Causalities of war: The connection between type VI secretion system and microbiota

Abstract Microbiota niches have space and/or nutrient restrictions, which has led to the coevolution of cooperation, specialisation, and competition within the population. Different animal and environmental niches contain defined resident microbiota that tend to be stable over time and offer protection against undesired intruders. Yet fluxes can occur, which alter the composition of a bacterial population. In humans, the microbiota are now considered a key contributor to maintenance of health and homeostasis, and its alteration leads to dysbiosis. The bacterial type VI secretion system (T6SS) transports proteins into the environment, directly into host cells or can function as an antibacterial weapon by killing surrounding competitors. Upon contact with neighbouring cells, the T6SS fires, delivering a payload of effector proteins. In the absence of an immunity protein, this results in growth inhibition or death of prey leading to a competitive advantage for the attacker. It is becoming apparent that the T6SS has a role in modulating and shaping the microbiota at multiple levels, which is the focus of this review. Discussed here is the T6SS, its role in competition, key examples of its effect upon the microbiota, and future avenues of research.

sheath, which propels the Hcp tube through the membrane complex.
At the tip of the Hcp tube is the puncturing device made of a trimer of VgrG proteins and PAAR protein cap, which facilitates puncture of both producing and target cell membranes. Loaded on the secretion machine are two classes of effectors ( Figure 1). The "evolved" or "specialised" class has effector domains fused to structural T6SS components, that is, Hcp, VgrG, or PAAR proteins (Coulthurst, 2019;Hachani, Wood, & Filloux, 2016). The "cargo" effectors interact with these aforementioned proteins either directly or through specific adaptors proteins (Coulthurst, 2019). These effectors are delivered extracellularly, into eukaryotic target cells such as amoeba, animal cells, or fungal competitors, or more commonly into bacterial competitors to kill or suppress growth (Figures 1 and 2  These effectors are encoded in effector-immunity pairs (also called toxin/immunity or effector modules) with the immunity/antidote preventing intoxication from T6SS active siblings (Coulthurst, 2019).
The effectors secreted by the T6SSs are classified according to their targets and functions (Russell, Peterson, & Mougous, 2014). Amongst antibacterial effectors targets are nucleic acids cleaved by nucleases (Ma, Hachani, Lin, Filloux, & Lai, 2014;Pissaridou et al., 2018), cell wall degraded by peptidoglycan amidases or hydrolases (Russell et al., 2012), membranes permeabilised by lipases/phospholipases Whitney et al., 2013), or pore-forming toxins (LaCourse F I G U R E 1 Schematic of the T6SS from host to effector delivery: (I) effectors coupling for T6SS delivery. "Cargo" effectors noncovalently interact with structural components (VgrG/PAAR/Hcp), whereas "specialised" or "evolved" effectors contain additional effector domains within one protein (VgrG/PAAR/Hcp). (II) Protein subcomplexes of the T6SS. (III) Model of contraction-based firing. The T6SS assembles in an extended conformation, fires, and is then disassembled for recycling. Delivery of effectors (purple) into competing cells results in growth stasis or death. (IV) Antibacterial activity of the T6SS in the intestine. Top panel shows T6SS-mediated defence by the microbiota (blue) against pathogen (red). Bottom panel shows invasion and killing of microbiota by pathogen. Dotted outlines and pale colour indicate dead cells. (V) Model of human with outline of the intestine focusing on region of interest Miyata, Unterweger, Rudko, & Pukatzki, 2013). Other targeted processes are bacterial cell division (e.g., blocking FtsZ polymerisation; Ting et al., 2018;Wood et al., 2019), protein synthesis , or availability of dinucleotides such as NADP+ and ATP to cripple prey cells (Ahmad et al., 2019;Whitney et al., 2015). Some T6SS effectors function as metal ion scavenging proteins, which may result in starvation instead of cell damage and impact the community more broadly (Figure 2; Lin et al., 2017).
The delivered effector repertoire determines the T6SS efficacy.
