Several species of mole salamanders in the genus Ambystoma are targeted by various state, provincial, and federal agencies for conservation. These salamanders have specific wetland and forested upland habitat requirements that render them vulnerable to environmental alteration. The blue-spotted salamander, Ambystoma laterale (LL) and the Jefferson salamander, A. jeffersonianum (JJ) have both been listed for protection in various parts of their ranges, but the identification of these salamanders is confusing because they often coexist with unisexual individuals that are mostly polyploid and use the sexual species as sperm donors. We used isozyme electrophoresis, blood erythrocytes, and chromosome counts in a continued effort to identify sexual and unisexual individuals in eastern North America. We examined 1377 salamanders from 118 sites in Connecticut, Massachusetts, New Jersey, New York, Pennsylvania, and Virginia. Most Pennsylvania salamanders were A. jeffersonianum (JJ) but A. laterale (LL), previously unknown from Pennsylvania, were found in that state. The two sexual species were never found together. We found diploid (LJ), triploid (LLJ; LJJ), and tetraploid (LLLJ; LJJJ; LLJJ) unisexuals. At most collecting sites, unisexuals were more numerous than sexual individuals. The association of sexual and unisexual individuals support a kleptogenic reproductive system in which the unisexuals steal genomes from their sympatric sexual sperm donors.
Introduction
There are about 30 living species of mole salamanders in the North American family Ambystomatidae (Petranka, 1998). Most species occur in the continental United States, but there has been a Mexican invasion of the Ambystoma tigrinum complex (Shaffer and McKnight, 1996) and some species have northern ranges that include parts of Canada (Petranka, 1998). Mole salamanders often have specific habitat requirements that make them vulnerable to anthropogenic environmental alterations and most species are listed as candidates for conservation in some or all parts of their range. The blue-spotted salamander, A. laterale and the Jefferson salamander, A. jeffersonianum present problems for conservationists because they both coexist with unisexual individuals that normally do not have a conservation status (Kraus, 1995). Connecticut lists Ambystoma jeffersonianum complex and A. laterale complex as state species of special concern because it is difficult to distinguish the unisexuals from the sexual species (Klemens, 2000). Connecticut also lists the pure diploid populations of A. laterale in the eastern portion of the state as a threatened species (Klemens, 2000). Connecticut is the only state in the northeastern United States that has differentially protected the pure diploid A. laterale populations that also occur in southeastern Massachusetts and on the eastern tip of Long Island, New York. The conservation status of these salamanders does vary over their ranges (table 1).
TABLE 1
Conservation status of Ambystoma laterale and A. jeffersonianum populations in Connecticut, Massachusetts, New Jersey, New York, Pennsylvania, and Rhode Island E = endangered, T = threatened, SC = special concern.
We (Bogart and Klemens, 1997) provided a detailed historical perspective on the unisexual salamanders and their associated sperm donor species. We also identified and characterized populations of A. laterale, A. jeffersonianum, and their unisexual associates in the New England states and parts of New York. Based on 1002 individuals from 106 sites, we found the unisexuals to be common and widespread. The majority of the populations that contained either A. laterale or A. jeffersonianum also contained unisexual individuals. Seventy percent of the salamanders in that study were found to be unisexuals. Ten of the 106 sites were found to contain only A. laterale, but six of these sites were represented by one or two individuals. Samples obtained from two of the three sites where only A. jeffersonianum were found consisted of single specimens and only three individuals were sampled from the other A. jeffersonianum site. Diploid, triploid, and tetraploid unisexuals were found throughout New England. Most unisexual salamanders were polyploid, but, because diploid unisexuals were found at 21 scattered sites, ploidy is not a dependable distinguishing character to separate unisexual and sexual individuals without corroborative data. Unisexual salamanders are expected to be female, but rare “unisexual” males were also found; thus, knowing that an individual is a male does not always confirm a sexual species identification.
Clarification of the evolution of unisexual salamanders and the reproductive mode of these females has been further elucidated using gene sequences and microsatellite DNA loci (Bogart, 2003; Bogart et al., 2007). Mitochondrial DNA analysis show that the unisexuals arose about three to four million years ago from a hybridization event involving a female that shared its most recent common ancestor to individuals of A. barbouri from Kentucky. Ambystoma barbouri has a current, restricted range in Ohio, Kentucky, and Tennessee and is a recently uncovered (Kraus and Petranka, 1989) sister species (Petranka, 1998; Niedzwiecki, 2005) to A. texanum. Subsequently, unisexuals rapidly dispersed in eastern North America using a reproductive mode that appears to be unique (Bogart et al., 2007). All unisexuals maintain a similar A. barbouri–like mitochondrial DNA but can incorporate and exchange nuclear genomes with sympatric males. Because unisexual females can steal genomes from a variety of sympatric males, the suggested reproductive mode is kleptogenesis (Bogart et al., 2007) and has given rise to at least 20 nuclear genomic combinations, or kleptogens, that are diploid, triploid, tetraploid, and even pentaploid (Bogart, 2003).
To expand the coverage of populations and to include more of the range of these salamanders in New Jersey, New York, and Pennsylvania the present study used isozyme electrophoresis to identify additional individuals of sexual A. laterale, A. jeffersonianum, and unisexual kleptogens that use these species as sperm donors. We also sampled new populations in New England and included a few samples from populations that were previously sampled by Bogart and Klemens (1997) to increase the sample size and to confirm our findings in previously sampled populations. Accurate data on the distribution of this salamander complex should assist conservation efforts to protect these salamanders. Such data are also necessary to explore the evolutionary significance of kleptogenesis.
