Detection and characterization of novel luchacoviruses, genus Alphacoronavirus, in saliva and feces of meso-carnivores in the northeastern United States

INTRODUCTION

The Coronaviridae family encompasses diverse viruses with single-stranded, positive-sense RNA genomes. Coronaviruses (CoVs) are currently classified into four genera [Alphacoronavirus, Betacoronavirus, Gammacoronavirus, and Deltacoronavirus (1)], that group viruses known to infect birds (mostly gammacoronaviruses and deltacoronaviruses) and mammals (mostly alphacoronaviruses and betacoronaviruses), including domesticated and wild animal species and humans (2). The potential host range of CoVs is coordinated by the spike (S) protein, which is involved in receptor binding via its S1 domain (3). The S protein also mediates membrane fusion with target cells through its S2 domain (4). In general, cleavage of S can occur during virion biosynthesis at a cleavage site between the S1 and S2 domains (S1/S2 cleavage site), which activates the S protein exposing the receptor binding region within the S1 domain, promoting receptor binding (5). After receptor binding, S is also cleaved in the S2 domain close to the fusion peptide (S2′ cleavage site), promoting membrane fusion (6). Coronavirus entry is often regulated through the S cleavage sites. In the Betacoronavirus genera, MERS-CoV (from the Merbecovirus clade) can be cleaved by furin at both cleavage sites [S1/S2 and S2′ (5)], while SARS-CoV-2 is the only known member of the Sarbecovirus clade that has a furin cleavage site in S1/S2 (7) , which is a key determinant of its emergence (8). In the Alphacoronavirus genera, CoVs like feline and canine CoV type 2 (FCoV-2 and CCoV-2, respectively), transmissible gastroenteritis virus and human CoV 229E do not have an identifiable S1/S2 cleavage site (9), while swine acute diarrhea syndrome CoV (SADS-CoV), an emergent and highly pathogenic CoV that possibly originated from bat CoV HKU2, does have an S1/S2 cleavage site (10). The S protein of SADS-CoV also has an additional cleavage site in the S1 domain whose cleavage together with that on the S1/S2 site is essential for cell-cell fusion (11). Overall, multiple cleavage sites may contribute to membrane fusion and virus entry, processes that are crucial determinants for an efficient infection and the disease outcome.

The advancement in sequencing technologies has contributed to the surge of viral discovery which together with comparative genetic analyses can help increase our understanding of how CoV exploit their genetic variation for host adaptation and pathogenesis (3). As the S protein of several CoVs has been crystalized, it is possible to identify and map regions within this protein that interact with the host for receptor binding (12) and cleavage-mediated activation or fusion (13). Comparative genetic analyses of the S protein between CoVs from different hosts can help guide experimental assays to evaluate the importance of these regions and assess the impact of certain mutations on CoV host range and/or pathogenicity. For example, furin, tryspin-like proteases, and cathepsins, among other proteases, can cleave the S protein after recognizing certain amino acid sequence motifs (14). Mutations and/or insertions/deletions in these motifs can impact the virus-host interactions. For example, the S1/S2 cleavage site of FCoV type 1 (FCoV-1) and SARS-CoV-2 is located in exposed loops easily accessible to proteases (15 – 17). Mutations in these loops that disrupt the furin cleavage motif result in highly pathogenic variants for FCoV-1 (18) and variants with decreased pathogenicity for SARS-CoV-2 (19). Likewise, mutations in the S2′ cleavage site can inhibit viral fusion, which, in MERS-CoV, results in reduced viral infectivity (20). Therefore, identifying and mapping in the tertiary structure of S potential cleavage sites in CoVs detected in different species can shed light on the processes CoVs may use to acquire cleavage sites (21) and its relation to host range, pathogenicity, and evolution of CoVs (22).

