Abstract
The genus Yersinia is composed of 11 species, three of which are pathogenic in humans. The three pathogens, Y. pestis, Y. enterocolitica, and Y. pseudotuberculosis, cause a broad spectrum of disease ranging from pneumonic plague to acute gastroenteritis. Each of the three requires a large, well-defined plasmid for full virulence, as well as many chromosomally encoded virulence factors (CEVF). This review will describe these CEVF and their roles in virulence. In addition, a possible model for key events in Y. enterocolitica pathogenesis is described based on information revealed by analysis of several of the CEVF.
1 Introduction
“One can't forget everything, however great one's wish to do so; the plague was bound to leave traces, anyhow, in people's hearts.” Albert Camus [1]
Certainly Yersinia pestis, the causative agent of bubonic and pneumonic plague, is the most infamous member of the genus Yersinia; however, of the additional 10 species in this genus, two, Y. enterocolitica and Y. pseudotuberculosis, are also serious human pathogens. The study of these Yersinia species has led to many important insights into bacterial pathogenesis, including the characterization of a large virulence plasmid and the type III secretion system encoded by this plasmid. A critical asset to the study of Yersinia species as model organisms for bacterial pathogenesis has been the mouse infection model. Y. pestis, Y. enterocolitica, and Y. pseudotuberculosis each efficiently infect mice, and Y. pseudotuberculosis and Y. enterocolitica infections reproduce much of the pathology seen during human infections. Y. pestis is spread by either fleabite or aerosol, while Y. enterocolitica and Y. pseudotuberculosis are primarily enteric pathogens transmitted through contaminated food or water. Y. pseudotuberculosis and Y. enterocolitica cause a broad spectrum of disease, ranging from self-limiting enterocolitis to life-threatening septicemia [2]. Importantly, not all Y. enterocolitica strains are pathogenic in humans, and those that are can be divided into high pathogenicity strains that cause systemic infections in humans and kill mice at relatively low doses, and low pathogenicity strains that are non-lethal in the mouse model. Each of the three pathogens requires a well-defined 70-kb virulence plasmid (pYV) for full virulence, but early studies by Heesemann et al. showed that while pYV is necessary, it is not sufficient to cause disease [3,4]. This observation led to the search for additional chromosomally encoded virulence factors (CEVF). This review will focus on these factors, primarily those from Y. enterocolitica, because the CEVF have been more extensively studied in this organism. Where appropriate, however, virulence factors from Y. pestis and Y. pseudotuberculosis will also be included. The plasmid-encoded virulence factors will not be discussed here but several comprehensive reviews focusing on various aspects of pYV and the factors it encodes have been recently published [5,6].
2 Chromosomally encoded virulence factors
2.1 Functional complementation screens
Studies using functional complementation of Escherichia coli led to the identification of three chromosomally encoded virulence genes, yst, psa and inv[7,8]. yst encodes the heat-stable enterotoxin Yst, a 30-amino acid peptide. Yst is structurally and functionally similar to the heat-stable (STI) toxin of E. coli; both toxins activate the particulate form of guanylate cyclase, thus increasing cGMP levels in the intestine. The activity of the toxin eventually leads to fluid accumulation in the intestine. This activity led to the hypothesis that Yst is involved in inducing diarrhea during Yersinia infection. Diarrhea is not a major symptom in the mouse infection model; however, oral inoculation of the young rabbit induces clear diarrhea as well as systemic infection. Using the young rabbit model, a yst mutant strain was defective in inducing diarrhea, weight loss, and death relative to wild-type Y. enterocolitica[9]. These data suggest that Yst may be a mediator of the diarrhea observed in infants infected with Y. enterocolitica.
psaA encodes the main structural component of the pH6 antigen fimbriae in Y. pestis. The pH6 antigen is a fibrillar structure with morphology similar to the CS3 fibrillin from enterotoxigenic E. coli. A mutation in the psaA gene resulted in a 200-fold increase in the intravenous LD50 of Y. pestis in the BALB/c mouse model [10]. The Y. pseudotuberculosis homolog is also called psaA and the Y. enterocolitica homolog is called myfA[11,12]. Although neither of these two loci have been tested for their role in virulence, the Y. pseudotuberculosis pH6 antigen has been shown to mediate binding to cultured human cells [13].
