![[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}
Department of Microbiology, University of Guelph, Guelph, Ontario, Canada N1G 2W1
Received 21 May 2002/ Accepted 26 August 2002
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionAddendum in proofReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAddendum in proofReferences |
|---|
E. coli group 1 CPS are assembled via a Wzy-dependent pathway. The current model (reviewed in reference 41) is based extensively on evidence gathered from the parallel system for lipopolysaccharide (LPS) O-antigen assembly, in which repeat units are assembled on undecaprenol phosphate at the cytoplasmic face of the inner membrane by the sequential activity of glycosyltransferases. The repeat units are then believed to be flipped to the periplasmic face of the inner membrane by an unknown mechanism involving the Wzx protein. Polymerization of the lipid-linked repeat units is then carried out by Wzy, the polymerase. In E. coli K30, the K antigen can undergo one of two fates at this stage. High-level polymerization can occur to generate high-molecular-weight (HMW) CPS, which is then translocated to the cell surface (K30CPS). Alternatively, one to a few repeat units of the K antigen can be ligated onto lipid A-core and expressed on the cell surface as KLPS (11, 25).
In the prototype K30 strain, E69, the genes responsible for the synthesis and cell surface assembly of the K30 polymer are found in an operon located near his and rfb (42). The products of the first four genes of the K30 operon (orfX, wza, wzb, and wzc) are believed to be involved in high-level polymerization and surface expression of the K30CPS polymer. These genes are separated from a block of downstream genes encoding enzymes for repeat unit synthesis by a transcriptional attenuator (32; A. Rahn and C. Whitfield, unpublished data). The OrfX protein has only one known homologue, found in Klebsiella K2 strains (ORF3), but its exact function is unknown (1, 2). Wza is an outer membrane lipoprotein that multimerizes to form ring-like structures resembling secretins for type II and type III protein secretion (13). HMW CPS is believed to cross the outer membrane through this complex. WzcCPS (involved in group 1 CPS assembly) is a tyrosine autokinase, while WzbCPS is its cognate phosphatase. The biochemical activities of these two proteins in a number of systems representing group 1 CPS and related EPS have been confirmed (8, 14, 17, 21, 30, 31, 37, 43). In E. coli K30, these proteins are essential for the assembly of a HMW capsular layer on the cell surface, though not for low-level polymerization or assembly of KLPS (12, 43). There are second copies of the wza, wzb, and wzc genes on the chromosomes of E. coli K30 and K-12 strains. The products of these genes have been shown to participate with low efficiency in K30 CPS and colanic acid production (38, 43). To distinguish these genes from those in the K30 cluster, they have been named wza22min, wzb22min, and wzc22min.
Interest in the function of Wzb and Wzc homologues arises from the limited distribution of tyrosine phosphorylation in prokaryotes (reviewed in references 3 and 36) and the involvement of equivalent autokinase and phosphatase proteins in the assembly of capsules in gram-positive bacteria such as Streptococcus pneumoniae (18, 26), Streptococcus agalactiae (33, 44), and Staphylococcus aureus (24, 35). In these systems, the Wzc homologue consists of two separate polypeptides. One protein contains the two transmembrane domains and the periplasmic loop (corresponding to the Wzc N terminus), while the other protein corresponds to the C terminus and contains the Walker A ATP-binding motif as well as the C-terminal tyrosine-rich domain (18, 35).
Previous studies from our laboratory have shown that the last 17 amino acids of WzcCPS include the site(s) of phosphorylation of the protein and that phosphorylation at this site is essential for assembly of HMW K30 CPS (43). This C-terminal region is tyrosine rich, with seven tyrosine residues in the last 17 amino acids. Similar features were reported for a number of Wzc homologues (16, 27, 37). In E. coli K-12, five of the six C-terminal tyrosine residues can be phosphorylated in vitro (16). However, it is not clear at this point whether all of the Wzc C-terminal tyrosine residues or one or more specific tyrosine residues are accessible for phosphorylation in vivo. The ability of Wzc molecules to participate in transphosphorylation reactions was suggested by the finding that phosphorylated WzcCPS could be obtained from incubation of a WzcCPS mutant that could not bind ATP with a WzcCPS mutant that could bind ATP but that lacked the C-terminal site of phosphorylation (43). This transphosphorylation function was later confirmed and extended by a study of the WzcCA (colanic acid) protein (16). Work on the WzcCA protein in E. coli K-12 has revealed that a tyrosine residue outside of the C-terminal tyrosine-rich domain (Y569) was modified by autophosphorylation only, while the C-terminal domain was accessible to transphosphorylation by other WzcCA molecules (16). Furthermore, it was shown in the same study that phosphorylation of the C-terminal region was greatly enhanced in the presence of Y569. The authors proposed a two-step process (autophosphorylation and transphosphorylation) for the phosphorylation of WzcCA and suggested that Y569 may modulate the level of phosphorylation of the C-terminal tyrosine residues. However, the biological role of this modulation has not been established.