Materials and Methods
Methods were mostly identical to those used by Bogart and Klemens (1997). Salamanders were collected with dip nets and minnow traps from breeding ponds early in the spring, while crossing roads at night during spring rains, or from under rocks and logs during the day from March through October. Egg mass surveys (Bogart, 1982) were not done because data obtained from such surveys may not provide a random sample of the population as most eggs laid by unisexual females do not result in hatched larvae and many hatched larvae do not survive to transformation. In some cases, when adults were not collected from a breeding pond, advanced larvae were randomly collected using a dip net. They were maintained through transformation in the laboratory and the juveniles were considered to be representative of the population or site. Our present study is a continuation of the collections done for our previous study (Bogart and Klemens, 1997). Collections for the present study were conducted over a nine-year period, from 1996 to 2004. Collections were concentrated in New York, New Jersey, and Pennsylvania, but collections also included sites in Connecticut, Massachusetts, and Virginia. Salamanders (1377 individuals) from 118 sites (appendix 1) were collected and shipped to Guelph for identification. Salamanders were injected with colchicine (0.25 to 0.75 ml of a 0.1 mg/ml solution) two days before they were killed by prolonged anesthesia in a 7% solution of buffered (pH 7.0) tricaine methane sulfonate (MS222). The heart was exposed and blood was collected from the conus arteriosis in heparinized microhematocrit tubes for ploidy determination. The intestine, including the cloaca, was dissected from each individual, immersed in de-ionized water for 15 min, fixed in 3:1 ethanol:acetic acid in 1.5 ml Eppendorf microtubes, and stored at −20°C for chromosome analyses. Tissues required for isozyme electrophoresis were liver and a combination of the heart, skeletal muscle, and spleen. The tissues were removed from freshly killed salamanders and stored with an equal volume of de-ionized water in 1.5 ml Eppendorf microtubes in an ultracold freezer (−80°C). After the tissues were removed, the specimens were preserved and deposited at the American Museum of Natural History (AMNH) (appendix 1).
Isozyme Electrophoresis
Just prior to electrophoresis, the frozen tissues were ground using a sharp glass rod and spun for two minutes in a microcentrifuge. Filter-paper wicks were used to soak up a portion of the supernatant. The wicks were air dried on filter paper and inserted in starch gels. Horizontal starch-gel electrophoresis followed the procedures outlined by Selander et al. (1971), Bogart (1982), and Bogart and Klemens (1997). The buffer systems used were described by Selander et al. (1971) and Clayton and Tretiak (1972). Electrophoretic loci for each enzyme system were numbered on the gel from the most anodally migrating locus. As in previous studies (Bogart et al., 1987; Bogart, 1989; Bogart and Klemens, 1997), alleles, or allozymes, were designated by their relative mobilities compared with the mobility of the most common allele in A. laterale, which was assigned a mobility of 100. We chose 11 of the 21 isozyme loci used by Bogart and Klemens (1997) in our present study. These loci proved to be most useful in identifying the sexual and unisexual individuals and included loci that demonstrated some homozygosity and reversals in our previous study. The loci that were assayed, buffer systems used, and the tissues examined are provided in table 2. Nine loci were dimeric enzymes that have been the most useful in visualizing staining intensities of the bands (dosage) for assessing genome composition in polyploid amphibians (Danzmann and Bogart, 1982). Lactate dehydrogenase is a tetrameric enzyme that also expresses differential staining of the heterotetrameric bands. The only monomeric enzyme, phosphoglucose mutase, often demonstrated different staining intensities for two loci (pgm-1; pgm-2) and was previously found to posses a few rare heterozygotes at pgm-2 in both sexual and unisexual salamanders (Lowcock and Bogart, 1989; Bogart and Klemens, 1997).
TABLE 2
Presumptive structural gene loci examined in Ambystoma
Ploidy Determination
A small drop of blood from the hematocrit tube, taken from every individual, was mixed with an approximately equal amount of diluted (25 ml H2O added to 100 ml) Hanks Balanced Salt Solution (Sigma) and photographed under phase-contrast and bright-field optics. The area of the blood cells was determined using a sonic digitizer on a rear-projection image of the negatives (Austin and Bogart, 1982). Six blood cells from each individual were measured to obtain an average erythrocyte area. During our investigation, a flow cytometric method (FCM) was found to be a more accurate method to determine ploidy and that method was introduced into our study in 2003 following the methods in Ramsden et al. (2006). Blood cell data provided the first clear indication of diploid and tetraploid unisexuals that may not be easily distinguished from triploids using only the electrophoretic patterns. Chromosome numbers were the absolute proof of ploidy in the salamanders and those data were required if a discrepancy existed between the electrophoretic genotype, blood cell area, or flow cytometric analyses. When necessary, chromosomes were obtained from the gut epithelial tissue using procedures outlined by Kezer and Sessions (1979) and Sessions (1982). Unstained, or conventionally stained, chromosomes of A. laterale and A. jeffersonianum are virtually indistinguishable (Taylor and Bogart, 1990), so the chromosomes obtained for most individuals and those observed in our previous study (Bogart and Klemens, 1997) could not be used to assign genomic constitution in the unisexuals. New chromosome methods that use fluorescent genomic probes and in situ hybridization clearly differentiate genomes of these two species in the unisexuals (Bi and Bogart, 2006). Microsatellite DNA loci that are amplified using primers developed for A. jeffersonianum (Julian et al., 2003) also proved to be an accurate method to determine ploidy and genomic constituents in unisexual individuals (Ramsden et al., 2006; Bogart et al., 2007). Although we employed all of these methods to confirm ploidy in individual salamanders, the analyses of chromosome and microsatellite data from tissue collected from individuals in the present study will be the subject of future investigations that use these new techniques.
Site Designation
Because we were interested in knowing the sexual and unisexual associations in breeding populations, we tried to analyze the individuals that were expected to breed in the same population. In most cases, salamanders could be assigned to distinct breeding ponds because they were collected as adults or larvae in a pond or as newly transformed juveniles at the edge of a pond. Sometimes, however, individuals were collected on roads or in upland wooded areas and could not be assigned to any particular breeding pond. Additionally, salamanders were found at different locations entering or leaving extensive swamps that may be partitioned into subpopulations. These factors required the use of site designations based on the assumption that individuals from the same site shared a potentially common breeding area. A site may or may not be equated with a breeding population. In some instances, we combined individuals into a single site even though they were collected from extensive, yet ecologically similar, and contiguous habitat. Some other collections that were close geographically were treated as separate sites if they were distinctly separated by habitat and, in some instances, elevation. Klemens (1993) found that topography plays an important role in determining wetland type, which in turn influences the distribution of the sexual species. This was most apparent in areas of close contact where breeding ponds of the two sexual species were separated by as little as 100 meters. We grouped our sites into drainage basins because Klemens (1978, 1993) demonstrated that there were significant differences in the herpetofauna of New England's drainage basins, as a result of post-Pleistocene dispersion of amphibians into the interior of New England.
