| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 280, Issue 36, 31378-31389, September 9, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||
From the
Laboratoire d'Ingénierie des Systèmes Macromoléculaires, UPR9027, IBSM/CNRS, 31 Chemin Joseph Aiguier, 13402 Marseille Cedex 20, France and the ¶Department of Structural Biology, Max Planck Institute of Biochemistry, Am Klopferspitz 18a, D-82152 Martinsried, Germany
Received for publication, May 27, 2005 , and in revised form, July 11, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Additional components of the type II secretion machineries share extensive similarities with proteins involved in the type IV piliation system. Already identified are the secretins PilQ, XcpQ, and HxcQ (11), the "traffic ATPases" PilB, XcpR, and HxcR (12), and the polytopic inner membrane proteins PilC, XcpS, and HxcS. The striking similarities were functionally supported when several groups demonstrated that the pseudopilin of the type II secretion systems could be assembled in multifibrillar structures, called the type II pseudopilus (13–15). Most interestingly, besides the major type IV pilin PilA, minor pilins have been described, which did not appear to be part of the fibrillar structure and for which no particular function was ascribed. The Xcp pseudopilins can be divided into one major component, XcpT, and minor components, XcpU–X (9, 10). Among the latter, XcpX is atypical (9). Indeed, it is three times longer than XcpT, and it lacks the highly conserved glutamate residue at position +5 in mature pseudopilins and pilins. This residue has been proposed to play a role for the helical assembly of type IV pilins into the pilus structure (16).
Several major issues remain to be understood in the biogenesis of the type II pseudopilus, most importantly is the determination of its function in the secretion process. In this study, we discovered several important aspects concerning pseudopilus assembly. We demonstrate that only a subset of the Xcp components is specifically dedicated to the assembly of the pseudopilus. We confirmed that among the five pseudopilins, only one, XcpT, appears to have the characteristics that allow the assembly in a fibrillar structure. Finally, we reveal that the atypical pseudopilin XcpX is a key component in controlling the pseudopilus elongation process. We propose that this control can be exerted via a direct interaction between XcpX and XcpT.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
Plasmid Construction for Overproduction of the Full-length XcpT and XcpX Pseudopilins in E. coli—The xcpT gene was amplified using pMTWT as matrix and the primers 5'Rev(–48), 5'-AGC GGA TAA CAA TTT CAC ACA GGA-3', and 3' XCPThC (His6), 5'-ATT CAG TGG TGG TGG TGG TGG TGG TTG TCC CAG TTG CCG AT-3'. The resulting gene encoded a C-terminally His6-tagged XcpT. The amplicon was first cloned into the pCR2.1 vector (Invitrogen) and sequenced. A BamHI/HindIII 600-bp insert was then subcloned into the broad host range vectors pMMB190 or pBBR1MCS2, yielding pMMB-TH and pBBR2-TH, respectively.
The xcpXWT and xcpXT+5E alleles were amplified using pSB28 (XcpXWT) and pRV20 (XcpXT+5E) as matrix, respectively, and the primers 5'AF024* (BamHI, SD), 5'-ACT GGA TCC TAA GGG AAG CCG GAA TGA GG-3' and 3'XCPX2 (HindIII, TGA–), 5'-ATA AAG CTT TCG CTC GTC CTT CTT-3'. "SD" indicates that we have slightly modified the corresponding Shine-Dalgarno sequence upstream of the xcpX gene to closely resemble the consensus. The amplicons were cloned into the pCR2.1 vector (Invitrogen) and sequenced. 1000-bp BamHI/HindIII inserts were subcloned into the pET22b plasmid in-frame with the sequence encoding a His6 tag but out-of-frame with the pelB sequence, yielding the plasmids pET-XWTH and pET-XT+5EH. In these plasmids the genes are expressed under the control of the T7 promoter.