The Pseudomonas aeruginosa H1-T6SS has nine experimentally tested effectors that target nucleic acids, peptidoglycan, the inner membrane, protein synthesis, NADP+, and the ATP pools (Ahmad et al., F I G U R E 2 Key functional impact and roles of the T6SS. See text for additional details 2019; Hachani, Allsopp, Oduko, & Filloux, 2014;Hood et al., 2010;LaCourse et al., 2018;Nolan et al., 2019;Pissaridou et al., 2018;Russell et al., 2011;Whitney et al., 2014;Whitney et al., 2015). The requirement for many distinct T6SS effectors is debatable. An arsenal of effectors with synergistic biochemical activities may be more potent than higher levels of one specific effector. Producer cells also increase the odds of killing different competitors and overcoming resistance by increasing effectors diversity and their targets. Additionally, certain effectors may have conditional efficacy and be more impactful under certain environmental conditions (e.g., low pH; LaCourse et al., 2018). Finally, trans-kingdom effectors target both eukaryotic and prokaryotic cells, enabling greater functionality for lower production cost (Jiang, Waterfield, Yang, Yang, & Jin, 2014).
Bioinformatic approaches focusing on conserved motifs, proximity to known T6SS components, or global genomic and proteomic approaches continue to uncover additional effectors, indicating that we are underestimating their number and thus their functions/cellular targets (Coulthurst, 2019;Jana, Fridman, Bosis, & Salomon, 2019;Nolan et al., 2019). T6SS-harbouring microorganisms also have multiple independent T6SS clusters; for example, P. aeruginosa encodes three and Burkholderia up to six (Schell et al., 2007). Overall, this suggests that we are likely scratching the surface of what these systems are capable of.
With this broad range of functions, it stands to reason that the T6SS will be employed by bacteria to help gain control of their environment/host or to facilitate community balance by eliminating cheaters and foes ( Figure 2).

| THE T6SS MODULATES THE EVOLUTION OF POLYMICROBIAL COMMUNITIES
A community of bacteria or even multiple strains of one bacterium can be shaped by the T6SS on the basis of the effector modules present. As opposed to the core T6SS structural components, which are typically highly conserved, sequence divergence is frequently observed for effector modules even within species (Chatzidaki-Livanis, Geva-Zatorsky, & Comstock, 2016;Pissaridou et al., 2018;Unterweger et al., 2014).
Antibacterial effector-immunity proteins enable social interactions through recognition of self when the same sets of effector-immunity proteins are expressed in different strains but can drive antagonism if different ( Figure 2). Opposing strain swarms of Proteus mirabilis form T6SS-dependent "battle lines" of segregation called Dienes lines (Alteri et al., 2017). In Vibrio cholerae, related families of effectors (>30% identity) are associated with immunity proteins whose variation is far greater (Kirchberger, Unterweger, Provenzano, Pukatzki, & Boucher, 2017;Unterweger et al., 2014). Combinations of distinct effector modules within a strain further contribute to interstrain incompatibility and confer a significant competitive advantage over rivals (Unterweger et al., 2014). T6SS armament can be increased by acquisition of individual effector modules, or small operons encoding vgrG/hcp/paar and cognate effector modules or in the case of Bacteroidetes via horizontally acquired entire T6SS operons on integrative conjugative elements with near 100% identity amongst Bacteroides species Coyne, Zitomersky, McGuire, Earl, & Comstock, 2014;Unterweger et al., 2014). Maintaining ancestral immunity genes or acquiring immunity but not effector genes is a good strategy to cope with T6SS-active strains in a population at lower cost, such as the C-type immunity gene in the aux-1 cluster of V. cholerae or the recently described acquired interbacterial defence gene clusters, which encode arrays of immunity genes (Alteri et al., 2017;Kirchberger et al., 2017;Ross et al., 2019;Wexler et al., 2016).
The advantage of the T6SS in polymicrobial populations extends beyond the simple killing of competitors, preventing their use of nutrients ( Figure 2). T6SS-mediated killing that results in lysis of prey cells leads to the release of nutrients and cellular content. Studies in V. cholerae (Borgeaud, Metzger, Scrignari, & Blokesch, 2015) and Acinetobacter baylyi (Ringel, Hu, & Basler, 2017) show that regulation of the T6SS and DNA uptake is linked. This enables coordinated killing of prey and DNA acquisition from lysed bacteria, which fosters horizontal gene transfer. Incorporation of this genetic material enables the acquisition of new antimicrobial resistance genes, pathogenicity islands, or T6SS effectors/immunity pairs, all of which could provide an evolutionary advantage (Blokesch, 2017;Borgeaud et al., 2015;Kirchberger et al., 2017;Thomas, Watve, Ratcliff, & Hammer, 2017).
