Nurturing scientific mutualism: A report from the ‘Young Microbiologists Mini-Symposium on microbe signalling, organisation and pathogenesis’

Authors


*E-mail r.ryan@ucc.ie; Tel. (+353) 21 4901331; Fax (+353) 21 4275934.

Summary

In April 2009, over one hundred microbiologists, primarily early career scientists, from 17 different countries met to discuss their work, under the broad heading of ‘Microbe signalling, organization and pathogenesis’. The meeting took place at University College Cork, Ireland and was supported by the British Council, Society for General Microbiology, American Society for Microbiology, EMBO and others. The key and relatively unusual feature of this meeting was that it was specifically aimed to provide a platform for junior scientists to present their work to a broad audience. In this review, we have tried to summarize and highlight a number of particular areas covered during the meeting, including bacterial intracellular signalling and regulation; microbe–microbe communication; biogenesis; structure and transport of the bacterial cell envelope; and pathogenic versus probiotic microbe–host interactions. We draw attention to new findings, highlight unanswered questions and reveal the anticipated future directions of a variety of areas, as described in both oral and poster presentations. Overall, this meeting provided high-quality science, with many intriguing findings being eloquently reported, in a setting that fostered interactions between diverse young and talented microbiologists.

Introduction

Microbes display an amazing range of responses to maximize their exploitation of the surrounding environment. Our understanding of microbial life has changed rapidly since the advent of molecular microbiology. As with other areas of the life sciences, the technical advances in ‘omics’, imaging, biochemistry and structural biology have had a tremendous impact on our understanding of microbial gene expression, regulation, metabolism, physiology and, importantly, pathogenicity and host interaction.

On 21–22 April 2009, over one hundred microbiologists congregated at University College Cork, Ireland, for a ‘Young Microbiologists Mini-symposium’ with the wide-ranging title ‘Microbe signalling, organisation and pathogenesis’, which was organized by Robert Ryan (University College Cork) and by Sarah Coulthurst (University of Dundee). The meeting had broad theme-based sessions discussing microbial intracellular signalling, intercellular signalling between microbes, microbial–host interactions and structure and transport across bacterial membranes. The key and relatively unusual aspect of this meeting was that it was specifically aimed to provide a platform for junior scientists to present their work and receive constructive feedback, as well as to interact with other graduate students, postdocs and young principal investigators in the field, hopefully facilitating future collaborations. Each of the four main sessions was chaired by a distinguished academic who has made considerable contribution to an area covered by the session, but the majority of the talks were given by young researchers from leading laboratories around the world. The meeting highlighted the remarkable breadth of the microbial field, and the range of approaches being utilized in addressing some important questions about microbial life. The overall outcome of the meeting, of course, was that we still have many more questions than answers!

In this Meeting Report, we emphasize some of the most significant advances and exciting developments that were presented in the talks and posters during the meeting. Due to space limitations, we can only summarize some of the highlights of the meeting and we apologize to participants whose excellent work could not be mentioned here.

Choosing a lifestyle: global regulators and environmental adaptation

Despite their deceptive simplicity, most microbes lead exceptionally complex lives. They are able to recognize changes in their surroundings and respond by rapidly altering their phenotypic properties and behaviour in ways that permit them to better survive stresses, prepare for hard times, compete with other microbes in the environment, interact successfully with eukaryotic hosts and maintain growth and cellular homeostasis. To do this they use complex regulatory networks that encompass pathways and cascades of factors that affect expression of dozens or even hundreds of genes. Several presentations from the meeting highlighted regulatory circuits and mechanisms that are used by bacteria to co-ordinate gene expression on a global level and thereby permit them to choose between alternative lifestyles that may be best suited for the conditions at hand.

A prominent theme at the meeting was the role of the nucleotide secondary messenger cyclic di-GMP in bacterial regulation (reviewed recently by Hengge, 2009; Jonas et al., 2009; Pesavento and Hengge, 2009). First identified as a regulator of bacterial cellulose synthesis, this nucleotide controls the transition between a motile or planktonic and a sessile or biofilm mode of growth in many Gram-negative species. It is produced from two GTP molecules by enzymes containing a GGDEF domain and is hydrolysed to linear di-GMP by EAL- and HD-GYP-domain containing proteins. It acts by serving as an allosteric effector of membrane-bound polysaccharide polymerases, and by binding to transcription factors and riboswitches located at the 5′-end of mRNAs.

Roger Simm, newly established at the University of Oslo, described the signalling role of cyclic di-GMP in Salmonella typhimurium, a topic that he was studying in Ute Römling's group at the Karolinska Institute. He discussed the role of cyclic di-GMP in inversely co-ordinating biofilm formation and motility, in part through its effects on expression of the transcription factor CsgD and a cyclic di-GMP-binding protein, YcgR, respectively. S. typhimurium carries genes for some 20 proteins with GGDEF and/or EAL domains. Interestingly, the EAL domain protein STM1344 was found to regulate CsgD expression through an unorthodox mechanism. This protein does not bind to or cleave cyclic di-GMP, but indirectly activates production of cyclic di-GMP in the cell by somehow controlling the expression of other cyclic di-GMP metabolizing proteins. This project illustrates a theme of broad interest that was well represented at this meeting, namely the evolution of GGDEF and EAL domains to assume novel mechanisms of action beyond cyclic di-GMP synthesis and turnover.

Christina Pesavento, a graduate student in Regine Hengge's group (Freie Universität Berlin, Germany), presented fascinating studies on the inverse co-ordination of motility and curli-fimbriae expression in Escherichia coli during transition from post-exponential phase to stationary phase (Pesavento et al., 2008). Motility and curli production are each regulated by multitier cascades, which are governed in a top-down fashion by the DNA-binding protein FlhDC and the stationary phase sigma factor σs respectively (Fig. 1). Furthermore, a number of components of each cascade exert negative influences on the opposite system, apparently to avoid the non-productive simultaneous presence of adhesive factors and flagellar motility. FlhDC activates cyclic di-GMP turnover via its effects on the EAL domain protein YhjH, while σs activates expression of CsgD, an activator of curli synthesis, as well as the expression of two GGDEF domain proteins that produce cyclic di-GMP, YegE and YedQ. FlhDC also activates expression of FliZ protein, which was found to antagonize the σs activity by binding directly to the −10 region of σs-controlled promoters. Apparently, this serves as a timing device to transiently inhibit σs activity in post-exponential phase and delay the expression of the curli activator mlrA until FhlDC activity and motility cease.

Figure 1.

