INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The genus Yersinia contains three species that are pathogenic for humans: Y. pestis, the causative agent of bubonic plague, and Y. pseudotuberculosis and Y. enterocolitica, which cause gastroenteritis and mesenteric lymphadenitis. A requirement for the pathogenesis of yersiniae is the ability to both express and secrete Yersinia outer proteins (Yops) and V antigen (also called LcrV) (10, 11, 50). The genes that encode Yops and V antigen, the proteins which regulate their expression, and the proteins of the Yop-specific secretion apparatus (Ysc) are carried on a ca. 70-kb plasmid found in all virulent isolates of pathogenic yersiniae (referred to as pCD1 in Y. pestis KIM) (6, 19). Yops and V antigen have antihost functions that subvert innate defenses (e.g., phagocytosis) and cellular signaling and cytokine expression necessary for the mobilization of an effective immune response (14). They are crucial for the survival and multiplication of yersiniae in lymphoid tissues (14, 50, 74). Yops exert their antihost effects within eukaryotic cells. When yersiniae contact a eukaryotic cell, at least four Yops (YopE, YopH, YopM, and YpkA) are vectorially targeted into the cell at the point of contact (9, 28, 52, 59, 71) by a process requiring YopB, YopD, and YopK (29, 31, 59, 71). In Y. pestis, the secreted Yops and V antigen are subject to degradation by Pla, a plasminogen activator protease, present on the bacterial surface and encoded by the Y. pestis-specific plasmid pPCP1 (69, 70). Despite its ability to rapidly degrade Yops on the cell surface (60, 61, 69), Pla has little effect on the expression, secretion, or vectorial targeting of Yops (46, 55, 56, 66-68, 72).

High-level expression of Yops and V antigen, concomitant with secretion, is called the low-Ca²⁺ response (LCR). In vitro this occurs at 37°C in the absence of Ca²⁺ and is accompanied by growth cessation (termed restriction). Ca²⁺ is required in millimolar concentrations for growth in certain media at temperatures above 34°C (80). The LCR is believed to be manifest in vivo upon yersinial contact with a eukaryotic cell (14). It is believed that growth restriction does not occur in vivo (23) but that Yops and V antigen expression and secretion are regulated and that the absence of Ca²⁺ in vitro mimics cell contact (14, 73). The genes that are coordinately downregulated (73) by Ca²⁺ are referred to as the LCRS (low-Ca²⁺ response stimulon) (72).

A cluster of genes occupying ca. 25 kb on the Yersinia LCR plasmid is responsible for the regulation of expression, secretion, and postsecretion targeting of Yops (13, 50). Most of these (at least 22 gene products) encode the type III, or contact-dependent Yop secretion mechanism, Ysc (13, 14, 43). This is one of the best characterized of a relatively new class of secretion mechanisms present in a wide variety of gram-negative pathogens of mammals and plants (37). The Ysc mechanism is environmentally modulated in its activity at 37°C, and this modulation is indirectly responsible for the Ca²⁺ regulation of yop transcription seen in vitro. A current model for this effect (14) holds that under conditions that do not activate the Ysc (presence of Ca²⁺ and absence of contact with a eukaryotic cell), proteins necessary for negative regulation (LcrQ [53, 58] and YopD [77]) are retained in the bacterial cytoplasm, where they permit only a low expression of Yops and V antigen. However, this small pool of virulence proteins is available for immediate targeting upon imposition of Ysc-activating conditions (5). Upon cell contact, the Ysc is activated locally between the eukaryotic and bacterial cells and the effector Yops are vectorially targeted into the eukaryotic cell by the postsecretion delivery mechanism containing YopB, YopD, and YopK. The negative regulator LcrQ also is secreted out of the bacterial cell (53, 58), and this permits upregulation of yop transcription (14, 53). In vitro, the absence of Ca²⁺ is thought to activate all secretion channels and cause strong yop induction, through the massive secretion of YopD and LcrQ. Upregulation of translation of at least some Yops may also occur upon the activation of the Ysc and be coordinated with targeting of the newly expressed Yops to the Ysc (5). This also would ensure a rapidly available supply of additional Yops for targeting and a greater degree of amplification of Yops expression than that expected from transcriptional induction alone (5).

