|Home | About | Journals | Submit | Contact Us | Français|
Here we demonstrate that flagellar secretion is required for production of secreted lipase activity in the fish pathogen Yersinia ruckeri and that neither of these activities is necessary for virulence in rainbow trout. Our results suggest a possible mechanism for the emergence of nonmotile biotype 2 Y. ruckeri through the mutational loss of flagellar secretion.
Yersinia ruckeri is the etiologic agent of enteric redmouth disease, a disease of salmonid fish species that is found worldwide in areas where salmonid fish species are farmed (3, 6, 18, 20). Vaccines for enteric redmouth disease have been used successfully for nearly 3 decades and consist of immersion-applied, killed whole-cell preparations of motile serovar 1 Y. ruckeri strains (22). Recently though, outbreaks have been reported in vaccinated fish at trout farms in the United Kingdom (2), Spain (9), and the United States (1). The Y. ruckeri strains isolated from these outbreaks are uniformly atypical serovar 1 isolates lacking both flagellar motility and secreted lipase activity. These variants have been classified as Y. ruckeri biotype 2 (BT2) and are believed to have a reduced sensitivity to immersion vaccination (2). The objective of this study was to obtain a better understanding of the emergence of BT2 Y. ruckeri by identifying genetic elements necessary for expression of the Y. ruckeri flagellum and determining the role that the flagellum plays in virulence by using a rainbow trout infection model.
To identify genes involved in flagellar motility, random mutant clones were generated using the transposon Tn5-RL27 as previously described (12) and screened on motility medium (tryptic soy broth, 0.3% agar). By using this approach, several nonmotile mutants were identified. The motility phenotype of one such mutant, designated BTF1, is shown in Fig. Fig.1.1. The location of the transposon insertion in strain BTF1 was determined by transposon-directed cloning (12) and revealed a transposon insertion (Fig. (Fig.2)2) in an open reading frame (ORF) that encodes a predicted protein with significant homology to FlhA, an essential component of the flagellar secretion apparatus (4). Immediately downstream we identified six ORFs encoding predicted flagellar structural proteins (FlgB and FlgC), flagellar secretion chaperones (FlgN and FlgA), the flagellar regulatory protein FlgM, and FlhE, a protein of unknown specific function (4, 13). Immediately upstream of FlhA an additional flagellar secretion protein, FlhB, was identified. These predicted proteins were greater than 80% similar and 60% identical to analogous proteins of Yersinia enterocolitica, Yersinia pestis, and Yersinia psuedotuberculosis. An additional ORF immediately upstream of FlhB was identified that encodes a protein with significant similarity (59% identical, 74% similar) to HreP, a virulence-associated protease present in Y. enterocolitica (10, 26).
The organization of the flhBAE-flgNMABC cluster in Y. ruckeri is identical to that of other Yersinia species (Fig. (Fig.2)2) with two noteworthy exceptions. First, hreP, present upstream of flhB in Y. ruckeri, Y. enterocolitica, and Y. frederiksenii, is absent at this position in the genomes of Y. pseudotuberculosis and Y. pestis KIM (Fig. (Fig.2)2) as well as in the genomes of Yersinia intermedia, Yersinia bercovieri, and Yersinia mollaretii (data not shown). Furthermore, hreP-like ORFs could not be found in the genome sequences of these other Yersinia species, indicating that this gene is likely unique to Y. ruckeri, Y. enterocolitica, and Y. frederiksenii. Additionally, the inv gene encoded between flhE and flgN in Y. enterocolitica and Y. pseudotuberculosis is absent at this position in Y. ruckeri. This gene encodes a virulence factor in Y. enterocolitica and Y. pseudotuberculosis that plays a critical role in intracellular invasion (11, 14, 17).
The flagellar export apparatus of Y. enterocolitica has previously been shown to function as a secretion system for the transport of several nonflagellar proteins, in addition to flagellar secretion targets, including the virulence-associated lipase YlpA (27). To determine whether lipase production by Y. ruckeri similarly requires an intact flagellar secretion apparatus, strain BTF1 and its parent were assessed for lipase production and secretion as previously described (21, 27). Lipase activity seen in the wild-type strain was absent in mutant strain BTF1 (Fig. 1B and C). Transcomplementation experiments were performed to verify that the Tn5 insertion in BTF1 was responsible for the lack of both motility and secreted lipase activity, as opposed to the result of one or more other mutations. Plasmids were created for this analysis by directly cloning PCR products using the pBAD TOPO TA kit (Invitrogen) and transferred to Y. ruckeri by electroporation (5). Plasmids containing either flhA or flhE alone failed to restore motility or lipase production, while a construct containing both flhA and flhE (pJE10) restored both activities (Fig. (Fig.11 and Fig. Fig.3A).3A). These data show that the flhA insertion in strain BTF1 exerts a downstream effect on flhE and that both of these genes are necessary for flagellar motility and lipase production. In related enteric bacteria, expression of flagellar secretion targets, such as flagellin, is contingent on production of the complete flagellar secretion apparatus (19). Western sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis using an antiflagellin monoclonal antibody, specific to a conserved epitope (8), was used to investigate the effect of the flhA::Tn5 insertion on flagellin production. A band of 40 to 45 kDa was detected in both wild-type CSF07-82 and in the complemented mutant but was absent in the mutant strain BTF1 (Fig. (Fig.3B).3B). These results indicate that mutational loss of Y. ruckeri flagellar secretion eliminates expression of flagellin.