The role of cyclic di-GMP signalling, the flagellar regulator FliZ and the Csr system in the co-ordination of motility and biofilm-associated functions in Escherichia coli. The figure summarizes the various levels of communication between the σS-controlled curli cascade and the FlhDC/σ70-controlled motility cascade. Another cyclic di-GMP control module (YciR/YdaM) acts in parallel to control curli expression, while the CsgD-dependent GGDEF protein YaiC and the σS-dependent EAL protein YoaD control cellulose synthesis at later time points in stationary phase. The direct and indirect effects of the Csr system on motility and PGA synthesis are also illustrated. GGDEF and EAL proteins are indicated by red and blue symbols respectively. Further demonstrated and potential signalling and regulatory input is indicated by black bolts. Regulatory connections requiring further characterization are indicated by dashed lines. For further details please refer to the text and to Pesavento et al. (2008). This figure was kindly provided by Christina Pesavento and Regine Hengge.

Natalia Tschowri, also a student of Regine Hengge, presented her studies on the role of YcgF as a blue light-activated antirepressor protein in E. coli. YcgF is a BLUF-EAL domain protein that has the ability to sense blue light via its FAD-containing BLUF domain (Tschowri et al., 2009). The degenerate EAL domain of YcgF was unable to hydrolyse or bind to cyclic di-GMP. Clues to the novel mechanism of this protein were found in the genetic context of the ycgF gene. The conserved location of this gene is near ycgE, encoding a MerR-family DNA-binding protein that regulates a set of genes encoding several small proteins. These proteins appear to act as connectors that affect regulation via the Rcs signal transduction system. YcgF was found to antagonize the repressive effects of YcgE on transcription by the binding of its EAL domain to the N-terminus of YcgE causing it to release bound operator DNA in the presence of blue light. Expression of both ycgE and ycgF was induced at low temperatures. The consequence of this regulatory pathway is that genes encoding the small regulatory proteins needed for capsule (colanic acid) production and antagonising curli production are turned on in the presence of blue light, which may affect biofilm formation in the aquatic environment.

The ASM lecture was presented by Tony Romeo (University of Florida) on the workings of the Csr (Rsm) system and its role in E. coli biofilm formation. Csr is a post-transcriptional global regulatory system. It is based on a small dimeric RNA-binding protein CsrA (RsmA) that activates or represses gene expression by binding to the leaders of regulated mRNAs and altering their translation and/or decay (reviewed in Babitzke and Romeo, 2007; Lapouge et al., 2008). The Csr system has antagonistic effects on opposing processes and pathways; for example, in E. coli CsrA represses biofilm formation and activates motility (Fig. 1). Its effect on motility is mediated via binding to the flhDC mRNA leader, which stabilizes this transcript and activates the flagellum cascade. Its role in biofilm formation is mediated primarily through the production of a polysaccharide, poly β-1,6-N-acetyl-D-glucosamine (PGA), which acts as a biofilm adhesin. The formation and secretion of PGA in E. coli depends on the four-gene operon pgaABCD, whose production is repressed by CsrA binding to the pgaA leader (Itoh et al., 2008). CsrA also indirectly regulates levels of cyclic di-GMP by controlling expression of cyclic di-GMP synthesising proteins, for example, YdeH and YcdT that stimulates PGA synthesis (Jonas et al., 2008). Additional Csr components include non-coding regulatory RNAs (CsrB and CsrC) that sequester and antagonize CsrA, and a specificity factor for turnover of CsrB/C RNAs, CsrD, which belongs to a growing list of GGDEF-EAL domain proteins that do not metabolize cyclic di-GMP. Expression of the regulatory RNAs requires a two-component signal transduction system, BarA/UvrY in E. coli. These Csr components are connected by complex feedback circuits that appear to be designed to serve as a homeostatic mechanism for fine control of CsrA activity, which may reflect its regulatory role in numerous metabolic pathways and physiological processes.

Two presentations further highlighted the regulatory role of RsmA (CsrA) in Pseudomonas aeruginosa. Hazel O'Connor, a graduate student from Fergal O'Gara's group (University College Cork) described global effects of RsmA and particular role as a post-transcriptional activator of the type III secretion system (T3SS) and its interplay with the MexT DNA-binding protein in accomplishing this task. The presentation on signalling given by Paul Williams (University of Nottingham) also described the diverse roles that RsmA plays in global regulation, and considered the function of RsmA in swarming motility in detail. Based on genetic observations, he introduced a novel hypothesis that RsmA disruption inhibits swarming motility by increasing cyclic di-GMP levels. In turn, cyclic di-GMP binding to two PilZ domain-containing proteins activates transcription of the non-coding RNA RsmZ, which titrates out RsmA and thereby inhibits swarming. It will be interesting to determine the molecular bases for this observed circuitry.

Charles Dorman (Trinity College Dublin, Ireland) described studies on the role and mechanisms of the nucleoid-binding protein H-NS in Shigella flexneri and other Enterobacterial species. This DNA-binding protein targets and bends A-T-rich regions of DNA, a feature that often typifies DNA that has been recently acquired by horizontal transfer (Dorman, 2009; Dorman and Kane, 2009). The global regulatory role of H-NS includes motility, production of fimbriae, stress resistance and regulation of virulence factors and type III secretion. H-NS is a dimeric protein containing two DNA binding domains on opposite sides of the protein, which it uses to form repression loops. In turn, H-NS repression can be antagonized via alteration of local DNA structure, by antagonistic DNA binding by proteins such as FIS and CspA, or by altering the supply of H-NS by transcriptional autorepression or post-transcriptional repression by the antisense RNA DsrA, which pairs with the hns message and inhibits translation. Molecular genetic examples, including VirB activation of Shigella virulence genes, were presented to illustrate H-NS antagonism. The audience was left with the intriguing notion that bacteria possess simple mechanisms by which H-NS repression can be antagonized, which must be invoked during the evolutionary adaptation of horizontally transferred genes, traits and regulatory circuits (Stoebel et al., 2008).

Socializing in microbial communities: microbe–microbe interactions and communication

Any microorganisms living in close proximity are probably interacting and communicating with each other. Microbes have been shown to communicate and co-ordinate their behaviour through the use of diffusible signalling molecules that spread through the extracellular environment or through physical interaction between neighbouring cells (see recent reviews Waters and Bassler, 2005; von Bodman et al., 2008; Shank and Kolter, 2009). A number of presentations at this meeting described work aimed at understanding these interactions, in what appears to be an area of ever-increasing complexity. Many microbes use intercellular signalling mediated by diffusible signal molecules to monitor their population density or confinement within niches, and to modulate their behaviour in response to these aspects of their environment, a process often termed quorum sensing.

