Comprehensive Investigation on the Interplay between Feline APOBEC3Z3 Proteins and Feline Immunodeficiency Virus Vif Proteins

ABSTRACT As the hosts of lentiviruses, almost 40 species of felids (family Felidae) are distributed around the world, and more than 20 feline species test positive for feline immunodeficiency virus (FIV), a lineage of lentiviruses. These observations suggest that FIVs globally infected a variety of feline species through multiple cross-species transmission events during a million-year history. Cellular restriction factors potentially inhibit lentiviral replication and limit cross-species lentiviral transmission, and cellular APOBEC3 deaminases are known as a potent restriction factor. In contrast, lentiviruses have evolutionary-acquired viral infectivity factor (Vif) to neutralize the APOBEC3-mediated antiviral effect. Because the APOBEC3-Vif interaction is strictly specific for viruses and their hosts, a comprehensive investigation focusing on Vif-APOBEC3 interplay can provide clues that will elucidate the roles of this virus-host interplay on cross-species transmission of lentiviruses. Here, we performed a comprehensive investigation with 144 patterns of a round robin test using 18 feline APOBEC3Z3 genes, an antiviral APOBEC3 gene in felid, and 8 FIV Vifs and derived a matrix showing the interplay between feline APOBEC3Z3 and FIV Vif. We particularly focused on the interplay between the APOBEC3Z3 of three felids (domestic cat, ocelot, and Asian golden cat) and an FIV Vif (strain Petaluma), and revealed that residues 65 and 66 of the APOBEC3Z3 protein of multiple felids are responsible for the counteraction triggered by FIV Petaluma Vif. Altogether, our findings can be a clue to elucidate not only the scenarios of the cross-species transmissions of FIVs in felids but also the evolutionary interaction between mammals and lentiviruses. IMPORTANCE Most of the emergences of new virus infections originate from the cross-species transmission of viruses. The fact that some virus infections are strictly specific for the host species indicates that certain “species barriers” in the hosts restrict cross-species jump of viruses, while viruses have evolutionary acquired their own “arms” to overcome/antagonize/neutralize these hurdles. Therefore, understanding of the molecular mechanism leading to successful cross-species viral transmission is crucial for considering the menus of the emergence of novel pathogenic viruses. In the field of retrovirology, APOBEC3-Vif interaction is a well-studied example of the battles between hosts and viruses. Here, we determined the sequences of 11 novel feline APOBEC3Z3 genes and demonstrated that all 18 different feline APOBEC3Z3 proteins tested exhibit anti-feline immunodeficiency virus (FIV) activity. Our comprehensive investigation focusing on the interplay between feline APOBEC3 and FIV Vif can be a clue to elucidate the scenarios of the cross-species transmissions of FIVs in felids.

FIVfca Vif can degrade not only the A3 proteins of its host, domestic cat, but also those of the other felids such as puma, lynx, lion, and tiger (23). On the other hand, we have reported that the Vif protein of FIVfca subtype B cannot degrade the antiviral A3 proteins of its host (24). Because FIVfca subtype B seems to be less pathogenic and divergent than the other FIVfca subtypes (25,26), it is assumed that the loss of Vif's ability to counteract antiviral A3 protein may be a way for FIVfca to adapt to the host through the attenuation of its virulence (24).
