Single-Point Mutations within the Coxsackie B Virus Receptor-Binding Site Promote Resistance against Soluble Virus Receptor Traps

The emergence of resistant viruses is one of the most frequent obstacles preventing successful therapy of viral infections, representing a significant threat to human health. We investigated the emergence of resistant viruses during treatment with sCAR-Fc, a well-studied, highly effective antiviral molecule against CVB infections. Our data show the molecular aspects of resistant CVB3 mutants that arise during repetitive sCAR-Fc usage. However, drug resistance comes at the price of lower viral fitness. These results extend our knowledge of the development of resistance by coxsackieviruses and indicate potential limitations of antiviral therapy using soluble receptor molecules.

1 g/ml sCAR-Fc. This amount represents an sCAR-Fc concentration that inhibited viral replication between 90% and 99% (22), allowing survival of enough virus progeny to enable passaging at a constant MOI. In two independent experiments, HeLa cells were subsequently infected after sCAR-Fc incubation, and replication was quantified after 24 h. The replication inhibition efficiency of sCAR-Fc was calculated relative to CVB3 Nancy replication after preincubation in the absence of sCAR-Fc. For serial passaging, viral progeny from each round of sCAR-Fc treatment was incubated again, with or without sCAR-Fc, and the efficiency of inhibition was calculated as previously explained.
In the first round, sCAR-Fc inhibited the replication of CVB3 Nancy by more than 99% (Fig. 1A). After the third round of sCAR-Fc incubation, the inhibitory effect significantly declined in both experiments, but each in a different manner. In the first experiment, the inhibition of CVB3 replication suddenly fell from 99.9% to 65% (Fig. 1A, left graph), while in the second experiment the antiviral effect of sCAR-Fc declined in a stepwise manner from 99.9% in the first round to 76.7% after four rounds of sCAR-Fc incubation (Fig. 1A, right graph). During repeated exposure to sCAR-Fc, virus mutants arose that were less susceptible or even resistant to sCAR-Fc-induced neutralization.
Interestingly, for another commonly used CVB3 strain, CVB3-H3, no decrease in the antiviral efficacy of sCAR-Fc during serial passaging under various experimental conditions was observed (Table 1).
For subsequent analysis, eight clones from each serial passaging experiment were isolated by plaque purification from cell lysates after the third round of sCAR-Fc incubation. At this time point of the experiment, the detection of reduced inhibition efficiency by sCAR-Fc indicated the emergence of resistant mutants (Fig. 1A). Cell-killing assays were performed to determine the susceptibility of the isolated clones, CVB3 M1 to CVB3 M8 from the first experiment and CVB3 K1 to CVB3 K8 from the second experiment, to sCAR-Fc. HeLa cells were infected with each clone after preincubation with 1 g/ml or 8 g/ml or without sCAR-Fc, and cell viability was assessed by crystal violet staining. While the cell-killing activity of the CVB3 Nancy wild type was significantly inhibited by 1 g/ml sCAR-Fc and completely prevented by 8 g/ml (Fig. 1B), none of the CVB3 M1 to CVB3 M8 isolated clones were inhibited by even the high-dose sCAR-Fc, as indicated by complete virus-induced cell lysis (Fig. 1B, left). On the other hand, the cell-killing assay of CVB3 K1 to CVB3 K8 showed a few living cells after preincubation with the lower-dose sCAR-Fc (1 g/ml), while treatment with high-dose sCAR-Fc almost completely protected HeLa cells from virus-induced cell lysis (Fig. 1B, right). This indicates that in both serial passaging experiments involving sCAR-Fc treatment, CVB3 Nancy mutants arise with reduced susceptibility to sCAR-Fc. While CVB3 K clones seemed to be less susceptible to sCAR-Fc, the CVB3 M clones from the first experiment were completely resistant.
Next, we performed neutralization assays to quantify the remaining inhibition efficiency of sCAR-Fc more precisely. To this end, 10,000 PFU of each analyzed CVB3 M clone and CVB3 K clone was preincubated with a high concentration of sCAR-Fc (8 g/ml) at 37°C. Interaction of sCAR-Fc with the virus at this physiological temperature immediately leads to the formation of noninfectious A-particles, thereby increasing the antiviral efficiency (11,22). While the parental CVB3 Nancy strain was completely neutralized by sCAR-Fc ( Fig. 1C and D), it did not impair the infectivity of the isolated clones CVB3 M2 to CVB3 M5 , confirming their complete resistance. As observed in the cell-killing assay, the isolated CVB3 K clones showed a decreased susceptibility to sCAR-Fc-induced neutralization, but, in contrast to the CVB3 M clones, they were not completely resistant (Fig. 1D). The neutralization efficacy for all four analyzed CVB3 K clones was about 98%.
