INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

An increase in the mutation rate during replication of RNA viruses can result in a decrease of viral infectivity and occasional virus extinction (11, 34, 39, 40, 60). Studies with the important animal pathogen foot-and-mouth disease virus (FMDV)a member of the Picornaviridae family (53, 63)have shown that a small replicative population size and low viral fitness favored virus extinction (60). This was documented with single and multiple passages of FMDVs differing up to 10⁶-fold in relative fitness in the absence or presence of the mutagenic base analogs 5-fluorourocil (FU) or 5-azacytidine, individually or in combination (59, 60). Mutagenic treatments resulted in occasional, not systematic, viral extinction, while parallel passages in the absence of mutagens never led to loss of infectivity, no matter how low the initial viral population size and fitness were (59, 60). These results suggested the possibility that a combination of an antiviral inhibitor, to reduce the replicative load of virus, and a mutagenic agent could be more effective in producing viral extinction than a mutagenic agent used in isolation. To test this possibility we have studied the effect of the mutagenic base analog FU and the antiviral inhibitors guanidine hydrochloride (G) and heparin (H) on the infectivity of FMDV clones and populations depicting a wide range of relative fitness values. FU has been shown to be mutagenic for a number of RNA viruses (6, 20, 31, 34, 51, 71), including FMDV (59, 60). G at millimolar concentrations blocks the replication of picornaviruses (5, 7, 15, 49, 52, 55), arboviruses (27), and several plant viruses (13, 67). In poliovirus, the target of G is the ATPase activity of nonstructural protein 2C (49), a protein involved in viral replication and encapsidation. In FMDV, amino acid substitutions at 2C have also been associated with resistance to G (56). Heparins are sulfated polysaccharides (9) which bind FMDV when the virus has been passaged in cell culture and has adapted to using heparan sulfate (HS) as a receptor (2, 35, 54). Adaptation to use of HS as a receptor has been associated with substitutions which lead to positively charged amino acids at exposed positions of the capsid (2, 54).

Here we report that high-frequency extinctions of FMDV of low and intermediate fitness values can be achieved with a combination of FU and G but not with either drug alone. Extinction of high-fitness FMDV populations required a triple combination of G and H together with FU. Mutation frequencies in the mutant spectrum of three genomic regions of a preextinction population obtained by the combined action of an inhibitor and a mutagen were compared to values in genomes passaged in standard conditions and also to values previously determined for FMDV populations subjected to one or multiple passages in the presence of FU (60). We found that mutation frequencies increased up to fourfold. There was also a statistically significant increase in the number of mutations in preextinction populations with respect to control populations, and the frequency of mutation types that can be attributed to the mutagenic action of FU was significantly higher than the frequency of other types of mutations.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells and viruses. The origins of baby hamster kidney 21 cells (BHK-21) and Chinese hamster ovary cells (CHO) have been previously described (3, 18, 25, 62). Both cell types were grown in Dulbecco's modified Eagle's medium (DMEM) (Gibco) supplemented with nonessential amino acids (Gibco), 50 mg of gentamicin (Sigma)/ml, and 5% fetal calf serum (Gibco).

Mutagenic and antiviral treatments. To prepare culture medium containing FU (Sigma), the appropriate amount of analog was dissolved in DMEM to yield a 2.5-mg/ml solution, which was diluted in DMEM as needed for the experiments. To prepare medium containing G (Sigma), the appropriate amount was dissolved in DMEM to yield a 4 mM solution (0.38 mg/ml). To prepare medium containing a mixture of both FU and G, the appropriate amount of G was added to DMEM containing FU (200 µg/ml) to yield FUG, a solution containing FU (200 µg/ml) and G (4 mM). H (Heparin sodium salt; Sigma) was dissolved in water and diluted in DMEM or FUG (to yield FUGH, a solution containing FU [200 µg/ml], G [4 mM], and H [1 mg/ml]) as needed for the experiments. All solutions were sterilized by filtration. Media were supplemented with 2% fetal calf serum and stored at 4°C for a maximum of 14 days. The effect of FU, inhibitors, or their combination on cell viability was monitored by trypan blue exclusion as described previously (60). Before counting, cells were washed with DMEM, detached by trypsin treatment, and resuspended in DMEM supplemented with 2% fetal calf serum. Viable BHK-21 cells comprised at least 20% of the total (Fig. 1), under the doses and times of exposure used for the experiments. CHO cells were more resistant than BHK-21 cells to the treatments, and after exposure to FUG, approximately 80% of the cells were viable (N. Pariente et al., unpublished data).