The essential requirement in group 1 capsule assembly for an active autokinase capable of phosphorylation at the C terminus and an active phosphatase led us to propose that the cycling of WzcCPS between phosphorylated and nonphosphorylated states may be necessary for its function (43). In E. coli K-12 and S. pneumoniae, both the kinase and the phosphatase are also required for polysaccharide production (27, 38). However, in these particular systems, decreases in the phosphorylation of WzcCA (caused by overexpression of WzbCA) and CpsD (caused by removal of the site of phosphorylation) still allowed wild-type levels of polysaccharide expression. This led the authors to suggest that phosphorylated WzcCA and CpsD act as negative regulators of colanic acid and capsule production. In Sinorhizobium meliloti, succinoglycan is produced in HMW and low-molecular-weight (LMW) forms, similar in some respects to KCPS and KLPS in E. coli. Mutations in ExoP (Wzc homologue) result in a decrease in the amount of HMW polymer and an increase in the amount of LMW material, suggesting that polymerization itself is not prevented (5, 30). This phenotype resembles that of a wzcCPS mutant, where K30LPS production is increased (12, 43). Data on CPS expression in group B streptococci point to a potential role for Wzc homologues in chain length determination (9). It is also interesting that the topology of the N terminus of Wzc resembles that of Wzz, a protein that influences chain length of LPS O antigens (reviewed in references 28 and 39). Current hypotheses suggest that Wzc proteins may interact with the polymerization machinery to control the level of polymerization of the EPS, but it is not yet clear what the specific roles of the Wzc proteins in CPS and EPS production are.
Although several studies have begun to decipher the impact of phosphorylation of particular residues on the in vitro activity of the autokinase, the importance of this phosphorylation in the biology of the organism has not been addressed. The purpose of this research was to study the importance of phosphorylation of specific tyrosine residues in the C-terminal domain of WzcCPS in CPS assembly in E. coli O9a:K30.
(These data were presented in preliminary form at the 101st General Meeting of the American Society for Microbiology [Abstr. Annu. Meet. Am. Soc. Microbiol., abstr. B-275, p.102, 2001].)
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAddendum in proofReferences |
|---|
|
by electroporation with a Gene Pulser from Bio-Rad (Hercules, Calif.) (6).
For the creation of pWQ214 to pWQ219, the wzc gene from pWQ212 was amplified with a forward primer that introduced an EcoRI site for cloning and a reverse primer that reintroduced the desired tyrosine codon and provided a PstI site for cloning. The PCR products were cloned back into pBAD18-Km, and the resulting plasmids were transformed into E. coli DH5.
Sequencing. Site-directed mutations were confirmed and plasmid sequences were verified by sequencing at the Guelph Molecular Supercentre (University of Guelph, Guelph, Ontario, Canada).
Expression of WzcCPS and derivatives. WzcCPS and mutant derivatives were expressed from pBAD arabinose-inducible expression vectors. Cells were grown to mid-exponential growth phase, and expression of the WzcCPS derivative was induced by addition of 0.02% L-arabinose. After induction, cells were grown for 2.5 h and then harvested. Cell pellets corresponding to a cell optical density at 600 nm of 1.0 were resuspended in 2x sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer and heated at 100°C for 10 min. These cell lysates were then analyzed by SDS-PAGE and Western blotting.
