An atlas of sarbecovirus infectivity across diverse ACE2 orthologs Authors: Yeqing Sun, Ziqi Cheng, Xi Wu, Kefang Liu, Lin Wang, Dongchun Yang, Weijin Huang, Jianhui Nie Published: 2026 ABSTRACT SummarySarbecoviruses pose a substantial zoonotic threat, yet the determinants governing their host range remain incompletely defined. To systematically investigate host entry mechanisms, we generated a panel of vesicular stomatitis virus (VSV) pseudoviruses representing 46 sarbecoviruses and assessed their infectivity across 66 angiotensin-converting enzyme 2 (ACE2) orthologs. This analysis produced a high-resolution host-entry landscape with direct implications for cross-species transmission risk. We found that all severe acute respiratory syndrome coronavirus 1 (SARS-CoV-1) isolates, WIV1, and pangolin-derived sarbecoviruses demonstrated broad ACE2 usage, whereas no sarbecovirus utilized alternative entry factors including neuropilin-1, CD147, AXL, or TMEM106B. Notably, ACE2 orthologs from Rhinolophus affinis and Miniopterus schreibersii supported infection by nearly all clade 1 sarbecoviruses, identifying highly permissive host receptors. We also identified residues 27 and 42 of ACE2 as critical determinants of sarbecovirus host tropism. Together, these findings define host entry constraints relevant to receptor-based risk assessment and inform surveillance and pandemic preparedness.Keywords: sarbecovirus, ACE2 orthologs, host range, receptor usage, viral entry, cross-species transmission Introduction Coronaviruses constitute a diverse group of RNA viruses that circulate widely in nature; their extensive genetic variability and ability to cross-species barriers continue to create global public health challenges.1,2 In the past two decades, three zoonotic coronaviruses—severe acute respiratory syndrome coronavirus (SARS-CoV-1), Middle East respiratory syndrome coronavirus (MERS-CoV), and SARS-CoV-2—have caused major outbreaks and demonstrated the pandemic potential of viruses of animal origin.2,3,4,5 These events underscore the importance of understanding the molecular and ecological factors that enable viral spillover across species. Both SARS-CoV-1 and SARS-CoV-2 belong to the sarbecovirus subgenus of Betacoronaviruses.6,7 Most sarbecoviruses, including these two pathogens, use angiotensin-converting enzyme 2 (ACE2) as their cellular receptor.8,9,10,11 However, certain viral clades rely on distinct or unidentified receptors, indicating alternative entry mechanisms.12,13,14 A species’ susceptibility to sarbecovirus infection is primarily determined by molecular compatibility between the viral spike (S) protein receptor-binding domain (RBD) and host ACE2.15 This RBD-ACE2 interaction, defined by its affinity for species-specific ACE2 orthologs, remains the principal determinant of viral host range and tissue tropism.16 Previously, sarbecoviruses were classified into three major branches and at least four sub-branches based on their RBD characteristics and ACE2 utilization efficiency.10,17 Clade 1 includes ACE2-utilizing viruses and is further divided into clade 1a (SARS-CoV-1 and bat- or civet-derived strains)18,19,20 and clade 1b (SARS-CoV-2 and related bat or pangolin viruses),21,22,23 both of which can efficiently infect cells that express human or other mammalian ACE2.11 Clade 2 comprises most known sarbecoviruses,13 characterized by two deletions in the RBD and uncertain receptor usage,8 although some can enter human cells under high-trypsin conditions.17,24 Clade 3 includes SARS-like coronaviruses from African and European bats,25,26,27 featuring smaller RBD deletions and weaker ACE2 binding; certain members can partially use ACE2, and mutations such as residue 493 in BtKY72 may enable engagement of human ACE2.28,29 However, a systematic understanding of sarbecovirus cross-species infectivity has not been achieved. Bats represent the most important natural reservoir of coronaviruses and harbor extensive genetic diversity across multiple lineages of Alphacoronaviruses and Betacoronaviruses. Surveillance studies in China, Southeast Asia, and Africa have identified hundreds of sarbecovirus-related sequences from diverse Rhinolophus (bat) species, many of which show close genetic similarity to SARS-CoV-1 or SARS-CoV-2.17,23,30 Experimental evidence further indicates that many bat ACE2 orthologs can function as entry receptors for sarbecoviruses that infect humans.18,22 Spillover events likely occur when bats interact with intermediate hosts, enabling viral recombination or adaptive mutations—particularly within the RBD—that broaden host range and permit infection of human cells.30 Civets, for example, served as amplifying hosts for SARS-CoV-1.2,19 Similar zoonotic dynamics contributed to the emergence of SARS-CoV-2.31,32 These findings illustrate the central role of bats in the long-term maintenance and evolutionary diversification of sarbecoviruses; they underscore the need for a systematic, comparative analysis of receptor compatibility across bat species. Despite extensive research, a comprehensive landscape of sarbecovirus infectivity across diverse ACE2 orthologs remains unavailable. Here, we systematically mapped the ability of 46 representative sarbecoviruses to utilize 66 ACE2 orthologs from different species—including 48 from bats—using a pseudovirus-based infection system.10,33 This work establishes an infectivity atlas that defines cross-species receptor usage, provides insights into the receptor-usage potential of bat-derived sarbecoviruses, and supports surveillance strategies that target high-risk