Cnemidophorus neomexicanus, Lowe & Zweifel, 1952

TAYLOR, HARRY L., COLE, CHARLES J., HARDY, LAURENCE M., DESSAUER, HERBERT C., TOWNSEND, CAROL R., WALKER, JAMES M. & CORDES, JAMES E., 2001, Natural Hybridization Between the Teiid Lizards Cnemidophorus tesselatus (Parthenogenetic) and C. tigris marmoratus (Bisexual): Assessment of Evolutionary Alternatives, American Museum Novitates 3345, pp. 1-1 : 1-

parthenogen C. neomexicanus and the bisexual C. tigris have the greatest color pattern

resemblance to their paternal parent, C. tigris , in an area where C. tigris is highly polymorphic (Dessauer et al., 2000). These hybrids, like those from Arroyo del Macho, have two sets of tigris chromosomes, with C. neomexicanus rather than C. tesselatus contributing one of the two sets.

KARYOTYPIC ANALYSIS

Only one clearly resolved karyotype of the allodiploid C. tesselatus has been published (Dessauer and Cole, 1989: 57). However, additional studies of karyotypes (Wright and Lowe, 1967b; Lowe et al., 1970b) and of protein electrophoresis (Neaves and Gerald, 1968, 1969; Neaves, 1969; Parker and Selan­ der, 1976; Dessauer and Cole, 1989) reveal that the diploid unisexual C. tesselatus had a hybrid origin involving C. tigris marmoratus X C. gularis septemvittatus .

Cnemidophorus tigris marmoratus and C. gularis septemvittatus belong to the tigris and sexlineatus species groups, respectively. Each species group has a diagnostically distinctive karyotype (Lowe et al., 1970b). The C. tigris marmoratus complement (n = 23) consists of three large set I biarmed macrochromosomes + eight smaller set II biarmed intermediate­sized macrochromosomes + 12 set III microchromosomes. The second largest chromosome in set I of C. tigris has a dotlike satellite on the end of one arm, which is often difficult to see, and the third largest chromosome is the sex chromosome (Cole et al., 1969; Bull, 1978), of which the X­chromosome is recognizable in the karyotype of C. tesselatus . The complement from C. gularis septemvittatus (n = 23) consists of only one large set I metacentric macrochromosome (with a subterminal secondary constriction on one arm followed by an elongate satellite) + 12 smaller set II intermediate­sized telocentric or subtelocentric macrochromosomes + 10 set III microchromosomes. The sex chromosomes of C. gularis septemvittatus are not morphologically recognizable. The secondary constrictions on the set I chromosomes are the nucleolar organizer regions (Ward and Cole, 1986).

As expected, three representatives of C. tesselatus from Arroyo del Macho had a diploid karyotype consisting of one normal tigris group haploid complement and one normal sexlineatus group haploid complement of chromosomes, or 2 n = 46 ( fig. 9A ). The other seven C. tesselatus from Arroyo del Macho that were karyotyped were all of a slightly derived karyotypic clone (2 n = 47) in which the X­chromosome of C. tigris had apparently undergone centric fission, as it was represented by two telocentric chromosomes, each being the size of one of the arms of the ancestral X. Such karyotypic variants among parthenogenetic species are known to perpetuate themselves through cloning (Cole, 1979).

Nine suspected hybrids of C. tesselatus X C. tigris marmoratus from the Arroyo del Macho site were karyotyped. Eight of these were triploids having 3 n = 69 chromosomes,

including the full diploid karyotype of C. tesselatus with the ancestral tigris X­chromosome unfissioned plus a second haploid complement of tigris chromosomes ( fig. 9B ). One individual was a modified triploid having 3 n = 70 chromosomes, with the fissioned tigris X­chromosome inherited from C. tesselatus . The four female hybrids had the tigris X­chromosome in both tigris complements, whereas the five male hybrids had the tigris Y­chromosome in the second tigris complement ( fig. 9B ). Clearly, these individuals all appeared to be F 1 hybrids between C. tesselatus X C. tigris marmoratus .