Acquisition of new effectors may allow successful lineages of bacteria to dominate particular environments. One example of this is within the bee, where the microbiota are dominated by nine bacterial species (including Gilliamella apicola and Snodgrassella alvi, which are T6SS + ); however, multiple strains coexist within individual bees, hives, host species, and geographic locations. These two species contain a massive diversity of putative effectors with 1,112 identified bioinformatically within 77 strains (Steele, Kwong, Whiteley, & Moran, 2017). A subset of these effector genes was found to be upregulated in two S. alvi strains when grown within the bee gut, suggesting a role of the T6SS in either intrastrain competition (between S. alvi) or interspecies competition with other microbiota (Steele et al., 2017). This maintenance of the T6SS and diverse effectors suggests that T6SS provides an advantage and that coevolution amongst bacterial species is occurring in the bee gut. A second example is in the Hawaiian bobtail squid where intraspecies competition by Vibrio fischeri isolates facilitates colonisation of the light-organ crypts. Here, T6SS + strains outcompete T6SS − strains within a single crypt. Contrastingly, two T6SS + strains formed spatially segregated but interspersed microcolonies with a crypt. Incompatible effector modules and bacterial killing will prevent mixing similar to the Dienes lines observed in vitro for P. mirabilis and will highlight that T6SS fighting is occurring and shaping the population within a host (Speare et al., 2018). The T6SS can also function as a policing mechanism to enforce cooperation. In Burkholderia thailandensis, T6SS-1 and a range of public goods that benefit the overall population are controlled through quorum sensing (QS).
However, QS mutants have a growth advantage over wild type, as they do not produce public goods (e.g., extracellular proteases) but gain the benefits of them acting as cheaters in the population. This growth advantage can be observed in liquid coculture where the cheater population rises from 1% of the population to 50% over 3 days. However, the same population grown on solid surfaces stayed at 1% through the action of the T6SS (Majerczyk, Schneider, & Greenberg, 2016). By linking expression of public goods with antibacterial T6SS effectors, B. thailandensis has evolved as a hardwired regulatory system to control cheaters under defined environmental conditions.

| T6SS IN PLANT-ASSOCIATED BACTERIA
It is increasingly becoming apparent that the T6SS provides an advantage in several model systems particularly in an agricultural context. tumefaciens is outcompeted by the T6SS of P. aeruginosa. However, in a plant coinfection assay, the T6SS of A tumefaciens prevails over P. aeruginosa to gain a competitive advantage (Ma et al., 2014). Plants species also modulate their environment to promote beneficial microbes in the root microbiome to protect against soil pathogens (Lebeis et al., 2015), and future experiments will define the contribution of the T6SS in this niche.

| T6SS IN RESIDENT GUT MICROBIOTA
Given the role of the T6SS, it is reasonable to consider that this system could be capable of deeply transforming the gut microbiota, which may have a great impact on human health. The mammalian gastrointestinal tract harbours the densest microbial community currently known with dysbiosis associated with obesity, inflammatory bowel disease, and cancer. This community is numerically dominated by Gram-positive Firmicutes and Gram-negative Bacteroidales and is an attractive niche to prospect for T6SS interactions. One of the first studies to discover T6SS clusters in human gut microbiota strains was by Coyne et al. (2014). They identified the majority of the T6SS structural genes and a putative effector encoding a recombination hot spot (Rhs) protein with a deaminase domain and two putative immunities.
