In silico analysis of available bacterial genomes revealed the phylogenetic proximity levels of enzymes responsible for biosynthesis of lipopolysaccharide (LPS) of Yersinia pestis, the cause of plague, to homologous proteins of closely related Yersinia spp. and some other bacteria (Serratia proteamaculans, Erwinia carotovora, Burkholderia dolosa, Photorhabdus luminescens and others). Isogenic Y. pestis mutants with single or double mutations in 14 genes of LPS biosynthetic pathways were constructed by site-directed mutagenesis on the base of the virulent strain 231 and its attenuated derivative. Using high-resolution electrospray ionization mass spectrometry, the full LPS structures were elucidated in each mutant, and the sequence of monosaccharide transfers in the assembly of the LPS core was inferred. Truncation of the core decreased significantly the resistance of bacteria to normal human serum and polymyxin B, the latter probably as a result of a less efficient incorporation of 4-amino-4-deoxyarabinose into lipid A. Impairing of LPS biosynthesis resulted also in reduction of LPS-dependent enzymatic activities of plasminogen activator and elevation of LD(50) and average survival time in mice and guinea pigs infected with experimental plague. Unraveling correlations between biological properties of bacteria and particular LPS structures may help a better understanding of pathogenesis of plague and implication of appropriate genes as potential molecular targets for treatment of plague.

The LcrV protein is a multifunctional virulence factor and protective antigen of the plague bacterium and is generally conserved between the epidemic strains of Yersinia pestis. We investigated the diversity in the LcrV sequences among non-epidemic Y. pestis strains which have a limited virulence in selected animal models and for humans. Sequencing of lcrV genes from 19 Y. pestis strains belonging to different phylogenetic groups (“subspecies”) showed that the LcrV proteins possess four major variable hotspots at positions 18, 72, 273, and 324–326. These major variations, together with other minor substitutions in amino acid sequences, allowed us to classify the LcrV alleles into five sequence types (A-E). We observed that the strains of different Y. pestis “subspecies” can have the same type of LcrV, including that conserved in epidemic strains, and different types of LcrV can exist within the same natural plague focus. Therefore, the phenomenon of “selective virulence” characteristic of the strains of the microtus biovar is unlikely to be the result of polymorphism of the V antigen. The LcrV polymorphisms were structurally analyzed by comparing the modeled structures of LcrV from all available strains. All changes except one occurred either in flexible regions or on the surface of the protein, but local chemical properties (i.e. those of a hydrophobic, hydrophilic, amphipathic, or charged nature) were conserved across all of the strains. Polymorphisms in flexible and surface regions are likely subject to less selective pressure, and have a limited impact on the structure. In contrast, the substitution of tryptophan at position 113 with either glutamic acid or glycine likely has a serious influence on the regional structure of the protein, and these mutations might have an effect on the function of LcrV. The polymorphisms at positions 18, 72 and 273 were accountable for differences in the oligomerization of LcrV.

Two isogenic sets of Yersinia pestis strains were generated, composed of wild-type strains 231 and I-1996, their non-polar pH 6" mutants with deletions in the psaA gene that codes for its structural subunit or the whole operon, as well as strains with restored ability for temperature-and pH-dependent synthesis of adhesion pili or constitutive production of pH 6 antigen. The mutants were generated by site-directed mutagenesis of the psa operon and subsequent complementation in trans. It was shown that the loss of synthesis or constitutive production of pH 6 antigen did not influence Y. pestis virulence or the average survival time of subcutaneously inoculated BALB/c naïve mice or animals immunized with this antigen.

Sequencing of lcrV genes and comparison of the deduced amino acid sequences from ten Y. pestis strains belonging mostly to the group of atypical rhamnose-positive isolates (non-pestis subspecies or pestoides group) showed that the LcrV proteins analyzed could be classified into five sequence types. This classification was based on major amino acid polymorphisms among LcrV proteins in the four "hot points" of the protein sequences. Some additional minor polymorphisms were found throughout these sequence types. The "hot points" corresponded to amino acids 18 (Lys --> Asn), 72 (Lys --> Arg), 273 (Cys --> Ser), and 324-326 (Ser-Gly-Lys --> Arg) in the LcrV sequence of the reference Y. pestis strain CO92. One possible explanation for polymorphism in amino acid sequences of LcrV among different strains is that strain-specific variation resulted from adaptation of the plague pathogen to different rodent and lagomorph hosts.

It has been shown previously that several endemic Y. pestis isolates with limited virulence contained the I259 isoform of the outer membrane protease Pla, while the epidemic highly virulent strains possessed only the T259 Pla isoform. Our sequence analysis of the pla gene from 118 Y. pestis subsp. microtus strains revealed that the I259 isoform was present exclusively in the endemic strains providing a convictive evidence of more ancestral origin of this isoform. Analysis of the effects of the I259T polymorphism on the intrinsic disorder propensity of Pla revealed that the I259T mutation slightly increases the intrinsic disorder propensity of the C-terminal tail of Pla and makes this protein slightly more prone for disorder-based protein-protein interactions, suggesting that the T259 Pla could be functionally more active than the I259 Pla. This assumption was proven experimentally by assessing the coagulase and fibrinolytic activities of the two Pla isoforms in human plasma, as well as in a direct fluorometric assay with the Pla peptide substrate. The virulence testing of Pla-negative or expressing the I259 and T259 Pla isoforms Y. pestis subsp. microtus and subsp. pestis strains did not reveal any significant difference in LD50 values and dose-dependent survival assays between them by using a subcutaneous route of challenge of mice and guinea pigs or intradermal challenge of mice. However, a significant decrease in time-to-death was observed in animals infected with the epidemic T259 Pla-producing strains as compared to the parent Pla-negative variants. Survival curves of the endemic I259 Pla+ strains fit between them, but significant difference in mean time to death post infection between the Pla−strains and their I259 Pla+ variants could be seen only in the isogenic set of subsp. pestis strains. These findings suggest an essential role for the outer membrane protease Pla evolution in Y. pestis bubonic infection exacerbation that is necessary for intensification of epidemic process from endemic natural focality with sporadic cases in men to rapidly expanding epizootics followed by human epidemic outbreaks, local epidemics or even pandemics.