1530 NOTES AND COMMENTS

Evolution. 45(6). 1991. pp. 1530-1533

POECILIA MEXICANA IS THE RECENT FEMALE PARENT OF THE UNISEXUAL FISH P. FORMOSA

JOHN C. AVISE,' JOEL C. TREXLER,> JOSEPH TRAVIS,3 AND WILLIAM S. NELSON' 'Department ofGenetics. University ofGeorgia, Athens, GA, 30602 USA

Key words.-Asexuallineages, clonal evolution, gynogenesis, mitochondrial DNA.

Received September 21, 1990. Accepted January 23, 1991.

Poeciliaformosa, a small live-bearing fish native to these Poecilia fishes, and show unequivocally that P. northeastern Mexico, was the first recognized verte­ mexicana was indeed a recent female parent ofP. for­ brate with unisexual reproduction (Hubbs and Hubbs, mosa. 1932). This all-female "species" produces diploid apo­ mictic eggs (Rasch et aI., 1982; Monaco et al., 1984), and embryogenesis is subsequently activated by sperm MATERIALS AND METHODS from a related bisexual species, normally either P. la­ Fish were collected from the following locales (Table tipinna or P. mexicana (Turner et aI., 1980a, I980b). I): (I) Live Oak Island, Wakulla Co., FL; (2) Escondido The process apparently takes place without syngamy, Creek, Kleberg Co., TX; (3) Resaca del Rancho Viejo, such that genetic transmission is predominantly or ex­ Cameron Co., TX (Rio Grande drainage); (4) south of clusively clonal (Balsano et al., 1989). San Benito, Cameron Co., TX (Rio Grande drainage); The gynogen P. formosa almost certainly arose via (5) Rio Tigre, Tamaulipas, Mexico; (6) coastal lagoon hybridization of P. latipinna and P. mexicana (Hubbs drainage north of Altamira, Tamaulipas, Mexico. All and Hubbs, 1932; Turner, 1982). It is sympatric with assayed specimens of P. formosa were diploid. Al­ these species (Darnell and Abramoff, 1968), possesses though triploid P.formosaalso occurin some drainages an intermediate morphology (Hubbs and Hubbs, 1932; (Rasch and Balsano, 1989; Balsano et al., 1989), most Meyer, 1938), and exhibits nearly fixed heterozygosity notably the Rio Sota la Marina, none was represented at numerous protein and allozyme loci that distinguish in our collections. or are polymorphic in P. latipinna and P. mexicana To enhance mtDNA yields, most work was con­ (Abramoffet al., 1968; Balsano et al., 1972; Simanek, ducted with large, gravid females. Embryos, liver, and 1978; Turner, 1982). There is a relative paucity of muscle from individual fish provided the tissue source allozyme variation among the unisexuals, and the best for isolation ofmtDNA by either of two methods: (a) fit to the particular P. formosa alleles occurs in pop­ ultracentrifugation in cesium chloride gradients (Lans­ ulations ofP. mexicana in the Rio Tigre drainage near man et al., 1981); or (b) an "alkaline lysis" procedure Tampico, Mexico. This led Turner (1982) to hypoth­ (Tamura and Aotsuka, 1988). Mitochondrial DNA was esize that the gynogen may have arisen recently in this then digested with the following restriction enzymes area (perhaps from a single hybridization event), and that proved to produce multiple cuts in most samples: that its current geographic distribution reflects a rapid (I) AvaI; (2) AvaIl; (3) BamHI; (4) Bcll; (5) Bgll; (6) northward expansion via efficient colonizing ability. A EstEll; (7) EcoRI; (8) HincIl; (9) HindIII; (10) PvuIl; relatively recent origin for P. formosa is also consistent (II) StuI; and (12) XbaL The fragments were end­ with the finding that the species retains a capacity to labeled with 35S nucleotides, separated by molecular express male-specific genes and to undergo normal weight through 1% agarose gels, and revealed by au­ spermatogenesis upon experimental induction with ex­ toradiography, all according to standard procedures ogenous androgens (Turner and Steeves, 1989). (Brown, 1980; Lansman et al., 1981; Maniatis et aI., One open evolutionary question concerns the direc­ 1982). No attempt was made to score fragments less tion ofthe cross that produced P. formosa. Unlike the than about 0.6 kilobases in length. situation in hybridogenetic Poeciliopsis fishes from For unknown reasons, Poecilia proved rather re­ western Mexico, where results of laboratory hybrid­ fractory to the mtDNA isolation procedures that we ization experiments have been critical in deciphering have employed successfully with numerous other ver­ the male and female parents ofthe unisexuals (Schultz, tebrates (Avise et al., 1987). The mtDNA yields were 1973, 1989; Wetherington et aI., 1989), attempts to highly variable, often heavily contaminated with nu­ synthesize unisexual P. formosa through laboratory clear DNA, and many isolations were unsuccessful en­ crosses thus far have failed (Turner, 1982). Balsano et tirely. Altogether, among 187 individuals attempted, al. (1989) suspect that P. mexicana was the female only 44 (24%) provided adequate mtDNA for complete parent, but note that direct evidence from maternally scoring by all (or all except one) ofthe 12 informative transmitted mitochondrial DNA (mtDNA) might set­ endonucleases. These form the heart of our analysis. tle the issue. Here we survey mtDNA genotypes in An additional 71 individuals (38%) were scored for a NOTES AND COMMENTS 1531

