Widely Conserved Recombination Patterns among Single-Stranded DNA Viruses

The combinatorial nature of genetic recombination can potentiallyprovide organisms with immediate access to many more positionsin sequence space than can be reached by mutation alone. Recombinationfeatures particularly prominently in the evolution of a diverserange of viruses. Despite rapid progress having been made inthe characterization of discrete recombination events for manyspecies, little is currently known about either gross patternsof recombination across related virus families or the underlyingprocesses that determine genome-wide recombination breakpointdistributions observable in nature. It has been hypothesizedthat the networks of coevolved molecular interactions that definethe epistatic architectures of virus genomes might be damagedby recombination and therefore that selection strongly influencesobservable recombination patterns. For recombinants to thrivein nature, it is probably important that the portions of theirgenomes that they have inherited from different parents workwell together. Here we describe a comparative analysis of recombinationbreakpoint distributions within the genomes of diverse single-strandedDNA (ssDNA) virus families. We show that whereas nonrandom breakpointdistributions in ssDNA virus genomes are partially attributableto mechanistic aspects of the recombination process, there isalso a significant tendency for recombination breakpoints tofall either outside or on the peripheries of genes. In particular,we found significantly fewer recombination breakpoints withinstructural protein genes than within other gene types. Collectively,these results imply that natural selection acting against virusesexpressing recombinant proteins is a major determinant of nonrandomrecombination breakpoint distributions observable in most ssDNAvirus families.

INTRODUCTION

Top

INTRODUCTION

Genetic recombination is a ubiquitous biological process thatis both central to DNA repair pathways (10, 57) and an importantevolutionary mechanism. By generating novel combinations ofpreexisting nucleotide polymorphisms, recombination can potentiallyaccelerate evolution by increasing the population-wide geneticdiversity upon which adaptive selection relies. Recombinationcan paradoxically also prevent the progressive accumulationof harmful mutations within individual genomes (18, 35, 53).Whereas its ability to defend high-fitness genomes from mutationaldecay possibly underlies the evolutionary value of sexualityin higher organisms, in many microbial species where pseudosexualgenetic exchange is permissible among even highly divergentgenomes, recombination can enable access to evolutionary innovationsthat would otherwise be inaccessible by mutation alone.

Such interspecies recombination is fairly common in many virusfamilies (8, 17, 27, 44, 82). It is becoming clear, however,that as with mutation events, most recombination events betweendistantly related genomes are maladaptive (5, 13, 38, 50, 63,80). As genetic distances between parental genomes increase,so too does the probability of fitness defects in their recombinantoffspring (16, 51). The viability of recombinants is apparentlylargely dependent on how severely recombination disrupts coevolvedintragenome interaction networks (16, 32, 51). These networksinclude interacting nucleotide sequences that form secondarystructures, sequence-specific protein-DNA interactions, interproteininteractions, and amino acid-amino acid interactions withinprotein three-dimensional folds.

One virus family where such interaction networks appear to havea large impact on patterns of natural interspecies recombinationare the single-stranded DNA (ssDNA) geminiviruses. As with otherssDNA viruses, recombination is very common among the speciesof this family (62, 84). Partially conserved recombination hotand cold spots have been detected in different genera (39, 81)and are apparently caused by both differential mechanistic predispositionsof genome regions to recombination and natural selection disfavoringthe survival of recombinants with disrupted intragenome interactionnetworks (38, 51).

Genome organization and rolling circle replication (RCR)—themechanism by which geminiviruses and many other ssDNA virusesreplicate (9, 67, 79; see reference 24 for a review)—seemto have a large influence on basal recombination rates in differentparts of geminivirus genomes (20, 33, 39, 61, 81). To initiateRCR, virion-strand ssDNA molecules are converted by host-mediatedpathways into double-stranded "replicative-form" (RF) DNAs (34,67). Initiated by a virus-encoded replication-associated protein(Rep) at a well-defined virion-strand replication origin (v-ori),new virion strands are synthesized on the complementary strandof RF DNAs (28, 73, 74) by host DNA polymerases. Virion-strandreplication is concomitant with the displacement of old virionstrands, which, once complete, yields covalently closed ssDNAmolecules which are either encapsidated or converted into additionalRF DNAs. Genome-wide basal recombination rates in ssDNA virusesare probably strongly influenced by the specific characteristicsof host DNA polymerases that enable RCR. Interruption of RCRhas been implicated directly in geminivirus recombination (40)and is most likely responsible for increased basal recombinationrates both within genes transcribed in the opposite directionfrom that of virion-strand replication (40, 71) and at the v-ori(1, 9, 20, 69, 74).

