Yersinia enterocolitica Type III Secretion

Yersinia enterocolitica inject toxic proteins (effector Yops) into the cytosol of eukaryotic cells by a mechanism requiring the type III machinery. Previous work mapped a signal sufficient for the targeting of fused reporter proteins to amino acids 1–100 of YopE. Targeting requires the binding of SycE to YopE residues 15–100 in the bacterial cytoplasm. We asked whether SycE functions only to stabilize YopE in the bacterial cytoplasm, or whether the secretion chaperone itself contributes to substrate recognition by the type III machinery. Fusions of glutathioneS-transferase to either the N or C terminus of SycE resulted in hybrid proteins that bound YopE but prevented targeting of the export substrate into HeLa cells. As compared with wild-type SycE, glutathione S-transferase-SycE bound and stabilized YopE in the bacterial cytoplasm but failed to release the polypeptide for export by the type III machinery. Thus, it appears that SycE functions to deliver YopE to the type III secretion machinery. A model is presented that accounts for substrate recognition of effector Yops, a group of proteins that do not share amino acid sequence or physical similarities.

Yersinia enterocolitica inject toxic proteins (effector Yops) into the cytosol of eukaryotic cells by a mechanism requiring the type III machinery. Previous work mapped a signal sufficient for the targeting of fused reporter proteins to amino acids 1-100 of YopE. Targeting requires the binding of SycE to YopE residues 15-100 in the bacterial cytoplasm. We asked whether SycE functions only to stabilize YopE in the bacterial cytoplasm, or whether the secretion chaperone itself contributes to substrate recognition by the type III machinery. Fusions of glutathione S-transferase to either the N or C terminus of SycE resulted in hybrid proteins that bound YopE but prevented targeting of the export substrate into HeLa cells. As compared with wild-type SycE, glutathione S-transferase-SycE bound and stabilized YopE in the bacterial cytoplasm but failed to release the polypeptide for export by the type III machinery. Thus, it appears that SycE functions to deliver YopE to the type III secretion machinery. A model is presented that accounts for substrate recognition of effector Yops, a group of proteins that do not share amino acid sequence or physical similarities.
Yersiniae infect human and animal hosts to cause a variety of intestinal and septicemic diseases. Three pathogenic species, Y. enterocolitica, Y. pseudotuberculosis, and Y. pestis, share a tropism for lymphoid tissues but differ in their mode of host entry and the sites of the resulting pathological lesions (1,2). To evade phagocytic killing by immune cells, Yersiniae bind to the surface of macrophages and inject several proteins (effector Yops) into the eukaryotic cytosol via a mechanism requiring the type III secretion machinery (3)(4)(5). Genes specifying the type III machinery and Yop proteins are located on a 70kilobase pair virulence plasmid (6). In the absence of calcium, Yersiniae carrying this plasmid are triggered to secrete massive amounts of all Yops into the extracellular medium (7). Some 14 secreted Yops have been identified, and their specific role during animal infection is currently being investigated. Several different assays have been established that permit localization of Yops during infection of either HeLa cells or cultured macrophages (5,8,9). YopE, YopH, YopM, YopN, YopO (YpkA), YopP (YopJ), and YopT are injected into the eukaryotic cytosol (5, 10 -14), whereas YopB, YopD, and YopR are found in the extracellular milieu (9,15). Other type III secretion substrates such as YopQ (YopK) and LcrV remain associated with the bacteria during infection, and these proteins have been implicated in regulating the injection of effector Yops (16,17). Yop proteins display neither amino acid sequence homology nor physical similarity (2).
Yersinia mutants lacking any one of the structural components of the type III machinery are defective for the export of all Yops both during tissue culture infection and when induced by low calcium (18 -20). In contrast, Yersinia mutants that lack a small cytoplasmic protein, SycE, can not inject YopE into eukaryotic cells but retain the ability to secrete YopE into the medium of low calcium-induced cultures (9,21). Mapping of the signals for YopE secretion revealed that this polypeptide can be exported in two ways (21). One pathway recognizes a signal located in YopE mRNA specifying the first 15 amino acids of the polypeptide (22)(23)(24). The second signal is provided by amino acid residues 15-100 and is absolutely dependent on the binding of SycE to YopE polypeptide (21). Although each signal is sufficient for the secretion of fused reporter proteins, neither one is absolutely necessary for the low calcium-induced secretion of YopE (21). During tissue culture infection, targeting of YopE into the cytosol of HeLa cells is absolutely dependent on the presence of SycE (9). The mRNA encoded signal of YopE does not appear to result in secretion under these conditions because polypeptides generated by fusions of this signal to a reporter are located in the bacterial cytoplasm (9).