TABLE 1. MtDNA haplotype descriptions in Poecilia fishes. Upper-case letters from left to right refer to the multi-fragment mtDNA digestion profiles for the 12 endonucleases listed in same order as in Materials and Methods. Locales are also described in Materials and Methods. No. of Collection MtDNA clone individuals locale Genotype description P. latipinna a 6 I C C C C C C C C C D C C b 2 I C D C C C C D C C C C C c 8 2, 3 C G C D C C D C C C D D P. mexicana d 9 6 X X D X X D X X D B X X e 2 6 X X D X X D X X E B X X f I 5 X X y D X D X X D B X X g I 5 X X D X X D X y D B X X Piformosa d 15 2-4 X X D X X D X X D B X X

small subset ofenzymes (usually 1-7), and hence pro­ in several other vertebrate vided ancillary groups (about 2% sequence information only. divergence per million years-Brown For most endonucleases, et al., 1979; the mtDNA digestion pro­ Shields and Wilson, 1987; Wilson files for P. mexicana et aI., 1985), then and P. latipinna were sufficiently these species may last have shared different to preclude a common ancestor simple site interpretations. There­ roughly three to four million years ago. fore, overall estimates of sequence divergence were All fully scored specimens ofP. formosa were based on the "fragment" comparison iden­ approach ofNei tical in assayed genotype to the most common mtDNA and Li (1979). A matrix ofgenetic distances between clone in P. mexicana. This finding is particularly mtDNA "clones" was clustered im­ phenetically using the pressive because these P. formosa were from unweighted pair-group three lo­ method with arithmetic means cales in Texas near the northern (Sneath and Sokal, 1973). distributional limit of the species, whereas the P. mexicana were from the REsULTS southern portion of P. formosa range near Tampico, Mexico (close to the Composite putative site of origin of P. for­ mtDNA designations for the 44 individ­ mosa-Turner, 1982). In Texas, uals fully assayed P. formosa uses P. are listed in Table I. An average of latipinna as sexual host exclusively 50 mtDNA (Turner, 1982). fragments was monitored per individual. Thus, beyond reasonable This represents doubt, P. mexicana rather approximately 283 base pairs (bp) of than P. latipinna contributed recognition the mtDNA now present sequence (or 1.7% ofthe Poecilia mtDNA in northern populations genome, ofP. formosa, and hence was which we estimate to be about 16.5 kb, in the female parent in the length). original cross or crosses pro­ ducing these unisexuals. Three A phenogram summarizing mtDNA haplotypes were observed among the relationships among 16 the fully assayed Poecilia forms P. latipinna that were fully scored. Two of these is presented in Figure I. characterized the Florida sample, while the third was Although we had great difficulty purifying mtDNA confined to the collection sites in Texas. The Texas from other P. formosa, we provisionally typed and Florida genotypes differed in at least four an ad­ digestion ditional 18 individuals from the Mexican locales. In a profiles, with mean estimated sequence divergence p single scorable digest (HincH), all of these specimens = 0.012. In P. mexicana, four mtDNA haplotypes were exhibited the diagnostic, high-molecularweight distinguished among the 13 mtDNA specimens from two sur­ bands characteristic of P. mexicana. Thus, veyed locales in Mexico. All differences they too among these probably had a P. mexicana (rather than P. latipinna) P. mexicana genotypes were minor, attributable to one matriarchal ancestry, although we cannotspeculate or two restriction site changes. An fur­ additional 40 spec­ ther as to their clonal diversity or age. imens ofthis species from these same Mexican locales were scored for only one to six enzyme patterns, but in all cases they too exhibited the "P. mexicana" DISCUSSION mtDNA profile. As recently as 1978, in referring to the hybrid-de­ In comparison to the moderate mtDNA differences rived parthenogenetic grasshopper Warramaba virgo, observed within either species, the distinction between the late M. J. D. White lamented that "we are never the P. latipinna and P. mexicana was dramatic: diges­ likely to know which species was the female tion patterns parent." for all endonucleases were distinct, and Within a year, the female parent mean mtDNA oftwo parthenogenetic sequence divergence was p = 0.070. lizards (Cnemidophorus tesselatus After and C. neomexican­ correction for within-species polymorphism (Nei, us) was determined unambiguously 1987), the net using recently in­ nucleotide divergence between P. lati­ troduced mtDNA methods (Brown pinna and P. mexicana and Wright, 1979). remainedp = 0.066. IfmtDNA Since then, studies utilizing the female-transmitted in Poecilia evolves at a rate considered "conventional" mtDNA molecule have helped resolve questions con- 1532 NOTES AND COMMENTS cerning the maternal origins and ages of several uni­ sexual or clonally reproducing vertebrates, including additional Cnemidophorus and Heteronotia lizards, Ambystoma salamanders, and Menidia, Poeciliopsis, andPhoxinus fishes [see Dawley and Bogart (1989) and references therein, and also several papers in Evolution 43(5) (1989)]. An emerging generalization from these studies is that most vertebrate parthenoforms are uni­ directional in