View larger version (22K): [in this window] [in a new window]   FIG. 1.   BHK-21 cell viability in the presence of FU, G, and H, alone or in combination. The origin of BHK-21 cells and FMDV C-S8c1, as well as procedures for cell growth, antiviral treatment, quantification of cell viability, and infections with FMDV, are detailed in Materials and Methods. Percent cell viabilities (trypan blue exclusion) were calculated relative to those in parallel, untreated cell cultures, counting 300 to 700 cells per sample. Hours of exposure refers to the time elapsed between the addition of the mutagen or antiviral agent to cells and the determination of cell viability. Standard deviations (not included in the plots) never exceeded 20%.

Viral infections in the presence of FU, inhibitors, or their combination. Different standard single infections were carried out in parallel. (i) In control infections (absence of mutagen and inhibitors), confluent cell monolayers were infected with FMDV (adsorption for 45 min at 37°C), washed for 1 min with 0.1 M phosphate buffer (pH 6.0) (to inactivate unadsorbed virions), and washed again extensively with DMEM. The infection was allowed to proceed in DMEM supplemented with 2% fetal calf serum. (ii) For infections in the presence of FU, confluent cell monolayers were pretreated for 13 h with 200 µg of FU/ml and then washed with DMEM; the procedure was done as for the control infections, and the infection was allowed to proceed in the presence of the same concentration of FU. (iii) For infections in the presence of G, no pretreatment of the cell monolayer was performed; after addition of FMDV and washing of the cell monolayers, the infection was allowed to proceed in the presence of 4 mM G. (iv) For infections in FUG, confluent cell monolayers were pretreated for 13 h with 200 µg of FU/ml and then washed with DMEM. After virus adsorption and washing of cell monolayers, the infection proceeded in the presence of FUG. (v) For infections in H, no pretreatment of the cell monolayer was performed; FMDVs were preincubated with 1 mg of H/ml for 30 min at room temperature. After inoculation and washing of the cell monolayers, the infection was allowed to proceed in the presence of 1 mg of H/ml. (vi) For infections in the presence of FUGH, confluent cell monolayers were pretreated for 13 h with 200 µg of FU/ml and then washed with DMEM. Viruses were preincubated with 1 mg of H/ml for 30 min at room temperature. After virus adsorption and washing of the cell monolayers, the infection proceeded in the presence of FUGH. All infections were allowed to continue for approximately 24 h.

