Skip to main content

Time and time again: unisexual salamanders (genus Ambystoma) are the oldest unisexual vertebrates



The age of unisexual salamanders of the genus Ambystoma is contentious. Recent and ancient evolutionary histories of unisexual Ambystoma were proposed by a few separate studies that constructed phylogenies using mitochondrial DNA markers (cytochrome b gene vs. non-coding region). In contrast to other studies showing that unisexual Ambystoma represent the most ancient unisexual vertebrates, a recent study by Robertson et al. suggests that this lineage has a very recent origin of less than 25,000 years ago.


We re-examined the phylogenetic relationship of the unisexuals to A. barbouri from various populations using both mitochondrial markers as well as the complete mitochondrial genomes of A. barbouri and a unisexual individual from Kentucky. Lineage dating was conducted using BEAST and MultiDivTime on a complete mitochondrial genome phylogeny. Our results support a monophyletic lineage for unisexual Ambystoma that shares its most recent common ancestor with an A. barbouri lineage from western Kentucky. In contrast to the Robertson et al.'s study, no A. barbouri individual shared an identical or almost identical cytochrome b haplotype with any unisexual. Molecular dating supports an early Pliocene origin for the unisexual linage (~5 million years ago). We propose that a unisexual-like cytochrome b numt (or pseudogene) exists in the controversial A. barbouri individuals from Kentucky, which was likely the cause of an erroneous phylogeny and time estimate in Robertson et al.'s study.


We reject a recent origin of unisexual Ambystoma and provide strong evidence that unisexual Ambystoma are the most ancient unisexual vertebrates known to exist. The likely presence of an ancient cytochrome b numt in some Kentucky A. barbouri represents a molecular "fossil" reinforcing the hypothesis that these individuals are some of the closest extant relatives to unisexual Ambystoma.


Because of the absence of sex, a brief evolutionary lifespan is generally expected for unisexual and asexual organisms [1]. Nevertheless, ancient unisexuals and asexuals that persist millions of years have been discovered in various taxa among plants, fungi and animals [2]. With recent advances in molecular genetics and phylogenetics, our knowledge of reproductive systems and evolutionary histories of many unisexual and asexual lineages has been quickly improved. Recent evidence reveals that many unisexuals are capable of utilizing modified reproductive modes and/or mitotic or meiotic mechanisms to incorporate "a bit of sex" [37], or they can incorporate additional genetic material to compensate for the suspected lethal effects caused by the accumulation of deleterious mutations [710]. Thus, although populations may be all female, unisexual are not necessarily equivalent to asexual populations [5]. The discovery of ancient unisexual and asexual lineages not only poses a dilemma for evolutionary theoreticians but also provides an opportunity to address questions that relate to the prevalence and maintenance of sexual reproduction.

Unisexuality in vertebrates has been discovered in about 90 lineages of fresh water fish, amphibians and reptiles [11], most of which are recently spun off from sexual relatives via interspecific hybridization [7, 12]. North American unisexual mole salamanders of the genus Ambystoma co-evolve with five distinct sexual ambystomatids (A. laterale, A. jeffersonianum, A. tigrinum, A. texanum, and A. barbouri) across the entire unisexual distribution [9, 1315]. Unisexual Ambystoma persist as a "parasitic entity" by stealing and incorporating sperm from sympatric sexual species via a complex reproductive mode, kleptogenesis, to generate nearly 30 genomic combinations or biotypes, with ploidy levels ranging from diploid to pentaploid [5, 15]. Genomes in unisexuals may not be transmitted unaltered. Recent studies using genomic in situ hybridization (GISH) demonstrate complex intergenomic exchanges in unisexual populations [4, 1618].

Despite the complexity of their nuclear genomes, all unisexuals contain a highly conserved mitochondrial genome which is derived from A. barbouri, suggesting that A. barbouri may be the maternal ancestor of the unisexual lineage [5, 9, 19]. Although the ancestry of unisexual Ambystoma is not contested by these studies, the age of the unisexual lineage is controversial. Unisexual Ambystoma have been proposed as the most ancient unisexual vertebrates known to exist [5, 20, 21]. Based on a phylogeny constructed by the mitochondrial intergenic spacer and control region (mitochondrial non-coding region or NCR), Bogart et al. [5] estimated that the unisexual lineage and the closest relatives, an A. barbouri lineage from Kentucky, are descended from the most recent common ancestor 2.4-3.9 million years ago (Ma). Their results are based on the observation of a 3.91% pairwise difference in the control region between unisexuals and a few Kentucky A. barbouri individuals. On the contrary, Robertson et al. [19] constructed a phylogeny using a different mitochondrial marker cytochrome b (cyt-b) gene and suggested that unisexual Ambystoma have a very recent origin which could be less than 25,000 years ago. Strikingly, they found a few A. barbouri specimens that were also used by Bogart et al. [5] to have an identical or almost identical cyt-b haplotype to unisexual individuals. Such a large discrepancy is difficult to understand because the two different mitochondrial DNA markers (NCR vs. cyt-b) that came from the same A. barbouri individuals demonstrated distinctly different evolutionary relationships to the unisexual lineage (~3.91% vs. ~0%). Therefore, the question: "are unisexual Ambystoma ancient or recent?" is unresolved.

To provide a clearer answer to this question, we re-examined the phylogenetic relationship of unisexuals and A. barbouri from various populations using both mitochondrial cyt-b gene and NCR as markers. We especially focused on A. barbouri samples (JPB34337, JPB34342, JPB34343, JPB34356) that demonstrated distinctly different sequence divergences to the unisexuals in the two studies [5, 19]. Given that the rate of substitution likely varies for different genes/regions in the mitochondrial genome, we compared the substitution rates of complete mitochondrial genomes (mtgenomes) of unisexual and A. barbouri. We sequenced the mtgenome of a Kentucky A. barbouri specimen (JPB34342) that was used in both studies and provided conflicting results. We chose to sequence the mtgenome of one tetraploid unisexual A. laterale - 3 jeffersonianum (or LJJJ) from a northern Kentucky population as a representative unisexual. This individual was chosen because it was found to have the most common mitochondrial NCR haplotype among unisexuals (haplotype B in [5]), it was collected in a state (Kentucky) where unisexuals were previously unknown, and it was geographically close to the A. barbouri individuals that were deemed to be the closest relatives to unisexual Ambystoma. For comparison, we also sequenced mtgenomes of one A. barbouri and one A. texanum from Ohio to examine their phylogenetic relationships to the Kentucky unisexual and A. barbouri individuals. The time to the most recent common ancestor (TMRCA) for unisexual lineage and its closest relative was re-calculated based on the mtgenome phylogeny.


