Perturbation of the Oxidizing Environment of the Periplasm Stimulates the PhoQ/PhoP System in Escherichia coli

Abstract

The PhoQ/PhoP two-component system is repressed by divalent cations, such as Mg²⁺ and Ca²⁺, in the growth medium and stimulated by low pH and certain cationic antimicrobial peptides. In Escherichia coli, it was recently shown that the histidine kinase PhoQ is also modulated by at least two additional factors, the small membrane proteins SafA and MgrB. This raises the possibility that the PhoQ/PhoP circuit has additional regulatory components and integrates additional input signals. We screened E. coli transposon insertion mutants to look for proteins that modulate the activity of the PhoQ/PhoP system, and we uncovered a role for DsbA, a periplasmic oxidant that facilitates the formation of disulfide bonds. Deletion of dsbA or dsbB, which maintains a pool of oxidized DsbA, leads to increased transcription of at least two PhoP-regulated genes. Addition of the reducing agent dithiothreitol to wild-type cells had a similar effect, and treatment of a dsbA null strain with the oxidant Cu²⁺ rescued the reporter gene expression phenotype. We also demonstrated that expression of an MgrB mutant that lacked cysteines blocked the effect of a dsbA null mutation on PhoQ/PhoP activity, suggesting that MgrB acts downstream of DsbA in this pathway. Taken together, these results demonstrate that a decrease in the oxidizing activity of the periplasm stimulates PhoQ/PhoP and may reveal a new input stimulus for this important two-component system.

INTRODUCTION

Many two-component signaling systems depend on more than their two core protein components, the histidine kinase and response regulator, for proper signal transduction. Additional proteins may modulate various steps in signaling pathways, such as input sensing, phosphoryl transfer, dephosphorylation, or output regulation. Some of these auxiliary proteins have well-understood functions, such as detecting specific input signals and relaying this information to the sensory domain of a histidine kinase, or connecting two otherwise distinct circuits (1, 7, 14, 20). However, in many cases the physiological significance of these proteins remains poorly understood.

One example of a well-studied two-component system whose activity is modulated by additional protein components is the PhoQ/PhoP system, which is found in Escherichia coli, Salmonella species, and related bacteria. The sensor kinase PhoQ controls the phosphorylation state of PhoP through PhoQ autophosphorylation, followed by phosphoryl transfer to PhoP and PhoQ-mediated phosphatase activity against phosphorylated PhoP (PhoP-P). Thus, the level of PhoP-P is set by the balance between the kinase and phosphatase activities of PhoQ. For simplicity, we refer to conditions that increase or decrease the ratio of PhoQ kinase to phosphatase activity, which lead to increased or decreased PhoP-P, as stimulating or repressing PhoQ activity, respectively. PhoQ activity is repressed by growth medium containing high concentrations of Mg²⁺ and is stimulated by low pH and certain cationic antimicrobial peptides (4, 16, 25, 28). Recent work also uncovered two small membrane proteins that modulate PhoQ activity: SafA (B1500) (10), which to date has only been identified in E. coli, and MgrB, which is found in numerous Enterobacteriaceae (18). SafA stimulates PhoQ and functions as a connector between the EvgS/EvgA and PhoQ/PhoP two-component systems (10). MgrB, on the other hand, represses PhoQ activity and, as it is part of the PhoP regulon, is part of a negative feedback loop (18). The function of this feedback loop, however, is not understood.

A protein that regulates a signal transduction pathway may provide an entry point for additional input signals, since factors that modulate the protein's activity will modulate the circuit output. We therefore looked for additional factors that affect PhoQ/PhoP signaling by screening E. coli transposon-insertion mutants for increased expression of a PhoP-regulated reporter gene. From this screen we found that disruption of pathways associated with disulfide bond formation in the periplasm leads to activation of the PhoQ/PhoP system.

Disulfide bond formation in the periplasm is catalyzed by the periplasmic protein DsbA and the membrane protein DsbB (reviewed in references 13, 15, and 23). DsbA oxidizes cysteine residues to form disulfide bonds through a thiol-disulfide exchange reaction. The reduced form of DsbA resulting from this reaction is reoxidized by the membrane protein DsbB. We found that deletion of dsbA results in activation of several PhoP-regulated genes in a PhoQ-dependent manner. In addition, deletion of dsbB as well as treatment with the reducing agent dithiothreitol (DTT) has a similar effect on PhoQ/PhoP signaling. Furthermore, we provide evidence suggesting this redox-sensitive pathway acts through MgrB.

