INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Pathogenic members of the Yersinia genus include the etiologic agent of plague, Y. pestis, and the enteropathogenic species Y. pseudotuberculosis and Y. enterocolitica. The ability of these facultative intracellular pathogens to cause disease in mammalian hosts is conferred, in part, by components of the coordinately regulated low-Ca²⁺ response stimulon (LCR) encoded on a common 70-kb virulence plasmid. The virulence of Y. pestis is enhanced by two additional, unique plasmids, pMT1 (encoding murine toxin and capsular fraction 1 protein) and pPCP1 (encoding the plasminogen activator serine protease Pla). The proteinaceous capsule enhances resistance to phagocytosis by monocytes (6), while Pla is an outer membrane protease (66) that is required for full virulence from peripheral infectious routes (4). LCR components encoded on the conserved plasmid (pCD1 in Y. pestis) include a set of secreted antihost proteins termed Yops and V antigen (LcrV) and a specialized apparatus for the secretion and deployment of these proteins during infection. LcrV and most Yops are absolutely essential for virulence in a mouse model (reviewed in references 9, and 44), and the yersiniae employ an elegant mechanism which links expression of these proteins to their subsequent secretion and delivery to respective host targets.

The LCR is thermally regulated by the transcriptional activator LcrF (28), with maximal expression possible at 37°C. Maximal induction of Yops and LcrV, however, can occur only when bacteria are activated for secretion of these proteins by cultivation in contact with eukaryotic cells (10) or in a Ca²⁺-deficient medium (9, 44, 70, 80), reflecting an intimate link between regulation of the LCR and the ability to secrete proteins. Secretion occurs via a type III mechanism and is mediated by a Yop secretion apparatus (Ysc) consisting of at least 22 gene products (reviewed in references 9 and 44). Numerous studies demonstrated that Y. pestis ysc strains could not be fully induced at 37°C, and it was subsequently surmised that the Ysc mediates secretion of a negative regulator. In a currently accepted model, Ysc secretion channels remain blocked in the presence of Ca²⁺ or in the absence of cell contact by at least three regulatory proteins. Blockage mediated by LcrE (also called YopN) (15) and TyeA (24) at the outer surface and by LcrG (43) at the inner surface enforces a negative feedback mechanism by causing retention of the secretable negative regulator LcrQ (49, 54, 68) and an accessory protein, YopD (77). Under conditions that relieve the Ysc secretion block, LcrQ is purged from the bacteria, resulting in derepression of Yops and LcrV expression and secretion.

Yops and LcrV lack cleavable signal sequences and are secreted without processing by the Ysc (10, 33). Work by Anderson and Schneewind (1) demonstrated that YopE and YopN (LcrE) secretion can be mediated by a signal contained within their respective mRNAs. In addition, specific Yop chaperones (Sycs) have been identified for some Yops. These are necessary for the secretion of their respective Yops (75) and may function by donating the Yop to the secretion machinery (7). Whether this theme applies to all Yops remains unclear. LcrV has not been systematically mapped for secretion determinants, and no Syc has been identified for it. However, LcrV does bind to the cytoplasmic Ysc gate protein LcrG (43). During contact-induced release, at least six Yops (YopE, YopT, YopH, YpkA, YopM, and YopJ) are targeted, without being released into the surrounding medium, into the cytoplasm of associated eukaryotic cells, where they function as direct antihost effectors. Several of these effector Yops are susceptible to the Y. pestis-specific Pla protease and do not accumulate on the surface of the bacteria (44). Targeting of the secreted effector Yops across the eukaryotic plasma membrane is accomplished by at least three additional Yops (YopB, YopD, and YopK) and LcrV. YopK influences the size (22) of the YopB-containing pore which inserts into the eukaryotic membrane. Although both YopD and LcrV are required for targeting, their precise roles are unclear. YopD has been shown to bind several Yops, including YopB (40), and LcrV may be required for YopB to reach the plasma membrane and form a pore. Yop targeting has been characterized mainly for the enteropathogenic yersiniae, but it is clear that Y. pestis, which has an essentially identical set of LCR genes (9, 46), targets Yops into eukaryotic cells in a similar, contact-activated manner (see, e.g., references 13, 41, and 63). The specific adhesin(s) that functions to promote contact has not been identified, but the resulting attachment lasts at least 4 h. Importantly, the Pla protease, unique to Y. pestis, has not been found to prevent Yop targeting (63), although it can degrade surface-exposed Yops (67). Once targeted, three effector Yops disrupt host cell signaling by interfering with intracellular cascades (YopJ) (57, 58), dephosphorylating focal adhesion components (YopH) (2, 47), and phosphorylating unidentified host proteins (YpkA) (18). YopE (55, 56) and YopT (23) cause disruption of actin microfilaments. The activities of YopE, YopT, and YopH can be directly visualized as the rounding-up or cytotoxic phenotype of infected eukaryotic cells (10). The intracellular role of YopM is unidentified, but the protein does localize to the nuclei of infected cells (63).

