Search for polyoma-, herpes-, and bornaviruses in squirrels of the family Sciuridae

Squirrels (n = 126) of five species belonging to the family Sciuridae were collected in four different countries (Canada, USA, Italy and UK), dissected according to standard protocols and brain, lung and spleen samples of these animals were available for screening for all three viruses. In addition, lung and spleen samples from another 112 squirrels belonging to another three species were available from previous bornavirus studies [17, 21, 22]. Thus, a total of 361 organ samples (243 spleen samples and 118 lung samples) were available for PyV and HV analyses (Tables 1, 2 and 3).

RNA extraction was performed using the KingFisher™ Flex Purification System (Thermo Fisher) in combination with the Nucleo Mag Vet Kit (Macherey Nagel) following the instructions of the manufacturer. DNA extraction was performed using EURx GeneMatrix Tissue DNA Purification Kit (Roboklon), Qiagen DNeasy Blood & Tissue-Kit (Qiagen) and the KingFisher™ Flex Purification System (Thermo Fisher) in combination with the Nucleo Mag Vet Kit (Macherey Nagel). DNA preparations were stored at − 20 °C and RNA preparations at − 80 °C. During the nucleic extraction processes negative controls were carried along to monitor for potential contaminations. Morphological species identification was confirmed for all samples by cytochrome b PCR and sequencing according to a previously described protocol ( [50], data not shown). Prior to nucleic acid extraction no attempts concerning virus isolation in cell culture were done.

Brain samples of the newly dissected 126 squirrels were screened with broad-range and VSBV-1 specific RT-qPCRs (Additional file 1) as described previously [17, 28]. The other animals (n = 112) have already been investigated before [17, 21, 22].

For identification of PyVs and HVs generic nested PCRs were performed, that broadly detect a partial VP1 coding sequence (CDS) of PyVs (Additional files 1 and 2) or a fragment of the DPOL gene of HVs (Open reading frame 09 (ORF09) of GHV; ORF UL54 of BHV) (Additional file 1 and Additional file 3) in the second PCR round. Both nested PCRs were performed as carried out previously [46, 48].

For amplification of gB sequences of BHVs (ORF UL55) and GHVs (ORF08), we used subfamily-specific nested primer sets (Additional file 1) essentially as described previously [51, 52]. The scheme of multi-level PCR analysis used for squirrel BHVs and GHVs is shown in Additional file 3.

For all PyVs, specific nested primers (Additional file 1) were selected tail-to-tail from the sequences amplified with generic PCR. They were used for the amplification and sequencing of the remaining parts of the circular genomes (approximately 5 kbp). For all HVs, for which both gB and DPOL sequences could be amplified, nested primers (Additional file 1) were selected that were specific for each virus and used in long-distance PCR (LD-PCR) for the amplification and sequencing of the sequence that spans the gap between the partial gB and DPOL sequences. LD-PCR was performed with the TaKaRa-Ex PCR system (Takara Bio Inc.), according to the manufacturer’s instructions. After the first PCR round, a 2 μl aliquot of the reaction mix was used as template in the second-round reaction.

In cases where only the generic HV PCR was successful (and not gB PCR and/or LD-PCR), the short DPOL sequence was extended in upstream direction with hemi-nested PCR, using the outer sense primer (285 DFA) of the generic HV PCR and two virus-specific antisense primers for amplification of approximately 480 base pairs (bp) (Additional file 1). The hemi-nested DPOL PCR was carried out as described above for generic HV PCR.

Based on these extended sequences we were able to design again virus-specific primers that were used for re-screening of all samples of the respective species.

For all novel PyVs, for which full genomes could be assembled, nested primer sets (Additional file 1) were selected. They were used for amplification of sequences of approximately 800 bp that encompass the short overlap between the sequences generated from the generic PCR fragments and the LD-PCR fragments of the respective PyV genome. PCR was performed in a total volume of 25 μl with 0.4 μl (2 units) Applied Biosystems AmpliTaq Gold DNA Polymerase (Thermo Fisher), 25 pmol of each primer, 200 μM dideoxynucleoside triphosphates (dNTPs), 2 mM MgCl₂, 5% dimethyl sulfoxide (DMSO) and 250 ng of sample DNA as template. A thermocycler from Biometra was used under the following cycling conditions: activation of the polymerase at 95 °C for 12 min and 45 cycles of denaturation at 95 °C for 30 s, annealing at 61 °C for 30 s, and elongation at 72 °C for 5 min, followed by a final extension step at 72 °C for 30 min. For the second PCR round, a 1 μl aliquot of the first-round reaction mix was used as a template. The primers were also used for more sensitive re-screening all samples of the respective host species.

