This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Tully, D. C.
Right arrow Articles by Fares, M. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Tully, D. C.
Right arrow Articles by Fares, M. A.

 Previous Article  |  Next Article 

Journal of Virology, January 2009, p. 781-790, Vol. 83, No. 2
0022-538X/09/$08.00+0     doi:10.1128/JVI.01500-08
Copyright © 2009, American Society for Microbiology. All Rights Reserved.

Shifts in the Selection-Drift Balance Drive the Evolution and Epidemiology of Foot-and-Mouth Disease Virus {triangledown}

Damien C. Tully and Mario A. Fares*

Evolutionary Genetics and Bioinformatics Laboratory, Smurfit Institute of Genetics, Trinity College Dublin, Dublin 2, Ireland

Received 17 July 2008/ Accepted 31 October 2008


arrow
ABSTRACT
 
Foot-and-mouth disease virus (FMDV) is the causative agent of an acute vesicular disease affecting wild and domesticated animals. Despite the economic burden of the disease and all efforts to eradicate it, FMD outbreaks continue to emerge unexpectedly in developed and developing countries. Correlation of the mutational dynamics of the virus with its epidemiology remains unexplored. Analysis of 103 complete genomes representing the seven serotypes shows the important role that selection plays in the genomic evolution of viral isolates for serotypes. We identified selection and relaxed constraints due to genetic drift through analyses of synonymous sites. Finally, we investigated interactions between mutations that showed coevolving patterns and analyzed, based on protein structures, slightly deleterious and compensatory mutational dynamics. Specifically, we demonstrate that structurally exposed capsid proteins present a greater number of adaptive mutations and relaxed selection than nonstructural proteins. Such events have been magnified during the evolution of the southern African virus types (SATs). These shifts in selection-drift balance have generated the great antigenic and genetic diversity observed for SAT serotypes and that are responsible for epizootics on the continent of Africa. The high number of slightly deleterious and compensatory mutations in SAT serotypes in structural proteins is testament to such balance plasticity. The significant accumulation of these coevolving mutations in African serotypes supports their contribution in generating adaptive immune-escaping mutants and in establishing persistent infections. The reverse of this pattern in nonstructural proteins reveals the neutral fixation of mutations in the more widely spread and commonly studied Euro-Asiatic serotypes.


arrow
INTRODUCTION
 
Foot-and-mouth disease virus (FMDV) is the causative agent of an acute, systemic disease that affects farm animals and therefore has economically devastating effects (26). FMDV belongs to the Aphthovirus genus in the family Picornaviridae. Its positive-sense, single-stranded RNA genome is 8.5 kb in length and is immediately translated into a polyprotein upon virus entry into a host cell. This polyprotein is posttranslationally cleaved to yield 12 mature proteins (37, 42). The 12 proteins comprise, from 5' to 3', a set of four structural capsid proteins (1A, 1B, 1C, and 1D, also known as VP4, VP2, VP3, and VP1, respectively), and eight nonstructural proteins (Lpro, 2A, 2B, 2C, 3A, 3B, 3Cpro, and 3Dpol) (38). The structural proteins, 1A to 1D, assemble to form the capsid of the virion and, with the exception of 1A, are the first line of contact between the host cell and the virus. The interaction between both biological systems occurs specifically between the capsid protein VP1 and the RGD-dependent integrins and heparan sulfate proteoglycan receptors (23). Nonstructural proteins are involved in replicatory and other functions (18, 22, 30). The 2B protein is known to be involved in membrane rearrangements required for viral RNA replication and capsid assembly (18). Both nonstructural proteins 3A and 3B appear to play a role in virulence and host range (34), with a single amino acid substitution mediating adaptation of the virus to a guinea pig (33). The 3C protease is responsible for most of the proteolytic cleavage in the viral polyprotein and RNA replication (51), while the 3D protein is the core subunit of the picornavirus RNA-dependent RNA polymerase (17). Functions for the remaining nonstructural proteins are less well understood.

The high mutation rates (2, 10) and population bottlenecks experienced during the transmission of RNA viruses such as FMDV (reviewed in reference 13) account for their high genetic and antigenic variability. This genetic variability is translated into the existence of seven immunologically diverse serotypes distributed around the world (southern African type [SAT] 1 to 3 and Asia 1, A, O, and C) with varied degrees of selective genetic variability (19, 24, 50). Uncovering the selective patterns of this fixed variability is instrumental to the understanding of the epidemiological behavior of these viruses and in monitoring possible outbreak sources. To date, most of the evolutionary studies performed in FMDV have focused on the structural surface-exposed capsid protein VP1 due to its essential role in cell-receptor recognition and escape from the immune response (45). Some of these studies have been aimed at describing the evolutionary parameters responsible for the emergence of specific serotypes (12). Others have yielded information on the evolutionary constraints acting on FMDV strains subjected to different experimental passage regimens (16) or naturally isolated (19, 50). Despite the clear role of adaptive evolution in generating variability in the VP1 capsid protein, many other factors may have shaped its evolution, including changes in the balance between selection and drift (50).

Most of the analyses performed in FMDV have been limited in different ways. Indeed, data have been biased by the phylogenetic nature of the analyses conducted (reviewed in reference 24). The number of studies utilizing full-length FMDV genomes for comparative evolutionary analyses has been even less significant and mainly concentrated on intra- or intertypic comparisons of a limited number of isolates (27, 36). These studies have been also biased by the fact that only a few serotypes have been considered (notably O, A, and C) due to the lack of resolution of full-length genomes for some of the serotypes, especially for the sub-Saharan serotypes SAT1 and SAT3. The analysis of 103 complete genomes isolated from the seven different serotypes (4) is one of the most complete evolutionary analyses so far conducted in the comparative genomics of FMDV. However, the main objective of that earlier study was to identify specific genomic regions under strong purifying selection. In their analysis of the complete full genomes, including new SAT3 and SAT1 genomes, those authors reported novel viral genomic motifs with possible biological importance due to their strongly constrained evolution. They mainly analyzed these constraints through the ratios between the rates of nonsynonymous replacements and synonymous changes. These strong constraints were operating in approximately 58% of the amino acid sites, with nonstructural proteins presenting on average greater constraints to amino acid replacements than structural proteins (for example, see Fig. 2C of reference 4). In addition, evidence for the action of adaptive Darwinian selection was only detected in some structural proteins, while most of the genome was under strict purifying selection. Despite the insightful work of these and other authors, further analyses of selective constraints operating in FMDV have ignored selection on synonymous sites, which is expected to be highly significant in RNA viruses due to constraints imposed over the secondary RNA structure (5). These constraints may have led to high nonsynonymous-to-synonymous ratios due to low synonymous substitution rates at these synonymous sites. This ratio is generally used as an indicator of selection based on the assumption that synonymous sites evolve neutrally, which is generally not the case for constrained RNA molecules.


