Molecular Phylogeny of Philippine Freshwater Sardines Based on ...

Molecular Phylogeny of Philippine Freshwater Sardines Based on Mitochondrial DNA Analysis I. E. Samonte, R. C. Pagulayan, and W. E. Mayer

The commercially important Sardinella species (family Clupeidae or herrings) usually thrive in marine environments. An exception is Sardinella tawilis of Taal Lake, Batangas, Philippines, the only known freshwater sardine. This species is believed to have immigrated from Balayan Bay to the lake when it was formed in the course of volcanic eruptions some 240 years ago. To determine the relationship of S. tawilis to the marine species S. albella, S. fimbriata, and S. longiceps from the Balayan Bay we sequenced 358 bp of the cytochrome b gene and the mitochondrial control region. The cytochrome b gene was highly conserved and contained little phylogenetic information. The control region sequences, however, demonstrated two highly diversified main haplotypes grouping S. tawilis with S. albella, as shown by maximum parsimony and neighbor-joining analysis. The haplotypes are characterized by the presence of an 81 bp indel and up to eight 35 bp tandem repeat elements. The repeat copy number varied within individuals of S. tawilis and S. albella, thus showing heteroplasmy in these two species only. The analysis of two subpopulations of S. tawilis revealed restricted substitutions that may indicate the beginning of genetic differentiation of the two subpopulations.

From the Max-Planck-Institut fu¨r Biologie, Abteilung Immungenetik, Corrensstrasse 42, D-72076 Tu¨bingen, Germany (Samonte and Mayer), and the Institute of Biology, University of the Philippines–Diliman, Diliman, Quezon City, Philippines (Pagulayan). We thank Dr. Jan Klein for his support and stimulating discussions. I.E.S. was supported by a short-term scholarship from the Deutscher Akademischer Austauschdienst ( DAAD) and the Engineering and Science Education Program ( ESEP) of the Department of Science and Technology ( DOST ), Philippines. Address correspondence to Werner E. Mayer at the address above or e-mail: werner.mayer@ tuebingen.mpg.de. 2000 The American Genetic Association 91:247–253

Members of the family Clupeidae, or herrings (order Clupeiformes, subclass Actinopterygii), are fish with a compressed streamlined body, a single soft-rayed dorsal fin, and protruding scales. They are found worldwide in temperate and tropical waters where they mainly move in large schools feeding on plankton. Most of the species live in the marine environment, but some, like the American gizzard shad (Dorosama cepedianum), occur only in freshwater systems. The family Clupeidae includes the subfamilies Alosinae or shads, the Pellonulinae or freshwater herrings, and the Clupeinae with the genera Clupea, Sardina, and Sardinops. Little is known about the phylogenetic relationships among the commercially important Sardinella species ( Family Clupeidei, Suborder Clupeoidei), especially at the molecular level. In the Philippines, eight members of this genus were identified and collectively called sardines (Conlu 1986). All of these species are thriving in the marine environment except Sardinella tawilis (Whitehead 1985), the only known freshwater sardine, endemic to Taal Lake, Batangas, Philippines. Known locally as tawilis, this species is believed to have immigrated from Balayan Bay to Taal Lake. The said lake was a caldera lake that formed partly

by the collapse of a large volcanic crater and by subsidence, and whose morphometry has been modified by subsequent volcanic activities, especially by a powerful eruption in 1754 ( Hargrove 1991). The 10th century lake, connected to Balayan Bay through a wide channel, has now rearranged its shape and narrowed its outlet to form the Pansipit River, the sole outlet from its southwest corner ( Kira 1995). Like other members of the genus, tawilis has an oblong to nearly elongated body, equal jaws, large mouth, lacks lateral lines, and has thin deciduous scales (Munro 1955). Because of its small size (10–20 cm) it does not command much interest economically, but its year-round presence makes it important to fishers who depend solely on the lake for their daily subsistence. Interest in this species has arisen mainly because of a recent report of human intervention in the lake which threatens its fishery resources ( Villanueva et al. 1996). Also the recent reclassification of the species from Harengula to Sardinella on the basis of morphoanatomic features makes it an interesting subject of a molecular study ( FishBase 1995). Analyses of the mitochondrial DNA (mtDNA) sequences have proved to be a valuable tool in addressing the problems