For all novel HVs, for which extended sequences could be determined, nested primer sets were selected for re-screening all samples of the respective host species (Additional file 1). They were used as described above for specific PyV amplification (annealing temperatures listed in Additional file 1).

For each round of all nested PCRs and the RT-qPCR of our screening, no-template controls (PCR-grade H₂O) and extraction controls were carried along to detect any possible contaminations. These analyses were negative for all PCRs and RT-qPCRs. As positive PCR controls, DNA extracts of samples were used that tested positive for BHVs, GHVs or PyVs in previous studies.

All PCR products were purified with MSB® Spin PCRapace (Stratec), according to the manufacturer’s instruction and directly sequenced using the BigDye Terminator v3.1 system (Life Technologies) on an Applied Biosystems 3500xL DX Genetic Analyzer (Thermo Fisher). The LD-PCR products were sequenced by a classical primer walking strategy (primers not listed).

Early or late region plus flanking sequences (approximately 3.1 kbp) of Sciurus carolinensis polyomavirus 1 (ScarPyV1; GenBank accession number MK671101) were commercially synthesized and delivered as recombinant plasmids (Biomatik). They were named pScarPyV1early and pScarPyV1late and transformed into competent Escherichia coli DH5 alpha cells (Thermo Fisher). Plasmid DNA was extracted with Invisorb® Spin Plasmid Mini Two (Stratec) according to the manufacturer’s instruction.

Vero cells C1008 (monkey kidney cells; European Collection of Authenticated Cell Cultures (ECACC) # 85020206) were cultured in standard high-glucose Dulbecco’s minimal Eagle medium (DMEM, Thermo Fisher) containing 10% fetal calf serum (FCS) (PAN Biotech) and 1% penicillin/streptomycin (Thermo Fisher). For NIH/3T3 cells (mouse embryo fibroblast cells; American Tissue Culture Collection, ATCC®, CRL-1658™) the same medium was used, except that 5% FCS (PAN Biotech) was added. Both cell lines were cultivated at 37 °C and 5% CO₂. DNA extracts of cell aliquots were tested with PCR for absence of mycoplasma contamination [53]. Primers are listed in Additional file 1.

As described previously [54], cells were transfected with 1 μg DNA of plasmid pScarPyV1early or plasmid pScarPyV1late, using X-tremeGENE HP DNA Transfection Reagent (Roche Applied Biosciences). Before transfection, Vero cells were seeded in a volume of 500 μl cell culture medium in tissue culture plates with 24 wells (Sarstedt). For NIH/3T3 cells, Cell+ plates with 24 wells and a special Cell+ growth surface for sensitive adherent cells (Sarstedt) were used. Transfection procedures were performed 24 h after seeding according to the manufacturer’s instructions.

Total RNA was isolated on 0, 1, 2.5, and 6 days post transfection of recombinant plasmid DNA using the NucleoSpin® RNA-Kit (Macherey Nagel) according to the manufacturer’s instructions. DNA was removed by an additional Turbo DNA-free DNase treatment (Thermo Fisher). RNA concentrations were determined with the NanoDrop 8000 device (Thermo Fisher) at 260 nm. Synthesis of cDNA was carried out with 500 ng RNA using SuperScript II Reverse Transcriptase (Thermo Fisher) and oligo(dT)16 primers (Roche Applied Bioscience).

To identify introns in the early and late region of ScarPyV1, PCR was performed with cDNA as a template, using primers specific for ScarPyV1 that bind in flanking regions of predicted introns (Additional file 1). PCR was performed with 2.5 μl cDNA in a total volume of 25 μl with 0.2 μl (1 unit) Applied Biosystems AmpliTaq Gold DNA Polymerase (Thermo Fisher), 1 μM of each primer, 200 μM dNTP PCR Mix (Metabion), 2 mM MgCl₂ and 5% DMSO (Merck). All PCR products were purified with MSB® Spin PCRapace (Stratec), according to the manufacturer’s instructions. Sequencing reactions were performed as described above.