On the viral side, Troyer et al. have conducted a comprehensive surveillance for FIV infections on 35 feline species and revealed that at least 23 feline species tested were positive for either anti-FIV antibodies or FIV nucleotides (27). This comprehensive study indicates that FIVs globally infect a wide variety of host species. Specifically, the vif sequences have been determined in 5 FIV lineages in 5 felids: FIVfca in domestic cats (Felis catus), Pallas's cats (Otocolobus manul; FIVoma), pumas (Puma concolor; FIVpco), bobcats (Lynx rufus; FIVlru), and lions (Panthera leo; FIVple) (28)(29)(30)(31)(32). In investigating the origin of HIV, previous studies showed SIV infection in a variety of Old World monkey species as well as nonhuman great apes (reviewed in reference 33). However, the habitat of the primates infected with SIVs is limited to Africa. In sharp contrast, FIVs are distributed globally and have been isolated from felids of multiple genera. Two felids, puma and bobcat, naturally live in North America, and interestingly, two different classes of FIVs, FIVpco and FIVlru, circulate between these two felids belonging to different genera (28,34,35). More intriguingly, we have previously demonstrated that all FIVlru Vifs tested can degrade both puma and bobcat A3Z3s, while a subclass of FIVpco Vifs can degrade puma A3Z3 but not bobcat A3Z3, suggesting that bobcat A3Z3 can potentially be a "species barrier" that hampers the cross-species transmission of a subclass of FIVpco from pumas to bobcats (36).
Several classes of FIVs and their host felids are distributed globally. Therefore, a comprehensive investigation focusing on the interplay between FIV Vif proteins and feline A3Z3 proteins may provide clues that will elucidate the roles of the Vif-A3 interplay on cross-species transmission of lentiviruses and the coevolution of lentiviruses and mammals. In this study, we newly determine the feline A3Z3 sequences of 11 species and investigate the anti-FIV activities of the A3Z3 proteins of 18 felid species. We also determine a novel FIV vif sequence from leopard (Panthera pardus; FIVppa) and comprehensively test the ability of the 8 different FIV Vif proteins from 6 FIV classes against 18 different feline A3Z3 proteins.

RESULTS
Complicated evolution of the A3Z3 genes in felids. The 18 feline species including focused in this study are listed in Table 1. In this study, we newly determined the A3Z3 sequences of 11 felids: Pallas's cat (Otocolobus manul), leopard cat (Prionailurus bengalensis), fishing cat (Prionailurus viverrinus), ocelot (Leopardus pardalis), caracal (Caracal caracal), serval (Leptailurus serval), Asian golden cat (Catopuma temminckii), jaguar (Panthera onca), leopard (Panthera pardus), snow leopard (Panthera uncia), and clouded leopard (Neofelis nebulosa) (Table 1). We generated the multiple-sequence alignments (MSAs) of these A3Z3 genes, constructed the maximum likelihood (ML) phylogenetic tree, and compared the phylogenetic topology of feline A3Z3 genes with that of felid species (15). As shown in Fig.  1, the A3Z3 genes of the Panthera lineage (including jaguar, lion, leopard, snow leopard, tiger, and clouded leopard) corresponded well to the phylogenetic relationship of species classification. On the other hand, the A3Z3 genes of non-Panthera lineages were incongruent with the species phylogeny (Fig. 1). Particularly, the phylogenetic positions of the A3Z3 genes of lynx and bobcat are clearly different from those in the species phylogeny (Fig. 1). We then assessed the possibility of geographical convergence of feline A3Z3 genes according to a previous study (15). However, there were no associations between the diversification of feline A3Z3 genes with the biogeographical region of feline habitat (Fig. 1).
Next, we performed evolutionary analysis on feline A3Z3 genes. The MSAs of feline A3Z3 proteins showed that an amino acid, asparagine, is inserted at position 24 of all A3Z3 proteins belonging to the Panthera lineage. In the non-Panthera lineages, only cheetah possessed an additional amino acid, tyrosine, at position 24. The ratio of nonsynonymous to synonymous evolutionary changes (dN/dS ratio) analyses by two independent methods, fixed-effects likelihood (FEL) (37) and mixed-effects model of evolution (MEME) (38), revealed that the amino acid at position 65 of feline A3Z3 is under strong diversifying selection ( Fig. 2A and B; note that the amino acid position of feline A3Z3 protein mentioned in this study is based on that of domestic cat A3Z3 protein). Because the dN/dS analysis is not applicable for singleton indel sites, a one-amino-acid insertion at position 24 in the Panthera lineage and cheetah is removed from these analyses). These results correspond well to our previous findings showing that the amino acid at position 65 of domestic cat A3Z3 is under strong diversifying selection (22). Additionally, the analysis by FEL showed that the residues at positions 32, 49, 65, 74, 127, and 186 are under diversifying selection ( Fig. 2A). By plotting these residues on the protein homology model of domestic cat A3Z3 protein (22), we found that residue 65 and the four additional residues, positioned at 32, 49, 74, and 127, are located on the protein surface ( Fig. 2C and D).