Luchacovirus is a subgenus of Alphacoronavirus (also known as rodent or murine Alphacoronavirus), which includes viruses infecting rodents (order Rodentia) and rabbits (order Lagomorpha). Although currently classified as alphacoronaviruses, the S gene of luchacoviruses is more similar to and groups with betacoronaviruses and not alphacoronaviruses, likely as a result of an early recombination event (23). To date, luchacoviruses have only been detected in at least 13 species of rodents (including rats, mice, and voles) and 2 species of lagomorphs (including rabbits and plateau pika, Table 1). Even though luchacoviruses are highly diverse and have been reported from various rabbit and rodent species in different countries in Asia, Europe, and the USA (24, 25), they form a monophyletic group in the phylogenies involving the S and other genes and share similar genetic features in their genomes, suggesting a long association with rodents and rabbits (23). In general, the S gene of these viruses is highly variable, and the sequence of one of these viruses (JC34, accession number KX964649) obtained from a Chevrier’s field mouse (Apodemus chevrieri) in China in 2011 (26) has a furin cleavage motif (S-R-R-A-R | A, where S is serine, R is arginine, A is alanine, and | is where the cleavage occurs) in its S1 domain. Interestingly, this motif has a very similar amino acid sequence to the S1/S2 cleavage site of SARS-CoV-2 [P/H-R-R-A-R | S, where P is proline, H is histidine (22)], which is an atypical furin recognition motif (the canonical furin cleavage motif is X-R-R-X-K/R-R | S, where K is lysine and X is any residue). Although experimentally the Luchacovirus JC34 cleavage motif is not cleaved by human furin (27), continued surveillance of these viruses focusing on targeted regions is essential to understand if there is variation in regions with motifs of interest and their relevance in the origin and emergence of new CoVs.

The meso-carnivores are a group of small- to mid-sized carnivores of economic and ecologic relevance, which include domesticated and farm animals (dogs, cats, and American mink, Canis lupus familiaris, Felis catus, Mustela vison, respectively) as well as various wild species. In the northeastern United States (U.S.), there are 14 species of meso-carnivores (28), including ubiquitous species like raccoons (Procyon lotor), red foxes (Vulpes vulpes), and striped skunks (Mephitis mephitis). Due to their diverse behavior and ecology, meso-carnivores can thrive in various habitats including urban and rural areas (29). In rural areas, they may act as reservoirs for pathogens that may spill over into other species of wild carnivores. Indeed, skunks and raccoons in the eastern U.S. can maintain the sylvatic cycle of rabies independently of domesticated animals (30). In urban areas, they may get exposed to pathogens from or spread infectious diseases to domesticated species, especially free-ranging stray or feral dogs and cats. In the U.S., early estimates suggest there may be more than 30 million free-ranging feral/stray cats (31). Feral/stray cats often live in large groups (known as feral cat colonies) in which some viruses are highly prevalent (32), and large-scale outbreaks can occur (33). In many cities in the U.S., including New York City (NYC), humans often provide food for stray cats through feeding stations. These stations also attract wild species, facilitating the interspecific contact between domestic cats and other urban meso-carnivores as well as with rodents. Despite their potential importance in the spread of infectious diseases, our knowledge of the viral diversity hosted by free-ranging wild and domesticated meso-carnivore species in the U.S. remains limited (34).

Alphacoronaviruses and betacoronaviruses can infect diverse species of meso-carnivores causing a range of diseases. Alphacoronaviruses like mink CoV (35) and ferret enteric CoV (36) cause enteric diseases in mink and ferret, respectively, while ferret systemic CoV (36) and FCoV (37) can cause highly fatal diseases in ferrets and domesticated cats, respectively. Betacoronaviruses like canine respiratory CoV are widely spread throughout the U.S. causing respiratory signs in domesticated dogs (38). Like their domesticated counterpart, wild meso-carnivores also host different CoVs, including the divergent novel alphacoronaviruses identified in free-ranging Asian leopard cats (Prionailurus bengalensis) and Chinese ferret badgers (Melogale moschata) in southern China (39). Meso-carnivores also play a role in the epidemiology of CoVs relevant to humans including SARS-CoV-2 which can infect minks and ferrets (40), species that can also infect humans (41). Likewise, recent studies have found CCoV-2, a common Alphacoronavirus of domesticated dogs and wild canids, in humans (42). This zoonotic event was related to extensive recombination with other alphacoronaviruses and the possible loss of function of the N-terminal region of the S1 domain (43). Meso-carnivores may also be susceptible to CoVs from domesticated and farmed meso-carnivores, due to their close genetic relatedness. For example, the sequence of the receptor for CCoV-2 and FCoV-2 (aminopeptidase N) is highly similar (>92.7%) between members of the Felidae and Canidae families (44). Overall, both urban and rural wild meso-carnivores carry a high diversity of viruses (45) that need to be further explored to understand the complex epidemiology of CoVs in wild populations, especially those highly adapted to anthropogenically disturbed areas.