Mitochondrial cyt-b and NCR trees

Primers Glu14100L, MLM651 and M2R (abbreviated as GMM) were used to amplify and sequence a ~ 2200 base pairs (bp) mitochondrial fragment that included cyt-b and NCR. A total of 46 cyt-b sequences, including one from an outgroup species (Ambystoma laterale) downloaded from GenBank, were compared (Table 1). Of the 1141 bp resolved, 228 sites were variable and 143 sites were phylogenetically informative. A total of 42053 most parsimonious trees resulted from the parsimony analysis with 335 steps, a consistency index (CI) of 0.758 and a retention index (RI) of 0.939. For NCR, 46 sequences including one A. laterale from GenBank were recovered (Table 1). Of the 1065 bp resolved, 215 sites were variable and 111 sites were phylogenetically informative. A total of 90 most parsimonious trees were obtained from the parsimony analysis with 310 steps, a CI of 0.768 and a RI of 0.944. For both markers, a GTR+I+G model was selected as the best-fit model. The cyt-b strict consensus tree (not shown) was nearly identical to the Bayesian tree, but NCR strict consensus (not shown) and Bayesian trees yielded slightly different topologies. The Bayesian trees with the BPP and BSP from the parsimony analyses are presented in Figure 1.

Table 1 Specimens used for phylogenetic construction by mitochondrial cyt-b and NCR in this study with GenBank accession numbers.
Figure 1

Bayesian gene trees of mitochondrial cytochrome b (cyt-b) gene from Robertson et al. [19][A] and the present study [B], and Bayesian gene tree of non-coding region (NCR) by the present study [C]. The tree arrangement from Robertson et al. [19] has been slightly modified. Left numbers along branches represent the Bayesian posterior probabilities (BPP) and right numbers represent bootstrap proportions (BSP) derived from the parsimony analysis. The detailed information of specimens that were sequenced is listed in Table 1. *A unisexual-like cyt-b numt likely exists in A. barbouri specimens JPB34337, JPB34342, JPB34343 and JPB34356 [A] so they were grouped within the unisexual clade.

Both cyt-b and NCR trees clustered all specimens into four clades (Figure 1). All the unisexuals, irrespective of their biotypes and ploidy levels, formed a monophyletic clade A. Clade A was sister to clade B that contained A. barbouri individuals from southern Ohio and west of the Kentucky Rivers in central Kentucky. This clade corresponded to the A. barbouri "western clade" in Niedzwiecki's study [22]. Clade C contained A. barbouri from north or east of the Kentucky River and north of the Ohio River, as well as the disjunct populations in western Kentucky. This clade corresponded to the "northern clade" that was identified by Niedzwiecki [22]. Clade D included A. texanum and A. barbouri from Tennessee. In the NCR tree, the relationship of clade D in respect to three other clades was, however, not well resolved but the monophyly of clade A+ clade B was strongly supported (BSP = 99; BPP = 100).

In Robertson et al.'s study [19], four specimens (JPB34337, JPB34342, JPB34343, JPB34356) of A. barbouri were found to be clustered in the unisexual clade and three of them (JPB34337, JPB34342, JPB34343) shared the same cyt-b haplotype with some of the unisexuals (Figure 1A). We sequenced the cyt-b gene from the very same individuals but we found that none of them shared an identical or almost identical haplotype to any unisexual. We examined the sequence chromatogram of cyt-b in these individuals and only single signal peaks were observed at every site. All four samples were grouped in clade B with other A. barbouri where the average cyt-b sequence divergence between A. barbouri and unisexuals was 5.16% (Figure 1B). Likewise, no A. barbouri shared the same or even a very similar NCR sequence with unisexuals (Figure 1C), and the ingroup relationship in the NCR tree was concordant with that obtained by Bogart et al. [5].

Amplification of cyt-b gene by MVZ15 and MVZ16

Bogart et al. [5] used primers MLMTHR and MLM651 to obtain a ~1100 bp intergenic spacer and control region from unisexual and A. barbouri specimens. We compared our sequencing results using the same primers and confirmed that MLMTHR and MLM651 produced the identical sequences to those by using primers GMM. Primers MVZ15 and MVZ16 were used by Robertson et al. [19] to amplify an ~800 bp cyt-b fragment from ambystomatids. In their study, A. barbouri individuals JPB34337, JPB34342, JPB34343 and JPB34356 were found to contain an identical or almost identical haplotype to unisexuals which placed them in the unisexual clade. Using the same primers, we failed to duplicate the same sequencing results in JPB34342 and JPB34343 as discovered by Robertson et al. [19]. Cyt-b failed to amplify in JPB34337 when using MVZ15 and MVZ16 but primers GMM confirmed that JPB34337 did not contain a unisexual haplotype. The results showed that these three individuals shared A. barbouri cyt-b haplotypes that were grouped with other A. barbouri in clade B (Figure 1B). We found, however, that JPB34356 did have a unisexual-like cyt-b haplotype as was found by Robertson et al. [19] using MVZ15 and MVZ16 as PCR and sequencing primers. This contradictory result was unexpected. We further examined the sequence chromatogram of JPB34356 and found that its sequence contained many heterozygous sites (multiple signal peaks) in positions where the variable sites between unisexuals and clade B A. barbouri were located (Figure 2). At these heterozygous sites, A. barbouri's sequence signals were generally weaker than unisexual-like sequence signals so the latter would be read by the program by default (Figure 2). We ruled out the possibility of DNA cross contamination and mitochondrial heteroplasmy in JPB34356 because, using primers GMM which targeted a much longer fragment for PCR amplification and sequencing, we consistently found a pure A. barbouri haplotype from JPB34356. In the sequence chromatogram, unisexual-like sequence signals were no longer detectable (Figure 2A). We used various primer combinations to amplify mitochondrial genes of JPB34342 using a new DNA extraction in the present study as well as an old DNA extraction used by Robertson et al. [19]. None of the amplicons showed any signs of DNA contamination or mitochondrial heteroplasmy. Additionally, through sequencing the mtgenome of A. barbouri JPB34342, unisexual JPB39932 and other specimens we showed that cyt-b gene did not have any duplicated component residing in the mitochondrial genome. A plausible explanation was that JPB34356 contained a nuclear copy (or copies) of a unisexual-like mitochondrial cyt-b gene fragment (numt or pseudogene). When using whole genomic DNA as template, both the actual mitochondrial cyt-b gene and its nuclear numt could have been co-amplified by MVZ15 and MVZ16. To further test this prediction we used the mitochondrial DNA fragment as a template and conducted sequencing PCR using MVZ15 and MVZ16 as primers. The results showed that no heterozygous sites existed in the sequence chromatogram (Figure 2A) and JPB34356 did not contain a unisexual-like cyt-b haplotype in its mitochondrial genome. Another primer combination using MVZ15 and MLM651 (targeted a ~2200 bp fragment too) was also used to amplify cyt-b using JPB34356 whole genomic DNA as template and no unisexual-like signals were detected in that sequence chromatogram.

Figure 2

Cyt-b fragment chromatograms showing the presence of unisexual-like cyt-b numt in Ambystoma barbouri individual JPB34356. [A] Amplification and sequencing of cyt-b from JPB34356 using primers GMM, or using MVZ15/16 with mitochondrial DNA as template. [B] Amplification and sequencing of cyt-b from JPB34356 using primers MVZ15/16 with whole genomic DNA as template. [C] Amplification and sequencing of cyt-b in unisexual individual JPB30066, either using primers GMM or MVZ15/16 with whole genomic DNA as template. Arrows point to four sites that differ between A. barbouri and unisexual Ambystoma.

