Heredity 79 (1997)178—189 Received21 August 1996

Mitochondrial DNA variation in the (L.) assessed by a PCR-RFLP method

CRAIG S. WILDING*, ANDY R. BEAUMONT & JOHN W. LATCH FORD School of Sciences, University of Wa/es (Bangor), Menal Bridge, Gwynedd LL59 5EY, UK.

Two PCR—RFLP mitochondrial DNA (nitDNA) markers were developed through the cloning and sequencing of mtDNA from the scallop Pecten maximus, and were used to study genetic differentiation of UK and Atlantic coast populations of this . Although no distinct pattern of mtDNA haplotype frequencies was apparent and no diagnostic haplotypes were identified for any population, sequence divergence data provided convincing evidence that a P maximus sample taken from Muiroy Bay, Eire, a semi-enclosed sea lough, was genetically differentiated from all other samples. However, this could not be unequivocally attributed to a restriction in gene flow, as the sample consisted of an ongrown single spatfall, which may not have been representative of the wild population. Despite the inability to separate populations on the basis of haplotype frequency, it was noteworthy that the frequency pattern of the commonest haplotype varied between sampling sites in a manner similar to that of allozyme allele frequencies in Aequipecten opercularis, a scallop species with a similar distribution and life history, for which there is evidence of population subdivision. Pecten maximus from St Brieuc Bay, reasoned to be a self-recruiting population from reproductive and physiological evidence, could not be separated froni other populations using mtDNA markers. Further investigation of this population with alternative markers is warranted.

Keywords:lengthvariation, mitochondrial DNA, Pecten maximus, PCR, population genetics, scallop.

Introduction evidence that larval movement between beds, and Thescallop Pecten maximus (L.) has an extensive hence gene flow, may be constrained. The evidence distribution throughout the coastal waters of for this claim ranges from differences between popu- northern , ranging from Norway south to the lations in growth rate (Mason, 1983) and colour Iberian Peninsula, and supports an important and (Minchin, 1991) to potential hydrographic restric- valuable fishery (Ansell et al., 1991; Brand, 1991). tions to gene flow between exploited populations Pecten maximus are most common between 20 m detected using the North Sea water parcel following and 45 m on bottoms of sand, fine gravel or sandy (NORSWAP) model (Dare et at., 1994). gravel (Brand, 1991) in areas of favourable bottom from a population in St Brieuc Bay, Brittany, stress (Dare et a!., 1994). Such restrictions on suit- present the clearest indication that gene flow able habitat result in a patchy distribution, with between beds may be constrained, displaying a dense aggregations (beds) where scallops occur in markedly different spawning cycle from neighbour- sufficient densities for fishery exploitation (Sinclair ing populations, the characteristics of which are et at., 1985). Although P maximus larvae and post- maintained following reciprocal transplantation larvae have a considerable planktonic dispersal (Paulet et at., 1988; Cochard & Devauchelle, 1993; capability (Beaumont & Barnes, 1992), there is Mackie & Ansell, 1993). Past efforts to identify differences between P maximus populations using allozyme frequency data *Correspondence: Department of Biology, University of Leeds, have, however, proved unfruitful (Beaumont et at., Leeds LS2 9JT, U.K. E-mail: [email protected] 1993). In contrast, for Aequipecten opercularis, a

