Following oral inoculation of mice, Salmonella face the challenge of colonizing the intestinal lumen prior to invasion and dissemination to systemic sites. An uncontrived gut microbiota limits nutrient availability and provides colonization resistance toward this first infection step (Patel and McCormick, 2014; Erhardt and Dersch, 2015). Considerable research has investigated how Salmonella overcomes such host restrictions within the intestinal lumen, revealing a key role for horizontally acquired genes in endowing Salmonella with the ability to outcompete commensal microbes to establish an intestinal niche. A recent study identified direct killing of commensal microbes by Salmonella through the expression of a type 6 secretion system (T6SS), a widely-distributed nano-machine used primarily for inter-bacterial antagonism (Sana et al., 2016). The Salmonella T6SS, encoded on the horizontally acquired island SPI-6—formerly Salmonella centisome 7 genomic island (SCI 7; Blondel et al., 2009)—encodes an antibacterial amidase (Tae4) that can induce bacterial lysis in non-immune target cells (Sana et al., 2016). This ability seems to confer a fitness benefit to Salmonella at later stages of colonization of the gut lumen, but not systemic sites of infection (Sana et al., 2016, 2017). Interestingly, accessory genes within the SPI-6 island also contribute to intracellular replication and systemic dissemination in the permissive mouse line, C57BL/6 (Mulder et al., 2012). These data suggest that in addition to a role in antibacterial antagonism in the gut, the Salmonella T6SS appears to also have a secondary role in an unknown facet of intracellular cell biology. In other bacteria, the T6SS is involved in dampening host immunity (Chen et al., 2017), promoting intracellular growth in macrophages (Eshraghi et al., 2016), and interacting with the host microtubule network (Sana et al., 2015), all of which point to an expanding role for T6SS in host-pathogen interactions.

Induction of inflammation is another host defense mounted in response to Salmonella in the lumen. Surface proteins on Salmonella, like flagellin and LPS, are detected by pattern recognition receptors (PRRs) on epithelial and immune cells to activate innate immune responses (Gewirtz et al., 2001). This pro-inflammatory environment is associated with a rapid neutrophil influx, leading to induction of oxidative and nitrosative stress that targets both pathogenic and commensal bacteria in the lumen (Tükel et al., 2006). Salmonella actively induces intestinal inflammation, and this inflammatory response is important for successful luminal colonization (Stecher et al., 2007). One of the key mediators of this intestinal inflammation are the gene products of the Salmonella pathogenicity island 1 (SPI-1; Lee et al., 2000). This island encodes a type 3 secretion system (T3SS-1) that translocates effector proteins into host cells to modulate their function. Effector proteins encoded within SPI-1 and in other horizontally acquired islands are translocated through the T3SS-1, and can induce inflammation within the gut. The interaction of the SPI-1 translocated proteins SopB, SopE, and SopE2 with epithelial cell Rho GTPases activates NF-kB and eventually leads to the release of pro-inflammatory cytokines such as IL-1β and IL-23 (Santos et al., 2009). Certain SPI-1 secreted proteins, such as SipC, SipD, and SopE have been shown to induce iNOS expression in macrophages, and recent research has demonstrated that this induction, specifically by SopE, promoted Salmonella colonization (Cherayil et al., 2000; Lopez et al., 2012). The mechanistic basis for the enhanced growth of Salmonella in the inflamed gut appears to be largely metabolic fitness. Much work has demonstrated that Salmonella thrives in inflamed environments because it can better utilize metabolic by-products generated during inflammation. For example, nitrates produced from reactive oxygen and nitrogen products within the lumen, and tetrathionate produced from reduced thiosulfate, provide alternative energy sources for Salmonella, which can be metabolized in the anaerobic gut environment (Winter et al., 2010b; Lopez et al., 2012). Salmonella-specific genes such as the tetrathionate utilization operon (ttr) allow the bacteria to utilize these alternate metabolites to outcompete commensal microbes (Hensel et al., 1999; Rivera-Chávez et al., 2013). Although nitrate respiration is conserved among Proteobacteria, the existence of a secretion system effector that manipulates the host in order to preferentially generate nitrates has only been reported in Salmonella (Lopez et al., 2012). In addition to the metabolic benefits of oxidative and nitrosative stress, Salmonella encodes redundant systems of defense against these reactive oxygen and nitrogen intermediates, acting to mitigate damage under high concentrations (Vazquez-Torres et al., 2000; Aussel et al., 2011; Henard and Vázquez-Torres, 2011).

