In 2015 and 2016, at least 10 abandoned weak juvenile red squirrels were submitted to a sanctuary near Berlin, Germany. The animals had exudative and erosive-to-ulcerative dermatitis with serocellular crusts at auricles, noses, digits, tails, and genital/perianal regions. Skin specimens from affected animals were investigated by electron microscopy (EM) and PCR. Three animals that died under care were submitted for necropsy. We obtained samples of all organs for histological and PCR examination. We used 1 sample of a skin lesion for virus propagation in cell culture.

Figure 1

Figure 1. Ultrastructure of Berlin squirrelpox virus particles from skin lesions on a red squirrel in Berlin, Germany, visualized by negative staining (uranyl acetate) (original magnification ×68,000).

EM-negative staining of skin lesions from all animals led to the discovery of brick-shaped poxvirus particles with irregular threadlike surface fibers and an average size of 294 nm × 221 nm (Figure 1). Pathological findings of corresponding skin lesions were consistent with poxvirus infection (ballooning degeneration of epidermal keratinocytes, numerous intracytoplasmic inclusion bodies, epidermal ulceration with suppurative inflammation, and secondary bacterial infection). All inner organs had either no pathological changes or lesions unrelated to poxvirus infection.

To confirm the morphologic diagnosis, we extracted DNA from skin lesions and performed various PCRs. An orthopoxvirus (OPV)–specific PCR showed negative results (6); a parapoxvirus (PPV)–specific PCR (6), a leporipoxvirus-specific PCR (A. Nitsche and L. Schrick, unpub. data), and a poxvirus-screening PCR (7) were positive for some samples. Obtained sequence fragments indicated poxviral relatedness but did not allow for the assignment to a poxvirus genus. Thus, we performed massively parallel sequencing. We directly subjected DNA extracted from a skin lesion on the foot of a dead animal to Nextera XT Library preparation and sequenced it on an Illumina HiSeq 1500 instrument (Illumina, San Diego, CA, USA), yielding 7,242,301 paired-end reads (150 + 150 bases, rapid run mode). Mapping (8) the obtained reads to all poxvirus reference sequences available in GenBank in high-sensitivity mode provided no notable results, which pointed to a virus with a highly deviant genome. Therefore, we separated poxviral reads from background data using RAMBO-K version 1.2 (9) and assembled the resulting 1,520,811 reads (10), yielding 1 single contig of 142,974 bp with ≈460-fold coverage after manual iterative mapping and scaffolding. We confirmed the genomic sequence by resequencing (Illumina MiSeq) of a Vero E6 cell-culture isolate obtained from a different skin specimen of the same animal. We named the new virus Berlin SQPV (BerSQPV), and uploaded the combined sequence information to GenBank (accession no. MF503315). Direct sequencing of DNA from skin samples of three other animals from the same origin yielded sequences with >99.9% identity to BerSQPV.

Figure 2

Figure 2. Phylogenetic position of BerSQPV (bold) from a red squirrel in Berlin, Germany, within the Chordopoxvirinae. We used MAFFT (12) to perform multiple alignments of all complete genome sequences within...

We compared characteristics of BerSQPV to related viruses and found that the EM structure shows features typical for OPV but the genome size of ≈143 kb is more consistent with PPV or SQPV from the United Kingdom (11) than with the large genome of OPV, whereas the guanine-cytosine (GC) content of 38.5% is more consistent with OPV and leporipoxvirus than with PPV and SQPV from the United Kingdom. Therefore, we explored the genomic relationship of BerSQPV to other chordopoxviruses. Pairwise alignments of each of the chordopoxvirus genomes available in GenBank with the BerSQPV genome resulted in a pairwise identity of at most 47% to tanapox virus isolate TPV-Kenya (accession no. EF420156.1). The retrieved phylogenetic tree (Figure 2) demonstrates that BerSQPV cannot be assigned to any of the known poxvirus genera; moreover, it does not cluster with the only other squirrel poxvirus with a published genome sequence (11). Further phylogenetic analyses based on conserved single genes frequently used for poxvirus tree calculations (A3L, F10L+F12L, F13L, E13L, E9L [VACV Copenhagen nomenclature]) showed similar results (A56R was not used for tree calculations because this open reading frame is too divergent among the Chordopoxvirinae), with BerSQPV forming a unique branch (data not shown). In addition, any partial sequences of SQPV available in GenBank were aligned to BerSQPV, showing a maximal sequence identity of 64.3% to gene E9L (GenBank accession no. AY340976.1), further emphasizing the uniqueness of this newly identified virus.

We designed a BerSQPV-specific quantitative PCR based on the genome sequence as a tool for future investigations (primer BerSQPV_F: ggAAgTTTTCCCATACCAACTgA, primer BerSQPV_R: ATCTCAAACCgCAgACggTA, probe BerSQPV_TM: FAM-ACTgTTATTCTTAgCgTAATT). Sensitivity was <10 genome equivalents per reaction amplifying plasmid dilution rows. We first validated the specificity in silico during the design process, revealing the highest identity of 88% to cowpox virus Kostroma (GenBank accession no. KY369926.1), with mismatches in crucial positions in the primer and probe binding sites. Squirrel poxvirus strain Red squirrel UK (GenBank accession no. NC_022563.1) showed only 84% identity, with additional mismatches in amplification-relevant positions. Practical PCR testing using DNA from cowpox, monkeypox, ectromelia, parapox-ORF, myxoma, avipox, and molluscipox viruses showed no cross-reactivity.

The new specific quantitative PCR was subsequently applied to DNA from skin lesions archived from 1 squirrel found dead in 2014 in the Berlin area, 2 live squirrels from 2015, and 5 live squirrels from 2016, as well as various organs from 3 affected squirrels necropsied in 2015 (Table). Organ tissues yielded high BerSQPV DNA loads in the affected skin but low viral DNA loads for inner organs, findings in concordance with pathological findings, indicating the detection of viral DNA in the blood homogenously distributed throughout the organs with specific tropism for the skin. Low virus loads in inner organs are usually observed in poxvirus infections that do not generalize. PCR results indicate that this virus has been circulating in the Berlin area over the past 10 years.

We describe a new poxvirus, BerSQPV, isolated from red squirrels in Berlin, Germany, that causes pathological changes consistent with other epidermal poxvirus infections. Genome analysis revealed a unique sequence within the poxvirus family, as BerSQPV is not clustering to other poxvirus genera, including UK SQPV from red squirrels from Great Britain. In contrast to UK SQPV, which resembles PPV ultrastructurally (2), the ultrastructure of BerSQPV is comparable to that of OPV. Two other poxviruses from tree squirrels with ultrastructural appearance similar to BerSQPV have been reported: a Eurasian red squirrel from Spain with epidermal poxvirus lesions (5) and American red squirrels from Canada (15). Although no sequence information is available for the SQPV from Spain, the partial sequence analysis of SQPV from Canada showed the virus to also be distinct from all known mammalian poxviruses but most closely related to PPV, followed by UK SQPV (15).

BerSQPV is suspected to have been circulating for several years among Eurasian red squirrels in the greater Berlin area. Although diseased animals in care were handled in close contact, caretakers have remained asymptomatic, suggesting a negligible risk for human infection. Further detailed characterization of the isolated virus is ongoing.