hosts. Results Systematic evolution and sequence characteristics of sarbecoviruses and ACE2 orthologs To establish an evolutionary framework for subsequent functional analyses, we first examined the phylogenetic relationships and sequence characteristics of representative sarbecoviruses and ACE2 orthologs across host species. In total, we selected 46 sarbecoviruses, along with 66 ACE2 orthologs (18 from non-bat mammals and 48 from bats) and four host factors previously reported to participate in sarbecovirus entry (AXL, CD147, TMEM106B, and NRP1).34,35,36,37 Phylogenetic reconstruction based on the RBD of the S protein divided the sarbecoviruses into three major clades (clades 1–3), and clade 1 was further separated into subclades 1a and 1b (Figure 1A). Among the 46 viruses analyzed, 31 originated from bats, two arose from pangolins, 10 were derived from humans, and three originated from civets. Consistent with the S protein similarity matrix, SARS-CoV-2-related viruses (RaTG13 and pangolin-CoVs) formed a tight high-identity cluster (90%–97%), SARS-CoV-1-related viruses showed extremely high internal identity (∼99%), whereas HKU3-like clade 2 viruses displayed the lowest RBD identity relative to clade 1 viruses (∼74%–76%). These patterns reinforce the presence of three robust evolutionary groups (Figure 1A). The sources of all sarbecovirus S proteins used in this study are summarized in Table S1. Phylogenetic and sequence landscape of representative sarbecoviruses and ACE2 orthologs (A) Phylogenetic relationships and pairwise amino acid identity among 46 representative sarbecoviruses based on the RBD. Left image shows a maximum-likelihood phylogenetic tree that divides sarbecoviruses into three major clades (clades 1–3); clade 1 is further separated into subclades 1a and 1b. Colored boxes indicate the natural host of each virus (human, pangolin, civet, or distinct bat species). Heatmap (right) displays pairwise S protein sequence identity, where red indicates higher identity. (B) Sequence logo comparison of the receptor-binding motif (RBM) across the four sarbecovirus clades. Red boxes highlight distinct deletions and conserved residues. Clade 2 viruses harbor a double-segment deletion and clade 3 viruses display a single-segment deletion; clade 1 viruses retain a complete RBM with diversified key residues. (C) Phylogenetic tree of 66 ACE2 orthologs from 18 non-bat mammals and 48 bat species representing 11 bat families. Inner ring depicts ACE2 phylogeny, and outer colored sectors indicate the taxonomic distribution of host species. Comparative sequence alignment showed that clade-specific variation is primarily concentrated within the receptor-binding motif (RBM). Clade 3 viruses contain a single-segment deletion, whereas clade 2 viruses exhibit a double-segment deletion, both located in the RBM loop that contacts ACE2 (Figure 1B). In addition to these deletions, clade 2 and clade 3 viruses share low S protein identity with ACE2-using clade 1 viruses (∼75%), consistent with markedly reduced or altered ACE2 usage relative to clade 1 viruses. In contrast, clade 1 retains an intact RBM with diverse amino acid residues at key binding sites, consistent with its broader ACE2 usage noted in previous studies. To contextualize viral evolution with host receptor diversity, we next analyzed ACE2 orthologs from 66 species representing 11 globally distributed bat families33 and 18 non-bat mammals (Figure 1C). The ACE2 phylogeny revealed extensive evolutionary divergence among bat lineages, providing a broad structural basis for cross-species infection mapping. The source information for all ACE2 proteins from different species used in this study is provided in Tables S2 and S3. An infectivity atlas of 46 sarbecoviruses across 66 ACE2 orthologs Building upon the phylogenetic framework described previously, we next characterized the functional infectivity profiles of 46 sarbecoviruses across 66 ACE2 orthologs to define sequence-function relationships. Flow-cytometric analysis was used to assess transfection-associated receptor expression across target cells, providing a reference for comparison among receptor-virus combinations (Figures S1, S2A, and S2B). Western blot analysis further confirmed the expression of human, Rhinolophus affinis (R. affinis), Miniopterus schreibersii (M. schreibersii), and Mastomys natalensis (M. natalensis) ACE2 in transfected BHK-21 cells (Figure S2C). A panel of 46 sarbecovirus pseudoviruses was successfully generated, and western blot analysis of representative strains selected from the major phylogenetic branches showed broadly comparable cellular spike expression and pseudovirus-associated spike incorporation efficiencies among the tested constructs (Figure S2D). Pseudovirus input was further quantified by quantitative reverse-transcription PCR (RT-qPCR) (Figure S2E), enabling a systematic infection screen across 66 ACE2 orthologs and four additional proposed entry factors (NRP1, CD147, AXL, and TMEM106B). Functionally, clade 1 viruses displayed the broadest cross-species ACE2 usage (Figure 2A). Within clade 1b, two pangolin-associated sarbecoviruses, pangolin coronavirus (PCoV)-GD (GD/1/2019) and PCoV-GX (GX/P2V/2017), share high S protein identity with SARS-CoV-2 (89.6% and 92.3%, respectively) and exhibit broad ACE2 compatibility. Notably, PCoV-GD shows the broadest tropism, efficiently infecting nearly all tested non-bat mammalian ACE2s and most bat ACE2-expressing cells (Figure 2A). Despite sharing 97.4% identity across the full S protein and 89.3% within the RBD with SARS-CoV-2, RaTG13 displayed