247

Table 1. Primers used for the amplification of the cytochrome b gene and the control region of the sardines Primer designation

Sequence

Cytochrome b: cytbLa cytbHa

5⬘-CCCCTCAGAATGATATTTGTCCTCA-3⬘ 5⬘-CCATCCAACATCTCAGCATGATGAAA-3⬘

Control region: L15926a H16498b Q B J N A88c A87c

5⬘-TTACACCAGTCTTGTAAACC-3⬘ 5⬘-CCTGAAGTAGGAACCAGATG-3⬘ 5⬘-GGGCGGATCCCACCACTAGCTCCCAAA-3⬘ 5⬘-ACGCTGGAAAGAACGCCCGGCATGG-3⬘ 5⬘-TTTGGTTCCTATTTCAGGGCCA-3⬘ 5⬘-GGGGCGCGGATCCCATCCTAATATCTTCAG-3⬘ 5⬘-GCAAAACTCTCACCAACTCATC-3⬘ 5⬘-GGGGTTTCATGGTGAACT-3⬘

Primer designed by Kocher et al. (1989) except that the restriction site was deleted. Primer designed by Meyer and Wilson (1990). c Internal primers designed from the generated sequences. The remaining primers are those used by Lee at al. (1995).

a

b

that S. tawilis faces. Features of this molecule, such as its compact organization (Russell 1994), its elevated rate of mutation ( Brown 1983), its primarily maternal inheritance, and the absence of recombination (Moritz et al. 1987), make it a valuable tool in evolutionary and population studies. Fish mtDNA ranges in length from 15.2 to 19.8 kb ( Billington and Hebert 1991) and contains information specifying the two rRNAs (16s and 12S), 22 tRNAs, 13 polypeptides, as well as the dissociation ( D) loop, also referred to as the control region. The proteins it encodes include NADH dehydrogenase, cytochrome oxidase, cytochrome b (cytb), and ATPase 6 and 8 (Gyllensten and Wilson 1987). Of these, cytochrome b ( Barlett and Davidson 1991) together with the control region ( Zhu et al. 1994) have become widely used molecular markers. In vertebrates, the rate of the cytb gene evolution is estimated to be between 1% and 2.5% substitutions per one million year ( Irwin et al. 1991; Martin et al. 1992) which means that approximately one hundred thousand years are needed before one might confidently expect a substitution to differentiate closely related mtDNA types (Palumbi and Kessing 1991). This relatively slow evolution of the cytb gene makes it a suitable marker for resolving deeper phylogenetic relationships among different taxa (Meyer and Wilson 1990). Unlike cytb, the mitochondrial control region does not code for proteins (Avise et al. 1987) and has only short sequence blocks conserved among distant taxa ( Brown et al. 1993; Clayton 1982; Lee et al. 1995; Saccone et al. 1987), presumably because of relaxed functional constraints (Alves-Gomes et al.

248 The Journal of Heredity 2000:91(3)

1995). The variable portions of the control region contain varying numbers of tandem repeat sequences ( Broughton and Dowling 1994; Lee et al. 1995) which are the targets of frequent mutations, particularly deletions, insertions, and duplications ( Brown 1983; Harrison 1989). Its fast evolutionary rate makes the control region a suitable marker for resolving relatively recent divergences of closely related species and populations ( Zischler et al. 1995). Analyses of mitochondrial genes have been facilitated by the discovery of the polymerase chain reaction (PCR), which enables specific sections of the genes to be routinely investigated (Carr and Marshall 1991; Kocher et al. 1989) from minute quantities of tissue samples (McVeigh et al. 1991) to yield high resolution information about any variation that may exist in the DNA ( Hartley et al. 1992). The aim of this study was to determine the phylogenetic relationship of S. tawilis to the marine sardines S. albella (Whitehead and Wongratana 1986), S. fimbriata (Whitehead 1985), and S. longiceps (Whitehead 1985), and the genetic variability between populations of S. tawilis from the mtDNA sequences. The latter information was analyzed for fish management purposes.