maternal source. Furthermore, most uni­ 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 sexuals exhibit only a limited subset of the mtDNA Sequence divergence (%) diversity present in their female sexual progenitors (an exception involves Poeciliopsis-Quattro et al., 1991), FIG. I. UPGMA phenogram sumrnanzmg dis­ suggesting relatively recent origins through a small tances among the Poecilia mtDNA clones. number of successful hybridization events. Ironically, the first unisexual vertebrate discovered have found mtDNA haplotype heterogeneity within P. (Hubbs and Hubbs, 1932) has not previously been the formosa to be less than that within P. mexicana (and subject of mtDNA analysis. Here we have employed P. latipinna). mtDNA assays to determine that the hybridization(s) This study contributes to the growing catalogue of producing the gynogenetic fish Poecilia formosa in­ unisexual vertebrates for which the bisexual female volved P. mexicana as the female parent. Furthermore, ancestor has now been determined. It also contributes the close genetic similarity of the P. formosa mtDNA to the emerging view that most unisexual vertebrate genome to the most common mtDNA haplotype ob­ species are evolutionarily young, and in terms of ma­ served in P. mexicana indicates that the assayed uni­ triarchal phylogeny are embedded within the broader sexuals arose recently through one or a few hybridiza­ matriarchal diversity of their female sexual progeni­ tion events involving closely related females. tors. How recent may these hybridization(s) have been? Ourassays with 12 restriction endonucleases produced ACKNOWLEDGMENTS some 50 mtDNA fragments per individual. A change Work was supported by NSF grant 8805360 to JCA, in one such fragment translates into an estimate of grants BSR 84-15529 and 88-18001 to JT, and by sequence divergence ofabout 0.2%, or about 100,000 grants from Eckerd College and the Southern Regional years of lineage separation under the "conventional" Education Board to JCT. Logistical support for field mtDNA clock calibration mentioned above. Since no collections was provided by J. Reynolds. We thank the fragment changes distinguished the mtDNA in P. for­ Secretaria de Pesca, Mexico for permission to collect mosa from the common haplotype in P. mexicana, a fish, and B. J. Turner for help in identifying possible literal interpretation indicates a time of P. formosa triploids. origin less than 100,000 years ago. Of course, more intensive molecular assays and broader geographic sampling (particularly within the LITERATURE CITED southern range) might likely reveal additional mtDNA ABRAMOFF, P., R. M. DARNELL, AND J. S. BALSANO. haplotypes in Poecilia formosa, and perhaps lead to 1968. Electrophoretic demonstration ofthe hybrid more refined estimates ofclonal ages and the possibility origin of the gynogenetic teleost Poecilia formosa. of multiple origins. Genetic variation within P. for­ Am. Nat. 102:555-558. mosa clearly does exist, as judged by differences in AVISE, J. C., J. ARNOLD, R. M. BALL., JR., E. BER­ allozymes (Turner, 1982), nuclear DNA fingerprints MINGHAM, T. LAMB, J. E. NEIGEL, C. A. REEB, AND (Turner et al., 1990), ribosomal DNA (Monaco et al., N. C. SAUNDERS. 1987. Intraspecific phylogeog­ 1988), and histocompatibility response (Kallman, raphy: The mitochondrial DNA bridge between 1962). However, a difficulty in interpreting genetic population genetics and systematics. Annu. Rev. variation within P. formosa or any other parthenogen Ecol. Syst. 18:489-522. involves determining whether the differences accu­ BALSANO, J. S., R. M. DARNELL, AND P. ABRAMOFF. mulated within a monophyletic lineage after separation 1972. Electrophoretic evidence of triploidy asso­ from a perhaps distant bisexual progenitor, or alter­ ciated with populations of the gynogenetic teleost natively whether they reflect multiple hybridization Poeciliaformosa. Copeia 1972:292-297. events and retentions of polymorphisms from more BALSANo, J. S., E. M. RAsCH, AND P. J. MONACO. recent ancestors that mayor may not still exhibit these 1989. The evolutionary ecology of Poecilia for­ genotypes today (Turner, 1982). mosa and its triploid associate, pp. 277-297. In G. Most of the known allozyme alleles within P. for­ K. Meffe and F. F. Snelson, Jr. (eds.), Ecology and mosa have also been observed in P. mexicana or P. Evolution of Livebearing Fishes (Poeciliidae). latipinna (Turner, 1982); and in terms ofnuclear DNA Prentice Hall, Englewood Cliffs, NJ. fingerprints. the maximum genetic distance (propor­ BROWN, w. M. 1980. Polymorphism in mitochon­ tional band dissimilarity) within P. formosa appears drial DNA of humans as revealed by restriction to be much less than the average distance within bi­ endonuclease analysis. Proc. Nat. Acad. Sci. 77: sexual fish species (Turner et al., 1990). Such obser­ 3605-3609. vations led these authors to conclude that unisexual P. BROWN, W. M., M. GEORGE, JR., AND A. C. WILSON. formosa clones "are ultimately descended from hy­ 1979. Rapid evolution of animal mitochondrial bridization events involving the same or very closely DNA. Proc. Nat. Acad. Sci. 76:1967-1971. related individuals" (Turner et al., 1989). Similarly, we BROWN, W. M., AND J. W. WRIGHT. 1979. Mito- NOTES AND COMMENTS 1533