V antigen has been associated with full virulence of Y. pestis for decades (5, 52). LcrV is a 327-residue soluble protein (53) whose only known homolog is PcrV of Pseudomonas aeruginosa (78). Unlike most Yops, LcrV is relatively resistant to the Pla protease (69). Evidence that implicates LcrV as a multifunctional protein has accumulated. Within Y. pestis, LcrV is required for induction of the LCR (52, 65), and this activity is manifested by the noted ability of LcrV to associate with cytoplasmic LcrG (43) and titrate the LcrG-mediated inner Ysc secretion block (41). Nilles et al. (41) and Sarker et al. (60) have also reported that LcrV functions in Yop targeting through an effect on YopB. This requirement was demonstrated in the absence of LcrG, thereby separating LcrV's function in LCR induction from that in targeting of Yops (41).

The observation that LcrV-specific antibodies confer resistance to experimental plague (3, 73, 74) prompted the hypothesis that LcrV also promotes virulence outside the bacteria, and it has been proposed that LcrV acts directly to compromise innate host immune responses. In support of this idea, an LcrV-containing fusion protein (in the absence of yersiniae) was shown to prevent production of the proinflammatory cytokines tumor necrosis factor alpha and gamma interferon in mice (37). This immunosuppressive effect was not limited to Yersinia infections, since exogenous LcrV exacerbated heterologous infections of mice with Listeria monocytogenes or Salmonella typhimurium (37). Nedialkov et al. (38) have proposed that this effect arises from the ability of LcrV to upregulate the negatively acting cytokine interleukin-10. Interestingly, others have shown that LcrV inhibits chemotaxis of neutrophils (76).

This complex array of functions has made the elucidation of LcrV's mechanisms of action during infection difficult. It is likely that different domains of LcrV are required for its various functions. Genetic analysis indicated that residues 224 to 266 are required for binding to LcrG (60), and a protective epitope lies between residues 176 and 276 (21, 35). The N terminus of LcrV is dispensable for the immunosuppressive activity, because the fusion protein used by Nakajima et al. (37) lacked residues 1 to 67 of LcrV. However, the first 125 residues likely are important for secretion of LcrV (65). How LcrV carries out its multiple functions remains obscure, and its molecular targets are unknown.

This study focuses on LcrV's functions in Yop targeting and virulence. We found that both the N and C termini of LcrV are required for its role in Yop targeting but that the N terminus was dispensable for LcrV's role in activation of the Ysc for Yop secretion. LcrV localized to the surface of non-contact-activated Yersinia, but the truncated versions did not, suggesting the hypothesis that LcrV may exist on the bacterial surface in a pretargeting complex. However, LcrV-specific antibody did not prevent Yop targeting, suggesting that antibody can be protective against plague without blocking the delivery of Yops.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, eukaryotic cell lines, and growth conditions. The strains and plasmids used in this study are listed in Table 1. This study employed the human HeLa epithelium-derived cell line. Unless otherwise noted, Escherichia coli strains were cultivated in Luria-Bertani medium (34) at 37°C or on Luria-Bertani agar. Y. pestis strains were grown in the defined medium TMH (70) as described for physiological studies. Briefly, Y. pestis cultures were cultivated with shaking at 200 rpm overnight at 26°C for about eight generations. Cultures were diluted into fresh 26°C TMH to an optical density at 620 nm (OD₆₂₀) of 0.1, initially incubated at 26°C, and shifted to 37°C when cultures had reached an OD₆₂₀ of ca. 0.2. Cultures then were harvested 6 h after the shift to 37°C. For infection of eukaryotic cell lines, Y. pestis was grown in heart infusion broth (HIB) (Difco Laboratories, Detroit, Mich.) at 26°C for at least six generations in exponential phase. During construction of the lcrG3 strain, Y. pestis was cultivated on tryptose blood agar (Difco) supplemented with chlorotetracycline HCl (Sigma) and fusaric acid (32), as described by Nilles et al. (41). When appropriate, bacteria were grown in the presence of antibiotics, which were used at 15 µg/ml for tetracycline, 25 µg/ml for chloramphenicol, 50 µg/ml for kanamycin, or 100 µg/ml for ampicillin and streptomycin (Sigma, St. Louis, Mo.). The HeLa cell line was maintained in RPMI (GIBCO-BRL, Grand Island, N.Y.) supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum (FBS) (GIBCO-BRL) at 37°C, with CO₂ maintained at 5%. During experiments to determine the partitioning of LcrV and Yops within cultures of infected HeLa cells, infection was done in RPMI lacking FBS. When arabinose induction was required, infection was done in Leibovitz's L15 medium (L15) (GIBCO-BRL) lacking FBS.