Figure 1. Geography of Taal Lake showing the Pansipit River and portions of Balayan Bay and Batangas Bay. The location of the Talisay and Agoncillo parts is indicated.

tions of S. tawilis were obtained from the Talisay and Agoncillo portions of Taal Lake ( Figure 1). S. albella and S. fimbriata from Balayan Bay were procured from fish vendors in Lemery, Batangas. S. longiceps, caught in Cebu Bay, was obtained from the Malabon Fish Port in Malabon, Metro Manila. Muscle and gonads were dissected out and preserved in 85% ethanol. These were then brought to the Max-Planck Institute for Biology for analysis. DNA Extraction Total DNA of the ethanol-preserved samples was extracted according to the NucleoSpin C⫹T protocol (Macherey-Nagel, Du¨ren, Germany). MtDNA was extracted using the modified phenol extraction procedures of Gonzales Villasenor et al. (1990), Chapman and Powers (1984), and Bernatchez et al. (1988). The extracted mtDNA was precipitated with ethanol and resuspended in 50 ␮l 10 mM TE buffer (pH 8.5). Yields were typically 800 ␮g DNA/ml of TE. PCR Amplification The primer sets used to PCR amplify (Saiki et al. 1988) a segment of the mitochondrial cytochrome b gene (cytb) and the mito-

Materials and Methods Collection of Fish Samples Four Philippine sardine species, S. tawilis, S. albella, S. fimbriata, and S. longiceps, were collected and transported to the Natural Sciences Research Institute’s laboratory for identification and dissection. Rau and Rau (1980) and Conlu (1986) served as bases for identification. Two popula-

Figure 2. Schematic representation of the mitochondrial control region. Primers used for PCR and sequencing are indicated by arrows.

chondrial control region are shown in Table 1 and in Figure 2. Hot-Start PCR was performed in a 50 ␮l reaction containing Mg2⫹-free PCR buffer [60 mM Tris, 15 mM ( NH4)2SO4, pH 8.5 ( Invitrogen, Leek, The Netherlands)], 250 ␮M dNTPs, 0.5 ␮M of each primer, 1.0 ␮l DNA template, and 2.5 units Taq polymerase (Pharmacia Biotech). Magnesium was supplied to the reactions by adding HotWax-Mg2⫹ beads ( Invitrogen). Amplification was performed in the PTC-200 thermal cycler (MJ Research, Watertown, MA, and Biozym, Hess. Oldendorf, Germany). The amplification profile consisted of a 3-min preliminary denaturation at 94⬚C, followed by 30 cycles of denaturation at 94⬚C for 30 s, primer annealing at 48⬚C–55⬚C for 45 s, and primer extension at 72⬚C for 2 min. An extension step at 72⬚C for 5 min completed the reaction. After thermal cycling, the PCR products were separated on 1% agarose gel ( NEEO ultra-quality, Roth, Karlsruhe, Germany) and retrieved using the QIAquick Gel Extraction Kit (QIAGEN, Hilden, Germany).

ware (version 0.6, written by Don Gilbert, Indiana University, Bloomington, IN) and the alignments were then corrected manually. For the control region analyses, Alosa sapidissima, Alosa pseudoharengus, and Clupea harengus as the most closely related sequences in the GenBank database served as outgroups. Maximum parsimony (MP) analysis was performed using the branch and bound search algorithm of the PAUP 3.1.1 software (Swofford 1993). The topology of the consensus tree was evaluated by 500 bootstrap replications. Neighbor-joining ( NJ) trees (Saitou and Nei 1987) were constructed from a Kimura two-parameter distance matrix ( Kimura 1980) using the MEGA program ( Kumar et al. 1993). Input orders of the taxa were varied to test the robustness of the trees. The reliability of the tree topology was assessed by 500 bootstrap replications ( Felsenstein 1985). The number of transitions and transversions was calculated using the MEGA program.

