The phosphate regulon and bacterial virulence: a regulatory network connecting phosphate homeostasis and pathogenesis


  • Martin G. Lamarche,

    1. Groupe de Recherche sur les Maladies Infectieuses du Porc (GREMIP), Faculté de médecine vétérinaire, Université de Montréal, Saint-Hyacinthe, Québec, Canada
    Search for more papers by this author
  • Barry L. Wanner,

    1. Department of Biological Sciences, Purdue University, West Lafayette, Indiana, USA
    Search for more papers by this author
  • Sébastien Crépin,

    1. Groupe de Recherche sur les Maladies Infectieuses du Porc (GREMIP), Faculté de médecine vétérinaire, Université de Montréal, Saint-Hyacinthe, Québec, Canada
    Search for more papers by this author
  • Josée Harel

    1. Groupe de Recherche sur les Maladies Infectieuses du Porc (GREMIP), Faculté de médecine vétérinaire, Université de Montréal, Saint-Hyacinthe, Québec, Canada
    Search for more papers by this author

  • Editor: Diego de Mendoza

Correspondence: Josée Harel, Groupe de Recherche sur les Maladies Infectieuses du Porc (GREMIP), Faculté de médecine vétérinaire, Université de Montréal, C. P. 5000, Saint-Hyacinthe, Québec, Canada J2S 7C6. Tel.: +450 773 8521, ext 8233; fax: +450 778 8108; e-mail:


Bacterial pathogens regulate virulence factor gene expression coordinately in response to environmental stimuli, including nutrient starvation. The phosphate (Pho) regulon plays a key role in phosphate homeostasis. It is controlled by the PhoR/PhoB two-component regulatory system. PhoR is an integral membrane signaling histidine kinase that, through an interaction with the ABC-type phosphate-specific transport (Pst) system and a protein called PhoU, somehow senses environmental inorganic phosphate (Pi) levels. Under conditions of Pi limitation (or in the absence of a Pst component or PhoU), PhoR activates its partner response regulator PhoB by phosphorylation, which, in turn, up- or down-regulates target genes. Single-cell profiling of PhoB activation has shown recently that Pho regulon gene expression exhibits a stochastic, ‘all-or-none,’ behavior. Recent studies have also shown that the Pho regulon plays a role in the virulence of several bacteria. Here, we present a comprehensive overview of the role of the Pho regulon in bacterial virulence. The Pho regulon is clearly not a simple regulatory circuit for controlling phosphate homeostasis; it is part of a complex network important for both bacterial virulence and stress response.


Phosphate is an essential nutrient for cell functions and life. It is found in lipids, nucleic acids, proteins and sugars, and is involved in many biochemical reactions that rely on transfer of phosphoryl groups. The Pho regulon is a global regulatory circuit involved in bacterial phosphate management (Wanner, 1993, 1996). It has been best studied in model cells like Escherichia coli and Bacillus subtilis. In E. coli, the Pho regulon is comprised of at least 47 genes (Wanner, 1996; Han et al., 1999; Suziedeliene et al., 1999; Harris et al., 2001; Baek & Lee, 2006). Recently, additional genes with putative Pho operator containing Pho boxes have been identified (Yuan et al., 2006b). Moreover, proteome profiles of cells grown under phosphate-rich or phosphate-limited conditions revealed that the overall phosphate response of E. coli could be constituted of up to 400 genes (VanBogelen et al., 1996). Such a high number, which represents almost 10% of the E. coli genome, reflects the importance of phosphate in cellular processes. It seems likely that such a network influences the virulence of bacterial pathogens. Indeed, as we discuss in this review, it is now clear that the Pho regulon influences bacterial virulence and, in some bacteria, directly controls virulence gene expression. We will first present a brief overview of the role of the Pho regulon. We will then describe strategies involving the Pho regulon that bacteria have evolved to survive adverse conditions. Finally, we will focus on the evidence that links the Pho regulon to bacterial virulence. The role played by the Pho regulon in stress responses explains how the Pho regulon can be involved both in phosphate homeostasis and in pathogenesis. To illustrate this, we will present bacterial mechanisms involved in tolerance to acidity and biofilm formation, both of which are responses to harsh environmental conditions but have also been related to the Pho regulon and virulence. Also, we will discuss how the stochastic properties of Pho regulon gene expression can have a major impact on bacterial pathogenesis.

