The southern sea otter (Enhydra lutris nereis) is a threatened marine sentinel. During postmortem investigations of stranded sea otters from 2004 to 2013 in California, US, papillomas were detected in the oral cavity of at least seven otters via necropsy and histopathology. Next-generation sequencing of viral particles purified from a single papilloma revealed a novel papillomavirus, Enhydra lutris papillomavirus 1 (ElPV-1). The genome of ElPV-1 was obtained, representing the first fully sequenced viral genome from southern sea otters. Phylogenetic analysis of the entire L1 gene, as well as a concatenated protein identities plot of all papillomaviral genes revealed that ElPV-1 is a λ-papillomavirus, related to a raccoon papillomavirus (Procyon lotor papillomavirus type 1) and a canine oral papillomavirus. Immunohistochemical staining, using a cross-reactive bovine papillomavirus antibody, suggested that ElPV-1 is present in intranuclear inclusions and intracytoplasmic keratin granules. Virus-infected cells were scattered throughout the stratum granulosum and stratum spinosum of the gingival and buccal papillomas. Using ElPV-1–specific PCR, we confirmed viral DNA in oral papillomas from all seven stranded sea otters, with identical L1 sequences. This virus is associated with the development of oral papillomatosis in southern sea otters.
The southern sea otter (Enhydra lutris nereis) is a federally listed, threatened species found only in California, US. Because of overharvest, the population declined to approximately 50 individuals in the 1910s and has slowly increased during the past century to approximately 2,800 animals in 2012 (Hatfield and Tinker 2012). Because this population has experienced a significant genetic bottleneck, reducing genetic diversity, a potential for enhanced susceptibility to infectious disease has been postulated (Aguilar et al. 2008). Southern sea otters serve as a keystone species; by controlling sea urchin populations, they promote the growth and survival of the kelp forest, providing critical habitat for other marine species (Riedman and Estes 1990).
Despite more than a century of legal protection, ongoing high mortality has limited southern sea otter population recovery (Kreuder et al. 2003). A high proportion of animals have died from infectious diseases caused by fungi and parasites that have been traced to terrestrial sources (Miller et al. 2001; Kreuder et al. 2003; Conrad et al. 2005). Following years of careful sea otter necropsy and histopathologic screening, potential infections by uncharacterized viruses have been identified (M.A.M. unpubl.), but, until this year, the only publicly available genetic information for sea otter viruses was a 200 base-pair (bp) herpesviral sequence (GenBank GU979535).
Occasionally, small, raised, fleshy, oral plaques were observed on the gingiva and buccal mucosa of stranded southern sea otters. Histopathologic characteristics of these masses were similar to papillomaviral lesions described in other mammals (Demonbreun and Goodpasture 1932; Le Net et al. 1997; Robles-Sikisaka et al. 2012). Papillomaviruses (PVs) are nonenveloped, double-stranded DNA viruses most often associated with benign, mucosal and cutaneous, epithelial proliferations, such as warts and condylomas. Some PVs are also capable of inducing malignant tumors (Franceschi et al. 1996). We used next-generation sequencing of enriched viral particle–associated nucleic acids (Ng et al. 2011, 2013) to sequence any virus(es) present in the lesions, followed by PCR, electron microscopy, and immunohistochemical staining for PV.
MATERIALS AND METHODS
Between 2004 and 2013, stranded southern sea otters were necropsied as part of ongoing research on causes of mortality. Systematic necropsies were completed at the Marine Wildlife Veterinary Care and Research Center (MWVCRC) as described by Kreuder et al. (2003), including microscopic examination of all major tissues, bacterial culture, and biochemical testing for biotoxins. Additional samples were cryoarchived at −80 C for future testing. For suspect oral papillomas, immunohistochemical staining was performed on 4-µm, formalin-fixed, paraffin sections on glass slides using an anti–BPV-1 polyclonal antibody (B0580, Dako, Carpinteria, California, USA) at 1∶800 dilution. Anti–BPV-1 polyclonal antibody was labeled with polymer-based immunoperoxidase (Dako Envision). Postincubation, postwash antibody binding was visualized using Vector Nova Red chromogen (Vector Laboratories, Burlingame, California, USA).