Within the lumen, salmonellae also encounter antimicrobial peptides secreted by epithelial cells and infiltrating immune cells, which work to disrupt membrane integrity (Bevins and Salzman, 2011). Salmonella has a number of genes that modify the outer membrane to protect against this host defense activity. Many of these genes are regulated by the PhoPQ and PmrAB two-component systems (Gunn and Miller, 1996). Horizontally acquired genes have also been implicated in the defense to antimicrobial peptides. For example, Salmonella can inhibit antimicrobial peptide production by Paneth cells in a T3SS-1-dependent manner (Salzman et al., 2003). This is likely mediated through the modification of host cell signaling pathways induced by SPI-1 effector proteins.

The attachment and invasion of Salmonella to host epithelial cells sets up conditions that eventually favor luminal colonization and systemic infection. Epithelial cells in the gut, specifically goblet cells, produce mucins that form a physical barrier between the apical surface of these cells and the lumen (McGuckin et al., 2011). Zarepour et al. showed that this mucus barrier is an important host defense against Salmonella (Zarepour et al., 2013). Virulent Salmonella strains can bind to mucin preparations better than avirulent strains, suggesting the genetic factors involved in Salmonella virulence mediate this binding, although the interacting partners have not been identified (Vimal et al., 2000). As Salmonella approaches the epithelial barrier within the intestine, it can be endocytosed by M cells (Clark et al., 1994). Entry into the lamina propria allows for TLR activation by flagellin on the basolateral surface, uptake by macrophages, and activation of B and T cells (Johansson et al., 2006). In addition to luminal sampling by M cells, Salmonella can mediate uptake by host epithelial cells through a cooperative interaction among several virulence factors. Genes on SPI-4, which encode a Type 1 secretion system and major adhesin protein SiiE, are important for initiating epithelial cell contact in the intestine (Gerlach et al., 2007). The activity of SiiE is important for maximal bacterial entry into polarized epithelial cells because it allows for clusters of bacteria in close proximity to the invasion site to be pulled into the membrane ruffle at the host cell surface (Lorkowski et al., 2014). In addition to SiiE in SPI-4, the horizontally acquired gene misL, encoded on SPI-3, binds fibronectin and is required for long-term colonization (Dorsey et al., 2005). Along with the protein ShdA, which is not encoded on a genomic island, binding to surface structures like fibronectin appears to be important for intestinal persistence (Kolachala et al., 2007). The activation of intestinal inflammation can increase levels of fibronectin on epithelial cells, and MisL has been suggested to bind fibronectin at areas of epithelial cell erosion, thus providing another example of the link between intestinal inflammation and efficient bacterial invasion.