Pseudomonas aeruginosa quorum sensing was the main focus of the presentation by Paul Williams (University of Nottingham, UK). He focused his presentation on how P. aeruginosa integrates two very distinct quorum-sensing signalling systems, 2-alkyl-4-quinolones (AQs) and N-acyl-homoserine lactones (AHLs) (for an excellent recent review of this system see Williams and Camara, 2009). The core of the P. aeruginosa quorum-sensing system consists of two AHL-based quorum-sensing systems. The las system includes LasI, the AHL synthase that generates the signal molecule N-(3-oxododecanoyl)-L-homoserine lactone (3-oxo-C12-HSL) and the transcriptional regulator, LasR. The rhl system consists of RhlI, another AHL synthase, which generates the signal molecule N-butanoyl-L-homoserine lactone (C4-HSL) and the transcriptional regulator RhlR. The las and rhl systems directly or indirectly regulate over 10% of the P. aeruginosa genome, including genes involved in virulence and biofilm development. In addition to 3-oxo-C12-HSL and C4-HSL, P. aeruginosa produces a third signalling molecule, 2-heptyl-3-hydroxy-4(1H)-quinolone, the pseudomonas quinolone signal (PQS). Structural genes for quinolone production have been identified (phnAB, pqsABCD and pqsH), together with a transcriptional regulator (pqsR or mvfR) and a response effector (pqsE). This system also contributes significantly to P. aeruginosa virulence.

The production of las- and rhl-regulated virulence determinants, including elastase, lectin A, pyocyanin and biofilm, is advanced and enhanced by the exogenous addition of the PQS to the growth medium. These observations suggest that these cell–cell signalling systems are connected by multiple regulatory pathways, likely to operate at both transcriptional and post-transcriptional levels. Paul Williams explained how it emerged that the las system sits at the top of a hierarchy, as it positively regulates both the rhl quorum-sensing system and the AQ signalling system. The rhl and AQ systems are also interconnected, in that C4-HSL negatively affects PQS production, while PQS positively regulates rhlR. Conversely, RhlR negatively regulates both pqsR and pqsA. However, mutational inactivation of las generally results in a delay rather than the complete loss of rhl-dependent and AQ-dependent quorum sensing (Diggle et al., 2003; Duan and Surette, 2007; Dekimpe and Deziel, 2009). Nevertheless, the las/rhl and AQ systems are intimately linked, and mutational inactivation of either system attenuates P. aeruginosa virulence. This work demonstrates that the quorum-sensing circuitry in P. aeruginosa is much more complex than simple las/rhl hierarchy. Understanding the molecular intricacies of such systems in P. aeruginosa and other bacteria, and their integration into global regulatory networks that respond to diverse environmental signals is a major challenge for the postgenomic era.

Yvonne McCarthy, from the Dow Group (University College Cork, Ireland), reported that Burkholderia cenocepacia produces a diffusible fatty acid signal molecule, cis-2-dodecenoic acid (BDSF) that has been implicated in interspecies and inter-kingdom communication (Boon et al., 2008). She showed that BDSF also acts as an intraspecies signal in the major opportunistic pathogen B. cenocepacia to control factors contributing to virulence. The role of BDSF in regulation in B. cenocepacia was initially investigated by comparative proteomic studies of the wild type and rpfF mutant. Mutation of rpfF reduced the abundance of several proteins. These proteins included the chaperone DnaK, CblD, which is implicated in cable pilus production, and FliJ, a soluble component of the type III flagellar protein export system. The results from this proteomic analysis were confirmed by quantitative RT-PCR. The effect of mutation of rpfF on the abundance of CblD and FliJ suggested that BDSF signalling was involved in regulating of bacterial motility and biofilm formation (Ryan et al., 2009). The rpfF mutation showed reduced motility and reduced adherence to porcine mucin. The work also demonstrated that BDSF contributes to virulence, as an rpfF mutant was attenuated in a Galleria mellonella larval infection model. Overall, these findings suggest as well as its involvement in inter-kingdom signalling BDSF functions as an intraspecies signal for B. cenocepacia.

Neil Williamson, from the Salmond Group (University of Cambridge, UK), explained that Serratia sp. ATCC 39006 has a complex regulatory hierarchy that controls antibiotic production, swarming and virulence, and involves a SmaIR-dependent AHL-based quorum-sensing system; a GGDEF/EAL domain protein; the GacA/Rsm system, Rap (regulator of antibiotic and pigment production); and a master regulator, PigP. The PigP and quorum-sensing regulons overlap with both systems modulating the expression of some common targets (Fineran et al., 2005; Williamson et al., 2006). Swarming in Serratia sp. 39006 requires a RhlA-dependent biosurfactant, which also acts to disperse the antimicrobial red pigment, prodigiosin. There is significant overlap in the regulation of prodigiosin biosynthesis and swarming, suggesting a synergistic function for the two compounds, possibly facilitating surface colonization by Serratia (Williamson et al., 2008). The master regulator PigP regulates transcription of the prodigiosin biosynthesis gene cluster by modulating the expression of at least six other regulators (Fineran et al., 2005). However, it is uncertain if this regulation is direct or indirect.

Tamzin Gristwood, also from the Salmond group, reported on another regulator of secondary metabolite production in Serratia sp. 39006, PigS, which shares homology with the ArsR/SmtB of family of metalloregulatory transcriptional repressors. The pigS gene is the first gene of a predicted operon encoding three putative membrane proteins (pmp123), which may represent a novel transport system. A mutation in pigS resulted in decreased production of prodigiosin. PigS autoregulates its own production via direct binding to the pigS promoter, and also regulates transcription of the divergently transcribed gene, blhA, encoding a putative metallo-beta-lactamase. Gristwood and colleagues also demonstrated that PigP activates transcription of both pigS and blhA. Although the genes in the pigS locus do not encode proteins predicted to be involved in metal resistance, the genomic context of pigS does share significant similarity with that of the recently described bigR locus in the plant pathogen Xylella fastidiosa 9a5c. Therefore, PigS appears to fall into the newly proposed BigR subfamily of ArsR/SmtB proteins. This work illustrates the complexity of regulatory inputs into the co-ordinated swarming, biosurfactant and prodigiosin phenotypes of Serratia species.

