What a Dinner Party! Mechanisms and Functions of Interkingdom Signaling in Host-Pathogen Associations

Adrenaline and NA are the most abundant catecholamines in the human body and are primarily involved in the fight-or-flight response. Adrenaline and NA contribute important functions to intestinal physiology, including modulation of intestinal smooth muscle contraction, submucosal blood flow, and chloride and potassium secretion (9). Several sources contribute to the pool of intestinal adrenaline and NA, which can reach micromolar concentrations (10). In addition to the central nervous system and adrenal medulla, the adrenergic neurons that are present in the enteric nervous system are major sources of NA in the intestine (11, 12). Furthermore, the resident microbiota, as well as pathogens, influences adrenaline/NA concentrations in the colon. The resident commensal microbiota stimulates the production of NA and, to some extent, adrenaline in the GI tract through an as-yet-unknown mechanism (13). Additionally, T cells, macrophages, and neutrophils generate and secrete adrenaline and NA (14). Therefore, pathogens may augment adrenaline and NA concentrations in the GI tract, not only because of the stress of infection but also by stimulating an immune response.

Adrenaline and NA were the first hormones recognized as influencing bacterial growth and virulence gene expression. Notably, these studies were conducted with bacteria that colonize the human GI tract (7, 15–28), such as enterohemorrhagic Escherichia coli (EHEC). Adrenaline and NA have been reported to enhance the growth of EHEC (19, 22, 24), possibly by functioning as siderophores that aid in iron uptake or by upregulation of iron uptake systems (19, 29). More recent work supports this idea, as NA was shown to induce enterobactin expression and iron uptake in E. coli (15).

Adrenaline and NA influence virulence gene expression in EHEC, independently of their influence on growth. Adrenaline and NA do not cross the cellular membrane; therefore, EHEC relies on the two-component systems (TCSs) QseBC and QseEF to relay the adrenergic signals to the intracellular environment. TCSs are typically composed of a histidine sensor kinase (HK) located in the cytoplasmic membrane that senses a specific signal and a cytoplasmic response regulator (RR) that controls the output (30). Upon sensing a particular environmental cue, the kinase autophosphorylates and subsequently transfers this phosphate to its cognate RR. The majority of RRs are transcription factors that, upon phosphorylation, mediate the output of this signaling cascade by binding DNA to promote or repress gene expression (30).

The sensor kinase QseC senses adrenaline and NA (17) and is a global regulator that influences the expression of more than 400 genes (31), including a second QS TCS, QseEF. To mediate these responses, QseC phosphorylates its cognate RR QseB, as well as the RRs QseF and KdpE (31–33). These RRs coordinate the expression of genes encoding flagella and motility, attaching-and-effacing (AE) lesion formation on enterocytes, stress responses, potassium uptake, and osmolarity (16, 26, 31–36) (Fig. 1).

QseC is critical for EHEC virulence in vivo, as a qseC deletion strain was highly attenuated in rabbit and bovine infection models (17, 37) and, significantly, QseC has been identified in at least 25 animal and plant pathogens and contributes to virulence in all of the examples characterized. QseC signaling enhances the expression of type III secretion systems (T3SSs), adhesins, biofilm formation, toxin production, intramacrophage replication by intracellular pathogens, and iron uptake, among others (38–45) (Table 1; Fig. 1). Although, QseC-dependent regulation has been directly linked to adrenaline/NA in only a subset of these studies, the cumulative results of these findings suggest that this interkingdom signaling is a conserved trait that allows bacteria to modulate virulence gene expression.

QseEF is a second TCS that plays a role in adrenaline- and NA-dependent interkingdom signaling (26). This TCS was discovered through experiments that demonstrated that inhibition of QseC-dependent signaling was effectively inhibited only by an α-adrenergic antagonist (phentolamine) (17), whereas AE lesion formation was also blocked by a β-adrenergic antagonist (propranolol) (7). Adrenergic receptors in mammalian cells are GPCRs that belong to two broad categories, α and β, and can be specifically blocked by α-adrenergic and β-adrenergic antagonists. Further examination confirmed that QseEF influences EHEC formation of AE lesions on epithelial cells (36). Additionally, QseEF regulates the expression of genes involved in the SOS response and Shiga toxin production, as well as genes encoding other TCSs, including RcsBC and PhoPQ (26, 46, 47). The QseC and QseE regulatory cascades have distinct, as well as overlapping, targets (46). Thus, adrenaline/NA signaling through QseC and QseE may lead to the enhanced expression of certain virulence genes, or perhaps QseC and QseE are important for EHEC to effectively coordinate virulence gene expression in response to distinct microenvironments within the host. EHEC colonizes the human colon, and it has to sense when it reaches this GI compartment, as well as whether it is in the lumen or at the interface with the epithelium, where it forms AE lesions. Different compartments of the GI tract have different concentrations of adrenaline and NA. Hence, the ability to sense these adrenergic hormones may be important for the optimal expression of EHEC’s virulence repertoire to ensure colonization of its host.

