|Home | About | Journals | Submit | Contact Us | Français|
Pathogenic Yersinia species use a virulence-plasmid encoded type III secretion pathway to escape the innate immune response and to establish infections in lymphoid tissues. At least 22 secretion machinery components are required for type III transport of 14 different Yop proteins, and 10 regulatory factors are responsible for activating this pathway in response to environmental signals. Although the genes for these products are located on the 70-kb virulence plasmid of Yersinia, this extrachromosomal element does not appear to harbor genes that provide for the sensing of environmental signals, such as calcium-, glutamate-, or serum-sensing proteins. To identify such genes, we screened transposon insertion mutants of Y. enterocolitica W22703 for defects in type III secretion and identified ttsA, a chromosomal gene encoding a polytopic membrane protein. ttsA mutant yersiniae synthesize reduced amounts of Yops and display a defect in low-calcium-induced type III secretion of Yop proteins. ttsA mutants are also severely impaired in bacterial motility, a phenotype which is likely due to the reduced expression of flagellar genes. All of these defects were restored by complementation with plasmid-encoded wild-type ttsA. LcrG is a repressor of the Yersinia type III pathway that is activated by an environmental calcium signal. Mutation of the lcrG gene in a ttsA mutant strain restored the type III secretion of Yop proteins, although the double mutant strain secreted Yops in the presence and absence of calcium, similar to the case for mutants that are defective in lcrG gene function alone. To examine the role of ttsA in the establishment of infection, we measured the bacterial dose required to produce an acute lethal disease following intraperitoneal infection of mice. The ttsA insertion caused a greater-than-3-log-unit reduction in virulence compared to that of the parental strain.
Three pathogenic Yersinia species, Y. pestis, Y. enterocolitica, and Y. pseudotuberculosis, cause human disease and use a 70-kb virulence plasmid-encoded type III secretion pathway to subvert the innate immune response during host infection (18, 60, 61, 65). The Yersinia type III pathway transports 14 polypeptides across the bacterial envelope (50), and this mechanism allows bacterial multiplication and spread within lymphoid tissues (66). During infection of tissue culture cells, Y. enterocolitica secretes type III substrates either into the extracellular medium (YopB, YopD, YopR, and LcrV) (46, 47) or into the cytosol of host cells (YopE, YopH, YopM, YopN, YopO, YopP, YopT, YscM1, and YscM2) (8, 10, 32, 37, 44, 52, 53, 57, 67). Although many different types of tissue culture cells serve as targets for type III injection, it is thought that only some cell types of an infected host are injected by Yersinia (9).
Yersinia type III secretion is activated by environmental signals (42, 59, 79). During bacterial growth in laboratory media, yersiniae secrete most, but not all, Yop proteins upon chelation of calcium from the extracellular medium (50). When yersiniae secrete large amounts of Yop proteins, bacterial growth is slowed in the absence of calcium, a phenomenon that is referred to as the low-calcium response (Lcr) (73, 74). The critical threshold for activation is <80 μM, well below the calcium concentration in extracellular fluids of mammalian hosts (1.2 mM) but also well above the intracellular calcium concentration of mammalian cells (low nanomolar range) (45, 59). Yersiniae that adhere to the surface of mammalian cells catalyze the type III injection process (67, 68), and recent work suggested that yersiniae measure the intracellular calcium concentration in host cells (45). Bacterial growth in chemically defined media, for example, Dulbecco's minimal Eagle medium, does not lead to type III secretion, even under low-calcium conditions (45). Two additional signals, glutamate and host serum proteins, must be provided with the chemically defined media to activate the type III pathway (45).
