To investigate the differences among RV receptors in different species, protein sequences were retrieved from NCBI (Agarwala et al., 2018) by BLAST search. All sequences for each of the proteins were downloaded and collated in a FASTA format.

Multiple sequence alignment for protein sequences of host receptors were performed using DNASTAR Laser Gene version 17.0.2.1 (Burland, 2000), using the MUSCLE method (Edgar, 2004).

The protein sequences of ITGB1, mGluR2, nAChR, and NCAM receptors in different host species (humans, dogs, and black fruit bats) were retrieved from GenBank (Agarwala et al., 2018). Physicochemical properties of proteins, molecular weight, theoretical isoelectric point (pI), instability index, and grand average of hydropathicity (GRAVY) were identified by ProtParam (Gasteiger et al., 2005).

Domain organization for different RV receptors in different species were analyzed by Pfam (http://pfam.xfam.org) (Finn et al., 2016), InterPro (http://www.ebi.ac.uk/interpro/) (Mitchell et al., 2015), and CDD databases (Marchler-Bauer et al., 2015). Schematic diagrams for domains in the different protein receptors have been represented by PROTTER software (Omasits et al., 2014). Signal peptide, transmembrane, and low complexity regions within receptors in different species were retrieved from SMART (Letunic et al., 2015) and SignalP databases (Armenteros et al., 2019).

The 3D structure models for RV receptors (ITGB1, mGluR2, nAChR, and NCAM proteins), RV-G protein of Egyptian strain (QEU57979.1), and RV-G protein of bat-related group (BAE95290.2) were generated using I-TASSER (Yang et al., 2010). The 3D structures generated by I-TASSER were based upon threading, fragment assembly, and iteration. The best model was selected according to the confidence score (C-score) which represented the quality of predicted models by I-TASSER. The C-score range was between (−5 and −2), where the higher the C-score, the higher the confidence of a model and vice versa. After predicting the protein model, structure and stereochemical analyses were performed and the predicted 3D structures were visualized and annotated using PyMOL software (DeLano, 2002).

The predicted structures were used for protein-protein docking studies using GRAMM-X software (Tovchigrechko and Vakser, 2006). Docking studies were performed for the Egyptian RV-G protein against each of RV receptors from human, dog, and black fruit bat. In addition, in silico interactions between RV-G protein related to bat strain were mapped with different receptors.

The docking complexes obtained from GRAMM-X were uploaded to PDBsum (Laskowski et al., 2018) and PDBePISA (Battle, 2016) servers for analysis of the protein-protein interactions. Identification of hydrogen bonds, interacting interfaces, nonbonded contacts, salt bridges, Gibb’ free energy of binding (ΔG^int, kcal/mol), pores, and tunnels in protein complexes were carried out. Mapping of the docking complexes was performed using PYMOL software (DeLano, 2002).

The physicochemical properties of ITGB1, mGluR2, nAChR, and NCAM were compared among different species (human, dog, and black fruit bat) based on their amino acid (a.a.) composition as summarized in Table 1 . Investigation of the hydrophobic nature of the RV receptors was assessed by GRAVY values. The range of GRAVY values for ITGB1 and NCAM were negative, revealing their overall hydrophilic nature. In contrast, positive GRAVY values identified in mGluR2 and nAChR proteins indicated their hydrophobic nature. To test the stability of different receptors, the instability index values were estimated. Instability indices for mGluR2, nAChR, and NCAM were all below 40, indicating the stability of these proteins in all species examined. However, the ITGB1 in human and dogs were considered relatively unstable, compared with ITGB1 from black fruit bat which had an instability index below 40. The number of amino acid residues, signal peptide, transmembrane, and low complexity regions were predicted and are summarized in Table 1 . Interestingly, the signal peptide in black fruit bat ITGB1 was missing according to SMART and SignalP databases in the sequence that is available in the current genomic annotation.

