Esterase-D and chromosome patterns in Central Amazon piranha ( Serrasalmus rhombeus Linnaeus , 1766 ) from Lake Catalão

This study presents additional genetic data on piranha (Serrasalmus rhombeus Linnaeus, 1766) complex previously diagnosed due to the presence of distinct cytotypes 2n = 58 and 2n = 60. Three esterase-D enzyme loci (Est-D1, Est-D2 and Est-D3) were examined and complemented with chromosomal data from 66 piranha specimens collected from Lake Catalão. For all specimens the Est-D1 and Est-D2 loci were monomorphic. In contrast, the Est-D3 locus was polymorphic with genotypes and alleles being differentially distributed in the previously described cytotypes and served as the basis for detecting a new cytotype (2n = 60 B). In cytotype 2n = 58 the Est-D3 locus was also polymorphic and presented Mendelian allelic segregation with four genotypes (Est-D3, Est-D3, Est-D3 and Est-D3) out of six theoretically possible genotypes, presumably encoded by alleles Est-D3 (frequency = 0.237), EsT-D3 (0.710) and Est-D3 (0.053). A Chi-squared ( ) test for Hardy-Weinberg equilibrium was applied to the Est-D3 locus and revealed a genetic unbalance in cytotype 2n = 58, indicating the probable existence in the surveyed area of different stocks for that karyotypic structure. A silent null allele (Est-D3) with a high frequency (0.959) occurred exclusively in the 2n = 60 cytotype. On the other hand, the new cytotype 2n = 60 B described here for the first time was monomorphic for the presumably fixed Est-D3 allele. The data as a whole should contribute to the better understanding the rhombeus complex taxonomic status definition in the Central Amazon.

Electrophoretic investigations of genetic markers such as proteins and enzymes, especially allozymes and isoenzymes, have been decisive in determining the taxonomic and population status of many organisms (Ferguson, 1980), with allozymes having been particularly useful for identifying fish species and their hybrids in natural and artificial populations (Ferguson et al., 1995) and have been especially useful for identifying cryptic species (Allendorf and Utter, 1979;Lavery and Shaklee, 1991;Musyl and Keenan, 1996).Many enzymes, such as esterases, show pronounced differentiation in isoenzymatic patterns in many organisms, including plants (Anti, 2000), phytonematoids (Alonso and Alfenas, 1998), mollusks (Richardson et al., 1982) and fish (Payne et al., 1972;Reinitz, 1977;Solomon and Child, 1978;Ferguson, 1980).
Although Serrasalmus is one of the most widespread and specious South American piranha genera (Machado Allison and Fink, 1995) only a few allozyme genetic stud-ies have been carried out on this piscine group since the study of Serrasalmus spilopleura lactate dehydrogenase (LDH), malate dehydrogenase (MDH) and glucose phosphate isomerase (GPI) isoenzyme patterns by Cestari (1996) on fish from the Paraná and Paraguai river basins.
Recent studies on Amazonian Serrasalmus species have revealed karyotypic divergence between and within populations of S. spilopleura and Serrasalmus rhombeus (Nakayama et al., 2000(Nakayama et al., , 2001(Nakayama et al., , 2002;;Centofante et al., 2002).Nakayama et al. (2001) has suggested that S. rhombeus cryptic species may exist based on the two cytotypes (2n = 58 and 2n = 60) found at Lake Catalão located near the confluence of the Negro and Solimões rivers in the Brazilian state of Amazonas.In addition to being identified by their karyotypes (Nakayama et al., op. cit.) fish belonging to the S. rhombeus complex are also moderately distinguishable by parasite analysis ( Van Every and Kritsky, 1992) but not by their 16S mitochondrial DNA (Ortí et al., 1996) but as yet there have been no isoenzyme studies on this complex.
The work described in the present paper used karyotype and esterase-D isoenzyme patterns to provide addi-tional genetic information on the S. rhombeus complex in order to complement studies on the taxonomic status of the Central Amazon rhombeus complex.