SDS-PAGE and Western immunoblot analysis. Whole-cell lysates were analyzed by SDS-PAGE (23) with 7.5 or 10% polyacrylamide resolving gels. Proteins were transferred to Westran polyvinylidene difluoride membranes (Schleicher & Schuell) for phosphotyrosine analysis or BioTrace NT nitrocellulose membranes (Gelman Laboratory) for Wzc expression analysis. The transfer buffer used contained 3 mM Na2CO3, 10 mM NaHCO3, and 20% methanol. Wzc expression was detected with a WzcCPS rabbit polyclonal antiserum (43) and a goat anti-rabbit secondary antibody (Caltag, Burlingame, Calif.), while the level of phosphorylated Wzc protein was assessed with the PY20 antiphosphotyrosine monoclonal antibody (Transduction Laboratories, Lexington, Ky.) and a goat anti-mouse secondary antibody (Jackson ImmunoResearch Laboratories Inc., West Grove, Pa.). Both secondary antibodies were conjugated to alkaline phosphatase, which allowed colorimetric detection using nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolylphosphate.
SDS-PAGE analysis of cell surface polysaccharides. Cell surface polysaccharide samples were prepared by the method of Hitchcock and Brown (20). Briefly, whole-cell lysates were prepared in 1x SDS-PAGE sample buffer lacking ß-mercaptoethanol. These lysates were heated at 100°C for 30 min, proteinase K was added to 0.5 mg/ml, and the protein content was digested at 55°C overnight. These samples were then analyzed by electrophoresis on NuPAGE 4 to 12% N,N-methylenebisacrylamide-Tris gels (Invitrogen Life Technologies Inc.) and then transferred to BioTrace NT nitrocellulose membranes (Gelman Laboratory) for immunoblotting. K30 polysaccharide was detected with a rabbit anti-K30 polyclonal antiserum (11) and an alkaline phosphatase-conjugated goat anti-rabbit secondary antibody.
Bacteriophage sensitivity assays. The sensitivity of strains to lysis by bacteriophage K30 was determined in order to assess the presence or absence of K30 antigen on the cell surface. Bacteriophage K30 lyses cells bearing a full or partial K30 capsule, while the absence of K30 polymer makes strains resistant to attack by this phage (40).
CPS purification and quantitation. Cell-free and cell-associated CPS were purified as described previously (22). Briefly, cells were harvested from colonies grown for 18 h on LB-kanamycin (or LB-ampicillin) containing 0.02% L-arabinose. The colonies were scraped into phosphate-buffered saline, and viable counts were determined from this suspension. Cells were harvested by centrifugation, leaving cell-free polysaccharide in the supernatant. The supernatant was treated with DNase (Roche), RNase A (Roche), and proteinase K (Sigma) and dialyzed against water. The pellet was extracted with phenol, allowing recovery of cell-associated CPS in the aqueous phase. The aqueous phase was dialyzed against water, treated with enzyme as described for the cell-free CPS, and redialyzed. For both cell-free and cell-associated samples, LPS was removed by ultracentrifugation (105,000 x g, 1 h, 15°C).
Purified CPS samples were analyzed for uronic acid content by the modified carbazole assay of Bitter and Muir (7). Briefly, 0.5 ml of an appropriate dilution of the sample was added to 3 ml of 0.025 M sodium tetraborate (BDH Chemicals Ltd., Poole, United Kingdom) in concentrated H2SO4. Samples were heated to 100°C for 10 min and then cooled to room temperature. Next, 0.1 ml of 0.125% (wt/vol) carbazole (Sigma) in absolute ethanol was added to the samples, and the samples were heated to 100°C for an additional 15 min. After the samples were cooled, the A530 was determined. A standard curve of glucuronolactone was generated for determination of the amount of uronic acid in the samples.