Control of the Pho regulon

The PhoR/PhoB two-component regulatory system (TCRS)

In bacteria, many external conditions are sensed by TCRSs responsible for signal transduction and gene regulation. The TCRSs are important for the adaptation, survival and the pathogenic responses of bacteria (Hoch, 2000; West & Stock, 2001; Beier & Gross, 2006). The Pho regulon is controlled by the PhoR/PhoB TCRS. PhoR is an inner-membrane histidine kinase (HK) sensor protein that appears to respond to periplasmic orthophosphate (Pi) concentration variations through interaction with a phosphate transport system. In some microorganisms, e.g., Saccharomyces cerevisiae, intracellular phosphate also plays a role in the control of the Pho regulon (Auesukaree et al., 2004). Even though intracellular phosphate is likely to influence Pho regulon activity, this phenomenon has not yet been demonstrated in bacteria. PhoB is a response regulator (RR) that acts as a DNA-binding protein to activate or inhibit gene transcription (Smith & Payne, 1992; Wanner, 1996; Harris et al., 2001; Blanco et al., 2002). The activation signal, a phosphate concentration below 4 μM, is transmitted by a phospho-relay between the two partner proteins, from PhoR to PhoB (Fig. 1). PhoR is auto-phosphorylated on a histidine residue and transfers this phosphoryl group to an aspartate residue of the PhoB regulator through a transitory interaction. Phospho-PhoB, in turn, controls Pho regulon gene expression.

Figure 1.

 Control of the Pho regulon and transmembrane signal transduction by environmental inorganic phosphate (Pi). Adapted from Wanner (1997) (Fig. 5.6, p. 116), with kind permission of Springer Science and Business Media. Small squares mark a Pi-binding site on PstS and a hypothetical regulatory site on PhoR. Notation: PhoB, RR; PhoR, sensor protein (A, autophosphorylated; R, repression); PstA and PstC, integral membrane channel proteins; PstB, traffic ATPase; PstS, periplasmic Pi-binding protein; PhoU, regulatory protein. See text for details. This model is adopted from one that was originally illustrated in Wanner (1990) and modified in Wanner (1995, 1996).

The phosphate-specific transport (Pst) system and inhibition of the Pho regulon

The Pst system belongs to the Pho regulon. Moreover, the Pst system is linked to the molecular mechanisms that lead to turning off the Pho regulon. Four proteins (PstS, PstC, PstA and PstB) form an ABC transporter important for the high-affinity capture of periplasmic inorganic phosphate (Pi) and its low-velocity transport into the cytosol (Van Dien & Keasling, 1998). These proteins are encoded together with a protein called PhoU in the polycistronic pstSCAB–phoU operon. PstS is a periplasmic protein that binds Pi with high affinity. PstC and PstA are inner-membrane channel proteins for Pi entry, while PstB is an ATP-dependent permease component that releases the energy necessary for Pi transport from the periplasm to the cytosol. When phosphate is in excess, the Pst system is thought to form a repression (inhibition) complex with PhoR that prevents activation of PhoB. Phosphate limitation leads to conversion of PhoR into an activated form in which PhoR is autophosphorylated and transfers a phosphoryl group to PhoB. A return to the repression state occurs during growth shift to a phosphate-rich environment; the activation signal is interrupted by PhoR acting as a phosphatase on phospho-PhoB. PhoU and PstB are also required for dephosphorylation of phospho-PhoB (Fig. 1) (Wanner, 1997). Indeed, PhoU is essential for the repression of the Pho regulon under a high-phosphate condition (Wanner, 1996). While the precise mechanism of how PhoU acts is not yet understood, it is reasonable to suggest that PhoU acts by binding to PhoR, PhoB or a PhoR/PhoB complex as a chaperone to promote dephosphorylation of phospho-PhoB or by inhibiting formation of the PhoR–PhoB transitory complex (Oganesyan et al., 2005). In many bacterial species, mutations in the Pst system result in constitutive expression of the Pho regulon, regardless of environmental phosphate availability (Wanner, 1996).


