Abstract
The type VI secretion system (T6SS) is a mechanism evolved by Gram-negative bacteria to negotiate interactions with eukaryotic and prokaryotic competitors. T6SSs are encoded by a diverse array of bacteria and include plant, animal, human and fish pathogens, as well as environmental isolates. As such, the regulatory mechanisms governing T6SS gene expression vary widely from species to species, and even from strain to strain within a given species. This review concentrates on the four bacterial genera that the majority of recent T6SS regulatory studies have been focused on: Vibrio, Pseudomonas, Burkholderia and Edwardsiella.
This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
The type VI secretion system (T6SS) is a multi-functional virulence mechanism utilized by Gram-negative bacteria to kill prokaryotic and eukaryotic organisms. The T6SS is hypothesized to form a needle-like appendage that exports toxins across the bacterial cell envelope and into an adjacent target cell. Structural T6SS proteins form a spring-loaded inverted bacteriophage-like structure within the bacterial cytoplasm that is hypothesized to eject proteins across the cell envelope (Basler et al., 2012). The T6SS needle complex proteins share similarity with the T4 bacteriophage tail spike and include the haemolysin co-regulated protein (Hcp) and valine glycine repeat G (VgrG) proteins (Williams et al., 1996; Pukatzki et al., 2006, 2007; Leiman et al., 2009). Hcp is predicted to form a nanotube that is surrounded by VipA and VipB proteins (Bönemann et al., 2009; Basler et al., 2012). Upon contraction of the cytoplasmic VipA/VipB tube, the Hcp tube (postulated to be capped with VgrG proteins) is presumably ejected from the bacterium (Mougous et al., 2006; Pukatzki et al., 2006, 2007; Bönemann et al., 2009; Leiman et al., 2009; Basler et al., 2012). ClpV is required to disassemble the contracted VipA/VipB sheath (Basler & Mekalanos, 2012). These proteins are highly conserved among T6SSs and are imperative for the secretion of toxins from the bacterial cell (Ma et al., 2009).
Initially, the T6SS was described as a virulence mechanism specifically targeting eukaryotic cells. Virulence towards phagocytic cells like macrophages or the amoeboid host model Dictyostelium discoideum was described in a number of bacterial species including Vibrio cholerae (Pukatzki et al., 2006, 2007; Miyata et al., 2011), Burkholderia species (Shalom et al., 2007; Aubert et al., 2008; Burtnick et al., 2010; Chen et al., 2011), Yersinia pestis (Robinson et al., 2009) and Salmonella enterica (Parsons & Heffron, 2005). Phenotypic consequences of T6SS-based interactions with eukaryotes include actin cross-linking (Pukatzki et al., 2007), formation of actin protrusions (Aubert et al., 2008, 2010), limitation of bacterial colonization and intracellular replication, and adaptation to deoxycholic acid (Parsons & Heffron, 2005; Robinson et al., 2009; Chow & Mazmanian, 2010; Lertpiriyapong et al., 2012).
More recently, it was discovered that the T6SS enables T6SS+ bacteria to kill prokaryotic target cells in addition to eukaryotes (Hood et al., 2010; MacIntyre et al., 2010; Schwarz et al., 2010; Murdoch et al., 2011; Russell et al., 2011). V. cholerae uses its T6SS to outcompete a variety of Gram-negative bacteria (MacIntyre et al., 2010; Ishikawa et al., 2012). Pseudomonas aeruginosa encodes three bacterial toxins, two of which target the peptidoglycan layer and one that is probably introduced into the cytoplasm of target cells (Hood et al., 2010; Russell et al., 2011). Burkholderia thailandensis encodes five specialized T6SS clusters, one of which is important for inter-bacterial competition (Schwarz et al., 2010). Antibacterial T6SS effector proteins are encoded by a wide array of Gram-negative bacteria, implying that the T6SS is commonly used to mediate inter-bacterial competition (Russell et al., 2012).
T6SS gene clusters are present in numerous Gram-negative proteobacteria, encompassing an impressive array of pathogens that infect humans (Mougous et al., 2006; Pukatzki et al., 2006), animals (Schell et al., 2007; Burtnick et al., 2011) and plants (Bladergroen et al., 2003; Liu et al., 2008; Records & Gross, 2010). Given the diversity of bacteria that utilize the T6SS, it is not surprising that numerous regulatory mechanisms govern T6SS gene expression. Some common trends in regulatory mechanisms of T6SS include quorum sensing (QS), changes in temperature and pH, and two-component regulatory systems (TCSs) (Zheng et al., 2005; Aubert et al., 2008; Liu et al., 2008; Ishikawa et al., 2009; Khajanchi et al., 2009; Lesic et al., 2009; Pieper et al., 2009; Wang et al., 2009b; Chakraborty et al., 2010, 2011; Records & Gross, 2010; Zheng et al., 2010; Gode-Potratz & McCarter, 2011; Moscoso et al., 2011; Rogge & Thune, 2011; Zhang et al., 2011; Ishikawa et al., 2012; Sheng et al., 2012). Other environmental cues such as the concentration of iron, phosphate and magnesium have also been implicated in T6SS regulation (Mueller et al., 2009; Wang et al., 2009b; Chakraborty et al., 2010, 2011; Rogge & Thune, 2011; Lv et al., 2012). A summary of T6SS regulatory mechanisms in select Gram-negative bacteria is provided in Table 1. Although these mechanisms exhibit tight control over the T6SS of some bacterial species, other species or strains possess constitutively active T6SSs. It is curious why bacteria with constitutive T6SSs dedicate a vast energy expenditure to the production of T6SS proteins. Recently, much knowledge of T6SS regulatory mechanisms has been gained regarding environmental signals and the ensuing regulatory cascades that trigger T6SS gene activation. Here, we review recent advances in the field of T6SS regulation focusing on the genera Vibrio, Pseudomonas, Burkholderia and Edwardsiella. Furthermore, we provide a rationale for the disparity that some bacterial species possess constitutively active T6SSs, while others maintain strict regulation over T6SS genes.