The aim of this study was to perform a longitudinal surveillance of CoV in free-ranging wild meso-carnivore in the northwestern region of the U.S. Using next-generation sequencing and comparative genetic and phylogenetic analyses, we provide the first molecular characterization, focusing on the S gene, of luchacoviruses detected in three species of meso-carnivore in the U.S. and show that the luchacoviruses detected in the U.S. to date are highly diverse when compared to those reported elsewhere.

MATERIALS AND METHODS

RESULTS

From 184 individual meso-carnivores, we collected 321 samples: 152 feces (or rectal swabs), 7 saliva swabs, 7 nasal swabs, and 155 tissues samples (Table 2) including liver (43%), lung (44%), brain (2.8%), kidney (6%), small intestine (2.8%), and heart (1.4%) from 9 species (Table 3). Using a nested PCR pancoronavirus screening method, we found five samples positive for CoV RNA, four from feces, and one from a saliva swab (Table 2) from three different meso-carnivore species (Table 3). The species with positive samples included a bobcat collected in October 2021 (W291/bobcat/2021), a fisher collected in January 2022 (W145/fisher/2022), a red fox (W317/red fox/2022) collected in February 2022, and a feral domestic cat collected in August 2022 from which we found a positive sample in the saliva swab (DC6837 S/domestic cat/2022) and in the rectal swab (DC6837 F/domestic cat/2022) (Table 4). The bobcat and red fox samples were collected in New Hampshire, and the fisher and feral domestic cat samples were collected in New York State. Additional tissues or swabs from the four animals that were positive for CoV were screened, and all were negative for the presence of CoV RNA. The other tissues included the lung from the bobcat (W291/bobcat/2021) and red fox (W317/red fox/2022); the lung, liver, kidney, and heart from the fisher (W145/fisher/2022); and the nasal swab from the feral cat (DC6837 F and S/domestic cat/2022). Of the 30 fecal samples from raccoons and skunks sampled in 2021 in New York City, none was positive for any CoV. For rodents, we collected 36 samples in 2021, including feces (50%) and lung (50%). We did not find CoV in any of these samples (Table S1).

For the five positive samples obtained from the meso-carnivores, we initially sequenced a partial region of the RdRp gene (401 bp). The percentage sequence similarity of this region between these three sequences was 82.3%. The CoV sequences obtained in the feral cat in NYC (DC6837 F and S/domestic cat/2022) were identical to the one obtained from the red fox (W317/red fox/2022). A phylogeny including these five sequences plus 60 others from CoVs in the four genera revealed that all sequences obtained in this study are within the Luchacovirus clade (Fig. 1). Sequences W145/fisher/202, W317/red fox/2022, and DC6837 F and S/domestic cat/2022 clustered in a group together with viruses from rabbit and vole reported in Europe, while W291/Bobcat/2021 did not group with any specific group of luchacoviruses. The Luchacovirus cluster grouped together with SADS-CoV and bat CoV HKU2 (Fig. 1, emphasized with blue arrows).