The Ysc secretes Yops, without processing, through both bacterial membranes. Although some Yops can be obtained in small amounts in periplasmic fractions (46), there is no evidence for a true, obligatory periplasmic phase of Yop secretion. By analogy to the type III secretion system of centisome 63 in Salmonella typhimurium (35), the Ysc probably spans both membranes as a supramolecular complex. Three operons encode the secretion mechanism proper: lcrDR (55), yscA to yscM (27, 43), and yscN to yscU (3, 7, 20, 78); VirG appears to promote insertion of the secretin YscC into the outer membrane (1, 34). Additional genes, including those that encode LcrE (also called YopN), LcrG, LcrV, and TyeA, modulate the activity of the Ysc in response to environmental inputs such as Ca²⁺ and cell contact (22, 32, 46, 66, 67, 79). Finally, members of the lcrGVH-yopB-yopD operon function in the postsecretion targeting of Yops (14).

Homologies to components of the flagellar basal body have been noted for LcrD (54, 55) and all except YscP of the products of the eight-member yscN to yscU operon (3, 7, 20, 78). All of these except YscO, YscP, and YscQ have been shown or assumed on the basis of predicted sequences to be inner membrane proteins or inner membrane-associated proteins (in the case of the putative energizer protein YscN). The only other known inner membrane component is YscD (not a flagellar component homolog) (43, 56). It has been speculated that these homologies could reflect a supramolecular analogy between the flagellar basal structure and the inner membrane and membrane-spanning regions of the type III mechanism, and this is supported by the recently demonstrated shape of the Salmonella type III mechanism (35).

Other than VirG, which probably is located in the outer membrane, the only proven outer membrane component of the Ysc is the secretin YscC, which oligomerizes to form a ring-shaped structure with a central pore (34). This has been assumed to serve as a channel for Yop secretion through the outer membrane (34). Interestingly, the homolog in Salmonella, InvG, was found to be one of three main components of a needle-like projection from the putative outer membrane ring of the isolated type III complex (35).

In Y. pestis, the LcrD, YscC, YscD, YscG, and YscR products have been shown to be required for Yop secretion (20, 27, 56). In Y. pseudotuberculosis or Y. enterocolitica, YscC to YscG, YscI to YscL, YscN, YscQ, YscR, and YscU have been shown to be essential for the secretion of Yops (2, 3, 7, 43, 78). All are thought to be core (essential) components of the Ysc and to exert an indirect inductive effect on yop operon expression through their requirement for the secretion of negative regulators such as LcrQ and YopD out of the bacterial cell (14, 53, 77). Mutations in any of the essential Ysc components abolish the Ca²⁺ requirement for growth (a phenotype referred to as Ca²⁺ independence), limit the induction of LCRS operons at 37°C to the level typical for the presence of Ca²⁺, and block the secretion of LCRS proteins (20, 25, 55, 56).