We numbered the sites starting with 107 as a continuation of the 106 sites sampled in our earlier study (Bogart and Klemens, 1997). Eight sites that were re-sampled from our earlier study were not given new site numbers in the present study.
Nomenclature
Identifying and naming unisexual individuals is a formidable challenge for taxonomists and conservationists. The unisexuals do have a hybrid nuclear genomic constitution, but they are not hybrids that have resulted from the crossing of A. laterale and A. jeffersonianum or any combination of the four species whose genomes might be found in a unisexual. Genomes are gained and lost by kleptogenesis, which is driven entirely by male sperm donors. In this flexible genetic system, a single unisexual female can produce offspring that have differing genotypes and genomes (Bogart et al., 1987, 2007) and, because of intergenomic interaction (Bi and Bogart, 2006; Bi et al., 2007), genomes found in offspring from a particular unisexual might evolve within that unisexual independently of male genomic contribution. Lowcock et al. (1987) suggested an informal descriptive system for unisexual Ambystoma that was used by Schultz (1969) to describe the genetic composition of hybridogenetic unisexual fish of the genus Poeciliopsis. Letter designations for species that contributed genomes to unisexual Ambystoma have previously been used for convenience in a number of previous studies (Uzzell, 1964; Bogart et al., 1985, 1987; Bogart and Klemens, 1997): J for A. jeffersonianum, L for A. laterale, T for A. texanum, and Ti for A. tigrinum. The proposed name, for example, for a triploid unisexual that has a nuclear genome consisting of one A. laterale genome and two A. jeffersonianum genomes would be Ambystoma laterale – (2) jeffersonianum, or LJJ.
Genotype and Genomotype Analysis
The unisexuals are mostly fixed heterozygotes for a number of isozyme alleles and their mode of reproduction is kleptogenesis. Therefore, population genetic models that are based on randomly interbreeding individuals do not apply to the unisexual kleptogens. When unisexuals occur in a population, they usually outnumber individuals of the sexual species (Bogart and Klemens, 1997). If the sample of individuals obtained from a site contained unisexuals, but neither A. laterale nor A. jeffersonianum individuals were found, we suspected that the sample was too small to have encountered these species. In order to obtain some measure of the influence, or the presence of the sexual species, we calculated the genomic percentage of A. laterale in each site to obtain values ranging from 0% (all A. jeffersonianum) to 100% (all A. laterale). Under this scheme, a triploid LLJ would be 66.7% (A. laterale) and triploid LJJ would be 33.3% (A. laterale). If the average percentage of all the individuals from a site was below 50 then A. jeffersonianum is assumed to be the sperm donor in that population even though the sample of salamanders may not have included A. jeffersonianum. We use the term genomotype (Lowcock, 1994) to describe the template genomic contributions of A. laterale (LL) and A. jeffersonianum (JJ) in diploid and polyploid unisexual individuals while recognizing that the combination of genomes in unisexuals may be slightly restructured by intergenomic recombinations and translocations (Bi and Bogart, 2006; Bi et al., 2007).
Results
Electrophoresis
As in our previous study (Bogart and Klemens, 1997), A. laterale could be easily distinguished from A. jeffersonianum individuals based on the presence of alternate electrophoretic alleles that were mostly homozygous with a gene frequency (p) > 0.90 in both species. The genotypes for A. laterale (appendix 2-1) and A. jeffersonianum (appendix 2-2) individuals and their allele frequencies for each locus (table 3) demonstrate considerable intraspecific homozygosity for most of the 11 loci examined, but Sod-1 was the only locus that was alternately fixed (p = 1.00) for Sod-1100 in A. laterale and Sod-137 in A. jeffersonianum. Gene frequencies at each of the 11 loci were calculated for the sexual genotypes and unisexual genomotypes from all sites (table 3) based on the data for all individuals (appendix 2). The common allele (p) > 0.95 for A. jeffersonianum was not found in any A. laterale specimen for Aat-179, Aat-2−180, Idh-1142, Mdh-1176, and Mpi120. Based on these diagnostic isozyme alleles and ploidy determination (below), the unisexual specimens were identified as A. laterale – jeffersonianum (LJ) (appendix 2-3); A. (2) laterale – jeffersonianum (LLJ) (appendix 2-4); A. laterale – (2) jeffersonianum (LJJ) (appendix 2-5); A. (3) laterale – jeffersonianum (LLLJ) (appendix 2-6); A. laterale – (3) jeffersonianum (LJJJ) (appendix 2-7); and A. (2) laterale – (2) jeffersonianum (LLJJ) (appendix 2-8). The sites (appendix 1) where the sexual and unisexual individuals were found are included for each individual in appendix 2. A summation of the sexual and unisexual genomotypes at each site is provided in table 4. We found more unisexual individuals (44%) than either of the sexual species (23% LL and 34% JJ) and most unisexuals were triploid. Tetraploids (3%) and unisexual males (1%) were rare and were found with unisexual diploid and/or triploid females in scattered populations.
