. Ca2+-influx activates a calpain protease, which cleaves the cytoskeletal element talin and thus releases the integrin-CyaA complex and permits lateral diffusion within the membrane. Recruitment of the complex into lipid rafts is followed by translocation into the host cytoplasm where the adenylate cyclase activity of CyaA causes a rapid collapse of host cell signalling (Bumba et al., 2010).

Dan Andersson from Uppsala University, Sweden, introduced the concept of gene amplification and how this phenomenon contributes to the fixation of antibiotic resistance in bacterial populations. Many antibiotic resistance determinants confer a fitness disadvantage in the absence of the selective pressure (i.e. the antibiotic). Nevertheless, such determinants can be fixed in populations by the occurrence of secondary compensatory mutations that can suppress the deleterious effect of the antibiotic resistance determinant. Because occurrence of these compensatory mutations (typically point mutations) is predicted to be a very rare event it was unclear how they emerge with high frequencies in natural populations. Andersson showed examples, where compensation is initially based on gene amplification (polyploidy of a part of the chromosome). In contrast with the occurrence of point mutations, gene amplification with up to dozens of copies of a specific chromosomal region is a common phenomenon with frequencies in the range of 10−5 to 10−2 per cell (almost every cell has an amplification somewhere). Increased expression of an amplified gene can cause partial compensation of the disadvantage of antibiotic resistance and thus ‘buys the bacteria some time’ to acquire a genetically stable point mutation. After acquisition of such a point mutation, selection for the amplification is removed and copy number of the amplified region reverts back to one (Pranting and Andersson, 2011). This principle can not only explain how antibiotic resistance can be fixed in a population, but also might account for efficient integration of horizontally acquired genes into the molecular context of recipient cells (Lind et al., 2010).

Ulrich Vogel from the Institute for Hygiene and Microbiology of the University of Würzburg was the only local speaker presenting at this meeting. He showed an excellent example of how a combination of epidemiological and in vitro studies can inform the molecular basis of a virulence trait. It is known that different clonal complexes of Neisseria meningitidis show pronounced geographical distribution differences. The so-called ‘settler’ clonal complexes persist in defined regional clusters (e.g. in a city or county) over long periods of time, while the so-called spreaders display rapid geographic spreading. To colonize the nasopharynx N. meningitidis forms biofilm communities on the tonsillar tissues of carrier individuals. Vogel discovered that the settler type clones, but not the spreader clones, are capable of utilizing extracellular DNA as a component of their biofilm matrix. This extracellular DNA provides enhanced shear force tolerance and thus is likely to improve colonization abilities of the settler clones. The spreaders are not able to release DNA and thus are poor colonizers as they cannot form competitive biofilms in the nasopharynx. This disadvantage might be compensated for by enhanced transmission properties, which might in turn explain their virulence profile (Lappann et al., 2010).

Michael Glickman from the Sloan-Kettering Cancer Center in New York described novel findings on the functioning of the ‘Site two’ protease (S2P) Rip1 from Mycobacterium tuberculosis (Makinoshima and Glickman, 2005). The membrane spanning S2P proteases are involved in unleashing ECF σ factors by proteolytic processing of their cognate anti-σ factors, after the anti-σ has been cleaved by a ‘Site 1’ protease. Rip1 has at least three different anti-σ factor substrates (Sig K, M and L), enabling it to regulate several downstream pathways including defence against oxidative stress or the control of lipid metabolism (Sklar et al., 2010). It is not clear how S2P substrate specificity is achieved, in particular why substrates are only recognized after Site 1 cleavage. Potentially, the extracytoplasmic PDZ domain of Rip1 and the newly discovered negative regulator of the Rip1 pathway, Ppr1, are involved in the discrimination between site 1 processed anti-sigma factors and their unprocessed precursors.

In a selected short talk, Stephanie Shames from Brett Finlay's lab, University of British Columbia, Vancouver, Canada, introduced the audience to a proteomics approach aiming at the identification of host interaction partners of type III effector proteins. With the help of this screen she identified two host factors that interact with the zinc-metalloprotease NleC from enteropathogenic E. coli, which is known to interfere with the NF-κB pathway (Shames et al., 2011).