Results Cloning The purified PCR products were ligated into pGEM-T Easy Vector (Promega, Mannheim, Germany) and transformed into E. coli JM109 according to manufacturer’s instructions. Plasmid DNA from recombinant clones was isolated using the QIAGEN Plasmid Mini Kit. Sequencing Single-stranded plasmid DNA was sequenced using ALF DNA sequencer (Pharmacia Biotech) and/or LI-COR Automated Sequencer 4200 (MWG-Biotech, Ebersberg, Germany). For ALF, sequencing reactions were prepared using either the AutoRead 200 Sequencing Kit (Pharmacia Biotech) or the SequiTherm EXCEL Long-Read Premix DNA Sequencing KitALF ( Epicentre Technologies, Biozym). LICOR sequencing reactions were prepared using the SequiTherm Long-Read DNA Sequencing Kit-LC ( Epicentre Technologies, Biozym) with primers labeled with two different fluorescent dyes ( IRD 700 and IRD 800, MWG-Biotech). Phylogenetic Analysis/Sequence Analysis Only sequences obtained by at least two independent PCRs were considered. The cytb and the control region sequences were aligned together with related sequences available from the GenBank database with the help of the SeqPup soft-

Nucleotide Variation in the Cytochrome b Gene The amplification of a 359-nucleotide long cytb fragment from fishes of two populations of the freshwater sardine species S. tawilis from Taal Lake and three marine sardine species from the Balayan Bay produced identical sequences except for single individual-specific substitutions (GenBank accession codes AF1014944– AF104948). Only a single transition at position 166 was shared by two individuals from the northern population of S. tawilis ( Talisay). It occurred at a nonsynonymous site leading to the conservative amino acid replacement Ile to Val. Nucleotide Variation in the Control Region For the amplification of the mitochondrial control region of the different sardine species, distinct primer sets had to be applied, suggesting haplotypic variation between species. The primer pair L15926 and H16498 amplified the first two-thirds of the control region of S. fimbriata and S. longiceps to produce a fragment of 600 bp, whereas the primer combination Q and B generated the corresponding product in S. albella and the two populations of S. tawilis. The remaining portions were amplified with the primer pairs A87/A88 for the two populations of S. tawilis, J/A88 for all the sardines, and J/N for S. albella and the two

populations of S. tawilis, as shown in Figure 2. The complete control region sequence of the Philippine sardines is 804 nucleotides long for S. fimbriata and S. longiceps, while it ranges from 1231 to 1418 nucleotides in S. albella and the two populations of S. tawilis due to the presence of an 81 bp insertion and several copies of a 35 bp repetitive element near the tRNAPro region ( Figure 3). At its 5⬘ end the 81 bp insertion repeats the 32 bp preceding the insertion site. The copy number of the 35 bp repetitive element varied from five to eight in the sequenced fragments. This variation was also observed within individuals, thereby producing more than one form of mtDNA per fish. To verify this phenomenon of heteroplasmy, the PCR products from tissue samples of all species and from cloned control region fragments were electrophoresed through a 2.5% NuSieve 3: 1 agarose gel with high resolving power. The analysis revealed three to five bands ranging from 800 to 1100 bp in mtDNA samples from S. albella and S. tawilis, but only single bands in S. longiceps or S. fimbriata and in the samples of cloned control region fragments from either species, thus confirming the presence of heteroplasmic mtDNA in S. albella and S. tawilis and excluding PCR artifacts. Comparison of the sardine control region sequences in Figure 3 shows low similarity between the two main haplotypic sequences. The highest variability resides in the first 940 nucleotides followed by low variability in the rest of the sequence. Conserved sequence blocks (CSBs), CSB-D and CSB-2, homologous to those of other vertebrates ( Lee et al. 1995), are present together with the GTGGG-box. Sequence-Based Phylogeny Control region. Only mitochondrial control sequences were informative for phylogenetic analysis. The control region analysis was divided into two parts. First, the segment between the tRNAPro gene and the conserved sequence blocks was used for the assessment of the phylogenetic position of the sardines because of the large number of informative sites and because corresponding sequences from related fish species were available in the GenBank sequence database. Second, all homologous segments of the control region were used for a more detailed analysis within the genus Sardinella. In the first part, the analysis was restricted to the alignable segments (positions 699 to 961 in Figure