Kidney cell mitotic chromosomes were prepared and analyzed using the air-drying technique (Bertollo et al., 1978) and skeletal muscle protein extracts and starch-gel electrophoresis were used to detect the esterase-D Est-D1, Est-D2 and Est-D3 loci using standard gel and electrode electrophoretic buffers (Ridgway et al., 1970) and the staining procedure described by Hopkinson et al. (1973).
Hardy-Weinberg expectations were calculated using the Chi-square (c 2 ) test to verify the population gene balance for the 2n = 58 karyotype based on allelic segregation on the Est-D3 polymorphic locus.This locus was also used to estimate the 2n = 60 karyotype null (recessive) allele frequency using the square root of the null genotype frequency as calculated using the Tools for Population Genetic Analyses" (TFPGA) program developed by Miller (1997).
We found three electrophoretic activity S. rhombeus esterase-D zones, presumably coded for by the three loci Est-D1, Est-D2 and Est-D3 (Table 1).The Est-D1 and Est-D2 loci were monomorphic in all specimens and pre-sented genotypes Est-D1 11 and Est-D2 11 , presumably encoded by the fixed alleles Est-D1 1 and Est-D2 1 , while the Est-D3 locus had polymorphic genotype and allele distributions which were differentiated and highly congruent with the identified distinct cytotypes (Table 1, Figure 2).
We found that for the 2n = 58 (46M-SM+2ST+10A) cytotype (Nakayama et al., 2001) S. rhombeus specimens the Est-D3 locus presented Mendelian polymorphism and allelic segregation and showed four genotypes (Est-D3 11 , Est-D3 12 , Est-D3 22 and Est-D3 33 ) out of the six theoretically expected genotypes.The four genotypes were presumably encoded by the Est-D3 1 , Est-D3 2 and Est-D3 3 alleles (Table 1), of which Est-D3 2 was the most commonly observed (f = 0.710).On the other hand we found that in S. rhombeus specimens with the 2n = 60 (44M-SM+6ST+10A) cytotype (Nakayama et al. 2001) the Est-D3 locus presented a silent null allele (Est-D3 0 ) at the very high frequency of 0.959 (with no electrophoretic bands being detected exclusively for this cytotype) in addition to the low frequency Teixeira et al. 499  Geophagus brasiliensis (Cichlidae, Perciformes) and far higher as compared with the low frequencies (under 5%) of null alleles at allozyme loci in natural Drosophila melanogaster populations (Voelker et al., 1980;Langley et al., 1981).Despite the excess of Est-D3 22 homozygotes and a corresponding deficiency of predictable Est-D3 20 heterozygotes on the Est-D3 locus, the observed number of the Est-D3 00 null genotype for the 2n = 60 cytotype showed good agreement with the statistical expectation, discarding the possibility of the Est-D3 0 silent null allele being interpreted as a technical artifact (Table 1).Although there are some examples in the literature associating the presence of null alleles with possible mildly deleterious effects to its carriers (see Aquino-Silva et al.,1998), this kind of association involving the silent Est-D3 0 allele in the 2n = 60 cytotype could only be effectively tested for by crossing experiments with Est-D3 22 and Est-D3 00 homozygotes.
The new S. rhombeus cytotype reported here for the first time, 2n = 60 B (44M-SM+4ST+12A) revealed monomorphism for the presumably fixed allele Est-D3 3 , and was detected in all four S. rhombeus specimens examined, although this allele can only be definitively reported as fixed following the screening of the Est-D3 locus in a larger population of the S. rhombeus 2n = 60 cytotype.
The Chi-squared (c 2 ) test for Hardy-Weinberg equilibrium used to check the genetic balance in the S. rhombeus 2n = 58 population revealed a highly significant statistical difference (c 2 (3) = 44.443for p £ 0.001) due to an excess of homozygotes and a corresponding deficiency of heterozygotes at the Est-D3 locus (Table 1).Cestari (1996) also found highly significant departures from Hard-Weinberg equilibrium regarding the allele frequency distributions of two polymorphic glucose phosphate isomerase loci (GPI-A* and GPI-B*) examined in the S. spilopleura `a' cytotype, which appears to be an endemic cytotype of the Brazilian upper Paraná River basin, this genetic disequilibrium also being due to homozygote excess and heterozygote deficiency as was the case for the S. rhombeus Est-D3 locus studied by us.