The T3SS-1-mediated invasion of Salmonella has been extremely well-studied and the specific role of SPI-1 associated effectors in this process have been reviewed elsewhere (Lostroh and Lee, 2001; McGhie et al., 2009; Que et al., 2013). The net result of effector translocation upon host cell contact is the induction of major cytoskeletal changes in the epithelial cell, leading to membrane ruffling and bacterial internalization. The effectors involved in this process are encoded in SPI-1 and elsewhere in the genome and are coordinated at the level of gene regulation for synchronized expression. Novel functions and regulatory interactions are still being uncovered for even well-known SPI-1 effectors. For example, the phage-encoded effector, SopE, can exert an inside-out response on host epithelial cells by enhancing their expression of iNOS, which in turn liberates host-derived nitrate that luminal Salmonella can respire anaerobically (Lopez et al., 2012). The SPI-5 genomic island is another mosaic island that enables different stages of the Salmonella infection process. The SPI-5 effector, SopB, is expressed and translocated by the T3SS-1 early in infection to promote membrane ruffling and invasion, whereas PipB (also encoded in SPI-5), is activated when Salmonella is within host cells and is translocated by the SPI-2-encoded type 3 secretion system (T3SS-2) to promote intra-macrophage survival (Knodler et al., 2002). Recently, SPI-14 was identified as being involved in Salmonella virulence by mediation of bacterial invasion, however the gene responsible appears to play a role in the activation of SPI-1 genes, rather than directly interacting with host processes (Jiang et al., 2017).

Shortly after bacterial invasion, effectors secreted by SPI-1 control reversion of the host epithelial cell to its normal state, inducing de-ruffling of the membrane and repression of the pro-inflammatory signals. This is typically mediated by host ubiquitination pathways that target SPI-1 effector proteins for degradation. However, there are also specific effectors secreted through SPI-1 that modify the Salmonella containing vacuole (SCV). For example, the phage-encoded effector SspH2, translocated by SPI-1, helps to down-regulate pro-inflammatory responses once Salmonella is in the SCV (McGhie et al., 2009). SPI-1 effectors can recruit host factors to the vacuolar membrane to redirect the vacuole in the endosomal trafficking pathway, thus limiting lysosomal fusion. SPI-1 associated genes are also implicated in escape from the SCV in non-phagocytic cells, allowing for Salmonella survival and replication in the cytosolic compartment. This has been recently characterized in a number of epithelial cell lines (Knodler et al., 2010, 2014). The hyper-replication seen in the cytosol is associated with Salmonella induced pyroptosis of host cells, releasing primed bacteria into the lumen where they can infect other cells, or be picked up by macrophages. A subset of SCV's are subject to destabilization by SPI-1 associated genes, leading to a dual lifestyle of Salmonella within epithelial cells (Knodler, 2015).

Whether through bacterial mediated pyroptosis, trafficking of the SCV to the basolateral surface of epithelial cells, or by direct sampling of luminal bacteria by dendritic cells, Salmonella eventually arrives in the lamina propria where it is taken up by phagocytic cells through a combination of bacterial-mediated endocytosis and immune cell driven phagocytosis (Haraga et al., 2008). Within immune cells, several horizontally acquired virulence determinants are deployed in the creation of the SCV and the evasion of host immune defenses. The virulence genes necessary for bacterial survival within macrophages are predominantly located on the SPI-2 genomic island that encodes the T3SS-2 and two translocated effectors, SseF and SseG (Ochman et al., 1996; Shea et al., 1996). We will not review these genes in detail as they have been comprehensively reviewed elsewhere (Cirillo et al., 1998; Hensel et al., 1998; Haraga et al., 2008; Figueira and Holden, 2012). However, a number of phage associated genes and genes encoded on other genomic islands are paramount to the survival of Salmonella within macrophages. For example, SPI-12 encodes sspH2, which is secreted by the T3SS-2. SPI-12 is a pathogenicity island relevant for immune evasion that was first linked to virulence in a calf model of S. Typhimurium infection (Hansen-Wester and Hensel, 2001, 2002). Transcription of the sspH2 gene is induced in an SsrB-dependent manner during survival within RAW264.7 macrophages, and sspH2 mutants do not induce lethal infection of calves (Miao et al., 1999; Tomljenovic-Berube et al., 2013). Also present within SPI-12 is oafA, which acetylates O-antigen on its 2-hydroxyl group to generate the O5 serotype, which confers antigenic variation beneficial for pathogenesis and immune evasion (Slauch et al., 1996). Similarly, SPI-5 encodes pipA, another effector translocated by SPI-2 that if deleted, reduces virulence in a mouse infection (Knodler et al., 2002).

