Genetic variability of wild and captivity populations of Colossoma macropomum (Cuvier, 1818)

Tambaqui (Colossoma macropomum) is among the most important fish species of the Amazon and one of the most cultivated in Brazil. In the present work we have evaluated the genetic variability of wild and captivity populations of C. macropomum. Enzymatic markers were used to estimate the genetic variability of 41 specimens from a wild group; and 30, 33 and 45 from three captivity groups, which came from Pentecostes (Ceará State), Jaboticabal (São Paulo State) and Itacoatiara (Amazonas State), respectively. Nine isoenzymic systems were used to evaluate the genetic variability of these populations. Using zimogram data we obtained the polymorphism level, allele number, allelic frequency, observed and expected heterozigosity, Wright F statistics (FIS, FST), genetic distance, level of similarity and group analysis. The isoenzymic data showed that, from the nine systems, six presented polymorphic loci (Fbp-2, G6pdh-2, G6pdh-3, Pgi-1, Pgi-2 and Pgm-1). The populations from Pentecostes and Jaboticabal presented loss of genetic variability and low heterozigosity, compared to the wild population and to the artificial population acquired at Itacoatiara fish farm. Based on these results and on fish farmer information we could consider the population from Itacoatiara as recently derived from a wild population. Concluding, we suggest that the artificial populations of tambaqui, which contain animals originated from this funding population at Pentecostes, should be renewed with the introduction of a new group of individuals with genetic variability equivalent to the wild population.


Introduction
Fish culture is a branch of agribusiness and has been growing in Brazil.The reason for this phenomenon is the decrease of environmental stocks of the main species with commercial value.Within the natural species found in the Brazilian watershed, we can emphasize the tambaqui (Colossoma macropomum), a species that can reach 30 kg and has been cultivated in several Brazilian regions (VAL;HONCZARIK, 1995).
There are evidences that wild populations of tambaqui are depleted by the excessive fishing effort and not by their habitat reduction.The idea of creating fish artificially became stronger since the environmental stocks of these commercially valuable fishes decreased by overfishing (VAL; ALMEIDA-VAL, 1995).This idea has been contributing for the raise of aquaculture intake in the global production of fish and is becoming an alternative to the increase of food production of aquatic origin (ARBELÁEZ-ROJAS et al., 2002).
However, if previous genetic planning is not included for the activity, aquaculture process will not result in the desired genetic integrity of captivity populations, and some management will be required to allow genetic breeding to the stocks.To achieve good results and genetic improvement of the progeny, high genetic variability is needed in founder populations, assuring the existence of sufficient variety that will account for survival and improvements of consecutive generations.
Studies comparing genetic variability of artificial and natural populations of fish are required to understand how the variations on mating choices, survival, reproduction, and growth rate contribute to alterations in genetic and genotypic frequency, and if these alterations improve fish adaptation to changing environments.Almeida-Val and Farias (1996) suggested that the environmental pressure that was imposed to this and other species along their evolutionary history, gives this group a high genotypic variability and, consequently, phenotypic plasticity.
Studies based on protein markers are among the first methodologies used to investigate genetic variability between the 70s and 80s, when few polymorphic loci were discovered (PARKER et al., 1998).A number of allozyme surveys examined the amount of variation in different populations of the same or related species (HEDRICK, 2005).Isozymic markers were widely applied to detect genetic variability losses of fish artificial populations compared to natural populations (RYMAN;STAHL, 1980).Protein markers were also used in phylogenetic studies (TRINGALI et al., 1999) and to detect natural fish population variability (CALCAGNOTTO; TOLEDO-FILHO, 2000).
According to Allendorf and Ryman (1987) and Allendorf et al. (1987), these markers require a great extent of loci to be used in the study of population variability, since the occurrence of mutational events on the genetic material may not always result in alterations of protein structure.Amino acid replacements in the sequence of a protein may not result in changes of its electrophoretic mobility, underestimating allele variability levels (SOLFERINI; SCHEEPMAKER, 2001).However, these systems have been recently used as powerful tools in the analysis of population genetic structure and ecological mechanisms of dispersal, isolation and recolonization in wild fish (PHILLIPS et al., 2009).
Considering the decrease in the survival of artificial populations of tambaqui in the South of Brazil, this study has aimed to analyze the genetic variability of wild and captivity populations of Colossoma macroporum.
The specimens were collected and conserved in liquid nitrogen at the Laboratory of Ecophysiology and Molecular Evolution from National Institute for Amazon Research (LEEM-INPA) in Manaus, Amazonas State, Brazil.Tissue samples as muscle, liver and heart were homogenized with a plastic stick in propylene tubes (1.5 mL) with 100 μL of Tris-HCl 0.02 M, pH 7.5 buffer and centrifuged at 44,000 g at 5°C during 30 min.Due to the presence of a great amount of fat in the liver, 100 μL of carbon tetrachloride (CCl 4 ) was added to the tubes (PASTEUR et al., 1988).
Gels were prepared with 13% corn starch (VAL et al., 1981).Two buffer solutions were used: Tris 0.135 M/Citric acid 0.043 M pH 7.0 (TC) (SHAW; PRASAD, 1970) and Tris 0.18 M/Boric acid 0.1/EDTA 0.004 M pH 8.6 (TBE) (BOYER et al., 1963).The gel was made with TC diluted 1:15 and TBE diluted 1:4.The enzyme extract was applied to the gel using Kodabromide (Kodak ® ) paper strips (4 mm x 8 mm) soaked with the samples.A voltage gradient of 200 V was applied to the gel during 16 hours (TC buffer), and a voltage gradient of 400 V was applied to the gel during 10 hours for TBE buffer.After electrophoreses, the gels were horizontally sliced lengthwise into two slabs, which were incubated with specific staining solutions according to Murphy et al. (1996) (Table 1).Allele and loci designations followed the International Union of Biochemistry and Molecular Biology (IUBMB, 1992) nomenclature.

