Chromosomes of Cichlasoma istlanum (Perdformes: Cichlidae) and karyotype comparison oí two presumed subspecies

Abstrad: Cichlasoma istlanum (lordan & Snyder, 1900) is a freshwater cichlid from the Balsas river province in the Pacific Basin. Two subspecies: C. istlana istlana, from lhe Ixtla river in ¡he state of Morelos, Mexico, and C. istlana fusca, from the Huámito river near the town of La Huacana, in the state of Michoacán, were named half a century ago on the basis of merisüe characters. In this work, the karyotype of the species was eslablished by conventional and G-banding cytogenetic procedures and a comparative analysis of karyotypes from the two populations, previously proposed as subspecies, was performed. Ten females were collected in the Amacuzac river, and nine specimens (two females), were eollected in the Huámito river. Based on the eount of 264 mitot­ ie fields of the former and 203 of the latter, the modal number of 2n=48 was established in every sample and considered as ¡he diploid number of both populations. The karyotype analysis was based on ten karyotypes pre­ pared from Morelos and eight from Michoacán, which included ¡hree from females and five from males. The chromosome formula thus established was of 8sm+40stt. The G-banding pattem was similar in both populations and the comparison of the mean Jengths of lhe chromosome pairs did not reveal statistically significant differ­ ences between both populations. The presence of a practically identical karyotype does not support the subspe­ cific division. The morphometric analyses made by other authors, which detected overlappíng in the characters that were proposed as distinctive of the two subspecies, agree with the results of this study: not enough diver­ gence has been found lo substantiate ¡he $ubdivision of C. istlanum. The lack of population divergence might have been brought about by an abated geographic isolation caused by gene flow among contiguous populations along their continuous distribution in the Balsas Basin regions, by the relatively small divergence time since their distribution in these regions, andlor, less probably, by a recent historie al replacement of one popuiation by the other. The absence of karyotype differences might also be attributed lo characteristics inherent to the genome organization in the genus Cichlasoma still to be identified and understood.

The family Cichlidae is represented in Mexico by the two autochthonous genera: Petenia Günther 1862 and Cichlasoma Swainson 1839, and by the introduced Tilapia Smith 1840 and Oreochromis Günther 1889.
The genus Cichlasoma is found only in the New World where it comprises approximately 49 species, whereas the genus Petenia incJudes only one (Nelson 1994).These fishes have been the subject of severa!regional studies Cichlasoma istlanum (Jordan & Snyder, 1900) is a freshwater cichlid from the Balsas River province in the Pacific Basin.It has been collected from the high river sources to the mouth river.Its distribution area ineludes also the small basins of the Armería river in the sta tes of Colima and Jalisco, and sorne other small basins of the state of Guerrero, such as those of the Nexapa, Papagayo, and Coyuca ri vers (De Buen 1946, Miller 1966, Alvarez del Villar 1970, 1972a, b, 1977).Recent reports have shown its drastic displacement by the exotic cichlid C. nigrofasciatum (Contreras McBeath 1997).
De Buen (1946) based on differences of the fin-rays formulas and on the number of spinesdivided the species C. istlanum into two subspecies: C. istlana istlana, from the Ixtla river in the state of Morelos, and C. istlana fusca, from the Huárnito river near the town of La Huacana, in the state of Michoacán.
The karyotype might characterize a species through the determination of the num ber, size and structure of its chromosomes.Although in vertebrates it constitutes a partic ularly stable feature because it does not seem to be subject to direct selective pressures, chromosome variability has been• shown in sorne species through the detection of poly morphisms in the chromosome diplóid num ber, like in Noturus albater Taylor 1969 (LeGrande 1981), Gymnotus carapo L. 1758 (Foresti et al. 1984), Gobius niger L. 1758 (Vítturi & Catalano 1989), Gobius paganellus L. 1758 (Amores et al. 1990) and in the cich lid C. friedrichsthali (Heckel, 1840) (Salas & Boza 1991) or by the presence of supemumer ary chromosomes, like in Prochilodus scrofa Steindachner 188 1 (Pauls & Bertollo 1983) and in the cichlid Gymnogeophagus balzanii (Perugia 1891) (Feldberg & Bertollo 1984).These antecedents justify the search of chro mosome differences that rnight be coincident with the spatial separation of the subspecies proposed by De Buen (1946).
The purpose of this study was to establish the karyotype and the G-banding pattem of C. istlanum, obtained from organisms of the same regions where De Buen (1946) collected the samples that led him to establish the subspe cific division of the species, and to compare the chromosome diploid number, the morphol ogy, relative size and banding pattems of the chromosomes from both samples, in order to provide arguments to verify or deny the subdi vision of the species.