The importance of bacteriophages to the spread of virulence determinants through horizontal gene transfer is well-established (Barksdale and Arden, 1974; Cheetham and Katz, 1995; Waldor, 1998). In line with this, several pathogenicity islands in the Salmonella genome are proximal to either phage-encoding genes or phage attachment sites (Blanc-Potard and Groisman, 1997; Wood et al., 1998). Of pathogenic importance are two prophage-like genomic regions harbored within S. Typhimurium, Gifsy-1 and 2. Both of these genetic elements influence the interactions of Salmonella with host cells, as was identified due to their transcriptional induction upon exposure to reactive oxygen species. Notably, Gifsy-2 encodes the sodC periplasmic superoxide dismutase that confers resistance to oxidative stress (Figueroa-Bossi and Bossi, 1999). Variants of S. Typhimurium that are cured of these prophage become rapidly re-lysogenized upon exposure to hydrogen peroxide, indicating the importance of these genetic elements for pathogenesis and survival in vivo (Figueroa-Bossi and Bossi, 1999; Figueroa-Bossi et al., 2001). The Gifsy-1 prophage encodes gogB, an effector that can be translocated by both SPI-1 and 2 (Coombes et al., 2005a). GogB targets the host SCF E3 type ubiquitin ligase through an interaction with Skp1 and the human F-box only 22 (FBXO22) protein to inhibit the pro-inflammatory response once Salmonella is in the intracellular niche (Pilar et al., 2012). Another phage relevant for infection is Fels-1, which carries a third sodC gene, again contributing to the resistance response to oxidative stress (Brüssow et al., 2004).

Within the macrophage environment, both sequestration of essential metals as well the fusion of granules containing antimicrobial peptides to the SCV act together to restrict bacterial replication. SPI-11 encodes several genes that are critical for pathogenesis and survival at this stage of infection (Miller et al., 1989). This pathogenicity island is present in both the Typhimurium and Typhi serovars, however it has undergone degradation to some extent within Typhimurium. Of particular importance are a set of genes that encode the envelope proteins, pagD, envE, envF (Gunn et al., 1995). These genes participate in outer membrane remodeling that improves resistance to the immune defenses encountered intracellularly. Additional important SPI-11 genes are pagC and pagD, which are part of the PhoPQ regulon and were likely acquired via horizontal gene transfer (Gunn et al., 1995). Transposon insertions within these genes have been shown to attenuate virulence and restrict survival within macrophages (Miller et al., 1989).

The systemic spread of Salmonella to the liver and spleen is contingent upon survival within the macrophage environment; mutants in Salmonella that cannot survive within this environment are avirulent in a systemic mouse model of infection. Salmonella-containing macrophages enter the mesenteric lymph node, where they are shuttled to the liver and spleen (Watson and Holden, 2010). Replication within macrophages is also essential for colonization at the liver and spleen, as analyses of cell types within these tissues containing Salmonella identify macrophages, and to a degree neutrophils, as the primary cells at infection foci (Geddes et al., 2007; Thöne et al., 2007). Genes encoded on SPI-13 are important for intracellular viability (Shi et al., 2006). SPI-13 contains a cluster that encodes putative lyase, hydrolase, oxidase, arylsulphatase regulators, and deletion of this island renders Salmonella attenuated for virulence (Haneda et al., 2009). STM3118 (a homolog of acetyl-coA hydrolase) is an important virulence factor contained within this island that hydrolyzes acetyl-coA to acetate in order to modify peptidoglycan as a protective mechanism against degradative enzymes found in macrophages. Additionally encoded in SPI-13 is STM3119, a monoamine oxidase that converts aminoacetone (degradation product of L-threonine) to a peptidoglycan precursor (Shi et al., 2006). Further characterizing the function of genes encoded on SPI-13 is likely to broaden our understanding of how Salmonella modulates the intracellular environment and interacts with the host during systemic infection.