Samonte et al • Philippine Freshwater Sardine Phylogeny 249

Figure 3. Sequence alignment of the sardine mitochondrial control region. Representative sardine sequences were aligned to control region sequences of A. sapidissima, A. pseudoharengus, and C. harengus. The sardine sequences are distinguished by a number for the individual specimen, separated by a dot from the sequence number. The two subpopulations of S. tawilis are indicated by a T ( Talisay) and an A (Agoncillo). A dash (–) indicates identity to the sequence on top, an asterisk (*) an indel, a dot (.) missing information, a ( N) undetermined nucleotide at the primer annealing site. The 32 bp segment preceding the 81 bp insertion site is underlined, solid bars indicate the 35 bp repeat units, the shaded bar the central conserved region. Conserved sequence blocks are boxed. The sequences reported in this article have been deposited in the GenBank database (accession codes AF104949–AF104969).

3) of the two main sardine haplotypes and the sequences of the most closely related Clupeidae ( herrings), Alosa sapidissima (GenBank accession U12061), A. pseudoharengus (accession U12067), and Clupea harengus (accession U12062), excluding all repeat elements. The alignment of the remaining 263 nucleotides was used to reconstruct phylogenetic trees by the neighbor-joining and maximum parsimony methods ( Figure 4). Both trees show essentially the same topology and reveal the divergence of the sardine control region into two major haplotypes evident already from the presence or absence of the repeat elements: S. fimbriata and S. longiceps


cluster together in one branch, and the freshwater species S. tawilis and the marine species S. albella group in a separate branch. The clades are supported by high bootstrap values. To evaluate the relationships within the genus Sardinella, the nonrepeat portions of the control region of S. tawilis and S. albella were aligned (positions 45–160, 241–487, 594–1058) . A total number of 30 variable nucleotide sites was found in the 626 nucleotides considered, some of which distinguish S. albella from S. tawilis (substitutions at positions 100, 263, 279) or the two subpopulations of S. tawilis from each other (positions 778, 846),

while others seem to represent intraspecific variants.

Discussion Structural Analysis The amplified segments of the Sardinella mtDNA showed features characteristic of the respective regions. The selective pressure exerted on the protein-coding cytb gene kept its sequence conserved among the sardines. In stark contrast, in the control region, even primer sets known to amplify mtDNA from species as distantly related as teleost fishes and primates failed to amplify two of the sardine species. The

Figure 4. (A) Maximum parsimony tree. The tree was constructed from alignable segments of clupeid control regions using the branch-and-bound option of the PAUP program. Numbers at the nodes indicate the percent recovery in 500 bootstrap replications.

due to multiple copies of repeat units. The first of these repeat units, an 81 bp insertion, was found in some S. tawilis sequences and in S. albella. The second repeat element, a tandem repeat of 35 bp, was present in up to eight copies in S. tawilis and S. albella. The single repeat units showed low or no sequence variation from each other indicating a recent divergence, whereas the similarity to the monomeric element in S. fimbriata and S. longiceps was rather low. To calculate the ratio of transitions to transversions in the control region, 626 repeat-free nucleotides from S. tawilis and S. albella were aligned. Six to 17 transitions and zero to five transversions were observed. The mean ratio of transitions to transversions in the control region thus was 4.3, suggesting no saturation effects in the Sardinella sequences. Heteroplasmy Copy number polymorphism ( heteroplasmy) of the 35 bp repeat element was found

Figure 3. Continued.

structural organization of the control region in S. longiceps and S. fimbriata is dramatically different from that in S. albella and S. tawilis, thus defining two different major haplotypes, although homologous segments clearly identify all four species as closely related to each other.

