• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of geneticsGeneticsCurrent IssueInformation for AuthorsEditorial BoardSubscribeSubmit a Manuscript
Genetics. Dec 2006; 174(4): 1983–1994.
PMCID: PMC1698637

Nonneutral Evolution of Organelle Genes in Silene vulgaris


Knowledge of mitochondrial gene evolution in angiosperms has taken a dramatic shift within the past decade, from universal slow rates of nucleotide change to a growing realization of high variation in rates among lineages. Additionally, evidence of paternal inheritance of plant mitochondria and recombination among mitochondrial genomes within heteroplasmic individuals has led to speculation about the potential for independent evolution of organellar genes. We report intraspecific mitochondrial and chloroplast sequence variation in a cosmopolitan sample of 42 Silene vulgaris individuals. There was remarkably high variation in two mitochondrial genes (atp1, atp9) and additional variation within a third gene (cob). Tests for patterns of nonneutral evolution were significant for atp1 and atp9, indicative of the maintenance of balanced polymorphisms. Two chloroplast genes (matK, ndhF) possessed less, but still high, variation and no divergence from neutral expectations. Phylogenetic patterns of organelle genes in both the chloroplast and mitochondria were incongruent, indicating the potential for independent evolutionary trajectories. Evidence indicated reassociation among cytoplasmic genomes and recombination between mitochondrial genes and within atp1, implying transient heteroplasmy in ancestral lineages. Although the mechanisms for long-term maintenance of mitochondrial polymorphism are currently unknown, frequency-dependent selection on linked cytoplasmic male sterility genes is a potential candidate.

UNTIL recently (Städler and Delph 2002; Cho et al. 2004; Parkinson et al. 2005), the coding regions of the organelle genomes of angiosperms were viewed as highly constrained, with low rates of mutation and strict maternal inheritance (Wolfe et al. 1987). This has resulted in the assumption that intraspecific variation in organelle genes should be low or even absent. Contradictory patterns of extremely high mtDNA variation have recently been observed in Pelargonium (Parkinson et al. 2005), Plantago (Cho et al. 2004), and Silene (Städler and Delph 2002). Two of these genera (Plantago and Silene) are model systems for studies of the evolution of gynodioecy in plants (De Haan et al. 1997; Charlesworth and Laporte 1998). The elevated rate of synonymous substitutions in Pelargonium has been attributed to an increase in mutation rate or a reduction in the efficacy of DNA repair mechanisms (Parkinson et al. 2005). In Plantago, high diversity at the mitochondrial cytochrome oxidase I (cox1), adenosine 5-triphosphate (atp1), and small subunit RNA-encoding DNA (SSU rDNA) loci has been attributed to similar causes (Cho et al. 2004). Silene acaulis (L.) Jacq. possessed high diversity of mitochondrial apocytochrome b (cob) haplotypes, but this was attributed to ancient coalescence due to long-term maintenance of a balanced polymorphism (Städler and Delph 2002).

If selection maintains a balanced polymorphism in mitochondrial types for long periods, one candidate for the target of selection is cytoplasmic male sterility (CMS) genes that influence sex expression. Sex expression in many gynodioecious species (those with females and hermaphrodites in natural populations) is controlled by a complex joint cytonuclear polymorphism between CMS genes and their associated restoration of male fertility (Rf) genes (Saumitou-Laprade et al. 1994; Schnable and Wise 1998). Theoretical studies predict that through time CMS and Rf frequency cycles are generated from negative frequency-dependent selection assuming both female compensation (females produce more seeds than hermaphrodites; Darwin 1887) and a cost to carrying Rf genes (cost of restoration, Charlesworth 1981; Delanney et al. 1981). CMS/Rf cycles are stabilized by specific relationships between the fitnesses of particular CMS types, by the costs of carrying Rf genes (Frank 1989; Gouyon et al. 1991), and by low levels of paternal leakage of CMS genes (Wade and McCauley 2005), potentially providing mechanisms whereby the joint CMS/Rf polymorphism is maintained over long periods. CMS genes are ubiquitous in plants and have utility within breeding programs of nearly every major crop species. For reasons that are yet to be discovered, the male-sterile phenotype associated with the expression of CMS is uncommon in natural populations of the majority of plant species, with gynodioecy found in only 7% of angiosperm taxa (Richards 1997).

Even if some form of balancing selection maintains CMS/Rf polymorphism on an ecological timescale, there remains considerable contention over whether polymorphism in CMS genes is maintained over longer time periods (Frank 1989; Gouyon et al. 1991) or whether newly arisen types more likely spread as epidemics within gynodioecious populations (Ingvarsson and Taylor 2002). For instance, CMS and their associated Rf genes may spread to fixation in the absence of costs of restoration (Charlesworth 1981; Frank 1989; Gouyon et al. 1991), or demographic effects and drift may cause the loss/fixation of a particular CMS type when CMS genes are at a low/high frequency during a cycle. Both of these outcomes are consistent with an epidemic model.

A comparative molecular population genetic approach may show promise for discriminating among the three alternatives proposed to shape mitochondrial nucleotide diversity. For instance, elevated mutation rates would be expected to increase diversity within lineages in all genes in the genome, whereas selection may target only specific genes or genomic regions. Furthermore, elevated mutation rates are not expected to influence either the relative lengths of internal and external branches or the expected numbers of haplotypes, whereas selection will (Fu and Li 1993; Depaulis and Veuille 1998). An example of the influence of long-term balancing selection is found in plant self-incompatibility alleles wherein negative frequency-dependent selection may have preserved variants for 27–50 million years, leading to trans-specific polymorphism (Ioerger et al. 1990; Uyenoyama 1995). Finally, epidemics generate short internal branches in gene genealogies as all alleles coalesce to the allele responsible for the epidemic (Hatcher 2000; Ingvarsson and Taylor 2002). If CMS genes either were maintained over long periods or were characterized by spreading as epidemics, a naive prediction might be for similar patterns of long or short time to coalescence, respectively, for CMS alleles. However, studies of species wherein multiple CMS genes have been identified have shown that CMS variants are not allelic, but rather each is constructed from different sets of genes to create a novel chimeric open reading frame (Schnable and Wise 1998). Nonetheless, some researchers have posited that since the mitochondrial genome is inherited as a single linkage unit and the chloroplast and mitochondria are co-inherited, these genomes may harbor signals of past balancing selection or epidemics via hitchhiking (Olson and McCauley 2000; Ingvarsson and Taylor 2002; Cho et al. 2004).