Whereas most ssDNA virus families replicate via either a rollingcircle mechanism (the Nanoviridae, Microviridae, and Geminiviridae)(3, 23, 24, 31, 59, 67, 74) or a related rolling hairpin mechanism(the Parvoviridae) (25, 76), among the Circoviridae only theCircovirus genus is known to use RCR (45). Although the Gyrovirusgenus (the other member of the Circoviridae) and the anelloviruses(a currently unclassified ssDNA virus group) might also useRCR, it is currently unknown whether they do or not (78). Additionally,some members of the Begomovirus genus of the Geminiviridae eitherhave a second genome component, called DNA-B, or are associatedwith satellite ssDNA molecules called DNA-1 and DNA-Beta, allof which also replicate by RCR (1, 47, 68).

Recombination is known to occur in the parvoviruses (19, 43,70), microviruses (66), anelloviruses (40, 46), circoviruses(11, 26, 60), nanoviruses (30), geminivirus DNA-B components,and geminivirus satellite molecules (2, 62). Given that most,if not all, of these ssDNA replicons are evolutionarily relatedto and share many biological features with the geminiviruses(22, 31, 36), it is of interest to determine whether conservedrecombination patterns observed in the geminiviruses (61, 81)are evident in these other groups. To date, no comparative analyseshave ever been performed with different ssDNA virus familiesto identify, for example, possible influences of genome organizationon recombination breakpoint distributions found in these viruses.

Here we compare recombination frequencies and recombinationbreakpoint distributions in most currently described ssDNA virusesand satellite molecules and identify a number of sequence exchangepatterns that are broadly conserved across this entire group.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Sequence data sets. All publicly available full-length circovirus, microvirus, andparvovirus genome sequences, full-length nanovirus genome componentsequences, anellovirus sequences that were >50% of full genomesize, and geminivirus DNA-B and DNA-1 sequences were obtainedfrom public sequence databases by using TaxBrowser (http://www.ncbi.nlm.nih.gov/)between October and December 2007. An alignment of geminivirusDNA-Beta sequences has been described previously (6). With theexception of the anelloviruses, parvoviruses, and microviruses,sequences were linearized at the site that is nicked duringvirion-strand replication. In the case of the anelloviruses,sequences were linearized at the first nucleotide of eitherthe conserved AGGGCGGTGCCG sequence (Torque teno viruses [TTV])or the T/AGGGCGGGAGC sequence (Torque teno mini viruses [TTMV]).With the parvoviruses, the VP-NS intergenic region (5' $->$ 3') wasexcluded from analyses (because it was largely unalignable)and sequences were linearized at the first codon of the nonstructuralprotein gene. Microvirus sequences were linearized at position1469 relative to the sequence of isolate M14428. Sequence alignmentswere constructed using poa (37) and edited both by eye and usingthe ClustalW-based (77) alignment tool implemented in Mega4(75). Highly divergent sequences (i.e., those sharing <60%genome-wide sequence identity to any other sequences in a dataset) were discarded. Finally, to ensure that sequences couldbe aligned properly, data sets were split into groups of sequencesall sharing >60% genome-wide sequence identity. The BegomovirusDNA-A/DNA-A-like and Mastrevirus genome sequence alignmentsanalyzed here were described previously (38, 81).

Four "population-level" data sets that were used to detect evidenceof recombination rate differences within the complementary-and virion-strand geminivirus and circovirus genes were assembledprecisely as outlined previously (61).

Details of all analyzed data sets are given in Table S1 in thesupplemental material, and sequence alignments are availableupon request from the authors and/or within RDP3 project filesprovided as supplemental material.