For negative-staining transmission electron microscopy (TEM), 1 mL of distilled water was added to about 1 g of thawed, cryopreserved oral papilloma tissue from a southern sea otter and homogenized using a hand-held homogenizer. The resulting slurry was centrifuged at 12,000 × G in a microfuge for 2 min. One drop (30–50 µL) of the supernatant was placed onto parafilm, and a formvar, carbon-coated/Parlodion-subbed electron microscopy (EM) grid (film side down) was placed on top of the drop for 30 min. The grid was then removed, blotted with filter paper, and placed onto a drop of 3% phosphotungstic acid (PTA), pH 7.0, for 30 s. Excess PTA was removed, and the EM grid was allowed to dry on clean filter paper (film side up) before being examined and photographed in a TEM at 25,000× magnification.
Metagenomic sequencing of the papillomavirus was performed according to previously described protocols (Victoria et al. 2008; Ng et al. 2013), consisting of the following steps: filtration of tissue homogenate to enrich viral particles; depletion of host nucleic acid in filtrate using nucleases; unbiased, sequence-independent amplification using random priming; and deep sequencing using Illumina MiSeq (Illumina, Inc., San Diego, California, USA). The resulting sequences were compared with GenBank virus and nonredundant database using BLASTn and BLASTx (NCBI 2014).
To obtain the complete genome, we performed PCR using two overlapping PCR primer sets. The first set, ElPV-AF 5′-AGGCAATGTGCCATGTGTAA-3′ and ElPV-AR 5′-ATGTGCATGCTGATGAGAGG-3′, amplified a 7,587-bp region. The second set, ElPV-BF 5′-CATCGTCACCCTCGTCTTCT-3′ and ElPV-BR 5′-TGCTCTGTTGACTGGGTTGT-3′, amplified a 1,192-bp region. Six additional samples were screened, using a diagnostic PCR developed to target the entire L1 protein, with primers ElPV-CF 5′-CAGAAGTTGGGTTTGGGAAA-3′ and ElPV-CR 5′-CCTGCCAACAATGAAGACAA-3′. We performed PCR using a previously described protocol (Ng et al. 2013). The resulting amplicons were sequenced by Sanger methods, and the complete genomes were assembled using Geneious version R6 (Biomatters, Auckland, New Zealand). Sequence alignment was performed using MAFFT software (Katoh et al. 2005) with the E-INS-I alignment strategy and previously described parameters (Ng et al. 2011, 2012). Bayesian phylogenetic analysis was performed using the amino acid sequences of the L1 protein of related papillomaviruses. For concatenated protein identities plots, gene sequences (in terms of open reading frames) were translated computationally and concatenated by each genome. Pairwise comparison was performed on the amino acid alignments, and the identities were plotted with a step window size of 36 amino acids.
During 2004–13, seven southern sea otters receiving detailed necropsies at the MWVCRC exhibited multifocal to coalescing papillomas. These lesions were most commonly observed incidentally in the mouths of immature to subadult sea otters (aged 1–3 yr) during routine postmortem examination. Representative samples were placed in 10% neutral-buffered formalin for histopathology, and adjacent papillomas were cryopreserved at −80 C for viral characterization.
Sea otter papillomas were characterized grossly as single or multiple, often coalescing, variably pigmented (pale pink to black), flat-topped, exophytic masses ranging from 2 mm to 5 mm in diameter, which were scattered across the rostral gingiva or buccal mucosa (Fig. 1A) and, less commonly, the ventral tongue. Microscopic examination of 4-µm, formalin-fixed, paraffin-embedded, hematoxylin and eosin–stained sections revealed focally extensive epithelial hyperplasia and hyperkeratosis, with an intact basement membrane and prominent, broad, downward-projecting rete pegs, separated by slender bands of fibrovascular stroma (Fig. 1B). At higher magnification, moderate epithelial hyperplasia was accompanied by diffuse, marked parakeratotic hyperkeratosis, hypergranulosis, and intracytoplasmic, large, irregularly shaped keratin granules. Numerous cells, especially in the stratum granulosum and stratum spinosum, exhibited moderate ballooning degeneration and mild spongiosis, and contained both small, round to oval, central, homogenous, lightly basophilic intranuclear inclusions, and large, finely granular, lightly eosinophilic to amphophilic intracytoplasmic inclusions consistent with aggregated keratin (Fig. 1C). Intranuclear and intracytoplasmic inclusions were similar to those described for canine viral papillomas (Le Net et al. 1997), except that the cytoplasmic inclusions in our case were less brightly eosinophilic. Koilocytes were also common, characterized by cytoplasmic ballooning degeneration and eccentrically placed, hyperchromatic nuclei.