Peptides constitute an extremely diverse class of hormones. They are processed from a prohormone and exported out of the cell (8). Bacteria sense several of these hormones (Table 1). Helicobacter pylori colonizes the stomachs of 50% of the human population. This bacterium can cause duodenal ulcers and gastric cancer (48). Gastrin, a peptide hormone secreted by stomach cells, stimulates the release of gastric acid (49) and has been shown to increase the growth rate of H. pylori, and this gastrin-induced phenotype can be outcompeted by the synthetic analogue pentagastrin and another peptide hormone, cholecystokinin. Although the sensor for gastrin in H. pylori has not yet been identified, the fact that this response can be blocked by gastrin analogs suggests the presence of a bacterial receptor for this hormone. It is also noteworthy that this response is restricted to human gastrin, and this may explain the specificity of H. pylori to humans (50). Conversely, another gastric peptide hormone, somatostatin, which inhibits the secretion of gastric acid, suppresses H. pylori growth. Somatostatin has been shown to directly bind to bacteria, and this binding can be blocked by an antisomatostatin antibody, again suggesting that H. pylori harbors an as-yet-unidentified receptor for this peptide (51). Moreover, H. pylori infection also changes the levels of these two hormones, increasing and decreasing the levels of gastrin and somatostatin, respectively (52, 53).

Burkholderia pseudomallei is the causative agent of melioidosis, and the most common risk factor for melioidosis is diabetes mellitus. B. pseudomallei and the related species B. cepacia are both capable of binding and responding to insulin, the peptide hormone secreted by the pancreas that is responsible for the regulation of blood glucose levels, to increase growth and virulence by reducing the activity of phospholipase C and acid phosphatase (54). However, the bacterial insulin receptor remains unknown.

Another class of peptide hormones sensed by bacteria is natriuretic peptides, which are involved in the osmoregulation of blood. This family contains three structurally related peptides, atrial natriuretic peptide (ANP), brain natriuretic peptide (BNP), and C-type natriuretic peptide (CNP), which in mammalian cells signal through membrane-bound guanylyl cyclase receptors. Pseudomonas aeruginosa and P. fluorescens increase their cytotoxicity and modify their lipopolysaccharide (LPS) in response to BNP and CNP. This LPS modification causes it to be more necrotic to eukaryotic cells than the unmodified LPS is. These hormones increase the levels of cyclic GMP (cGMP) and cyclic AMP (cAMP) in Pseudomonas through two cytoplasmic adenylate cyclases, and cAMP binds to Vfr to engender this response. Cyclic nucleotides are important intracellular signaling molecules that promote biofilm formation and the expression of virulence traits in P. aeruginosa. Vfr bound to cAMP activates the expression of virulence genes that promote acute P. aeruginosa infection, while cGMP signaling represses the genes involved in acute infection and facilitates the expression of genes involved in biofilm formation, which causes chronic infections. The sensing of these natriuretic peptides may allow Pseudomonas to sense the blood environment. The receptor(s) that senses these peptides has not yet been identified, but the shared use of cyclic nucleotides as second messengers downstream of natriuretic peptides converges in both prokaryotes and eukaryotes (55, 56) (Fig. 1).

