Yersinia pestis lineages in Mongolia

Abstract

Background: Whole genome sequencing allowed the development of a number of high resolution sequence based typing tools for Yersinia (Y.) pestis. The application of these methods on isolates from most known foci worldwide and in particular from China and the Former Soviet Union has dramatically improved our understanding of the population structure of this species. In the current view, Y. pestis including the non or moderate human pathogen Y. pestis subspecies microtus emerged from Yersinia pseudotuberculosis about 2,600 to 28,600 years ago in central Asia. The majority of central Asia natural foci have been investigated. However these investigations included only few strains from Mongolia.

Methodology/principal findings: Clustered Regularly Interspaced Short Prokaryotic Repeats (CRISPR) analysis and Multiple-locus variable number of tandem repeats (VNTR) analysis (MLVA) with 25 loci was performed on 100 Y. pestis strains, isolated from 37 sampling areas in Mongolia. The resulting data were compared with previously published data from more than 500 plague strains, 130 of which had also been previously genotyped by single nucleotide polymorphism (SNP) analysis. The comparison revealed six main clusters including the three microtus biovars Ulegeica, Altaica, and Xilingolensis. The largest cluster comprises 78 isolates, with unique and new genotypes seen so far in Mongolia only. Typing of selected isolates by key SNPs was used to robustly assign the corresponding clusters to previously defined SNP branches.

Conclusions/significance: We show that Mongolia hosts the most recent microtus clade (Ulegeica). Interestingly no representatives of the ancestral Y. pestis subspecies pestis nodes previously identified in North-western China were identified in this study. This observation suggests that the subsequent evolution steps within Y. pestis pestis did not occur in Mongolia. Rather, Mongolia was most likely re-colonized by more recent clades coming back from China contemporary of the black death pandemic, or more recently in the past 600 years.

Figure 1. MLVA clustering and SNP branch assignment of 66 previously published Y. pestis microtus and pestis 0, 1 and 3 branches.

Microtus and strains from the 0 and 1 branches so far investigated by MLVA25 and by SNP analysis are shown , . Three Ulegeica, two Hissarica and nine Altaica strains not investigated by SNP analysis are also included. For completion, Table 1 gives further information about assignment of biovar, genotype, and origin. Colors reflect MLVA clustering as suggested by Li et al. . The SNP branch assignment of each strain as defined by Morelli et al. is indicated (column Morelli2010) together with the strain ID and biovar designation . The results of CRISPR analysis according to Cui et al. are shown in column group . Bootstrap support values are indicated. The figure shows the satisfying terminal branches clustering achieved by MLVA but the sometimes incorrect and usually low bootstrap values of deep branching nodes illustrating the complementarity of the two methods.

Figure 2. MLVA clustering and SNP branch assignment of 68 previously published Y. pestis pestis branches 1 and 2.

Sixty-eight strains from the 1 and 2 branches previously investigated by both MLVA25 and SNP analysis are displayed , . For completion, Table 1 gives further information about assignment of biovar, genotype, and origin. Colors reflect MLVA clustering as suggested by Li et al. . The SNP branch assignment of each strain as defined by Morelli et al. is indicated (column Morelli2010) together with the strain ID and biovar designation . Bootstrap support values are indicated for each node. The results of CRISPR analysis according to Cui et al. are given in column group . * This strain shows a Medievalis phenotype due to a different mutation in the napA gene compared to the mutation causing the Medievalis phenotype in the Medievalis biovar, as demonstrated by Pourcel et al. .

Figure 3. Minimal spanning tree of the strains as shown in Figures 1 and 2 using the same color code.

The figure is based on the same data set as Fig. 1 and 2. Table 1 gives further information about assignment of biovar, genotype, and origin. Basic correlation and grouping of genotypes is similar compared to previously published Fig. 2 in Morelli et al. .

Figure 4. Sampling sites in Mongolia and observed genotypes.

The map of Mongolia shows the sampling sites 1 to 35 and associated clusters (in color). Sizes of circles correlate to the number of collected strains. The exact geographic position of plague foci, and further background data of strains is listed in Table S1. Genotypes are explained in Table 1. Some previously published natural plague foci are shown in ovals. Colors match the corresponding lineage found in Mongolia: Mountain-Altai, Russia (#36), a reservoir for 0.PE1/Altaica strains. Tuva (Mongun-Taigin), Russia (#37), populated with 3.ANT genotype strains. Khentii province (KP), Mongolia , associated with the 2.ANT3 lineage. Bayanölgie province (BP), Mongolia associated with the Ulegeica biovar. Chinese natural plague foci are present in Ningxia, Hebei, Shanxi, and Inner Mongolia, (I), Inner Mongolia, Jilin, Heilongjiang (H), and Inner Mongolia (L). Different lineages have been isolated here as shown by Li et al. .

Figure 5. MLVA25 assignment of four clusters of the investigated Mongolian Y. pestis strains.

MLVA25 tree of 16 investigated Mongolian Y. pestis strains (marked with color and boxes) representing four of the 6 clusters, and various Y. pestis strains originating from microtus and pestis biovars. For each strain, the tentative SNP branch or node according to Morelli et al. as deduced by the presence of a linking strain in the same MLVA cluster is indicated by a question mark. Strain name, CRISPR profile as investigated in this study, and the sampling site (Focus) are listed.

Figure 6. MLVA25 assignment of two clusters of the Mongolian Y. pestis strains.

MLVA25 tree of two clusters comprising 84 investigated Mongolian Y. pestis strains (marked with color and boxes) compared to various previously typed Y. pestis strains. For each strain, the tentative SNP branch or node according to Morelli et al. as deduced by the presence of a linking strain in the same MLVA cluster is indicated by a question mark. Strain name, CRISPR profile as investigated in this study, and the sampling site (Focus) are listed.

Figure 7. MST tree of the main Mongolian clade.

MST tree of the 78 Mongolian Y. pestis strains, determined as 3.ANT genotype and showing CRISPR spacer b48 (blue). The seven strains previously described by Li et al. and associated to the Mongolian cluster were included (yellow) , suggesting distinct clustering. Numbers refer to the sampling sites given in Fig. 4.