We next determined the long-term stability of sCAR-Fc-resistant mutants against sCAR-Fc-catalyzed conversion into A-particles. Therefore, we incubated one representative of each of the resistant mutants, CVB3 M2 and CVB3 K4 , as well as the susceptible wild-type CVB3 Nancy, with sCAR-Fc over a period of about 10 h at 37°C. The sCAR-Fc-catalyzed loss of infectivity was quantified by determination of infectious particles using plaque assays. As shown in Fig. 1E, the parental CVB3 Nancy, as a highly sCAR-Fc-susceptible strain, rapidly lost its infectivity with a calculated half-life (t 1/2 ) of 0.8 h. Compared to short-term incubation with sCAR-Fc ( Fig. 1B and C), where the FIG 1 Resistant CVB3 Nancy mutants emerge after repetitive sCAR-Fc exposure. (A) Inhibition efficacy of sCAR-Fc during serial passaging of CVB3 Nancy with sCAR-Fc in two independent experiments. Virus was preincubated with 1 g/ml sCAR-Fc at 4°C for 30 min, and HeLa cells were subsequently infected for 24 h. The amount of virus was determined by plaque assay, and inhibition efficacy was calculated as a percentage relative to virus replication without sCAR-Fc preincubation. Student's t tests were performed. *, P Ͻ 0.05; ***, P Ͻ 0.001. (B) Eight clones were selected from the third round of sCAR-Fc passaging (A) by plaque purification, and antiviral efficacy of sCAR-Fc was analyzed in a cell-killing assay. HeLa cells were infected with each isolated clone (MOI, 0.1) after preincubation with 1 g/ml, 8 g/ml, or no sCAR-Fc (30 min, 4°C), and cells were stained 36 h p.i. with crystal violet. Cells infected with the wild-type CVB3 Nancy, uninfected cells, and mock-treated cells served as controls. (C) Isolated virus mutants from the first serial passaging experiment under sCAR-Fc treatment (A, left graph) are completely resistant to sCAR-Fc-induced neutralization. The plaque-purified CVB3 mutants CVB3 M2 to CVB3 M5 and 10,000 PFU of the original CVB3 Nancy strain were preincubated with 8 g/ml sCAR-Fc for (Continued on next page) CVB3 M clones revealed complete resistance, long-term sCAR-Fc incubation of the parental and mutant strains showed that CVB3 M2 is not completely resistant to sCAR-Fc per se. However, the sCAR-Fc-triggered conversion of CVB3 to noninfectious particles is greatly slowed in the CVB3 M2 mutant, with a calculated half-life of 14.9 h. Moreover, the half-life of the CVB3 M2 mutant was much longer than the half-life of CVB3 K4 (t 1/2 ϭ 3.9 h), providing additional evidence for the higher resistance of the CVB3 M mutants compared to that of the CVB3 K mutants (Fig. 1E). In conclusion, during repeated incubation of CVB3 Nancy with low concentrations of sCAR-Fc, viral mutants emerge that are less susceptible or nearly completely resistant to sCAR-Fc. Our data also revealed that CVB3 M2 is completely resistant to short-term sCAR-Fc incubation, while, during long-term incubation, a greatly slowed conversion of CVB3 M2 to noninfectious particles occurs.