In another short talk, Melanie Blokesch from the EPFL, Lausanne, Switzerland focused on how specificity of DNA uptake is achieved during natural competence of V. cholerae. Many naturally competent bacterial species have a preference for taking up DNA from closely related species, and this is commonly achieved by the recognition of specific DNA sequences by the uptake machinery. Blokesch showed that V. cholerae uses an alternative mechanism. Competence in this pathogen is induced by the intraspecies quorum sensing system, which may serve to ensure that it is only developed in the presence of closely related bacteria, thereby increasing the chances to acquire DNA that is compatible with Vibrio's molecular context (Suckow et al., 2011).

Cell biology

The cell biology session commenced with a talk by Lotte Søgaard-Andersen from the Max Planck Institute for Terrestrial Microbiology in Marburg (Germany). She showed how the novel Ras-type G-protein, MglA, and its cognate GTPase-activating protein (GAP), MglB, orchestrate a switch in cell polarity in gliding Myxococcus xanthus. This bacterium moves on surfaces uni-directionally with the help of two simultaneously operating gliding motility systems, a type IV pili-based system located at the leading pole and the so-called ‘A-engine’, which resides at the lagging cell pole and mediates locomotion by a recently discovered mechanism (Sun et al., 2011). Cells reverse their polarity axis in a stochastic fashion, which causes a swapping of the polar localization of the proteins from both locomotion systems and consequently to a reassignment of the leading and lagging ends of the cell (Bulyha et al., 2009). MglA and MglB lie at the heart of defining polarity and mediate reversals of gliding direction. In the current model of polarity switching only MglA in its GTP-bound form is capable of forming clusters at the cell pole, while GDP-bound MglA is freely diffusible in the cytoplasm. However, if the concentration of MglA-GTP reaches a threshold, the clusters are destabilized and dispersed. In moving cells MglA-GTP forms a cluster at the leading end of the cell where it orchestrates the behaviour of motility proteins. MglB resides at the lagging cell pole where it stimulates MglA GTPase activity and by this excludes MglA from this pole. Crystallographic studies showed that, in contrast with the majority of well studied GAPs of eukaryotic Ras-like G-proteins, MglB does not contribute directly to active site formation in order to stimulate GTP hydrolysis. Instead formation of a novel trimeric MglAMglB2 complex triggers a marked repositioning of active site residues in MglA, thereby increasing GTPase activity (Miertzschke et al., 2011) (Fig. 2). Upon perception of a signal from the ‘Frz’ chemosensory machinery – which might function like a eukaryotic type guanine nucleotide exchange factor (GEF) – MglA-GTP ‘hyper-accumulates’ at the leading pole; this causes cluster instability and dispersal of MglA-GTP. Due to the GTPase-stimulating activity of MglB the concentration of MglA-GTP is lowered in the vicinity of the lagging pole, allowing formation of a new stable MglA-GTP cluster at this pole. Somehow this triggers release of MglB from the lagging pole and re-association at the old leading pole, which has now become the new lagging pole. Importantly, Ras-like G-proteins used to be regarded as specific to the eukaryotic kingdom, but this must be revised with the discovery of this system in myxobacteria (Leonardy et al., 2010).

Figure 2.

Structure of a bacterial G-protein in complex with its GAP. The crystal structure of the trimeric MglAMglB2 complex from Thermus thermophilus (PDB ID 3T1Q) in complex with the non-hydrolysable GTP analogue GppNHp (stick representation). MglA is shown in blue and the MglB protomers are shown in green and yellow respectively. The T. thermophilus proteins serve as a model for the related MglAB proteins from Myxococcus xanthus (see Miertzschke et al., 2011).

