| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Molecular and Cell Biology, University of Cape Town, Private Bag, Rondebosch 7701, South Africa
2 Institute of Infectious Disease and Molecular Medicine, University of Cape Town, Anzio Rd, Observatory 7925, South Africa
3 Department of Crop Science, Faculty of Agriculture, Makerere University, PO Box 7062, Kampala, Uganda
4 Electron Microscope Unit, University of Cape Town, Private Bag, Rondebosch 7701, South Africa
Correspondence
Arvind Varsani
arvind.varsani{at}uct.ac.za
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The GenBank/EMBL/DDBJ accession numbers for sequences of MSV isolates determined in this study are EF547063–EF547124 (also shown individually in Table 1).
Supplementary figures, Excel tables and alignment files are available with the online version of this paper.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
There is good evidence that, throughout Africa, MSD is caused by a group of viruses that all share >97 % genome-wide sequence identity with one another (Briddon et al., 1994
; Martin et al., 2001). Although diversity between these so-called MSV-A viruses is low, there is strong phylogenetic support for their classification into at least six lineages or subtypes (named MSV-A1–MSV-A6; Martin et al., 2001). There is also some evidence of variation in the subtype composition of MSV-A populations in different parts of Africa (Martin et al., 2001).
Importantly, low continent-wide MSV diversity should vastly simplify the development of resistant maize genotypes. The situation is quite different from that experienced by biotechnologists and breeders attempting to develop cassava with resistance to the related cassava mosaic disease (CMD)-causing geminiviruses. At least seven distinct/tentative species of cassava-infecting geminivirus (CGV), each sharing <90 % genome-wide sequence identity with the others, cause CMD in sub-Saharan Africa (Fauquet et al., 2003; Ndunguru et al., 2005a; Bull et al., 2006). Additionally, circulating inter-species recombinants of these viruses are common (Ndunguru et al., 2005a; Bull et al., 2006).
Although inter-species MSV recombinants have been detected, the scale of recombination, in terms of both the size of sequence tracts transferred and the genetic distances between parental viruses, appears to be much smaller than that observed in CGVs (Martin et al., 2001). It is possible that the apparently striking evolutionary and demographic differences between MSV and CGVs are due to CGVs having been sampled far more thoroughly. Another explanation is that, despite many common biological features (insect transmission, largely overlapping geographical ranges and similar molecular biology), differences in the epidemiological and population genetic characteristics of the two groups are responsible for the apparently large differences in their evolutionary trajectories.
In this report, we describe the fine-scale population structure of MSV isolates sampled from Ugandan maize in 2005, and compare this with that of CGVs sampled from Kenya between 2001 and 2002. We use analysis of MSV PCR–restriction fragment-length polymorphism (PCR-RFLP) and full-genome sequence data, first to identify the major circulating MSV variants and then to determine the distribution of these in different regions of Uganda. In an attempt to identify the underlying factors responsible for vast differences in patterns of MSV and CGV recombination, we construct matched population-scale datasets and use these to infer and compare various estimates of evolutionary and population genetic parameters of MSV and two CGV species.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
). The fields were approximately 10 km or more apart and contained plants ranging from 1 to 3.5 months old. Global positioning system (GPS) coordinates were captured for each sample (see Supplementary Table S1, available in JGV Online).
DNA extractions.
Total DNA was extracted from 305 fresh MSV-infected leaf samples using a modified CTAB method (Kiprop et al., 2002; Owor et al., 2007). In addition, DNA was extracted from 155 leaf-pressed samples on FTA cards as described previously (Ndunguru et al., 2005b; Owor et al., 2007).
PCR-RFLP analysis.
MSV DNA was detected in total plant DNA extracts by PCR using a pair of degenerate primers described previously (Willment et al., 2001). Each PCR amplification product was digested with RsaI, HpaII, HaeIII, CfoI, HindIII, BamHI and Sau3AI as described previously (Willment et al., 2001). In addition, in silico RFLP analysis was performed on sequenced genomes. Derived in vitro and in silico restriction patterns were compared and categorized according to a comparative panel of all currently published MSV restriction-pattern types (Martin et al., 2001; Willment et al., 2001; Owor et al., 2007), with newly discovered patterns being added to the comparative panel of pattern types (see Supplementary Fig. S1, available in JGV Online).