Antibiotics are probably the best-known bacterial secondary metabolites. An intriguing idea is that antibiotics are not solely bacterial weapons but, at sub-inhibitory concentrations, act as signalling molecules that regulate the homeostasis of microbial communities. Sub-inhibitory antibiotic concentrations increase expression of genes encoding factors that influence interaction with host cells, induce biofilm formation, activate the SOS-response and alter virulence factor production (see recent reviews by Fajardo and Martinez, 2008; Shank and Kolter, 2009). Many recent papers exploring this hypothesis have used semi-synthetic or synthetic antibiotics, making it difficult to argue that the observed effects are representative of natural microbial interactions. Jerry Reen, from the O'Gara group at University College Cork, Ireland, used a combination of genomic and functional assays to demonstrate that sub-inhibitory concentrations of the natural antibiotic colistin (polymyxin E), a cationic antimicrobial peptide produced by certain strains of Bacillus polymyxa var. colistinus, trigger expression of genes influencing the virulence of the major opportunistic bacterial pathogen P. aeruginosa, for example, PQS biosynthesis genes and the phenazine biosynthesis operon. This induction of one of the central components of the quorum-sensing network might represent a switch to a more robust population, with increased fitness in the competitive environment of the cystic fibrosis lung.

The bulk of research in this field has focused on microbial signalling mediated by diffusible signal molecules. However, there is a growing appreciation of the role of contact-dependent signalling. Angela Nobbs, from the Jenkinson group (University of Bristol, UK), highlighted this when she reported on the interactions between Streptococcus gordonii and Candida albicans in oral cavity colonization. Streptococci are among the first microorganisms to bind to saliva-coated surfaces in the oral cavity. Their deposition provides a layer onto which further microbes can attach and accumulate, so paving the way for the development of complex polymicrobial communities (Jenkinson and Lamont, 2005). These might then serve as a source of infection for diseases of oral tissues and at other sites throughout the body following systemic dissemination (Okuda et al., 2004). C. albicans is a pleiomorphic fungus that can be found as a commensal in the oral cavity in 20–60% of the human population. As an opportunistic pathogen, however, it is associated with > 90% of oral fungal diseases and is becoming an increasingly serious problem in nosocomial infections (Perlroth et al., 2007; Samaranayake et al., 2009). Co-adhesion between S. gordonii and C. albicans has been proposed to facilitate oral carriage and persistence of C. albicans in mixed-species biofilms, but a detailed understanding of this association remains to be defined. Nobbs and colleagues found that S. gordonii adheres to all morphological forms of C. albicans, but mainly to hyphae. Candida receptors for streptococci were identified by using a number of biochemical and genetic approaches, in conjunction with fluorescence and atomic force microscopy. These receptors include hyphal-specific protein ALS3p, a member of the ALS family of C. albicans adhesins involved in biofilm formation (Hoyer et al., 2008).

Richard Silverman, also from the Jenkinson group, reported that interactions of S. gordonii with candidal protein ALS3p involve the bacterial surface proteins SspA and SspB, members of the streptococcal antigen I/II family of polypeptide adhesins. SspB on the surface of the heterologous bacterium Lactococcus lactis conferred on it the ability to adhere to C. albicans hyphae. SspB+L. lactis adhered less well to C. albicans als3D/als3D hyphae lacking ALS3p. Furthermore, the expression of ALS3p on the surface of recombinant Saccharomyces cerevisiae increased S. gordonii adherence to this yeast. These results indicate that SspB interacts with ALS3p, and might account for the higher levels of streptococci bound to ALS3p+ filaments.

Finally, Nobbs and colleagues have also demonstrated that coadhesion is affected by glycosylation defects in Candida. Building on this, a deletion mutant lacking the mannosyltransferase Mnt1, but not two other mannosyltransferases, Mnt2 or Mnn4, exhibited significantly reduced streptococcal adhesion (Fig. 2). These data imply that S. gordonii adheres to C. albicans through the recognition of specific protein receptors, possibly utilizing glycosylated epitopes, and that modulation of these receptors can strongly influence the interactions of streptococci and Candida in mixed-species communities.

Figure 2.

Interactions of Streptococcus gordonii DL1 with isogenic glycosylation mutants of Candida albicans CAI-4/CIp10. Yeast cells were grown at 37°C, 220 r.p.m. in YNB/20 mM phosphate buffer supplemented with 0.1% tryptone and 0.4% glucose for 3 h, and incubated for a further 1 h with FITC-labelled streptococci. Cells were then harvested and visualized by fluorescence and light microscopy. S. gordonii bound strongly to (A) wild-type yeast cells and (C) cells lacking mannosyltransferase Mnt2. By contrast, adhesion of S. gordonii was significantly impaired to (B) mutant cells lacking mannosyltransferase Mnt1. This figure was kindly provided by Angela Nobbs and Howard Jenkinson.

The attenuation of microbial antibiotic resistance and virulence using small-molecule inhibitors (the antipathogenic drug principle) is likely to play a role in future treatment strategies of any infections (recently reviewed in Rasmussen and Givskov, 2006; Janssens et al., 2008a). Hans Steenackers, from the De Keersmaecker group (Katholieke Universiteit Leuven, Belgium), provided exciting new insights into the treatment of biofilms of Salmonella, one of the most prevalent food-borne pathogens, using synthetic brominated furanones. He reported that these halogenated furanones, a class of molecules originally derived from the red alga Delisea pulchra (de Nys et al. 2006), has a strong inhibitory effect on Salmonella biofilm formation at non-growth-inhibiting concentrations. A microarray analysis of the gene expression profile of Salmonella in the presence of (Z)-4-bromo-5-(bromomethylene)-3-ethyl-2(5H)-furanone identified induced genes involved in metabolism, stress response and drug sensitivity. Most of the repressed genes were involved in metabolism, the T3SS and flagellar biosynthesis. Pretreatment with furanones rendered Salmonella biofilms more susceptible to antibiotic treatment. This work demonstrates that particular brominated furanones might be able to prevent biofilm formation by S. typhimurium (see also Janssens et al., 2008b). Stijn Robijns, also from the De Keersmaecker group, reported a second strategy in the search for Salmonella biofilm inhibitors. They are currently screening a library of 17 000 small molecules for biofilm inhibitors that are active at temperatures ranging from 16°C to 37°C, and therefore active outside and/or inside the host.