TABLE 3
Allele frequencies of diploid and polyploid salamanders for each of the genomotypes in appendix 2
TABLE 4
Ambystoma laterale (LL), A. jeffersonianum (JJ), and nuclear hybrid genomes found at each site
New and Rare Alleles
Sod-1 was fixed for one allele in A. laterale and one in A. jeffersonianum but we found additional alleles at the other loci in a few individuals of the two species and the unisexuals. Some rare alleles (p) < 0.05 were also found in our earlier study in individual A. laterale (Aat-1110, Aat-2−50, Ldh-188, Ldh-255, Mdh-1135, Pgi325, Pgi115, Pgm-197, Pgm-2120, Pgm-275) and A. jeffersonianum (Aat-1100, Aat-2−100, Aat-2−50, Idh-1100, Ldh-1115, Mdh-1100, Mpi140, Mpi100, Pgi100, Pgm-182). In the present study we found most of these same rare alleles and we also found nine new alleles that were not encountered by Bogart and Klemens (1997). A new Ldh-178 allele (D) with a frequency (p) of 0.109 in A. laterale (table 3) is not a rare allele. This allele was found in both A. laterale and LLJ unisexual specimens in northern Pennsylvania where it was found in a homozygous condition in A. laterale and unisexual LLJ individuals (site 189). Unisexual LLJ from site 148 in western New York were also homozygous for Ldh-178. Ambystoma laterale was not collected at that site but is presumed to be the sperm donor (67.8% L, table 4). The other eight new alleles that were found were rare (p) < 0.05. Aat-1 86 was found only as a heterozygous allele in one LLJ (site 145) in western New York. Two new Idh-1 alleles were found (Idh-1160 and Idh-150). The Idh-1160 allele migrated more anodally on the gel than the Idh-1142 (A) allele and is designated Q in appendix 2 and table 3. That allele was found in a heterozygote condition in A. laterale (sites 113 and 189), A. jeffersonianum (sites 118 and 204), and in unisexual LJ and LJJ individuals (site 130). The Idh-150 allele was found in only one A. laterale QC heterozygous individual (site 113). A new Ldh-2160, also a “Q” allele was found in A. jeffersonianum heterozygotes (sites 2 and 120) in western New York. This allele was not found in the two individuals from site 2 that were sampled by Bogart and Klemens (1997). Both site 2 and site 120 are in Tompkins County. One LLJ heterozygote from site 208 in eastern New York also possessed the Ldh-2160 Q allele. A new Mdh-1200 allele was found in A. jeffersonianum and an LJJ unisexual from site 158 in Morris County, New Jersey. A new Mpi80 allele was found in A. laterale from distant localities (sites 108 and 150), A. jeffersonianum (site 138), LJ (site 112), and LJJ (site 126). A new Pgi380 Q allele was found in scattered populations that included A. jeffersonianum (sites 118, 170, 185) and one LLJ unisexual (site 155). A new Pgm-1115 allele was found in A. laterale from western (site 127) and northern New York (sites 150 and 151) as well as in A. jeffersonianum from southeastern New York (site 121) and in one unisexual LLJ individual from site 148 in western New York.
Homozygosity and Reversed Genotypes in Unisexuals
The majority of the unisexuals were heterozygous and were identified by the observed diagnostic alleles at several loci (above). There were, however, a few unisexual individuals that were homozygous for A. laterale or A. jeffersonianum alleles at some loci. In addition, a few polyploid unisexuals had a genotype that, based on observed electrophoretic dosage, included some reversed patterns. In our earlier investigation (Bogart and Klemens, 1997), we also found homozygous alleles and reversed dosage patterns for some individuals at some loci. Of the six diagnostic loci where homozygous unisexuals were observed, 6.93% of the unisexuals were homozygous for Idh-1 alleles. There were fewer homozygous unisexual individuals at the other loci. The number of unisexual individuals that were homozygous for alleles at one or more of the six diagnostic loci is included in table 5. Again, more unisexual polyploid individuals had a reversed electrophoretic pattern for Idh-1 diagnostic alleles (4.8%), but the frequency of reversals found for any diagnostic allele at the seven loci involved very few individuals (table 6).
TABLE 5
Unisexual individuals that demonstrated an unexpected homozygous condition for alleles at loci diagnostic for A. laterale and A. jeffersonianum Sample sizes in parentheses (n) are the number of individuals of each genomotype that were examined and scored for each locus (from appendix 2) (see footnotes in table 3).
TABLE 6
Unisexual individuals that demonstrated a reversed condition for diagnostic alleles For example, the expected genotype for an LLJ individual for Aat-1 would be Aat-1100/100/79 or BBD and a reversed genotype at that locus would be Aat-1100/79/79 or BDD which would be the expected genotype for LJJ. A reversal in a tetraploid LLLJ could be Aat-1100/100/79/79 or Aat-1100/79/79/79. Sample sizes in parentheses (n) are the number of individuals of each genomotype that were examined and scored for each locus (from appendix 2) (see footnotes in table 3).
Ploidy
Data for blood cell measurements for individuals are included in appendix 2 and are summarized in table 7 and fig. 1. Although the ranges for the erythrocyte area measurements were large in all of the genomic classes, a T-test revealed a significant difference (p < 0.01) between the ploidy classes when they were grouped as diploid, triploid, or tetraploid. We found a significant difference between measurements of A. laterale and A. jeffersonianum but no significant difference between the measurements of LJJ triploids and LLLJ tetraploids. These data show that A. jeffersonianum has larger blood cells than do A. laterale and raise the possibility that measurements of some triploid or tetraploid individuals might not accurately distinguish these ploidy classes. We have more confidence in the ploidy determinations based on flow cytometric data, but those data were not quantified as absolute picograms of DNA. Blood cells from individuals were compared with a diploid A. jeffersonianum standard diploid peak (Ramsden et al., 2006) to determine ploidy. Ploidy determinations for such individuals are signified by FCM in appendix 2.
TABLE 7
Summation statistics for blood cell measurements for A. laterale (LL), A. jeffersonianum (JJ), and unisexual individuals prior to the use of flow cytometric methods (FCM) for ploidy determination Average erythrocyte values for the individuals are in appendix 2.
Fig. 1
Graph of mean (± SD) erythrocyte area for diploid Ambystoma jeffersonianum (JJ), diploid A. laterale (LL), diploid unisexuals (LJ), triploid unisexuals (LJJ, LLJ), and tetraploid unisexuals (LLLJ, LJJJ) from the data in table 7.