yscN and yscQ through yscT have a high sequence similarity to counterparts in loci for virulence protein secretion in Salmonella pathogenicity island 1 (SPI1) at centisome 63 (spa or inv), SPI2 at centisome 30 (ssa), Shigella flexneri (spa), and enteropathogenic Escherichia coli (esc) (18, 24, 26, 30, 63). The Yersinia and both Salmonella loci are similar in the size and arrangement of the yscN to yscU homologs. However, the predicted sequences of yscO and yscP have at best weak similarity to homologs in type III systems and no similarity to any non-type III-related proteins outside the Yersinia spp. For example, YscO is 21% identical (37% similar) to SpaM/InvI of the Salmonella SPI1 and 20% identical (23% similar) to Shigella Spa13. This suggests that YscO and YscP may play a specific role in the recognition, secretion, or targeting of the Yersinia-specific Yops or V antigen and prompted us to initiate their characterization. In this study, we have focused on yscO. We constructed a mutant with essentially a complete deletion of yscO and characterized it for the expression and secretion of V antigen and Yops. The data show yscO to be essential for high-level expression and secretion of V antigen and Yops, with a direct effect on secretion. We show that under LCR-inductive conditions, YscO is weakly expressed and found in all bacterial fractions. To our knowledge, this is the first mobile core component of the Ysc.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacteria, plasmids, and growth conditions. The bacterial strains and plasmids used in this study and their relevant properties are listed in Table 1, and some are shown in Fig. 1. Y. pestis KIM5-3001 (LCR⁺) was the parent strain (wild type [wt]) used in these studies. It contains three naturally occurring Y. pestis plasmids: pCD1 (containing the genes encoding the LCR phenotype) (19, 25), pPCP1 (encoding the plasminogen activator [Pla] responsible for the degradation of Yops [69]), and pMT1 (encoding the F1 capsular protein) (57). Y. pestis KIM8-3002 (LCR⁺, Pla: wt Pla) was used in experiments in which degradation of proteins would have affected the analysis of the results. E. coli strains were typically grown in Luria-Bertani (LB) broth or on LB agar (15). Y. pestis strains were routinely grown in heart infusion broth or on tryptose blood agar base plates (Difco Laboratories, Detroit, Mich.) at 26°C. For physiological studies, Y. pestis strains were grown in TMH defined liquid medium (72) supplemented with 2.5 mM CaCl₂ as indicated. The medium was inoculated to an optical density at 620 nm of ca. 0.1 from a culture that had been growing exponentially at 26°C with shaking at 200 rpm for about seven generations. Cultures were started at 26°C and then shifted to 37°C when the optical density reached ca. 0.2. Cells and secreted proteins were harvested at 5 or 6 h after the temperature shift. All bacteria with antibiotic resistances were grown in the presence of the appropriate antibiotic(s) (100 µg/ml for ampicillin and streptomycin).

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Physical and genetic map of the region of pCD1 that encompasses yscO and yscP. (A) Coding regions for yscO and yscP and parts of yscN and yscQ carried on pYscOP.2, as well as selected restriction sites. Asterisks denote restriction sites introduced by site-directed mutagenesis. (B) Regions included in selected clones.

DNA methods. Cloning methods, including the use of restriction endonucleases, T4 DNA ligase, isolation of plasmid DNA (8, 33, 41), and electroporation (51), were as previously described. The PCR technique (44) was performed with 20 to 30 cycles of amplification; the denaturing, annealing, and extending conditions were 94, 55, and 72°C for 30 s each in a model 480 thermocycler (Perkin-Elmer Cetus, Norwalk, Conn.). Double-stranded DNA was sequenced by the method of Sanger et al. (62) with the Sequenase 2.0 sequencing kit (United States Biochemical, Cleveland, Ohio) and [-³⁵S]dATP (NEN Research Products, Boston, Mass.) or by the Macromolecular Structure Analysis Facility (University of Kentucky, Lexington).