A single mutation in the virus-receptor recognition site is sufficient for sCAR-Fc resistance. To gain insight into underlying mechanisms of emerging sCAR-Fc resistance, the complete capsid region of the isolated resistant clones was sequenced and compared to that of the CVB3 Nancy strain. As expected from the different phenotypes, the virus mutants from the first (CVB3 M ) and second (CVB3 K ) serial passaging experiments showed nucleotide changes that result in different single-site amino acid substitutions ( Fig. 2A). All analyzed clones from one passaging experiment revealed the same substitution. The completely resistant CVB3 M clones are mutated in the coding sequence of the VP2 protein, leading to an amino acid exchange from Asn (N) to Ser (S) at position 139, here referred to as N2139S. The less susceptible CVB3 K clones are mutated at amino acid residue 150 in the VP1 protein, changing Val (V) to Ala (A), here referred to as V1150A. Both amino acids, N2139 and V1150, are predicted to be located within the receptor footprint of the viral capsid (6,9), most likely directly affecting the virus-receptor interaction. 30 min at 37°C, infected for 30 min at 37°C, and subsequently overlaid with agar. Plaques were counted 2 to 3 days after infection. Values are given as means Ϯ SD from three independent experiments. Student's t tests were performed. ns, not significant. (D) Isolated virus mutants from the second serial passaging experiment under sCAR-Fc treatment (A, right graph) are less susceptible to sCAR-Fc-induced neutralization. Plaque-purified CVB3 K clones were incubated with 8 g/ml sCAR-Fc for 30 min at 37°C, and infectious virus was measured as described for panel C. The original CVB3 Nancy strain served as a control. Values are given as means Ϯ SD from three independent experiments. Student's t tests were performed. **, P Ͻ 0.01; ***, P Ͻ 0.001. (E) The loss of infectious CVB3 over time at 37°C in the presence of sCAR-Fc has first-order decay kinetics. Wild-type strain CVB3 Nancy and the sCAR-Fc resistant mutants CVB3 M2 and CVB3 K4 were incubated at 37°C with 32 ng/ml (0.25 nM) sCAR-Fc, and the conversion to noninfectious A-particles was analyzed over time by plaque assay. V 0 is the concentration of infectious virus at the start of the incubation at 37°C, and V t is the amount of infectious virus that remains after the indicated time points. Virus infectious half-life was determined with the equation t 1/2 ϭ ln 0.5/Ϫk, with k as the first-order rate constant. which had been preincubated with 1.0 g/ml sCAR-Fc when incubated at 4°C or 0.5 g/ml sCAR-Fc when incubated at 37°C. Virus replication was assessed 24 h later by plaque assay. Viral progeny resulting from the first round of sCAR-Fc treatment was used for the second round, again incubated with sCAR-Fc and propagated as described above. Inhibition efficacy was calculated in percentage relative to virus replication without sCAR-Fc preincubation. Values are given as means Ϯ SD from two independent experiments. In further investigations, we focused on the completely resistant CVB3 M clones and examined whether the single-amino-acid exchange in VP2 is responsible for resistance to sCAR-Fc. To analyze whether substitution N2139S indeed induces complete sCAR-Fc resistance, we changed the coding sequence for amino acid N2139 by site-directed mutagenesis into the coding sequence for serine (S) in the cDNA clone of CVB3-H3, generating CVB3-H3 N2139S . A replication inhibition assay demonstrated that both viruses with the N2139S substitution, CVB3 M2 selected during sCAR-Fc exposure of CVB3 Nancy and CVB3-H3 N2139S generated by mutagenesis of CVB3-H3, were completely resistant to sCAR-Fc, whereas both parental viruses were efficiently inhibited (Fig. 2B). Moreover, as shown for CVB3 M2 , long-term incubation of CVB3-H3 N2139S with sCAR-Fc revealed a greatly slowed and less pronounced transition to noninfectious particles over time compared to the sCAR-Fc-susceptible parental strain CVB3-H3 (Fig. 2C). Altogether, these data demonstrate that the N2139S mutation, acquired by CVB3 M2 during viral passaging in the presence of sCAR-Fc, indeed determines the development of resistance to sCAR-Fc.
Given the complete resistance of CVB3 M2 and CVB3-H3 N2139S to sCAR-Fc-induced conversion to noninfectious A-particles (11,12), we asked whether CVB3 strains with the N2139S mutation still use CAR as a cellular receptor for binding and uptake. To address this issue, we infected the CAR-deficient cell line CHO-K1 (Fig. 2D) with the resistant mutants CVB3 M2 and CVB3-H3 N2139S as well as with their parental strains CVB3 Nancy and CVB3-H3. Determination of infectious viruses by plaque assay directly after incubation (0 h postinfection [p.i.]) and 24 h later (24 h p.i.) showed that the sCAR-Fcresistant mutants, just as the parental strains, were unable to infect CHO-K1 cells lacking CAR (Fig. 2E). Additional infection experiments with CHO-K1 cells stably expressing CAR on the cell surface (Fig. 2D, CHO-CAR) revealed a markedly increased virus concentration 6 h after infection for all virus strains (Fig. 2F), illustrating that both the parental viruses and the sCAR-Fc-resistant variants use CAR as a cellular receptor, independent of the efficiency of their neutralization by sCAR-Fc.
