The mechanisms of evolution of plant viruses are being unraveled, yet the timescale of their evolution remains an enigma. To address this critical issue, the divergence time of plant viruses at the intra- and inter-specific levels was assessed. The time of the most recent common ancestor (TMRCA) of Rice yellow mottle virus (RYMV; genus Sobemovirus) was calculated by a Bayesian coalescent analysis of the coat protein sequences of 253 isolates collected between 1966 and 2006 from all over Africa. It is inferred that RYMV diversified approximately 200 years ago in Africa, i.e., centuries after rice was domesticated or introduced, and decades before epidemics were reported. The divergence time of sobemoviruses and viruses of related genera was subsequently assessed using the age of RYMV under a relaxed molecular clock for calibration. The divergence time between sobemoviruses and related viruses was estimated to be approximately 9,000 years, that between sobemoviruses and poleroviruses approximately 5,000 years, and that among sobemoviruses approximately 3,000 years. The TMRCA of closely related pairs of sobemoviruses, poleroviruses, and luteoviruses was approximately 500 years, which is a measure of the time associated with plant virus speciation. It is concluded that the diversification of RYMV and related viruses has spanned the history of agriculture, from the Neolithic age to the present.
The timescale of the evolution of plant viruses is an enigma, and even its order of magnitude is unknown. This critical issue is addressed here by calculating the age of plant viruses. An accurate estimate of the age of Rice yellow mottle virus (RYMV) was obtained by statistical analysis of a set of dated sequences. The age of RYMV provides a reliable calibration of related viruses, applying recently developed relaxed molecular clock models. It was found that RYMV diversified approximately 200 years ago, and that inter-specific diversification ranged from 500 years to 9,000 years. Altogether, plant virus diversification has spanned the history of agriculture from the Neolithic age to the present. This suggests that the Neolithic was a period of epidemiological transition for plant virus diseases, as already proposed for infectious human diseases. Intrinsically, it is for the same reason: increased contacts between hosts, pathogens, and vectors. This is consistent with the view that RNA viruses have a recent origin, and that humans have become the world's greatest evolutionary force.
Citation: Fargette D, Pinel-Galzi A, Sérémé D, Lacombe S, Hébrard E, Traoré O, et al. (2008) Diversification of Rice Yellow Mottle Virus and Related Viruses Spans the History of Agriculture from the Neolithic to the Present. PLoS Pathog 4(8): e1000125. doi:10.1371/journal.ppat.1000125
Editor: Edward C. Holmes, The Pennsylvania State University, United States of America
Received: March 31, 2008; Accepted: July 14, 2008; Published: August 15, 2008
Copyright: © 2008 Fargette et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was funded by the ECOGER action of the French National Program ANR “Ecosphère continentale: processus, modélisation et risques environnementaux.” Support by INRA is also acknowledged.
Competing interests: The authors have declared that no competing interests exist.
The mechanisms of evolution of plant viruses are being progressively unraveled –, yet the timescale of their evolution remains an enigma. Even the order of magnitude is unknown . Several viruses showed few genetic changes between isolates separated in space and time, sometimes for centuries –. In contrast, recent evidence from statistical analyses of sequences of dated isolates of Tomato yellow leaf curl virus (genus Geminivirus) , Rice yellow mottle virus (genus Sobemovirus) (RYMV)  and Zucchini yellow mosaic virus (genus Potyvirus)  indicated rapid evolution, similar to that of most animal viruses. The paradox is addressed here by calculating the divergence time of plant viruses at the intra- and inter-specific levels using RYMV and related viruses.
Molecular-dating techniques provide insights into the history of lineages that have a poor or non-existent fossil record, such as viruses ,. These techniques were originally based on the assumption of a strict molecular clock reflecting steady accumulation of genetic changes over time. Recently, new methods enable the incorporation of variable rates into molecular dating . Here, we applied a Bayesian Markov Chain Monte-Carlo method for performing relaxed phylogenies that is able to co-estimate phylogeny and divergence times under uncorrelated relaxed-clock models .
RYMV causes an emergent disease that was first observed in 1966 in Kenya. Since then, it has been reported in nearly all rice-growing countries of sub-Saharan Africa. RYMV is transmitted by coleopterous insects and is also disseminated abiotically. It has a narrow host range limited to wild and cultivated rices and a few related grasses . There is no evidence of recombination between RYMV isolates ,. The rate of evolution of RYMV was recently evaluated using the coat protein (CP) sequences of 253 isolates collected between 1966 and 2006 from all over Africa . The same group of sequences is analyzed here to assess the time of their most recent common ancestor (TMRCA), which is a measure of the divergence time of RYMV. The TMRCA was calculated by a Bayesian coalescent analysis of the sequences using several molecular clock and population genetic models .