Mtgenome phylogeny

A total of 17 mtgenome sequences, including 13 downloaded from GenBank, were used for phylogenetic construction (Table 2). The combined DNA dataset contained 9647 bp with 3464 variable sites and 2361 were phylogenetically informative. Parsimony analysis generated a single tree 8803 steps in length (CI = 0.569, RI = 0.614). For Bayesian analysis, a GTR+I+G model was selected as the best-fit model. The tree topologies derived from parsimony and Bayesian analyses were similar, with slight differences in the relationships of some outgroup species, and we present the Bayesian tree in Figure 3. The monophyly of Kentucky A. barbouri JPB34342 and unisexual JPB39932 was strongly supported by both analyses (BSP = 100; BPP = 100).

Table 2 Species and unisexual biotype LJJJ used for mtgenome phylogeny and molecular dating analyses in this study with their GenBank accession numbers.
Figure 3

A close phylogenetic relationship of unisexual Ambystoma (JPB39932) and Kentucky A. barbouri (JPB34342) inferred from mtgenome phylogeny. Other species are incorporated to introduce calibration points for molecular dating analyses, and to serve as outgroups. Left numbers along branches represent the Bayesian posterior probabilities (BPP) and right numbers represent bootstrap proportions (BSP) derived from the parsimony analysis. BPP and BSP below 50 are not shown. Branch lengths are estimated by the Bayesian inference.

The overall sequence divergence of mtgenome between the unisexual sample JPB39932 and A. barbouri JPB34342 was 4.42% (Table 3). We divided the mtgenome into 18 partitions and the pairwise difference of each partition ranged from 2.27% (concatenated tRNAs) to 7.03% (intergenic spacer). All 13 protein-coding genes were rather variable between genomes of JPB39932 and JPB34342. Except for ND4L (3.03%) and ATP8 (3.57%), the pairwise differences of all other genes were no less than 4.50%. JPB34342 was found to have an identical cyt-b haplotype to unisexuals by Robertson et al. [19]. Our study clearly demonstrated that neither the cyt-b gene nor any other genes, tRNAs or NCR throughout the mitochondrial genome were the same between A. barbouri JPB34342 and the unisexual.

Table 3 Sequence pairwise divergence of 18 genes/regions of mtgenome between the unisexual sample JPB39932 and A. barbouri JPB34342

Molecular dating

The relaxed lognormal clock analysis of the mtgenome sequences by BEAST produced the same topology as the Bayesian analysis. The divergence times estimated by BEAST and MultiDivTime are listed in Table 4. In general, the time estimates by BEAST and MultiDivTime were largely congruent. We present a time-calibrated tree from BEAST in Figure 4. The split between unisexual Ambystoma and Kentucky A. barbouri (Node A) took place about 5.3 Ma (CI 2.4, 8.7) by BEAST, and 5.1 Ma (CI 2.7, 9.0) by MultiDivTime. Both analyses agreed that the origin of the unisexual linage may date back to early Pliocene.

Table 4 Divergence time means and 95% confidence intervals calculated by BEAST and MultiDivTime*.
Figure 4

A time-calibrated phylogeny inferring the origin of unisexual Ambystoma , fitted to a geological timescale. Times for nodes [A-F] are estimated by BEAST. The horizontal gray bars through these nodes indicate 95% credibility intervals from the BEAST analyses. Non-ambystomatid specimens are not shown in the tree. Detailed time estimates can be found in Table 4. An early Pliocene origin for unisexual Ambystoma is proposed.


No Ambystoma barbourishares an identical mitochondrial cyt-b gene with unisexuals

The phylogenies generated using cyt-b, NCR as well as mtgenome all support the hypothesis that unisexual Ambystoma form a monophyletic group and share the most recent common ancestor with an A. barbouri lineage from Kentucky [5, 9, 19]. The age of unisexual Ambystoma was in high disagreement because four A. barbouri individuals (JPB34337, JPB34342, JPB34343, JPB34356) examined by two previous studies demonstrated distinctly different phylogenetic relationships to the unisexual lineage. Robertson et al. [19] found that these four A. barbouri were grouped with the unisexual cyt-b clade that suggested a very recent origin of unisexual lineage which was less than 25,000 years ago. On the other hand, Bogart et al. [5] found a 3.91% sequence pairwise distance in the control region between unisexuals and a few Kentucky A. barbouri individuals (including JPB34342, JPB34343, JPB34356) and claimed an ancient evolutionary history of unisexual Ambystoma, approximately 2.4-3.9 Ma. When using primers targeting a long mitochondrial fragment (~2200 bp), we found that no A. barbouri, including the four controversial A. barbouri individuals used by Robertson et al. [19], shared the same cyt-b gene sequence with unisexuals. The average cyt-b sequence pairwise distance between unisexuals (clade A) and their closest A. barbouri relatives (clade B) is as high as 5.16%. The examination of mtgenome of the representative specimens further demonstrated that neither the cyt-b gene, nor other genes/regions in the mitochondrial genome were identical or almost identical between unisexuals and A. barbouri. The overall substitution rates throughout the mitochondrial genome, especially in protein coding genes, were found to be rather consistent. Through a thorough sampling of A. barbouri across its distribution, Niedzwiecki (personal communication) did not find any haplotype that was more similar to the unisexuals than western clade Kentucky A. barbouri including specimens JPB34337, JPB34342, JPB34343 and JPB34356. Therefore, all our evidence clearly demonstrates that there is no known A. barbouri that shares an identical cyt-b gene with unisexuals.

Molecular dating demonstrates an ancient origin of unisexual Ambystoma

In our BEAST analysis, the rate covariance among adjacent branches was 0.13 which was close to zero and their 95% confident intervals spanned zero (CI -0.20, 0.46). This result suggests that rate autocorrelation is insignificant [23] and BEAST (without rate-autocorrelation assumption), rather than MultiDivTime (with rate-autocorrelation assumption), would be more appropriate for the data and calibration choices in our study. Nevertheless, our results show that the time estimations between BEAST and MultiDivTime are largely congruent, especially within the ambystomatid salamanders (Table 4). Both analyses show that unisexual Ambystoma and Kentucky A. barbouri descended from their most recent common ancestor about 5 Ma so this dichotomy likely occurred in early Pliocene. More conservatively, if we take the lowest value at 95% CI inferred by both analyses, the unisexual lineage originated at least 2 Ma (early Pleistocene). Generally, our estimate was consistent with previous time estimates of the origin of unisexual Ambystoma [5, 20, 21] which validates them as the most ancient unisexual vertebrate lineage known to exist [2].