The role of SycE in targeting YopE into the eukaryotic cytosol has not yet been established. Because YopE is not exported by the type III machinery in the absence of SycE, we wondered whether this chaperone contributes to substrate recognition. SycE itself remains in the bacterial cytoplasm and, after delivery of YopE to the type III machinery, must be released from the bound substrate (9,25,26). Thus, identification of SycE mutants that abolish targeting of YopE but do not interfere with polypeptide binding would provide evidence that the chaperone contributes to substrate recognition. Here we demonstrate the existence of such mutants and suggest a molecular explanation for their defect in substrate recognition.

EXPERIMENTAL PROCEDURES
Bacterial Strains and Plasmids-Y. enterocolitica strains W22703 (wild-type), LC2 (sycE1), and KUM1 (lcrD1) have been described elsewhere (21). Wild-type SycE, GST 1 -SycE, and SycE-GST were cloned on a low copy number vector and transformed by electroporation into * This work was supported by United States Public Health Service Grant AI 42797 (National Institutes of Health-National Institute of Allergy and Infectious Diseases Branch). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  amplifying wild-type SycE with abutted NdeI and BamHI sites using SycE-B (5Ј-AAGGATCCTCAACTAAATGACCGTGGT-3Ј) and SycE-N (5Ј-AACATATGTATTCATTTGAACAAGATATCA-3Ј). The DNA fragment was cut with NdeI-BamHI and cloned into pET9a to yield pHTT1. One liter of E. coli BL21 (DE3), pHTT1 was grown to mid-log phase and induced with 1 mM IPTG for 2 h at 37°C. Cells were harvested by centrifugation (6,000 ϫ g for 15 min), suspended in F buffer (50 mM Tris-HCl, 20% sucrose, and 1 mM dithiothreitol, pH 7.5), and lysed by one passage through a French press at 14,000 p.s.i. Extracts were centrifuged twice at 33,000 ϫ g for 15 min, and proteins in 20 ml of supernatant were precipitated with 45% ammonium sulfate by incubation at 4°C for 2 h and centrifugation at 33,000 ϫ g for 15 min. The sediment was suspended in 4 ml of 50 mM Tris-HCl, pH 7.5, and subjected to chromatography on a Sephacryl S-200HR (Amersham Pharmacia Biotech) column. The purification of GST-SycE protein has been described elsewhere (21).
Binding of YopE to GST-SycE Hybrid Proteins-Overnight cultures of Yersinia LC2 (sycE1) carrying pGST-SycE, pSycE-GST, pGST-SycE ⌬C , or pGST-SycE ⌬N grown in TSB supplemented with 20 g/ml chloramphenicol were diluted into fresh media (10 ml of culture:500 ml of TSB). Bacteria were grown for 2 h at 26°C and induced for 3 h at 37°C. Cells were harvested and suspended in 10 ml of F buffer and broken by a single passage through a French pressure cell at 14,000 p.s.i. Unbroken cells, debris, and membranes were removed by centrifugation at 6,000 ϫ g for 15 min. Supernatants were subjected to affinity chromatography on glutathione-Sepharose (Amersham Pharmacia Biotech) pre-equilibrated with F buffer. The column was washed with 30 column volumes of wash buffer (50 mM Tris-HCl, 150 mM NaCl, and 15% glycerol, pH 7.5), and proteins were eluted with 4 ml of the same buffer containing 10 mM glutathione. Proteins in the eluate were precipitated with 10% trichloroacetic acid using 50 l of 2% deoxycholate/ml as a carrier. Precipitates were washed with acetone and air-dried, and proteins were suspended in 100 l of sample buffer containing 3 M urea.