FLAG-tagged derivatives of xcpX were also constructed. In a first PCR step, the xcpXWT and xcpXT+5E alleles were amplified using pSB28 (XcpXWT) or pRV20 (XcpXT+5E) as matrices, respectively, and the primers 5'AF024* (SD), 5'-ACT GGA TCC TAA GGG AAG CCG GAA TGA GG-3', and 3'XFLAG1 (FLAG1), 5'-TTT TCA CTT GTC ATC GTC GTC CTT GTA GTC TCG CTC GTC CTT CTT CCA-3', introducing a C-terminal FLAG-tag. In a second PCR step, these amplicons were re-amplified using the primers 5'AF024* (SD), 5'-ACT GGA TCC TAA GGG AAG CCG GAA TGA GG-3', and 3'FLAG2 (FLAG), 5'-TAA TCA CTT GTC GTC GTC ATC CTT GTA GTC TAA CTT GTC ATC GTC GTC-3', generating a second FLAG tag after and in-frame with the first one. It has been shown that tandem FLAG tags increase the sensitivity of the detection in Western blotting experiments. The PCR products were cloned into the pCR2.1 vector (Invitrogen) and sequenced. 1200-bp EcoRI inserts were subcloned into the pET22b plasmid out-of-frame with both the pelB sequence and the DNA region encoding the His6 tag, yielding the plasmids pET-XWTF and pET-XT+5EF. In these plasmids the genes are expressed under the control of the T7 promoter.
Preparation of Pseudopili—A P. aeruginosa PAK mutant strain with mutations in the pilA and fliC genes was used as genetic background to prepare pseudopili samples free from type IV pili and flagella. The plasmid pMTWT was introduced in this strain by mobilization to allow overproduction of XcpT. The PAKpilA/fliC mutant containing pMTWT was streaked on large square plates containing LB agar supplemented with 300 µg ml–1 of carbenicillin and 2 mM of IPTG. After overnight incubation at 37 °C, bacterial colonies were scraped off 8 plates and resuspended in 10 mM Tris-HCl buffer, pH 7.5, containing 150 mM KCl. After homogenization, the cells were pelleted by low speed centrifugation (6,000 x g), and the pseudopili-enriched supernatant was collected. The pseudopili were further pelleted by high speed centrifugation (158,000 x g). The resulting pseudopili pellet was resuspended in a small volume of 10 mM Tris-HCl buffer, pH 7.5, containing 150 mM KCl. These preparations were directly observed by classical TEM or cryo-EM.
TEM and Immunogold Labeling—The negative-staining procedure and TEM observation were performed as described previously (15).
Cryo-electron Microscopy—A 5-µl droplet of the pseudopili preparation was placed on a carbon-coated copper grid (Plano GmbH, Wetzlar, Germany). The biological sample was embedded in vitreous ice by plunge freezing the grids into liquid ethane (T =–186 °C) as described by Dubochet et al. (17). The grid was then transferred into a ±70° tilt cryo-holder (Gatan Inc., Pleasanton, CA) cooled with liquid nitrogen (T = –196 °C) and inserted into the microscope. Two-dimensional images were recorded using an FEI CM200 TEM operating at 160 kV. The instrument was equipped with a field emission gun. The defocus level was about 2 µm. The pixel size at the specimen level was 0.3 nm.
Image Processing—Cryo-electron micrographs were treated with the Eman software package (18). EMAN is semi-automated software for high resolution single-particle reconstructions (18). Different filaments were boxed (64 x 64 pixels) and rotated using Boxer. No CTF correction was performed. The boxed filaments were then grouped into self-similar classes, and good classes were selected for modeling.
Immunofluorescence Microscopy—Bacteria were grown in a similar manner as the one described previously (15) for shearing experiments. After overnight incubation at 37 °C, bacteria were scraped from the plate with a toothpick and resuspended into PBS buffer. All further incubations were performed at room temperature. 25 µl of this suspension was deposited onto a glass slide. After 5 min, bacteria were fixed for 10 min by adding 100 µl of paraformaldehyde (1%) directly onto the bacterial drop. After three washing steps with PBS, the preparations were saturated with 5% bovine serum albumin in PBS for 10 min. The glass slides were washed three times with PBS and incubated with a 1:400 dilution of the primary antibodies against XcpT (rabbit) or against the His tag (mouse) in PBS, 0.5% bovine serum albumin for 1 h. After three washing steps, the preparations were incubated with a 1:400 dilution of the fluorescein/anti-rabbit (Vector Laboratories) or fluorescein/anti-mouse (Vector Laboratories) conjugates for 1 h. After three washing steps, the preparations were covered by a drop of Vectashield mounting medium (Vector) and with a small glass slide. Samples were observed on a Zeiss Axio imager microscope.