Here we assess patterns of nucleotide diversity in a cosmopolitan sample of S. vulgaris (Monech.) Garcke individuals, a well-studied gynodioecious plant species (Marsden-Jones and Turrill 1957; Pettersson 1992; Olson et al. 2005). We estimate diversity in three mitochondrial and two chloroplast genes to determine whether genomewide hitchhiking can indeed influence genetic diversity throughout the mitochondrial genome and potentially extend to the chloroplast genome, which is often co-inherited. The presence of intraspecific diversity in the mitochondrial and chloroplast genomes also offers an opportunity to explore the transmission dynamics and role of recombination on the cytoplasmic genomic compartments. With rare exceptions (Reboud and Zeyl 1994), the mitochondria and chloroplast in angiosperms are thought to be primarily maternally inherited (Birky 2001; McCauley et al. 2005). Phylogenetic patterns among co-inherited genes or genomes are expected to be congruent, due to a process akin to linkage disequilibrium (Olson and McCauley 2000). The absence of such congruence implies independent transmission histories perhaps resulting from horizontal transfer (Bergthorsson et al. 2003) or paternal leakage (McCauley et al. 2005), producing heteroplasmic individuals. Over generations, heteroplasmy is expected to be reduced through random genetic drift (Albert et al. 1996; Roze et al. 2005). As heteroplasmy is a necessary precondition for recombination among divergent genes or genomes, the presence of recombination within or among cytoplasmic genomes not only indicates that cytoplasmic genes may have the potential to evolve as independent units, but also indicates that transmission dynamics may be much less constrained than previous studies have indicated. S. vulgaris is an excellent candidate for better understanding the mitochondrial and chloroplast evolution as previous studies have identified high levels of mitochondrial polymorphism within small populations (Olson and McCauley 2000; Štorchová and Olson 2004), and small samples have indicated that sequence diversity at coding regions may also be elevated relative to that in other taxa (McCauley et al. 2005).


S. vulgaris (Moench.) Garcke (Caryophyllaceae) seeds were collected from four regions (Virginia and New York/Vermont; central Europe, Czech Republic/Germany; and Siberia, Russia; Figure 1) and grown in the Institute of Arctic Biology greenhouse at the University of Alaska (Fairbanks, AK). Eight populations were located in Giles and Craig counties, Virginia (628, GR, CO, AJ, MV, AR, WR, and EB, Figure 1; listed as populations 3, 4, 5, 6, 11, 12, 14, and 15, respectively, in Olson and McCauley 2002), and four populations were located in New York (NY) and Vermont (Vt) (Vt17b, Vt100, and VtLnG, Figure 1; McCauley et al. 2003). Of the central European populations, three were located near Prague, Czech Republic (Kovary Meadow, KM; Institute Garden, InG; Sedlac railway station, SD), two were in northern Moravia, Czech Republic (Mount Keprnik, MK4; Rychleby Mountains, RM), two were from southern Moravia (Klentnice, KL; Lednice, Ld), and the eighth population was located near Gemsteltal (GM) in the Allgauer Alps, Austria (Figure 1; Štorchová and Olson 2004). One population was collected from Krasnojarsk (Krn) in south central Russia near Lake Baikal (Figure 1). One plant was arbitrarily selected for study from each of one to five maternal families from each of these 21 populations of S. vulgaris. Two additional Silene species were included as potential outgroups and to determine the extent of interspecific divergence. Seeds from S. uniflora Roth (also called S. maritima), a species closely related to and not entirely reproductively isolated from S. vulgaris (Marsden-Jones and Turrill 1957; Runyeon-Lager and Prentice 2000), originated in Saint Catherine's Bay, United Kingdom. Seeds from S. latifolia Poir. were collected in the courtyard at the Abbey of St. Thomas, Brno, Czech Republic; S. vulgaris and S. latifolia are not interfertile.

Figure 1.
Map of the locations of Silene vulgaris samples collected for this study. The large dots represent multiple populations from within the same locality, whereas the small dots represent the locations of single populations.

Total genomic DNA was isolated and purified from leaf tissue using a QIAGEN (Valencia, CA) DNeasy plant mini kit. We PCR amplified three mitochondrial (adenosine 5′-triphosphate subunit 1, atp1; adenosine 5-triphosphate subunit 9, atp9; apocytochrome b, cob) and two chloroplast genes (maturase K, matK; NADH dehydrogenase F, ndhF) from each individual. PCRs were performed in 35-μl reactions, with 15–40 ng of genomic DNA template and a 50° annealing temperature using standard protocols (Štorchová and Olson 2004). Sequencing reactions were carried out using the ABI Big-Dye terminator sequencing kit and sequenced directly on an ABI 3100 automated sequencer. Sequences had no indications of heterozygosity (e.g., overlapping chromatograms), suggesting that if heteroplasmy was present, we were sequencing the most common allele. PCR primer sequences and aligned data sets are available upon request.

