|Home | About | Journals | Submit | Contact Us | Français|
The phage shock protein (Psp) system is induced by extracytoplasmic stress and thought to be important for the maintenance of proton motive force (PMF). We investigated the contribution of PspA to Salmonella virulence. A pspA deletion mutation significantly attenuates the virulence of S. Typhimurium following intraperitoneal inoculation of C3H/HeN (Ityr) mice. PspA was found to be specifically required for virulence in mice expressing the Nramp1 (Slc11a1) divalent metal transporter, which restricts microbial growth by limiting the availability of essential divalent metals within the phagosome. Salmonella competes with Nramp1 by expressing multiple metal uptake systems including the Nramp-homolog MntH, the ABC transporter SitABCD, and the ZIP family transporter ZupT. PspA was found to facilitate Mn2+ transport by MntH and SitABCD, as well as Zn2+ and Mn2+ transport by ZupT. In vitro uptake of 54Mn2+ by MntH and ZupT was reduced in the absence of PspA. Transport-deficient mutants exhibit reduced viability in the absence of PspA when grown under metal-limited conditions. Moreover, the ZupT transporter is required for S. Typhimurium virulence in Nramp1-expressing mice. We propose that PspA promotes Salmonella virulence by maintaining PMF, which is required for the function of multiple transporters mediating bacterial divalent metal acquisition during infection.
Integrity of the bacterial cell envelope is essential for fundamental physiological processes including energy generation, nutrient uptake, protein folding and secretion, cell wall synthesis, motility and signal transduction. Extracytoplasmic stress in gram-negative bacteria is sensed by multiple regulatory systems that serve to maintain cell envelope and periplasmic homeostasis, including the alternative sigma factor σE (RpoE), the CpxA-CpxR and BaeS-BaeR two-component regulatory systems, and the phage shock protein (Psp) system (Ades, 2004; Darwin, 2005; Duguay and Silhavy, 2004; Raivio and Silhavy, 2001; Rowley et al., 2006; Ruiz and Silhavy, 2005).
CpxA-CpxR and σE control overlapping but distinct stress responses (Raivio, 2005; Raivio and Silhavy, 1999). The importance of σE for the virulence of Salmonella enterica serovar Typhimurium (S. Typhimurium) has been amply demonstrated. The σE regulon is required for S. Typhimurium to withstand multiple stresses encountered within mammalian hosts, including acid and oxidative stress, and exposure to antimicrobial peptides (Bang et al., 2005; Crouch et al., 2005; Humphreys et al., 1999; Muller et al., 2009; Testerman et al., 2002). As a result, σE-deficient S. Typhimurium strains are impaired in their ability to survive within macrophages and cause progressive systemic infection in mice (Cano et al., 2001; Humphreys et al., 1999; Rowley et al., 2005). In addition to activation by the presence of misfolded outer membrane proteins (Alba and Gross, 2004), recent work has shown that σE can be activated by acid stress encountered within the macrophage phagosome (Muller et al., 2009). The Cpx system is also required for Salmonella virulence. A cpxA mutation reduces S. Typhimurium adherence to and invasion of cultured epithelial cells, and impairs the ability of bacteria to replicate in mice (Humphreys et al., 2004). This may be because the Cpx system is required for the assembly and expression of surface organelles such as pili. In addition, studies in a related enteric pathogen, Yersinia pseudotuberculosis, indicate that both σE and CpxA-CpxR are important for the function of Type III Secretory Systems (T3SS) (Carlsson et al., 2007), which play a vital role in Salmonella pathogenesis as well (Carlsson et al., 2007; Humphreys et al., 2004; Humphreys et al., 1999). The BaeS-BaeR system has not yet been characterized with respect to Salmonella virulence, but this regulatory system is also induced by envelope stress and may provide some compensation for the loss of Cpx function (Raffa and Raivio, 2002).
The Psp system was first discovered by Model as a response of E. coli to infection with filamentous bacteriophages (Brissette et al., 1990). The six phage shock proteins are encoded by an operon (pspABCDE) and an unlinked gene (pspG) (Brissette et al., 1991; Lloyd et al., 2004). A divergently transcribed locus called pspF encodes a transcriptional activator that is required along with the sigma factor σN for psp operon transcription (Jovanovic et al., 1996). PspA is abundantly expressed during filamentous bacteriophage infection of E. coli (Brissette et al., 1990) and is believed to have both regulatory and effector roles. PspA negatively regulates psp operon expression by binding to PspF (Dworkin et al., 2000), but PspB and PspC bind to PspA to promote psp gene expression (Adams et al., 2003).