Shearing of Pseudopili—The method to purify cell surface-associated pseudopili was described previously (15).
Biofilm Formation Assay on Glass Slides—The methodology used is adapted from Ref. 19. Bacterial strains were grown in 5 ml of M63 minimal medium supplemented with 0.2% glucose, 0.5% casamino acids, 1 mM MgCl2,and2mM IPTG, in a 50-ml Corning tube containing a semi-immersed glass cover slide. After a short period of incubation for1 h at 30 °C without shaking, the slide was removed and rinsed. The attached bacterial cells were visualized by phase-contrast microscopy using an Axiovert 200M microscope (magnification x100). Images were captured with a Hamamatsu-type Orca ER camera.
Production and Purification of Full-length Pseudopilins—E. coli BL21(DE3) cells bearing the plasmids pBBR2-TH, pET-XWTF, or pETXT+5EF were grown in Luria Broth (LB) at 37 °C supplemented with kanamycin 25 µgml–1 (pBBR1MCS2 or derivatives) or ampicillin 50 µg ml–1 (pET22b or derivatives). When the bacterial culture reached an A600 of 0.5, production of the recombinant proteins was induced by adding 50 µM IPTG for a period of 2.5 h. Unless otherwise stated, the following steps were performed on ice. The bacterial cells were pelleted and resuspended in 50 mM sodium phosphate buffer, pH 8, containing 2 mg of lysozyme and the protease inhibitor mixture Complete EDTA-free (Roche Applied Science) diluted 1:100. The cells were broken by sonication. After a short spin down to remove unbroken cells (centrifugation 5 min at 6,000 x g), the total membranes were pelleted by centrifugation at 100,000 x g for 45 min. Proteins from the total membrane pellet were then extracted by resuspension into a 50 mM sodium phosphate buffer, pH 8, containing 150 mM NaCl, 1.2% of the sulfobetaine detergent n-dodecyl-n,n-dimethyl-3-ammonio-1-propanesulfonate (Sigma), and Complete EDTA-free 1:100. Solubilization was performed at room temperature for 1 h with gentle shaking. Insoluble material was removed by centrifugation at 100,000 x g for 30 min. The detergent-soluble supernatants containing extracted membrane proteins were saved and stored at –80 °C with 10% glycerol.
Proteinase K Sensitivity Assay—Detergent-solubilized proteins were used for this assay. Equal volumes of extracts enriched in XcpTH, XcpXWTF, and XcpXT+5EF were mixed and preincubated at 4 °C for 10 min. 1.2 µg of proteinase K was added, and the samples were incubated at 37 °C for the selected times. The protease activity was stopped by adding 10 mM phenylmethylsulfonyl fluoride at 4 °C for 10 min. Samples were solubilized in Laemmli SDS-buffer and heated for 7 min at 95 °C.
