|Home | About | Journals | Submit | Contact Us | Français|
The acid-dependent disassembly of foot-and-mouth disease virus (FMDV) is required for viral RNA release from endosomes to initiate replication. Although the FMDV capsid disassembles at acid pH, mutants escaping inhibition by NH4Cl of endosomal acidification were found to constitute about 10% of the viruses recovered from BHK-21 cells infected with FMDV C-S8c1. For three of these mutants, the degree of NH4Cl resistance correlated with the sensitivity of the virion to acid-induced inactivation of its infectivity. Capsid sequencing revealed the presence in each of these mutants of a different amino acid substitution (VP3 A123T, VP3 A118V, and VP2 D106G) that affected a highly conserved residue among FMDVs located close to the capsid interpentameric interfaces. These residues may be involved in the modulation of the acid-induced dissociation of the FMDV capsid. The substitution VP3 A118V present in mutant c2 was sufficient to confer full resistance to NH4Cl and concanamycin A (a V-ATPase inhibitor that blocks endosomal acidification) as well as to increase the acid sensitivity of the virion to an extent similar to that exhibited by mutant c2 relative to the sensitivity of the parental virus C-S8c1. In addition, the increased propensity to dissociation into pentameric subunits of virions bearing substitution VP3 A118V indicates that this replacement also facilitates the dissociation of the FMDV capsid.
Foot-and-mouth disease virus (FMDV) is a member of the Aphthovirus genus in the family Picornaviridae. FMDV displays epithelial tropism and is responsible for a highly contagious disease of cloven-hoofed animals (23, 60). FMDV populations are quasispecies and exhibit a high potential for variation and adaptation, one consequence of which is the extensive antigenic diversity of this virus, reflected in the existence of seven serotypes and multiple antigenic variants (reviewed in references 17 and 60). Different cellular receptors, including αvβ integrins and heparan sulfate (HS) glycosaminoglycans, have been described for natural isolates and tissue culture-adapted FMDVs (3, 4, 6, 28-31, 56). However, viruses that are infectious in vivo use integrins as receptors (28). The interaction between FMDV and the integrin molecule is mediated by an Arg-Gly-Asp (RGD) triplet located at the G-H loop of capsid protein VP1 (9, 47). FMDV isolates interacting with integrins gain entry into the cell following clathrin-mediated endocytosis (8, 39, 52). On the other hand, it has been described that a genetically engineered HS-binding mutant uses caveolae to enter into cultured cells (51). After internalization, FMDV must release its genomic RNA molecule of positive polarity into the host cell cytoplasm to establish a productive infection. Early work showed that a variety of lysosomotropic agents, such as weak bases and ionophores that block acidification of endosomes, inhibit FMDV infection (5, 11-13), indicating that genome release is dependent on endosomal acidification. In addition, internalized FMDV particles colocalize with markers from early and recycling endosomes (8, 51, 52) and FMDV infection is reduced by expression of a dominant negative mutant of Rab5 (33), suggesting that FMDV may release its genome from these compartments.
The FMDV capsid comprises 60 copies of each of the four structural proteins (VP1 to VP4) arranged in an icosahedral lattice of 12 pentameric subunits. FMDV particles are highly acid labile and disassemble at pH values slightly below neutrality (13). Acid lability is not a feature of the capsids of other picornaviruses, such as Enterovirus. Pentameric subunits are intermediates of FMDV assembly and disassembly (64). A high density of His residues is found close to the interpentameric interface. Protonation of these residues at the acidic pH in the endosomes has been proposed to trigger acid-induced capsid disassembly by electrostatic repulsion between the protonated His side chains (1). His 142 (H142) in VP3 of type A FMDV is involved in a His-α-helix dipole interaction, which is likely to influence the acid lability of FMDV (13). In silico predictions suggested that H142 and H145 in VP3 may have the greatest effect on this process (63). Experimental evidence of the involvement of H142 of VP3 in acid-induced disassembly of FMDV has also been reported (20). Concomitantly with capsid disassembly into pentameric intermediates, internal protein VP4 and viral RNA are released. VP4 is a highly hydrophobic and myristoylated protein (7) whose release has been suggested to mediate membrane permeabilization and ion channel formation, thus facilitating the endosomal exit of viral RNA (15, 16, 34).