Connecticut and Massachusetts
Unisexuals were found in 14 of the 19 sites sampled in Connecticut and Massachusetts in the present study. We found one site (196) in Hartford County, Connecticut, that only had A. jeffersonianum (n = 3), but the sample size is too low to confirm a pure A. jeffersonianum population. We found only one A. laterale at another Hartford County site (199). Windham County sites (60: n = 2; 204: n = 20) in Connecticut as well as a Bristol and Plymouth County site (66: n = 2) in Massachusetts are probably pure A. laterale sites. Only A. laterale was found in our earlier study (Bogart and Klemens, 1997) at site 60 (n = 17) and site 66 (n = 4). In our present study, only A. laterale and LLJ individuals were collected in Massachusetts. Many more sites in these states were sampled by Bogart and Klemens (1997), and the distribution provided in figure 6 includes our earlier findings. Both A. laterale as well as A. jeffersonianum and their unisexual associates are widespread in these states.
New Jersey
Ambystoma laterale, A. jeffersonianum, and unisexuals have previously been reported to occur in New Jersey (Uzzell, 1964; Anderson and Giacosie, 1967). Unisexuals were found in 15 of the 17 New Jersey sites that we sampled (appendix A). The only site where unisexuals were not collected was an A. laterale site in Morris County (site 154: n = 13) in northern New Jersey. Even though more A. jeffersonianum individuals (n = 31) were collected from more sites in New Jersey than were those of A. laterale (n = 28) (table 4), no site was found for A. jeffersonianum that did not also have unisexual individuals (fig. 4).
New York
Unisexuals were found in most (43 of 57) New York sites (appendix 1) (fig. 5). The northern New York sites in St. Lawrence County (sites 149 to 152) were dominated by A. laterale, and only A. laterale (n = 17) was found at site 150. Unisexual LLJ were found with A. laterale in sites 151 (LL: n = 7; LLJ: n = 2) and 152 (LL: n = 8; LLJ: n = 2). A single LLJ individual was collected at site 149. Other sites where we found only A. laterale were in western New York counties of Erie (site 146: n = 1) and Niagara (site 147: n = 7); in the northeastern county of Essex (site 153: n = 1); and in the southeastern counties of Dutchess (site 206: n = 2) and Putnam (site 210: n = 5), but with such small sample sizes, these sites could also contain unisexual individuals. Ambystoma jeffersonianum were found in populations throughout New York (fig. 5). Populations containing only A. jeffersonianum were Tompkins County (site 2: n = 38; site 120: n = 14), Sullivan County (site 121: n = 20), and Otsego County (site 118: n = 16) in east-central New York. The single A. jeffersonianum specimens from site 116 in Ulster County is probably not a pure A. jeffersonianum site because site 116 is in the same drainage basin and is close to site 117 where LJJ and LJJJ individuals were found. Other New York sites where only A. jeffersonianum were found were represented by only one or two individuals: Cattaraugus County in western New York (site 135: n = 1) and Columbia County (site 212: n = 2), as well as Westchester County (site 205: n = 1) in southeastern New York.
Pennsylvania and Virginia
Ambystoma laterale and unisexuals have not been reported to occur in Pennsylvania. As expected, most of the Pennsylvania sites contained only A. jeffersonianum, but unisexual LJJ (n = 7) and LJJJ (n = 1) were found with A. jeffersonianum (n = 9) at site 179 in Monroe County in eastern Pennsylvania. Ambystoma jeffersonianum, LJJ, and LJJJ from site 179 are shown in figure 3. We also found one LJ and three LLJ unisexual individuals at site 174 in Northampton County, which is also in eastern Pennsylvania. Individual A. laterale were not found at site 174 but is assumed to be the sperm donor for these unisexuals (62.5% L, table 4). Ambystoma laterale (n = 45) was found in McKean County (site 189), in north-central Pennsylvania, with LLJ (n = 20) unisexuals. Ambystoma laterale, LJ, and LLJ individuals from these sites are provided in figure 2. The distribution of sexual and unisexual individuals that we found in Pennsylvania is plotted in fig. 4. Only four individuals were sampled from one site (site 190) in Virginia. They were all A. jeffersonianum.
Discussion
We found a lower ratio of unisexual to sexual individuals than were found by Bogart and Klemens (1997) in populations from New England and mostly eastern New York. We can account for the discrepancy because, in the present study, we sampled large numbers of A. jeffersonianum from populations in Pennsylvania where unisexuals probably do not exist and we sampled some more northern populations in New York where unisexuals were found to be at a lower frequency than A. laterale. But, in keeping with our previous study, in most populations where unisexuals coexist with sexual individuals, the unisexuals are still more numerous. During the course of this study, we visited some of the same sites in different years to increase our sampling of populations to find rare genomotypes, which can confirm the absence or presence of sexual and unisexual individuals. However, this was not always possible, especially if the salamanders were rare or difficult to collect and many sites are represented by one or two individuals. Because the unisexuals require a sperm donor to successfully reproduce, if a sexual individual was not encountered, we predicted the presence of A. laterale or A. jeffersonianum based on the genomotypes of the unisexuals. It was not possible to predict the absence or the presence of unisexuals. The possibility exists that a population might change over time or that our sampling was biased. Eight of the 118 sites (marked † in table 4) sampled in the present study were also sampled for our earlier study (Bogart and Klemens, 1997). The species and genomotypes of the unisexuals were similar for most of the resampled populations. Increasing the sample size from three (Bogart and Klemens, 1997) to 38 individual A. jeffersonianum at site 2 in western New York supported our earlier contention that unisexuals are absent at that site. We did find A. jeffersonianum in Hartford County at Granby (site 195) and King Phillip Mountain near Simsbury (site 196), but we also found A. laterale at East Granby (site 197) and Farmington (site 198) near Rattlesnake Mountain. The Burnt Hill site (site 199), also near Farmington, yielded a single LJJ. A site near Wethersfield (site 201) was unusual as we found only seven diploid LJ unisexuals and, therefore, we can not predict a putative sperm donor for that site. Evidently, both sexual species are present and in close proximity in this region of Hartford County.
