BIOCHEMICAL GENETIC DIFFERENTIATION AMONG WILD POPULATIONS OF MILKFISH ( Chanos chanos ) IN INDONESIA

Four populations of milkfish (Chanos chanos ) were collected (N=100) from coastal water of  Aceh. Bali, East Java and South Sulawesi and were examined electrophoretically at 29 loci to deterurine the genetic vuriability and the population structure.


INTRODUCTION
The milkfrsh (Chanos chanos) is one of the rnost important species extensively culttued in brackish- water Indonesia ponds.Recently, this species has become more important as frozen and live bait for tuna long-line capture.The supply of milkfrsh fry comes mainly from the wild.High demands on the supply of fry emphasize the need for artificial propagation to supplement the supply.The technology for artificial seed production of rnilkfish has been developed in Gondol Research Station for Coastal Fisheries, Agency for Agriculture Research and Developrnent, and the technology has been adopted by private hatcheries who rnass produced.egg and fry to supply brackishwater pond of milkfrsh culture (Sugarna et aI.,Igg7).
The broodstocks for seed production have been collected frorn several regions in the country.An understanding of rnilkfish population structure particularly in this region is therefore a requirernent for broodstock developrnent since the existence of subpopulations or stock, heterogeneity will necessarily influence the forrnulation of stocking and./or conserva tion practices ( Macarana s et al.,1990; Taniguchi & Sugarna, lgg0).
Although various studies have been conducted on reproduction, larva rearing, nutrition and disease of milkfish, none study directed toward genetic improvement.Therefore, genetic studies of this species were conducted to obtain basic information on the amount of genetic variability and genetic differentiation among wild populations.Electrophoresis was used for allozyme analysis.The results may be important for the future development of milkfish breeding and proper management of natural populations.
Whole fish were freighted to Gondol Research Station for Coastal Fisheries, Bali, where they were held frozen at-20,,C until electrophoresed.1l *) Resear.he',fllt'st'rrr.lrStrrt.i.rr firr o,^st.lFislrt:ri.s. (irrrrkrl.Four tissues (skeletal rnuscle, liver, heart and eye) were dissected, tissue extracts were prepared and starch gel electrophoresis was performed following the methods of Sugama et al. (1996).Fifteen enzyrnes and protein systern were evaluated (Table l).The staining procedures followed the method of Show & Prasad (1970).Two buffer systems Tris-citric acid pH-8 (TC-8) and (citric acid N-(3-Amino propylmorpholine pH:6 (CAMP-6) were used.Extracts from all four tissues (skeletal muscle, heart, liver and eye) were tested on all combination of buffer systems and stained in the search for optimal electrophoretic activity and resolution.The list of enzymes and proteins examined, tissue specificity, Iocns detected and buffer systems giving clearest resolution are given in Table 1.A capitalized abbreviation represents each enzyme.The same abbreviation with only the first letter capitalized represents the locus coding for that enzyme.
When multiple loci code for an enzyme, the locus with the most anodal migration is designated one, the next two and so on.Allelic variant are designated according to their relative mobility.
The most cornmon alleles is designated 100 and other alleles are given numbers that indicate their mobility relative to that of the common allele.Cathodal systems are designated in a similar way but given a negative sign (Shaklee et a|..1990) (Table 2).Sorbitol deh5'dr'eg"uase (SDH ) Superoxide dismutase (SOD)   Sarcoplasrnic protein (SP)   Adh

Data Analysis
A log-likelihood ratio for goodness of fit (Sokal  & Rohlf, 1981) was used to compared observed frequencies of phenotypic classes with those expected under Hardy-Weinberg equilibrium.A locus was considered to be polymorphic if the most cornmon allele was equal or less than 0.99 (l%) at one or more sarnples.Heterozygosity was defined by h= 1-pi' , where pi3 is the frequenc5' of the i-th allele.H is the average of h over loci including monomorphic.Observed main allele frequencies at polyrnorphic loci were compared a cross All localities using heterogeneity test (lf ).
The genetic distance arnong populations were calculated using Nei ( t972) formula and dendrogram was constructed from the matrix genetic distance using Unweighted Pair Group Method with Arithmetic Averase (UPGMA) (Sneath & Sokal. 1973).