DSP cross-linking. Fifty-milliliter cultures were grown to mid-exponential growth phase, and the expression of wild-type and mutant WzcCPS proteins in pBAD-based constructs was induced by addition of 0.02% L-arabinose. Growth was continued for an additional 2 h, and cells were harvested and washed with buffer A (150 mM NaCl in 20 mM NaH2PO4, pH 7.5). Cell pellets were resuspended in 5 ml of buffer A, and samples were collected prior to cross-linking. Cross-linking was carried out for 30 min by the addition of 1 mM dithiobis(succinimidylpropionate) (DSP; Lomant's reagent; Pierce, Rockford, Ill.) from a 25 mM stock prepared in anhydrous dimethyl sulfoxide (Aldrich, Oakville, Ontario, Canada). The reaction was quenched for 15 min by the addition of Tris-HCl (pH 7.5) to a final concentration of 100 mM. Cells were harvested and resuspended in 20 mM Tris-HCl (pH 7.5) and lysed by sonication. Sonication is required for efficient cell lysis of cross-linked samples. Samples of cross-linked material were prepared by addition of NuPAGE 4x lithium dodecyl sulfate sample buffer (Invitrogen Life Technologies Inc.) to the lysates. In one set of samples, the cross-links were cleaved by addition of dithiothreitol (DTT) to a final concentration of 50 mM. Samples were heated to 100°C before being run on SDS-7.5% PAGE gels and transferred onto polyvinylidene difluoride membranes for antiphosphotyrosine (PY20) blots or onto nitrocellulose membranes for anti-Wzc blots.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAddendum in proofReferences |
|---|
wzc strain. In S. pneumoniae, cpsCD are also essential for capsule assembly. However, in this system, phosphorylation of CpsD acts as a negative regulator of CPS production (27). One possible interpretation of the different result for E. coli O9a:K30 is that the WzcCPS truncation alters protein folding or inner membrane interactions with other proteins. In this case, the phenotype would not be attributable solely to the removal of the seven tyrosine residues in the C-terminal region. To address this issue, a WzcCPS mutant in which the tyrosine residues at amino acid positions 705, 706, 708, 713, 715, 717, and 718 were each replaced by phenylalanine was constructed. This was done by sequential site-directed mutagenesis of pWQ130 (pBAD18Km-WzcCPS), and the resulting plasmid (pWQ212) was transformed into CWG285, a strain in which both chromosomal copies of the wzc gene have been knocked out. The mutated WzcCPS protein (WzcY705-718
F) was expressed at levels similar to those for the native WzcCPS protein, as assessed by Western blotting with antiserum against WzcCPS (Fig. 1A). However, WzcY705-718F was not phosphorylation competent, since no reactivity with the antiphosphotyrosine antibody PY20 was detected (Fig. 1B). The in vivo activity of this protein was assessed by determining the ability of the plasmid to restore HMW K30 CPS expression in CWG285. This is detected by Western immunoblot analysis of polysaccharide preparations probed with a polyclonal antiserum specific against the K30 capsule. WzcY705-718F could not restore any detectable HMW CPS (K30CPS) production (Fig. 1C). The higher degree of polymerization of K30LPS (seen as a ladder of slower-migrating molecules) was evident in CWG285 with or without the plasmid. This is a characteristic feature of wzc mutants in E. coli O9a:K30 (12, 43). The absence of an HMW polymer was confirmed by the resistance of this strain to bacteriophage K30, which requires the K30 antigen as its receptor.
|
F) did not enhance the base level of polymer production in CWG285 (Table 2). In contrast, introduction of pWQ130 (encoding the wild-type WzcCPS) resulted in an elevated amount of cell-associated capsular K30 antigen (Table 2). The complementation with plasmid-encoded WzcCPS derivatives never restores wild-type levels of CPS production (compare to CWG258 [Table 2]), which is consistent with previous studies (12, 43).
|
F reflected a loss of protein function (e.g., due to improper folding) caused by the replacement of seven tyrosine residues with seven phenylalanine residues rather than a direct effect of tyrosine phosphorylation. To confirm the importance of WzcCPS phosphorylation for CPS expression, WzbCPS was overexpressed in CWG258 and the amounts of cell-free and cell-associated K30 polymer produced by this strain were determined. Overexpression of Wzb would be expected to greatly decrease the amount of phosphorylated Wzc, providing a different means to examine the role of Wzc phosphorylation. Overexpression of WzbCPS was confirmed by SDS-PAGE of cell lysates (data not shown) and did indeed cause a 3.7-fold decrease in the amount of cell-associated uronic acid but had no dramatic effect on the amount of cell-free polymer produced (Table 2). These experiments were performed in a background (CWG258) lacking the second copies of wza, wzb, and wzc (at 22 min on the E. coli K-12 chromosome) in order to simplify their interpretation. Is any single tyrosine residue alone sufficient to maintain phosphorylation of WzcCPS and/or capsule assembly? To determine whether any single tyrosine residue in the C-terminal tyrosine-rich domain is sufficient to maintain phosphorylation of the WzcCPS protein and/or restore HMW K30 CPS, seven WzcCPS mutant proteins that each retained only one of the seven tyrosine residues were created. The anti-WzcCPS blot shows that all of these mutant WzcCPS proteins were expressed at levels comparable to that for the native WzcCPS protein (Fig. 2A). However, none of these proteins showed detectable phosphorylation by PY20 Western blotting (Fig. 2B), and none could restore the capsular phenotype to CWG285 (Fig. 2C), though each retained elevated levels of K30LPS. These strains were also resistant to the K30 phage, confirming the absence of K30 CPS on the cell surface.