No Apparent Syntopic Association of the Sexual Species
Ambystoma laterale and A. jeffersonianum were not found together in any of the 118 sites and there were only four sites where both LLJ and LJJ were found (sites 108 and 113 in New York, site 161 in New Jersey, and site 197 in Connecticut) (appendix 1; table 4). Both New York sites were in Orange County. At site 108 we found LJJ (n = 6), LLJ (n = 17), LLLJ (n = 3), and LJJJ (n = 3) unisexual individuals. Ambystoma laterale (n = 14) was the only sperm donor collected at that site, but the LJJ triploids and the LJJJ tetraploids indicate the sympatric presence of A. jeffersonianum. Based on egg mass data (Bi and Bogart, 2006; Bi et al., 2007; Bogart et al., 2007) tetraploids are often found among the triploid offspring from individual triploid unisexuals through ploidy elevation that may be mediated by elevated temperature (Bogart et al., 1989). The additional J genome in the LJJJ tetraploids would be an indication that A. jeffersonianum was the sperm donor if the LJJJ unisexual individuals were derived from a triploid LJJ female, but the tetraploids could have been gynogenetic offspring from a tetraploid LJJJ female. Ambystoma laterale and associated LLJ and LLLJ unisexuals were found in other Orange County sites 107, 110, 111, and 113. Ambystoma jeffersonianum, LJ, and LJJ were found in Orange County (site 114). If A. jeffersonianum does occur at site 108, it must be less common than A. laterale and that population might demonstrate the phenomenon of one sexual species displacing the other over time. Making a case for the presence of A. jeffersonianum in the other Orange County site that is dominated by A. laterale (site 113) is more tenuous because, of the 43 specimens collected at that site, there was only one LJJ individual and no LJJJ tetraploids. Dispersal of LJJ unisexuals into an A. laterale/unisexual site would be a possible alternate explanation.
Neither sexual species was collected at site 161 in Warren County, New Jersey, where LLJ (n = 3) and LJJ (n = 7) were found. The tetraploid LJJJ that was also collected at site 161 provides similar evidence that A. jeffersonianum is most likely the sperm donor. In that same county, both A. laterale (n = 5) (site 162) and A. jeffersonianum (site 159: n = 2; site 163: n = 3) were also collected. Based on our collection, the Hartford, Connecticut, site (198) is an A. laterale/unisexual site because A. laterale (n = 4) and three LLLJ tetraploids were collected, but the three LJJ unisexuals that were also collected could indicate that this site also has A. jeffersonianum as have been found in other Hartford County sites (195 and 196).
From the relatively few examples of the sympatric occurrence of LLJ and LJJ, it is evident that all such instances exist only where A. laterale and A. jeffersonianum are parapatric and possibly sympatric over time and space. In a southern Ontario population where A. laterale and A. jeffersonianum were found to be breeding in the same pond, LJJ and LLJ unisexuals shared microsatellite alleles (Bogart et al., 2007). In that study, A. laterale was found to be an acceptable sperm donor for LJJ unisexual females because LLJJ larvae were encountered in the same egg mass with LJJ larvae. Through genome replacement, the switch from LJJ to LLJ or vice versa is probably a rapid and widespread phenomenon within the unisexual complex (Bogart, 2003) and would explain our observations in the few populations where this process is probably occurring.
Diploid and Tetraploid Unisexuals
Most unisexuals in this and previous studies (Uzzell, 1964; Bogart et al., 1987; Bogart and Klemens, 1997) are triploid. Tetraploids can be produced in the laboratory (Bogart et al., 1989) and are found to be derived in nature from triploid individuals (Bogart et al., 2007) through the process of ploidy elevation where a sperm nucleus is incorporated in an unreduced, triploid egg. We found tetraploids in scattered populations (table 4) and triploids were always found in the same populations. Diploid unisexuals are much more difficult to explain. They, too, are scattered in various populations (table 4) and, in a few populations, they were more numerous than the triploids (e.g., sites 123, 125, 137, 214) or were the only unisexual genomotype found (sites 134, 201). They occur in sites with either A. laterale or A. jeffersonianum. Diploid LJ and triploid LLJ or LJJ unisexuals share the same microsatellite alleles when they are sympatric and diploid LJ unisexuals can produce triploids through ploidy elevation events (Bogart et al., 2007). Ploidy reductional events probably occur from triploid to diploid unisexuals, but empirical evidence is lacking. Two very interesting sites (sites 123 and 125) in Schoharie County, New York, have LJ and LLJ unisexuals as well as Ambystoma jeffersonianum. The other Schoharie site (124), which also has A. jeffersonianum and LJ unisexuals, has the expected LJJ unisexual genomotype. We did not find any A. laterale in Schoharie County. Nor did we find A. laterale in Albany (site 126) or Otsego (site 118) counties that are, respectively, north and south of Schoharie County. However, in our 1997 paper we found A. laterale and its associated hybrids in Albany County (site 3). Only A. jeffersonianum (n = 16) was found at site 118. We found A. jeffersonianum (n = 4), as well as LJ (n = 5), LJJ (n = 18), and LJJJ (n = 2) unisexuals at site 126.
“Unisexual” Males
We found a very low frequency of “unisexual” males in this investigation and in our earlier study (Bogart and Klemens, 1997). They were diploid LJ (sites 114, 122), triploid LLJ (sites 108, 112, 152, 162), triploid LJJ (Site 165), and tetraploid LLLJ (site 148). Such males are probably sterile based on the limited cytological evidence of mieotic chromosomes (Bogart, 2003). It is possible that an answer to the rare finding of “unisexual” males may be found among the chromosome translocations and recombinations (Bi and Bogart, 2006; Bi et al., 2007) that may involve sex-determining genes, but sex-determining genes have not been mapped in unisexuals to any chromosome region or, indeed, to any chromosome. The possibility exists that these rare males, even if they are sterile, could stimulate gynogenetic development of unisexual eggs. This might be important if other male sexual sperm donors are not available in a breeding pond. Obviously, natural selection should favor the production of female unisexuals.