Goguen, Yother, and Straley used Mu-d1(Ap lac) transposon mutagenesis in Y. pestis and were the first to isolate mutants that are defective in the Lcr pathway (30, 80). Wolf-Watz, Cornelis, and colleagues demonstrated that such mutations block Yop protein secretion (ysc mutations) across the bacterial envelope (2, 3, 5). Goguen et al. isolated 206 mutants, and further analysis revealed that 47 transposon insertions had occurred on the virulence plasmid, 16 insertions were accompanied by loss of the virulence plasmid, and 143 insertions were mapped to the bacterial chromosome (30). Subsequent research focused on the mutations in the virulence plasmid. DNA sequencing and mutational analysis has revealed the genes for a type III secretion pathway on the virulence plasmid; however, the nature of the mutations on the bacterial chromosome has hitherto not been described (19). Knockout mutations of ysc genes, which encode the secretion machinery, lead to an lcr phenotype, abolishing type III secretion and allowing bacterial growth at 37°C even in the absence of calcium (19). Knockout mutations in lcrE (yopN), tyeA, sycN, yscB, and lcrG result in a calcium-blind temperature-sensitive growth phenotype, as the mutant yersiniae massively secrete Yop proteins at 37°C even in the presence of 5 mM calcium (14, 15, 27, 36, 38, 80). Knockout mutations in yopD, lcrH, and lcrQ (yscM1 and yscM2 in Y. enterocolitica) yield a different phenotype in which some, but not all, Yop proteins are secreted in the presence of calcium (4, 6, 64, 72).
Recent work suggested that yopN, tyeA, sycN, yscB, lcrG, lcrV, yopD, lcrH, yscM1, and yscM2 encode negative regulators of the type III pathway (45). One subset, yopD, lcrH, yscM1, and yscM2, is required to prevent the expression of yop genes by a posttranscriptional control mechanism that targets the 5′ untranslated regions of yop mRNAs (11). Mutations that block the function of any one of these four genes bypass the requirement for glutamate to activate type III secretion (45). The Yersinia type III pathway secretes YopD into the extracellular medium by a mechanism that requires binding of YopD to LcrH in the bacterial cytoplasm (4, 77). YscM1 and YscM2, on the other hand, are injected into the cytoplasm of eukaryotic cells, a transport reaction that requires not only Ysc proteins but also SycH, a cytoplasmic protein that binds to YscM1 and YscM2 and whose overexpression results in the activation of the Lcr (10, 11, 78). Thus, one can view YopD, LcrH, YscM1, YscM2, and SycH as regulators that activate the type III pathway in response to an extracellular glutamate signal by promoting distinct transport reactions and derepression of a posttranscriptional control mechanism for yop genes (61).
A partially overlapping subset of regulatory genes, yopN, tyeA, sycN, yscB, lcrG, and lcrV, affect the response to calcium (23, 24, 39, 47). YopN is initiated into the type III pathway even in the presence of calcium. This process requires SycN and YscB, which form a heterodimer that binds to YopN between residues 15 and 100 (14, 23). TyeA functions as a repressor of YopN secretion in the presence of calcium and binds a more distal portion of YopN (residues 100 to 215) (14, 15). A drop in calcium concentration results in the injection of YopN into eukaryotic cells and the activation of the type III pathway; the regulatory factors YscB, SycN, and TyeA presumably detach from YopN and continue to reside in the bacterial cytoplasm (14, 44). In this model of regulation, YopN acts as a modifier of the type III machinery that hinders the transport of specific sets of Yop proteins in the presence of calcium.
The mechanism by which LcrG regulates yop gene expression is not yet known (70). LcrG binds to LcrV, an antirepressor that is transported by the type III pathway; initiation of LcrV into the type III pathway is dependent on its binding to LcrG (54, 55, 69). Environmental glutamate and serum proteins trigger secretion of LcrV into the extracellular medium (45). In contrast to the case for most other Yops, the removal of calcium blocks LcrV transport, resulting in intrabacterial sequestration of LcrG by LcrV and in the activation of Lcr (47). Knockout mutations in lcrG or yopN bypass the Yersinia requirement for a calcium signal to activate the type III pathway without affecting bacterial dependence on glutamate or serum protein signals (45).