DNA methods. Template DNA was isolated with midi-prep or spin-prep columns from Qiagen, Inc. (Studio City, Calif.), and cloning procedures were performed essentially as described previously (31). Amplification of specific DNA was achieved by PCR in a Perkin-Elmer Cetus (Foster City, Calif.) GeneAmp model 2400 thermocycler with Pfu (Stratagene, La Jolla, Calif.) or Vent (New England Biolabs, Beverly, Mass.) DNA polymerase and oligonucleotide primers synthesized by Genosys Biotechnologies (The Woodlands, Tex.) or the Macromolecular Structure Analysis Facility (University of Kentucky). Typical PCR conditions included a 5-min preincubation at 94°C followed by 30 amplification cycles. Denaturation, annealing, and extension were done at 94, 55, and 72°C, respectively, for 15 to 30 s each. Restriction fragments excised from agarose gels and PCR products were purified by using a Qiaquick gel extraction or DNA purification kit (Qiagen), respectively. Transformation of E. coli was achieved by the CaCl₂ method (31) or the frozen-storage-based protocol as described previously (19), and that of Y. pestis was achieved by electroporation as described by Perry et al. (45). Double-stranded DNA from pHT-V was sequenced by the method of Sanger et al. (59) with the Sequenase version 2.0 kit (United States Biochemical Corp., Cleveland, Ohio) and -³⁵S-dATP (NEN Research Products, Boston, Mass.).

Plasmid construction. The plasmids pHT-V (encoding N-terminally histidine-tagged LcrV) and pHT-VN68 (encoding N-terminally histidine-tagged, N-terminally truncated LcrV) were constructed by insertion of portions of lcrV (PCR amplified with Pfu polymerase) into the expression plasmid pPROEX-1. For pHT-V, lcrV was amplified from pES6-1 DNA. The 5' end of the sense primer (5'-TATATAGGCGCCATGATTAGAGCCTACGAACAAAACCC-3') was a non-Yersinia-encoded sequence containing a NarI restriction site; the antisense primer (5'-CCCCCTCCTTTTAGG-3') annealed immediately downstream of lcrV. PCR products were 5' end phosphorylated with T4 polynucleotide kinase (Promega, Madison, Wis.) and then digested with NarI. The resulting DNA was ligated into NarI- and StuI-cut pProEX-1 and transformed into E. coli DH5. PCR fidelity was verified by sequencing the cloned lcrV. pHT-VN68 was derived from pHT-'V2, in which the encoded 'LcrV lacks the first 67 amino acids of LcrV. A sense primer (5'-ACGATATCCCAACGACCG-3') complementary to pPROEX-1 DNA and an antisense primer (5'-CCCCCTCCTTTTAGG-3') positioned 68 nucleotides (nt) into lcrH (also called sycD) were used to amplify only 'lcrV and the polyhistidine tag-coding region. The PCR product was digested with KasI and 5' end phosphorylated with T4 kinase. Treated fragments were ligated into KasI- and StuI-digested pProEX-1 and used to transform E. coli DH5. The resulting leader sequences fused to LcrV and VN68 were similar (HT-V, MGHHHHHHDYDIPTTENLY FQGA; HT-VN68, MGHHHHHHDYDIPTTENLYFQGAHMGIQR).

Strain construction. Y. pestis strains lacking the plasminogen activator protease (Pla) were created by curing strains of the encoding plasmid pPCP1. Y. pestis KIM5-3241 and KIM5-3001.12 were transformed with a plasmid, pOS50 (a gift of J. Goguen, University of Massachusetts, Worcester), containing pPCP1's origin of replication. Transformants were transferred on tryptose blood agar containing tetracycline to allow curing of pPCP1 via competition with pOS50. Potential plasmid-cured strains were screened for the loss of pPCP1 by analysis of plasmid DNA as previously described (25) and for the lack of plaminogen activator activity by assaying for activation of thrombin-mediated cleavage of fibrin film as previously described (71). Strains cured in this way then retained the pOS50 plasmid. Y. pestis KIM8-3002.7 (lcrG3) was created by allelic exchange of the parent lcrG on pCD1 with the lcrG3 allele carried on pMNlcrG3. Introduction of pMNlcrG3 into Y. pestis, selection for crossover events, and screens for successful replacements were done as described previously (41). For physiological studies characterizing the phenotype of Y. pestis KIM8-3002.7, bacteria were cultivated in TMH as described above (70), and the cultures were separated into whole cells and cell-free culture supernatants as done previously (41).

Antibody preparations. LcrV-specific antibodies (-HTV) were raised against purified HT-V in female New Zealand White rabbits as previously described (50). A nonspecific, irrelevant antibody pool (-NS) was prepared from rabbit preimmune serum harvested prior to hyperimmunization with HT-V. -Yersinia was a preparation of polyclonal rabbit antibodies that had been raised against whole cells of Y. pestis KIM6 grown at 26°C in HIB and partially purified by ammonium sulfate precipitation. In one set of experiments YopM-specific rabbit polyclonal antibody (-YopM) (39) was used as an irrelevant antibody control. All antibodies other than -Yersinia were purified from serum by using an Econo Pac protein A cartridge (Bio-Rad Laboratories, Hercules, Calif.) and subsequently dialyzed into phosphate-buffered saline (PBS) (135 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, pH7.4) as described previously (39). For some experiments, -HTV or -NS was dialyzed into buffer containing 100 mM sodium acetate (pH 5.5) and used to generate antibody Fab fragments. Antibodies were digested with papain as described previously (20) for 18 h, and Fab fragments were purified by collecting flowthrough fractions from a protein A cartridge. Fab preparations were dialyzed into PBS and concentrated by centrifugation in a Centricon 30 concentrator (Amicon, Inc., Beverly, Mass.), and the protein content was quantitated by the bicinchoninic acid assay (Pierce Chemical, Rockford, Ill.). When resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and visualized by immunoblot analysis, Fab fragments were of predicted sizes, and when used as primary antibodies in immunoblots, they were able to detect LcrV (data not shown). -HTV was absorbed twice with nonspecific antigens for use in immunostaining of samples to be analyzed by confocal microscopy. Briefly, spleens were excised from female BALB/c mice (Harlan Sprague-Dawley, Indianapolis, Ind.), minced, and extracted with ice-cold acetone. Five hundred micrograms of precipitated, dried material was combined with 500 µl of -HTV in the presence of a protease inhibitor cocktail (Pefabloc, leupeptin, and aprotinin at 20 µg/ml each; all from Boehringer Mannheim Corp., Indianapolis, Ind.) and incubated overnight on ice. Solid material was then pelleted at 4°C by centrifugation at 20,800 × g and discarded.

