![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Previous Article | Next Article 
Journal of Bacteriology, January 2005, p. 672-686, Vol. 187, No. 2
0021-9193/05/$08.00+0 doi:10.1128/JB.187.2.672-686.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Department of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada
Received 15 July 2004/ Accepted 4 October 2004
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Recent studies indicated the involvement of the Cpx envelope stress response in multiple aspects of bacterial physiology and the cellular response to stress. Bioinformatic approaches suggest that the Cpx pathway may regulate a host of additional genes with diverse functions, many of which are part of other stress response pathways (21, 22). Additionally, the Cpx envelope stress response was recently linked to adhesion processes. In stationary phase, the Cpx response is activated by and required for efficient attachment to abiotic surfaces (67). Further, the Cpx signal transduction pathway is involved in the assembly and regulation of at least one type of pilus.
P or Pap pili are attachment organelles expressed by strains of uropathogenic E. coli (UPEC). P pili are the prototypes of the chaperone-usher class of pili, which are all thought to be expressed on the cell surface by means of a similar assembly mechanism (91). P pili consist of multiple subunits that make up a thin tip fibrillum containing the adhesin anchored to a thick, helical rod (8, 32, 53). The subunits are initially synthesized in the cytoplasm and transported to the periplasm through the general secretory machinery. Upon entry into the periplasm, the subunits are met by a P pilus-specific chaperone protein which serves to prevent premature aggregation into toxic aggregates in the periplasm, assist folding, and foster assembly into the growing pilus at an outer membrane assembly platform or usher (3, 80, 81). The platform consists of a single, pilus-specific usher protein assembled into ring-shaped oligomeric complexes containing a central pore through which pilus subunits are secreted and incorporated into mature pili (23, 92).
During P pilus expression, some subunits leave the pathway (off-pathway) or fail to interact with the periplasmic chaperone, thereby forming in the periplasmic toxic misfolded aggregates that associate with the inner membrane (39, 47). In laboratory strains of E. coli, the expression of all of the pap genes required for P pilus expression induces the Cpx envelope stress response, which in turn activates the expression of at least two genes that encode products required for efficient P pilus assembly (47). DegP degrades the otherwise toxic, off-pathway aggregated subunits (47), while DsbA is required for the proper folding of several P pilus subunits and assembly proteins (43). In the absence of the Cpx pathway, laboratory strains of E. coli that are provided with the pap gene cluster synthesize aberrant P pili that are far shorter than those synthesized by the isogenic wild-type strain (40). In addition, phosphorylated CpxR binds to the control region of the pap operon that mediates the phase-variable expression of these structures, indicating that Cpx regulators also act to affect the transcription of the pap operon (40). Accordingly, at least in laboratory strains of E. coli, one role of the Cpx pathway is to enable the proper expression of P pili.
Interestingly, a number of Cpx-regulated Dsb and DegP homologues have been implicated in the assembly of envelope-localized virulence determinants in a variety of pathogens. These include type IV pili (TFP) in Vibrio cholerae (70) and Neisseria meningitidis (93), type III secretion systems (TTSS) in Shigella flexneri (99, 107-109) and Salmonella enterica serovar Typhimurium (27, 60), and virulence determinants in Burkholderia cepacia (37), Erwinia carotovora (97), Pseudomonas aeruginosa (7, 36, 58), Bordetella pertussis (89), S. enterica serovar Typhimurium (27, 46), S. flexneri (73), and Yersinia pestis (42, 101). Further, DegP homologues have been implicated in virulence in Klebsiella pneumoniae (11) and Yersinia enterocolitica (57) and shown to be involved in the intracellular survival of Bartonella henselae (78), Y. enterocolitica (105), Legionella pneumophila (69), S. enterica serovar Typhimurium (4, 29), and Haemophilus influenzae (13). In addition, an E. coli K1 strain was defective in causing systemic disease in an infant rat model when the dsbA gene was obliterated (33). Thus, it seems possible that the Cpx envelope stress response plays a role in the expression of diverse types of envelope structures that share a common need for Cpx-regulated folding and/or degrading factors. In this study, we have tested this hypothesis by analyzing the role of the Cpx envelope stress response in the expression of the type IV bundle-forming pili (BFP) of enteropathogenic E. coli (EPEC).