Niedzwiecki [22] suggested that the Kentucky and Ohio Rivers likely act as barriers between western and northern A. barbouri mtDNA clades (corresponding respectively, to clades B and C in our study, Figures 1B & 1C). The Kentucky River might have been responsible for the divergence between A. barbouri from A. texanum as well as for the divergence of geographically distinct A. barbouri mtDNA lineages. Using a rough calibration approach (1.5% per million years for the A. tigrinum complex) [24], Niedzwiecki [22] estimated that various A. barbouri mitochondrial lineages were separated around 3.5-5 Ma and a similar time would have elapsed since the split between the main A. barbouri clades and A. texanum. Our study suggests that both divergence times are possibly more ancient than these estimates. In the BEAST analysis, for example, the separation between Kentucky A. barbouri (represented by clade B, Figures 1B & 1C) and Ohio A. barbouri (represented by clade C, Figures 1B & 1C) likely took place 8.1 Ma (CI 4.4, 12.4). The split of western and northern A. barbouri from A. texanum likely occurred 10.4 Ma (CI 6.0, 15.4). Our results indicate that both events may have taken place in late to middle Miocene or at least in early Pliocene, suggesting a longer evolutionary history for A. barbouri than expected.

Is a unisexual-like cyt-b numt present in A. barbouri?

Nuclear copies of mitochondrial genes or numts, which evolve independently as paralogous copies of the original mitochondrial DNA segment, are relatively common and have been reported in various taxa [2530]. Failure to discriminate real mitochondrial sequences and numts often confuses the genetic diversity of species and produces erroneous phylogenies [31]. Because the nuclear genome has an overall slower mutation rate than the mitochondrial genome, estimates on divergence times can be problematic [28, 32]. Numts can originate from all parts of the mitochondrial genome including the cyt-b gene [33] and the majority of numts are fairly short sequences [34].

When using whole genomic DNA as template and primers MVZ15 and MVZ16 for PCR amplification and sequencing [19], A. barbouri individual JPB34356 possessed two different PCR products. One of the two was not detectable in the sequence chromatograms when using mitochondrial DNA as template with MVZ15/MVZ16 as sequencing primers, or using total genomic DNA as template with other various primer combinations. DNA cross contamination can yield sequence heterozygosity but contamination should have affected each gene/region across the entire mtgenome. We tested the old DNA extractions of A. barbouri individuals (JPB34337, JPB34342, JPB34343, JPB34356) used by Robertson et al. [19], and we did not find any signs of DNA contamination in any other genes/regions. Mitochondrial heteroplasmy may also be responsible for mixed mitochondrial amplicons [35, 36]. The possibility that these A. barbouri individuals embrace A. barbouri and complete unisexual-like mtgenomes is minimal as proven by the experiments we conducted above. Mitochondrial heteroplasmy could also be caused by regional mutations or duplications of some mitochondrial copies [37, 38]. If these A. barbouri individuals have two types of mtgenomes which are heteroplasmic only in the cyt-b gene, we should have detected the heterozygous signals in the cyt-b gene when using old/new total genomic DNA extractions or mitochondrial DNA as templates with any primer combinations for PCR and sequencing. The most reasonable explanation is that A. barbouri individual JPB34356 contains a unisexual-like cyt-b numt. This unexpected discovery in JPB34356 has resolved the mystery and provides a reasonable explanation for the data that were obtained by Robertson et al. [19]. We believe that a unisexual-like cyt-b numt was present in some A. barbouri individuals, including JPB34337, JPB34342, JPB34343, and JPB34356. It is very possible, in their study, that the cyt-b sequences recovered from these four A. barbouri were in fact unisexual-like numts. Therefore, the subsequent phylogenetic construction and divergence time estimates that were based on these putative numts were erroneous. Although we did not detect stop codons in these numts, it is known that numts do not always include stop codons or indels and sometimes are indistinguishable to their mitochondrial orthologs [35]. It is possible, however, that the stop codons were situated up/downstream beyond the region amplified by the MVZ15/MVZ16 primers that we used. Whether numts can be amplified may also depend on the quality and quantity of nuclear DNA template in the PCR reaction. In our study we failed to find or amplify the same sequences in three other controversial specimens JPB34337, JPB34342 and JPB34343, which likely resulted from the more highly degraded nuclear genomic DNA from these samples.

Implications of having an ancestral cyt-b numt in A. barbouri

Dating the origin of unisexual/asexual lineages is usually based on the extent of genetic divergence from their nearest sexual relatives [7, 39]. In general, time estimates are more robust if both mitochondrial and nuclear genomic markers are used as inferences. Uncoupling evolutionary trajectories between their mitochondrial and nuclear genomes, however, make molecular dating of unisexual Ambystoma depend solely on the mitochondrial genome. All unisexual Ambystoma share one mitochondrial origin but the evolutionary histories of their nuclear genomes are much more complicated and dynamic. It is clear that no recurrent hybridization between any of the sexual sperm donors produces new unisexual lineages [5, 9] otherwise unisexual individuals would contain a mitochondrial genome derived from one or more of the extant sexual, sperm - donating species of Ambystoma. Bogart et al. [15] showed that A. barbouri can serve as a sperm donor for unisexuals in a single pond from southern Ohio. The cyt-b and NCR haplotypes recovered from those A. barbouri were found to be grouped in clade C (Figures 1B & 1C), and are distantly related to the unisexuals. Without recurrent hybridization, unisexual individuals have evolved an extremely flexible reproductive mode by which they rely on sexual sperm donors to perpetuate. Unisexuals replace their nuclear genomes with those from sexual sperm donors, historically and contemporarily, throughout their entire distribution [40]. Consequently, it is impossible to use nuclear genome sequences for molecular dating of unisexual Ambystoma.

Little and Hebert [41] have criticized the traditional dating strategy used for unisexual and asexual organisms [39] because dating, not only requires a thorough search of extant taxa, but also related species that may have gone extinct since unisexuality/asexuality originated. Neiman et al. [42] reviewed asexual lineage longevity among invertebrate and vertebrate taxa and criticized the notion that there existed young and "ancient" asexual lineages. They found that the distribution of asexual lineages followed a regular and linear age distribution. The tabulated ages of asexual vertebrate lineages (fish, amphibians, and reptiles) were all less than 300,000 years old and could even be considered young compared with the estimated age of a sexual species. The unisexual Ambystoma lineage was included in their table of asexual lineages with an age <25,000 years based on the incorrect time estimate provided by Robertson et al. [19]. Our present study demonstrates that unisexual Ambystoma have persisted for about 5 million years and we attribute this longevity to the fact that they are unisexual, not asexual, and have a unique reproductive mode (kleptogenesis)[5].

Because A. barbouri may have an ancient evolutionary history [[22], this study], it is reasonable to assume that there may have existed a closer common ancestor to the unisexuals which became extinct at some unknown time [5]. Finding a unisexual-like cyt-b numt in some A. barbouri individuals suggests that its origination must have taken place at about the same time as the hybridization event that gave rise to unisexuals. Once a numt is established in its descendants' nuclear DNA, it would be much more conserved than the mitochondrial haplotype because the mitochondrial cyt-b mutates faster than its numt. The presence of an ancestral cyt-b numt in the western Kentucky clade of A. barbouri would also suggest that no other population of A. barbouri, including any potential extinct lineages, needs to be considered more closely related to the unisexuals.


