During 2002 through 2008, a total of 366 Apodemus mice were trapped at 7 different sites in Germany (Figure 1). Serologic screening of transudates collected from these rodents by using an in-house DOBV immunoglobulin (Ig) G-ELISA, with a yeast-expressed nucleocapsid protein of DOBV-Af as antigen, identified 16 reactive and 5 equivocal samples of 114 A. agrarius trapped at 7 trapping sites in 3 federal states of Germany (Figure 1; Technical Appendix). Additionally, of 237 A. flavicollis mice, 1 equivocal sample and 4 DOBV-reactive samples were detected at 4 trapping sites (Figure 1; Technical Appendix). In contrast, of 15 wood mice (A. sylvaticus) originating from 3 trapping sites, none were found to be DOBV-seroreactive. A subsequent focus-reduction neutralization test showed a higher endpoint titer with DOBV-Aa than DOBV-Af (Technical Appendix) for 6 of the 8 investigated transudates independently, whether originating from A. agrarius or A. flavicollis.

An initial screening by a large (L) segment–specific nested reverse transcription–PCR (RT-PCR) (3) of 67 lung samples, representing all seroreactive (n = 20) and equivocal (n = 6) as well as 36 selected seronegative and 5 serologically not-analyzed animals, showed a 390-nt amplification product for 21 samples representing 16 seroreactive, 4 seronegative, and 1 serologically not-investigated animals (Technical Appendix). To enable a comparison with the only available DOBV sequence from Germany (H169), an S segment portion of 559 nt was amplified by RT-PCR from 11 lung tissues (Technical Appendix). In the phylogenetic analyses, all sequences from Germany formed 1 well-supported (PUZZLE [www.tree-puzzle.de]) and bootstrap support values >90%) monophyletic group consisting of 2 clusters. The first cluster contained S segment sequences from district Güstrow (trapping sites Pe1 and Pe3), Lüneburg (trapping site WG), Nordvorpommern (trapping site H), and the previously published DOBV sequence from an HFRS patient from northern Germany (H169; [8]; Figure 2, panel A). A second cluster was formed by S segment sequences originating from districts Ostprignitz-Ruppin (trapping sites Ka, To) and Demmin (trapping site K/A1). Notably, the A. flavicollis–derived sequences from sites Pe1, H, Ka, and K/A1 clustered together or were completely identical with A. agrarius–derived sequences from the same or neighboring trapping sites, suggesting multiple spillover infections (Figure 2, panel A, Technical Appendix Table 1).

The sequences from Germany share a common ancestor with the DOBV-Aa sequences originating from Slovakia and Russia. Together, they form a monophyletic group (DOBV-Aa lineage) that is clearly separated from A. flavicollis–borne (DOBV-Af) and A. ponticus–borne (DOBV-Ap) sequences and from A. agrarius–borne Saaremaa virus sequences. Subsequent analysis of nucleotide sequences of the entire nucleocapsid (N) protein– and glycoprotein precursor (GPC)–encoding regions confirmed these findings (Figure 2, panel B; Technical Appendix). A pairwise comparison between nucleotide and amino acid sequences of the complete N and GPC open reading frames from the novel German DOBV strains showed divergences of 1.5%–8.8% (0.3%–1.4%) and 2.1%–8.3% (0.8%–1.8%), respectively (Technical Appendix Table 2). The highest identity values on the nucleotide and amino acid sequence level (91.2%–91.7% and 99%–99.7%) were found for an S segment sequence from Denmark (Lolland/1403; GenBank accession no. AJ616854; Technical Appendix). The nucleotide and amino acid sequence divergence to other DOBV sequences was much higher, reaching 10.1%–14.3% (1%–3.3%) and 12.6%–20.7% (2.9%–9.4%), respectively.

Morphologic species determination for all DOBV-seroreactive and RT-PCR–positive rodents was confirmed by a mitochondrial cytochrome b gene-specific PCR (9,10), sequence determination, and comparison with available GenBank sequences from A. agrarius and A. flavicollis (Technical Appendix Table 1).

Based on a large panel of the entire N- and GPC-encoding DOBV sequences, we report direct molecular evidence that DOBV in Germany is represented by a genetic lineage associated with A. agrarius (DOBV-Aa). In contrast, we found no evidence for the occurrence of DOBV-Af in A. flavicollis or other Apodemus species from Germany. Consistent with the geographic distribution of A. agrarius (11) and the report of human DOBV disease exclusively in northern and northeastern Germany, this finding may confirm DOBV-Aa as the sole causative agent of DOBV infections in Germany (4; Robert Koch-Institut, SurvStat, www.rki.de).

Previously A. agrarius–associated Saaremaa virus was experimentally shown to be able to infect A. agrarius and A. flavicollis mice (12). We report multiple natural spillover infections of A. flavicollis by a DOBV strain originally hosted by A. agrarius. The observed spillover infections represent a crucial prerequisite for genetic reassortment. This observation is in contrast to other reports from Slovenia and Slovakia where, although A. agrarius and A. flavicollis are occurring sympatrically, A. flavicollis is exclusively carrying the DOBV-Af and A. agrarius the DOBV-Aa lineage (4,13). In contrast to our observations, single DOBV-Af spillover infections of A. sylvaticus and Mus musculus have been reported previously (14).

The phylogenetic analyses demonstrated 2 well-separated clusters within the DOBV-Aa lineage. These rodent-derived DOBV sequences in Germany represent a major contribution to the DOBV genomics and phylogenetics. Future investigations should help to identify specific features of these DOBV-Aa strains resulting in its frequent spillover to A. flavicollis and to prove a putative adaptation of DOBV-Aa on A. flavicollis after spillover, as well as possible reassortment processes.