Cell Fractionations-Overnight cultures of Yersinia were diluted 1:50 into 250 ml of fresh TSB media, grown for 2 h at 26°C, and induced at 37°C for 3 h. Cells were harvested at 6,000 ϫ g for 15 min and suspended in 10 ml of lysis buffer (20 mM HEPES, 100 mM KOAc, 2 mM MgOAc, and 1 mM dithiothreitol, pH 7.5). Bacteria were broken in a French pressure cell at 14,000 p.s.i., and intact cells were removed by centrifugation at 6,000 ϫ g for 10 min. Two 3-ml aliquots of crude bacterial extract were centrifuged at 180,000 ϫ g for 30 min. The supernatant (ucs1) was separated from the membrane sediment, which was suspended in 3 ml of lysis buffer (ucp1). The sediment of the other centrifuged aliquot was salt extracted with 3 ml of 1 M KOAc and centrifuged at 180,000 ϫ g for 30 min, and the supernatant (ucs2) was separated from the sediment (ucp2). At each step of the fractionation, 500-l aliquots were withdrawn and precipitated with 500 l of 10% trichloroacetic acid and washed with acetone. The precipitate was solubilized by adding 50 l of buffer B and 50 l of sample buffer before boiling. Samples were separated on 15% SDS-PAGE and immunoblotted with anti-YopE and anti-SycE, and the developed signals were quantified by densitometry scanning.
K d Determination-500 ng of purified, denatured YopE 6His in 10 l of eluate (6 M GnHCl, 0.1 M NaH 2 PO 4 , and 0.01 M Tris-HCl, pH 8.0) were diluted into 1 ml of buffer (50 mM Tris, pH 7.0, 100 mM KOAc, 5 mM MgOAc, 0.25% Tween 20, and 50 g/ml bovine serum albumin) containing 50% slurry Ni-NTA resin and incubated at room temperature for 20 min. Sepharose beads were washed three times and suspended at a concentration of 1 pmol of YopE 6His /l of beads. Increasing amounts of SycE (0 -150 pmol) were added to a 200-l suspension of YopE 6His :Ni-NTA-Sepharose (100 pmol of YopE). Purified GST-SycE was first bound to glutathione-Sepharose (Amersham Pharmacia Biotech) at a concentration of 1 pmol of protein/l of beads. Increasing concentrations of YopE (0 -120 pmol) were then added to a 200-l suspension of GST-SycE:glutathione-Sepharose (100 pmol of GST-SycE). Samples were incubated for 20 min and centrifuged at 15,000 ϫ g for 5 min. An aliquot of the supernatant was removed, mixed with sample buffer, separated on 15% SDS-PAGE, and immunoblotted with anti-SycE and anti-YopE (1:2000 dilution). Immune complexes were detected with 12 Ci of 125 I-labeled protein A/blot, and signals were quantified by PhosphorImager. Each SDS-PAGE blot was calibrated with a dilution series of known concentrations of YopE or SycE. elements required for YopE binding and/or targeting, we constructed both N-and C-terminal fusions of GST with SycE and asked whether the hybrid proteins were functional in delivering polypeptide to the type III machinery. This was tested by infecting HeLa cells with either wild-type Y. enterocolitica strain W22703 or strain LC2, a mutant carrying a knockout mutation of the sycE gene (sycE1) (21). Infected HeLa cells were fractionated by first decanting and then centrifuging the media. Nonadherent bacteria were sedimented into the pellet, whereas the extracellular medium remained in the supernatant. HeLa cells with adherent bacteria were extracted with digitonin, a detergent known to disrupt the cholesterol-containing plasma membrane of HeLa cells but not the bacterial envelope (9). The digitonin extract was centrifuged to sediment bacteria as well as insoluble debris, whereas the soluble contents of the eukaryotic cytosol were separated with the supernatant. Proteins in all fractions were precipitated with chloroform/methanol and analyzed by SDS-PAGE and immunoblotting.