a more restricted host range. These findings imply that subtle differences within the RBM underlie the expanded host range of SARS-CoV-2. An infectivity atlas of sarbecoviruses across ACE2 orthologs (A) Cross-species ACE2 usage profiles of sarbecoviruses. Vesicular stomatitis virus (VSV)–based reporter viruses pseudotyped with the S proteins of 46 sarbecoviruses were used to infect BHK-21 cells overexpressing 66 ACE2 orthologs along with four additional host factors (NRP1, CD147, AXL, and TMEM106B). Viruses are ordered according to phylogenetic relationships inferred from RBD sequences, and bat species from different families are color-coded. (B) Correlation between ACE2 usage in different species and human ACE2. The heatmap shows one of two independent experiments as a representative result. In each experiment, pseudoviruses were generated in an independent round of production and used to infect separately prepared target cells, with two technical replicates per condition. Pearson’s correlation coefficients (r) were calculated from mean relative luminescence units (RLU). Although S protein sequence identity exceeded 87% within clade 1a, members segregated into two distinct ACE2-usage patterns. Human SARS-CoV-1 isolates (GZ-C, Urbani, HGZ8L1-A, and GD01) efficiently used ACE2 orthologs from Pteropodidae, Myotis, Phyllostomidae, and Rhinolophidae species (Figure 2B). In contrast, most bat- or civet-derived SARS-like sarbecoviruses displayed much narrower ACE2 usage across bat species relative to human SARS-CoV-1 strains. Among these, WIV1 exhibited a distinct ACE2-recognition profile compared with other clade 1a members such as LYRa11: in addition to infecting through the ACE2 of its natural host Rhinolophus sinicus (R. sinicus), WIV1 efficiently used ACE2 orthologs from multiple Rhinolophus and Myotis species (Figure 2A), bridging human and bat SARS-like lineages. Clade 2 viruses, which constitute most known sarbecoviruses, contain two deletion segments within the RBD (Figure 1B) and do not mediate ACE2-dependent entry, consistent with prior reports.10,13 An ACE2-usage heatmap confirmed the absence of infection by clade 2 viruses across all ACE2 orthologs and other entry factors. Although high-trypsin conditions can permit limited entry of certain clade 2 pseudoviruses, no alternative receptor has been identified to date.14,17 Clade 3 viruses demonstrated intermediate phenotypic features, with RBD sequence identities of 72.3%–82.7%. BtKY72 used ACE2 orthologs from only a few Rhinolophus species and did not engage human ACE2, whereas Khosta-2 exhibited broader ACE2 tropism and particularly strong infection through rabbit ACE2—over 10-fold higher than that observed through human ACE2 (Figure 2A), consistent with lineage-specific adaptation. Several host factors—including NRP1, CD147, AXL, and TMEM106B—have previously been proposed to facilitate SARS-CoV-2 entry based on biochemical binding assays, overexpression analyses, or single-virus reports.34,36,37,38 These studies suggested that certain host factors, beyond ACE2, may modulate viral entry in specific viral or cellular contexts. However, across all sarbecoviruses tested in our pseudovirus panel, none of these host factors supported detectable infection, indicating that none of them functions as an independent entry receptor for any of the sarbecoviruses tested under our assay conditions (Figure 2A). The source information for the four human host factors used in this study is provided in Table S4. Correlation analyses showed that ACE2 orthologs from rabbits, cattle, sheep, raccoon dogs, badgers, Siberian chipmunks, and pigs exhibited a strong correlation with human ACE2 usage (r > 0.8; Figure 2B), suggesting that these domesticated or semi-wild species may represent candidate intermediate hosts for further investigation. Among bats, only R. affinis and M. schreibersii ACE2 orthologs showed a strong correlation with human ACE2 usage. Given their broad distribution across Southeast Asia, Europe, and North Africa—regions that overlap with human habitats33—these species represent important candidates for ongoing surveillance. Differential analysis of ACE2 usage among bat species ACE2 serves as the primary receptor for sarbecoviruses, including SARS-CoV-1 and SARS-CoV-2. Variations in ACE2 sequence and conformation among bat species directly influence viral recognition and thus determine host range. Hierarchical clustering of infectivity profiles across 48 bat ACE2 orthologs revealed pronounced heterogeneity, even within individual families (Figure 3A). Almost all sarbecoviruses capable of using human ACE2 also efficiently engaged M. schreibersii and R. affinis ACE2 orthologs; ACE2 orthologs from Rhinolophus pearsonii, Rhinolophus pusillus, and Rhinolophus ferrumequinum supported only limited infection. Comparison of ACE2 usage and sequence determinants among bat species (A) Hierarchical clustering of sarbecovirus infectivity across 48 bat ACE2 orthologs. Heatmap shows infectivity of sarbecovirus pseudotypes in BHK-21 cells overexpressing ACE2 orthologs from different bat species; bat families are color-coded. (B and C) Correlation of entry efficiencies of clade 1-pseudotyped sarbecoviruses between human ACE2 and bat ACE2 orthologs (n = 20). Scatterplots show relative luminescence units measured in BHK-21 cells expressing human ACE2 versus Rhinolophus affinis 9479 ACE2 (B) or Miniopterus schreibersii ACE2 (C). (D and E) Surface plasmon resonance (SPR) sensorgrams for M. schreibersii ACE2 binding to the SARS-CoV-1 RBD (D) and SARS-CoV-2 RBD (E). (F and G) SPR