Many environmental conditions can modulate the activity of a RR by a nonpartner HK via a mechanism known as cross-regulation. PhoB activity can be induced in the absence of signal transduction from its cognate partner protein PhoR. For example, CreC, a sensor protein of the CreC/CreB TCRS implicated in the regulation of genes involved in carbon catabolism, can induce PhoB activity in the absence of PhoR (Wanner & Wilmes-Riesenberg, 1992). The phosphotransacetylase-acetate kinase (Pta-AckA) pathway, belonging to the Cre regulon (Avison et al., 2001), generates an intermediary product named acetyl phosphate that can also cause activation of PhoB (Kim et al., 1996; Wanner et al., 1996). Curiously, genes of the Pta-AckA pathway belong to the Pho regulon in the symbiotic bacteria Sinorhizobium meliloti (Summers et al., 1999). These are only two examples from a variety of situations known to result in modification of the phosphorylation status of PhoB in the absence of PhoR (Wanner et al., 1988; Wanner, 1992, 1995, 1996; Wanner & Wilmes-Riesenberg, 1992; Wilmes-Riesenberg & Wanner, 1992; Fisher et al., 1995; Haldimann et al., 1996; Kim et al., 1996; Verhamme et al., 2002; Nishino et al., 2005; Zhou et al., 2005). Nevertheless, because of cross-regulation, particular TCRS responding to a variety of environmental signals may influence expression of genes belonging to the Pho regulon. For example, cross-regulation (phosphorylation) of PhoB by a nonpartner HK may result in normal cells when the amount of PhoR protein is limiting (Wanner & Wilmes-Riesenberg, 1992). In this regard, cell-to-cell variations in the amount of an HK-like PhoR can lead to cells with no PhoR protein stochastically because the amount of PhoR normally constitutes only 10 or so molecules per cell (Wanner, 1996). Such mechanisms are likely to contribute to adaptation to changing environments. In B. subtilis, however, whose Pho regulon genes are controlled by three known TCRS (Sun et al., 1996), the induction of Pho is abolished in the absence of the HK PhoR. This is an illustration of a multiple-signal input system required for the control of a precise set of genes.

Stochastic activation of the RR PhoB

Single-cell profiling of PhoB activation has recently revealed an unforeseen stochastic, ‘all-or-none,’ character of Pho regulon gene expression (Zhou et al., 2005). Such a variation in gene expression, also referred as noise, occurs within a population of genetically identical bacterial cells (Raser & O'Shea, 2005). Recent studies showed that at least six noncognate HKs (QseC, ArcB, CreC, KdpD, BaeS and EnvZ) can activate the RR PhoB, in the absence of PhoR and acetyl phosphate biosynthesis. In all cases, PhoB activation results in a bimodal distribution of cells expressing the Pho regulon (Zhou et al., 2005). In other words, in the presence of a PhoB activation signal, a proportion of the cell population remains uninduced, showing a threshold instead of a graded activation response. What factor or combination of factors determines the activation threshold is unknown. However, it is likely that components of the phosphorylation cascade, such as the amount of HK, phospho-HK, RR and phospho-RR as well as other factors, contribute to the threshold phenomenon (Zhou et al., 2005). Moreover, the autoregulation of PhoR/PhoB synthesis, forming a positive feedback loop, is likely to contribute to the stochastic behavior in gene expression and to the emergence of multiple stable phenotypes within a population of genetically identical cells (Smits et al., 2006). It also contributes to the propagation and autoamplification of the noise to downstream genes (Pedraza & van Oudenaarden, 2005). This kind of ‘noise’ was shown to have great consequences on the phenotypic diversity within a population of cells (Elowitz et al., 2002; Ozbudak et al., 2002). Such a diversity contributes to the fitness of cell subpopulations in changing environments and is likely to have a huge impact on adaptation of bacterial pathogens within a host (Balaban et al., 2004; Kaern et al., 2005). Importantly, stochastic activation of gene expression was also found to occur upon activation of PhoB by PhoR in response to Pi limitation (L. Zhou and B.L.Wanner, unpublished data), suggesting that stochastic activation of Pho regulon gene expression occurs in normal cells. Consequently, stochastic effects of this kind must be taken into account as part of the equation connecting bacterial pathogenesis and the Pho regulon, as discussed below.

Phosphate and the adaptation/survival response

Adaptation of bacterial cell surface components to phosphate deprivation

To deal with phosphate starvation conditions, bacteria rely on different mechanisms to optimize the acquisition and the bioavailability of phosphate and to maintain essential biochemical reactions. For instance, some bacteria change their shape to enhance the surface to volume ratio, coming more in contact with the surrounding environment and enhancing chances to capture essential nutrients including phosphate. This is exemplified by the stalk elongation of Caulobacter crescentus, which is regulated by a phosphate-limiting environment in a Pho regulon-dependent manner (Gonin et al., 2000; Wagner et al., 2006). Under phosphate starvation conditions, many bacteria modify their cell surface components. In B. subtilis, the replacement of the cell wall phosphate-rich teichoic acid with the phosphate-free teichuronic acid occurs in phosphate-limiting environments. In these bacteria, the teichoic acid synthesis genes tagA and tagD and the teichuronic acid synthesis operon tua are under the control of the B. subtilis PhoR/PhoP TCRS (Pragai et al., 2004). In S. meliloti mutants of phosphate uptake system PhoCDET, which is homologous to PstSCAB, phosphate-free lipids substitute for phospholipids (Geiger et al., 1999). These mutants also failed to fix nitrogen and form smaller white nodules on alfalfa. These phenotypes are due to the effects of PhoB on gene expression because a mutation in phoB restores the wild-type phenotype and the Pho regulon is constitutively activated in phoCDET mutants. These observations strongly suggest an involvement of the Pho regulon in the control of genes that are part of the adaptation and survival response (Bardin et al., 1996; Yuan et al., 2006a). Moreover, in S. meliloti, PhoB positively regulates the synthesis of some exopolyssacharides, which are involved in the bacterial invasion of root nodules, when phosphate is scarce in the environment (Ruberg et al., 1999; Mendrygal & Gonzalez, 2000).