Different anti-FIV capacities of feline A3Z3 proteins. We next investigated the anti-FIV ability of these 18 different feline A3Z3 proteins. Because the amino acid sequence of the snow leopard A3Z3 protein is identical to that of the tiger A3Z3 protein, we excluded the snow leopard A3Z3 protein from the cell culture experiments. Instead, we included two haplotypes of domestic cat A3Z3, hap I and hap V (22). The hemagglutinin (HA)-tagged expression plasmids for 18 feline A3Z3 proteins were constructed and were respectively cotransfected with vif-deficient FIV-based reporter plasmids at three different doses as previously described (22,24,36). As shown in Fig. 3A, all feline A3Z3 proteins were expressed and incorporated into the released virions in a dose-dependent manner. Feline A3Z3 expression did not affect the expression level of FIV Gag precursor in virus-producing cells or the amount of released viral particles (Fig. 3A). Furthermore, the virion-incorporated feline A3Z3 proteins suppressed viral infectivity in a dose-dependent manner (Fig. 3B). Because the antiviral activities of feline A3Z3 proteins were different from each species (Fig.  3B), we assessed the association of the antiviral activity of feline A3Z3 proteins with the lineage and the habitat of felids. As shown in Fig. 3C, the five A3Z3 proteins of the Panthera lineage (lion, jaguar, leopard, tiger, and clouded leopard) and the two A3Z3 proteins of the Puma lineage (puma and cheetah) tended to exhibit relatively higher anti-FIV activity. On the other hand, the antiviral activities of the two A3Z3 proteins of the Lynx lineage (lynx and bobcat) were relatively low (Fig. 3C). Correlation analysis with the felid habitat showed that the A3Z3 proteins of the felids living in the Oriental (leopard cat, fishing cat, Asian golden cat, tiger, and clouded leopard), Ethiopian (cheetah, caracal, serval, and lion) and Neotropical (puma and ocelot) regions tended to exhibit relatively higher anti-FIV activity (Fig. 3D).
In addition to the differences in the antiviral activities of each feline A3Z3 protein (Fig. 3B), the expression levels of feline A3Z3 proteins in the cells and the efficacy for virion incorporation of these proteins are different among species (Fig. 3E). As shown in Fig. 3F, there was a significant correlation between the levels of feline A3Z3 proteins in the cells and the virions (P = 0.0037, r = 0.647 by Spearman's rank correlation coefficient). Particularly, the A3Z3 proteins of some felids of the Panthera lineage such as leopard and tiger were expressed relatively highly and incorporated efficiently into the released virions (Fig. 3F). In contrast, Pallas's cat A3Z3 protein was incorporated into the released virions relatively efficiently irrespective of its relatively low level of  (15), and the distribution of jaguar is from the IUCN RED LIST (https://www.iucnredlist.org). Based on the previous report (15) and the IUCN RED LIST, the habitat of leopards is broad and spans African and Eurasian continents (Afro-Eurasia). (Bottom right) The phylogenetic relationship of felid species. The data are from a previous study (15). MYA, million years ago. Scale bar indicates 0.005 nucleotide substitutions per site. The colors of felid names and symbols are identical to those of their biogeographic distribution (top right). expression in cells (Fig. 3F). When we assessed the correlation between the level of virion-incorporated A3Z3 protein and the infectivity of the released virions, there was a significant and negative correlation between the level of virion-incorporated A3Z3 protein and viral infectivity (Fig. 3G). As expected, the A3Z3 proteins of leopard and tiger, which were incorporated efficiently into the released virions, exhibited higher antiviral activity (Fig. 3G). In contrast, Pallas's cat A3Z3 showed relatively low antiviral activity per the amount of A3Z3 protein in the virion (Fig. 3G).