Besides providing information about the endosomal pH requirements for the release of virus genomes, drugs modifying endosomal acidification can reveal the molecular changes associated with viral resistance to their action. These analyses may also address whether the balance between acid lability and capsid stability required for completion of virus replication allows FMDV, which disassembles at a pH close to neutrality, to escape inhibition by drugs raising the endosomal pH. In this work, we have isolated and characterized FMDV mutants that are able to escape from the inhibition of endosomal acidification exerted by NH4Cl, a lysosomotropic weak base that raises endolysosomal pH and impairs uncoating and infection of viruses that require transit through acidic endosomal compartments for penetration (5, 26, 53). These mutants showed an increased acid lability, which is likely to allow them to uncoat at more-alkaline pH values. A single amino acid substitution close to the interpentameric interfaces in the capsid of one of these mutants was responsible for a total resistance to the elevation in endosomal pH caused by NH4Cl treatment and for the acid-labile phenotype.
The origin and culture procedures for BHK-21, IBRS-2, and CHO cells have been described previously (39, 40). C-S8c1 is a biological clone from a type C FMDV isolate (59). FMDV MARLS is a monoclonal antibody (MAb)-resistant mutant selected with MAb SD6 from a population of highly tissue culture-passaged C-S8c1 (4). Bovine enterovirus (BEV) (32) and FMDV stocks were prepared by amplification in BHK-21 cells.
FMDV structural proteins were detected using mouse MAb SD6, which recognizes antigenic site A located in the G-H loop of VP1 (46), and MAb 5C4, which recognizes the discontinuous antigenic site D (36). Rabbit polyclonal serum 163, raised against the purified recombinant FMDV 3A protein expressed in Escherichia coli (55), was used to detect FMDV nonstructural protein 3A. Rabbit EEA1 antibody was from Cell Signaling. Mouse MAb anti-β-actin AC-15 was from Sigma. Anti-mouse and anti-rabbit IgG secondary antibodies coupled to Alexa Fluor 488 (AF488) or AF555, used in immunofluorescence assays, and AF488-labeled phalloidin were from Molecular Probes (Invitrogen). Enhanced chemiluminescence (ECL) anti-mouse IgG horseradish peroxidase-coupled secondary antibodies used in Western blot assays were from GE Healthcare. NH4Cl (Merck) was dissolved in water (1 M). A stock solution of 28.8 μM concanamycin A (ConA) (Sigma) was prepared in dimethyl sulfoxide (DMSO).
Infections were carried out as previously described (39). Cells were infected at various multiplicities of infection (MOI) (defined as PFU/cell), and after 1 h, the virus inoculum was removed and fresh medium containing 5% fetal calf serum (FCS) was added. This time point was considered 1 h postinfection (p.i.). After 8 h of infection, the total virus yield (intracellular and extracellular, recovered from monolayers frozen and thawed three times) was determined by a plaque assay using BHK-21 cells (39, 40, 59). Virus titers were determined by counting the number of lysis plaques that developed after 24 h (FMDV) or 48 h (BEV) of infection.
Cell monolayers grown on glass coverslips were washed and incubated for 30 min at 37°C with the RGD-containing peptide SARGDLAHLTTTHAR (positions 139 to 153 of VP1 in C-S8c1) or control peptide PTAYHKGPVTRLALP (positions 104 to 118 of VP1) diluted in phosphate-buffered saline (PBS) completed with Ca2+ and Mg2+. The final peptide concentration was 1 mM. Cells were washed again and incubated with viruses diluted in complete PBS for 25 min at 37°C. Then, cells were fixed and processed for immunofluorescence.
To determine FMDV acid sensitivity, a modification of a procedure previously described (34) was used. Briefly, an aliquot (10 μl) containing about 2 × 106 PFU of the virus tested was mixed with 300 μl of PBS solutions (50 mM NaPO4 and 140 mM NaCl) of different pHs for 30 min at room temperature. Then, the solution was neutralized by adding 100 μl of 1 M Tris (pH 7.6), and the remaining PFU in each sample was determined by a plaque assay using BHK-21 cells. The number of PFU developed was counted and expressed as the percentage of infectivity compared to that obtained using PBS at pH 7.5.
A previously published procedure (42, 44) was followed for radioactive labeling and purification of virions. Briefly, FMDV virions (C-S8c1 and mutants) were metabolically labeled with [35S]methionine (EasyTag Express protein labeling mix; PerkinElmer) during infection of BHK-21 cells. The total virus produced was centrifuged through a cushion of 20% sucrose in TNE buffer (10 mM Tris-HCl [pH 7.5], 0.1 M NaCl, 1 mM EDTA) in an AH-625 rotor (Sorvall) at 25,000 rpm for 2.5 h at 4°C. The viral pellet was resuspended in TNE buffer and centrifuged in 7.5% to 30% sucrose density gradients in the same buffer at 37,000 rpm for 1 h at 4°C in an SW40 rotor (Beckman). The fractions containing full virions (sedimentation coefficient, 140S) were pooled and extensively dialyzed against PBS. The integrity of virions was analyzed in 10% to 45% sucrose density gradients at 18,000 rpm for 18 h at 4°C in an SW40 rotor (Beckman). The relative amounts of full virions (140S) and pentameric subunits (12S) were estimated from the associated radioactivity as determined with a liquid scintillation counter.