RELATIONSHIP BETWEEN KARYOTYPIC CYTOTYPE AND MORPHOLOGY

There was an unequal distribution of the two cytotypes between C. tesselatus and the hybrids. The fissioned X­chromosome was found in a majority of C. tesselatus karyotyped (7 of 10), but only one hybrid had the fissioned X­chromosome (1 of 9). This disparity might reflect different effects of the fissioned X in the two groups. Although the type of X­chromosome inherited does not affect the expression of meristic characters in C. tesselatus , the presence of a fissioned X may exaggerate the expression of certain meristic characters in hybrids—with potentially negative effects. Evidence supporting this possibility comes from both univariate and multivariate analyses. At the univariate level, the two karyotypic clones in C. tesselatus (2 n = 46 [intact X] and 2 n = 47 [fissioned X]) are similar (t ­tests; P s> 0.38) in all 12 meristic characters (see Materials and Methods section for character descriptions), supporting a hypothesis that the two cytotypes confer equivalent reproductive success in C. tesselatus . Whether the 3:7 ratio reflects the true frequencies of the two cytotypes in the population of C. tesselatus is unknown; it could simply represent a deviation from a 1: 1 ratio based on sampling variation (Chisquare = 1.600, 1 df, P = 0.21). Nevertheless, the ratio of 8 hybrids with intact Xchromosomes to 1 hybrid with a fissioned Xchromosome is a significant deviation from the 3:7 ratio in C. tesselatus (Chi­square [with Yate’s correction for small samples] = 4.539, 1 df, P = 0.03).

To summarize the morphological differences associated with the two cytotypes, we used 11 meristic characters (table 3) in a principal components analysis to compare the hybrid with a fissioned X­chromosome to hybrids with intact X­chromosomes. Principal components scores for the only hybrid with a fissioned X­chromosome lie outside the 95% confidence ellipse for the sample of the other eight hybrids ( fig. 10 ). All significant differences between the two hybrid cytotypes were conveyed by the first principal component (t = 12.804, 1 df, P <0.00005; table 4). This component provided a contrast between the three color pattern characters and number of granules around midbody (all with positive loadings) and two characters associated with the hindlimb, number of subdigital lamellae on the fourth toe and number of femoral pores (both with negative loadings; table 3).

The disjunct position of the hybrid with a fissioned X­chromosome in the ordination of PC scores ( fig. 10 ) was based on striking differences between the two cytotypes in univariate scores (table 4). We used one­sample t ­tests on individual meristic characters to determine if the unusual hybrid was significantly different from the sample of hybrids with intact X­chromosomes. The hybrid with a fissioned X had fewer pale segments in the vertebral field (t = 6.804, 1 df, P = 0.0003), fewer interruptions in the dorsolateral stripes (t = 6.065, 1 df, P = 0.0005), fewer interruptions in the paravertebral stripes (t = 6.174, 1 df, P = 0.0005), fewer granular scales around midbody (t = 12.477, 1 df, P <0.0005), more femoral pores (t = 7.332, 1 df, P = 0.0002), more circumorbital scales (t = 4.314, 1 df, P = 0.0035), more subdigital lamellae on the fourth toe (t = 3.000, 1 df, P = 0.0199), and more scales contacting the outer perimeter of the parietal and interparietal scales (t = 7.483, 1 df, P = 0.0001) (table 4). Whether these deviations, or others undetected by us, translate into higher mortality in hybrids inheriting a fissioned Xchromosome is unknown. The low frequency of the fissioned X cytotype in our sample of hybrids compared to its higher frequency in C. tesselatus indicates that this might be the case. However, alternative hypotheses, such as a lower susceptibility of individuals of C. tesselatus with fissioned X­ chromosomes to hybridization, cannot be tested with our evidence.

EVIDENCE FROM BIOCHEMICAL GENETICS

Based on genotypes detected at 34 loci, we obtained evidence bearing on the following questions: (1) how many electrophoretically detected clones of C. tesselatus are present in our sample from the Roswell area; (2) were the suspected hybrids actually hybrids; (3) were the male parents of the hybrids representatives of C. tigris marmoratus ; (4) was there evidence for separate fertilization events in the origin(s) of these hybrids; and (5) are the parental taxa at the hybridization

Derived from Meristic Variation Between Two Cytotypes of Hybrids Derived from the Parthenogenetic Species Cnemidophorus tesselatus X the Bisexual Species C. tigris marmoratus

site genetically similar to individuals representing the same taxa from other localities?