Cloning and Expression of the Soluble Periplasmic Domains of the Xcp Pseudopilins—The DNA fragments encoding the N-terminal truncated xcpT and xcpX genes from position +25 and +23 relative to the leader peptide cleavage site, respectively, were amplified by standard PCR. Primers used for amplification were 5'TpD1(BamHI;His6)5'-ATA GGA TCC ACA CCA CCA CCA CCA CCA CAT GAG CCG TCC CGA CCA G-3' and 3'TpD2(HindIII) 5'-ATA AAG CTT ATC AGT TGT CCC AGT TGC CGA TGT CGG CGT CGT TGT C-3' for xcpT; 5'XpC1(BamHI;His6) 5'-ATA GGA TCC A CAC CAC CAC CAC CAC CAC CGC CAG CAG TTG GCG ATA-3' and 3'XpC2(HindIII) 5'-ATA AAG CTT A TCA TCG CTC GTC CTT CTT CCA ATC GTC GCC GCC CGT-3' for xcpX. The PCR introduced a region encoding an N-terminal His6 tag together with BamHI/HindIII cloning sites. PCR products were first subcloned into the pCR2.1 vector (Invitrogen) and sequenced. BamHI/HindIII DNA fragments were then generated and subcloned into the pET22b (Novagen). The cloning created an in-frame fusion of the xcp genes with the pelB region encoding the N-terminal signal sequence. In this way the recombinant protein could be produced in the periplasm. The resulting plasmids were called pET-Tp25–148NH and pET-Xp23–333NH. Each gene fusion was expressed from the T7 promoter of the pET22b vector in E. coli BL21(DE3) host cells. The recombinant proteins were called XcpTp25–148NH and XcpXp23–333NH, where "p" indicates periplasmic, "NH" indicates N-terminal His6 tag, and the numbers correspond to the amino acid position in the primary sequence of the protein starting from the first residue after the leader-peptidecleavage site (F + 1 for XcpT or V + 1 for XcpX). Recombinant BL21 strains carrying the plasmids pET-TpNH and pET-XpNH were grown in ZYP-5052 auto-inducing medium developed by Studier (BNL: rich medium containing yeast extract, tryptone, phosphate-buffered; 0.05% glucose, 0.5% glycerol, and 0.2% lactose). After 4 days of growth, the periplasmic fractions containing the soluble recombinant proteins were obtained from osmotically shocked bacteria and dialyzed against 50 mM sodium phosphate buffer, pH 8, overnight at 4 °C in dialysis tubing (Sigma).
Affinity Purification of the Pseudopilin Soluble Domains—Dialyzed periplasmic fractions were applied on a 5-ml Hi-Trap chelating column (Amersham Biosciences), loaded with nickel, operated by an AKTA Prime liquid chromatography system (Amersham Biosciences). After equilibration of the column with buffer A (50 mM sodium phosphate, pH 8, and 500 mM NaCl) supplemented with 10 mM imidazole, the periplasmic fractions were applied to the column. After several washing steps with buffer A supplemented with 20–50 mM imidazole, His6-tagged proteins were eluted with buffer A gradually supplemented with 50–500 mM imidazole. Xcp proteins-containing eluted fractions were pooled and dialyzed overnight at 4 °C against 50 mM sodium phosphate, pH 8, and 150 mM NaCl. After dialysis proteins were concentrated using Centricon devices (Amicon; Biomax-5) with a cut-off size of 5 kDa. The concentrations reached were 10.4 mg ml–1 for XcpTp25–148NH and 2.82 mg ml–1 for XcpXp23–333NH, as evaluated by UV spectra and Bradford colorimetric measurement. Purity, before and after concentration, was checked by analysis on 15% SDS-PAGE and Coomassie Blue staining.
Chemical in Vitro Cross-linking—Purified periplasmic soluble domains (PD), XcpTp25–148NH and XcpXp23–333NH, were used for in vitro chemical cross-linking. 5 µM of each protein was mixed and diluted in 35 µl of 50 mM sodium phosphate, pH 8, and 150 mM NaCl. After 20 min of incubation at room temperature, 4.3 µl of freshly prepared paraformaldehyde (1% final) was added to the mixture. Incubation was continued at room temperature for 20 min, after which the reaction was blocked by adding 14.3 µl of 4-fold concentrated SDS-
-mercaptoethanol buffer. 20 µl of each mix was directly analyzed on a 4–15% gradient SDS-PAGE (Bio-Rad). The gel was either directly stained with Coomassie Blue or immunoblotted with the Penta-His horseradish peroxidaseconjugated mouse IgG1 (Qiagen) following the instructions of the kit. The membranes were developed by chemiluminescence (Pierce). Interesting bands revealed in the first dimension and stained by Coomassie Blue were cut out from the 4–15% acrylamide gel, boiled or not for 20 min in plastic films immersed in a water bath, and placed at the bottom of large wells prepared in a 12% and 1.5-mm-thick SDS-containing gels. After electrophoresis, the proteins were immunoblotted on nitrocellulose membranes, probed with the Penta-His horseradish peroxidaseconjugated mouse IgG1 (Qiagen), and revealed by chemiluminescence using the Super Signal kit (Pierce).