GST-SycE and SycE-GST
When infected with wild-type Yersinia strain W22703, YopE and YopH were located in the supernatant of digitonin extracts, indicating that 35% (YopE) and 48% (YopH) of these proteins had been injected into the cytosol of HeLa cells (Fig. 1). SycE remained inside of bacterial cells and was found only in the sediment of digitonin extracts. The sycE Ϫ strain LC2 injected 54% of YopH into the eukaryotic cytosol. However, YopE was found only in the bacterial sediment of digitonin extracts, indicating that this polypeptide had not been targeted into HeLa cells. Complementation of the sycE Ϫ mutation with plasmid-encoded sycE restored YopE targeting (29%). Neither plasmid-encoded gst-sycE nor sycE-gst complemented the YopE targeting defect of strain LC2. The targeting defect of strains carrying the gst-sycE and sycE-gst alleles was limited to YopE because the mutant Yersiniae injected YopH in a manner sim-ilar to wild-type bacteria.
GST-SycE and SycE-GST Bind YopE and Do Not Affect Low Calcium-induced Secretion of YopE-One explanation for the targeting defect of GST-SycE and SycE-GST could be that the hybrids do not bind YopE and thus cannot deliver substrate to the type III machinery. To examine this possibility, we asked whether YopE could co-purify with GST-SycE. Bacterial extract supernatants representing soluble cytoplasmic contents of Yersiniae were subjected to affinity chromatography on glutathione resin. Eluted samples were analyzed by immunoblotting for the presence of hybrid SycE proteins and bound YopE. YopE co-eluted with both GST-SycE and SycE-GST, indicating that the hybrid proteins bound to YopE ( Fig. 2A). Syc proteins contain a C-terminal domain with similarity to leucine zipper (seven residue periodicity) sequences that are known to impose coiled-coil dimerization onto ␣-helices (26). It is conceivable that this domain might be involved in either binding to YopE or recognition of SycE:YopE by the type III machinery. Truncation of 45 C-terminal residues of SycE, including part of the conserved leucine zipper domain, abolished the ability of the mutant GST-SycE ⌬C to bind YopE polypeptide ( Fig. 2A). Deletion of 30 N-terminal residues of SycE reduced the stability of GST-SycE ⌬N , and only small amounts of this fusion protein were eluted from the glutathione resin. Even on overexposed films, no binding of GST-SycE ⌬N to YopE could be detected. Thus, truncations of either the N or C terminus of SycE abolished binding to YopE. To test whether the GST-SycE and SycE-GST hybrids interfered with the function of the mRNA secretion signal, we measured YopE secretion in low calciuminduced cultures. The sycE Ϫ mutant strain LC2 secreted 55% of all YopE into the culture medium (Fig. 2B). When complemented by plasmid-encoded wild-type SycE, YopE expression was increased; however, steady-state secretion remained at the same level (55%). Plasmid-encoded GST-SycE and SycE-GST did not affect low calcium-induced secretion of YopE, indicating that the hybrid proteins did not interfere with the function of FIG. 1. Type III targeting of YopE polypeptide into the cytoplasm of HeLa cells. HeLa cells were infected with wild-type Y. enterocolitica W22703 or strain LC2 (sycE1) carrying either no plasmid, pSycE, pGST-SycE, or pSycE-GST. After incubation for 3 h at 37°C, the tissue culture medium (M) was decanted and centrifuged to separate secreted proteins from those present within nonadherent bacteria. HeLa cells as well as adherent Yersinia were extracted with digitonin (D), a detergent that solubilizes the eukaryotic plasma membrane but not the bacterial envelope. Extracts were centrifuged to separate proteins solubilized from the HeLa cytoplasm from those that sediment with the bacteria. Proteins in each fraction were precipitated with chloroform/methanol and analyzed by SDS-PAGE and immunoblotting with antibodies directed against YopE, SycE, or YopH. Targeting was measured as the percentage amount of protein solubilized by digitonin extraction divided by the total amount of protein. As a control for solubilization of all membranes, HeLa cells and adherent bacteria were extracted with SDS (S). P, pellet. the YopE mRNA signal.