Passive immunization of mice with -HTV. Female BALB/c mice (7 to 8 weeks old) were passively immunized with 500 µg of -HTV or with 333 µg of -NS Fab fragments as a negative control. There were four mice per group in one experiment and five per group in a repetition. Single antibody doses were given intraperitoneally in 100- to 250-µl volumes of PBS. Forty-eight hours after passive immunization, mice were challenged with a lethal dose of Y. pestis KIM5-3001. Bacteria were diluted into PBS and injected intravenously into the retro-orbital sinus, and mice were observed for 14 days postinfection. The actual infecting doses were determined by enumerating CFU from samples of the doses given to the mice. These were 746 CFU in one experiment and 215 in a replicate experiment.

Infection assays. Prior to infection, eukaryotic cells were subcultured into six-well 35-mm-diameter tissue culture plates in RPMI with 10% (vol/vol) FBS and incubated at 37°C with 5% CO₂ for roughly 72 h or to a density of 5 × 10⁵ to 8 × 10⁵ per well. Cells were washed twice with RPMI or L15 medium lacking FBS immediately prior to infection. Bacteria were cultivated at 26°C in HIB and harvested at an OD₆₂₀ of ca. 1.0. Arabinose or isopropyl--D-thiogalactopyranoside (IPTG) was added to 0.2% or 0.1 mM, respectively, at 30 to 60 min prior to harvest for strains harboring constructs with inducible promoters. Harvested bacteria were diluted directly into 37°C L15 or RPMI lacking FBS (containing arabinose or IPTG where appropriate) and transferred to duplicate wells containing eukaryotic cells. Bacteria were added at a 10-fold excess per well over the number of eukaryotic cells (nominal multiplicity of infection [MOI] = 10). The plates were then centrifuged at 200 × g for 5 min to achieve contact between bacteria and target cells and incubated at 37°C with 5% CO₂ (for RPMI) or at 37°C (for L15) for 4 h. After infection, one replicate well per infecting bacterial strain was treated for 5 min at 37°C (5% CO₂) with trypsin (100 µg/ml in RPMI or L15), and the trypsin treatment was terminated by addition of the protease inhibitor cocktail described above. The contents of the wells were subsequently harvested and fractionated following lysis of the HeLa cells in ice-cold water containing protease inhibitors at 2 µg/ml, essentially as described previously (13, 41). Three fractions were collected: (i) large cellular debris plus yersiniae recovered by centrifugation at 20,800 × g, (ii) proteins precipitated with 10% (wt/vol) trichloroacetic acid (TCA) from filtered media, and (iii) TCA-precipitated soluble proteins released by water lysis plus small organelles not removed by centrifugation at 20,800 × g. The fractions were solubilized in electrophoresis sample buffer containing 2.3% (wt/vol) SDS, 5% (vol/vol) -mercaptoethanol, 60 mM Tris (pH 6.8), and 25% (vol/vol) glycerol.

Protein electrophoresis and immunoblot analysis. Proteins from fractionated bacterial cultures and from infection assays were resolved in polyacrylamide gels (12% [wt/vol] acrylamide) by SDS-PAGE (27). Samples were loaded so that lanes containing different culture fractions represented equivalent volumes of the original cultures. Resolved proteins were transferred to Immobilon-P (Millipore, Corp., Bedford, Mass.) in carbonate buffer (pH 9.9) (64) when LcrG was analyzed or in Tris-glycine (72) in all other cases. Specific proteins were detected by using polyclonal antibodies specific for His-tagged LcrV (-HTV), His-tagged YopD (-YopD) (76), His-tagged YopE (-YopE) (a gift of G. Plano, University of Miami) or glutathione S-transferase (GST)-tagged LcrG (-GST-G) (43). Alkaline phosphatase or horseradish peroxidase-conjugated goat anti-rabbit IgG (Sigma) was used to visualize proteins by development with nitroblue tetrazolium-5-bromo-4-chloro-3-indolylphosphate (NBT-BCIP) (GIBCO-BRL) or enhanced chemiluminescence (ECL) substrate (Pierce), respectively.