chondrial DNA analyses and the origin and relative SHIELDS, G. F., AND A. C. WILSON. 1987. Calibration age of parthenogenetic lizards (genus Cnemidoph­ of mitochondrial DNA evolution in geese. J. Mol. orus). Science 203:1247-1249. Evol. 24:212-217. DARNELL, R. M., AND P. ABRAMOFF. 1968. Distri­ SIMANEK, D. E. 1978. Population genetics and evo­ bution ofthe gynogenetic fish Poeciliaformosa, with lution in the Poeciliaformosa complex (Pisces: Poe­ remarks on the evolution of the species. Copeia ciliidae). Ph.D. Diss., Yale University, New Haven, 1968:354-361. CN. DAWLEY, R. M., AND J. P. BOGART. 1989. Evolution SNEATH, P. H. A., AND R. R. SOKAL. 1973. Numerical and Ecology of Unisexual Vertebrates. New York Taxonomy. W. H. Freeman, San Francisco, CA. State Museum, Albany, NY. TAMURA, K., ANDT.AOTSUKA. 1988. Rapidisolation HUBBS, C. L., AND L. C. HUBBS. 1932. Apparent par­ method of animal mitochondrial DNA by the al­ thenogenesis in nature, in a form of fish of hybrid kaline lysis procedure. Biochem. Genet. 26:815­ origin. Science 76:628-630. 819. KALLMAN, K. D. 1962. Population genetics ofa gyno­ TURNER, B. J. 1982. The evolutionary genetics of a genetic teleost, Mollienesiaformosa (Girard). Evo­ unisexual fish, Poecilia formosa, pp. 265-305. In lution 16:497-504. C. Barigozzi (ed.), Mechanisms ofSpeciation. Alan LANSMAN, R. A., R. O. SHADE, J. F. SHAPIRA, AND J. R. Liss, N.Y. C. AVISE. 1981. The use of restriction endonu­ TURNER, B. J., B. H. BRETT, AND R. R. MILLER. 1980a. cleases to measure mitochondrial DNA sequence Interspecific hybridization and the evolutionary or­ relatedness in natural populations. III. Techniques igin of a gynogenetic fish, Poecilia formosa. Evo­ and potential applications. J. Mol. Evol. 17:214­ lution 34:917-922. 226. TURNER, B. J., B. H. BRETT, E. M. RAsCH, AND J. S. MANIATIS, T., E. F. FRITSCH, AND J. SAMBROOK. 1982. BALSANO. 1980b. Evolutionary genetics ofa gyno­ Molecular Cloning. Cold Spring Harbor Lab, NY. genetic fish, Poeciliaformosa, the Amazon molly. MEYER, H. 1938. Investigations concerning the re­ Evolution 34:246-258. productive behavior of Mollienesia formosa. J. TURNER, B. J., J. F. ELDER, JR., AND T. F. LAUGHLIN. Genet. 36:329-366. 1989. Clonal diversity and evolutionary origin of MONACO, P. J., E. M. RASCH, ANDJ.S. BALSANO. 1984. a unisexual fish: Assessment by DNA "fingerprint­ Apomictic reproduction in the Amazon molly, Poe­ ing." Genetics 122:s31. ciliaformosa, and its triploid hybrids, pp. 311-328. TURNER, B. J., J. F. ELDER, JR., T. F. LAUGHLIN, AND In B. J. Turner (ed.), Evolutionary Genetics ofFish­ W. P. DAVIS. 