Site-directed mutagenesis of yscO. To study an effect of yscO on the LCR, we used a variation of two site-directed in vitro mutagenesis protocols (11a, 56a) to construct an in-frame deletion in yscO that effectively created a null mutation. A single-stranded DNA (ssDNA) template of pYscOP.2 (Table 1; Fig. 1) coding strand was prepared by using the helper phage M13K07, as described previously (41). The mutagenesis reaction involved annealing a selection oligonucleotide and two mutagenic oligonucleotides to the ssDNA template. The 5'-phosphorylated selection oligonucleotide (Trans Oligo AlwNI/SpeI; Clontech) changed AlwNI to SpeI in the vector region of pYscOP.2. The two mutagenic oligonucleotides used to introduce AvrII sites into yscO at bp 881 and 1321 were 5'-ACTTTAACCCGGCCTAGGCGGCGTATC-3' and 5'-TCATTAGGCGTTCCTAGGATGCTGTAG-3', respectively. (The underlined portions represent AvrII sites; the bold sequences represent bases that were changed during mutagenesis; nucleotide numbering corresponds to the numbering of GenBank accession no. L25667). The mutagenic oligonucleotides were 5'-end phosphorylated (Promega) and annealed to ssDNA at 100°C for 3 min with a 200-fold molar excess of each of the three oligonucleotides over the ssDNA template (Clontech). The mutant strand was synthesized with T4 DNA polymerase to prevent strand displacement and allow the annealing of all three oligonucleotides in one reaction (42, 48). After ligation (Clontech), any nonmutagenized parental pYscOP.2 DNA remaining in the mixture was selectively linearized by digestion with AlwNI to decrease its transformation efficiency, and the DNA was transformed into a repair-defective strain of E. coli (BMH 71-18 mutS). The use of BMH 71-18 mutS prevented repair of the newly synthesized unmethylated strand and provided high mutation efficiency. A 100-µl volume of the transformation mix was subjected to 60 min of incubation in LB broth without antibiotics and then to overnight selection in the presence of ampicillin, and plasmid DNA was isolated and subjected to a second digestion with AlwNI to enrich for mutant plasmid DNA (undigested). Following desalting with a Qiaex kit (Qiagen Inc., Studio City, Calif.), the plasmid digestion mix was introduced into E. coli DH5. Plasmid DNA from selected transformants was sequenced to verify the introduction of AvrII sites at bp 881 and 1321. Plasmid DNA containing both AvrII sites was digested with AvrII, filled in with Klenow, and religated to form pyscO, which encoded only 9 of the original 154 amino acids of yscO but retained the ribosome binding site of the downstream yscP gene. This was transformed into E. coli DH5. Again, use of the correct in-frame deletion was confirmed by sequencing the double-stranded DNA (dsDNA) of transformants.

Antibody preparation. To determine the location of YscO in bacterial fractions, antibody was raised in rabbits against a fusion protein of glutathione S-transferase (GST) and amino acids 12 to 154 of YscO. Plasmid pGST-YscO was transformed into E. coli DH5 (Table 1; Fig. 1). The soluble GST-YscO fusion protein was expressed and purified as specified by the manufacturer for use of the pGEX-3X vector (Pharmacia Biotech Inc., Piscataway, N.J.) and used to raise antibodies in New Zealand White rabbits as previously described (54).

Cell fractionation and immunoblot analysis. Cells were pelleted by centrifugation, and the top half of the culture medium containing secreted proteins was removed from the tubes. The bacterial pellet was washed and resuspended in 100 mM Tris-HCl (pH 7.4)-1 mM EDTA and lysed by a single passage through a chilled French pressure cell at 20,000 lb/in². Unlysed cells and cellular debris were removed by centrifugation at 8,800 × g for 5 min at 4°C. Total soluble proteins (cytoplasmic plus periplasmic) were separated from membranes of the cleared lysates by ultracentrifugation at 417,000 × g for 15 min in a TLA 100.4 rotor (Beckman, Inc., Palo Alto, Calif.). The membranes were resuspended in 100 mM Tris-HCl (pH 7.4)-1 mM EDTA. Secreted proteins were precipitated with 5% (vol/vol) trichloroacetic acid for 2 h to overnight on ice. After centrifugation (14,000 × g for 30 min at 4°C) to pellet the precipitated proteins, the pellet was neutralized with 1 M Tris-HCl (pH 8.0), resuspended in electrophoresis sample buffer (60 mM Tris-Cl [pH 6.8], 2.3% [wt/vol] sodium dodecyl sulfate [SDS], 20% [vol/vol] glycerol, 5% [vol/vol] -mercaptoethanol, 0.01% [wt/vol] bromophenol blue), and stored at 20°C. Bacterial fractions were analyzed on denaturing SDS-polyacrylamide gels (12 to 15% [wt/vol] acrylamide) (36) followed by transfer to Immobilon-P (Millipore Corp., Bedford, Mass.) using Towbin transfer buffer (75). The gels were loaded such that each lane contained proteins corresponding to equal numbers of bacteria. Y. pestis LCR proteins were visualized on the membranes by using protein-specific primary antibody and a secondary antibody (goat anti-rabbit or goat anti-mouse [Sigma]) conjugated to either alkaline phosphatase or horseradish peroxidase. The primary rabbit polyclonal antibodies included anti-ECP, against total proteins in the culture medium of Y. pseudotuberculosis 43 (pCD1 yopKL::Mu d1 [Ap^r lac]) (45); anti-HTV, against full-length V antigen having a 19-residue leader containing 6 His residues fused to its N terminus (21, 45); anti-YopM, against purified YopM (45); anti-LcrD, against an LcrD peptide coupled to the carrier protein bovine serum albumin (54); anti-YscD, against a fusion protein of GST and YscD (56); and anti-YscP, against a fusion protein of GST and amino acids (aa) 328 to 455 of YscP (49). Rabbit anti-YopE was a generous gift from Gregory V. Plano (University of Miami, Miami, Fla.). Mouse anti-YopE and anti-YopH were a generous gift from Gerard P. Andrews and Arthur M. Friedlander (U.S. Army Medical Research Institute for Infectious Diseases, Ft. Dietrick, Md.).