Binding of SycE and GST-SycE to YopE-We used purified protein components to measure the dissociation constants of wild-type SycE and GST-SycE for YopE. YopE carrying a Cterminal histidine tag was expressed in E. coli, purified under denaturing conditions by affinity chromatography on Ni-NTA resin, and eluted with 6 M GnHCl (Fig. 3A). The six histidyl tag is located outside of the SycE binding site (residues 15-100 of YopE) and affects neither chaperone binding nor targeting of YopE 6His (data not shown). Wild-type SycE was expressed in E. coli and purified from cell extracts by a combination of ammonium sulfate precipitation and gel filtration chromatography (Fig. 3). Purified, denatured YopE 6His was insoluble when diluted into aqueous buffer. However, when diluted into buffer containing 0.25% Triton X-100 and bovine serum albumin, YopE 6His did not sediment after centrifugation at 160,000 ϫ g for 30 min, suggesting that the detergent prevented aggregation of denatured (unfolded) YopE 6His . Soluble YopE 6His was added to Ni-NTA resin and incubated with increasing amounts of SycE. After sedimenting Ni-NTA-Sepharose/SycE:YopE 6His , the concentration of unbound SycE in the supernatant was determined. Fig. 4A shows a Scatchard plot of the collected data (30). Wild-type SycE bound to YopE 6His with a K d of 3.3 ϫ 10 Ϫ10 M (S.D., 0.9 ϫ 10 Ϫ10 M). The intercept of the Scatchard plot was at 2.3 molecules of SycE. This measurement is in good agreement with previous observations that purified SycE is retained on size exclusion resin with the mass of a homodimeric complex (25). Together, these data suggest that SycE:YopE 6His complexes consist of one molecule of YopE 6His and two molecules of SycE (SycE 2 :YopE).
GST-SycE was purified by affinity chromatography on glutathione-Sepharose. Purified GST-SycE bound to Ni-NTA resin, and our protocol was modified to permit affinity measurements via glutathione-Sepharose. Increasing amounts of YopE 6His were added to GST-SycE bound to glutathione-Sepharose. The GST-SycE:YopE 6His complexes were sedimented by centrifugation, and unbound YopE 6His in the supernatant was measured by immunoblotting (Fig. 4B). The dissociation constant for GST-SycE binding to YopE 6His was measured to be 3.6 ϫ 10 Ϫ10 M (S.D., 1.1 ϫ 10 Ϫ10 M), similar to that observed for wild-type SycE. The intercept of the Scatchard plot was at 0.25 YopE 6His molecules, suggesting that four GST-SycE molecules were bound per YopE 6His polypeptide (GST-SycE 4 :YopE 6His ). GST is known to form a homodimer (31), suggesting that GST-SycE binds not only YopE but also other GST-SycE molecules in the GST-SycE 4 :YopE complex.
Secretion and Stability of YopE Bound to SycE, GST-SycE, or SycE-GST-Previous work showed that YopE is rapidly degraded in the absence of SycE and that GST-SycE can complement this defect of sycE Ϫ mutant Y. pseudotuberculosis (32). We wished to test whether SycE-GST could also stabilize YopE polypeptide. Yersiniae were pulse-labeled with [ 35 S]methionine, and culture aliquots were analyzed at timed intervals by

FIG. 4. Affinity of SycE and GST-SycE for YopE secretion substrate. A, purified SycE was incubated with purified YopE 6His bound to
Ni-NTA. YopE 6His /SycE and Ni-NTA complexes were sedimented by centrifugation, and the binding of wild-type SycE was determined as the amount that sedimented with YopE 6His :Ni-NTA. The collected data are shown as a Scatchard plot. SycE bound to YopE 6His with a K d of 3.3 ϫ 10 Ϫ10 M. Each YopE 6His molecule appears to bind two SycE polypeptides. B, for the binding of GST-SycE to YopE 6His , purified polypeptide was added to GST-SycE bound to glutathione-Sepharose, and the bound complexes were sedimented by centrifugation. GST-SycE bound to YopE 6His with a K d of 3.6 ϫ 10 Ϫ10 M. In contrast to wild-type SycE, about four molecules of GST-SycE appeared to be bound to each YopE 6His polypeptide.
precipitating proteins with ice-cold trichloroacetic acid. As reported previously for wild-type Yersiniae (21), about 50% of YopE was degraded within 10 min, and the remaining 50% of YopE was stable over a 60-min period (Fig. 5). Proteolysis of YopE in strain LC2 (sycE Ϫ ) occurred much more rapidly (halflife, 3.8 min), and 6% of all YopE remained stable over a 60-min period. When complemented by wild-type SycE, the half-life of YopE was increased to 17.5 min. Both GST-SycE and SycE-GST increased the half-life of YopE to 5 and 6.8 min, respectively; 15% (GST-SycE) and 25% of YopE (SycE-GST) were stable over a 60-min period. Thus, our results corroborate previous findings and show that GST-SycE and SycE-GST increased the stability of YopE, albeit not to the same level as wild-type SycE.
