Reproductive strategies and genetic variability in tropical freshwater fish

We estimated the genetic variability of nine fish species from the Brazilian upper Paraná River floodplain (Astyanax altiparanae, Hoplias malabaricus, Leporinus lacustris, Loricariichthys platymetopon, Parauchenipterus galeatus, Pimelodus maculatus, Rhaphiodon vulpinus, Roeboides paranensis and Serrasalmus marginatus) based on data for 36 putative allozyme loci obtained using corn starch gel electrophoresis of 13 enzymatic systems: aspartate aminotransferase (EC 2.6.1.1), acid phosphatase (EC 3.1.3.2), esterase (EC 3.1.1.1), glycerol-3-phosphate dehydrogenase (EC 1.1.1.8), glucose-6-phosphate dehydrogenase (EC 1.1.1.49), glucose-6-phosphate isomerase (EC 5.3.1.9), Iditol dehydrogenase (EC 1.1.1.14), isocitrate dehydrogenase – NADP (EC 1.1.1.42), L-lactate dehydrogenase (EC 1.1.1.27), malate dehydrogenase (EC 1.1.1.37), malate dehydrogenase-NADP (EC 1.1.1.40), phosphoglucomutase (EC 5.4.2.2) and superoxide dismutase, (EC 1.15.1.1). The mean expected heterozygosity varied from zero to 0.147. When data from the literature for 75 species of tropical fish were added to the nine species of this study, the heterozygosity values differed significantly among the groups of different reproductive strategies. The highest mean heterozygosity was for the non-migratory without parental care, followed by the long-distance migratory, and the lowest mean was for the non-migratory with parental care or internal fecundation.


Introduction
The reactions that species offer to environmental selective pressures is reflected in their physiological strategies, which can reduce maintenance energy, increase the efficiency of energy acquisition and, consequently, optimize reproductive efficiency.The success of such strategies may be estimated by the individual ability of being genetically present in the offspring (Agostinho and Júlio Jr., 1999).
In fish, reproductive strategies are extremely diversified, encompassing variations in the way they meet and attract partners, in spawning locations and parental care (Wootton, 1990).External fertilization and embryonic development are the dominant pattern for the majority of teleost fish, although internal fertilization and external embryonic development occurs in some Brazilian freshwater species such as Parauchenipterus galeatus.Blumer (1982) described parental care as any type of investment that enhances the survival probability of offspring (reproductive success) and which vary from pre-fertilization ac-tivities such as nest building (e.g. in Hoplias malabaricus) to mouth brooding and live-bearing.
Adaptation to heterogeneous environments depends upon the genetic variability of species, reduction in genetic variability resulting in reduced feeding and reproductive performance leading to inefficient exploration of the habitats (Kirpichnikov, 1992).Several authors have tried to explain the mechanisms which maintain genetic variability in natural populations.Nevo (1988) analyzed the relationship between heterozygosity and ecological and biological features such as climate (arctic, temperate and tropical), geographic range (wide, narrow, endemic), habitat type (underground, terrestrial, aquatic), habitat range (specialist, generalist) aridity (arid, sub-arid, sub-humid, mesic with a moderate or well-balanced moisture supply) and territoriality (territorial, non-territorial) in 1,111 species (vertebrates, invertebrates, and plants) and concluded that environmental heterogeneity is the major factor in maintaining and structuring genetic diversity in natural populations.Ward et al. (1994) compared the heterozygosity of marine, freshwater and anadromous (fish which live mostly in the sea but breed in fresh water) fish and concluded that marine fish have greater genetic variability than the other types.Heithaus and Laushman (1997) investigated the effects of ecology, life history and water quality on genetic variation of three stream-dwelling fish species and revealed that the genetic variability tends to decrease as species become more ecologically specialized.
Allozyme electrophoresis has been extensively used to estimate genetic variability in natural populations (Nevo, 1988;Van Der Bank et al., 1989;Ward et al., 1994) and our study employed this technique to estimate the genetic variability of nine species of fish from the Brazilian upper Paraná River floodplain and, jointly with literature data, verify the relationships between genetic variability and reproductive strategies.Immediately after capture, white skeletal muscle, gill, heart and liver tissue were removed from each specimen and frozen in liquid nitrogen.Tissues were homogenized with plastic sticks in 1.5 mL microcentrifuge tubes containing Tris/0.02 M HCl (1:1 w:v) buffer, pH 7.5.Carbon tetrachloride (CCl 4 ) was added to the homogenized liver samples (1:2 v:v) due to the large amounts of fat present in the tissues (Pasteur et al., 1988).The homogenized samples were centrifuged at 45,114 x g for 30 min at temperatures between 1°a nd 5 °C and the supernatants submitted to horizontal electrophoresis in 15% corn starch gel (Val et al., 1981).
The species were organized in three distinct groups of reproductive strategy according to Vazzoler and Menezes (1992): group 1 (G1), containing the long-distance (> 100 km) migratory species P. maculatus and R. vulpinus; group 2 (G2), containing the sedentary or short-distance (< 100 km) migratory species A. altiparanae, L. lacustris, R. paranensis which provide no parental care; and group 3, containing the sedentary or short-distance migratory species H. malabaricus, L. platymetopon, P. galeatus, S. marginatus with internal fertilization or which provide parental care.Since it is unlikely that the species analyzed by us were representative of all tropical fish, we added data from the literature for a further 75 tropical fish species (61 from Brazil and 14 from Africa) for which the heterozygosity values had been estimated using allozyme data (Table 2), hence the total number of species analyzed was 84.
The significance of differences between groups was evaluated using a null models analysis of variance and the EcoSim 7 program (Gotelli & Entsminger, 2006).