DNA sequence analysis. DNA and predicted protein sequences were analyzed with PCGene (IntelliGenetics, Inc., Mountain View, Calif.), IntelliGenetics Suite (IntelliGenetics, Inc.), and Coils 2.1 (39, 40). The deduced amino acid sequences were compared with available sequences in the GenBank database via the National Center for Biotechnology Information BLAST (4) mail server. The yscQRS nucleotide sequence of Y. pestis (20) has been updated to include the sequences of yscO and yscP (accession no. L25667).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Analysis of yscO of Y. pestis KIM. We extended the sequence upstream of yscQRS (20) in Y. pestis. This encompassed the 3' end of yscN and the 5' end of yscQ. The predicted amino acid sequences of YscO and YscP were 98 and 100% identical to those predicted for Y. pseudotuberculosis but not significantly similar to those outside the Yersinia spp. Computer analysis of yscO predicted a 154-residue protein with a molecular mass of 18.9 kDa and a predicted isoelectric point of 7.89. The predicted protein sequence does not contain either a putative signal sequence or any hydrophobic domain. Secondary-structure analysis predicted a predominantly alpha-helical structure with a high probability of existing as a coiled-coil protein. Interestingly, Salmonella SpaM (also called InvI) (12) and Shigella Spa13 are also predicted by the same analysis to have a large percentage of charged residues and potentially exist as coiled-coil proteins. The similarity of size, charges, and potential secondary structure suggests that these proteins may have similar functions in their respective bacterial strains but have no similarity in sequence because of the differences in the proteins with which they interact.

Mutational analysis. To determine the role of YscO in the LCR and its presumed role in secretion, we compared the growth (Fig. 2) and secretion (Fig. 3) of Y. pestis KIM5-3001 (wt), KIM5-3001.16 (yscO), and KIM5-3001.16 carrying yscO or yscO and yscP in trans (yscO/O or yscO/OP) in TMH defined medium under both inductive and noninductive conditions of the LCR.

View larger version (14K): [in this window] [in a new window]   FIG. 2.   Growth of Y. pestis KIM5-3001 (wt); KIM5-3001.16, the yscO mutant (yscO); and the mutant carrying yscO in trans (yscO/O) and (yscO/OP). Y. pestis strains were grown at 37°C in the presence or absence of Ca2+ in TMH defined medium. The temperature was shifted from 26 to 37°C (temperature shifts are denoted by arrowheads). Symbols: , +Ca2+; , Ca2+.

View larger version (60K): [in this window] [in a new window]   FIG. 3.   Secretion profile of yscO Y. pestis grown in the presence or absence of Ca2+. Shown is an immunoblot analysis of proteins expressed and secreted from Y. pestis KIM5-3001 (wt); KIM5-3001.16, the yscO mutant (yscO); and the mutant carrying pYscO.2 or pYscOP.2 in trans (yscO/O and yscO/OP). Bacteria were grown in TMH with (+) or without () Ca2+, and proteins from bacterial fractions were separated by SDS-PAGE (A, B, and D, 12% [wt/vol] acrylamide; C, 15% [wt/vol] acrylamide). Proteins from soluble (s), membrane (m), whole-cell (c), and culture medium (e) fractions were visualized with polyclonal antibodies specific to YopE, YopM, V antigen, and YscP and with antibody raised to a mixture of extracellular Yersinia proteins (ECP). The secondary antibody used was conjugated to alkaline phosphatase. Arrows denote the positions of proteins. (A) YopE was visualized with mouse anti-YopE (YopE and its Pla-generated degradation products are enclosed in brackets). (B) YopE was visualized with rabbit anti-YopE. (C) The complexity of the protein pattern reflects degradation products from multiple Yops due to the Pla protease. (D) YscP was visualized with antibody to a GST-YscP fusion protein. Molecular masses (in kilodaltons) of prestained molecular mass standards (Bio-Rad) are denoted to the right in panels A and C.