EPEC is the leading cause of human infantile diarrhea (65). Although the actual causative mechanism behind the diarrheal symptoms is not clear, the events comprising infection have been well characterized at the molecular level. Initially, clumps, or microcolonies, of EPEC held together through interactions between BFP on adjacent cells loosely adhere to intestinal epithelial cells in a pattern known as localized adherence (25, 31). This event is followed by the injection of virulence determinants (E. coli secreted proteins) through a TTSS made up of E. coli secreted proteins and E. coli secretion complex proteins into the host cell cytoplasm (28, 44, 51, 52, 98, 103) and more intimate adherence mediated by the bacterial outer membrane protein intimin and an EPEC-encoded receptor known as the translocated intimin receptor (50). The translocated intimin receptor and other TTSS-translocated effector proteins mediate gross changes in the cytoskeleton of the eukaryotic cell that lead to the formation of actin-rich pedestals beneath tightly adhering EPEC bacteria, resulting in the disruption of the host cell surface (for reviews, see references 9, 26, 30, 65, and 95).
In this study, we have examined the effect of the Cpx envelope stress response on the expression of the type IV BFP. The TFP, including the BFP of EPEC, have different topologies and are thought to utilize an assembly mechanism very different from that of the chaperone-usher-type pili (104). They consist predominantly of a single pilin subunit and are thought to be assembled from pools of pilin subunits in the inner membrane by pilus-specific assembly complexes that span the entire envelope. Genetic and biochemical studies of the 14 genes found in the bfp gene cluster (90) suggest that the BFP assembly apparatus spans the envelope and is comprised of distinct subassembly complexes made up of BFP proteins that reside in the inner and outer membranes and are connected by a periplasmic component containing BfpU (1, 41, 77, 83). Recent work suggests that the inner membrane subassembly complex consists of the integral inner membrane proteins BfpE (6) and BfpC, together with the putative cytoplasmic ATPases BfpD and BfpF (88, 90), which are proposed to power pilus extension and retraction (2, 5, 15, 77), respectively. BfpB is an outer membrane secretin (82) which, together with BfpG, constitutes the outer membrane subassembly complex (77). The bfp genes that are sufficient for expression and assembly are carried on a large plasmid known as the EPEC adherence factor (EAF) plasmid (25, 31) and are subject to transcriptional regulation by the BfpTVW/PerABC regulatory locus and a variety of environmental signals (59, 72, 94).
Curiously, although a cluster of 14 bfp genes was previously shown to be sufficient for BFP expression in E. coli laboratory strain HB101 (90), we found that the same gene cluster failed to permit BFP assembly on the surface of laboratory strain MC4100 unless the Cpx pathway was constitutively activated by the presence of cpxA* gain-of-function mutations. This observation led us to investigate the role of the Cpx envelope stress response in the expression of BFP in an EPEC background. Our experiments indicate that an intact Cpx pathway is required for efficient BFP expression and attachment to epithelial cells. This is the first study to address the role of the Cpx envelope stress response directly in a pathogenic isolate of E. coli.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Antibiotics were purchased from Sigma and used at the following concentrations: ampicillin, 100 µg/ml; kanamycin, 50 µg/ml; amikacin, 3 µg/ml; spectinomycin, 20 µg/ml; and tetracycline, 25 µg/ml.
Construction of E2348/69 cpxR::Kn mutants (ALN88, TR1042, and TR1043).
The cpxR gene was amplified from the chromosome of MC4100 by using restriction enzyme-tagged primers finpho (5'-CATTAACAGGATCCTGTTCGTGCC -3') and cpxR3'Eco (5'-CGGAATTCCGGTTAAGCTGCCTATCATG -3') (underlining indicates the tags). The PCR product was purified by using spin columns according to the manufacturer's recommendations (Qiagen), digested with EcoRI and BamHI (Invitrogen), and cloned into the same restriction sites in pUC19. A kanamycin resistance cassette was excised from pUC4K (New England Biolabs) by using restriction enzyme SalI (Invitrogen) and cloned into the unique XhoI site in cpxR by standard techniques (79). The resultant cpxR::Kn fragment was excised from the plasmid by using BamHI and EcoRI and purified by using a GeneClean kit (QBIOgene) according to the manufacturer's instructions. The mutated cpxR::Kn allele was used to electroporate E2348/69 carrying plasmid pKD46 or pTP223, each of which encodes the 
Red recombination system, as previously described (62, 106). Transformants resistant to kanamycin were selected, and the Red-encoding plasmids were cured. Plasmid pKD46 was cured by virtue of its temperature-sensitive origin of replication, while transformants that had spontaneously lost plasmid pTP223 were screened by virtue of the loss of tetracycline resistance (62, 106). The newly generated mutation was confirmed by analysis of PCR products derived from amplification of the cpxR gene with primers that flanked the point of insertion, Western blot analysis (see below) and, in some experiments, Southern analysis with standard techniques and an AlkPhos direct nonradioactive labeling kit (Amersham) according to the manufacturer's instructions.