Vibrio species
The majority of Vibrio T6SS research has focused on V. cholerae, the aetiological agent of the diarrhoeal disease cholera. This bacterial genus is extremely diverse and is classified into more than 200 serogroups. Of the seven recorded cholera pandemics, the strains responsible for the last two (and possibly three) pandemics are believed to belong to the O1 serogroup, including the El Tor strains C6706 and N16961 (responsible for the current seventh pandemic) (Cvjetanovic & Barua, 1972; Kamal, 1974). These strains utilize the toxin co-regulated pilus (TCP) and cholera toxin (CT) as their prominent virulence factors. Conversely, non-O1/non-O139 serogroup strains cause small-scale outbreaks of gastroenteritis, often independently of the CT and TCP. The overwhelming majority of sequenced V. cholerae strains possess the full complement of T6SS genes, suggesting that the T6SS is part of the V. cholerae 1500-gene core genome and plays a crucial role in V. cholerae’s lifestyle. However, T6SS gene regulation differs between strains, as some strains constitutively express T6SS genes and others restrict T6SS gene expression. Numerous publications have reported T6SS regulation in V. cholerae (Pukatzki et al., 2006; Ishikawa et al., 2009, 2012; Mueller et al., 2009; Zheng et al., 2010, 2011; Kitaoka et al., 2011) and the T6SS regulon in this pathogen is complex. Both the environmental cues and the ensuing regulatory cascades leading to T6SS gene expression will be reviewed here.
QS regulates V. cholerae virulence factors such as CT, TCP and chemotaxis genes in C6706 (Zhu et al., 2002). QS communication is mediated via sensing autoinducer (AI) concentration in the environment. At low-cell density, the QS response regulator LuxO is phosphorylated and induces the transcription of small quorum-regulatory RNAs (Qrrs). Qrrs bind and destabilize mRNA of the central output regulator, HapR (LuxR in Vibrio alginolyticus, or OpaR in Vibrio parahaemolyticus). Conversely, at high-cell density, LuxO is inactivated, allowing expression of the master regulator. LuxO, in combination with the regulator TsrA (VC0070), represses the T6SS in C6706 and mutations in both genes are required for Hcp secretion in this strain (Zheng et al., 2010). In addition to T6SS regulation, TsrA represses TCP and CT, but activates expression of the haemagglutinin protease HapA (Zheng et al., 2010). Thus, TsrA acts as a global regulator of virulence genes in the El Tor strain C6706, exerting both positive and negative influences on gene expression.
In some V. cholerae strains, T6SS genes are upregulated at a high-cell density. Expression of Hcp in O1 V. cholerae strains A1552, AJ3 and AJ5 (but not C6706) is dependent on QS networks, as the deletion of hapR or the AI synthase genes (cqsA and luxS) resulted in a loss of Hcp production (Ishikawa et al., 2009). Hcp is also important for pellicle formation in V. parahaemolyticus, implying that hcp gene expression occurs under high-cell density (Enos-Berlage et al., 2005). Interestingly, OpaR oppositely regulates two distinct T6SS gene clusters in V. parahaemolyticus (Gode-Potratz & McCarter, 2011), implying that one cluster is active at high-cell density and the other is active at low-cell density. More specifically, OpaR represses transcription and expression of Hcp1 indirectly as OpaR does not bind the hcp1 promoter region (Ma et al., 2012). OpaR was also found to positively regulate cyclic-di-GMP levels; a phenotype associated with increased T6SS gene expression in Pseudomonas aeruginosa (Moscoso et al., 2011). V. alginolyticus also encodes two T6SS gene clusters; however, only one of these is regulated via QS (Sheng et al., 2012). Hcp1 expression in V. alginolyticus peaks at mid-exponential phase and mutation of luxO resulted in lower Hcp transcript levels, while the opposite was true for a luxR mutant. These results are opposite to those observed for V. cholerae A1552, AJ3 and AJ5, and imply that V. alginolyticus upregulates Hcp expression under low-cell density. Similarly, Vibrio anguillarum upregulates T6SS genes under low AI concentrations (Weber et al., 2009, 2011). It is plausible that differences in T6SS QS regulation between strains arise from variations in the signal transduction pathways of QS.