As ancient gene transfers from the mitochondrial to the nuclear genome or nuclear pseudogenes are possible in plants (Mower et al. 2004), we confirmed maternal inheritance of the atp1 and atp9 regions to ensure that they were indeed of cytoplasmic origin. For atp1, a reciprocal cross was conducted between two plants with types C and F (see Table 1) polymorphic for SmaI restriction sites in atp1. Atp1 was amplified from 16 progeny from each mother (32 total) and allelic inheritance was determined after separation of digested PCR products on a 3% agarose gel stained with ethidium bromide. All progeny possessed the same haplotype as their mother. Under a null hypothesis of biparental inheritance as expected for nuclear copies, the probability of finding only the single haplotype in the progeny from each maternal parent was 0.00002, with a power of 99%. For atp9, a reciprocal cross was conducted between 62812 and a plant with type C (see Table 1) that differed in a MvnI restriction site. Ten progeny from the maternal parent with type C and 7 progeny from 62812 as a maternal parent were genotyped. Atp9 restriction polymorphisms exhibited strict maternal inheritance; the probability of all 17 progeny inheriting the maternal haplotype under a null hypothesis of biparental inheritance and nuclear localization was 0.001, with a power of 91%. As it is known that low levels of biparental inheritance occur in S. vulgaris, one might also be interested in the probability of detecting biparental inheritance given our sample sizes. Using the 95% confidence interval of the zero term of the Poisson distribution to estimate the minimum detectable frequency of biparental inheritance, a total sample size of 49 (combined across atp1 and atp9) allows detection of not less than 6.08% (2.98/49) biparental progeny. In addition to verifying the maternal inheritance of these subunits, Southern blots with atp1-specific probes hybridize strongly, indicating that atp1 is present in high copy number as is expected for genes located in the mitochondria. Atp1, atp9, and cob, like most other respiratory genes, have never been found to be lost in the mitochondrial genome and transferred to the nuclear genome in plants, making this very unlikely in this taxon (Adams et al. 2002). Finally, a phylogeny of the atp1 data set from S. vulgaris and that from a much larger study (Bergthorsson et al. 2003) were combined to rule out the possibility of horizontal transfer via parasitic plants. In this analysis, atp1 haplotypes from S. vulgaris formed their own clade with high bootstrap support and were not associated with any other genera; thus, there is no evidence for horizontal transfer of organelles or genes as has been shown in other groups with high organelle genomic variation (Mower et al. 2004).

Polymorphic sites in the three mitochondrial (atp1, atp9, cob) and two chloroplast (matK, ndhF) coding regions of S. vulgaris

We reamplified and resequenced a subset of samples (SD6, NY006, Ld5, EB3AG, 62812, NY007, Vt17b, and Gm7) to reconfirm the patterns of diversity and recombination within and between atp1 and atp9. Although we feel it is unlikely that cross-sample contamination could have produced the results reported here, it cannot be discounted with complete certainty.

Sequences lacked indels, easing manual alignment. The following summary statistics were calculated using DNAsp version 4.0 (Rozas et al. 2003): θ, per site estimated from the total number of mutations (Watterson 1975); θs, per site, including only synonymous nucleotide differences: π, diversity, the average number of nucleotide differences per site between a pair of randomly chosen sequences (Nei 1987); Ks, average number of pairwise synonymous substitutions per synonymous site; Ka, average number of pairwise nonsynonymous substitutions per nonsynonymous site; Tajima's D (Tajima 1989), based on the differences between the number of segregating sites (θ) and the average number of nucleotide differences per sequence (π); Fu and Li's D with an outgroup (Nicotiana tabacum) (Fu and Li 1993), the test statistic is based on the differences between the total number of mutations in the external branches of the genealogy and the total number of mutations; K-test (Depaulis and Veuille 1998), the observed number of haplotypes vs. the range of the expected number of haplotypes (calculated using 1000 neutral coalescent simulations); and the four-gamete test (for recombination) (Hudson and Kaplan 1985). Recombination was not included in the confidence interval simulations for the K-test to generate a conservative test for the minimum number of haplotypes. Specific hierarchical hypotheses concerning differences between chloroplast and mitochondrial genes and differences among mitochondrial genes were constructed and evaluated using log-likelihood tests as implemented in MLHKA (Wright and Charlesworth 2004). N. tabacum was chosen as an outgroup because its shared ancestor with S. vulgaris has little chance of also sharing Silene mitochondrial haplotypes; also it was the closest relative of Silene for which all five genes were available in GenBank. It has been demonstrated that the presence of population subdivision can lead to overestimates of deviations from neutrality using the HKA test (Ingvarsson 2004). Population subdivision within our study was minor; the same alleles were present in different sample regions. Additionally, all loci are organellar, and no comparisons are made with nuclear data that are under a different population dynamic (cf. Ingvarsson and Taylor 2002).

Phylogenetic hypotheses were constructed using PAUP* with an HKY model of substitution (Swofford 2003). The best-fit substitution model was chosen using Modeltest (Posada and Crandall 1998). Partition homogeneity tests, to assess the null hypothesis of congruence between the phylogenies of each gene, were carried out using PAUP* (Swofford 2003). RDP version 2.0 was used to detect potential recombination, using the default settings (Martin and Rybicki 2000). RDP implements both phylogenetic and substitution methods to test for recombination. Phylogenetic methods (RDP, Bootscan) compare phylogenies constructed from different parts of the genomes—if the phylogenies have different topologies then it is assumed that recombination has occurred. Substitution methods (GENECONV, Chimaera, Max-Chi, SiScan) look for clustering within sites with substitutions or whether substitutions occur within the window of analysis more frequently than a calculated statistical distribution. In addition, the four-gamete test (Hudson and Kaplan 1985) was applied to the data. This test relies on the infinite-sites model and deems that where we find a combination of all four possible combinations of a pair of variable sites the most parsimonious explanation is recombination. It also is possible that such a pattern can be generated by recurrent mutations (homoplasy) at an individual site.