Confocal analysis. Bacteria were tested for surface localization of LcrV by cultivating Yersinia strains at 26°C in HIB as done in infection assays to an OD₆₂₀ of ca. 1.0. Bacteria were diluted into 37°C RPMI lacking FBS (containing 0.1 mM IPTG where appropriate) to a density of 10⁶/ml, and 1.0-ml aliquots were centrifuged at 200 × g for 5 min onto 13-mm-diameter glass coverslips, each bearing ca. 10⁵ HeLa cells, in 24-well cluster dishes (Costar, Cambridge, Mass.). After incubation for 2 h at 37°C with 5% CO₂, wells were washed once with 1.0-ml volumes of Hanks balanced salt solution (GIBCO-BRL) and subsequently treated with 2% (wt/vol) paraformaldehyde (Sigma) in PBS (pH 7.4) for 30 min. Fixed bacteria were blocked for 1 h at room temperature with 10% (vol/vol) FBS plus 1% MS in PBS and then washed and incubated for 1 h in PBS plus 1% (vol/vol) MS with -HTV, -YopE, or -Yersinia. Samples were then washed with PBS and incubated for 1 h with anti-rabbit IgG antibodies coupled to the fluorochrome Oregon green (Molecular Probes, Eugene, Oreg.). After staining, coverslips were mounted by using SlowFade mounting medium (Molecular Probes) and analyzed by using laser-scanning confocal microscopy. Specimens were visualized by differential interference contrast with a Leica DM IRB/E inverted microscope with Nomarski optics and a 100× objective. Fluorescence images were derived from a single 0.5-µm-thick optical plane with the Leica TCS NT confocal system (Ar-Kr laser), and image size was recorded at 1,024 by 1,024 pixels with an additional zoom factor of between 4.5 and 5.5.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

LcrV is essential for Yop targeting, while LcrG has a facilitative role. The goal of this study was to gain insight into how LcrV promotes virulence of Y. pestis. We began by clarifying the essentiality of LcrV for Yop targeting. Our previous report had shown that LcrV was required for the targeting of Yops into infected eukaryotic cells, but our data (41) and those of Sarker et al. (60) suggested that LcrG also may be needed. Importantly, LcrV and LcrG have been shown to interact (43, 60), raising the possibility that they may function synergistically to promote Yop targeting. To determine how LcrG might influence LcrV's activity in targeting, we constructed a Y. pestis strain possessing an in-frame, nonpolar deletion within lcrG that did not remove the ribosome-binding site for the downstream gene (lcrV) and therefore was predicted not to decrease LcrV expression. Consistent with the case for other Yersinia strains lacking lcrG (41, 61, 64), the lcrG3 Y. pestis KIM8-3002.7 displayed a "Ca²⁺-blind" growth phenotype in defined medium (79): i.e., the bacteria ceased growth prematurely at 37°C whether Ca²⁺ was present or not (data not shown). This phenotype is often indicative of strong induction of the LCR and secretion of LcrV and Yops, regardless of the presence of Ca²⁺. In agreement with these predictions, the levels of proteins detected in whole-cell samples (Fig. 1A) from the lcrG3 strain were elevated irrespective of the presence of Ca²⁺ (lanes 3 to 6), whereas the parent possessing wild-type pCD1 displayed significant induction of LcrV, YopD, and YopE only in the absence of Ca²⁺ (lanes 1 and 2). Importantly, the induced levels of protein encoded by the adjacent downstream gene (lcrV) and by the last gene of the polycistronic operon (yopD) confirmed the lack of polarity of the lcrG mutation in lcrG3 Y. pestis. When complemented with pAraG18K in the absence of arabinose induction, LcrG was not detectable in the lcrG3 strain by this assay, but the small amount present was sufficient to restore near-wild-type Ca²⁺ regulation of all gene products tested (Fig. 1A, lanes 7 and 8). Addition of arabinose, however, resulted in detectable levels of LcrG and complete restoration of Ca²⁺ regulation (lanes 9 and 10), demonstrated by the decreased levels of LcrV and Yops D and E when Ca²⁺ was present. Providing excess LcrV (lanes 11 to 14) did not appreciably alter the constitutively induced levels of YopD and YopE, supporting the observation by Nilles et al. (43) that the inductive role of LcrV in Ca²⁺ regulation is dependent on its interaction with LcrG. Finally, providing both lcrG and lcrV in the lcrG3 mutant restored the wild-type regulation of expression in the presence (Fig. 1A, lanes 17 and 18) but not the absence (lanes 15 and 16) of arabinose induction. Unlike the presumed small amount of LcrG present in the absence of arabinose (lane 7), LcrG was unable to mediate a significant reduction of YopD or YopE in the presence of Ca²⁺ (lane 15), possibly because excess LcrV provided on the same construct could prevent LcrG-mediated blockage of the Ysc secretion channel.