Figure 2
View larger version (45K):
[in this window]
[in a new window]

  		FIG. 2. Genomic distribution of nucleotide substitutions per synonymous site (dS). (A) Average dS variation along the FMDV genome. This distribution was highly heterogeneous when we compared structural proteins with nonstructural proteins (separated in the plot by a dashed line). The median of dS for structural proteins (indicated by a blue horizontal bar) is qualitatively higher than that for nonstructural proteins (indicated by an orange horizontal bar). The difference between the mean dS values between structural and nonstructural proteins is significant at the 5% significance level (box in the upper right corner of the plot). (B) Comparison of the distribution of dS values between southern African serotypes (blue bars) and non-SAT serotypes (red bars) along the genome. These values are qualitatively greater in the SATs compared to the other serotypes along the genome. This difference is mainly due to more relaxed constraints on synonymous sites of structural (S) proteins compared to nonstructural proteins (NS) in the SAT serotypes (box in the upper right corner of the plot).

Despite the effort invested in identifying evolutionary forces in FMDV, many questions remain to be answered. For instance, understanding why serotypes differ in their evolutionary and epidemiological dynamics remains a critical research question. Herein, we were able to examine, for the first time, how different genomic regions are under distinct selective constraints, as well as elucidating the role that shifts in the selection-drift balance play in the emergence of variability within FMD viruses. We describe how the effects of genetic drift are compensated for in the different serotypes and genomic regions and finally reveal how the neutral fixation of mutations may provide information about epidemiological patterns of the virus. Our main conclusions are the following: (i) different viral types present evidence of different evolutionary dynamics throughout their phylogeny and genome; (ii) these differences are mostly due to a changing selection-drift balance; (iii) this changing balance has led to the differential accumulation of slightly deleterious mutations and compensatory changes between genomes and serotypes; (iv) lastly, we postulate that these mutational dynamics may explain the diverse epidemiological patterns observed for the different FMDV serotypes.


arrow
MATERIALS AND METHODS
 
We conducted extensive comparisons of the genomes of the seven serotypes to understand the FMDV mutational dynamics and correlated these with the epidemiological characteristics of the virus. We first identified fixation of amino acid substitutions in each of the proteins and serotypes. Then we tested for selection and relaxed selective constraints in these proteins and serotypes by studying the accumulation of mutations at synonymous sites. We supported these analyses by the identification of slightly deleterious mutations (SDMs). Finally, we tested the shift in the balance between selection and drift by studying the phylogenetic distribution of these SDMs and by identifying compensatory mutations (CMs) through the analysis of the distribution of mutations in the structures available for FMDV proteins.

Genome sequences and phylogenetic analyses. All complete genome sequences of FMDV were downloaded from GenBank. A total of 47, 6, 7, 27, 9, 3, and 4 full-length FMDV sequences were available for all seven serotypes, serotypes A, Asia 1, C, O, SAT1, SAT2, and SAT3, respectively, collected worldwide during the period from 1951 to 2002. GenBank accession numbers for all sequences used in the study are available upon request.

We first constructed multiple protein sequence alignments for all genomes using MUSCLE (11), and then we built protein-coding nucleotide sequence alignments based on their corresponding protein sequence alignments. The total length of the alignment was 7,239 nucleotides. Maximum likelihood phylogenetic trees were inferred for each of the seven serotypes using the PAUP* package (47). In each case, the best-fit model of nucleotide substitution was identified by MODELTEST (39) as the general reversible GTR + I + {Gamma}4 model, with the frequency of each substitution type, proportion of invariant sites (I), and the gamma distribution of among-site rate variations with four rate categories ({Gamma}4) estimated from the empirical data (parameter values are available from the authors upon request). Despite our efforts to infer the most accurate phylogenetic relationships between the different FMDV genomes, recombination between specific lineages and RNA structural constrains may produce artifactual relationships. We, however, did not have evidence of such events occurring among serotypes, although intraserotype recombination events may still take place. Following the study of Carrillo and colleagues (4), only a few isolates presented evidence for incongruent phylogenies due to possible recombination events at specific limited genome regions. Our phylogenetic tree, however, is based on the phylogeny built upon the entire genome and, consequently, incongruent artifactual phylogenies caused by local recombination events may have meaningless effects in our phylogeny.

Identifying adaptive evolution in FMDV genomes. There are two types of nucleotide mutations, based on their effects on the protein amino acid composition: synonymous (dS) and nonsynonymous (dN) nucleotide replacements. In general, the dS rate accumulates neutrally because it has no effect on the amino acid composition of proteins and is hence unseen by selection, with the accumulated number of such mutations being proportional to time. In contrast, dN involves an amino acid replacement and is therefore subjected to selection. The intensity of selection ({omega}) can then be measured by the ratio between these two rates ({omega} = dN/dS). {omega} values of <1, 1, or >1 indicate negative selection, neutral evolution, and positive selection, respectively (1, 7, 44). However, it has been shown that the stability of the RNA molecule secondary structure as well as translational selection may impose constraints on synonymous sites, leading to lower dS values and consequently to inflated {omega} estimates (5, 28, 35, 41). To identify selective constraints at single lineages and codon regions, we used the sliding window approach (15) implemented in the program SWAPSC version 1 (14). This program isolates all those genomic regions that show significantly lower dS values than expected under neutrality and then yields unbiased estimates of {omega} (15). A coding region will therefore be detected to have undergone adaptive evolution if it presents a significantly greater number of nonsynonymous substitutions per nonsynonymous site while dS accumulates following the neutral evolution model. Window sizes for the analyses were statistically optimized so as to provide the lowest possible false discovery rate, as detailed elsewhere (15). Briefly, to estimate the optimum window size we built subalignments by randomly sampling from the initial alignments. Then we slid a window of a specific size (for example, sizes ranged between 1 and 20 codons) along the subalignment, and the distributions of dS and dN were calculated. The final optimum window size is one for which the dS and dN distribution tails (for example, the 5% extreme values under the tails of the density distribution curve) are above 5%. We tested for the presence of adaptive evolution at each one of the lineages leading to each one of the seven FMDV serotypes and at each one of the FMDV proteins. We finally conducted all the statistical analyses to compare selection results between serotypes and genomic regions by normalizing the proportion of codon sites under adaptive evolution.