To our knowledge, these results are the first demonstration that disruption of the oxidizing environment of the periplasm stimulates PhoQ. The results also raise the possibility that the periplasmic redox state may be a physiologically important input stimulus for the E. coli PhoQ/PhoP system whereby reducing conditions or possibly other factors that affect disulfide bonding status in the periplasm stimulate PhoQ by modulating the negative feedback loop mediated by MgrB.

MATERIALS AND METHODS

Strain and plasmid construction.

Strains used are derivatives of MG1655 (E. coli Genetic Stock Center, Yale University) and are listed in Table 1. Plasmids used in this work and the oligonucleotides used to construct plasmids or sequence transposon insertion sites are provided in Table 2 and Table 3, respectively.

Table 1.

Strains used in this study

Strain	Relevant genotype	Source, reference, or constructiona
MG1655	λ−rph-1	E. coli Genetic Stock Center no. 7740
JW3832	rrnB3 ΔlacZ4787 hsdR514 Δ(araBAD)567 Δ(rhaBAD)568 rph-1 (ΔdsbA)::kan	3
JW5182	rrnB3 ΔlacZ4787 hsdR514 Δ(araBAD)567 Δ(rhaBAD)568 rph-1 (ΔdsbB)::kan	3
TIM63	MG1655 λatt::(PmgrB-yfp)	22
TIM92	MG1655 λatt::(PmgrB-yfp) HKatt::(PtetA-cfp)	22
TIM100	MG1655 ΔphoQ λatt::(PmgrB-yfp) HKatt::(PtetA-cfp)	22
TIM148	MG1655 λatt::(PphoPQ-yfp cat) HKatt::(PtetA-cfp)	22
TIM229	MG1655 ΔphoQ λatt::(PphoPQ-yfp cat) HKatt::(PtetA-cfp)	22
AML20	TIM92 ΔmgrB::FRT	18
AML22	TIM148 (ΔmgrB)::kan	18
AML41	TIM92 (ΔdsbA)::kan	P1(JW3832) × TIM92
AML42	TIM100 (ΔdsbA)::kan	P1(JW3832) × TIM100
AML43	TIM92 (ΔdsbB)::kan	P1(JW5182) × TIM92
AML44	TIM100 (ΔdsbB)::kan	P1(JW5182) × TIM100
AML45	TIM92 (ΔrdoA)::kan	P1(JW3831) × TIM92
AML47	TIM148 (ΔdsbA)::kan	P1(JW3832) × TIM148
AML48	TIM229 (ΔdsbA)::kan	P1(JW3832) × TIM229
AML53	TIM148 ΔmgrB::FRT	AML22/pCP20
AML54	TIM229 ΔmgrB::FRT	18
AML55	TIM148 ΔmgrB::FRT (ΔdsbA)::kan	P1(JW3832) × AML53
AML56	TIM229 ΔmgrB::FRT (ΔdsbA)::kan	P1(JW3832) × AML54
AML65	TIM92 ΔdsbA::FRT	AML41/pCP20
AML92	TIM148 (ΔlacY)::aadA	P1(EAL181) × TIM148
AML93	AML22 (ΔlacY)::aadA	P1(EAL181) × AML22
AML95	AML55 (ΔlacY)::aadA	P1(EAL181) × AML55
EAL181	MG1655 (ΔlacY)::aadA	E. A. Libby and M. Goulian, unpublished data

^aP1(AAA) × BBB denotes P1 transduction from strain AAA into strain BBB. AAA/pCP20 denotes removal of kanamycin resistance from strain AAA by expression of FLP recombinase via pCP20 and subsequent curing of the plasmid.

Table 2.

Plasmids used in this study

Plasmid	Description	Citation
pTrc99a	lacIq; Ptrc-MCS; Ampr	2
pEB52	pTrc99a with NcoI site removed by cutting with NcoI, blunting with mung bean nuclease, and religation; Ampr	E. Batchelor and M. Goulian, unpublished data
pAL8	pEB52 Ptrc-mgrB; Ampr	18
pAL14	pEB52 Ptrc-mgrB(C16A); Ampr	This work
pAL15	pEB52 Ptrc-mgrB(C28A); Ampr	This work
pAL16	pEB52 Ptrc-mgrB(C39A); Ampr	This work
pAL47	pEB52 Ptrc-mgrB(C28A C39A); Ampr	This work
pAL52	pEB52 Ptrc-mgrB(C16A C28A C39A) [MgrB(3C→A)]; Ampr	This work
pAL54	pTrc99a Ptrc-6×His-mgrB; Ampr	This work
pAL59	pTrc99a Ptrc-6×His-mgrB(C16A C28A C39A) [6×HIS-MgrB(3C→A)]; Ampr	This work
pCP20	ori(Ts), FLP recombinase expression plasmid; Ampr Cmr	8
pRL27	oriR6K, Tn5-RL27; Kmr	17

Table 3.