Another example of interkingdom signaling that involves peptide hormones is the sensing of EGFs, which are small proteins widely used as signals by animal cells (57). Secretion of EGFs requires proteolytic activation by the eukaryotic membrane serine protease rhomboid (RHO). RHO homologues in prokaryotes, such as the Aar protein encoded by Providencia stuartii, provide a link between eukaryotic signaling and host-microorganism communication. Aar can functionally substitute for RHO in Drosophila melanogaster; moreover, eukaryotic RHO can complement a P. stuartii aar mutant. RHO is involved in signaling in animal cells, and Aar is also required for the production of a bacterial QS signal whose structure is still unknown (58, 59). QS in P. stuartii represses the expression of an acetyltransferase that acetylates peptidoglycan (an intrinsic component of the bacterial cell wall). The increased acetylation of peptidoglycan changes the morphology of cells, causing them to round up as coccobacilli and form chains (60, 61). Although eukaryotic RHO and prokaryotic Aar are interchangeable, suggesting an interkingdom signaling link, it is unknown whether P. stuartii responds to the EGF eukaryotic signal or whether the P. stuartii Aar-dependent signal influences eukaryotic cells.

A final class of peptide reported to mediate interkingdom signaling comprises opioid hormones. Intestinal damage that may occur during surgery or biopsy-related procedures causes the opioid dynorphin to be released into the intestinal lumen. The opportunistic pathogen P. aeruginosa can internalize dynorphin, which activates the QS pathway through the Pseudomonas quinolone signal (PQS). The PQS is a QS signal that binds to the multiple virulence factor regulator (MvfR) or PQS regulator (PqsR) to enhance the expression of genes necessary for its own production, as well as promote virulence gene expression. MvfR/PqsR is involved in the action of dynorphin (Fig. 1) (62).

Steroid hormones control metabolism, immunity, sexual development, and many other physiological functions in mammals. They are generally synthesized from cholesterol and are lipids that can pass through the cell membrane. Their receptors are usually located within the nucleus or in the cytosol (8). It is noteworthy that one of the best-characterized classes of bacterial QS signals, the acyl-homoserine lactones (AHLs), are themselves lipidic and are also known to cross membranes, with their bacterial receptors, the LuxR-type family of proteins, being cytosolic (63). While screening Agrobacterium tumefaciens and P. aeruginosa for AHL-based-QS inhibitors, Beury-Cirou et al. identified three mammalian estrogen steroid hormones (estrone, estriol, and estradiol) as QS inhibitors (Fig. 1). They decreased AHL accumulation and the expression of QS-dependent genes (64).

Cytokines are the chief signaling molecules employed by the immune system, which makes them a prime target to be recognized as cues by pathogens. Several cytokines (such as interleukin-1 [IL-1], IL-2, and granulocyte macrophage colony-stimulating factor) increase the growth of pathogenic E. coli, suggesting that bacteria sense these cytokines and change their behavior (65). IL-1 was reported to bind to pathogenic E. coli and to Yersinia pestis (the causative agent of plague) (66). IL-1 binds to the F1 antigen of the Y. pestis capsule through the capsule antigen F1 assembly (Caf1A) protein, which is an outer membrane protein responsible for its assembly (66). Caf1A shows significant homology to the human IL-1β receptor, suggesting that the gene encoding Caf1A may have been horizontally acquired from mammals. Additionally, several Gram-negative bacteria, including E. coli, Shigella, and Salmonella enterica, bind to and alter their virulence in response to another proinflammatory cytokine, tumor necrosis factor alpha (TNF-α). The exact mechanisms of TNF-α recognition by these bacteria are unknown. TNF-α binding to Shigella enhances the invasion of epithelial cells and uptake and replication within macrophages (67). Another cytokine that has been reported to be sensed by bacteria is gamma interferon (IFN-γ). IFN-γ binds to outer membrane porin F (OprF) (Fig. 2A) of P. aeruginosa, leading to activation of the QS gene rhlI, which synthesizes the signal C₄-homoserine lactone (C₄-HSL), resulting in induction of the expression of lectin PA-1, which leads to disruption of the intestinal barrier, and pyocyanin (68). This IFN-γ enhancement of P. aeruginosa pathogenesis is especially intriguing in light of the observation that IFN-γ treatment worsens P. aeruginosa infections in several mouse models and clinical studies (69).