Cloning and sequencing of full-genome sequences.
From the 155 fields sampled, 30 fields were chosen at random using a random-number generator and, for each chosen field, one sample with a positive PCR was selected for full-genome sequencing. An additional 32 samples were chosen for full-genome sequencing based on their unique RFLP profiles. Viral samples were amplified using phi29 DNA polymerase (TempliPhi; GE Healthcare) as described previously (Owor et al., 2007). Briefly, the amplified concatemers were digested with either BamHI or SalI to yield approximately 2.7 kb, potentially full-length linearized viral genomes, which were gel-purified (Invisorb spin DNA extraction kit; Invitek) and cloned into pGEM-3Zf(+) (Promega). Both strands of cloned genomes were sequenced commercially (Macrogen Inc., Korea), using the primer set described previously (Owor et al., 2007). GenBank accession numbers are shown in Table 1.
|
). All available MSV and CGV genome sequences were obtained from GenBank. Sequence alignments were constructed using POA (Grasso & Lee, 2004) and edited both by eye and using the CLUSTAL_W-based (Thompson et al., 1994) sub-sequence realignment tool implemented in MEGA. Phylogenetic trees were constructed by the neighbour-joining (Jules–Cantor distances, 1000 bootstrap replicates) and maximum-likelihood (Hasegawa–Kishino–Yano model, transition/transversion ratio inferred from the data and 100 bootstrap replicates) methods implemented in MEGA and PHYML (Guindon & Gascuel, 2003), respectively. Recombination was analysed using the RDP (Martin & Rybicki, 2000), GENECONV (Padidam et al., 1999), BOOTSCAN (Martin et al., 2005a), MAXCHI (Maynard Smith, 1992), CHIMAERA (Posada & Crandall, 2001), SISCAN (Gibbs et al., 2000) and LARD (Holmes et al., 1999) methods, implemented in RDP3 (Martin et al., 2005c). Default settings were used throughout and only potential recombination events detected by two or more of the above methods coupled with phylogenetic evidence of recombination were considered significant. Also, the severity of Bonferroni correction during detection was minimized by only searching for recombination signals in a single sequence within groups of sequences sharing >99.7 % sequence identity. Composite likelihood estimates (CLEs) of population-scaled recombination rates and estimates of population-scaled mutation rates were inferred using the PAIRWISE component of LDHAT [finite-sites version of the Watterson 
inferred from the data, a minimum minor-allele frequency of either 0.05 (for datasets containing 28 and 68 sequences) or 0.1 (for a dataset containing 14 sequences), a grid size of 100 and a maximum 
of 100, gene-conversion model of recombination with an average tract length of 1000 nt; McVean et al., 2002]. Site-to-site variation in the CLE of the population-scaled recombination rates was assessed using the INTERVAL component of LDHAT [using pre-computed likelihoods for population-scaled mutation rate=0.01, a minimum minor-allele frequency cutoff of 0.05 or 0.01 (decided as described above), a gene-conversion model with an average tract length of 1000 nt, block penalty of 10, starting of 5, 107 Markov chain Monte Carlo updates with sampling every 2000 updates and the first 500 samples discarded; McVean et al., 2004]. For variable recombination-rate analysis with LDHAT, we avoided analysis inaccuracies at the edges of alignments by simulating circular genome sequences. Alignments were constructed from tandem repeats of full-genome sequences, with the origins of virion-strand replication situated at positions 25 and 75 % along the alignments. Results of analyses with these alignments were then processed to exclude the first 25 % and last 25 % of point recombination-rate estimates. Departures from expectations of neutral evolution in full-genome sequences were inferred by using Tajima's D (Tajima, 1989) and Fu and Li's D* (Fu & Li, 1993) statistics, calculated by using the CONVERT component of LDHAT with no minor-allele frequency cutoff. The significance of these statistics contingent on inferred population-scaled mutation and recombination rates was tested by coalescent simulation using DnaSP (Rozas & Rozas, 1999). |