The Yersinia type III pathway can be viewed as a developmentally controlled secretion system, requiring signal input from the environment as well as signal transduction cascades that activate bacterial defense against the host's immune system. In an attempt to identify Yersinia proteins that receive or transmit such signals, we isolated mutants defective in the low-calcium response. One of these mutations, ttsA, mapped to a chromosomal gene encoding a polytopic membrane protein. The ttsA mutant yersiniae were completely defective in low-calcium-induced type III secretion of Yop proteins and were impaired in bacterial motility and in the expression of flagellar filament subunits. These defects were restored by complementation with plasmid-encoded wild-type ttsA. Mutation of the lcrG gene in a ttsA mutant strain restored the type III secretion of Yop proteins, although the double mutant strain secreted Yops in the presence and absence of calcium, similar to the case for mutants that are defective in lcrG gene function alone (24). Together these results suggest that ttsA encodes a regulatory factor of the Yersinia type III pathway.
Y. enterocolitica O:9 strain W22703 has been previously described (20). Plasmid pKD29 was constructed by PCR amplifying the lamB gene of Escherichia coli, using primers LamB-Ase (5′-AAATTAATGTGATGTGAAAAAAGAAAAGCAA-3′) and LamB-Bam (5′-AAGGATCCTTACCACCAGCTTTCCATCTG-3′) with abutted AseI and BamHI restriction sites. The PCR product was digested and cloned between the NdeI and BamHI sites of pKD15 (24). A genomic region spanning 240 bp upstream of the start codon and the ttsA open reading frame was PCR amplified by using TtsA-Eco (5′-AAGAATTCCCTTGCTATCCTTATTAGTCTA-3′) and TtsA-Bam (5′-AAGGATCCCTAACGTAGGAAAGGCGCTG-3′) with abutted NdeI and BamHI restriction sites. The fragment was cloned into pCR2.1 (Invitrogen), creating pNG7.
Y. enterocolitica strain KUM1 (ΔlcrD) and MC2 (ΔlcrG) have been described elsewhere (13, 24). The E. coli S17-1 strain (25) harboring the suicide plasmid pLC28 (16) or pMC8 (24) was used to construct the ΔlcrG mutation in Y. enterocolitica NG15307 (ΔttsA). Allelic exchange following mating of E. coli S17-1 and Y. enterocolitica strains has been previously described (16).
Y. enterocolitica W22703(pKD29) was grown overnight in 50 ml of TBMM (1% tryptone, 0.5% NaCl, 0.2% maltose, 10 mM MgSO4) supplemented with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) and 20 μg of chloramphenicol per ml. Cultures were centrifuged at 10,000 × g for 10 min, and the bacterial sediment was suspended in 10 mM MgSO4. Cells were mixed with λNK1098 at a multiplicity of infection of 10 and incubated at room temperature for 15 min. Samples were shifted to 37°C and incubated for an additional 90 min. The reaction mixtures were mixed in 3 ml of melted top agar at 50°C and poured over tryptic soy agar (TSA) plates supplemented with 12 μg of tetracycline per ml and 5 mM CaCl2. Plates were allowed to solidify and were then incubated at 37°C for 2 days. Colonies were streaked onto TSA with 12 μg of tetracycline per ml (TSATET) and onto TSA treated with 20 mM sodium oxalate (TSAOX). TSATET plates were incubated at 26°C for 2 days, whereas TSAOX plates were incubated at 37°C overnight (14 h). Y. enterocolitica strain W22703 (wild type) was used as a positive control, and KUM1 (ΔlcrD, type III secretion mutant) was used as a negative control for the Lcr growth phenotype.
Overnight cultures of yersiniae were diluted 1:20 into 4 ml of fresh tryptic soy broth (TSB) supplemented with either 5 mM calcium or 5 mM EGTA. Cultures were incubated for 2 h at 26°C and then switched to 37°C and incubated for an additional 3 h. Bacterial cultures were centrifuged at 15,000 × g for 15 min. The culture supernatant was separated from the bacterial sediment (pellet). Proteins in both fractions were precipitated with ice-cold 10% trichloroacetic acid (TCA). Samples were centrifuged at 15,000 × g for 15 min, and the precipitated sediments were washed for 15 min on ice with acetone. Samples were again centrifuged at 15,000 × g for 15 min and air dried after aspiration of most of the supernatant. Proteins were suspended in 50 μl of 500 mM Tris-HCl (pH 8.0)-4% sodium dodecyl sulfate (SDS) and boiled for 5 min. Sample buffer, i.e., 50 μl of YSB (3 M urea, 0.0625 M Tris-HCl, 4% β-mercaptoethanol, 2% SDS, 20% glycerol [pH 6.8]), was added to each sample. Proteins were separated by SDS-10 or 15% polyacrylamide gel electrophoresis (PAGE) and analyzed by immunoblotting with specific rabbit antisera with chemiluminescent detection.