These segments include the first 13 nucleotides after the tRNAPro gene and the stretch from the central conserved region to the tRNAPhe gene. The variable portion, which encompasses sequences without similarity between the two haplotypes, is mainly characterized by length variation

Figure 4. (B) Neighbor-joining tree. A Kimura two-parameter distance matrix of the control region sequences was used to construct a neighbor-joining tree using the MEGA program and the pairwise deletion option for missing information. Numbers at the nodes indicate the percent recovery in 500 bootstrap replications.


to be present within 17 tested individuals of both populations of S. tawilis and 7 individuals of S. albella (data not shown). Heteroplasmy has been noticed before in various fish species, among them Atlantic cod (A´rnason and Rand 1992), sturgeon ( Brown et al. 1992; Miracle and Campton 1995), and redfishes ( Bentzen et al. 1998). The occurrence of heteroplasmic mtDNA in only two of the sardine species and the virtual identity of the repeat sequences suggest a recent origin of the tandem repeat array. Phylogenetics Because of lack of variation, cytb sequences are of little use for the phylogenetic analysis of the sardine species. However, a single nonsynonymous substitution was found that may prove valuable in the recognition of two subpopulations of S. tawilis in Taal Lake. As expected the phylogenetic signal present in the control region is more useful. Among 626 nucleotides, excluding repeat-induced indels, eight to 22 substitutions were detected in pairwise comparisons in S. tawilis and S. albella. A similarly high rate variation was observed by McMillan and Palumbi (1997) in Pacific butterflyfishes, where the control region evolved 33–43 times faster than the cytb gene. Furthermore, the control region harbors a set of repeat-related indels that is useful as a cladistic marker. The aligned sardine sequences were used to construct neighbor-joining and maximum parsimony phylogenetic trees to determine the closest relative of the freshwater sardine species S. tawilis ( Figure 4). Both trees are congruent and support the recent reclassification of S. tawilis into the genus Sardinella. Two additional conclusions can be drawn. First, the closest marine relative to S. tawilis is S. albella. Its grouping within the tawilis branch and the small genetic distance of the tested individual to freshwater sardines identifies it as the most likely ancestor to the Taal Lake sardines. Its control sequence differs by 1.3–3.5% from the S. tawilis sequences which show an intraspecific variation of 1.4–3.0%. Ancestral polymorphism and incomplete lineage sorting, perhaps in combination with an accelerated mutation rate in the control region, may therefore account for the observed DNA diversification of S. tawilis from S. albella in the time span of some 240 years since the formation of Taal Lake. The possibility, however, that S. tawilis and S. albella are ecomorphs of the same species and not truly differentiated spe-


cies cannot be dismissed. Second, the two subpopulations of S. tawilis show indications of genetic differentiation. The presence of a population-restricted substitution in the control region is a first hint for genetic separation, possibly caused by reduced fish migration across a shallow barrier in the middle of the lake. A more extended investigation with large sample sizes is needed to answer these questions.