sensorgrams for human ACE2 binding to the SARS-CoV-1 RBD (F) and SARS-CoV-2 RBD (G), providing a benchmark for comparison with M. schreibersii ACE2. Members of the family Rhinolophidae—particularly R. affinis and R. sinicus—are well recognized as natural reservoirs of sarbecoviruses. Our results reinforce this recognition by showing that ACE2 orthologs from both species function as efficient entry receptors. Human-derived pseudotypes such as SARS-CoV-1 efficiently used ACE2 orthologs from Pteropodidae species and, to a lesser extent, from Vespertilionidae and Phyllostomidae, suggesting that these groups could act as bridge hosts in cross-species transmission. R. affinis has long been regarded as a major natural host of sarbecoviruses, whereas M. schreibersii has received far less attention.39,40,41 In our ACE2 usage screen, nearly all ACE2-dependent pseudotyped sarbecoviruses efficiently used M. schreibersii ACE2 for entry. The infectivity profiles of pseudotyped sarbecoviruses using R. affinis and M. schreibersii ACE2 orthologs showed strong correlations with the profiles of pseudotyped sarbecoviruses using human ACE2 (r = 0.8314 and 0.8777, p < 0.0001; Figures 3B and 3C). To assess whether these functional similarities reflect molecular binding compatibility, we performed surface plasmon resonance assays, which demonstrated high-affinity interactions between M. schreibersii ACE2 and the RBDs of SARS-CoV-1 and SARS-CoV-2 (KD = 8.14 × 10−8 M and 8.06 × 10−7 M; Figures 3D and 3E). As a benchmark, we further measured the interactions of SARS-CoV-1 and SARS-CoV-2 RBDs with human ACE2 (KD = 1.69 × 10−9 M and 1.34 × 10−9 M, respectively; Figures 3F and 3G), which placed the binding affinities of M. schreibersii ACE2 in the context of known human receptor usage. Key residues of Miniopterus ACE2 tropism for sarbecoviruses Despite its close evolutionary relationship with M. schreibersii, M. natalensis ACE2 was entirely non-permissive to all pseudotyped sarbecoviruses tested (Figure 2A). Sequence alignment revealed only seven amino acid differences between these two Miniopteridae ACE2 orthologs; three substitutions (positions 27, 31, and 42) were located directly at the predicted RBD-binding interface (Figure 4A). Accordingly, we selected these residues for functional interrogation via reciprocal single-site mutagenesis. Single-residue determinants of Miniopterus ACE2 tropism for clade 1 sarbecoviruses (A) Alignment of M. schreibersii and M. natalensis ACE2 sequences. Positions 27, 31, and 42 (arrowheads) lie within the predicted RBD-binding interface. (B and C) Circular heatmaps of pseudotyped sarbecovirus entry mediated by wild-type and mutant M. schreibersii and M. natalensis ACE2. Concentric rings represent ACE2 orthologs, and sectors correspond to the 46 pseudotyped viruses. Color intensity (blue for M. schreibersii, red for M. natalensis) indicates relative infectivity. Comparison of the two images shows that residues 27 and 42 have the strongest influence on Miniopterus ACE2 usage. The heatmap shows one of two independent experiments as a representative result. In each experiment, pseudoviruses were generated in an independent round of production and used to infect separately prepared target cells, with two technical replicates per condition. (D) Molecular mechanics generalized Born surface area binding energy calculations for M. schreibersii ACE2-RBD and corresponding mutant complexes. (E and F) AlphaFold 3-based structural model of the M. schreibersii ACE2-RBD interface. Image in (E) shows overall view of SARS-CoV-1 RBD (orange) bound to M. schreibersii ACE2 (green). Image in (F) presents an enlarged interface view highlighting key ACE2 residues and their interacting SARS-CoV-1 RBD residues. Molecular dynamics (MD) simulations indicated that the binding-free energy between the SARS-CoV-1 RBD and M. schreibersii ACE2, calculated using the molecular mechanics generalized Born surface area method, was −30.85 ± 2.55 kcal/mol. Replacement of I27 with K27 in M. schreibersii ACE2 raised the binding-free energy to −25.86 ± 2.87 kcal/mol, an increase of approximately 5 kcal/mol relative to the energy of the wild-type complex. Similarly, replacement of K42 with E42 produced a binding-free energy of −23.99 ± 2.09 kcal/mol, corresponding to an increase of approximately 6.86 kcal/mol compared with the energy of the wild-type complex. In contrast, the N31G mutation had minimal influence on binding, yielding a free energy similar to that of the wild type. These results suggest that the I27K and K42E substitutions in M. schreibersii ACE2 both reduce its binding affinity for the SARS-CoV-1 RBD (Figure 4B). Furthermore, the representative model from MD simulations showed that M. schreibersii ACE2 N31 forms a hydrogen bond with SARS-CoV-1 RBD Y442 (Figure 4C). Additionally, a salt bridge between M. schreibersii ACE2 D38 and K42 appears to stabilize their respective interactions with SARS-CoV-1 RBD Y436. When K42 is replaced with E42, however, the charge reversal—together with a conformational shift in the D38 side chain—disrupts the interaction network involving M. schreibersii ACE2 D38 and E42, and SARS-CoV-1 RBD Y442. This disruption reduces the binding affinity of M. schreibersii ACE2 for SARS-CoV-1 (Figure 4D). Upon further investigation concerning the role of these residues in sarbecovirus infection, entry assays demonstrated clear residue-specific and directionally consistent effects (Figures 4E and 4F). Consistent with the MD simulation results, I27K in M. schreibersii ACE2 selectively impaired the entry of civet-derived