Overnight cultures of yersiniae were diluted 1:50 into 30 ml of T medium (1% tryptone) and incubated at 26°C for 6 h. After centrifugation at 8,000 × g for 10 min, culture supernatants were separated from cell pellets. The protein in both fractions was precipitated with ice-cold 10% TCA. Samples were centrifuged at 15,000 × g for 20 min, and the precipitate was washed with ice-cold acetone. Samples were again centrifuged at 15,000 × g for 15 min. Supernatants were aspirated, and the pellets were allowed to air dry. Proteins were suspended in 100 μl of YSB separated by SDS-15% PAGE, and analyzed by immunoblotting with specific mouse antisera with chemiluminescent detection.
The coding sequence for ttsA379-661 was PCR amplified with abutted NdeI and BamHI restriction sites by using TtsACyNde (5′-AACATATGGGTGCGCAAAGCCAGCAG-3′) and TtsACyBam (5′-AAGGATCCGAGTGTCACGATATCCGGTT-3′), and the product was cloned into pET16b (Novagen) cut with the same enzymes. The recombinant plasmid was transformed into E. coli BL21(DE3). Expression of six-histidyl-tagged polypeptide was induced with 1 mM IPTG, and the polypeptide was purified by Ni-nitrilotriacetic acid affinity chromatography. The purified polypeptide was mixed with complete Freund's adjuvant and injected subscapularly into rabbits for antibody production. Mouse monoclonal antibody 15D8 recognizes E. coli flagellin and cross-reacts with Yersinia flagellin (Igen International).
Yersinia strains were grown in TSB at 26°C with aeration in a roller drum at 50 rpm overnight. Two microliters of culture (106 cells) was spotted onto the center of T medium plates (1% tryptone, 0.35% Difco agar). The plates were incubated at 26°C for 16 h and examined for motility (growth) by capturing images on an alpha imager system.
Bacteria were grown on motility agar (1% tryptone, 0.35% Difco agar) overnight at 26°C, scooped off the plates with a wire loop, and suspended in phosphate-buffered saline. Five microliters of sample was applied to grids, negatively stained with 1% uranyl acetate, and examined by electron microscopy with a Philips CM120 transmission electron microscope.
BALB/c adult female mice were injected intraperitoneally with 0.1 ml of a solution containing 100 mg of iron dextran per ml and 5 mg of desferrioxamine B mesylate per ml. The next day, overnight bacterial cultures were diluted 1:30 in fresh TSB and incubated at 26°C for 3 h. The optical density at 600 nm was measured, and cultures were diluted in sterile phosphate-buffered saline to concentrations ranging from 103 to 109 per 0.1 ml and injected intraperitoneally into 10 mice for each dilution. Aliquots of the diluted cultures were also plated on TSA to determine the CFU injected into mice. Infected mice were observed for 5 days at 3-h intervals. Yersinia-infected animals exhibiting ruffled fur, weight loss, immobility, and labored respiration were judged to suffer from acute lethal infection and were euthanatized according to institutional guidelines and recommendations for euthanasia by the American Veterinary Medical Association. The 50% acute disease dose was determined from a 10-fold series of bacterial dilutions administered to 10 mice per dilution (63).