Western Blot Quantification—Densitometric quantification of bands on Western blot were carried out using the Scion image 1.62c software (for MacOS).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
hRpQ = PAORQ). In this context, the Xcp machinery is the only system available to handle XcpT assembly. We confirmed that this mutant was affected both in Hxc-dependent secretion and type IV piliation but presented a wild type level of Xcp-dependent secretion (data not shown). Each of the 12 xcp genes was then individually deleted in the PAOhRpQ strain. The 12 mutants were fully complemented for the Xcp-dependent secretion by introducing a plasmid that carried the corresponding xcp gene, as seen by the halo formed around colonies grown on a skim milk plate (data not shown). We further introduced the plasmid pMTWT in these 12 mutants. The bacteria were grown in conditions that induce pseudopilus formation (see "Materials and Methods"). XcpT extracellular exposition was analyzed by shearing, and pseudopilus assembly was checked by TEM. The results are summarized in Table III, and in all cases both techniques correlate. If each of the Xcp components is required for efficient secretion, it appeared not to be the case for pseudopilus assembly. Indeed, the mutant strains lacking either xcpP, -Q, -Z, -T, -U, -W, or -X genes were still able to assemble pseudopili. It should be noted that the xcpP mutant formed much less pseudopili (Table II and data not shown). In contrast the mutant strains lacking either the xcpA or -R, -S, -V, or -Y genes no longer exposed XcpT on the cell surface nor assembled a pseudopilus. We concluded that among the 12 Xcp proteins required for the secretion process, only a restricted number of secreton components are essential for pseudopilus assembly. The absence of requirement for XcpQ in pseudopilus biogenesis was unexpected. It is, however, very likely that the XqhA secretin replaces XcpQ for pseudopilus assembly as it does for Xcp-dependent secretion (24). Moreover, in the P. aeruginosa genome (25), the xqhA gene (PA1868) is organized in tandem with another gene, PA1867. We found that the product of this gene shares homology with the XcpP protein (46% similar). We suggest that the PA1867 gene product could replace XcpP during pseudopilus assembly.
|
|
hxcpilA-C. In this genetic background we constructed single deletions of the xcpP, xcpR, xcpU, or xcpV gene. These mutants were analyzed for pseudopilus assembly and behave similarly as the mutants constructed in the PAOhRpQ genetic background. Indeed, pseudopilus assembly was observed for the PAOhxcpilA-CxcpP and PAOhxcpilA-CxcpU mutants but not for the PAOhxcpilA-CxcpR and PAOhxcpilA-CxcpV mutants (data not shown), confirming that XcpR and XcpV are essential components for the pseudopilus biogenesis, whereas XcpP and XcpU are not.
XcpV Is the Only Minor Pseudopilins Required for XcpT Incorporation into the Pseudopilus—Despite being involved in pseudopilus ultrastructure, the minor pseudopilins XcpU–X could be involved in a subtle control of its biogenesis. In our systematic analysis of the involvement of xcp genes in pseudopilus biogenesis, we demonstrated that among the four minor pseudopilin genes, only the xcpV gene is essential for pseudopilus assembly (Table II). Indeed, the analysis of the mutant PAOhRpQxcpV overproducing XcpT did not allow the observation of any pseudopili by TEM (data not shown). Moreover, shearing the cells also did not allow any release of XcpT in the supernatant (Fig. 5A, lane 4). Conversely, our TEM results demonstrated that XcpU, XcpW, and XcpX are not essential for pseudopilus assembly (Table II), and the shearing analysis revealed that XcpT could be recovered in the supernatant fraction (Fig. 5A, lanes 3, 5, and 6, respectively) of the corresponding mutants. We concluded that XcpV is the only minor pseudopilin required for the assembly of XcpT into pseudopilus.