View larger version (20K): [in this window] [in a new window]   FIG. 1.   KIM8-3002.7 (lcrG3) is defective for negative regulation of LcrV and Yops, and LcrG is required for efficient secretion of LcrV. Y. pestis KIM8-3002 (parent) containing plasmid pBAD18-Kan (vector) (lanes 1 and 2) or Y. pestis KIM8-3002.7 possessing pBAD18-Kan (lanes 3 to 6), pAraG18K (+LcrG) (lanes 7 to 10), pAraV18K (+LcrV) (lanes 11 to 14), or pAraGV18K (+LcrGV) (lanes 15 to 18) was grown in TMH at 37°C in the presence (odd-numbered lanes) or absence (even-numbered lanes) of Ca2+. Arabinose was added to 0.2% (wt/vol) prior to a shift to 37°C for vector controls (lanes 1, 2, 5, and 6) or to achieve induction of lcrG (lanes 9 and 10), lcrV (lanes 13 and 14), or lcrG and lcrV (lanes 17 and 18). Cultures were harvested after 6 h of growth at 37°C, and samples were fractionated into whole cells (A) and cell-free culture supernatants (B). Material corresponding to 0.03 OD620 unit · ml was resolved by SDS-PAGE in a 12% polyacrylamide gel and analyzed by immunoblotting with an antibody cocktail containing -HTV, -YopD, -YopE, and -GST-G. Proteins were visualized by probing with alkaline phosphatase-conjugated secondary antibodies followed by development with NBT-BCIP.

View larger version (93K): [in this window] [in a new window]   FIG. 2.   LcrG facilitates targeting of Yops into infected HeLa cells. Y. pestis KIM8-3002 (parent) or Y. pestis KIM8-3002.7 (lcrG3) possessing pBAD18-Kan, pAraG18K (lcrG3 pG), or pAraV18K (lcrG3 pV) was used to infect HeLa cell monolayers at an MOI of 10. Expression of LcrG was induced by addition of arabinose to 2% (wt/vol), and LcrV was induced by addition of IPTG to 0.1 mM. Cell morphology was visualized by phase-contrast microscopy and recorded at 3.5 or 6 h after infection by photography through a green filter.

Both N and C termini of LcrV are required for Yop-targeting activity. We initiated the identification of regions of LcrV that are necessary for targeting of Yops by testing the ability of several altered versions of LcrV to support this function. The proteins used in these assays are shown diagrammatically in Fig. 3. HT-V represents full-length LcrV, while HT-VN68 and E15VN68 represent N-terminal truncations which lacked the first 67 residues of LcrV. VC216 was truncated at the C terminus and lacked the final 111 residues of LcrV, a domain which displays a high degree of sequence similarity (27.9% similarity and 49.5% identity) to PcrV of P. aeruginosa.

View larger version (38K): [in this window] [in a new window]   FIG. 3.   Schematic representation of LcrV proteins. Portions of lcrV were subcloned into expression vectors possessing IPTG-inducible promoters. HT-V (full-length LcrV) and HT-VN68 (N-terminally truncated LcrV) possesses an N-terminal leader sequence containing six His residues. Fusion of the YopE mRNA secretion signal to LcrV in pE15VN68 resulted in the addition of amino acids 1 to 15 of YopE to N-terminally truncated LcrV, yielding the protein E15VN68. It was expressed from the yopE promoter. VC216 represents a C-terminal truncation and lacks the last 111 residues of LcrV.

View larger version (14K): [in this window] [in a new window]   FIG. 4.   Yop and V antigen expression and secretion in defined medium by lcrV yopJ Y. pestis KIM5-3241 complemented with pHT-V and pHT-VN68. Yersiniae pregrown at 26°C in the defined medium TMH containing (+) or lacking () 2.5 mM CaCl2 were shifted to 37°C, and IPTG was added to cultures containing pHT-V and pHT-VN68 to induce expression of the complementing lcrV genes. After 5 h, cultures were harvested, and proteins in whole cells (C) and the cell-free culture medium (M) were analyzed in immunoblots probed with a mixture of antibodies specific for LcrV and YopE (A) or anti-YopM (B). Proteins were visualized by probing with alkaline phosphatase-conjugated secondary antibodies followed by development with NBT-BCIP.

View larger version (32K): [in this window] [in a new window]   FIG. 5.   HT-VN68 is not secreted by Y. pestis and cannot complement Yop targeting. Y. pestis KIM5-3241 (lcrV yopJ) (lane 1) or Y. pestis KIM5-3241 carrying pHT-V (lanes 2 to 5) or pHT-VN68 (lanes 6 to 9) was used to infect HeLa cells, in duplicate, at an MOI of 10 in the presence of 0.1 mM IPTG. After 4 h, one replicate for each was treated with trypsin (lanes 4, 5, 8, and 9) at 100 µg/ml to assess protease accessibility of HT-V, HT-VN68, and YopE. The presence of proteins in cell-free medium (M) or H2O-released HeLa cell soluble (S) fractions was analyzed by resolving samples representing 20% of the original cultures in a 12% polyacrylamide gel followed by immunoblotting. Blots were probed with -HTV and -YopE, and proteins were visualized by probing with alkaline phosphatase-conjugated secondary antibodies followed by development with NBT-BCIP.