The same data set used here has previously been subjected to recombination analyses, with the authors of the previous study coming to the conclusion that the capsid proteins undergo recombination infrequently while nonstructural proteins may indeed undergo complex recombination events in some serotypes (4). Indeed, other studies have confirmed the observation that recombination has occurred more often among nonstructural than structural genes (23, 52). Hence, we did not test for recombination, given that any observed patterns have been previously described using standard methodologies. However, all the regions detected here as being under adaptive evolution were those showing no phylogenetic incongruence and were hence considered to be true positives.

Identifying constraints at synonymous sites. To determine the degree of relaxed selective constraints in each one of the proteins for each FMDV serotype, we estimated dS by the method of Li (25) as implemented in SWAPSC. Due to the limited evolutionary signal contained in one codon site, we slid a window of 20 codons along the genome and estimated dS for each sliding step. We then tested whether dS was homogeneously distributed along the genome, as would be expected under a neutral evolution model for synonymous sites. In particular, we were interested in knowing if dS accumulates differently in structural proteins than in nonstructural proteins and whether this effect is serotype specific. Such differences could indicate differences in the rates of changes in the drift and selection forces between structural and nonstructural proteins as well as between serotypes. If these differences correlated with epidemiological or infectivity differences, then we could speculate about the role that drift has in driving the evolution and epidemiology of particular serotypes. In our analyses, we compared different genomic regions and hence differences in the time of isolation of sequence that had no effect on the estimates of dS. We performed dS comparisons of structural versus nonstructural proteins because of their clearly differentiated roles. We also tested whether these differences were serotype specific.

Identifying and quantifying slightly deleterious mutations. To quantify the number of SDMs, we developed a new and simple method. Briefly, we first identified sites under strong purifying selection (highly conserved sites in the alignment). We then calculated the amino acid transition scores for the pairwise sequence comparisons at these sites. We calculated transition scores following the appropriate block substitution matrices (BLOSUM) method (20) given the average pairwise distance for the alignment. Once the distribution of BLOSUM values was determined for a site in the alignment (for example, the alignment containing the 103 sequences), we tested each lineage for the accumulation of SDMs and identified those lineages within and between serotypes accumulating changes with transition values showing significant values (highly negative transition scores) compared to the distribution of values of the rest of the alignment. The rationale behind this analysis is that highly conserved amino acid sites (for example, those amino acid sites showing BLOSUM values highly positive or close to 0) are expected to be functionally important because of their high level of conservation. A strong change at these sites in particular lineages (for example, at terminal lineages within a serotype) is more likely to be an SDM fixed by genetic drift rather than an adaptive amino acid transition. As an observable effect of the method, SDMs were then those that were lineage specific and not spread throughout the phylogenetic tree. SDMs hence were those fixed in the population because of the population structure rather than due to adaptive processes (for example, they were fixed due to genetic drift effects). Consequently, the percentage fixation of SDMs indicates the epidemiological or population history of the serotype or group of serotypes. A greater percentage of SDMs in one serotype would provide evidence of that serotype having undergone genetic drift, while lower percentages would point to strong selective constraints operating within the serotype. Following this procedure, we identified SDMs in each one of the proteins at lineages leading to each one of the serotypes and within serotypes. As in all the cases, we normalized the proportion of sites showing evidence of being SDMs so as to make possible the direct comparison of these proportions between genomic regions and serotypes. This normalization consisted of the transformation of these per-gene and per-serotype proportions in the ratio, taking into account serotypes and genes together.

Identifying compensatory mutations in the FMDV proteome. We searched for conditional advantageous mutations by performing a study of the distribution of SDMs in the available crystal structures for FMDV proteins. Seven proteins have so far been crystallized (with corresponding PDB identifications in parentheses): 1A to 1D (virus capsid 1QGC), 3C (2BHG), 3D (1UO9), and the leader protein (1QOL). For each one of the SDMs identified following the procedure detailed above, we searched for mutations showing the same phylogenetic distribution pattern. For example, if the SDM showed a phylogenetic pattern of mutation in the seven serotypes of 0000100, with 1 indicating strong amino acid change and 0 conservative change, we searched for all those mutations showing that exact pattern, 0000100. Then, both mutations were plotted in the crystal structure of the corresponding protein and the Euclidean distance between them calculated. We calculated this distance as the average distance between the atoms of the amino acid sites to which both mutations belong. Two mutations were considered to have compensated each other if, in addition to presenting the same phylogenetic pattern of transition scores, they were located at a distance less than 4 Å from each other in the protein crystal structure. Also, two amino acid sites can compensate each other indirectly. For example, if site A and site B are more than 8 Å apart but surround (at less than 4 Å) an important functional site C, then changes at site A may affect site C, which have to be compensated by changes at site B. We also considered instances where site A and B were compensating each other. To account for this last situation, we identified those SDMs showing the same phylogenetic patterns but located at a distance greater than 4 Å from each other in the protein structure. Then we identified sites between them in the structure that showed very low divergence levels in comparison with the rest of the molecule. We measured divergence levels per site by estimating the Poisson amino acid distances for each amino acid site in the multiple sequence alignments. We normalized all the results as explained above to make direct comparisons between serotypes and genomic regions.


arrow
RESULTS
 
Different evolutionary forces among genomic regions of FMDV serotypes. Selection analysis identified differences in the degrees of selection between proteins as well as serotypes (Fig. 1A). We detected adaptive evolution (diversifying selection) not only in the most exposed structural protein 1D but also in all of the structural proteins analyzed (Fig. 1A). Among these proteins, 1D presented the highest percentage (15%) of codon sites undergoing adaptive evolution, affecting four out of the seven serotypes. Proteins 1B and 1C have undergone adaptive evolution in almost all the FMDV serotypes, whereas 1A presented evidence of diversifying selection only in the African serotype SAT3 (Fig. 1A). Interestingly, the proportion of codon sites under adaptive evolution in structural proteins correlated with the percentage of a protein's exposed surface ({rho}Spearman = 0.966; P = 0.025).


Figure 1
View larger version (18K):
[in this window]
[in a new window]

  		FIG. 1. Distribution of the percentage of codon regions under adaptive evolution along the genome of FMDV for serotypes. (A) Plot of the percentages of codons under adaptive evolution in the different proteins of FMDV. Different serotypes are color coded. (B) Comparison of the mean percentages of codon regions under adaptive evolution in each serotype between structural proteins and nonstructural proteins.