For the datasets we selected reference viral genomes representing all currently recognized species in the family/subfamily considered as well as additional viruses whose genomes represented distinct viral lineages discussed in the literature but still not integrated into the official taxonomy (sensu International Committee on Taxonomy of Viruses (ICTV)). For PyVs, this represented 109 viruses; for BHVs, 21 viruses; and for GHVs, 39 viruses. We extracted the LTAg and VP1 (PyV) or DPOL and gB (BHV and GHV) coding sequences from these genomes as well as from the novel viruses we identified in squirrels using Geneious v11.1.5 [55]. For each coding sequence, sequences were translated into amino acid sequences and aligned using Muscle [56] as implemented in Seaview v4 [57]. Conserved amino acid blocks were then selected using Gblocks as implemented in Seaview, using options for a less stringent selection: allow smaller final blocks, allow gap positions within the final blocks and allow less strict flanking positions [58]. The final amino acid sequence alignments comprised 260 (VP1), 517 (LTAg), 628 (DPOL BHV), 298 (gB BHV), 622 (DPOL GHV) and 282 (gB GHV) positions, respectively.

For phylogenetic analyses of all datasets we first ran Maximum-Likelihood (ML) analyses using PhyML v3 with smart model selection (PhyML-SMS) using the Bayesian information criterion and a tree search using subtree pruning and regrafting [57, 59, 60]. Branch robustness was estimated using Shimodaira-Hasegawa-like approximate likelihood ratio tests (SH-like aLRT) [61]. The PyV and GHV ML trees were rooted with TempEst v1.5 by minimizing the variance of root-to-tip distances [62]; the BHV ML trees were rooted using roseolovirus outgroups. We then ran Bayesian Markov chain Monte Carlo (BMCMC) runs using BEAST v1.10.4 [63]. For each alignment, we used the amino acid substitution model identified by PhyML-SMS, an uncorrelated relaxed clock (lognormal) model and a speciation model (birth-death) as a tree prior. The output of multiple BMCMC runs was examined for convergence and appropriate sampling of the posterior using Tracer v1.7.1 [64], before being merged using LogCombiner v1.10.4 (distributed with BEAST). The maximum clade credibility (MCC) tree was identified from the posterior set of trees (PST) and annotated with TreeAnnotator v1.10.4 (also distributed with BEAST). Branch robustness was estimated based on their posterior probability in the PST.

In previous studies, brain and oral swab samples of 112 squirrels belonging to various species were investigated with a broad-range orthobornavirus RT-qPCR and VSBV-1-specific RT-qPCR. VSBV-1-RNA was detected in brain and/or oral swab samples of 7/7 individuals from holdings of variegated squirrels, in samples of 10/17 Prevost's squirrels and a few other species [17, 21, 22], but not in any brain samples of the 77 tested Eurasian red squirrels and in any of the 11 Eastern grey squirrels [17, 21, 22]. Here we investigated a total of 126 brain samples from wildlife-derived squirrels from four countries, but did not find any positive specimens (Table 1).