Since Model’s original observations, the phage shock response has been shown to be induced by conditions in addition to bacteriophage infection that perturb the integrity of the cytoplasmic membrane, including the protonophore CCCP (carbonylcyanide m-chlorphenylhydrazone) and overproduction of the secretin component of a Type II or Type III Secretory System (Becker et al., 2005; Brissette et al., 1991; Guilvout et al., 2006; Kleerebezem et al., 1996; Maxson and Darwin, 2004; Weiner and Model, 1994). Defects in secretion can also induce expression of the psp genes (Jones et al., 2003), as can ethanol, organic solvents, alkaline pH, osmotic or heat shock, and stationary phase (Brissette et al., 1991; Kobayashi et al., 1998; Weiner and Model, 1994). Mutations in rpoE strongly induce expression of the CpxA-CpxR and Psp regulons (Bang et al., 2005; Becker et al., 2005; Connolly et al., 1997), suggesting complementarity of these extracytoplasmic stress responses. Accordingly, mutations in both rpoE and pspA have a greater effect on stress-related phenotypes than single mutations in either locus alone (Becker et al., 2005). Psp-inducing conditions appear to share the common feature of membrane damage resulting in the impairment of proton motive force (PMF).
Model suggested that the primary function of the phage shock response is to maintain PMF during extracytoplasmic stress (Model et al., 1997). As a function of membrane potential (Δψ) and the transmembrane pH gradient (ΔpH), PMF (Δp) is sensitive to alterations in membrane permeability (Harold, 1996). PspA has been shown to play a role in the maintenance of PMF when protein export is impaired (Kleerebezem et al., 1996). More recently, Kobayashi et al. have shown that PspA can oligomerize and prevent proton leakage across damaged membrane vesicles in vitro (Kobayashi et al., 2007). The formation of extensive scaffold structures has been proposed as a mechanism by which PspA might stabilize membrane integrity (Standar et al., 2008).
Activation of the phage shock response has been observed under a variety of complex biological conditions of potential importance in bacterial virulence, including swarming motility, biofilm formation, activation of type III secretion, epithelial cell invasion, and bacterial internalization by macrophages (Beloin et al., 2004; Darwin and Miller, 1999; Darwin and Miller, 2001; Eriksson et al., 2003; Lucchini et al., 2005; Wang et al., 2004). Given the putative role of the phage shock response in the maintenance of PMF and the importance of PMF for a number of processes important for pathogenesis, including motility, resistance to host-derived antimicrobial mediators, secretion and nutrient acquisition, we investigated the contribution of the phage shock response to Salmonella virulence.
To determine whether PspA is essential for Salmonella virulence, six week-old C3H/HeN (Ityr, Nramp1/Slc11a1+) mice were injected intraperitoneally with 3000 CFU wild-type or isogenic mutant S. Typhimurium strains carrying an in-frame pspA deletion mutation (ΔpspA) and monitored for signs of illness (see Experimental procedures). The S. Typhimurium strain lacking pspA was severely attenuated for virulence in C3H/HeN mice (Fig. 1). All mice infected with wild-type Salmonella succumbed to infection within 15 days, whereas no mouse infected with the ΔpspA mutant died during this time period, and most animals appeared completely healthy throughout the course of the experiment. The promoter and coding region of pspA were cloned into plasmid vector pRB3 (see Experimental procedures), and this construct, pJK664, was introduced into the S. Typhimurium ΔpspA strain. PspA levels of the ΔpspA strain carrying pJK664 were comparable to wild-type by Western blot analysis (data not shown), indicating that pJK664 restored PspA production. Salmonella virulence in C3H/HeN mice was fully complemented by introduction of pJK664 into the ΔpspA strain, demonstrating that the virulence defect in the ΔpspA mutant resulted from the loss of pspA (Fig. 1).
Expression of the phage shock operon is induced when Type III secretion is activated in Y. enterocolitica (Darwin and Miller, 2001) or Salmonella (Wang et al., 2004). Additional studies have shown the importance of PMF for T3SS function (Minamino and Namba, 2008; Paul et al., 2008; Wilharm et al., 2004). As T3SSs play an essential role in Salmonella motility, host cell invasion, intracellular survival and virulence, we investigated whether a pspA mutation affected T3SS-dependent phenotypes. As shown in Fig. 2A, the loss of pspA affected neither motility nor swarming behavior under conventional laboratory conditions. Thus, flagellar Type III secretion appears to be unaffected by the loss of PspA.
Salmonella invasion of host epithelial cells requires effector proteins secreted by the T3SS encoded on Salmonella Pathogenicity Island 1 (SPI1) (Hansen-Wester and Hensel, 2001; Lostroh and Lee, 2001). As SPI1 and flagellar secreted proteins can be detected in culture supernatants (Kimbrough and Miller, 2000; Lee and Galan, 2004), culture supernatants were obtained from wild type and ΔpspA strains under conditions favoring SPI1 expression, subjected to SDS-PAGE, and stained with Coomassie blue (see Experimental procedures). No differences in SPI1 or flagellar T3SS secreted proteins were observed between wild-type and ΔpspA bacteria (Fig. 2B). Further, the loss of PspA did not abrogate the ability of Salmonella to invade a non-phagocytic epithelial cell line (HeLa) compared to wild-type (Supplemental Fig. S1).