V. cholerae spends the majority of its life cycle in the ocean or in brackish waters, where it constantly experiences fluxes in environmental conditions. Fittingly, V. cholerae alters its gene expression in response to changes in temperature and salinity. Ishikawa et al. (2012) demonstrated that the osmoregulator OscR (VCA0029) represses the expression of Hcp in A1552. V. cholerae El Tor strains A1552, E7946 and 93Ag49 produce and secrete Hcp when grown in medium with high salt concentration, as opposed to standard laboratory conditions where Hcp is produced (in a QS-dependent manner) but not secreted (Ishikawa et al., 2012). Furthermore, A1552 kills Escherichia coli MC4100, and this T6SS-mediated killing is enhanced with high salt conditions (Ishikawa et al., 2012). This implies that the V. cholerae T6SS is important for inter-bacterial competition in the species’ native environment. Interestingly, E7946 was also used in a study employing in vivo expression technology, which determined that T6SS genes were upregulated as the bacterium traversed the human and murine gastrointestinal tract (Lombardo et al., 2007). Furthermore, T6SS gene expression in strain C6706 was increased during an infection of mouse intestine (Mandlik et al., 2011). This adds additional complexity to the T6SS regulatory cascade in O1 V. cholerae strains, as T6SS genes appear to be upregulated not only in their native environment, but also during host infection.
Many environmental cues feed into the T6SS regulatory networks of V. cholerae. Once an external signal is received, the bacterium translates this signal to alter gene expression accordingly. A number of transcriptional regulators have been identified in V. cholerae that positively or negatively influence T6SS gene expression. Microarray and qPCR data by Syed et al. (2009) indicated that V. cholerae O395 flagellar mutants had increased T6SS gene expression. This supports the hypothesis that virulence and flagellar genes are inversely regulated. Zheng et al. (2011) identified the transcriptional regulator VCA0122 (encoded within the large V. cholerae T6SS gene cluster on the small chromosome) which, when deleted, results in reduced hcp expression and attenuated D. discoideum killing, but is dispensable for killing E. coli. Another positive regulator of the T6SS in V. cholerae is the alternate σ54 factor RpoN (Pukatzki et al., 2006; Ishikawa et al., 2009; Kitaoka et al., 2011; Dong & Mekalanos, 2012). RpoN and one of its activators, VasH, are crucial for V. cholerae virulence towards D. discoideum and for killing E. coli (Pukatzki et al., 2006; Kitaoka et al., 2011; Zheng et al., 2011; Dong & Mekalanos, 2012). VasH acts as a bacterial-enhancer-binding protein (bEBP) and is required for recruiting σ54 to −24/−12 promoter sequences. σ54-dependent T6SS promoters were found upstream of putative T6SS operons in V. cholerae, Pseudomonas aeruginosa and Aeromonas hydrophila, implying that regulation by bEBPs is common in T6SS gene clusters (Bernard et al., 2011). However, there exists a discrepancy in the literature regarding the specific operons regulated by RpoN in V. cholerae (Bernard et al., 2011; Dong & Mekalanos, 2012). Using electromobility shift assays and β-glucuronidase fusions as a reporter of transcriptional activity, Bernard et al. (2011) determined that VasH/RpoN regulates the large T6SS gene cluster, as well as the small auxiliary clusters which encode T6SS secreted proteins in V. cholerae O395. Contradictory to this, RNA-seq, ChIP-seq and qPCR data indicate that in strain V52, RpoN strictly regulates the smaller hcp-encoding clusters and not the large T6SS cluster (Bernard et al., 2011; Dong & Mekalanos, 2012). Differences in the promoter sequences of El Tor and classical strains may explain these different findings. VasH has an N-terminal regulatory domain, a central σ54-activating domain and a C-terminal helix–turn–helix DNA-binding domain similar to other bEBPs (Bernard et al., 2011; Kitaoka et al., 2011). Importantly, polymorphisms in VasH are not responsible for the differential regulation of T6SS genes in strains like V52 (constitutively active T6SS) and N16961 (repressed T6SS under laboratory conditions), as vasH from N16961 is functional when expressed in trans in both V52ΔvasH and wild-type N16961 (Kitaoka et al., 2011). Sequence alignments of vasH from V52, N16961 and four V. cholerae isolates from the Rio Grande indicate that vasH is conserved among not only pandemic and endemic clinical strains, but also those isolates from environmental sources (Unterweger et al., 2012). Taken together, a universal role for VasH in several T6SS+ bacterial species has been established.
T6SS regulation in V. cholerae is complex and can differ depending on the strain studied. It is currently unclear how all the internal regulators (VasH, HlyU, VCA0122, TsrA, OscR) function in concert within the cell to influence gene expression. At present, it appears that different extracellular signals (osmolarity, cell density, salinity, temperature) are responsible for regulating T6SS gene expression within the genus Vibrio.