An analysis of the core T6SS iii clusters of 205 human gut Bacteroidales revealed three distinct genetic architectures (GA1, GA2, and GA3) distinguished by conserved orientation and organisation within T6SS clusters (Coyne et al., 2016). The role of GA1 and GA2 is unknown, but putative effector-immunity pairs (e.g., several evolved PAAR proteins with putative nuclease domains) hint at it being involved in interbacterial competition (Coyne et al., 2016). Supporting this notion, multiple species encode the same effector genes within a single microbiome (Verster et al., 2017). GA1 and GA2 are shared amongst Bacteroides (Coyne et al., 2016;Coyne et al., 2014), whereas GA3 is exclusively found in some Bacteroides fragilis strains (Coyne et al., 2014). Metagenomic and strain analyses have shown a monophyletic group of B. fragilis within individuals, suggesting single colonisation or out-competition events dependent on the effectorimmunity pair genotypes (Verster et al., 2017). An analysis of infant microbiomes containing B. fragilis is significantly more likely to harbour GA3 structural genes (92%) compared with those of adults (74%), suggesting that GA3 mediates strain level competition early on in gut microbiota development (Verster et al., 2017). The presence of GA3 in B. fragilis is also associated with higher relative abundance of B. fragilis and negatively correlated with Gram-positive Firmicutes genera (Faecalibacterium, Oscillospira, and Ruminococcus). As Grampositive bacteria have not been shown to be targets of the T6SS, the decreased abundance of these genera in GA3-containing microbiomes is most likely due to a fitness advantage conferred to GA3 + B. fragilis and not direct action against Firmicutes (Verster et al., 2017). Several studies have also demonstrated a role of the GA3 T6SS iii in the gut microbiota of mice. In competitive colonisation experiments of germ free mice, a human strain of B. fragilis was able to outcompete in a T6SS iii -dependent manner, the gut commensal Bacteroides thetaiotaomicron, which lacks the T6SS (Figure 1; Russell, Wexler, et al., 2014). Chatzidaki-Livanis et al. (2016) confirmed the expression of the system and a role in intrastrain killing as the parent outcompeted an effector-immunity pair mutant. Furthermore, a strain of B. fragilis can restrict colonisation of the murine host by two pathogenic strains of enterotoxigenic Bacteroides fragilis and reduced pathogen damage, which is associated with inflammatory bowel disease, sepsis, and colon cancer (Figure 1; Hecht et al., 2016). This T6SS antagonism between a gut resident and incoming pathogens likely contributes to gut homeostasis.

| THE T6SS ARMS RACE IN THE CONTROL OF GUT MICROBIOTA/ PATHOGEN BALANCE
If commensals use the T6SS to restrict invasion by intruders, it seems obvious that pathogens deploy their T6SS to invade the resident microbiota facilitating colonisation (Figure 1). Salmonella enterica Typhimurium is a major causative agent of human gastroenteritis. The Salmonella Pathogenicity Island Six encodes a T6SS cluster that contributes to the colonisation of mice and chickens. Recent work confirmed this role in mice and demonstrated that S. Typhimurium exhibits T6SS antibacterial killing of Klebsiella oxytoca and Klebsiella variicola (Sana et al., 2016). However, no killing of other Gramnegative bacteria was observed (e.g., Enterobacter cloacae, Escherichia coli, or commensals such as Prevotella copri, Parabacteroides distasonis, and Bifidobacterium longum; Sana et al., 2016). This suggests the T6SS has a defined role in colonisation, which is dependent upon antibacterial activity, but S. Typhimurium only kills specific members of the microbiota to invade the gut. The T6SS of pathogenic Shigella sonnei is also required to colonise the mouse gut (Anderson, Vonaesch, Saffarian, Marteyn, & Sansonetti, 2017). This T6SS mediates out-competition of commensal E. coli or Shigella flexneri both in vitro and in mice (Anderson et al., 2017). This confirms a role of the T6SS in host colonisation where, via direct killing, the colonisation resistance of the microbiota is overcome (Figure 1). This work also shows that S. sonnei has a T6SS-mediated competitive advantage over S. flexneri (which lacks the T6SS) and may be contributing to its increasing global dominance over S. flexneri as the leading cause of Shigellosis (Anderson et al., 2017). Thus, the T6SS is an antibacterial weapon used by pathogens as a virulence factor.
Although pathogens can use the T6SS to promote colonisation, this may not be solely via killing of the microbiota and niche takeover but may also combine the host response. Drosophila infected with T6SS + V. cholerae have a reduced lifespan, enhanced intestinal damage, and higher diarrheal symptoms than has T6SS − infection (Fast, Kostiuk, Foley, & Pukatzki, 2018). However, removal of commensal bacteria attenuated disease, as did deletion of the fly IMD antibacterial pathway, highlighting that T6SS killing leads to secondary events that enhance pathogenicity (Fast et al., 2018). In the mammalian gut, several studies have suggested a role for the T6SS in facilitating colonisation by influencing the host response (Figures 1 and 2).