SARS-CoV-1, N31G had minimal impact, and K42E broadly eliminated usage by nearly all tested pseudotypes. Conversely, in M. natalensis ACE2, K27I and E42K conferred measurable entry competence for multiple SARS-CoV-1 pseudotypes, whereas G31N remained non-functional. Taken together, these results identify residues 27 and 42 as the principal molecular determinants that govern Miniopterus ACE2 recognition by clade 1 sarbecoviruses. Reciprocal substitutions at these positions either restore viral usage to M. natalensis ACE2 or abolish usage of M. schreibersii ACE2, underscoring their decisive roles in shaping host-range restriction within Miniopteridae. Discussion A subset of clade 1 viruses—including SARS-CoV-1, WIV1, PCoV-GD, and PCoV-GX—displayed broad ACE2 compatibility, effectively engaging receptors from humans, along with multiple mammalian and bat species. This observation supports prior reports that bat-derived sarbecoviruses such as WIV1 and SHC014 can directly infect primary human airway epithelial cells,42 suggesting that these viruses possess substantial cross-species entry potential at the receptor-engagement level, rather than directly demonstrating productive cross-species infection without adaptive evolution. Such broad receptor tropism suggests that several circulating bat sarbecoviruses harbor receptor-usage features compatible with cross-species entry and therefore warrant closer surveillance, rather than directly indicating preadaptation for zoonotic emergence. The strong correlation we observed between bat ACE2 permissiveness—particularly regarding ACE2 orthologs from R. affinis and M. schreibersii—and human ACE2 usage (r > 0.8) suggests that compatibility with certain bat ACE2 orthologs might serve as a proxy for human ACE2 usage at the receptor level. This interpretation aligns with recent observations that bat sarbecoviruses capable of using ACE2 from multiple bat species tend to exhibit greater cross-species potential.22 Monitoring ACE2 sequence diversity in wildlife may therefore inform receptor-based surveillance prioritization for zoonotic sarbecoviruses. Given the extensive diversity and receptor variability among bats, identifying species with ACE2 structures resembling human ACE2 may help prioritize hosts for further receptor-based evaluation. Notably, although the binding affinities of M. schreibersii ACE2 for the SARS-CoV-1 and SARS-CoV-2 RBDs were lower than those of human ACE2, they remained consistent with the efficient receptor usage observed in our entry assays. Consistent with prior phylogenetic and ecological evidence, we identified R. affinis and M. schreibersii as key host species with ACE2 orthologs that can be used by nearly all clade 1 strains. These bats may represent host contexts in which spike variants with broad ACE2 recognition are more likely to be maintained. Previous surveillance has also implicated Rhinolophus bats as reservoirs of SARS-related coronaviruses with high sequence identity to both SARS-CoV-1 and SARS-CoV-2,23 reinforcing their central role in coronavirus ecology. At the molecular level, our mutational analyses identified residues 27 and 42 in Miniopteridae ACE2 as critical determinants of sarbecovirus entry, highlighting structural features that underlie species-dependent receptor usage. In addition to ACE2, NRP1, CD147, AXL, and TMEM106B have been proposed to participate in SARS-CoV-2 entry, but their reported roles differ substantially from that of ACE2.34,35,36,37,43,44,45,46 NRP1 has mainly been described as an ACE2-associated host factor in furin-cleavage-dependent settings, whereas AXL has also been interpreted as a phosphatidylserine receptor that enhances infection in low-ACE2 cells rather than acting as an ACE2-like primary receptor.34,35,45 TMEM106B was first identified as a lysosomal host factor and only later proposed to mediate ACE2-independent entry in specific SARS-CoV-2 contexts, while the role of CD147 remains controversial because follow-up studies did not support direct CD147/basigin-spike binding.36,37,43,44,46 By contrast, our study used a vesicular stomatitis virus (VSV)-based comparative pseudovirus system to test whether these factors are sufficient, upon overexpression in heterologous cells, to independently support detectable entry across a broad sarbecovirus panel. Under these conditions, NRP1, CD147, AXL, and TMEM106B did not support detectable entry for any sarbecovirus tested here, indicating that none of these factors functions as an independent entry receptor in our assay system. However, context-dependent accessory or alternative roles in specific SARS-CoV-2 settings cannot be excluded. The receptor-usage atlas presented here offers a mechanistic framework for receptor-based risk assessment and surveillance prioritization of sarbecoviruses. This pseudovirus-based system complements field surveillance by identifying species with ACE2 profiles that support viral entry, helping to prioritize wildlife species for monitoring and to identify candidate hosts for follow-up ecological and virological investigation. Limitations of the study A key limitation of this study is its reliance on pseudoviruses constructed from available sequences, which may not fully represent the diversity of natural sarbecoviruses. Although receptor compatibility constitutes a major barrier to cross-species transmission, additional host factors—including protease availability, replication competence, and immune evasion—also influence susceptibility. Thus, efficient ACE2 engagement alone does not ensure productive infection in vivo. Future work should