Extracellular Yersinia adhering at the surface of a eukaryotic cell translocate effector Yops across the plasma membrane of the cell by a mechanism requiring YopD and YopB, the latter probably mediating pore formation. We studied the role of SycD, the intrabacterial chaperone of YopD. By producing GST–YopB hybrid proteins and SycD in Escherichia coli, we observed that SycD also binds specifically to YopB and that this binding reduces the toxicity of GST–YopB in E. coli. By analysis of a series of truncated GST–YopB proteins, we observed that SycD does not bind to a discrete segment of YopB. Using the same approach, we observed that YopD can also bind to YopB. Binding between YopB and YopD occurred even in the presence of SycD, and a complex composed of these three proteins could be immunoprecipitated from the cytoplasm of Yersinia. In a sycD mutant, the intracellular pool of YopB and YopD was greatly reduced unless the lcrV gene was also deleted. As LcrV is known to interact with YopB and YopD and to promote their secretion, we speculate that SycD prevents a premature association between YopB–YopD and LcrV.
In this paper, we address the question of the role of SycD. We show that SycD binds to multiple sites on YopB and that YopB binds to YopD. SycD does not prevent the cytosolic association of YopB and YopD but, rather, is associated with the complex. This complex does not include LcrV, although LcrV can bind to YopB and to YopD.
SycD is not a secreted protein
To localize SycD, Yop synthesis was induced, and the secreted proteins were collected. After sonication of the bacteria, the soluble and insoluble fractions were separated by centrifugation. The three fractions were then analysed by Western blotting. As shown in Fig. 1, SycD was only seen in the soluble bacterial fraction and not in the insoluble fraction containing bacterial membranes or among the secreted proteins (Fig. 1). This is in agreement with the observations made by Williams and Straley (1998) for Y. pestis. SycD is thus an intrabacterial protein, most probably cytosolic like the other Yersinia Syc chaperones.
SycD binds to YopB
As the absence of SycD abolishes secretion of YopB as well as of YopD (Wattiau et al., 1994), we wondered whether SycD could be a bifunctional chaperone serving both YopD and YopB. To test the capacity of SycD to interact with YopB, we took advantage of the glutathione-S-transferase (GST) hybrid protein expression and purification system (Smith and Johnson, 1988). pCN13, a plasmid that encodes both GST–YopB and SycD was constructed, and both proteins were overproduced simultaneously in E. coli XL1-blue. To serve as a control, pCN14 encoding GST and SycD was constructed by deleting the yopB gene from pCN13. As an additional control, pCN46, which encodes GST–YopE and SycD, was constructed. After overproduction of the proteins of interest in E. coli, cleared extracts were mixed with glutathione–Sepharose beads, and proteins absorbed on the beads were analysed by Western blotting. As shown in Fig. 2, SycD co-purified with GST–YopB, but not with GST alone and not with GST–YopE. From this experiment, we conclude that SycD specifically binds to YopB.
In the course of our experiments, we observed that overproduction of GST–YopB led to the death of E. coli XL1-blue, while E. coli XL1-blue producing GST–YopB and SycD simultaneously survived much better, suggesting that SycD binds to YopB in vivo. To investigate this effect in more detail, we monitored the growth curve and the survival of E. coli overproducing GST–YopB in the absence and in the presence of SycD. Overexpression of the gst–yopB hybrid gene alone led to lysis of the culture (Fig. 3A), the number of colony-forming units was reduced 104-fold (Fig. 3B) and the GST–YopB protein could not be overproduced (Fig. 3C). However, when GST–YopB and SycD were produced simultaneously, lysis was less pronounced and delayed (Fig. 3A), the viable count was only reduced twofold (Fig. 3B) and the hybrid protein could be overproduced (Fig. 3C). As a control, we monitored the growth curve and the survival of E. coli XL1-blue overproducing GST: no lysis was observed (Fig. 3A) and the viable counts were not reduced (Fig. 3B). These data thus show that the presence of SycD makes the GST–YopB hybrid protein less toxic in vivo, supporting the fact that SycD binds specifically to YopB.