Feline A3Z3 antagonism by a variety of FIV Vif proteins. To reveal the functional relationship between feline A3Z3 and the Vif proteins of a variety of FIV classes, we constructed a phylogenetic tree of the Vif proteins of the 70 different strains of FIV listed in Table 2. In addition to the vif genes from the 5 FIV lineages in 5 felids, we newly determined an FIV vif sequence from a leopard (FIVppa, strain Rasheed). This is the first FIV vif sequence obtained from leopard, and the phylogenetic tree showed that FIVppa Vif is similar to FIVoma Vif (Fig. 4). This is consistent with a previous study (39) showing that the sequence of the reverse transcriptase (RT)-pol region of FIVoma is close to that of FIVppa.
From these 70 FIV vif sequences, we picked up eight representative FIV Vif proteins from six FIV lineages: FIVfca strains Petaluma (subtype A) and TM2 (subtype B) (22,24), FIVoma strain Oma3, FIVpco strains PLV1695 and Pco5 (36,40), FIVlru strain Lru1 (36), FIVple strain 1027, and FIVppa strain Rasheed. We prepared His-tagged Vif expression plasmids and used them for cell culture experiments ( Fig. 5 and 6). Western blotting showed that all eight FIV Vif proteins expressed efficiently (Fig. 5). Also, the expression of FIV Vif tested did not affect the expression level of Gag precursor in transfected cells or the amount of released viral particles (Fig. 5). However, two FIV Vif proteins, FIVfca TM2 and FIVpco PLV1695, were unable to degrade any feline A3Z3 proteins (Fig. 5). These results are reminiscent of the previous findings that FIVfca TM2 (24) and FIVpco PLV1695 (40) Vifs cannot degrade the APOBECs of their respective host felids (i.e., domestic cat and puma). FIV reporter assays showed that the A3Z3 proteins of some felids such as serval, lion, and leopard were relatively resistant to the  (37) (A) and MEME (38) (B) are shown. FEL and MEME are specialized for detecting the selections that occurred in the entire phylogenetic tree and the specific clades, respectively. (C and D) Structure modeling of domestic cat A3Z3 (hap I). The data of the structure homology model are from our previous study (22). Cartoon (C) and surface (D) models of the structure of domestic cat A3Z3 hap I are shown. In panels C and D, the residue under diversifying selection estimated by both FEL and MEME (residue 65) and those by FEL (residues 32, 49, 74, and 127) are represented in red and orange, respectively. Residue 66 is labeled in blue. Note that the residue 186 is not plotted, because this residue is located outside this homology model. counteraction by FIV Vif proteins (Fig. 6). However, the four FIV Vif proteins tested (i.e., except for FIVfca TM2 and FIVpco PLV1695) were able to counteract the antiviral activity of most feline A3Z3 proteins (Fig. 6).