One of the front lines of defense of the innate immune system is the production of cationic antimicrobial peptides (CAMPs). CAMPs interact with and disrupt the bacterial membrane. Several bacterial pathogens modify the lipid A of their LPS to prevent CAMPs from binding to their outer membrane. One of the best-characterized examples of lipid A modifications that lead to CAMP resistance has been described in the intracellular pathogen S. enterica. These lipid A modifications are controlled by the PhoPQ TCS, where PhoQ is the HK and PhoP is the RR. PhoPQ regulation of genes encoding lipid A modification enzymes is vital for S. enterica virulence (70). PhoQ autophosphorylation is repressed by divalent metal ions such as Ca²⁺ and Mg²⁺, and their depletion induces virulence gene expression. These ions act as bridges between acidic residues in the periplasmic domain of PhoQ and the negatively charged phospholipids on the membrane, locking the protein in a repressed state (71). Conversely, CAMPs activate PhoQ by directly binding to the same metal binding sites on PhoQ (Fig. 2A), even in the presence of physiologically relevant concentrations of divalent cations. CAMPs alter the conformation of PhoQ, activating its kinase activity, leading to the phosphotransfer to PhoP and the activation of downstream genes necessary for invasion of epithelial cells and intramacrophage replication (72).

Inasmuch as bacteria sense host signaling molecules, it is not surprising that the host is also capable of sensing molecules important in bacterial signaling. The bacterial QS AHL signals have immunomodulatory effects (73). P. aeruginosa produces two AHLs, C₁₂-HSL and C₄-HSL. Extensive studies showed that C₁₂-HSL mediates both pro- and anti-inflammatory responses. To mediate an anti-inflammatory response, C₁₂-HSL inhibits the production of IL-4 and IFN-γ, thereby inhibiting the proliferation of both Th1 and Th2 T helper cells (74–76). Furthermore, C₁₂-HSL prevents LPS- or TNF-mediated activation of NF-κB, suggesting that this AHL modulates NF-κB-dependent gene expression induced by proinflammatory stimuli (77). In contrast, to elicit a proinflammatory response, C₁₂-HSL acts as a chemotactic cue for polymorphonuclear leukocytes (78) and induces IL-8 production through the mitogen-activated protein kinase pathway via the NF-κB and AP-2 transcription factors (79). IL-8 is a potent chemokine for neutrophils, and C₁₂-HSL NF-κB activation also leads to the activation of cyclooxygenase (Cox-2), which is a proinflammatory enzyme responsible for the production of prostaglandin (80). The strong induction of inflammation assists in the clearance of P. aeruginosa by healthy individuals but exacerbates the disease in cystic fibrosis patients (81). The anti- and proinflammatory roles of AHLs are probably reflective of the complex interaction of these molecules with different sets of immune cells during the infection process.

In addition to immune modulation, AHLs (specifically, C₁₂-AHL) induce apoptosis in various mammalian cell types, including macrophages, neutrophils, and some epithelial cell lines. AHLs suppress signaling through the signal transducer and activator of transcription 3 (STAT3) pathway (82, 83) and by releasing calcium from the endoplasmic reticulum (84).

AHLs are lipidic signals; hence, it was shown that C₁₂-HSL enters mammalian cells across the cell membrane in a fashion similar to that of mammalian steroid hormones (85). The observation that AHLs penetrate the mammalian cell membrane, combined with comparative structural analyses of LuxR-type bacterial receptors, led to the speculation that intracellular nuclear receptor proteins may be AHL receptors (86). Indeed, the peroxisome proliferation activator receptor (PPAR) family of nuclear receptors was shown to be mammalian AHL receptors (87) (Fig. 2B). However, PPAR recognition of AHLs does not fully account for all of the biological effects of this bacterial signal on mammalian cells, suggesting that other receptors exist (87).

Another facet of this interkingdom communication includes the fact that mammals evolved means to interfere with bacterial AHL signaling through a family of enzymes called paraoxonases (PONs), which act as lactonases that degrade AHLs. These enzymes are not secreted into the extracellular environment, where they could interfere with QS signaling, but they degrade AHLs within mammalian cells (88–91). There are three PONs (PON1, PON2, and PON3), and studies using knockout mice have determined that PON2 is the most likely candidate for an active AHL-degrading enzyme (92). Interestingly, in these arms races of interkingdom chemical signaling, a C₁₂-HSL-mediated increase in cytoplasmic calcium in mammalian cells leads to the degradation of PON2 mRNA and a decrease in the hydrolytic activity of this enzyme (93) (Fig. 2B).