|
wzc strain and reduce the amount of K30LPS assembled (Fig. 3C). These strains were also sensitive to infection by the K30 phage.
|
F, WzcY708/713F, WzcY708/715F, and WzcY715/717F,) were still phosphorylated, as seen by PY20 immunoblotting (Fig. 4B, lanes 3 to 6, respectively), and these strains still assembled the HMW K30 polymer, as shown by the anti-K30 immunoblot and the phage sensitivity data (Fig. 4C, lanes 3 to 6). However, the relative amounts of polymer produced by these strains as assessed by immunoblotting appear to vary depending on the combination of tyrosine residues remaining. Specifically, WzcY706/717F (Fig. 4C, lane 3), WzcY708/715F (Fig. 4C, lane 5), and WzcY715/717F (Fig. 4C, lane 6) show larger amounts of immunoreactive K30 antigen than WzcY708/713F (Fig. 4C, lane 4). Furthermore, the K30 phage does not produce clear plaques when WzcY708/713F is expressed. For the three strains that express a WzcCPS protein lacking three of the terminal tyrosine residues, only WzcY713-717F (Fig. 4C, lane 9) showed a detectable K30CPS polymer by immunoblotting (Fig. 4C, lanes 7 to 9), although Wzc derivatives from all three strains are expressed and phosphorylated (Fig. 4A and B, lanes 7 to 9). WzcY708-715F showed only turbid plaques when subjected to the K30 phage, while WzcY706-713F was not sensitive to the K30 phage at all. It is again of interest to note the variable amounts of immunoreactive K30CPS material produced by these derivatives. Two WzcCPS proteins that lacked four of the seven terminal tyrosine residues and two others that lacked five of these seven residues were constructed. All of these proteins were expressed at similar levels (Fig. 4A, lanes 12 to 15) and were phosphorylated (Fig. 4B, lanes 12 to 15), although WzcY708-718F (Fig. 4B, lane 14) and WzcY706-717F (Fig. 4B, lane 15) showed a greatly reduced level of phosphorylation. All of these proteins were unable to restore HMW CPS expression in CWG285, as shown by immunoblotting and resistance to the K30 phage (Fig. 4C, lanes 12 to 15).
|
F protein was expressed at levels similar to those for the wild-type protein (Fig. 5A) and was still phosphorylated, as assessed by PY20 immunoblotting (Fig. 5B). The phosphorylation level of WzcY569F appeared to be reduced slightly compared to that of wild-type WzcCPS. Analysis of K30 capsule assembly by immunoblotting revealed that HMW capsular material was still assembled when this protein was expressed in CWG285 (Fig. 5C), and the phage sensitivity data revealed no difference from wild-type WzcCPS expressed in CWG285. To determine whether there is any subtle difference in the amount of K30 polymer produced by this derivative, cell-free and cell-associated CPS were purified from CWG285 expressing WzcCPS and WzcY569F. There was no apparent difference in the amounts of cell-free K30 CPS produced by WzcCPS and WzcY569F or in the amounts of cell-associated CPS produced by these strains (Table 2).
|
Do mutations in the C-terminal tyrosine-rich region affect the ability of WzcCPS proteins to interact?