Characterization of individual recombination events. Detection of potential recombinant sequences, identificationof likely parental sequences, and localization of recombinationbreakpoints was carried out with the RDP (48), GENECONV (62),BOOTSCAN (49), MAXCHI (54), CHIMAERA (64), SISCAN (21), LARD(29), and 3SEQ (4) methods implemented in RDP3 (52) (see theRDP project files submitted as supplemental material for fulldetails of program settings). Default settings were used throughout,and only potential recombination events detected by three ormore of the above methods coupled with phylogenetic evidenceof recombination were considered significant. Our choice ofusing the consensus of three or more methods was determinedempirically based on false-positive rates encountered duringanalyses of the simulated data sets of Posada and Crandall (64).Simultaneously analyzing these data sets with seven methods(all of those mentioned above except for LARD), using a consensusof three or more methods with a Bonferroni-corrected P valuecutoff of 0.05, resulted in false-positive rates below one falselyinferred recombination event per 100 data sets analyzed whileat the same time ensuring a good degree of analysis power. Toachieve maximum analysis power, we minimized the severity ofBonferroni correction during exploratory recombination analysesby either removing from analyzed alignments or masking withinthem (a setting in RDP3) all but one sequence within groupsof sequences sharing >98% genome-wide sequence identity.The exact program settings can be accessed within the RDP projectfiles provided as supplemental material.

Analysis of genome-wide recombination patterns. Recombination breakpoint density plots and recombinant regioncount matrices were constructed using RDP3 as described previously(27, 38). The matrices represent the numbers of times that recombinationalmovements of sequence tracts between genomes separate pairsof nucleotide sites. This representation of detectable recombinationevents highlights the differential "exchangeability" of sequencetracts between genomes. Whereas highly exchangeable genome regions(i.e., those represented by warm colors in the matrices dueto their frequent movement into foreign genetic backgrounds)are expected to be most modular, the less exchangeable regions(i.e., those represented by cool colors due to their infrequentmovement into foreign genetic backgrounds) are expected to bethe least modular.

Recombination hot and cold spot tests. Recombination breakpoint hot and cold spots were identifiedfrom breakpoint distribution plots by use of previously describedpermutation-based linear "local" and "global" tests (27). Thestatistical significance of potential recombination region hotand cold spots in recombinant region count matrices was testedusing a two-dimensional version of the linear local recombinationhot and cold spot permutation test of Heath et al. (27). Briefly,this involved the same procedure as the linear test except thatrather than plotting breakpoints in permuted and real data setson a linear genome map, the genomic regions bounded by breakpointswere plotted on recombination region count matrices as describedpreviously (45). The score of each cell in a particular realrecombinant region count matrix was ranked relative to correspondingcells recorded in 1,000 permuted matrices to identify cellsin the real matrix that had either higher or lower values than95% or 99% of corresponding cells in the permuted matrices.It should be stressed that the permutation P values were notcorrected for multiple testing and that one would, for example,expect a false-positive rate of 5% of the cells in the matrixat a P value cutoff of 0.05. Nevertheless, the test does providea reasonable quantitative assessment of the least and most transmissibleportions of genomes that takes into account the influence thatsequence diversity has on the detectability of recombinationevents (64).

Comparison of recombination breakpoint densities between different genome regions. We used another modification of the local permutation test ofHeath et al. (27) to specifically test for clustering of recombinationbreakpoints in different genome regions. In this test, ratherthan partitioning the alignment with a moving window of setlength, the alignment was partitioned in various other ways.For example, to test for significant clustering of breakpointsin the intergenic regions, alignments were partitioned intocoding and noncoding regions and tested to determine whethermore/fewer breakpoints were detectable in the intergenic regionsthan could be accounted for by chance. Other similar tests included(i) discounting breakpoints falling outside coding regions anddetermining whether individual genes contained significantlymore/fewer detectable breakpoints than the remainder of thecoding regions and (ii) again discounting breakpoints fallingoutside coding regions and determining whether the middle 50%of all genes collectively contained significantly more/fewerdetectable breakpoints than those collectively observed in thebeginning 25% and ending 25% of the genes.

