Effects of two amino acid substitutions in the capsid proteins on the interaction of two cell-adapted PanAsia-1 strains of foot-and-mouth disease virus serotype O with heparan sulfate receptor

Foot-and-mouth disease (FMD) is an acute, highly contagious vesicular disease of cloven-hoofed animals including cattle, swine, sheep, and goats[1]. Its etiological agent, foot-and-mouth disease virus (FMDV), is a member of the Aphthovirus genus of the family Picornaviridae[2]. The virus genome consists of a single-stranded, positive-sense RNA molecule of approximately 8,500 nucleotides that encodes four structural viral proteins (VP1–VP4) and nine non-structural proteins (Lab/Lb, 2A, 2B, 2C, 3A, 3B_1–3, 3C, and 3D)[3, 4]. The FMDV virion has a non-enveloped, icosahedrally symmetric capsid, which is composed of 60 copies each of the surface-exposed VP1, VP2, and VP3 as well as the internal VP4[5].

The first step during FMDV infection is the recognition of target cells by binding to cellular receptors[6]. It has been reported that FMDV could utilize four members of the αV subgroup of integrins (αVβ1, αVβ3, αVβ6, and αVβ8) as receptors in tissue culture[7‐12], via a highly conserved arginine-glycine-aspartic acid (RGD) motif located within the G-H loop of VP1[8, 13‐15]. In addition to integrins, FMDV likewise utilizes multiple membrane molecules as receptors. These alternative receptors include heparan sulfate (HS)[16], the Fc receptor (FcR)[15, 17], the synthetic scab/ICAM1 protein (a FMDV-specific, single-chain monoclonal antibody fused to the intercellular adhesion molecule 1)[18], and certain still unknown molecules on the cell surface that are neither integrins nor HS[19‐24].

HS is a glycosaminoglycan polymer of disaccharide repeats of L-iduronic acid (Idu) and D-glucosamine (GlcN), which is highly sulfated and thus negatively charged[25, 26]. The ability of FMDV to utilize HS as a receptor enables the virus to grow more efficiently in cultured cells, whereas viruses that have a high affinity for heparin (one of HS-like substances) are attenuated for cattle[27]. Studies of FMDV-HS interaction have identified the acquisition of positively charged residues during HS adaptation at widely spaced locations on the outer capsid surface of FMDV serotypes O[16, 24, 27, 28], A[24, 29], C[20‐22], SAT1[30, 31], and SAT2[30]. Here, several chimeric viruses and site-directed mutants were generated to identify the molecular determinant(s) of the HS-binding that contribute to establishing an efficient infection by two cell-adapted PanAsia-1 viruses in wild-type Chinese hamster ovary (WT-CHO) cells. Furthermore, phenotypic properties of these viruses were explored with focus on the particular effect of the addition of heparin and RGD-containing peptide in FMDV infection of baby hamster kidney 21 (BHK-21) cells and the potential consequences of the indicated FMDVs to bind the cell surface HS of WT-CHO cells. Our results are useful for improving present knowledge of the molecular basis of FMDV-HS interaction in cultured cells.