Secretion of V antigen and YopM expressed from plasmids with non-LCR, inducible promoters. To test the hypothesis that yscO affects secretion directly and to show that low levels of substrate available in the mutant were not responsible for the lack of secretion, we analyzed secretion in Y. pestis KIM5-3001 (wt) and Y. pestis KIM5-3001.16 (yscO), each carrying V antigen or YopM expressed in trans from an inducible, non-LCR-regulated promoter. Plasmid pHTV, which encodes histidine-tagged V antigen (HTV), and pTRCM.2, which encodes YopM, were introduced into wt and yscO Y. pestis strains. Both plasmids contain an isopropyl--D-thiogalactopyranoside (IPTG)-inducible promoter. At 5 h before harvest, the expression of YopM or HTV was induced by the addition of IPTG (1 mM). The expression and secretion of YopM and HTV were monitored by immunoblot analysis (Fig. 4).

View larger version (24K): [in this window] [in a new window]   FIG. 4.   Immunoblot analysis of YopM, V antigen, and HTV in the soluble (s), membrane (m), and culture medium (e) fractions from Y. pestis KIM5-3001 (wt) and KIM5-3001.16 (yscO) with (+) or without () pHTV or pTRCM.2. The Y. pestis strains were grown at 37°C in the absence of Ca2+. Expression of HTV from pHTV and YopM from pTRCM.2 was induced by the addition of IPTG to 1 mM 5 h prior to harvest. The proteins were separated by SDS-PAGE (12% [wt/vol] acrylamide). Polyclonal antibody to HTV or YopM was used to detect HTV and V antigen, and YopM, respectively. Secondary antibody was conjugated to alkaline phosphatase. In the lower panel, YopM was detected as two closely migrating species.

View larger version (8K): [in this window] [in a new window]   FIG. 5.   Detection of LcrD and YscD in yscO Y. pestis. Y. pestis strains were grown in TMH at 37°C in the absence of Ca2+. The proteins in membrane fractions were separated by SDS-PAGE (12% [wt/vol] acrylamide), transferred to Immobilon P, and analyzed by immunoanalysis with antibodies specific to LcrD or YscD. (A) Y. pestis KIM5-3001 (wt) and Y. pestis KIM5-3001.16 (yscO) were analyzed with anti-LcrD. (B) Y. pestis KIM8-3002 (wt Pla) and Y. pestis KIM8-3002.3 (yscO Pla) were analyzed with anti-YscD. The secondary antibody used was conjugated to alkaline phosphatase. Arrows indicate each protein.

Identification of the yscO gene product. Since we had established a direct role for yscO in secretion, we wanted to gain insight into its possible function by determining its bacterial location. YscO was visualized on immunoblots of SDS-polyacrylamide gel electrophoresis (PAGE)-separated proteins from fractionated Y. pestis cultures by using a polyclonal antiserum that was directed against the GST-YscO fusion protein containing aa 13 to 154 of YscO (Fig. 6).