λNK1098 carries a selectable tetRA marker within the inverted repeats of a mini-Tn10 transposon, while its transposase gene is located immediately adjacent to the mobile element (41). Following λNK1098 infection of a suitable host that does not replicate the bacteriophage, mini-Tn10 may be mobilized and inserted into host DNA. Coliphage λ cannot adsorb to Y. enterocolitica because these microbes express a structurally distinct LamB maltoporin (43). The E. coli lamB gene (49) was cloned into the low-copy-number plasmid vector pHSG576 (76), and the recombinant plasmid pKD29 was transformed into Y. enterocolitica W22703. Transformants were selected on TSA supplemented with chloramphenicol and examined for λNK1098 absorption and plaque formation. Y. enterocolitica W22703(pKD29) adsorbed λNK1098; however, infection did not result in plaque formation, suggesting that the Yersinia strain does not support lytic replication of λNK1098 (data not shown). After λNK1098 absorption on Y. enterocolitica W22703(pKD29), tetracycline-resistant colonies arose on agar plates (TSATET) at a frequency of 10−7, suggesting that the injection of λNK1098 DNA and transposon insertion mutagenesis had occurred. Transposon mutants were plated on agar medium in the presence of 5 mM calcium, and 27,221 colonies were examined for a temperature-resistant phenotype on medium lacking calcium (TSAOX) at 37°C. One hundred sixty-eight mutants that displayed a defect in the Lcr were isolated, as these strains formed colonies at 37°C on TSAOX. The mutants were analyzed by Southern hybridization to determine whether the mini-Tn10 had inserted into plasmid or chromosomal DNA. The transposon insertions were mapped further by using PCR amplification and direct sequencing of the amplified DNA. Most of the mini-Tn10 insertions that led to Lcr defects occurred on the virulence plasmid (92%); 68 insertions were mapped to 17 different genes (virF, virG, lcrD, lcrE, lcrF, sycN, yscB, yscC, yscD, yscF, yscI, yscJ, yscN, yscP, yscR, yscU, and yscY). Of these mutations, 40 mapped to identical sites in virG (1), lcrF (79), and yscU (3), suggesting that the mini-Tn10 of λNK1098 has some insertion site bias. Eight transposon insertions mapped to the bacterial chromosome, one of which carried an insertion in a hitherto-unidentified Yersinia gene that was named ttsA (for type three secretion A) (Fig. (Fig.1A).1A). Blast searches revealed that ttsA is 49% similar to igaA, a Salmonella enterica gene that is required for growth attenuation within cultured fibroblasts (12). Further, Proteus mirabilis umoB, which is 43% similar to igaA and 42% similar to ttsA, is required for flhDC activity and bacterial swarming on agar plates, a phenotype that requires flagellar motility (26, 29).
To assess the role of ttsA in Yersinia type III secretion, bacteria were grown for 3 h at 37°C in TSB supplemented with 5 mM calcium or 5 mM EGTA. The cultures were centrifuged, and the extracellular medium containing secreted proteins (supernatant) was separated from the bacterial sediment (pellet). Proteins in both fractions were precipitated with TCA, separated by SDS-PAGE, and analyzed by immunoblotting (Fig. (Fig.1B).1B). As expected, the wild-type parent strain Y. enterocolitica W22703 secreted YopD into the medium both in the presence and in the absence of calcium (45), although YopD expression and secretion were substantially reduced in the presence of calcium (Fig. (Fig.1B).1B). YopE, on the other hand, was secreted into the medium only in the absence of calcium (45) (Fig. (Fig.1B).1B). Mini-Tn10 insertion into ttsA blocked the secretion of YopD and YopE (Fig. (Fig.1B).1B). Mini-Tn10 insertion into ttsA also caused a reduction in the intrabacterial concentrations of YopD and YopE. As a control, the alpha subunit of RNA polymerase (RpoA) was not secreted by the type III pathway, and mini-Tn10 insertion into ttsA did not affect the expression of rpoA. The wild-type ttsA gene was cloned on a plasmid vector and transformed into NG15307, which restored both the temperature-sensitive growth phenotype in the absence of calcium (Lcr) and the type III secretion of YopD and YopE (Fig. (Fig.1B1B).
