Possible Mechanisms of Action of Norepinephrine During E coli Infection

Though evidence exists that catecholamines partly exert their effects by acting on host tissues (above), it is clear that during culture in vitro they promote the growth of Gram-negative bacteria of several genera and expression of their virulence factors.

E. coli O157:H7, Salmonella enterica and Yersinia enterocolitica vary markedly in their ability to grow in response to catecholamines (Freestone et al. 2007b). The response of Y. enterocolitica was limited to NE and DA, and these proved to be more potent inducers of growth than epinephrine in E. coli and Salmonella. The impact of catecholamines on bacterial growth is strictly dependent on inoculum density and media composition, being most prominent in a minimal salts medium supplemented with 30% (v/v) adult bovine serum (serum-SAPI) from inocula of 102-103 colony-forming units (CFU; reviewed in Chap. 3). From low inocula such media are bacteriostatic in the absence of catecholamines and are proposed to mimic the nutrient poor and iron-limited conditions encountered in vivo. The effect of catecholamines on growth is ameliorated during culture in rich media or at high inoculum densities. Growth stimulation is unlikely to be due to provision of a metabolite, as supplementation with the NE derivative normeta-nephrine, which contains one more methyl group than NE, failed to produce the effect (Lyte et al. 1996a, 1997a, b). Remarkably, a- (but not b-) adrenergic receptor antagonists are able to block NE- and epinephrine-induced growth and dop-aminergic receptor antagonists inhibit the growth response to DA in the absence of effects on cell viability (Freestone et al. 2007a). These data imply that Gramnegative bacterial pathogens may possess elements that specifically interact with catecholamines and/or that antagonists can interfere with the ability of cate-cholamines to liberate factors required for bacterial growth.

Two major hypotheses have been put forward to explain the ability of NE to promote growth and virulence gene expression, and these are not mutually exclusive. The first posits that NE and related catecholamines facilitate the supply of iron to Gram-negative bacteria under iron-limiting conditions. As many virulence genes in Gram-negative bacteria are growth-phase regulated, this may explain downstream effects. The second hypothesis posits that virulence gene expression may result from microbial detection of catecholamines via specific receptors that initiate a signal transduction cascade on receipt of the signal leading to altered gene expression (Fig. 6.3). These hypotheses are considered in more detail subsequently.

The ability to liberate ferric iron from host storage proteins such as lactoferrin (Lf) and transferrin (Tf) is a key requirement in bacterial pathogenesis. Many Gram-negative bacteria secrete low molecular weight catecholate or hydroxamate siderophores to acquire iron, which are then imported via specific receptors. Significantly, NE, epinephrine and DA appear to facilitate iron supply from Lf and Tf, likely via the ability of the catechol moiety to complex ferric ion thereby lowering its affinity with Lf and Tf and releasing iron for siderophores to capture (Freestone et al. 2000, 2002, 2003). Recent data indicate that the formation of NE complexes with Lf and Tf leads to reduction of Fe(III) to Fe(II), for which Lf/Tf have a lower affinity (Sandrini et al. 2010). Siderophore synthesis and transport are required for NE-stimulated growth of E. coli, as strains with mutations affecting enterobactin synthesis (entA) or ferric-enterobactin transport (fepA or tonB) do not respond to NE in iron-limited serum-rich medium from a low inoculum (Burton et al. 2002; Freestone et al. 2003). This implies that NE does not act as a sidero-phore per se, although tritiated NE is taken into E. coli cells (Freestone et al. 2000;

Enterobactin Synthesis

Fig. 6.3 Catecholamine sensing and signal transduction in E. coli O157:H7. QseC is an adrenergic sensor kinase that autophosphorylates on detection of epinephrine/NE/AI-3 and transfers the phosphate moiety to its cognate response regulator QseB, thereby activating transcription of the flagellar regulon (Clarke et al. 2006). The signalling cascade downstream of QseC in E. coli O157:H7 is known in some detail and is discussed elsewhere (Hughes et al. 2009). Transcription of genes encoding a second two-component system (QseEF) is sensitive to epinephrine and QseC (Reading et al. 2007). QseEF influences AE-lesion formation via activation of genes encoding a Type III secretion system and the Tir-cytoskeleton coupling protein (Reading et al. 2007), and recent evidence indicates that it directly senses catecholamines (Reading et al. 2009). Catecholamines also mediate iron supply by an ill-defined mechanism and promote bacterial replication in serum-rich iron-limited media from low inoculum densities by a mechanism that may involve induction of a heat-stable autoinducer (AI). NE stimulates production of Shiga toxins in E. coli O157:H7 however the signal transduction cascade leading to this event is unclear at the time of writing. Adapted from Reading et al. (2007)