All 11 C. tesselatus examined electrophoretically, including a laboratory­reared offspring and her mother, had identical genotypes at each of the 34 loci (table 5). Seventeen loci (50%) had alleles in the heterozygous state, attesting to the origin of this taxon from a hybridization event, and the specific alleles at each locus were consistent with the parental taxa being C. tigris marmoratus and C. gularis septemvittatus (Neaves, 1969; Parker and Selander, 1976; Dessauer and Cole, 1989). The presence of identical genotypes in the laboratory­reared specimen, its mother, and every other individual of these C. tesselatus showed that they reproduce by parthenogenetic cloning, as do other individuals of different pattern classes of this taxon (Dessauer and Cole, 1986). Only one electrophoretic clone of tesselatus was detected at the hybridization site. Because individuals examined included both of the local karyotypic clones of C. tesselatus (see above), no molecular marker was correlated with the two cytotypes. In addition, these representatives of C. tesselatus were electrophoretically identical to other speci­

for 11 Meristic Characters and Scores of Four Principal Components for Two Cytotypes in Cnemidophorus tesselatus X C. tigris marmoratus Hybrids Scores for one hybrid with a fissioned X­chromosome are compared with scores (mean ± SE and range) from a sample of eight hybrids with intact X­chromosomes.

mens of pattern class E from several other localities (work in progress), although more than one clone does exist among specimens examined from elsewhere (also see Parker and Selander, 1976).

The alleles found in the four specimens of C. tigris marmoratus from the hybridization site (table 5) were also the same ones commonly found in specimens of this taxon from other localities (Dessauer et al., 2000; however, we did not cross­correlate allele designations for the uncommon alleles found in their large samples from southwestern New Mexico). Despite our small sample from the Roswell area , four loci ( IDDH, PEPA [by deduction from hybrids], PEPD, and GPI) showed local polymorphism in this bisexual species that reproduces with a Mendelian pattern of inheritance .

TABLE 5 Genotypes or Allelesa at 34 Gene Locib in Samples of Cnemidophorus from the Hybridization Site

Of the 34 loci analyzed, 15 showed no variation among all individuals of each taxon

and the hybrids examined (the same alleles were shared universally), but 19 loci were particularly informative for identifying hybrids and their parental species (table 5). For each locus, all 10 suspected hybrids (based on morphology and karyotypes) had electrophoretic banding patterns consistent with triploids bearing a combination of alleles that included the two alleles found in the diploid C. tesselatus plus a third allele from the local C. tigris marmoratus . This is consistent with a cloned tesselatus ovum having been fertilized by a haploid marmoratus spermatozoan (table 5).

Although we did not examine, electrophoretically, representatives of C. inornatus from the hybridization locality, this was not necessary. Previous studies (e.g., the review by Cole et al., 1988) have shown that C. tigris marmoratus and C. inornatus are electrophoretically distinguished from each other at many loci, including the following eight in table 5: ADH, LDH1, sSOD, sAAT, mAAT, PEPE, ADA, and TF. In addition to the third haploid set of chromosomes being diagnostic of marmoratus in the hybrids reported here (see above), the alleles detected at these loci were also those specifically of marmoratus , not inornatus .

The presence of the extra marmoratus allele in hybrids was detected at most loci based on allele dosage effects on band densities (isozyme activities) in electrophoretic phenotypes. For example, at GPI, specimens of the diploid C. tesselatus , which are heterozygotes, show a three­banded pattern. Hybrids show the same three bands, but the relative band densities differ from those of tesselatus . Phenotypes (gel patterns) predicted for the ab genotype of this dimeric enzyme by the expansion of (a + b) 2, which equals a 2 + 2ab + b 2, consist of three isozymes with a ratio of activities (band densities on gels) approximating 1:2:1, as observed for tesselatus ( fig. 11 ). Phenotypes predicted for the abb genotype predicted by the expansion of (a + 2b) 2, which equals a 2 + 4ab + 4b 2, consist of the same three isozymes but with a ratio of activities approximating 1:4:4, as observed in triploid hybrids ( fig. 11 ). Consequently, tesselatus and the hybrids had three­ banded patterns for GPI, but the ratio of the band densities differed between them.

Banding patterns for IDDH and PEPA, however, involved differences in both the number of isozymes and the band densities present in C. tesselatus versus hybrids. All hybrid genotypes at IDDH included the ballele of marmoratus ( fig. 12). Phenotypes predicted for the aab genotype of this tetrameric enzyme by the expansion of (2a + b) 4, which equals 16a 4 + 32a 3 b + 24a 2 b 2 + 8ab 3 + b 4, consist of five isozymes with a ratio of activities approximating 16:32:24:8:1. The least active isozyme (b 4) is not visible in the hybrids in figure 12.

Although dosage effects were not as clear for PEPA as for IDDH, comparison of the tesselatus and hybrid patterns in figure 13 clearly shows that the banding pattern of the hybrids varies according to which allele was inherited from marmoratus . Nine of the 10 hybrids had phenotypes of five isozymes. Dosage effects for the triple heterozygote bcd of this dimeric enzyme predict a sixbanded phenotype, estimated by the expansion of (b + c + d) 2, which is b 2 + 2bc + c 2 + 2bd + 2cd + d 2, with band intensities approximating the ratio of 1:2:1:2:2:1. Five of the six isozymes were resolved in these triple heterozygotes ( fig. 13 ). Note that the middle band of the triple heterozygotes is the most dense, suggesting that the c 2 and 2bd bands are superimposed to produce a fivebanded pattern with a ratio of 1:2:3:2:1.