1990. Genetic variation in clonal es. Plenum Press, N.Y. vertebrates detected by simple-sequence DNA fin­ MONACO, P. J., E. M. RAsCH, AND P. R. MUSICH. 1988. gerprinting. Proc. Nat. Acad. Sci. 87:5653-5657. Polymorphisms in ribosomal DNA ofa unisexual TURNER, B. J., AND H. R. STEEVES, III. 1989. Induc­ fish. Copeia 1988:774-777. tion ofspermatogenesis in an all-female fish species NEI, M. 1987. Molecular Evolutionary Genetics. Co­ by treatment with an exogenous androgen, pp. 113­ lumbia Univ. Press, N.Y. 122. In R. M. Dawley and J. P. Bogart (eds.), Evo­ NEI, M., AND W.-H. LI. 1979. Mathematical model lution and Ecology of Unisexual Vertebrates. New for studying genetic variation in terms ofrestriction York State Museum, Albany, NY. endonucleases. Proc. Nat. Acad. Sci. 76:5269-5273. WETHERINGTON, J. D., R. A. SCHENCK, AND R. C. VRI­ QUATTRO, J. M., J. C. AVISE, AND R. C. VRUENHOEK. JENHOEK. 1989. The origins and ecological success 1991. Molecular evidence for multiple origins of of unisexual Poeciliopsis: The frozen-niche varia­ hybridogenetic fish clones (Poeciliidae: Poeciliop­ tion model, pp, 259-275. In G. K. Meffe and F. F. sis). Genetics 127:391-398. Snelson, Jr. (eds.), Ecology and Evolution of Live­ RAsCH, E. M., AND J. S. BALSANO. 1989. Trihybrids bearing Fishes (Poeciliidae). Prentice Hall, Engle­ related to the unisexual molly fish, Poecilia for­ wood Cliffs, NJ. mosa, pp. 252-267. In R. M. Dawley and J. P. WHITE, M. J. D. 1978. Cytogenetics of the parthe­ Bogart (eds.), Evolution and Ecology of Unisexual nogenetic grasshopper Warramaba (formerly Mo­ Vertebrates. New York State Museum, Albany, NY. raba) virgo and its bisexual relatives III. Meiosis of RAsCH, E. M., P. J. MONACO, AND J. S. BALSANO. 1982. male "synthetic virgo" individuals. Chromosoma Cytophotometric and auto radiographic evidence for 67:55-61. functional apomixis in a gynogenetic fish, Poecilia WILSON, A. c., R. L. CANN, S. M. CARR, M. GEORGE, formosa, and its related triploid hybrids. Histo­ JR., U. B. GYLLENSTEN, K. M. HELM-BYCHOWSKJ, chemistry 73:515-533. R. G. HIGUCHI, S. R. PALUMBI, E. M. PRAGER, R. SCHULTZ, R. J. 1973. Unisexual fish: Laboratory syn­ D. SAGE, AND M. STONEKING. 1985. Mitochon­ thesis ofa "species." Science 179:180-181. drial DNA and two perspectives on evolutionary ---. 1989. Origins and relationships ofunisexual genetics. BioI. J. Linn. Soc. 26:375-400. poeciliids, pp. 69-87. In G. K. Meffe and F. F. Snelson, Jr. (eds.), Ecology and Evolution of Live­ Corresponding Editor: J. J. Bull bearing Fishes (Poeciliidae). Prentice Hall, Engle­ wood Cliffs, NJ.