In nonstructural proteins, we also detected adaptive evolution affecting many serotypes, but the number of serotypes undergoing adaptive evolution was significantly reduced compared to structural exposed proteins. For example, except for the case of protein 1A, the average number of serotypes having undergone adaptive evolution was five out of seven in the case of structural proteins, whereas in nonstructural proteins the average was two serotypes. Notably, the detection of positive selection was not dependent upon the variability of the proteins, because nonstructural proteins 3A and 3B exhibited similar divergence levels as those shown by structural capsid proteins (4), and yet the percentage of sites under diversifying selection was lower than the rest of nonstructural proteins (Fig. 1A). In general, the percentage of protein sites under adaptive evolution in structural proteins ranged between 7.05% and 15.32%, whereas nonstructural proteins showed percentages ranging between 3.51% and 11.11%. When accounting for serotypes, comparison of the percentage of codons under adaptive evolution showed a greater mean for structural than for nonstructural proteins, although the difference was not significant (Fisher's exact test, P = 0.188). When taking into account serotypes individually, structural proteins presented a higher percentage of sites under adaptive evolution compared to nonstructural proteins (Fig. 1B). The only two exceptions were serotypes A, with a slightly greater percentage of adaptively selected sites in nonstructural proteins, and Asia 1, which was the most represented serotype under adaptive evolution when analyzing individual proteins (Fig. 1A).

Relaxed selective pressures at structural capsid proteins in SAT serotypes. Analysis of dS showed that its distribution was highly heterogeneous throughout the genomes. Structural capsid proteins presented significantly greater dS values than nonstructural proteins (one-way analysis of variance test; F[1,2331] = 590.370, P < 0.001) (Fig. 2A). In addition, we also noticed significant differences in dS values between serotypes, with SAT serotypes presenting the highest dS values compared to the other serotypes (Fig. 2B). An analysis of the contribution of serotype and structural versus nonstructural genes to the difference in dS values shows that both factors as well as the interaction between them contribute significantly to the differences in dS. For example, SAT serotypes show greater differences in dS values compared to non-SAT types (Euro-Asiatic serotypes) (F = 251.713; P < 0.001), structural proteins show greater dS values than nonstructural ones (F = 592.201; P < 0.001), and the interaction between serotype and type of protein is also significant (F = 92.394; P < 0.001).

Differential fixation of slightly deleterious mutations among FMDV serotypes. The percentage of SDMs varied along the genomes and phylogeny, with most of these mutations having been fixed at the terminal phylogenetic branches of each serotype. This supports previous tests suggesting that there is a correlation between the age of mutations and their adaptive value (40). The average percentage of mutations showing evidence of being SDMs over the genomes and serotypes was 4.8% (Table 1). All the serotypes presented evidence of SDMs in most of the proteins except in the structural protein 1A (Table 1). In general, among variable amino acid sites the proportion of SDMs was higher in structurally exposed capsid proteins than in nonstructural proteins, although the difference was only marginally significant (Fisher's exact test, P = 0.077). SAT serotypes also presented significantly greater percentages of SDMs than other serotypes (Fisher's exact test, P = 0.0027). In addition, SAT serotypes presented a higher proportion of SDMs at structural proteins than at nonstructural proteins (Fisher's exact test, P = 0.0032). Interestingly, 1A presented no evidence for accumulation of SDMs. The next question we asked was whether SDMs are heterogeneously distributed throughout the FMDV phylogeny.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Fractions of SDMs in FMDV genomes and serotypesa

In Fig. 3 we represent the phylogenetic distribution of SDMs in FMDV in the 103 full-length genomes. This figure highlights differences in the patterns of distribution of SDMs along the FMDV genome and along the phylogeny. First, structural proteins accumulate most of the SDMs in the SAT serotypes (the difference with the proportion of SDMs in non-SATs is significant; Fisher's exact test, P = 9.5 x 10–4), whereas no significant difference was observed in the case of nonstructural proteins between SATs and non-SATs (Fisher's exact test, P = 0.319). Second, most of the SDMs accumulated in structural proteins occurred in the ancestor of the serotypes (Fig. 4a and Table 2). Nonstructural proteins showed, however, the opposite trend, with most of the SDMs accumulating in the terminal branches (Fig. 4b and Table 2). This pattern indicates that structural proteins may have accumulated slightly harmful amino acid substitutions at the basal lineages of the serotypes, especially in SAT serotypes, and some of these mutations became advantageous later within these serotypes undergoing fixation. On the other hand, nonstructural proteins may have accumulated noncompensated SDMs.


Figure 3
View larger version (33K):
[in this window]
[in a new window]

  		FIG. 3. Phylogenetic distribution of the percentages of SDMs in each serotype. Horizontal bars represent the genome of FMDV. Horizontal bars in the lineage leading to each serotype as well as in the tip of each serotype indicate percentages of SDMs in the ancestral serotype lineage and within serotypes, respectively. The percentages of SDMs are normalized along the genome and phylogeny and are therefore comparable between serotypes, between proteins, and even between temporal sampling in the phylogeny (ancestral versus terminal distributions). The normalized percentages of SDMs are color coded, as is the serotype group according to the average percentage of SDMs observed per serotype.


Figure 4
View larger version (9K):
[in this window]
[in a new window]

  		FIG. 4. Comparison of the percentage of slightly deleterious mutations between SAT and non-SAT serotypes in structural (a) and nonstructural (b) proteins. Percentages have been normalized along genomes and phylogenies so as to make them comparable. (a) Black bars refer to the percentage of SDMs identified in the lineages leading to the ancestors of serotypes (root of serotype), whereas gray bars refer to those percentages within serotypes (terminal branches). Standard errors have been averaged for root and terminal lineages.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Distribution of SDMs in the different phylogenetic levels in SAT compared to non-SAT serotypesa

Compensatory mutations generate immune-escaping mutants in SAT serotypes. We considered a pair of SDMs to compensate each other if they presented evidence to be at a distance of less than 4 Å in the protein structure or, alternatively, directly surrounding a highly conserved amino acid site (assumed to be conserved because of its functional importance; see Materials and Methods for details). To measure the amount of advantageous mutations compared to SDMs fixed for each one of these two patterns, we searched for the existence of evidence of CM events in each one of the serotypes and lineages following the procedure described in Materials and Methods. The mean percentages of SDMs compensated for in each one of the analyzed proteins were 13.7%, 32.83%, 20%, 21%, 30.9%, and 9.03% for the leader, 1B, 1C, 1D, 3C, and 3D proteins, respectively. In general, structural proteins showed on average a greater percentage of compensated SDMs (24.61% of the SDMs) compared to nonstructural proteins (17.9% of the SDMs). Furthermore, SAT serotypes presented more CMs than non-SAT serotypes. An average of 56% of SDMs in SAT serotypes were compensated, whereas only around 7% of the SDMs in non-SAT serotypes were compensated. We also analyzed the phylogenetic distribution of these CMs and found that, in the case of structural proteins, most of the CMs took place in the lineages leading to the ancestors of the southern African serotypes, whereas nonstructural proteins presented most of these mutations in the terminal branches within serotypes (Table 3).