cDNA synthesis, PCR amplification, and nucleotide sequencing. Viral RNA was extracted as previously described (60) using 150 µl of culture medium; for preextinction populations (virus in the passage prior to extinction), the volume used was 1 ml due to the low viral load present in these samples. cDNA synthesis and PCR amplification (reverse transcription-PCR [RT-PCR]) were performed as previously described (23). In all cases cDNA synthesis was carried out using avian myeloblastosis virus reverse transcriptase (Promega). Amplification was performed with Taq polymerase (Perkin-Elmer) for determination of the consensus sequences and for the extinction test. Amplifications intended for molecular cloning and nucleotide sequencing of individual clones were carried out with the Expand High Fidelity PCR System (Roche) (4). To ensure an excess of template molecules for amplification, only the samples for which RT-PCR amplification of 1/10 of the initial RNA template was also positive were used for molecular cloning of the cDNA. Three FMDV genomic regions were subjected to RT-PCR amplification: residues 3193 to 3869 (spanning the VP1-coding region), residues 4280 to 5349 (spanning the 2C-coding region), and residues 6609 to 8035 (spanning the entire 3D [polymerase]-coding region). Numbering of FMDV genomic nucleotides is that used in reference 65. The oligonucleotide primers used to amplify the VP1- and 3D-coding regions have been previously described (60). The following primers were used for 2C amplification: 5'-TCGGAGCTCCGATTCTGTTGGCCGGGTTG-3', termed 2BR3SacI (sense, 5' position 4253), and 5'-AAAGAATTCAATTGCTGCCTCGTGTTG-3', termed 3AD4EcoRI (antisense, 5' position 5376); they include restriction enzyme sites for SacI and EcoRI, respectively (underlined). Amplified cDNAs coding for VP1, 2C, and 3D were digested with the appropriate restriction enzymes, ligated to plasmid pGEM-4Z (Promega) previously digested with the same restriction enzymes, and cloned into Escherichia coli DH5 as described previously (60). DNA from bacterial colonies was obtained by direct PCR amplification from colony lysates using the commercial primers SP6 and T7 (Promega), which are complementary to positions flanking the polylinker site of pGEM-4Z. Nucleotide sequences were determined using the Big Dye Terminator Cycle Sequencing kit (ABI Prism; Perkin-Elmer) and the automated sequencer ABI373. Each mutation was confirmed by two independent sequencing assays using primers of different orientation. Sequences were analyzed with the DNA Star 4.0, Genedoc, and GCG package programs.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Systematic extinction of C-S8c1 by a combination of FU and G. G at a concentration of 4 mM in the culture medium causes a 10- to 10²-fold reduction in the production of FMDV C-S8c1 with no significant effect on BHK-21 cell viability. The combination of 200 µg of FU/ml and 4 mM G (FUG) resulted in a decrease of cell viability to about 40% of that of the untreated cells (Fig. 1). To test whether FUG was more effective than either FU or G alone in inhibiting viral production, FMDV MARLS, C-S8c1, and five low-fitness subclones derived from H were used to infect BHK-21 cells in the absence or presence of FU, G, or FUG. The FMDVs used differed up to 10⁶-fold in relative fitness value (see Materials and Methods). The results (Fig. 2) show that FUG led to reductions in viral production that were 10²- to 10³-fold larger than those observed in the presence of FU or G alone. Yields of the subclones H3, H4 and H5 in the presence of FU were undetectable, but virus reemerged after one passage. However, extinction of the subclones H1 to H5 was observed in all cases in one passage in the presence of FUG.

View larger version (35K): [in this window] [in a new window]   FIG. 2.   Effect of viral fitness on the decrease of infectivity in a single round of infection in the absence or presence of FU, G, or the mixture of both (FUG). FMDV MARLS, C-S8c1, and H clones are described in Materials and Methods. Empty columns indicate viral production in the absence of mutagen or antiviral agent; gray columns indicate viral production in the presence of 200 µg of FU/ml; striped gray columns indicate viral production in the presence of 4 mM G; and black columns show viral production in the presence of FUG. In this experiment 9 × 105 BHK-21 cells were infected with 5 × 104 PFU of virus. Extinction of the five H FUGp1 clones was confirmed by three additional blind passages, in the absence of mutagen and inhibitor with no evidence of infectivity, and no RT-PCR-amplifiable material in the supernatant of the third passage. Yields of H 3, H 4, and H 5 clones in the presence of FU were undetectable (<10 PFU/ml), but virus reemerged after one blind passage in the absence of the mutagen. However, these clones were extinguished upon a second passage in the presence of FU. Titrations were done in triplicate, and standard deviations are indicated. Procedures for chemical mutagenesis, antiviral treatment, and determination of viral infectivity are described in Materials and Methods.