Finally in this section, Liang Yang, from the Tolker-Nielsen group (Technical University of Denmark), described the application of a structure-based virtual screening method to search for putative quorum-sensing inhibitors from a database comprising approved drugs and natural compounds. The database was built from compounds that showed structural similarities with previously reported quorum-sensing inhibitors. In this study, the ligand of the quorum-sensing receptor LasR in P. aeruginosa, a well-studied receptor for the signal molecule, 3-oxo-C12-HSL (discussed above), was targeted. Six compounds were identified from this database using a ‘molecular docking’ approach and tested for inhibitory activity against LasR-controlled expression of a lasB::gfp reporter (Yang et al., 2009). Three compounds, salicylic acid, nifuroxazide and chlorzoxazone, showed significant inhibition of quorum sensing-regulated gene expression and related phenotypes. These results indicate that structure-based virtual screening might be an efficient tool with which to search for novel compounds that combat bacterial infections.

The bacterial cell envelope: biogenesis, structure and transport

The bacterial cell envelope has been known for many years to provide bacterial cells with the structural support, protection and selective permeability barrier required for their survival and proliferation. Three talks on the components of the cell envelope provided an interesting reminder that the simplistic view of the cell envelope as consisting of ‘membrane(s) plus cell wall’ is far from complete story. Following on from this, we were elegantly reminded that the envelope is not an impenetrable boundary, with three talks describing recent advances in understanding the molecular mechanisms by which three diverse systems move proteins across the cell envelope.

Jeff Errington, from the University of Newcastle, delivered the EMBO Lecture, describing exciting recent work on so-called ‘L-form’ bacteria (Leaver et al., 2009). L-forms are cell wall-deficient mutants of many common bacteria that have been implicated in disease and persistence due to their high resistance to cell wall-targeting antibiotics such as penicillin. However, they are hitherto proven difficult to generate and maintain, meaning that the genetic basis of the phenotype and the means by which they grow and divide remain elusive. The Errington group developed a robust method to generate viable L-forms of Bacillus subtilis strains (utilizing a strain conditionally unable to synthesize peptidoglycan, in combination with penicillin selection). This was then utilized, first to identify a single-point mutation predisposing the cells for proliferation-competent L-forms (in a gene-involved isoprenoid biosynthesis), and second to demonstrate that the ‘essential’ gene, ftsZ, is no longer required for cell division in L-forms. This very surprising finding indicated that L-forms undergo cell division in a fundamentally different manner from normal, cell wall-containing bacteria, as indeed proved to be the case. Beautiful time-lapse microscopy images showed that L-forms proliferate by several variations on an ‘extrusion and resolution’ mechanism, including the formation of long protrusions that resolve into a chain of progeny (Fig. 3), as well as ‘binary fission’-like and multiple surface ‘budding’-like events. A very interesting idea from these findings is that such ‘extrusions’, perhaps driven by active chromosome segregation, might represent an ancient proliferation mechanism, predating the cell wall itself.

Figure 3.

Time lapse microscopy of a growing and dividing Bacillus subtilis L-form. The large cell in the centre of the image undergoes a series of shape changes before forming stable protrusions that resolve to produce a progeny cell. Time interval 20 min, duration 100 min. Scale bar = 2 μm. This figure was kindly provided by Mark Leaver and Jeff Errington.

Robert Fagan, from the Fairweather group (Imperial College London, UK), described the first structural insights into the S-layer of a bacterial pathogen, namely the notorious nosocomial pathogen, Clostridium difficile. The majority of bacterial species, although not the most common model organisms, possess S-layers: 2D paracrystalline protein arrays that surround the bacterial cell and play an important role in adhesion. The C. difficile S-layer is composed of two proteins, both derived from the post-translational cleavage of SlpA, termed LMW and HMW (low- and high-molecular-weight S-layer proteins respectively). This study (Fagan et al., 2009) demonstrated that these two proteins interact ‘end-to-end’ via the C-terminus of LMW and the N-terminus of HMW to form an elongated complex, as visualized by small-angle X-ray scattering analysis. This structural model was further refined by determining the high-resolution crystal structure of a truncated LMW protein (Fig. 4). The results suggest that the LMW contains a variable surface-exposed region, presumably involved in adhesion and host interaction, together with a conserved interaction domain for interaction with HMW, which is itself predicted to mediate attachment to the cell wall. Another member of the Fairweather group, Catherine Reynolds, described the novel discovery that another C. difficile surface protein, the cell wall protein CwpV, is produced in a phase variable manner. This phase variation operates by the inversion of a region of DNA between the transcriptional and translational start sites of cwpV, and is catalysed by a C. difficile-specific recombinase. Moving onto transport of proteins across the cell envelope, Tracy Palmer from University of Dundee, UK, spoke about the Tat export system that moves folded proteins out across the cytoplasmic membrane. Although the Tat system, with just three protein components, appears simple, attempts to isolate a TatABC complex that functions efficiently in vitro or to obtain a high-resolution structure for any of the components have yet to be successful. However, recent work on TatA, a small transmembrane protein, which multimerizes to form large complexes and is believed to form the transport channel, has shed light on the mechanism of this export system. Electron microscopy (Gohlke et al., 2005) revealed that the ‘donut’-shaped TatA complexes resemble a cavity with a lid and have dimensions consistent with the presence of a channel. Moreover, the complexes formed are of variable sizes, consistent with the ability to accommodate different-sized, folded substrates while maintaining a seal across the membrane. These findings were extended using single-molecule in vivo imaging of TatA (Leake et al., 2008). This revealed that the complexes assemble with variable stoichiometries (median ∼25 TatA) from basic tetrameric units. These TatA complexes can only assemble in the presence of the other Tat components, TatBC. Thus, it appears that the Tat transport channel forms by polymerization, when and with the size required. The physiological role of Tat was also discussed, in particular the fact that Tat in Streptomyces coelicolor, very many more, varied proteins are exported by Tat than is the case in E. coli, a feature that is even more pronounced in the plant pathogen, Streptomyces scabies. Indeed, several virulence proteins secreted by the latter that apparently acquired by gene transfer from other plant pathogens use the Tat system for export, although the reason for this is not clear.

Figure 4.

On the left is a representation of the biogenesis of the Clostridium difficile S-layer. The S-layer protein precursor, SlpA, includes an N-terminal signal peptide (SP) that is cleaved (1) upon translocation across the cytoplasmic membrane. This is followed by a second cleavage event (2) separating the LMW and HMW S-layer proteins (LMW SLP, HMW SLP). These two proteins interact via their C- and N-termini, respectively, to form the H/L complex (3).The H/L complex then self-associates to form the two-dimensional paracrystalline S-layer that surrounds the entire cell. On the right is a low-resolution structure of the H/L complex obtained by small-angle X-ray scattering. Onto which is superimposed a 2.4 Å X-ray crystal structure of two domains of the LMW SLP. This figure was kindly provided by Robert Fagan and Neil Fairweather.