View larger version (14K): [in this window] [in a new window]   FIG. 6.   Both the N and C termini of LcrV are required to mediate Yop targeting and for release of LcrV from bacteria. Y. pestis KIM8-3002 (parent), KIM5-3241 (lcrV yopJ), and KIM8-3002.8 (lcrG lcrV) alone (lanes 1, 2, 7, 8, 13, and 14), carrying pVC216 (lanes 5, 6, 11, 12, 17, and 18) or carrying pE15VN68 (lanes 3, 4, 9, 10, 15, and 16) were used to infect HeLa cells, in duplicate, at an MOI of 10 in the presence of 0.2 mM IPTG. After 4 h, trypsin was added to replicate wells at 100 µg/ml to assess protease accessibility of LcrV and YopE. Samples were fractionated, and portions representing 12% of the original cultures were resolved in 12% polyacrylamide gels. Native LcrV, truncated LcrVs, and YopE in the HeLa cell soluble fraction (A) or released into culture medium (B) were detected by probing blots with -HTV and -YopE. Proteins were visualized by probing with alkaline phosphatase-conjugated secondary antibodies followed by development with NBT-BCIP.

Surface localization of LcrV on Y. pestis. Nilles et al. (41) and Sarker et al. (60) had speculated that LcrV is a dynamic participant in the secretion process and may even mediate assembly of a surface structure for the delivery of Yops. According to this hypothesis, LcrV may be present on the outer surface of Y. pestis. We tested this possibility by staining bacteria with -HTV and analyzing LcrV distribution by laser-scanning confocal microscopy (Fig. 7). Y. pestis was incubated in RPMI, a medium that does not promote secretion of LcrV or Yops in the absence of contact with eukaryotic cells (13). Surface deposition of LcrV was tested by probing fixed but unpermeabilized bacteria with -HTV. Staining with antibodies specific for Yersinia lacking pCD1 was used as a positive control and produced an even, homogeneous surface staining of the bacteria. However, probing with -HTV showed that LcrV was present in punctate zones on the surface of the bacteria. This deposition was not affected by Pla, since this surface-associated pattern of LcrV was also detected in Y. pestis lacking the plasmid pPCP1 (data not shown). We also tested localization of LcrV in a strain lacking the outer membrane secretin YscC, which is essential for Ysc function. A control experiment verified that the amounts of LcrV in the bacteria were the same for the parent and yscC strains after 2 h in RPMI at 37°C with 5% CO₂; i.e., the yscC mutation did not prevent thermal induction of LcrV expression under these conditions (data not shown). No fluorescence signal was detected on the surface of yscC Y. pestis KIM8-3001.12 (Fig. 7A), indicating that LcrV requires the Ysc machinery to gain access to the bacterial surface.

View larger version (28K): [in this window] [in a new window]   FIG. 7.   LcrV can localize to the surface of Y. pestis. Surface deposition of LcrV and YopE was tested in Y. pestis strains lacking pPCP1. Parent Y. pestis (KIM8-3002), secretion-negative Y. pestis (yscC KIM8-3001.12), and the lcrV-null strain (KIM8-3241) complemented with pHT-VN68 were fixed on coverslips after incubation for 2 h at 37°C in RPMI. Bacteria were stained with -Yersinia or -HTV (A) or -YopE (B). The respective proteins were then visualized by treatment with Oregon green-conjugated secondary antibody followed by confocal laser-scanning microscopy. Immunofluorescence (IF) confocal and differential interference contrast (DIC) images are shown.

LcrV-specific antibody does not block targeting of Yops and LcrV. Our previous results suggested that LcrV may need to be secreted to have its role in targeting of Yops, raising the possibility that the surface-localized LcrV may be important for Yop targeting. Since we were able to detect surface-localized LcrV with antibody, we tested whether LcrV-specific antibodies would interfere with Yop targeting (Fig. 8). In one experimental design, -HTV or -YopM was present at an estimated 2,000-fold molar excess over the amount of LcrV that typically is expressed in the infected monolayer (see Materials and Methods). The -YopM served as an irrelevant antibody against a target not expected to be at the bacterial surface to control for nonspecific effects of immunoglobulin on the assay. The antibody was added to Y. pestis KIM8-3002 just prior to infection of washed HeLa cells, but the bacteria were allowed to settle by gravity onto the monolayer, instead of being rapidly brought into contact by centrifugation, to allow the antibody time to interact with bacteria and HeLa cells. In some wells, the HeLa cells were pretreated for 30 min with antibody from preimmune serum (-NS), and -NS was also present along with -HTV or -YopM during the infection. The monolayers were observed for retraction and rounding up (cytotoxicity), which are indicative of targeting of YopE. Figure 8A shows that by 3 h, all infected monolayers were showing clear signs of cytotoxicity (data not shown for wells containing -NS). By 4 h, the infected cells were further rounded up. This test indicated that LcrV-specific antibody did not block targeting of Yops by Y. pestis.