Resistance of the A3Z3 proteins of ocelot and Asian golden cat to FIVfca Petaluma Vif. For the sake of simplicity, we summarized the data as a heat map (Fig. 7). This heat map showed that FIVlru Vif (strain Lru1) is able to counteract the antiviral activity of all feline  A3Z3 proteins tested (Fig. 7). We also found that the Vif proteins of FIVpco Pco5, FIVple 1027, and FIVppa Rasheed can counteract most of the feline A3Z3 proteins tested (Fig. 7). Consistent with our previous study (22), domestic cat A3Z3 hap V protein was resistant to counteraction by FIVfca Petaluma Vif, while domestic cat A3Z3 hap I protein was sensitive ( Fig. 5 and 7). Interestingly, we found that the A3Z3 proteins of ocelot and Asian golden cat were also resistant to FIVfca Petaluma Vif-mediated counteraction (Fig. 7). It has been demonstrated that only one amino acid at position 65 is different between domestic cat A3Z3 hap I (65A) and hap V (65I) proteins (22). This previous finding indicated that residue 65 of domestic cat A3Z3 protein determines the sensitivity to FIVfca Petaluma Vif. When we assessed the MSA of feline A3Z3 proteins, we found that only ocelot A3Z3 protein possesses arginine (R) at position 65 while only Asian golden cat A3Z3 protein possesses R at position 66 (Fig. 8A). To address the possibility that the residues at positions 65 to 66 of the A3Z3 proteins of ocelot and Asian golden cat determine their resistance to FIVfca Petaluma Vif-mediated degradation, we prepared expression plasmids for the three A3Z3 mutants: ocelot A3Z3 R65A and R65S and Asian golden cat A3Z3 R66L. As shown in Fig. 8B, these three A3Z3 protein mutants were degraded by FIVfca Petaluma Vif, although parental ocelot and Asian golden cat A3Z3 proteins were not. Furthermore, the FIV reporter assay revealed that the anti-FIV ability of these three A3Z3 mutants was canceled by FIVfca Petaluma Vif (Fig. 8C). To further test this issue, we prepared expression plasmids for the three domestic cat A3Z3 mutants: hap I A65R, hap I L66R, and hap II L66R. As expected, these three mutated A3Z3 proteins of domestic cats were resistant to FIVfca Petaluma Vifmediated degradation (Fig. 8D). These A3Z3 proteins were incorporated into the released virions even in the presence of FIVfca Petaluma Vif (Fig. 8D) and significantly suppressed viral infectivity (Fig. 8E). Taken together, these results suggest that the residues at positions 65 and 66 of feline A3Z3 proteins determine sensitivity to FIVfca Petaluma Vif.

DISCUSSION
In this study, we determined the sequences of 11 novel feline A3Z3s and demonstrated that all 18 different feline A3Z3 proteins tested exhibit anti-FIV activity ( Fig. 1  and 3). Because the A3Z3 proteins of the felids that did not show FIV infections in the past (e.g., leopard cat, fishing cat, serval, and caracal) (27) also significantly suppressed the infectivity of vif-deleted FIV (Fig. 3), our data suggest that the anti-FIV activity of A3Z3 proteins is broadly maintained in all felids. Additionally, we used the eight FIV Vif expression plasmids from six different classes (FIVfca, FIVoma, FIVpco, FIVlru, FIVple, and FIVppa) and, particularly, newly determined one sequence (FIVppa) (Fig. 4). Using these materials, we performed 144 patterns of the round robin test using 18 feline A3Z3 proteins and 8 FIV Vif proteins (Fig. 5 and 6) and derived a matrix showing the interplay between feline A3Z3 and FIV Vif (Fig. 7). To our knowledge, this is the first study comprehensively investigating the functional interplay between feline A3Z3 proteins and FIV Vif proteins. Based on these analyses, particularly, on the interplay between the A3Z3 proteins of domestic cat, ocelot, and Asian golden cat and FIVfca Petaluma Vif, we revealed that residues 65 and 66 of the A3Z3 proteins of multiple felids are responsible for the counteraction triggered by FIVfca Petaluma Vif (Fig. 8).