YopB binds to YopD
The similarity between SycD and IpgC, which prevents a premature cytoplasmic association between its two partners IpaB and IpaC (Ménard et al., 1994), suggested that SycD could play a similar role by preventing a premature association between its cognate YopB and YopD. However, YopB and YopD have never been shown to bind to each other. To demonstrate such an association, we first used the glutathione-S-transferase purification method. The sycD gene was replaced by the yopD gene in pCN13, yielding plasmid pCN22, which encodes GST–YopB and YopD. As a negative control, pCN29, which encodes GST and YopD separately, was used. As shown in Fig. 4, incubation of the cleared extract from E. coli XL1-blue(pCN22) with glutathione beads led to the recovery of GST–YopB and YopD (lane 2), whereas no YopD was recovered on the beads incubated with the cleared extract of E. coli XL1-blue(pCN29) (lane 4). Western blot analysis confirmed that the protein that co-purified with GST–YopB is indeed YopD (data not shown). This result suggested that YopB and YopD can bind to each other.
SycD does not prevent the association between YopB and YopD
SycD can thus bind to YopB as well as to YopD, and these last two proteins also have the capacity to bind to each other. We then tried to determine whether these two associations are mutually exclusive. We again used the GST hybrid co-purification system in which the partners are produced simultaneously in the same bacterium, encoded by the same plasmid. Either GST–YopB plus YopD (pCN22) or GST–YopB plus YopD and SycD (pCN18) were overproduced in E. coli XL1-blue. After incubation of the cleared extracts with glutathione beads, the purified proteins were analysed by SDS–PAGE. Figure 4 shows that YopD co-purified with GST–YopB in the absence as well as in the presence of SycD (lanes 1 and 2). Moreover, when SycD was present, it co-purified with GST–YopB and YopD (lane 1). This suggested that SycD does not prevent binding of YopD to YopB when they are produced in similar amounts. However, one could argue that, as YopD and SycD are produced simultaneously, they could compete for binding to YopB or they could bind to different molecules of GST–YopB.
To clarify this point, the experimental protocol was modified so that GST–YopB would not be produced simultaneously with SycD and YopD. Either GST–YopB + SycD or GST alone serving as a control were overproduced in E. coli XL1-blue. The cleared extracts were mixed with glutathione beads, and the beads loaded with GST–YopB + SycD or GST were washed and incubated overnight with a cleared extract containing YopD [E. coli LK111(pGP1-2)(pBB5)], YopD plus SycD [E. coli LK111(pGP1-2)(pBB4)] or YopD deleted from its hydrophobic segment [E. coli LK111(pGP1-2)(pBB9)]. As shown in 5Fig. 5A, the presence of SycD on GST–YopB did not prevent binding of YopD, even when YopD itself was bound to SycD. Beads covered with GST alone did not bind any protein from the cleared extracts. We assume that the GST–YopB molecules on glutathione beads were saturated with SycD because an additional incubation of the beads with a cleared extract containing SycD [LK111(pGP1-2)(pPW71)] did not increase the quantity of SycD bound to the beads (Fig. 5B). The amount of SycD in the latter extract was in large excess, as it contained at least 500 times the amount of SycD already fixed on the beads (Fig. 5B, lane 3). Likewise, we checked that the YopD molecules obtained from E. coli LK111(pGP1-2)(pBB4) were also saturated with SycD. After immunoprecipitation of the YopD–SycD complexes with an anti-YopD antibody on protein A–Sepharose beads, the beads were reincubated with a cleared extract containing SycD [LK111(pGP1-2)(pPW71)], and no increase in the amount of SycD bound to the beads could be observed (Fig. 5C). We can thus assume that both GST–YopB and YopD were saturated with SycD before contact, and we conclude that, even if both partners are bound to SycD, YopB and YopD can still interact.