178 1997The Genetical Society of Great Britain. MITOCHONDRIAL DNA VARIATION IN PECTENMAX/MUS179 scallop species with a similar distribution and life Methods history, there are significant allele frequency differ- ences between populations, and separation of the St Sampling Brieuc population on the basis of allozyme data Twelvepopulations of P maximus were sampled by (Beaumont, 1982). As A. opercularis has an dredge fishing or scuba diving (see Fig. 1). All were increased capacity for larval dispersal relative to P samples of adult . Mulroy Bay individuals maximus owing to its enhanced byssal drifting had been collected as wild spat, ongrown to approxi- capability (Beaumont & Barnes, 1992), gene flow in mately 2 cm in Mulroy Bay, then transplanted to P maximus is a prioriexpectedto be lower and, thus, Kilkieran Bay. Mortality was negligible (Dr G. genetic differentiation greater. If I? maximus is Burnell,UniversityCollege,Cork,personal indeed genetically structured, then the inability to communication). Animals were received live and separate populations using allozyme methodologies maintained in running seawater until dissection, or may be explainable by the action of balancing selec- dissected and frozen (—70°C)until use. The tion at allozyme loci, which has been suggested as Polperro and Lyme Bay samples were received being capable of maintaining similar population frozen at —20°Cand used after thawing of tissue. allele frequencies in spite of negligible gene flow (Karl & Avise, 1992). DNA-based methodologies targeted at essentially neutral loci are now being used to challenge some of the results based on allozyme data (Mitton, 1994). The marker of choice for the majority of population level comparisons has been the mitochondrial DNA (mtDNA), and restriction fragment length poly- morphism (RFLP) analysis of the mtDNA molecule has been widely applied to marine species (Ovenden, 1990). However, recent work has shown that in many scallops (Snyder et a!., 1987; Gjetvaj et al., 1992; Repin & Brykov, 1992; Boulding et al., 1993), including P maximus (Rigaa et al., 1993), mtDNA length differences between individuals and species are extensive. Pecten maximus mtDNA is variable in length as a result of varying numbers of a tandemly repeated 1.55 kb element (Rigaa et al., 1993, 1995). This may complicate mtDNA RFLP studies, as the separation of fragment pattern differ- ences because of restriction site losses and gains from those caused by length differences can prove difficult (Moritz et a!., 1987; Boulding et al., 1993; Wilding, 1996). This problem can be avoided p through examination of RFLPs in defined sections of mtDNA amplified by the polymerase chain reac- tion (PCR). PCR—RFLP has proved to be a useful technique for systematic studies and species identi- fication (e.g. Chow & Inoue, 1993) and in popula- tion analysis (Cronin et al., 1993, 1994; Bartlett & Davidson, 1995; Hall & Nawrocki, 1995). To date, there has been no study examining popu- Fig. 1 Collection locations for Pecten maximus samples. lation level DNA variability within P maximus. DOU, Douglas; PEE, Peel; CHI, Chicken Rock; STO, Here, we demonstrate that specific primers designed Stonehaven; MUL, Mull; LYM, Lyme Bay; POL, for PCR—RFLP can be useful in the examination of Polperro; MRY, Mulroy Bay (transplanted to and population genetic structure and we apply the ongrown in Kilkieran Bay); KIL, Kilkieran Bay; STB, method to P maximus from Britain, Eire and Nerput, St Brieuc Bay; BRE, Tinduff, Daoulas Bay, Rade northern France. de Brest; LAT, La Trinité sur Mer.

The Genetical Society of Great Britain, Heredity, 79, 178—189. 180 C. S. WILDING ETAL.