Primers used in this study

Primer designation	Primer sequence (5′–3′)
P1	GCTAGAATTCTGACATAAGGTAGGTG
P2	AAAGGATCCTCACCACGGGATAAACTGG
P3	GTGGTGTTGGCTGCCTTGCTGCTTTGGGC
P4	GCCCAAAGCAGCAAGGCAGCCAACACCAC
P5	CGCAGGTATTCAACATGATGGCCGATCAGGATGTACAATTTTTCAGC
P6	GCTGAAAAATTGTACATCCTGATCGGCCATCATGTTGAATACCTGCG
P7	GGATGTACAATTTTTCAGCGGAATTGCTGCCATTAACCAGTTTATCCC
P8	GGGATAAACTGGTTAATGGCAGCAATTCCGCTGAAAAATTGTACATCC
P9	ATACCATGGGTCACCATCACCACCATCACATGAAAAAGTTTCGATGGGTCGTTCTGGTTGTCGTGG
P10	CATCCGCCAAAACAGCCAAG
P11	GACACAGGAACACTTAACGGC
P12	GAGCATTACGCTGACTTGAC

Mutations and reporter constructs were transferred between strains by transduction with P1_vir (19). Deletions transduced from strains in the Keio Collection (3) were confirmed by PCR using primers that flank the gene. When necessary, kanamycin resistance markers were removed with FLP recombinase by transforming with pCP20 (8) and subsequently curing the plasmid, as described in reference 9.

The plasmids pAL14, pAL15, and pAL16 contain point mutations that generate substitutions of alanine for cysteines C16, C28, and C39 of MgrB, respectively. The plasmids were constructed by overlap-extension PCR, using pAL8 as template and primers P1, P4, P3, and P2 (for pAL14); P1, P6, P5, and P2 (for pAL15); and P1, P8, P7, and P10 (for PAL16). The resulting PCR products were digested with EcoRI and BamHI and ligated to the EcoRI and BamHI sites of pEB52. pAL47 carries mgrB(C28A C39A) and was constructed by DpnI-mediated site-directed mutagenesis of pAL15 with primers P7 and P8 (26). pAL52 carries mgrB(C16A C28A C39A) (3C→A) and was constructed by site-directed mutagenesis of pAL47 using primers P3 and P4. pAL54 and pAL59 encode N-terminally 6×His-tagged (6×HIS) mgrB and mgrB(3C→A) proteins, respectively. The plasmids were constructed by PCR using primers P9 and P2 with templates pAL8 and pAL52, respectively. The resulting PCR products were digested with NcoI and BamHI and ligated into the NcoI and BamHI sites of pTrc99a.

Growth conditions.

Liquid cultures were grown at 37°C in minimal A medium (19) supplemented with 0.2% glucose, 0.1% Casamino Acids (Difco), and the indicated concentrations of MgSO₄. The trc promoter was induced with isopropyl β-d-1-thiogalactopyranoside (IPTG) at the concentrations indicated. When IPTG is not mentioned in the description of culture conditions, basal transcription from the trc promoter was used to drive expression.

Fluorescence images of agar plates.

Strains were grown overnight on LB Miller agar (Difco) plates at 37°C. Yellow fluorescent protein (YFP) and cyan fluorescent protein (CFP) fluorescence images were acquired with a fluorescence illuminator, as described previously (27). Images were adjusted for brightness and contrast using ImageJ (National Institutes of Health; http://rsbweb.nih.gov/ij/).

Transposon mutagenesis screen.