Pseudomonas species
Bacteria of the genus Pseudomonas belong to the family Pseudomonadaceae which has nearly 200 described species (Euzéby, 1997). They inhabit a wide variety of environments such as soil, water and host organisms, and therefore have very versatile biology (Palleroni, 1992). Genomic information has been compiled for 90 strains (Silby et al., 2011), and recently in silico genomic analysis of 34 Pseudomonas genomes identified 70 T6SSs with at least one locus per strain (except Pseudomonas stutzeri A1501) (Barret et al., 2011). Six pathovars of the plant pathogen Pseudomonas syringae carry one or two T6SS gene clusters in addition to a T3SS (Records & Gross, 2010; Sarris et al., 2010; O’Brien et al., 2011). In contrast, the human opportunistic pathogen Pseudomonas aeruginosa encodes nearly all secretion systems described thus far, including three evolutionarily distinct T6SS clusters (designated Hcp secretion island (HSI) I–III) and an orphaned hcp2/vgrG2 locus (Mougous et al., 2006; Lesic et al., 2009; Termine & Michel, 2009; Bleves et al., 2010). Toxins critical for inter-bacterial competition are co-regulated with HSI-I genes (Hood et al., 2010; Russell et al., 2011); however, HSI-I also plays a crucial role in chronic infections as shown in the rat model of chronic respiratory infections, and in chronically infected cystic fibrosis patients (Potvin et al., 2003; Mougous et al., 2006). For HSI-I, PppA and PpkA (serine–threonine phosphatase and kinase, respectively) exert their effects on the T6SS scaffold protein Fha1, which mediates the export of T6SS proteins (Mougous et al., 2007; Casabona et al., 2012). HSI-II and HSI-III are important for pathogenesis in the Arabidopsis thaliana model (Lesic et al., 2009) and furthermore, HSI-II deletion mutants are attenuated in mouse infection and the HSI-III gene cluster is upregulated in the presence of epithelial cell extracts (Chugani & Greenberg, 2007; Lesic et al., 2009; Starkey & Rahme, 2009). HSI-I and HSI-III clusters are also induced during different stages of biofilm development (Southey-Pillig et al., 2005), which implies differential regulation via QS systems. Therefore, strict regulation of the different T6SS loci in Pseudomonas species seems to be crucial for a functional virulence mechanism (Mougous et al., 2006; Lesic et al., 2009; Bernard et al., 2010; Leung et al., 2011).
QS systems regulate ~350 genes in Pseudomonas aeruginosa and are involved in a wide variety of cellular processes (Veesenmeyer et al., 2009). The two major QS systems las and rhl are each composed of an N-acyl-l-homoserine synthase (LasI or RhlI) and its cognate transcriptional regulator (LasR or RhlR) (Déziel et al., 2005). Pathogenic Pseudomonas aeruginosa also synthesizes a quinolone-signalling molecule 2-heptyl-3-hydroxy-4-quinolone (PQS), which is a regulatory link between the Las and Rhl systems. Regulation of PQS biosynthesis genes is mediated by the transcriptional regulator MvfR (Déziel et al., 2004; Déziel et al., 2005; Venturi, 2006). Importantly, QS systems play a role in differentially regulating gene expression in all three T6SS loci in Pseudomonas aeruginosa (Lesic et al., 2009; Bernard et al., 2010). HSI-I gene expression is suppressed by LasR and MvfR, whereas HSI-II and HSI-III are positively regulated by both.
RpoN (discussed above for V. cholerae) positively regulates synthesis of RhlI and is crucial in the regulation of virulence genes in Pseudomonas aeruginosa (Buck et al., 2000; Hendrickson et al., 2001; Leung et al., 2011). Bioinformatic analysis suggests that the Pseudomonas syringae and Pseudomonas aeruginosa HSI-II and HSI-III clusters contain σ54-binding sites; however, it remains to be shown conclusively whether σ54-dependent gene regulation in Pseudomonas is required for T6SS-mediated virulence (Hendrickson et al., 2001; Alarcón-Chaidez et al., 2003; Bernard et al., 2011). Using microarray analysis, the TetR-like transcription factor PsrA was shown to upregulate the HSI-II locus in stationary phase (Kang et al., 2008) and in a separate report, microarray analysis identified MvaT (a histone-like nucleoid structuring protein) as a negative regulator of clpV3, hcp2 and vgrG2 in the HSI-II and HSI-III clusters (Castang et al., 2008).
The RetS/LadS/GacS pathway coordinately regulates gene expression of HSI-I and the T3SS during acute and chronic Pseudomonas aeruginosa infections (Goodman et al., 2004; Coggan & Wolfgang, 2012). RetS and LadS are orphan hybrid sensor kinases that counteract each other’s effects. RetS downregulates biofilm formation and T6SS gene expression, but induces low levels of the second messenger c-di-GMP and upregulates T3SS genes, whereas LadS has the reciprocal effect (Goodman et al., 2004; Ventre et al., 2006; Bordi et al., 2010). LadS also has a role in Pseudomonas aeruginosa PA14 cytotoxicity towards HeLa cells (Mikkelsen et al., 2011). Regulation of both secretion systems is mediated by the GacS/GacA TCS. Phosphorylation of GacA leads to transcription of the small RNAs rsmZ and rsmY, which sequester the post-transcriptional regulator RsmA, resulting in opposite regulation of the T3SS and T6SS. The same regulatory effect can be achieved by artificial modulation of c-di-GMP levels by overexpressing phosphodiesterases or the diguanylate cyclase WspR, suggesting that the pathways are linked (Goodman et al., 2004; Ventre et al., 2006; Bordi et al., 2010; Moscoso et al., 2011). The HptB signalling pathway also controls biofilm formation and the T3SS via a response regulator and an anti-σ factor and has an impact on the expression of RsmY. However, regulation of T6SS genes seems to be specifically controlled via the RetS pathway, as overproduction of RsmY alone does not have an impact on T6SS gene expression (Bordi et al., 2010). RetS and LadS sensor kinases also antagonistically regulate the T6SS and virulence factors in Pseudomonas syringae, but the RetS/LadS regulatory pathway seems to be parallel to, rather than upstream of, the GacS/GacA TCS (Records & Gross, 2010). Also, in Pseudomonas fluorescens Pf-5 and Pseudomonas brassicacearum the GacS/GacA TCS plays a critical role for the expression of T6SS proteins, as gacA and gacS mutants reduce T6SS gene expression compared with that of the wild-type (Hassan et al., 2010; Lalaouna et al., 2012). Therefore, it is conceivable that there is a connection between GacS/GacA TCS and T6SS in Pseudomonas species.