Environmental stresses, stringent response and the Pho regulon

In bacteria, metabolism changes that allow cells to survive occur during environmental stress conditions (Weber et al., 2005; Jain et al., 2006; Klauck et al., 2007). Many lines of evidence suggest that the Pho regulon and the stress response are interrelated (Spira et al., 1995; Spira & Yagil, 1998; Ruiz & Silhavy, 2003; Taschner et al., 2004, 2006; Schurdell et al., 2007). Figure 2 illustrates interconnections between the Pho regulon and some key factors involved in nutritional and other stress responses. The role of these factors is briefly discussed in this section. Polyphosphate (poly P) is a known stress response molecule synthesized when environmental stresses occur, e.g., nutrient starvation and high salt concentrations, among others (Kornberg et al., 1999). Cells having an imbalance in their poly P reserves have been shown to elicit a defective response to adverse environmental conditions (Brown & Kornberg, 2004). Poly P can serve both as a phosphate reservoir and as an energy source in different biological processes. For instance, poly P can substitute for ATP in phosphorylation reactions by kinases. Poly P also acts as a metal-chelating agent or as a buffer in alkali environments (Kornberg et al., 1999). In some bacterial species, such as Klebsiella aerogenes and Streptomyces lividans, the poly P kinase gene ppk, which is involved in the poly P synthesis, has been reported to be under the control of the PhoR/PhoB TCRS (Kato et al., 1993; Ghorbel et al., 2006). Similarly, PhoB has been shown to be required for poly P accumulation in E. coli (Fig. 2) (Rao & Kornberg, 1999).

Figure 2.

 Scheme illustrating interactions between the Pho regulon, Poly P biosynthesis pathway, RpoS, and the stringent response major factors. Dashed lines represent interactions that could be either direct or indirect. aIntergenic region between the pstA and pstB genes (iAB) of the pst operon-processed mRNA (Ruiz & Silhavy, 2003; Schurdell et al., 2007). bRpoS contributes in turning off the Pho regulon by directly stimulating, in an IHF-dependent manner, the expression of the pst operon (Spira & Yagil, 1999; Taschner et al., 2004, 2006). cBasal level of cellular ppGpp is essential for proper Pho regulon expression. Moreover, induction of the Pho regulon is observed concomitantly with ppGpp synthesis, until a certain threshold is reached; i.e. a ppGpp intracellular concentration that induces inhibition of protein synthesis (Spira et al., 1995; Spira & Yagil, 1998). d,ePhosphate limitation induces SpoT-dependent ppGpp accumulation (Spira et al., 1995; Spira & Yagil, 1998). This promotes transcription of an iraP gene coding for the antiadaptor IraP that interferes with the delivery of RpoS to the ClpXP protease by blocking the action of RssB, an adaptor protein for RpoS degradation (Bougdour et al., 2006; Bougdour & Gottesman, 2007). Under a phosphate-limiting condition, a phoB mutant accumulates twofold less iraP transcript, but still, RpoS is stabilized in an IraP-dependent manner (Bougdour et al., 2006; Bougdour & Gottesman, 2007). Information about the Pho regulon, RpoS, Poly P and ppGpp is given in the text.