View larger version (31K): [in this window] [in a new window]   FIG. 6.   Localization of YscO by immunoblot analysis. Y. pestis KIM5-3001 (wt), KIM5-3001.16 (yscO), KIM8-3002 (wt Pla), and KIM8-3002.3 (yscO Pla) are shown. Strains carrying plasmids are denoted /O or /OP for pYscO.2 or pYscOP.2, respectively. Bacteria were grown at 37°C in TMH with (+) or without () Ca2+ for 5 h prior to harvest. Proteins from bacterial fractions were separated by SDS-PAGE (15% [wt/vol] acrylamide). Antibody raised against a GST-YscO fusion protein was used to detect YscO in soluble (s), total membrane (m), whole-cell (wc), or culture medium (e) proteins. All panels were analyzed with alkaline phosphatase. Arrows indicate the various proteins that were potential candidates for YscO.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The present study characterized yscO of the Y. pestis yscN to yscU operon. The lack of significant similarity of YscO to the corresponding type III proteins in Salmonella and Shigella suggests that YscO could function in recognition of Yersinia secreted effector proteins, and our data point to a role for YscO in Yop secretion.

At 37°C in the absence of Ca²⁺, our yscO mutant failed to undergo growth restriction and was unable to export LCRS proteins (V antigen and Yops). We believe that this phenotype was an effect of the loss of yscO only, since we observed expression of YscP in the mutant and since provision of only yscO in trans complemented the mutant. A His-tagged fusion of V antigen was not secreted when overexpressed from pHTV in the yscO mutant, and only a small amount of YopM was seen in the culture medium when overexpressed in the mutant from pTRCM.2. These results indicated that the yscO deletion directly prevented secretion of LCRS proteins in Y. pestis. This type of defect is similar to those described for other Yersinia proteins found to be components of the Ysc secretion mechanism, including the known or predicted inner membrane proteins, LcrD (55), YscD (43, 55), and YscR, YscS, and YscU (3, 7, 20); the outer membrane PulD homolog YscC (56); the lipoprotein YscJ (2); and the cytoplasmic protein YscN (78), thought to be associated with the inner membrane and a likely candidate for providing the energy needed for secretion. Other proteins that are essential for secretion but whose function and subcellular location are not known include YscE to YscG and YscI to YscL (2, 27, 43). For all of these essential Ysc components, mutations block the secretion of LCRS proteins and prevent full transcriptional induction of LCRS operons, probably by blocking release of the LCR negative regulator(s) from the cytosol through the type III secretion system. The similarity in growth and secretion defects in the yscO mutant suggests that yscO also is a core (essential) component of the Y. pestis LCRS export apparatus.

Surprisingly, we found that YscO was present in the culture medium in the absence of Ca²⁺. This differs from the localization of the corresponding type III protein in Salmonella called SpaM or InvI. SpaM/InvI was not detected in the culture medium but is essential for Salmonella entry into cultured epithelial cells (12). The corresponding protein in Shigella, Spa13, has been shown to be essential for invasion and presumably for the secretion of Ipa proteins, but its bacterial location has not been determined (63).

Secretion and translocation of Yops have been shown to occur in distinct steps (64) and to require recognition by some component of the type III secretion apparatus. There is no picture yet for how the ca. 22 Ysc components are arranged and function to move Yops through the two bacterial membranes. We would expect that each type III system would have unique proteins or adapted domains of conserved proteins for the recognition and/or transport of the system-specific secreted proteins. If YscO serves this purpose, it is unlikely that it interacts and moves with Yops in a 1:1 ratio, because YscO does not appear to be abundant enough for such a role. YscO evidently is a peripheral protein that is largely associated with or released from the membrane, suggesting that its function does directly correlate with Yop conduction through the membranes. Consistent with this idea, no YscO was found in the culture medium at 26°C, a situation in which no bulk Yop secretion was taking place (data not shown). YscO was sometimes observed in culture medium of the Pla parent at 37°C in the presence of Ca²⁺, a condition in which low levels of Yops were also seen in the culture medium. Our findings suggest the possibility that the type III Ysc mechanism has a dynamic moving core and that YscO is part of this. Such an image is consistent with the developing picture for protein secretion through the Sec system, in which cycles of the piston-like pumping of SecA and the orientation inversion of SecG are essential to the mechanism of protein secretion (17, 47). We do not yet know the role of YscO release in the Yops secretion process. It is conceivable that this is an artifact of our in vitro system, which lacks the localized activation of secretion upon contact with a mammalian cell and the coupling of secretion with translocation into the cell. However, it is also possible that the release of YscO into the medium in our studies reflects a role that links the secretion of Yops with their subsequent targeting into a host cell.