incorporate authentic virus systems and animal challenge models to validate these receptor-level findings. Resource availability Lead contact Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Jianhui Nie (niejianhui@nifdc.org.cn). Materials availability All unique reagents generated in this study are available from the lead contact with a completed materials transfer agreement. Data and code availability •All sequence data used in this study are publicly available through GenBank. All data supporting the findings of this study are available within the article and supplemental information. The original data generated in this study have been deposited at Figshare (https://doi.org/10.6084/m9.figshare.32981705) and are publicly available as of the date of publication. The corresponding accession numbers and DOI are listed in the key resources table. All sequence data used in this study are publicly available through GenBank. All data supporting the findings of this study are available within the article and supplemental information. The original data generated in this study have been deposited at Figshare (https://doi.org/10.6084/m9.figshare.32981705) and are publicly available as of the date of publication. The corresponding accession numbers and DOI are listed in the key resources table. •No custom code was generated in this study. No custom code was generated in this study. •Software and algorithms used for data analysis, visualization, molecular modeling, MD simulations, sequence alignment, phylogenetic reconstruction, and flow cytometry analysis are listed in the key resources table. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Software and algorithms used for data analysis, visualization, molecular modeling, MD simulations, sequence alignment, phylogenetic reconstruction, and flow cytometry analysis are listed in the key resources table. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Acknowledgments This work was supported by the Natural Science Foundation of China (grant no. 82172244) and the Major Project of Guangzhou National Laboratory (grant no. GZNL2024A01019). Author contributions Conceptualization, J.N., D.Y., and W.H.; methodology, J.N., Y.S., and D.Y.; investigation, Y.S., Z.C., J.N., D.Y., and K.L.; formal analysis, Y.S., D.Y., and L.W.; data analysis, J.N., Y.S., L.W., D.Y., and Z.C.; visualization, J.N., Y.S., D.Y., and K.L.; writing – original draft, Y.S., Z.C., and X.W.; writing – review and editing, J.N., Y.S., K.L., and L.W.; supervision, J.N., D.Y., and W.H.; funding acquisition, J.N. and W.H. Declaration of interests The authors declare no competing interests. STAR★Methods Key resources table REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesGAPDH Mouse mAbABclonalCat# AC033; RRID: AB_2769570Mouse anti-DDDDK-Tag mAbABclonalCat# AE005; RRID: AB_2770401Mouse anti-His-Tag mAbABclonalCat# AE003; RRID: AB_2728734Anti-VSV-M [23H12] antibodyKerafastCat# EB0011; RRID: AB_2734773Bacterial and virus strainsG∗ΔG vesicular stomatitis virusKerafastN/AChemicals, peptides, and recombinant proteinsLipofectamine 3000InvitrogenCat# L3000015High-glucose DMEMHyCloneCat# SH30243.01BFetal bovine serumHyCloneCat# SH30406.05Penicillin-streptomycinGibcoCat# 15140122PBSHyCloneCat# SH30256.01B0.05% Trypsin-EDTAGibcoCat# 25300054Bright-Glo™PerkinElmerCat# 6066769Opti-MEMGibcoCat# 31985088PrimeSTAR® Max DNA PolymeraseTakaraCat# R045BSARS-CoV Spike/RBD ProteinSino BiologicalCat# 40150-V05HSARS-CoV-2 (2019-nCoV) Spike RBD-mFc Recombinant ProteinSino BiologicalCat# 40592-V05HHuman ACE2 ProteinSino BiologicalCat# 10108-H08HM. schreibersii ACE2 recombinant proteinThis paper/Sino BiologicalN/ABD® Shutdown SolutionBDCat# 334224BD® FACS FlowBDCat# 342003BD® FACS Cleaning SolutionBDCat# 340345Critical commercial assaysQIAamp Viral RNA Mini KitQiagenCat# 52906QIAGEN Plasmid Plus Maxi KitQiagenCat# 12965SuperScript III First-Strand Synthesis System and RT-qPCR KitInvitrogenCat# 18080–051TB Green Premix Ex Taq IITaKaRaCat# RR820ADeposited dataData generated in this studyThis paper; Figsharehttps://doi.org/10.6084/m9.figshare.32981705Sequence datasets used in this studyGenBank/NCBIGenBank accession numbers: Tables S1–S4Experimental models: Cell Lines293T cellsATCCCRL-3216BHK-21 cellsATCCCCL-10OligonucleotidesMiniopterus ACE2 point mutation primer(s)This paper/SinoGenoMaxTable S5Recombinant DNApcDNA3.1-based sarbecovirus S-expression plasmidsThis paper/General BiolTable S1pIRES2-EGFP-based ACE2 ortholog expression plasmidsThis paper/General BiolTables S2 and S3pIRES2-EGFP-based NRP1, CD147, AXL, and TMEM106B expression plasmidsThis paper/General BiolTable S4Software and algorithmsFlowJo v10.6TreeStar Inc.https://www.flowjo.com/Prism 9GraphPadhttps://www.graphpad.com/Amber24AmberMDhttps://ambermd.org/AlphaFold 3Google DeepMindhttps://alphafoldserver.com/MUSCLEMUSCLEhttps://www.drive5.com/muscle/MEGAMEGA Softwarehttps://www.megasoftware.net/RAxMLRAxMLhttps://cme.h-its.org/exelixis/web/software/raxmlEvolviewEvolviewhttps://www.evolgenius.info/evolview/#/treeview Experimental model and study participant details Cell lines and culture conditions 293T cells(ATCC CRL-3216™) were used for pseudovirus production, whereas BHK-21 cells (ATCC CCL-10™) were used for transient receptor expression and pseudovirus infection assays. Cells from various species overexpressing ACE2 and four other CoV-associated receptors were prepared as follows: transiently expressed cell lines were generated by transfecting BHK-21 cells with 30 μg of ACE2 plasmid using Lipofectamine 3000 (Invitrogen) in T-75 flasks. All cells were maintained in