Additional sequences were obtained for W317/red fox/2022 and W291/bobcat/2021 but not for W145/fisher/2022 or DC6837 F and S/domestic cat/2022 due to low quality and quantity of RNA. For W317/red fox/2022, we obtained a partial genome sequence (12,214 nt) that includes the complete S gene, and for W291/Bobcat/2021, we obtained a partial sequence of S gene (2,561 nt) containing the complete sequence of the S1 domain and a partial sequence of the S2 domain. A phylogeny constructed using the nucleotide sequences of the S1 domain of the two variants obtained in this study and 36 other CoVs from the four families revealed that our sequences also grouped within the Luchacovirus cluster and that this group was not closely related to the Alphacoronavirus group (Fig. 2). In general, there were three main groups (bootstrap support of 100%) within the Luchacovirus cluster: group 1 includes viruses from voles and rat from China, group 2 includes viruses from rat from China and the UK, and group 3 includes viruses from mice and rat from China (Fig. 2). There were four luchacoviruses, including the two reported in this study, which did not cluster within these three groups. W317/red fox/2022 was placed in a group with one of the other divergent luchacoviruses obtained from a pika (P83/plateau pika/China), outside group 3, while W291/Bobcat/2021 clustered outside groups 1 and 2. The fourth divergent luchacovirus, UMN2020/mouse/2018/USA, collected from a mouse in the U.S., did not group with any other luchacovirus (Fig. 2). Each of the three defined groups has an overall higher percentage of sequence similarity in the S1 domain (>92%) than when compared to the other groups (group 1 vs 2, 50%; group 1 vs 3, 56.2%; and group 2 vs 3, 51.8%) or to the four luchacoviruses that were not within these groups, including the two viruses obtained in this report, W317/red fox/2022 and W291/Bobcat/2021 (<64%), as well as P83/plateau pika/China (<69%) and the most divergent one, UMN2020/mouse/2018/USA (<52%). Like in the phylogeny of the partial RdRp gene, the Luchacovirus cluster grouped together with SADS-CoV and bat CoV HKU2 (Fig. 2, emphasized with blue arrows). Phylogenetic trees of the E, M, and N genes (Fig. S1) also show the three well-supported groups within the Luchacovirus clade. In these trees, W317/red fox/2022 did not cluster with any of these groups, and only in the tree of the M gene did it cluster with UMN2020/mouse/2018/USA, the luchacovirus detected in a mouse in the U.S.

The amino acid sequences of the partial S protein (853 amino acids), containing the complete S1 domain and a partial region of the S2 domain of W317/red fox/2022 and W291/bobcat/2021, were 52.1% similar. We compared the amino acid sequences of the four regions in S known in other CoVs to have protease cleavage sites between the two viruses detected in this study, 15 other luchacoviruses, 2 alphacoronaviruses, and 3 betacoronaviruses. The genetic comparison of these four regions revealed that two of these regions are located in the C-terminal domain (CTD) of S1 (Fig. 3). Region 1 includes the furin cleavage site of SADS-CoV (R-Y-V-R | I; Fig. 3A), whose sequence differs in only one residue (residue 458 in Fig. 3A) with the respective sequence from luchacoviruses from groups 1 and 2, the divergent variants UMN2020/mouse/2018/USA and P83/plateau pika/China ( S -Y-V- R | I), and bat CoV HKU2 ( K -Y-V-R | I). For these variants, a single nucleotide change is sufficient to revert the residue change detected between these viruses (Fig. 3B). Compared to group 3 luchacoviruses and the two viruses obtained in this study, it differs in two residues (458 and 459, Fig. 3A). In the tertiary structure of SADS-CoV, region 1 is located on a beta sheet within the CTD (Fig. 3C). Region 2 is highly variable and includes a weakly predicted furin cleavage motif detected in group 3 luchacoviruses (R-R-A-R | A, ProP score 0.28) and motif F-R | S in W317/red fox/2022/USA consistent with a cathepsin L cleavage site (Fig. 3A). In this region, divergent luchacoviruses UMN2020/mouse/2018/USA and W291/Bobcat/2021 have highly different sequences (Fig. 3A). Region 3 includes the S1/S2 cleavage site, which is observed in betacoronaviruses like SARS-CoV-2 and two rodent betacoronaviruses (ProP scores >0.5), but not in luchacoviruses (Fig. 3A). Region 4 includes the S2′ cleavage site in which furin cleavage motifs were not predicted for any Luchacovirus. Regions 2, 3, and 4 are located in exposed loops in the tertiary structure of SADS-CoV (Fig. 3C).