A functional inv gene is present on the chromosome of Y. pseudotuberculosis and Y. enterocolitica, while Y. pestis has a disrupted form of inv. The inv gene of Y. enterocolitica encodes invasin, a 92-kDa outer membrane protein. Invasin is the primary invasion factor of Y. enterocolitica in tissue culture invasion models, and is required for efficient translocation of the bacteria across the intestinal epithelium in mice [14,15]. The Y. pseudotuberculosis invasin was shown to bind to host cell β1 integrins that are expressed on the apical surface of the antigen-sampling M-cells that overlie the Peyer's patches (PP) [16]. Analysis of an inv mutant strain of Y. enterocolitica showed no change in virulence in the BALB/c mouse model as assessed by oral LD50[15]. More detailed analysis revealed that although the inv mutant does not have a defect in virulence as measured by LD50, it does have a defect in the early stages of infection as well as altered dissemination to the deeper tissues. The inv mutant is less efficient at colonizing the PP early in infection, but by day 4 post infection the inv null strain has colonized these tissues as well as wild-type. In addition to the defect in colonization of the PP, the inv null strain does not colonize the mesenteric lymph nodes (MLN) as frequently as does wild-type; however, the inv mutant does colonize the deeper tissues such as liver and spleen as effectively as wild-type Y. enterocolitica[15].
Transposon mutagenesis was used to identify the inv regulator, rovA (regulator of virulence). RovA was found to be required for inv expression both in vitro and in vivo. Additional characterization showed that a mutation in rovA had a more severe impact on virulence than the loss of inv alone. Although the inv mutant strain did not alter the oral LD50 in the BALB/c mouse model, the rovA mutant had a 70-fold increase in oral LD50 in the same model [17]. In addition, the rovA mutant strain was 300-fold more attenuated in the C57BL/6 mouse model. Kinetic analysis revealed that the rovA mutant colonized the PP more efficiently than the inv mutant, perhaps due to the small amount of invasin still synthesized by the rovA mutant. However, the loss of rovA resulted in a more significant defect in the ability to colonize the MLN, as well as a defect in the ability to effectively colonize the deeper tissues. Importantly, although the rovA mutant did not colonize the deeper tissues well, occasional colonization of the spleen did occur, but without prior or concurrent colonization of the lymph nodes. The failure of the rovA mutant to colonize the MLN and spleen was likely not due to an inability to survive in these tissues, because when the bacteria were delivered intraperitoneally the rovA mutant survived and replicated to high numbers in both the MLN and spleen. Additionally, when the rovA mutant did rarely reach the deeper tissues via the oral infection route it was capable of replicating to high numbers.
Interestingly, although the rovA mutant colonizes the PP, the patches are not as inflamed as in infections with wild-type Y. enterocolitica. Analysis of the pro-inflammatory cytokine profile of mice infected with the rovA mutant or wild-type Y. enterocolitica revealed that most cytokines (interferon-γ, tumor necrosis factor-α, and interleukin (IL)-1β) are expressed to similar levels in mice infected with either strain, with one striking exception. IL-1α is expressed at high levels in mice infected with wild-type Y. enterocolitica; however, in the rovA infected mice there is no detectable IL-1α expressed [18]. This loss of IL-1α appears to play a role in the virulence defect of the rovA mutant. When mice infected with wild-type Y. enterocolitica are depleted of IL-1α, there is very little inflammation of the PP, similar to that of the rovA-infected mice. In addition, the wild-type Y. enterocolitica are less virulent in IL-1α-depleted mice, as assessed by oral LD50[19]. These data suggest that the host inflammatory response during infection with wild-type Y. enterocolitica may be a key factor in virulence.