Inference of population-scaled mutation and recombination frequencies. Variations in site-to-site composite likelihood estimates ofpopulation-scaled recombination frequencies were assessed withthe INTERVAL (56) component of ldhat (55). The program settingsfor these analyses were a precomputed likelihood lookup tablefor a population-scaled mutation frequency of 0.001, a minimumminor allele frequency cutoff of 0.05 (for data sets containing20 or more sequences) or 0.01 (for data sets with between 11and 19 sequences), a block penalty of 10, a starting recombinationfrequency of 5, and 10⁷ Markov chain Monte Carlo updates, withsampling every 2,000 updates and the first 500 samples discarded(56). The average genome-wide recombination frequency estimateobtained after a first run with these parameter settings wasthen used for a second run, using the same parameters but withthe starting recombination frequency replaced by that estimatedfrom the first run. We avoided analysis inaccuracies at theedges of alignments by simulating circular genome sequencesas described previously (61). Briefly, this involved constructingtandemly repeated alignments of full genome sequences and thenexcluding the point recombination frequency estimates for therepeated ends of the alignments.

RESULTS AND DISCUSSION

Top

RESULTS AND DISCUSSION

Increased recombination frequencies in complementary-sense genes are common but not absolutely conserved among ssDNA replicons. The geminiviruses and circoviruses are the only known ssDNAviruses that both utilize RCR and have complementary-sense genesthat are transcribed in the opposite direction and on the samestrand as virion-strand synthesis. For geminiviruses, it hasbeen speculated that this genome organization may result inclashes between transcription and replication complexes andresult in increased basal recombination rates within the complementary-sensegenes (61). Such a pattern has in fact been observed for thegeminiviruses maize streak virus, East African cassava mosaicvirus (Fig. 1a and b) (61), and East African cassava mosaicKenya virus. Here we show that this pattern is also apparentwithin the complementary-sense genes of the circoviruses porcinecircovirus 1 (PCV-1) and porcine circovirus 2 (PCV-2) (Fig.1c and d). Also in common with the geminiviruses is that recombinationfrequencies in these circovirus genomes decrease sharply onthe virion-sense gene side of the v-ori.

View larger version (19K):
[in this window]
[in a new window]

FIG. 1. Variable recombination rates across the genomes of maize streak virus (MSV) (61) (a), DNA-A genome components of East African cassava mosaic virus (EACMV) (61) (b), porcine circovirus 2 (PCV-2) (c), porcine circovirus 1 (PCV-1) (d), and DNA-B genome components of East African cassava mosaic virus (EACMV-DNA B) (e and f). The black plots represent average estimates of point recombination rates determined by the reversible-jump Markov chain Monte Carlo (RJMCMC) approach implemented in the INTERVAL component of ldhat (64). Gray regions represent the 95% credibility intervals of point recombination rate estimates from the RJMCMC chain. Genome cartoons above the plots indicate the starting and ending alignment positions of various genes (cp, CP gene; rep, replication-associated protein gene). Vertical lines beneath the genome cartoons indicate the locations of polymorphic sites used for the analysis.

The only other known classes of rolling circle replicons wheretranscription is known to occur in the opposite direction fromthat of virion-strand replication are the begomovirus DNA-Bgenome component and begomovirus-associated DNA-Beta satellitemolecules. Whereas we were unable to assemble a suitable DNA-Betadata set for our recombination frequency analyses, two smallEast African cassava mosaic virus DNA-B data sets (containingonly 12 and 11 sequences) displayed no obvious differences inrecombination frequencies between their complementary- and virion-sensegenes (Fig. 1e and f). Whereas no genome-wide fluctuations inrecombination frequencies were detectable in one of these datasets (Fig. 1e), the other (Fig. 1f) displayed increased recombinationfrequencies within the intergenic region around the v-ori, apattern that is superficially similar to that seen for the DNA-Acomponents of these viruses. In contrast to the case for DNA-Acomponents, however, there was a notable decrease in detectablefrequencies at the v-ori, with recombination frequency peaksinstead occurring at the boundaries of the so-called "commonregion," a stretch of sequence containing the v-ori that ishighly conserved between the DNA-A and DNA-B components of bipartitesequences. Although it is possible that these small DNA-B datasets are not representative of DNA-B sequences in general, thisevidence suggests that complementary-strand transcription runningcounter to virion-strand RCR is perhaps not absolutely associatedwith increased recombination rates in complementary-sense genes.