Altogether, the complex regulation of the T6SS in Pseudomonas is controlled at many levels and by different regulatory systems, highlighting the importance of this virulence factor in the life cycle of different Pseudomonas species. Currently, Pseudomonas entomophila and Pseudomonas mendocina are species used as simplified bacterial models for understanding/identifying regulatory mechanisms in silico (Sarris & Scoulica, 2011).
Burkholderia species
The Burkholderia genus encompasses several pathogenic organisms including Burkholderia mallei (Glanders disease), Burkholderia pseudomallei (melioidosis) and Burkholderia cepacia (human lung infections). Burkholderia thailandensis is a close relative of B. pseudomallei but rarely infects humans or animals (Glass et al., 2006). Burkholderia species possess multiple T6SS gene clusters: B. mallei and B. thailandensis have five T6SS clusters, while B. pseudomallei has six (Schell et al., 2007). The T6SS-5 of B. thailandensis targets eukaryotes and is important for murine pneumonic melioidosis, whereas T6SS-1 mediates inter-species bacterial competition (Schwarz et al., 2010). It is currently unclear what the other T6SS gene clusters in Burkholderia are important for and whether each one targets different species. Encoding multiple T6SS gene clusters that potentially target different species would require strict regulation over the various gene clusters. Recent advances in the regulation of Burkholderia T6SS clusters are discussed here.
The response regulator AtsR negatively regulates T6SS genes in B. cenocepacia (Aubert et al., 2008). AtsR shares similarity and functions similarly to the Pseudomonas aeruginosa regulator RetS, which is important for switching between acute and chronic stages of lung infection in cystic fibrosis patients (Mougous et al., 2006). Inactivation of AtsR results in increased biofilm production and expression, and hypersecretion of Hcp (Aubert et al., 2008).
The T6SS-1 gene cluster in both B. mallei and B. pseudomallei are strictly regulated, becoming active when the VirAG TCS is overexpressed or upon internalization by phagocytic cells (Schell et al., 2007; Shalom et al., 2007; Burtnick et al., 2010, 2011; Chen et al., 2011). The VirAG TCS is crucial for B. mallei virulence and regulates genes involved in actin tail formation, capsule production and the T6SS (Schell et al., 2007). Recently, a T3SS cluster 3 (T3SS-3) regulator, BspR, was identified in B. pseudomallei that also regulates the T6SS-1 cluster (Sun et al., 2010). Mutation of bspR reduced T6SS-1 gene expression compared with that of the parent strain when grown in liquid culture (Sun et al., 2010). BspR is a TetR family regulator involved in a complex regulatory cascade that includes downstream regulators BprP, BsaN, BicA and BprC. BprP is a membrane-localized DNA-binding protein that binds the promoter region of the AraC-type regulator bsaN. Consequently, BsaN (encoded within the T3SS-3 gene cluster) and its co-activator BicA positively influence the expression of BprC, which strictly regulates T6SS-1 and not T3SS-3 in B. pseudomallei (Sun et al., 2010). Interestingly, BprC is the critical regulator of T6SS-1 when B. pseudomallei is grown in liquid culture; however, following phagocytosis by RAW 264.7 macrophages, the crucial regulator of T6SS-1 genes becomes the VirA–VirG TCS (Chen et al., 2011). VirG and BprC have unique transcriptional start sites, with VirG acting directly at the hcp1 promoter and the VirG regulatory cascade acting upstream of tssA (Chen et al., 2011). Delineation of this regulatory cascade implies there is cross-talk between the T3SS-3 and T6SS-1 (via BsaN) and that both systems contribute to virulence in the B. pseudomallei mouse infection model.
In B. pseudomallei, deletion of hcp1 resulted in an increased LD50 in the Syrian hamster infection model, and the hcp1 mutant was impaired for growth and in the formation of giant multi-nucleated cells when used to infect RAW 264.7 macrophages (Burtnick et al., 2011). Importantly, Hcp1 expression was not detected when cells were grown in liquid culture, supporting the idea that the T6SS-1 gene cluster is important for intracellular growth/survival and is not active when grown in the absence of eukaryotic host cells. Conversely, Hcp6 was constitutively expressed (but not secreted) when B. pseudomallei was grown in liquid culture, and microarray analysis indicated that T6SS-6 genes were upregulated 100-fold compared with those of T6SS-1–T6SS-5 when grown in rich medium (Burtnick et al., 2011). Thus, T6SS gene clusters in B. pseudomallei are under control of different regulators and are expressed at different times during the bacterium’s life cycle. It will be interesting to determine what environmental cues activate T6SS-2–T6SS-5 and whether T6SS-6 is important for microbial competition.