YopB and YopD are associated in Yersinia
The observation that YopB and YopD can associate in the presence of SycD suggests that they are probably associated in the Yersinia cytoplasm. To demonstrate such an association, we tried to co-immunoprecipitate the complex. Yop synthesis was induced in wild-type Y. enterocolitica MRS40 and in a yopB mutant [E40(pPW401)]. The bacteria were then washed with xylene (Michiels et al., 1990) or incubated with proteinase K to eliminate the secreted Yop proteins that could be absorbed to the outer membrane. After sonication, the cleared extracts were incubated with anti-YopB polyclonal antibodies, and the complex was then recovered on protein A–Sepharose beads. After washing of the beads, the purified proteins were analysed by Western blotting. Figure 6 shows that YopD and SycD were recovered from the extracts containing YopB but not or only poorly from the extracts that did not contain YopB. We also probed the complex for LcrV and did not detect any signal (data not shown). This indicates that YopB, YopD and SycD form a complex in the Yersinia cytoplasm. SycD thus does not prevent the association of YopB and YopD, but it is associated with the YopB–YopD complex.
LcrV interacts with the YopB–YopD complex in the absence of SycD
Wattiau et al. (1994) showed that YopB and YopD are not secreted in a sycD mutant, presumably because they are degraded. We wondered whether this degradation would not depend on the presence of their partner for secretion, LcrV (Sarker et al., 1998a). The intrabacterial content of YopB and YopD was analysed in various mutants 2 h after the induction of Yop synthesis. As shown in Fig. 7, the intracellular pool of YopB and YopD was strikingly reduced when SycD was missing (lane 3). This phenotype could be complemented by overexpressing sycD from plasmid pPW64, indicating that the mutation was indeed non-polar (lane 4). However, if LcrV was also missing, the intracellular pool of YopB and YopD was maintained (lane 2) and, if LcrV was overproduced in trans in the lcrV sycD mutant, the intracellular content in YopB and YopD was again reduced (lane 7). This indicated that LcrV is involved in the reduction of the intracellular pool of YopB and YopD. As LcrV is known to be able to interact with YopB and YopD (Sarker et al., 1998a), we could speculate that SycD prevents a premature association between YopB–YopD and LcrV. In this regard, it is, however, noteworthy that, in the sycD mutant, the intracellular pool of LcrV is increased.
Localization of the SycD binding segment of YopB
In an attempt to understand the mechanism of action of SycD better, we investigated its binding to YopB. Starting from the hypothesis that SycD is a chaperone like SycE and SycH, we tried to identify a discrete binding site on YopB. A series of eight in-frame deletion mutants spanning the entire yopB was constructed. The deletions, called Δ1 to Δ8 (Fig. 8), remove all the characteristic segments of the protein, such as the two hydrophobic segments (Håkansson et al., 1993) and the hypothetical coiled coils spanning residues 103–168 and residues 330–385 (Lupas et al., 1991).
The mutated genes were then substituted for yopB in pCN13, and all the truncated proteins, fused to GST, were overproduced with SycD in E. coli XL1-blue. Cleared extracts were mixed with glutathione–Sepharose beads, and the purified proteins were analysed by SDS–PAGE. Mutants Δ1, Δ2, Δ4, Δ5, Δ6 and Δ7 were seen to bind to SycD (Fig. 9A). Mutants Δ3 and Δ8 were recovered in very low amounts because of lysis of the culture (data not shown), making the analysis impossible. Half of the hydrophobic segment (Δ5) of these mutants was then removed in order to avoid lysis, as YopB without the hydrophobic segment is not toxic (C. Neyt and G. R. Cornelis, unpublished). To our surprise, GST–YopBΔ3Δ5 and GST–YopBΔ5Δ8 could still bind SycD (Fig. 9B). The fact that mutants Δ3 and Δ8 induced lysis of the culture even in the presence of SycD is interesting because, as we have shown above, SycD normally prevents lysis. This suggested that both segment 3 and segment 8 could be binding sites for the chaperone. In an attempt to clarify this, a triple deletant GST–YopBΔ3Δ5Δ8 was constructed and binding to SycD was monitored. Unexpectedly, the hybrid protein was recovered in high amounts, and it could bind SycD. To investigate this point further, segment 3 and segment 8 were fused to GST, and binding of these fusions to SycD was analysed. No binding could be seen (data not shown), but this could be the result of an altered conformation of these small segments fused to GST. From all these results, we conclude that SycD does not bind to a unique clearly defined segment of YopB, but rather binds at different sites of the protein. Chaperone SycD thus appears to be quite different from SycE and SycH. Not only does it bind two cognate Yops rather than one, but it binds several segments rather than a discrete segment.