Patterns of inter- and intraspecific diversity and divergence:

Considerable diversity was found within the three mitochondrial and two chloroplast genes in S. vulgaris (Tables 1 and and2),2), with remarkably high levels of mitochondrial intraspecific variation for angiosperms (Table 2). The mean pairwise numbers of synonymous substitutions per total synonymous sites (Ks) for the eight atp9 haplotypes found in S. vulgaris were similar to those found in comparisons of atp9 haplotypes from rice and maize (Table 3), two species that are believed to have diverged 50 MYA (Gaut 1998). Ks for the seven atp1 haplotypes in S. vulgaris also was high, approximately half the value found in comparisons of atp1 haplotypes from rice and maize (Table 3). Ks for the six apocytochrome b (cob) haplotypes was not as high as that for atp1 and atp9 and was an order of magnitude lower than that found both between rice and maize and the strikingly high value found among S. acaulis haplotypes (Table 3).

Diversity measures in organelle genes of Silene vulgaris
Intraspecific average interhaplotype pairwise synonymous and nonsynonymous substitutions per site in organelle genes of Silene vulgaris relative to other taxa and the rice–maize divergence

Plant mitochondrial and chloroplast transcripts are known to undergo post-transcriptional C–U editing at nonsynonymous sites (Gray 1996). Such editing may result in C–T DNA polymorphism not being reflected as a polymorphism in the mRNA; thus although the site is predicted to be nonsynonymous from the DNA, with editing the mutation will not alter the amino acid. In our data set, four nonsynonymous variable sites are potentially influenced by C–U RNA editing: site 21 in atp9, site 796 in cob, and sites 19 and 140 in matK. Many of the analyses presented herein do not utilize information on the coding state of the site; however, we have presented adjusted Ks/Ka-values for these three genes by scoring these sites as synonymous rather than nonsynonymous substitutions, as well as the nonadjusted values (Table 3).

For both atp1 and atp9, one S. vulgaris haplotype (represented by Vt17b and Ld5, Table 1) was identical to the haplotypes found in S. latifolia and S. uniflora, whereas other haplotypes were extremely diverged, for atp1, for atp9, or for both (Figure 2, Table 1). The majority of mitochondrial variation occurred at synonymous codon positions. Four of the 25 variable sites in atp1 resulted in amino acid changes, with Ka/Ks = 0.0590 (Rozas et al. 2003 following Nei and Gojobori 1986). For atp9, 4 of the 21 variable sites in S. vulgaris resulted in amino acid changes, with Ka/Ks = 0.0791. Three of the four variable sites in cob resulted in amino acid changes (Ka/Ks = 1.2222). Ka/Ks-values presented for atp9 and cob are unadjusted for potential RNA editing; adjusted values can be seen in Table 3.

Figure 2.
Maximum-likelihood phylograms of the three mitochondrial (atp9, cob, atp1) and two chloroplast (matK, ndhF) genes. Bootstrap support values >50 are shown above the branches. Individuals with five-character-long names are Silene vulgaris from North ...

Multivariate HKA tests indicated that intraspecific and interspecific divergence was not correlated among all five genes. Specifically, intraspecific diversity of mitochondrial genes was significantly greater than that of chloroplast genes (likelihood-ratio test = 24.22, d.f. = 2, P < 0.0001; using N. tabacum as an outgroup). Moreover, among mitochondrial genes intraspecific diversity for atp1 and atp9 was greater than that of cob (likelihood-ratio test = 9.655, d.f. = 1, P < 0.002).

The seven atp1 haplotypes observed were significantly fewer than expected assuming 25 segregating sites as determined by neutral coalescent simulations of the expected distribution of haplotypes (K-test, Table 2). For atp9, haplotype numbers were within the 95% confidence intervals for the expected distribution given the observed numbers of segregating sites, whereas in cob there were more haplotypes than expected (Table 2). For these tests we assumed no recombination, which leads to conservative estimates of the expected number of haplotypes. If recombination is allowed, then the number of haplotypes present in cob is not significantly different from neutral expectations (expected number of haplotypes is four to seven). Because recombination was detected in cob using the four-gamete test (see below), we suspect that the elevated number of haplotypes results from recombination. Fu and Li's D-statistic detected a significant departure from neutral expectations for atp1 and atp9 (Table 2) as can be seen via relatively long internal branches for these genes (Figure 2). No other tests of nonneutrality using Tajima's D and Fu and Li's D were significant (Table 2).

Interspecific divergence in matK and ndhF was substantial between S. vulgaris and S. latifolia (Table 1). Intraspecific variations in synonymous sites among the distinct haplotypes of these two S. vulgaris chloroplast genes were both two orders of magnitude lower than that found between rice and maize (Table 3). In matK, there was a bias toward nonsynonymous substitutions, with unadjusted Ka/Ks = 2.9500. Two of the four variable sites in ndhF were nonsynonymous, with Ka/Ks = 0.3030. No tests of nonneutrality (K-test, Tajima's D, Fu and Li's D) for either chloroplast gene were significant.

Reassociation and recombination:

Evidence for reassociation between genomes and recombination within genomes was detected at three levels of organization: between the chloroplast and mitochondrial genomes, between genes within the mitochondrial genome, and between sites within genes in the mitochondria. Strict co-inheritance of genomes is expected to generate congruent phylogenetic histories between genes in different genomes, whereas the absence of congruence implies at least somewhat independent histories (Desplanque et al. 2000; Olson and McCauley 2000). Pairwise partition-homogeneity tests (Farris et al. 1994) indicated incongruence between the pairwise histories of matK and atp1 (P = 0.003), matK and atp9 (P = 0.001), and atp1 and atp9 (P = 0.001). No additional pairwise comparisons detected incongruence using a Bonferroni-adjusted critical value of 0.005 (α = 0.05/10 comparisons), although cob and ndhF phylogenies did not contain sufficient structure to allow for a powerful test.