Edwardsiella species
Edwardsiella tarda and Edwardsiella ictaluri are both fish pathogens in which the T6SS is crucial for virulence (Rao et al., 2004; Zheng et al., 2005; Zheng & Leung, 2007; Wang et al., 2009a, 2009b; Chakraborty et al., 2010, 2011; Rogge & Thune, 2011; Lv et al., 2012). E. tarda is an emerging pathogen that causes septicaemia in fish and gastrointestinal illness in humans. This versatile organism encodes 33 response regulators (Wang et al., 2009a), allowing it to respond to a variety of environmental stimuli. The TCSs EsrA-EsrB, PhoP-PhoQ and PhoB-PhoR have been associated with T6SS regulation, along with EsrC and the ferric-uptake regulator protein Fur (Zheng et al., 2005; Wang et al., 2009a, 2009b, 2010; Chakraborty et al., 2010, 2011; Lv et al., 2012). The response regulators EsrB and PhoP are crucial for E. tarda pathogenesis in zebrafish and the Japanese flounder (Wang et al., 2009b; Lv et al., 2012). E. ictaluri causes enteric septicaemia in catfish and also regulates its T6SS via EsrA-EsrB (Rogge & Thune, 2011).
The AraC-type transcriptional activator EsrC is encoded within the E. tarda and E. ictaluri T3SS and is directly regulated by EsrB. EsrC in turn regulates the T3SS and T6SS (Chakraborty et al., 2011), which implies a co-dependent regulatory mechanism for the T3SS and T6SS, given that EsrC is encoded within the T3SS gene cluster. Mutants in esrB are markedly attenuated in zebrafish (Lv et al., 2012) and mutation of esrC significantly increases LD50 levels in blue gourami fish (Zheng et al., 2005), implicating the T3SS and the T6SS as crucial virulence factors. In support of this, an evpP (which encodes a crucial T6SS effector protein) mutant is attenuated in zebrafish and Japanese flounder infection models (Wang et al., 2009b) and is defective for internalization by epithelial papilloma of carp cells (Wang et al., 2009b). Severe replication deficiencies in head kidney-derived macrophages (HKDM), and an avirulent phenotype in the channel catfish model in esrA and esrB mutants of E. ictaluri was also reported (Rogge & Thune, 2011). Interestingly, an esrC mutant maintained the ability to replicate within HKDM, but lost its virulence in catfish (Rogge & Thune, 2011). Therefore, the T3SS appears important for establishing infection and replication within macrophages, while the T6SS is required for virulence. Alternatively, the T6SS may limit intracellular replication within macrophages similar to Salmonella, since EsrB shares similarity with the Salmonella TCS response regulator SsrB (Parsons & Heffron, 2005).
Several triggers have been identified that activate the expression of T3SS and T6SS genes in Edwardsiella species including temperature, pH and the extracellular concentrations of magnesium, iron and phosphate (Zheng et al., 2005; Wang et al., 2009b; Chakraborty et al., 2010, 2011; Rogge & Thune, 2011; Lv et al., 2012). In E. ictaluri, T6SS genes are activated under conditions that mimic those of a phagosome; that is, low pH and limited phosphate (Rogge & Thune, 2011). In contrast, the E. tarda T3SS (and therefore the T6SS) is active at neutral pH and is repressed under acidic conditions (Srinivasa Rao et al., 2003; Rao et al., 2004; Zheng et al., 2005; Okuda et al., 2009). The PhoP-PhoQ TCS responds to changes in temperature and magnesium. Secretion of EvpC (Hcp homologue) and EvpP was observed at 23–35 °C, with maximal secretion at 30 °C and no secretion at 20 or 37 °C (Chakraborty et al., 2010). Between 23 and 35 °C, PhoQ autophosphorylates and the phosphate is transferred to the response regulator PhoP, which binds the PhoP box within the esrB promoter (Chakraborty et al., 2010). In turn, phosphorylated EsrB upregulates EsrC, leading to T6SS gene expression. Interestingly, PhoP has varying degrees of regulation over esrB depending on the strain. In strain EIB202, mutation of phoP reduced the expression of EsrB and EsrC by ~one third (Lv et al., 2012). This is in contrast with PPD130/91, where PhoP strictly regulates EsrB (Chakraborty et al., 2010). Elevated magnesium concentration (10 as opposed to 1 mM) reduces the expression of T6SS genes (Chakraborty et al., 2010), whereas elevated copper levels have no significant effect on T6SS gene expression (Hu et al., 2010). Importantly, temperature and magnesium concentrations conducive to T6SS gene expression are physiologically relevant and this implies that T6SS genes are upregulated during infection (Chakraborty et al., 2010).
Phosphate and iron are sequestered within the host and low concentrations can prompt invading pathogens to activate virulence genes. In PPD130/91, high-iron concentration lowered the expression levels of EsrC, EvpP and EvpC (Chakraborty et al., 2011). Assuming that iron depletion results in the opposite phenotype, this coincides with the hypothesis that virulence genes become activated once inside the host. Conversely, EIB202 increases evpP expression when grown in iron-rich media (Wang et al., 2009b). Obviously there are marked differences in T6SS regulatory cascades between these two strains and it is currently unclear why EIB202 upregulates EvpP under iron-rich conditions.