In this study, we attempted to clarify the role of SycD in the Yop translocation mechanism. We showed first that SycD is cytosolic and that it binds the translocator YopB as well as YopD. As SycD was already known to bind to YopD, we have shown here that SycD is bifunctional, serving both the YopB and the YopD translocators. This observation is in perfect agreement with the observations of Wattiau et al. (1994) that SycD is necessary for the secretion of both YopB and YopD. Interestingly, the Shigella homologue of SycD also serves two secreted proteins, namely IpaB and IpaC, by binding to each of them and preventing their premature association (Ménard et al., 1994).
The similarity between SycD and IpgC suggested that SycD could play a similar role and could thus prevent a premature association between YopB and YopD. We showed first that YopB and YopD are indeed able to associate. However, this association occurred even if both partners were bound to SycD, which suggested that YopB, YopD and SycD could be associated in the bacterial cytoplasm. Co-immunoprecipitation experiments carried out on a Yersinia cleared extract indeed allowed recovery of YopB associated with YopD and SycD. This result indicates that SycD not only protects the two proteins separately, but can also protect a complex composed of these two proteins. This could explain why, at variance with SycE and SycH, SycD is a bivalent chaperone.
The cytoplasmic association of YopB and YopD suggests that they can be secreted as a YopB–YopD heteromer, and SycD would presumably be released at the secretion step. The fact that YopB and YopD could be secreted as a complex can be correlated with previous observations related to the secretion apparatus. The YscC secretin inserts in the outer membrane and assembles as a very stable polymeric complex that forms a large channel (Koster et al., 1997). The lipoprotein VirG, which stabilizes YscC in the outer membrane (Koster et al., 1997), is necessary for normal secretion of YopB, YopD and LcrV, but not for secretion of the other Yops (Allaoui et al., 1995a). Furthermore, truncation of YscF, a protein from the secretion apparatus, reduces secretion of YopB and YopD, but not of the other Yops (Allaoui et al., 1995b). These observations thus indicate that the secretion of YopB and YopD is more sensitive to small alterations in the secretion machinery than the secretion of the other Yops, supporting the hypothesis that YopB and YopD are secreted as a bulky complex, thus needing a well-formed secretion apparatus.
If YopB and YopD are secreted as a complex, as suggested by our results, one could predict that YopD is also part of the pore, which is consistent with the transmembrane nature of this protein. However, YopD has recently been shown to be translocated into the eukaryotic cell cytosol (Francis and Wolf-Watz, 1998). To reconcile these two elements, one could imagine that YopD is only a transient component of the pore that is then translocated with the Yop effectors, acting as a kind of extrabacterial chaperone needed for the translocation.