View this table:
[in this window]
[in a new window]

  		TABLE 3. Complete list of all compensatory mutations detected in each serotype and in each gene with respect to its location in the root or in the terminal branches of the FMDV phylogenya


arrow
DISCUSSION
 
Previous studies have focused on identifying the evolutionary forces operating at particular protein-coding genes mostly involved in host-cell receptor recognition and evasion from the host immune response. Some of these studies were aimed at identifying the origin of specific serotypes of FMDV, such as the CS serotype (12). We have previously shown that selective constraints operating within the VP1 protein are heterogeneous among the seven FMDV serotypes and are due to the differential action of adaptive evolution (50). This observation was in agreement with previous results in natural as well as in experimental isolates when using a more limited number of sequences (16, 19). In addition, because adaptive evolution affected the antigenic regions in VP1, we speculated that genetic variability was associated with the emergence of immune escape and persistent mutants that were previously observed (19, 29, 48). Covariation between the different amino acid sites within and between antigenic regions has been also linked to the ability of FMDV to escape from the immune response (50). Only a few evolutionary analyses have examined the evolutionary constraints in regions other than the VP1 protein or its precursor, P1 (3, 4, 27). However, variation of adaptive evolution between genomic regions and serotypes has not been previously examined.

Diversifying selection has driven the evolution of FMDV genomes. Our results indicate that most of the FMDV genomes show some evidence of the accumulation of adaptive mutations. These results are in contrast to a previous report that showed protein 1D to be the only structural protein under adaptive evolution (4). These differences, however, may be due to the fact that those investigators failed to screen for lineage-specific adaptive evolution. In their work, for instance, Carrillo and colleagues (4) conducted a codon-specific analysis of the ratios of substitutions per synonymous and nonsynonymous mutations using maximum parsimony and maximum likelihood-based methods. Although their analyses were correct, they were based on the simplistic assumption that lineages evolve at a constant rate (for example, differences in the selective constraints among lineages were not accounted for). Furthermore, the lack of signal for adaptive evolution in 1A in most serotypes may be due to its conserved nature and to its important role in the conformational changes of the capsid upon binding of cell receptors as well as in the generation of ions that allow RNA entry into the cell (21). In addition, 1A is the only structural protein with no exposed surface and hence has no role in generating immune escape mutants. The effect of the percentage of surface-exposed amino acids in structural proteins and their importance for the biology of FMDV can be glimpsed from the positive correlation between the absolute proportion of codons under adaptive evolution and the amount of exposed surface we observed. This correlation may also reflect the evolutionary dynamic of capsid proteins in relation to their ability to interact with the host cell.

All nonstructural proteins were determined to be under adaptive evolution, although the fact that only a few serotypes showed evidence for selective forces may explain the previously undetected selection in these proteins when all serotypes have been analyzed together (4). Among the serotypes, type A is the one showing the lowest proportion of codon sites under adaptive evolution. FMD caused by serotype A is endemic in India, where this serotype coexists with types O and Asia 1. This serotype is the most divergent (48, 49) and has been responsible for several outbreaks in India in recent years (29). However, adaptive evolution in this serotype has occurred in the protease 3C and in the transmembrane protein 2B, but very modestly in the capsid exposed proteins (showing very low percentages of codons under adaptive evolution in only proteins 1B and 1C, but not in 1D). 3C protease is involved in RNA replication and processing of the polyprotein and is considered an important drug target (see, for examples, references 8 and 46) Further, protein 2B is involved in membrane rearrangements required for viral RNA replication and capsid assembly (18). Hence, these results lend support to the hypothesis of selection for faster replication in this serotype, which may generate spontaneous outbreaks.

Varied selective constraints at synonymous sites between FMDV genomic regions. The main conclusion resulting from the analysis of synonymous nucleotide substitutions is that these are heterogeneously distributed along the genome as well as between serotypes. The enhanced heterogeneous genomic distribution of dS values in SAT compared to Euro-Asiatic serotypes pinpoints the complex evolutionary dynamics of the different FMDV proteins and the possible dependency between these dynamics and the epidemiological behavior of the different serotypes. SAT viruses are confined to the African continent and are known for their ability to establish persistent infections and for being the source of spontaneously emerging highly infective variants (53). Indeed, it is widely known that the African buffalo (Syncerus caffer) is instrumental in the maintenance of the disease for the SAT serotypes, for which it acts as a long-term carrier, lasting for up to 5 years. One report has shown that in a small isolated free-living herd the disease was maintained for at least 24 years and through several generations (6). Relaxed selective constraints in these serotypes at structurally exposed proteins may permit the emergence of new mutants that may eventually enable the appearance of immune-escaping viruses that become rapidly fixed in the population. Even though generation of immune-escaping mutants is also important in other serotypes, as reflected also by the faster accumulation of mutations in structural proteins, the more relaxed constraints in SAT serotypes may be related to the high persistence in comparison with the other serotypes. This is supported by the magnified difference between the rates of evolution of structural and nonstructural proteins when comparing SATs to non-SAT serotypes. In conclusion, genetic drift in SATs may have been an important factor in the generation of polymorphic immune-escaping mutants.

High rates of fixation of SDMs in SAT serotypes. Deleterious mutations are those that are eliminated by selection because of their harmful effect on a protein's function and hence on the organism fitness. These mutations are removed from the population by selection regardless of the population dynamics, since the organism is deemed nonfunctional. If mutations are slightly deleterious (SDMs), these can be fixed in the population by genetic drift when the effective population size is small. Viral populations generally present significant sizes within their hosts, but they undergo strong bottlenecks during the infection of other host individuals or in the switch to different stages during the viral life cycle. This dynamic favors the fixation of slightly deleterious mutations. Consequently, analyses of adaptive evolution may be misleading since SDMs can be confounded by adaptive mutations. Furthermore, RNA viruses present a highly active mutational dynamic (9). The fitness consequences of such mutations can be measured generally through fitness assays, which have been used to demonstrate that RNA viruses are prone to accumulate slightly deleterious mutations (31, 43). A recent study analyzed the distribution of SDMs in different proteins from 143 different RNA viruses, with the authors concluding that RNA viruses are indeed prone to fix such mutations (40). Their assumption was that the average age of nonsynonymous mutations increases with their selective advantage (32, 40). Because constraints at synonymous sites may inflate {omega} values at internal branches, we used a different approach that does not rely on the assumption of neutrality at synonymous nucleotide sites.