References Alves-Gomes JA, Orti G, Haygood M, Heiligenberg W, and Meyer A, 1995. Phylogenetic analysis of the South American electric fishes (Order Gymnotiformes) and the evolution of their electrogenic system: a synthesis based on morphology, electrophysiology and mitochondrial sequence data. Mol Biol Evol 12:298–318. A´rnason E and Rand DM, 1992. Heteroplasmy of short tandem repeats in mitochondrial DNA of Atlantic cod, Gadus morhua. Genetics 132:211–220. Avise JC, Reeb CA, and Saunders N, 1987. Geographic population structure and species differences in mitochondrial DNA of mouthbrooding marine catfishes (Ariidae) and demersal spawning toadfishes ( Batrachoididae). Evolution 41:991–1002. Barlett SE and Davidson WS, 1991. Identification of Thunnus tuna species by the polymerase chain reaction and direct analysis of their mitochondrial cytochrome b genes. Can J Fish Aquat Sci 48:309–317. Bentzen P, Wright JM, Bryden LT, Sargent M, and Zwanenburg KCT, 1998. Tandem repeat polymorphism and heteroplasmy in the mitochondrial control region of redfishes (Sebastes: Scorpaenidae). J Hered 89:1–7. Bernatchez L, Savard L, Dodson JJ, and Pallotta D, 1988. Mitochondrial DNA sequence heterogeneity among James–Hudson Bay anadromous coregonines. Finn Fish Res 9:17–26. Billington N and Hebert PDN, 1991. Mitochondrial DNA diversity in fishes and its implications for introductions. Can J Fish Aquat Sci 48:80–94. Broughton RE and Dowling TE, 1994. Length variation in mitochondrial DNA of the Minnow Cyprinella spiloptera. Genetics 138:179–190. Brown WM, 1983. Evolution of animal mitochondrial DNA. In: Evolution of genes and proteins ( Nei M and Koehn RK, eds). city?: publisher?; 62–88. Brown JR, Beckenbach AT, and Smith MJ, 1992. Mitochondrial DNA length variation and heteroplasmy in populations of white sturgeon (Acipenser transmontanus). Genetics 132:221–228.

FishBase, 1995. Fishbase: a biological database on fish, version 1.2. Manila, Philippines: ICLARM. Gonzalez Villasenor LI, Burkhoff AM, Corces V, and Powers DA, 1990. Characterization of cloned mitochondrial DNA from the teleost Fundulus heteroclitus and its usefulness as an interspecies hybridization probe. Can J Fish Aquat Sci 43:1866–1872. Gyllensten U and Wilson AC, 1987. Mitochondrial DNA of salmonids: inter- and intraspecific variability detected with restriction endonucleases. In: Population genetics and fishery management. Seattle: University of Washington Press; 302–317. Hargrove TR, 1991. The mysteries of Taal. Flagstaff, AZ: AGCOM International. Harrison RG, 1989. Animal mitochondrial DNA as a genetic marker in population and evolutionary biology. Trends Ecol Evol 4:6–11. Hartley SE, Barlett SE, and Davidson WS, 1992. Mitochondrial DNA analysis of Scottish populations of Arctic charr, Salvelinus alpinus ( L). J Fish Biol 40:219–224. Irwin DM, Kocher TD, and Wilson AC, 1991. Evolution of the cytochrome b gene of mammals. J Mol Evol 32: 128–144. Kimura M, 1980. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J Mol Evol 16: 111–120. Kira T (ed), 1995. Data book of world lake environments—a survey of the state of world lakes—Asia & Oceania. Otsu, Japan: Miyagawa Printing; 545. Kocher TD, Thomas WK, Meyer A, et al., 1989. Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers. Proc Natl Acad Sci USA 86:6196–6200. Kumar S, Tamura K, and Nei M, 1993. MEGA: molecular evolutionary genetic analysis, version 1.0. University Park, PA: Pennsylvania State University. Lee W-J, Conroy J, Howell WH, and Kocher TD, 1995. Structure and evolution of teleost mitochondrial control regions. J Mol Evol 41:54–66. Martin AP, Naylor GJP, Palumbi SR, 1992. Rates of mitochondrial DNA evolution in sharks are slow compared with mammals. Nature 357:153–155. McMillan WO and Palumbi SR, 1997. Rapid rate of control-region evolution in Pacific butterflyfishes (Chaetodontidae). J Mol Evol 45:473–484. McVeigh HP, Barlett SE, and Davidson WS, 1991. Polymerase chain reaction/direct sequence analysis of the cytochrome b gene in Salmo salar. Aquaculture 95:225– 233. Meyer A and Wilson AC, 1990. Origin of tetrapods inferred from their mitochondrial DNA affiliation to lungfish. J Mol Evol 31:359–64.