Starting from the hypothesis that the Yop translocation pore is initially formed by YopB and YopD, we would like to compare it with the pore formed by perforin, which is produced by cytolytic lymphocytes (for a review, see Liu et al., 1995). The similarity between Yersinia and cytolytic lymphocytes is indeed particularly striking. In both cases, an attacker cell contacts a target cell, forms a pore and delivers proteins, some of which induce apoptosis (Liu et al., 1995; Mills et al., 1997; Monack et al., 1997; Ruckdeschel et al., 1997; Schesser et al., 1998). The similarity at the global level might also apply to the mechanism of pore formation. Perforin monomers insert into the target membrane and polymerize into a central opening (pore), which grows in diameter through the progressive recruitment of additional monomers (for review, see Liu et al., 1995) allowing granzyme B delivery into the target cell. According to this model, the Yersinia translocation pore would be formed by gradual polymerization of YopB–YopD heteromers.
Wattiau et al. (1994) have already described degradation of YopD in a sycD mutant. We observed here that the intracellular pool of YopB is also reduced in a sycD mutant. This reduction could be caused by degradation of the proteins. However, we also observed that this reduction in YopB and YopD does not occur if LcrV is also absent, suggesting that LcrV is involved in the reduction of the intracellular pool of YopB and YopD. Sarker et al. (1998a) have shown that LcrV has the capacity to bind to YopB and to YopD, and that LcrV is necessary for their secretion. In contrast, SycD does not bind to LcrV (Fields et al., 1997; Sarker et al., 1998a). Combining these data with our observations, we would like to suggest the following model. YopB, YopD and SycD would be associated in the bacterial cytoplasm. At the moment of secretion, SycD would be released and LcrV would bind to YopB and YopD helping them to be secreted and to assemble into the translocation apparatus. The binding of YopB–YopD with LcrV is probably only transitory, occurring during secretion through the bacterial membrane. If SycD is missing, the LcrV–YopB–YopD association could occur prematurely, leading to the degradation of YopB and YopD. However, if both SycD and LcrV are missing, YopB and YopD are still present in the bacteria, but they are not secreted. SycD would thus prevent a premature association between LcrV and the YopB–YopD complex, a role reminiscent of that of IpgC from Shigella (Ménard et al., 1994). Further efforts need to be devoted to the study of the hypothetical antagonism between LcrV and SycD.
In an attempt to understand the mechanism of action of SycD better, we tried to identify a discrete binding site on YopB. We constructed a series of in frame deletion mutants spanning the entire protein. Surprisingly, all the mutants could bind SycD. Furthermore, truncated YopB proteins missing two or three segments could still bind to SycD. Together, these observations indicate that SycD is able to bind to several segments on YopB, which is consistent with the roles we have described. This property evokes SecB, a molecular chaperone, from the general secretory pathway in E. coli, which is dedicated to the export of newly synthesized proteins (Khisty et al., 1995) and binds to multiple sites on its targets (Kumamoto and Beckwith, 1985). This situation is thus different from that of SycE and SycH, which are known to bind to their cognate Yop at a well-defined segment (Schesser et al., 1996; Woestyn et al., 1996). SycD thus appears to be a chaperone different from the SycE and SycH chaperones with regard to its roles and the way it acts. It serves a complex consisting of two proteins rather than a single protein as do the other Syc chaperones, and it binds to several segments of the protein, while the other chaperones bind a single well-defined segment. Thus, the Syc family seems to comprise two subfamilies: one including SycE, SycH and SycT (Wattiau et al., 1993, 1994; Iriarte and Cornelis, 1998), which act on the Yop effectors (YopE, YopH and YopT), and the other one, represented by SycD, which acts on translocators (YopB and YopD). There are representatives of this SycD subfamily in several other type III systems, namely PcrH in Pseudomonas aeruginosa (Yahr et al., 1997), IpgC in Shigella (see above) (Ménard et al., 1994), SicA in Salmonella (Kaniga et al., 1995) and CesD in EPECs (Wainwright and Kaper, 1998). Genetic organization and sequence analyses indicate that some of these type III systems, such as the Pop system of P. aeruginosa, are very close to the Yersinia Yop virulon (Yahr et al., 1997), whereas other type III systems, such as the Shigella Ipa system and the EPEC Esp system, appear to be more distantly related. In the latter system, the mode of action of the CesD chaperone could be somewhat different as, unlike SycD and IpgC (Ménard et al., 1994), a substantial amount of CesD from EPEC is associated with the inner membrane (Wainwright and Kaper, 1998). However, the presence of a chaperone of the SycD subfamily in many type III systems reinforces the idea that it is a key player in the synthesis and assembly of the translocation apparatus. The results presented here concerning SycD could also benefit the understanding of the other type III secretion systems.