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Most of the RFLPs observed have been reported previously for MSV-A isolates (Willment et al., 2001; Owor et al., 2007). Whilst no new patterns were found for HindIII, CfoI, BamHI and HpaII, we observed one new pattern each for RsaI and HaeIII, and five new patterns for Sau3AI (see Supplementary Fig. S1, available in JGV Online). Of the 387 samples analysed, 6.20 % (24/387) had evidence of mixtures of previously described RFLP patterns. Although these samples possibly represented mixed infections, only 14 of them were included in further analyses, as it was possible to identify unambiguously the RFLP profiles of the two viruses present in these (this was achievable for these 14 samples because only one of the seven restriction-enzyme digests used per sample yielded evidence of multiple restriction patterns). In addition to in vitro RFLP analysis, cloning and sequencing of 62 full-length genomes (see below) allowed us to perform in silico RFLP analysis of these sequences. This led to the identification of seven and one previously unpublished RFLP patterns for HpaII and HindIII, respectively. These patterns have subtle differences that would be difficult to distinguish on an agarose gel, but are nonetheless different variants (see Supplementary Fig. S1, available in JGV Online). Therefore, by using a combination of in vitro and in silico RFLPs, we observed new patterns for five of the seven enzymes.
In total, 49 different RFLP-pattern combinations were observed (see Supplementary Table S1, available in JGV Online). By the convention described previously (Willment et al., 2001), the four most prevalent RFLP-pattern groups were AABBDAD, BCABBAA, AAABDAD and ABABBAA, respectively representing 59.34, 11.00, 3.07 and 2.56 % of the 391 isolates in Supplementary Table S1 (available in JGV Online).
Recombination and phylogenetic analyses of full-genome sequences
To analyse the range of Ugandan MSV diversity more accurately, we cloned and completely sequenced 62 MSV genomes. Amongst these were isolates representing 45 of the 49 unique RFLP-pattern combinations that we observed. These were aligned together with six previously described Ugandan MSV genome sequences (Owor et al., 2007), 21 MSV genome sequences sampled elsewhere in Africa, two Panicum streak virus sequences, a sugarcane streak virus sequence, a sugarcane streak Reunion virus sequence and a sugarcane streak Egypt virus sequence (94 sequences in total). Preliminary phylogenetic analysis indicated that, as expected, all of the Ugandan MSV sequences were of the maize-adapted MSV-A type. More specifically, all clustered with either the MSV-A1 or MSV-A5 subtypes (Fig. 1a) identified previously in a 1999 survey of African MSV diversity (Martin et al., 2001).
|
), we analysed the above alignment for evidence of recombination between the 84 MSV-A viruses and the other African streak viruses (including both other MSV types and non-MSV African streak viruses). Amongst all of the MSV-A sequences, we detected only two previously reported recombination events involving MSV-A viruses as recipients of sequence from non-MSV-A sources. Both of these recombination events, described by Martin et al. (2001), involve exchanges of small fragments (approx. 100 bp) between MSV-A- and MSV-B-like viruses to yield MSV-VM in one case and, in the other, the ancestor of MSV-SA, MSV-Kom and MSV-VM. Clearly, therefore, neither inter-type nor inter-species recombination is currently a major factor in Ugandan MSV-A diversification. This does not, however, discount the possibility that intra-MSV-A recombination might be an important feature of MSV-A evolution. To increase statistical power during the analysis of intra-MSV-A recombination, both the MSV-B-like portions of the MSV-Kom, -VM and -SA genomes and all of the non-MSV-A sequences were removed from the MSV alignment before rescreening for recombination.