View larger version (33K): [in this window] [in a new window]   FIG. 8.   LcrV-specific antibody does not neutralize Yop-targeting activity by nonpreinduced Y. pestis. (A) -HTV or -YopM antibody was added at 175 µg/ml to warm RPMI containing Y. pestis KIM8-3002, and the mixture was added without centrifugation to monolayers of HeLa cells at an MOI of 10. After incubation for 3 h at 37°C with 5% CO2, the cultures were visualized by phase-contrast microscopy and photographed through a green filter. (B) HeLa cell monolayers were untreated (UI) (lane 1) or infected with Y. pestis KIM8-3002 at an MOI of 10 in the presence (lanes 6 to 9) or absence (lanes 2 to 5) of LcrV-specific Fab antibody fragments (-HTV Fab). For this experiment, Fab antibody fragments were present at 26.5 µg/ml, corresponding to a 1,000-fold molar excess over estimated LcrV levels. After 4 h, replicate wells were treated with trypsin at 100 µg/ml prior to harvesting (lanes 4, 5, 8, and 9) or were harvested directly. Samples were fractionated into cell-free supernatants (lanes 3, 5, 7, and 9) and HeLa cell soluble fractions (lanes 1, 2, 4, 6, and 8). To verify that released LcrV could be quantitatively bound by the Fab fragments, LcrV bound by -HTV Fab was immunoprecipitated (Immunoppt) from a nontrypsinized supernatant fraction by using anti-IgG and protein A-agarose beads. The combined void and protein A column washes (FT), elution (E) fractions, and fractionated culture samples were analyzed by immunoblotting with -HTV and -YopE. Preparations of whole antibody (IgG) (lane 12) and Fab fragments (lane 13) were also resolved as references. Proteins were visualized by probing with horseradish peroxidase-coupled secondary antibody, developed with ECL reagent, and exposed to film.

View larger version (34K): [in this window] [in a new window]   FIG. 9.   LcrV-specific antibody does not neutralize Yop-targeting activity of Y. pestis preinduced for expression of LcrV and Yops. Y. pestis KIM8-3002 was grown overnight at 26°C, diluted into warm RPMI, centrifuged in six-well dishes, and incubated for 2 h at 37°C with 5% CO2 to allow surface expression of LcrV. After 1.5 h of the incubation, the bacteria were resuspended, and -YopM or -HTV was added to some wells. The yersiniae were used to infect HeLa cells at an MOI of 2, and some wells were photographed at hourly intervals, up to 4 h, through phase-contrast optics and a green filter. (A) Cultures after 2 h of infection. (B) After 4 h of infection, replicate cultures were treated (+) or not () with trypsin, harvested, and fractionated as described in Materials and Methods, and TCA-precipitated proteins from cell-free medium (Med) and HeLa cell soluble (Sol) fractions were analyzed by immunoblotting with a mixture of -HTV and -YopE antibodies as a probe. The rightmost two lanes contain 0.2 and 2.0 µg of -HTV as a reference for the size of the predominant band due to antibody. The proteins were visualized by probing with horseradish peroxidase-coupled secondary antibody followed by development with ECL reagent and exposure to film.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, we have found that LcrV's N and C termini are required for Yop targeting. We also demonstrated a novel surface localization of LcrV that did not occur for truncated forms of LcrV unable to sponsor Yop targeting, and we speculate that the secretion of LcrV is required for targeting to occur. However, LcrV's function in targeting of Yops was not blocked by our protective LcrV-specific polyclonal antibody.

Our initial tests for requirements of the targeting reaction stemmed from our previous finding that LcrV was required for targeting of Yops into eukaryotic cells but that LcrG was necessary for maximal hemolysis by Y. pestis, which was an indirect assay for a functional Yop delivery channel (41). In that study we were not able to clearly demonstrate LcrG's role in targeting, because the LcrG strain we used expressed LcrV only weakly, and supplying LcrG in trans without also supplying LcrV caused downregulation and blockage of secretion. Accordingly, we wanted to know whether Y. pestis was like Y. enterocolitica (61) in that LcrV may function more efficiently in Yop targeting in the presence of LcrG. Our new, nonpolar lcrG3 strain indeed induced cytotoxicity significantly more slowly than did the parent Y. pestis strain, but its Yop-targeting defect could be alleviated by complementation with lcrG or by providing excess LcrV, indicating that LcrG is not essential for Yop targeting but is indirectly involved. Curiously, LcrV was secreted at significantly lower levels by the lcrG3 mutant than by the parent strain. Release of LcrV into the medium could be increased to near-wild-type levels when lcrV or lcrG was expressed in trans. Clearly, LcrG is not essential for LcrV secretion, but it may function to make secretion of LcrV more efficient. Future studies will determine whether it mediates this effect within the bacteria by modulating Ysc activity or at the bacterial surface, where it conceivably could be acting in association with LcrV.

To begin to dissect how LcrV carries out its multiple functions, we tested whether two versions of LcrV could support Yop secretion and targeting. HT-VN68 was able to support secretion of YopE, consistent with the idea that LcrV participates in Ysc activation by binding to the inner-gate protein LcrG, since this N-terminally truncated version retains a domain required for association with LcrG (60), and cross-linking experiments have shown that HT-VN68 and LcrG can interact (12). However, HT-VN68 did not support targeting of YopE into the HeLa cell cytoplasm. We believe that this defect in targeting was probably not due to the His-containing N-term