Sobemoviruses infect both monocotyledonous and dicotyledonous plants, but the host range of each virus species is narrow and confined to a few plant species of the Poaceae or Fabaceae. Sobemoviruses are transmitted by beetle vectors, seeds and direct contact . They share a common genomic organization, as found after re-sequencing some of the virus species ,. Ten sobemovirus species have been fully sequenced, nine of them are currently registered by ICTV  and a tentative one, Imperata yellow mottle virus (IYMV), was recently isolated from Imperata cylindrica in Africa . Their genomes contain four open reading frames (ORFs). ORF1, located at the 5′ end of the genome, encodes a protein involved in virus movement and gene silencing suppression. ORF2 comprises two overlapping ORFs. ORF2a encodes a serine protease and a viral-genome-linked protein. ORF2b is translated through a -1 ribosomal frameshift mechanism through a fusion protein. It encodes the RNA-dependent RNA polymerase (RdRp). The coat protein gene (ORF4) is expressed by a sub-genomic RNA at the 3′ end of the genome. No evidence of recombination between sobemoviruses has been found either in phylogenetic , or experimental studies .
The genus Sobemovirus is not assigned to a family. However, the RdRp of the sobemoviruses is phylogenetically related to that of the poleroviruses and enamoviruses (family Luteoviridae) , and to Poinsettia latent virus (PnLV), a putative polerovirus-sobemovirus hybrid . Sobemoviruses, luteoviruses (family Luteoviridae) and dianthoviruses (family Tombusviridae) are more distantly related. The CPs of sobemoviruses are related to those of the necroviruses (family Tombusviridae), and the CPs of the poleroviruses to those of the luteoviruses . Recombination events between ancestors of these genera are the likely causes of the present situation ,. Altogether, this led to the proposal of a “supergroup” to include these related genera . The RdRp of the sobemoviruses also shows similarities with that of Mushroom bacilliform virus (MBV) (genus Barnavirus, family Barnaviridae) which infects mushrooms .
The divergence time of sobemoviruses was assessed from the full-length sequences using the age of RYMV under relaxed molecular clock models for calibration. The divergence time of the sobemoviruses with members of related genera was inferred from RdRp sequences with the same methodology. The time associated with plant virus speciation was assessed by calculating the TMRCA of closely related pairs of sobemoviruses, poleroviruses and luteoviruses. Collectively, these studies provide estimates of the diversification time of a plant virus species, the time associated with plant virus speciation, and the TMRCA of plant viruses of the same genus and of different genera. The intra- and inter-specific plant virus diversification was found to span the history of agriculture from the Neolithic age to the present.
TMRCA of RYMV
The estimates of the TMRCA of RYMV inferred from the 253 dated CP sequences were dependent on both molecular clock and demographic models. Models enforcing relaxed molecular clocks performed better than the strict clock model, whatever the population genetic model selected (Table 1). The average substitution rates ranged from 5.1×10−4 to 12.3×10−4 nucleotides (nt)/site/year among the models (data not shown). The highest marginal likelihood was obtained with the model implementing the relaxed uncorrelated exponential molecular clock and the exponential growth model. The Bayes Factor (BF) gave strong support to this model when compared to other clock and population models. Under this model, the average TMRCA of RYMV was 195 years and the substitution rate was 11.7×10−4 nt/site/year. The median was 182 years. The highest density probability (HPD) interval ranged from 107 years to 308 years, with an approximate lognormal distribution of the estimates. Subsequently, a lognormal distribution with a lognormal mean of 5.2 and a standard deviation of 0.3 was applied as the prior distribution of TMRCA of RYMV for the upward calibration of nodes of sobemoviruses and related viruses.
TMRCA of sobemoviruses and related viruses
The two most divergent RYMV isolates, Ma10 and Tz202, were collected 5,000 km apart in Mali and Tanzania, respectively, and differed by 10.5% in the full genome. They were subsequently referred to as isolates RYMV-1 and RYMV-2. The distribution of the estimates of the TMRCA of RYMV calculated from the dated CP sequences was taken as the prior of their divergence time (node 1 in all figures). The full sequences of these two RYMV isolates and of nine other sobemoviruses were considered (Table 2). A total of 4,798 characters was analyzed, 3,432 of them (72%) being parsimony-informative. The lognormal clock model performed better than the strict model (marginal likelihoods in loge units were −50237 and −50260, respectively), whereas the exponential model failed to converge. The deviation from the hypothesis of a strict clock was limited (coefficient of variation = 0.23). The TMRCA and the substitution rates of the sobemoviruses under the lognormal and the strict clock models were close: 3,137 vs. 3,326 years, and 4.0×10−4 vs. 3.7×10−4 nt/site/year, respectively. The Yule speciation process and the constant population size coalescent model as tree priors yielded similar estimates. Among sobemoviruses, RYMV is most closely related to IYMV (node 2). The TMRCA of RYMV and IYMV was 1262 years (523–2248) (Figure 1). Cocksfoot mottle virus (CfMV, genus Sobemovirus) also infects monocotyledonous plants but without overlap in host or geographical range. CfMV is the species the most closely related to RYMV and IYMV (node 3). The TMRCA of CfMV, IYMV and RYMV was 2,317 years (921–3,929). The root height of all sobemoviruses (node 4) was 3,137 years (1,133–5,295).