Fig. 6.3 Catecholamine sensing and signal transduction in E. coli O157:H7. QseC is an adrenergic sensor kinase that autophosphorylates on detection of epinephrine/NE/AI-3 and transfers the phosphate moiety to its cognate response regulator QseB, thereby activating transcription of the flagellar regulon (Clarke et al. 2006). The signalling cascade downstream of QseC in E. coli O157:H7 is known in some detail and is discussed elsewhere (Hughes et al. 2009). Transcription of genes encoding a second two-component system (QseEF) is sensitive to epinephrine and QseC (Reading et al. 2007). QseEF influences AE-lesion formation via activation of genes encoding a Type III secretion system and the Tir-cytoskeleton coupling protein (Reading et al. 2007), and recent evidence indicates that it directly senses catecholamines (Reading et al. 2009). Catecholamines also mediate iron supply by an ill-defined mechanism and promote bacterial replication in serum-rich iron-limited media from low inoculum densities by a mechanism that may involve induction of a heat-stable autoinducer (AI). NE stimulates production of Shiga toxins in E. coli O157:H7 however the signal transduction cascade leading to this event is unclear at the time of writing. Adapted from Reading et al. (2007)

Kinney et al. 2000) and the role played by NE-ferric ion complexes in growth induction and the mechanism of NE import remain unclear. NE has also been found to induce the production of the ferric-enterobactin receptor (FepA), indicating that it may facilitate iron acquisition by multiple mechanisms (Burton et al. 2002).

Iron supply may not be the only mechanism by which catecholamines induce bacterial growth. Supernatants of E. coli cultures collected after induction of growth by NE in iron-limited serum-rich medium from low inocula contain a heat-stable autoinducer of growth that stimulates replication of naive bacteria to a comparable extent as catecholamines (Lyte et al. 1996b; Freestone et al. 1999). The autoinducer is produced rapidly after exposure to NE and is able to activate growth of Gram-negative bacteria of several genera (Lyte et al. 1996b; Freestone et al. 1999). It has also been implicated in the ability of NE to resuscitate viable but non-culturable E. coli and Salmonella and appears to act in manner independent of iron-supply from Lf and Tf (Reissbrodt et al. 2002).

The ability of catecholamines to promote interactions between pathogenic E. coli and intestinal mucosa may reflect not only bacterial outgrowth, but also the induction of virulence factors. In ETEC O9:K30:H-, NE promotes expression of the K99 pilus adhesin (Lyte et al. 1997a, b). K99 pili are important in the pathogenesis of ETEC-induced diarrhoea in pigs, and such regulation may partially explain the effect of stress on the outcome of ETEC infection in pigs (Jones et al. 2001). Increased expression of Type I pili has also been described in commensal E. coli following stress induced by partial hepatectomy or short-term starvation in mice (Hendrickson et al. 1999). Such fimbriae are vital in the pathogenesis of ascending urinary tract infections by E. coli and were proposed to mediate important interactions with intestinal epithelia following catabolic stress (Hendrickson et al. 1999). However, this does not offer an explanation for the increased adherence of E. coli O157:H7 to bovine intestinal mucosa in the presence of NE (Vlisidou et al. 2004; Fig. 6.2b), as E. coli O157:H7 fail to elaborate functional Type I fimbriae owing to mutations in the fimA promoter (Roe et al. 2001) and FimH adhesin (Bouckaert et al. 2006). Indeed, expression of Type I pili is negatively selected during colonization of the bovine intestines by EHEC O26:H- (van Diemen et al. 2005).