To identify PyVs and HVs in squirrels, we analyzed 238 squirrels (spleen, n = 208; lung, n = 118) of eight different species (Tables 2 and 3). First we screened with generic nested PCR that targets a short fragment of the major capsid VP1 CDS of mammalian PyVs (Additional file 1 and Additional file 2) [43, 47‐49] and generic nested PCR with specificity for the highly conserved DPOL gene of mammalian HVs (Additional file 1 and Additional file 3) [44‐46]. Products of expected length were sequenced and sequences analyzed using Nucleotide Basic Local Alignment Search Tool (BLASTn). Thirteen samples of twelve animals were positive for yet unknown PyVs (Table 2) and 104 samples of 89 animals were positive for yet unknown HVs (Table 3) as revealed by BLASTn analysis (data not shown). In total, four novel PyVs, and 10 novel HVs (four BHVs and six GHVs) were identified and tentatively named: Sciurus carolinensis polyomavirus 1 (ScarPyV1), Sciurus carolinensis betaherpesvirus 1 (ScarBHV1), Sciurus carolinensis gammaherpesvirus 1 and 2 (ScarGHV1 and ScarGHV2), Sciurus variegatoides polyomavirus 1 (SvarPyV1), Sciurus vulgaris betaherpesvirus 1 (SvulBHV1), Callosciurus erythraeus polyomavirus 1 (CeryPyV1), Callosciurus erythraeus betaherpesvirus 1 (CeryBHV1), Callosciurus erythraeus gammaherpesvirus 1 (CeryGHV1), Callosciurus prevostii polyomavirus 1 (CprePyV1), Callosciurus prevostii betaherpesvirus 1 (CpreBHV1), Callosciurus prevostii gammaherpesvirus 1 (CpreGHV1), Urocitellus richardsonii gammaherpesvirus 1 (UricGHV1) and Tamias striatus gammaherpesvirus 1 (TstrGHV1).

In PyV detection with generic PCR, one of 118 spleen samples and one of 88 lung samples from Eastern grey squirrels (Sciurus carolinensis) were positive for ScarPyV1, 1/7 spleen samples of variegated squirrels (Sciurus variegatoides) was positive for SvarPyV1, 8/44 tested samples (6x spleen and 1x spleen and lung) from 35 Pallas’s squirrels (Callosciurus erythraeus) originating from Italy were positive for CeryPyV1, and 3/17 spleen samples of Prevost’s squirrels (Callosciurus prevostii) were positive for CprePyV1 (Table 2).

In HV tests of 118 spleen and lung samples of Sciurus carolinensis from the UK, Italy and USA with generic PCR, we identified 20 animals from the UK as ScarGHV1-positive (16x in spleen and 4x in spleen and lung), three Eastern grey squirrels from the USA as ScarGHV1-positive in lung and two individuals as ScarGHV2-positive, one from the UK in spleen and one from the USA in lung. ScarBHV1 was detected in lung samples of two animals from Italy. Testing 77 Sciurus vulgaris revealed 18 SvulBHV1-positive individuals from Germany (1x in spleen, 12x in lung and 5x in spleen and lung) and 13 from the UK (1x in spleen, 9x in lung and 3x in spleen and lung). We identified 10 out of 11 Richardson’s ground squirrels (Urocitellus richardsonii) to be positive for UricGHV-1 in spleen and 2/2 Eastern chipmunks (Tamias striatus) positive for TstrGHV1, one in spleen and one in lung. In wild Callosciurus erythraeus from Italy as well as Callosciurus prevostii from a holding in Germany, we discovered both a BHV and a GHV. Upon testing a total of 35 Callosciurus erythraeus, we identified 16 individuals as CeryBHV1-positive (12x in spleen, 2x in lung and 2x in spleen and lung) and two animals as CeryGHV1-positive (1x in spleen and 1x in lung). Within the 17 Prevost’s squirrels, one spleen each was detected to be positive for CpreBHV1 or CpreGHV1 (Table 3).

We did not detect any PyVs in 77 Eurasian red squirrels, 11 Richardson’s ground squirrels and two Eastern chipmunks and no HVs in seven spleen samples of variegated squirrels from German holdings. In one lung sample of an American red squirrel (Tamiasciurus hudsonicus) from the USA, neither PyVs nor HVs were detected (Tables 2 and 3).