RovA is a member of the MarR family of transcriptional regulators, yet a rovA mutant is more attenuated than a mutation in the gene, inv, that it is known to regulate. Thus it is likely that the increased attenuation of a rovA mutant is due to the loss of expression of other virulence genes regulated by RovA. The fact that a mutation in rovA had no effect on expression of the virulence plasmid-encoded factors suggests that these additional genes are most likely chromosomally encoded [17].
2.2 Targeted disruption of presumed virulence genes
The targeted disruption of genes suspected to be important for virulence, based on studies in other organisms, has also led to the identification of factors involved in Yersinia pathogenesis. The O-antigen (O-Ag) of Y. enterocolitica and the Mn-cofactored superoxide dismutase, SodA, were both shown to be involved in virulence in this manner. Zhang et al. [20] showed that a spontaneous mutant of the O-Ag biosynthesis operon resulted in a 100-fold increase in LD50 in the DBA/2 mouse model. Furthermore, genes involved in O-Ag biosynthesis were identified several times in a signature-tagged mutagenesis (STM) screen, and in a screen for in vivo expressed genes (IVET) [21,22]. Competition assays also revealed that these mutants were less virulent than wild-type Y. enterocolitica in a mixed infection.
SodA, the Mn-cofactored superoxide dismutase, was shown to be a virulence factor by intravenous and intragastric kinetic assays in the mouse model. The sodA mutant strain colonized the PPs less effectively than wild-type Y. enterocolitica, and by either route of inoculation, the sodA mutant was not recovered from deeper tissues such as the liver and spleen [23]. These data suggest a role for SodA in bacterial survival during the systemic phase of infection. One potential mechanism for this decrease in survival in the mouse tissues is an increased sensitivity to reactive oxygen during the inflammatory process. This possibility was addressed by measuring the oxidative burst of neutrophils in response to contact with wild-type or sodA mutant Y. enterocolitica. The level of superoxide detected in neutrophils exposed to the sodA mutant was higher than the level of superoxide detected in neutrophils exposed to wild-type Y. enterocolitica. This suggests that the absence of SodA results in an increased exposure of the bacteria to reactive oxygen, which correlates well with the observed increase in susceptibility of the sodA mutant to killing by neutrophils in vitro.
2.3 Functional screens
Large-scale mutant screens for potential virulence determinants also have been performed. Specifically, enriching for genes that are expressed in vivo has been very useful for identifying chromosomally encoded virulence factors. One such technique is IVET. The first IVET screen in Y. enterocolitica was done using an oral route of inoculation and identified genes expressed in the PPs early during infection [21]. The result of this screen was the identification of 61 allelic groups that had insertions in genes that were expressed in vivo (i.e. the mouse), but not in vitro (i.e. the laboratory). Of these 61, 48 were sequenced and classified according to their probable function and sequence homology. This search resulted in the identification of previously characterized virulence genes, as well as genes not previously known to be involved in virulence. Known genes include those that encode proteins involved in iron acquisition. Two of these genes, fyuA and irp2, are found only in the highly pathogenic Yersiniae. The chromosomal region that contains these genes is known as the high pathogenicity island (HPI) and is required for full virulence. A recent review by E. Carniel comprehensively covers the HPI and its role in virulence [24].