Norepinephrine has also been reported to stimulate the production of Shiga toxins by E. coli O157:H7 during growth from a low inoculum in serum-containing medium (Lyte et al. 1996a, 1997a). Toxin production could also be stimulated by the catecholamine-induced autoinducer of growth, indicating that NE may partially act via this secreted intermediary (Voigt et al. 2006). Enhanced production of Shiga toxins is significant in the context of EHEC pathogenesis in humans, as the toxins may cause acute renal and neurological sequelae via damage to microvascular endothelial cells. Shiga toxins have also been reported to influence persistence of E. coli O157:H7 in the intestines of mice (Robinson et al. 2006) and rabbits (Ritchie et al. 2003), and to deplete a subset of intraepithelial lymphocytes in the bovine intestines (Menge et al. 2004). However, it is unlikely that this would explain the ability of NE to augment adherence and E. coli O157:H7-induced enteritis in bovine ligated ileal loops as a non-toxigenic strain was used and Shiga toxin 1 does not play a significant role in EHEC-induced enteritis in this model (Stevens et al. 2002).

A more plausible explanation for the ability of NE to stimulate adherence of E. coli O157:H7 to mucosa surfaces was afforded by the finding that it increases the production of several factors required for the formation of AE lesions (Sperandio et al. 2003; reviewed in Chap. 12). AE lesions are characterised by intimate bacterial attachment to enterocytes and localised destruction of microvilli (Fig. 6.2c), and their formation relies on a Type III protein secretion system (T3SS) encoded by the locus of enterocyte effacement (LEE). This apparatus injects a set of bacterial proteins into enterocytes, one of which (the translocated intimin receptor, Tir) becomes localised in the apical leaflet of the host cell plasma membrane where it serves as a receptor for the bacterial outer membrane protein intimin. Both intimin, Tir, and the T3SS play pivotal roles in colonization of the bovine intestines by E. coli O157:H7 (Dziva et al. 2004; van Diemen et al. 2005; Vlisidou et al. 2006). A mutant lacking both intimin and Tir did not adhere to the surface of bovine ileal loops in response to NE 12 h after loop inoculation (Vlisidou et al. 2004). However, studies in Ussing chambers have indicated that intimin and the T3SS component EspA are not required for NE-stimulated early non-intimate adherence to porcine intestinal explants (Chen et al. 2006). Production of LEE-encoded proteins is regulated by the growth phase, however induction of the expression and secretion of Type III secreted proteins, as well as induction of transcription of LEE operons, was reported to occur in the absence of effects on growth (Walters and Sperandio 2006). Epinephrine and NE also promote motility of E. coli O157:H7 (Sperandio et al. 2003), by a mechanism that appears to involve induction of the flagella regulon via the master regulators FlhDC (Clarke et al. 2006). In a bovine model, an E. coli O157:H7 flhC mutant was impaired in its ability to persist in the intestines, but a mutant lacking the flagellin subunit FliC was not (Dobbin et al. 2006), indicating that if NE acted via this circuit to promote adherence in the bovine intestines, it may have required flhDC-regulated genes other than flagella genes for the effect.

A key contribution to our understanding of host-microbe communication was the identification of an adrenergic receptor in E. coli O157:H7 (Clarke et al. 2006; reviewed in Hughes and Sperandio 2008). This arose from the finding that the qseBC genes, which regulate flagella-mediated motility (Sperandio et al. 2002), are required for the ability of epinephrine to stimulate motility (Sperandio et al. 2003). QseBC exhibit homology to two-component systems, which typically comprise a sensor kinase (SK) that autophosphorylates at a conserved histidine on receipt of a specific signal and a response regulator (RR) to which the phosphate moiety is transferred from the cognate SK. Phosphorylation of the RR alters its activity, leading to altered expression of genes under its direct or indirect control. It was subsequently proven that QseC binds tritiated NE and that it autophosphorylates in response to epinephrine and NE when reconstituted in lipid micelles (Clarke et al. 2006). Transfer of the phosphate moiety to QseB after epinephrine stimulation of QseC could be detected, and this is believed to modulate transcription of genes under the control of QseB, which include flhDC (Clarke et al. 2006). The signalling cascade downstream of QseC is now known is some detail (Hughes et al. 2009). Catecholamine binding to QseC and autophosphorylation could be blocked by the a-adrenergic receptor antagonist phentolamine, but not by the ß-adrenergic antagonist propanolol (Clarke et al. 2006), implying that QseC may structurally mimic eukaryotic adrenergic receptors. QseC also senses a bacterial autoinducer (AI-3) that provides a measure of population density (Sperandio et al. 2003). The interplay between such 'quorum sensing' systems and detection of host-derived cate-cholamines is reviewed in detail in Chap. 12.