ß-Galactosidase assays. ß-Galactosidase assays were performed as previously described (76, 86).
Western blot analysis. Whole-cell lysates were made by resuspending cell pellets derived from 1 ml of culture in 50 µl of 2x sodium dodecyl sulfate-polyacrylamide gel electrophoresis loading buffer (125 mM Tris [pH 6.8], 20% glycerol, 10% ß-mercaptoethanol, 6% sodium dodecyl sulfate, 0.2% bromophenol blue). Lysates (10 µl) were subjected to electrophoresis and Western blotting as previously described (75). The blots were reacted with a 1:30,000 dilution of rabbit polyclonal antibody directed against bundlin (a kind gift from M. Donnenberg) or a 1:5,000 dilution of rabbit polyclonal antibody directed against a maltose-binding protein (MBP)-CpxR fusion protein (75), followed by incubation with a 1:20,000 dilution of secondary anti-rabbit immunoglobulin G-alkaline phosphatase conjugate (Sigma). The blots were developed with an enhanced chemiluminescence kit according to the specifications of the manufacturer (Bio-Rad). Proteins analyzed by Western blotting were quantified with an AlphaEase software package and a FluorChem IS-5500 imaging system (Alpha Innotech, Fisher Scientific). The densities along each lane of the Western blot were measured by using the 1D-Multi autogrid function. The peak area corresponding to the level of CpxR, BfpA, or a nonspecific loading control was normalized to either E2348/69 or E2348/69(pBAD18), which was arbitrarily assigned a value of 100.
Autoaggregation assays. Autoaggregation assays were performed as previously described (1) but with the following modifications. Briefly, overnight cultures were subcultured 1:50 in either LB broth containing 1 mM IPTG or DMEM/F-12 containing 0.1 M Tris (pH 7.4) and appropriate antibiotics and grown at 37°C with shaking to an OD600 of 0.2 to 0.6 (3 to 4 h). The OD600 of each culture was measured, the samples were allowed to remain at room temperature for 5 min and then were vortexed for 2 to 5 min, and the OD600 was remeasured (OD600V). The aggregation index was calculated as [(OD600V – OD600)/OD600] x 100. Each assay was done in triplicate and repeated at least twice.
Transmission electron microscopy. E2348/69 and MC4100 derivatives were grown as described above for BFP expression. One drop of culture was placed on a 300-mesh Formvar-carbon-coated copper grid and allowed to dry. The grids were stained with 0.5% phosphotungstic acid in phosphate-buffered saline and examined at a magnification of x14,000 by using a Morgagni 268 Philips transmission electron microscope. Three grids were analyzed for each strain, and representative images are shown.
Localized-adherence assays. Analysis of the localized-adherence phenotype exhibited by E2348/69 and ALN88 was performed as previously described (96).
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In a wild-type MC4100 background, we found that the overexpression of bundlin (BfpA), the major subunit of BFP, increased the expression of ß-galactosidase from the spy-lacZ promoter approximately fivefold in a CpxR-dependent fashion (Fig. 1, compare lanes 1 and 2 to lanes 4 and 5). Similarly, in a strain carrying the cpxA101 allele, which consitutively activates the Cpx signal transduction pathway but is still sensitive to inducing cues (76), the overexpression of BfpA alone caused increased expression of spy-lacZ (Fig. 1, compare lanes 7 and 8). In the absence of the bfp genes required for the processing and incorporation of bundlin or BfpA into the mature pilus, BfpA accumulates in an unprocessed form in the inner membrane (1, 77, 90). Consequently, these observations suggest that the Cpx signal transduction pathway is sensitive to the accumulation of prebundlin in the inner membrane.