Insertional mutagenesis was performed on the mgrB transcriptional reporter strain TIM63 by using the Tn5-RL27 transposon derivative pRL27 (17), whose transposable element bears an oriR6k and a kanamycin cassette. Mutants that appeared to have increased YFP expression, as determined by fluorescence images of colonies on LB agar plates supplemented with 25 μg/ml kanamycin, were transduced into the P_mgrB-yfp transcriptional reporter strain TIM92, which also constitutively expresses CFP, as an internal control. Transductants were selected on medium containing 35 μg/ml kanamycin. If mutants retained increased YFP expression, plasmids containing transposon insertions were recovered as described in reference 5, except that the DNA was sheared by sonication. Plasmid DNA was sequenced using primers P11 and P12 to identify the location and orientation of the transposon insertion.

Fluorescence measurements.

Cellular fluorescence was measured by microscopy, essentially as described previously (18). Briefly, overnight cultures grown in minimal A medium with 1 mM MgSO₄, and with 50 μg/ml ampicillin to maintain plasmids when necessary, were diluted 1:1,000 into prewarmed (37°C) minimal medium containing 100 μM, 1 mM, or 10 mM MgSO₄, as indicated. When the reducing agent DTT was used, it was added to the prewarmed medium to a final concentration of 3 mM. Similarly, when the oxidizing agent CuSO₄ was used, it was added to a final concentration of 50 μM. Cultures were then grown to an optical density at 600 nm between 0.2 and 0.3 (approximately 4.5 h) and cooled quickly in an ice water slurry, and streptomycin was added to a final concentration of 250 μg/ml to inhibit further protein synthesis. Samples were imaged by fluorescence microscopy and analyzed as described in references 21 and 22. For each culture, fluorescence was determined from the mean fluorescence of at least 40 cells.

RESULTS

Screen for negative regulators of the PhoQ/PhoP system.

We screened for negative regulators of the PhoQ/PhoP two-component system by using transposon mutagenesis. The host strain contained the PhoP-regulated mgrB promoter driving YFP expression (P_mgrB-yfp) integrated at the phage lambda attachment site. Growth on LB is a moderately activating condition for the PhoQ/PhoP system (16), likely due to low concentrations of Mg²⁺ and Ca²⁺ in this growth medium (24). Therefore, for cells grown on LB, inactivation of a gene encoding a negative regulator should produce an increase in P_mgrB-yfp transcription (18). From approximately 12,000 kanamycin-resistant colonies, we identified 10 colonies that displayed increased YFP fluorescence on LB agar plates. We determined the insertion site for six of these mutants. Four proved to be independent insertions in dsbA; one was in the 3′ end of rdoA, a gene upstream of and in an operon with dsbA (6), and one was located between the promoter and start codon of mgrB. The remaining four mutants could be complemented with a plasmid expressing DsbA and therefore were not analyzed further. Deletion of rdoA (replacement with a kanamycin resistance gene by P1 transduction from a strain in the Keio collection [3]) had no effect on YFP fluorescence, whereas a similar deletion of dsbA resulted in a large increase in fluorescence (Fig. 1A). Deletion of dsbA in a strain containing a reporter for a different PhoP-regulated promoter (P_phoPQ-yfp) similarly showed an increase in fluorescence relative to the dsbA⁺ strain under both activating (100 μM) and repressing (10 mM) levels of Mg²⁺ (Fig. 1B). We also noted that deletion of dsbA in a ΔphoQ strain did not increase P_mgrB-yfp or P_phoPQ-yfp transcription (Fig. 1A and B). That deletion of dsbA leads to a PhoQ-dependent increase in expression for at least two PhoP-regulated genes suggests that DsbA may act at an upstream pathway common to the entire PhoP regulon.

Fig 1.

Deletion of dsbA increases transcription of mgrB and the phoPQ operon. The strains contained a chromosomal copy of the mgrB promoter (P_mgrB) (A) or the autoregulated phoPQ promoter (P_phoPQ) (B) driving yfp expression and a constitutive promoter driving cfp expression as an internal reference. (A) YFP (left) and CFP (right) fluorescence images of a transcriptional reporter strain with the indicated gene deletions, grown on LB agar. The strains are, clockwise starting with the wild type (WT), TIM92, AML20, AML41, AML42, AML43, AML44, and AML45. (B) Cells were grown in minimal glucose medium with 100 μM or 10 mM MgSO₄. For each condition, the means and ranges for two independent cultures are shown. The strains are, from left to right, TIM148, TIM229, AML47, AML53, AML55, and AML56.

Conditions that affect disulfide bonding in the periplasm regulate the activity of PhoQ/PhoP.