Brown JR, Beckenbach AT, and Smith MJ, 1993. Intraspecific DNA sequence variation of the mitochondrial control region of white sturgeon (Acipenser transmontanus). Mol Biol Evol 10:326–341.

Miracle AL and Campton DE, 1995. Tandem repeat sequence variation and length heteroplasmy in the mitochondrial DNA D-loop of the threatened gulf of Mexico sturgeon, Acipenser oxyrhynchus desotoi. J Hered 86: 22–27.

Carr SM and Marshall HD, 1991. Detection of intraspecific DNA sequence variation in the mitochondrial cytb gene of Atlantic cod (Gadus morhua) by polymerase chain reaction. Can J Fish Aquat Sci 48:48–52.

Moritz C, Dowling TE, and Brown WM, 1987. Evolution of animal mitochondrial DNA: relevance for population biology and systematics. Annu Rev Ecol Syst 18:269– 292.

Chapman RW and Powers DA, 1984. A method for rapid isolation of mtDNA from fishes. Technical report no. UM-SG-TS-84-05. College Park, MD: Maryland Sea Grant Program.

Munro ISR, 1995. The marine and freshwater fishes of Ceylon. Canberra: Department of Environmental Affairs.

Clayton DA, 1982. Replication of animal DNA. Cell 28: 693–705. Conlu PV, 1986. Guide to Philippine flora and fauna. Vol. IX. Manila, Philippines: National Resources Management Center, Ministry of Natural Resources, and University of the Philippines. Felsenstein J, 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783–791.

Palumbi SR and Kessing BD, 1991. Population biology of the trans-arctic exchange: mtDNA sequence similarity between Pacific and Atlantic sea urchins. Evolution 45:1790–1815. Rau N and Rau A, 1980. Commercial marine fishes of the central Philippines ( bony fishes). German Agency for Technical Cooperation; 25, 196, 203, 205. Russell PJ, 1994. Fundamentals of genetics. New York: Harper Collins.

Saccone C, Attimonelli M, and Sbisa E, 1987. Structural elements highly preserved during the evolution of the D-loop-containing region in vertebrate mitochondrial DNA. J Mol Evol 26:205–211. Saiki RK, Gelfand DH, Stoffel, et al., 1988. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487–491.

parsimony, version 3.1. Champaign, IL: Illinois Natural History Survey. Villanueva LS, Luistro AP, and Calabig CS, 1996. Assessment of Taal Lake capture fisheries with emphasis on the exploitation of Harengula tawilis. STIARC.

Saitou N and Nei M, 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4:406–425.

Whitehead PJP, 1985. FAO species catalogue. Vol. 7. Clupeoid fishes of the world. An annotated and illustrated catalogue of the herrings, sardines, pilchards, sprats, shads, anchovies and wolf-herrings. Part 1—Chirocentridae, Clupeidae and Pristigasteridae. FAO; 303.

Swofford DL, 1993. PAUP: phylogenetic analysis using

Whitehead PJP and Wongratana T, 1986. Clupeidae. In:

Smith’s sea fishes (Smith MM and Heemstra PC, eds). Berlin: Springer-Verlag; 199–204. Zhu D, Jamieson BGM, Hugall A, and Moritz C, 1994. Sequence evolution and phylogenetic signal in controlregion and cytochrome b sequences of rainbow fishes (Melanotaeniidae). Mol Biol Evol 11:672–683. Zischler H, Geisert H, von Haeseler A, and Pa¨a¨bo S, 1995. A nuclear ‘‘fossil’’ of the mitochondrial D-loop and the origins of modern humans. Nature 378:489–492. Received November 24, 1998 Accepted November 30, 1999 Corresponding Editor: Stephen J. O’Brien