A classical explanation for homozygote excess in population samples is the Wahlund effect caused by the mixture of genetically distinct populations (Wahlund, 1928).Our S. rhombeus cytotype 2n = 58 data indicates the probable existence of different stocks of this karyotypic structure within the Lake Catalão.Teixeira et al., (2002) have shown highly statistically significant departures from genetic equilibrium due to homozygote excess at the transferrin locus (Tf) in seven out of eight Central Amazon population samples of the piscine Plagioscion squamosissimus (pescada in Portuguese), including three out of the four Lake Catalão P. squamosissimus population samples collected which showed three genetically discreet subpopulations of this species.Lake Catalão is an ecotone formed by the mixture of acid and dark water from the Negro river clear waters from the Solimões river but may also be viewed as an area of mixing of genetically distinct fish populations since this area is widely known as a stopping place and passage corridor in the migratory route of several Central Amazon fish species.
Although esterase isoenzyme patterns are usually species-specific, especially in fish (Payne et al., 1972;Reinitz, 1977;Solomon and Child, 1978;Ferguson, 1980), our investigation showed no cytotype-specific fixed allele in the three esterase-D loci for the three piranha karyotypic structures examined.Generally, species are typically fixed for different alleles on the same locus, while co-specific populations typically differ in regard to the same allele frequencies (Smith et al., 1981).
Gradual frequency differences in the A*125 and B*210 alleles at two GPI loci detected in S. spilopleura caught between the upper Paraná River (cytotype `a') and the lower Paraná River (cytotype `b' and cytotype `c') led Cestari (1996) to suggest that there may be interbreeding between fish from these two sites, supporting the hypothesis of a hybrid origin for the `c' cytotype.However, our esterase-D and chromosome data do not support the existence of different S. rhombeus piranha species in Lake Catalão and there was no indication of hybridization among the S. rhombeus cytotypes examined.Thus once cytotypespecific fixed alleles are detected in any S. rhombeus isoenzyme patterns different taxonomic units with species status will have to be formally recognized.
Several cases have been described in the literature where genetic polymorphism seems to be shared between a pair of species while closely related species might be expected to show higher levels of shared polymorphism (see Clark, 1997).Nakayama et al. (2001) considered S. rhombeus to be a cryptic species, with imperceptible morphological differences among the three cytotypes examined by us and it follows that the occurrence of a Est-D3 locus polymorphism shared among these cytotypes would reasonably be expected to follow the same pattern as that seen for the 2n = 58 cytotype i.e. segregation of alleles following a Mendelian model which did not occur.Regarding our study, recent and ongoing differentiation of the distinct diploid number of piranhas might explain the very high frequency (95.90%) of the Est-D3 0 silent null allele only seen in cytotype 2n = 60, the high frequency (71%) of the Est-D3 2 allele only occurring in cytotype 2n = 58, the apparent fixation (despite the small number of specimens) of the Est-D3 3 allele only detected in the new 2n = 60 B cytotype and the absence of different cytotype-specific fixed alleles at the Esterase-D loci (Table 1, Figure 2).Additionally, our data may suggest that these Central Amazon piranhas karyotypic groups partially represent isolated populations, or populations which have been isolated for an insufficient period of time for the fixation of different cytotype-specific alleles.
A character applied for identifying taxonomic units with species status should occur in all members of the species and not in other species, i.e., be a unique fixed allele or its product.Consequently, various distinct genetic and molecular techniques such as chromosome, DNA and protein studies should be complemented with meristic-morphometric studies in order that the taxonomic status of Central Amazon rhombeus complex can be elucidated.

Figure 1 -
Figure 1 -Location of Lake Catalão in the Brazilian Central Amazon where specimens of the piranha Serrasalmus rhombeus were sampled.