Based on extended sequences that we generated for three PyVs and eight HVs (see below) we selected specific nested primers (Additional file 1) for more sensitive screening and re-tested all samples of the respective species. By this approach we observed the following prevalences: ScarPyV1 was detected in 35.2% (31/88 squirrels, 27x positive in spleen, 2x lung and 2x in spleen and lung) of Sciurus carolinensis, SvarPyV1 in 14.3% (1/7 spleen samples positive) of Sciurus variegatoides, CeryPyV1 in 42.9% (15/35 squirrels, 12x positive in spleen and 3x in spleen and lung) of Callosciurus erythraeus and CprePyV in 23.5% (4/17 squirrels positive in spleen) of Callosciurus prevostii (Table 2). For the herpesviruses, the ScarGHV1 prevalence of the Sciurus carolinensis from the UK was 40.8% (20/49 squirrels, 16x positive in spleen and 4x in spleen and lung) and for Sciurus carolinensis from the USA prevalence was 60.0% (3/5 spleen samples positive). The second gammaherpesvirus of Sciurus carolinensis (ScarGHV2) was detected in 2.0% (1/49 squirrels positive in spleen) and 20.0% (1/5 lung samples) of the samples from Britain and the USA, respectively. The average prevalence of ScarBHV1 amounts 5.7% (5/88 squirrels positive in spleen) and was only found in Eastern grey squirrels from Italy. The Sciurus vulgaris from the UK revealed a SvulBHV1-prevalence of 68.4% (13/19 squirrels, 1x positive in spleen, 9x in lung and 3x in spleen and lung) and those from Germany, 31.0% (18/58 squirrels, 1x positive in spleen, 12x in lung and 5x in spleen and lung). In Callosciurus erythraeus from Italy, the prevalence of CeryBHV1 was 65.7% (23/35 squirrels, 19x positive in spleen, 1x in lung and 3x in spleen and lung) and that of CeryGHV1 was 14.3% (6/35 squirrels, 5x positive in spleen and 1x in spleen and lung). Richardson’s ground squirrels from Canada showed a very high prevalence (90.9%, 10/11 spleen samples positive) of UricGHV1. The herpesvirus prevalence in Prevost’s squirrels was 5.9%: Two of seventeen samples were positive, one for CpreBHV1 and the other for CpreGHV1 (Table 3).

In this study we identified one Sciurus variegatoides (sample #10291) from a holding in Germany with a coinfection of VSBV-1 and SvarPyV1 and four Callosciurus prevostii (samples #10295, #10296, #10303 and #10304) from German holdings that were infected with VSBV-1 and CprePyV1.

Complete genomes of ScarPyV1, CeryPyV1, and CprePyV1 were generated by nested LD-PCR using specific tail-to-tail primers (Fig. 1, Additional file 1 and Additional file 2). The LD-PCR products of around 5.2 kbp were sequenced by classical primer-walking (primers not listed) and sequences of the (i) initial generic PCR products, (ii) LD-PCR products, and (iii) specific PCR fragments mentioned above were assembled to final circular genome sequences (Fig. 1). Full PyV genomes or partial PyV sequences are listed with countries of origin and GenBank accession numbers in Table 4.

ORF analysis with Geneious 11.1.5. software revealed the typical PyV genome organization with CDS for the viral capsid proteins VP1, VP2 and VP3, the regulatory proteins STAg and LTAg (but not middle T antigen (MTag)) on the opposite strand, and the NCCR (Fig. 1). As the LTAg CDS of all other mammalian PyVs and some of the STAg CDS are interrupted by an intron, we searched for such introns in the squirrel PyV genomes in silico and experimentally, as described in detail below. In addition, the VP2 ORF of CprePyV1 and ScarPyV1 was found to be interrupted by two (CprePyV1) or three (ScarPyV1) stop codons (not shown). Therefore, splicing of the VP2-encoding late mRNA was predicted and analyzed experimentally in cell culture. In upstream direction of the VP2 coding sequence ScarPyV1 and CprePyV1 display an additional ORF in their genome. In a few PyVs of the genus Betapolyomavirus, e.g. SV40 and BK virus, the so-called agnoprotein is encoded at this position. Agnoprotein was identified as a regulatory protein required for efficient virus proliferation. However, the proteins putatively encoded by ScarPyV1 and CprePyV1 do not show similarities with the known agnoproteins, and their function is currently unknown.

Splice donor and acceptor sites with high Human Splicing Finder (HSF) rating (75–95) and conserved in sequence and position compared to related PyVs were identified in LTAg CDS of all three squirrel PyVs. In addition, splice sites were predicted for STAg mRNA of CprePyV1, with the splice donor site interrupting the stop codon. The introns of STAg mRNA of ScarPyV1 and CeryPyV1 were found to be located after the stop codon. Finally, an intron was predicted for the VP2 CDS of CprePyV1 and ScarPyV1 explaining the occurrence of an interrupted CDS (Fig. 2 and Table 5).