|
Expression of BFP in MC4100 requires constitutive activation of the Cpx envelope stress response. To determine whether the bfp genes present on pKDS301 and pKDS302 were expressed in the strain MC4100 background, we performed Western blot analysis of whole-cell lysates with antibody directed against bundlin or BfpA (Fig. 2A). As expected, in MC4100(pKDS301), which carries bfpA but lacks most of the bfp gene cluster, including the genes for the processing peptidase, the expression of unprocessed prebundlin was detected (Fig. 2A, lane 2). However, in MC4100 transformants carrying pKDS302, which carries the entire bfp gene cluster, we could not detect any BfpA-specific product, either processed or unprocessed (Fig. 2A, lane 3). These observations suggest that the presence of pKDS302 is unable to confer BFP elaboration in an MC4100 background.
|
To rule out the possibility that a chance mutation or rearrangement of plasmid pKDS302 was responsible for our observations, we performed the same experiments with strain HB101 and were able to detect processed bundlin by Western blotting and aggregation of transformants in liquid cultures, as previously described (data not shown). Accordingly, we speculated that the lack of expression of BFP in MC4100 transformants carrying pKDS302 was due to the strain background. Since the Cpx response controls the expression of folding factors that are involved in the efficient assembly of P pili (40), we hypothesized that the increased activity of the Cpx pathway might enable BFP assembly in the strain MC4100 background.
To test this idea, we transformed pKDS301 and pKDS302 into a variety of cpx* backgrounds in which the Cpx envelope stress response is constitutively activated due to mutations in the cpxA gene, encoding the sensor kinase (76). The resulting transformants were subjected to anti-BfpA immunoblot analysis and aggregation assays to test for the presence of BFP. Western blot analysis indicated that, in the presence of four different cpx* mutations, mature, processed bundlin could be detected in strains carrying the entire bfp gene cluster (Fig. 2A, lanes 6, 9, 12, and 15). Further, these strains displayed aggregation in liquid cultures (Fig. 2B, lanes 6, 9, 12, and 15), indicating that BFP are expressed on the surface of cpx* mutants carrying plasmid pKDS302.
To confirm that BFP were expressed on the cell surface of cpx* mutants carrying pKDS302, we performed transmission electron microscopy on negatively stained preparations of an MC4100 cpxA101(pKDS302) transformant (Fig. 2C). We observed characteristic BFP on three of three grids examined (Fig. 2C), while no BFP were seen on grids fixed with negative control strain JPN15, which lacks the BFP-encoding EAF plasmid (see Fig. 4B). Together, these data suggest that BFP are expressed in MC4100 only when the Cpx pathway is constitutively activated. Further, when BFP expression is enabled in this fashion, it is sensed by the Cpx pathway, as indicated by the activation of the spy-lacZ fusion (Fig. 1, lane 9).
|
To ensure that this technique worked as expected with EPEC, we analyzed the resultant E2348/69 cpxR::Kn mutants by PCR (data not shown) and Western blot analysis (Fig. 3A). PCR analysis with primers that flanked the point of the kanamycin resistance cassette insertion showed that the selected mutants carried an insertion of the expected size in the cpxR gene (data not shown). Similarly, Southern blot analysis of one of the mutant chromosomes showed bands of the expected sizes when a cpxR probe was used for hybridization (data not shown). Finally, we performed Western blot analysis of whole-cell lysates of mutant and wild-type E2348/69 strains with polyclonal antisera directed against CpxR to confirm that we had eliminated CpxR from the mutant strains (Fig. 3A, top). As anticipated, the band corresponding to the predicted molecular weight for CpxR was absent from whole-cell lysates of the newly constructed E2348/69 cpxR::Kn mutants (Fig. 3A, top, compare lane 1 with lanes 2 to 4), indicating that we had successfully disrupted cpxR in these mutants.
|
To determine whether the diminished levels of processed bundlin associated with whole cells in the cpxR mutants reflected lower levels of cell surface-associated BFP, we performed aggregation assays (Fig. 3C). In agreement with the results of the Western blot analysis, we found that in liquid cultures, our three independently isolated E2348/69 cpxR::Kn mutants aggregated weakly relative to the wild-type parent strain (Fig. 3C, compare lane 1 with lanes 2 to 4), which displayed the formation of large, plentiful bacterial aggregates that were visible with the naked eye (data not shown). We observed that, similar to the BfpA levels detected by Western blot analysis, autoaggregation was diminished approximately threefold in the E2348/69 cpxR::Kn mutants relative to the wild-type parent strain (Fig. 3C).