MATERIALS AND METHODS
Ten females were collected in the Amacuzac river, near the town of La Fundición, in the state of Morelos, Mexico, (18.52°N and 99.22° W) and nine organisms: two females and seven males in the Huámito river near the town of La Huacana, in the state of Michoacán (18.95°N and 101.87°W).These sample sizes fa11 within the range used in several cytogenetic studies, due to the kary otype stability generally found.
The cytogenetic methods fo11owed to ob tain the mitotic fields have been described by Uribe-Alcocer et al. (1983) and Maldonado Monroy et al. (1985).Briefly, they consist of a pretreatment with calcium chloride and colchi cine, gentle scraping of the gill epithelium fo-110wed by a hypotonic shock, fixation in a met hanol-acetic acid solution, air slide drying and Giemsa staining.The method of Levan et al. (1964), based on the centromere position, as sessed in this work through the determination of the centromeric index (short arm length (p)/total chromosome length (p+q», was fo11o wed to elassify chromosomes.The G-banding patterns were obtained by incubation of the slides in 2xSSC at 60" C for 2 hr, followed with digestion with a mixture of trypsin 0.03% and Giemsa stain 3.3% in buffer phosphate pH 7.2 (Ruiz-Carús, 1995).

RESULTS
Chromosomes of 467 mitotic fields, derived from both populations, were counted to identify the modal number, 48, which was considered as the diploid number of the species.Numbers under the mode were con sidered as chromosome losses during the cyto genetic technique rather than aneuploidies, due to the decreasing frequencies of chromo some numbers below the modal diploid num-   .a,.

••
• 1 ber in all specimens studied and to the lack of high frequencies indicative of a different chro mosome number in any particular specimen.
Eighteen mitotic fields were studied to determine the karyotype of the species.In both samples it was composed by four submeta centrics pairs (sm), and 20 subtelo-acrocentric (stt) chromosome pairs.The chromosome for mula derived from both samples was 8sm+40stt.The fundamental number found (number of chromosome arms: two for sub metacentrics and one for subtelocentri c and acrocentric chromosomes) was 56.Figs. 1 and  2 show the chromosome groups as well as their G-banding pattern.A digitally composed karyotype, which shows an element of each population for every chromosome pair, is depicted in Fig. 3.  Mean relative length of the chromosome pairs corresponding to both samples and their standard deviations are shown in Table l.They were derived from the measurements of ten karyotypes from the sample of the state of Morelos and eight from the state of Michoacán.No dependence among the chromosome diploid number counts and the specimens analyzed nor the collecting sites was found (Contingency table p>O.05).Data corresponding to chromo sorne and arrn lengths were assayed for good ness of fit to a normal distribution and no statis tical differences were found (Shapiro Wilk W p>O.05).Similarly no differences were found between the natural logarithms of the standard ized whole length chromosome (p+q) of every pair, nor between the short (p) and long arms (q) of the four bíarrn ed chromosome pairs of both populations (Two-way ANOVA p>O.05).
Consistently with the lack of reports on sex chromosomes in all other cichlid species studied, heteromorphic chromosome ele ments, indicative of differentiated sex-chro mosomes, were not detected in the sample of the Huámíto river, the only site in which both sexes were represented.