We detected ten potential intra-MSV-A recombination events (Bonferroni-corrected P value <0.05 for at least two different recombination-detection methods coupled with phylogenetic evidence of sequence exchange), nine of which have apparently occurred in sequences ancestral to the Ugandan MSV-A sequences determined in this study (Figs 1
, 2). Six of these nine events (events 1, 3, 4, 7, 8 and 10 in Fig. 2) apparently involved sequence exchanges between MSV-A1 and viruses related most closely to those in the MSV-A2, -A3 and -A4 subtype groups. Due to generally low sequence diversity amongst African MSV-A sequences and the comparatively poor sampling of all MSV-A subtypes other than MSV-A1, it was not possible to identify the exact origins of the recombinant regions convincingly for three of the five inter-subtype recombination events (events 4, 7 and 10 in Fig. 2).
|
]. This reduction in tree length is exactly what would be expected following removal of the most significant recombination signals within the dataset (Schierup & Hein, 2000).
Importantly, the sequences for MSV-MakD and MSV-MatC, formerly classified as belonging to the MSV-A5 subtype (Martin et al., 2001), cluster clearly with the MSV-A1 subtype in both of the trees. Although both maximum-likelihood and neighbour-joining reconstructions with only the MSV-A sequences used in the study by Martin et al. (2001) confirmed the tree topology determined in that study (data not shown), when using our much larger MSV-A sequence dataset, both tree-construction methods indicate that the MSV-A5 subtype is in fact a sublineage of the MSV-A1 subtype. Although the reason for this discrepancy is unclear, MSV-MakD and MSV-MatC are both recombinants containing large tracts of MSV-A1-like sequence [identified here and by Martin et al. (2001)], which would be expected to compromise the correct placement of these sequences in phylogenetic trees (Schierup & Hein, 2000; Posada & Crandall, 2002; Awadalla, 2003). Owing to the large proportion of MSV-A1-like sequences within these two isolates and their close Ugandan relatives, we have opted to reclassify MSV-A5 as a recombinant sublineage of MSV-A1.
Distribution of major MSV genotypes in Uganda
Based on detected recombination patterns, the Ugandan MSV sequences were classified into eight haplotypes, named MSV-A1UgI to MSV-A1UgVIII (Fig. 1b, c), with MSV-A1UgV representing the MSV-A1 sequences that are not detectably recombinant. We should specify here that our haplotype-classification scheme is not intended as a serious taxonomic proposal. It is simply a convenient and evolutionarily relevant way of splitting the Ugandan virus isolates into distinguishable groups.
By using RFLP data for the remaining 321 Ugandan MSV samples [197 from this study and 124 from Owor et al. (2007)] that were not analysed by full-genome sequencing, it was possible to classify each of these into one of the eight haplotype groupings.
The sampling area in Uganda was split into seven zones and the relative proportions of the different haplotypes were determined for each of these zones (Fig. 3a). Importantly, there was no significant difference in the population frequencies of different haplotypes across the seven zones (P=0.2964; 
2 test with eight haplotypesxseven sampling zones), indicating that, generally, the diversity of samples collected from any one of the zones was not significantly unrepresentative of country-wide MSV diversity.
|
2 test with 15 three-sample haplotype combinationsxseven sampling zones). This indicates that MSV diversity observed within sampling areas of a few hundred square metres is also generally not significantly unrepresentative of that observed country-wide.
Showing that the distribution of MSV diversity at these different sampling scales is not significantly different is not, however, the same thing as showing that the distribution of diversity at the different scales is significantly similar. More careful examination of the samples collected in individual farms revealed five instances where the haplotype combinations observed were reasonably improbable (P<0.05; 2x2 2 test). Although none of these comparisons were significant following Bonferroni correction of P values (to account for the multiple tests made), we could not discount the possibility that including all three samples from each farm might introduce a sampling bias into our analysis of country-wide Ugandan MSV population structure. Therefore, only one example of each haplotype sampled in each location was considered in subsequent analyses. Whilst this method biased estimates of population representation slightly against the two most common haplotypes, it enabled more sensitive analysis of the distributions of the six rarer haplotypes. This selection process did not, however, grossly distort the overall representation of the respective haplotypes, as illustrated in Supplementary Fig. S2 (available in JGV Online).