Effector proteins secreted by the T3SS encoded on Salmonella Pathogenicity Island 2 (SPI2) are required for bacterial survival in macrophages (Hansen-Wester and Hensel, 2001; Kuhle and Hensel, 2004). Expression of SPI2 genes is dependent on the SsrA-SsrB two-component regulatory system (Feng et al., 2003; Feng et al., 2004; Garmendia et al., 2003; Walthers et al., 2007). RAW264.7 macrophages were infected with wild-type Salmonella or isogenic ΔpspA, phoP, ssrA or ssrB mutant strains, and the numbers of surviving bacteria after 18h of infection were determined by lysis, serial dilution and plating (see Experimental procedures). Survival of the ΔpspA mutant strain was comparable to wild-type, whereas the SPI2-deficient ssrA and ssrB mutant strains, as well as the attenuated phoP mutant control strain, exhibited intracellular survival defects (Fig. 2C). Thus, SPI2-dependent intracellular bacterial survival does not require an intact phage shock response. In addition, a ΔpspA mutant was no more sensitive than wild-type Salmonella to host-derived antimicrobial mediators including hydrogen peroxide, nitric oxide, and the antimicrobial peptides cryptdin-4 and P2 (data not shown). Collectively, these data indicate that the attenuated virulence of ΔpspA mutant S. Typhimurium does not result from defective Type III secretion, nor to increased sensitivity to phagocyte killing, oxidative stress, nitrosative stress, or antimicrobial peptides encountered in the host environment.
In contrast to our observations in C3H/HeN mice (Fig. 1), a ΔpspA mutant S. Typhimurium strain was unexpectedly found to be as virulent as wild-type in C57BL/6 (Itys) mice, which lack a functional Nramp1/Slc11a1 locus (Fig. 3A). We hypothesized that PspA might only be able to promote Salmonella virulence in mice possessing an intact Nramp1 locus (also designated ity/lsh/bcg) on chromosome 1. This locus, encoding natural-resistance-associated macrophage protein 1 (Nramp1), is associated with enhanced susceptibility to intracellular pathogens including S. Typhimurium, Leishmania major and Mycobacterium bovis BCG (Bradley, 1977; Gros et al., 1981; Plant and Glynn, 1976). Nramp1, also called Slc11a1, is now established to be a proton-dependent divalent metal transporter expressed by phagocytic cells and localized to the phagosomal membrane (Forbes and Gros, 2003; Goswami et al., 2001; Gruenheid et al., 1997). Nramp1 enhances host resistance to intracellular pathogens by limiting essential metal availability within the phagosome (Cellier et al., 2007; Forbes and Gros, 2001; Mackenzie and Hediger, 2004). C57BL/6 and many other strains of inbred mice are phenotypically Nramp1− due to a G169D mutation (Malo et al., 1994). To investigate the importance of Nramp1 for the attenuated virulence phenotype of the ΔpspA S. Typhimurium mutant strain, we compared the virulence of wild-type and ΔpspA mutant S. Typhimurium in C57BL/6 mice and transgenic C57BL/6 Nramp1G169 mice that are phenotypically Nramp1+ (Govoni et al., 1996). The ΔpspA mutant bacteria were avirulent in mice expressing the Nramp1G169 transgene (Fig. 3B), indicating that the requirement for PspA in Salmonella pathogenesis is dependent on the presence of a functional Nramp1 locus in the host. In addition, we investigated the survival of isogenic wild-type and ΔpspA S. Typhimurium strains in primary macrophages isolated from C57BL/6 (Nramp1−) and C57BL/6 Nramp1G169 (Nramp1+). S. Typhimurium lacking PspA survived similarly to wild-type in C56BL/6 (Nramp1−) macrophages (Supplemental Fig. S2), whereas a ΔpspA S. Typhimurium mutant was defective for survival in C57BL/6 Nramp1G169 (Nramp1+) macrophages in comparison to wild-type (Supplemental Fig. S2). These results support the in vivo observation that PspA promotes Salmonella virulence in mice expressing the Nramp1 (Slc11a1) divalent metal transporter.
Nramp1 enhances host resistance presumably by limiting the availability of essential metals required for microbial growth in the phagosomal compartment (Cellier et al., 2007; Forbes and Gros, 2001). Previous studies have shown that murine Nramp1 facilitates the transport of divalent cations including Mn2+, Fe2+ and Zn2+ along a proton gradient (Forbes and Gros, 2003; Goswami et al., 2001). Salmonella possesses a high-affinity manganese transporter called MntH, which is a homolog of Nramp1, as well as an ATP-Binding Cassette transporter known as SitABCD, which can mediate both high affinity manganese and low affinity iron uptake at alkaline pH (Kehres et al., 2002; Kehres et al., 2000; Zhou et al., 1999). Cation uptake by MntH requires a proton gradient, whereas ABC-type transport systems such as SitABCD utilize ATP hydrolysis (Courville et al., 2004; Davidson and Chen, 2004; Higgins, 1992; Makui et al., 2000). The MntH and SitABCD transporters are required for S. Typhimurium virulence in Nramp1+ but not in Nramp1-deficient mice (Zaharik et al., 2004), suggesting that metal acquisition becomes of vital importance to intracellular pathogens when the host is able to actively evacuate divalent metals from the phagosome via Nramp1. The putative role of PspA in the maintenance of PMF, the proton-dependence of MntH (Makui et al., 2000), and the Nramp1-dependence of the pspA virulence phenotype (Fig. 3), as well as the reduced survival of a ΔpspA Salmonella mutant in C57BL/6 Nramp1G169 (Nramp1+) primary macrophages (Supplemental Fig. S2), led us to investigate a possible role of PspA in Salmonella metal transport. PspA induction during Salmonella growth in chelated media (LB+dipyridyl, Supplemental Fig. S3) further suggested a possible role of PspA under metal-limited conditions.