The tree was reconstructed from the full sequences by Bayesian inference under an uncorrelated lognormal relaxed molecular clock model. The age of RYMV was used for calibration (node 1). Nodes 2–4 are associated with more internal nodes. External node “a” gathers SeMV and SBMV, the two most closely related sobemoviruses. The posterior probabilities are below the nodes (italics). The divergence times (in years) are positioned at the nodes, and the 95% HPD intervals are indicated in brackets. The species names and the sequence accession numbers are given in Table 1.
The divergence time of sobemoviruses and related viruses was assessed from the RdRp sequences (Table 2). Again, the distribution of the estimates of the TMRCA of RYMV calculated from the dated CP sequences was taken as the prior of the divergence time of RYMV-1 and RYMV-2 (node 1). A total of 2,199 characters were analyzed, 1,607 being parsimony-informative (73%). The model enforcing the lognormal clock model performed better than the strict model (marginal likelihoods in loge units were −29663 and −29677, respectively), whereas the exponential model failed to converge. Again, the deviation from the hypothesis of strict clock was limited (coefficient of variation = 0.28) and the estimates were close. For instance, the basal root height under the lognormal and the strict clocks were 8,772 vs. 10,440 years, and the substitution rates were 3.2×10−4 and 2.8×10−4 nt/site/year, respectively. However, the HPD interval was wider with the lognormal model (2,929–15,671 years) than with the strict model (4,971–18,060 years), i.e., a 1∶5.4 ratio for the lognormal model vs. a 1∶3.6 ratio for the strict clock model.
The age of sobemoviruses (node 4) calculated on the full genome and on the RdRp sequences were similar (3,137 and 3,056 years, respectively) despite the difference in number of parsimony-informative characters considered (3,432 vs. 1,607 characters). The age of sobemoviruses calculated on the CP sequences was similar (2,884 years) although the dN/dS ratios of the RdRp and of the CP genes were 0.18 and 0.39, respectively, reflecting the differences in functional constraints operating on the two genes. The TMRCA of the sobemoviruses and MBV (node 5) was 4,418 years (1,480–8,092) (Figure 2). The divergence time of sobemoviruses, poleroviruses, and MBV (node 6) was 5,118 years (1,840–9,050). The root height of sobemoviruses, MBV, poleroviruses, and luteoviruses (node 7) was 8,772 years (2,929–15,671). The TMRCA of these viruses and Red clover necrotic mosaic virus (RCNMV) (genus Dianthovirus) was 9,059 (3,370–16,260) (node 8), a value not substantially different from node 7 (Figure 3).
The tree was reconstructed from the RdRp sequences by Bayesian inference under an uncorrelated lognormal relaxed molecular clock model. The age of RYMV was used for calibration (node 1). Nodes 4–7 are associated with more internal nodes. External node “b” gathers CYDV-RPS and CYDV-RPV, the two most closely related poleroviruses. External node “c” gathers BYDV-PAS and BYDV-MAV, the two most closely related luteoviruses. The posterior probabilities are below the nodes (italics). The divergence times (in years) are positioned at the nodes, and the 95% HPD intervals are indicated in brackets. The species genus is indicated alongside the vertical line. The species names and the sequence accession numbers are given in Table 1.
The divergence times and the 95% HPD intervals are in brackets and framed. Nodes 1 to 8 encompass plant virus diversification at the intra-specific, intra- and inter-generic levels, as indicated by the vertical lines. Nodes “a,” “b,” and “c” gather closely related pairs of viruses. The time axis spreads from the beginning of the Neolithic period to the present.