Guanosine tetraphosphate (ppGpp) is a key phosphate-containing molecule that initiates changes in global metabolism under nutritional stress, which is known as the stringent stress response (Chatterji & Ojha, 2001; Magnusson et al., 2005; Paul et al., 2005). The ppGpp nucleotide induces RpoS accumulation (Hengge-Aronis, 2002). RpoS is a sigma factor implicated in the cellular response to many stresses such as osmotic, oxidative and acid stresses. It is also implicated in the stationary phase and the induction of genes in a nutrient-limiting environment (Loewen et al., 1998). Repression of the Pho regulon occurs in E. coli mutants that cannot accumulate ppGpp, such as relA and spoT mutants, whereas a mutation in Pst down-regulates ppGpp synthesis (Fig. 2) (Spira et al., 1995; Spira & Yagil, 1998). This suggests a regulatory link between ppGpp and the Pho regulon. The presence of an inverted Pho box in the -10 region of the spoT promoter has led to the suggestion that SpoT activity could depend on the Pho regulation (Sarubbi et al., 1989; Spira & Yagil, 1998). Although direct evidence is lacking, other results link the Pho regulon to the stringent stress response. For example, Pst mutants accumulate RpoS when grown to the exponential phase. This phenomenon is dependent on Hfq, a posttranscriptional regulator that interacts with various RNAs, and requires intact PhoB (Fig. 2) (Ruiz & Silhavy, 2003). Furthermore, Schurdell et al. (2007) have shown recently that the 3′ end of the processed pstCA transcript is likely to mediate RpoS accumulation through a direct interaction with the 5′- untranslated leader region of rpoS, stimulating translation of rpoS (Fig. 2). Bougdour & Gottesman (2007) have shown recently that ppGpp accumulation during phosphate starvation promotes transcription of the iraP gene coding for a small protein that interferes with RssB-dependent degradation of RpoS (Bougdour et al., 2006; Bougdour & Gottesman, 2007). Thus, the Pho regulon influences the production of poly P, ppGpp and RpoS, all of which are necessary for proper adaptation and survival response in bacteria. Accordingly, changes in the levels of a number of effector molecules may occur in E. coli phoBR and pst mutants, resulting in an aberrant stress response.

Pho regulation and virulence

A correlation between inactivation of the Pst system, constitutive expression of the Pho regulon and bacterial virulence has been established in a number of cases. Table 1 lists some important studies in which pst, phoBR and Pho regulon genes have been shown to be important in bacterial virulence from different approaches.

Table 1.   Identification of in vivo expressed PhoB-dependent genes using infection models
SpeciesInfection models*SelectionReferences
  • *

    Animal strains or cell lines used in the study are given in parentheses.

  • RTI, respiratory tract infection; DFI, differential fluorescence induction; IVET, in vivo expression technology; SCOTS, selective capture of transcribed sequences.

Brucella melitensisMouse (BALB/c) septicemiaSTMLestrate et al. (2000)
Burkholderia cenocepaciaRat (Sprague–Dawley) chronic lungSTMHunt et al. (2004)
Campylobacter jejuniChicken (SPF Light Sussex) intestinalSTMGrant et al. (2005)
Human cells (INT407)MicroarrayGaynor et al. (2005)
Corynebacterium pseudotuberculosisMouse macrophage-like cells (J774A.1)DFIMcKean et al. (2005)
Escherichia coliMouse (CBA) UTISTMBahrani-Mougeot et al. (2002)
Mouse septicemia (ICR)IVETKhan & Isaacson (2002)
Chicken (Leghorn SPF) extraintestinalSCOTSDozois et al. (2003)
Erwinia amylovoraImmature pears necrotic diseaseIVETZhao et al. (2005)
Erwinia chrysanthemiSpinach leaf diseaseIVETYang et al. (2004)
Listeria monocytogenesMouse (ICR Swiss) septicemiaIVETDubail et al. (2000)
Murine macrophage cells (P388D1)MicroarrayChatterjee et al. (2006)
Klebsiella pneumoniaeMouse (C57Bl/6j) intranasalSTMLawlor et al. (2005)
Mycobacterium tuberculosisMouse (BALB/c and BALB/cSCID/SCID) LungMicroarrayTalaat et al. (2004)
Proteus mirabilisMouse (CBA) UTISTMBurall et al. (2004)
Salmonella TyphimuriumMouse macrophage-like cells (RAW 264.7)DFIValdivia & Falkow (1997)
Mouse macrophage-like cells (J774-A.1)MicroarrayEriksson et al. (2003)
Salmonella TyphiHuman monocyte cells (THP-1)SCOTSFaucher et al. (2006)
Shigella flexneriHuman intestinal cells (INT407)DFIRunyen-Janecky & Payne (2002)
Streptococcus pneumoniaeMouse (BALB/c) pneumonia and septicemiaSTMPolissi et al. (1998)
Mouse (Swiss Webster)
bacteremia and nasopharyngeal carriage
STMHava & Camilli (2002)
Mouse (CD-1) RTIDFIMarra et al. (2002)
Vibrio choleraeMouse (CD-1) sucklingSTMMerrell et al. (2002)
Yersinia enterocoliticaMouse (BALB/c) intestinal and septicaemiaSTMDarwin & Miller (1999)
Yersinia pestisMouse (C57BL/6) RTIMicroarrayLathem et al. (2005)

