Search

The genus Yersinia comprises species that differ widely in their pathogenic potential and public-health significance. Yersinia pestis is responsible for plague, while Yersinia enterocolitica is a prominent enteropathogen. Strains within some species, including Y. enterocolitica, also vary in their pathogenic properties. Phenotypic identification of Yersinia species is time-consuming, labour-intensive and may lead to incorrect identifications. Here, we developed a method to automatically identify and subtype all Yersinia isolates from their genomic sequence. A phylogenetic analysis of Yersinia isolates based on a core subset of 500 shared genes clearly demarcated all existing Yersinia species and uncovered novel, yet undefined Yersinia taxa. An automated taxonomic assignment procedure was developed using species-specific thresholds based on core-genome multilocus sequence typing (cgMLST). The performance of this method was assessed on 1843 isolates prospectively collected by the French National Surveillance System and analysed in parallel using phenotypic reference methods, leading to nearly complete (1814; 98.4 %) agreement at species and infra-specific (biotype and serotype) levels. For 29 isolates, incorrect phenotypic assignments resulted from atypical biochemical characteristics or lack of phenotypic resolution. To provide an identification tool, a database of cgMLST profiles and reference taxonomic information has been made publicly accessible (https://bigsdb.pasteur.fr/yersinia). Genomic sequencing-based identification and subtyping of any Yersinia is a powerful and reliable novel approach to define the pathogenic potential of isolates of this medically important genus.

Yersinia pestis is the aetiologic agent of plague. Without appropriate treatment, the pathogen rapidly causes septicaemia, the terminal and fatal phase of the disease. In order to identify bacterial genes which are essential during septicaemic plague in humans, we performed a transcriptome analysis on the fully virulent Y. pestis CO92 strain grown in either decomplemented human plasma or Luria-Bertani medium, incubated at either 28 or 37 degrees C and harvested at either the mid-exponential or the stationary growth phase. Y. pestis genes involved in 12 iron-acquisition systems and one iron-storage system (bfr, bfd) were specifically induced in human plasma. Of these, the ybt and tonB genes (encoding the yersiniabactin siderophore virulence factor and the siderophore transporter, respectively) were induced at 37 degrees C, i.e. under conditions mimicking the mammalian environment. Growth in human plasma also upregulated genes involved in the synthesis of five fimbrial-like structures (including the Psa virulence factor), and in purine/pyrimidine metabolism (the nrd genes). Genes known to play a role in the virulence of several bacterial pathogens (such as those encoding the Lpp lipoprotein and non-iron metal-uptake proteins) were induced in human plasma, during either the exponential or the stationary phase. Finally, 120 genes encoding proteins of unknown function were upregulated in human plasma. Eleven of these genes were specifically transcribed at 37 degrees C and may thus represent new virulence factors that are important during the septicaemic phase of human plague.

Some adenine methyltransferases have been shown not only to protect specific DNA restriction sites from cleavage by a restriction endonuclease, but also to play a role in various bacterial processes and sometimes in bacterial virulence. This study focused on a type I restriction-modification system (designated yrmI) of Y. pseudotuberculosis. This system is composed of three adjacent genes which could potentially encode an N6-adenine DNA methylase (YamA), an enzyme involved in site-specific recognition (YrsA) and a restriction endonuclease (YreA). Screening of 85 isolates of Y. pestis and Y. pseudotuberculosis indicated that the yrmI system has been lost by Y. pestis and that yamA (but not yrsA or yreA) is present in all Y. pseudotuberculosis strains tested, suggesting that it may be important at some stages of the epidemiological cycle of this species. To further investigate the role of yamA in Y. pseudotuberculosis survival, multiplication or virulence, a DeltayamA mutant of Y. pseudotuberculosis IP32953 was constructed by allelic exchange with a kanamycin cassette. The fact that DeltayamA mutants were obtained indicated that this gene is not essential for Y. pseudotuberculosis viability. The IP32953DeltayamA mutant strain grew as well as the wild-type in a rich medium at both 28 degrees C and 37 degrees C. It also grew normally in a chemically defined medium at 28 degrees C, but exhibited a growth defect at 37 degrees C. In contrast to the Dam adenine methyltransferase, a mutation in yamA did not impair the functions of DNA repair or resistance to detergents. However, the DeltayamA mutant exhibited a virulence defect in a mouse model of intragastric infection. The in silico analysis indicated that the chromosomal region carrying the Y. pseudotuberculosis yrmI locus has been replaced in Y. pestis by a horizontally acquired region which potentially encodes another methyltransferase. YamA might thus be dispensable for Y. pestis growth and virulence because this species has acquired another gene fulfilling the same functions.