A schematic representation of the domain organization of human ITGB1, mGluR2, nAChR, and NCAM receptors are shown in Figures 1A – D , respectively. The 3D structures of the receptors highlighting each domain are also shown ( Supplementary Figures S1A – D ). Analysis of the human ITGB1 a.a. sequence ( Figure 1A ) revealed six distinct domains. The integrin plexin domain, a short disulfide-rich domain (a.a. from 25 to 76) located at the N-terminus of integrin beta chains, was also present in dog ITGB1, but interestingly was the only domain absent in black fruit bat ( Supplementary Figure S1A ). The longest domain within ITGB1 is the von Willebrand A (VWA) domain, encompassing the region from a.a. 36 to 464. The smallest domain in ITGB1 is the integrin beta epidermal growth-like factor domain 1 (a.a. 466–495). In addition, an epidermal growth factor (EGF)-like domain 2 (a.a. 599–630) and integrin beta tail domain (a.a. 640–728) were also identified. The most distal domain was the cytoplasmic domain (a.a. 752–798). The mGluR2 receptor was structurally divided into three domains ( Figure 1B ). The large extracellular region is the ligand binding domain of the group II metabotropic glutamate receptor (a.a. 6–458). The cysteine-rich domain (CRD) (a.a. 469–546) is characterized by highly conserved cysteine residues forming disulfide bridges. Linked to the CRD domain is a transmembrane domain composed of seven transmembrane helices (7 TMD) (a.a. 567–833 a.a.). All domains were similarly present in dogs and black fruit bat. The domain organization analysis of the nicotinic acetyl-choline receptor nAChR ( Figure 1C ) demonstrated a large conserved extracellular domain (a.a. 22–256). Four transmembrane regions named as neurotransmitter gated ion channel (a.a. 263–468) were followed by a cytoplasmic loop (Albuquerque et al., 2009). A similar domain arrangement was also present in dog and black fruit bat within the nAChR. Analysis of the neural cell adhesion molecule (NCAM) domain architecture revealed an extracellular portion of NCAM which is composed of five N-terminal Ig-like domains and two fibronectin type III domains which form a dimeric glycoprotein composed of disulfide-linked subunits. NCAM domain arrangement was the same in dog and black fruit bat ( Figure 1D ).

Analysis of sequence similarities of receptors among different species was crucial for mapping the differences within the interaction site of the receptors with the RV-G protein. For each of the four RV receptors, protein sequences from human, dog, and black fruit bat were aligned ( Figures 2A – D ). Analysis of the ITGB1 protein sequences from human, dog, and black fruit bat ( Figure 2A ) revealed that the VWA domain represented the most variable region among species, where dog showed 15 a.a. differences in comparison with human and 18 a.a were variable between black fruit bat and human. The area of greatest homology among the ITGB1 protein was the EGF-1 domain where only one a.a. varied between species. There were five different a.a. residues in the integrin plexin domain between human and dog. Similarly, within the EGF-2 domain, four and five a.a. residues showed differences in black fruit bat and dog, compared with human. Both dog and black fruit bat showed nine a.a. residue difference in the tail domain relative to human. The differences in these residues (1–728 a.a) may be crucial, since the ITGB1 ectodomain constitutes the interaction site with the RV-G protein, as was pointed out in a previous study (Shuai et al., 2019). Our results highlight the sequence homology within the mGluR2 protein alignment among the different species ( Figure 2B ). The ligand-binding domain showed the least sequence similarity between species where eight and 10 a.a. residues were different in black fruit bat and dog, respectively, in comparison with human. The CRD domain only showed three a.a. differences between black fruit bat, dog, and human. Only four amino acid residues were different in the transmembrane domain between human, dog, and black fruit bat. Considering that the binding site of RV-G with nAChR was previously mapped to be on the αx subunit between residues 173 and 204 (Lafon, 2005), one major finding in our study ( Figure 2C ) was the high conservation of this region among human, dog, and black fruit bat. The interaction site of NCAM with RV-G protein has not been previously determined. Our results showed the a.a. differences of NCAM domains among the different species. As shown in Figure 2D , the first immunoglobulin domain and second and third immunoglobulin I-set domains were the most conserved regions among all species. On the other hand, the first immunoglobulin I-set domain showed the most variable region among the three species where eight and 10 a.a. residues differ in black fruit bat and dog, respectively, compared with human. Black fruit bat NCAM displayed the highest variability in its fibronectin type III domain where 26 residues were variable, while only one residue differed in dog with respect to human.

The 3D structures of different receptors were predicted using the ITASSER online server. The best predicted model structures were chosen according to the maximum confidence score which was calculated according to threading templates significance ( Supplementary Figures S1A – D ). The C-score ranged from −5 to −2 (the higher the value, the higher the confidence and vice versa). For human ITGB1, a model with a C-score of 0.40 was selected; for dog ITGB1, a model with a C-score of 0.37 was chosen; and for bat, a model with C-score of 0.23 was selected. For mGluR2, 3D structures of the dog (C-score, −0.13) human (C-score, −0.13), and black fruit bat (C-score, −0.19) were chosen. For nAChR, the best model in human had a C-score of −0.37, dog with a C-score of 0.05, and black fruit bat with a C-score of 0.35. NCAM C-scores of 3D structures were as follows: human, −0.69; dog, −0.79; and black fruit bat, 0.10.