Stephen Trent from the University of Texas at Austin reported how the machinery that modifies LPS is essential for flagellar biogenesis in Campylobacter jejuni. Gram-negative bacteria possess amazing abilities to adjust the composition of their LPS by decorating it with various modifications. This protects them from challenges by cell envelope-targeting antimicrobials and immune responses (Herrera et al., 2010). A well-known example is the modification of LPS sugars with phosphoethanolamine (pEtN). Surprisingly, C. jejuni harbours a pEtN transferase (CJ0256) that not only targets the lipid A core sugars, but also modifies the flagellar rod protein FlgG (Cullen and Trent, 2010). Thus Cj0256 is a rather promiscuous protein that can transfer pEtN to a sugar moiety of LPS or an amino acid side chain of a polypeptide. Modification of FlgG occurs at a threonine residue after the protein is exported to the periplasm and this modification is required for flagellar assembly and thus motility. The targeted threonine residue is conserved among FlgG homologues of Campylobacter and Helicobacter, while homologues of CJ0256 occur in even more distantly related bacterial species, such as Pseudomonas aeruginosa. pEtN-modification of flagellar proteins might be more common than previously thought.

Zemer Gitai from Princeton University (USA) presented data on the response of P. aeruginosa to the presence of solid surfaces and introduced the new concept of bi-functional proteins that are both metabolic enzymes and cytoskeletal components. Their work on the regulation of type IV pili discovered that P. aeruginosa mounts a specific response upon detection of a solid surface, most likely by detecting a mechanical rather than a chemical property of surfaces (Cowles and Gitai, 2010). Surface association controls expression of the exopolysaccharide alginate, rhamnolipids – involved in motility and biofilm formation – and a number of virulence factors. Thus, surface associated P. aeruginosa are more virulent than planktonic cells. The question of how mechanosensation in bacteria is accomplished on the molecular level revealed a novel cytoskeletal element in a different bacterial model, Caulobacter crescentus. This element can be visualized with the help of electron-cryotomography and consists of filaments that are composed of CTP synthase (CtpS), a universally conserved protein that, as the name suggests, is involved in cytosine-triphosphate synthesis in addition to forming cytoskeletal elements (Ingerson-Mahar et al., 2010). Purified CtpS can spontaneously form filaments in vitro and interacts with the intermediate filament-like protein Crescentin in vivo to control curvature of the crescent shaped Caulobacter cells. The ability of CTP synthase to form filaments appears to be an evolutionary conserved trait, since the CtpS orthologue of E. coli also forms fibres and can even cross-complement the cytoskeletal defect of Caulobacter ctpS mutants.

Peter Graumann from the University of Freiburg (Germany) presented work on ‘structural maintenance of chromosome’ (SMC) proteins, which play an important role for chromosome condensation in many prokaryotic and eukaryotic species (Gruber, 2011). SMC proteins harbour a long coiled-coil domain that connects an ATPase ‘head’ domain and the ‘hinge’ region. The proteins form V-shaped dimers via the hinge region. It is believed that SMC proteins embrace two different DNA regions from the same chromosome to keep them in close proximity during condensation of the chromosome. Proper activity of the SMC protein from B. subtilis in vivo depends on ATPase activity of the head domains and is negatively regulated by two auxiliary factors, ScpA and ScpB (Graumann and Knust, 2009). Graumann showed that DNA binding in vitro or the ability to compact DNA in a heterologous host, unexpectedly, does not require the ability of SMC to bind ATP. Likewise, the inhibitory action of ScpA/B does not depend on ATP binding, since ATP binding mutant SMC still responds to the presence of ScpA/B. Moreover, a truncated version of SMC lacking the head domains can still bind to DNA in vitro and DNA binding can still be downregulated by ScpA/B. Tracking of single SMC-YFP molecules and FRAP analysis showed that the protein is highly dynamic and moves around the quarter positions within cells. In summary this suggests that the primary role of the ATPase headgroups is in regulation of the SMC proteins.

The cell biology session finished off with two short talks that highlighted the versatility of the model C. crescentus to address questions about cellular architecture and cell cycle control. Caulobacter undergoes a characteristic asymmetric cell division into a motile ‘swarmer’ cell and a non-motile stalked cell. The stalk is a long tubular extension of Caulobacter cells that carries an adhesive organelle at its tip which is essential for surface adherence (while the stalk itself is not). The exact role of the stalk remains enigmatic, but one idea is that it might serve to increase the surface to volume ratio of the cell under low nutrient conditions and thereby facilitate nutrient transport (Wagner and Brun, 2007). Only the stalked cell is capable of DNA replication and cell division while the swarmer cell has to differentiate into a stalked cell to be able to progress through the cell cycle.