Experimental confirmation of the splice sites predicted for ScarPyV1 was performed with an approach that was used previously for splice site analysis of two human PyVs [54]. First we transfected either the early or late region with flanking sequences of ScarPyV1 (from sample #9804) into monkey Vero cells and murine NIH/3T3 cells, isolated mRNA at the day of transfection and at days 1, 2.5 and 6 after transfection, and synthesized cDNA. Thereafter, a nested PCR was performed with primers that flank the putative introns (Additional file 1). By sequencing of the PCR products, the splice sites predicted for both the early (STAg and LTAg CDS) and the late region (VP2 CDS) were confirmed in Vero and NIH/3T3 cells as described in detail below.

With two nested PCRs that respectively span 300 bp and 500 bp in the early region around the predicted STAg intron, a spliced mRNA was detected in Vero and NIH/3T3 cells that displays an unspliced CDS of 519 bp and encodes the predicted STAg of 172 amino acids (aa). Behind the stop codon (nucleotide (nt) 517–519), a short intron of 72 nt was observed (Fig. 2a and Table 5). With PCR that spans 800 bp in the early region around the predicted LTAg intron, a spliced mRNA was detected in Vero and NIH/3T3 cells from which a spliced CDS of 1.935 kbp (exon 1 and 2 in frame + 1) was inferred. It encodes an LTAg of 644 aa (Table 5 and Fig. 2a). The ScarPyV1-encoded TAgs share the 81 N-terminal aa.

Finally, we performed a PCR that spans 400 bp in the late region around the predicted VP2 intron. A spliced CDS was amplified from cDNA that displays two exons (Table 5 and Fig. 2b).

The experimental results generated for ScarPyV1 were completely in line with the theoretical predictions, i.e., all three predicted introns were detected in both tested cell lines and at different time points of harvesting. The splice sites predicted for LTAg, STAg, and VP2 CDS of CprePyV1 (Table 5) were not studied experimentally because the CprePyV1 genome is closely related to that of ScarPyV1 (see phylogenetic tree analysis below) and the splice sites of both PyVs are conserved in position and sequence. Likewise, splice sites predicted for CeryPyV1 were not studied experimentally as they are conserved in position and sequence with those of the closely related Philantomba monticola PyV1, whose sites were experimentally confirmed earlier [65].

To extend the DPOL gene sequence information for each discovered HV in upstream direction into the adjacent gB CDS, all samples HV-positive in the DPOL PCR were re-evaluated with generic PCRs that were reported earlier [44, 51, 52, 66] to target the gB sequences of either BHVs or GHVs (Additional file 1 and Additional file 3). These generic gB PCRs were less sensitive compared to the generic DPOL PCR, as we amplified a partial gB sequence of ScarGHV1 from only one lung sample of 20 spleen and seven lung samples that had been positive in DPOL PCR. SvulBHV1 gB sequence was amplified from two spleen and three lung samples of 39 DPOL PCR-positive samples and UricGHV1 gB sequence was amplified from all 10 DPOL PCR-positive spleen samples. Next, we closed the sequence gap between the gB and the DPOL sequence (approximately 3.2 kbp) for each of the three HVs with LD-PCR and sequenced the gB-to-DPOL product by classical primer walking. This led to five continuous SvulBHV1 sequences with a length of 3.327–3.442 kbp, three ScarGHV1 sequences of 3.334 kbp, and five UricGHV1 sequences of 3.212 kbp (Table 4).

For ScarBHV1, ScarGHV2, TstrGHV1, CpreBHV1, CpreGHV1, CeryBHV1 and CeryGHV1 the generic gB PCR did not produce the desired gB fragment. Therefore we extended the short DPOL sequences from around 170 bp to > 400 bp by using the outer sense primer of the generic DPOL PCR (285-S DFA; Additional file 1) in combination with two virus-specific antisense primers (Additional file 1) in hemi-nested PCR format. This approach was successful for ScarBHV1 (412 bp; from one sample), CeryBHV1 (478 bp; from three samples), CeryGHV1 (472 bp from one sample), CpreBHV1 (478 bp; from one sample) and TstrGHV-1 (443 bp from two samples), and failed for CpreGHV1 and ScarGHV2. The partial HV sequences are listed with countries of origin and GenBank accession numbers in Table 4.