The sCAR-Fc resistant viruses exert reduced viral fitness. Having shown that N2139S amino acid substitution does not prevent CAR-mediated virus uptake, the replication of both sCAR-Fc resistant mutants, CVB3 M2 and CVB3-H3 N2139S , was compared to those of their respective parental wild-type viruses in one-step growth curves in HeLa cells. Both resistant virus mutants replicated with slower kinetics and generated fewer progeny than their respective parental strains (Fig. 3A), confirming results obtained in CHO-CAR cells (Fig. 2F). Complementarily, RNA replication (Fig. 3B) and expression of viral proteins (Fig. 3C) were impaired for the resistant mutants. Both mutants had a diminished plaque size compared to their parental viruses when assayed on HeLa cells 2 days p.i. (Fig. 3D). As a consequence of the decreased viral replication, resistant CVB3 mutants were less cytopathic (Fig. 3E). Moreover, virus competition assay confirmed that the resistant mutants are inferior to their parental strains in terms of their replication efficiency (Fig. 4A). In fact, the sequencing chromatograms 8 h after infection with initial equal amounts of CVB3/CVB3 M2 or CVB3-H3/CVB3-H3 N2139S (0 h p.i.) show that the wild-type virus sequences are overwhelmingly prevalent, while the mutated nucleotide, inducing the amino acid substitution Asn/N to Ser/S, is minimally present (CVB3/CVB3 M2 ) or completely absent (CVB3-H3/CVB3-H3 N2139S ). These results point out that the amino acid substitution N2139S in CVB3 M2 and CVB3-H3 N2139S generated a virus variant with an inferior ability to replicate.
Besides replication efficiency, the infectious half-life of a virus at physiological temperature is another indicator for viral fitness (32). At 37°C, dependent on the

DISCUSSION
Initially, it was suggested that viruses resistant to soluble receptors could not emerge, because mutations that interfere with virus-receptor binding would compromise the ability of the virus to infect host cells, leading to its rapid extinction from the viral population (27). In this study, we investigated the development of resistance of CVB3 against the soluble receptor molecule sCAR-Fc, which has been suggested as a potent antiviral compound against various CVB serotypes. Contrary to the assumption that soluble receptor-resistant viruses cannot evolve, we isolated different sCAR-Fcresistant CVB3 mutants after in vitro serial passaging under low-dose sCAR-Fc exposure. It is particularly remarkable that resistant mutants could emerge among picornaviruses, because, on top of the simple competition for receptor binding sites, the sCAR-Fcinduced formation of noninfectious A-particles potentiates the antiviral effect of soluble virus receptors (11,22). On the other hand, the low fidelity of RNA virus polymerases leads to very error-prone replication, contributing to the generation of high population diversity, which facilitates adaption of the virus to its current environment (33). Here, we report on two different mutations in the capsid proteins of CVB3, which were responsible for decreased susceptibility (V1150A, CVB3 K clones) or nearly complete resistance (N2139S, CVB3 M clones) of CVB3 to sCAR-Fc. Cryo-EM reconstructions predict that the identified amino acid exchanges that mediate their resistance are located within the CVB3-CAR interaction footprint. Hence, the defined mutations in the viral capsid proteins are most likely involved in the interaction of the virus with its cellular receptor CAR (6,9), explaining their major impact on sCAR-Fc susceptibility. In accordance with our findings, the mutations in the majority of PV mutants known to be resistant to soluble PV receptor (PVR), and also Pleconaril-resistant CVB, were detected in the virus-receptor interaction site, verifying the biological relevance of this area for resistance development against binding inhibitors (25,26,30).
Interestingly, the position mutated in CVB3 M2 , N2139, has been found to be one of the amino acids contributing to the CVB3-CAR footprint. It is the only conserved amino acid in the otherwise highly variable VP2 puff region among different CVB serotypes (9). The conserved nature of N2139 was also shown by next-generation sequencing of CVB3 during long-term passaging in cell culture studies. During 40 passages, no minor or major variants with amino acid exchange in position 2139 were identified (33). Given the nearly complete resistance of the viruses CVB3 M2 and CVB3-H3 N2139S investigated here , N2139 seems to be especially important for receptor binding and uptake by A-particle formation (9).