One would expect a sample of 10 hybrids from the Roswell area to include individuals with different genotypes at IDDH, PEPD, and GPI because the local population of marmoratus is polymorphic at these loci (table 5). For both IDDH and GPI, all 10 hybrids inherited the b­allele, which has a frequency of 0.75 in the local marmoratus . For PEPD, six hybrids inherited the b­allele and four inherited the c­allele, which have frequencies

as follows in the local marmoratus : b = 0.25, c = 0.75. In addition, for PEPA, nine hybrids inherited the b­allele (bcd genotype), whereas one inherited the c­allele (ccd genotype) from marmoratus . The marmoratus in our small sample from the Roswell area expressed only the b­allele at PEPA, but the callele is present in marmoratus from other sites (frequency of about 0.08; Dessauer et al., 2000).

Finally, the apparent sterility of the triploid female hybrids (see Histological Analysis, below) also is consistent with hypothesizing that a new triploid parthenogenetic clone has not been generated from the triploid hybrids at this site. If such a new clone were present, our sample of 10 triploids would have included a preponderance of females bearing identical genotypes, which is not the case. In fact, considering the combination of genotypes together with the presence of the Y­chromosomes observed, we have evidence for nine separate fertilization events (nine different combinations of eggs and sperm) among the 10 hybrids examined electrophoretically.

HISTOLOGICAL ANALYSIS

Figures

Fig. 9. Karyotypes of two whiptail lizards (Cnemidophorus) from Arroyo del Macho. A. Diploid unisexual C. tesselatus, AMNH R­146638 (2n = 46). This unisexual taxon is a clone with its ultimate ancestor being an F1 hybrid between C. tigris marmoratus (haploid complement of chromosomes in the upper two rows) X C. gularis septemvittatus (haploid complement in the lower two rows). Large arrow points to the centromere of the X­chromosome ultimately inherited from C. tigris marmoratus; this chromosome apparently had undergone centric fission into two telocentric chromosomes in the other karyotypic clone of C. tesselatus found at the same locality (2n = 47). B. Triploid hybrid between C. tesselatus X C. tigris marmoratus (male), AMNH R­146690 (3n = 69). The four rows of chromosomes above the scale bar represent the two haploid complements inherited from C. tigris marmoratus, whereas the lower two rows represent the single haploid complement from C. gularis septemvittatus. The Ychromosome was inherited from the most recent backcross hybridization between C. tesselatus and C. tigris marmoratus. Bar represents 10 µm. Small arrows in both A and B illustrate distinctive secondary constrictions (nucleolar organizer regions) from C. gularis septemvittatus.

Fig. 10. Pattern of morphological variation expressed by the distribution of scores on the first two principal components extracted from a correlation matrix of 11 meristic characters. Symbols depict scores of nine karyotyped hybrids collected in the vicinity of Arroyo del Macho, Chaves Co., New Mexico: •, 8 hybrids of 2n = 46; -, one hybrid of 2n = 47 (fissioned X­chromosome). The ellipse defines the 95% confidence limits for the eight hybrids with intact X­chromosomes.

Fig. 11. Electrophoretic phenotypes of GPI, a dimeric enzyme, from erythrocytes of 12 lizards. Letters below gel identify allozymes based on alleles present (table 5). Lanes for individual lizards are labeled beside the gel (with genotype) as follows: TESE, C. tesselatus pattern class E; SEP, C. gularis septemvittatus; MAR, C. tigris marmoratus; and HYB, C. tesselatus X C. tigris marmoratus hybrid. Although the SEP and MAR are identical on this gel (bb), the a­allele occurs in other individuals of MAR. All specimens except the SEP are from the hybridization site near Roswell. Anode is to the right; arrow indicates position of sample applications.

Fig. 13. Electrophoretic phenotypes of PEPA, a dimeric enzyme, for 12 lizards; letters and labeling as in figure 11, except asterisk marks hybrid (AMNH R­145146) that received the c­allele from MAR instead of the b­allele (coincidentally, this is the only hybrid karyotyped that has a fissioned X­chromosome). The c­allele occurs in other individuals of MAR, the d­allele in others of both MAR and SEP. Anode is to the right.