Bacterial strains, plasmids, growth conditions and genetic conjugations
Y. enterocolitica KNG22703 and MRS40 are the blaA mutants of strains W22703 and E40, respectively, in which the gene encoding β-lactamaseA was replaced by the luxAB genes (Kaniga et al., 1991; Sarker et al., 1998b). Escherichia coli XL1 blue (Stratagene) was used for standard genetic manipulations and to produce the glutathione-S-transferase (GST) fusion proteins. E. coli LK111, received from M. Zabeau (Ghent, Belgium) was used to overproduce proteins with the T7 polymerase system (see below). E. coli CJ236 was used for site-directed mutagenesis (Kunkel et al., 1987). E. coli SM10 lambda pir+ constructed by Miller and Mekalanos (1988) was used to deliver the mobilizable plasmids into Y. enterocolitica. The plasmids used in this study are listed in Table 1.
Bacteria were grown routinely in tryptic soy broth (Oxoid) and plated on tryptic soy agar. For the induction of the yop regulon, Y. enterocolitica was grown in brain–heart infusion (BHI) broth supplemented with 4 mg ml−1 glucose, 20 mM MgCl2 and 20 mM sodium oxalate (BHIOX). Yersinia grown overnight at room temperature in BHI were inoculated to an optical density at 600 nm of 0.1 in 10 ml of BHIOX. Cultures were incubated for 2 h at room temperature and then shifted to 37°C for 4 h (Cornelis et al., 1987; Michiels et al., 1990). Conjugations were carried out as described by Sarker et al. (1998a).
SDS–PAGE analysis of proteins and immunoblotting
Proteins were analysed by SDS–PAGE and Western blotting as described previously (Cornelis et al., 1987; Allaoui et al., 1995b). For analysis of the whole-cell extract, 8 × 108 bacteria were applied to SDS–PAGE. Immunoblotting was carried out with rat monoclonal antibodies 7A7 (anti-YopD), 9B7 (anti-YopB), 6G1 (anti-YopE) and 7C1 (anti-LcrV) and with purified rabbit polyclonal antibodies against SycD. The amount of proteins transferred onto the membrane was checked by staining the membrane with Ponceau Red 0.2% (Sigma). Supersignal chemiluminescent substrates (Pierce) were used for chemiluminescence detection. Alternatively, detection was carried out with 4-chloro-1-naphthol (Sigma) and H2O2 (Merck).
Construction of pGEX-derived recombinant plasmids
The yopB gene was amplified by polymerase chain reaction (PCR) with pYVe227 as a template using amplimers MIPA123 (5′-CGTCTAGACATGAGTGCGTTGATAA-3′) and MIPA124 (5′-ATGTCGACTTAAACAGTATGGGGTC-3′) and was cloned into the XbaI–SalI sites of pGEX-KG, yielding plasmid pCN8 (encoding GST–YopB). The sycD gene preceded by the Shine–Dalgarno sequence of gene Φ10 was amplified from pPW71 with MIPA293 (5′-CGCAAGCTTAAGAAGGAGATATACATATGCAAC-3′) and MIPA287 (5′-CTCAAGCTTAGCGGTCATGGGTTATCAA-3′) and introduced into the HindIII site of pCN8, giving plasmid pCN13, encoding GST–YopB and SycD. Plasmid pCN14 was constructed by digestion of pCN13 with EcoRI and SalI restriction enzymes, followed by Klenow treatment and religation, leading to the elimination of the yopB gene. To construct plasmid pCN46, encoding GST–YopE and SycD, the yopE gene was amplified by PCR with pYVe227 as a template using amplimers MIPA498 (5′-CGTCTAGACGTCATGAAAATATCATCA-3′) and MIPA499 (5′-ACGTCGACTCACATCAATGACAGTAA-3′) and introduced into the XbaI–SalI sites of pCN13.