TMRCA of closely related virus species
Several isolates of Subterranean clover mottle virus (SCMoV, genus Sobemovirus), which caused a disease restricted to southwest Australia, were fully sequenced . The highest divergence between two isolates collected in 1991 and 1996, respectively, was 1.2%. Accordingly, the divergence time of SCMoV was estimated to be 20 years with a HPD interval of 6–44 years, indicating a date of diversification between 1952 and 1990 (Figure 3). Southern bean mosaic virus (SBMV, genus Sobemovirus) and Sesbania mosaic virus (SeMV, genus Sobemovirus) differed by 31.6% in their complete genome and thus are the two most closely related sobemoviruses (Figure 1). Their divergence time (node “a”) was 526 years (169–938). Cereal yellow dwarf virus CYDV-RPV and CYDV-RPS, two closely related poleroviruses, differed by 22% in their RdRp sequences (Figure 2). Their TMRCA was 531 years (180–1,018) (node “b”). Barley yellow dwarf virus BYDV-MAV and BYDV-PAS, two closely related luteoviruses, differed by 21.1% in their RdRp. Their divergence time (node “c”) was 451 years (141–813). Altogether, the TMRCA of these closely related pairs of sobemoviruses, poleroviruses and luteoviruses ranged from approximately 450 to 550 years.
The 253 RYMV isolates collected in 16 countries represent the diversity of the species ,. Accordingly, the TMRCA of these 253 isolates provides a reliable estimate of the divergence time of RYMV. By contrast, the 10 sobemovirus species probably underestimate the number of sobemoviruses in cultivated and wild plants . However, theoretical studies indicated that numerous samples are not necessary to date old coalescent events. It was calculated that the coalescence time of a sample of 10 taxa was 90% of the expected coalescent time of the entire population . Consequently, although calculated on a limited number of species, the TMRCA of sobemoviruses and members of related genera provide reliable estimates of their divergence times.
Relaxed molecular clock models incorporate the rate variation among lineages in estimates of divergence time. Accordingly, any punctuated evolution, as might occur in species jump, should be accounted for in the relaxed clock models. Results from relaxed clocks should be evaluated in relation to those of strict clocks . In our study, the lognormal relaxed clock model performed better at the inter-specific level than the strict clock model. However, the deviation from a strict clock model was limited. This explained why the TMRCA estimates under strict and relaxed clock models were close.
There was, however, a 1∶3 ratio between the lower and upper bounds of the HPD intervals of the TMRCA of RYMV (308 and 107 years, respectively). The variance of this estimate, further enlarged after relaxation of the molecular clock assumption, accounted for the large HPD intervals of divergence times at the inter-specific level. However, the HPD of RYMV divergence time is still substantially narrower than those of the other plant viruses studied with dated sequences ,. This is likely to be due to the larger number of isolates used and the wider range of dates encompassed with RYMV. This could also reflect the fact that the RYMV isolates were collected, sequenced and analyzed by the same group of scientists, subsequently reducing the uncertainties associated with the use of data sets from various and heterogeneous sources.
Assessing the divergence time of RYMV from dated sequences does not suffer from the limitations of alternative approaches. Measuring RYMV evolution rate from experimental studies or from old virus specimens was previously found to be inappropriate . Applying epidemiological evidence is not adequate either. Symptoms of RYMV were first described in 1966, i.e., 40 years ago, a value inconsistent with the 107 to 308 years of the HPD interval for RYMV diversification. This means that RYMV diversified decades before the disease symptoms were reported. It also suggests that RYMV caused epidemics long before it was recognized as a disease. The first report of symptoms should better be considered as a lower bound of virus diversification, i.e., the minimum time since the virus diversified. Exceptions are viruses in localized and well-surveyed regions such as SCMoV in southwest Australia. From dated sequences, SCMoV diversification was estimated to occur between 1952 and 1990. This interval includes 1979, the year when the first symptoms were reported . Biogeographical evidence to estimate divergence time can be misleading too. Madagascar was separated from mainland Africa approximately 100 millions years ago. The timescale of evolution of RYMV excludes the possibility that the divergence between isolates from Madagascar and from East Africa reflects vicariance events . Altogether, the set of CP sequences of 253 dated isolates of RYMV currently provides the most reliable approach to date plant virus diversification.
The divergence time of RYMV was approximately 200±100 years, whereas symptoms were reported for the first time in 1966 in East Africa  and in 1975 in West Africa . The African rice Oryza glaberrima was domesticated in West Africa approximately 3,000 years ago, whereas the Asiatic rice O. sativa was introduced in the 10th and 16th centuries in East and West Africa, respectively ,. Consequently, RYMV diversified centuries after rice was domesticated or introduced in Africa, and decades before epidemics were reported. The 19th century was a period of extension of the rice culture in Africa . This may have favored the spread of RYMV from is primary host to rice, followed by its dissemination throughout Africa.
The divergence time between sobemoviruses and related viruses was estimated to be approximately 9,000 years, that between sobemoviruses and poleroviruses approximately 5,000 years, and that among sobemoviruses approximately 3,000 years (Figure 3). The estimates of the age of sobemovirus diversification did not depend on the sequence length or on the gene considered. Even considering their HPD, these time-scales encompassed the Neolithic “agricultural revolution.” This period was the transition from nomadic hunting and gathering communities to agriculture and settlement. It occurred independently in several prehistoric human societies between 10,000 and 4,000 years before present (BP) ,. Ancient peoples completed the domestication of all major plant species upon which human survival depends ca. 4,500 years BP ,.