Similar to the observations on the A3 genes in primates (11) and bovids (13), the phylogenetic relationship of most of the feline A3Z3 genes was incongruent with the species phylogeny of felids (Fig. 1). However, the six A3Z3 genes of the Panthera lineage (jaguar, lion, leopard, snow leopard, tiger, and clouded leopard) form an independent cluster, and the phylogeny of the A3Z3 genes of the Panthera lineage was identical to their species phylogeny (Fig. 1). It was also intriguing that the amino acid sequences  of tiger A3Z3 and snow leopard A3Z3 proteins were identical. Although intraspecies polymorphisms in A3 genes have been reported in human A3H (19,(41)(42)(43)(44), African green monkey A3G (11,12), and domestic cat A3Z3 (16,22,36), this is the first report showing that the amino acid sequences of the A3Z3 proteins from different species are  Feline APOBEC3 versus FIV Vif Journal of Virology identical. Moreover, the anti-FIV activity of the A3Z3 proteins of the Panthera lineage tended to be commonly higher than those of non-Panthera lineages (Fig. 2C). These findings suggest that the antiviral activity of the A3Z3 proteins of the Panthera lineage has been maintained and sophisticated in a long evolutionary history of approximately 10.8 million years. On the other hand, the phylogenetic relationship of the A3Z3 genes of the Lynx lineage (lynx and bobcat) was clearly inconsistent with the species phylogeny (Fig. 1), and the antiviral activity of these A3Z3 proteins was commonly lower than those of the other feline A3Z3 proteins (Fig. 3C). Regarding this issue, a previous report showed that the common ancestor of the Lynx lineage diversified from the other felid species approximately 7.2 MYA, and this ancestor lived in North America until the diversification of lynx and bobcat as independent species approximately 3.2 MYA (15). A possible explanation for the unique properties of the A3Z3 proteins of the Lynx lineage is that the antiviral capacity of the A3Z3 proteins of the Lynx lineage has degenerated during this period, because the common ancestor of the Lynx lineage was free from pathogenic FIV infection. As another feature of the felids belonging to the Lynx lineage, bobcats and lynxes are the only felids with a short "bobbed" tail (45).
Although the reason why their tails were shortened during the evolution and its association with the unique feature of the A3Z3 genes in terms of the phylogeny (Fig. 1) and the antiviral ability ( Fig. 3C) remain unclear, the feature of the A3Z3 genes may be one of the consequences of the unique evolution of the Lynx lineage.
Here, we determined the novel sequence of FIVppa Vif and revealed that FIVppa Vif is phylogenetically similar to FIVoma Vif (Fig. 4). Our finding is consistent with a previous report showing that the RT-pol region of FIVppa is similar to that of FIVoma (39). These observations suggest that FIVppa and FIVoma share an origin. To assume the direction of the cross-species transmission, there might be two possibilities: first, Pallas's cats, the host of FIVoma, only live in Asia, while the habitat of leopards, the host of FIVppa, extends across both Africa and Eurasia. Therefore, there would be relatively more frequent opportunities for leopards to be infected with a variety of pathogens, including FIVs, and one possibility is that FIVoma has emerged by the cross-species transmission of FIVppa from leopards to Pallas's cats. Another possibility may be explained by the prey-predator relationship. This relationship partly explains the direction of cross-species transmission of SIVs from Old World monkeys to chimpanzees in the wild: because small Old World monkeys are the prey of chimpanzees in the wild, chimpanzees are frequently exposed to various SIVs that infect their prey species such as Old World monkeys (46,47). Similarly, because Pallas's cats are relatively small and can be the prey of leopards in the wild, leopards can be frequently exposed to FIVoma that infect Pallas's cat, and this prey-predator relationship may have led to the emergence of FIVppa in leopards. In either case, the Vif proteins of both FIVoma and FIVppa are able to counteract the antiviral A3Z3 proteins of Pallas's cats and leopards ( Fig. 5  and 7). Therefore, our findings suggest that the A3Z3 proteins of Pallas's cat and leopard cannot be the "species barrier" that hampers the cross-species transmission of FIVs.
Molecular evolutionary analyses revealed that at least six amino acid residues are under diversifying selection (Fig. 2). Particularly, residue 65 of domestic cat A3Z3 (identical to residue 66 of the MSA in this study) is the site that was determined as a strongly selected site in polymorphic domestic cat A3Z3 proteins and is responsible for resistance to degradation by FIVfca Petaluma Vif (22). In addition to residue 65, the other sites that are under diversifying selection by FEL ( Fig. 2A) were exposed to the protein surface (Fig. 2D). Therefore, it would be possible to assume that an FIV Vif-like factor(s) has been the selective pressure of feline A3Z3 during the evolution of felids.