Previous research has shown that Wzc proteins have the ability to participate in transphosphorylation reactions (16, 43). This provides evidence that Wzc proteins can interact both in vivo and in vitro. Chemical cross-linking of whole cells by using DSP was performed to investigate the possible interactions of WzcCPS proteins. Cross-linking in E69, which expresses the protein from the chromosomal copies of wzc, was first examined. Cells were cross-linked and lysed, and the lysates were examined by Western immunoblotting with both the anti-WzcCPS polyclonal antiserum and the PY20 phosphotyrosine antibodies. After cross-linking, Wzc was detected in a single HMW complex that barely migrated into the gel (Fig. 6A). Upon reduction, monomeric Wzc was detected. This complex may represent the association of Wzc into oligomers or reflect interactions with other components of the K30 biosynthesis, assembly, and translocation machinery. To examine whether smaller complexes could be formed in the absence of the outer membrane protein Wza, which has been shown to assemble into large multimeric structures, cross-linking in CWG281, an E69 derivative lacking both chromosomal copies of wza, was carried out. In this strain, WzcCPS is expressed from a chromosomal copy only. The results were the same as those for E69, with a single large complex containing Wzc detected by immunoblotting after cross-linking (Fig. 6A). To isolate the effects of cross-linking on the multimerization state of Wzc alone, further experiments with whole cells of E. coli DH5 expressing WzcCPS encoded by pWQ130 were performed. An E. coli K-12 derivative (DH5) was used in these experiments to eliminate any possible complicating associations of Wzc with other components of the capsule translocation apparatus in E69 derivatives. E. coli DH5 contains the wza, wzb, and wzc genes in the colanic acid biosynthesis cluster, but these are not expressed in E. coli K-12 at the growth temperatures used in this study (15). Western immunoblot analysis using PY20 showed that WzcCPS was present in at least two HMW complexes of greater than 
150 kDa (Fig. 6A). Cleavage by DTT resulted in detection of monomeric Wzc and of two complexes, one smaller than 150 kDa and another 250 kDa (Fig. 6A). The same result was obtained when the immunoblot was probed with WzcCPS polyclonal antiserum, but the lower reactivity of the antibody generated a weaker signal (data not shown). Further studies were then undertaken to investigate whether WzcCPS derivatives bearing changes in the C-terminal tyrosine-rich region can still interact. Loss of this interaction is one possible explanation for why some of the mutations impair the phosphorylation of WzcCPS and/or assembly of HMW K30 CPS. DSP cross-linking was carried out in DH5 expressing various WzcCPS constructs in trans. When a WzcCPS Walker A box mutant (WzcK540R; pWQ132) was expressed in DH5, a cross-linking profile similar to that obtained with WzcCPS was observed (Fig. 6B). The same was seen when either the truncated WzcCPS protein lacking the C-terminal tyrosine-rich domain (Wzc1-704; pWQ133) or the site-directed mutant lacking all the C-terminal tyrosine residues (WzcY705-718F) was expressed in DH5 (data not shown).
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAddendum in proofReferences |
|---|
Previously we showed that the C-terminal tyrosine-rich region of WzcCPS from E. coli O9a:K30 is the site of phosphorylation of the protein and that the removal of these tyrosine residues by truncation results in a protein whose phosphorylation cannot be detected by an antiphosphotyrosine antibody. Furthermore, this protein could not restore synthesis of HMW CPS in a wzc strain. It was therefore suggested that phosphorylation at one or more of the tyrosine residues in the C-terminal region was important for the ability of WzcCPS to function in capsule assembly (43). Similar findings were reported from studies on WzcCA and ExoP, which showed that the C-terminal domains of these proteins were phosphorylation competent (16, 30, 38).
One concern in interpreting data from truncated mutants is that the removal of the last 17 amino acids of WzcCPS might have altered the folding of the protein and thus altered its phosphorylation and capsule assembly functions. However, the WzcY705-718F mutant lacking only the seven C-terminal tyrosine residues showed the same phenotype as the truncated version of the protein. This supported the proposal that phosphorylation of one or more of the C-terminal tyrosine residues is essential for HMW CPS expression in group 1 E. coli strains, though not essential for KLPS assembly. The importance of the phosphorylation of WzcCPS in CPS production was further confirmed by findings that overexpression of WzbCPS caused a 3.7-fold reduction in the amount of cell-associated uronic acid produced. This observation provides additional support for the hypothesis that WzcK30 must undergo cycles of phosphorylation and dephosphorylation for K30 CPS to be expressed. This finding is in contrast to the situation in E. coli K-12 and S. pneumoniae, where phosphorylated WzcCA and CpsD are believed to act as negative regulators of polysaccharide production (27, 38). It is unclear whether differences in the E. coli K30 antigen and colanic acid situations reflect the fact that one is primarily capsular (K30) and the other is a cell-free EPS (colanic acid).