Immunohistochemical staining of formalin-fixed, paraffin-embedded sea otter papilloma tissue sections using polyclonal antibodies directed against bovine papillomavirus type 1 (BPV-1, a δ-papillomavirus) yielded positive immunostaining of both the intranuclear and cytoplasmic inclusions within regions of epithelial hyperplasia, especially throughout the stratum granulosum and spinosum (Fig. 1D). Strong BPV-positive labeling occurred only at sites of epithelial hyperplasia, but not in adjacent, normal epithelium, confirming spatial associations between ElPV-1 viral protein expression in cytoplasmic and intranuclear inclusions, and regions of epithelial hyperplasia associated with the formation of grossly apparent oral papillomas in sea otters. Negative-staining electron microscopy also confirmed papillomaviral virions in subsamples of cryopreserved sea otter papilloma tissue (Fig. 1E). One of the cryopreserved papillomas was processed for deep sequencing, revealing sequences consistent with a novel papillomavirus. By designing two overlapping PCR reactions, we obtained the entire 8,194-nucleotide genome of this virus (Enhydra lutris PV 1, or ElPV-1, GenBank KJ410351). Phylogenetic analysis using the L1 protein confirmed that ElPV-1 belongs to the genus Lambdapapillomavirus (Fig. 2A). The closest relatives are raccoon papillomavirus (Procyon lotor papillomavirus type 1, or PlPV-1) and canine oral papillomavirus (COPV), the type species of Lambdapapillomavirus. All known hosts of λ-papillomaviruses, including sea otters, belong to the order Carnivora and ElPV-1 is a λ-papillomavirus detected in a marine carnivore.
Consistent with genome organization of λ-papillomaviruses, ElPV-1 encodes five early (E) proteins (E1, E2, E4, E6, and E7), and two late (L) capsid proteins (L1 and L2) (Fig. 1). The ElPV-1 genome shares 51–60% nucleotide identity with previously reported λ-papillomavirus genomes. Pairwise comparison of all open reading frames (ORFs) among sea otter, dog, and raccoon papillomaviruses (Fig. 2C) indicate that L1 is most conserved, followed by E1 and L2, whereas all other ORFs are divergent, sharing <50% amino acid identity.
Using specific PCR, we identified the same DNA sequence in oral papillomas from two southern sea otters that stranded during 2012, and four otters that stranded during 2013. Additional screening of a cryoarchived sample from 2004 was also positive for ElPV-1. DNA sequences of the entire L1 ORF from all seven cases, spanning 9 yr, were 100% identical.
We describe papillomaviral infection in sea otters. Based on preliminary necropsy and histopathologic characterization of oral papillomas in seven stranded sea otters from 2004 to 2013, a full genome of ElPV-1 was obtained by viral metagenomics from one sample. Subsequent ElPV-1-specific PCR confirmed presence of the same virus in other stranded otters with oral papillomatosis. Immunohistochemical staining and electron microscopy further illustrated the link between ElPV-1 infection and expression of oral papillomatosis in sea otters.
As illustrated in Figure 2C, ElPV-1 shares significant protein identity with both COPV and PlPV-1, especially in the L1 ORF. Together with the phylogenetic analysis, this supports the placing ElPV-1 into the genus Lambda papillomavirus, because L1 is the most conserved papillomavirus gene and, therefore, the primary factor for distinguishing between genera (Bernard et al. 2010). The conservation of L1 is further attested to by the 100% nucleotide identities in seven EIPV-1 cases spanning >9 yr. Similarly, when a large number of human papillomavirus (HPV) sequences of the same subtype were compared, few or no mutations were observed at the L1 locus within the HPV types 31, 35, 52, and 58 (Calleja-Macias et al. 2005).