View larger version (22K): [in this window] [in a new window]   FIG. 3.   Serial passages of C-S8c1 in the absence or presence of FU, G, or FUG. In this experiment, 9 × 105 BHK-21 cells were infected with 5 × 104 PFU of FMDV C-S8c1, and subsequent infections were carried out with 0.1 ml of supernatant of the previous passage (the multiplicity of infection at each passage can be calculated from the titers shown in ordinate). The origin of FMDV C-S8c1, conditions for mutagenic and antiviral treatment, and those for determination of FMDV infectivity are described in Materials and Methods. The preextinction population is indicated by an arrow. Viral extinction was confirmed by three additional blind passages in the absence of mutagen and inhibitor with no evidence of infectivity, and no RT-PCR-amplifiable material was found in the supernatant of the third passage. All titrations were done in triplicate. Standard deviations (not included in the plots) never exceeded 15%.

Efficient extinction of MARLS in CHO cells. MARLS displays a 130-fold-greater fitness than C-S8c1 in BHK-21 cells (59, 60), as a result of 213 serial passages in BHK-21 cells (3). However, its fitness in CHO cells is at a much lower level, as judged from a 10-fold-lower progeny production, than in BHK-21 cells. Therefore, if difficulties for extinction of MARLS in infections in the presence of FUG were due to its high fitness (Fig. 2), MARLS should manifest an increased susceptibility to FU and FUG in infections of CHO cells. The results (Fig. 4) show that extinction of MARLS occurred after three passages in CHO cells with FU and after two passages with FUG, suggesting that resistance to extinction of MARLS in BHK-21 cells was due to its high fitness in the environment provided by BHK-21 cells and not to an intrinsic property of the MARLS genome to be refractory to extinction.

View larger version (22K): [in this window] [in a new window]   FIG. 4.   Infectivity values upon passage of FMDV MARLS in the absence or presence of FU, G, or FUG in BHK-21 and CHO cells. Conditions for mutagenic and antiviral treatments and for determination of FMDV infectivity are detailed in Materials and Methods. Left panel: FMDV MARLS was tested by infecting 9 × 105 BHK-21 cells with 5 × 104 PFU of virus, and the next passages were carried out by infecting cells with 0.1 ml of supernatant from the previous passage. Right panel: 9 × 105 CHO cells were infected with 9 × 105 PFU of FMDV MARLS, and passages were carried out as for the infections of BHK-21 cells. The multiplicity of infection at each passage can be calculated from the titers shown in ordinate. Preextinction populations are indicated by arrows. Viral extinction was ascertained by three additional blind passages in the absence of mutagen and inhibitors, with no evidence of infectivity and RT-PCR-amplifiable material in the supernatant of the third passage. All titrations were done in triplicate. Standard deviations (not included in the plots) never exceeded 15%.

High-fitness virus can be extinguished with the combination of a mutagen and two inhibitors. The effect of FUG on FMDV C-S8c1 suggested that extinction of high-fitness MARLS in BHK-21 cells could be favored by decreases in the viral load achieved with the combination of two antiviral inhibitors targeted to different steps in the virus life cycle. To test this possibility, the the effect of combination of FU, G, and H (FUGH) on production of MARLS in BHK-21 cells was tested. The FUGH treatment resulted in a decrease of cell viability to about 20% of the untreated cells (Fig. 1). Control experiments (described in Materials and Methods) ruled out cytotoxicity as being responsible for the absence of viral production. The results (Fig. 4) show that a combination of two inhibitors and a mutagen can lead to extinction of MARLS in BHK-21 cells in three passages. In order to elucidate how systematic the extinctions were, 47 repetitions of the passages in the presence of FUGH were performed. In 40 cases extinction of MARLS was obtained in the third passage and in 7 cases in the fourth passage. Therefore, a combination of a mutagen and multiple inhibitors is effective in promoting systematic extinction of high-fitness FMDV.