The yopD gene preceded by the Φ10 Shine–Dalgarno sequence was amplified from pBB5 using MIPA342 (5′-ATGTCGACTCAGACAACACCAAAAGC-3′) and MIPA343 (5′-AT-
GTCGACAAGAAGGAGATATACATATGACA-3′) and cloned into the XhoI site of pCN13, yielding plasmid pCN18 encoding GST–YopB, YopD and SycD. Partial digestion of pCN18 with HindIII and religation eliminated the sycD gene, giving plasmid pCN22 encoding GST–YopB and YopD. Digestion of pCN18 with XbaI and SalI restriction enzymes followed by Klenow treatment and religation led to the deletion of the yopB gene, giving plasmid pCN28 encoding GST + YopD + SycD.
Construction of yopB deletion mutants
The EcoRI–SacI fragment from pCN13 containing the yopB gene was subcloned into the corresponding sites of pBluescript SK+, yielding plasmid pCN15. In frame deletion mutants in yopB were constructed by site-directed mutagenesis as described by Kunkel et al. (1987) using E. coli CJ236 dut ung, giving plasmids pCN150, pCN151, pCN152, pCN153, pCN154, pCN155 and pCN156 (Table 1). Large deletions (yopBΔ4, yopBΔ6, yopBΔ8) were made in two or three steps, using smaller deletions as intermediates (pCN157, pCN158 and pCN159; Table 1).
The double deletion mutants yopBΔ3Δ5 and yopBΔ5Δ8 and the triple deletion mutant yopBΔ3Δ5Δ8 were constructed by the same method, giving plasmids pCN1513, pCN1563 and pCN15163 respectively (Table 1).
Construction of the GST–YopBΔ recombinant plasmids
Mutant yopBΔ1 was constructed by amplifying by PCR the yopB gene from codon 21 until the 3′ end using pYVe227 as a template and amplimers MIPA460 (5′-ACTCTAGACATCGAGACACCAGCG-3′) and MIPA124 (5′-ATGTCGACTTAAACAGTATGGGGTC-3′). This XbaI–SalI PCR fragment was then cloned into the corresponding sites of pCN13, replacing the wild-type yopB gene and yielding plasmid pCN37, which encodes GST–YopBΔ1 and SycD.
The yopBΔ deletion mutants (see above) were subcloned from the pBluescript derivatives (plasmids pCN150, pCN151, pCN152, pCN153, pCN154, pCN155, pCN156, pCN1513, pCN1563 and pCN15163) into pCN13 using EcoRI and SacI sites, thereby replacing yopB by yopBΔ and giving plasmids pCN130, pCN131, pCN132, pCN133, pCN134, pCN135, pCN136, pCN1313, pCN1363 and pCN13163 respectively. These plasmids encode various GST–YopBΔ + SycD.
Production and purification of hybrid GST fusion proteins
The production and purification of the fusion proteins was performed basically as described by Smith and Johnson (1988) and according to Pharmacia Biotech.
Overproduction of YopD, YopD + SycD, YopD deleted from its hydrophobic segment (YopDΔ121–165) and SycD
Overproduction of YopD, YopD + SycD, YopDΔ121–165 and SycD was carried out by the T7 RNA polymerase system described by Tabor and Richardson (1985), using strains LK111(pGP1-2)(pBB5), LK111(pGP1-2)(pBB4), LK111(pGP1-2)(pBB9) and LK111(pGP1-2)(pPW71) respectively.