Susan Schlimpert from Martin Thanbichler's laboratory at the Max Planck Institute for Terrestrial Microbiology in Marburg (Germany), reported on her discovery of a novel molecular mechanism underlying the synthesis of the so-called ‘cross-bands’ in the stalk of Caulobacter. The stalk has long been known to contain these cross-bands, which are structures of unclear function and identity that separate the stalk into several irregular segments. Schlimpert showed that cross-bands contribute to the intracellular organization of Caulobacter, which might have implications for the proposed role of the stalk under low nutrient conditions.

Kristina Jonas from Michael Laub's laboratory, Massachusetts Institute of Technology, Cambridge (USA) showed that DNA replication periodicity and asymmetry are governed by two separated control circuits (Jonas et al., 2011). DNA replication initiation in Caulobacter swarmer cells is prevented by the cell cycle master regulator CtrA by binding to the origin of replication. However, the intrinsic periodicity of DNA replication initiation was found to be independent of CtrA, but dependent on the activity of the replication initiation protein DnaA. Increased DnaA activity caused faster replication cycles, while lower DnaA activity caused slow replication cycles. Thus the DnaA-dependent control module appears to be responsible for replication periodicity, while the CtrA module controls replication asymmetry (Chen et al., 2011). This is consistent with the phylogenetic distribution of these two important cell cycle regulators. While CtrA appears to be confined to α-proteobacteria, which often display asymmetrical cell division, DnaA occurs in nearly all bacterial species, including species that divide symmetrically.

Signalling session

Two presentations, one given by Judith Armitage from the University of Oxford, UK, and one short talk given by Simon Ringgaard from Matthew Waldor's lab (Harvard Medical school, Boston, USA), focused on the mechanism of chemoreceptor cluster localization in the bacterial cell. Motile bacteria respond to chemical gradients with the help of chemoreceptors. These sensory proteins can form large cytoplasmic or membrane associated clusters, containing thousands of copies of individual receptor proteins and auxiliary proteins for receptor adaptation or downstream signalling, which ultimately control the switching behaviour of the flagellar motor and thereby chemotaxis. While many model organisms used to study bacterial chemotaxis possess only one type of receptor cluster, other species, such as Rhodobacter sphaeroides, contain multiple chemotaxis operons for different receptor clusters, all of which have a similar architecture yet with different sensory roles and distinctive subcellular localizations (Wadhams et al., 2003; Scott et al., 2010). Ringgaard now show that the main V. cholerae chemotaxis cluster displays a cell cycle specific polar localization, forming a cluster at the older cell pole (where the flagellum is located) which splits into two polar clusters prior to cell division. The correct subcellular positioning of the chemotaxis clusters in both Rhodobacter and Vibrio depends on a ParA-like ATPase (PpfA from R. sphaeroides and ParC from V. cholerae) that is encoded in a chemotaxis operon conserved in both species (Thompson et al., 2006). Mutation of the respective genes leads to mislocalization of receptor clusters and chemotaxis defects. In Rhodobacter, ParA appears to interact with one specific receptor protein, TlpT, which harbours a ParB-like N-terminus that is essential for cluster localization. Subcellular partitioning of the different receptor clusters is believed to minimize cross-talk between clusters dedicated two disparate functions and/or to ensure faithful propagation of receptor clusters to daughter cells during cell division (Sourjik and Armitage, 2010; Ringgaard et al., 2011).