Recombination between mitochondrial genes can be inferred from some particularly striking patterns wherein all four possible combinations of genes were observed in different individuals (Table 1). Examination of Table 1 reveals that the most common combined atp1–atp9 haplotype in our sample (represented by SD6) is differentiated from the atp1–atp9 haplotype carried by NY006 and NY007 in 12 changes localized in atp9. Ld5 and Vt17b possess the atp9 haplotype of NY006 and NY007, but differ from SD6, NY006, and NY007 by 8 changes in atp1. Finally, individuals such as EB3AG possess the atp1 haplotype of Ld5 and Vt17b, but have the atp9 type of SD6. Given the large numbers of differences between the two atp1 haplotypes (8) and the two atp9 haplotypes (12), it is highly unlikely that all four combinations of atp1–atp9 haplotypes arose via independent mutational events. This conclusion is supported by the detection of a recombination event between atp1 and atp9 haplotypes by all three nonphylogenetic substitution models implemented in RDP (Chimaera, Max-Chi, and SiScan; RDP version 2.0). Finally, the four-gamete test was applied to all pairwise comparisons between segregating sites within each gene and detected three potential intragenic recombination events within atp1 (between positions 22 and 147, 415 and 507, and 507 and 978; Table 1), one event within atp9 (between 34 and 148; Table 1), and one event within cob (sites 516 and 796; Table 1). The event between sites 22 and 147 of atp1 was the only recombination event that was detected using all six recombination detection models implemented in RDP (RDP, GENECONV, Bootscan, Chimaera, Max-Chi, and SiScan).


Many aspects of cytoplasmic diversity in S. vulgaris, from the range of nucleotide diversity to the nonneutral patterns of diversity in atp1 and atp9, to the absence of complete linkage disequilibrium, indicate that the cytoplasmic genomes are evolving in a manner unlike those of most previously studied taxa. Our results bolster the findings of Laroche et al. (1997), Cho et al. (2004), and Parkinson et al. (2005), which showed highly variable rates of mitochondrial gene evolution among different lineages and genes in interspecific and intergeneric comparisons, and extend these observations to include the presence of high diversity within species. Here we offer several possible explanations for these widespread and enigmatic patterns in Silene.


We observed extremely high intraspecific nucleotide diversity at atp9, perhaps the highest average number of pairwise differences among intraspecific mitochondrial haplotypes observed for any mitochondrial gene to date (Table 2). The most divergent atp9 haplotypes (e.g., Vt17b vs. 6283B; Ks = 0.37) exhibited five times the level of divergence found between rice and maize and one-third the level found for atp1 within the entire genus Plantago, which has the highest documented rate of intrageneric mitochondrial evolution (Cho et al. 2004; Mower et al. 2004). A comparable study of intraspecific mitochondrial nucleotide diversity showed that S. acaulis also exhibited high intraspecific variation at the cob mitochondrial coding region (Städler and Delph 2002). The common ancestor of S. acaulis and S. vulgaris is represented by the earliest members of the genus (Oxelman et al. 1997), indicating that high mitochondrial variation may be a characteristic of the entire genus, although the amount of variation in particular genes such as cob is not consistent across taxa (Table 3).

Three potential mechanisms may be responsible for the high rates of variation within plants: (1) lineage-specific increased mutation rates of mitochondrial genes throughout the genome (Cho et al. 2004; Parkinson et al. 2005); (2) divergence in allopatry or in separate species and recent hybridization or introgression; and (3) the maintenance of ancient mitotypes resulting from long-term balancing selection, possibly due to the presence of cytoplasmic male-sterility genes (Städler and Delph 2002). For the first hypothesis to be correct for Silene, different mitochondrial lineages within S. vulgaris would be predicted to have different mutation rates, an explanation that is possible but perhaps unlikely at the intraspecific scale (but see Cho et al. 2004). Also, we found that different genes within the S. vulgaris mitochondrial genome evolve at different rates. Such a pattern is unlikely if elevated diversity results from a genomewide increase in mutation rate, but could result if the mutation rate is elevated for particular genes and not for others.

Population structure, hybridization, and introgression can generate high levels of diversity such as were exhibited in S. vulgaris. Although S. vulgaris exhibits population structure at an ecological timescale (McCauley 1998; Olson and McCauley 2002), populations in the United States and Europe have been isolated for only a few hundred years (McCauley et al. 2003) and are found in disturbed areas that are not conducive to long-term population persistence. The lack of allopatric isolation and the presence of diverged haplotypes in similar geographic locations (Figure 1, Table 1) suggest that recent divergence in allopatry was not likely to generate the observed patterns of diversity. Moreover, if more ancient allopatric isolation among European and/or Asian populations is responsible for high contemporary variation, a genomewide influence would be expected; so under this hypothesis it is puzzling why elevated diversity was not found among all mitochondrial and chloroplast genes. S. vulgaris is widespread throughout Europe and known to hybridize with some closely related sympatric Silene species such as S. uniflora (Runyeon-Lager and Prentice 2000); thus, hybridization and introgression from different species is a potential mechanism generating some diversity. S. vulgaris does not hybridize with highly divergent species such as S. latifolia and S. acaulis, however, so introgression is unlikely to explain the presence and persistence of the extremely divergent haplotypes in S. vulgaris.

Haplotype diversity in atp1 and atp9 is characterized by long internal and short external branches, a pattern consistent with the maintenance of divergent haplotypes over long time periods due to either balancing selection or population structure. As discussed above, it is unlikely that this pattern was generated by population structure. Although atp1 and atp9 genealogies exhibited nonneutral evolutionary patterns, it seems unlikely that the target of selection was either atp1 or atp9 because most polymorphisms in these genes occurred at synonymous sites (Table 1). Instead, allelic variation in mitochondrial genes may be maintained via hitchhiking effects with selection on nearby linked loci. The most likely candidates are CMS genes that theoretical models suggest may be maintained via negative frequency-dependent selection (Frank 1989; Gouyon et al. 1991).