ApoB is a component of low-density lipoprotein (LDL) and very LDL present in large concentrations in serum. Nox2 is an NADPH oxidase expressed in phagocytes that generates reactive oxygen species. Both Nox2 and ApoB antagonize the QS signal autoinducing peptide (AIP), which activates virulence gene expression in Staphylococcus aureus (94, 95). ApoB directly binds to AIP, sequestering this signal from its bacterial receptor AgrC and preventing QS and virulence activation in S. aureus (95). Nox2 oxidizes LDL, facilitating a conformational change in ApoB and allowing it to bind and sequester all classes of AIPs (there are four different AIPs, which differ in their amino acid sequences, that are produced by different S. aureus strains) (96). This constitutes an example of host defense by interference with bacterial QS.

Phosphatidylethanolamine (PE) is an essential membrane component of mammalian and bacterial cells. Besides contributing to membrane architecture (100), PE plays important roles in cell division and cell signaling and is also an important supplier of biologically active molecules (100, 101). The turnover and exfoliation of intestinal cells, including bacterial cells and enterocytes, provide a continuous supply of ethanolamine (EA) in the intestine that bacteria can use as a carbon and/or nitrogen source. Although EA is abundant in the GI tract, the resident microbiota does not efficiently metabolize EA (102), and thus, pathogens, including EHEC, S. enterica, and Enterococcus faecalis, exploit EA as a noncompetitive metabolite to outgrow the resident microbiota and establish infection (102–104). Notably, recent evidence suggests that EA metabolism confers a growth advantage on pathogens outside the intestinal environment, as both uropathogenic E. coli and Listeria monocytogenes rely on EA metabolism for robust growth during murine urinary tract and systemic infections, respectively (105, 106). Bacterial pathogens link the ability to sense EA with the genes required to catabolize EA. The EA utilization (eut) locus contains genes encoding the transport and breakdown of EA, as well as genes for a microcompartment that recycles cofactors and toxic intermediates generated during EA metabolism (reviewed in reference 107). Although EA metabolism is a trait associated with diverse bacterial phyla, the number and organization of the genes contained in the eut operon can vary considerably among species (44). The mechanism of regulating eut expression can broadly be categorized into the Gram-negative or Gram-positive classification scheme based on studies with S. enterica and EHEC (Gram negative) or E. faecalis and L. monocytogenes (Gram positive) (105, 108–112). In these examples, the eut locus is positively autoregulated not only in response to EA but also in response to the cofactor vitamin B₁₂ (detailed in Fig. 3).

In addition to enhancing growth, EA controls virulence gene expression in EHEC and S. enterica serovar Typhimurium (S. Typhimurium) (110, 113–115). In EHEC, EA activates the expression of genes important for colonization of the GI tract, including those encoding fimbrial adhesins and the locus of enterocyte effacement (LEE) genes that promote AE lesion formation on enterocytes, as well as genes encoding Shiga toxin (114, 115). EHEC encodes 16 distinct fimbrial loci (116, 117), and these fimbriae may be crucial for the initial adherence of EHEC to epithelial cells that precedes intimate, LEE-dependent adherence (118). However, many fimbrial loci have been hypothesized to be cryptic because of the difficulties of expressing fimbrial genes in vitro (119). Thus, the finding that the biologically relevant molecule EA promotes the expression of EHEC fimbriae suggests that these fimbriae play critical roles in the ability of EHEC to establish infection. The transcription factor EutR directly regulates LEE expression (110), and genetic data suggest that EutR also regulates a subset of EHEC fimbriae (114).

The in vivo importance and relevance of EA-dependent virulence gene regulation were recently demonstrated in S. enterica. S. enterica senses EA within macrophages to activate the expression of genes required for intracellular survival and replication and thus enhance dissemination. Importantly, EA-dependent virulence gene regulation is independent of EA metabolism, as a strain deficient in the EA catalytic enzymes is able to respond to EA and activate the expression virulence genes in EHEC and S. enterica (115, 113).

The mammalian intestinal epithelium is protected from direct contact with bacteria by the mucus layer. This mucus is composed of mucin, antimicrobial peptides, glycoproteins, glycolipids, epithelial cell debris, and 50% polysaccharides (121). The major structural component of mucus is mucin, a glycoprotein that has a protein backbone connected to hydrophilic and hygroscopic oligosaccharide side chains that form a gel-like tridimensional structure (122). A diverse collection of monosaccharides decorates mucin: arabinose, fucose, galactose, gluconate, glucuronate, galacturonate, mannose, glucosamine, N-acetylglucosamine, galactosamine, N-acetylgalactosamine, N-acetylneuraminic acid, and ribose. These sugars are made available to the microbiota through the polysaccharide-degrading activity of glycolytic commensal anaerobes. Hence, the mucus layer is an important habitat and source of carbohydrates for bacterial communities that colonize mucosal surfaces, especially in the colon (122). Moreover, gut commensals influence the glycan composition of mucin, with the prominent member of the GI microbiota Bacteroides thetaiotaomicron inducing the fucosylation of host mucins (123).