All other reports of λ-papillomavirus infection are from terrestrial carnivores (Fig. 2A). Enhydra lutris PV 1 is most closely related to COPV and PlPV-1. Canine oral papillomavirus generally infects the oropharyngeal mucosa of young dogs, whereas nasal, genital, and subglottal mucosae were resistant to experimental infection (Demonbreun and Goodpasture 1932). In extreme cases, oral papillomatosis in dogs can interfere with the host's ability to feed. Of the 14 canine papillomaviruses currently known, two belong to the genus Lambdapapillomavirus: COPV and canine papillomavirus 6 (Luff et al. 2012; Lange et al. 2013). Other Carnivora known to carry λ-papillomaviruses include raccoons (Procyon lotor), spotted hyenas (Crocuta crocuta), and members of the family Felidae (Fig. 2A). Although oral papillomas have not been classified in raccoons, λ-papillomavirus infection in these animals (PlPV-1) is typically associated with dermal proliferation on the palmar surfaces of the feet (Rector et al. 2005).
Other genera of papillomaviruses have been reported from terrestrial mustelids and marine mammals. A mustelid papillomavirus of the genus Tau was identified from ferret (Mustela putorius furo) feces using metagenomic sequencing (Smits et al. 2013). In cetaceans and manatees (Sirenia), several papillomaviral genera were associated with oral and genital papillomas (Rector et al. 2004, 2008; Robles-Sikisaka et al. 2012). Some marine mammals, such as the harbor porpoise (Phocoena phocoena), carry multiple types of papillomaviruses (Gottschling et al. 2011). In humans, where papillomaviruses are more completely characterized, more than 180 HPV types exist, with varying pathogenicity and site specificity (Bernard et al. 2010).
In the sea otter cases to date, oral papillomatosis was considered an incidental finding. However, severe oral papillomatosis could impair sea otter fitness by hindering their ability to forage and consume prey, or it could result in secondary bacterial infection. Severe ElPV-1 infection could also compromise immune function.
Lambdapapillomaviruses described in this study have not been linked with cancer. Other types of papillomaviruses can induce neoplastic transformation of infected cells. In humans, nearly all cervical cancers are caused by α-papillomaviruses, with HPV-16 and HPV-18 accounting for more than 70% of cases (Li et al. 2011). Squamous cell carcinoma have been linked in humans with human papillomavirus (Human alphapapillomavirus type 56 [Mii et al. 2012; Kato et al. 2013; Murao et al. 2013]), as well as in domestic cats (Felis catus) and ferrets with animal papillomaviruses (Baer and Helton 1993; Rodrigues et al. 2010; Ravens et al. 2013). Further investigation of papillomavirus diversity in sea otters, especially screening for other papillomavirus types in tumor samples, could provide insight on potential papillomavirus-associated neoplasia of otters. The PCR test that we developed for ElPV-1 DNA detection will facilitate future monitoring of ElPV-1 infection in sea otters. The southern sea otter population has experienced a significant genetic bottleneck, and it is possible that this reduced genetic diversity could enhance susceptibility to papillomaviral infection. Future investigations could trace the incidence of papillomaviral infection in sea otters through time to determine whether there is an increasing trend. Such information will be useful for monitoring the health status of this threatened population.
During detailed postmortem examinations of stranded sea otters during the past two decades, viral infections have often been suspected. However, until now, tests to diagnose viral infections have been unavailable, mainly because of a lack of genetic information on, and biological understanding of, sea otter viruses. Few systematic studies have been reported on viral infections of sea otters. Northern sea otters (Enhydra lutris kenyoni) have been found to be PCR-positive for morbilliviruses, and antibody-positive for influenza A (Goldstein et al. 2009; Li et al. 2014). Both northern and southern sea otters were reported with suspected herpesvirus infection via histopathology and electron microscopy (Reimer and Lipscomb 1998). A poxvirus was recently discovered and sequenced in northern and southern sea otters in a concurrent study this year (Tuomi et al. 2014). In this study, we demonstrated that viruses suspected on routine histopathology can be rapidly identified through next-generation sequencing, followed by development and application of specific PCR assays. This cutting-edge approach will enhance our understanding of viral pathogens infecting sea otters and other marine species. Because southern sea otters have been shown to be important sentinels for land-sea transfer of terrestrial pathogens, including parasites, fungi, and bacteria (Miller et al. 2001; Kreuder et al. 2003; Conrad et al. 2005), continued studies of sea otter viral flora may also reveal land-sea transmission patterns for viruses originating from terrestrial animals and humans.
We thank the staff at CDFW-MWVCRC for assistance with project completion and all collaborating organizations and individuals that have submitted stranded southern sea otters for postmortem examination.
These authors contributed equally to this work.