Widespread exposure to SARS-CoV-2 in wildlife communities

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), one of the greatest threats to public health, continues to evolve and raises concerns about spillover into wildlife populations where a sylvatic cycle could become established and potentially serve as a source for new variants. While transmission to captive animals is well documented, detections in free-ranging wildlife have been limited to relatively few species (e.g., white-tailed deer, feral mink, Eurasian river otters). Experimental infections and ACE2 receptor modeling suggest numerous wildlife species may be competent hosts, but it remained unclear whether diverse wildlife species are infected under natural, typically indirect, exposure conditions. Since 2019, multiple SARS-CoV-2 variants have emerged in humans and animals, including lineages in deer that suggest both human-to-deer spillover and deer-to-deer transmission, implying minimal adaptation may be necessary for onward transmission in some species. These dynamics raise questions about the extent of human-to-wildlife transmission and the potential for other wildlife species to sustain transmission. Establishment of SARS-CoV-2 in wildlife could generate novel mutations affecting virulence, transmissibility, or immune escape, complicating public health responses and vaccine development. To address these gaps, the study examined how widespread SARS-CoV-2 exposure has been in wildlife communities (May 2022–September 2023) across Virginia and Washington, D.C., using RT-qPCR of 789 swab samples from 23 species, serology (PRNT) on 126 serum samples from six species collected before and after SARS-CoV-2 emergence, and analyses of urbanization/human use effects and viral genomics from positive wildlife samples.

Prior work documented SARS-CoV-2 infection in a limited set of free-ranging wildlife, notably white-tailed deer, with evidence for multiple human-to-deer spillovers and onward deer-to-deer transmission. Experimental infection studies and ACE2 modeling indicate susceptibility and potential viral shedding in several peridomestic species (e.g., deer mice, skunks), suggesting broader host competence than field detections had shown. Variants dominant in humans often carry mutations that enhance infectivity and may alter host range, raising concerns that minimal adaptation may suffice for transmission in certain wildlife (e.g., deer). Collectively, the literature underscores the need to quantify the breadth of wildlife exposure, the role of human activity, and to track viral evolution and cross-species transmission in natural settings.