The effect of the replacement of six of the seven tyrosine residues in the C-terminal tyrosine-rich region was then studied. None of these proteins were detectably phosphorylated by immunoblot analysis, and the absence of detectable phosphorylation correlates with the loss of the biological activity of the protein, i.e., the loss of HMW CPS expression. It remains to be established whether the absolute level of phosphorylation in these mutants is zero or whether the sensitivity of the PY20 antibody is an issue. However, it is important to note that phosphorylation is readily detected by the PY20 antibody in some constructs harboring only two tyrosine residues. This does not detract from the conclusion that phosphorylation at more than one residue in this region is essential for capsule assembly.
The removal of a single tyrosine residue in this region could have revealed whether any one residue was essential to the phosphorylation of WzcCPS or to capsule assembly. Furthermore, if a specific sequence of phosphorylation events was required, removal of the first tyrosine residue in a fixed reaction cascade should also have eliminated phosphorylation of the WzcCPS protein entirely. Neither scenario occurred, as any single tyrosine residue could be replaced without the loss of phosphorylation of WzcCPS or capsule assembly. From the data currently available it appears unlikely that phosphorylation in this region needs to occur in any given order. However, we cannot rule out cooperative effects among phosphorylated tyrosines by using this approach. Such effects have been suggested by Niemeyer and Becker (30) in their studies of ExoP.
Mutants that lacked two to five of the tyrosine residues in the C-terminal domain were then created to determine whether any specific combination of tyrosine residues was important for the activity of WzcCPS. These altered WzcCPS proteins could also provide some information about the minimal number of phosphorylation sites required for the biological activity of WzcCPS. The loss of three or four tyrosine residues in this region was enough to eliminate capsule assembly, though phosphorylation of these WzcCPS derivatives was still detected by immunoblotting. The requirement of at least four tyrosine residues in the C-terminal region for HMW CPS assembly raises the possibility that specific combinations of these tyrosine residues are more important than others. However, the available data do not allow the determination of exactly which combination(s) of residues is most important for WzcCPS activity. Given the number of permutations involving seven residues, the generation of mutants representing every possible combination of tyrosine alterations was impractical. Furthermore, as indicated above, it is conceivable that phosphorylation of tyrosine residues in this region occurs cooperatively and that certain residues in this region may not be as accessible to phosphorylation. Based on the phenotypes of wzc and wzb mutants, we previously proposed that the level of phosphorylation in this region appears to be important (43). This is supported by the fact that Wzc phosphorylation and HMW CPS assembly can be uncoupled, for example, when only two or three tyrosine residues remain in this region. It is therefore important in functional studies of Wzc to distinguish between derivatives that are competent for phosphorylation and those that are capable of supporting capsule assembly.
Studies of ExoP from Sinorhizobium meliloti have shown that deletion of the C terminus of the protein or replacement of certain tyrosine residues causes a reduction in the amount of EPS produced (5, 30). In contrast, studies of CpsD have shown that there is no change in the amount of CPS produced when the C-terminal tyrosine residues are replaced, though the colonies appear to be more mucoid (27). In this study, analysis of K30CPS by Western immunoblotting revealed apparent differences in the amounts of CPS produced by strains expressing certain mutant WzcCPS derivatives. In S. agalactiae, the Wzc homologues CpsC and CpsD were shown to affect capsule chain length. In the absence of CpsC or CpsD, CPS chain length was reduced (9). Examination of K30CPS by SDS-PAGE did not reveal any difference in the sizes of the polymers produced when different WzcCPS derivatives were expressed. However, it is possible that SDS-PAGE profiles do not reflect the entire picture, as some molecules may be too large to enter the resolving gel. Attempts were made to determine whether K30 CPS chain length was affected by the expression of mutant WzcCPS proteins in CWG285. The K30 capsular material produced by a representative group of these mutants was purified, and sizing was performed by gel filtration chromatography using a column with a fractionation range for globular proteins of 5,000 to 5 x 106 Da. All of the KCPS samples tested eluted at the void volume, with no additional peaks representing smaller CPS material (data not shown). From these limited data, there are certainly no K30CPS molecules with significantly smaller molecular weights formed by a biosynthesis machinery containing mutant WzcCPS proteins.