Pst mutants and virulence

Several studies link the Pst system with virulence. A Pst mutant of porcine extraintestinal pathogenic E. coli ExPEC strain was shown to be avirulent in a colostrum-deprived newborn pig infection model. Insertional inactivation of pst has multiple effects on virulence traits such as an increased sensitivity to the bactericidal effect of serum and a significant reduction in the amount of capsular antigen at the cell surface (Harel et al., 1992; Ngeleka et al., 1992; Daigle et al., 1995). More recently, deletion of the pst genes in an ExPEC strain belonging to the avian pathogenic E. coli APEC group (Lamarche et al., 2005) was shown to reduce virulence in a chicken infection model. The pst mutation affects multiple virulence attributes; it reduces resistance to the bactericidal effect of serum, to acidity and to cationic antimicrobial peptides (Lamarche et al., 2005). These results are suggestive, in part, of bacterial cell surface modifications. In the uropathogenic E. coli UPEC strain, Bahrani-Mougeot et al. (2002) identified a phoU mutant by signature tag mutagenesis (STM), an approach that allows finding genes that are important for infection, using a murine model of ascending urinary tract infection (UTI) (Bahrani-Mougeot et al., 2002). The ability of the phoU mutant to colonize the murine urinary tract was impaired in a competition infection assay with the wild-type strain (Buckles et al., 2006). A pst mutant was also identified by STM in Proteus mirabilis using a UTI model (Burall et al., 2004). Lastly, transposon mutants of pst in Edwardsiella tarda, a pathogen known to cause extraintestinal diseases in fish and mammals, were shown to exhibit a reduced capacity to multiply within phagocytes, a lower replication rate in serum and an overall reduced virulence (Srinivasa Rao et al., 2003). Proteome comparative studies of E. tarda wild-type strain and its pst mutant revealed that translocon proteins of type III secretion system (T3SS) as well as other virulence proteins were absent in the pst mutant (Rao et al., 2004). The Pst system also influences other virulence mechanisms that are not limited to extraintestinal diseases. For example, a pst mutant of an intestinal E. coli pathogenic strain was shown to cause significantly less attaching and effacing lesions in a pig ileal explant model (Batisson et al., 2003). In Shigella flexneri, a pst mutation reduced the capacity to form plaques on Henle cell monolayers in a Pho regulon-dependent manner. The capacity to form plaques is a characteristic associated with Shigella virulence (Runyen-Janecky et al., 2005). In contrast, an entero-invasive E. coli pst mutant showed hyperinvasiveness of Hep-2 cells (Sinai & Bavoil, 1993). Lastly, as observed with E. tarda, the PhoR/PhoB TCRS is involved in the expression of T3SS-related genes in Salmonella enterica serovar Typhimurium (Lucas et al., 2000; Baxter & Jones, 2005). Those studies indicate that the Pst system can function as a regulator of bacterial pathogenicity.

PhoB/PhoR inactivation and virulence

As the Pst mutations result in a constitutive activation of the Pho regulon, it is likely that their effects on virulence are mediated through the PhoR/PhoB TCRS. Therefore, inactivation of phoBR should also have an effect on virulence. Indeed, this was shown to be the case. It has, for instance, been shown that ChvI, a PhoB homolog, is essential for expression of full virulence of Agrobacterium tumefaciens (Mantis & Winans, 1993). In comparative microarray analyses of the responses of Corynebacterium glutamicum and its isogenic phoRS (phoBR homolog) mutant after a shift to Pi-limiting conditions, virulence-related genes such as siderophore systems are expressed differentially (Kocan et al., 2006). Many gene products that may be associated with virulence are overexpressed in the Vibrio cholerae O1 wild-type strain compared with its isogenic phoB mutant, e.g., hemolysin (von Kruger et al., 1999, 2006). Furthermore, a V. cholerae phoB mutant showed a decreased capacity for colonization of adult rabbit ligated ileal loops in competitive colonization assays with the wild-type strain. This result is somewhat surprising because the small intestine has a high phosphate concentration and the Pho regulon would be expected to be repressed in wild-type strains under those conditions. This raises the possibility that stimuli encountered in the small intestine, other than the phosphate signal, may be responsible for those observations. Lastly, phosphate has been demonstrated recently to be a major signal for sporulation of Clostridium perfringens and for C. perfringens endotoxin production (Philippe et al., 2006). Taken together, these studies reinforce the notion that PhoB may act not only as a RR for phosphate homeostasis but also as a modulator of virulence attributes.

How can the Pho regulon influence pathogenesis?

The information summarized above illustrates that complex networking events occur when control of the Pho regulon is altered. Orchestrated regulation of the Pho regulon seems to occur during different steps of bacterial pathogenesis. Furthermore, some studies suggest that the PhoR/PhoB TCRS responds to stimuli other than phosphate and that a subset of Pho regulon genes is activated or repressed by other transcriptional regulators. The next sections shed light on strategies implicating the Pho regulon that allow survival under adverse conditions encountered within hosts as well as in the environment. This can be illustrated by responses of the Pho regulon to pH variation and its role in biofilm formation. Moreover, as reported above, the Pho regulon has regulatory links with cell surface components such as lipids and exopolysaccharides, which may have important consequences on bacterial virulence.