The higher rate of fixation of SDMs in structural compared to nonstructural proteins as tested in our study lends support to the greater permissibility of structural proteins to accumulate mutations by genetic drift. This pinpoints the idea that the higher diversity of structural proteins compared to nonstructural ones may in part be due to the neutral fixation of SDMs. Interestingly, 1A presented no evidence for accumulation of SDMs. 1A is the only capsid protein that presents no solvent-accessible amino acid regions, which may impose a higher constraint over the permissibility of amino acid sites to accumulate disruptive mutations. Accumulation of SDMs in capsid-exposed proteins may have affected the immune escape dynamics of FMDV, which would explain the ability of SAT serotypes to establish persistent infections (53) much more efficiently than any of the non-SAT serotypes. This is in agreement with experimental results showing that persistent infections of African buffaloes and the genomic and antigenic diversity of SAT serotypes are positively correlated (53). However, we also found SDMs in proteins 3C and 3D, which would suggest also a genetic drift effect enabling the fixation of SDMs in all the proteins of FMDV. Because a greater effect of genetic drift would relax constraints in synonymous and nonsynonymous sites, we should observe higher dS values at synonymous sites, when genetic drift is more important. Mayrose and colleagues elegantly showed that synonymous sites are usually under constraints and that these constraints may be the cause for the inflated {omega} ratios in protein-coding sequences (28). This effect may be more dramatically enhanced by the fact that RNA viruses may be subjected to strong secondary structure constraints at nucleotide sites depending on their location in loops or stems. Our analyses of synonymous nucleotide substitutions clearly demonstrate that synonymous sites are much more relaxed in SAT serotypes than in other serotypes and significantly more relaxed in structural proteins than in nonstructural proteins along the genome. The reason therefore for the increasing number of SDMs in SAT serotypes and in structural proteins very likely may be the result of relaxed constraints in the SAT serotypes. These relaxed constraints may have permitted the accumulation of SDMs that, once compensated for by the fixation of conditional advantageous mutations in nearby structural regions, became advantageous to generate immune escape mutants (explained by the increase of SDMs in structural capsid exposed proteins). To test this hypothesis we examined the phylogenetic distribution of these SDMs in each one of the serotypes as well as possible compensatory mutations being fixed in nearby structural regions on the same branches.

Compensatory mutations generate immune-escaping mutants in SAT serotypes. In this study we have developed and applied a method to identify compensatory mutations (advantageous mutations conditional to their compensatory effects on SDMs). The main purpose of this test was to determine whether SDMs have accumulated as a result of either a founder effect (genetic drift), as a result of a selective process, or due to a change in the selection-drift balance in each serotype. In the former case, we would expect SDMs to accumulate stochastically in terminal branches of the tree and to be located in isolated structural regions, presenting no evidence for nearby compensatory mutations. In the second case we would expect most of the SDMs accumulating in the branches lead to the serotypes (for example, they are ancestral changes that became fixed within serotypes due to its adaptive value) and present also signals of compensation in nearby structural regions. Finally, in the latter case, we should observe a balanced compromise of SDM distribution along the phylogeny and the protein structure. The fact that most CMs in structural proteins take place in the lineages leading to the southern Africa serotypes whereas nonstructural proteins presented most of these mutations within serotypes supports that SDMs in the ancestral SAT serotypes may have been compensated to generate immune-escaping mutants establishing persistent infections. The low proportion of CMs in nonstructural proteins confirms that most mutations within serotypes are real SDMs fixed by genetic drift. We also observed a significant compensatory effect when we compared serotypes confined to southern African to those distributed around the globe, further suggesting that SDMs combined with CMs may have allowed SAT serotypes to generate the genetic and antigenic diversity needed to escape the immune response and to establish persistent infections while maintaining their protein structural/functional stability, as previously suggested (50).

The results and conclusion discussed here indicate that as a consequence of its evolution, eradication of FMD from Africa as a whole is not a prospect for the foreseeable future. To complicate matters further, most countries in sub-Saharan Africa are ill-equipped to face the disease because of lack of infrastructure and financial resources. As a consequence, animal movement and migration will inevitably play a major factor in the epidemiology of the disease. The long-term result of this is that, despite continued efforts and extensive use of vaccines, FMD in Africa is likely to constitute a rapidly increasing problem with more sporadic outbreaks. The danger of this is that the more-developed world may be vulnerable to more undesirable transcontinental introductions of the disease due to the possibility of illegal movement of livestock or exportation of livestock products. Therefore, it is imperative that adequate surveillance systems and routine sampling be maintained to monitor foot and mouth disease virus. This study has far-reaching implications for the evolution and epidemiology of FMD viruses and ultimately in the control of the disease.


arrow
ACKNOWLEDGMENTS
 
This work is supported by a grant from Science Foundation Ireland to M.A.F.

We thank Santiago F. Elena for valuable discussions and critical reading of the manuscript. We also thank the reviewers of the manuscript, who contributed to the improvement of the analyses presented.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Evolutionary Genetics and Bioinformatics Laboratory, Smurfit Institute of Genetics, Trinity College Dublin, Dublin 2, Ireland. Phone: 353 1 8963521. Fax: 353 1 6798558. E-mail: faresm{at}tcd.ie Back