To measure the secretion of YopE bound to SycE or GST-SycE, we incubated pulse-labeled Yersiniae for 10 min to allow type III secretion of polypeptide via the mRNA pathway. The bacteria were collected by centrifugation, suspended in fresh medium, and incubated over a period of 20 min. Post-translational export of YopE was measured on a PhosphorImager after SDS-PAGE separation of pulse-labeled YopE that had been immunoprecipitated from culture supernatants. Wild-type Yersiniae secreted 18% of all pulse-labeled YopE, whereas sycE Ϫ mutants did not secrete YopE. This secretion defect could not be rescued by expressing GST-SycE in sycE Ϫ mutants cells, indicating that although the mutant chaperone binds and stabilizes the polypeptide, GST-SycE can not release YopE for secretion by the type III secretory pathway.
YopE Requires SycE or GST-SycE for Solubility within the Cytoplasm of Yersiniae-We wished to determine the fate of SycE 2 :YopE complexes in bacterial cells and fractionated Yersiniae grown under low calcium conditions (Fig. 6). Bacteria were collected by centrifugation and lysed in a French pressure cell. Extracts were ultracentrifuged to sediment membranes, whereas soluble cytoplasmic components remained in the supernatant. In wild-type Yersiniae, 95% of all YopE was soluble in the bacterial cytoplasm. The subcellular distribution of YopE changed dramatically when analyzed for sycE Ϫ cells: all YopE (100%) sedimented with the membranes, suggesting that these polypeptides were insoluble in the bacterial cytoplasm. Expression of GST-SycE and SycE-GST in sycE Ϫ cells restored YopE solubility to 75% and 47%, respectively. Thus, the binding of hybrid GST-SycE and SycE-GST to YopE in the Yersinia cytoplasm solubilized the polypeptide.
Membrane fractions were extracted with 1 M KOAc to release proteins peripherally associated with the lipid bilayer (Fig. 6). In wild-type Yersiniae, most YopE species that sedimented with the membranes could not be extracted with 1 M KOAc (80%). Membrane-bound YopE species might represent polypeptides in the process of translocation by the type III machinery in a manner that cannot be extracted with salt. If so, mutants of the type III pathway that cannot export Yop proteins should not contain YopE species that are resistant to salt extraction. This was tested, and Yersiniae carrying a null allele of lcrD (strain KUM1) (21), a cytoplasmic membrane protein absolutely required for all type III export (33), were fractionated as described above. Almost all YopE (94%) was soluble after French press lysis. Of the 6% of polypeptide that sedimented with the membranes, 83% could be extracted with 1 M KOAc. Thus, in the lcrD1 strain KUM1, 99% of YopE is soluble, suggesting that the salt-resistant, membrane-associated YopE species represent an export intermediate of the type III pathway. Export intermediates were observed in LC2 cells (sycE Ϫ ) (36% of all YopE) because the mRNA signal initiates YopE into the export pathway, even in the absence of SycE. Similarly, GST-SycE-and SycE-GST-expressing Yersiniae also harbored significant amounts of membrane-associated, salt-resistant YopE species.