The 3′-end of the genome of luchacoviruses has a general organization that includes four structural proteins S, envelope (E), matrix (M), and nucleocapsid (N) and three to five non-structural ORFs, one or two between ORF1b and the S protein (ORF2 and 2b), ORF3 between S and E, ORF6 M and N, ORF8 within N and ORF9 after N (Fig. 4). In this region, the three luchacovirus groups differed in the number of ORF2: group 1 has ORF2 and 2b, group 2 only has ORF2, and group 3 does not have any ORF2 (Fig. 4). The obtained partial genome of W317/red fox/2022 had a main similar organization to group 2 luchacoviruses with only one ORF2 (Fig. 4) but differed in that it does not have an ORF8 or its internal transcription signal and ORF9 was shorter than those from the other luchacoviruses, due to a 69-nucleotide deletion (Fig. 4). W317/red fox/2022 shared low overall genetic similarity at the 3′-end of the genome when compared to other luchacoviruses (<86%; Fig. 5A). In comparison, the luchacoviruses within groups 1, 2, and 3 shared an overall high nucleotide similarity (pairwise percentage similarity >94.2%). Divergent luchacoviruses UMN2020/mouse/2018/USA and P83/plateau pika/China had the same genomic organization as luchacoviruses from groups 3 and 2, respectively. However, genetic comparison revealed an overall low similarity throughout the 3′-end of the genome between these luchacoviruses and those in groups 3 and 2 (UMN2020/mouse/2018/USA vs group 3 <82% and P83/plateau pika/China vs group 2 <88%; Fig. S2). We did not find any recombination breakpoints in this region of the genome between the 16 luchacoviruses included in the analysis (see Materials and Methods). The partial S gene of W291/Bobcat/2021 was overall genetically different from other luchacoviruses, including W317/red fox/2022 (<80%; Fig. 5B).

DISCUSSION

Coronaviruses are important pathogens that have a wide host range and can result in subclinical infection or cause a variety of diseases in both domesticated and wild mammals. Within the Alphacoronavirus genus, the Luchacovirus subgenus includes CoVs reported to date only in rodents and rabbits. In this study, we report for the first time luchacoviruses shed in feces and saliva of three species of wild meso-carnivores including bobcat (W291/bobcat/2021), fisher (W145/fisher/2022), and red fox (W317/red fox/2022) and a feral domestic cat (DC6837 F and S/domestic cat/2022) living freely in NYC (Fig. 1). The phylogeny of the S1 domain of the S protein and the general genomic organization of the rodent and rabbit luchacoviruses reported to date showed that there are three main groups (groups 1–3; Fig. 2 and 4; Fig. S1), two of which were already observed using fewer sequences (58). Phylogenetic and genetic comparison of the S1 domain of two of the viruses obtained in this study (W291/bobcat/2021 and W317/red fox/2022) revealed that this region is genetically different from the luchacoviruses reported to date in rodents and rabbits (<75%; Fig. 5) and that these viruses did not group within the three groups that comprise most of the luchacoviruses reported to date (Fig. 2). Although in general CoVs have highly variable S1 regions, further genetic comparison of the 3′-end of the genome revealed that luchacoviruses within the three established groups are highly similar (>94.2% similarity), but W317/red fox/2022 is overall divergent (<86% nucleotide similarity; Fig. 5). Likewise, the genetic organization of the 3′-end of the genome of W317/red fox/2022 had certain differences with the other luchacoviruses including a shorter ORF9 and the absence of ORF8 (Fig. 4). We found that the only available luchacovirus reported from a mouse in the U.S. (UMN2020/mouse/2018/USA) has the same genetic organization as group 3 luchacoviruses, which includes viruses detected in rodents in China (Fig. 4). These results support the idea that, with the sequences available to date, at least two of the viruses reported in this study (W291/bobcat/2021 and W317/red fox/2022), from which we obtained partial genome sequences, are genetically different from other luchacoviruses reported in rodents and rabbits.