Homozygosity and Reversals of Diagnostic Alleles in Unisexual Individuals
Identification of unisexual genomotypes is based on observations of allozymes that are diagnostic for A. laterale and A. jeffersonianum and demonstrate the appropriate dosage in diploids and polyploids. Homozygous allozyme patterns at these loci identify the sexual species and a reversed dosage pattern of allozymes at a locus would misidentify a LLJ as an LJJ or vice versa. Misidentification and ambiguity is alleviated by using several loci that possess diagnostic allozymes, so it is possible to recognize that reversals can and do occur, but if an individual had a reversed pattern for several loci it could be misidentified. If unisexuals were identical, genetic clones, as has been previously proposed (Uzzell, 1964; Macgregor and Uzzell, 1964; Uzzell and Goldblatt, 1967), a unisexual lineage could not revert from a homozygous to a heterozygous condition and reversed patterns, once attained, would be fixed (Asher and Nace, 1971). Fluorescent genomic in situ hybridization (GISH) of unisexual chromosomes (Bi and Bogart, 2006; Bi et al., 2007) and microsatellite analyses (Bogart et al., 2007) clearly show that unisexuals are not clones and, based on chromosomal translocations and recombinations, may be expected to demonstrate gene restructuring. Thus, we may be surprised that observed homozygous and reversed-allozyme alleles are not more common and widespread than the few instances that we found in the present study (tables 5 and 6) and in our earlier study (Bogart and Klemens, 1997). We suspect that natural selection eliminates many such mutational events and helps to explain the embryonic mortality that is ubiquitous in unisexual populations. Unisexual kleptogens swap genomes with sexual sperm donors (Bi et al., 2008). This process may be essential for unisexuals to recover from detrimental genetic restructuring and consequentially, maintain a fixed heterozygous genotype that is most commonly found in unisexual individuals and likely has a selective advantage.
Rare Alleles
Loci chosen to identify the sexual and unisexual individuals were loci that were mostly monomorphic for alternate alleles in A. jeffersonianum and A. laterale, but these same loci are not considered to be very conservative in other vertebrates. When sampling individuals from many populations over a large range we should expect to find polymorphism and population specific “private” alleles. The very low frequency of some rare alleles may be attributed to mutational events and especially if a single rare allele is found in only one individual. Finding rare alleles among sexual and unisexual individuals in the same populations adds credence to kleptogenesis.
Some rare alleles that were found in both A. jeffersonianum and A. laterale such as Aat-2−50, Mdh-1200or Mpi80 (table 3) may be alleles that have been passed on from a common ancestor and are maintained at a low frequency in both species. A rare allele demonstrates that a diploid-triploid unisexual relationship exists in site 130. The only unisexuals that were found to have Idh-1160 were LJ at site 130 and LJJ from sites 130 and 131. That allele was not found in A. jeffersonianum from those sites but was found in individual A. jeffersonianum in sites 118 and 204 and in A. laterale from sites 113 and 189. Ldh-178 was not found by Bogart and Klemens (1997). We found this allele in A. laterale populations (sites 127, 134, 136, 143, 145) in western New York and in site 189 in north-central Pennsylvania. The unisexual LLJ individuals from the Pennsylvania site that also had this allele were homozygous for Ldh-178. LLJ homozygotes for this allele were also found in site 148 close to Lake Erie in New York. The relatively high frequency of Ldh-178 in western New York A. laterale individuals is a clear indication that A. laterale genomes that contain this allele have been used to replace L genomes in sympatric unisexuals. The homozygous condition of Ldh-178 in the unisexuals demonstrates that the diagnostic A. jeffersonianum Ldh-188 allele was probably lost through genome translocations or recombination or is no longer functional in the homozygous unisexuals.
Summary and Conclusions
Identification of Ambystoma jeffersonianum and A. laterale and their breeding ponds in the northeastern United States in this, and in our earlier, study should benefit conservation efforts to preserve and maintain these species. Populations that we examined as many as 20 years ago may no longer exist. The interaction of unisexuals with the sexual species is a fascinating evolutionary phenomenon that is probably unique. Unisexual kleptogens take genomes from either sexual species and are usually more numerous in sympatric association with a sexual species. Unisexuals were not found in most of the A. jeffersonianum populations in Pennsylvania and are less numerous than A. laterale in northern New York populations. Perhaps the unisexuals are recent immigrants in those New York populations and have not yet dispersed throughout Pennsylvania. The fact that no unisexuals were found with A. laterale in Long Island (New York), eastern Connecticut, and southeastern Massachusetts (Bogart and Klemens, 1997) could also be the result of isolation of those populations prior to unisexual dispersion. Although no new samples were obtained from Vermont, New Hampshire, and Maine, we have included the ranges of known sexual and unisexual individuals from those states in figure 7 from our earlier investigation (Bogart and Klemens, 1997). Figure 8 includes the known ranges for these salamanders in the northeastern United States from the present and previous study.
Fig. 6
Distribution of Ambystoma jeffersonianum, Ambystoma laterale, and unisexuals in Connecticut, Massachusetts, and Rhode Island.
Fig. 7
Distribution of Ambystoma jeffersonianum, Ambystoma laterale, and unisexuals in Maine, New Hampshire, and Vermont.
Fig. 8
Distribution of Ambystoma jeffersonianum, Ambystoma laterale, and unisexuals in northeastern U.S.