Inhibition of endosomal acidification with NH4Cl was performed as described previously (5, 39). Cells were treated for 1 h prior to infection with 25 or 50 mM NH4Cl. To buffer the extracellular pH, 25 mM HEPES at pH 7.4 was added to the culture medium. The drug was maintained throughout the infection to avoid cellular recovery. To determine the possible effects of NH4Cl on virus replication, the drug was added 3 h p.i. and maintained throughout the rest of the assay. In the case of concanamycin A (ConA) treatment, cells were pretreated for 30 min with 100 nM ConA and the drug was maintained only during the first hour of infection, as reported previously (8, 39). Control cells were treated in parallel, with the same amount of DMSO.
Cells grown on glass coverslips were infected as described above, fixed, and processed for immunofluorescence (22, 39, 40). Samples were examined using an Olympus BX61 epifluorescence microscope coupled to a DP71 digital camera (Olympus) or with an Axioskop (Zeiss) epifluorescence microscope coupled to a Coolsnap FX digital monochrome camera (Roper Scientific). The percentage of 3A-positive cells was determined by scoring triplicates of at least 300 cells per coverslip per condition. For confocal microscopy, an LSM510 META laser scanning confocal microscope (Zeiss) was used, and images were acquired using Zeiss LSM510 4.2 Sp2 software and processed using an LSM image browser (Zeiss) and Adobe Photoshop CS2.
Western blotting was performed as described previously (22, 39, 40). Briefly, cells were lysed on ice into lysis buffer (10 mM EGTA, 2.5 mM MgCl2, 1% NP-40, 20 mM HEPES, pH 7.4), and equal amounts of protein as determined by the Bradford assay (8a) were mixed with Laemmli sample buffer and resolved by SDS-PAGE. Proteins were transferred onto a nitrocellulose membrane, blocked with skimmed milk, and probed with primary antibodies and subsequently with secondary antibodies coupled to horseradish peroxidase. Signals were detected using Western Lighting chemiluminescence reagent (Perkin Elmer).
Viral RNA was extracted from supernatants of infected cell cultures by using TRI reagent (Sigma) as described by the manufacturer. cDNA was synthesized by reverse transcription of viral RNA using murine leukemia virus (MuLV) reverse transcriptase (Roche) and previously described oligonucleotide primers (2, 49, 61). cDNA was amplified by PCR using the desired oligonucleotide primers and BioTaq DNA polymerase (Bioline) supplemented with 10% Expand high-fidelity polymerase (Roche) for proofreading activity. PCR products were purified with Wizard SV gel and a PCR clean-up system (Promega) and sequenced by automatic DNA sequencing at the Genomic Unit of the Madrid Science Park at the Universidad Autónoma de Madrid. DNA sequences were confirmed by at least two independent sequencing reactions performed using different oligonucleotide primers. Nucleotide positions correspond to those previously described for the FMDV C-S8c1 isolate (61).
Plasmid pMT28 (21), which contains the full-length cDNA of FMDV C-S8c1, was used to construct plasmid pMT28-VP3 A118V/VP1 N47D bearing the mutations found in the capsid of NH4Cl-resistant c2 FMDV (Table (Table1).1). First, cDNA obtained from reverse transcription of RNA extracted from the c2 virus was digested with SfiI and AvrII, which cut at nucleotides 2828 and 3758 in the C-S8c1 sequence within the VP3 and VP1 coding regions, respectively (61). The resultant fragment was purified by agarose gel electrophoresis and ligated into pMT28 digested with the same enzymes by use of T4 DNA ligase (Promega). Infectious clone pMT28-VP1 N47D was generated by triple ligation of the following DNA fragments: (i) a 490-bp fragment resulting from digestion of pMT28 with SfiI and AgeI (restriction site located at nucleotide 3317), (ii) a fragment containing a VP1 N47D mutation resulting from digestion of infectious clone pMT28-VP3 A118V/VP1 N47D with AgeI and AvrII, and (iii) pMT28 digested with SfiI and AvrII. Thus, this infectious clone carried only mutation N47D at VP1.