Interspecies and interstrain recombination is common in ssDNA viruses. We used a battery of eight recombination detection methods anda series of manual recombination signal evaluation tools implementedin the program RDP3 (52) to identify and characterize 663 uniquerecombination events detectable within 27 different ssDNA virusfull genome/genome component data sets (see Table 1 for a summaryof these events and Table S1 in the supplemental material fordetails of each individual event, as well as the interactiveRDP3 project files provided as supplemental material).

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

TABLE 1. Summary of recombination signals detectable in ssDNA virus full-genome data sets

Surprisingly, most sequences in most of the data sets were apparentlyrecombinant. For example, 38 events were detectable in 37 microvirussequences and 38 events were detectable in 39 begomovirus-associatedDNA-1 satellite sequences. Whereas in some of the data setssequence diversity was relatively low and the detected recombinationevents were all between members of the same species (the PCV,goose circovirus, and gyrovirus data sets), in other data setsmost, and in some cases all, detectable recombination eventswere between viruses sharing <90% genome-wide nucleotidesequence identify (Table 1). Although for some groups, suchas the gyroviruses, goose circoviruses, and erythroviruses,only a few recombination events were detected, this was probablya reflection of both the low diversity and small sizes of thesedata sets (see Table S1 in the supplemental material; also seereference 64 for a discussion on the inherent difficulty ofidentifying and characterizing individual recombination eventsin data sets with low diversity).

A v-ori recombination hot spot is partially conserved across diverse rolling circle replicons. For data sets in which more than 10 recombination events weredetectable, we tested for the presence of both recombinationbreakpoint and recombinant region hot and cold spots. Whereasbreakpoint distribution maps were generated and tested as describedpreviously (27), the tracts of sequence exchanged during recombinationevents were also mapped onto recombinant region count matrices.These matrices describe the relative frequencies with whichdifferent parts of the analyzed ssDNA replicons are separatedduring recombination. Recombinant region hot and cold spotswere identified as genome regions that were more or less frequentlyexchanged, respectively, during recombination than can be accountedfor by chance (with the null hypothesis that tracts of sequenceare randomly exchanged). Importantly, the permutation testsused for both the recombination breakpoint distribution plotsand the recombinant region count matrices account for recombinationbeing inherently easier to detect in more diverse genome regionsthan it is in less diverse regions.

The most conserved feature of detectable recombination breakpointdistributions was a statistically significant breakpoint clusterat the v-ori's (black arrows in Fig. 2) of circovirus (beak-and-featherdisease virus), microvirus, and geminivirus (mastrevirus andbegomovirus DNA A, DNA B, and DNA-Beta) genomes/genome componentsthat are known to use RCR. While it has been apparent for sometime now that the v-ori is a mechanistically predisposed recombinationhot spot in geminiviruses (73, 74) and circoviruses (9), ourresults indicate that the same is probably true for most otherrolling circle replicons. There are two probable reasons forv-ori sequences being recombination hot spots. Firstly, theyare the natural points at which recombinational repairs of double-strandedgenome breakages are resolved by the joining and nicking activitiesof Rep proteins (74), and secondly, they are the sites at whichunit-length genomes are replicationally released from high-molecular-weightgenomic concatemers that are produced by recombination-dependentreplication (33, 65).

View larger version (48K):
[in this window]
[in a new window]

FIG. 2. Distributions of recombination breakpoints detected within different ssDNA virus data sets. All detectable breakpoint positions are indicated by small vertical lines at the top of the graphs. A 200-nt window was moved along each of the represented alignments 1 nt at a time, and the number of breakpoints detected within the window region was counted and plotted (solid lines). The upper and lower broken lines indicate 99% and 95% confidence thresholds, respectively, for globally significant breakpoint clusters. Light and dark gray areas indicate local 99% and 95% breakpoint clustering thresholds, respectively, taking into account local regional differences in sequence diversity that influence the ability of different recombination detection methods to identify recombination breakpoints. Red areas indicate recombination hot spots, while blue areas represent recombination cold spots. Genes (horizontal arrows) are represented on the top of the graph (cp, CP gene; rep, replication-associated protein gene). The v-ori's of various rolling circle replicons are indicated with vertical black arrows.