The interaction of FMDV with HS has been discussed by structural studies of two tissue culture-adapted strains of FMDV (O₁BFS and A10₆₁) in a complex with heparin[28, 29]. The HS-binding sites (Lys-2134, Arg-2135 and Tyr-2138 in VP2; Arg-3056, Gly-3059, Gly-3060, Ser-3087 and Asn-3088; and His-1195 in VP1) occupy critical positions that act as ligands for the five sugar residues Idu 1 to 5 and come into contact with each other[28, 32]. The amino acid residues at positions 3056 (Arg) and 3084 (Lys) are involved in heparin binding of A10₆₁[29], and a single amino acid change Arg-3056 → His is critical to modulate the affinity of O₁BFS and A10₆₁ to heparin[28]. Arg-2135 in VP2 and Arg-3056 in VP3 were conserved between the serotype O and A FMDV-heparin complexes. Arg-2135 of VP2 makes a hydrophobic interaction in both complexes and appears to polarize Asn-3088 of VP3, increasing its affinity for GlcN-4-6-O-SO₃[29]. Some experimental evidence suggests that the alteration in the HS-binding may be mediated by only one or two amino acid changes in the three surface-exposed capsid proteins of FMDV[20‐22, 24, 27, 30]. Sa-Carvalho et al. have reported that the positively charged residues at both 2134 (Lys) in VP2 and 3056 (Arg) in VP3 were required for the binding of FMDV O₁Campos to heparin[27]. Amino acid substitutions were found in the HS-adapted SIR (soluble integrin resistant) isolates of A₂₄ Cruzeiro (A₂₄-SIR#45) and O₁Campos (O₁C-SIR#46): Leu-1150 → Arg (RGD + 4) in VP1 of A₂₄-SIR#45, His-3056 → Arg and Asp-3060 → Ala in VP3 and Glu-1113 → Val in VP1 of O₁C-SIR#46[24]. The selected variants of C₃Arg/85 (C₃-Rb) had acquired the ability to bind heparin and to infect WT-CHO cells, and Asp-3009 → Ala in VP3 and either Gly-1110 → Arg or His-1108 → Arg in VP1 were located at the five fold axis of the viral capsid[21]. The positively charged Lys-3173 in VP3 (the same position at 3174 in VP3 of O₁BFS) is essential and Ser-1144 (RGD + 1) in VP1 may also play a role for the binding of MARLS (selected from C-S8c1) to heparin[20]. Loss of heparin binding in C-S8c1p100c10 was associated with a single amino acid replacement which was either Lys-3173 → Glu in VP3 or Arg-1197 → His (the same position at 1195 in VP1 of O₁BFS) at the C-terminal region of VP1[22]. In serotype SAT1 FMDVs, a recombinant virus vSAT1tc (derived from SAT1/SAR/9/81tc) displayed a high affinity for WT-CHO cells and the VP1 Lys-1110 → Asn replacement could abrogate the ability to infect WT-CHO cells[31]. The introduction of amino acid mutations at position Gly-1111 → Arg and Gly-1112 → Arg in VP1 of cell culture-adapted SAT1/NAM/307/98 (vNAM(KRR)/SAT) is thought to correlate with the acquisition of the HS-utilizing ability for viral infection of WT-CHO cells[30]. Thus, these implied that FMDV has a high heparin affinity, which is accompanied by the fixation of not only positively charged residues but also neutral amino acid residues surrounding the G-H loop of VP1 and the fivefold axis of the virus. However, no amino acid differences at these potential HS-binding sites were found on the capsid of O/Tibet/CHA/6/99tc and O/Fujian/CHA/9/99tc, as compared to their non-HS-adapted field isolates (O/Tibet/CHA/6/99wt and O/Fujian/CHA/9/99wt), respectively. All the selected PanAsia-1 viruses carried Lys-2134, Arg-2135 and Tyr-2138 in VP2, Asp-3009, His-3056, Gly-3059, Asp-3060, Lys-3084, Ser-3087, Asn-3088 and Ala-3174 in VP3, and His-1108, Ala-1110, Pro-1111, Leu-1112, Thr-1113, Leu-1148 (RGD + 1), Leu-1151 (RGD + 4) and His-1195 in VP1. The neutral amino acid change Ala-1144 (substituted for Val) was located at position RGD-1 in VP1 of O/Tibet/CHA/6/99tc (Table 1), and the chimeric virus encoding VP1 of O/Tibet/CHA/6/99tc (rHN/TAR6-VP1) did not exhibit more efficient binding to HS for viral infection in WT-CHO cells (Figure 1).

Nonetheless, the 3D structural models of rHN/TAR6-VP0 and rHN/FJ9-VP1 indicated that Gln-2080 in the B-C loop of VP2 of O/Tibet/CHA/6/99tc and Lys-1083 within the D-E loop of VP1 of O/Fujian/CHA/9/99tc surround the G-H loop of VP1 and the pore at the fivefold axis of symmetry, respectively (Figure 5A, 5B)[35‐38]. The difference in phenotypic properties of the two chimeric viruses (rHN/TAR6-VP0 and rHN/FJ9-VP1) and their site-directed mutants (rHN/TAR6-VP0^Q2080L and rHN/FJ9-VP1^K1083E) on WT-CHO cells were consistent with our expectations (Figure 1). Yet, the introduction of a Leu-2080 → Gln mutation in VP2 of O/Fujian/CHA/9/99tc (pOFS/FJ9-VP0^L2080Q) was particularly deleterious for the generation of progeny virus in BHK-21 cells, and a Glu-1083 → Lys mutation in VP1 of O/Tibet/CHA/6/99tc did not result in the acquisition of the HS-utilizing ability of rHN/TAR6-VP1^E1083K to infect WT-CHO cells (Figure 1). Moreover, a second-site mutation (Glu-2136 → Gly) in VP2 of O/Fujian/CHA/9/99tc was unable to compensate for the lethal effect of the primary mutation Leu-2080 → Gln in the VP2 coding region of pOFS/FJ9-VP0^L2080Q and the site-directed mutant encoding Glu-1083 → Lys and Ser-1139 → Arg in VP1 of O/Tibet/CHA/6/99tc also could not utilize HS as a receptor to establish an efficient infection in WT-CHO cells (results not shown). There are a limited number of amino acid differences fixed in the VP2 and VP1 coding regions of the mutated plasmids as compared to rHN/TAR6-VP0 and rHN/FJ9-VP1, respectively (Table 1). In particular, the distinct amino acid residues at 2079 (Tyr) of VP2 lies adjacent to the RGD motif (Figure 5A), and (ii) Thr-1041 and Gln-1045 of VP1 are clustered around the fivefold axis (Figure 5B). The outcome of the substitution at residue 2079 of VP2 could potentially dislocate both antigenic sites 1 (1141–1160 residues of the VP1 G-H loop,[39]) and 2 (2070–2078 residues of the VP2 B-C loop,[40]), with a structural and/or functional change[33]. One of the amino acid mutations, Lys-1041 → Glu in VP1 was fixed in two heparin-binding derivatives of C-S8c1 (C-S8c1p100c10 and the MARLS mutant)[20, 22]. Accordingly, His-2079 → Tyr in VP2 of O/Fujian/CHA/9/99tc and Lys-1141 → Thr and/or Lys-1145 → Gln in VP1 of O/Tibet/CHA/6/99tc may potentially be able to restore infectious viral phenotypes in BHK-21 cells, and be helpful in gaining the HS-utilizing ability of the corresponding site-directed mutants for viral infection in WT-CHO cells. In other words, the specific amino acid residues at positions 1041 (Lys) and 1045 (Lys) in VP1 of O/Tibet/CHA/6/99tc and 2079 (His) in VP2 of O/Fujian/CHA/9/99tc (Table 1), could be detrimental to the genetic response of the indicated viruses to FMDV-HS interaction.