Using multiple sets of empirical evidence and rigorous statistical methodologies, we reject the conclusion of a recent origin of unisexual Ambystoma and support the hypothesis that unisexual Ambystoma is the most ancient lineage in unisexual vertebrates known to exist. The amplification of unisexual-like cyt-b sequence in a few Kentucky A. barbouri individuals is the possible cause of the erroneous phylogeny and time estimate determined by Robertson et al. [19]. The unisexual-like cyt-b sequence, presumably a numt, in a few Kentucky A. barbouri could represent a useful molecular "fossil" showing that the A. barbouri lineage from Kentucky represents the closest relatives to the unisexuals. Interestingly, our study shows that the unisexuals' mitochondrial genomes, which are descended from ancestral A. barbouri, seem to have changed little over time because their cyt-b genes are very similar to the ~800 bp ancestral cyt-b numt amplified sequence identified in some extant A. barbouri individuals. Low mitochondrial DNA variation among present unisexual populations have been described before [5, 20] and it might be attributed to population bottlenecking of unisexuals prior to their rapid expansion after the Last Glacial Maximum. The length, number of copies and distribution of the cyt-b numt in A. barbouri is not yet known and there may be other A. barbouri individuals that have similar cyt-b numts and/or other ancient signatures shared with unisexuals. The confirmation of an ancient ancestry of unisexual Ambystoma validates them as an excellent model system for studying the evolution and maintenance of unisexuality in vertebrates.



Ambystoma barbouri from various populations from Kentucky, Ohio and Tennessee were examined. Because unisexuals were found to share a highly conserved mitochondrial genome in previous studies [5, 9, 19, 20], we only sequenced a few unisexual individuals from several geographically distant populations to represent the unisexual lineage. The four specimens that were used for sequencing of their mtgenomes were: A. barbouri from Kentucky (JPB34342), unisexual LJJJ from Kentucky (JPB39932), A. barbouri from Ohio (JPB37710), and A. texanum from Ohio (JPB37892). To introduce calibration points into the molecular dating analyses, mtgenome sequences of 12 salamander species were retrieved from GenBank [4347]: A. californiense, A. laterale, A. tigrinum, Cryptobranchus alleganiensis, Cynops cyanurus, Dicamptodon atterimus, Euproctus platycephalus, Hynobius amjiensis, Notophthalmus viridescens, Paramesotriton laoensis, Taricha rivularis, and Triturus cristatus. A frog, Xenopus tropicalis, was used as an outgroup. Its mtgenome was also downloaded from GenBank. Detailed sampling and sequence information is listed in Tables 1 and 2.

Laboratory protocols

Total genomic DNA was extracted from frozen larvae, adult muscle, heart or liver tissues using a standard phenol-chloroform extraction method. We also tested some DNA extractions used by previous studies [5, 19]. A combination of 22 primers [45] was used to amplify contiguous and overlapping fragments that covered the entire mitochondrial genome. PCR was conducted in a 25 μl mix including 30-50 ng of template DNA, 1 U Taq DNA polymerase (TaKaRa), 1× PCR buffer, 1.5 mM MgCl2, 0.2 mM of each dNTP, and 10 pmol of each primer. PCR cycling parameters were 95°C for 5 min as initial denaturation, followed by 30 cycles of 95°C for 30 sec, 45-50°C [45] for 45 sec, 72°C for 30 sec to 2 min depending on the expected size of fragments (approx. 1 min/kb), and a final step at 72°C for 5 min. PCR products were verified on 1% agarose gels and purified using QIAquick PCR purification kits (Qiagen). Sequencing was performed with the corresponding PCR primers using BigDye 3.1 terminator sequencing chemistry with an ABI 3730 (Applied Biosystems). Sequences were assayed using Sequencher (version 4.5; Gene Codes). For some large PCR fragments that were more than 1600 bp, specific internal primers were designed to obtain their complete sequences.

Primers Glu14100L [45] and MLM651 [24] were used to amplify a ~2200 bp mitochondrial fragment that included cyt-b, intergenic spacer and control region. An internal primer M2R (reverse primer, 5'-GTTGGTGGTTTCTCGCCCTAAG-3') was designed for sequencing across the complete fragment. As a comparison, we also used the same primers MVZ15 and MVZ16 [48] that were used to amplify cyt-b genes by Robertson et al. [19] to examine some of the A. barbouri individuals, especially those that had an identical or almost identical cyt-b haplotype to unisexuals. These two primers targeted an ~800 bp partial cyt-b gene fragment [19]. Likewise, we used the same primers MLMTHR and MLM651 [24] that were used by Bogart et al. [5] to amplify NCRs from some specimens and the expected size of PCR products was approximately 1100 bp [5]. The PCR annealing temperature was set to 50°C for MVZ15 and MVZ16, and was set to 46°C for MLMTHR and ML651. The protocols of standard PCR reaction, purification, sequencing and sequence alignment were the same (above).

Phylogenetic methods

Three phylogenies were constructed using mtgenome, cyt-b, and NCR, respectively. When using mtgenomes for phylogenetic construction, we excluded all the tRNAs and non-coding regions and only included two rRNAs and 13 protein-coding genes. This was necessary because of the absence of several tRNAs and the lack of the intergenic spacer in mtgenomes of some outgroup species as well as the difficulty in sequence alignment of control regions. With the 13 protein-coding genes, the third codon positions were eliminated because of high substitution rates. Multiple substitutions likely produce noise in phylogenetic and dating analyses [45]. Ambiguous alignments in the two rRNA regions were also excluded. Additionally, all the gaps in the alignments were eliminated manually. Finally, a DNA dataset containing all 17 DNA alignments (each with two rRNAs and 13 protein-coding genes without third codon positions) was generated.

A maximum parsimony analysis was conducted using PAUP* (version 4.01b10) [49]. Each sequence was treated as a taxon and each nucleotide was treated as a character. All characters were weighted equally and unordered. A heuristic search method via tree-bisection-reconnection (TBR) branch swapping was used. Bootstrap proportions (BSP) [50] with 1,000 replicates were used to evaluate the nodal support. A Bayesian analysis was conducted using MrBayes (version 3.1) [51]. Model selection was based on the Akaike information criterion (AIC) as implemented in MrModeltest (version 2.2) [52]. The best-fit model was used in subsequent Bayesian phylogenetic analysis. Four Markov chains were used and the dataset was run for 10,000,000 generations to allow adequate time for convergence. Trees were sampled every 500 generations, and the last 5,000 trees were used to estimate the consensus tree and the Bayesian posterior probabilities (BPP). The overall sequence pairwise divergence between mtgenome of Kentucky A. barbouri and the unisexual was calculated using MEGA (version 4.0) [53]. The mtgenomes were then divided into 18 partitions according to genes and regions (one concatenated tRNAs, two rRNAs, two non-coding regions and 13 protein-coding genes) and sequence pairwise difference of each partition between Kentucky A. barbouri and unisexual respectively, was calculated. Pairwise differences of cyt-b and NCR between the major lineages of interest were also calculated by MEGA.