Infectious clone pMT28-VP3 A118V was generated from plasmid pMT28 by introducing a C/T change at position 2903, using the overlap extension method. First, two PCRs were performed using external sense primer ACCTCTACACACACAACCAACACC in combination with antisense mutated primer CGCCACCATGTACCGAACTTTCGCGTCGGTCG or sense mutated primer CGACCGACGCGAAAGTTCGGTACATGGTGGCG combined with antisense external primer GCATCTGGTTAATTGTCTCCACC (the mutated position is indicated in bold). Next, 50 ng of each PCR-purified product was mixed and subjected to PCR using the external primers. The PCR product of the expected size was purified by agarose gel electrophoresis, digested with SfiI and AvrII, and ligated into pMT28 digested with the same restriction enzymes. All restriction endonucleases used were from New England Biolabs. E. coli DH5α was used for amplification of plasmid DNA. The nucleotide sequences of the infectious clones were confirmed by DNA sequencing as described above.
Plasmid DNA from infectious FMDV clones was linearized by digestion with NdeI restriction endonuclease (New England Biolabs). Infectious viral RNA was transcribed from the linearized plasmids by using SP6 RNA polymerase (New England Biolabs). The remaining DNA was eliminated by digestion with RNase-free RQ1 DNase (New England Biolabs), and the amount and purity of the RNA synthesized were determined by agarose gel electrophoresis and UV spectrophotometry using NanoDrop equipment. BHK-21 cells grown on a 35-mm-diameter culture dish were transfected with 0.5 μg of infectious RNA by using Lipofectin reagent (Invitrogen), as indicated by the manufacturer. The viruses recovered from transfections were amplified by two rounds of infection of BHK-21 cells with undiluted transfection supernatants. The nucleotide sequence of the complete capsid region was determined from viral RNA extracted from infected cells, as described above.
The atomic coordinates for FMDV C-S8c1 previously obtained by X-ray crystallography (36) and deposited in the Protein Data Bank were used. The structural model of a pentameric subunit of the FMDV capsid was built using the oligomer generator from the VIPER database (58), and the structures were visualized using PyMOL (DeLano Scientific) (http://www.pymol.org/).
Analysis of variance (ANOVA) using the F Fisher-Snedecor distribution was performed with the statistical package SPSS 15 (SPSS, Inc.) for Windows. In the case of multiple comparisons, Bonferroni's correction was applied. Data are presented as means ± standard deviations (SD). Statistically significant differences are denoted in the figures by one asterisk for a P value of <0.05 or two asterisks for a P value of <0.005.
As previously described (5, 39), NH4Cl inhibited productive FMDV infection in a dose-dependent manner when added preinfection but had no significant effect when added postinfection (p.i.) (Fig. (Fig.1A).1A). This result confirmed that NH4Cl affected only an early step of infection. When the destination of FMDV particles was tracked early upon infection (25 min), fluorescent spots that corresponded to FMDV particles could be detected in control as well as in NH4Cl-treated cells (Fig. (Fig.1B),1B), indicating that NH4Cl did not inhibit virus binding and internalization. Conversely, fluorescence to nonstructural protein 3A, whose detection is diagnostic of virus replication, could be found at 4 h p.i. in control cells but not in NH4Cl-treated cells (Fig. (Fig.1C).1C). In addition, strong fluorescence was observed throughout control cells labeled with MAb 5C4, indicating the assembly of capsid intermediates and/or complete viral particles. MAb 5C4 staining was more intense and colocalized with that of the 3A protein at the perinuclear area, where the replication complex is organized (22, 35, 48). In NH4Cl-treated cells, MAb 5C4 displayed a scattered, endosome-like staining pattern (Fig. (Fig.1C),1C), suggesting that internalized virions were trapped in endosomal compartments. Colocalization between fluorescence against MAb 5C4 and that of an antibody against the early endosome marker protein EEA1 was observed in NH4Cl-treated cells (Fig. (Fig.1D).1D). This confirmed that FMDV particles were retained in early endosomes. The partial colocalization observed between fluorescence to FMDV capsids and to EEA1 could be explained because viral particles may also be retained in other endosomal compartments, such as recycling endosomes, which have also been implied in FMDV entry (33). Taken together, these results indicate that an elevation in endosomal pH, caused by NH4Cl addition, inhibited acid-dependent endosomal uncoating and blocked FMDV infection progress. Neither virus binding nor internalization was inhibited by this drug.