It should be noted especially that the acquisition of the ability to infect WT-CHO cells as a result of substitutions at the functionally critical residues in the capsid coding regions of FMDV does not necessarily mean HS is required for efficient infection[20, 22‐24, 34]. In the present study, the results from three different assays on two CHO cell lines are evidence for the HS-utilizing ability of O/Tibet/CHA/6/99tc, O/Fujian/CHA/9/99tc, rHN/TAR6-VP0 and rHN/FJ9-VP1, and the incapability of HS-dependent infection by the other eight viruses in WT-CHO cells (Figure 1, Figure 3A, Figure 4). Even so, a high affinity for heparin and the HS-binding of rHN, rHN/FJ9-VP0 and rHN/TAR6-VP0^Q2080L were supported by plaque reduction neutralization and inhibition assays on BHK-21 cells and virus adsorption assay on WT-CHO cells, respectively (Figure 3B,3C; Figure 4). Loss of the heparin-sensitivity of these three viruses was accompanied by a single amino acid mutation Lys-1083 → Glu of VP1 (results not shown). It is also worth noting that, while heparin sensitivity of O/Tibet/CHA/6/99tc was detected in WT-CHO cells and the HS-binding and RGD integrin-binding sites of FMDV function independently of each other[32] (Figure 1, Figure 3A, Figure 4), the virus did not appear to initiate its HS-utilizing ability to facilitate viral infection in BHK-21 cells pre-incubated with the RGD-containing peptide VR-17 (Figure 3C). The distinct heparin affinity of O/Tibet/CHA/6/99tc, O/Fujian/CHA/9/99tc, rHN/TAR6-VP0 and rHN/FJ9-VP1 suggested that the specific amino acid residues on the outer capsid proteins (VP2 of O/Fujian/CHA/9/99tc and VP1 of rHN) may be related to FMDV-HS interaction on BHK-21 cells (Figure 3B, Table 1). As for the plaque phenotypes of all the selected PanAsia-1 strains and rescued viruses (Figure 1), the effect of a His or Tyr at residue 2079 of VP2 and a Lys or Glu at residue 1083 of VP1 on the local charged regions might contribute to the differences in the size of plaques produced by the corresponding viruses on BHK-21 cells (Table 1). Further work is needed to consider these two issues in more detail. One goal is to evaluate the cooperative effect of several amino acid residues in the capsid proteins of the HS-utilizing FMDVs and heparin-sensitive FMDVs on the adsorption efficiency and penetration capability of these viruses in WT-CHO cells. Another is to identify the type of heparin-sensitive cellular receptor(s) utilized by the heparin-sensitive FMDVs to facilitate viral infection in BHK-21 cells. Further studies will help us to interpret the positive selection of amino acid substitutions in the viral capsid for HS adaptation during FMDV microevolution in cultured cells.

In conclusion, our results demonstrate that VP2 of O/Tibet/CHA/6/99tc and VP1 of O/Fujian/CHA/9/99tc function as the critical regions in the acquisition of the HS-utilizing ability to infect WT-CHO cells. Gln-2080 → Leu in VP2 of O/Tibet/CHA/6/99tc and Lys-1083 → Glu in VP1 of O/Fujian/CHA/9/99tc would lead to the loss of the HS-utilizing ability of two chimeric viruses in WT-CHO cells, respectively. The distinct amino acid substitutions at the other positions in VP0 of O/Tibet/CHA/6/99tc and VP1 of O/Fujian/CHA/9/99tc may also play important roles in retaining the infectivity and the HS-utilizing capability of the corresponding site-directed mutant(s) in cultured cells. Although the HS-utilizing ability of O/Tibet/CHA/6/99tc is required to establish an efficient infection in WT-CHO cells, the virus was unable to utilize the cell surface HS for viral infection in BHK-21 cells. The heparin-sensitive FMDVs could bind to HS, but might be insufficient to initiate HS-dependent infection in WT-CHO cells.