We obtained lung tissue samples from bank voles collected by the territorial Сenters of Hygiene and Epidemiology in the Arkhangelsk, Ivanovo, Moscow, Penza, Saratov, Tyumen, Ulyanovsk, and Udmurtia regions (Figure 1). We stored samples, analyzed hantavirus antibodies by immunofluorescence assay (Table 1), extracted RNA, determined samples with high viral RNA content by real-time reverse transcription PCR (Table 1), and sequenced them as described in previous research (12).

To determine bank vole sublineages, we obtained sequences of the cytochrome b gene by using a previously described protocol (13). In all, we obtained PUUV genomic sequences and cytochrome b sequences from 17 bank voles (Appendix Table 1). All voles belonged to the Eastern lineage of C. glareolus (Appendix Figure 1).

We mapped locations of the newly obtained PUUV sequences and all representative LAT, FIN, and RUS lineage sequences available in GenBank that had complete S segment coding sequences (Figure 1). We also included representative sequences of other PUUV lineages on the map.

Figure 2

Figure 2. Phylogenetic trees of small (S), medium (M), and large (L) segments in a study of evolutionary formation and distribution of Puumala virus genome variants, Russia. A) S segment based on...

For phylogenetic analysis, we added sequences from this study to GenBank representatives of the LAT, FIN, and RUS lineages. We included representatives of the other PUUV lineages as outgroup (6) (Figure 2). According to S-segment phylogeny, we propose dividing the RUS lineage into 3 phylogenetically based sublineages: W-RUS (from western Russia) (12), Balt-RUS (from the Baltic coast region), and Volga-RUS (from the Volga River Valley). The percentages of identical nucleotides and amino acids within and between the sublineages did not enable us to establish clear criteria for the delimitation of the sublineages according to the quantitative differences (Table 2).

The topology of our S-segment tree does not match those from previous studies (12). The sequences we obtained provide additional information for better resolution of PUUV. L and S segments of PUUV are comparably conservative (Appendix Figure 2), and their phylogenetic trees mostly repeat each other. The M-segment tree was the most distinct (Figure 2, panel B), a finding that corresponds with previous data (5). That finding confirms that the M segment is the most variable and most often involved in reassortment events (14).

In the M-segment tree, the W-RUS sublineage stood within the Volga-RUS sublineage, but the branch from Bashkiria formed an outgroup (Figure 2, panel B). Almost the entire Volga-RUS sublineage, except for the Bashkir branch, appears to be a product of the exchange of the М segment between ancestors of the W-RUS and the Bashkir branch. The S and L segments are related to the Bashkir branch, and the M segment was obtained from the ancestors of the W-RUS sublineage. Those findings confirm and extend previous results (12). Thus, strains from the most epidemiologically active areas have arisen through reassortment. However, whether reassortment has altered the pathogenic properties of those strains remains unclear.

The genome variant from Penza belongs to the group that obtained the M segment by reassortment, as we have described (Figure 2, panel B). On the S-segment tree, Penza/Cg8730 is located on the Volga-RUS sublineage (Figure 2, panel A). However, on the L-segment tree, Penza/Cg8730 belongs to the W-RUS sublineage (Figure 2, panel C). Thus, Penza/Cg8730 likely emerged from a 2-stage reassortment. Although L segment exchange is not a unique event, it occurs much less frequently than exchange among other segments (14). Geographically, the Penza region lies west of locations where viruses belonging to the Volga-RUS sublineage were isolated but far from representatives of the W-RUS sublineage. Thus, Penza could be a meeting point of those 2 lineages.

On the S-segment tree, the FIN branch was represented by 2 sublineages: West-FIN from Finland and Karelia and East-FIN from western Siberia (Figure 2). However, on the M-segment tree, West-FIN and East-FIN did not form a common branch (Figure 2, panel B). The current hypothesis generally accepts that the FIN branch split into 2 parts when the territories were recolonized by bank voles after the glacier melted (8).

The sequences we obtained from the Arkhangelsk region clustered with sequences of viruses collected in Omsk and Tyumen regions, despite their geographic distance (Figure 1; Appendix Figure 3). According to the phylogenetic trees, the Tyumen and Omsk clades divided after the Arkhangelsk clade split from their common ancestor (Figure 2). Such branching suggests 2 possible paths for spread of the East-FIN subclade: from northwest Russia to the southeast, across the Ural Mountains to western Siberia; or spreading from the Pre-Ural region in 2 directions, to western Siberia and to the Arkhangelsk region. The second possibility supposes a second vole migration wave to the north, after PUUV split off from the East-FIN and West-FIN groups during postglacial period (8). PUUV sequences from the area between Arkhangelsk and western Siberia might shed more light on spread of the FIN lineage.

The newly identified viruses from the Tyumen and Arkhangelsk regions of Russia and viruses from Omsk form a common East-FIN branch. This finding raises additional questions about the dispersal routes of the FIN lineage.

Here we considered the RUS lineage as 3 separate sublineages: Volga-RUS, W-RUS, and Balt-RUS. The PUUV genome variants from Tatarstan, Udmurtia, Samara, and Saratov probably emerged through reassortment because they contained S and L segments related to the Bashkir branch and an M segment derived from ancestors of the W-RUS sublineage. The Penza/Cg8730 genome variant might have arisen from a 2-stage reassortment. Continued surveillance is needed to assess PUUV pathogenicity in Russia, but we found PUUV reassortments in areas with the highest HRFS incidence, indicating reassortment might contribute to pathogenic properties of PUUV.