Grangeasse and coworkers (16) studied the importance of tyrosine residues throughout the entire cytoplasmic domain for the activity of the WzcCA protein in vitro. Purified protein was incubated with [
-32P]ATP, and the levels of phosphate incorporation in WzcCA were detected by autoradiography and scintillation counting. They showed that one tyrosine residue located between the Walker A and B ATP-binding motifs (Y569) greatly enhanced the level of phosphorylation of the WzcCA protein in vitro (45-fold). The authors suggested that phosphorylation of this site with ATP could provide a means of modulating the level of phosphorylation of the C-terminal domain of Wzc. Alignment of the cytoplasmic region of the WzcCA and the WzcCPS proteins shows that this Y569 residue and the regions around it are well conserved. It was therefore of interest to determine what the biological consequence of a Y569F mutation would be in the K30 system. In the C-terminally truncated form of Wzc (Wzc1-704) (43) and in WzcY705-718F, no in vivo phosphorylation was detected with the phosphotyrosine antibody, which can detect as few as two phosphorylated tyrosine residues in the C-terminal region. The replacement of Y569 with phenylalanine resulted in an apparent reduction in the amount of phosphorylation of the protein, as determined by antiphosphotyrosine immunoblotting. However, this protein could restore HMW CPS production in CWG285 to amounts very similar to those produced by WzcCPS. This suggests that phosphorylation of WzcCPS at position Y569 is not essential for in vivo phosphorylation of the C-terminal domain of the protein or for assembly of HMW CPS. This does not, however, preclude the involvement of this residue in modulating the phosphorylation of the C-terminal domain, as suggested by Grangeasse and coworkers (16).
Previous results from our laboratory have shown that WzcCPS molecules can participate in a transphosphorylation reaction. No phosphorylation was detected when WzcK540R, which bears a Walker A mutation that prevents ATP binding and/or hydrolysis, or Wzc1-704, which has a functional Walker A motif but which lacks the C-terminal tyrosine-rich region, was expressed alone in CWG285. Furthermore, neither protein was functional in capsule expression. However, when they were coexpressed, phosphorylated Wzc was detected and capsule expression was restored. This showed that Wzc1-704 was able to bind and hydrolyze ATP and that it could transfer phosphate to the C-terminal region of WzcK540R (43). This first evidence of a transphosphorylation reaction between Wzc molecules was supported by subsequent detailed in vitro work done with WzcCA (16). Chemical cross-linking studies were therefore performed to determine whether Wzc proteins could form complexes and to assess the possible role of phosphorylation in such interactions. Studies were carried out with E. coli DH5 in order to better study the interactions between Wzc molecules without the interference of other components of the capsule translocation apparatus, which potentially exist in a complex involving Wzc. Wild-type Wzc was found in several HMW complexes after DSP cross-linking. Cleavage of the cross-link eliminated most of these complexes and restored a large amount of monomeric Wzc. The patterns of complexes formed by Wzc1-704 (truncation of the Y-rich region), WzcK540R (Walker A mutation), and WzcY705-718F (all C-terminal Ys changed to Fs) were examined. All of these proteins are phosphorylation null and do not function in capsule assembly. However, the cross-linking profiles of strains expressing these proteins were indistinguishable from the profiles of those expressing wild-type WzcCPS. These data suggest that changes in the level of phosphorylation of the C-terminal tyrosine-rich region do not affect the interactions of WzcCPS proteins. The present study emphasizes the importance of Wzc phosphorylation in capsule assembly but does not shed light on the precise role played by Wzc in this biological process. One possibility is that Wzc coordinates other components in the assembly complex in a manner analogous to the proposed role of Wzz in the biosynthesis of O antigens (4, 29). A similar role is hypothesized for the ATP-binding protein PulE and its homologues in type II protein secretion (reviewed in reference 34). It is not unreasonable to assume that interactions between Wzc molecules or between Wzc and other proteins occur mainly through the periplasmic loop. This region has one or more predicted coiled-coil domains (28), and such elements are implicated in protein-protein interactions. In fact, the N-terminal region of Wzz was sufficient for the oligomerization of this protein (10). Perhaps changes in the phosphate load in the cytoplasmic region affect the ability of Wzc to interact with other members of the capsule assembly and translocation machinery. Further research is now required to determine whether the functional form of Wzc is as a monomer, dimer, or multimer and to dissect its possible interactions with other proteins.
|
ADDENDUM IN PROOF |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAddendum in proofReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
We thank Thomas Wugeditsch for help with the early stages of this work and Jutta Nesper for critical reading of the manuscript.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAddendum in proofReferences |
|---|
This article has been cited by other articles:
