MOLECULAR ZOOGEOGRAPHY OF FLORIDA SNOOK (PISCES: CENTROPOMIDAE) BY A SIMPLIFIED TECHNIQUE FOR THE PURIFICATION OF FISH MITOCHONDRIAL DNA by Maurizio A. Mangini A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Marine Science in the University of Douth Florida May, 1987 Major Professor: John C. Briggs, Ph.D.
INTRODUCTION Snook are estuarine fishes of the family Centropomidae, genus Centropomus. The genus comprises several species found in tropical and subtropical American waters (Rivas, 1986). In the Atlantic Ocean, snook range from the Gulf of Mexico off Texas and lower Florida to the Caribbean and Central and South America. They are occasionally found as far north as North Carolina and as far south as Rio de Janeiro, Brazil. The species of the genus Centropomus posess several overlapping meristic and morphometric characters. Rivas (1986) recently completed an extensive revision of Centropomus. He performed statistical examinations of 47 taxonomic characters, and recognized 12 valid species of the 30 nominal species that have been proposed since 1792. Rivas also concluded that six species occur in the Altantic, six in the Pacific, with no species occuring in.both oceans. C. pectinatus, C. undecimalis, poeyi, parallelus, mexicanus and C. ensiferus in the Atlantic; medius, nigriscens, viridis, unionensis, C. armatus and C. robalito in the Pacific. Although no single species are considered to occur both in the Atlantic and Pacific, several transisthmian species pairs were recognized: C. undecimalis and C. viridis, ensiferus and robalito, pectinatus and C. medius. 1
parallelus and mexicanus were reported as 2 closely related Atlantic sympatric species, with no close Pacific relationships. Previous workers proposed a different taxonomy: Fraser (1978) considered as Atlantic species Centropomus undecimalis (Bloch), C. poeyi Chavez, parallelus Poey, pectinatus Poey, and C. ensiferus Poey. A sixth Atlantic species, constantinus, may occur, but according to Fraser (1968), its diagnostic characters overlap those of f. parallelus and a revision is necessary before C. constantinus may be recognized as a separate species. In the Pacific Ocean, Chavez (1963) and others (Anon. 1976), reported that snook are found from Baja California to Peru, and include C. nigriscens Gunther, f robalito Jordan and Gilbert, and C. pectinatus. The latter is considered by earlier authors to be the only species found on both the Atlantic and Pacific coasts of Central America (Chavez, 1963, Anon. 1976). Therefore, considerable disagreement exists among authors who recently worked on Centropomus. Elucidation of centropomid systematics is complicated by considerable overlap of meristic and morphologic characters between species. To date, infraspecific genetic variation has not been inferred from phenotypic characters. In Florida's subtropical environment, spawning activity may occur throughout the year, but usually occurs from May to November, with maximum activity from June to August (Gilmore et al., 1983). Mature fish congregate at the mouths of estuaries and spawn off nearby beaches. In years of low rainfall, spawning adults can be found in the lower reaches of estuaries, where oceanic salinities occur (Personal observation during Department of Natural Resources 1985 collections in 2
Tampa Bay). Spawning activity is maximum at night, on a full moon, and incoming tides. The eggs are carried into the estuary, and, as noted by Gilmore et al. (1983), "larvae and early juveniles spend 15 to 30 days in transit from the spawning site to peripheral estuarine and freshwater nursery grounds". As juvenile fish increase in size, they migrate to seagrass beds (at circa 150-300mm standard length), and larger snook (>300 mm SL) are found in marine and estuarine habitats. All species of the genus Centropomus are stenothermic. Their northern and southern limits of distribution coincide with the coastal intercepts of the August and February 24 C surface isotherms of Sverdrup et al. (1942, in Rivas, 1968). These isotherms correspond to a range from North Carolina to southern Brazil. Shafland and Foote (1983) demonstrated that fingerling f. undecimalis may not survive at temperatures below 15 C. Although occasionally found north of Florida, the northern limit of the populations is mid Florida, where February nearshore water temperatures range between 10 and 15 C (Rivas, 1968). Such winter temperatures are low enough to stress the adult snook. The common snook, C undecimalis, is one of the most desired florida gamefish. Public concern about reduced catches throughout the range of the species resulted in the snook receiving the designation of "species of special concern" in Florida waters, and prompted studies on its life history and polulation dynamics. A survey of the Florida snook fishery was conducted by Marshall (1958). More recently, Bruger (1983) examined the population dynamics of snook in the Naples-Marco Island area. He estimated a 1977 3
4 population of 28,000 mature fish. In 1981, the population decreased by approximately 707., to an estimated 8,600 mature fish. Bruger concluded that, in his study area, one half to one third of the adult snook die annually. He considered fishing pressure a significant cause of mortality, exacerbated by habitat loss, fresh water flow alteration, and habitat contamination by pesticides. Besides the previously cited local studies However, little is known about the statewide population dynamics of the species, and no information exists on the degree of genetic isolation of the several estuarine populations. Snook are considered to be relatively nonmigratory fish, making only short coastal movements. Although f. undecimalis exibits migration within an estuary, as a function of life cycle stage and reproductive cycle (Volpe, 1959). The species has limited migration along the western Florida coastline. Volpe (1959) found that over a one year tracking period, of 57 returned tagged specimens, 797. were caught within 6 miles of the tagging site. Of the eleven specimens that traveled more than 6 miles from the tagging site, five made inshore-offshore movements, six moved along the coast, and only one a distance of 38 miles. Snook tagged on the eastern coast of Florida kept to the tagging area, with one specimen having traveled 13 miles in 29 days. Tagging returns from the Naples area reported by Bruger (1983) indicate that "snook are nonmigratory, but do make annual excursions from estuarine and fresh waters to open gulf beaches and inshore waters during the summer spawning season". If the geographic range of the Florida populations is limited to one estuary, and there is no significant migration between estuaries, then the efficient management
of this game fish requires a different approach than the current one, which treats the species as a continuous population. The determination of fish population dynamics has traditionally relied on physical tagging of individual specimens. More recently, genetic markers have been employed to determine the degree of overlap between populations, such as electrophoretic analysis of polymorphic proteins (Avise, 1974). However, when population 5 numbers are small, as in the case of rare or difficult to collect specimens, the data thus generated can be ambiguous due to difficulties in unequivocally differentiating the gene products of different alleles. Also, analytical problems are inherent in the interpretation of such data from diploid organisms. These difficulties are greatly reduced in data derived from uniparentally transmitted (clones) genetic material (Avise et al., 1979a). In recent years, molecular biology has offered another approach to population genetics. The discovery of deoxyribonucleic acid (DNA) restriction endonucleases that recognize and cut speacific base sequences in double stranded DNA, has provided a tool for the rapid analysis of DNA at the level of nucleotide sequences (Roberts, 1982). Animal mitochondrial DNA (mtDNA) consists of a single, rapidly evolving molecule of ca. 16,000 base pairs that is inherited maternally (Brown, 1981). The determination of mtDNA polymorphisms can yield data on genetic variation at the base sequence level by comparison of individual genomes. The possibility for individual heteroplasmy does exist (Densmore et al, 1985; Harrison et al., 1985). However, the study of restriction generated lenghth polymorphisms offers great potential for the characterization of population structure and
differentiation, and has been used to determine subgeneric and infraspecific genetic variation in fish populations (Avise et al., 1984; Berg and Ferris, 1984; Bermingham and Avise, 1986; Graves et -al., 1984; Gonzalez-Villasenor and Powers, 1986). Saunders et al. (1986) demonstrated that continuously distributed conspecific populations of marine organisms can posess geographically correlated mtDNA differentiation. The species in the genus Centropomus are morphologically very similar. Diagnostic meristic characters are few, and species identification is often dependent on modal values of morphological characters (Rivas, 1986). Therefore, infraspecific genetic variation can not be inferred by phenotypic (morphological) variation, and a biochemical or molecular approach is required, such as determination of mtDNA polymorphisms. The strength and advantages of mtDNA polymorphism analysis reside in the function and structure of the mtDNA molecule, in its mode of inheritance, and in the molecule's evolutionary mechanisms. Structure of mtDNA MtDNA is a circular, supercoiled molecule, common to all metazoans and most protozoans. Metazoan mtDNAs range in size from 15,700 to 19,500 base pairs (b.p.) (Brown, 1983). The molecules contain 35 genes, which encode for small (12s) and large (16s) ribosomal RNAs (rRNAs), transfer RNAs (tRNAs) and messenger RNAs (mRNAs). Five of the mRNAs are known to be translated into proteins that are subunits of the mitochondrial electron transport system: Cytochrome C oxidase subunits I, II, III, ATPase subunit 6, cytochrome b, and 8 unidentified proteins 6
(Avise and Lansman, 1983). Unlike nuclear DNA, mtDNA generally lacks intergenic spacer sequences; if present, these sequences consist of one to a few nucleotides. Complete colinearity exists between genes, primary transcripts, and mature gene products (Anderson et al., 1982; Montoya et al., 1981). Therefore, it is estimated that up to 90 7o of the mitochondrial genome is transcribed. The approximate ten percent that is not transcribed consists of a sequence block that contains a structure known as the D loop region. This structure may serve as the primer for the heavy strand replication, or as an initiator sequence fortranscription. Of special interest to population geneticists is the fact that the D loop region is the most variable region in the mitochondrial genome (Upholt and Dawid, 1977; Brown, 1981). The smallest animal nuclear genomes are 25,000 times larger than mitochondrial genomes. copy number from 1 to 1 x 10 The former contain sequences that vary in per haploid genome (Brown, 1983), while mtDNA contains sequences that occur only once per genome The mitochondrial genome is less susceptible to frequent sequence rearrangements, which are common among the chromosomes of the nuclear genome. Mitochondrial base sequences are highly conserved among taxonomic groups. There is a high degree of gene sequence homology between metazoan species belonging to different systematic orders, although the genes are not arranged in the same order in all metazoan taxa. Therefore, the genetic structure of mtDNA differs markedly from nuclear DNA. Mode of Inheritance of mtDNA 7
Relevant to the use of mtDNA in the screening of populations is the molecule's mode of inheritance. MtDNA is clonally inherited: In most groups studied, mtDNA molecules are inherited maternally, via the cytoplasm of the egg, and are homogeneous in nucleotide sequence composition within an organism (Avise and Lansman, 1983). Exceptions have been reported, such as individual heteroplasmy found in populations of fishes (Bermingham et al., 1986), lizards and crickets (Densmore et al., 1985; Harrison et al., 1985). This mode of inheritance obviates many of the genetic variables that are a consequence of sexual recombination in nuclear DNA. Therefore, the mechanisms of genetic variation possible in mtDNA are greatly reduced. As a conseqeunce, the evolution of the mtDNA molecule and its inheritance occur in a relatively simpler, more straightforward mode than in nuclear DNA. Mode of mtDNA evolution Most mtDNA variations are changes in the nucleotide sequence of the mtDNA, and arise mainly from nucleotide substitutions and sequence rearrangements, not deletions or additions (Cann et al., 1984). Both microvariation and macrovariation in mtDNA size do occur. Microvariation is due to additions of from one to a few base pairs (Aquadro and Greenberg, 1983; Crews et al., 1979). Macrovariations result from larger deletions and additions, but have been described exclusively in the nontranscribed region of the D loop and its adjacent regions (Upholt and Dawid, 1977). MtDNA nucleotide substitution rates are 3.5 to 5 times faster than in single copy nuclear DNA (Brown et al. 1979; Brown, 1983.). In closely related species pairs, rates of 8
9 change between mtDNA and nuclear DNA are even more divergent. Among species separated by 15 million years or the initial rate of change is up to 10 times greater in mtDNA than in nuclear DNA, but it involves only 25 to 30 7. of the mtDNA sequences. the determination of mtDNA polymorphisms is an adequate tool for estimating the degree of genetic divergence among closely related populations. (Cann et al., 1984). Comparison of Protein Polymorphisms and mtDNA Polymorphisms Protein electrophoresis has been used in the evaluation of infraspecific genetic variation in marine organisms. Heterogeneity estimates are derived from comparisons of allozyme frequencies between populations. If the populations exhibit significant it is assumed that it is caused by restricted gene flow with different selective pressures or random drift, or are due to different mortality rates of different genotypes as a function of geographic location and local selective pressures. A comparison of the two technologies can be obtained from the work of Avise et al., (1979b), who compared protein electrophoresis and mtDNA restriction analysis among populations of pocket gophers. MtDNA was found to be more sensitive in identifying genetic relationships between conspecifi c populations. In the mar ine environment, the horseshoe crab Limulus polyphenus has been found to comprise a northern and a southern population by analysis of mtDNA restriction demonstrating that a continuously distributed marine organism can show infraspecific genetic variation (Saunders et al., 1986). Studies of infraspecifi c protein polymorphisms in fishes with
10 pelagic larvae have shown genetic variation to occur only between widely separated populations (Skow and Chittenden, 1980), or have shown that the populations studied are homogeneous over large geographical areas (Winans, 1980). In fishes with pelagic larvae, mtDNA genotypes clearly differentiated between the European eel Anguilla anguilla and the American eel, rostrata, while allozyme data could not readily differentiate the two species (Avise et al, in press). However, Graves et al. (1984) found no geographical correlation in mtDNA genotypes of skipjack tuna, Katsowonus pelamis, with approximately 40 assayed restriction sites from populations collected in Hawaii, Puerto Rico and Brazil. Graves et al. concluded that since the skipjack tuna lack discrete spawning sites, interoceanic larval dispersal and gene flow are sufficient to prevent genetic differentiation of the Atlantic and Pacific populations. Protocols Used in Evaluating mtDNA Polymorphisms MtDNA nucleotide sequence polymorphisms can be detected by digesting the mtDNA molecule with restriction endonucleases. The enzymes cleave the DNA as a function of the specific base sequences they recognize. The fragments thus generated are separated by agarose gel electrophoresis. Comparisons of different specimens treated with the same restriction endonucleases will yield estimates of base sequence variation among the samples. Repeating the procedure with different enzymes, including six, five, and four base pair cutters, will increase the number of bases compared for potential sequence variation (polymorphism) along the mtDNA molecule. However,
11 the fact that different restriction fragments from different specimens migrated the same electrophoretic distance does not mean the fragments are of identical base sequence, but that they are of the same molecular size. Therefore, without sequencing the fragments, or conducting multiple restrictions, it cannot be concluded that differences between the two samples do not occur. A method used by fish population biologists minimizes the possibility of scoring two nonhomologous fragments as identical due to chance comigration by rerunning fragments of questionalbe identity side by side at different gel concentration (Bermingham and Avise, 1986). Mathematical models and algorithm programs are available for calculating sequence divergence estimates and evaluating clusters of sequence distance matrices (Dixon, 1981; Takahata and Palumbi, 1985; Nei and Li, 1979). Determination of interpopulation mtDNA polymorphisms is a technology that has not been more widely applied due mostly to technical difficulties. To obtain mtDNA of a purity adequate for restriction analysis, most methods require the use of fresh tissue as a source of DNA, and separation of mtDNA from nuclear DNA by ultracentrifugaton in cesium chloride gradients. This method is expensive and logistically unavailable to many laboratories. Chapman and Powers (1984) developed a mtDNA purification and enrichment protocol that does not require ultracentrifugation. However, that protocol depends on the separation of intact mitochondria from fresh tissue. Also, a swinging bucket refrigerated centrifuge is required, since the technique depends on separation of intact mitochondria by density gradient centrifugation.
The first objective of this research was to develop a methodology that facilitates the purification of mtDNA by permitting pelleting 12 of mtDNA from either frozen or fresh tissue, without depending on first pelleting intact mitochondria or on separating mtDNA by ultracentrifugation. Once developed, that protocol could be used to clone fish mtDNA and to produce homologous probes for the examination of infraspecific genetic variation among conspecific estuarine fish populations.
13 Objectives The objectives of the proposed research were threefold: First, to develop a methodology that would allow the purification, cloning, and radiolabelling of fish mtDNA from fresh or frozen tissue, without depending on ultracentrifugation or density gradient separation protocols; second, to characterize the genome size (molecular weight) of the mtDNA of f undecimalis; and third, to apply that protocol to test the hypothesis that the populations of snook separated by the Florida peninsula are genetically distinct. The proposed objectives were accomplished by: 1. Demonstrating the applicability of simplified molecular techniques to the extraction and purification of mtDNA from either frozen or fresh tissues. 2. Characterizing the mtDNA genome size of C. undecimalis from Florida waters. 3. Cloning the mtDNA genome of f. undecimalis in order to obtain a high specificity molecular probe for the determination of mtDNA population polymorphisms. 4. Utilizing restriction endonuclease analysis of mtDNA to estimate and compare the level of mtDNA base sequence polymorphisms within east and west Florida populations of C. undecimalis. 5. Using the mtDNA base sequence polymorphisms data to evaluate the degree of genetic differentiaiton and reproductive isolation among the east and west Florida populations of f. undecimalis.
14 METHODS Study Populations Specimens of f undecimalis were collected from Tampa Bay and Fakahatchee Bay (Everglades area) for the West Florida populations. Fishes from East Florida populations were collected from the Indian River coastal area: Sebastian River Bay and the St. Lucie River. The mtDNAs of two other snook specieswerealso examined. f parallelus (the fat snook) from the Indian River and f. pectinatus (the tarpon snook) from the Sebastian River. They were used as outgroup populations. The latter two species were extremely difficult to obtain alive or sufficiently fresh for the isolation of DNA. Their scarcity is reflected in the number of specimens sampled. Sample Preparation The most suitable tissue for extracting mtDNA was determined preliminary to the population assay. The traditional source of fish mtDNA is via extraction by density gradient or ultracentrifugation from liver, kidney, or mature oocyte tissues (Avise et al., 1984; Chapman and Powers, 1984). To optimize extraction efficiency of snook mtDNA, and its subsequent restriction, snook livers, brains, hearts, gonads and samples of skeletal muscles were assayed as sources of
mtDNA. The mtDNA thus obtained was digested with restriction endonucleases, and the mtDNA fragments identified with radiolabelled mtDNA probes. MtDNA Purification, Restriction and Characterization The Identification of mtDNA restriction fragments by the simplified methodology presented here was compared to the results obtained with the classical, cesium chloride purification protocol. Cesium chloride purified mtDNA was obtained by a modification of Maniatis (1982): liver and kidney tissues were homogenized in a motorized, Dounce homogenizer. This step ruptured cell membranes, leaving a suspension of cell nuclei, cytoplasmic debris and oganelles. Care was taken to execute two gentle passes with the homogenizer, in order to avoid breaki ng the mitochondrial membranes. For each sample, the homogenate was transferred to a SO ml centrifuge tubes, and centrifuged 10 minutes at 800 x g to remove cell debris and nuclei. The supernatant was removed, and EDTA added to a final concentration of 20 mM. to inhibit nuclease activity. The supernatant was centrifuged 20 minutes at 20,000 x g to pellet the mitochondrial fraction. Sodium dodecyl sulfate (SDS) was added to a final concentration of 1%, and the preparation incubated 5 minutes at 37 C. The SDS and proteinaceous debri s were precipitated by centrifuging at 9500 x g for 25 minutes. 5.0 micrograms of ethidium bromide were added per ml of supernatant. 1.1 grams of CsCl/ml were added to the supernatant, and the refractive index was adjusted to 1 .392. Samples were centrifuged 48 hours at 42,000 x g The mtDNA band was then removed 15
16 with an 18 gauge syrynge needle. The ethidium bromide was removed with repeated washes in isoamyl alcohol saturated with TE buffer: 10 mM Tris-HCl, 1 mM EDTA, pH = 8.0. This step minimized the precipitation of cesium chloride. The samples were diluted 2:1 with TE buffer, and 20 micrograms of tarnsfer RNA were added per ml of sample, to minimize loss of mtDNA during precipitation with 2.5 volumes of 957. ethanol. The precipitation was conducted at -20 C overnight. The mtDNA pellets were collected by centrifugation at 12,000 x g for 30 minutes. The pellets were then washed with 707. ethanol, and dried under vacuum. The samples were resuspended in 30 microliters of TE, and stored overnight over chloroform at 4 C. Long term storage was at -20 c. The mtDNA was quantified by fluorometry, and 1 microgram per sample were restricted with 16 to 50 units of the.endonuclease Hind III, which recognizes the nucleotide sequence AAGCTT. The fragments generated from CsCl purified mtDNA were separated by electrophoresis on the same agarose gel with fragments restricted from mtDNA purified by the simplidfied methodology presentd here. The methodology was as described in Mangini et al., (in press): One to two grams of tissue were homogenized on ice by first cutting into very small pieces with an electric knife or razor blade. Frozen tissue was more conveniently minced by grating on nylon screens, as described by Graves et al. (1984). The minced tissue was homogenized in a 50 ml Dounce homogenizer, on ice, in approximately 15 ml of chromatid isolation buffer (CIB): 0.1 M NaCl, 10 mM EDTA, 10 mM mercaptoethanol, 0.5 7. Triton X -100/L and 30 mM Tris-HCl, at pH 8.0. This step ruptured the cells, leaving a suspension from which the
nuclei and aggregated chromatin were removed by centrifugation at 5000 x g for 10 min at 4 C. The centrifugation was repeated until the supernatant was clear and no pellet was formed, which produced a suspension enriched in mtDNA and mitochondria. The mitochondrial membranes were lysed with 1 volume of 2.0 io sodium dodecyl sulfate (final concentration of 1.0 % SDS) for five min at 37 C. This step is especially important if fresh tissue is used, but should not be omitted even with frozen tissue preparations, in order to maximize the extraction of mtDNA from any mitochondrial membranes which may have survived freezing. The lysate was then extracted twice with 2 volumes of phenol:chloroform:isoamyl alcohol 25:24:1, to remove proteins that may interfere with subsequent restrictions. A final extraction with 24 volumes of chloroform to 1 volume of isoamyl alcohol was performed 17 to remove traces of phenol. The aqueous (upper) phase was precipitated with 2.5 volumes of 95 % ethanol or one volume of isopropanol for 2 h at -20 C. DNA was then pelleted by centrifugation at 12,000 x g for 30 min at 4 C, dried in vacuo, and resuspended in 30 microliters of TE buffer: 10 mM Tris-HCl, 1 mM EDTA, pH 8 .0. The DNA thus purified was stored overnight over chloroform at 4 C. Longer storage was at -20 c. The DNA was digested separately with the restriction endonuclease Hind III. Restrictions were conducted for 4 to 8 hours at 37 C. An excess of endonuclease was used, of 16 to 50 units of endonuclease per microgram of DNA extract. This amount maximized cutting efficiency. The restricted DNA was separated by electrophoresis (40 Volts, 64 mA), on 1 io agarose gels run in TAE
electrophoresis buffer: 0.4 M Tris-HCl, 0.2 M glacial acetic acid, 10 mM EDTA, pH 8.0 (Maniatis et al., 1982). Both the gel and electrophoresis buffer were stained with 2.5 micrograms of ethidium bromide per ml of gel and buffer. 18 The gel was photographed under long wave ultraviolet illumination, then washed in 0.25 M HCl for 15 minutes. This step facilitates the subsequent transfer of the DNA fragments to filters, by breaking the DNA into smaller fragments. The gel was then washed two times, 15 minutes each, in 0 5 M NaOH, 1.5 M NaCl, to denature the DNA (Lansman et al., 1981). Two more washes, of 15 minutes each were conducted in 0.5 M Trisma base, 1.5 M NaCl, pH = 8.0, to neutralize th NaOH prior to tranfering the gel to nitrocellulose filters. The section of the gel below the sample origin was removed, and the DNA was then transferred to DNA binding filters by the method of Southern (1979). The filters used were nitrocellulose filter, (from BRL). The blotting buffer was lOX SSC, prepared as per Maniatis (1982). After the Southern blotting proceeded overnight, the filters were air dried for one half hour unde a heat lamp, then baked in a vacuum oven for two hours at 80 C. The restricted mtDNA was identified by hybridization to radiolabeled heterologous mtDNA probes. The probes consisted of cloned plasmid DNA which contains the vector pBR322 and the complete bovine mtDNA genome (Hauswirth and Laipis, 1982) Probe DNA was labeled with 32P dCTP by nick translation (Maniatis et al., 1982). Hybridization conditions were a modification of those used by Maniatis et al., (1982): Filters were prehybridized one hour at 65 oC and hybridized overnight at the same temperature with 1 0 x 10 cpm of radiolabeled
19 probe DNA per ml of hybridization buffer. Autoradiographs were obtained by exposing the filters to Kodak X-Omat AR X-ray film. These experiments allowed direct comparison of the mtDNA restriction fragments obtained by cesium chloride purification versus by the simplified purification method presented here, and characterized the molecular weight of the C. undecimalis mtDNA genome. Cloning of Snook mtDNA The undecimalis mtDNA was cloned to produce a high resolution homologous probe. For cloning purposes, cellular DNA was isolated from fresh tissue according to the above protocol. The DNA was restricted with H ind III, the same enzyme tha was used in characterizing the C. undecimalis mtDNA. The linearized molecule or fragments thus generated were separated by electrophoresis on a 1 7. agarose gel in TAE buffer. The three DNA bands that corresponded to the snook mtDNA molecular weights were visualized by ethidium bromide staining (2.5 micrograms/ml of gel and buffer), cut from the gel, and the gel matrix removed by freeze extraction via repeated freezing and thawing (Tautz and Renzi, 1983) followed by centrifugation through a siliconized fiberglass wool filter. The fragments were combined with the DNA from the bacterial plasmid pT7-2 (Gene ScribeTM U. S. Biochemical Corp. Cleveland, Ohio). The plasmid DNA was also.restricted with Hind III, the same enzyme as used for the undecimalis DNA . This step generated compatible ends between the DNA of the plasmid and the snook mtDNA. The plasmid DNA was treated with calf intestinal alkaline phosphatase as per Maniatis et al., (1982). This step removed the S'phosphate group from both ends of the
linearized plasmid DNA. Therefore, neither strand can form a phosphodiester bond, and recircularization of the vector DNA is prevented. The snook mtDNA fragments contained the 5' terminal phosphate group, and in the presence of T4 DNA ligase the snook mtDNA was ligated to the dephosphorylated plasmid DNA. The ligation buffer used was modified from Maniatis et al. (1982): 0.05 M Tris (pH 7.4), 0.01 M MgC12, 0 .01 M Dithiothreitol, 1 mM ATP, 50 micrograms/ml bovine serum albumin, 5 Weiss units of T4 DNA ligase. 20 100 ng of pT7-2 vector DNA and 100 ng of snook fragment mtDNA were used per reaction. The ligation resulted in open circular DNA with 2 nicks, referred to as chimeric molecules containing the plasmid and the mitochondrial DNA fragments. These circular molecules transform much more efficiently than linearized molecules, so that most of the transformants contained recombinant plasmids. Three snook mtDNA fragments were to be cloned. Therefore, separate ligation reactions were performed, using the three fragments isolated by gel electrophoresi s The reactions were allowed to proceed overnight at 12 C. These molecules were used to transform Escherichia coli cells rendered competent by calcium chloride treatment (Maniati s et al., 1982). The protocol used differed from that of maniatis by 2 modif ications: 1. Doubling the CaC12 concentration, from 0.05 M to 0.10 M. 2. The CaCl solution was made fresh each time. (Dr. Lee Weber, Personal communication). Three controls were run, together with the three cloning experiments: A phosphatase free positive control for vector DNA recircularization. A phosphatase treated, negative control, for
21 linearized plasmid uptake by coli. An uncut plasmid uptake experiment, to control for transformation. Three experiments were run, to insert each of the three snook DNA fragments. Transformed cells were selected by ampicillin resistance, which indicates plasmid uptake. This procedure may result in the cloning of nuclear DNA as well as mtDNA, due to the comigration of the residual nuclear DNA in the mitochondrial extract. Confirmation of those plasmids containing mtDNA was obtained by preparing plasmid DNA from the transformed cells by alkaline lysis (Birnboin and Doly, 1979) and hybridization to bovine mtDNA probes. The cloning experiment was repeated, because only two of the three snook mtDNA fragments were cloned in the first attempt: One of the fragments cloned consisted of nuclear DNA. Final confirmation of the identity of the obtained C. undecimalis clones, and of their applicability as molecular probes, required the demonstration that they hybridize to the same bands identified by the bovine mtDNA probes. This experiment also the relative sensitivity of the homologous probes vs. the bovine probes, by determining if stronger hybridization signals and smaller mtDNA fragments are obtained with the homologous probes. Therefore, mtDNAs from five f undecimalis specimens were digested with four restriction enzymes that recognize six base sequences: Bam HI, Hind III, Pst and Pvu II. The resultant fragments were separated by electrophoresis and identically transferred to two nitrocellulose filters by a Southern sandwich blot transfer (Maniatis et al. 1982). The DNA on one filter was hybridized to a radiolabeled probe consisting of the complete bovine mitochondrial genome The DNA on the other
filter was hybridized to the radiolabeled C. undecimalis cloned mtDNA. Determination of mtDNA Population Polymorphisms 22 Upon completion of the cloning protocol, the population screening was initiated. MtDNA was extracted and digested with the following six-base-pair cutter restriction endonucleases: Hind III, Bgl I, Pst I, Eco RI, Bam HI, Pvu II, Hae II, Hind II and Xba I. The five-base-pair cutters used were Ava II and Hinf I. One four-base-pair cutter was used, Hae III. Digestion reactions were conducted at the manufacturer's specifications (BRL) for 4 to 8 hours. 16 to 50 units of restriction endonucleases were used per microgram of DNA. The high ratio of endonuclease to snook DNA was used to maximize cutting efficiency. Internal controls for the adequeacy of the restriction conditions were conducted by digesting lambda bacteriophage DNA with H i nd III. This served the dual purpose as control for restriction reaction conditions, and as molecular weight markers in the electrophoresis gels. The cutting efficiency of the other enzymes was evaluated by digestion of DNA from two other snook species: C pectinatus and f. parallelus. 9 specimens of C undecimalis from the East coast and twelve from the West coast were examined with Hind III. The amount of DNA extracted and specimen availability determined the number of specimens examined with the other restriction endonucleases. The adequate number of specimens examined must be determined empirically. However, previous workers have used as few as 6 specimens per site for studying infraspecific mtDNA variation between populations of vertebrate and
invertebrate species (Avise et al., 1984; Lansman et al, 1981; Gonzalez-Villasenor and Powers, 1985; Birt and Green, 1986). 23
24 RESULTS Application of the Protocol The purified cellular DNA is enriched in mtDNA and is sufficiently pure for direct quantification and restriction. The mtDNA restriction fragments obtained by this protocol (Figure 1, lanes 4 and 5) were compared with results obtained by cesium chloride purification of mtDNA (Figure 1, lanes 2 and 3). The calculated mtDNA fragment and genome sizes are given in Table 1. Some contamination by nuclear DNA occurs in the mtDNA purified by the cytoplasm enrichment method. Such contamination is inevitable, since the nuclear membranes break during freezing, and some membrane breakage occurs during homogenization of fresh tissue. The direct screening of total DNA with radiolabeled probes also yielded adequate results for population comparisons, (Figure 1, lane 5), but those results required 31 times the amount of DNA than was used in the other three lanes. The experiment represented by Figure 1 also demonstrates a potential genome and fragment heteroplasmy between the mtDNA obtained from the liver and from the kidney by cesium chloride extraction. The mtDNA genome size calculated from the CsCl purified mtDNA shows a 3.57. size difference between the liver and kidney mtDNA. The values for the mtDNA genome size and restriction fragments obtained from the liver,
23130 9416 6557 4373 2322 2027 580 Figure 1. 25 1 2 3 4 5 6 Hind III aeaerated restriction pattern of c undecimalia Lanee 1 and 6: Molecular weight standard of Hind III reatricted lambda bacteriophage DNA. Lane 2: CSC:i purified liver mtDNA. Lane 3: CsCl purified kidney mtDNA. Lane 4: liver extracted cellular DNA, enriched for mtDNA by the protocol. Lane 5: nuclear DNA fraction as by the protocol from the liver. 0.5 micrograms of per lane were used, except in lane 5, where 15.5 micrograms were used. the mtDNA bands were separated on lr. agarose, and identified by radiolabelled bovine mtDNA as molecular probes.
26 both by CsCl and total celluar DNA purification, are more uniform than the values obtained from the liver and kidney be CsCl isolation. Their maximum variation is 1 .87o. Such size variations may be due to either tissue specific heteroplasmy, or to tissue specific compounds that alter the migration rate of the purified mtDNA (see discussion). If the kidney derived mtDNA restriction data are omitted from genome calculations, the obtained molecular sizes are contained in the range of the mtDNA fragments and genome size established with data from all four samples. Therefore, the latter are used as the characterized values for the C undecimalis mtDNA. Table 1. Hind III generated restriction pattern of undecimalis mtDNA. Lanes 1 and 6: Molecular standards of Hind III restricted lambda bacteriophage DNA. Lane 2: CsCl purified liver mtDNA. Lane 3: CsCl purified kidney mtDNA. Lane 4: liver extracted cellular DNA, enriched in mtDNA by the above protocol. Lane 5: nuclear DNA fraction as extracted by the protocol. lane 2 3 4 5 Fragment mol. wt. 7.54 7 .78 7.31 7.47 6.20 6.53 6 10 6.10 3 .79 3 83 3 69 3 69 Genome mol. wt. 17.53 18 14 17.22 17.23 Mean molecular weights, Mean molecular weights, all four samples minus lane three 7 51 +0.17 7.44 +0.10 6 22 +0.18 6.13 +0.01 3.75 +0.06 3. 72 +0 .01 17.49 +0.41 17.33 +0.14
MtDNA Genome Size and Restriction Fragments in Three Centropomus Species Centropomus undecimalis The protocol allowed the identification of three mtDNA fragments generated by Hind III restricti on (Table 1 and Figure 1). These fragments consisted of 3.75 +0.06, 6.22 +0.18, and 7.51 +0.17 kilobase pairs (kbp), as determined by regression analysis, using Hind III restricted bacteriophage Lambda DNA as molecular weight standards (where Y = -44.4319 X + 202.4895, R squared = 0.9932, R = 0.9966). This suggests that the snook mtDNA genome is ca. 17.49 +0.41 kbp (table 1). Centropomus parallelus The mtDNA genome size was estimated by restriction with Hind III. 27 Two specimens were secured, which yielded the restriction fragments presented in Figure 2 and Table 3). The weight standards were Lambda DNA and undecimalis brain extracted mtDNA, both restricted with Hind III. The lambda DNA restriction fragments generated a molecular weight regression equation of Y = 35 0500 X + 166.670. R squared = 0.9910, R = 0.9950. The liver extracted mtDNA of undecimalis was previously determined to have a total molecular weight of 17,49 +0.41 kilobases. In this preparation, the brain extracted mtDNA of C. undecimalis yielded a higher mtDNA genome molecular weight. Therefore, the data obtained were presented in two forms in Table 3. The raw data column (a) represents the molecular weight of the C. parallelus brain extracted mtDNA, as obtained from the
1 2 3 4 23130 9416 6557 4373 2322 2027 Figure 2. MtDNA of Centropomus parallelus. Lane one: molecular weight standards. Lane 2: 5.0 micrograms of DNA extracted from f parallelus brain tissue. Lane 3: 0.5 micrograms of DNA from f parallelus brain tissue. Lane 4: 1.0 micrograms of DNA from f undecimalis brain tissue. All lanes were restricted with Hind III, separated on 1 % agarose, and mtDNA identified by radiolabeled bovine mtDNA. 28
Table 2. Hind III restriction of C. undecimalis used, with Lambda DNA, as internal standards for parallelus. Molecular weights are in kilobases. Spec imen 1 2 (a) (b) (a) (b) raw data X 0.914 raw data X 0.914 Molecular weight 7.85 7 18 8.34 7.63 6.45 5.90 7 24 6.62 4.00 3.66 4.40 4.02 Total mol. wt. 18.30 16.73 19.98 18.27 Mean values: Raw data: 19 14 +0.84 (4.4 % internal variation) Corrected data: 17.49 +0 .77 (5.4% internal variation) Table 3. Hind III restriction of c. parallelus. Specimen 1 2 (a) (b) (a) (b) raw data X 0 914 raw data X 0 914 Molecular weight 8.69 7 92 undetectable 6.69 6 12 6.60 6.03 3 .43 3.19 3.42 3.13 2.40 2.19 2.38 2.18 Total mol. wt. 21.27 19.45 >12.40 >11.34 Mean values: 29 Raw data: 8.69 +-na, 6 65 +0.05, 3 .46 +0.03, 2.39 +0.01 = 21.27 Corrected data: 7.95 +-na, 6.08 +0 .05, 3 16 +0.03, 2.19 +0.01 = 19 38 +0.09
lambda DNA molecular standards regression equation. The (b) column represents the same data corrected for the potential size artifact. These corrections are not necessary for comparisons between the three species, but are included to calculate the equivalent molecualr weight of the liver extracted mtDNA of the other species of snook. 30 The correction coefficient was obtained by dividing the liver characterized molecular weight of C. undecimalis mtDNA = 17.49 into the molecular weight obtained in this preparation, 19.14 +0.84, n = 2, = +4.4%). The resultant cofficient equaled 0.914. Therefore, the mtDNA genome of the two species is compared by inspection of the raw data and of the data corrected for variation between gels. Raw data: C. parallelus 21.27 I C. undecimalis 19.14 = 1.1085, which yields a 10.85 7o size difference. Compensated data: E 19.38 I C. u. 17.49 = 1.1105, this method yields an 11.05% size difference. The significant results of this experiment are: one, the resolution of four Hind III restriction fragments in parallelus, versus three in C. undecimalis. Two, the larger molecular weight obtained for the mtDNA of parallelus. This larger weight is not an artifact, because it is significantly larger before and after applying the correction coefficient to the data. Furthermore, the size differences are more than twice the variation obtained within the standard and experimental samples. DNA fron parallelus was also digested with Bam HI and Eco RI, and electrophoretically separated alongside undecimalis DNA also digested with those two enzymes. Eco RI recognized two CAATTC sites, by producing two fragments
31 (Figure 3): one of approximately 11.5 kbp, the other of 7.83 kbp. These values are consistent with the previoulsy determined molecular weight of the mtDNA molecule of f. parallelus. Bam HI produced one linearized molecule, as it did with C. undecimalis (Figure 4). The migration distance of this fragment indicated a molecualar weight in the nonlinear region of the standard DNA. However, that migration distance is consistent with that expected of a linearized mtDNA molecule of circa 19 kbp. Therefore, Bam HI recognized only one GGATCC site on the mtDNA molecule. The results are summarized in Table 5. Centropomus pectinatus Three juvenile fish were obtained, ranging from East Florida (Sebastian River). Their sizes ranged from 3.5 em to 5.2 em standard length. The gastrointestinal tract and the gills were removed, and mtDNA was extracted by processing the entire specimen. Molecular weight standards were obtained with Hind III restricted lambda DNA and C. undecimalis mtDNA (Fig. 5). The lambda DNA generated a regression equation of Y = -45.632 X+ 210.188, R2 = 0.989, R = 0.990. In this preparation, The C. undecimalis mtDNA that was used as an internal standard was also purified from entire, eviscerated juvenile specimens. That mtDNA genome size was estimated to be 15.60 +-0.29, n=10. Therefore, the correction coefficient, from comparison with the previously characterized C. undecimalis mtDNA, was 17.49 I 15.60 = 1.121. The untreated and corrected data are presented in Table 4. The size difference between the mtDNA genome size of two species species is estimated by
23130 9416 6557 4373 2322 2027 E E IR IR IR 6 CPa T T T T 12 Figure 3. Eco RI restriction fragments from 4 populations, 2 species of snook. Lanes 6 and 12: molecular weight standards. E= Everglades, IR= Indian River, T= Tampa Bay. CPa= Centropo mus parallelus. 2.0 micrograms of DNA were used per sample lane. The fragments were separated on 1 % agarose and identified with radiolabeled bovine mtDNA. 32
23130 9416 6557 4373 2322 2027 Figure 4. 33 1 E E IR IR IR 7 CPa T T T T 13 Bam HI restriction of 3 populations and two species of FlOrida snook. lanes 1, 7, 13: molecular weight standards. E= Everglades, IR= Indian River, parallelus, T= Tampa Bay. 2 0 raicrograms of DUA v1ere used per sample lane.
Table 4. Hind III restriction fragments of pectinatus. Molecular weights are in kilobases. Specimen 1 2 3 (a) (b) (a) (b) (a) (b) raw data X 1.121 raw data X 1.121 raw data X 1.121 M. w. 8.89 9.97 8.03 9.00 7.64 8.57 4.78 5.36 5.50 6.17 5.64 6.32 2.52 2.83 2.22 2.49 2.39 2.70 1.64 1.84 1.52 1. 70 1.60 1.79 Total 17.83 20.00 17.27 19.36 17.27 19.38 Mean molecular weight raw data: 8.19 +0.52, 5.31 +0.38, 2.38 +0.12, 1.59 +0.05 = 17.46 +0.27 Mean molecular weights of corrected data: 9.18 +0.59, 5.95 +0.42, 2.67 +0.14, 1.77, 0.05 = 19.58 +0.30 Table 5. Restriction endonucleases, (recognition sequneces), total number of fragments (and fragment patterns) revealed by each enzyme. Centropomus: Hind III Eco RI Bam HI (AAGCTT) (GAATTC) (GGATCC) undecimalis 3 (1) 1 (1) 3 (2) parallelus 4 (1) 2 (1) 1 (1) pectinatus 4 (1) n. a. n. a. 34
23130 9416 6557 4373 2322 2027 580 1 18 Figure 5. Tissue specific mtDNA migration rates. Lanes 1 to 6: East coast juvenile extracted mtDNA. Lane 7: molecular weight standards. Lanes 8 to 14: West coast juvenile extracted mtDNA. Lane 18: adult, brain extracted mtDNA. (All from f undecimalis). Lanes 15, 16, 17: C. pectinatus mtDNA, extracted from juveniles. All specimen lanes contained 1 microgram DNA, were restricted with Hind III, and separated on 1% agarose gel. mtDNA fragments were identified by radiolabeled bovine mtDNA. 35
36 comparison of the raw and compensated data. The raw data difference is 17.46 I 15.60 = 1.1192, or 11.92 % The corrected data difference is 19.58 I 17.49 = 1.1195, or 11.95%. The significant results of this experiment are: one, the determination of four Hind III restriction fragments from the mtDNA of each f. pectinatus. Two, the consistently higher molecular weights of the mtDNA of f pectinatus. The proof that the size variation is not an artifact is given by direct, side by side electrophoresis of the mtDNA obtained from juveniles of the two species. The mtDNA of juvenile f undecimalis, collected at the same time and in the same location as the juvenile f. pectinatus, reveal both a size and restriction difference in the untreated and corrected data. The three restriction generated genotypes are presented in Table 5. A phenogram constructed with the mtDNA genotypes of the three pecies is presented in Figure 6. Figure 6a presents a phylogenetic relationship based on the number of H i nd III restriction sites shared and the total mtDNA genome size. Figure 6b presents an alternate phenotype, based on the shared mtDNA fragments, rounded to the nearest kilobase.
Figure 6a 6 Figure 6b f. undecimalis 17.49 kb, 3 Hind III sites parallelus 19.38 kb, 4 Hind III sites pectinatus 19.58 kb, 4 Hind III sites C. undecimalis 8 6 4 f. parallelus 8 6 3 2 C. pectinatus 9 6 3 2. 37 Figure 6. Figure 6a: Estimated phylogenetic relationships based on the number of Hind III mtDNA restriction sites and the total genome molecular weights. Fig. 6b. Estimated phylogenetic relationships between the three species, based on shared mtDNA fragment size (from Tables 1, 3 and 4), calculated to the nearest kilobase.
Selection of Optimal Tissue for mtDNA Purification and Population Screening. Total DNA yields were consistent between tissue samples. We obtained a total DNA yield of about 0.9 micrograms DNA/gr. tissue. However, the mtDNA fraction varied as a function of the tissue type used in the extraction protocol. The estimated mtDNA components were highest from brain DNA extracts, lower from heart and skeletal muscle, and lowest from liver and kidney. The different yields may reflect not only the different mtDNA content of each tissue, but 38 also the differential tissue content of storage polysaccharides and RNA's, and their effect on the extraction of mtDNA. Brain tissue yielded mtDNA of sufficient purity and quantity to allow visualization of restriction generated mtDNA bands on 1.07. agarose gels stained with 2.5 micrograms of ethidium bromide per ml of gel and electrophoresis buffer. Figure 7 shows the mtDNA bands obtained from a digest of 150 nanograms of total cellular DNA extracted from brain and muscle tissues. Cloning of Snook mtDNA The cloning experiment resulted in the insertion, into the Hind III restriction site of the polylinker sequence of two pT7-2 plasmids, of two DNA fragments. These fragments had molecular weights corresponding to those of the snook mtDNA fragments generated by restriction with Hind III (Figure 8). The identity of these clones was confirmed by transfering the cloned fragments from the agarose gel to nitrocellulose filters, and hybridizing the filters to radiolabelled
23130 9416 6557 4373 2322 2027 580 l 39 2 4 5 Figure 1. f undectmalis restriction fragments visualized on agarose gel with ethidium bromide. Lane 1: DNA molecular weight standards. Lanes 2 and 3: 150 nan-ograms/lane of muscle extracted DNA. Lanee 4 and 5: 150 ftanograms/lane of brain extracted DNA. The mtDNA bands in lanes 4 and 5 are visible without the use of radiolabelled Molecular probee. All lanes were produced with Hind III restriction. The mtDNA bands in lanes 2 obscured by a background of nuclear DNA.
40 1 2 3 4 5 6 7 8 9 10 11 12 23130 9416 6557 4373 2322 2027 Figure 8. Cloned DNA inserts, visualized on 1% agarose gel stained with ethidium bromide. Lanes 6 and 12: Hind restricted lambda DNA. Lanes 1 and 3: 7.51 kbp---inserts. Lane 7: 6.22 kbp insert. Lanes 8, 10, 11: 3.75 kbp inserts. All lanes except standard lanes show the 2.8 kbp pT7-2 vector DNA.
41 bovine mtDNA probes. This substantiated the identity of two independent clones of the 3.75 kbp fragment, and two clones of the 7.51 kbp fragment. Final confirmation of the identity of the clones, and of their applicability as molecular probes, required the demonstration that they hybridize to the same bands identified by the bovine mtDNA probes. Conducting such an experiment also gave an indication of the relative sensitivity of homologous and heterologous molecular probes. Therefore, snook mtDNA was digested with Bam HI, Hind III, Pst I, and Pvu II. The fragments were separated by electrophoresis and identically transfered to two nitrocellulose filters by a Southern sandwich blot transfer (Maniatis et al., 1982). The DNA on one filter was hybridized to a radiolabelled probe consisting of the complete bovine mitochondrial genome. The DNA on the other filter was hybridized to the radiolabelled 3.75 and 7.51 kbp subunits of the snook mitochondrial genome. Results of this experiment are shown in Figure 9. Several points are worthy of note. First, our cloned probe does hybridize to the 3.75 and 7.51 kbp Hind III generated fragments of snook mtDNA (Figure 9b). Second, the signal intensity of that hybridization is consistently greater than that generated by the bovine probe. The autoradiograph in Figure 9b, which used the snook probe, was achieved with half the exposure time as that in Figure 9a, which used the bovine probe. Finally, the use of the homologous probe reveals a Pvu II fragment of ca. 0.67 +0.01 kbp (Figure 9b, lanes 15-18). that does not hybridize to the bovine probe. These results demonstrate that the use of cloned snook probes increased the sensitivity of the
23130 9416 6557 4373 2322 2027 580 23130 9416 6557 4373 2322 2027 580 2 3 Figure 9a Figure b Figure 9, Figure 9a: Filter probed with the two cloned snook mtDNA fragments. Figure 9b: Filter probed with the entire bovine mtDNA genome. Lanes 1, 10 and 19: Molecular weight standards. Lanes 2 and 3: Hind III restricted East coast samples, 4 and 5, t-lest coast samples. Lanes 6 and 7: Pst I restricted East coast samples, 8 and 9: West coas. t samples. 11 to 14, Bam HI restrictions, 15 to 18, Pvu II restrictions. 42
43 screening technique. The methodology presented here allows the use of frozen tissue in the rapid purification of mtDNA of sufficient purity for direct digestion with endonucleases. It permits the cloning of the mtDNA fragments obtained, and demonstrates the use of the cloned fragments as homologous probes with higher sensitivity than heterologous probes. The results demonstrate that the use of heterologous probes will allow the identification of fish mtDNA sequences, which permits obtaininig clones of homologous fragments. In turn, these clones, when employed as probes, increase the sensitivity of these methods. Of special significance is the accessibility of the technique and the convenience of isolating mtDNA from either fresh or frozen tissues. Screening of East and West Florida Populations of C. undecimalis Twelve restriction endonucleases were used in screening the populations for geographically correlated mtDNA restriction polymorphisms. These enzymes and the nucleotide sequences they recognized were: Hind III (AAGCTT), Eco RI (AorG AorG AT Tore TorC), Pst I (CTGCAG), Xba I (TCTAGA), Pvu II (CAGCTG), Bam HI (GGATCC), Bgl I (GCCNNNNNGGC), Ava II (GG AorT CC), Hae II (PuGCGCPy), Hind II (GTPyPuAC), Hinf I (GANTC), Hae III (GGCC). The mtDNA restriction fragments thus generated are reported in Table 6. Hind III Restriction Analysis Nine specimens from the east coast, and twelve from the west coast of Florida were compared by the above protocol. Results are
44 Table 6. Summary of the mtDNA restriction patterns of East and West coast undecimalis as determined by restriction with twelve endonucleases. Molecular weights are in kilobases. Numbers in parenthesis indicate individuals sampled. Hind III Pst I West Coast East Coast West Coast East Coast 7.61 0 42 7 72 0.56 117.49 117.49 5.91 0 .37 5.99 0.58 3.20 + 0.42 3.13 + 0.59 16.72 1.21(12) 16 83 1.58(9) 17 .49(7) 17.49(4) Hae III Hinfi West Coast East Coast West Coast East Coast 1.18 1.27 1. 27 1.11 1.20 1.02 1.02 0 93 0.93 0.86 0.78 0.78 0. 77 0. 77 0 65 0.65 0.66 0.66 0.58 0.61 0.55 0.55 0.56 0.56 0.52 0.51 0.51 0.49 0.49 0.43 0.43 0.40 0.40 0.36 0.36 0 35 0.35 17.49(2) 17.49(2) 17.49(2) 17.49(2) 3 3 Ava II Hind II West Coast East Coast West Coast East Coast 7 0 7.0 5.0 5.0 3 0 3.0 3.5 3.5 1.5 1.5 2.0 2.0 >11.5(2) >11. 5( 2) >10.5(2) >10.5(2) 1In nonlinear region of standard curve, estimated by inspection. 2In nonlinear region of standard curve, estimated by subtracting the lighter fragments. 3Molecular weights of fragments approximated by inspection, due to incomplete molecular standard lanes.
4-5 Table 6. (Continued) Summary of the mtDNA restriction patterns of East and West coast undecimalis as determined by restriction with twelve endonucleases. Molecular weights are in kilobases. Number in parenthesis indicate individuals sampled. Eco RI XBA I West Coast 117.49 17.49(9) East Coast 17.49(4) West Coast 17.49 17.49(2) Bam HI West Coast Tampa Bay Everglades West Coast 216.73 17.49 214.33 3.16.05 17.49(5) 17.49(2) Pvu II 0 .66 + 0.01 17.49(2) East Coast 216.71 0.68.00 17.49(2) East Coast Indian R. Indian R. 14.21 17.49 3.28.12 17.49(2) 17.49(2) West Coast 17.49>x>9.41 37.0 17.49(2) Bli West Coast East Coast 211.84 211.97 4.36 0.19 4 .21 0.18 1.29 + 0.08 1. 31 0. 09 17.49(10) 17.49(7) East Coast 17.49 17.49(2) East Coast 17.49>x9.41 37.0 17.49(2) 1In nonlinear region of standard curve, estimated by inspection. 2In nonlinear region of standard curve, estimated by subtracting the lighter fragments. 3Molecular weights of fragments approximated by inspection, due to incomplete molecular standard lanes.
Table 7. Hind III restriction screening of mtDNA polymorphisms between East coast, West coast, adult brain extracted mtDNA and juvenile extracted mtDNA. Molecular weights are reported in kilobases. East coast West coast juvenile specimens, n = 5 juvenile specimens, n = 5 7.25 +0.15 5.99 +0.58 3 13 +0.59 16. 83 +-1. 58 7.63 +0 47 5.81 +0.34 3.02 +0.31 16.46 +0.95 Adult, brain mtDNA, n = 4 8.30 +0.27 Adult, brain mtDNA, n= 3 8.28 +0.00 6.55 +0.40 3.59 +0.63 18.45 +0.91 Pooled samples, n = 9 7.72 +0.56 5.99 +0.58 3.13 +0.59 16.83 +-1.58 Tissue specific mtDNA Pooled brain East and West, n = 7 8.29 +0.20 6.38 +0.39 3.35 +0.57 18.03 +0 89 6.14 +0.19 3.04 +0.26 17.25 +0.16 Pooled samples, n = 12 7.61 +0 42 Liver and kidney Std. mtDNA, n = 4 7.52 +0.17 6.19 +0.154 3.76 +0 .02 17.49 +0.41 5.91 +0.37 3.20 +0 42 16.72 +-1.21 Pooled juveniles East and West, n = 10 7.27 +0.14 5. 53 +0 .11 2.81 +0.09 15 .60 +0.30 46
47 presented in Table 7. The molecular weights of the fragments are reported without compensation for tissue type variation. Direct comparisons are permitted because tissue types, age classes and specimes from both coasts were screened on the same gels, which also contained Lambda DNA molecular weight standards. These side by side comparisons obviated any potential artifacts from different electrophoretic runs. No variation was observed in the restriction patterns obtained from the two populations. The variation between specimens of all populations was within the calculated measurement error. The samples used in this study do not represent a statistically normal population. They were collected over a period of several months, and DNA was isolated from large adult and post planktonic juveniles. Therefore, the restriction fragment data was analyzed using Mann Whitney nonparametric tests (Zar, 1984). The analysis indicate the molecular size of the brain extracted mtDNA differs from that of the mtDNA extracted from entire juvenile specimens, at alpha (2) = 0.05. This variation was caused by differences in the two heavier mtDNA fragments resolved. The variation was observed in molecular size, but not in the number of restriction restriction fragments produced (Figure 4). A minor variation in fragment size was also observed in the CsCl purified liver and kidney mtDNA (Table 1 and Figure 1). It is concluded that the apparent size variation is a function of tissue specific organic molecules that coprecipitate with the DNA during purification. It is therefore necessary to compare only restriction patterns from mtDNA isolated from the same tissues. Furthermore, it is recommended
that, as conducted in this study, specimens be compared on the same gel, using DNA extracted from thesame tissue type and age class. 48 Eco RI, Pst I and Xba I produced only one band, located below the linearly migrating region of the DNA standards. These data imply that those enzymes recognized only one restriction site, producing one cut, which yielded a linearized molecule. The migration distance of theird digestion products is consistent with the expected position of a linearized mtDNA molecule of 17.49 kbp (figure 3, 9 and 10). Although the fragment's molecular weight can not be acurately determined from the autoradiographs, two results are obained from inspection of the data. First, both enzymes generated bands that migrated faster than the slowest fragments yet obtained (Pvu II digestion fig. 9). Second, the fragments from both the East and West coast specimens migrated at the same rate. It is possible that the apparently faster migration rates are a function of the nonlinear migration of the larger molecules. Another alternative is that several small fragments were produced, then lost from the gel in the migrating buffer front. Pvu II: Two fragments were resolved, a 0.67 kbp band and the heaviest band yet obtained (figure 9). This band is assumed to be approximately 16.78 kbp, or the balance of the mtDNA genome size. The lighter band showed no difference between the two populations. Ava II: Three fragments were obtained, of approximately 7.0, 3 .0, and 1.5 kbp, without any interpopulational differences (table 6). Hind II: Three fragments were obtained, of approximately 5 .0, 3.5, and 2.0 kbp (Table 6). No interpopulational differences were resolved. Hae II: Two fragments were obtained, of approximatley 7.0 and
23130 9416 6557 4373 2322 2027 580 23130 9416 6557 4373 2322 2027 580 1 2 3 4 5 6 7 8 9 10 Figure 10. Southern sandwich blot of 6, 5, and 4 base pair digested C. undecimalis mtDNA. Lanes 1, 10, 18: molecular-weight standards. Lanes 2,3: E.coast Eco RI 4,5: W. coast Eco RI. 6,7: E. I. I. 11,12: E. coast Hinf I. 13, 14: W. coast Hinf I. 15, 16: E. coast Hae .III 17: W. coast Hae III Top: fish ntDNA molecular probe. Bottom, bovi11e probe. 49
11.0kbp, without any interpopulational differences (Table 6). Bam HI: Ten specimens were examined (figures 4 and 9). Four Tampa Bay specimens and two East coast specimens yielded only one restriction band of circa 17.49 kbp band. A subset consisting of two West coast and two East coast specimens yielded other restriction fragments (Figure 4). The Everglades population (Fakahatchee bight, n = 2) and an East coast subpopulation (Indian River, n = 2) which produced a large band, of circa 14.30 kbp, and a ligher band of circa 3 20 kbp Four Tampa Bay specimens (Manatee River) run on the same gel produced only one band (ca. 17.49 kb). Note that the variation within populations is equal to that between populations. Bgl I: Pooling the East coast specimens, n = 7, and the West coast specimens, n = 10, yielded three restriction fragments, none significantly different between the two populations. Typical comparisons of East coast and West coast samples are presented in figure 10 and 11. Hinf I: The restricted DNAs were probed with both the snook and the bovine molecular probes, and the resultant fragments presented in table 6 and figure 10. 10 fragments were obtained from the East Coast population, with sizes ranging from 1.27 to 0.35 kb. The West coast samples yielded 9 fragments, ranging from 1.27 to 0.35 kbp. Of these, 8 were of homologous molecular weights. The West coast samples resolved one geographically unique fragment, while the East coast set produced 2 geographically unique fragments. Hae III: 10 fragments were resolved for the East coast, between 1.18 and 0.36 kb. (Table 6 and F igure 10). 8 fragments were 50
23130 4373 2322 2027 580 51 Figure 11. restriction of 3 populatons of f. undecimalis. Lanes 1, 7, 13: molecular weight standards. E= Everglades IR= Indiar River. T= Tampa Bay. 0.5 micrograms of DNA perlane, separated on 1% agarose gel.
resolved for the West Coast, between 1.02 and 0.36 kb. 7 of these fragments were considered of homologous molecular weights between the two populations. Therefore, the East coast population had 3 distinct fragments, and the West coast one distinct fragment. A summary of the resolved mtDNA genotypes is presented in Table 52 8. The most common genotypes are designated with the letter C. Polymorphic restriction fragments are represented by a different letter for the same enzyme.
Table 8. MtDNA genotypes, collection locales and sample size of the Centropomus undecimalis populations studied. MtDNA genotype 1 2 3 Letter designation, from left to right, refer to digestion profiles with: Hind III, Pst I, Eco RI, Xba I, Pvu II, Ava II, Hind II, Hae II, Bam HI, Bgl I, Hinf I and Hae III. composite letter No. scored collection sample description restriction sites locale size cccccccccccc 36 Tampa 10 CCCCCCCCBCCC 37 Everglades 2 ccccccccBcnn 40 East Coast 9 53
54 DISCUSSION AND CONCLUSIONS Application of the Protocol Comparisons of the values obtained for the snook mtDNA genome size and endonucleases generated fragments demonstrated consistent results between the current protocol and the CsCl method. The only variation observed between the two protocols was that obtained when different tissues were used as sources of mtDNA. It is significant that less variation was observed between the mtDNA genome sizes and restriction fragments obtained from liver tissue by the two techniques than by the values obtained by CsCl purification of mtDNA from the liver and kidney. The total cellular DNA extracted by the protocol is enriched in mtDNA and is sufficiently pure for direct quantification and restriction. Contamination by nuclear DNA is inevitable, since the nuclear memebranes break during freezing, and some membrane breakage occurs during homogenization of fresh tissue. However, the DNA extract is enriched by a factor of up to 31 (Figure 1, lane 5). The total DNA yield of the extractions was consistent between samples. The total cellular DNA yields were circa 0 9 micrograms DNA per gram of tissue. However, the mtDNA fraction varied as a function of the tissue used. The estimated mtDNA fractions were highest from brain DNA
55 extracts, lower from heart and skeletal muscle, and lowest from liver and kidney. The different yield may reflect not only the differential mtDNA content of each tissue, but also the differential tissue content of storage polysaccharides and RNAs and their effect on the extraction of mtDNA. Brain tissue yielded mtDNA of sufficient purity and quantity to allow visualization of restriction bands on 1.0% agarose gels stained with 2.5 micrograms of ethidium bromide per ml in the gel and electrophoresis buffer. Figure 7 shows the mitochondrial bands obtained from a digest of 150 nanograms of brain and muscle extracted total celluar DNA. It is concluded that the technique presented here is a simplified, valid alternative to current protocols for purification of mtDNA. Of special significance is that both fresh or frozen tissues can be used as sources of mtDNA. This greatly facilitates the logistics of sample collection and processing. Cloning of Snook mtDNA The cloning experiment resulted in the insertion, into the Hind III site of the polylinker sequence of plasmid pT7-2, of three DNA fragments. These fragments had moleclar weights corresponding to those of the snook mtDNA fragments generated by restriction with Hind III. The identity of two of these clones was confirmed by transferring the cloned fragments from the agarose gel to nitroclellulose filters, and hybridizing the filters to radiolabelled bovine mtDNA probes. Thus was substantiated the identity of two independent clones of the 3.75 kbp fragment, and two clones of the 7 51 kbp fragmnet. Repeated cloning experiments failed to obtain the 6.22 kbp
fragment of the snook mtDNA. A DNA fragment of circa 6 kbp was cloned (figure 8) but did not hybridize to the bovine mtDNA probe. Tthe bovine mtDNA probe consistently identified the 6.22 kbp fragment of Hind III restricted snook DNA. Therefore it is concluded that the cloned 6 kbp DNA fragment consisted of nuclear DNA. Final confirmation of the identity of the clones, and of their applicability as molecular probes was demonstrated by hybridizing them to the same bands identified by bovine mtDNA probes. This 56 experiment also evaluated the relative sensitivity of the homologous and heterologous probes. As seen in Figures 9 and 10, several conclusions are obtained. First, the cloned probes do hybridize to the 3.75 and 7.51 kbp Hind III fragments of snook mtDNA. Second, the signal intensity of that hybridization is considerably greater than that generated by the bovine probe: The homologous probed autoradiograph was obtained with a 36 hours exposure, versus a 72 hours exposure for the bovine probed filter. Finally, the use of the homologous probe. revealed a Pvu II fragment of 0.67 + 0.1 kbp that does not hybridize to the bovine mtDNA probe. These results demonstrate that the use of homologous probes increases the sensitivity of our techniques. These results were especially useful in resolving small molecular weight fragments generated by restriction with endonucleases that recognize five and four base pair sequences. The methodology presented here allows the rapid purification, from frozen or fresh tissues, of mtDNA of sufficient purity for direct digestion with endonucleases. It permits the cloning of mtDNA fragments extracted from fresh tissue, and demonstrates the use of the cloned fragments as homologous probes with higher sensitivity than
heterologous probes. Our results demonstrate that the use of heterologous probes will allow the identification of fish mtDNA which then permits us to obtain clones of homologous fragments. In these when employed as increase the sensitivity of these methods. Tissue Specific mtDNA Several tissue types were used as sources of due to the difficulty in obtaining adult snook from all the populations studied. 57 The purest mtDNA was obtained from brain tissue. MtDNA extracted from the and eviscerated entire specimens resulted in f. undecimalis tissue specific mtDNA genome size variations up to 167.. These results are supported by data from CsCl purified mtDNA extraced from liver and kidney tissues of the same specimen. two conclusions are possible. that C undecimalis may posess tissue specific heteroplasmic mtDNA molecules. An alternate conclusion is that mtDNA extracted from different tissues exhibit different electrophoretic migration rates. Reports of mtDNA heteroplasmy from fishes and lower vertebrates (Bermingham et al., 1986; Densmore et al., 1985) found variation in the total mtDNA genome size and in the restriction generated fragmnets. In this study, the tissue specific mtDNA (Table 7) always yielded 3 Hind III restriction fragments. The observed variation consisted of tissue specific fragment size and total mtDNA molecualr weight. The most parsimonious explanation is that the observed variation is a function of tissue specific organic
58 molecules that coprecipitate with the DNA during extraction, and affect the electrophoretic migration of the mtDNA restriction fragments. It is concluded that all the tissues assayed are adequate sources of mtDNA. However, for populational analysis, comparisons must be made only between mtDNA reatriction patterns obtained from the same tissue types. Mitochondrial DNA of three Centropomus species The protocol developed in this study was applied to the characterization of the Hind III endonuclease restriction pattern and genome size of three species of snook. That analysis revealed a C. undecimalis genome comparable to that reported for most fishes (Bermingham et al., 1986). The mtDNA genome size of f. parallelus and pectinatus are less typical, yet fall in the range reported for vertebrate species (Bermingham et al., 1986, Brown, 1983). The above data allow an evaluation of the evolutionary relationships among the three species. One phenogram (Fig. 6a) places undecimalis, which has 3 restriction sites and the smallest genome, as the most genetically distant from f. parallelus and f. pectinatus. The two latter species have 4 restriction sites and a larger mtDNA genotype. This conclusion is opposed to that proposed by Rivas (1986), who compared shared morphological character states. By assuming that the most derived species are also the most successful competitors in the range of a species complex, (sensu Briggs, 1974), it is suggested that the smaller genome size of C. undecimalis reflects loss of base sequences, probably in the non translated D loop region of the mtDNA molecule.
59 An alternate phenogram (Fig. 6b) is based on the size of the shared mtDNA fragments, and places f. undecimalis in the same group as C. parallelus. This phenogram is more in accordance with (Rivas, 1986) Intraspecific mitochondrial DNA polymorphism Lack of restriction site polymorphisms between the East and West coast populations imply the populations studied are panmictic, and that snook are not limited to local estuaries, but are capable of extensive migrations. Tagging studies should demonstrate the potential for extended coastal movement by adult snook. Of special interest is the unique occurrence of a second Bam HI restriction site in a sample subset consisting of two specimens from the Everglades and two from the Indian River. This finding would suggest a closer affinity of the two populations. If such an affinity is real, it would be consistent with the position of Cape Romano (Briggs, 1974) which, in the area studied, is the only major biogeographic barrier to fishes that evince the euryhaline and migratory potential of the snook.