Pho regulon and response to acid stress

The Pho regulon also includes genes that do not appear to be directly implicated in the cell phosphate management, including one related to acid stress, which bacteria must overcome to maintain homeostasis. For example, the acid-inducible asr gene is under the transcriptional control of the PhoR/PhoB TCRS in E. coli. Asr is thought to play a role similar to that of the E. coli periplasmic protein HdeA, which serves as a proton sink or a chaperone for protecting periplasmic proteins from the deleterious effects of low pH (Gajiwala & Burley, 2000). As another example, the PhoR/PhoB TCRS has been suggested to sense external acidity and to regulate the transcription of genes that are important for acid shock resistance (Suziedeliene et al., 1999; Tucker et al., 2002; Seputiene et al., 2003, 2004, 2006). In contrast, a pH shift from 4.5 to 5.8 increases pst transcription in Clostridium acetobutyliculum (Fischer et al., 2006). However, other members of the Pho regulon, such as phoA, are not induced under alkaline conditions, as demonstrated using a lacZ fusion and microarray experiments (Novak et al., 1999; Atalla & Schumann, 2003). This can be explained by a differential expression of the Pho regulon genes, depending on the nature of the signal. In light of the observations in these studies, the PhoR/PhoB TCRS is likely to be a part of a circuit involved in the response to acid stresses. During their life cycle, many bacterial pathogens face acid stress. Enteric pathogens must survive stomach acidity while invasive bacteria must be able to survive an endosomal-acidified environment. Thus, mutations in the Pst system or phoB can reduce virulence by affecting the response to acid stress.

Biofilms and Pho regulon

In natural environments, most microorganisms have a sessile life-style and live in complex structures called biofilms, which are mixed microbial populations typically embedded in a matrix of extracellular polymers. In pathogenesis, biofilms play many important roles. Biofilms lead to resistance to environmental stresses such as exposure to biocides and antimicrobial agents and to an increase in the concentration of virulence factors in a precise environment such as in a host tissue. Also, biofilms enhance the genetic exchange of virulence elements. Additionally, biofilms serve as a reservoir that favors the preservation and the spreading of bacterial pathogens (Davey & O'Toole, 2000). Some studies have shown a relationship between the Pho regulon and biofilm formation. For example, a Pst mutation abolishes biofilm formation in two rhizosphere-colonizing bacteria: Pseudomonas aerofaciens and Pseudomonas fluorescens (Monds et al., 2001). Recently, Monds et al. (2007) demonstrated that in P. fluorescens, RapA has a c-di-GMP phosphodiesterase activity mostly responsible for the inhibition of biofilm formation in Pst mutants. This in turn inhibits the secretion of LapA, a large adhesin required for biofilm formation by P. fluorescens. Indeed, Pho modulation of the c-di-GMP cellular level, via RapA activity, was shown to be linked to LapA adhesin-decreased secretion as well as its increased dissociation from the bacterial cell surface (Monds et al., 2007). In the plant pathogen A. tumefaciens, PhoR/PhoB TCRS was shown to induce the production of larger biofilms under phosphate-depleted conditions (Danhorn et al., 2004). Moreover, in E. coli, the Pho regulon was shown to be linked to quorum sensing (Ren et al., 2004), a key regulator mechanism for the formation of biofilms (Parsek & Greenberg, 2005). PhoR/PhoB TCRS may therefore play a role in the virulence of bacteria via regulatory links with biofilm formation and quorum-sensing circuits.

Bacterial cell surface components and the Pho regulon

The Pho regulon is involved in the regulation of some cell surface component modifications, i.e. teichuronic acid, phosphate-free lipids, phospholipids, exopolysaccharides and adhesin. Moreover, we have found that lipid A and fatty acid composition is modified in Pst E. coli mutants (Lamarche et al., unpublished data). Interestingly, Pst mutants showed a reduced resistance to different host defense elements that target bacterial cell surface components. For instance, pst mutations reduce resistance to serum, acid and cationic antimicrobial peptides (Daigle et al., 1995; Lamarche et al., 2005). Lipopolysaccharides, outer-membrane proteins and membrane lipids are known to be involved in the resistance to such host defense mechanisms. Thus, Pho-regulated modifications of these cell surface components may explain, at least in part, virulence attenuation of Pst mutants. Furthermore, as described above, many lines of evidence link the Pho regulon to the bacterial stress response. Indeed, because the cellular status of stress response molecules like poly P and ppGpp may be controlled by the PhoR/PhoB TCRS, virulence attributes, including some bacterial cell surface components, acid stress response and biofilm formation, must also require an intact Pho regulon (Chang & Cronan, 1999; Hammer & Swanson, 1999; Kornberg et al., 1999; Primm et al., 2000; Rashid & Kornberg, 2000; Rashid et al., 2000; Taylor et al., 2002; Haralalka et al., 2003; Pizarro-Cerda & Tedin, 2004; Gaynor et al., 2005; Jain et al., 2006).