Since Western blot analysis and aggregation assays suggested that BFP were expressed at much lower levels in the E2348/69 cpxR::Kn mutants, we examined one of the mutants, ALN88, together with parent strain E2348/69 and negative control strain JPN15 by transmission electron microscopy after negative staining for the presence of BFP on the cell surface (Fig. 4). In the wild-type E2348/69 background, BFP were present in thick, rope-like structures between bacteria on three of three grids examined (Fig. 4A). Conversely, in JPN15, although flagella and hair-like structures that were likely type I fimbriae were observed, no BFP were present (Fig. 4B). Suprisingly, even though we could detect bundlin by Western blotting, no structures reminiscent of BFP were observed on E2348/69 cpxR::Kn cells on three of three grids examined (Fig. 4C).
To confirm that the absence of BFP observed by transmission electron microscopy and the diminished levels of bundlin or BfpA revealed by Western blot analysis and autoaggregation assays were due to the absence of CpxR, we cloned MC4100 cpxR downstream of the inducible PBAD promoter on pBAD18. The resulting CpxR overexpression plasmid, pROX, was transformed into one of the E2348/69 cpxR::Kn mutants. Surprisingly, even though the PBAD promoter is tightly regulated by arabinose in laboratory strains of E. coli (35), we observed CpxR expression from plasmid pROX in the absence of arabinose (Fig. 5A and B). Thus, it appears that the regulation of PBAD by arabinose is somewhat leaky in E2348/69. At present, we do not know the basis for this leaky expression phenotype. Although the tissue culture medium that we used (DMEM/F-12) is not reported to contain arabinose, it is possible that there are trace amounts of arabinose in this medium that account for the expression that we observed. Alternatively, perhaps PBAD regulation occurs in a different manner in an EPEC strain background. Regardless, our data demonstrate that plasmid pROX permits the restoration of CpxR levels in an E2348/69 cpxR::Kn background in the absence of arabinose.
|
Elimination of the Cpx envelope stress response has deleterious effects on localized adherence. BFP mediate the distinctive initial stage of attachment during EPEC infection that is known as localized adherence (25, 31). Localized adherence is the attachment of microcolonies of multiple bacteria held together by BFP to the surface of epithelial cells. Strains lacking BFP are defective in this stage of infection and in causing disease in human volunteers (56). Since the elimination of CpxR affected the production of BFP by EPEC grown in cultures (Fig. 3, 4, and 5), we examined whether localized adherence might also be altered. Wild-type and cpxR E2348/69 strains were incubated with HEp-2 cells and stained, and the number and size of microcolonies per cell were quantified. Visually it was obvious that, even though the same numbers of bacteria were used in the experiments, the cpxR mutant exhibited fewer microcolonies per epithelial cell and fewer microbes within each microcolony than did the wild-type parent (Fig. 6). Statistical analysis confirmed these observations, revealing an approximate 33% decrease in the number of E2348/69 cpxR::Kn microcolonies per epithelial cell relative to those of the wild-type E2348/69 parent (Table 2). Further, the microcolonies formed by the cpxR mutant were less than half the size of those formed by the wild-type E2348/69 parent (Table 2). Thus, elimination of the Cpx envelope stress response has a significant effect on the first step in EPEC pathogenesis, attachment of microcolonies to epithelial cells.
|
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
BFP assembly intermediates are sensed by the Cpx signal transduction pathway. In this study, we discovered that the Cpx envelope stress response is activated by the overexpression of unprocessed BfpA as well as mature, surface-expressed BFP (Fig. 1). In laboratory strain MC4100, expression of unprocessed BfpA leads to more than a fivefold increase in the Cpx-dependent activation of spy expression. Further, when the Cpx response is constitutively activated by mutation, the Cpx signal transduction pathway is further activated by expression of the entire bfp gene cluster (Fig. 1). This result indicates that other assembly intermediates, beyond unprocessed BfpA, are sensed by the CpxA sensor kinase. Currently, we do not know the nature of the CpxA activating signal. It was recently shown that CpxA recognizes the off-pathway P pilus subunit PapE in the absence of the cognate PapD chaperone in a manner that requires the PapE N-terminal extension that mediates the assembly of PapE into the pilus fibrillum (54). Since Cpx activation by PapE did not correlate with protein stability, subunit association, or levels of protein, it has been proposed that the N-terminal extension is involved in facilitating a folding intermediate in the absence of the PapD chaperone that specifically activates the Cpx stress response (54). Although the TFP share no similarities with chaperone-usher-type pili at the level of structure or assembly, all TFP subunits contain a hydrophobic N-terminal alpha helix that mediates subunit-subunit interactions to promote fiber formation (14, 38, 48, 68). In addition, electrostatic and hydrophobic interactions between the globular head domains of pilin subunits are also important in TFP architecture (14). Thus, by analogy with PapE, perhaps these domains contribute to a distinct folding intermediate in the absence of other BFP assembly components. This intermediate might be generated when BfpA is expressed alone or when BFP elaboration is up-regulated, leading to activation of the sensor kinase CpxA and the production of factors that promote efficient BFP expression. Expression of the bfp structural and assembly genes are regulated by a multitude of signals thought to reflect the host environment via a regulatory factor consisting of the perABC and bfpTVW gene products (59, 72, 94). Accordingly, upon initiation of an infection, increased elaboration of BFP may be the trigger that activates the Cpx response, reinforcing BFP expression and adherence.