Study design and sites: Wildlife were sampled across 43 counties in Virginia and Washington, D.C., USA. Nasopharyngeal/oropharyngeal swabs were collected from free-ranging wildlife at eight field sites spanning rural-to-urban gradients (Giles, Montgomery, Roanoke, Wythe counties) during 2022 and 2023, and from three wildlife rehabilitation centers (Boyce, Roanoke, Waynesboro, VA). Trapping was conducted in 2–4 day sessions (May–July 2022) and 4-day sessions (June–September 2023). Trapping and processing: Animals were captured using Tomahawk and Sherman live traps. Larger animals were anesthetized in chambers (oxygen 2 L/min with 3–4% isoflurane); smaller animals were anesthetized via canister with isoflurane and then masked (except mice). Individuals were ear-tagged; morphometrics, sex, and reproductive status recorded. Swabs: polyester swabs for naso/oropharyngeal sampling; two swabs per animal collected after July 2023. Swabs were stored in transport media (DMEM + DNA/RNA Shield) on ice, then at 4 °C, and processed within five days. Blood for serology was collected (submandibular vein in mice; toenail quick in other species), stored at 4 °C. Biosafety and contamination control: Personnel used fit-tested N95s and gloves; equipment sterilized between animals and sessions. Personnel were regularly tested by RT-qPCR (1–2×/week; daily in fall 2023). RNA extraction and RT-qPCR: Total RNA from swab media was purified using 96-well spin columns. RT-qPCR (Bio-Rad CFX384) used a multi-target assay (N, E, S genes) with high analytical specificity and 100% detection probability at 10 copies/10 µL across targets; Cox1 (species-specific) and human RPP30 housekeeping genes served as sampling controls. Samples were considered conclusively positive when at least two viral targets and the housekeeping gene amplified below standard-curve thresholds (per EUA-FDA/CLIA protocols). The Virginia DCLS had confirmed assay performance against the CDC EUA diagnostic panel. One 2022 opossum positive was insufficient for USDA NVSL confirmation but was sequence-verified as opossum-derived. Serology (PRNT): Archived pre-2020 and 2022 wildlife serum samples were assayed by PRNT using SARS-CoV-2 Delta (USA-GNL-1205/2021) on VeroE6 hACE2-TMPRSS2 cells. Samples diluted 1:10; virus-serum mixtures (~40 PFU/well) incubated 1 h at 37 °C, inoculated, overlaid, fixed at 48 h, stained, and plaques counted. Neutralization thresholds: ≥90% (strong positive), ≥60% (weak positive), <60% (negative). Analyses of urbanization/human presence used a 60% cutoff. Urbanization and human presence metrics: Impervious surface data from NLCD 2019 and population density from 2020 Census were summarized within species-specific buffers around trap locations. Human presence at serology sites was estimated from trail counters, landowner reports, state park visitation records, facility usage, or census-based occupancy approximations, standardized to visits per month. Human clinical surveillance: 13,221 human samples (May 1, 2022–Sept 8, 2023) analyzed by RT-qPCR; 4,123 positives were sequenced to contextualize circulating lineages during wildlife sampling periods. Sequencing: 2022 opossum whole genome sequenced (Illumina MiSeq) using ARTIC V4.1 primers; 2023 wildlife positives sequenced on ONT GridION using Midnight-ONT/V3 scheme. Basecalling/demultiplexing via Guppy; analysis with EPI2ME wf-artic (Medaka, Nextflow, Nextclade, Pangolin). Sanger sequencing used ARTIC primers for some targets. Sequences deposited to NCBI and GISAID (accessions listed in supplements). Phylogenetics: Consensus sequences were placed with UShER against global trees (12.4M genomes in 2022; 16.3M in 2023). Nextclade used for clade/mutation calling. ML phylogenies were generated (DECIPHER/TreeLine) including 90 representative human sequences across 15 Pango lineages and lineage-focused regional trees (VA and neighboring states; or North America when sparse). Novel amino acid substitutions unique to wildlife sequences were identified. Molecular modeling: Structural modeling of BA.2 Spike bound to hACE2 (PDB 7X08) introduced E471V and G798D mutations (PyMOL), followed by minimization (Schrödinger Maestro, OPLS3e). MM/GBSA binding free energies and surface properties were computed (Bioluminate), and conformational sampling of RBD-hACE2 complexes assessed ΔGbind differences between BA.2 and E471V variants. Statistics: Prevalence and 95% CIs via Agresti-Coull (PropCIs). Univariate GLMMs (glmmTMB) with species as a random effect tested effects of imperviousness, human presence, and population density on seroprevalence. Two-tailed p-values reported.

• Sampling overview: 789 naso/oropharyngeal swabs from 23 wildlife species; 23 unique RT-qPCR positive individuals across six species (2.9% of samples; 26.1% of species). Slightly higher positivity in field samples vs rehab centers (4.04% vs 2.24%).
• Species-level RT-qPCR positives (positivity; N tested): deer mouse Peromyscus maniculatus (4.7%; 8/172), Virginia opossum Didelphis virginiana (2.9%; 4/140), raccoon Procyon lotor (4.8%; 4/84), Eastern cottontail Sylvilagus floridanus (2.5%; 3/118), groundhog Marmota monax (9.7%; 3/31), Eastern red bat Lasiurus borealis (8.3%; 1/12).
• Serology (2022 samples; ≥60% PRNT cutoff): antibodies detected in 5/6 species: Virginia opossum 37.5% (N=8), raccoon 36.4% (N=11), Eastern gray squirrel 57.1% (N=7), white-footed mouse 16.7% (N=6), deer mouse 7.1% (N=14). Four samples exceeded 80% neutralization (2 raccoons, 1 opossum, 1 white-footed mouse). Pre-2020 sera showed significantly lower neutralization than post-arrival (t = -10.774, p < 0.001).
• Urbanization and human presence: Positive relationship between imperviousness and seroprevalence (intercept -1.665 ± 0.48 SE; slope 0.039 ± 0.02 SE; p = 0.031). A low-imperviousness but high-visitation state park exhibited high seroprevalence (80%, N=5), similar to urban sites. Human presence positively associated with seroprevalence (intercept -1.132 ± 0.36 SE; coeff 0.705 ± 0.35 SE; p = 0.044); robust across neutralization cutoffs (40–65%).
• Genomics: Sequences obtained from 12/23 RT-qPCR positives; lineages assigned for 9 individuals across six species. 2022: opossum BA.2.10.1 with mutations in ORF1a/b, S, E, M; unique S:E471V within RBD and S:G798D within S2 fusion peptide region relative to reference and contemporaneous human sequences. 2023: all sequences were XBB-lineages circulating in humans in VA at the time, including XBB (deer mouse), XBB.1.5 (raccoon), XBB.1.5.10 (opossum), XBB.1.16 (Eastern cottontail), XBB.1.5.45 (groundhog), EG.5.1.1 (deer mouse, N=2), JD.1 (deer mouse). Three additional S-gene amplicons (two Eastern red bats, one deer mouse) matched SARS-CoV-2 (95–100% identity) but were too short for lineage assignment.
• Transmission inference: Wildlife sequences closely matched human contemporaneous sequences (99.1–100% similarity), suggesting at least seven recent independent human-to-animal introductions. Co-circulation of multiple lineages at the same site/time and clustering of two deer mouse EG.5.1.1 sequences suggest potential shared exposure or possible animal-to-animal transmission.
• Molecular modeling: E471V in RBD predicted to increase hACE2 binding favorability (MM/GBSA ΔGbind more favorable vs BA.2); G798D near N801 glycosylation site may alter glycosylation probability and affect Spike function.
• Additional unconfirmed detections: Numerous single-gene RT-qPCR amplifications across additional species (e.g., beaver, bobcat, black bear, red fox, white-tailed deer, skunk, Eastern gray squirrel) and some sequence matches suggest possible broader exposure, though conservative criteria required ≥2 genes for confirmation.