Significant evidence of v-ori recombination hot spots was, however,not found in the nanovirus and geminivirus DNA-1 satellite moleculedata sets, indicating that mechanisms underlying v-ori hot spotsmay not be conserved across all rolling circle replicons. Inthis regard, it is currently unknown whether recombination-dependentreplication, a process directly implicated in the v-ori recombinationhot spot found in geminiviruses (33), occurs in any other ssDNAviruses. It should also be pointed out, however, that our analysismay have simply lacked enough power to detect v-ori hot spotsin some of the analyzed data sets. For example, within eachof the six nanovirus data sets, the hot spot test lacked anyappreciable power due to the numbers of detectable recombinationbreakpoints being very low (<15).

Besides the v-ori hot spot, we found no other obviously conservedrecombination breakpoint patterns between ssDNA replicons fromdifferent families. We realized, however, that given the smallnumbers of recombination breakpoints detected in many of thedata sets, our analysis lacked sufficient power to find anybut the most obvious recombination hot and cold spots.

Despite this, we were encouraged by the fact that careful visualinspection of recombinant region count matrices (Fig. 3, uppertriangles) revealed what may have been subtle conserved recombinationpatterns that were missed by our breakpoint hot/cold spot test.In these matrices, whereas dark/light blue triangles and rectanglescorrespond with genomic regions that tend to be inherited fromthe same parental source during recombination events, orange/redtriangles and rectangles denote recombinant regions that tendto become separated during recombination events. The lower trianglesin Fig. 3 indicate whether individual pairs of sites representedin the upper triangles are separated more (red) or less (blue)often during recombination events than can be accounted forby chance. Note, however, that the statistical test used inthese "probability matrices" is not multiple comparison corrected,which means, for example, that for a P value threshold of <0.01one would expect a 1% false-positive rate per pair of sitesanalyzed.

View larger version (57K):
[in this window]
[in a new window]

FIG. 3. Recombination region count matrices (upper hemimatrices) and recombination region hot/cold spot matrices (lower hemimatrices) for 12 different ssDNA replicon data sets. Unique recombination events were mapped onto the matrices based on their estimated breakpoint positions. In the upper matrices, the shades displayed are a function of the number of times that pairs of nucleotides (plotted on the x and y axes) were separated by observable recombination events. Pairs of sites that appear to be the least and most separable by recombination (taking into account the genome-wide variation in sequence diversity and the impact that this has on recombination detection) are represented on the bottom matrices by blue and red, respectively. The positions of genes (horizontal arrows) are represented between the matrices (cp, CP gene; rep, replication-associated protein gene).

Nevertheless, these matrices indicated three potential recombinationpatterns shared by many of the analyzed viruses, as follows:(i) intergenic regions tended to be moved between genomes morefrequently than individual genes were (note red/orange/lightgreen diagonals in the upper matrices and red diagonals in thelower matrices originating on intergenic regions in the anellovirus,geminivirus, DNA-Beta, and nanovirus data sets); (ii) certaingenes, particularly those encoding coat proteins (CP), tendedto be moved by recombination as either complete or mostly complete(>50% of the middle regions) units (note light/dark bluetriangles in the upper matrices and blue patches in the lowermatrices associated in particular with CP genes in the anellovirus,microvirus, circovirus, parvovirus, and geminivirus data sets);and (iii) CP genes tended to contain fewer recombination breakpointsthan other genes did (note blue patches in the recombinationregion count matrices in Fig. 3 and the distribution of breakpointsindicated in Fig. 2).

Selection apparently disfavors recombinants with breakpoints in coding regions. To directly test for conserved features of recombination breakpointdistributions that might underlie the patterns we observed inthe recombination region count matrices, we modified our recombinationbreakpoint hot and cold spot test. The original test effectivelydetermined whether the numbers of breakpoints observed in particularsmall stretches of sequence (in the case of Fig. 2, a moving200-nucleotide [nt] window) were greater or less than couldbe accounted for by chance. Given the small numbers of breakpointsdetectable in many of the data sets, this test lacked powerprimarily because the average number of breakpoints per windowwas low (and often zero), regardless of whether windows wereover recombination cold spots or not. To remedy this problemin our new test, rather than partitioning sequence alignmentsusing a moving window, we simply partitioned them into two orthree large regions and used the same permutation test to determinewhether individual partitions contained more or fewer recombinationbreakpoints than could be accounted for by chance. Specifically,we compared breakpoint numbers for (i) coding versus noncodingregions, (ii) the middle 50% of genes versus the beginning andend 25% of genes, and (iii) CP genes versus other genes. Tofurther increase the power of our modified test, we merged someof our original 27 data sets and discarded five others in whichfewer than eight recombination breakpoints were detectable (Table2).