Enteric pathogens have evolved systems to exploit the metabolic properties of specific members of the microbiota as nutritional cues to gain a competitive advantage. S. enterica and Clostridium difficile exploit metabolic end products of B. thetaiotaomicron, including fucose, sialic acid, and succinate, and expand in the intestine following disturbances of the microbiota due to antibiotic treatment (124, 125). Furthermore, EHEC uses metabolic end products of B. thetaiotaomicron as nutritional cues to alter the expression of virulence genes important for host colonization. Specifically, glycophagic B. thetaiotaomicron generates fucose from host mucin, making fucose accessible to EHEC. EHEC senses fucose by using the TCS FusKR, which has been recently horizontally acquired by this pathogen, possibly from E. faecalis. Fucose sensing through the HK FusK initiates a signaling cascade through FusK’s RR, FusR, that regulates the expression of virulence and metabolism genes (Fig. 4). This regulatory circuit may function as a cue for EHEC to sense its location in the lumen, where B. thetaiotaomicron resides and makes this sugar source available and where the expression of EHEC’s virulence repertoire is onerous and not advantageous (126). Specifically, FusR represses the expression of the LEE genes, which are only necessary at the epithelial interface, and also the genes involved in fucose utilization. EHEC’s main competition for carbon sources in the lumen is commensal E. coli, which favors fucose as a carbon source within the intestine (126). Repression of the fucose utilization genes by FusR in EHEC prevents competition with commensal E. coli for fucose and rewires EHEC’s metabolism to favor d-galactose as a primary carbon source. Of note, d-galactose is also highly prevalent in this compartment, and EHEC utilizes this sugar source better than commensal E. coli does, gaining a metabolic competitive advantage (127). Thus, it is tempting to speculate that acquisition of FusKR enhances EHEC’s ability to successfully compete for a niche in the colon.

Using yet another nutrient-based environmental cue, EHEC also regulates virulence gene expression through recognition of glycolytic and gluconeogenic environments. The lumen is more glycolytic (rich in sugar) because of predominant glycophagic members of the microbiota degrading complex polysaccharides into monosaccharides that can be readily utilized by nonglycophagic bacterial species such as E. coli and Citrobacter rodentium (a murine pathogen that models EHEC mammalian infection). In contrast, the tight mucus layer between the lumen and the epithelial interface in the GI tract is devoid of microbiota; it is known as a “zone of clearance.” Enteric pathogens such as EHEC produce mucinases whose expression is enhanced by end products of B. thetaiotaomicron metabolism. As infection progresses, these mucinases completely degrade the mucus. In this scenario, both the pathogen and B. thetaiotaomicron are now closer to the epithelium in a gluconeogenic (poor sugar) microenvironment (no mucus left to extract sugar from), where B. thetaiotaomicron is producing and secreting succinate that is now sensed by the EHEC transcription factors Cra and KdpE to promote virulence gene expression, presumably at the interface with the epithelium. Importantly, succinate also influences host behavior, inhibiting polymorphonuclear leukocyte function (128, 129). Hence, the coupling of virulence regulation to optimal expression under gluconeogenic and low-fucose conditions mirrors the interface with the epithelial layer environment in the GI tract, suggesting a model in which EHEC will only express its virulence genes at optimal levels at the epithelial interface (35, 130) (Fig. 4).

Importantly, these sugar-sensing systems in EHEC are interconnected with the adrenergic sensing systems. The QseC and QseE kinases that sense adrenaline and NA through their RRs repress the expression of the fusKR genes, preventing FusKR’s repression of LEE gene expression (126). Another level of integration occurs with the sensing of gluconeogenesis and succinate, given that the QseC-phosphorylated RR KdpE directly interacts with the gluconeogenesis sensor Cra to promote virulence gene expression (35). Hence, at the epithelial interface, where adrenaline and NA are more likely to be found and the environment is gluconeogenic, QseC and QseE are triggered, and they promote LEE gene expression through the KdpE and Cra transcription factors and relieve the previous LEE repression by FusKR by repressing the expression of the fusKR genes.