{triangledown} Published ahead of print on 12 November 2008. Back


arrow
REFERENCES
 
    1
  1. Akashi, H. 1999. Within- and between-species DNA sequence variation and the ‘footprint’ of natural selection. Gene 238:39-51.[CrossRef][Medline]
    2. 2
  2. Batschelet, E., E. Domingo, and C. Weissmann. 1976. The proportion of revertant and mutant phage in a growing population, as a function of mutation and growth rate. Gene 1:27-32.[CrossRef][Medline]
    3. 3
  3. Carrillo, C., Z. Lu, M. V. Borca, A. Vagnozzi, G. F. Kutish, and D. L. Rock. 2007. Genetic and phenotypic variation of foot-and-mouth disease virus during serial passages in a natural host. J. Virol. 81:11341-11351.[Abstract/Free Full Text]
    4. 4
  4. Carrillo, C., E. R. Tulman, G. Delhon, Z. Lu, A. Carreno, A. Vagnozzi, G. F. Kutish, and D. L. Rock. 2005. Comparative genomics of foot-and-mouth disease virus. J. Virol. 79:6487-6504.[Abstract/Free Full Text]
    5. 5
  5. Chamary, J. V., J. L. Parmley, and L. D. Hurst. 2006. Hearing silence: non-neutral evolution at synonymous sites in mammals. Nat. Rev. Genet. 7:98-108.[CrossRef][Medline]
    6. 6
  6. Condy, J. B., R. S. Hedger, C. Hamblin, and I. T. Barnett. 1985. The duration of the foot-and-mouth disease virus carrier state in African buffalo (i) in the individual animal and (ii) in a free-living herd. Comp. Immunol. Microbiol. Infect. Dis. 8:259-265.[CrossRef][Medline]
    7. 7
  7. Crandall, K. A., C. R. Kelsey, H. Imamichi, H. C. Lane, and N. P. Salzman. 1999. Parallel evolution of drug resistance in HIV: failure of nonsynonymous/synonymous substitution rate ratio to detect selection. Mol. Biol. Evol. 16:372-382.[Abstract]
    8. 8
  8. Curry, S., N. Roque-Rosell, T. R. Sweeney, P. A. Zunszain, and R. J. Leatherbarrow. 2007. Structural analysis of foot-and-mouth disease virus 3C protease: a viable target for antiviral drugs? Biochem. Soc Trans. 35:594-598.[CrossRef][Medline]
    9. 9
  9. Domingo, E., and J. J. Holland. 1997. RNA virus mutations and fitness for survival. Annu. Rev. Microbiol. 51:151-178.[CrossRef][Medline]
    10. 10
  10. Drake, J. W., and J. J. Holland. 1999. Mutation rates among RNA viruses. Proc. Natl. Acad. Sci. USA 96:13910-13913.[Abstract/Free Full Text]
    11. 11
  11. Edgar, R. C. 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32:1792-1797.[Abstract/Free Full Text]
    12. 12
  12. Elena, S. F., F. Gonzalez-Candelas, and A. Moya. 1992. Does the VP1 gene of foot-and-mouth disease virus behave as a molecular clock? J. Mol. Evol. 35:223-229.[Medline]
    13. 13
  13. Escarmis, C., E. Lazaro, and S. C. Manrubia. 2006. Population bottlenecks in quasispecies dynamics. Curr. Top. Microbiol. Immunol. 299:141-170.[CrossRef][Medline]
    14. 14
  14. Fares, M. A. 2004. SWAPSC: sliding window analysis procedure to detect selective constraints. Bioinformatics 20:2867-2868.[Abstract/Free Full Text]
    15. 15
  15. Fares, M. A., S. F. Elena, J. Ortiz, A. Moya, and E. Barrio. 2002. A sliding window-based method to detect selective constraints in protein-coding genes and its application to RNA viruses. J. Mol. Evol. 55:509-521.[CrossRef][Medline]
    16. 16
  16. Fares, M. A., A. Moya, C. Escarmis, E. Baranowski, E. Domingo, and E. Barrio. 2001. Evidence for positive selection in the capsid protein-coding region of the foot-and-mouth disease virus (FMDV) subjected to experimental passage regimens. Mol. Biol. Evol. 18:10-21.[Abstract/Free Full Text]
    17. 17
  17. Flanegan, J. B., and D. Baltimore. 1977. Poliovirus-specific primer-dependent RNA polymerase able to copy poly(A). Proc. Natl. Acad. Sci. USA 74:3677-3680.[Abstract/Free Full Text]
    18. 18
  18. Grubman, M. J., and B. Baxt. 2004. Foot-and-mouth disease. Clin. Microbiol. Rev. 17:465-493.[Abstract/Free Full Text]
    19. 19
  19. Haydon, D. T., A. D. Bastos, N. J. Knowles, and A. R. Samuel. 2001. Evidence for positive selection in foot-and-mouth disease virus capsid genes from field isolates. Genetics 157:7-15.[Abstract/Free Full Text]
    20. 20
  20. Henikoff, S., and J. G. Henikoff. 1992. Amino acid substitution matrices from protein blocks. Proc. Natl. Acad. Sci. USA 89:10915-10919.[Abstract/Free Full Text]
    21. 21
  21. Hogle, J. M. 2002. Poliovirus cell entry: common structural themes in viral cell entry pathways. Annu. Rev. Microbiol. 56:677-702.[CrossRef][Medline]
    22. 22
  22. Inoue, T., S. Parida, D. J. Paton, W. Linchongsubongkoch, D. Mackay, Y. Oh, D. Aunpomma, S. Gubbins, and T. Saeki. 2006. Development and evaluation of an indirect enzyme-linked immunosorbent assay for detection of foot-and-mouth disease virus nonstructural protein antibody using a chemically synthesized 2B peptide as antigen. J. Vet. Diagn. Investig. 18:545-552.[Abstract/Free Full Text]
    23. 23
  23. Jackson, A. L., H. O'Neill, F. Maree, B. Blignaut, C. Carrillo, L. Rodriguez, and D. T. Haydon. 2007. Mosaic structure of foot-and-mouth disease virus genomes. J. Gen. Virol. 88:487-492.[Abstract/Free Full Text]
    24. 24
  24. Knowles, N. J., and A. R. Samuel. 2003. Molecular epidemiology of foot-and-mouth disease virus. Virus Res. 91:65-80.[CrossRef][Medline]
    25. 25
  25. Li, W. H. 1993. Unbiased estimation of the rates of synonymous and nonsynonymous substitution. J. Mol. Evol. 36:96-99.[CrossRef][Medline]
    26. 26
  26. Mahy, B. W. 2005. Introduction and history of foot-and-mouth disease virus. Curr. Top. Microbiol. Immunol. 288:1-8.[Medline]
    27. 27
  27. Mason, P. W., J. M. Pacheco, Q. Z. Zhao, and N. J. Knowles. 2003. Comparisons of the complete genomes of Asian, African and European isolates of a recent foot-and-mouth disease virus type O pandemic strain (PanAsia). J. Gen. Virol. 84:1583-1593.[Abstract/Free Full Text]
    28. 28
  28. Mayrose, I., A. Doron-Faigenboim, E. Bacharach, and T. Pupko. 2007. Towards realistic codon models: among site variability and dependency of synonymous and non-synonymous rates. Bioinformatics 23:i319-i27.[Abstract/Free Full Text]
    29. 29