To elucidate the mechanism by which the RV-G protein interacts with different receptors, protein-protein docking was performed using GRAMMX. Analysis of the docking complexes was resolved through PDBsum and PDBePISA servers. Additionally, mapping the interacting hydrogen bonds, salt bridges, and the ΔG^int was also performed ( Table 2 ). The ΔG^int value which expresses the solvation free energy gain upon assembly formation (total solvation energies of assembled structures-solvation energies of isolated structures) was determined as well (Pantsar and Poso, 2018). The RV-modeled 3D structure of Egyptian strain G protein was utilized to undertake the docking against different receptors. The docking complex of human ITGB1 and Egyptian RV-G protein indicated five interactions mediated by hydrogen bonds between residues of ITGB1 VWA and EGF-1-like domain with the RV-G protein in addition to formation of one salt bridge between Glu³⁴⁰ of human ITGB1 and His⁴³⁸ of the RV-G protein ( Figure 3A ). A relatively more stable docking complex between dog ITGB1 VWA domain and RV-G protein of ΔG^int −20.8 kcal/mol was mapped ( Figure 3B ). The stability of the docking complex might be due to formation of four hydrogen bonds between VWA domain of dog ITGB1. Four salt bridges between Lus¹⁵⁶, Asp²⁸⁷, Glu³⁴⁰, and Glu³⁴⁷ in dog ITGB1 and Asp⁴²⁹, His¹⁰⁵, His⁴³⁸, and Arg¹⁰³ in the RV-G protein were identified. Predicted interactions of the integrin beta tail domain from black fruit bat ITGB1 showed bonding with RV-G protein through three hydrogen bonds ( Figure 3C ). Our results showed that the G protein ectodomain is responsible for binding to ITGB1 in different hosts ( Figures 3A – C ). Our modeling of the interaction between mGluR2 in human and dog with RV-G protein showed that the interactions were only mediated by the hydrogen bonds in the seven-transmembrane domain of mGluR2 ( Figures 4A, B ). Intriguingly, 10 hydrogen bonds were at the interface between the ligand-binding domain of mGluR2 from black fruit bat and the RV-G protein ( Figure 4C ), along with formation of four salt bridges between Lys²⁴, Arg¹⁰⁷, and His¹²⁹ in black fruit bat mGluR2 and Glu⁴³⁰, Asp⁴²⁷, and Asp⁴²⁰ in the RV-G protein. Interestingly, the G protein ectodomain, transmembrane, and cytoplasmic domains all appear to play a role in interactions of G protein with mGluR2 in different hosts ( Figures 4A – C ). The nAChR extracellular domain of human and black fruit bat interacted with RV-G protein through three hydrogen bonds ( Figures 5A, B ). Surprisingly, the docking complex with dog nAChR showed neither hydrogen bonds nor any salt bridges with the RV-G protein which needs further investigation. It was observed that the G protein ectodomain was responsible for interactions with nAChR in human and bat ( Figures 5A, B ). Two hydrogen bonds mediate the interaction between human NCAM and RV-G in the docking complex ( Figure 6A ). A salt bridge was noticed between residues Arg¹⁷⁷ in human NCAM and Asp⁴²⁹ within the RV-G protein. Modeling of the dog NCAM-RV-G docking complex demonstrated nine hydrogen bonds ( Figure 6B ). A total of five hydrogen bonds were mapped in the docking complex between black fruit bat NCAM and RV-G protein ( Figure 6C ). The RV-G ectodomain was the interacting part with human and dog NCAM, while both the RV-G ectodomain and cytoplasmic domains may interact with bat NCAM.

Interestingly, we have performed further docking analysis and compared the interaction of bat ITGB1 with the G protein of the bat RV group as shown in Figure 7 . Our results show the formation of four hydrogen bonds between bat ITGB1 and bat-related RV-G protein ( Figure 7A ), while only two hydrogen bonds were mapped between bat ITGB1 and the G protein of the dog-related group ( Figure 3C ). To elucidate if the interaction of human ITGB1 will be stronger with bat- or dog-related RV-G proteins, we modeled the docking complex of human ITGB1 with bat RV-G protein ( Figure 7B ). The best model predicted an interaction mediated by only one hydrogen bond between human ITGB1 and bat RV-G protein, in contrast to the five hydrogen bonds created upon interaction of human ITGB1 with dog RV-G protein ( Figure 3A ). For checking the accuracy of the docking results, we have tested if nonsusceptible host to RV (chicken) will show any interacting residues with G protein. Two docking complexes were modeled as follows: chicken nAChR–dog G protein and chicken ITGB1–dog G protein. Unexpectedly, chicken nAChR–dog G protein complexes showed seven hydrogen bonds ( Figure 7C ). While in chicken ITGB1–dog G protein, four hydrogen bonds were mapped ( Figure 7D ).