The antioxidant glutathione (l-γ-glutamyl-l-cysteinyl-glycine; GSH) is a low-molecular-mass thiol that is involved in multiple physiological processes in plants and animals (reviewed in references 131 and 132). In prokaryotes, the ability to synthesize GSH is limited to cyanobacteria, proteobacteria, and a few strains of Gram-positive bacteria (133, 134). However, some bacterial pathogens that do not produce GSH, including Francisella tularensis (135) and Helicobacter pylori (136–138), have evolved mechanisms to scavenge and incorporate host-produced GSH into energy-generating or biosynthetic pathways. Moreover, the food-borne pathogen Listeria monocytogenes senses mammalian-produced GSH as an interkingdom signaling molecule. During infection, L. monocytogenes crosses the intestinal barrier and invades immune and epithelial cells, escapes from the phagosome, replicates in the host cell cytosol, and spreads from cell to cell, thereby escaping immune defenses. The transcription factor PrfA is a master regulator of virulence genes in L. monocytogenes and directs the expression of traits necessary for each of these steps of listerial infection (139–142). Recent genetic and biochemical data demonstrate that PrfA directly binds GSH to activate the expression of genes involved in L. monocytogenes pathogenesis. Importantly, PrfA binds to reduced GSH with higher affinity than oxidized GSH (143), demonstrating that sensing of GSH may be a mechanism for L. monocytogenes to specifically recognize the intracellular environment.

As signaling systems based on hormones, metabolites, and host defenses, which are used by mammalian pathogens, are discovered and characterized, additional research indicates that these interkingdom signaling mechanisms are conserved in bacteria that infect nonmammalian hosts. For example, in the fish pathogen Edwardsiella tarda, the QseBC TCS regulates the expression of genes encoding motility, fimbriae, and the T3SS (45) similarly to the role of interkingdom signaling in mammalian systems. Moreover, E. tarda responds to adrenaline to promote motility, and this effect was lost by a qseB or qseC deletion strain (45), suggesting that the mechanism of QseBC-dependent gene regulation is conserved in diverse pathogens. Additionally, the eut locus that encodes EA metabolism has been implicated as important to the infection strategies of the insect pathogen Photorhabdus luminescens and the plant pathogen Erwinia chrysanthemi (144, 145). Interestingly, a disruption of eutR in E. chrysanthemi results in impaired systemic infection of African violet plants (145). Whether this effect is due to the defect of EA metabolism and/or EA signaling through EutR-dependent regulation of virulence traits has not been reported. Finally, the coral pathogen Vibrio coralliilyticus senses the host metabolite dimethylsulfoniopropionate (DMSP) to regulate chemotaxis and chemokinesis in order to infect the mucus of its coral host, Pocillopora damicornis (146). Significantly, V. coralliilyticus does not metabolize DMSP, indicating that DMSP functions specifically as a signal for V. coralliilyticus to regulate virulence genes (146).

Although this review has focused on interkingdom signaling in host-pathogen interactions, interkingdom signaling also plays a role in mutualistic relationships between bacteria and hosts. One of the best-studied examples is that of rhizobia and host legume plants. The rhizobia are nitrogen-fixing bacteria that infect plant roots, eventually forming a new root organ or a nodule, under nitrogen-limiting conditions. In the nodule, the bacteria fix nitrogen that can be used by the host, and in return, the plant provides carbon that the bacteria use for growth. For this mutually beneficial relationship to be established, signals are exchanged between the bacteria and the host. Specifically, the legumes produce flavonoids that are sensed by the bacterial NodD protein. Subsequently, the NodD protein initiates a regulatory cascade essential for nodule formation. Furthermore, the bacteria also produce signals that influence the host response to infection. For example, the bacterial Nod factor may play a role in suppressing the plant immune response, thereby inducing nodule development (164). Studies on how the rhizobia influence plant innate immunity to establish symbiotic associations highlight that interkingdom signaling is not limited to host-pathogen interactions. Indeed, it is likely that additional studies that examine interkingdom signaling between a host and the resident microbiota will demonstrate that interkingdom signaling is a conserved mechanism shared by pathogens and commensal bacteria to form associations.