GST-SycE Competes with SycE for YopE Secretion Substrate-If GST-SycE cannot release polypeptide for membrane translocation, GST-SycE should compete with SycE binding to YopE, thereby preventing type III secretion in wild-type Yersiniae (Fig. 7). To test this prediction, we used plasmid pDA141 encoding a yopE-npt translational fusion with a frameshift mutation of codons 2-15 that abolishes the function of the secretion signal located in yopE mRNA (YopE ϩ1 -NPT). When expressed in wild-type cells, YopE ϩ1 -NPT protein was secreted into the extracellular milieu (Fig. 7A). In contrast, sycE Ϫ strain LC2 failed to export this polypeptide, indicating that secretion of the fusion protein is absolutely dependent on binding to SycE polypeptide. Expression of GST-SycE did not complement the secretion defect of the sycE Ϫ strain for YopE ϩ1 -NPT. When overexpressed in wild-type Yersiniae, GST-SycE interfered with the secretion of YopE ϩ1 -NPT, suggesting that GST-SycE competes with SycE for secretion substrate (Fig. 7A). If so, increased expression of GST-SycE via the IPTG-inducible tac promoter should lead to an increased ratio of GST-SycE/SycE concentration and to a decreased secretion of YopE ϩ1 -NPT by wild-type Yersinia. This was tested, and Yersiniae grown in the absence of IPTG secreted 47% YopE ϩ1 -NPT, whereas the addition of 0.1, 0.5, and 1 mM IPTG to the growth medium caused an increase in the ratio of GST-SycE/SycE and a stepwise reduction in the percentage amount of secreted fusion protein (Fig. 7B). As a control for the function of the yop mRNA signaling pathway, both wild-type and sycE Ϫ mutant Yersiniae exported wild-type YopE. Thus, when bound to GST-SycE, YopE is irreversibly retained from the type III secretory pathway.
If GST-SycE expression interferes with the SycE-dependent export of YopE ϩ1 -NPT in low calcium-induced cultures, it should also interfere with the targeting of YopE into the eukaryotic cytosol during Yersinia infection of HeLa cells. To examine this possibility, HeLa cells were infected with wildtype Yersiniae expressing GST-SycE from the IPTG-inducible tac promoter in either the absence or presence of increasing amounts of IPTG (Fig. 8). Increased expression of GST-SycE led to decreased targeting of YopE in the presence of wild-type SycE, indicating that GST-SycE competed with SycE for binding to YopE. Overexpression of GST-SycE did not affect the targeting of YopH into HeLa cells. Because the targeting of YopH is thought to be dependent on its binding to the SycH chaperone, this observation suggests that GST-SycE 4 :YopE complexes do not compete with YopH:SycH for recognition by the type III machinery. DISCUSSION To investigate the targeting pathway of YopE, we examined the role of SycE in binding to YopE polypeptide. Initial experiments aimed at the purification of YopE species that had been secreted into the extracellular medium of low calcium-induced Yersinia cultures (data not shown). We were surprised to find that most, if not all, extracellular YopE was insoluble after FIG. 7. Expression of GST-SycE interferes with the SycE-dependent type III secretion of YopE. Plasmid pDA141 encodes YopE ϩ1 -NPT, a fusion protein between full-length YopE and NPT. The first 15 codons (45 nucleotides) of YopE carry a ϩ1 frameshift mutation (insertion of A immediately after the AUG start codon) of the mRNA signal that is suppressed at codon 15 via the removal of a nucleotide. The defective mRNA signal is tethered to the nucleotide sequence (codons 16 -220) encoding the SycE binding site of YopE. The gst-sycE allele is located on a p15A replicon plasmid and expressed from the IPTG-inducible tac promoter. Bacteria were grown at 37°C in TSB supplemented with 0.01 M EGTA and 1 mM IPTG. Cells were harvested by centrifugation (P) and separated from the culture supernatant (S). Proteins in both fractions were precipitated with trichloroacetic acid and subjected to immunoblotting with ␣NPT, ␣YopE, ␣GST, and ␣SycE. A, expression of GST-SycE did not complement the secretion defect of sycE1 cells, and expression of GST-SycE in wild-type Yersiniae interfered with the SycE-dependent secretion of YopE ϩ1 -NPT. Expression of GST-SycE also reduced the amount of secreted YopE; however, secretion was still observed because YopE can be exported by the mRNA secretion signal located within codons 1-15. B, GST-SycE expression in Y. enterocolitica W22703, pDA141, pLC202 was increased by adding either 0, 0.1, 0.5, or 1 mM IPTG. Increasing the ratio of GST-SycE/SycE caused a corresponding reduction in the amount of YopE ϩ1 -NPT secretion.