SPI-16 is a pathogenicity island that contributes to immune evasion in the form of O-antigen variation. It includes the gene STM0557 that resembles bactoprenol-linked glucose translocases (for example gtrA/B), and is predicted to be involved in serotype conversion through glycosylation of O-antigen. The gtr cluster of genes encodes proteins that add glucose residues to repeating O-antigen subunits within LPS, and ultimately confer form variation at the O-12 antigen galactose to generate the 12-2 variant. Most commonly, this variant arises after exposure to or growth within macrophages (Bogomolnaya et al., 2008). It has been demonstrated that form variation within O-antigen is critical for persistence of Salmonella infection in the murine intestine, and so SPI-16 may play a role in the re-infection of the intestine from the gall bladder once Salmonella has established a niche in the liver in a chronic model of Salmonella infection.

Taken together, the suite of horizontally acquired virulence genes found across the SPIs act in a concerted fashion to drive infection, as Salmonella successfully outcompetes the host microbiota, colonizes the lumen, invades the intestinal epithelium, and resides within the replicative niches of neutrophils and macrophages. At various stages of the pathogenesis process, Salmonella encounters several host defense mechanisms deployed by the immune system. However, it appears that Salmonella has evolved to integrate these immune response cues into signaling pathways leading to adaptive gene expression that, ultimately, evades these very host defenses (Wong et al., 2009; Arpaia et al., 2011). In severe stages of systemic infection, the establishment of infection foci in the liver and spleen eventually leads to necrotic lesions and death of susceptible mice through lipid A induced toxic shock (García-Del Portillo, 2001; Santos et al., 2001).

Although, the acquisition of genes through horizontal transfer has been essential to the evolution of Salmonella pathogenesis, the majority of foreign DNA is detrimental to bacteria (Buckling and Rainey, 2002; Navarre et al., 2007). Insertion within functional genes, overexpression of energetically taxing genes, and activation of unfavorable gene products are all examples of the drawbacks to horizontal gene transfer. Thus, the integration of these genes into core regulatory circuitry is critical for appropriate temporal gene expression and to ensure that these genes do not antagonize existing cellular functions. While this integration slowly proceeds via cis-regulatory mutation (discussed later), transcriptional silencing of horizontally acquired genes, termed xenogeneic silencing, protects the bacteria from these laterally acquired genes (Singh et al., 2016).

Seminal work over the past decade has identified H-NS, a DNA binding protein conserved across Gram-negative bacteria, as a broad repressor of horizontally acquired genes (Lucchini et al., 2006; Navarre et al., 2006). The mechanism of this repression relies on the intrinsic curvature of DNA rich in adenine and thymine (AT-rich; Navarre et al., 2006), which allows for H-NS binding in the minor groove, followed by nucleation and bridging along the DNA that ultimately blocks RNA polymerase (Ali et al., 2012). The existence of H-NS as a global sentinel of horizontally acquired genes has been speculated to contribute to the preferential retention of AT-rich horizontally acquired genes, as H-NS can mitigate the immediate harmful effects of lateral gene transfer (Dorman, 2007; Higashi et al., 2016).

Without compensatory mutations, deletion of hns tends to be lethal in Salmonella but not in closely related bacteria like E. coli, indicating the importance of H-NS to the biology and virulence of Salmonella (Navarre et al., 2006; Ali et al., 2014). Given the importance of horizontally acquired genes to the host-adapted lifestyle of Salmonella, research on H-NS has focused on the protein as a regulator of the pathogenic lifestyle. H-NS binds to SPIs 1-6 as well as other virulence-associated islets, blocking RNA polymerase from activating transcription from these loci (Stoebel et al., 2008). The importance of H-NS in pathoadaptation is further highlighted in in vitro evolution experiments, where mutations in hns are associated with loss of virulence determinants such as SPI-1 (Ali et al., 2014).