The autotransporters, a simple secretion system that transports polypeptide segments across the outer membrane, was highlighted by Amanda Rossiter from Ian Henderson's group at the University of Birmingham, UK. One of these factors, plasmid encoded toxin (Pet), is a prototypical member of the serine protease autotransporter family of the Enterobactericeae (SPATE). While the function of this SPATE has been well characterized, its regulation and biogenesis remain poorly understood. In the hope of unravelling this, the Henderson group employed two methods. The first, a global transposon mutagenesis screen in enteroaggregative E. coli strain 042, identified over 150 transposon mutants deficient in Pet secretion. However, given the complex nature of the autotransporter secretion mechanism, this does not identify the exact point at which the process is defective, whether it be transcription, translation, inner membrane translocation or secretion across the outer membrane. In the second approach, the minimal region of the pet promoter was fused to the gene encoding for β-galactosidase. This revealed that transcription of pet was downregulated in mutants of the global transcriptional regulators, crp and fis. In contrast, in mutants of the periplasmic chaperones surA and skp, a surprisingly high pet promoter activity relative to the parental strain was observed. This might suggest a role for the chaperones in triggering the Cpx stress response that promotes the transcription of genes encoding virulence factors to sustain its existence in the adverse conditions of the host environment.

Guy Cornelis, from the University of Basel, Switzerland, gave a fascinating description of the current state of knowledge of the mechanism of T3SSs, in particular those of the type found in Yersinia spp. T3SSs are complex, envelope-spanning protein assemblies, found in many Gram-negative bacteria that secrete effector proteins across the bacterial inner and outer membranes and form an extracellular ‘needle complex’ to inject these effectors directly into target cells. T3SSs (and other secretion systems) have been shown to play a key role in virulence and host interaction in many pathogens. Thus, it is impossible to separate the topics of ‘host-microbe interactions’ and ‘protein secretion’, as reflected in the fact that professor Cornelis chaired the ‘Host-Microbe Interactions’ session of the meeting, helping to place both topics in their shared context. A very interesting recent finding from the Cornelis group addressed the unanswered question of whether the YscP protein, known to determine the length of the Yersinia T3SS needle, acts as a molecular ‘timer’ or a molecular ‘ruler’. This work, utilizing a combination of mutational analyses, microscopy and in silico modelling, provides compelling evidence that not only is YscP a ‘ruler’, but that the helical content of YscP is critical to needle length determination (Wagner et al., 2009).

Finally, Daniel Walker (University of Glasgow, UK) described a system by which the E family of colicins, bacteriocins produced by and active against E. coli, are imported across the envelope of the target cell, a process termed retrotranslocation. Crystallographic analysis of colicin E9 demonstrated that it binds to the periplasmic protein TolB via a short epitope in the disordered N-terminus of colicin. This binding competitively disrupts the binding of Pal, a major outer membrane lipoprotein, to TolB, thereby destabilizing the outer membrane and facilitating translocation of the complete toxin across the cell envelope (Loftus et al., 2006).

Microbe–host interactions, the complete spectrum from pathogenic to probiotic

Microbes, including viruses, fungi and bacteria, interact or communicate not only with each other but also, in many cases, with higher organisms such as plants, invertebrates (nematodes and insects) and vertebrates, including humans. The outcome of these microbe–host interactions covers a spectrum, from pathogenicity through commensialism to symbiosis and health promotion.

Secreted and/or surface-located macromolecules are crucial in establishing host–microbe interactions. Strain-specific differences in the presence/absence or subtle modifications of these molecules steer the outcome of the microbe–host interaction. For instance, Listeriolysin S is a novel peptide haemolysin exclusively produced by a subset of Listeria monocytogenes lineage I strains (those responsible for the majority of outbreaks of listeriosis and for which the genetic basis was for a long time unknown) (Cotter et al., 2008). Paul Cotter (Teagasc, Moorepark Food Research Centre, Ireland) told the exciting story of the discovery of Streptolysin-S-like modified virulence peptides in several pathogenic Gram-positive bacteria, including L. monocytogenes. This finding counters the long-held dogma that Streptolysin-S is only produced by Group A Streptococcus and that Listeria only produces one cytotoxin, Listeriolysin O. Listeriolysin S, encoded by the LIPI-3 pathogenicity island and induced by oxidative stress, is a haemolytic and cytotoxic factor that contributes to virulence, as evidenced by murine- and human polymorphonuclear neutrophil-based studies.

Another toxin-secreting pathogen Staphylococcus aureus also requires multiple surface-expressed proteins in order to adhere to substrates of the extracellular matrix (such as fibronectin, elastin, fibrinogen) and plasma proteins. Several different aspects of this requirement were reported by members of the Foster group (Trinity College, Dublin) (Foster and Hook, 1998; Fitzgerald et al., 2006). Emma Smith, for instance, demonstrated that the IgG-binding protein (Sbi) is associated with the S. aureus cytoplasmic membrane, despite lacking a hydrophobic domain characteristic of integral membrane proteins. Sbi is also secreted into the culture medium. In addition to binding to IgG, β2-glycoprotein 1 and complement factor C3, the Sbi protein exhibited a dose-dependent and saturable binding to the von Willebrand factor. von Willebrand factor is a blood glycoprotein present at sites of damage to the endothelium and therefore possibly involved in the initiation of intravascular infection. However, no evidence was found to confirm that Sbi is a virulence factor in a mouse model for septic arthritis and sepsis. Thus, its biological function remains to be determined. Nevertheless, the two IgG-binding proteins (Sbi and Spa) contribute to immune evasion by S. aureus, by impeding phagocytosis and classical pathway complement fixation (Foster, 2005).

Capsular polysaccharides and exopolysaccharides (EPS) were also highlighted as being involved in bacterial immune suppression, as reported for both pathogens (Sharma and Qadri, 2004) and commensals (Mazmanian et al., 2008). Sarah Lebeer (De Keersmaecker group, K.U.Leuven, Belgium) used knock-out mutants of the clinically well-documented probiotic strain Lactobacillus rhamnosus GG to identify the strain-specific long galactose-rich EPS molecules as important modulating factors of adhesion and immune-stimulating activity (Francius et al., 2008; Lebeer et al., 2008; 2009). She demonstrated that the L. rhamnosus GG EPS molecules appear to serve a shielding function, thereby preventing underlying adhesins, such as fimbriae (Lebeer et al., 2009) or other surface-exposed molecules, from interacting with the host.