We only detected luchacovirus RNA in feces and saliva. Therefore, it is possible that these may be viruses from the prey consumed by these animals, as has been shown for insectivorous bats which shed various insect viruses in their feces (59). However, one of the luchacoviruses was detected in feces and saliva of a feral cat living in NYC (Fig. 1). A previous study of cat colonies in NYC found that 94.3% of sampled cat scat contained cow, chicken, and fish taxa, indicating that most of the diet of these cats consists of cat food provided by humans (60). Additionally, it has been reported that enteric viruses like norovirus, astrovirus, and rotavirus can replicate in salivary glands and be transmitted through saliva (61). The sequences obtained from this feral cat (DC6837 F and S/domestic cat/2022) were identical to the one obtained from the red fox (W317/red fox/2022; Fig. 1), from which we were able to sequence the complete 3′-end of the genome (Fig. 4). Our phylogenetic analyses of the three structural genes (E, M, and N) revealed that W317/red fox/2022 did not cluster with any of the three luchacovirus groups and only grouped with the luchacovirus UMN2020/mouse/2018/USA, which was detected in a mouse bought at a pet store in the U.S. (62) in the M gene tree (Fig. S1). However, a genetic comparison of the M gene between these two viruses showed that their percentage of nucleotide similarity was low (<86%; Fig. 5). Overall, the luchacoviruses reported from the U.S., including those reported in this study (W291/bobcat/2021 and W317/red fox/2022) and the one previously reported from a mouse (UMN2020/mouse/2018/USA), were genetically highly variable and did not group within any luchacovirus group (Fig. 2; Fig. S1). Nevertheless, as stray cats also feed occasionally from mice (60) and cat food stations in urban areas attract rodents (especially house mice (63), it is necessary to screen both rodents and feral cats to unveil the importance of these interspecies gatherings as well as predator-prey interactions in the epidemiology and possible cross-species transmission of luchacoviruses. Moreover, since in the phylogeny of the partial RdRp gene W317/red fox/2022, W145/fisher/2022, and DC6837 F and S /domestic cat/2022 grouped with viruses of vole and rabbit from which only partial sequences exist (Fig. 1), additional sequences of luchacoviruses from these species will also help clarify the host range of these viruses. Whether luchacovirus can infect meso-carnivores or the detected viral RNA was derived from their prey, our results reveal that luchacoviruses circulating in the U.S. are more genetically diverse than those reported in Asia and Europe to date.

Some CoVs have wide host ranges, possibly due to the use of protein receptors like angiotensin converting enzyme 2, dipeptidyl peptidase 4, or aminopeptidase N, which are highly conserved between diverse species (44). As the S1 domain of the S protein of CoVs interacts with the receptor, signatures of viral adaptation to the receptor of different species have been identified in this region which has helped increase our understanding of the molecular mechanisms of the adaptation and emergence of CoVs in various species (12). Much less is known about the relation between the cleavage of S and the emergence of pathogenic CoV variants; however, the presence of a unique cleavage site in the S1/S2 of SARS-CoV2, which, to date, has not been found in closely related viruses (64), has highlighted the importance of studying the mechanisms that CoVs may use to acquire functional cleavage (22). For example, genetic comparison of the S1/S2 region between SARS-CoV-2 and closely related CoVs from European bats revealed that a single nucleotide substitution is sufficient to produce a functional furin cleavage site identical to that of SARS-CoV-2 in two bat CoV variants, an essential molecular determinant of their zoonotic potential (21). However, this substitution alone is not enough to enable furin cleavage in similar bat CoVs, possibly because it is in a short loop that may not be accessible to the protease (65). The Luchacovirus clade is closely related to SADS-CoV and bat CoV HKU2 (Fig. 1 and 2), which are also characterized for having a recombinant S gene that is more similar to betacoronaviruses than alphacoronaviruses (66). To date, the receptor of luchacoviruses, SADS-CoV, or bat CoV HKU2 remains unknown (67); in this study, we directed our genetic analysis of the S protein on finding regions with putative cleavage sites and the genetic comparison of these regions with other CoVs like SADS-CoV (11) and SARS-CoV-2 in which the cleavage of S has been experimentally studied.