The findings demonstrate widespread exposure and infection of SARS-CoV-2 across diverse peridomestic wildlife species beyond white-tailed deer, indicating that multiple common North American mammals encounter and can carry SARS-CoV-2 in natural settings. Patterns in deer mice, raccoons, and cottontails partly align with ACE2 modeling and some experimental infection studies, though discrepancies (e.g., raccoon and rabbit RNA positives despite limited shedding in early-variant lab studies) likely reflect viral evolution and altered host competence with Omicron-class variants. The strong association between human activity (urbanization and site visitation) and wildlife seroprevalence highlights human-wildlife interfaces—urban parks, recreational sites, and areas with human refuse—as plausible hotspots for cross-species transmission. Genomic analyses show multiple, recent human-to-animal introductions with lineages mirroring human circulation, while occasional clustering at sites suggests possible localized animal-to-animal spread. A unique opossum Spike mutation (E471V) predicted to enhance hACE2 binding points to potential host-associated adaptation, though its origin remains uncertain. Together, the results underscore ongoing bidirectional risks: wildlife exposure to evolving human lineages and the possibility of wildlife-adapted variants that could impact public and animal health. Targeted surveillance at high human-use sites and continued sequencing of both wildlife and human cases are essential to monitor adaptation, detect nascent sylvatic cycles, and assess spillback risk.

This study expands the known spectrum of wildlife species exposed to and infected with SARS-CoV-2, documenting positive RNA detections in six species and neutralizing antibodies in five of six species tested. Seroprevalence correlates with urbanization and human presence, indicating high human-use areas as important interfaces for cross-species transmission. Genomic data reveal multiple, recent human-to-animal spillovers involving contemporaneously circulating Omicron sublineages, with one wildlife-derived sequence carrying a unique Spike mutation predicted to enhance hACE2 binding. These findings emphasize the need for continued, broad wildlife surveillance; focused monitoring in urban and recreational hotspots; updated experimental infections incorporating current variants; and comprehensive genomic analyses to detect potential wildlife adaptation and assess risks of establishing sylvatic cycles and spillback to humans. Future research should elucidate transmission pathways (e.g., refuse, wastewater, contact with pets), evaluate variant-specific host competence across species, and employ molecular dynamics to probe the functional impacts of wildlife-associated mutations.

• Transmission source uncertainty: Reduced human sequence reporting in 2023 constrained fine-scale inference of human-to-animal versus animal-to-animal transmission for some events.
• Conservative diagnostic criteria: Requirement of ≥2 gene targets may undercount true positives with low viral loads, particularly in small-bodied species; many single-gene amplifications and partial sequences suggest broader exposure but were not classified as confirmed.
• Serology assay choice: PRNT provides high specificity but potentially lower sensitivity than binding-antibody assays (e.g., ELISA), possibly underestimating prior exposure.
• Sample size and coverage: Some species had small sample sizes, limiting precision of prevalence estimates and cross-site comparisons.
• Temporal scope: Sampling windows (2022–2023) may not capture seasonal or interannual variation in exposure dynamics.
• Modeling constraints: Structural predictions (MM/GBSA) do not capture full dynamic conformational behavior of Spike; functional consequences of mutations require experimental validation.