Infection of BHK-21 cell monolayers with FMDV was performed in semisolid agar medium in the presence or absence of 25 mM NH4Cl, a concentration of ammonia that inhibited FMDV yield by 90% without exerting significant cell damage (Fig. (Fig.1A).1A). After 30 h of infection, cells maintained in the presence of NH4Cl showed about 8% of the number of plaques that developed in untreated cells (data not shown). Three of these NH4Cl-resistant plaques were picked and amplified twice in BHK-21 cell monolayers grown in liquid medium containing 25 mM NH4Cl. These amplified viral clones were termed mutants c1, c2, and c3. The virus titers determined for c1 and c2 were about 2.4 × 106 and 1.5 × 107 PFU/ml, respectively, values similar to those found for C-S8c1 infections (see below). In addition, the sizes of the plaques recovered from c1 and c2 in the presence or in the absence of NH4Cl were similar to that for C-S8c1 (data not shown). On the other hand, the virus titer of mutant c3 was 2.8 × 106 PFU/ml, and plaques formed in the presence or absence of NH4Cl were significantly smaller (46% smaller on average) than those for c1 and c2 (P < 0.000). Thus, although mutant c3 showed a slightly decreased growth capacity, no severe effects on virus multiplication were found in the mutants resistant to NH4Cl.
Tissue culture-adapted FMDVs can use cellular receptors different from integrins (3, 4, 56, 66). This prompted us to test the possibility that FMDV clones c1, c2, and c3 had gained the ability to be internalized using a different cellular receptor that resulted in virus uncoating. To this end, BHK-21 cells were incubated with synthetic peptides containing the RGD motif in VP1 or a control peptide corresponding to a different VP1 region. Next, cells were incubated with the following virus: C-S8c1, c1, c2, c3, or MARLS, a derivative of C-S8c1 that is internalized using cellular receptors different from integrins (3, 4). For all viruses, fluorescence to capsid proteins, revealed with MAb 5C4, could be detected (at 25 min p.i.) in cells treated with non-RGD-containing control peptide (Fig. (Fig.2A).2A). However, among cells treated with RGD-containing peptide, only those infected with MARLS exhibited fluorescence to MAb 5C4. When these viruses were tested with CHO cells, which lack expression of the integrin receptors used by FMDV (24), neither C-S8c1 nor the mutant viruses induced a cytopathic effect or produced infective viruses, while MARLS was infectious (Fig. (Fig.2B2B and data not shown). Taken together, these results indicated that FMDV mutants c1, c2, and c3 use integrins as cellular receptors to gain entry into cultured cells.
Infection by mutants c1, c2, and c3 was less inhibited by NH4Cl than that by C-S8c1 (Fig. (Fig.3A).3A). However, differential sensitivity to the drug was observed among the mutants, with c2 being the only mutant that produced the same virus titer in the presence and in the absence of NH4Cl. In addition, c3 was more resistant to NH4Cl than c1. When the acid sensitivities of the infectious virions were compared using inactivation assays, all of them proved to be more acid labile than the parental C-S8c1 virion. Differential behavior among the three viruses was also noted (Fig. (Fig.3B).3B). Mutant c2 was the most acid labile, with a pH50 value (defined as the pH value leading to a 50% loss of infectivity) of 6.95. Mutants c1 and c3 displayed intermediate acid lability (pH50 values of 6.70 for c1 and 6.78 for c3), between that of c2 and that of C-S8c1 (pH50 of 6.58). BEV, a picornavirus with a capsid that is stable over a wide pH range (37), did not show any loss of infectivity within the pH range tested. Taken together, these results indicate that resistance to inhibition of endosomal acidification by NH4Cl correlates with an increase in acid sensitivity of the FMDV virion.
Mutant c2 was completely resistant to NH4Cl, as no significant differences in virus yield were noticed when c2 was grown in the presence or absence of this drug (Fig. (Fig.3A).3A). Therefore, this virus was selected for further analysis. The lack of effect of NH4Cl treatment on the viral growth of the c2 mutant was consistent with the lack of effect on the amount of viral protein expressed in infected cells. Thus, contrary to what was observed for C-S8c1, neither the number of cells positive for fluorescence to the 3A nonstructural protein nor the amount of VP1 protein detected by Western blotting was reduced by the treatment with NH4Cl (Fig. 4A and B). This confirms that mutant c2 infection is resistant to inhibition of endosomal acidification by NH4Cl.
In porcine-derived IBRS-2 cells, infection with C-S8c1 was inhibited by treatment with NH4Cl (Fig. (Fig.4C),4C), as previously reported (39). In contrast, this treatment did not affect infection with mutant c2. In this cell line, infection with FMDV C-S8c1 has been reported to be sensitive to ConA (39), a potent and specific inhibitor of V-ATPase that results in blockage of endosomal acidification (19). As shown in Fig. Fig.4D,4D, infection of IBRS-2 cells by C-S8c1, but not by c2 virus, was inhibited by ConA. Taken together, these results indicate that mutant c2 is able to escape from inhibition of endosomal acidification by different drugs and in different cell types.