high-glucose Dulbecco’s Modified Eagle Medium (DMEM; HyClone, Logan, UT, USA) supplemented with 10% fetal bovine serum (Gibco, Billings, MT, USA) and 1% penicillin–streptomycin (Gibco) under 5% CO2 in a humidified incubator at 37°C. Transiently transfected ACE2-expressing cell lines were used for experiments after 36 h of incubation at 37°C with 5% CO2. Method details Plasmid construction Codon-optimized genes encoding the S proteins of 46 representative sarbecoviruses were synthesized according to mammalian codon usage. To enhance pseudotyping efficiency of VSV, the C-terminal 19 amino acids of each S protein were deleted (Δ19) before cloning into the BamHI and XhoI sites of the pcDNA3.1 expression vector.47 Codon-optimized genes encoding 66 ACE2 orthologs and four host factors previously implicated in sarbecovirus entry (NRP1, CD147, AXL, and TMEM106B) were similarly cloned into the pIRES2-EGFP vector between the BglII and BamHI restriction sites for transient expression in mammalian cells. All genes, mutation sites, and corresponding primers used in this study are listed in Tables S1–S5. Pseudovirus production Pseudotyped viruses incorporating S proteins from sarbecoviruses were generated using a procedure described in our previous work.48 One day before transfection, 293T cells were prepared at a concentration of 5–7 × 105 cells/mL. A 15-mL aliquot of the cell suspension was transferred into a T-75 flask and incubated overnight at 37°C in a humidified incubator with 5% CO2. When the cells reached 70%–90% confluence, the medium was removed. The cells were then infected with 15 mL of G∗ΔG vesicular stomatitis virus (VSV G pseudotyped virus, Kerafast) at 7.0 × 104 TCID50/mL. Simultaneously, 30 μg of sarbecovirus S protein expression plasmid were transfected using Lipofectamine 3000 (Invitrogen), in accordance with the manufacturer’s instructions. Cells were then incubated at 37°C with 5% CO2. After 4–6 h, the medium was removed, and the cells were washed twice with phosphate-buffered saline containing 2% fetal bovine serum. Fresh complete DMEM (15 mL) was added, and the flasks were returned to 37°C with 5% CO2. Supernatants containing pseudotyped virus were collected at 24 and 48 h, filtered through 0.45-μm membranes (Millipore, Cat# SLHP033RB), and stored at −80°C in 2-mL aliquots until use. Quantification of pseudotyped viral particles by reverse transcription quantitative polymerase chain reaction (RT-qPCR) Viral RNA was extracted from 140 μL of pseudotyped virus using the QIAamp Viral RNA Mini Kit (QIAGEN, Cat# 52906). Reverse transcription was performed using the SuperScript III First-Strand Synthesis System and RT-qPCR Kit (Invitrogen, Cat# 18080-051). Viral RNA quantification was conducted by RT-qPCR with TB Green Premix Ex Taq II (TaKaRa, Cat# RR820A), in accordance with the manufacturer’s instructions. The VSV P protein gene was cloned into the pcDNA3.1(+) vector and used as a standard to calculate viral copy numbers. Western blot analysis For ACE2 expression analysis, BHK-21 cells transiently expressing human, R. affinis, M. schreibersii, and M. natalensis ACE2 were harvested at 36 h post-transfection. For spike protein expression analysis, 293T producer cells transfected with representative sarbecovirus S protein plasmids from the major phylogenetic branches were collected at 36 h post-transfection. For pseudovirus-associated spike detection, pseudovirus-containing supernatants were collected and processed for western blot analysis. Proteins were separated by SDS–PAGE, transferred onto PVDF membranes, and probed with GAPDH Mouse mAb (ABclonal, Cat# AC033, 1:100000), Mouse anti-DDDDK-Tag mAb (ABclonal, Cat# AE005, 1:10000), Mouse anti-His-Tag mAb (ABclonal, Cat# AE003, 1:5000), or Anti-VSV-M [23H12] Antibody (Kerafast, Cat# EB0011, 1:1000). ACE2 expression was detected using the DDDDK-Tag antibody, spike protein was detected using the His-Tag antibody, GAPDH served as an internal control for cell lysates, and VSV-M served as an internal control for pseudovirus preparations. Receptor transfection and flow cytometry assays pIRES2-EGFP vectors encoding ACE2 genes or plasmids expressing four host factors previously associated with sarbecovirus entry (NRP1, CD147, AXL, and TMEM106B) were transiently transfected into BHK-21 cells using Lipofectamine 3000 (Thermo Fisher Scientific). After 36 h, cells were harvested and washed three times with phosphate-buffered saline. Flow cytometry was performed on a BD FACSCalibur cytometer (BD Biosciences). Enhanced green fluorescent protein (EGFP) fluorescence served as a transfection-associated reference across constructs, rather than a strict quantitative surrogate for ACE2 protein abundance. Data were analyzed and visualized using FlowJo v10.6 software (TreeStar Inc., Ashland, OR, USA). Pseudotype virus infection experiment After quantification by RT-qPCR, pseudovirus stocks were diluted to equal particle numbers and dispensed into 96-well plates (100 μL per well). ACE2-overexpressing cells were detached with trypsin and seeded into each well at 4 × 105 cells/mL. Non-transfected BHK-21 cells served as negative controls. After 24 h of incubation at 37°C with 5% CO2, infection efficiency was measured by chemiluminescence. Procedures followed the method in our previous report.47 Briefly, 100 μL of luciferase substrate (PerkinElmer) were added to each well, and the plates were gently shaken for 2 min at room temperature. Luminescence was measured using a luminometer (PerkinElmer) after the mixture had been transferred to a white assay plate. All samples were analyzed with