The precise role of the QseBC system in activation of LEE genes remains to be determined. Interestingly, NE-stimulated adherence, production of LEE-encoded proteins and AE-lesion formation can be inhibited by the ß-adrenergic receptor antagonist propanolol (Chen et al. 2003; Sperandio et al. 2003), even though this treatment does not impair QseC autophosphorylation in the presence of NE (Clarke et al. 2006). This finding, taken together with the fact that some enteric pathogens respond to NE but lack a QseC homologue (e.g. Campylobacter jejuni and Yersinia enterocolitica), may reflect the presence of other bacterial catecholamine receptors or alternative modes of action. In the case of E. coli O157:H7, it has recently been reported that a second two-component sensory system (QseEF), that is regulated by QseBC and required for AE-lesion formation (Reading et al. 2007), acts as a secondary receptor for epinephrine (Reading et al. 2009). It has been established that QseC is required for the induction of enteritis by rabbit enteropathogenic E. coli in infant rabbits (Clarke et al. 2006). However, mutations in two-component systems often have pleiotropic effects, and it remains to be shown that QseC is required for NE-induced phenotypes of E. coli O157:H7 in vivo. A specific inhibitor of QseC signalling (LED209) has recently been reported to interfere with actin nucleation by E. coli O157:H7 and the transcription of flhDC and stx2A; however, studies of its ability to control infection in infant rabbits were complicated by rapid adsorption of the inhibitor from the gut (Rasko et al. 2008).

In an attempt to define the global transcriptional response of enteric pathogens to catecholamines, microarray studies have been undertaken by several laboratories. Bansal et al. defined the transcriptome of E. coli O157:H7 during biofilm formation on glass wool during 7 h culture in Luria Bertani (LB)-glucose medium supplemented with 50 mM epinephrine or 50 mM NE. A total of 938 and 970 genes were differentially transcribed in response to epinephrine and NE, respectively, with 411 genes exhibiting the same pattern of transcription in response to both catecholamines (Bansal et al. 2007). The fact that epinephrine and NE exert different effects at the level of transcription is consistent with the observation that they do not stimulate the growth of E. coli O157:H7 to the same extent (Freestone et al. 2007a, b). Several genes involved in iron acquisition were upregulated in response to epinephrine and NE, however Shiga toxin and LEE genes were not found to be activated under the conditions used (Bansal et al. 2007), in contrast with observations at the protein level. Following growth in serum-SAPI medium in the presence of 50 mM NE from an inoculum of c. 106 CFU, Dowd detected differential regulation of 101 genes, including induction of genes involved in iron acquisition and the genes for intimin, T3SS components and Shiga toxins (Dowd 2007). Such changes mirror those seen at the protein level (Lyte et al. 1996a, 1997a; Voigt et al. 2006; Sperandio et al. 2003); however, other genes in the same operon as eae and espAB were not observed to be differentially regulated. The fepA gene was found to be repressed, in contrast with other studies (Burton et al. 2002), whereas induction of feoB and fhuD was also reported by Bansal et al. Culture of an E. coli O157:H7 luxS mutant (unable to synthesize autoinducer) in Dulbecco's modified Eagle's medium with 50 mM epinephrine has also been reported to induce transcription of LEE, stx2 and flagella genes (Kendall et al. 2007). A key consideration with studies of this nature is the effect of exogenous catecholamines on bacterial growth. If control and treated cultures are collected at different points in the growth cycle, it is difficult to interpret whether altered virulence gene expression is a direct consequence of catecholamine-mediated regulation (e.g. via QseBC and QseEF) or associated with the change of growth phase. A further consideration is that at high concentrations catecholamines may sequester iron leading to indirect activation of iron acquisition genes. Such aspects are reviewed in detail elsewhere (Freestone et al. 2008).

Was this article helpful?

+1 0

Responses

  • Lily Bolger
    How to act e.coli on ocp?
    3 years ago

Post a comment