More intriguingly, we found that the A3Z3 proteins of ocelot and Asian golden cat are resistant to counteraction by FIVfca Petaluma Vif (Fig. 7) and demonstrated that the resistance of these A3Z3 proteins is determined by the amino acid residues at positions 65 and 66 (Fig. 8). Together with our previous findings (22), these results suggest that these two naturally occurring feline A3Z3 proteins confer resistance to infection with FIVfca. Because the lineage and biogeographical distribution of these three felids, domestic cat, ocelot, and Asian golden cat, are different (Fig. 1), the ability of A3Z3 to be resistant to FIVfca Petaluma Vif has been acquired independently and is a convergent evolution. Furthermore, Troyer et al. isolated an FIV from ocelot (27). Although its vif sequence was not determined, the phylogeny of the FIV pol-RT sequence showed that the FIV from ocelot is phylogenetically similar to FIVfca (27). These observations suggest that the FIVs from domestic cats and ocelots share their origin. Although FIV sequences have not been detected in Asian golden cats so far (27), future investigations on the interplay between the A3Z3 protein of ocelots and Asian golden cats and the Vif proteins of the FIV from these two felids will further reveal the interplay between felids and FIV with higher resolution.
Here, we performed a comprehensive investigation using expression plasmids for a variety of feline A3Z3 proteins and FIV Vif proteins and revealed new aspects on the roles of feline A3Z3 and FIV Vif on the cross-species transmission of lentiviruses. However, some issues could not to be elucidated due to technical limitations. For example, we found that the A3Z3 proteins of the Panthera lineage commonly exhibit relatively higher anti-FIV activity (Fig. 3C). Additionally, Pallas's cat A3Z3 protein was efficiently incorporated into released virions but exhibited relatively lower antiviral activity ( Fig. 3F and G). Moreover, some feline A3Z3 proteins exhibited resistance to counteraction by certain FIV Vif proteins (e.g., FIVple Vif versus domestic cat hap I A3Z3 and serval A3Z3, and FIVppa Vif versus domestic cat A3Z3) (Fig. 8). Although we revealed the residues of feline A3Z3 proteins that conferred resistance to FIVfca Petaluma Vif (Fig. 8), the MSA of feline A3Z3 proteins did not reveal which residues were responsible. It is possible that multiple amino acid residues of feline A3Z3 proteins contribute to its antiviral activity and resistance to FIV Vif. Furthermore, the amino acid residues on FIV Vif proteins that determine ability to counteract feline A3Z3 proteins were not addressed because of their higher diversity. Although we previously revealed that the residues at positions 167, 239, 242, and 243 of FIVfca subtype B Vif are responsible for losing the counteracting activity against domestic cat A3 proteins (24), the residue(s) that determines the loss of FIVpco PLV1695 Vif activity against feline A3 proteins remains veiled. To fully elucidate the complicated interplay between feline A3Z3 proteins and FIV Vif proteins, more detailed investigations will be needed. In addition to A3Z3, other intrinsic restriction factors such as tetherin (48,49) and SAMHD1 (50,51) in felids can be a barrier(s) against cross-species transmission of FIVs, while FIVs may possess viral counterparts to overcome these hurdles. Further investigations will be needed to unveil the interplay between FIV and felids.
In summary, here we performed a comprehensive investigation of the interplay between antiviral A3 proteins and lentiviruses in felids. As previously reported, the evolution of mammalian A3 genes is complicated and is driven by certain selective pressures, including retroviruses (52)(53)(54). Our findings may provide an insight into both the scenarios of cross-species transmission of FIVs in felids and the evolutionary interactions between mammals and lentiviruses.
FIVppa was isolated from a sample of leopard (Panthera pardus) blood submitted to the University of Glasgow for FIV diagnosis from the Breeding Centre for Endangered Arabian Wildlife in Sharjah, United Arab Emirates. Ethical approval for the use of residual feline blood submitted for routine diagnostic testing was granted by the ethics committee of the University of Glasgow School of Veterinary Medicine, UK.