One likely consequence of agricultural expansion is the dramatic increase of opportunities for encounters between wild and cultivated plant species, between cultivated plants at various stages of domestication, and between plants and potential insect vectors. These new encounters must have facilitated the emergence of plant viruses. This is still apparent nowadays when crop species are moved from their center of origin into new regions. They are exposed to infection by indigenous viruses to which they have not previously been adapted ,,. Further crowding of plants associated with agricultural development, especially monoculture, facilitated the build-up of vector populations and the disease spread, as is still apparent at the present time . Similarly, the Neolithic age was critical for the emergence of infectious human diseases, a period referred to as the first epidemiologic transition . This was attributed to the increased contacts between humans and wild fauna, and among humans themselves. Our results suggest that the Neolithic age was also a period of epidemiological transition for plant pathogens such as viruses, intrinsically for the same reason: increased contacts between hosts, pathogens and vectors. The hypothesis that the emergence of plant viruses is linked to the development of agriculture is consistent with the view that RNA viruses have a recent origin , and also that humans have become the world's greatest evolutionary force .
The divergence time of the RdRp of sobemoviruses and poleroviruses bounded the dates of the recombination events between the genera. They must have occurred after the diversification of the common ancestor of the RdRp of sobemoviruses and poleroviruses approximately 5,000 years ago, and before the diversification of each of the two genera approximately 3,000 years ago. These recombination events, which necessarily involved the co-existence of different genomes in the same plant, must have been favored by the increased opportunities of co-infections associated with agricultural expansion that started during the Neolithic age. Events occurring at this period also possibly led to virus diversification outside the plant kingdom, as suggested by the divergence time of the sobemovirus and MBV estimated to be approximately 4,500 years.
Much effort has been recently devoted to the numerical taxonomy of plant viruses to set thresholds in percentage of nucleotide divergence for demarcation criteria at the intra- and inter-specific levels . In this study, nucleotide divergence illuminates the timescales associated with these demarcation criteria (Figure 3). The limited deviation from the strict clock model allowed the comparison of these timescales. The inter-generic divergence time between sobemoviruses, poleroviruses and luteoviruses exceeded approximately 3,000 years. The inter-specific divergence of sobemoviruses ranged from approximately 500 to 3,000 years. Consistent divergence times of approximately 500 years were obtained between closely related pairs of sobemoviruses, luteoviruses and poleroviruses, which were first considered as strains and later ranked as different species. This provides an estimate of the time associated with speciation of plant viruses. The intra-specific divergence time of RYMV was approximately 200 years, which is 2 to 3 times less than the speciation time of plant viruses. Overall, this range of values revealed that plant diversification at the intra- and inter-specific levels occurred within the Holocene, and has spanned the entire history of agriculture, from the Neolithic age to the present.
Materials and Methods
The CP genes (720 nucleotides) of 253 isolates from 16 countries in Africa collected over a 40-year period, and the complete genome of two isolates of RYMV were previously sequenced ,. The complete sequences of the sobemoviruses, the sequences of the RdRp of the poleroviruses, luteoviruses, PnLV, and MBV were downloaded from GenBank (Table 1). The sequences were aligned using CLUSTAL W with default parameters . The parameters of interest were estimated within a Bayesian coalescent framework by a Markov Chain Monte Carlo (MCMC) method using the Bayesian Evolutionary Analysis by Sampling Trees (BEAST) program (http://beast.bio.ed.ac.uk/) . The Bayesian MCMC method estimates a parameter as the mean of its posterior distribution while simultaneously incorporating uncertainty in the underlying genealogy or phylogeny and other parameters.
The length and number of MCMC chains were chosen so that the effective sample size for the root height parameter and other parameters was >200, indicating that the parameter space was sufficiently explored. The convergence of the parameters to a stationary distribution was assessed with TRACER , and the statistical uncertainties were summarized in the 95% HPD intervals. Comparison of models was performed by calculating the Bayes Factor (BF), which is the ratio of the marginal likelihood of each model . A value of loge(BF) >2.3 was taken as evidence of a strong support for the model with the highest marginal likelihood. The coefficient of variation of the evolution rates calculated under the uncorrelated lognormal relaxed clock model was used to assess the degree of deviation from the strict molecular clock model.