Results and discussion
The especimen's isozymic profiles were analyzed through corn starch gel electrophoresis using nine enzymatic systems (Table 1).Table 2 shows the allelic frequencies obtained for the 18 loci to the population from each locality.Despite the small quantity of isoenzyme systems analyzed, the loci/allele numbers were adequate for the population analysis and genetic parameters obtained.The electrophoretic banding pattern was similar to that described by Limeira et al. (2009) in two populations of Rineloricaria for the average polymorphism and allelic frequency.
Low polymorphism rate was also detected as the number of alleles per locus (Table 3).We did not detect deviation from Hardy-Weinberg equilibrium at any of the polymorphic loci in C. macropomum populations (Table 4).
Mean value for observed and expected heterozigosities for all loci of C. macropomum were: Wild population, Ho = 0.125 and He = 0.120; Pentecoste, Ho = 0.053 and He 0.106; Jaboticabal, Ho = 0.059 and He = 0.118; and Itacoatiara, Ho = 0.134 and He = 0.131.Wright's F statistics (F IS and F ST ) for estimating the excess of heterozygotes and population structure (WRIGHT, 1978) was calculated for all loci.Mean values were F IS = -0.04 and F ST = 0.07 for wild population; F IS = 0.50 and F ST = 0.62 for Pentecoste; F IS = 0.50 and F ST = 0.56 for Jaboticabal; and F IS = -0.02and F ST = 0.02 for Itacoatiara.Wright`s F statistics shows that populations of Pentecost and Jaboticabal lack heterozygotes (F IS > 0) and show high structure (F ST > 0.25).However, we found that the wild and Itacoatiara populations presented excess of heterozygotes (F IS < 0) low levels of structure (F ST < 0.05).Nei's genetic identity and distance values (NEI, 1978) values are presented in table 4. Figure 1 shows UPGMA dendrogram using Nei's (1978) similarity indices for analyzed populations.2006) enzyme polymorphic loci are successfully used to differentiate populations from diverse fish species.The efficiency in using isoenzyme as genetic markers and as a tool for diagnosis, identification and limitation of fish stocks or any other live stock has also been recently reaffirmed in literature (PHILLIPS et al., 2009;SILVA et al., 2008).
Table 2 shows the allelic frequency of nine isoenzyme systems.In the present study, only six, out of the analyzed 18 loci were polymorphic.Fbp-2 presented polymorphism in all analyzed populations of Tambaqui; G6pdh-2 presented monomorphism in Pentecostes and Jaboticabal populations; G6pdh-3 and Gpi-1 were monomorphic for Pentecostes specimens and the Gpi-2 and Pgm-1 were monomorphic for Jaboticabal specimens.These data indicate a loss of genetic variability in artificial populations that may be the result of an extensive period of population isolation from nature and a high degree of inbreeding.This may explain these findings with such a small amount of variable loci.
We also verified that the number of alleles per locus was practically the same in all populations.It has been observed that the wild and Itacoatiara populations present the same polymorphism index and that the Pentecostes and Jaboticabal populations present similar values indicating similar origin for the two groups (Table 3).According to Avise and Aquadro (1982) populations differing at the level of 10% generally represent distinct species.
The isoenzyme systems can reveal the differential expression of duplicated loci among different tissues and in different developmental stages of the animal.Therefore, the same isoenzyme system can show different patterns of enzymatic activity in each cell when present.According to Fisher and Whitt (1978) these differences cause regulation on gene's tissue and ontogenetic expression.Regulatory gene variation may be operative in determining the number of molecules of enzyme in a cell.Higher or lower levels of isozymes are also related to different metabolic activity of the tissues, as well as to the respective subcellular compartmentalization of enzymes (ZAWADZKI et al., 2001).
The genetic variability analysis revealed the occurrence of a wide variation in heterozygosity when tambaqui populations are compared to the average heterozygosity for freshwater fish (He = 0.051) established by Ward et al. (1992).Wild population (0.125), Pentecostes (0.053), Jaboticabal (0.059) and Itacoatiara (0.134) presented higher values of heterozygosity.According to Ward et al. (1994) the values of expected heterozygosity for 107 species of fishes analyzed by starch gel electrophoresis, range from 0 to 0.05 in 54% of them, from 0.05 to 0.10 in 30%, from 0.10 to 0.15 in 12%, and higher than 0.15 in 4% of them.Tambaqui wild population follow in the 4% of highest heterozigosity levels.The loss presented in captivity populations is directly related to the size of breeding stocks and to the increased inbreeding in fish farms along the generations.Thus, the time of isolation of these populations (number of generations since the beginning of that stock) contributed to the loss of heterozigosity.It is important to emphasize that the population from Jaboticabal has been funded by the Pentecostes DNOCS population.
We consider the isolation time (ZAWADZKI et al., 2005) and the decrease in population size the main cause of the increase in the rate of inbreeding and in the consequent loss of genetic variability of populations of Pentecoste and Jaboticabal.
According to Frankham et al. ( 2004) captive populations of threatened species lose genetic diversity due to the bottleneck at found, small subsequent population size and because Ne (population effective size) is lower than N (population size).Consequently, loss of genetic Environment Itacoatiara Pentecoste Jaboticabal 0.95 0.96 0.97 0.98 0.99 1.00 Similarity diversity is minimized by using an adequate number of founders, minimizing the number of generations by breeding from older animals, or using cryopreservation, and maximizing both the population size and Ne/N ratio.
We suggest either a replacement or renovation in the parental couples of Pentecostes and Jaboticabal fish stocks as the best way to recover their genetic variability.Populations utilized in fish culture, or restocking programs, must be founded from a sufficiently bigger number of specimens (ALLENDORF; RYMAN, 1987) in order to avoid an important decrease in Ne.However, big populations originated from small groups or from unequal progenitor contribution can exhibit a low Ne (BRISCOE et al., 1992), and low Ne is probably amongst the most important causes of genetic variability loss, resulting in genetic drift (FRANKHAM, 1996).
The genetic effect of stock formation in fish culture has been documented in several fish species and has shown that even for stocks formed by 100 specimens, no population will absorb more than 45% of these specimens' contributions without suffering a drastic and potentially dangerous fall in the genetic variability (TANIGUCHI, 2003).Consequently, isolated populations, especially small ones, may have their heterozygosity eroded by inbreeding process or by chance (SOLÉ-CAVA, 2001).
The values of genetic distance and genetic identity among tambaqui populations (Table 4) showed that the individuals from the Itacoatiara population probably came from the environment population (wild population) or were recently founded by adults from nature.This assumption is based on the fact that the values of genetic distance and genetic similarity between these populations showed no difference (genetic distance = 0 and genetic similarity = 1.0).A similar pattern was also observed for Pentecostes and Jaboticabal populations and we can presume that both belong to the same founder stock.
Population genetic analysis, using different types of genetic molecular markers, complemented with other fishery biology approaches, is urgently needed for the current management of captivity populations and for future health of cultured fish generations.
Our following plans are to keep searching for appropriate enzyme or isozyme variant alleles, which can conveniently diagnose tambaqui and its genetic stocks, to help in the long-term management of this species in the intensive and extensive aquaculture activities.
The specimens from Pentecostes and Jaboticabal presented loss of genetic variability, what requires an immediate replacement or renewal of these stocks.
This study showed that the population of Itacoatiara was founded recently from wild and variable specimens and that the number of found stocks is, probably, bigger than the number of founding stocks of the artificial populations here studied.