ttsA encodes a presumptive polytopic membrane protein of 715 amino acids. The SOSUI transmembrane algorithm (35) predicts four transmembrane helices with the topology shown in Fig. Fig.2A.2A. By using PCR amplification with specific primers, the coding sequence for a portion of ttsA (codons 379 to 661) was cloned into the expression vector pET16b, and the recombinant plasmid was transformed into E. coli BL21(DE3) (75). After overexpression of the recombinant gene product via IPTG-induced T7 RNA polymerase, TtsA379-661 was purified from crude cell lysates by affinity chromatography on nickel-nitrilotriacetic acid and injected into rabbits to raise specific antiserum. Proteins in crude cell lysates were separated by SDS-PAGE and analyzed by immunoblotting (Fig. (Fig.2B).2B). Anti-TtsA revealed a 70-kDa immunoreactive species in crude lysates of the wild-type strain Y. enterocolitica W22703 as well as the lcrD (yscV) mutant strain KUM1 (Fig. (Fig.2B).2B). Mini-Tn10 insertion into ttsA in NG15307 abolished the immunoreactive signal generated by anti-TtsA serum (Fig. (Fig.2B).2B). The signal was restored by transformation of NG15307 with plasmid-encoded wild-type ttsA (Fig. (Fig.2B)2B) Further, the TtsA signal of plasmid-complemented NG15307 was significantly stronger than that of the wild-type strain, consistent with the notion that ttsA expression from multicopy plasmids is increased due to a gene dosage effect (Fig. (Fig.2B2B).
To determine whether ttsA expression resembles the pattern observed for plasmid-borne type III genes, Y. enterocolitica W22703 was grown in the presence and absence of calcium at 26 and 37°C. As expected, growth of yersiniae at 37°C induced the expression of type III genes for YscD (a secretion machinery component) (58) and LcrG (a regulatory factor) (56) (Fig. (Fig.2C).2C). In the absence of calcium, the expression of LcrG increased, whereas the expression of YscD remained unaltered (Fig. (Fig.2C).2C). As a control, neither the temperature nor the calcium concentration affected the expression of RpoA (Fig. (Fig.2C).2C). Immunoblotting with anti-TtsA revealed that the relative amount of this membrane protein was not altered by bacterial growth at 26 or 37°C or by the presence or absence of environmental calcium (Fig. (Fig.2C2C).
Y. enterocolitica strain 8081 employs at least three different type III pathways: the virulence plasmid-encoded pathway, a chromosomally encoded pathway, and the flagellar assembly pathway. Recent work suggests that some proteins, for example, the phospholipase YplA, may be transported by more than one type III pathway (83). We wondered whether the ttsA mutant is defective for more that one type III pathway. PCR analysis of chromosomal DNAs from strains 8081 and W22703 revealed that the former, but not the latter, strain harbors genes for a chromosomal type III pathway (33) (data not shown). These results are consistent with immunoblotting experiments revealing the expression of YspA, a chromosomally encoded type III protein (28), in crude extracts of strain 8081 but not in W22703 (data not shown). We conclude that the serotype O:9 European clinical isolate Y. enterocolitica W22703 does not harbor the chromosomally encoded type III pathway reported for the American isolate Y. enterocolitica O:8 strain 8081, and this finding is consistent with another recently published report (28).
To assess the involvement of TtsA in the flagellar type III pathway, we measured bacterial motility by inoculating yersiniae on soft agar (motility) plates and observing the formation of a concentric ring surrounding the inoculation site (40, 84) (Fig. (Fig.3).3). The wild-type parent strain W22703 and the type III mutant strain KUM1 were both motile; however, the ttsA::Tn10 mutant NG15307 showed very little bacterial growth beyond the inoculation site (Fig. (Fig.3).3). A similar defect was observed for the 8081 variant lacking flhD, a mutation that is known to inhibit bacterial motility (40, 84). Transformation of NG15307 with plasmid-carried ttsA largely restored bacterial motility, although the ring of growth did not reach the same radius observed for strain W22703 or KUM1 (Fig. (Fig.33).