Yop synthesis was induced in Y. enterocolitica as described above, but bacteria were incubated for only 2 h at 37°C instead of 4 h as in the initial protocol. After induction, the bacteria were treated with either xylene (Michiels et al., 1990) or proteinase K (Iriarte et al., 1998). For the proteinase K treatment, cultures were washed with phosphate-buffered saline (PBS), and proteinase K (Serva) was then added to a final concentration of 500 μg ml−1 for 30 min at room temperature. After incubation, phenylmethylsulphonyl fluoride (PMSF) was added to a concentration of 1 mM, and cells were washed three times with cold PBS containing 1 mM PMSF. After treatment, the bacterial pellet (5 × 109 bacteria) was resuspended in 1.5 ml of 10 mM Tris-HCl, pH 8.0, 5 mM MgCl2, 1 mM PMSF and sonicated. The soluble fraction was recovered by a 30 min centrifugation step (10 000 r.p.m.) at 4°C, and absorbed polyclonal anti-YopB antisera was added at a 1:500 dilution of the antibody. After 90 min of gentle rocking at 4°C, the immunocomplexes were harvested by the addition of 40 μl of protein-A CL-4B Sepharose (50% slurry in 10 mM Tris-HCl, pH 8.0, 5 mM MgCl2). The complexes were gently rocked for 2 h at 4°C, collected by centrifugation at 3000 r.p.m., washed five times in 10 mM Tris-HCl, pH 8.0, 5 mM MgCl2 and finally eluted from the beads by boiling in sample buffer. The purified proteins were then analysed by SDS–PAGE.
Immunoprecipitation of YopD and SycD from LK111(pGP1-2) (pBB4) was performed according to the same protocol, using absorbed polyclonal anti-YopD antisera at a 1:500 dilution and PBS instead of 10 mM Tris-HCl, pH 8.0, 5 mM MgCl2.
Cellular localization of SycD
After induction of Yop synthesis, the culture was harvested by 10 min centrifugation at 10 000 r.p.m. at 4°C. Secreted proteins were prepared by precipitation of the supernatant with TCA (10% final). The bacterial pellet was resuspended in cold 10 mM Tris-HCl, pH 8.0, 5 mM MgCl2 and sonicated. The intact cells were pelleted by a 5 min centrifugation step at 10 000 r.p.m. The supernatant was then centrifuged for 30 min at 10 000 r.p.m., allowing separation of the membrane fraction (pellet) from the soluble fraction (cytosol). For each fraction, the equivalent of 4 × 108 bacteria were analysed by Western blotting.
To establish the growth curves, overnight cultures were diluted to an OD600 of 0.1 in 10 ml of broth containing 200 μg ml−1 of ampicillin and 0.2% glucose. The cultures were incubated at 37°C with shaking until an OD600 of 0.8–1.0 was reached. IPTG was then added to a 1 mM final concentration, and the OD600 was measured at different times. For the viability tests, the colony-forming units ml−1 were determined just before and 45 min after the addition of IPTG.
We acknowledge D. Desnoeck and C. Kerbourch for excellent technical assistance and Marie Monteforte for help with the figures. We also thank Sophie Bleves and Maite Iriarte for critical reading of the manuscript. We are grateful to Béatrice Bernier and Pierre Wattiau for the construction of plasmids pBB5, pBB9 and pPW71, and to J.-F. Rémy and P. Gildemeester for participating in the formation of some of the genetic constructs and for protein analysis. C.N. is a research assistant funded by the Belgian ‘Fonds National de la Recherche Scientifique’. This work was supported by the Belgian ‘Fonds National de la Recherche Scientifique Médicale’ (Convention 3.4595.97), the ‘Direction Générale de la Recherche Scientifique–Communauté Française de Belgique’ (Action de Recherche Concertée 94/99-172) and the ‘Interuniversity Poles of Attraction Program — Belgian State, Prime Minister's Office, Federal Office for Scientific, Technical and Cultural Affairs’ (PAI 4/03).