Of the loci that had no previously known role in virulence, two have been characterized further. The first locus, initially designated hre-22, now hreP, encodes a subtilisin/kexin-like protease [25]. When this gene was disrupted and its role in virulence assessed, the results showed no change in kinetics of infection but did show a 33-fold increase in oral LD50 in the BALB/c mouse model. The second locus, initially designated hre-20, encodes a LysR-type transcriptional regulator. Disruption of this locus resulted in altered kinetics of infection, and a five-fold increase in oral LD50[21]. A screen for genes regulated by this factor was done using transposon mutagenesis and resulted in the identification of rscA, a homolog of the HmwA adhesin of Haemophilus influenzae. A mutation in rscA resulted in a similar kinetic phenotype to that of the original hre-20 mutant, which is now designated rscR (regulator of systemic colonization). Mice infected with either the rscR mutant or the rscA mutant had increased bacterial dissemination to the spleen compared to wild-type Y. enterocolitica; colonization of the PP and MLN appeared normal [26]. This is an unusual phenotype and it will be interesting to determine how RscA, presumably the effector, influences dissemination of the bacteria.
Another IVET screen done in Y. enterocolitica was done using an intraperitoneal route of inoculation, and identified genes expressed in the spleen during infection [27]. Thirty-one loci were identified that were expressed in vivo but not in vitro, and each of these loci was given a sif designation for systemic invasion factor. As with the previous IVET screen, known virulence genes were identified as well as previously uncharacterized genes potentially involved in virulence. The previously known genes include manB, a gene involved in O-Ag biosynthesis, and fyuA, a gene on the HPI that encodes an iron siderophore precursor. Only one of the additional sifs has been characterized further for its role in virulence. sif15 is homologous to HP0694 in Salmonella, a putative outer membrane protein. Intraperitoneal competition infections with wild-type Y. enterocolitica and a sif15 mutant showed that a disruption of sif15 resulted in attenuation relative to wild-type Y. enterocolitica. However, the same competition assays done using an oral route of infection did not result in an attenuated phenotype. These data suggest that sif15 is involved in systemic colonization, but not early stages of infection. Further characterization of this locus as well as the remaining sifs will be necessary to elucidate their roles in virulence.
STM was successfully used in both Y. enterocolitica, and Y. pseudotuberculosis to identify several CEVF. This approach allows one to screen pools of mutants together to identify mutants unable to survive in vivo [28]. The STM screen in Y. enterocolitica identified 27 loci that when disrupted were not able to survive in vivo [22]. Nine of these loci were on the virulence plasmid, the remaining 18 were located on the chromosome. Of the 18 chromosomal genes, nine were genes involved in O-Ag biosynthesis. Disruptions of each of these genes resulted in attenuation relative to wild-type Y. enterocolitica in the mouse model, as measured by intraperitoneal competition assay. Another chromosomally encoded gene, irp1, which is encoded on the HPI, was also identified by this screen, and was attenuated by the same competition assay. Of the eight additional chromosomal loci identified and shown to be involved in virulence in this screen, several were involved in stress response (Table 1). The disruption of one of these chromosomally encoded genes, pspC, resulted in a level of attenuation similar to that seen with a virulence plasmid-cured strain. The psp locus is found in all enterobacteria whose genomes have been sequenced to date. The physiological role of the E. coli psp operon is not known, but it has been proposed that this operon plays a role in protecting the bacteria from various stresses associated with membrane perturbations. In Y. enterocolitica the severely attenuated phenotype of the pspC mutant was shown to be due to a growth defect when the type III secretion system encoded on the virulence plasmid is active [29].