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

TABLE 2. Imbalances in recombination breakpoint locations between different genome regions

It was apparent from our earlier recombination breakpoint distributionanalyses that whereas hot spots tended to occur within intergenicregions (70% of 23 detected hot spots), cold spots tended tooccur within coding regions (94% of 18 detected cold spots).Applying our modified method, we confirmed that in 8 of the14 analyzed data sets intergenic regions had a significantlyhigher density of detectable recombination breakpoints (P <0.05) than coding regions did (Table 2). The exceptions werethe begomovirus DNA-A and DNA-1 satellite, nanovirus, circovirus(PCV), microvirus, and dependovirus data sets. Whereas breakpointdensities were clearly highest in the noncoding regions of thebegomovirus DNA-A (35 versus 24 breakpoints per 100 nt), DNA-1(8.7 versus 5.2 breakpoints per 100 nt), microvirus (5.05 versus2.7 breakpoints per 100 nt), and nanovirus (3.1 versus 2.6 breakpointsper 100 nt) data sets, the circovirus (PCV) and dependovirussequences had only extremely small intergenic regions includedin the analysis, and therefore neither had any detectable breakpointsoutside genes. Despite these two exceptions, the trend is clear:across all ssDNA virus families, there is a significant tendencyfor detectable recombination breakpoints to fall outside codingregions.

While this tendency might be due to recombination breakpointswithin genes being less tolerable than those that fall betweengenes, it is difficult to discount the fact that the tendencymight instead be caused by the occurrence within intergenicregions of mechanistically predisposed recombination hot spotssuch as v-ori's. However, even when intergenic regions werediscounted, we detected a tendency for recombination breakpointsto occur within the beginning and ending 25% of genes ratherthan in the middle 50%. This tendency was significant (P <0.05) for all but the geminivirus DNA-1, anellovirus (TTV),circovirus (PCV), parvovirus, and dependovirus data sets. Amongthese exceptions, the anellovirus (TTV), parvovirus, and dependovirusdata sets displayed higher densities of recombination breakpointsat the edges of genes than in the middle 50% of genes. Also,when we merged the dependovirus and parvovirus data sets (bothare members of the Parvoviridae), the increased clustering ofbreakpoints at the edges of genes became significant (Table2).

Collectively, the relatively low abundance of breakpoints withincoding regions and the tendency for breakpoints to occur towardthe edges of genes are consistent with the hypothesis that breakpointsare less tolerable when they occur within genes (5, 38, 83)because there is a relatively high probability that recombinantproteins will not fold properly (14; see reference 7 for a review).

CP genes contain fewer detectable breakpoints than other genes. While recombination region count matrices (Fig. 3) indicatedthat the entire CP genes of various ssDNA virus groups tendedto be inherited from the same parental source, recombinationbreakpoint distribution analyses indicated that for data setscontaining a CP gene, 8 of 13 (61%) had statistically significantrecombination cold spots within these genes. We therefore specificallytested whether CP genes tended to have significantly fewer recombinationbreakpoints than other genes. Importantly, the test we useddiscounted breakpoints that fell within intergenic regions andwas therefore an unbiased comparison between the genes themselves.

As expected, relative to the other genes, we found significantlyfewer (P < 0.05) recombination breakpoints within CP genesfor 6 of 10 analyzed data sets (3 of the original 14 data setsdid not contain a CP gene, and 1 contained only a CP gene).Among the four data sets where CP genes did not have significantlyfewer breakpoints than other genes, the circovirus (PCV) andparvovirus data sets had lower densities of breakpoints withintheir CP genes than in their Rep genes (the only other geneon these replicons), and in the case of the parvovirus dataset, this lower density was marginally significant (P = 0.0932).The other two exceptional data sets, the anellovirus (TTMV)and dependovirus data sets, were both unusual in that relateddata sets (TTV and parvovirus data sets, respectively) displayedeither a significant or marginally significant tendency towardhaving lower breakpoint numbers within their CP genes.