TMRCA of RYMV
In earlier studies, the evolution rate was the target parameter , whereas here the TMRCA or the root height was the parameter of interest. It was taken as a measure of the divergence time of RYMV. The root height was estimated by enforcing strict and relaxed (uncorrelated lognormal and uncorrelated exponential) molecular clocks as implemented in BEAST . Four demographic models were applied as coalescent priors: constant population size, exponential growth, expansion growth, and a piece-wise Bayesian skyline plot . Default values were used for the other priors. The uncertainty in the TMRCA of RYMV is summarized by the highest posterior density interval that contains 95% of the marginal posterior distribution.
TMRCA of sobemoviruses and related viruses
The full sequences of 10 sobemoviruses were considered for the intra-generic analysis (Table 2). The RdRp sequences of related viruses were added for the inter-generic analysis. The total number of characters and the number of parsimony informative characters were calculated with PAUP . The dN/dS ratios were calculated under the MG94 model  as implemented in Hyphy (http://www.hyphy.org/) . The poleroviruses listed by ICTV , Pea enation mosaic virus (genus Enamovirus) and PnLV were screened for recombination signals. Putative recombinant genomes were searched using the RDP3 package (http://darwin.uvigo.es/rdp/rdp.html). It implements six recombinant detection programs: RDP, GENECONV, MaxChi, Chimera, Bootscan and Siscan . The default detection thresholds were applied. Five poleroviruses showing no signals of recombination were subsequently selected: Beet chlorosis virus (BchV), Beet mild yellowing virus (BMYV), Potato leaf roll virus (PLRV), CYDV-RPS and CYDV-RPV (Table 2). Similarly, the RdRP sequences of two luteoviruses were chosen: BYDV-PAS and BYDV-MAV.
The best-fitting nucleotide substitution model was evaluated by hierarchical likelihood ratio testing , as implemented in HyPhy . The best-fitting model was the HKY model with gamma rate heterogeneity. The dates of isolation of the virus species were considered as contemporaneous as they differed by a few years only, whereas our study dealt with inter-specific divergence times ranging from hundreds to thousands of years. The maximum clade credibility tree was reconstructed by Bayesian inference under the relaxed molecular clock models as implemented in BEAST. A Yule speciation process was selected as a tree prior. The distribution of the estimates of the TMRCA of RYMV was subsequently used as the prior of the RYMV node for upward calibration of the nodes of the trees. The HPD intervals of the TMRCA of sobemoviruses and related viruses subsequently summarized both the uncertainties of the phylogenetic signal and of the prior (the RYMV age). A uniform distribution with bounds of 5×10−5 and 5×10−3 nt/site/year was applied as the prior of the uncorrelated lognormal relaxed clock mean. A similar prior was applied for the Yule speciation process birth rate. A uniform distribution with bounds of 0.2 and 5 was applied as the prior of the gamma shape parameter. A Jeffrey prior with initial value of 1 was applied for the HYK transition-transversion parameter.
We thank J.M. Thresh, A.L. Haenni, André Fargette, and three anonymous reviewers for constructive criticisms of the manuscript; J. Berthaud, C. Brugidou, F. Fabre, A. Ghesquière, J.F. Guegan, B. Lafay, B. Moury, J.C. Pintaud, G. Serpantié, and members of the recently established Plant Virus Ecology Network for helpful discussions; and R.A.C. Jones and G.I. Dwyer for supplying dates of isolation of SCMoV.
Conceived and designed the experiments: DF APG DS SL EH OT GK. Analyzed the data: DF APG DS SL EH OT GK. Wrote the paper: DF. Advised on epidemiological and ecological aspects of the study: APG DS SL.
- 1. Duffy S, Shackelton LA, Holmes EC (2008) Rates of evolutionary change in viruses: patterns and determinants. Nat Rev Genet 9: 267–276.
- 2. Elena SF, Agudelo-Romero P, Carrasco P, Codoner FM, Martin S, et al. (2008) Experimental evolution of plant viruses. Heredity 100: 478–483.
- 3. Roossinck MJ (2005) Symbiosis versus competition in plant virus evolution. Nature Rev 3: 917–924.
- 4. Lovisolo O, Hull R, Rösler O (2003) Coevolution of viruses with hosts and vectors and possible paleontology. Adv Virus Res 62: 325–379.
- 5. Garcia-Arenal F, Fraile A, Malpica JM (2001) Variability and genetic structure of plant virus populations. Annu Rev Phytopathol 39: 157–186.
- 6. Block J, Mackensie A, Guy P, Gibbs A (1987) Nucleotide sequence comparisons of turnip yellow mosaic virus isolates from Australia and Europe. Arch Virol 97: 283–295.
- 7. Fraile A, Escriu F, Aranda MA, Malpica JM, Gibbs AJ, Garcia-Arenal F (1997) A century of tobamovirus evolution in an Australian population of Nicotiana glauca. J Virol 71: 8316–8320.