Robert Ryan from University College Cork (Ireland) presented an example for modularity of signalling proteins in bacteria. Several bacterial species employ variants of cis-unsaturated fatty acids as cell–cell signals to regulate the expression of factors contributing to virulence and biofilm formation (Ryan and Dow, 2011). A cis-unsaturated fatty acid family was first described in the plant pathogen Xanthomonas campestris and called DSF (for diffusible signal factor); subsequently, a different variant of this diffusible signal factor, referred to as BDSF, was identified in the opportunistic human pathogen Burkholderia cenocepacia. Homologous proteins produce both these signal molecules and the presence of the DSF/BDSF signals is perceived by sensor kinases of the two-component system class. While the DSF/BDSF sensory kinases share a similar domain architecture in their cytoplasmic parts, they have different periplasmic sensory domains. Nevertheless, both sensor kinases are able to sense both DSF/BDSF; elegant domain swapping experiments showed that a chimeric protein consisting of Burkholderia's periplasmic domain and Xanthomonas' cytoplasmic part mediated downstream signalling in Xanthomonas in response to both cell–cell signalling molecules (McCarthy et al., 2010). Interestingly, although the highly similar sensory kinases control similar physiological processes (virulence and biofilm formation), signalling immediately downstream of them relies on two different molecular mechanisms. While the cognate response regulator of the Xanthomonas system harbours a HD-GYP domain, which degrades the second messenger cyclic dimeric GMP, the Burkholderia system appears to make use of a classical response regulator with a DNA binding output domain (Ryan and Dow, 2011).

Yi-Ping Wang from Peking University (China) was the last speaker at the meeting. He presented his work on transmembrane signal transduction in two-component systems. Two-component systems are among the best studied sensory and signalling proteins, yet a comprehensive model for the events that relay a periplasmic binding event across the inner membrane and trigger autophosphorylation of the sensor kinase is still lacking (Casino et al., 2010). To this end, Yi-Ping Wang solved the crystal structures of the periplasmic domains of the apo and ligand-bound forms of the succinate-sensing DctB sensor kinase from Sinorhizobium meliloti and complemented this structural approach with biochemistry and genetics experiments (Zhou et al., 2008; Nan et al., 2010). The periplasmic portion of DctB consists of a tandem repeat of a PAS-related domain. Comparison of the two crystal structures revealed profound differences on both the monomeric and the dimeric level (the functional unit of sensor kinases is the dimer). Succinate binding, which occurs only in the membrane distal PAS-like domain not only causes a tightening of the succinate binding pocket on the level of the DctB monomer, but also triggers an ∼ 20° opening of the membrane proximal domains in the DctB dimer. This inter-subunit switch is accompanied by the formation of an ∼ 50% smaller dimerization interface between the periplasmic domains in crystal structures. Thus opening and closure (or association and dissociation) of the membrane proximal PAS-like domain is proposed to be the basis of signal transduction to the kinase domain. These data are difficult to reconcile with an earlier proposed ‘piston’ model for the functioning of sensor kinases. Because the piston model is based on data derived from the study of different sensor kinases it appears possible that different sensor kinases employ different mechanisms for downstream signalling.

Closing remarks

Following positive feedback from the audience, it was quickly decided to continue this meeting in the same format, with the keynote lecture and conference dinner on the first evening, lectures and evening poster session during the second day and a half day of lectures to close the event. Many attendees not only appreciated the scientific programme, but also the accompanying social events, like the conference dinner and wine tasting in the cellars of the UNESCO world heritage ‘Würzburger Residenz’. The modern lecture hall and lobby at the Institute for Molecular Infection Biology provided ample space to host the poster session, coffee and lunch breaks as well as a barbecue party on the second evening. These new buildings are integrated together with laboratories into the historical walls of the old surgical institute where the famous German physician Rudolf Virchow taught and did his research in the mid-19th century. (Ironically, Virchow opposed the idea that bacteria would be infectious agents, because he thought this would be at odds with his own theory of cytopathology.)

The organizers are currently in the process of inviting the speakers for the second Mol Micro Meeting Würzburg (M3W2). Several participants of the first meeting, as well as members of the Molecular Microbiology editorial team, have already confirmed their participation in the upcoming conference. Thus we are looking forward to welcome you in Würzburg in 2012.

Acknowledgements

We thank all speakers for their contribution to the meeting and for providing feedback on this report. We are very grateful to Hilde Merkert and all members of the Institute for Molecular Infection Biology and the Centre for Infectious Disease Research who contributed to the organization and realization of the meeting.