Chromosomally encoded Yersinia virulence factors
| Gene or gene homolog | Function or property | Required for virulence in mice | Reference |
| inv | Invasin | + | [8,14,15] |
| yst | Yst toxin | +a | [7,9] |
| rovA | Transcriptional regulator (regulates inv) | + | [17] |
| sodA | Mn-cofactored superoxide dismutase | + | [23] |
| hreP | Subtilisin/kexin-like protease | + | [21,25] |
| rscA | Homolog of outer membrane adhesin | + | [26] |
| rscR | Regulator of rscA | + | [21] |
| pspC | Regulation of phage shock protein operon | + | [22,29] |
| yifH | Enterobacterial common antigen synthesis | + | [22] |
| dnaJ | Heat shock response | + | [22] |
| pstC | Inorganic phosphate importer | + | [22] |
| topA | DNA topoisomerase I | + | [22] |
| nlpD | Outer membrane lipoprotein | + | [22] |
| yibP | Unknown | + | [22] |
| cls | Cardiolipin synthesis | + | [30] |
| ksgA | Kasugamycin resistance | + | [30] |
| sif15 | Putative outer membrane protein | + | [27] |
| pH 6 antigen | Fimbriae | + | [10,,,13] |
| O-Ag biosynthesis operon | Lipopolysaccharide O-Ag production | + | [20] |
| HPI | Iron acquisition | + | [24] |
| ail | Attachment and invasion locus | − | [14,34,,,37] |
| ysa locus | Chromosomal type III secretion system | − | [31] |
| hre loci | Unknown | ND | [21] |
| sif loci | Unknown | ND | [27] |
| Gene or gene homolog | Function or property | Required for virulence in mice | Reference |
| inv | Invasin | + | [8,14,15] |
| yst | Yst toxin | +a | [7,9] |
| rovA | Transcriptional regulator (regulates inv) | + | [17] |
| sodA | Mn-cofactored superoxide dismutase | + | [23] |
| hreP | Subtilisin/kexin-like protease | + | [21,25] |
| rscA | Homolog of outer membrane adhesin | + | [26] |
| rscR | Regulator of rscA | + | [21] |
| pspC | Regulation of phage shock protein operon | + | [22,29] |
| yifH | Enterobacterial common antigen synthesis | + | [22] |
| dnaJ | Heat shock response | + | [22] |
| pstC | Inorganic phosphate importer | + | [22] |
| topA | DNA topoisomerase I | + | [22] |
| nlpD | Outer membrane lipoprotein | + | [22] |
| yibP | Unknown | + | [22] |
| cls | Cardiolipin synthesis | + | [30] |
| ksgA | Kasugamycin resistance | + | [30] |
| sif15 | Putative outer membrane protein | + | [27] |
| pH 6 antigen | Fimbriae | + | [10,,,13] |
| O-Ag biosynthesis operon | Lipopolysaccharide O-Ag production | + | [20] |
| HPI | Iron acquisition | + | [24] |
| ail | Attachment and invasion locus | − | [14,34,,,37] |
| ysa locus | Chromosomal type III secretion system | − | [31] |
| hre loci | Unknown | ND | [21] |
| sif loci | Unknown | ND | [27] |
aVirulence assayed in young rabbit model.
Chromosomally encoded Yersinia virulence factors
| Gene or gene homolog | Function or property | Required for virulence in mice | Reference |
| inv | Invasin | + | [8,14,15] |
| yst | Yst toxin | +a | [7,9] |
| rovA | Transcriptional regulator (regulates inv) | + | [17] |
| sodA | Mn-cofactored superoxide dismutase | + | [23] |
| hreP | Subtilisin/kexin-like protease | + | [21,25] |
| rscA | Homolog of outer membrane adhesin | + | [26] |
| rscR | Regulator of rscA | + | [21] |
| pspC | Regulation of phage shock protein operon | + | [22,29] |
| yifH | Enterobacterial common antigen synthesis | + | [22] |
| dnaJ | Heat shock response | + | [22] |
| pstC | Inorganic phosphate importer | + | [22] |
| topA | DNA topoisomerase I | + | [22] |
| nlpD | Outer membrane lipoprotein | + | [22] |
| yibP | Unknown | + | [22] |
| cls | Cardiolipin synthesis | + | [30] |
| ksgA | Kasugamycin resistance | + | [30] |
| sif15 | Putative outer membrane protein | + | [27] |
| pH 6 antigen | Fimbriae | + | [10,,,13] |
| O-Ag biosynthesis operon | Lipopolysaccharide O-Ag production | + | [20] |
| HPI | Iron acquisition | + | [24] |
| ail | Attachment and invasion locus | − | [14,34,,,37] |
| ysa locus | Chromosomal type III secretion system | − | [31] |
| hre loci | Unknown | ND | [21] |
| sif loci | Unknown | ND | [27] |
| Gene or gene homolog | Function or property | Required for virulence in mice | Reference |