For evolutionary conclusions and taxonomical classification phylogenetic analyses were done. The ML and MCC trees based on PyV LTAg aa sequences (Fig. 3 and Additional file 4) are quite similar and allow the following conclusions: the novel PyVs can be tentatively assigned to different genera within the family Polyomaviridae, as ScarPyV1 and CprePyV1 nest within the diversity of genus Betapolyomavirus and CeryPyV1 in the genus Alphapolyomavirus. Within the betapolyomaviruses the three ScarPyV1 (GenBank Accession numbers MK671096, MK671097 and MK671101) and the three CprePyV1 (GenBank Accession numbers MK883808 - MK883810) cluster together in a well-supported monophyletic group, which also comprises another rodent polyomavirus, Glis glis polyomavirus 1 (GgliPyV1, GenBank Accession number MG701352), and Delphinus delphis polyomavirus 1 (DdelPyV1, GenBank Accession number KC594077). The close evolutionary relationship of these viruses is strengthened by the observation that their VP2 CDS is interrupted by an intron. This splicing event within the VP2 CDS has been either experimentally verified (Sciurus carolinensis polyomavirus 1 (this study) and Glis glis polyomavirus 1 [65]) or predicted in silico for Delphinus delphis polyomavirus 1, and Callosciurus prevostii polyomavirus 1 as the VP2 CDS of all clade members comprise highly conserved splice donor and acceptor motifs with high score (> 75) in the Human Splicing Finder 3.1.

All three representatives of the novel CeryPyV1 cluster together within the genus Alphapolyomavirus. Although their exact phylogenetic placement is uncertain, they belong to a well-supported clade of polyomaviruses infecting hosts of the orders Artiodactyla, Chiroptera, Primates and Scandentia (Fig. 3 and Additional file 4).

As expected and reported many times, the VP1-based ML and MCC trees (Additional file 5 and Additional file 6) support clades that differ from those delineating genera in the LTAg-based analyses [41]. In these trees, ScarPyV1 and CprePyV1 also formed a weakly supported monophyletic group with GgliPyV1 and DdelPyV1 but this group also included Canis familiaris polyomavirus 1 (GenBank Accession number KY341899). This virus may also have a spliced VP2 as it shares conserved VP2 splice sites with the other four polyomaviruses. The evolutionary position of SvarPyV1 (from the variegated squirrel) had to be allocated in the phylogenetic trees based on VP1 (Additional file 5 and Additional file 6), because only a partial VP1 sequence was identified. The single SvarPyV1 sequence clusters within the genus Alphapolyomavirus and is a sister virus to a PyV group consisting of an organ-utan and two chimpanzee PyVs. This phylogenetic allocation and the tentative host association of SvarPyV1 will need confirmation, once additional SvarPyV1 sequences from several animals and a complete SvarPyV1 LTAg sequence are available.

The ML and MCC trees based on partial DPOL sequences of BHVs (Fig. 4 and Additional file 7) are very similar and show that the four identified squirrel BHVs form a separate cluster that is associated with moderate support to a clade comprising rodent HVs of the genus Muromegalovirus and other currently unclassified rodent HVs. The ML and MCC trees based on partial gB (Additional file 8 and Additional file 9) are very similar to the DPOL-based trees but only the Sciurus vulgaris betaherpesvirus 1 sequences are included as for the other three BHVs gB sequences were not available.

The ML and MCC phylogenetic trees based on partial DPOL (Fig. 5 and Additional file 10) and partial gB of GHVs (Additional file 11 and Additional file 12) are also quite similar and show that the squirrel GHVs form two separate groups. One group nests within GHVs of the genus Rhadinovirus and forms a weakly supported clade with a group of rodent HVs (in all trees) and a rhadinovirus of the South American tapir (only in DPOL-based trees). The other squirrel HV group nests within a clade that comprises ungulate GHVs of the genus Macavirus, a GHV of African elephant and the human Epstein-Barr virus (species Human herpes virus 4).