Salmonella strains defective for various divalent metal transporters were constructed to determine the effect of a pspA mutation on the MntH and SitABCD manganese transport systems (see Experimental procedures). To assess the individual contribution of each manganese transport system, the growth of wild-type and transport-deficient mutant strains was examined in chelated medium (LB+dipyridyl) to which Mn2+ was or was not added back. The contribution of MntH transporter for growth in the same chelated medium was assessed in an SFZ (sitA feoB ΔzupT) triple mutant strain lacking the ABC transporter SitABCD, the major Fe2+ transporter FeoB, and the ZIP family metal permease ZupT. The SFZ mutant grew poorly in chelated medium, but growth was restored close to wild-type levels by the addition of MnCl2 (Fig. 4A), consistent with MntH-dependent acquisition of Mn2+. MntH dependence was confirmed in a SFZM (sitA feoB ΔzupT mntH) mutant strain, for which growth in chelated medium was not restored by the addition of MnCl2 (Fig. 4A). An SFZP (sitA feoB ΔzupT ΔpspA) mutant strain exhibited an even greater impairment of growth in chelated medium (Fig. 4A), but growth was restored by complementation with the pspA-expressing plasmid pJK664, indicating that PspA supports MntH-dependent Mn2+ uptake (Supplemental Fig. S4A). The ABC-type metal transporter SitABCD was similarly assessed in an MFZ (mntH feoB ΔzupT) triple mutant strain. The MFZ mutant strain grew poorly in chelated medium, but growth was restored by the addition of MnCl2, consistent with SitABCD-mediated uptake of Mn2+ (Fig. 4B). Somewhat unexpectedly, PspA was also found to facilitate SitABCD-mediated Mn2+ transport. Like the SFZP (sitA feoB ΔzupT ΔpspA) mutant, an MFZP (mntH feoB ΔzupT ΔpspA) mutant strain grew poorly in chelated medium (Fig. 4B), and growth was restored by the pspA-expressing plasmid pJK664 (Supplemental Fig. S4B). Acquisition of Mn2+ by MntH and Mn2+ and possibly Fe2+ by SitABCD is important for survival in macrophages as well as virulence in S. Typhimurium (Boyer et al., 2002). To assess SitABCD-mediated Fe2+ uptake, an ΔentC mutation was introduced into transport-deficient mutant strains to eliminate siderophore-mediated iron uptake (Crouch et al., 2008). SitABCD-dependent growth of an EMZF mutant (ΔentC mntH ΔzupT feoB) in chelated medium was enhanced by the addition of FeSO4 (Fig. 4C). The addition of a ΔpspA mutation significantly impaired growth of the EMZF strain in chelated medium (Fig. 4C).
Collectively, these observations suggest that PspA can facilitate divalent metal transport by both Nramp-type (MntH) and ABC-type (SitABCD) transport systems. Furthermore, at high dipyridyl concentrations, a pspA mutation alone results in growth impairment (Supplemental Fig. S5), supporting the notion that PspA plays an integral role in metal acquisition by Salmonella. In the growth assays, we noted that transport-deficient mutant strains lacking PspA may exhibit a decline in OD600 following initial growth (Figs. 4B, 4C and Supplemental Figs. S4 and S5), suggesting that PspA is required for cell viability during essential metal deprivation. We therefore determined the viability of MntH-deficient mutants under metal-limiting conditions. Salmonella strains WT, P (ΔpspA), SFZ (sitA feoB ΔzupT), and SFZP (sitA feoB ΔzupT ΔpspA), were grown in chelated medium and monitored for 24h by OD600, with aliquots obtained in parallel to determine colony forming units (CFU). After 24h growth in metal-depleted conditions, cultures of WT and an isogenic SFZ mutant were found to contain ~40-fold more CFU than at T=0 (Supplemental Fig. S6), and a ΔpspA mutant showed only a two-fold reduction in survival compared to wild-type (Supplemental Fig. S6). However, the transport-deficient SFZ mutant strain lacking PspA (SFZP) was essentially non-viable after 24h (Supplemental Fig. S6). These data suggest that PspA not only facilitates metal transport, but is also required for Salmonella viability during metal deprivation.
The ZIP metal transporter family was first described in Arabidopsis, but members have subsequently been identified in all eukaryotic kingdoms (Eng et al., 1998; Guerinot, 2000). ZupT is the first bacterial ZIP homolog to be discovered and was recently shown to transport a broad range of divalent cations in Escherichia coli (Grass et al., 2005; Grass et al., 2002). Although the energy requirements of ZIP transporters are not known, family members share some sequence similarity to a cation diffusion facilitator family transporter that is driven by PMF (Guerinot, 2000). The S. Typhimurium genome sequence is predicted to encode a ZupT protein that is 86% identical to ZupT of E. coli W3110. We investigated whether ZupT in S. Typhimurium is also a divalent metal permease and whether PspA supports ZupT-dependent transport.
The contribution of ZupT to S. Typhimurium growth in chelated medium was examined in an MSFN (mntH sitA feoB ΔznuABC) strain background. The MSFN mutant was unable to grow in chelated medium, but growth was restored when the medium was supplemented with ZnCl2 (Fig. 5A). In addition, an MSFNZ (mntH sitA feoB ΔznuABC ΔzupT) mutant was unable to grow in chelated medium, even with ZnCl2 supplementation, unless the zupT gene was provided on plasmid pJK671 in trans (Fig. 5B). Thus, ZupT mediates Zn2+ transport in Salmonella. MSFN and MSFNZ/pJK671 mutant strains lacking PspA exhibited impaired growth in chelated medium (Figs. 5A, 5B). Growth of the MSFN mutant in chelated medium was restored by the introduction of the pspA-expressing plasmid pJK664 (Fig. 5A). Parallel studies performed with MnCl2 instead of ZnCl2 supplementation suggest that ZupT may mediate transport of Mn2+ as well (data not shown). Thus, ZupT exhibits broad cation substrate specificity in S. Typhimurium as shown previously in E. coli (Grass et al., 2005). Together, these observations demonstrate that PspA facilitates the ZupT-mediated transport of Zn2+ and Mn2+ in Salmonella.
Transport assays were performed to determine the dependence of MntH and ZupT transport function on PspA by measuring uptake of 54Mn2+ in wild-type and transport-deficient mutant strains (see Experimental procedures). The activity of the MntH transporter was assessed in an SFZ (sitA feoB ΔzupT) triple mutant background in which uptake of 54Mn2+ is primarily mntH-dependent (Fig. 6A). In the absence of PspA, MntH-mediated uptake of 54Mn2+ was reduced by 45% (Fig. 6A). The ZupT transporter was assessed in an MSFNZ (mntH sitA feoB ΔznuABC ΔzupT) mutant strain carrying the ZupT plasmid pJK671, in which 54Mn2+ uptake is primarily ZupT-dependent (Fig. 6B). The absence of PspA reduced ZupT-mediated uptake of 54Mn2+ by 40% (Fig. 6B). These data provide direct evidence that PspA facilitates the uptake of divalent cations by metal transport systems in Salmonella.
Uropathogenic E. coli carrying a zupT mutation was found to be defective for competitive infections in CBA/J mice only when the high-affinity ZnuABC zinc transporter was absent, suggesting a supportive role for ZupT in E. coli pathogenesis (Sabri et al., 2009). We compared the contributions of ZupT and MntH to Salmonella virulence. Previous studies showed that an mntH S. Typhimurium mutant is moderately attenuated for virulence in Nramp+ mice (Zaharik et al., 2004). Therefore, we compared the ability of zupT and mntH mutant strains to cause progressive infection in Nramp+ C3H/HeN (Ityr) mice following intraperitoneal inoculation. The S. Typhimurium ΔzupT mutant exhibited an attenuated virulence phenotype similar to that of an isogenic mntH Nramp+ mutant in C3H/HeN mice, with 50% of mice infected with the ΔzupT mutant surviving at 25 days post-infection in comparison to only 20% of mice infected with wild-type S. Typhimurium (Fig. 7). Attenuated virulence was not observed in Nramp− C57BL/6 mice infected with the S. Typhimurium ΔzupT mutant, and animals infected with either wild-type or the ΔzupT mutant succumbed equally to infection (Supplemental Fig. S7). These results demonstrate that ZupT is required for full Salmonella virulence in Nramp+ mice.
The phage shock response is a regulated response to extracytoplasmic stress that serves to maintain PMF and preserve essential cellular functions. PspA, the most abundant phage shock protein, is required for most of the phenotypes associated with the phage shock response (Darwin, 2005). The goal of this study was to investigate the contribution of the phage shock response to Salmonella virulence. We have demonstrated that PspA-deficient mutant S. Typhimurium is profoundly attenuated for virulence following intraperitoneal inoculation of C3H/HeN mice. However, although the phage shock response is strongly induced by the activation of type III secretion, Salmonella motility and the SPI1 and SPI2 T3SSs appear to be unimpaired in the absence of PspA. Moreover, a pspA mutant strain appears no more susceptible than wild-type S. Typhimurium to extracellular stresses encountered in the host environment, including exposure to reactive oxygen or nitrogen species, and antimicrobial peptides.
An important clue to the role of PspA in Salmonella pathogenesis came from the observation that a pspA mutation only attenuates virulence in mice expressing Nramp1 (Fig. 3). Nramp1 (also called Slc11a1) was originally discovered as a genetic locus (ity/lsh/bcg) of inbred mice that is associated with increased susceptibility to phylogenetically diverse intracellular pathogens including S. Typhimurium, Leishmania major and Mycobacterium bovis BCG (Bradley, 1977; Gros et al., 1981; Plant and Glynn, 1976). The Nramp1 locus was shown to encode a 56kDa membrane protein that contains a nonconservative glycine to aspartic acid substitution at residue 169 in animals with deficient Nramp1 function (Malo et al., 1994). Subsequent studies showed that Nramp1 is a metal ion transporter that can mediate the transport of divalent cations including Mn2+, Fe2+, Zn2+ and Co2+ by a proton-dependent mechanism. (Forbes and Gros, 2003; Goswami et al., 2001). Although some investigators have suggested that Nramp1 might transport metals into the phagosome (Blackwell et al., 2001; Goswami et al., 2001; Kuhn et al., 1999; Kuhn et al., 2001; Techau et al., 2007; Zwilling et al., 1999), the preponderance of evidence indicates that Nramp1 leads to the depletion of divalent metal ions in the phagosome (Barton et al., 1999; Cellier et al., 2007; Forbes and Gros, 2001; Forbes and Gros, 2003; Fritsche et al., 2007; Jabado et al., 2000; Mulero et al., 2002; Nevo and Nelson, 2006), and a recent study directly measuring total cellular iron content by atomic absorption spectrometry has demonstrated that Nramp1-expressing macrophages contain lower quantities of iron (Nairz et al., 2009). Depletion of metal ions serves as a defense mechanism against intracellular pathogens, as microbes require metals such as iron, zinc and manganese for a variety of essential metabolic processes (Cellier et al., 2007; Forbes and Gros, 2001; Mackenzie and Hediger, 2004). The selective PspA requirement of S. Typhimurium in Nramp1-expressing mice and peritoneal macrophages (Fig. 3 and Supplemental Fig. S2) suggested a possible role of PspA in Salmonella metal acquisition.
Bacteria utilize multiple types of transporter with overlapping specificities to acquire metal ions within the host environment. S. Typhimurium utilizes the Nramp-like transporter MntH, the ATP-binding cassette SitABCD transport system, the ZIP family transporter ZupT, and the Feo transporter to mediate the uptake of divalent cations such as Mn2+, Fe2+ and Zn2+. These transporters have been shown in this and earlier studies to play in important role in Salmonella pathogenesis (Boyer et al., 2002; Campoy et al., 2002; Janakiraman and Slauch, 2000; Zaharik et al., 2004). Here we have also shown that the phage shock protein PspA facilitates Mn2+ transport by MntH, Mn2+ and Fe2+ transport by SitABCD (Fig. 4 and Supplemental Fig. S4), and Zn2+ transport by the ZIP family transporter ZupT (Fig. 5). Like MntH and SitABCD (Zaharik et al., 2004), we have found that ZupT and PspA are only required for Salmonella virulence in mice possessing a functional Nramp1 locus, consistent with the idea that host Nramp1 and bacterial MntH, SitABCD and ZupT transporters compete for metals within the phagosomal vacuole.
Accumulating evidence continues to support Model’s original hypothesis that the phage shock response preserves PMF under extracytoplasmic stress conditions (Becker et al., 2005; Kleerebezem et al., 1996; Kobayashi et al., 2007; Standar et al., 2008). Our observations further reinforce this model, showing that PspA promotes Mn2+ uptake by the MntH transporter during growth in chelated medium (Fig. 4 and Supplemental Fig. S4A) as well as 54Mn2+ uptake in transport assays (Fig. 6A). As MntH is a proton symporter in the Nramp family (Papp-Wallace and Maguire, 2006), MntH function would be predicted to be highly dependent on PMF (Fig. 8). Although PspA is most likely to promote MntH-mediated transport by facilitating the maintenance of PMF, an effect of PspA on transport protein expression or activity cannot be excluded. Somewhat less anticipated was our discovery that PspA also facilitates metal uptake by the ABC-type SitABCD transporter (Fig. 4 and Supplemental Fig. S4B). However, although ABC-type transporters are energized by ATP hydrolysis (Davidson and Chen, 2004; Higgins, 1992), PMF can drive the generation of ATP via oxidative phosphorylation by the F0F1 ATPase (Harold, 1996; Senior, 1990). The reverse reaction can occur under anaerobic conditions when the proton gradient is low, allowing the F0F1 ATPase to support PMF as well as utilize it for ATP generation (Senior, 1990). Interestingly, mutations in atpB encoding the a subunit of the proton-translocating F0 complex increase the expression of pspA (Becker et al., 2005; Maxson and Darwin, 2004), suggesting a complex interplay between the F0F1 ATPase, PspA and PMF. Furthermore, overexpression of a thermophilic F0F1 ATPase has been shown to cause transmembrane proton leakage and the induction of pspA expression in E. coli (Kobayashi et al., 2007). PspA can oligomerize and inhibit proton leakage across damaged membranes in vitro (Kobayashi et al., 2007). In addition, the formation of extensive scaffold structures by PspA is proposed to stabilized membrane integrity (Standar et al., 2008). Although the precise mechanism by which PspA maintains PMF is uncertain, we suggest that a reduction in PMF resulting from the loss of PspA may limit ATP generation by the F0F1 ATPase, thereby limiting transport by SitABCD (Figs. 4 and and8).8). Defective survival of strains lacking PspA in chelated medium (Supplemental Fig. S6) may reflect reduced activity of other ATP-dependent processes and could suggest a possible role of PspA in maintaining the overall energy status of the cell.
In this work, we have also shown for the first time that a bacterial ZIP family transporter plays an essential role in pathogenesis. S. Typhimurium lacking the ZIP transporter ZupT is attenuated for virulence in Nramp1+ mice following intraperitoneal inoculation (Fig. 7). It is interesting to note that LIT1, a ZIP family Fe2+ transporter, is required for the growth of the protozoan Leishmania amazonensis in macrophages (Huynh et al., 2006), providing an example of a common mechanism utilized by a bacterial and a eukaryotic pathogen to promote growth within the host environment. Little is currently understood regarding the energy requirements for metal transport by the ZIP protein family. However, ZIP transporters share sequence homology with the cation diffusion facilitator family, which mediate metal efflux from eukaryotic cells (Kambe et al., 2006). Transporters in the cation diffusion facilitator family appear to function as proton antiporters, thus requiring a proton gradient for metal efflux (Haney et al., 2005). We have shown that PspA facilitates Zn2+ uptake via ZupT in chelated medium (Fig. 5) as well as uptake of 54Mn2+ by transport assays (Fig. 6B), although additional studies will be required to establish whether ZupT-mediated transport is directly PMF-dependent. Collectively, our observations demonstrate that the phage shock response plays an essential role in Salmonella pathogenesis by facilitating high-affinity metal acquisition within the nutrient-limited host cell environment.
Salmonella enterica serovar Typhimurium strains, plasmids and primers used in this study are listed in Table 1. Strains were routinely grown in LB at 37°C. Antibiotics were used at the following concentrations as required: ampicillin (100µg·ml−1), kanamycin (50µg·ml−1), chloramphenicol (20µg·ml−1), and tetracycline (25µg·ml−1). Deletion mutants were constructed by λ-Red-mediated methods as described (Datsenko and Wanner, 2000; Karlinsey, 2007). In-frame deletion of pspA was constructed with primers JP18, JP21, JP61 and JP62 using the λ-Red tetRA replacement method (Karlinsey, 2007). An additional in-frame deletion of pspA (pspA::FRT) was constructed with primers MLC165 and MLC166 using the λ-Red method as described (Datsenko and Wanner, 2000). Both pspA deletions were phenotypically identical as determined by susceptibility to hydrogen peroxide and Sper/NO, growth in chelated medium with divalent metal add-back, and mouse virulence assays. In-frame deletions of zupT and znuABC were constructed using primers JP272/JP273 and JP316/JP317, respectively, using the λ-Red method (Datsenko and Wanner, 2000). A SPI1 deletion was constructed using primers KMp161 and KMp162 (Main-Hester et al., 2008), except that the FRTkanFRT cassette was not flipped out. All mutations were transduced into wild-type S. Typhimurium using bacteriophage P22HT105/int (Davis, 1980). Plasmids with pspA or zupT used for complementation in mouse virulence assays and for assays of growth in chelated medium were constructed as follows. Primers JP5/JP62 and JP279/JP280 were used to PCR-amplify the promoter and coding regions of pspA and zupT, respectively. The genes were cloned into the stable low-copy cloning vector pRB3 (Berggren et al., 1995) to create pJK664 (pRB3::pspA) and pJK671 (pRB3::zupT). To purify His8-PspA protein for subsequent anti-PspA antibody production, plasmid pJK654 (pBAD24-His8-PspA) was made using primers JP143 and JP144 to PCR-amplify and engineer the addition of eight histidines to the N-terminal coding region of pspA before cloning into pBAD24 (Galan and Curtiss, 1989).
Moribund animals were euthanized. The following inocula were used for infection in various mouse backgrounds: 3×103 CFU in six to eight week-old C3H/HeN (Nramp+, Ityr) mice (Charles River Laboratories), 1×102 CFU in six to eight week-old C57/BL6 (Nramp−) mice (Charles River Laboratories), and 105 CFU in six to eight week-old congenic C57/BL6 Nramp1G169 mice (bred and housed in Modified Specific Pathogen Free facilities at the University of Washington animal facilities).
Motility assays were performed at 37°C for 8.5 h in motility medium as described (Silverman and Simon, 1973), except the agar concentration was reduced to 0.3%. Swarming assays were performed on swarm medium (Wang et al., 2004) at 37°C for 9h. SPI1-secreted proteins were purified from overnight cultures essentially as described (Pegues et al., 1995). Fifty ml of overnight culture supernatants were TCA-precipitated and fractionated on a 4–15% polyacrylamide gel in Tris-Glycine-SDS buffer (Biorad, Hercules, CA) and the proteins detected by Coomassie blue staining.
HeLa cells were grown in RPMIc (RPMI 1640 with L-glutamine, 25mM HEPES (Mediatech Inc.), 10% heat-inactivated fetal calf serum (Hyclone) and 1:100 Pen/Strep (Mediatech Inc)). Cells were seeded at 2×105 cells per well in a 48 well microtiter dish. The cells were changed into fresh RPMIc without antibiotic 24h before the assay. S. Typhimurium strains were grown in LB without shaking at 37°C for 18–24h. HeLa cells were infected with an MOI of ~25:1 and centrifuged at 215g for 10min at 25°C and incubated for 20min at 37°C in 5% CO2. Cells were washed 3× in RPMIc (without antibiotics) then incubated in RPMIc + 100µg/ml gentamicin sulfate (MP Biochemicals) for 1h at 37°C in 5% CO2. At indicated times the cells were lysed in 1% Triton X-100 for 5min and bacteria quantified by plating on LB agar. % invasion was calculated as follows: (CFU T1h/CFU Tinitial) × 100.
S. Typhimurium survival within primary Nramp1− and Nramp1+ macrophages and RAW264.7 cells lines was determined as previously described (Zaharik et al., 2004). Cultures were grown with shaking in LB at 37°C for 18–24h. Salmonella strains were opsonized with mouse serum, infected at a multiplicity of infection of 5:1 and incubated with macrophages for 15 min. After phagocytosis, the cells were washed in RPMI with 12 µg·ml−1 gentamicin to kill any remaining bacterial that were not internalized. At designated time intervals, the cells were lysed in 1% Triton X-100 and surviving bacteria quantified by plating on LB agar.
Peptide killing of cryptdin-4 and P2 were performed as described (Crouch et al., 2005). Sensitivity to hydrogen peroxide (Fisher Chemical, Fairlawn, NJ) or nitric oxide donor spermine NONOate (Sper/NO, Calbiochem, La Jolla, CA) were performed as follows. Overnight cultures were diluted to 2×106 CFU/ml in LB with the addition of either hydrogen peroxide (ranging up to 1mM) or spermine NONOate (ranging up to 2mM) in a final volume of 300µl. Growth kinetics were determined by measuring the optical density at 600nm at 37°C on a Bioscreen C Microbiology Microplate reader (Growth Curves USA, Piscataway, NJ).
His-PspA was purified from strain JK451 using His-Bind resin (Novagen, Gibbstown, NJ) according to the manufacturer’s instructions. Rabbit anti-PspA antisera were produced at Scantibodies Laboratories, Inc. (Santee, CA) and non-specific antibodies adsorbed as described (de Wet et al., 1984). Goat peroxidase-conjugated anti-rabbit HRP was obtained through Bio-Rad (Hercules, CA). Culture sample preparation, SDS electrophoresis, and western blot hybridization and detection were performed as described (Muller et al., 2009).
Overnight cultures were diluted to 4×105 CFU/ml in LB chelated with 2,2’-dipyridyl (Sigma-Aldrich, St. Louis, MO). Divalent metals (all acquired from Sigma-Aldrich, St. Louis, MO) were added back to chelated LB medium at concentrations indicated in the figure legends. Growth kinetics were determined by measuring the optical density at 600nm at 37°C on a Bioscreen C Microbiology Microplate reader (Growth Curves USA, Piscataway, NJ). Three to five biological replicates were performed on each strain, and the averages and standard deviations of the optical density measurements were calculated for each time point.
Uptake of 54Mn2+ by whole cells was measured essentially as described (Maguire, 2007). Overnight cultures were diluted in fresh LB broth and grown to OD600 0.6 to 0.7. The cells were spun and washed in 1× cold N-minimal medium (Maguire, 2007) and resuspended at OD600 2.0 in cold N-minimal medium with no additions. Cells (100 µl) were added to 900 µl pre-warmed N-minimal medium containing 0.5 to 1 µCi 54Mn2+ at a final total Mn2+ concentration of 0.1 µM (equal to the K0.5 for Mn2+ uptake by MntH and SitABCD (Kehres et al., 2000), and uptake allowed to proceed for 10 min. Four ml of cold wash buffer (1× N-minimal medium, 0.5 mM EDTA and 10 mM MgSO4) were added and immediately filtered onto a BA85 filter (Schleicher and Schuell, Keene, NH) before washing once with cold wash buffer. Filters were placed in vials and counted in a Beckman scintillation counter at any efficiency of ~80%. At least three biological replicates were assayed for each strain.
We are grateful to Stephen Libby, Kara Main-Hester, Samuel Miller and Barry Wanner for strains used in this study. This work was supported by research grants (AI44486, AI77629) to F.C.F. and a training grant (AI54052) to M.L.C. from the National Institutes of Health.