The CpxA sensor may also recognize distinct signals from PerABC and BfpTVW and/or the accumulation of bundlin intermediates in vivo. Alterations in pH, which are encountered by EPEC in transit through the intestinal tract, are known to activate the Cpx response (17, 64). Further, recent experiments showed that the Cpx signal transduction cascade senses and mediates attachment to hydrophobic surfaces via the lipoprotein NlpE (67). Thus, another in vivo signal that may activate the Cpx response could be contact with hydrophobic host cell surfaces in vivo. In this respect, initial BFP binding of host cells might lead to further induction of the Cpx response via contact with hydrophobic host cell surfaces. Interestingly, a number of studies on TFP expression have implicated chemotaxis-related signal transduction proteins in this process (for a review, see reference 102). Part of the Cpx response involves the inhibition of motility and chemotaxis genes (21; unpublished data). In consequence, the Cpx envelope stress response may exert effects on BFP assembly and adherence via this route as well.
The Cpx envelope stress response is involved in the assembly of diverse envelope-localized structures. The Cpx envelope stress response was recently shown to sense and assist in the assembly of UPEC P pili in laboratory strain MC4100 (40, 47). P pili are multisubunit pili (8, 32, 53) that are elaborated by means of a specific periplasmic chaperone, PapD, and outer membrane secretion platform, or usher, formed by the secretin PapC (3, 80, 81). CpxA is thought to detect off-pathway PapG and PapE subunits that fail to associate with PapD and thus become misfolded (47, 54). The Cpx-regulated gene products DsbA (43) and DegP (47) are involved in P pilus biogenesis, and the abolition of CpxR leads to a short-pilus phenotype (40). Consequently, activation of the Cpx envelope stress response upon pilus expression appears to be an important mechanism in ensuring efficient assembly of the P pilus, likely by providing necessary folding and degrading factors. TFP structure and assembly are significantly different from those of the chaperone-usher class of pili (104). Recent studies suggest that TFP are elaborated from inner membrane pools of the mature pilus subunit via distinct subassemblies of biogenesis proteins that span the envelope (1, 15, 41, 77, 104). Despite extensive mutational analysis of the EPEC BFP biogenesis machinery, no periplasmic assembly intermediates have been identified, and the pilus is thought to originate in the inner membrane and form a structure that spans the entire envelope (1, 77). Accordingly, since the P pilus assembly requirement for the Cpx pathway is thought to reflect the need for folding and degrading factors that act on periplasmic assembly intermediates, it is somewhat surprising that a similar requirement appears to exist for the BFP of EPEC.