The intra-MSV-A1 recombinant haplotype MSV-A1UgIII (Fig. 1) comprised 50 % or more of the MSV samples in all seven zones (Fig. 3d) and is clearly the dominant MSV variant throughout Uganda. The MSV-A1UgV haplotype, containing all of the MSV-A1 sequences that are not obviously recombinant, is the only other haplotype that was detected in all seven sampling zones (Fig. 3e). All of the other haplotypes were either absent or present below detectable levels in two or more of the zones. As most of these were present at close to the detection limits in the zones where they were observed, it is possible that they are all present throughout the country.
There is some indication of slight variation in MSV demography in different zones. Whilst this is particularly true for some of the rarer haplotypes, such as MSV-A1UgVI, MSV-A1UgVII and MSV-A1UgVIII (Fig. 3f–h), which display >5-fold variations in population representation in different zones, there is also evidence of variation in relative population representation of the more common haplotype MSV-A1UgV. This haplotype is at its highest prevalence in four of the five eastern zones and at its lowest prevalence in the two western zones. Statistically significant deviation from country-wide population frequencies was, however, only evident for MSV-A1UgVI in zones 5 and 6 (P=0.015 and 0.042, respectively; Bonferroni-corrected 2x2 2 test; Fig. 3f).
Despite the possibility of slight geographical variations in haplotype frequencies across Uganda, the fact that MSV haplotype distributions do not differ substantially over sampling scales ranging from 0.1 to 100 000 km2 suggests strongly that MSV population structure (in Uganda at least) is highly homogeneous. This in turn implies that there are no substantial impediments to the movement of viruses throughout the country.
Comparison of the population genetic characteristics of MSV and African CGVs
To compare the MSV diversity and recombination data with those of CGVs, the only other substantially sampled African geminivirus group, we obtained all 118 African cassava-infecting begomovirus DNA-A sequences available in GenBank on 30 November 2006, aligned these using POA, edited and realigned subsections of the alignments in MEGA and identified 23 major inter-species and four intra-species recombination events using RDP3 (see the supplementary files ‘Cassava.rdp’ and ‘Cassava.csv’, available in JGV Online, for detailed results of this analysis).
Relative to the MSV dataset, the CGV sequences contain more evidence of recombination involving larger fragments of sequence between more distantly related parental viruses (Fig. 4). Unsurprisingly, these quite striking differences have been noted elsewhere (Padidam et al., 1999; Martin et al., 2001; Schnippenkoetter et al., 2001). However, the cause(s) of these differences remain unexplored. We propose that there are three main reasons that MSV and CGVs might have such different patterns of recombination: (i) the biochemical recombination rate in MSV may be significantly lower than that found in CGV species; (ii) although mixed infections, a prerequisite for detectable recombination, have been observed in both MSV and CGVs (this study; Willment et al., 2001; Vanitharani et al., 2004), they may be much more common amongst CGVs than they are amongst MSVs; (iii) whilst the CGV species are all cassava-adapted, only one MSV strain is maize-adapted, and differences seen in the extents and prevalence of recombination in CGVs and MSV may therefore be the result of purifying selection eliminating greater proportions of MSV recombinants, particularly when these contain large tracts of sequence from viruses that are not maize-adapted (Martin & Rybicki, 2002; Martin et al., 2005b).
|
.
|
. It is important to make a distinction here between ‘population-scaled’ and ‘biochemical’ mutation and recombination rates. Whilst biochemical mutation and recombination rates are the rates at which nucleotide substitutions and strand-invasion events leading to recombination occur over time (such as per replication cycle or per year), the population-scaled estimates of these do not have an easily definable unit of measurement. The population-scaled mutation and recombination rates for haploid organisms evolving exclusively through genetic drift can be expressed as 2Neu and 2Ner, respectively, where Ne is the effective population size, u is the biochemical rate at which neutral mutations occur and r is the biochemical rate at which recombination events that have no fitness effects occur. Note that whilst these population-scaled values are not actual estimates of biochemical mutation and recombination rates, both expressions contain the value 2Ne and therefore their ratio should be the same as the biochemical recombination- and mutation-rate ratio (the 2Ne part of the equations and all accessory uncertainties associated with estimating effective population sizes of viruses are cancelled out when one considers only the ratios of population-scaled mutation and recombination rates). For this relationship to hold, however, it is important that all observable mutation and recombination events considered during calculation of the population-scaled mutation- and recombination-rate estimates are selectively neutral. As a direct correlation between the relative neutrality of recombination events and the relatedness of parental sequences has been observed for MSV (Martin & Rybicki, 2002; Martin et al., 2005b), we chose to use only very closely related groups of sequences to derive recombination- and mutation-rate estimates for both MSV and CGVs. Nevertheless, the proportions of mutation and recombination events that are neutral in these datasets are unknown and the values of the population-scaled mutation and recombination rates should be interpreted with caution.