The E. tarda PhoB-PhoR TCS responds to phosphate in the extracellular environment. The response regulator PhoB activates transcription of the pstSCAB-phoU operon which is responsible for phosphate acquisition (Chakraborty et al., 2011). In PPD130/91, PhoB and EsrC simultaneously bind the promoter region of evpA (Chakraborty et al., 2011). T6SS genes are repressed at high phosphate concentration and this effect was exacerbated when PPD130/91 was subjected to both high iron and phosphate concentrations, indicating a synergistic effect (Chakraborty et al., 2011). Interestingly, transposon insertions in pstB, pstC and pstS abolished the production of both T3SS and T6SS proteins (Rao et al., 2004; Zheng et al., 2005). Thus, not only the PhoB-PhoR TCS, but also the phosphate acquisition proteins are involved in the regulation of virulence factors.
Why do some bacteria possess constitutively active T6SSs, whereas others maintain strict regulation of T6SS genes?
T6SS gene clusters have been identified in nearly 100 different bacteria, implying that this virulence factor provides a competitive advantage and has been selected for throughout evolution (Boyer et al., 2009). As discussed above, the regulatory mechanisms governing T6SS gene expression vary widely from species to species, and even from strain to strain. Interestingly, some bacteria constitutively produce T6SS proteins when grown under standard laboratory conditions, whereas others maintain stricter regulation of T6SS genes. We propose that this disparity arose as an adaptation, as bacteria co-evolve with eukaryotic organisms.
Prokaryotic organisms have been in existence for approximately 3.8 billion years, whereas the first single-celled eukaryote appeared ~2 billion years ago (Sogin, 1991; Mojzsis et al., 1996). Thus, inter-bacterial interactions were a significant driving force behind prokaryotic evolution for the first ~1.8 billion years. With the advent of eukaryotic organisms ~2 billion years ago, bacteria were faced with an additional survival challenge: single-celled amoebae like D. discoideum began to prey upon bacteria, and with the emergence of higher eukaryotes, macrophages started to engulf and destroy bacteria. We postulate that constitutively active T6SSs initially provided a fitness advantage for bacteria during environmental inter-bacterial interactions and that, over time, the T6SS evolved to additionally target microbial flora inside a host environment and eukaryotic target cells, which coincided with the evolution of stricter T6SS regulatory mechanisms (summarized in Table 2).
Several studies have reported T6SS antibacterial effects in a variety of species (Hood et al., 2010; MacIntyre et al., 2010; Schwarz et al., 2010; Murdoch et al., 2011; Haapalainen et al., 2012), suggesting that the T6SS may have evolved as a prokaryotic-killing mechanism. Furthermore, strong similarity between T6SS structural proteins and the T4 bacteriophage-puncturing device (Pukatzki et al., 2006, 2007; Leiman et al., 2009; Bernard et al., 2010; Basler et al., 2012) further suggests that T6SSs evolved to target other bacteria. Notable bacteria that possess constitutively active T6SSs (under standard laboratory conditions) include V. cholerae V52, Serratia marcescens and B. thailandensis (T6SS-1) (Pukatzki et al., 2006; Schwarz et al., 2010; Murdoch et al., 2011). Importantly, each of these organisms has been shown to target other prokaryotes using their T6SSs (MacIntyre et al., 2010; Schwarz et al., 2010; Murdoch et al., 2011).
S. marcescens strain Db10 uses its constitutively active T6SS to target a closely related S. marcescens strain ATCC274; however, Db10 does not target the eukaryotic organisms D. discoideum, Caenorhabditis elegans or Galleria mellonella (Murdoch et al., 2011). Thus, Db10 encodes an unevolved or archaic T6SS according to our proposed model.
V. cholerae V52 encodes a multi-functional T6SS that can target both prokaryotic and eukaryotic cells using the same set of T6SS proteins. Other V. cholerae strains encode a similar T6SS gene cluster but lack constant production of T6SS proteins. Interestingly, V. cholerae strains responsible for cholera pandemics are those that lack constitutive T6SS expression. This may be indicative of a closer evolutionary relationship between pandemic strains and eukaryotic hosts compared with strain V52.
T6SSs are encoded by all sequenced V. cholerae genomes, unlike CT that is encoded only by toxigenic strains. Host cues are required for V. cholerae to produce toxin-coregulated pili that are recognized as receptors for the CTX phage, allowing the delivery of the CT genes to the bacterial genome (Waldor & Mekalanos, 1996). Thus, toxigenic conversion of V. cholerae in the host suggests that the acquisition of CT genes co-evolved with eukaryotes. In contrast, the T6SS appears to have been acquired earlier, probably before the advent of eukaryotic cells, as they are found in the ancestral V. cholerae genome of all environmental (non-toxigenic) and clinical (toxigenic) V. cholerae strains sequenced to date. Ma & Mekalanos (2010) previously suggested that the T6SS is a core virulence determinant and could have facilitated the acquisition of additional virulence factors like CT and TCP in V. cholerae, leading to the development of highly pathogenic V. cholerae strains. In this case, the acquisition and use of other key virulence factors such as CT lessen the need for a constitutively active T6SS, leading to the utilization of negative regulators such as TsrA in V. cholerae (Zheng et al., 2010).