| inv | Invasin | + | [8,14,15] |
| yst | Yst toxin | +a | [7,9] |
| rovA | Transcriptional regulator (regulates inv) | + | [17] |
| sodA | Mn-cofactored superoxide dismutase | + | [23] |
| hreP | Subtilisin/kexin-like protease | + | [21,25] |
| rscA | Homolog of outer membrane adhesin | + | [26] |
| rscR | Regulator of rscA | + | [21] |
| pspC | Regulation of phage shock protein operon | + | [22,29] |
| yifH | Enterobacterial common antigen synthesis | + | [22] |
| dnaJ | Heat shock response | + | [22] |
| pstC | Inorganic phosphate importer | + | [22] |
| topA | DNA topoisomerase I | + | [22] |
| nlpD | Outer membrane lipoprotein | + | [22] |
| yibP | Unknown | + | [22] |
| cls | Cardiolipin synthesis | + | [30] |
| ksgA | Kasugamycin resistance | + | [30] |
| sif15 | Putative outer membrane protein | + | [27] |
| pH 6 antigen | Fimbriae | + | [10,,,13] |
| O-Ag biosynthesis operon | Lipopolysaccharide O-Ag production | + | [20] |
| HPI | Iron acquisition | + | [24] |
| ail | Attachment and invasion locus | − | [14,34,,,37] |
| ysa locus | Chromosomal type III secretion system | − | [31] |
| hre loci | Unknown | ND | [21] |
| sif loci | Unknown | ND | [27] |
aVirulence assayed in young rabbit model.
The STM screen in Y. pseudotuberculosis identified five virulence plasmid loci, and seven chromosomal loci as being attenuated relative to wild-type Y. pseudotuberculosis[30]. As with the Y. enterocolitica screen, genes in the O-Ag biosynthesis operon were identified, and mutations in these genes were shown to attenuate Y. pseudotuberculosis in the mouse model. In addition, four other chromosomal loci were identified, including inv. Interestingly, the Y. pseudotuberculosis STM screen was done using an oral inoculation route, and the inv mutant was attenuated relative to wild-type in the PP and MLN. However, when a competition assay was done using an intraperitoneal infection route, the inv mutant was not attenuated relative to wild-type Y. pseudotuberculosis. These data support the earlier studies in Y. enterocolitica that demonstrated a role for inv early during infection, where invasin is required to translocate across the intestinal epithelial barrier, but not required for survival in the deeper tissues [15].
2.4 Chromosomal genes implicated in virulence
In addition to the previously described genes that clearly play a role in virulence in either the mouse or young rabbit model, there are other CEVF that have been implicated in pathogenesis. This class of genes would include the uncharacterized hre and sif loci, a chromosomally encoded type III secretion system in Y. enterocolitica, the yplA phospholipase and ail[14,21,27,31,32]. yplA encodes a phospholipase implicated in virulence; a yplA mutant strain showed reduced inflammation in the mouse infection model [32]. Interestingly, a newly described secretion pathway, the flagellar export apparatus, secretes this phospholipase [33]. This system shares significant homology with the contact-dependent type III system, and as with the type III system it is possible that additional virulence factors will also be secreted by this apparatus. There are also suggestive data that support a role for ail in virulence of Y. enterocolitica. First, ail was identified by its ability to confer an attachment and invasion phenotype to E. coli K-12, suggesting ail may be a secondary invasion factor during the infectious process. Secondly, it has been demonstrated that ail is required for resistance to killing by human serum [34,35]. This phenotype may be required for complete virulence in the human host but not in the mice, because Y. enterocolitica is not killed by mouse serum. Thirdly, the ail gene is found in pathogenic serotypes of Y. enterocolitica, but is not found in non-pathogenic serotypes of Y. enterocolitica nor in non-pathogenic species of Yersinia[36]. Interestingly, the Y. pseudotuberculosis ail gene does not confer the attachment and invasion phenotype, while it does promote serum resistance [13].