Our data clearly suggest that this is the case. In laboratory strain MC4100, BFP cannot be elaborated, even when the necessary genes are expressed from an exogenous promoter that is highly expressed in this strain background, unless the Cpx envelope stress response is constitutively activated (Fig. 2). Intriguingly, although no periplasmic assembly intermediates have been identified for BFP, DsbA is absolutely required for biogenesis (110). The first step in the assembly pathway is simultaneous cleavage of the prebundlin leader by the BfpP peptidase and introduction of an essential C-terminal disulfide bond, which is catalyzed by DsbA. Thus, it is possible that the Cpx requirement reflects a need for elevated levels of DsbA. Alternatively, additional Cpx-regulated assembly factors may also be required. For example, DegP is required in P pilus biogenesis to degrade toxic, off-pathway subunits (47). At this time, we are not certain what part of the assembly process is affected by the Cpx response. Prebundlin is clearly made in the absence of other Bfp biogenesis factors (Fig. 2); however, no bundlin, mature or otherwise, is detectable when the entire bfp biosynthetic gene cluster is present in laboratory strain MC4100 (Fig. 2). Hence, in the absence of elevated levels of a critical Cpx-regulated factor(s), some step after prebundlin production is prevented, leading to the degradation of bundlin and/or the assembly apparatus. In an EPEC strain background, reduced levels of bundlin are seen when the Cpx signal transduction pathway is removed (Fig. 3). Extensive mutational analysis indicates that cell-associated bundlin levels are unaltered in the presence of any one of a number of bfp mutations (except bfpP) (1, 77), suggesting that the effects of the Cpx envelope stress response on bundlin levels may be due to altered levels of folding and/or degrading factors that work directly on bundlin. Alternatively, the Cpx response may affect BfpP stability or multiple Bfp biogenesis proteins found in one or more of the envelope-localized subassemblies that are thought to mediate BFP elaboration. Recently, it was shown that the elimination of BfpE leads to reduced levels of bundlin (77). BfpE is thought to form an inner membrane scaffold required for the movement of assembly proteins across the inner membrane (77). Since the phenotype of the cpxR mutant with respect to bundlin levels is similar, it is possible that the Cpx response affects the levels or function of BfpE in EPEC.
Curiously, even though we detected cell-associated mature bundlin and weak aggregation in the EPEC cpxR mutant (Fig. 3), we were not able to visualize BFP by electron microscopy (Fig. 4). It is tempting to speculate that, as with P pili, the cpxR mutant elaborates shorter BFP that are not visible by transmission electron microscopy but still function weakly in aggregation and localized adherence (40). Alternatively, the residual localized adherence that we saw may reflect non-BFP adhesins (10, 66).
At this point, we cannot say whether the Cpx response affects the transcription of the bfp genes. In MC4100, the Cpx pathway must be constitutively active to permit BFP expression (Fig. 2). Since the bfp genes in this experiment were expressed from an exogenous, non-Cpx-regulated promoter in MC4100, some, if not all, of the effects of the Cpx envelope stress response on BFP expression must occur at posttranscriptional steps. Given the role of the Cpx envelope stress response in P pilus assembly, we favor the idea that the posttranscriptional effects occur at the level of assembly, likely facilitated by Cpx-regulated folding factors. However, since our experiments in the EPEC strain background all examined BFP expressed from the native promoter, we cannot say whether the effects of eliminating the Cpx response in this background are partly due to effects on bfp transcription. Experiments are under way to address this question.
Whatever the mechanism, our data clearly show that the Cpx envelope stress response influences the biogenesis of type IV BFP of EPEC in addition to chaperone-usher-type pili like the UPEC P pilus. Since these pili have very different structures and assembly pathways, it will be interesting to determine whether the Cpx stress response plays a role in the elaboration of other, varied envelope-localized virulence determinants in E. coli.
An EPEC cpxR mutant is only partially complemented by a plasmid encoding CpxR. The bfp phenotype of our EPEC cpxR::Kn mutants was only partially complemented by a plasmid encoding CpxR (Fig. 5). The EPEC cpxR mutant transformed with a plasmid carrying an intact cpxR gene expressed levels of bundlin and autoaggregation that were intermediate between those of the parent E2348/69 and the mutant E2348/69 cpxR::Kn (Fig. 5). There are several possible explanations for why we observed only partial complementation of BFP expression by pROX in our EPEC cpxR mutant. First, since cpxR is found upstream of cpxA in an operon, the cpxR::Kn mutant that we created disrupts both cpxR and cpxA expression. Therefore, it is possible that full complementation of BFP expression requires both CpxR and CpxA. Efforts to clone both genes together in the past have been unsuccessful (T. L. Raivio, unpublished data). We believe that this situation reflects the deleterious effects of the hyperactivation of the Cpx pathway, since the constitutively active cpx* mutants display many pleiotropic phenotypes (12). Perhaps in the absence of CpxA, the CpxR expressed from our complementing plasmid is not sufficiently phosphorylated to confer full complementation. Alternatively, since the cpxRA operon is autoregulated, it is possible that in uncoupling cpxR expression from its native promoter, our complementation plasmid does not express CpxR at the appropriate levels or time for full complementation. Finally, we used the cloned MC4100 cpxR gene in our complementation experiments. Our sequence analysis of the E2348/69 cpxRA operon indicates that there is a single amino acid difference between the MC4100 and E2348/69 cpxR genes (A. Z. Nevesinjac and T. L. Raivio, unpublished data). Accordingly, it is also possible that the lack of complete complementation reflects differences in the activities of the MC4100 and E2348/69 CpxR proteins. Whatever the reason, our complementation experiments clearly show that partial restoration of CpxR levels permits limited reinstatement of BFP expression. Plainly, the level, phosphorylation status, and/or presence of CpxA are important factors in the effect of Cpx on BFP elaboration.