DISCUSSION
The karyotypes of the C. istlanum popula tions of the rivers Amacuzac (Morelos) and Huámito (Michoacán) have a common corn plernent of 2n = 48.The assurnption that no chrornosornal change has occurred is support ed by the coincidence in the nurnber of sub rnetacentric and subtelo-acrocentric chromo sornes and in the centrornere position of the subrnetacentrics, by the lack of significant dif ferences in chrornosorne lengths, as well as by an oyeran sirnilarity of their G-banding pat terns.Alternatively, it could also be considered that actual rnodifications could be too small to be detected by the methodology followed in this work.The assumption of populational chromosome divergence between the popula tions would be based on frail evidence.G-band studies in fishes have been limited (Hartley & Thorne 1985, Gold et al. 1990) because, for reasons probably related to intrin sic chromosome structure or to their size, do not show the abundance nor the resolution found in other animal groups, like mammals, particularly when the chromosome material has been obtained through direct techniques, but for a few exceptional cases (e.g.Sola el al. 1984).In this study, the agreement in the G banding patterns of both populations supports the lack of chromosome diversification and, hence, of population divergence.
The morphometric analyses performed by Bonilla-Ruz (1982) and by Sánchez-Salazar (1984) detected overlapping in the characters that, according to De Buen (1946), were dis tinctive of the two proposed subspecies.These analyses agree with the results of this study in that not enough differences were found to sub stantiate the subdivision of C. istlanum.The absence of population diversification, as esti mated through chromosome and morphometric analyses, might have been brought about by an abated geographic isolation due to gene flow among contiguous populations along their con tinuous distribution in the Balsas Basin regions, by the relatively small divergence time sínce their distribution in these regions and/or, less probably, by a recent historical replacement of one population by the other, event of which there is no evidence.
The diploid chromosome number of C. istlanum (2n = 48) agrees with that of practi cally all the Cíchlasoma species studied.According to the c1assification proposed by Thompson (1979), based on the number of biarmed chromosomes, the karyotype of C. ist lanum is type "A" sínce ít has less than five meta-submetacentric chromosomes, differing from the karyotypes "B" that have five or more meta-submetacentric chromosome pairs.The common occurrence of karyotype "A" in some species is no evidence of a close phylogenetic relationship among them because, although they share this karyotype configuration, they belong to different species groups of the genus Cichlasoma (Regan 1905, Miller & Nelson 1961, Miller, 1966).If the classification method of Levan et al. (1964) is followed, the karyotype of C. istlanum is coincident with those of C. dovii (Günther, 1864) (Thompson, 1979, Salas & Boza, 1991) Furthermore, these species not only belong to different species groups, but some of them have also been considered to be long to the South American groups while others have been considered to have Central American affinities.
Kornfield et al. (1979), Thompson (1979), Vervoort (1980) and Uribe-Alcocer & Arre guín-Espinosa (1989) have pointed out that so me cichlid species, close or distant phylogene tically, do not show large karyotype differen ces.Existent variability might be due to the ac quisition or loss of small chromosome frag ments, to non-reciprocal translocations, to pe ricentric inversions, and to tandem duplica tions or differential crossing-overs (Uribe-Al cocer el al. 1992).It is not likely that these rea rrangements have been relevant to their evolu tion, sínce changes in the structure or size of chromosomes have not modified the diploid nor the fundamental chromosome numbers of their populations and, for this reason, their re productive isolation has been probably brought about by other means.Notwithstanding some repOlts of chromosome variability in cichlid species (Feldberg & Bertollo 1984, Salas & Boza 1991), which justified the search under taken in this study, the karyotype in this genus seems to be a conservative feature of low mo dificatíon rate that apparently accompanies the evolution of the species, but do es not bring about their reproductive isolation, nor the evo lution of other biological characters.The wi despread absence of significant karyotype dif ferences in the genus Cichlasoma might also be attributed to characteristícs inherent to the genome organization still to be identified and understood.
Genetic studies of the neotropical cichlids by methodologies that allow the detection of existing variations, for instance in their allozymes, their DNA (Sültmann el al. 1995), or in their microsatellite markers (Lee & Kocher 1996), are considered necessary to establish, confinn or modify our knowledge of the phylletic relationshíps of this important group of neotropical cichlids.

TABLE 1
Mean relative lengths of the chromosome pairs of the karyotypes of e. istlanumfrom the Amacuzac river (More/os) and the Huámito river (Michoacán)