XcpX Controls the Pseudopilus Assembly—Strikingly, the amount of XcpT recovered in the supernatant fraction is much more abundant with an xcpX mutant (Fig. 5A, compare lanes 1 and 6), which suggests a deregulation in the XcpT extrusion process and an abnormal elongation of the pseudopilus. Observations using immunofluorescence microscopy confirmed that the higher level of XcpT corresponded to the assembly of numerous (Fig. 5C compare panels 1–3 to panels 4–8) and longer (compare panels 1–3 to panel 9) pseudopili as compared with the parental strain. If the lack of XcpX allows an overassembly of XcpT into pseudopili, we reasoned that overproduction of XcpX could interfere with pseudopilus assembly. The xcpX gene was thus cloned into a broad host range plasmid (pBBR2-XhWTH) and introduced in a strain that contains the pMTWT. We checked pseudopilus formation by shearing analysis and TEM. As it is shown in Fig. 5B, the shearing analysis indicated that the release of XcpT into the extracellular medium was strongly decreased (62%) upon XcpX overproduction (Fig. 5B, lane 2). These results were confirmed by TEM observation, which showed that the decreased extracellular exposition of XcpT was associated with a significant decrease in the number of assembled pseudopili (data not shown). We thus concluded that the amount of XcpX produced could strongly contribute to the control of XcpT incorporation into pseudopili.
|
hxcRpilQ strain had a tendency to form bacterial clusters, whereas the non-overproducing strain was found attached as single bacterium (Fig. 6, compare A and D). Such a phenotype could not be seen with an xcpR mutant derivative overproducing XcpT (Fig. 6E), confirming that the microcolony formation phenotype is related to the Xcp-dependent biogenesis of the pseudopilus. More interestingly, when XcpT was overproduced in an xcpX mutant derivative (Fig. 6F), all the bacteria appeared to be organized in microcolonies, with only few isolated cells as compared with what was observed with a PAOhxcRpilQ strain (Fig. 6, compare D and F). Overall, by looking at pseudopilus-related functions such as adherence and biofilm formation, we could confirm that the lack of XcpX results in an increase in pseudopilus assembly. The E+5 Residue of XcpX Is Important for XcpX Function— The XcpX pseudopilin is atypical because of a larger size and the lack of the highly conserved glutamate residue at position +5 after the prepilin peptidase cleavage site. It has been proposed (16) that this residue may help the interaction between pilin subunits during type IV pilus assembly. We investigated whether introduction of a glutamate at position +5 may influence the function of XcpX. To test this hypothesis we used an xcpX allele mutation in which, at the corresponding +5 position, the threonine codon was changed for a glutamate codon (9). As for the other pseudopilins, the region encoding the leader peptide of XcpX was replaced with the 5' end of the xcpT gene, whereas the addition of a C-terminal His6 tag was engineered to yield a recombinant protein called XcpXhT+5EH. The mutated allele was cloned into the pBBR1MCS2 broad host range vector and mobilized in the PAO1 strain containing pMTWT. Co-overproduction of XcpT along with XcpXhT+5EH resulted in a much weaker exposition of XcpT. Indeed, upon pseudopili shearing, a decrease of about 95% in the amount of recovered XcpT was observed as compared with the PAO1 strain that overproduced XcpT alone (Fig. 5B, compare lanes 1 and 3). This reduction in the exposition of XcpT is more drastic in this case as compared with what was observed with the wild type XcpXhWTH (Fig. 5B, compare lanes 2 and 3). Moreover, with a strain that co-overproduces XcpXhT+5EH and XcpT, we could not observe any single pseudopilus using TEM (data not shown). We concluded that the presence of a glutamate at position +5 in XcpX might alter the function of this pseudopilin and strengthen the negative control exerted by XcpX on XcpT extrusion and further pseudopilus assembly.