Yersinia pestis is a species that emerged recently from Yersinia pseudotuberculosis and gained an exceptional pathogenicity potential. Among the major genetic differences between the plague bacillus and its ancestor is the acquisition of the pPla plasmid, which has been associated with the increased virulence of Y. pestis. In a previous study, introduction of pPla into Y. pseudotuberculosis did not lead to any modification of the virulence of the host bacterium. However, it was subsequently demonstrated that the presence of smooth lipopolysaccharide (LPS) inhibits the activity of Pla. In this study, pPla was introduced into a Y. pseudotuberculosis strain expressing smooth LPS, and into a variant in which a mutation that abrogates the formation of O-antigen (O-Ag) repeats (as in natural isolates of Y. pestis) was generated. It was found that in both strains, Pla was synthesized, exported to the bacterial membrane and processed as in Y. pestis. However, the ability of Pla to activate plasminogen was weak and observed only at 37 degrees C in the smooth strain, while this activity was similar to that of Y. pestis and expressed at both 28 and 37 degrees C in the O-Ag mutant strain. Similarly, Pla-mediated inactivation of the antiprotease alpha2-antiplasmin was not detected in the smooth Y. pseudotuberculosis strain grown at 28 degrees C, but was expressed at both temperatures in the O-Ag mutant strain. Despite the more efficient activity of Pla, the Y. pseudotuberculosis O-Ag mutant strain exhibited a lower pathogenicity upon subcutaneous infection of mice. The results thus indicate that, although abrogation of O side chain synthesis in a Y. pseudotuberculosis strain harbouring pPla potentiates the two proteolytic activities of Pla, this is not sufficient to confer to Y. pseudotuberculosis a higher pathogenicity potential. These results also suggest that acquisition of pPla may not have been sufficient to confer an immediate higher pathogenic potential to the ancestor Y. pestis strain.

The efficiency of pulsed-field gel electrophoresis (PFGE), ribotyping and restriction enzyme analysis of the virulence plasmid (REAP) for typing and subtyping strains of Yersinia enterocolitica was compared. All three techniques gave concordant results, and the strains studied could be separated into three distinct clusters: (1) heterogeneous strains of biotype 1A and serotype O5 (1A/O5); (2) one 3/O3 strain and all 2/O9 strains; and (3) all 4/O3, 2/O5 and two 3/O3 strains. Within cluster 3, the 2/O5 and 3/O3 strains were related more closely to each other than to the 4/O3 isolates. With ribotyping, PFGE and phage-typing, the 4/O3 isolates were subdivided into two homogeneous groups, corresponding to strains of phage type IXb and strains of other phage types, respectively. Similarly, ribotyping and PFGE subdivided the 2/O9 strains into two conserved groups (I and II), but REAP gave a different subdivision and identified a new REAP pattern (P3). The three techniques confirmed the clear distinction between the heterogeneous group of non-pathogenic 1A/O5 strains and the well conserved group of pathogenic 2/O5 strains. Additional plasmids were identified in some 3/O3 strains. Combined, the results indicated that REAP (with EcoRI) and ribotyping (with EcoRV) are valuable alternatives to bioserotype determination, and PFGE is the most suitable technique for epidemiological tracing.

Forty-eight human isolates of Yersinia enterocolitica of biotype 4, serotype O3 from various parts of the world were examined for susceptibility to 13 beta-lactam agents. The intracellular beta-lactamases of each of the 48 strains were examined. Isolates from Europe, Asia and Brazil (phage type VIII) or South Africa and Hungary (phage type IXa) produced both enzyme A and enzyme B, whereas isolates from New Zealand and Australia (phage type IXb) lacked the cephalosporinase enzyme B. Among the seven strains isolated in Canada belonging to phage type IXb, three strains expressed only enzyme A (group I) whereas the other four strains produced both enzymes A and B (group II). The high susceptibility to the combination of amoxycillin and clavulanate observed in one subtype was explained by the absence of the cephalosporinase enzyme B. A simple disk diffusion test with this antibiotic combination was shown to be effective in the detection of enzyme B in Y. enterocolitica 4/O3.