To identify the genotypic changes responsible for the NH4Cl-resistant and acid-labile phenotype of mutants c1, c2, and c3, the complete capsid coding regions of these mutants and the parental C-S8c1 virus were sequenced and compared (Table (Table1).1). Two nucleotide substitutions were found in the capsid of mutant c1, one synonymous in VP2 and a second leading to amino acid substitution A123T in VP3. Mutants c2 and c3 showed three nucleotide substitutions each, one (common) synonymous in VP3 and two leading to amino acid replacements: A118V in VP3 and N47D in VP1 in the c2 virus, and D106G in VP2 in combination with H151R in VP1 in the c3 virus. Only one replacement in each of the viruses (c1, VP3 A123T; c2, VP3 A118V; c3, VP2 D106G) was located at an invariant position of the FMDV capsid (10). Replacement N47D in VP1 found in the c2 virus was located within a region of positive selection (38), and mutation H151R in VP1 of c3 was found within the G-H loop of VP1, close to the antigenic site A (positions 130 to 150) (36, 45). In all cases, mutations were already present in the viral progeny recovered after the first amplification passage and were maintained after six additional passages in the absence of NH4Cl (data not shown).
The amino acid substitutions found in the c1, c2, and c3 capsids were mapped on the crystal structure of FMDV C-S8c1 (36) (Fig. (Fig.5A).5A). Interestingly, one substitution per mutant, namely, VP3 A123T in the c1 capsid, VP3 A118V in the c2 capsid, and VP2 D106G in the c3 capsid, involved amino acid residues located remarkably close to the interfaces between the pentameric subunits, into which the capsid is known to dissociate upon acidification. In addition, VP3 A123T and, to a lesser extent, VP3 A118V are located close to VP3 H140 and H143 (H142 and H145 in FMDV serotype A), which participate in interpentamer pH-dependent electrostatic interactions involved in virion disassembly (13, 20, 63) (Fig. (Fig.5B).5B). In contrast, the accompanying substitutions VP1 N47D in c2 and VP1 H151R in c3 are located in hypervariable protein loops exposed to solvent on the outer surface of the capsid, very far from the interfaces between pentameric subunits. The spatial locations of the amino acid replacements found in the c1, c2, and c3 mutants and the degrees of conservation among FMDVs of the residues involved suggest that VP3 A123T, VP3 A118V, and VP2 D106G are, respectively, responsible for the NH4Cl-resistant, acid-labile phenotype of mutants c1, c2, and c3. In the case of mutant c2, we have subjected this suggestion to experimental confirmation, as described below.
The above-described results show that mutant c2 is the most resistant to NH4Cl treatment and the most acid labile and suggest that this phenotype could be due to amino acid substitution VP3 A118V in the FMDV capsid. To dissect the molecular basis of the NH4Cl resistance and acid lability of mutant c2, plasmids bearing the amino acid substitutions found in this virus were derived from an infectious cDNA clone, plasmid pMT28, containing the complete coding region from FMDV C-S8c1. BHK-21 cells, either treated or not treated with NH4Cl, were infected with viruses derived from the plasmids, and the virus yield was determined by plaque assay (Fig. (Fig.6A).6A). Viruses carrying the two amino acid substitutions VP3 A118V/VP1 N47D or the single mutation VP3 A118V were able to escape from the inhibitory effect of NH4Cl. On the contrary, infections by the parental C-S8c1 virus derived from plasmid pMT28 as well as by the virus bearing replacement VP1 N47D were inhibited by NH4Cl. When the acid sensitivity of these viruses was tested (Fig. (Fig.6B),6B), those harboring the mutations VP3 A118V/VP1 N47D or the single replacement VP3 A118V displayed increased acid sensitivity (pH50 values of 6.94 and 6.92, respectively) compared with that of C-S8c1 and the virus carrying the single mutation VP1 N47D (pH50 values of 6.60 and 6.57, respectively). Resistance to NH4Cl of viruses carrying mutations VP3 A118V/VP1 N47D or only VP3 A118V was also confirmed by counting the number of infected cells (expressing 3A nonstructural protein) (Fig. (Fig.6C).6C). For C-S8c1 and VP1 N47D viruses, the percentage of infected cells was reduced by NH4Cl. Similarly, expression of viral proteins, as analyzed by Western blotting, was inhibited only in cells infected with C-S8c1 and the VP1 N47D mutant (Fig. (Fig.6D).6D). Taken together, these results show that mutation VP3 A118V is responsible for the increased acid sensitivity of the c2 virion and confers the ability to escape from the inhibitory effect of NH4Cl.