This shielding protects L. rhamnosus GG against the innate immune defence (immune evasion), thereby promoting survival in the gastrointestinal tract. On the other hand, the EPS barrier prevents binding of important bacterial ligands that induce host signalling pathways to host receptors that would normally lead to health-promoting effects (immune modulation). Therefore, the surface-exposed glycan structures are an ideal bacterial device to modulate host–microbe interactions; i.e. the production of the EPS most likely needs to be carefully balanced during gastrointestinal transit in order to obtain the anticipated health-promoting effect of L. rhamnosus GG. Identifying the L. rhamnosus GG cell-surface ligands that trigger specific host signalling pathways will be part of Lebeer's future work. The Toll-like receptors (TLR) will be important players; they are key mediators of the innate immune response because they recognize microbial-associated molecular patterns (MAMPs) of pathogens, commensals and probiotics. Indeed, Siobhan Cashman from the Morgan group (University College Cork, Ireland) reported that different intestinal epithelial cell types can express distinct TLRs.

Most attention has been focused on pathogen–host interactions. However, we live in close harmony with thousands of different microbes on our mucosal surfaces and these microbe–host interactions, which are mostly harmless, are essential for our daily life. The concept of probiotics is based on the observation that some of these microbes have health benefits. One of the possible mechanisms by which probiotics contribute to health is by excluding or inhibiting pathogens. Aisling Murphy from the Scallan group (University College Cork, Ireland) demonstrated that both L. rhamnosus GG and L. salivarius UCC118 antagonize rotavirus plaque formation on MA104 cell monolayers, with L. rhamnosus GG being more effective. Interfering activity was transferable in cell-free bacterial culture supernatant, suggesting that it might be lactic acid. Microarray analysis revealed that several host genes involved in regulation of apoptosis and immune responses were upregulated during rotavirus infection. UCC118 dampens host cell responses to infection, preventing the upregulation of many of these genes. Identifying the probiotic factors that interfere with these rotavirus-triggered host signalling pathways involved in apoptosis and immune responses will be essential in understanding the probiotic-mediated amelioration of rotavirus-induced diarrhoea (Majamaa et al., 1995).

In studying the tripartite interaction of probiotic bacteria, rotavirus and eukaryotic host cell in the above study, the development of a reproducible model system was required. During the meeting, it became clear that the availability of a representative model system is crucial in studying host–microbe interactions. Marlies Mooij (O'Gara group, University College Cork, Ireland in collaboration with the Bitter group, VU University Amsterdam, and Meijer group, Leiden University, the Netherlands) highlighted some benefits of using zebrafish embryos (Danio rerio) in the study of P. aeruginosa pathogenesis. This model allows for real-time analysis of the interactions, using fluorescent proteins such as mCherry. Bacterial mutants can be screened in a high-throughput way for a differential interaction with the innate immune system, as the embryos have macrophages, granulocytes, NK cells, a complement system, lectins, TLR and cytokines like IL-10, IFNgamma, TNFalpha and TGFbeta. The model also allows for an easy switch to adult zebrafishes to analyse the adaptive immune system. Using microarray analysis of the zebrafish embryos, Mooij demonstrated the upregulation of matrix metallo-proteases upon microinjection of P. aeruginosa, which corresponds to data obtained with cystic fibrosis patients. This clearly shows the reliability of the model system.

Several presentations were focused on the use of postgenomics-inspired technologies, including transcriptomics, proteomics and phenomics to gain insight into microbe–host interactions. Timothy Cairns of the Bignell group (Imperial College London, UK) described stage-specific gene expression analysis of both host and pathogen throughout a time-course of murine aspergillosis, aiming to identify areas of disease initiation for translation into therapeutic and diagnostic benefit in the battle against Aspergillus fumigatus. These transcriptome analyses revealed a biased distribution of host-adaptation genes in subtelomeric regions of chromosomes, which are hotbeds of genetic diversity undergoing frequent chromosomal rearrangements (McDonagh et al., 2008). This methodology was then utilized for probing the role of fungal secondary metabolites during invasive infection, by comparing transcriptional profiles throughout murine infection between the Af293 clinical isolate and a strain lacking the methyltransferase LaeA, a global transcriptional regulator of A. fumigatus secondary metabolite biosynthesis (Bok et al., 2005). This mutant is avirulent in the murine model. He found that genes downregulated in the ΔlaeA mutant, i.e. multiple secondary metabolite clusters and genes encoding secreted proteins, siderophore biosynthesis and calcium transportation, showed high levels of subtelomeric bias, suggesting that chromatin remodelling plays an important role in these areas during disease initiation.

A substantial number of the presentations dealt with Salmonella–host interactions, all reflecting aspects of the above-mentioned emerging themes, i.e. the use of high-throughput technologies, focus on secreted/surface compounds and the need for a representative model system. The food-borne pathogen Salmonella is one of the most common causes of food poisoning. During the critical pathogenic step of cell invasion, Salmonella uses many virulence factors that are encoded on Salmonella pathogenicity islands (SPI). Two studies reported on the use of proteomic approaches, first to identify novel secreted virulence factors by analysis of the secretome using type III secretion-inducing conditions and conditioned medium (Aileen Sherry, from the Roberts group, University of Glasgow, UK), and second to identify proteins important for adaptation in the host, by the use of in vivo mimicking conditions (Gwendoline Kint, K.U.Leuven, Belgium) (Sonck et al., 2009). Ellen Arena from the Finlay group (University of British Columbia, Canada) presented a new in vivo system for the study of S. typhimurium infection in epithelial cells, utilizing a meaningful infection model, the murine gallbladder (Menendez et al., 2009). She showed that murine gallbladders become heavily colonized, and that the Salmonella cells localize exclusively within the epithelium, in contrast to the intestine, where they are mainly translocated to the lamina propria (Fig. 5). The bacteria remain subnuclearly localized, colocalized with LAMP-1, in the epithelial cells, where they replicate in vacuolar compartments. This extreme intracellular infection of the gallbladder epithelium is SPI-1-dependent, as Salmonella invA invasion mutants remain extracellular in the lumen of the gallbladder. The bile-containing environment is highly supportive of the extracellular survival and replication of Salmonella.

Figure 5.