We asked whether the ttsA::Tn10 insertion in Y. enterocolitica NG15307 simply affected the motility (rotation) of the flagellar filament or whether the mutation abolished the assembly of the flagellar filament. Yersinia strains were grown on agar medium, and suspensions of bacterial colonies were stained with uranyl acetate and viewed by electron microscopy. One hundred cells of each strain were scored for the presence of flagellar filaments. As expected, Y. enterocolitica wild-type strain W22703 produced flagellar filaments (Fig. (Fig.4).4). In contrast, the ttsA::Tn10 mutant strain showed a severe defect in flagellar filament assembly, as less than 3% of the cells carried filaments (Fig. (Fig.4).4). This represents a 35-fold reduction in the assembly of the flagellar apparatus, suggesting that the motility phenotype of Y. enterocolitica NG15307 may be caused either by a defect in the assembly of the rotary filament or by an upstream defect in gene expression. Transformation of NG15307 with plasmid-encoded ttsA showed a partial complementation of the phenotype, as 55% of the cells carried flagellar filaments (Fig. (Fig.44).
A mouse model of infection was used to measure Yersinia virulence. BALB/c mice were injected intraperitoneally with a suspension of Yersinia and observed for the appearance of acute disease symptoms during the following 5 days (46) (Fig. (Fig.5).5). An average dose of 7.2 × 103 CFU of Y. enterocolitica W22703 caused an acute lethal disease in half of all experimentally infected animals, while for Y. enterocolitica NG15307, injection of 9.6 × 106 CFU was necessary to produce a similar effect (Table (Table1).1). As a control, the type III mutant strain KUM1 caused acute lethal disease in half of all experimentally infected animals when 3.1 × 108 CFU were injected intraperitoneally. It should be noted that nonvirulent E. coli or Yersinia produces similar disease symptoms after intraperitoneal injection of 0.5 × 109 to 1 × 109 CFU, a phenomenon attributed to endotoxin shock rather than to bacterial pathogenicity (46). We also compared animal survival in a time-to-disease study involving intraperitoneal injection of 104 CFU. Y. enterocolitica W22703 injection killed all animals examined within 3 days, whereas neither Y. enterocolitica KUM1 nor NG15307 caused symptoms during 5 days of infection (Fig. (Fig.5).5). Together these studies demonstrate that ttsA represents an essential virulence factor for the pathogenesis of Y. enterocolitica infections in mice.
We considered two alternative models to explain the multiple phenotypes of ttsA mutants. First, TtsA might be a regulatory factor required for activating multiple type III pathways, i.e., flagellar filament assembly and Yop secretion in strain W22703. Alternatively, TtsA might be required for substrate recognition or transport by all type III machines. If TtsA is a regulatory factor, the elimination of a regulatory factor that acts downstream should bypass the secretion defect of ttsA mutant cells and restore type III transport (45). Conversely, if TtsA is essential for substrate recognition or transport by type III machines, elimination of a downstream regulatory factor would not restore Yop secretion and motility (62). To test these possibilities, we mutated the lcrG gene of NG15307 and analyzed the double mutant strain, Y. enterocolitica KLD8, for type III secretion (Fig. (Fig.6).6). Mutation of lcrG alone in strain MC2 activates type III secretion in the presence of calcium and leads to the transport of large quantities of YopD and YopE into the extracellular media of laboratory cultures (24). Deletion of the lcrG gene in strain NG15307 activated type III secretion in the presence and absence of calcium (Fig. (Fig.6A).6A). When assayed in soft agar plates, the ttsA lcrG double mutant (KLD8) exhibited wild-type motility. These results suggest that ttsA encodes a regulatory factor involved in activating at least two type III pathways encoded by Yersinia. We examined the expression and secretion of flagellin by immunoblotting with a monoclonal antibody (Fig. (Fig.6D).6D). Y. enterocolitica W22703 (wild type) and the type III mutant KUM1 both expressed and secreted flagellin, whereas the flhDC and ttsA mutants did not. In contrast, the lcrG mutant strain MC2 expressed and secreted flagellin, as did the ttsA lcrG double mutant strain KLD8, albeit at a reduced level. These results are consistent with the notion that ttsA is required for flagellar gene expression in Y. enterocolitica.