Infection with the VP3 A118V/VP1 N47D double mutant or with the VP3 A118V single mutant was not affected by NH4Cl in IBRS-2 cells, while infection with C-S8c1 or the VP1 N47D mutant was significantly inhibited (Fig. (Fig.6E).6E). Likewise, treatment with ConA did not affect infection with the VP3 A118V/VP1 N47D or the VP3 A118V mutant but significantly inhibited infection with C-S8c1 or the VP1 N47D mutant (Fig. (Fig.6F).6F). These results indicate that viruses carrying the mutation VP3 A118V escape from inhibition of endosomal acidification induced in IBRS-2 cells by NH4Cl or ConA.
The above-described results show that the mutant virions c1, c2, and c3 were more susceptible to acid-induced inactivation than the parental C-S8c1 virion and that substitution VP3 A118V was responsible for this phenotype in the case of mutant c2. We next analyzed whether the original c2 mutant and the related VP3 A118V/VP1 N47D and VP3 A118V mutants are less resistant than the parental virus to dissociation into pentamers. Virions of C-S8c1 and the three mutants were radiolabeled and purified through sucrose gradients (Fig. (Fig.7A7A [results not shown for the c2 gradient]). Complete virions, with a sedimentation coefficient of 140S, could be obtained in all four cases, although yields were lower for the three mutants than for C-S8c1. The virus preparations were then dialyzed at neutral pH against a physiological buffer (PBS) and subjected to analytical sucrose gradient centrifugation. Remarkably, this treatment led to the complete dissociation into pentameric subunits of the three mutants, while the majority of the C-S8c1 virions remained intact (Fig. (Fig.7B7B [results not shown for the c2 gradient]). Dissociation into pentamers during dialysis at neutral pH has previously been observed with a few other, unrelated FMDV variants but was not observed with nonmutated FMDV and most variants analyzed (V. Rincón and M. G. Mateu, unpublished observations). The dialysis-induced dissociation is independent from the increased sensitivity of some of these variants to inactivation by lowering the pH. We attribute the dissociation of some mutant viral particles during dialysis to a reduced mechanical resistance against osmotic shock during sucrose removal.
This increased propensity of c2 and related mutants to dissociation into pentameric subunits is due, like their propensity to acid-induced inactivation, to the single substitution VP3 A118V, because this is the only genetic difference between one of these mutants and the C-S8c1 virus.
Evolution may have selected viral capsids that exhibit the right balance between stability (to ensure the extracellular protection of the viral genetic material) and lability (to enable the release of the genetic material within host cells). Different cellular mechanisms are exploited by viruses gaining entry through endocytosis to ensure specific nucleic acid delivery within the host cell. In the polyomavirus simian virus 40 (SV40), whose capsid is extensively cross-linked by interpentameric disulfide bonds, caveola-internalized virions are targeted to the endoplasmic reticulum and use disulfide isomerization reactions to initiate uncoating (57). Poliovirus ensures uncoating and genome release by a mechanism based mainly upon conformational rearrangements triggered by the interaction with its cell receptor (25, 27, 54). The FMDV uncoating mechanism takes advantage of endosomal acidification that dissociates the capsid into pentameric subunits, allowing the release of the genome within the host cell (11). Thus, FMDV is a model virus for the study of the acid-dependent disassembly of nonenveloped viruses (20).
In this study, we have shown that mutants with increased growth capacity in the presence of NH4Cl can easily be isolated from FMDV populations. NH4Cl acts as a proton sink within the internal space of acidic organelles, causing an elevation of endosomal pH. The three mutants analyzed—which use integrins as cell receptors, as previously described for the parental virus C-S8c1 (50)—showed different degrees of resistance to NH4Cl, which correlated with an increased sensitivity of the infectivity of their virions to acid treatment. These results suggest that mutant virions disassemble at pH values more alkaline than that of the parental C-S8c1 virus, thus allowing their escape from the inhibitory effect of NH4Cl. To our knowledge, these results provide the first evidence of FMDV mutants with enhanced acid sensitivity. To date, only FMDV mutants with increased acid resistance had been described (62).
Alkalinization of the pH threshold that initiates genome release from endosomes has also been described for influenza variants that escape inhibition of endosomal acidification, but the mechanisms in influenza virus and FMDV are entirely different. Mutant influenza viruses resistant to the lysosomotropic agent amantadine chloride trigger conformational changes in hemagglutinin protein at pH values higher that those of the parental virus, thus fusing with the endosomal membrane (14). Another variant of influenza virus that fuses at 0.2 pH units higher than the parental strain was reported to be less sensitive to the effects of NH4Cl (18).