In contrast to our results, studies of chloroplast intergenic spacer regions in S. vulgaris found patterns consistent with epidemic-like selective sweeps (Ingvarsson and Taylor 2002). Specifically, intraspecific variation at chloroplast intergenic spacer regions in S. vulgaris was depauperate relative to that in S. latifolia (a dioecious congener) after calibration against variation at a nuclear locus, consistent with epidemic-like selective sweeps (Ingvarsson and Taylor 2002). This is in opposition to patterns consistent with the long-term maintenance of high diversity in two of the mitochondrial genes that we studied (atp1 and atp9) in S. vulgaris. One explanation for these contrasting results is that chloroplast and mitochondrial genes may have sufficiently independent inheritance that selection in one genome does not influence patterns of diversity in the other.

Congruence, heteroplasmy, and recombination:

If organelles are strictly maternally inherited, we would expect the mitochondria and chloroplast genomes to contain congruent phylogenetic signals (Olson and McCauley 2000). We found an absence of congruence between mitochondrial and chloroplast genes and also between mitochondrial genes. The only pairs of coding regions for which congruence could not be rejected included ndhF or cob, where there was little phylogenetic signal. The absence of congruence illustrates the departure from the current paradigm of organelle gene evolution in S. vulgaris and indicates the presence of at least transient heteroplasmy, non-co-inheritance of organelles, and recombination.

The data reported here contrast with those by Olson and McCauley (2000), which found no evidence of incongruence between chloroplast and mitochondrial markers in S. vulgaris from Virginia. Both studies included individuals from some of the same populations in Virginia and in the current study, Virginia samples exhibited chloroplast–mitochondrial incongruence; thus, the difference in outcomes of the studies did not result from the smaller geographic scale sampled in Olson and McCauley (2000). The other major difference between the studies is that the current study included DNA sequences from coding regions only, whereas Olson and McCauley (2000) included cpDNA noncoding sequences and mtDNA RFLP Southern blots associated with regions flanking the cytochrome oxidase I locus. The RFLP flanking regions and intergenic spacers evolve relatively quickly compared to coding regions, and thus we suspect that each data set is capable of detecting reassociation among the chloroplast and mitochondria at different timescales. Thus, there may be no evidence for recent chloroplast–mitochondria reassociation as shown by Olson and McCauley (2000), but strong evidence for reassociation of the genomes over a longer timescale (shown here).

Our data exhibit strong evidence for the exchange of genetic information among mitochondrial genomes occurring within the timescale of species. Nonetheless, reciprocal recombination between different alleles from mitochondrial homologs must still be a relatively rare event; otherwise recombination would have generated more intermediate allelic variation between the divergent alleles we observed. If the atp1 and atp9 haplotypes are ancient, even low rates of recombination would be preserved in present-day haplotypes. The absence of high diversity and ancient haplotypes in cob is puzzling in light of the high diversity found in S. acaulis (Städler and Delph 2002), but possibly explained by the often-found close proximity of CMS genes and atp subunits in species where CMS genes are well characterized (Hanson and Bentolila 2004). We conjecture that atp1 and atp9 may be more tightly linked than cob to putative CMS genes. The influence of hitchhiking selection decreases rapidly with distance from the target of selection (Hudson and Kaplan 1988; Schierup et al. 2000) and thus may have only weak influence on diversity at cob. This linkage conjecture must be tempered, however, by our incomplete knowledge of the stability of the physical proximity of plant mitochondrial genes over time. For hitchhiking to be an explanatory mechanism, the physical linkage between genes must be maintained in the face of frequent rearrangements to which mitochondrial genomes are prone (Palmer and Herbon 1988).

The larger question of why few angiosperm species that carry CMS have high frequencies of females in populations (which characterizes gynodioecy) remains unresolved. The presence of CMS and associated male-fertility restorers in several agricultural crop species such as maize and rice, despite the absence of gynodioecy in their wild relatives, suggests that the rate of mutation of new CMS genes may not differ considerably among gynodioecious and nongynodioecious species. There appear to be two quite different fates of CMS genes after they arise within a species. In most species, such as the wild relatives of crops, CMS and their associated male-fertility restorer genes probably spread as epidemics, eventually homogenizing the mitochondrial genome within the ultimately hermaphroditic populations. However, in many of the species that we recognize as gynodioecious, a confluence of associations between the fitnesses of CMS types and their restorers, and perhaps even paternal inheritance, allows the long-term maintenance of CMS polymorphism within a species. A comprehensive theory of the complex evolutionary forces that lead to gynodioecy may one day unite our understanding of the evolution of CMS and mitochondrial molecular evolution within all angiosperms and promises to continue as one of the major themes in modern studies of plant breeding system evolution.


We thank D. Charlesworth and H. Štorchová for seeds. D. E. McCauley contributed seeds and inspiration. Accommodations during seed collection forays in Virginia were provided by the Mountain Lake Biological Station. Our manuscript benefited from discussions with D. E. Wolf, N. Takebayashi, the Olson lab, and Toby Bradshaw. Three anonymous reviewers provided useful comments on an earlier version of the manuscript. This work was supported by National Science Foundation grant no. DEB 03-17115 to M.S.O.


Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under accession nos. DQ841746DQ841789.