Importantly, neither the recombination rates nor recombination-rate : mutation-rate ratios of the different datasets are substantially different from one another (Table 2). It is important that the value of this ratio for the MSV dataset falls between that determined for the two CGV datasets. Assuming that the biochemical mutation rates of MSV and CGVs are not substantially different, these data imply that the biochemical recombination rates of the different groups of viruses are also not substantially different. This indicates, therefore, that the striking differences in the types of MSV and CGV recombination events detected in nature are probably not due to large differences in the biochemical recombination rates of these viruses.
Despite their apparently similar genome-wide recombination rates, we suspected that there might be differences in regional recombination-rate variation within the MSV genomes and the CGV DNA-A components. CLEs of regional variation in the population-scaled recombination rates of all three datasets were, however, surprisingly similar (Fig. 5). In all three populations, it seems that recombination rates are significantly higher in genomic regions encoding complementary-sense genes than they are in regions encoding virion-sense genes. A potentially important clue as to the mechanistic cause of this recombination-rate imbalance may be that, within 400 nt 3' of the virion-strand replication origin, recombination-rate estimates are at their lowest in all three datasets. Apart from indicating that similar mechanistic processes are possibly responsible for recombination-rate variation across MSV and CGV genomes, this result implies that these mechanistic processes may be features of the virion-strand replication and/or complementary-strand transcription systems.
|
). The values of these statistics were also marginally significant for the EACMKV dataset (P=0.09 for Tajima's D and P=0.06 for Fu and Li's D*). The negative values of the D and D* statistics indicate that, given an expectation of neutrality, there is an excess of low-frequency nucleotide polymorphisms in viruses sampled from these populations. Such departures from neutrality might be caused by a range of population phenomena including, for example, sporadic cycles of population collapse and expansion, such as those characteristic of both MSV and CGV epidemiology. It is important that both the MSV and CGV genome-sequence samples bear similar marks of population genetic processes, as this indicates that differences in the patterns of recombination events detected between the two groups are possibly not due to fundamentally different evolutionary forces acting on the viruses. It is plausible that, with respect to the evolutionary benefits of recombination, the primary differences between MSV and CGVs is that the diversity of high-fitness host-adapted genome constituents available for recombinational exchange is far greater for CGVs than it is for MSV.
Concluding remarks
Whilst our survey of Ugandan maize-infecting MSV-A isolates has revealed that the vast majority of MSD infections in the country are caused by a group of very closely related viruses, we have demonstrated that this low diversity does not necessarily equate to genetic uniformity. We have found that there is substantial evidence of genetic exchange between viruses within the MSV-A group and that a recombinant is in fact the most prevalent MSV-A variant within the country. By using recombination patterns as a means of haplotyping MSV variants, we determined that the diversity of Ugandan MSVs is remarkably constant over a wide range of sampling scales, such that viral diversity within individual farms is not significantly different from that detected across the entire country. The hypothesis that recombination is an important feature of geminivirus evolution is not new, but we have demonstrated that its characteristics are strikingly different in MSV and the related CGVs. We provide some evidence that the underlying cause of these differences is probably not that CGVs have a higher biochemical recombination rate than MSV, but rather that CGVs have recombinational access to a far greater diversity of appropriately host-adapted genome constituents. The data that we have provided will be useful in future studies involving either longitudinal monitoring of Ugandan MSV population turnover or comparative genetic analyses of large MSV population samples from different parts of the African continent. Agroinfectious constructs containing the virus genomes that we have cloned will also be useful for controlled challenges of new MSV-resistant maize genotypes currently being developed and tested for release in Uganda.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Briddon, R. W., Lunness, P., Chamberlin, L. C. & Markham, P. G. (1994). Analysis of the genetic variability of maize streak virus. Virus Genes 9, 93–100.[CrossRef][Medline]
Bull, S. E., Briddon, R. W., Sserubombwe, W. S., Ngugi, K., Markham, P. G. & Stanley, J. (2006). Genetic diversity and phylogeography of cassava mosaic viruses in Kenya. J Gen Virol 87, 3053–3065.