During this study, resistant mutants were isolated during serial passaging of the CVB3 Nancy wild type but not with CVB3-H3. Sequence comparison of the two strains revealed eight amino acid differences in the capsid region. Although not all of these sequence differences are necessary to allow manifestation of complete resistance, as proven by insertion of N2139S into CVB3-H3, it is conceivable that the Nancy genotype somehow supports the development of sCAR-Fc resistance, whereas CVB3-H3 does not. The sequence differences in position VP3 residue 180 of CVB3-H3 (A3180) and CVB3 Nancy (V3180) have previously been described as influencing the stability of the virus and probably affect receptor binding and capsid expansion (32). Furthermore, the leucine at position 1092, detected in CVB3 Nancy but not in CVB3-H3, is associated with lower viral stability. This polymorphism is, in general, rarely found among clinical CVB isolates but accumulates in the laboratory strains and strains with resistance to antiviral drugs. This indicates that a less stable virus, such as CVB3 Nancy, is favored under the selection conditions present in vitro (32,34). The VP2 residue 138, located right next to the amino acid that can confer resistance to sCAR-Fc (N2139S), seems to be of further interest. CVB3 Nancy with glutamic acid (E) in this position efficiently binds to the coreceptor DAF, while CVB3-H3, with asparagine (N) in this position, does not measurably bind to DAF (35,36). Thus, the interaction with coreceptors, such as DAF, appears to support the evolution and maintenance of resistant viruses by partially compensating for a decreased binding affinity to CAR.
CAR and sCAR are believed to be bound by CVB3 in the same way, so it is astonishing that the sCAR-Fc resistant mutants CVB3 M2 and CVB3 N2139S are still able to infect cells using CAR. Nevertheless, comparing the replication of the sCAR-Fcsusceptible CVB3 strains and the resistant viruses bearing the N2139S substitution, we observed decreased viral protein expression, RNA multiplication, and production of viral progeny for the sCAR-Fc-resistant strains, leading to smaller plaque size and a diminished virus-induced cytotoxicity. A virus competition assay revealed superior replication properties of the wild-type strains CVB3 Nancy and CVB3-H3 compared to their respective N2139S mutants. This finding differs from a previous observation in PV mutants that develop resistance against their soluble receptor. Here, virus propagation was not affected despite decreased virus receptor binding affinity (30). Both soluble receptor-resistant picornavirus mutants, PV and CVB3, still use their original receptor for cell infection (30). We assume that sCAR-Fc resistance is mainly based on delayed or inefficient transition to A-particles, an obligate stage during viral uptake into the host cell. This is supported by our virus/sCAR-Fc decay curves, where the mutants show a slowed transition to noninfectious particles. In addition, the one-step growth curves revealed up to 10-fold more cell-bound infectious CVB3 M2 and CVB3-H3 N2139S directly after virus incubation, while the cellular binding of the virus was similar for the mutant and parental strains (data not shown). This indicates that the parental viruses efficiently converted to A-particles during cellular uptake, while, as reflected by the larger amount of infectious viruses at this stage, the transition process was inefficient for the mutant viruses. Our results indicate an inefficient conversion of the mutant strains to noninfectious A-particles by both soluble and membrane-bound CAR. Thus, it seems conceivable that the restricted replication of the mutant strains CVB3 M2 and CVB3-H3 N2139S , as indicated by diminished RNA multiplication, viral protein expression, and less viral progenies, is mainly a result of the slowed infection of the host cells.
Additional virus binding to coreceptors like DAF or heparan sulfate (37) or other unidentified receptors (18) could explain the existence of virus mutants with impaired CAR-related uptake capacities. It has been shown that DAF interaction increases the CVB3 infection in CaCo cells, where CAR is located in the tight junctions and, therefore, is hard for CVB3 to access (38). Therefore, binding to DAF or other coreceptors might compensate for the compromised CAR-induced uptake mechanism by continuously facilitating the accessibility of the virus to CAR.