B. thailandensis is similar to V52 in that it has the ability to kill both prokaryotic and eukaryotic targets. However, unlike V52, B. thailandensis encodes multiple evolutionarily distinct T6SS gene clusters with different functions (Boyer et al., 2009). T6SS-1 is constitutively active and targets prokaryotes, whereas T6SS-5 strictly targets eukaryotes (Schwarz et al., 2010; Russell et al., 2012). The remaining T6SSs are uncharacterized at this time, but bioinformatic analyses suggest that these systems are more closely related to the T6SS-1 (Schwarz et al., 2010). It would seem that B. thailandensis uses its T6SS for inter-bacterial interactions and over time has acquired a new T6SS gene cluster which allows the organism to compete with eukaryotes as well.
In contrast to bacteria that constantly produce T6SS proteins and target prokaryotes are those which have developed stricter regulatory mechanisms (e.g. via the addition of a negative regulator) and use the T6SS only under specific conditions (such as following phagocytosis). Bacteria that fall into this category, all of which are pathogens, include Pectobacterium atrosepticum (Liu et al., 2008), Agrobacterium tumefaciens (Yuan et al., 2008), V. cholerae (e.g. strains C6706 and A1552) (Ishikawa et al., 2009, 2012; Zheng et al., 2010), B. cenocepacia (Aubert et al., 2008), B. pseudomallei (Shalom et al., 2007; Burtnick et al., 2011; Chen et al., 2011) and B. mallei (Schell et al., 2007; Shanks et al., 2009; Burtnick et al., 2010). We propose that these bacterial T6SSs evolved to target eukaryotes through close interaction with their host(s). This idea is supported by the case of B. mallei which, similar to B. thailandensis, encodes multiple T6SSs; however, the T6SS-1 (responsible for prokaryotic competition in B. thailandensis) cluster in B. mallei is significantly degraded and is unlikely to function (Schell et al., 2007). Furthermore, the T6SS-5 of B. mallei is crucial for competition against eukaryotic competitors and it was suggested that T6SS-5 could be the reason for B. mallei transitioning into an obligate pathogen, as opposed to B. thailandensis which has an active T6SS-1 (Schell et al., 2007; Schwarz et al., 2010).
Not all T6SS+ bacteria can be strictly segregated into either one of the two classifications we have just outlined. For example, Pseudomonas aeruginosa confers toxicity in eukaryotes and prokaryotes with its multiple T6SS gene clusters; however, none of these appear to be constitutively expressed under laboratory conditions (Mougous et al., 2006; Hood et al., 2010). Furthermore, organisms harbouring multiple T6SS gene clusters have just begun to be characterized, including those in V. parahaemolyticus, V. alginolyticus and Y. pestis (Pieper et al., 2009; Gode-Potratz & McCarter, 2011; Sheng et al., 2012). It will be interesting to determine whether the different T6SSs in these strains target different subsets of organisms. At the current time, we observe the trend that organisms with constitutively active T6SSs kill prokaryotes but not all organisms that kill prokaryotes have a constitutively active T6SS. Conversely, organisms with more strictly regulated T6SSs target eukaryotic organisms, but not all organisms with highly regulated T6SSs target eukaryotes. In the case of strictly regulated T6SSs, the incorporation of additional regulators may help the pathogen to coordinate the T6SS to prevent identification by the host immune system. This hypothesis is supported by the fact that infection of experimental animals with V. cholerae strains harbouring a constitutive T6SS causes major inflammation (Ma & Mekalanos, 2010; Zheng et al., 2010), further highlighting the need for tight T6SS control in a host environment. Based on these trends, we hypothesize that the T6SS evolved as a mechanism for bacteria to outcompete their prokaryotic neighbours and that constitutive expression of this molecular weapon conferred a survival advantage. Over time, the emergence of multicellular/eukaryotic hosts introduced bacteria to the immune system and made it crucial for T6SS+ pathogens to use their T6SS more wisely to compete with eukaryotic predators (phagocytic immune cells) and bacterial competitors (commensal flora) in a host environment, as was previously postulated (Miyata et al., 2010).
Concluding remarks
Recent work towards understanding T6SS regulatory cascades has demonstrated that the regulation of T6SS gene clusters is quite complex. As discussed here, the majority of recent work in T6SS regulation has focused on the bacterial genera Edwardsiella, Vibrio, Burkholderia and Pseudomonas. These organisms are found in a wide variety of environmental niches, and cause disease in a diverse array of eukaryotes. Thus, it is not surprising that these bacteria have evolved mechanisms, such as the T6SS, to promote their own survival. Depending on the bacterial species, the T6SS can serve as a eukaryotic virulence factor, a prokaryotic killing mechanism, or both. Regardless of the cellular target, the T6SS clearly provides a competitive advantage at some stage in the life cycle of T6SS+ bacteria. Further understanding of the T6SS regulatory networks will aid in discerning the forces driving T6SS evolution as a fitness tool for competing against both prokaryotic and eukaryotic organisms.
Acknowledgements
The authors thank Teresa Brooks, Daniel Unterweger and Marcia Craig for helpful discussions and the critical review of the manuscript. Work in S. P.’s laboratory was supported by the Canadian Institute for Health Research Operating grant MOP-84473, Alberta Innovates – Health Solutions, and the Canadian Foundation for Innovation. S. T. M. was supported by a PhD studentship from Alberta Innovates – Health Solutions.