The evolution of sexuality versus asexuality has been the focus of considerable attention for decades. Although sexuality is ubiquitous in animals, unisexual reproduction has independently evolved in various unrelated lineages [1–3]. The existence of ancient unisexuals, regardless of their rarity, poses a dilemma for evolutionary biologists because a lack of genetic recombination should, theoretically, cause unisexuals to be short-lived [4–10]. From a single hybrid origin, unisexual salamanders in the genus Ambystoma have survived for 2.4–3.9 million years , representing the most ancient known unisexual vertebrate lineage [12, 13]. Unisexual Ambystoma have a unique, while extremely flexible, reproductive system which is described as kleptogenesis . Female kleptogens usually produce unreduced eggs through premeiotic endomitosis , and sperm from sympatric sexual males is required to activate the eggs without fertilization (termed gynogenesis). But, a sperm nucleus can be incorporated to elevate the ploidy (termed ploidy elevation) or to replace one of the nuclear genomes in a developing offspring (termed genome replacement) . Unisexual individuals interact with and "steal" chromosomes from two to four sympatric sexual species (A. laterale (LL), A. jeffersonianum (JJ), A. texanum (TT), and A. tigrinum (TiTi)), and more than 20 distinct diploid, triploid, tetraploid and pentaploid genome compositions are known to exist . Among these reproductive mechanisms, genome replacement may be responsible for the apparent evolutionary longevity of unisexual Ambystoma [12, 16, 17].
Although the evolutionary significance of genome replacement in unisexual Ambystoma is acknowledged, empirical evidence for its prevalence in natural populations is absent. An ancient genome replacement event was proposed using a phylogenetic approach . All the unisexual Ambystoma, irrespective of their genome compositions, share an almost identical mitochondrial genome that is distinctly different from any of the four "parental" species, excluding all of them as candidates for a maternal ancestor. The sequence of unisexual mt DNA is most closely related to that in specimens of A. barbouri ("BB") from Kentucky, suggesting that A. barbouri and all unisexual salamanders share a close common maternal ancestor. However, the "B" nuclear genome was completely replaced, but left its mitochondrial genome behind . Nevertheless, the existence of genome replacement was questioned by Spolsky et al. [13, 16]. They found no evidence supporting genome replacement in a few A. texanum -dependent LJJ populations in central Indiana and northeastern Illinois so they believed that genome replacement in unisexual Ambystoma is very uncommon in nature. But, they acknowledged the potential importance of genome replacement for long-term survival of unisexual Ambystoma. Previous studies using isozyme electrophoresis, DNA microsatellite and cytogenetic analyses had been carried out to try to document the actual occurrence of genome replacement in a few unisexual populations [11, 18–24]. All these studies present discrete evidence that demonstrate the possible rare occurrence of genome replacement but the methods used were insufficient to document genome replacement across a large number of unisexual populations. As a consequence, none of those studies provide conclusive empirical evidence about the frequency or commonality of genome replacement in unisexual Ambystoma. To better document to what extent genome replacement takes place in this old lineage, nuclear gene genealogy between sexual species and their corresponding genomic contributions in various unisexual populations across the range could be the key to understand this process.
Despite the presence of genome replacement, all known unisexual Ambystoma, irrespective of their genome compositions, contain at least one "L" genome, even in those populations where no A. laterale have ever been discovered . The conservation of an "L" genome in every unisexual is a mystery. It was hypothesized that hybridization between a female A. barbouri and a male A. laterale gave rise to an ancestral genome composition "BL" that initiated a unisexual linage. Subsequently, early in the history of unisexual linage, the "B" genome was replaced by genomes from other sperm donors, but the ancient paternal "L" genome may have been maintained . Recent evidence contradicted this hypothesis . DNA microsatellite loci demonstrated that, in one population from southern Ontario, no common genome is transmitted in a clonal or hemiclonal fashion in unisexuals, and that there is not a single "L" genome that could be considered ancestral or frozen in all unisexuals . Nevertheless, additional empirical evidence is required for many other unisexual populations with respect to whether there exists a particular "L" genome that is always favored and trapped in unisexual individuals, and is somehow protected from genome replacement.
Directly examining the genealogical relationships of the "L" genomes within the unisexual and sexual lineages provides a rapid way to answer the following two fundamentally important questions. First, is genome replacement a common or sporadic event during the evolution of unisexual Ambystoma lineages? If it is rare, we would expect that most "L" genomes in unisexual lineages would be more closely related to each other than to respective sympatric A. laterale. Second, is there a particular " L" genome that is frozen in unisexual populations across their distributional range? If there is, we would expect that one "L" genome found in individual unisexuals from distant populations would group together and have a more basal position on the "L" genome tree. In this study, we chose a highly variable DNA fragment from a nuclear gene intron as a marker, designed specific primers that only amplified the fragment in the "L" genome in various unisexual genome compositions, and constructed a genealogy for "L" genomes in unisexuals and A. laterale.