While H-NS is the most widely known silencer of virulence gene expression, it does not act alone upon horizontally acquired genes. Co-factors such as Hha and YdgT, and other nucleoid associated proteins like Fis contribute to the expression of horizontally acquired genes in Salmonella, including those that regulate virulence (Schechter et al., 2003; Silphaduang et al., 2007; Vivero et al., 2008). In contrast with core genes, which can be silenced by H-NS alone, the silencing of horizontally acquired genes such as those on the SPI loci or the pSLT plasmid additionally involve the co-factors Hha and YdgT (Vivero et al., 2008). Hha is specific to enteric bacteria and YdgT is specific to E. coli and Salmonella, yet are both involved in stabilizing H-NS at horizontally acquired promoters by mediating additional DNA contacts and allowing for H-NS aggregation (Coombes et al., 2005b; Ali et al., 2014). Similar to H-NS, the expression of Hha facilitates acquisition of foreign genes, thus contributing to pathogenic adaptation (Aznar et al., 2013). In summary, H-NS has been well-characterized as a buffer to help the bacterial cell tolerate new horizontal gene transfer events. Over time, these acquired genes have undergone further regulatory refinement for optimal expression in the host environment.

De-repression of H-NS by environmental stimuli is one way through which horizontally acquired genes can be activated in the right environments. Changes to DNA curvature induced by osmolarity or temperature shifts can modify the bridging activity of H-NS at specific promoters, allowing for transcription to proceed (Hinton et al., 1992). However, directed activation of virulence genes predominantly occurs through regulatory rewiring of H-NS binding elements to put these genes under the control of virulence regulators, or to allow for transcription factor binding that displaces H-NS bridges from the DNA. Examples of these methods have been identified for counter-silencing H-NS at horizontally acquired virulence genes in Salmonella, and their significance is discussed below.

Given the extent of xenogeneic silencing in Salmonella, the pathogenic lifestyle is highly reliant upon the evolution of regulators to counter-silence H-NS and relieve this repression (Stoebel et al., 2008). Many horizontally acquired genes become incorporated into the regulatory framework of two-component sensory systems, which are embedded in genetic networks to rapidly detect and respond to environmental cues (Wallis and Galyov, 2000). The PhoP-PhoQ regulatory system is highly conserved across bacterial species and governs several aspects of the Salmonella virulence program (Miller et al., 1989) with the regulation of ~5% of genes (Zwir et al., 2005; Harari et al., 2010). Consisting of a sensor kinase in the inner membrane (PhoQ), and response regulator in the cytoplasm (PhoP), this system senses low Mg²⁺ (García Véscovi et al., 1996), low pH (Bearson et al., 1997), and cationic antimicrobial peptides (Bader et al., 2005) within the stomach and small intestine to mediate rapid adaptation of Salmonella to the host environment. The detection of these environmental cues results in the positive regulation of PhoP-activated (pag) gene expression, and negative regulation of PhoP-repressed (prg) gene expression (Groisman, 2001). Additionally, PhoPQ activity is involved in the upregulation of both the PmrA-PmrB and SsrA-SsrB two-component systems, which are both important for intracellular survival (Gunn and Miller, 1996; Deiwick et al., 1999).