In addition to A. laterale and A. jeffersonianum, unisexuals use other species (A. texanum, A. tigrinum) as sperm donors in Ohio, Illinois, Indiana, and Michigan (Kraus, 1995; Morris, 1985; Kraus and Petranka, 1989; Selander, 1994; Bogart, 2003). Distinguishing sexual and unisexual individuals in those states is complicated because A. texanum and A. laterale have the same allozymes for some of the diagnostic loci (Bogart et al., 1987). It is likely that isozyme electrophoresis will be replaced as a technique to identify these salamander species and their associated unisexuals. Sequences of mitochondrial DNA show all of the sexual species to be distinctive and distinctly different from the unisexuals (Hedges et al., 1992; Bogart, 2003; Bogart et al., 2007). Microsatellite DNA alleles can identify A. jeffersonianum, A. laterale and the unisexuals that use those species as sperm donors (Julian et al., 2003; Ramsden et al., 2006; Bogart et al., 2007). Because these DNA techniques do not require sacrificing individuals, there are obvious advantages when attempting to identify rare, threatened, or endangered salamanders. Identifying the species and unisexuals is a prerequisite to protecting the species and their habitat. Subsequent temporal monitoring will be necessary to understand the interactions of the species and the unisexual associates.
Acknowledgments
We wish to thank the many friends, colleagues, and students who helped us with the collecting of samples, for logistical support in the field, and assisting in the laboratory. Julian D. Avery, Michael Bernard, David H. Black, Randy Cassell, Amy R. Croushore, Peter Ducey, the late Leigh Gillett, Brian Gray, Hank J. Gruner, Christy Hicks, Art Hulse, Steve Kahl, Ed Kanze, Andy Keal, Daniel S. Klemens, Robert M. Klemens, Rick Koval, Mark Lethaby, George Logan,, Michael Losito, Tim Maret, Diane Murphy, Linda Ordiway, Damon Oscarson, Ed Pawlak, Dennis Quinn, Chris J. Raithel, Kenneth Roblee, Brandon M. Ruhe, Kevin J. Ryan, Jason Tesauro, Peter R. Warny, and Kristian Whiteleather assisted with the collection of salamanders. Dave Britton, Heather Lynn, Lesley Rye, Sara Wick, and Karlee Thomas all helped to process salamanders and assisted with the electrophoretic analyses. Karine Bériault measured blood cells and performed much of the flow cytometry. Dan Noble and Ke Bi made helpful comments on an earlier manuscript. We wish to especially thank Danielle T. LaBruna for creating the distribution maps and Kevin J. Ryan for assistance in rearing larval salamanders. We acknowledge use of the collections of the American Museum of Natural History (AMNH). All samples were collected under provisions of various state collecting permits and licenses and are on file at the AMNH. All laboratory procedures followed approved Animal Care Protocols from the University of Guelph Animal Care Committee. Financial support for portions of this study was provided by the Wildlife Conservation Society and the National Science and Engineering Research Council of Canada.
References
Appendices
Appendix 1
Additional Individuals Collected from Previous Collection Sites Examined by Bogart and Klemens (1997) Localities listed by site numbers where Ambystoma laterale (LL), A. jeffersonianum (JJ), and their diploid (LJ), triploid (LLJ or LJJ), and tetraploid (LLLJ, LLJJ, LJJJ) unisexual kleptogens were identified. New sites examined (sites 107–216) continue from the 106 sites examined by Bogart and Klemens (1997). Catalogue numbers refer to specimens that are deposited in the American Museum of Natural History (AMNH) and catalogue numbers of J. P. Bogart (JPB).
Appendix 2
Genotypes of Ambystoma laterale, A. jeffersonianum, and unisexual specimens ordered by site number
Genotypes of the specimens are ordered by site number (see appendix 1). Specimen numbers are American Museum of Natural History (AMNH) or catalogue numbers of J. P. Bogart (JPB). The sex of the specimen is female (F) or male (M). ?, M? or F? are juvenile individuals where the sex was difficult to determine, and L is a larval sample. The letters for the electrophoretic alleles, or allozymes, for each locus refer to relative mobilities of the stained enzyme on the gel (see table 4). A single letter is given for a homozygous condition so, for example, a B would be BB in a diploid, BBB in a triploid and BBBB in a tetraploid. Data that were not obtained for an individual at an isozyme locus is signified by –––. Data for “blood” are mean erythrocyte area determinations or flow cytometric (FMC) determinations for ploidy based on comparative fluorescence of blood cells from individuals against a diploid standard fluorescence (see Ramsden et al. 2006). Blood data that were not obtained for an individual is signified by –––. Tables in this appendix compare individuals that have the same genomotypes:
2-1: Ambystoma laterale (LL) diploids from all sites.
2-2: Ambystoma jeffersonianum (JJ) diploids from all sites.
2-3: Ambystoma laterale – jeffersonianum (LJ) diploids from all sites.
2-4: Ambystoma (2) laterale – jeffersonianum (LLJ) triploids from all sites.
2-5: Ambystoma laterale – (2) jeffersonianum (LJJ) triploids from all sites.
2-6: Ambystoma (3) laterale – jeffersonianum (LLLJ) tetraploids from all sites.
2-7: Ambystoma laterale – (3) jeffersonianum (LJJJ) tetraploids from all sites.
2-8: Ambystoma (2) laterale – (2) jeffersonianum (LLJJ) tetraploid from site 112.
Appendix 2-1: Genotypes of diploid (2n) Ambystoma laterale specimens ordered by site number.
Appendix 2‐2: Genotypes of diploid (2n) Ambystoma jeffersonianum specimens ordered by site number.
Appendix 2-3: Genotypes of diploid (2n) Ambystoma laterale–jeffersonianum LJ unisexual specimens ordered by site number.
Appendix 2-4: Genotypes of triploid (3n) Ambystoma 2 laterale – jeffersonianum (LLJ) specimens ordered by site number.
Appendix 2-5: Genotypes of triploid (3n) Ambystoma laterale – (2) jeffersonianum (LJJ) specimens ordered by site number.
Appendix 2-6: Genotypes of tetraploid (4n) Ambystoma (3) laterale – jeffersonianum (LLLJ) unisexual specimens ordered by site number.
Appendix 2-7: Genotypes of tetraploid (4n) Ambystoma laterale – (3) jeffersonianum (LJJJ) unisexual specimens ordered by site number.
Appendix 2-8: Genotypes of tetraploid (4n) Ambystoma (2) laterale – (2) jeffersonianum (LLJJ) unisexual specimen from site 18.