XcpX Increases Proteinase K Sensitivity of XcpT—The observation that strains overproducing both XcpT and XcpX had a severe defect in pseudopilus assembly suggested that XcpX interferes with XcpT assembly. Most interestingly, we observed by immunoblotting with whole cell extracts that the level of detectable XcpT is strongly decreased when XcpT and XcpX are co-overproduced in the E. coli strain BL21. In these conditions, XcpT could only be detected in the cells after 6 h of induction, whereas it is readily detected after 2 h when XcpT is produced alone (Fig. 7A, compare lanes 1–3 and lanes 4–6). Moreover, this time delay for reaching a detectable level of XcpT is even longer when co-producing the mutant form XcpXT+5E (Fig. 7A, compare lanes 4–6 and lanes 7–9). In contrast, the level of XcpX production was not strongly affected by the co-expression of XcpT (Fig. 7B), and more importantly, this level is strictly identical when considering either XcpX or the XcpXT+5E derivative (Fig. 7B, compare lanes 4–6 and 7–9). These observations suggested that XcpX could have a destabilizing effect toward XcpT. In order to check this possibility, we assessed the proteinase K sensitivity of XcpT produced in these different conditions. Upon overproduction in E. coli, both proteins were extracted from the inner membrane and solubilized with the zwitterionic sulfobetaine detergent n-dodecyl-n,n-dimethyl-3-ammonio-1-propanesulfonate. The extracts were used in a proteinase K sensitivity assay. As shown in Fig. 7C, XcpT is degraded by proteinase K in a time-dependent manner but is still detectable after 60 min of proteolysis, with a rate of degradation of 73% (Fig. 7C, lane 3). However, the proteinase K digestion kinetic of XcpT is strongly increased when XcpX is present (Fig. 7C, compare lanes 1–3 and lanes 4–6). In these conditions 60% of the XcpT protein is degraded after 10 min (Fig. 7C, lane 5), whereas no product is detected after 1 h of incubation (Fig. 7C, lane 6). When XcpT was produced in the presence of the XcpXT+5E mutant, the kinetic of XcpT degradation was further increased (Fig. 7C, compare lanes 4–6 and lanes 7–9). We concluded that the presence of XcpX and XcpT in the same biological sample could trigger conformational changes that lead to a greater instability of XcpT.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
In contrast to the work reported about the Pul secreton (13), we observed that XcpZ, the PulM homolog, is not essential for the assembly of XcpT into pseudopilus (Table II). XcpZ is a bitopic inner membrane protein (28), which interacts with XcpY (29, 30). Such an interaction results in a mutual stabilization between XcpY and XcpZ. Moreover, it has been demonstrated that membrane association of the putative ATPase XcpR required the interaction with the N-terminal domain of XcpY (31). In the absence of XcpZ, we thus might expect a lower level of XcpY within the cell and consequently a partial mislocalization of XcpR. Because XcpZ is dispensable for pseudopilus assembly, whereas XcpR and XcpY are not (Table II), this suggests that a lower level of XcpY and XcpR is sufficient to support pseudopilus assembly but not protein secretion. Because XcpZ is essential for protein secretion, it is thus likely that the function of this component is not only to stabilize XcpY, but it might also be related to exoprotein recognition, for example.
The observation that, in the absence of XcpQ, a normal amount of pseudopili could still be seen at the bacterial cell surface was intriguing (Table II and Fig. 5). Indeed, with type IV pili from Neisseria gonorrhoeae, it was shown that the filament crosses the outer membrane via the pore formed by PilQ (32). Because PilQ and XcpQ are both members of the secretin family (33), we anticipated that pseudopilus extrusion needed the XcpQ homomultimeric pore. Martinez et al. (24) have reported previously that in P. aeruginosa the secretin XqhA can replace XcpQ in the secretion process, and we suggest that it also replaces XcpQ for pseudopilus assembly.
Function of the Minor Pseudopilins in Pseudopilus Assembly—We demonstrated that none of the minor pseudopilins, XcpU–X, is able to form a pseudopilus such as XcpT does. Moreover, it seems that these proteins are not part of the extracellular pseudopilus (Fig. 4). Similarly, Vignon et al. (14) have shown that the purified PulG-pili did not contain any of the four other minor pseudopilins, PulH–K. We demonstrated that XcpV is the only minor pseudopilin required for pseudopilus assembly. Likewise, PulI, an XcpV homolog of the Pul secreton, appeared to be essential for the assembly of the PulGpili (14). One could imagine that XcpV acts at an early stage during fiber assembly, such as the initiation step, and is found at the base of the pseudopilus, in connection with the inner membrane components of the Xcp secreton.