DsbA requires the oxidizing activity of the transmembrane protein DsbB for recycling reduced DsbA to its active form (13, 15, 23). We therefore tested the effect of deleting dsbB in a P_mgrB-yfp reporter strain and observed a phoQ-dependent increase in mgrB transcription that was similar to the behavior of a dsbA deletion (Fig. 1A). We also found that addition of the reducing agent DTT increased YFP fluorescence for P_mgrB-yfp (Fig. 2A) and P_phoPQ-yfp (Fig. 2B), which was similar to the behavior of dsbA null strains that were not treated with a reducing agent. These results are consistent with a requirement for periplasmic disulfide bond formation in a pathway that represses transcription of PhoP-regulated genes.

Fig 2.

The reducing agent DTT increases transcription of mgrB and the phoPQ operon. Reporter strains for transcription of the PhoP-regulated genes mgrB (A) and phoPQ (B) were grown in minimal glucose medium with 1 mM MgSO₄ with no added DTT or 3 mM DTT. For each condition, the means and ranges for two independent cultures are shown. Strains are, from left to right, TIM92, TIM100, and AML65 (A) and TIM148, TIM229, and AML47 (B).

The redox active metal ion Cu²⁺ rescues defects in periplasmic disulfide bond formation in dsbA null strains (12). To determine whether copper could similarly suppress the increased PhoP-regulated transcription associated with a dsbA deletion, we grew ΔdsbA strains in the presence of CuSO₄ and measured the effect on YFP fluorescence. Addition of as little as 50 μM CuSO₄ resulted in levels of transcription of phoPQ (Fig. 3) and mgrB (data not shown) that were indistinguishable from the corresponding levels in a dsbA⁺ strain grown in the absence of CuSO₄, suggesting that the increased transcription of the PhoP regulon results from a deficiency in periplasmic disulfide bonding.

Fig 3.

Copper rescues the transcriptional phenotype of a dsbA deletion, but not an mgrB deletion. Reporter strains for phoPQ transcription were grown in minimal glucose medium with 1 mM MgSO₄ and either no added CuSO₄ or 50 μM CuSO₄. For each condition, the means and ranges for two independent cultures are shown. Strains are, from left to right, TIM148, AML47, AML22, AML55, TIM229, AML48, and AML56.

The inhibitory effect of DsbA on PhoP-regulated transcription is mediated by MgrB.

The periplasmic domain of E. coli PhoQ does not have any cysteines. However, MgrB, a small 47-amino-acid membrane protein that negatively regulates PhoQ (18), contains three conserved cysteines (C16, C28, and C39) (Fig. 4A). C16 is predicted to be in the transmembrane domain, and C28 and C39 are predicted to be in the periplasm (18). In a P_phoPQ-yfp reporter strain, deletion of both mgrB and dsbA resulted in only a slight increase in YFP fluorescence compared with the fluorescence levels in strains containing either deletion alone (Fig. 1B). This suggests that DsbA repression of PhoQ/PhoP signaling may act through the same pathway as that of mgrB.

Fig 4.

DsbA affects PhoQ/PhoP signaling through the cysteine residues of MgrB. (A) Amino acid sequence of MgrB. The underlined portion indicates the predicted transmembrane domain. (B) Reporter strains for mgrB transcription are either wild type for chromosomal mgrB (TIM92) or carry a deletion (AML20) and contain either a control plasmid (pEB52), a plasmid expressing MgrB (pAL8), or a plasmid expressing an MgrB mutant with the following amino acid substitutions: C16A (pAL14), C28A (pAL15), C39A (pAL16), C28A and C39A (pAL47), or C16, C28A, and C39A (3C→A; pAL52). (C) Reporter strains for phoPQ operon transcription with either wild-type chromosomal mgrB (AML92), a deletion of mgrB (AML93), or a deletion of both mgrB and dsbA (AML95) were transformed with a control plasmid (pTrc99a), a plasmid expressing 6×HIS-MgrB (pAL54), or a plasmid expressing 6×HIS-MgrB(3C→A) (pAL59). For each condition, the means and ranges for two independent cultures are shown. Cells were grown in minimal glucose medium with 100 μM MgSO₄ and without IPTG, except for the strain harboring the 6×HIS-MgrB(3C→A) plasmid, which was grown in medium with 68 μM IPTG.