Figure 1 .
Figure 1.UPGMA dendrogram from the similarity indices ofNei (1978) for the four Colossoma macropomum populations.According toChaplin et al. (1998) andHovgaard et al. (2006) enzyme polymorphic loci are successfully used to differentiate populations from diverse fish species.The efficiency in using isoenzyme as genetic markers and as a tool for diagnosis, identification and limitation of fish stocks or any other live stock has also been recently reaffirmed in literature(PHILLIPS et al., 2009;SILVA et al., 2008).Table2shows the allelic frequency of nine isoenzyme systems.In the present study, only six, out of the analyzed 18 loci were polymorphic.Fbp-2 presented polymorphism in all analyzed populations of Tambaqui; G6pdh-2 presented monomorphism in Pentecostes and Jaboticabal populations; G6pdh-3 and Gpi-1 were monomorphic for Pentecostes specimens and the Gpi-2 and Pgm-1 were monomorphic for Jaboticabal specimens.These data indicate a loss of genetic variability in artificial populations that may be the result of an extensive period of population isolation from nature and a high degree of inbreeding.This may explain these findings with such a small amount of variable loci.We also verified that the number of alleles per locus was practically the same in all populations.It has been observed that the wild and Itacoatiara populations present the same polymorphism index and that the Pentecostes and Jaboticabal populations present similar values indicating similar origin for

Table 1 .
Buffer systems used for separation of enzymes.
a EC: Enzyme Commission number; b H: Heart; L: Live; M: Muscle.

Table 3 .
Genetic variability of the 18 studied loci in Colossoma macropomum populations (mean ± standard error).

Table 4 .
Chi square contingency analysis for polymorphic loci: a DF: Degrees of Freedom; b P: Chi square probability.