The majority of genes identified to be regulated by PhoP-PhoQ have been acquired by horizontal gene transfer (Groisman, 2001) and integrated into the PhoP regulon over evolutionary time. This suggests that horizontally acquired genes co-evolve with the Salmonella genome to become assimilated into existing core regulatory architecture, such that the spatiotemporal expression of virulence determinants is tightly controlled. Interestingly, PhoP appears to act differentially to regulate promoters acquired by horizontal gene transfer relative to those predicted to be ancestral (Will et al., 2014). A comparison of the promoter architectures between foreign and ancestral genes suggests that those that were horizontally acquired bind PhoP flexibly with high variability at a number of positions, whereas those that are part of the core genome interact with PhoP in a conserved manner at one binding site (Zwir et al., 2005, 2012). Furthermore, PhoP is capable of activating ancestral promoters directly via RNA polymerase holoenzyme interaction (Will et al., 2014), but horizontally acquired promoters only by counter-silencing of H-NS (Will et al., 2015). The importance of PhoPQ in regulating horizontally acquired genes is also evident from the divergence in PhoP targets despite the conservation of this regulatory system across diverse taxa. For example, Yersinia pestis (Grabenstein et al., 2006), Shigella flexneri (Moss et al., 2000), Erwinia carotovora (Flego et al., 2000), Klebsiella pneumoniae (Cheng et al., 2010), and Sodalis glossinidius (Toh et al., 2006) all contain PhoPQ and are severely attenuated for fitness in its absence, but have strikingly different lifestyles and highly distinct PhoP regulons (Groisman, 2001). These findings indicate that PhoPQ is a broadly conserved regulatory system that can flexibly integrate ancestral and acquired genes to accommodate bacterial lifestyles ranging from endosymbiosis to parasitism.

The SsrA-SsrB two-component regulatory system activates the T3SS-2 and is essential for survival and replication within host immune cells (Fass and Groisman, 2009). SsrB can directly bind to genes within SPI-2, and outside of this island, and activate their transcription (Worley et al., 2000; Walthers et al., 2007). This activity is mediated by a flexible 18 base pair palindromic sequence in 7-4-7′ architecture upstream of SsrB-regulated genes, allowing SsrB to regulate ~5% of the Salmonella genome (Tomljenovic-Berube et al., 2010). Interestingly, in the absence of H-NS, SsrB is less required for SPI-2 gene expression, and relatively recent work has demonstrated that SsrB can directly displace H-NS polymers along DNA (Walthers et al., 2011). Additionally, SsrB mediates a key regulatory cross-talk between SPI-1 and SPI-2. For example, SPI-2-encoded SsrB downregulates SPI-1 genes by repressing HilA and HilD (Pérez-Morales et al., 2017), and reciprocally, SPI-1-encoded HilD upregulates SPI-2 gene expression by directly binding the ssrAB operon (Bustamante et al., 2008). This further demonstrates the cross-talk between pathogenicity islands in order to regulate the different lifestyles of Salmonella during infection.

The classical definition of cis-regulatory evolution rests upon the inevitable accumulation of mutations in non-coding DNA that drift in the nearly neutral range (Stone and Wray, 2001). Those mutations that generate a fitness-increasing quantitative output to alter gene expression may sweep to fixation, creating novel regulatory nodes that result in the flexible expansion of complex genetic networks (Wray, 2007). Adaptation within cis-regulatory elements is proposed to contribute to genetic tunability in response to environmental cues, a critical component of host colonization. However, up until recently, empirical evidence for this in the context of bacterial pathogenesis was largely lacking. Recent work has shown that mutations in non-coding DNA are targets for polymorphism-fixing selection to assimilate genes into the SsrB regulon, and that divergence in the regulatory patterns between S. enterica and S. bongori (which lacks SsrB) confers pathoadaptive fitness differences. For example, we demonstrated that following acquisition of a new regulatory system, the promoters that regulate ancestral genes can evolve responsiveness to the new transcription factor to fine-tune fitness in the host (Osborne et al., 2009). This involves rewiring the cis-regulatory element controlling the ancestral gene that generates phenotypic diversity among the bacterial population that is selective in the host setting. Regulatory evolution explains much of the organismal diversity among closely related animals in the context of developmental evolution (i.e., “evo-devo”) however these findings suggested that regulatory evolution might be a more broadly applicable evolutionary principle. Work by other groups has since verified and extended these findings, showing that regulatory evolution drives diverse bacterial traits including immune evasion (Tuinema et al., 2014), antibiotic resistance (Horii et al., 1999), and virulence (Li et al., 2009).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.