To explore the role of the cysteines in MgrB, we tested the effects of cysteine-to-alanine substitutions. MgrB, expressed from a plasmid at basal levels from the uninduced trc promoter, complemented a chromosomal deletion of mgrB, as determined by repression of PhoP-regulated transcription (18) (Fig. 4B). Substitution of C16 for an alanine in MgrB [MgrB(C16A)] similarly repressed reporter gene expression. MgrB(C28A), MgrB(C39A), MgrB(C28A C39A), and MgrB with all three cysteines replaced with alanine [MgrB(3C→A)], on the other hand, did not result in repression (Fig. 4B). However, when the trc promoter was induced with IPTG, these four MgrB C-to-A mutants showed significant repression (data not shown). We took advantage of this fact to determine whether DsbA exerts its effect on PhoP-regulated transcription through the cysteines of MgrB. We tested the effect of deletion of dsbA in a phoPQ reporter strain that lacked mgrB and contained either a plasmid expressing MgrB from the trc promoter without inducer, a plasmid expressing MgrB(3C→A) induced from the trc promoter with 68 μM IPTG, or an empty control plasmid. The inducing condition of 68 μM IPTG for MgrB(3C→A) was chosen because it was determined to give comparable levels of repression to that of (wild-type) MgrB without inducer. As expected, deletion of dsbA led to increased reporter transcription in the strain expressing MgrB (Fig. 4C). However, deletion of dsbA had no effect for the strain expressing MgrB(3C→A). This suggests that DsbA exerts its effect on the PhoQ/PhoP system by acting on MgrB or in a pathway upstream of MgrB. Consistent with this interpretation, we also found that Cu²⁺ failed to decrease reporter gene expression in a ΔmgrB strain as well as in a ΔdsbA ΔmgrB double mutant (Fig. 3).

DISCUSSION

The results presented here indicate that the redox state of the periplasm plays a critical role in the regulation of the E. coli PhoQ/PhoP circuit, with reducing conditions resulting in stimulation of this two-component system. Null mutations in dsbA or dsbB, or treatment with DTT, increased expression of PhoP-regulated genes in a PhoQ-dependent manner, while treatment with the oxidant Cu²⁺ suppressed the effects of a dsbA deletion. Taken together, these results suggest that disruption of one or more disulfide bonds in the periplasm stimulates the PhoQ/PhoP system.

We have also provided evidence that this redox sensitivity is mediated by the small membrane protein MgrB. In particular, expression of an MgrB mutant in which the three cysteines were replaced with alanines completely blocked the effect of a dsbA null mutation on PhoQ/PhoP signaling (Fig. 4C). There was, however, a small increase in PhoP-regulated transcription in a ΔdsbA ΔmgrB strain compared with a strain that had a deletion of only one of the two genes (Fig. 1B). This may indicate an additional effect of DsbA on PhoQ/PhoP signaling under conditions in which PhoP-P levels are in excess (due to the absence of MgrB). Our results are consistent with disulfide bonding by one or more MgrB cysteines playing a role in the repressive effect of MgrB on PhoQ. In addition, the behavior of specific MgrB cysteine-to-alanine mutants on PhoP-regulated transcription is consistent with the two cysteines in the periplasm (C28 and C39) forming either inter- or intramolecular disulfide bonds. The activity of MgrB mutants with C28A and/or C39A substitutions could be restored by increased expression of the mutants from an inducible promoter. This indicates that C28 and C39 are not absolutely required for MgrB to repress PhoQ, although these cysteine residues may be important for stabilizing a protein conformation or complex that is critical for MgrB's repressive action. Taken together, the results suggest a model in which DsbA affects PhoQ/PhoP signaling by oxidizing C28 and C39 of MgrB. At present, however, we cannot rule out alternative models in which these cysteines are in a reduced state and DsbA does not directly interact with MgrB. Further work will be required to establish the disulfide bonding status of the MgrB cysteines and to determine whether disulfide bonding of other proteins participates in this DsbA pathway.

We do not know whether E. coli encounters in its native environments conditions that are similar to the reducing conditions studied here. However, it is noteworthy that the oxidation potential of extracellular fluids in animals is quite variable and depends on numerous factors, including age, disease state and, for the gastrointestinal tract, diet (11). In addition, antimicrobial defenses may interfere with disulfide bond formation in the E. coli envelope through the production of reactive compounds that modify free sulfhydryls or disrupt disulfide bonding pathways. It is therefore possible that changes in periplasmic redox conditions or a related property of the cell envelope may be an important input for the sensor kinase PhoQ and may contribute to the ability of E. coli to adapt to environments in animal hosts.