two technical replicates per condition in two independent experiments. For the heatmap analyses shown in Figures 2 and 4, one of the two independent experiments is shown as a representative result. Protein expression and purification For M. schreibersii ACE2 protein expression, the recombinant plasmid and TF1 transfection reagent (Sino Biological, China) were mixed at a 1:10 ratio (DNA mass: TF1 volume) and transfected into HEK293 cells, in accordance with the manufacturer’s protocol. Cells were maintained at 37°C in a humidified incubator with 5% CO2 and shaken at 175 rpm for 5 days. Serum-free medium (Sino Biological, China) was added on days 1, 3, and 5 to sustain protein production. On day 6, the culture supernatant was collected by centrifugation and filtered through a 0.22-μm membrane. The clarified supernatant was subjected to affinity purification via Ni Sepharose resin (Cytiva) to isolate His-tagged ACE2 protein. Purified protein fractions were desalted into the appropriate buffer for downstream analyses, including surface plasmon resonance binding assays. MD simulation of SARS-CoV-1 ACE2 complexes The wild-type SARS-CoV-1_ACE2 complex structure and its optimal variants predicted by AlphaFold3 were used as the starting models for subsequent simulations. High-confidence conformations were subjected to MD simulations using the Amber24 software package with the Amber ff14SB force field. The systems were solvated in a cuboid box with the TIP3P water model under periodic boundary conditions (PBC), where the minimum distance between the solute and the box edges was set to 12 Å. Sodium ions were added to achieve charge neutrality. Initially, all systems underwent energy minimization: 10000 steps of the steepest descent algorithm, followed by the conjugate gradient algorithm until convergence. Subsequently, the temperature was gradually raised from 0 K to 300 K using the V-rescale thermostat under the canonical ensemble (NVT). Next, the system was equilibrated at 1 bar for 1 ns using the Berendsen barostat under the isothermal-isobaric ensemble (NPT). Following these equilibration steps, three independent replicates of 100 ns unconstrained production MD simulations were performed for each system to ensure statistical reliability. The SHAKE algorithm was employed to constrain bonds involving hydrogen atoms, while long-range electrostatic interactions were calculated using the particle mesh Ewald (PME) method. A cutoff distance of 10 Å was applied to short-range van der Waals interactions. The integration time step was set to 2 fs, and trajectory coordinates were saved every 10 ps. Phylogenetic tree analysis All gene sequences (66 ACE2 sequences and 46 sarbecovirus sequences) were downloaded from the National Center for Biotechnology Information. Multiple sequence alignments were performed using MUSCLE and visualized in MEGA. A maximum-likelihood phylogenetic tree was constructed using RAxML with 1000 bootstrap iterations. The resulting tree was edited using Evolview (https://www.evolgenius.info/evolview/#/treeview). Quantification and statistical analysis Most experiments were independently performed two to three times with similar results. Unless otherwise noted in the figure legends, data are presented as mean ± standard deviation. For the heatmap analyses shown in Figures 2 and 4, one of two independent experiments is shown as a representative result. Each experiment was performed with two technical replicates per condition using independently prepared pseudovirus batches and separately prepared target cells. All statistical analyses were performed using Prism 9 (GraphPad). Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.isci.2026.117769. Contributor Information Dongchun Yang, Email: dcyang@ipe.ac.cn. Weijin Huang, Email: huangweijin@nifdc.org.cn. Jianhui Nie, Email: niejianhui@nifdc.org.cn. Supplemental information Document S1. Figures S1 and S2, Tables S1–S5, and Data S1 References Associated Data Supplementary Materials Document S1. Figures S1 and S2, Tables S1–S5, and Data S1 Data Availability Statement •All sequence data used in this study are publicly available through GenBank. All data supporting the findings of this study are available within the article and supplemental information. The original data generated in this study have been deposited at Figshare (https://doi.org/10.6084/m9.figshare.32981705) and are publicly available as of the date of publication. The corresponding accession numbers and DOI are listed in the key resources table. All sequence data used in this study are publicly available through GenBank. All data supporting the findings of this study are available within the article and supplemental information. The original data generated in this study have been deposited at Figshare (https://doi.org/10.6084/m9.figshare.32981705) and are publicly available as of the date of publication. The corresponding accession numbers and DOI are listed in the key resources table. •No custom code was generated in this study. No custom code was generated in this study. •Software and algorithms used for data analysis, visualization, molecular modeling, MD simulations, sequence alignment, phylogenetic reconstruction, and flow cytometry analysis are listed in the key resources table. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Software and algorithms used for data analysis, visualization, molecular modeling, MD simulations, sequence alignment, phylogenetic reconstruction, and flow cytometry analysis are listed in the key resources table. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.