The Cpx envelope stress response affects early events required to initiate EPEC infection. Although other adhesins have been implicated (10, 66), BFP are thought to be the major facilitators of localized adherence to epithelial cells, a diagnostic trait of EPEC infections. Human volunteer studies show that mutants that fail to make BFP are much less effective at initiating an infection (56), emphasizing the importance of this initial step in the EPEC disease process. Our studies show that the Cpx envelope stress response is needed for the optimal expression of BFP (Fig. 3 and 4) and localized adherence (Table 2 and Fig. 6) in an EPEC background. Thus, it is probable that the activation of the Cpx envelope stress response optimizes early events in EPEC infection.
Interestingly, a role for the Cpx response in the regulation of virulence in the closely related microbe S. flexneri has been documented. Phosphorylated CpxR directly activates the transcription of the virF gene, a master regulator of virulence in Shigella spp. (63). Further, DsbA and DegP homologues are involved in the expression of TFP, TTSS, virulence, or intracellular survival in numerous pathogens (see above). These studies, together with our observations, suggest that the role of the Cpx envelope stress response in the expression of virulence determinants extends to other pili, envelope-localized virulence determinants, and regulatory proteins. Thus, as for many other stress responses, the Cpx stress response appears to play multiple roles within the cell. In addition to the regulation of virulence determinant expression, as shown here, these include adaptation to potentially lethal envelope stresses (12, 17), regulation of conjugation (34, 85), surface sensing (67), and likely others, as indicated by bioinformatics approaches (22).
BFP expression is strain dependent. Curiously, although the bfp genes carried on pKDS302 are sufficient to confer BFP elaboration in laboratory strain HB101 (90), the same plasmid fails to permit BFP expression in MC4100 unless the Cpx signal transduction pathway is constitutively activated by mutation (Fig. 2). In addition, BFP are elaborated by the EPEC type strain E2348/69 without any manipulation of the Cpx signal transduction cascade. These observations suggest that there are fundamental differences between these strains. For EPEC, it is easy to imagine that additional factors not found in laboratory strains permit BFP expression; however, the finding that HB101 can assemble these cell surface structures confounds this idea. An alternative explanation is that there may be differences in the basal levels of activity of the Cpx signal transduction pathway in these strains, such that HB101 and EPEC contain Cpx responses that are inherently more active.
Alternative explanations for the strain differences in BFP expression exist. Perhaps some critical Cpx-regulated factor is expressed at higher levels in HB101 and EPEC, permitting the assembly of BFP in these backgrounds but not in that of MC4100. The differences between the MC4100 and EPEC strain backgrounds could also reflect the fact that in the MC4100 background, we are observing only posttranscriptional effects, since the bfp genes are expressed from the exogenous trc promoter, while in EPEC, the effects of the Cpx response on BFP expression may be the result of both transcriptional and posttranscriptional events. Perhaps Cpx-mediated transcriptional events at the bfp locus lead to a different outcome in EPEC. Studies are currently under way to distinguish these possibilities.
Conclusions. Taken together, our data demonstrate that the Cpx envelope stress response is able to sense BFP assembly and is involved in the expression of these structures, at least at the posttranscriptional level. Additionally, we have shown that the Cpx response plays a role in maximizing the first step in EPEC infection, localized adherence to epithelial cells. This study allows us to generalize findings with regard to the role of the Cpx signal transduction pathway in UPEC P pilus assembly to other, distinct envelope structures with much different assembly pathways.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by a scholar award (T.L.R.) from the Alberta Heritage Foundation for Medical Research and an operating grant from the Canadian Institutes of Health Research.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
E and the Cpx signal transduction systems control the synthesis of periplasmic protein-folding enzymes in Escherichia coli. Genes Dev. 11:1183-1193.