Bacterial fitness and the Pho regulon

The capacity of the bacterial cell to adapt to different environments is often associated with its ability to occupy different ecological niches including the animal host. For instance, expression of the Pho regulon could be advantageous to adapt to some conditions encountered at some stage of the infectious process. PhoR/PhoB TCRS and Pho regulon are well documented as phosphate homeostasis circuitry but activation of PhoB by nonpartner HKs shows that cross-regulation leads to heterogeneity of the cell population (Zhou et al., 2005). In addition to specialized bacterial-sensing mechanisms that perceive environmental cues, phenotypic diversity within genetically identical cell population contributes to the presence of primed cells allowing a rapid response to such cues, increasing the chance to survive new stresses. In this regard, it is tempting to speculate that the stochastic behavior of the Pho regulon characterized by a bimodal, ‘all-or-none,’ pattern for gene expression can participate in an increased fitness in microenvironments encountered within human and animal hosts. Noise in gene expression and positive feedback regulation of the Pho regulon are likely to contribute to the emergence of phenotypes that would be favoured during pathogenesis as stochastic expression of virulence factor genes per se is observed in many pathogens (Zhou et al., 2005; Smits et al., 2006).

Concluding remarks and perspectives

As illustrated in this review, the Pho regulon controls and influences the expression of virulence traits in many bacterial species in addition to its role in phosphate homeostasis. Mutations in the Pst system or PhoB were shown to result in multiple effects, including reduction of virulence. The effects of pst mutations can be due to a deficiency in phosphate uptake under some conditions but, in all cases, PhoR/PhoB TCRS regulation is likely to generate the multiple effects that are observed in the Pst and PhoB mutants. Moreover, the regulation of cell surface components by the Pho regulon may contribute to bacterial virulence. Additionally, a strong network of links exists between the Pho regulon and some stress responses. These phenomena have been observed in many bacterial species. Information presented in this review reflects the need to redefine our vision concerning the role of the bacterial Pho regulon. Instead of being the simple regulatory circuit controlling phosphate homeostasis, the Pho regulon should be considered to be a part of a complex network important for both bacterial virulence and stress response. Moreover, expression of Pho regulon genes is induced under particular conditions regardless of the bioavailability of Pi. This can be explained, in part, by stochastic variation in Pho-dependent gene expression. However, it could also indicate that other signals are perceived by PhoR/PhoB TCRS or that other TCRS regulate genes belonging to the Pho regulon.

The interplay between the Pho regulon (including the Pst system) and virulence often has an unknown molecular basis, warranting the need for more comprehensive studies. High-throughput genomic and proteomic analyses are providing some clues to the relationship between the Pho regulon and virulence (Table 1, S. Crépin et al., unpublished, von Kruger et al., 2006). As mentioned previously, several lines of evidence suggest that some PhoR/PhoB-regulated genes are expressed within the host. However, little is known about the signals regulating the Pho regulon in vivo. The spatial and temporal patterns of Pho-dependent gene expression as well as the implication of nonpartner HK in their regulation, within microenvironments of the host, would provide a better understanding of the role of the Pho regulon in pathogenesis. In conclusion, TCRSs are absent in animal and human cells; thus, they represent choice targets for the development of therapeutics (Barrett et al., 1998; Stephenson & Hoch, 2004). A number of chemical inhibitors of TCRSs have been identified by screening libraries of synthetic compounds (Matsushita & Janda, 2002; Stephenson & Hoch, 2004). Drugs that interfere with the Pho regulon activity may also be useful as therapeutic agents that would compromise bacterial virulence and facilitate elimination of the pathogen by host defenses.


We thank Michaël Mourez (Université de Montréal) and France Daigle (Université de Montréal) for their insightful comments and critical reading of the manuscript. We are also grateful to Éliane Auger for her revisions of the manuscript. The authors gratefully acknowledge the financial help of Natural Sciences and Engineering Research Council of Canada (NSERC) (Discovery grant to J.H.), FQRNT studentship (M.G.L.) and National Institutes of Health (NIH GM62662 to B.L.W.).