The properties of Y. enterocolitica that are required for the establishment of animal infections have previously been examined. Miller and colleagues (31, 51, 82) used in vivo expression technology (48), a strategy to identify genes that are expressed during animal infection, as well as signature-tagged mutagenesis (34) to search for virulence genes. The signature-tagged mutagenesis experiments of Darwin and Miller identified virulence genes that caused defects in bacterial multiplication during mouse infection (21). Three categories of mutations were isolated (21). As expected from earlier work, mutations in virulence plasmid genes abolished type III secretion and bacterial multiplication in host tissues (17). Second, mutations in genes required for the functional assembly of lipopolysaccharide O antigen caused structural alterations in the bacterial envelope and diminished the ability of mutant yersiniae to survive complement-mediated killing within host tissues (71). A third category consisted of insertions in chromosomal genes with no previous assignment to virulence function (21). The phenotypes of signature-tagged mutants were examined further by using competition analysis between wild-type and mutant strains. Type III secretion as well as lipopolysaccharide O-antigen biosynthesis mutants displayed a 4- to 5-log-unit reduction in competitiveness (21). However, most signature-tagged mutants of the third category displayed only a small decrease in competitiveness (21). Mutations in the phage shock protein locus of Y. enterocolitica presented a notable exception to this rule. Y. enterocolitica pspA and pspC mutants are capable of growth under conditions that prevent type III secretion (22). However, induction of type III secretion, for example, during host infection, imposes an absolute requirement for pspC expression on yersiniae (22). Thus, even though Miller and colleagues did not score for an Lcr phenotype in their mutant screen (21), the spectrum of isolated genes emphasizes the importance of the Yersinia type III pathway as an essential element of pathogenicity.
To identify the genes and mechanisms that are required for the type III secretion of Y. enterocolitica, an experimental strategy that allows bacteriophage λNK1098-mediated delivery of mini-Tn10 was introduced. This scheme allows for the stable insertion of the mini-Tn10 mobile element into chromosomal and virulence plasmid genes of Y. enterocolitica. Similar to the Tn5 insertions studied by Miller and colleagues (31, 52, 82), mini-Tn10 insertions were nonrandom and occurred at preferred sites in Y. enterocolitica. Thus, the collection of 27,221 mini-Tn10 insertion mutants constructed here represented only 76 unique insertions, not a comprehensive collection. It is therefore not surprising that the search for Lcr mutants turned up only a limited number of insertional mutations in the chromosome of Y. enterocolitica. It is presumed that the number of genes involved in Lcr is significantly greater than the eight chromosomal mutations reported here.
Previous work predicted that chromosomal genes of Y. enterocolitica regulate the type III secretion genes carried by the virulence plasmid in response to environmental signals (30, 45). This hypothesis is corroborated with the identification of ttsA, encoding a membrane protein and positive regulator that activates the type III pathway by relieving the LcrG-mediated repression that occurs in the presence of calcium. This regulatory mechanism must play an important role during infection, as ttsA mutant yersiniae display a 3-log-unit reduction in virulence when measured in a mouse model of infection. Inactivation of ttsA dramatically reduced bacterial motility and flagellar assembly, suggesting that this regulatory factor activates more than one type III pathway. Several recent observations corroborate the notion of cross talk between secretion pathways, as some secretion substrates can be transported by more than one type III machine, for example, the virulence-plasmid-encoded and the flagellar pathways (81). Bleves et al. showed that flhDC, specifying the heterodimeric transcriptional activator necessary for flagellar assembly, exert a negative regulatory role on the expression and secretion of Yop proteins (7). Together these results suggest that multiple environmental signals influence the activity of type III machines, which can occur by both positive and negative regulatory mechanisms. We searched for ttsA genes in the genomes of other gram-negative bacteria and found homologs in S. enterica (50%), S. enterica serovar Typhi (50%), E. coli (49%), P. mirabilis (49%), Shigella flexneri (49%), and Y. pestis (78% identity). In contrast, Pseudomonas aeruginosa, another organism that employs type III secretion for pathogenesis, does not seem to carry ttsA. Thus, although TtsA-regulated type III secretion may occur in other gram-negative bacteria, this cannot represent a universal mechanism.
We gratefully acknowledge S. Minnich (University of Idaho) for sending us strains. We thank members of our laboratory for critical reading of the manuscript.
This work was supported by Public Health Service award AI42797 from the National Institute of Allergy and Infectious Diseases to O.S.