Motility is clearly not required for virulence in BT2 Y. ruckeri, given that emergent BT2 strains are virulent; however, it is possible that loss of motility or lipase production could alter virulence in the BT1 background. To test this, the virulence of mutant strain BTF1 was compared to its isogenic wild-type parent by using a bath challenge model that mimics natural waterborne exposure. For each bacterial strain examined duplicate groups of 25 and 15 fish weighing approximately 7 or 10 g, respectively, were exposed to 4 × 107 cells/ml in a volume of water that was 10 times the total weight of the fish in an aerated container maintained at 15°C. After 1 h fish were transferred into aquaria supplied with 15°C flowing water, and dead fish were removed and recorded daily. Mortality due to Y. ruckeri infection was confirmed by microbiological analysis of kidney tissue on 20% of the mortalities/day. Bacterial cells for the challenges were grown for 72 h at 15°C in tryptic soy broth, and viable cell numbers were quantified using direct plate counts. Both the wild-type and complemented mutant strains were motile when grown under these conditions and therefore relevant to the infection model. The duplicated challenge groups were not significantly different, and therefore data from these experiments were pooled; the results are shown in Fig. Fig.4.4. The flhA::Tn5 mutant was not significantly altered in virulence compared to its isogenic wild-type parent or the complemented strain. Additional experiments utilizing intraperitoneal injection to initiate the challenge also showed no significant difference between the flhA::Tn5 mutant and the wild-type parent strain (data not shown). It is important to emphasize that bath challenge necessitates host colonization, and thus this result demonstrates that flagellar secretion is unnecessary for virulence in this organism, including early steps in the process of infection. Moreover, these results imply that motility and lipase production could be lost with no consequence to virulence in emergent BT2 strains.
Recent pulsed-field gel electrophoresis analysis suggests that the BT2 phenotype has emerged independently in the United Kingdom and Europe from indigenous motile serovar 1 isolates (25). Our results, demonstrating that the BT2 phenotype may be caused by mutational loss of the flagellar secretion apparatus in a BT1 strain, is consistent with this hypothesis, and together these observations imply that BT2 strains likely will have intact, albeit cryptic, flagellar secretion genes. Therefore, PCR was used to assess the presence of the flhBAE-flgNMABC gene cluster in representative BT2 strains from the United States and Europe. Primers flhBF and flgCR (Table (Table1)1) were used to amplify a 5,911-bp fragment using the Qiagen LongRange PCR kit for isolates of BT2 Y. ruckeri from the United States, United Kingdom, and Denmark. All three strains tested yielded PCR products identical to those of BT1 strain CSF07-82. DNA sequencing revealed that the flhBAE-flgNMABC gene cluster of these three BT2 strains was identical to that of BT1 strain CSF07-82. The discovery of this completely conserved motility gene cluster in BT2 Y. ruckeri isolates is a strong indication that they are recent variants of motile Y. ruckeri strains. This result also demonstrates that in these BT2 strains the lesion causing loss of motility and lipase production is elsewhere in the genome.
The results presented here reveal a potential molecular basis for the concurrent absence of motility and lipase secretion in BT2 Y. ruckeri through the natural mutational loss of flagellar secretion. Loss of flagellar motility is not uncommon in the evolution of human pathogens. For example, Shigella species, as well as all strains of Y. pestis, are nonmotile, despite the presence of cryptic flagellar motility genes (7, 16, 23). Our results demonstrate that loss of the flagellum and flagellar secretion does not affect Y. ruckeri virulence as seen using a natural infection model, suggesting that the loss of these phenotypes in BT2 Y. ruckeri strains has likely had no deleterious effect on virulence. Finally, the correlation between vaccine failure and the recent emergence and dissemination of BT2 Y. ruckeri indicates that flagellar secretion and/or the flagellum itself may be a liability in a vaccinated host.
DNA sequences have been deposited in the GenBank database under accession number GQ217534.
We thank Jennifer Harper and Travis Moorland (National Center for Cool and Cold Water Aquaculture) and Bill Shewmaker and Robin Burkhart (Clear Springs Foods, Inc.) for technical assistance and Peter Feng (U.S. Food and Drug Administration) for kindly providing antiflagellin monoclonal antibody.
Published ahead of print on 21 August 2009.