FIG. 8. Expression of GST-SycE interferes with the SycE-dependent type III targeting of YopE into HeLa cells. Wild-type Y. enterocolitica W22703 expressing GST-SycE from the IPTG-inducible tac promoter was used to infect HeLa cells in the presence or absence of IPTG. Type III targeting was measured as described in the legend to Fig. 1 and detected by the immunoblotting of collected samples. Overexpression of GST-SycE interfered with the SycE-dependent type III targeting of YopE. In contrast, GST-SycE expression did not reduce the injection of YopH into the cytosol of eukaryotic cells. Results similar to those for YopH were obtained when samples were immunoblotted for YopM (data not shown). centrifugation at 100,000 ϫ g. This observation is consistent with the notion that some secreted Yops aggregate in the extracellular milieu to form macroscopically visible precipitates (7). Thus, it appears that YopE cannot fold properly unless the polypeptide is in the appropriate environment, i.e. the eukaryotic cytoplasm. To examine this prediction, we report here the purification of denatured YopE which, when diluted into aqueous buffer, did not acquire solubility even after prolonged incubation. Binding of purified SycE to denatured YopE resulted in solubility, even when the complex was diluted in the absence of detergent, suggesting that SycE binding is required for intrabacterial solubility. This hypothesis was tested in vivo; indeed, soluble YopE could not be detected in the cytoplasm of Yersiniae lacking the SycE chaperone. SycE remains in the bacterial cytoplasm, and once YopE is injected into eukaryotic cells, this polypeptide likely depends on host cell factors for folding and function.
SycE binding to YopE may not only provide solubility in the bacterial cytoplasm but may also allow recognition of the SycE 2 :YopE complex by the type III machine. We sought to separate polypeptide binding and type III delivery functions of the SycE chaperone by mutational analysis. Hybrid GST-SycE and SycE-GST fusion proteins were designed with the rationale to facilitate binding measurements with YopE. We were surprised to find that fusion of the GST domain to either the N or C terminus of SycE abolished type III targeting of YopE without affecting binding to the polypeptide. These results suggest that SycE does interact with the type III machinery during delivery of YopE and that fusion of the bulky GST domain may prevent this interaction. An alternative explanation is that type III machines recognize Yops while they are bound to Syc chaperones. However, this appears somewhat unlikely, because Yop proteins do not display sequence or physical similarity (2).
If type III machines recognize export substrate as a property of the SycE 2 :YopE complex, is there an element that is shared between SycE, SycH, SycT, SycN, and YscB (i.e. the cognate chaperones of targeted YopE, YopH, YopT, and YopN) (10, 25, 34 -36)? Wattiau et al. (26) noted the presence of a conserved C-terminal leucine zipper of Syc proteins. Such an element is found in many polypeptides that assume a coiled-coil conformation, in which the hydrophobic (leucine) side chains lie on one side of an ␣-helix to promote hydrophobic interactions with another helix of either the same (homodimer) or another polypeptide (heterodimer) (37). Dissociation of C-terminal helices within SycE 2 :YopE by another binding partner, say the type III machine, might separate the SycE 2 :YopE complex. This is supported by our observation that SycE ⌬C lacking the C-terminal leucine zipper cannot bind YopE. Although highly speculative, our hypothesis provides experimental opportunity, for example, a search for factors that displace SycE from YopE polypeptide.
This has been described for another secretion chaperone cycle that involves the SecB protein (38). SecB delivers precursor proteins to the Sec machine, which catalyzes the translocation of signal peptide-bearing polypeptides across the cytoplasmic membrane of E. coli and other bacteria (39). SecB binds unfolded polypeptide as a tetramer (40 -42). Initiation of polypeptide precursors into the secretory pathway requires interaction between SecB-substrate complexes and the Sec machinery. This is accomplished by the membrane-bound form of SecA (43), in that the 22 C-terminal amino acids of SecA bind to SecB (44). SecB-precursor complexes display a higher affinity for SecA than for SecB alone. ATP binding of SecA is thought to promote both translocation of the released precursor substrate and displacement of SecB, presumably via conformational changes of SecA that alter the SecB binding site (44). Released SecB is then once again available to bind another polypeptide precursor. Perhaps delivery of YopE to the type III machinery occurs by a similar cycle in which the SycE 2 :YopE complex is dissociated by a component of this machine, thereby initiating polypeptide substrate into the targeting pathway and releasing SycE chaperone.