Genetic and biochemical analyses have shown that the amino acid substitution VP1 N47D alone does not confer resistance to NH4Cl treatment or acid lability. Conversely, viruses with the single amino acid replacement VP3 A118V were completely resistant to NH4Cl, producing virus yields similar to those found for the VP3 A118V/VP1 N47D double mutant and the c2 virus. This result demonstrates that the single substitution A118V in VP3 is responsible for the complete resistance to NH4Cl and acid lability of mutant c2. Interestingly, the sizes of the plaques produced by the VP3 A118V virus were smaller (about 42% smaller) than those produced by the double mutant (data not shown), and virus titers were also lower (Fig. (Fig.6A),6A), suggesting that selection of substitution VP1 N47D could partially compensate for the effect of replacement VP3 A118V on viral growth.
Amino acid residues A118 of VP3 in c2, which showed the highest degree of resistance to NH4Cl, and A123 of VP3 in c1 are conserved in FMDV and located close to the capsid interpentamer interface. In addition, A123 is close to H140 and H143 of VP3, two residues that are likely involved in the acid-induced dissociation of the FMDV capsid into pentamers. The small side chains of A118 and A123 are completely buried, and both substitutions (A118V and A123T) introduce bulkier side chains whose accommodation inside VP3 probably requires some local structural deformation of the protein. In turn, this structural deformation could weaken some attractive pentamer-pentamer interactions, introduce a few steric clashes between pentamers, and/or lead to a closer approach and (on acidification) a stronger electrostatic repulsion between histidines belonging to neighboring pentamers. A debilitation of the net binding force between pentamers by substitution A118V is experimentally supported by the increased propensity for dissociation into pentamers of mutants carrying this substitution.
The amino acid residue D106 of VP2 in c3, which also showed resistance to NH4Cl, is also conserved in FMDV, and it is located not far from the capsid interpentamer interfaces. D106 is involved in a salt bridge with H157 of the same VP2 subunit. The substitution D106G, by completely removing the side chain, disrupts this interaction and could allow the main chain of VP2 to locally adopt a different conformation that could facilitate the filling of the cavity left by the removal of the aspartate side chain. Thus, by locally deforming VP2, the substitution D106G in c3 could indirectly lead to a decrease in the net binding force between pentamers, again facilitating dissociation.
Acid sensitivity analyses of mutants carrying substitution A118V (the c2, VP3 A118V/VP1 N47D, and VP3 A118V mutants) yielded pH50 values of 6.92 to 6.95. In the case of the C-S8c1 virus, the pH50 values were 6.58 to 6.60, very similar to those reported for type A FMDV (13). This indicates that an increase of only about 0.3 units of pH50 resulted in complete resistance to NH4Cl. Mutants displaying intermediate pH50 values (c1 and c3) escaped from the inhibitory effect of NH4Cl to a lesser extent. FMDV uncoating is supposed to occur within early and recycling endosomes (8, 52), and these are mildly acidic compartments (pH of about 6.5) (65). The pH50 values determined for mutants c2, c3, and c1 are consistent with FMDV uncoating occurring inside these compartments. Indeed, the high level of resistance to NH4Cl exhibited by mutant c2, which displays pH50 values over 6.9, suggests that NH4Cl raises the internal pH of early and recycling endosomes to values close to 6.9.
Mutant c2 and its derivatives carrying A118V in VP3 also displayed resistance to inhibition of endosomal acidification by ConA. This indicates that the mechanism of escape from NH4Cl also generated resistance to inhibition of endosomal acidification by other agents.
Contrary to what was previously observed for thermostable mutants, which are likely to be absent in FMDV populations (43), about 10% of the infectious C-S8c1 viruses were shown to be resistant to the NH4Cl concentration used in these experiments. This mutant frequency is also considerably higher than the values (about 10−5) reported for C-S8c1-derived MAb-resistant mutants (41). These observations suggest that acquisition of NH4Cl resistance does not associate with a severe loss of viral fitness, at least under the stable conditions of FMDV growth in cultured cells. Indeed, no major differences were found when c2 and C-S8c1 viruses were compared in a time course analysis of virus production (data not shown). Experiments are in progress to assess the frequency of NH4Cl-resistant mutants in FMDV populations recovered directly from infected hosts, as well as to address the effect of replacement A118V in VP3 on the infectivity of the virus in animal models.
We thank M. Sáiz for the CHO cell line, J. C. Sáiz for the BEV isolate, M. Herrera and E. Domingo for FMDV MARLS, C. Escarmís for infectious clone pMT28, E. Brocchi for MAb 5C4, and D. Andreu for the synthetic peptides used for cell binding assays.
V.R. is the recipient of an FPI fellowship from the Ministerio de Ciencia e Innovación. This work was supported by Spanish grants BIO2008-0447-C03-01 and CSD2006-0007 to F.S. and BIO2006-00793 and BIO2009-10092 to M.G.M. and by Fundación Severo Ochoa.
Published ahead of print on 6 January 2010.
¶Dedicated to the memory of Rosario Armas-Portela.