3 Infection model
Together with the virulence plasmid, the CEVF lead to the pathogenic phenotype of the Yersinia species. It is interesting to note that, although the complete loss of the type III secretion system and its secreted effectors encoded on the virulence plasmid results in the inability of the organism to cause any significant disease in the mouse model, this is more an exception than a rule for individual virulence factors. The phenotypes exhibited by the loss of various chromosomal genes are often subtle, but they offer valuable insights into the mechanism of disease that cannot be seen when a mutation totally abrogates virulence such as loss of pYV. For instance, the phenotype of the inv mutant in the mouse model gave important insight into the disease process even though there was no change in the ability of the mutant to kill the mice. Likewise, the differences in the phenotype of the inv mutant, the rovA mutant, and the rsc mutants may reveal critical information about the mechanism of dissemination, give insight as to how the bacteria actually kill the mouse, and reveal a distinct role for the IL-1α-mediated inflammatory response for full virulence of the bacteria.
During infection with wild-type Y. enterocolitica, at early times after infection the bacteria colonize the host PPs to high numbers, followed by the appearance of the bacteria in the MLN and subsequently the spleen. It is not clear whether the bacteria travel through the lymphatics to the MLN and then from there spread to the spleen and deeper tissues, or whether there are two parallel routes of spread, via the lymphatics and via the blood. The observation that the inv mutant does not colonize the MLN as well as wild-type but does colonize the deeper tissues to the same degree as wild-type Y. enterocolitica, along with the observation that the rovA mutant does not colonize the MLN but does infrequently colonize the spleen, suggests that there may indeed be two separate routes of dissemination (Fig. 1). Analysis of the rsc mutants indicated that while dissemination to the MLN is normal, dissemination to the spleen is altered, suggesting the primary routes of dissemination to the lymph nodes versus the deeper tissues such as the spleen may be distinct.
This is a pictorial representation of the proposed two routes of bacterial dissemination during Y. enterocolitica infection. The effect of each mutant on these pathways is also shown. Heavy solid lines represent significant bacterial spread, while lighter lines and dashed lines represent decreasing amount of bacterial spread. Red dashes represent an apparent block in dissemination to the tissues.
This is a pictorial representation of the proposed two routes of bacterial dissemination during Y. enterocolitica infection. The effect of each mutant on these pathways is also shown. Heavy solid lines represent significant bacterial spread, while lighter lines and dashed lines represent decreasing amount of bacterial spread. Red dashes represent an apparent block in dissemination to the tissues.
The inv mutant, although it does not colonize the MLN as well as wild-type Y. enterocolitica, colonizes the deeper tissues well, and kills the mice as well as wild-type Y. enterocolitica; in contrast, the rovA mutant does not colonize the deeper tissues as well as the inv mutant and does not readily kill the mice. These data suggest not only that there may be one route of spread through the lymphatics and one through the blood, but also that perhaps spread via the blood is the most critical in terms of death of the mouse.
The probable existence of two routes of dissemination and the key role of the host inflammatory response in Y. enterocolitica virulence probably would not have been identified without the analysis of the inv, rovA, and rsc mutants in the mouse model. Much remains to be learned from further characterization of the inv, rovA, and rsc mutants along with the other identified virulence factors and those yet to be identified.
Yersinia pathogenesis is a complex process that involves a multitude of genes both on the chromosome and on the virulence plasmid. This brief review describes several of the known chromosomally encoded factors and mentions several other genes potentially involved in virulence that await further characterization. Table 1 lists the genes described in this review and references their discovery. Much is yet to be done in order to completely understand the mechanism by which these bacteria cause disease, but it is clear that many additional virulence factors, both chromosomally encoded and plasmid-encoded, must be characterized in order to accomplish this goal.