Fauquet, C. M., Bisaro, D. M., Briddon, R. W., Brown, J. K., Harrison, B. D., Rybicki, E. P., Stenger, D. C. & Stanley, J. (2003). Revision of taxonomic criteria for species demarcation in the family Geminiviridae, and an updated list of begomovirus species. Arch Virol 148, 405–421.[CrossRef][Medline]
Fu, Y. X. & Li, W. H. (1993). Statistical tests of neutrality of mutations. Genetics 133, 693–709.[Abstract]
Gibbs, M. J., Armstrong, J. S. & Gibbs, A. J. (2000). Sister-scanning: a Monte Carlo procedure for assessing signals in recombinant sequences. Bioinformatics 16, 573–582.
Grasso, C. & Lee, C. (2004). Combining partial order alignment and progressive multiple sequence alignment increases alignment speed and scalability to very large alignment problems. Bioinformatics 20, 1546–1556.
Guindon, S. & Gascuel, O. (2003). A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol 52, 696–704.[CrossRef][Medline]
Holmes, E. C., Worobey, M. & Rambaut, A. (1999). Phylogenetic evidence for recombination in dengue virus. Mol Biol Evol 16, 405–409.[Abstract]
Kiprop, E. K., Baudoin, J. P., Mwang'ombe, A. W., Kimani, P. M., Mergeai, G. & Maquet, A. (2002). Characterization of Kenyan isolates of Fusarium udum from pigeonpea [Cajanus cajan (L.) Millsp.] by cultural characteristics, aggressiveness and AFLP analysis. J Phytopathol 150, 517–525.[CrossRef]
Kumar, S., Tamura, K. & Nei, M. (2004). MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignment. Brief Bioinform 5, 150–163.
Martin, D. & Rybicki, E. (2000). RDP: detection of recombination amongst aligned sequences. Bioinformatics 16, 562–563.
Martin, D. P. & Rybicki, E. P. (2002). Investigation of maize streak virus pathogenicity determinants using chimaeric genomes. Virology 300, 180–188.[CrossRef][Medline]
Martin, D. P., Willment, J. A., Billharz, R., Velders, R., Odhiambo, B., Njuguna, J., James, D. & Rybicki, E. P. (2001). Sequence diversity and virulence in Zea mays of maize streak virus isolates. Virology 288, 247–255.[CrossRef][Medline]
Martin, D. P., Posada, D., Crandall, K. A. & Williamson, C. (2005a). A modified bootscan algorithm for automated identification of recombinant sequences and recombination breakpoints. AIDS Res Hum Retroviruses 21, 98–102.[CrossRef][Medline]
Martin, D. P., van der Walt, E., Posada, D. & Rybicki, E. P. (2005b). The evolutionary value of recombination is constrained by genome modularity. PLoS Genet 1, e51[CrossRef][Medline]
Martin, D. P., Williamson, C. & Posada, D. (2005c). RDP2: recombination detection and analysis from sequence alignments. Bioinformatics 21, 260–262.
Maynard Smith, J. (1992). Analyzing the mosaic structure of genes. J Mol Evol 34, 126–129.[Medline]
McVean, G., Awadalla, P. & Fearnhead, P. (2002). A coalescent-based method for detecting and estimating recombination from gene sequences. Genetics 160, 1231–1241.