Evidence for FU-induced mutations in a preextinction population and moderate increases in mutation frequency. Consensus nucleotide sequences and sequences from molecular clones were obtained from the mixture of the supernatants of the 46 replicates of the C-S8c1FUGp1 preextinction population and C-S8c1DMEMp1 (Fig. 3). Three genomic regions were analyzed, the ones encoding the capsid protein VP1 and the nonstructural proteins 2C and 3D. Increases in mutation frequencies ranged between 1.5- and 4-fold; the maximum difference was found in 3D (Table 1) as in a previous report (60). The number of mutations observed in the C-S8c1FUGp1 population was significantly higher than that obtained with C-S8c1DMEMp1, as measured with the ² test (P < 0.005 and P > 0.001; ², 1 degree of freedom). The most abundant mutations found in the preextinction populations were the transitions AG and UC (Table 2), which have been associated with the mutagenic action of FU (37, 59, 72). The number of AG and UC transitions was found to be significantly higher than other mutation types with the ² test (P < 0.025 and P > 0.01; ², 1 degree of freedom). Increases were found in Shannon entropy, which is a measure of the different types of nucleotide sequences found in the mutant spectrum of the genomic region under study. Thus, in the three genomic regions analyzed, the mutant spectrum of the population near extinction was more complex than that in the corresponding control population (Table 1).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The possibility of a new antiviral strategy based on increasing the mutation rate above a threshold value (16, 34, 36, 40, 41) stems from the quasispecies dynamics of RNA virus populations (21, 57, 64). Error rates during RNA genome replication dictate that viral genomes consist of a complex mutant spectrum often dominated by a most fit genome, termed the master sequence (21, 22). Error-prone replication predicts an error threshold relationship of the form

(1)

in which v_max is the maximum length (information contents) that can be stably maintained in a viral genome, ₀ is the selective advantage (superiority) of the master sequence over the mutant spectrum, and is the average copying fidelity (in a scale from 0 to 1) during genome replication; therefore (1  ) is the average error rate (21, 22, 57, 64). According to the error threshold relationship, an increase in the error rate above a threshold value should result in the loss of information carried by the viral genome and therefore in a loss of infectivity. Experimental support for this theoretical prediction has been obtained by subjecting vesicular stomatitis virus, poliovirus (10, 11, 34, 39), and human immunodeficiency virus type 1 (40) to increased levels of mutagenesis. In the case of FMDV we showed previously that low fitness and reduced viral loads favored viral extinction upon replication in the presence of FU or 5-azacytidine or their combination (59, 60). Relative fitness among viral populations is conceptually parallel to relative fitness among components of the mutant spectrum within a viral population. The parameter ₀ in equation 1 represents the superiority of the master sequence over its mutant spectrum, and its value is larger the higher the relative fitness of the mutant distribution (21, 22, 57). Therefore, virus extinction should be favored in populations with low relative fitness, as observed experimentally (59, 60).

Further evidence that low fitness rather than a specific genomic structure was responsible for virus extinction has been obtained with infection of CHO cells with MARLS. This virus showed low fitness in CHO, as evidenced both by a 10-fold-lower progeny production than in BHK-21 cells and a 10-fold increase in progeny production after eight serial passages in CHO cells (N. Pariente et al., unpublished results). MARLS was prone to extinction with a combination of mutagen and one inhibitor and with the mutagen alone in CHO cells but not in BHK-21 cells (compare the graphics in Fig. 4). The viral load also has a definite influence in the adaptation (fitness increase) of viral quasispecies. When the viral population size is small, genetic drift and a decrease in fitness prevail (8, 23, 73, 74), while when the replicative size of the same virus clones and populations is large, competitive selection of increasingly adapted mutant swarms leads to exponential increases in viral fitness (24, 46, 47, 69).

These previous theoretical and experimental observations suggested that a combination of a mutagenic agent and antiviral inhibitors should increase the frequency of viral extinctions. This has indeed been documented in the present report by using FMDV clones and populations differing by about 10⁶-fold in relative fitness values. For the reference virus C-S8c1, whose relative fitness value is taken as 1 (23, 24), a combination of FU and G resulted in 47 extinction events out of 47 trials. For MARLS, which has a 130-fold-greater fitness than C-S8c1 in BHK-21 cells, two inhibitors were needed in combination with FU, to obtain viral extinction. This extinction was also obtained for each of 48 trials, in fewer than five passages. We term this occurrence systematic extinction, since chances of virus surviving for a larger number of passages under these adverse evolutionary conditions must be negligible. Not unexpectedly, clones whose fitness was 10³- to 10⁴-fold-lower than that of C-S8c1 did not survive even a single round of infection under the same conditions.