Molecular dating

To infer the TMRCA for the unisexual lineage and its closest relative, Kentucky A. barbouri, we incorporated 13 previously published mtgenome sequences to allow the calibration points to be introduced to our analysis. A combined DNA dataset including two rRNAs and 13 protein-coding genes without third codon positions were used for dating analyses. Gaps in the DNA alignment were manually excluded. The ingroup root of the tree (Salamandroidea and Cryptobranchoidea) was constrained to be 151 to 170 Ma. This constraint was based on the oldest salamander fossil known to exist, the salamandroid-like Iridotriton hechti (151 Ma) [54] and a proposed maximal bound for the origin of Caudata (170 Ma) [55]. The Taricha-Notophthalmus split was constrained to be at least 23 Ma (Taricha oligocenica from the upper Oligocene) [56]. The Ambystoma-Dicamptodon split was set to at least 55.8 Ma (Dicamptodon antiquus from the late Paleocene) [57]; The Euproctus -Triturus split was set to 20-30 Ma (disjunction of the Corsica-Sardinia microplate from the Iberian Peninsula) [46]; The Cynops-Paramesotriton split was constrained to be greater than 15 Ma (Procynops miocenicus from the upper Miocene) [56]. The split of A. tigrinum- A. californiense was set to be around 5 Ma (the beginning of Sierran uplift) [24]. Because most of the fossil record we used in this study only provided a minimum age for the origin, when setting the priors we used the fossil ages as the lower bounds.

Molecular dating was conducted by a Bayesian MCMC approach in the program BEAST (version 1.5.1) [58] and MultiDivTime [59]. Using BEAST, the topology and divergence times can be estimated simultaneously from the data. BEAST input files were generated with BEAUTi (version 1.5.1). A GTR+I+G was used to describe the substitution model, a Yule process was used to describe speciation, and an uncorrelated lognormal (UCLN) model was used to describe the relaxed clock [23]. We used a lognormal distribution for fossil calibrations, and a normal distribution for the biogeographic calibrations. BEAST was run for 80,000,000 generations with samples taken every 1,000 generations. Five independent MCMC runs were conducted and the log and time tree files were combined using LogCombiner (version 1.5.1). The results were examined by Tracer (version 1.4.1) to confirm stationary distribution and adequate effective sample sizes (ESS) that had been obtained for all parameters. TreeAnnotator (version 1.5.1) was then used to summarize a best-supported tree and annotate the tree with the mean age and posterior probabilities of the nodes under investigation. FigTree (version 1.2.3) was used to display the estimated tree with node ages and the 95% confidence intervals. Programs BEAST, BEAUTi, LogCombiner, Tracer, TreeAnnotator and FigTree were downloaded from In the MultiDivTime analysis, parameters of the substitution model were first estimated by program Baseml in the PAML package (version 4.3) [60]. The output from Baseml was then used in the Multidistribute package to estimate the maximum likelihood of the branch lengths and a variance-covariance matrix, and to perform a MCMC Bayesian analysis for estimating the posterior distributions of substitution rates and divergence dates. The tree presented in Figure 3 was used as the reference topology for molecular dating analysis. The priors for the ingroup root age mean (rttm) and standard deviation (rttmsd) (Salamandroidea - Cryptobranchoidea split, 151-170 Ma) were set to 1.60 and 0.1, respectively. The mean and standard deviation of the prior distribution for the rate of molecular evolution at the ingroup root node (rtrate and rtratesd) were both set to 0.12. The prior mean and standard deviation for the Gamma distribution of the parameter controlling rate variation over time (i.e. brownmean and brownsd) were both set to 0.5. The Markov chain was run for 1,000,000 generations and sampled every 100 generations with an initial burn-in of 200,000 generations. Three independent runs were performed to ensure the convergence. The PAML package was downloaded from, and the Multidistribute package from


  1. 1.

    Maynard Smith J: The Evolution of Sex. 1978, Cambridge: Cambridge University Press

    Google Scholar 

  2. 2.

    Normark BB, Judson OP, Moran NA: Genomic signatures of ancient asexual lineages. Biol J Linn Soc. 2003, 79: 69-84. 10.1046/j.1095-8312.2003.00182.x.

    Article  Google Scholar 

  3. 3.

    Scali V, Passamonti M, Marescalchi O, Mantovani B: Linkage between sexual and asexual lineages: genome evolution in Bascillus stick insects. Biol J Linn Soc. 2003, 79: 137-150. 10.1046/j.1095-8312.2003.00172.x.

    Article  Google Scholar 

  4. 4.

    Bi K, Bogart JP: Identification of intergenomic recombinations in unisexual slamanaders of the genus Ambystoma by genomic in situ hybridization. Cytogenet Genome Res. 2006, 112: 307-312. 10.1159/000089885.

    Article  CAS  PubMed  Google Scholar 

  5. 5.

    Bogart JP, Bi K, Fu J, Noble D, Niedzwiecki J: Unisexual salamanders (genus Ambystoma) present a new reproductive mode for eukaryotes. Genome. 2007, 50: 119-136. 10.1139/G06-152.

    Article  CAS  PubMed  Google Scholar 

  6. 6.

    Mable BK: Sex in the postgenomic era. Trends Ecol Evol. 2007, 22: 559-561. 10.1016/j.tree.2007.07.006.

    Article  PubMed  Google Scholar 

  7. 7.

    Avise JC: Clonality: The Genetics, Ecology, and Evolution of Sexual Abstinence in Vertebrate Animals. 2008, Oxford: Oxford University Press

    Google Scholar 

  8. 8.

    Schartl M, Wilde B, Schlupp I, Parzefall J: Evolutionary origin of a parthenoform, the Amazon molly Poecilia formosa, on the basis of a molecular genealogy. Evolution. 1995, 49: 827-835. 10.2307/2410406.

    Article  CAS  Google Scholar 

  9. 9.

    Bogart JP: Genetics and systematics of hybrid species. Reproductive Biology and Phylogeny of Urodela. Volume 1. Edited by: Sever DM. 2003, Enfield: M/s Science, 109-134.

    Google Scholar 

  10. 10.

    Gladyshev EA, Meselson M, Arkhipova IR: Massive horizontal gene transfer in bdelloid rotifers. Science. 2008, 320: 1210-1213. 10.1126/science.1156407.

    Article  CAS  PubMed  Google Scholar 

  11. 11.

    Janko K, Bohen J, Lamatch D, Flajshans M, Epplen JT, Ráb P, Kotlík P, Šlechtová V: The gynogenetic reproduction of diploid and triploid hybrid spined loaches (Cobitis: Teleostei), and their ability to establish successful clonal lineages--on the evolution of polyploidy in asexual vertebrates. Genetica. 2007, 131: 185-194. 10.1007/s10709-006-9130-5.

    Article  PubMed  Google Scholar 

  12. 12.

    Dawley RM, Bogart JP: Evolution and ecology of unisexual vertebrates. New York State Mus Bull. 1989, 466: 1-302.

    Google Scholar 

  13. 13.

    Bogart JP, Klemens MW: Hybrids and genetic interactions of mole salamanders (Ambystoma jeffersonianum and A. laterale) (Amphibia: Caudata) in New York and New England. Am Mus Novit. 1997, 3218: 1-78.

    Google Scholar 

  14. 14.

    Bogart JP, Klemens MW: Additional distributional records of Ambystoma laterale, A. jeffersonianum (Amphibia: Caudata) and their unisexual kleptogens in eastern North America. Am Mus Novit. 2008, 3627: 1-58. 10.1206/604.1.

    Article  Google Scholar 

  15. 15.

    Bogart JP, Bartoszek J, Noble DWA, Bi K: Sex in unisexual salamanders: discovery of a new sperm donor with ancient affinities. Heredity. 2009, 103: 483-493. 10.1038/hdy.2009.83.

    Article  CAS  PubMed  Google Scholar 

  16. 16.

    Bi K, Bogart JP, Fu J: Intergenomic translocations in unisexual salamanders of the genus Ambystoma (Amphibia, Caudata). Cytogenet Genome Res. 2007, 116: 289-297. 10.1159/000100413.

    Article  CAS  PubMed  Google Scholar 

  17. 17.

    Bi K, Bogart JP, Fu J: Two rare aneutriploids in the unisexual Ambystoma (Amphibia, Caudata) identified by GISH indicating two different types of meiotic errors. Cytogenet Genome Res. 2007, 119: 127-130. 10.1159/000109628.

    Article  CAS  PubMed  Google Scholar 

  18. 18.

    Bi K, Bogart JP, Fu J: Genealogical relationships of southern Ontario polyploid unisexual salamanders (genus Ambystoma) inferred from intergenomic exchanges and major rDNA cytotypes. Chromosome Res. 2008, 16: 275-289. 10.1007/s10577-007-1192-4.

    Article  CAS  PubMed  Google Scholar 

  19. 19.

    Robertson AV, Ramsden C, Niedzwiecki J, Fu J, Bogart JP: An unexpected recent common ancestor of unisexual Ambystoma. Mol Ecol. 2006, 15: 3339-3351. 10.1111/j.1365-294X.2006.03005.x.

    Article  CAS  PubMed  Google Scholar 

  20. 20.

    Hedges SB, Bogart JP, Maxson LR: Ancestry of unisexual salamanders. Nature. 1992, 356: 708-710. 10.1038/356708a0.

    Article  CAS  PubMed  Google Scholar 

  21. 21.

    Spolsky CM, Phillips CA, Uzzell T: Antiquity of clonal salamander lineages revealed by mitochondrial DNA. Nature. 1992, 356: 706-708. 10.1038/356706a0.

    Article  CAS  PubMed  Google Scholar 

  22. 22.

    Niedzwiecki J: Evolutionary history and hybridization of two mole salamander sister species from different habitats. Ph.D. thesis. 2005, University of Kentucky, Department of Biology

    Google Scholar 

  23. 23.

    Drummond AJ, Ho SY, Phillips MJ, Rambaut A: Relaxed phylogenetics and dating with confidence. PLoS Biol. 2006, 4: e88-10.1371/journal.pbio.0040088.

    PubMed Central  Article  PubMed  Google Scholar 

  24. 24.

    Shaffer HB, McKnight ML: The polytypic species revisited: genetic differentiation and molecular phylogenetics of the tiger salamander Ambystoma tigrinum (Amphibia: Caudata) complex. Evolution. 1996, 50: 417-433. 10.2307/2410811.

    Article  CAS  Google Scholar 

  25. 25.

    Gellissen G, Bradfield JY, White BN, Wyatt GR: Mitochondrial DNA sequences in the nuclear genome of a locust. Nature. 1983, 301: 631-634. 10.1038/301631a0.

    Article  CAS  PubMed  Google Scholar 

  26. 26.

    Fukuda M, Wakasugi S, Tsuzuki T, Nomiyama H, Shimada K, Miyata T: Mitochondrial DNA-like sequences in the human nuclear genome-characterization and implications in the evolution of mitochondrial DNA. J Mol Biol. 1985, 186: 257-266. 10.1016/0022-2836(85)90102-0.

    Article  CAS  PubMed  Google Scholar 

  27. 27.

    Lopez JV, Yuhki N, Masuda R, Modi W, O'Brien SJO: Numt, a recent transfer and tandem amplification of mitochondrial DNA to the nuclear genome of the domestic cat. J Mol Evol. 1994, 39: 174-190.

    CAS  PubMed  Google Scholar 

  28. 28.

    Quinn TW: Molecular evolution of the mitochondrial genome. Avian Molecular Evolution and Systematics. Edited by: Mindell DP. 1997, Orlando: Academic Press, 3-28. full_text.

    Google Scholar 

  29. 29.

    Richly E, Leister D: NUMTs in sequenced eukaryotic genomes. Mol Biol Evol. 2004, 21: 1081-1084. 10.1093/molbev/msh110.

    Article  CAS  PubMed  Google Scholar 

  30. 30.

    Spinks PQ, Shaffer HB: Conservation phylogenetics of the Asian box turtles (Geoemydidae, Cuora): mitochondrial introgression, numts, and inferences from multiple nuclear loci. Conserv Genet. 2007, 8: 641-657. 10.1007/s10592-006-9210-1.

    Article  Google Scholar 

  31. 31.

    Bensasson D, Zhang DX, Hartl D, Hewitt GM: Mitochondrial pseudogenes: evolution's misplaced witnesses. Trends Ecol Evol. 2001, 16: 314-321. 10.1016/S0169-5347(01)02151-6.

    Article  PubMed  Google Scholar 

  32. 32.

    Ibarguchi G, Friesen VL, Lougheed SC: Defeating numts: Semi-pure mitochondrial DNA from eggs and simple purification methods for field-collected wildlife tissues. Genome. 2006, 49: 1438-1450. 10.1139/G06-107.

    Article  CAS  PubMed  Google Scholar 

  33. 33.

    Arctander P: Comparison of a mitochondrial gene and a corresponding nuclear pseudogene. Proc R Soc B-Biol Sci. 1995, 262: 13-19. 10.1098/rspb.1995.0170.

    Article  CAS  Google Scholar 

  34. 34.

    Woischnik M, Moraes CT: Pattern of organization of human mitochondrial pseudogenes in the nuclear genome. Genome Res. 2002, 12: 885-893.

    PubMed Central  Article  CAS  PubMed  Google Scholar 

  35. 35.

    Song H, Buhay JE, Whiting MF, Crandall KA: Many species in one: DNA barcoding overestimates the number of species when nuclear mitochondrial pseudogenes are coamplified. Proc Natl Acad Sci USA. 2008, 105: 13486-13491. 10.1073/pnas.0803076105.

    PubMed Central  Article  CAS  PubMed  Google Scholar 

  36. 36.

    Brower AVZ: Problems with DNA barcodes for species delimitation: 'Ten species' of Astraptes fulgerator reassessed (Lepidoptera: Hesperiidae). Syst Biodivers. 2006, 4: 127-132. 10.1017/S147720000500191X.

    Article  Google Scholar 

  37. 37.

    Maté ML, Di Rocco F, Zambelli A, Vidal-Rioja L: Mitochondrial heteroplasmy in control region DNA of South American camelids. Small Rumin Res. 2007, 71: 123-129. 10.1016/j.smallrumres.2006.04.016.

    Article  Google Scholar 

  38. 38.

    Abbott CL, Double MC, Trueman JWH, Robinson A, Cockburn A: An unusual source of apparent mitochondrial heteroplasmy: duplicate mitochondrial control regions in Thalassarche albatrosses. Mol Ecol. 2005, 14: 3605-3613. 10.1111/j.1365-294X.2005.02672.x.

    Article  CAS  PubMed  Google Scholar 

  39. 39.

    Judson OP, Normark BB: Ancient asexual scandals. Trends Ecol Evol. 1996, 11: 41-46. 10.1016/0169-5347(96)81040-8.

    Article  CAS  PubMed  Google Scholar 

  40. 40.

    Bi K, Bogart JP, Fu J: The prevalence of genome replacement in unisexual salamanders of the genus Ambystoma (Amphibia, Caudata) revealed by nuclear gene genealogy. BMC Evol Biol. 2008, 8: e158-10.1186/1471-2148-8-158.

    Article  Google Scholar 

  41. 41.

    Little TJ, Hebert PDN: Ancient asexuals: scandal or artifact?. Trends Ecol Evol. 1996, 11: 296-10.1016/S0169-5347(96)91640-7.

    Article  CAS  PubMed  Google Scholar 

  42. 42.

    Neiman N, Meirmans S, Meirmans PG: What can asexual lineage age tell us about the maintenance of sex?. Ann NY Acad Sci. 2009, 1168: 185-200. 10.1111/j.1749-6632.2009.04572.x.

    Article  PubMed  Google Scholar 

  43. 43.

    Samuels AK, Weisrock DW, Smith JJ, France KJ, Walker JA, Putta S, Voss SR: Transcriptional and phylogenetic analysis of five complete ambystomatid salamander mitochondrial genomes. Gene. 2005, 349: 43-53. 10.1016/j.gene.2004.12.037.

    Article  CAS  PubMed  Google Scholar 

  44. 44.

    Mueller RL, Macey JR, Jaekel M, Wake DB, Boore JL: Morphological homoplasy, life history evolution, and historical biogeography of plethodontid salamanders inferred from complete mitochondrial genomes. Proc Natl Acad Sci USA. 2004, 101: 13820-13825. 10.1073/pnas.0405785101.

    PubMed Central  Article  CAS  PubMed  Google Scholar 

  45. 45.

    Zhang P, Wake DB: Higher-level salamander relationships and divergence dates inferred from complete mitochondrial genomes. Mol Phylogenet Evol. 2009, 53: 492-508. 10.1016/j.ympev.2009.07.010.

    Article  CAS  PubMed  Google Scholar 

  46. 46.

    Zhang P, Papenfuss TJ, Wake MH, Qu L, Wake DB: Phylogeny and biogeography of the family Salamandridae (Amphibia: Caudata) inferred from complete mitochondrial genomes. Mol Phylogenet Evol. 2008, 49: 586-597. 10.1016/j.ympev.2008.08.020.

    Article  CAS  PubMed  Google Scholar 

  47. 47.

    Zhang P, Chen YQ, Zhou H, Liu YF, Wang XL, Papenfuss TJ, Wake DB, Qu LH: Phylogeny, evolution, and biogeography of Asiatic Salamanders (Hynobiidae). Proc Natl Acad Sci USA. 2006, 103: 7360-7365. 10.1073/pnas.0602325103.

    PubMed Central  Article  CAS  PubMed  Google Scholar 

  48. 48.

    Moritz C, Schneider CJ, Wake DB: Evolutionary relationships within the Ensatina eschscholtzii complex confirm the ring species interpretation. Syst Biol. 1992, 41: 273-291.

    Article  Google Scholar 

  49. 49.

    Swofford DL: PAUP*: Phylogenetic Analysis Using Parsimony (and Other Methods). 2002, Sunderland: Sinauer Associates

    Google Scholar 

  50. 50.

    Felsenstein J: Confidence limits on phylogenies: an approach using the bootstrap. Evolution. 1985, 39: 783-791. 10.2307/2408678.

    Article  Google Scholar 

  51. 51.

    Huelsenbeck JP, Ronquist F: MrBayes: Bayesian inference of phylogeny. Bioinformatics. 2001, 17: 754-755. 10.1093/bioinformatics/17.8.754.

    Article  CAS  PubMed  Google Scholar 

  52. 52.

    Nylander JAA: MrModeltest (Version 2.2). 2004, Uppsala: Uppsala University, Computer program distributed by the author

    Google Scholar 

  53. 53.

    Tamura K, Dudley J, Nei M, Kumar S: MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol. 2007, 24: 1596-1599. 10.1093/molbev/msm092.

    Article  CAS  PubMed  Google Scholar 

  54. 54.

    Evans SE, Lally C, Chure DC, Elder A, Maisano JA: A Late Jurassic salamander (Amphibia: Caudata) from the Morrison Formation of North America. Zool J Linn Soc. 2005, 143: 599-616. 10.1111/j.1096-3642.2005.00159.x.

    Article  Google Scholar 

  55. 55.

    Marjanović D, Laurin M: Fossils, molecules, divergence times, and the origin of lissamphibians. Syst Biol. 2007, 56: 369-388. 10.1080/10635150701397635.

    Article  PubMed  Google Scholar 

  56. 56.

    Estes R: Gymnophiona, Caudata. Handbuch der Paläoherpetologie. 1981, 2: 1-115.

    Google Scholar 

  57. 57.

    Naylor BG, Fox RC: A new ambystomatid salamander, Dicamptodon antiquus n. sp. from the Paleocene of Alberta. Can J Earth Sci. 1993, 30: 814-818. 10.1139/e93-067.

    Article  Google Scholar 

  58. 58.

    Drummond AJ, Rambaut A: BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol. 2007, 7: e214-10.1186/1471-2148-7-214.

    Article  Google Scholar 

  59. 59.

    Thorne JL, Kishino H: Divergence time and evolutionary rate estimation with multilocus data. Syst Biol. 2002, 51: 689-702. 10.1080/10635150290102456.

    Article  PubMed  Google Scholar 

  60. 60.

    Yang Z: PAML: a program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci. 1997, 13: 555-556.

    CAS  PubMed  Google Scholar 

Download references


We thank J Bartoszek, J Ferner, MW Klemens, J Niedzwiecki, and J Davis for providing samples, SB Hedges for pointing out the possibility of a numt, A Hollis and J Gross for managing the sequencing facility at Guelph, J Fu and D Noble for helpful discussions. This work was supported by an NSERC (Canada) grant to J Bogart.

Author information



Corresponding author

Correspondence to James P Bogart.

Additional information

Authors' contributions

KB detailed the experimental design and performed most of the lab work, data analyses and manuscript preparation. JPB conceived and directed the study. Both authors contributed equally to this work in discussing research strategy and development. Both authors read and approved the final manuscript.

Authors’ original submitted files for images

Rights and permissions

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and Permissions

About this article

Cite this article

Bi, K., Bogart, J.P. Time and time again: unisexual salamanders (genus Ambystoma) are the oldest unisexual vertebrates. BMC Evol Biol 10, 238 (2010).

Download citation


  • Mitochondrial Genome
  • Consistency Index
  • Asexual Lineage
  • Mitochondrial Heteroplasmy
  • Unisexual Population