Genetic variability
The calculated genetic variability values for each species are presented in Table 3, from which it can be seen that the greatest frequency of polymorphic loci, average number of alleles per locus and heterozygosity were obtained for A. altiparanae, followed by H. malabaricus.No polymorphic locus was detected in R. vulpinus for the 29 loci analyzed.Except for P. maculatus the expected heterozygosity were higher than obtained heterozygosity, which indicates, on average, an excess of homozygotes for all loci.

Heterozygosity and reproductive strategies
The lowest heterozygosity values in our sample of nine species were for the G1 long-distance migratory species (R. vulpinus = 0 and P. maculatus = 0.011), which therefore also presented the lowest group average (He = 0.005).The highest average heterozygosity (He = 0.110) occurred in the G2 sedentary or short-distance migratory species without parental care, with A. altiparane having the highest heterozygosity (0.152) of all the species analyzed.The G3 sedentary or short-distance migratory species with internal fertilization or parental care exhibited intermediate heterozygosity (0.067).
When we considered the nine tropical fish species analyzed by us plus the 75 species from the other studies we found that the number of species in each reproductive strategy group and the average heterozygosity (He) ± the standard error for each group was as follows: G1 = 7 (Ho = 0.064 ± 0.021), G2 = 8 (He = 0.081 ± 0.016) and G3 = 69 (He = 0.046 ± 0.002) (Table 3, Figure 2), with He varying from zero to 0.142 for G1, from 0.006 to 0.152 for G2, and from zero to 0.143 for G3 (Tables 2 and 3).These data show that the highest average heterozygosity was presented by the G2 sedentary or short-distance migratory species without parental care, the G1 long-distance migratory species had and intermediate average heterozygosity and the G3 species with parental care or internal fertilization presented the lowest average heterozygosity.
A Shapiro-Wilk test for normality showed that the He values were not normally distributed (W = 0.9199, p < 0.0001) and the Levene test for homogeneity of variances showed a significant value (F = 4.843, p = 0.0103), indicating that the variances were heterogeneous.We conducted ANOVA but since the assumptions of normality and homogeneity of variances were not met by the data the significance of the ANOVA was tested by null models using the EcoSim 7 programs.The ANOVA detected a significant difference (F 2, 81 = 4.02; p < 0.05) between the three reproductive strategy groups.

692
Reproductive strategies in fish  Reproductive strategies in fish

Discussion
The genetic variability estimated in nine fish species from the upper Paraná River floodplain using Nei's gene diversity (He) varied from zero to 0.1518 with an average of 0.066, very near to the average of 0.051 described for 195 piscine species from several world-wide localities reviewed by Ward et al. (1992).
It has been generally accepted that populations of long-distance migratory species have higher genetic variability than non-migratory species because of their higher gene flow.Although this is true for several taxa (Ward et al., 1992), our data are not in agreement with this hypothesis.In our study, no polymorphic locus was found in a sample of 29 analyzed loci of R. vulpinus, a long-distance migratory species.This is surprising, since other cases of absence of genetic variability have been found in two sedentary species, Neoplecostomus paranensis from the upper Paraná River (Zawadzki et al., 2004b) and Hypostomus sp. 2 from the third-order Paraná River tributary Ribeirão Maringá (Paiva et al., 2005).Furthermore, P. maculatus, the other long-distance migratory species analyzed in our study, showed low genetic variability (He = 0.011), con-trasting with the value of He = 0.132 estimated for other migratory species such as Prochilodus lineatus (Revaldaves et al., 1997) and Leporinus friderici (Chiari and Sodré, 1999).
In the nine species analyzed by us the average heterozygosity of species with different reproductive strategies was not statistically different when compared amongst each other, which could lead to the conclusion that reproductive strategies are not related to genetic variability.However, our results could have been biased because the group of migratory species analyzed by us had only two species, so to compensate for this we carried out an analysis of our set of nine species combined with data on 75 tropical fish species from the literature.When the data for the 84  tropical fish species were analyzed, the relationship between heterozygosity and reproductive strategies were changed and there was a significant difference between the reproductive strategies groups (Figure 2).With this analysis, the average heterozygosity values which differed most were those between the G3 parental care group with and the G2 group without parental care.The heterogeneity of the heterozygosity variances between groups may have been due to the number of species analyzed, since the minimum and maximum heterozygosity values were similar among the groups.
The number of surviving offspring is likely to be greater in species with parental care than in species without parental care (Vazzoler, 1996), and species exhibiting parental care tend to be less variable.This can be observed in L. platymetopon, which brood their fry inside the male's mouth and is the most abundant species of the Paraná River floodplain.In species without parental care, few offspring are likely to survive, and a greater genetic variability is important to face environmental challenges.Nei (1987) pointed out three main factors related to heterozygosity level are the quaternary structure of proteins, the molecular weight of the protein subunit and the species population size.In addition to these factors, we found that parental care is another factor that should be taken into account to maintain gene diversity.However, since a large number of factors influence the amount of genetic variability, we believe that further studies will be necessary to elucidate the relationship between genetic variability and ecological features of fish species.

Figure 1 -
Figure 1 -Paraná River floodplain.Sample localities are indicated by a black circle.

Figure 2 -
Figure2-Mean heterozygosity values for each reproductive strategy group of tropical fish species (Brazil and Africa).Group 1 (G1) = migratory species; Group 2 (G2) = short-distance migratory or sedentary species without parental care; and Group 3 (G3) = short-distance migratory or sedentary species with parental care or internal fertilization.

Table 2 -
Heterozygosity (He) values estimated for 75 tropical fish species from Brazil and Africa based on enzymatic data.Group 1 = Migratory species, Group 2 = sedentary species or short-distance migratory without parental care; Group 3 = sedentary species or short-distance migratory with internal fertilization or parental care.Africa = Af, Brazil = Br.