  • Adams, K. L., Y.-L. Qiu, M. Stoutmeyer and J. D. Palmer, 2002. Punctuated evolution of mitochondrial gene content: high and variable rates of mitochondrial gene loss and transfer to the nucleus during angiosperm evolution. Proc. Natl. Acad. Sci. USA 99: 9905–9912. [PMC free article] [PubMed]
  • Albert, B., B. Godelle, A. Atlan, R. DePaepe and P. H. Gouyon, 1996. Dynamics of plant mitochondrial genome: model of a three-level selection process. Genetics 144: 369–382. [PMC free article] [PubMed]
  • Bergthorsson, U., K. L. Adams, B. Thomason and J. D. Palmer, 2003. Widespread horizontal transfer of mitochondrial genes in flowering plants. Nature 424: 197–201. [PubMed]
  • Birky, C. W., 2001. The inheritance of genes in mitochondrial and chloroplasts: laws, mechanisms and models. Annu. Rev. Genet. 35: 125–148. [PubMed]
  • Charlesworth, D., 1981. A further study of the problem of the maintenance of females in gynodioecious species. Heredity 46: 27–39.
  • Charlesworth, D., and V. Laporte, 1998. The male-sterility polymorphism of Silene vulgaris: analysis of genetic data from two populations and comparison with Thymus vulgaris. Genetics 150: 1267–1282. [PMC free article] [PubMed]
  • Cho, Y., J. P. Mower, Y.-L. Qiu and J. D. Palmer, 2004. Mitochondrial substitution rates are extraordinarily elevated and variable in a genus of flowering plants. Proc. Natl. Acad. Sci. USA 101: 17741–17746. [PMC free article] [PubMed]
  • Darwin, C., 1887. The Different Forms of Flowers on Plants of the Same Species. University of Chicago Press, Chicago.
  • De Haan, A. A., H. P. Koelewijn, M. P. J. Hundsheid and J. M. M. Van Damme, 1997. The dynamics of gynodioecy in Plantago lanceolata L. II. Mode of action and frequencies of restorer alleles. Genetics 147: 1317–1328. [PMC free article] [PubMed]
  • Delanney, X., P.-H. Gouyon and G. Valdeyron, 1981. Mathematical study of the evolution of gynodioecy with cytoplasmic inheritance under the effect of a restorer gene. Genetics 99: 169–181. [PMC free article] [PubMed]
  • Depaulis, F., and M. Veuille, 1998. Neutrality tests based on the distribution of haplotypes under an infinite-site model. Mol. Biol. Evol. 15: 1788–1790. [PubMed]
  • Desplanque, B., F. Viard, J. Bernard, D. Forcioli, P. Saumitou-Laprade et al., 2000. The linkage disequilibrium between cpDNA and mtDNA haplotypes in Beta vulgaris subsp maritima (L.): the usefulness of both genomes for population genetic studies. Mol. Ecol. 9: 141–154. [PubMed]
  • Farris, J. S., M. Källersjö, A. G. Kluge and C. Bult, 1994. Testing significance of incongruence. Cladistics 10: 315–319.
  • Frank, S. A., 1989. The evolutionary dynamics of cytoplasmic male sterility. Am. Nat. 133: 345–376.
  • Fu, Y.-X., and W. H. Li, 1993. Statistical tests of neutrality of mutations. Genetics 133: 693–709. [PMC free article] [PubMed]
  • Gaut, B. S., 1998. Molecular clocks and nucleotide substitution rates in higher plants. Evol. Biol. 30: 93–120.
  • Gouyon, P.-H., F. Vichot and J. M. M. Van Damme, 1991. Nuclear-cytoplasmic male sterility: single-point equilibria versus limit cycles. Am. Nat. 137: 498–514.
  • Gray, M. W., 1996. RNA editing in plant organelles: a fertile field. Proc. Natl. Acad. Sci. USA 93: 8157–8159. [PMC free article] [PubMed]
  • Hanson, M. R., and S. Bentolila, 2004. Interaction of mitochondrial and nuclear genes that affect male gametophyte development. Plant Cell 16: S154–S169. [PMC free article] [PubMed]
  • Hatcher, M. J., 2000. Persistence of selfish genetic elements: population structure and conflict. Trends Ecol. Evol. 15: 271–277. [PubMed]
  • Hudson, R. R., and N. L. Kaplan, 1985. Statistical properties of the number of recombination events in the history of a sample of DNA sequences. Genetics 111: 147–164. [PMC free article] [PubMed]
  • Hudson, R. R., and N. L. Kaplan, 1988. The coalescent process in models with selection and recombination. Genetics 120: 831–840. [PMC free article] [PubMed]
  • Ingvarsson, P. K., 2004. Population subdivision and the Hudson-Kreitman-Aguade test: testing for deviations from the neutral model in organelle genomes Genet. Res. 83: 31–39. [PubMed]
  • Ingvarsson, P. K., and D. R. Taylor, 2002. Genealogical evidence for epidemics of selfish genes. Proc. Natl. Acad. Sci. USA 99: 11265–11269. [PMC free article] [PubMed]
  • Ioerger, T. R., A. G. Clark and T.-H. Kao, 1990. Polymorphism at the self-incompatibility locus in Solanaceae predates speciation. Proc. Natl. Acad. Sci. USA 87: 9732–9735. [PMC free article] [PubMed]
  • Laroche, J., P. Li, L. Maggia and J. Bousquet, 1997. Molecular evolution of angiosperm mitochondrial introns and exons. Proc. Natl. Acad. Sci. USA 94: 5722–5727. [PMC free article] [PubMed]
  • Marsden-Jones, E. M., and W. B. Turrill, 1957. The Bladder Campions. The Ray Society, London.
  • Martin, D., and E. Rybicki, 2000. RDP: detection of recombination amongst aligned sequences. Bioinformatics 16: 562–563. [PubMed]
  • McCauley, D. E., 1998. The genetic structure of a gynodioecious plant: nuclear and cytoplasmic genes. Evolution 52: 255–260.
  • McCauley, D. E., R. A. Smith, J. D. Lisenby and C. Hsieh, 2003. The hierarchical spatial distribution of chloroplast DNA polymorphism across the introduced range of Silene vulgaris. Mol. Ecol. 12: 3227–3235. [PubMed]