RESULTS
Enzymes and protein exarnined, locus detected, tissues assayed and buffer systerns giving the clearest resolution are presented in Table 1.
Surnmary of genetic variability in four populations of rnilkfrsh which were indicated by the proportion of polyrnorphic loci the number of alleles per locus and the average observed (Ho)   and expected (He) heterozygosities are presented in Table 3.The proportion of polymorphic loci ranged from 0.276 to O.4I4 with an average of 0.336.The average number of alleles per locus ranged from 1.482 to 1.689 with an average of L.577.The observed heterozygosity (Ho) ranged from 0.062 to 0.074 with an average of 0.068 and average expected value He rangeC of 0.062 to 0.0?3 with an average of 0.068.The ratio of Ho/He ranged of 1.000 to 1.105 with an average of 1.007.
Since this value has nearly unity, there is no indication of inbreeding in these milkfish populations.In general, the value of genetic variability for Aceh sample was slightly higher than those of other samples.
The allele frequency differences among populations were tested by Chi-Square for heterozygosity at main allele of polymorphic loci (Table 4).In some loci, aIIeIic differences between two populations were significant such as in main alleles-100 of Est-1 and -Gpd loci for Aceh vs other populations, Adh and Est-2 for South Sulawesi vs other populations, also Adh, Est-1, Est-2, -Gpd and Pgm-2 for Aceh vs South Sulawesi, while no allellic differences found between Bali and E. Java populations (Table 4).This allelic differences indicated that these populations is independence to one and others, except for Bali and E. Java populations should be considered as a single population.The valrres of Nei's genetic distance (D)   between each pair of poprrlations ranged between 0.00028 to 0.00270 with an average of 0.00155 (Table 5).To stunmarize the relations arnong populations, a UPGMA dendrograln was conducted on the basis of a matrix at Nei s genetic distance (Fig. 2).The UPGMA cluster analysis of genetic distance revealed very clear relationship between genetic distarrce and geographic distance.
The result showed three main groupings of llopulations that were Aceh, S. Sulawesi with Bali and E. Java considered as one population.

DISCUSSION
The best estirnated of genetic variation in natural population is the average heterozygosity (H) tAllendorf & Utter, 1979).The mean observed heterozygositl' lH) of four ynpulations for rnilkfish in the present studv is 0.068.Winans (1980) cornpiled data for 82 species offrsh and calculated a rnean heterosigositv of O.478.Surnantadinata and Taniguchi (1982) found the average heterozygosit5r sg 0.066 for black sea bream, Taniguchi & Sugarna (1990) estimated heterozygosity ofred sea brearn at 0.058 and Fujio (1981) calculated rnean heterozygosity for 103 species of marine fish at 0.053.Thus, the heterozygosity of rnilkfish from the wild population is slightly higher than other marine frsh.
The rnain interest of the present study was to search if there is any genetic differentiation arnong populations of milkfish within Indonesia.The heterogeneity test revealed statistically significant differences in allele frequencies between pairs of populations except for Bali and E. Java.
According to Grant et al. (1987), the allele frequency differences between populations of marine fishes resulted from the action of three forces, i.e. rnigration, randorn genetic drift and natural selection.They also suggested that little lrrblc -j, Matrix Nci's gurctic distrrnce bctr.vceneverv parr of firur cod Gobus rnorh,tta (Mork et al.. 1985)   The genetic differentiation observed between natrrral populations of rnilltfish in the present study could be explained by the lack of rnigration or gene flow particularll' between populations of Aceh aud S. Sulawesi which were isolated frotrt each other due to geographical distance.
Consequentl5' no genetic differentiation observed betrveeu Rali and E. .Java poprrlations indicated free gene flow of the two polnlations due to verv close geographic distance.
The average genetic distance arnong populations is 0.00155.This value is typical of div er gence b etween conspecifrc populations.The UPGMA cluster analysis sf genetic distance forrned three geographical groups (gror.rpBali and E. Java.Aceh and S. Srrlawesi), and very clear relationship between the genetic distance and geographic distance (Figrrre 2).The values of genetic distance and the hiehly sigrrificant differerrce irr allelic frequencies arnong populations, suggested the independence of the subpopulation ofAceh frorn others, except Bali and E. Java was cousidered as one population.This canbe understandable since Aceh is geographicallv rnuch farer frorn the other populations.It was concluded that the natural poprrlation of rnilhfish in Indonesia could be divided into 3 geographical groups located in Western part of Indonesia (Aceh), Miclle part of Indonesia (Rali and Java ) and Easterrr part of Indonesia (Sorrth Srrlawesi).
Figure 2. Dendrogram drawrt frorn the genetic distance between every pair of poprrlations of rnilhfrsh

Table 2 .
Allele frequencies of polymorphic loci in four populations of milkfish, Chanoe chanos.Sarnple size for each population is 100 and )F = values of Chi-Square test for goodness of fir.