Immunostaining of a gallbladder 120 h post infection with wild-type Salmonella typhimurium SL1344 reveals that bacteria localize to the epithelial cells, but not the underlying lamina propria, from a representative orally infected mouse; bacteria are in red and cell nuclei are in blue. This figure was kindly provided by Ellen Arena and Brett Finlay.

Expression of SPI-1 is tightly regulated by transcriptional activators encoded on SPI-1 (HilC, HilD, HilA and InvF), and by additional regulators encoded outside SPI-1. One such regulator that is encoded outside of the SPI-1 locus is DNA adenine (Dam) methylation (Balbontin et al., 2006). Javier López-Garrido from the Casadesús group (University of Seville, Spain) showed, through epistasis analysis, that SPI-1 regulation by Dam depends on HilD. However, regulation of hilD by Dam occurs at the post-transcriptional level. For his insightful work López-Garrido was presented with Molecular Microbiology poster prize. Using a genetic screen of transposon insertions that suppressed the hilD-mediated expression defect of SPI-1 in Dam mutants revealed insertions in rcsB. The gene rcsB encodes the cytoplasmic response regulator of the Rcs phosphorelay an important signalling pathway that responds to envelope stress. These observations raise the possibility that a cellular factor under dual control by Dam methylation and the Rcs system might control hilD mRNA stability.

The meeting also reinforced the idea that, in many ways, pathogenicity and mutualism or even probiotic interactions are based on very similar types of interactions within the host. The study of the Photorhabdus–host interaction provides a unique opportunity for studying the prokaryotic contribution in two different modes of interaction, pathogenicity and mutualism, while also studying the role of the host in determining the outcome of association with the bacteria (Clarke, 2008). The Photorhabdus life cycle involves alternating pathogenic and mutualistic relationships with invertebrate hosts. Some species of Photorhabdus form specific mutualistic associations with nematodes of the family Heterorhabditidae. Together, the nematode and associated bacteria are virulent pathogens of a wide range of larval stage insects and, hence, are considered to be effective bio-control agents. The Photorhabdus life cycle is complex but key to the successful continuation of the mutualism with the nematode is the colonization of the infective juvenile (IJ), a specialist free-living stage in the nematode life cycle. Evidence suggests that Photorhabdus forms a biofilm in the gut of the IJ, implying that biofilm formation is required for IJ colonization. Catherine Easom from the Clarke group (University College Cork, Ireland) used a mutant screen to identify genes required for both P. luminescens biofilm formation and colonization of the IJ. In addition to genes involved in lipopolysaccharide biosynthesis, she found a mutation in hfdR, encoding a LysR-family transcriptional regulator. In E. coli, HfdR represses flhDC expression, although Easom has shown that HfdR does not affect P. luminescens motility. Therefore, HdfR is predicted to play a novel role in both biofilm formation and IJ colonization in P. luminescens. Although the term ‘probiotic’ is generally reserved for its application in the gastrointestinal tract of animals and humans, the use of antagonistic plant-associated microorganisms in the bio-control of plant pathogens could be classified as probiotics for plants, since preventing pathogenesis promotes plant health. This was exemplified for the sugar beet rhizosphere by Christin Zachow from the Berg group (Graz University of Technology, Austria). The pathogen Rhizoctonia solani was suppressed by a mixture consisting of three, for sugar beet autochthonous, and four, for sugar beet allochthonous), plant-associated microorganisms belonging to Trichoderma, Pseudomonas and Serratia genera (Fig. 6).

Figure 6.

Confocal laser scanning microscopy (Leica Microsystems, Heidelberg, Germany) of the main root of 2-week-old sugar beet seedling colonized by DsRed2-labelled Pseudomonas trivialis RE*1-1-14 (red colonies) and GFP-labelled Trichoderma sp. G1/8 (green hyphae). Sugar beet seeds were primed in a suspension containing bacterial cells and fungal spores. This figure was kindly provided by Christin Zachow and Gabriele Berg.

Concluding remarks

Overall, this symposium provided an excellent environment for those interested in microbial sciences to interact with each other and to learn about a diverse array of topics. The aim of the organizers was to enable young microbiologists working in all areas of the discipline and from around the globe to communicate about their work. The energetic and revealing discussions involving scientists who are just beginning their careers suggested that this aim had been successfully fulfilled.

There were many highlights of this meeting, ranging from descriptions of the molecular mechanisms underlying the integration of the Csr signalling cascade and the global second messenger cyclic di-GMP, through the first structural characterization of the S-layer of a bacterial pathogen, to the elucidation of a novel contact-dependent signalling mechanism between S. gordonii and C. albicans and the use of a novel murine model to visualize bacterial gallbladder infection. Moreover, not only was the diversity of current microbiology research beautifully illustrated, but the meeting was also able to bring to the forefront some excellent examples of the successful collaboration of scientists from very different disciplines. Ultimately, we anticipate that advances, such as those described at this meeting, towards further understanding microbial life might indirectly enable the development of the next generation of therapeutics for the treatment of disease.

This meeting also provided an opportunity to reflect on current progress in microbiology. In his closing remarks, Fergal O'Gara (University College Cork, Ireland) highlighted how cross-disciplinary molecular techniques are now being utilized in microbiology, as mentioned by many of the speakers. He also speculated on the future direction of the field, suggesting that it is likely to rely heavily on integrated ‘systems’ and ‘omics’ approaches to elucidate the complex network of pathways involving gene regulation, signalling, metabolism and behaviour of microbes in both synthetic and natural environments. He stressed his view that future studies must avail and be guided by bio-mathematical modelling and predictions based on systematic simulations.

In conclusion, this meeting provided a valuable and enjoyable opportunity for young microbiologists to meet their peers and learn about diverse areas of the field. Moreover, the meeting emphasized to all participants that the field of microbiology is blossoming and enjoys a bright future.

Acknowledgements

Special thanks must go to all those participants approached during the preparation of this report for providing images and helpful remarks. We would like to thank Karen O'Donovan, Yvonne McCarthy, Aileen O'Connell and Phrueksa Lawongsa for their excellent support during the organization of the meeting. We also thank the British Council, the Genetics Society, European Molecular Biology Organization, the Society for General Microbiology and the American Society for Microbiology for financial support of the meeting. The authors' research is supported as follows: R.P.R., in part by Science Foundation Ireland (SFI 03/IN3/B373 and 07/IN.1/B955); T.R., funded in part by the National Institutes of Health (GM066794; GM059969); S.C.J.D.K. is a postdoctoral researcher of the FWO-Vlaanderen; and S.J.C., by the BBSRC.

Ancillary