Another minor pseudopilin, XcpX, seems to play a crucial function, albeit not essential, in pseudopilus assembly. Indeed, our results showed that, in the absence of XcpX, pseudopilus assembly still occurs but the number and length of pseudopili increased markedly (Table II and Fig. 5). Reciprocally, overproduction of XcpX interfered with or abolished pseudopili formation (Fig. 5). We hypothesized that XcpX could control pseudopilus assembly, by limiting the incorporation of XcpT, and we began to investigate the molecular mechanism involved in this control. First, we demonstrated that XcpX can be cross-linked with XcpT (Fig. 8). Second, we showed that co-overproduction of XcpX with XcpT diminished the amount of XcpT normally recovered (Fig. 7A). Third, it appeared that XcpT was more sensitive to proteinase K digestion, in an in vitro assay, when incubated together with XcpX (Fig. 7C). These observations suggest that XcpX interacts with XcpT and that interaction induces conformational changes that destabilize XcpT and increase its sensitivity to proteolytic degradation. Remarkably, the effects observed with the wild type XcpX are strengthened when using the mutant protein XcpXT+5E (Fig. 5B and Fig. 7, A–C). We suggest that XcpXT+5E interacts more strongly with XcpT than XcpX, increasing the conformational changes and the destabilization of this protein.
How Is Fiber Elongation Controlled in Other Extracellular Appendage Assembly Machineries?—The CS5 pili produced by the enterotoxigenic E. coli involves a chaperone usher-type assembly system. CsfA is the major subunit that comprises mostly the extracellular pilus and requires the specific chaperone CsfB, whereas CsfD and CsfE are minor subunits interacting with the chaperone CsfF (34). Pilus biogenesis is thought to be initiated by the binding of the CsfD-CsfF complex to the outer membrane usher CsfC, resulting in the translocation of CsfD across the outer membrane and leaving CsfC in an assembly-competent state. Pilus elongation proceeds via incorporation of multiple copies of the CsfA major subunit and sometimes incorporation of fewer copies of CsfD, adding flexibility to the fiber. Duthy et al. (34) also demonstrated that pilus termination relies on the interaction of the CsfE-CsfF complex with the usher CsfC. Indeed, the csfE mutant assembled pili three times longer than those of the wild type strain. Moreover, production of CsfE in the csfE mutant cannot restore the wild type pili length. The authors concluded that CsfE did control pilus length but was not rate-limiting. Furthermore, the csfF mutant exhibited the same alteration in pilus length as the csfE mutant, but in this case the default was fully rescued by expression of CsfF, showing that the chaperone CsfF was rate-limiting for the control of pilus length. Duthy et al. (34) proposed a model in which the CsfE-CsfF complex is targeted to CsfC, resulting in the irreversible association of CsfE with CsfC and preventing further polymerization of the pilus. This mechanism has been originally described for the P pili and type 1 pili expressed by uropathogenic E. coli (35, 36). In the XcpT-pseudopilus assembly, the lack of XcpX or its overproduction led, respectively, to an increase of the pseudopilus assembly (number and length) or to a total abolition. It thus suggests that XcpX does have a rate-limiting function in the regulation of pseudopilus biogenesis and might also indicate that XcpX operates alone.
Based on these observations, we proposed a model in which XcpX could play a central role in controlling the assembly of XcpT into pseudopili. In the case of P. aeruginosa or N. gonorrhoeae type IV pili, which are long and filamentous fibers, two antagonistic "traffic warden" ATPases are required for pilus function. One, PilB/PilF, is involved in the assembly of the major pilin PilA/PilE into pili, and the other, PilT, is required for disassembly of the filament (32, 37). The extracellular XcpT-containing pseudopilus is only observed upon XcpT overproduction. The physiological size of the pseudopilus in the type II secreton might thus not exceed the dimension of the periplasmic space (20 nm = 5 x repeat units = 10 x XcpT (15)). It has been proposed that the pseudopilus must retract to allow protein secretion (2, 15). It is thus reasonable to suggest that the energetic cost to assemble and disassemble the pseudopilus might not be as high as for type IV pili. In support of this hypothesis, type II secretion systems are known to use one single traffic warden ATPase, the GspE. Either GspE might support both pseudopilus assembly and disassembly or GspE might only be required for assembly, whereas disassembly is achieved via another mechanism. XcpX is a minor and non-abundant pseudopilin as compared with XcpT. We have shown that XcpX and XcpT interact and that the XcpT/XcpX ratio appeared to be crucial for pseudopilus elongation. In physiological conditions, the frequency of contact between XcpX and XcpT is low, and XcpT-XcpT interactions might be favored to allow pseudopilus growth. The nonfrequent XcpX interaction with an XcpT