60 LITERATURE CITED Anderson, S., M.H.L. de Bruihin, A R. Coulson, I .C. Eperon, F. Sanger, and I.G. Young 1982. Complete sequence of bovine mitochondrial DNA: conserved features of the mammalian mitochondrial genome. J. Mol. Biol. 156 : 683-717. Anon 1976. Catalogo de Peces Merinos Mexiconos. Secretaria de Industria y Comericio, Subsecretaria de Pesca, Institute Macional de Pesca, Mexico, 1976. 462 pp. Aquadro, C F and B D Greenberg. 1983. Human mitochondrial DNA variation and evolution: analysis of nucleotide sequences from seven individuals. Genetics 103: 287-312. Avise, J.C., 1974. S ystematic value of electrophoretic data. Syst. Zool. 23(4): 465-481. Avise, J .C., C. Giblin-Davidson, J. Laerm, J C. Patton, and R.A Lansman. 1979a. Mitochondrial DNA clones and matriarchal phylogeny within and among geographic populations of the pocket gopher, Geomys pinetis. Proc. Natl. Acad Sci. 76 : 6694-6698. Avise, J.C., R .A. Lansman, and R.O. Shade. 1979b. The use of restriction endonucleases to measure mitochondrial DNA sequence relatedness in natural populations. I Population structure and evolution in the genus Peromyscus. Genetics 92: 279-295. Avise, J C and R.A. Lansman 1983. Polymorphism of mitochondrial DNA in populations of higher animals. In: Evolution of Genes and Proteins, M. Nei and R .K. Kohen, eds. Sinauer Associates, Inc., Sunderland, Massachusetts. Avise, J.C., E. Bermingham, L .G. Kessler, and N. Saunders. 1984. Characterizatio of mitochondrial DNA variability i n a hybrid swarm between subsepecies of bluegill sunfish (Lepomis machrochirus). Evolution, 38, 1984, pp 931-941. Avise, J.C. G.S. Helfman, N.C. Saunders, and L.S. Hales. In press. Mitochondrial DNA and life history pattern in North American eels. Proc. Nat. Acad. Sci. (Submitted). Bermingham, E. and J C Avise. 1986 Molecular zoogeography of freshwater fishes in the South Eastern United States. Genetics 113: 939-965.
61 Bermingham, E., T. Lamb, and J C. Avise. 1986. Size polymorphism and heteroplasmy in the mitochondrial DNA of lower invertebrates. Jour. of Heredity 77: 249-252. Berg, W.J. and S.D. Ferris. 1984. of salmonid mitochondrial DNA. 1041-1047. Restriction endonuclease analysis Can. J. Fish. Aguat. Sci. 41: Birnboim, H.C., and J. Doly 1979. A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nuc. Acids Res. 7: 1513-1523 Birt, T.P. and J.M. Green. 1986. Analysis of mitochondrial DNA in allopatric anadromous and nonanadromous Atlantic salmon, Salmo salar. Can. J. Zool. 64: 118-120. Briggs, J.C. 1974. Marine Zoogeography. McGraw-Hill, New York. 475 pp. Brown, W.M. M. Georgek Jr., and A.C. Wilson. 1979. Rapid evolution of animal mitochondrial DNA. Proc. Natl. Acad. Sci. 76 : 1967-1981. Brown, W.M. 1981. Mechanisms of evolution of animal mitochondrial DNA. Ann. N.Y. Acad. Sci. 361: 119-134. Brown, W.M. 1983 Evolution of animal mitochondrial DNA. In: Evolution of genes and Proteins, M Nei and R.K. Kohen, eds. Sinauer Associates, Inc., Sunderland, Massachusetts. Bruger, G. 1983. Overview of Snook Population Dynamics. Summary of Proceedings, Snook Symposium. Florida DNR and IGFA, Nov. 20, 1982, Ft. Lauderdale, Florida. p. 2. Cann, R .L., W.M. Brown, and A.C. Wilson. 1984. Polymorphic sites and the mechanism of evolution in human mitochondrial DNA. Genetics. 106: 479-499. Chapman, R .W. and D .A. Powers 1984 A method for the isolation of mitochondrial DNA from fishes. Tech. Rpt., Maryland Sea Grant Prgm., #UM-SG-TS-84-84-05. Chavez, H. 1963. Contribucion al conocimiento de la biologia de los robalos chucunite y constantine (Centropomus spp.) del Estado de Veracruz (Pisc. Centrop.). Ciencia, Mexico. XXII (5): 141-161. Crews, S.D. Ojala, J. Posakony, J. Nishiguchi, and G. Attardi. 1979. Nucleotide sequence of a region of human mitochondrial DNA containing the precisely identified origin of replication. Nature. 277: 192-198.
62 Densmore, L.D., S.W. Wright, and W.M. Brown. 1985. Length variation and heteroplasmy are frequent in mitochondrial DNA from parthenogenetic and bisexual lizards (genus Cnemidophorous). Genetics 110: 689-707. Dixon, W.J. 1981. Statistical Software. University of California Press Berkeley. Frazer, T.H. 1968. Comparative osteology of the Atlantic Snooks (Pisces, Centropomus). Copeia 1968(3): 433-460. Gilmore, R. G., C. J. Donohoe, and D. W. Cooke. 1983. Observations on the distribution and biology of east-central Florida populations of the common snook Centropomus undecimalis (Block). Florida Scientist 46(314) 313-336. Gonzalez-Villasenor, L.I. and D.A. Powers. 1985. MtDNA polymorphisms in popualtiosn of the teleost Fundulus heteroclitus. Amer. Zool., 25: 137A. Graves, J.E., S.D. Ferris, and A.E Dizon. 1984. Close genetic similarity of Atlantic and Pacific skipjack tuna (Katsuwonus pelamis) demonstrated with restriction endonuclease analysis of mitochondrial DNA. Mar. Biol. 79: 315-319. Hauswirth, W.W., and P.J. Laipis. 1982. Mitochondrial DNA polymorphisms in a maternal lineage of Holstein cows. Proc. Nat. Acad. Sci. (USA) 79: 4686-4690. Harrison, R.G., D.M. Rand and W.C. Wheeler 1985. Mitochondrial DNA size variations within individual crickets. Science 228: 1446-1448. Lansman, R.A., R.O. Shade, J.F. Shapira, and J.C. Avise. 1981. The use of restriction endonucleases to measure mitochondrial DNA sequence relatedness in natural populations. J. Mol. Evol. (1981) 17: 214-226. Mangini, M .A., B.J. Cochrane, and J .H. Paul. In Press. A readily accessible method for cloning fish mitochondrial DNA and its application as a molecular probe. Proc. V Congr. Europ. Ichthyol. Stockholm. 1985. Maniatis, J., E.F. Fritsch, and J. Hanbrook. 1982. Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, 545 pp. Marshall, A. R. 1958. A Survey of the Snook Fishery of Florida, with Studies of the Biology of the Principal Species, Centropomus undecimalis (Block). Florida State Board Conserv. Tech. Ser. No. 22:39.
Montoya, J., D. Ojala, and G. Attardi. 1981. Distinctive features of the 5' -terminal sequences of the human mitochondrial mRNAs. Nature 290: 465-470. Nei, M. and W. Li. 1979 Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Natl. Acad Sci. 75(10): 5269-5273. 63 Rivas, L. R. 1986. Systematic Review of the Perciform Fishes of the Genus Centropomus. Copeia, 1986 (3), pp. 579-611. Rivas, L. R. 1968. Fishermen's Atlas of Monthly Sea Surface Temperatures for the Gulf of Mexico. XFWC-A300 103 (1968), U.S. Fish Wildlife Serv. Circ. Roberts, R. 1982. Restriction and modification enzymes and their recognition sequences. Nuc. Acid Res. 10: p. 117. Saunders, N.C., L.G. Kessler, and J.C. Avise. 1986. Genetic variation and geographic differentiation in mitochondrial DNA of the Horseshoe crab, Limulus polyphemus. Genetics. 112: 613-627. Shafland, P. L. and K. J. Foote. 1983. A lower lethal temperature for fingerling snook, Centropomus undecimalis. Northeast Gulf Science 6(2): 175-177. Skow, L.C., and M.E. Chittenden. 1981. Differences in hemoglobin phenotypes among Spanish mackerel, Scomberomorus maculatus. Northeast Gulf Sci. 5(1): 67-70. Southern, E.M. 1979. Gel electrophoresis of restriction fragments. In R. Wu. (ed.), Methods in enzymology, Vol. 68 recombinant DNA. Academic Press, New York, pp. 152-182. Takahata, N., and S R. Palumbi. 1985. Extranuclear differentiation and gene flow in the finite island model Genetics 109: 441-457. Tautz, D and M. Renzi. 1983 An optimized freeze-squeeze method for the recovery of DNA fragments from agarose gels. Anal. Biochem. 132: 14-19. Upholt, W.B., and I .G. Dawid. individual sheep and goats: Cell 11 : 571-583. 1977. Mapping of mitochondrial DNA of Rapid evolution in the D Loop region. Volpe, A. V. 1959. Aspects of the Biology of the Common Snook, Centropomus undecimalis (Block) of Southwest Florida. Florida State Board Conserv. Ser. No. 31:37. Winans, G A 1980. Geographic variation in the milkfish Chanos Chanos. I. Biochemical evidence. Evolution 34: 558-574. Zar, J.H. Cliffs. 1984. Biostatistical Analysis. 718 pp Prentice-Hall, Englewood