- 8. Gibbs AJ, Keese PL, Gibbs MJ, Garcia-Arenal F (1999) Plant virus evolution: Past, present and future. In: Domingo E, Webster R, Holland J, editors. Origin and evolution of viruses. London: Academic Press. pp. 263–285.
- 9. Duffy S, Holmes EC (2008) Phylogenetic evidence for rapid rates of molecular evolution in the single-stranded DNA begomovirus Tomato yellow leaf curl virus (TYLCV). J Virol 82: 957–965.
- 10. Fargette D, Pinel A, Rakotomalala M, Sangu E, Traoré O, et al. (2008) Rice yellow mottle virus, an RNA plant virus, evolves as rapidly as most RNA animal viruses. J Virol 82: 3584–3589.
- 11. Simmons HE, Holmes EC, Stephenson AG (2008) Rapid evolutionary dynamics of zucchini yellow mosaic virus. J Gen Viro 89: 1081–1085.
- 12. Holmes EC (2003) Molecular clock and the puzzle of RNA virus origins. J Virol 77: 3893–3897.
- 13. Welch JJ, Bromham L (2005) Molecular dating when rates vary. Trends Ecol Evol 6: 320–327.
- 14. Drummond AJ, Ho SY, Phillips MJ, Rambaut A (2006) Relaxed phylogenetics and dating with confidence. PLoS Biol 4: e88. doi:10.1371/journal.pbio.0040088.
- 15. Kouassi N, N'Guessan P, Albar L, Fauquet C, Brugidou C (2005) Distribution and characterization of Rice yellow mottle virus: a threat to African farmers. Plant Dis 89: 124–133.
- 16. Chare ER, Holmes EC (2006) A phylogenetic survey of recombination frequency in plant RNA viruses. Arch Virol 15: 933–946.
- 17. Fargette D, Pinel A, Abubakar Z, Traoré O, Brugidou C, et al. (2004) Inferring the evolutionary history of Rice yellow mottle virus from genomic, phylogenetic and phylogeographic studies. J Virol 78: 3252–3261.
- 18. Hull R, Fargette D (2005) Sobemovirus. In: Fauquet CM, Mayo MA, Maniloff J, Desselberger U, Ball LA, editors. Virus taxonomy. Classification and nomenclature of viruses. Eight report of the International Committee on Taxonomy of Viruses. Amsterdam: Elsevier/Academic Press. pp. 885–890.
- 19. Balke I, Resevica G, Zeltins A (2007) The Ryegrass mottle virus genome codes for a sobemovirus 3C-like serine protease and RNA-dependent RNA polymerase translated via –1 ribosomal frameshifting. Virus Genes 35: 395–398.
- 20. Meir M, Truve E (2006) Sobemoviruses possess a common CfMV-like genomic organization. Arch Virol 152: 635–640.
- 21. Lokesh GL, Gopinath K, Satheshkhumar PS, Savithri HS (2001) Complete nucleotide sequence of Sesbania mosaic virus: a new virus species of the genus Sobemovirus. Arch Virol 146: 209–223.
- 22. Meer M, Truve E (2006) An attempt to identify recombinants between two sobemoviruses in doubly infected oat plants. Environ Biosaf Res 5: 47–56.
- 23. D'Arcy CJ, Dommier LL (2005) Luteoviridae. In: Fauquet CM, Mayo MA, Maniloff J, Desselberger U, Ball LA, editors. Virus taxonomy. Classification and nomenclature of viruses. Eight report of the International Committee on Taxonomy of Viruses. Amsterdam: Elsevier/Academic Press. pp. 891–900.
- 24. Siepen M, Pohl JO, Koo BJ, Wege C, Jeske H (2005) Poinsettia latent virus is not a cryptic virus, but a natural polerovirus-sobemovirus hybrid. Virology 336: 240–250.
- 25. Gibbs M (1995) The luteovirus supergroup: rampant recombination and persistent partnerships. In: Gibbs AJ, Calisher CH, Garcia-Arenal F, editors. Molecular basis of virus evolution. Cambridge: Cambridge University Press. pp. 351–368.
- 26. Miller WA, Liu S, Beckett R (2002) Barley yellow dwarf virus: Luteoviridae or Tombusviridae? Mol Plant Pathol 3: 177–183.
- 27. Wright P, Revill P (2005) Family Barnaviridae. In: Fauquet CM, Mayo MA, Maniloff J, Desselberger U, Ball LA, editors. Classification and nomenclature of viruses. Eight report of the International Committee on Taxonomy of Viruses. Amsterdam: Elsevier/Academic Press. pp. 1125–1128.