In our analysis, we observed that the sCAR-Fc-resistant CVB3 mutants are slightly less stable with respect to their conversion to noninfectious particles than their respective parental strains. While greater stability is generally a useful trait for CVB3, it has been shown that this is not always the case (32). Under optimal cell culture conditions, the survival of weaker virus mutants or viruses with impaired fitness may be facilitated by the high abundance of virus receptors or additional molecules for virus binding (32). However, with slowed cellular uptake, restricted replication, and decreased stability in vitro, the virulence and pathogenicity of the sCAR-Fc-resistant mutants in vivo are questionable. Previous studies have shown that sCAR-Fc can be successfully employed for treatment of CVB3 infections in vivo, with no sign of the development of resistance (17,20,21). While we found sCAR-Fc-resistant isolates can emerge under our laboratory conditions, it remains unclear whether such sCAR-Fcresistant isolates could emerge under clinical conditions. Should that prove to be the case, it would be important to better understand the limitations of sCAR-Fc as an anti-CVB therapeutic. CVB3 quantification by plaque assay. The amount of infectious CVB3 was determined on a monolayer of HeLa cells by plaque assay. Cell culture samples or agar pieces, including isolated virus in phosphate-buffered saline (PBS), were subjected to three freeze/thaw cycles and centrifuged to remove cell debris or agar. The number of infectious virus particles in the samples was quantified using a plaque assay, as previously described (18). Instead of neutral red, the plaques were stained using 0.5% MTT-PBS [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; Merck KGaA].
One-step growth curve. HeLa cell monolayers were infected in triplicates at an MOI of 2 (30 min, 37°C) for each virus. After inoculation, the virus was removed, cells were washed twice with PBS, and culture medium was added. Cells were frozen directly after incubation (0 h p.i.) or 2, 4, 6, 8, and 10 h after infection, and infectious virus was quantified by plaque assay.
Virus neutralization assay. Various amounts of virus were preincubated with or without sCAR-Fc at various concentrations at either 4°C or 37°C. If not otherwise described, after a preincubation period of 30 min, HeLa cells were infected for 30 min at 37°C and subsequently overlaid with agar. Emerging plaques were counted after 2 to 3 days, and percent neutralization efficiency was calculated relative to virus preincubated without sCAR-Fc. For virus plaque purification, single plaques were isolated under sterile conditions using a truncated 1,000-l pipette tip. Agar pieces, including virus and cell debris, were resuspended in PBS, and infectious virus was quantified by plaque assay after three freeze/thaw cycles, as already described.
Virus stability assay. Virus stabilities and receptor-mediated conversion to noninfectious A-particles were assessed in terms of the first-order rate constant for inactivation at 37°C, as described previously in detail (9). Briefly, 1 ϫ 10 7 PFU of each analyzed virus was incubated without or with sCAR-Fc (32 ng/ ml ϭ 0.25 nM) at 37°C. Aliquots were taken at different time points, and the amount of infectious virus was quantified by standard plaque assay, as described above. The conversion of CVB3 to noninfectious A-particles with or without receptor was described as ln(V t /V 0 ) over time, where V t is the concentration of infectious virus particles at the time (t) of incubation and V 0 is the infectious virus concentration at the beginning. The curves were fitted by linear regression and slopes were calculated using GraphPad Prism 7.00 (GraphPad Software, San Diego, CA, USA). The decay over time was further expressed by calculated half-life (t 1/2 ϭ ln 0.5/Ϫk, with k as the first-order rate constant).
Virus replication inhibition assay. Virus was preincubated with or without sCAR-Fc for 30 min at 4°C or 37°C, and HeLa cells were infected as described before. After infectious virus was removed, cells were washed with PBS and cultured for an additional 8 h to 24 h with HeLa cell growth medium.
Virus competition assay. HeLa cells were incubated with the virus pairs CVB3 Nancy/CVB3 M2 and CVB3-H3/CVB3-H3 N2139S at an MOI of 1 of each virus for 1 h at 37°C. After two wash steps with PBS, RNA was isolated directly after incubation (0 h p.i.) and 8 h later (8 h p.i.). RNA was reverse transcribed using MLV-RT in combination with random hexamer primers. Viral genomic fragments containing the resistance-inducing mutation were amplified by PCR with the primers CVB3-seq1 (5=-CCG ATG CTT TGT CGA ACT TAG G-3=) and CVB3-seq2 (5=-CCT CTG TAC CAA CTT GTT GGA CC-3=). Fragments were analyzed by Sanger sequencing.
Statistical analysis. Statistical analysis of the results was performed with GraphPad Prism 7.00 (GraphPad Software). Results are shown as means Ϯ standard deviations (SD) for each group. Unpaired Student's t test was used for two-group comparison. If values were normalized to an internal control, one-sample t test was applied. For multiple-group comparison, two-way analysis of variance (ANOVA) was performed. Differences were considered significant at a P value of Ͻ0.05.