SADS-CoV is a highly pathogenic CoV that infects pigs and has a unique cleavage site (R-Y-V-R | I) in the S1 domain (region 1 in Fig. 3) that is essential for cell fusion and may have an impact in the pathogenicity of this virus and its observed broad cell range (68). Comparative genetic analyses have shown that bat CoV HKU2 is closely related to SADS-CoV, suggesting that bat CoVs may also be involved in the emergence of pathogenic CoVs of veterinary importance (56). It has been previously shown that a substitution of a single residue in this cleavage site (residue 458 in Fig. 3) is enough to abrogate the cleavage of S (11). In this study, we show that several luchacoviruses, including those in groups 1 and 2 and divergent variants UMN2020/mouse/2018/USA and P83/plateau pika/China, have a cleavage motif in this region that differ only in this essential residue to that of SADS-CoV (R458S; Fig. 3A) and that a single nucleotide substitution is sufficient to revert this residue (Fig. 3B). Likewise, bat CoV HKU2 also differs only in this residue when compared to SADS-CoV (R458K; Fig. 3A and B). However, we show that unlike SADS-CoV and bat CoV HKU2, luchacoviruses do not have an S1/S2 cleavage site (Fig. 3A). As cleavage in both S1 and S1/S2 sites is required for cleavage and cell fusion of SADS-CoV (11), more than a single non-synonymous change would be necessary for luchacoviruses to acquire a cleavage mechanism like the one of pathogenic SADS-CoV. In contrast to luchacoviruses, most rodent betacoronaviruses have a predicted S1/S2 cleavage site [78% of available S sequences (22)], more than 13 times of what was predicted for bat betacoronaviruses [6% of available sequences (22)]. In vitro experiments have shown that SADS-CoV can successfully infect and replicate in several rodent cell lines (67); therefore, surveillance of rodents, meso-carnivores, pigs, and bats targeting luchacoviruses, rodent betacoronaviruses, and SADS-CoV and characterizing their S genes is essential to detect if there is any recombination or variation in these key regions and its relationship with pathogenicity and host range.

A previous study found that Luchacovirus JC34 had a predicted furin cleavage site (R-R-A-R | A) in the S1 domain similar to that of the S1/S2 cleavage site SARS-CoV2 (22). In this study, we identified two additional luchacoviruses (all from group 3; Fig. 2) that also have this cleavage site (Fig. 3). As it has been reported that this cleavage motif is not cleaved by human furin (27), we conducted a genetic comparison of this region among luchacoviruses, which revealed that this region is overall genetically highly variable. Aside from the furin cleavage motif, we also identified possible motifs for cathepsin L or V cleavage (F-R | S) in divergent luchacovirus P83/plateau pika/China (Fig. 3A). As cleavage motifs need to be accessible to proteases, we mapped this region in the tertiary structure of SADS-CoV and show that it is located in an exposed loop (region 2 in Fig. 3), and it is, like region 3 (S1/S2 cleavage site) and region 4 (S2′ cleavage site; Fig. 3C). Therefore, identifying cleavage sites in divergent CoVs together with their mapping on crystalized tertiary structures can help find regions that may be critical in the context of the virus-host interactions and which can be targeted for future laboratory testing. Aside from the two viruses reported in this study, two additional luchacoviruses [UMN2020/mouse/2018/USA (62) and P83/plateau pika/China (69)] are highly divergent (Fig. 2; Fig. S2), thus showing that luchacoviruses are highly diverse and surveillance targeting both conserved and variable regions like the S1 may be essential to understand CoV diversity and evolution. With the results shown in this study, we increased the data set of sequences from luchacoviruses extending the known host range of these viruses and showing possible regions of interest in the S protein for future surveillance.

We did not find recombination breakpoints in the genomic region analyzed between the luchacovirus sequences available to date. However, possible recombination breakpoints have been detected in the ORF1ab among viruses of groups 1, 2, and 3 (58). This may indicate that these viruses are infecting different hosts at the same time. In this study, we are reporting divergent luchacoviruses in three species of widely spread meso-carnivore, whose diet includes rabbits and rodents, and thus, it would be interesting to assess if the predator-prey interactions play a role in the spread of CoVs in wild populations. However, we did not find luchacoviruses in the rodent and rabbit samples screened in this study (Table S1) possibly due to our smaller sample size (36 samples) compared to that of the meso-carnivore (300 samples) in which we found luchacoviruses in 1% (3 positive samples). In previous studies from China (26) and Europe (70), it has been shown that luchacoviruses are less prevalent than rodent and rabbit betacoronaviruses (0.8% to 1.1% of positive samples were luchacoviruses compared to 6.1% to 11.8% for rodent betacoronaviruses). It is necessary to continue the surveillance of wildlife (including meso-carnivores, rodents, and lagomorphs) in the U.S. to increase our knowledge of the viral diversity hosted in native species. Since all the recovered viral sequences herein reported were obtained from fecal samples, we show that it is possible to perform a successful surveillance of alphacoronaviruses in meso-carnivore in a non-invasive manner.

DATA AVAILABILITY

Sequences obtained in this study are deposited in GenBank (accession numbers OQ756331-35 and OR428266-7).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jvi.00829-23.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.