Indeed, V. cholerae T6SS-mediated antagonism and killing of mouse commensals was shown to drive a more acute host innate immune response with higher induction of pro-inflammatory factors (Zhao, Caro, Robins, & Mekalanos, 2018). RNA-seq analysis corroborates this with an upregulation of 14 host immune genes including multiple targets in the NF-κB pathway. The authors suggest that killing of the intestinal microbiota releases microbe-associated molecular patterns driving this response. This study also reported higher host diarrheal symptoms in mice challenged with wild type compared with the T6SS − .
Thus, in this case, V. cholerae is using its antibacterial capacity to clear its niche, and this killing in turn induces a host response potentially enhancing pathogenicity and transmission (Zhao et al., 2018). In contrast, a study using the zebrafish (Danio rerio) model demonstrated that V. cholerae could manipulate the host leading to ejection of its resident microbiota (Figure 2). Delivery of the actin crosslinking domain of vgrG-1 promoted host intestinal contractions, dispelling the microbiota leading to enhanced V. cholerae colonisation (Logan et al., 2018). Overall, these studies show that the T6SS can modulate the host gut environment for the pathogen's benefit ( Figure 2).

| FUTURE PERSPECTIVES AND CONCLUSIONS
For many decades, the microbial composition of the mammalian gut has been considered to play a critical role in human health, mediating everything from immune system education, regulating endocrine function, vitamin synthesis, to protecting from pathogens (Lynch & Pedersen, 2016). The examples above show that the T6SS has a role in both pathogen defence and pathogen invasion. It is attractive to consider that specific cocktails of symbiotic bacteria or even engineered T6SS + commensal strains may be used as the prebiotics or probiotics in the mammalian gut, or as biocontrol agents in plants in the future. However, careful consideration is needed, as these strains may kill the microbiota as much as the pathogen. Realistically, T6SS biocontrol is something that is obviously best ready for plant pathogen control, and promising work in this area has been performed (Bernal et al., 2017;Decoin et al., 2014; Figure 2).
One additional consideration is that the control of the T6SS in many bacteria and within the intestinal conditions is likely highly complex. Experimental studies have shown that the activity of the T6SS can be influenced by host or dietary components. Bile salts, for instance, amplify S. Typhimurium T6SS-mediated bacterial killing (Sana et al., 2016). Mucins or chitin have been shown to promote T6SS expression and activity in V. cholerae (Bachmann et al., 2015;Borgeaud et al., 2015). Intriguingly, the microbiota modify bile acids to inhibit the T6SS of pandemic V. cholerae (Bachmann 2015). This suggests that bacterial pathogenicity could be controlled by the addition of certain dietary components to modulate the behaviour of the microbiota.
In summary, the T6SS is present in many Gram-negative bacteria and found in both pathogens and nonpathogens. It can perform a wide variety of functions, and mounting evidence shows a clear role in modulating the microbiota (Figures 1 and 2). This can be achieved through the direct injection of effector proteins into competing bacteria, which results in their growth inhibition or death. This death may in turn have consequences for the surrounding bacteria and the host.
T6SS effectors can also directly subvert the host for their gain, and recent examples show that this can trigger changes to the microbiota.
As bacteria in the human gut achieve the greatest densities recorded for any ecosystem, with an estimated 100 trillion cells, many more examples of T6SS-associated activities are likely to be identified.
Recent discoveries of the T6SS in killing fungus, that is, Serratia marcescens T6SS effectors Tfe1 and Tfe2 killing Candida albicans, Candida glabrata, and Saccharomyces cerevisiae (Trunk et al., 2018), or altering the behaviour of Gram-positive bacteria, that is, Pseudomonas chlororaphis influencing sporulation of Bacillus subtilis (Molina-Santiago et al., 2019), are opening unprecedented scope to this remarkable molecular weapon (Figure 2). However, the T6SS is but one of the specialised secretion systems deployed by bacteria for combat. For instance, diffusible colicins, the type IV secretion system, the type V secretion system/contact-dependent inhibition systems, and diverse type VII secretion systems have all been shown to mediate bacterial killing in confined areas (Aoki et al., 2010;Cao, Casabona, Kneuper, Chalmers, & Palmer, 2016;Gonzalez, Sabnis, Foster, & Mavridou, 2018;Souza et al., 2015). More research will undoubtedly discover additional effectors and systems that function as bacterial weapons for the unabating microbial conflict, which results in innumerable bacterial causalities of war.

ACKNOWLEDGEMENT
We apologise to authors' work not cited due to space restrictions.