  • McCauley, D. E., M. F. Bailey, N. A. Sherman and M. Z. Darnell, 2005. Evidence for paternal transmission and heteroplasmy in the mitochondrial genome of Silene vulgaris, a gynodioecious plant. Heredity 95: 50–58. [PubMed]
  • Mower, J. P., S. Stefanović, G. J. Young and J. D. Palmer, 2004. Gene transfer from parasitic to host plants. Nature 432: 165–166. [PubMed]
  • Nei, M., 1987. Molecular Evolutionary Genetics. Columbia University Press, New York.
  • Nei, M., and T. Gojobori, 1986. Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Mol. Biol. Evol. 3: 418–426. [PubMed]
  • Olson, M. S., and D. E. McCauley, 2000. Linkage disequilibrium and phylogenetic congruence between chloroplast and mitochondrial haplotypes of Silene vulgaris. Proc. R. Soc. Lond. Ser. B 267: 1801–1808. [PMC free article] [PubMed]
  • Olson, M. S., and D. E. McCauley, 2002. Mitochondrial DNA diversity, population structure, and gender association in the gynodioecious plant Silene vulgaris. Evolution 56: 253–262. [PubMed]
  • Olson, M. S., D. E. McCauley and D. Taylor, 2005. Genetics and adaptation in structured populations: sex ratio evolution in Silene vulgaris. Genetica 123: 49–62. [PubMed]
  • Oxelman, B., M. Liden and D. Berglund, 1997. Chloroplast rps16 intron phylogeny of the tribe Sileneae (Caryophyllaceae). Plant Syst. Evol. 206: 393–410.
  • Palmer, J. D., and L. A. Herbon, 1988. Plant mitochondrial-dna evolves rapidly in structure, but slowly in sequence. J. Mol. Evol. 28: 87–97. [PubMed]
  • Parkinson, C. L., J. P. Mower, Y.-L. Qiu, A. J. Shrik, K. Song et al., 2005. Multiple major increases and decreases in mitochondrial substitution rates in the plant family Geraniaceae. BMC Evol. Biol. 5: 73–84. [PMC free article] [PubMed]
  • Pettersson, M. W., 1992. Advantages of being a specialist female in the gynodioecious Silene vulgaris s.l. (Caryophyllaceae). Am. J. Bot. 79: 1389–1395.
  • Posada, D., and K. A. Crandall, 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14: 817–818. [PubMed]
  • Reboud, X., and C. Zeyl, 1994. Organelle inheritance in plants. Heredity 72: 132–140.
  • Richards, A. J., 1997. Plant Breeding Systems. Chapman & Hall, London.
  • Rozas, J., J. C. Sánchez-Delbarrio, X. Messeguer and R. Rozas, 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19: 2496–2497. [PubMed]
  • Roze, D., F. Rousset and Y. Michalakis, 2005. Germline bottlenecks, biparental inheritance and selection on mitochondrial variants: a two-level selection model. Genetics 170: 1385–1399. [PMC free article] [PubMed]
  • Runyeon-Lager, H., and H. C. Prentice, 2000. Morphometric variation in a hybrid zone between the weed, Silene vulgaris, and the endemic, Silene uniflora ssp. Petraea (Caryophyllaceae), on the Baltic island of Öland. Can. J. Bot. 78: 1384–1397.
  • Saumitou-Laprade, P., J. Cuguen and P. Vernet, 1994. Cytoplasmic male sterility in plants—molecular evidence and the nucleocytoplasmic conflict. Trends Ecol. Evol. 9: 431–435. [PubMed]
  • Schierup, M. H., D. Charlesworth and X. Vekemans, 2000. The effect of subdivision on variation at multi-allelic loci under balancing selection. Genet. Res. 76: 63–73. [PubMed]
  • Schnable, P. S., and R. P. Wise, 1998. The molecular basis of cytoplasmic male sterility and fertility restoration. Trends Plant Sci. 3: 175–180.
  • Städler, T., and L. Delph, 2002. Ancient mitochondrial haplotypes and evidence for intragenic recombination in a gynodioecious plant. Proc. Natl. Acad. Sci. USA 99: 11730–11735. [PMC free article] [PubMed]
  • Štorchová, H., and M. S. Olson, 2004. Comparison between mitochondrial and chloroplast DNA variation in the native range of Silene vulgaris. Mol. Ecol. 13: 2909–2919. [PubMed]
  • Swofford, D. L., 2003. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods), Version 4. Sinauer Associates, Sunderland, MA.
  • Tajima, F., 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123: 585–595. [PMC free article] [PubMed]
  • Uyenoyama, M. K., 1995. A generalized least-squares estimate for the origin of sporophytic self-incompatibility. Genetics 139: 975–992. [PMC free article] [PubMed]
  • Wade, M. J., and D. E. McCauley, 2005. Paternal leakage sustains the cytoplasmic polymorphism underlying gynodioecy but remains invasible by nuclear restorers. Am. Nat. 166: 592–602. [PubMed]
  • Watterson, G. A., 1975. On the number of segregating sites in genetical models without recombination. Theor. Popul. Biol. 7: 256–276. [PubMed]
  • Wolfe, K. H., W.-H. Li and P. M. Sharp, 1987. Rates of nucleotide substitution vary greatly among plant mitochondrial, chloroplast and nuclear DNAs. Proc. Natl. Acad. Sci. USA 84: 9054–9058. [PMC free article] [PubMed]
  • Wright, S. I., and B. Charlesworth, 2004. The HKA test revisited: a maximum-likelihood-ratio test of the standard neutral model. Genetics 168: 1071–1076. [PMC free article] [PubMed]

Articles from Genetics are provided here courtesy of Genetics Society of America
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Gene (nucleotide)
    Gene (nucleotide)
    Records in Gene identified from shared sequence links
  • MedGen
    Related information in MedGen
  • Nucleotide
    Published Nucleotide sequences
  • PopSet
    Published population set
  • Protein
    Published protein sequences
  • PubMed
    PubMed citations for these articles
  • Taxonomy
    Related taxonomy entry
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...