  29. Mittal, M., C. Tosh, D. Hemadri, A. Sanyal, and S. K. Bandyopadhyay. 2005. Phylogeny, genome evolution, and antigenic variability among endemic foot-and-mouth disease virus type A isolates from India. Arch. Virol. 150:911-928.[CrossRef][Medline]
    30. 30
  30. Moffat, K., G. Howell, C. Knox, G. J. Belsham, P. Monaghan, M. D. Ryan, and T. Wileman. 2005. Effects of foot-and-mouth disease virus nonstructural proteins on the structure and function of the early secretory pathway: 2BC but not 3A blocks endoplasmic reticulum-to-Golgi transport. J. Virol. 79:4382-4395.[Abstract/Free Full Text]
    31. 31
  31. Montville, R., R. Froissart, S. K. Remold, O. Tenaillon, and P. E. Turner. 2005. Evolution of mutational robustness in an RNA virus. PLoS Biol. 3:e381.[CrossRef][Medline]
    32. 32
  32. Nielsen, R., and D. M. Weinreich. 1999. The age of nonsynonymous and synonymous mutations in animal mtDNA and implications for the mildly deleterious theory. Genetics 153:497-506.[Abstract/Free Full Text]
    33. 33
  33. Nunez, J. I., E. Baranowski, N. Molina, C. M. Ruiz-Jarabo, C. Sanchez, E. Domingo, and F. Sobrino. 2001. A single amino acid substitution in nonstructural protein 3A can mediate adaptation of foot-and-mouth disease virus to the guinea pig. J. Virol. 75:3977-3983.[Abstract/Free Full Text]
    34. 34
  34. Pacheco, J. M., T. M. Henry, V. K. O'Donnell, J. B. Gregory, and P. W. Mason. 2003. Role of nonstructural proteins 3A and 3B in host range and pathogenicity of foot-and-mouth disease virus. J. Virol. 77:13017-13027.[Abstract/Free Full Text]
    35. 35
  35. Parmley, J. L., J. V. Chamary, and L. D. Hurst. 2006. Evidence for purifying selection against synonymous mutations in mammalian exonic splicing enhancers. Mol. Biol. Evol. 23:301-309.[Abstract/Free Full Text]
    36. 36
  36. Pereda, A. J., G. A. Konig, S. A. Chimeno Zoth, M. Borca, E. L. Palma, and M. E. Piccone. 2002. Full length nucleotide sequence of foot-and-mouth disease virus strain O1 Campos/Bra/58. Brief report. Arch. Virol. 147:2225-2230.
    37. 37
  37. Pereira, H. G. 1981. Foot and mouth disease. Academic Press, London, United Kingdom.
    38. 38
  38. Porter, A. G. 1993. Picornavirus nonstructural proteins: emerging roles in virus replication and inhibition of host cell functions. J. Virol. 67:6917-6921.[Free Full Text]
    39. 39
  39. Posada, D., and K. A. Crandall. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14:817-818.[Abstract/Free Full Text]
    40. 40
  40. Pybus, O. G., A. Rambaut, R. Belshaw, R. P. Freckleton, A. J. Drummond, and E. C. Holmes. 2007. Phylogenetic evidence for deleterious mutation load in RNA viruses and its contribution to viral evolution. Mol. Biol. Evol. 24:845-852.[Abstract/Free Full Text]
    41. 41
  41. Resch, A. M., L. Carmel, L. Marino-Ramirez, A. Y. Ogurtsov, S. A. Shabalina, I. B. Rogozin, and E. V. Koonin. 2007. Widespread positive selection in synonymous sites of mammalian genes. Mol. Biol. Evol. 24:1821-1831.[Abstract/Free Full Text]
    42. 42
  42. Rueckert, R. R. 1996. Picornaviridae: the viruses and their replication. Lippincott-Raven, Philadelphia, PA.
    43. 43
  43. Sanjuan, R., A. Moya, and S. F. Elena. 2004. The distribution of fitness effects caused by single-nucleotide substitutions in an RNA virus. Proc. Natl. Acad. Sci. USA 101:8396-8401.[Abstract/Free Full Text]
    44. 44
  44. Sharp, P. M. 1997. In search of molecular Darwinism. Nature 385:111-112.[Medline]
    45. 45
  45. Sobrino, F., M. Saiz, M. A. Jimenez-Clavero, J. I. Nunez, M. F. Rosas, E. Baranowski, and V. Ley. 2001. Foot-and-mouth disease virus: a long known virus, but a current threat. Vet. Res. 32:1-30.[CrossRef][Medline]
    46. 46
  46. Sweeney, T. R., N. Roque-Rosell, J. R. Birtley, R. J. Leatherbarrow, and S. Curry. 2007. Structural and mutagenic analysis of foot-and-mouth disease virus 3C protease reveals the role of the beta-ribbon in proteolysis. J. Virol. 81:115-124.[Abstract/Free Full Text]
    47. 47
  47. Swofford, D. 1998. PAUP*: phylogenetic analysis using parsimony (*and other methods). Sinauer Associates, Sunderland, MA.
    48. 48
  48. Tosh, C., D. Hemadri, A. Sanyal, and S. K. Bandyopadhyay. 2003. Genetic and antigenic analysis of two recently circulating genotypes of type A foot-and-mouth disease virus in India: evidence for positive selection in the capsid-coding genes. Arch. Virol. 148:853-869.[CrossRef][Medline]
    49. 49
  49. Tosh, C., A. Sanyal, D. Hemadri, and R. Venkataramanan. 2002. Phylogenetic analysis of serotype A foot-and-mouth disease virus isolated in India between 1977 and 2000. Arch. Virol. 147:493-513.[CrossRef][Medline]
    50. 50
  50. Tully, D. C., and M. A. Fares. 2006. Unravelling selection shifts among foot-and-mouth disease virus serotypes. Evol. Bioinform. 2:237-251.
    51. 51
  51. Vakharia, V. N., M. A. Devaney, D. M. Moore, J. J. Dunn, and M. J. Grubman. 1987. Proteolytic processing of foot-and-mouth disease virus polyproteins expressed in a cell-free system from clone-derived transcripts. J. Virol. 61:3199-3207.[Abstract/Free Full Text]
    52. 52
  52. van Rensburg, H., D. Haydon, F. Joubert, A. Bastos, L. Heath, and L. Nel. 2002. Genetic heterogeneity in the foot-and-mouth disease virus Leader and 3C proteinases. Gene 289:19-29.[CrossRef][Medline]
    53. 53
  53. Vosloo, W., A. D. Bastos, E. Kirkbride, J. J. Esterhuysen, D. J. van Rensburg, R. G. Bengis, D. W. Keet, and G. R. Thomson. 1996. Persistent infection of African buffalo (Syncerus caffer) with SAT-type foot-and-mouth disease viruses: rate of fixation of mutations, antigenic change and interspecies transmission. J. Gen. Virol. 77:1457-1467.[Abstract/Free Full Text]

Journal of Virology, January 2009, p. 781-790, Vol. 83, No. 2
0022-538X/09/$08.00+0     doi:10.1128/JVI.01500-08
Copyright © 2009, American Society for Microbiology. All Rights Reserved.




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Tully, D. C.
Right arrow Articles by Fares, M. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Tully, D. C.
Right arrow Articles by Fares, M. A.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»