McVean, G. A. T., Myers, S. R., Hunt, S., Deloukas, P., Bentley, D. R. & Donnelly, P. (2004). The fine-scale structure of recombination rate variation in the human genome. Science 304, 581–584.
Ndunguru, J., Legg, J. P., Aveling, T. A., Thompson, G. & Fauquet, C. M. (2005a). Molecular biodiversity of cassava begomoviruses in Tanzania: evolution of cassava geminiviruses in Africa and evidence for East Africa being a center of diversity of cassava geminiviruses. Virol J 2, 21[CrossRef][Medline]
Ndunguru, J., Taylor, N. J., Yadav, J., Aly, H., Legg, J. P., Aveling, T., Thompson, G. & Fauquet, C. M. (2005b). Application of FTA technology for sampling, recovery and molecular characterization of viral pathogens and virus-derived transgenes from plant tissues. Virol J 2, 45[CrossRef][Medline]
Owor, B. E., Shepherd, D. N., Taylor, N. J., Edema, R., Monjane, A. L., Thomson, J. A., Martin, D. P. & Varsani, A. (2007). Successful application of FTA Classic Card technology and use of bacteriophage phi29 DNA polymerase for large-scale field sampling and cloning of complete maize streak virus genomes. J Virol Methods 140, 100–105.[CrossRef][Medline]
Padidam, M., Sawyer, S. & Fauquet, C. M. (1999). Possible emergence of new geminiviruses by frequent recombination. Virology 265, 218–225.[CrossRef][Medline]
Posada, D. & Crandall, K. A. (2001). Evaluation of methods for detecting recombination from DNA sequences: computer simulations. Proc Natl Acad Sci U S A 98, 13757–13762.
Posada, D. & Crandall, K. A. (2002). The effect of recombination on the accuracy of phylogeny estimation. J Mol Evol 54, 396–402.[Medline]
Rozas, J. & Rozas, R. (1999). DnaSP version 3: an integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics 15, 174–175.
Schierup, M. H. & Hein, J. (2000). Consequences of recombination on traditional phylogenetic analysis. Genetics 156, 879–891.
Schnippenkoetter, W. H., Martin, D. P., Willment, J. A. & Rybicki, E. P. (2001). Forced recombination between distinct strains of maize streak virus. J Gen Virol 82, 3081–3090.
Tajima, F. (1989). Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123, 585–595.
Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22, 4673–4680.
Vanitharani, R., Chellappan, P., Pita, J. S. & Fauquet, C. M. (2004). Differential roles of AC2 and AC4 of Cassava geminiviruses in mediating synergism and suppression of posttranscriptional gene silencing. J Virol 78, 9487–9498.
Willment, J. A., Martin, D. P. & Rybicki, E. P. (2001). Analysis of the diversity of African streak mastreviruses using PCR-generated RFLPs and partial sequence data. J Virol Methods 93, 75–87.[CrossRef][Medline]
Received 5 May 2007;
accepted 10 July 2007.
This article has been cited by other articles:
|
|
 |
|
  P. Lefeuvre, J.-M. Lett, A. Varsani, and D. P. Martin Widely Conserved Recombination Patterns among Single-Stranded DNA Viruses J. Virol., March 15, 2009; 83(6): 2697 - 2707. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. van der Walt, E. P. Rybicki, A. Varsani, J. E. Polston, R. Billharz, L. Donaldson, A. L. Monjane, and D. P. Martin Rapid host adaptation by extensive recombination J. Gen. Virol., March 1, 2009; 90(3): 734 - 746. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Varsani, D. N. Shepherd, A. L. Monjane, B. E. Owor, J. B. Erdmann, E. P. Rybicki, M. Peterschmitt, R. W. Briddon, P. G. Markham, S. Oluwafemi, et al. Recombination, decreased host specificity and increased mobility may have driven the emergence of maize streak virus as an agricultural pathogen J. Gen. Virol., September 1, 2008; 89(9): 2063 - 2074. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | MICROBIOLOGY | J GEN VIROL |
| J MED MICROBIOL | ALL SGM JOURNALS | |