Decreased recombination breakpoint densities have been notedwithin the structural protein genes of picornaviruses (27, 42,72), adenoviruses (41), human immunodeficiency viruses (bothcapsid and envelope proteins) (17), and hepatitis B viruses(85). This implies either that viral CP genes generally experiencelow basal recombination rates or that they are generally lesstolerant of recombination than most other genes. For geminiviruses,there is good evidence from both experimental and computationalanalyses that CP genes both experience lower basal recombinationrates (33, 61) and have a low degree of recombination tolerance(38, 51). This evidence suggests that whereas recombinationbreakpoints within the CP gene frequently disrupt protein folding(38), even breakpoints bounding the gene can disrupt properCP function by interfering with coevolved interactions betweenthe CP and the remainder of the genome (51). In addition tothis, recombination might compromise coevolved interactionsbetween different CP molecules and disrupt virus particle assembly.Our results suggest that natural selection acting against disruptionof these various CP interactions is operational across all ofthe ssDNA viruses and is possibly a feature of virus evolutionin general.

Shared mechanistic and selective processes probably underlie shared recombination patterns. Although it has been known for some time now that large numbersof recombination events are detectable within the full genomesequences of most ssDNA virus groups (2, 11, 26, 30, 40, 46,60, 62, 66), we have shown here that these events have occurredin patterns which are broadly conserved across the ssDNA viruses.As has been suggested for the geminiviruses (38, 81), theseconserved patterns imply that an interplay between selectiveand mechanistic processes determines the general distributionsof recombination breakpoints detectable in ssDNA viruses.

We have found evidence of what appear to be similar mechanisticpredispositions to recombination between the geminiviruses andcircoviruses. Most obviously, we and others have found evidenceconsistent with the hypothesis that the v-ori's of diverse rollingcircle replicons are mechanistically predisposed recombinationhot spots (9, 73, 74). It also appears likely that complementary-sensegenes in both geminiviruses and circoviruses experience increasedrecombination rates relative to virion-sense genes, possiblydue to mechanistic interferences between the transcription andreplication complexes during RCR.

Higher mechanistic predispositions to recombination in particularparts of genomes do not, however, necessarily translate intoincreased numbers of breakpoints detectable in those regionsin naturally sampled genomes. If they are to survive, newlyproduced recombinants must be able to compete productively withtheir parents. Our results suggest that among the ssDNA repliconswe analyzed, natural selection in general tends to (i) penalizebreakpoints within coding regions more harshly than it doesbreakpoints in intergenic regions, (ii) favor recombinants withbreakpoints on the edges of genes more than recombinants withbreakpoints within the centers of genes, and (iii) stronglydisfavor recombinants with breakpoints within CP genes.

The notion that all ssDNA replicons might be experiencing approximatelyequivalent evolutionary processes is further supported by therecent finding that microviruses, circoviruses, parvoviruses,and geminiviruses (and probably other ssDNA replicons too) areunusual among DNA viruses in that they are subject to nucleotidesubstitution rates that are as high as those of some RNA viruses(see reference 15 for a review). Although high mechanistic mutationand recombination frequencies do not necessarily translate intohigh evolution rates, our recombination results emphasize thehuge evolutionary potential of these viruses. Whereas in thepast this potential no doubt facilitated the dispersal and adaptationof ancestral ssDNA replicons to bacterial, plant, and animalhosts (12, 22), in recent times it has no doubt contributeddirectly to the emergence of many of these viruses as seriousplant and animal pathogens (15, 58, 71, 84).

ACKNOWLEDGMENTS

P.L. is supported by the French Ministère de la Rechercheet de l'Enseignement Supérieur. J.-M.L. is funded byCIRAD and the Conseil Régional de la Réunion.A.V. is supported by the Carnegie Corporation of New York. D.P.M.is supported by the Wellcome Trust.

FOOTNOTES

* Corresponding author. Mailing address: Institute of Infectious Diseases and Molecular Medicine, University of Cape Town, Observatory 7925, South Africa. Phone: 27-21-406 6366. Fax: 27-21-698 1528. E-mail: Darrin.Martin{at}uct.ac.za

Published ahead of print on 30 December 2008.

Supplemental material for this article may be found at http://jvi.asm.org/.

REFERENCES

Top