- 28. Dwyer GI, Njeru R, Williamson S, Fosu-Nyarko J, Hopkins R, et al. (2003) The complete nucleotide sequence of Subterranean clover mottle virus. Arch Virol 148: 2237–2247.
- 29. Wren JD, Roossinck MJ, Nelson RS, Scheets K, Palmer MW, et al. (2006) Plant virus biodiversity and ecology. PLoS Biol 4: e80. doi:10.1371/journal.pbio.0040080.
- 30. Templeton AR (2006) Population genetics and microevolutionary theory. Hoboken: Wiley-Liss.
- 31. Renner SS (2005) Relaxed molecular clocks for dating historical plant dispersal events. Trends Plant Sci 10: 550–558.
- 32. Jones R, Fosu-Nyarko J, Jones M, Dwyer G (2001) Subterranean clover mottle virus. AAB Description of Plant Viruses, No. 387. www.dpvweb.net/dpv.
- 33. Abubakar Z, Ali F, Pinel A, Traoré O, N'Guessan P, et al. (2003) Phylogeography of Rice yellow mottle virus in Africa. J Gen Virol 84: 733–743.
- 34. Bakker W (1974) Characterization and ecological aspects of rice yellow mottle virus in Kenya. Agricultural research report no. 829. Wageningen Agricultural University, Wageningen. The Netherlands.
- 35. Fauquet C, Thouvenel JC (1977) Isolation of the rice yellow mottle virus in Ivory Coast. Plant Dis Reporter 61: 443–446.
- 36. Chang TT (1976) The origin, evolution, cultivation, dissemination and diversification of Asian and African rices. Euphytica 25: 425–441.
- 37. Porteres R (1950) Vieilles agricultures de l'Afrique intertropicale. Agron Afr 9: 489–507.
- 38. Diamond J (2002) Evolution, consequences and future of plant and animal domestication. Nature 418: 700–707.
- 39. Mazoyer M, Roudart L (2006) A history of world agriculture; from the Neolithic age to the current crisis. New-York: New York University Press.
- 40. Doebley JF, Gaut BS, Smith BD (2006) The molecular genetics of crop domestication. Cell 127: 1309–1321.
- 41. Harlan JR (1992) Crops and Man. 2nd edition. American Society of Agronomy.
- 42. Buddenhagen IW (1977) Resistance and vulnerability of tropical crops in relation to their evolution and breeding. Ann New York Acad Sci 287: 309–326.
- 43. Thresh JM (1982) Cropping practices and virus spread. Annu Rev Phytopathol 20: 193–218.
- 44. Barret R, Kuzawa CW, McDade T, Armelagos GJ (1998) Emerging and re-emerging infectious diseases: the third epidemiologic transition. Annu Rev Anthropol 27: 247–271.
- 45. Palumbi SR (2001) Humans as the world's greatest evolutionary force. Science 293: 1786–1790.
- 46. Van Regenmortel MHV (2007) Virus species and virus identification: past and current controversies. Infect Genet Evol 7: 133–144.
- 47. Thompson JD, Higgins DJ, Gibson TJ (1994) CLUSTAL W. Improving the sensitivity of the progressive multiple sequence alignment through sequence weighting, position gap penalties and weight matrix choice. Nucleic Acids Res 22: 4673–4680.
- 48. Drummond AJ, Rambault A (2007) BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol 7: 214. doi :biomedcentral.com/1471-2148/7/214.
- 49. Drummond AJ, Rambaut A, Shapiro B, Pybus OG (2005) Bayesian coalescent inference of past population dynamics from molecular sequences. Mol Biol Evol 22: 1185–1192.
- 50. Suchard MA, Weiss RE, Sinsheimer JS (2001) Bayesian selection of continuous time Markov chain evolutionary models. Mol Biol Evol 18: 1001–1013.
- 51. Swofford DL (2000) PAUP: phylogenetic analysis using parsimony, version 4. Sunderland: Sinauer Associates.
- 52. Yang Z (2006) Computational molecular evolution. Oxford: Oxford University Press.
- 53. Kosakovsky Pond SLK, Frost SDW, Muse SV (2005) HyPhy: hypothesis testing using phylogenies. Bioinformatics 21: 676–679.
- 54. Martin DP, Williamson C, Posada D (2005) RDP2: recombination detection and analysis from sequence alignments. Bioinformatics 21: 260–262.
- 55. Posada D, Crandall K (1998) MODELTEST: testing the model of DNA substitution. Bioinformatics 14: 817–818.
- 56. Sérémé D, Lacombe S, Konaté M, Pinel-Galzi A, Traoré E, et al. (2008) Biological and molecular characterization of a putative new sobemovirus infecting Imperata cylindrica and maize in Africa. Arch Virol. In press.