The mutant spectrum of an FMDV C-S8c1 preextinction population, subjected to passages in the presence of FUG, has been compared with the mutant spectrum of a C-S8c1 population passaged in DMEM. With the RT-PCR amplification system used, the mutation frequency values found in the mutant spectra analyzed cannot be affected by mutations introduced during the amplification procedures (1, 4, 59, 60). There was a significant bias toward the type of mutations induced by FU expected from the ambiguous reading of fluorouridine in template molecules (37, 59, 72). In the VP1-, 2C-and 3D-coding regions, 50, 67, and 42%, respectively, of the mutations were nonsynonymous (Table 1). The locations of amino acid replacements in VP1, 2C, and 3D are given in Table 3. Two amino acid replacements were found in the carboxy-terminal region of VP1 (antigenic site C [43]) (Q191H and H197R). In 2C, replacement R109H affected the A domain, a Walker nucleoside triphosphate binding motif (49). In 3D, except for the T365A change in  strand 4, which is a very conserved position among picornavirus polymerases (75), all other replacements were located on loops linking  helices and  strands, assuming that structural elements of poliovirus 3D (30) coincide with those of FMDV 3D upon alignment of the two amino acid sequences (42).

One of the possible drawbacks of the combination therapy presented herein would be the appearance of guanidine-resistant FMDV mutants due to the action of FU. There have been several guanidine-resistant mutants of poliovirus described, and some of FMDV serotype O, and most of them involve transition mutations presumably favored by FU (50, 56). Nevertheless, none of the clones analyzed from a preextinction FMDV population showed mutations described as responsible for a guanidine-resistant phenotype, despite the fact that FMDV replication in the presence of G alone can result in selection of resistant mutants (55, N. Pariente et al., unpublished results). In case resistant mutants arose during treatments with mutagen and inhibitors, there would be a competition between the selection of the inhibitor-resistant mutant and the mutational force towards extinction. When such a force dominates, chances of selecting resistant mutants diminish, and this may explain why no G-resistant mutants were detected in the mutant spectra analyzed. The option of using two inhibitors that act on different steps of the viral life cycle reduces the possibility of the mutagen generating escape mutants. The FMDV-FUG or -FUGH combination system provides a simple cell culture model system for studying the dynamics of mutational pressure in relation to G-resistant or H-resistant mutant generation and efficiency of extinction. These studies are currently in progress.

The need of combination therapy with two or more inhibitors to limit the chances of generating and selecting inhibitor-resistant mutants in viral quasispecies has been emphasized (17, 19, 29, 33). Early experiments suggested a more effective control of influenza virus by a combined vaccine-antiviral strategy (32, 45, 70). In a conceptually parallel study, ribavirin (1--D-ribofuranosyl-1H-1,2,4-triazole-3-carboxamide) treatment prevented the appearance of neutralization-resistant mutants of lymphocytic choriomeningitis virus in mice (58). Recent reports (10, 11) have documented that ribavirin can be mutagenic for poliovirus and that in this system reduction of infectious virus production correlated with mutagenic activity of the nucleotide analog. These results were interpreted as suggesting that the antipoliovirus activity of ribavirin could be due to the transition of viral replication into error catastrophe (10, 11), although viral extinction was not observed. Ribavirin is used in the therapy of a number of viral infections (12, 14, 48, 61). Certainly the contribution of inhibition versus mutagenesis in the antiviral action of ribavirin deserves further investigation. Also, the possible contribution of inhibition of virus-specific RNA synthesis in the extinction by FUsuggested by the reduction in viral production in a single round of infection in the presence of FU (59, 60)requires further investigation. Our observations on the benefits of a combination of a mutagenic agent and one or several antiviral inhibitors are consistent with current views on adaptability of viral quasispecies (8, 16, 21-24, 26, 46, 47, 69). The lesser the opportunity of a virus to replicate, the lower the chances of survival in the face of an antiviral response, be it natural (an immune response) or induced externally (through immune stimulation, immunotherapy, or drug administrations).