DNA extraction temperature (60°C). Primers were synthesized by MitochondrialDNA of individual animals was Cruachem, Glasgow. Primer sequences were as extracted from approximately 5 g of adductor muscle follows. Primer pair Pmal, target 2 kb EcoRI—HindIII using the technique of Snyder et a!. (1987) adapted to include the modifications of Boulding et al. (1993) fragment (5'—3'): and a CTAB (hexadecyltrimethyl ammonium TTTTAAGGAGGTAATCGCTAFCG bromide) incubation to remove polysaccharides CAGCAATCTGTATGGGTAGAACC (Fisher & Skibinski, 1990). MtDNA was resuspen- ded in 50 L of dH2O. Primer pair Pma2, target 3.85 kb Hindlil—Hindlil Genomic DNA was isolated from approximately fragment (5'—3'): 50 mg of tissue, homogenized in 300 1uL of CTAB buffer (100 mrvi Tris-HCI, pH 8.0, 1.4 M NaCl, 20 mM TTTTATGTGGTGGTATTGTTGAGG EDTA, 2 per cent CTAB, 0.2 per cent 2-mercapto- ATAGGATAGTCTGGAATACGACGC ethanol), then incubated at 65°C for 30—60 mm after the addition of 15 uL of 10 mg mL proteinase K. The lysate was then extracted once with chloroform— Amplificationand restriction digestion of PCR isoamyl alcohol (24:1), once with phenol—chioro- products form—isoamyl alcohol (24:24:1) and once with Amplificationusing the Pmal primer pair was chioroform—isoamyl alcohol. DNA was precipitated undertaken in a 50 jiL total volume of 50 mivi KCI, with 0.1 volumes of 3 M NaOAc and 1 volume of 10 mM Tris-HCI (pH 9.0), 2.5 mrvi MgC12, 0.1 per isopropanol, air-dried and resuspended in 100 uL of cent (v/v) Triton X-100, 200 M each dNTP, 25 pmol dH2O. of each primer and 1 U Taq (Promega). Cycle para- meters of 3 mm at 94°C, followed by 35 cycles of 45 s at 94°C, 30 s at 52°C and 2 mm at 72°C with a Cloning,sequencing and primer design final extension time of 7 mm at 72°C were used. The Anumber of restriction enzymes were applied to R Pma2 fragment was amplified in a 50 jiL volume maximus mtDNA in order to identify fragments suit- with 50 mrvi KC1, 10 m Tris-HC1 (pH 8.8), 2.5 mrvi able for cloning, sequencing and primer design MgCl2, 0.1 per cent (v/v) Triton X-100, 200 /1M each (Wilding, 1996). Two fragments of mtDNA, which dNTP, 25 pmol of each primer, 1 U Tbr polymerase showed no evidence of length variation after restric- (NBL) and 1 U Taq extender (Stratagene). Cycle tion digest screening of a number of individuals, parameters of 3 mm at 94°C and 40 cycles of 1 mm were chosen for sequencing; a 2 kb EcoRl—Hindlil at 94°C, 1 mm at 56°C and 5 mm at 72°C with a final fragment and a 3.85kbHindIII—HindIII fragment. extension time of 10 mm at 72°C were used. DNA MtDNA for ligation was purified as above from a templates consisted of 1 ,uL of a 1:10—1:100 dilution single except that two bouts of sucrose of mtDNA or 1 pL (100—500 ng) of genomic DNA gradient centrifugation (Snyder et al., 1987) were (gDNA). Restriction digestion was performed on 5 used to ensure complete separation of mitochondria uL aliquots of successful amplifications in 20 uL from nuclei. Restriction enzyme-treated mtDNA was total volume using 1—2 U of enzyme. The enzymes, ligated into pBluescript KS (Stratagene) using AluI, CfoI, Dral, HaeIII, Hinfl, MspI, RsaI, Sau3AI standard protocols (Sambrook et a!., 1989) and used and TaqI, were used in trial screenings to assess to transform competent XL1-blue E. coli. Recombi- variability. Digestion products were separated on 1.4 nant colonies were identified using fJ-galactosidase per cent agarose gels. complementation, and inserts of the correct size were detected via restriction enzyme digestion of mini-prepped DNA (Holmes & Quigley, 1981). Statisticalanalysis Vector DNA containing the desired insert was Foreach individual, a composite haplotype further purified using Promega's Wizard mini-preps, described the patterns generated by multiple restric- and the ends were sequenced using Pharmacia's T7 tion endonucleases and, for each RFLP pattern, the sequencing kit following the manufacturer's specifi- type of site change responsible could be ascertained cations. PRIMER (The Lander Laboratory, from digestion patterns and attributed to one, or Cambridge, MA, U.S.A.) was used to identify PCR occasionally two, site changes. Restriction mapping primers from these sequences, accepting default was not performed. Using information on the conditions except primer length (20—25) and melting number of composite haplotypes within each popu-

The Genetical Society of Great Britain, Heredity, 79, 178—189. MITOCHONDRIAL DNA VARIATION IN PECTEN MAX/MUS 181 lation and a presence/absence matrix of restriction (CfoI, Hinfl, MspI, RsaI, Sau3AI and TaqI). Seven sites, data were analysed using REAP(McElroyet al., enzymes produced polymorphic patterns when 1992). The DA option was used to compute haplo- screened against the Pma2 fragment (DraI, HaeIII, type diversity and nucleotide diversity both within Hinfl, MspI, RsaI, Sau3AI and TaqI). RFLP patterns populations and between populations, corrected for are detailed in Wilding (1996). No differences were within-population diversity (Nei, 1987). Nucleotide detected in RFLP patterns generated from ampli- diversity values were used as input for the SAHN and fication using gDNA or pure mtDNA templates, TREE options of NTSYS (Rohif, 1990) in order to suggesting that nuclear copies of the studied construct a UPGMA dendrogram. Sequence diver- mtDNA regions (Zhang & Hewitt, 1996) had not gence estimates between haplotypes were used as been amplified. For the Pmal-amplified fragment, a input for NEIGHBOR in PHYLIP to produce a UPGMA total of 30 variable restriction sites were recognized dendrogram of haplotype relatedness. The MONTE and 59 different composite haplotypes were identi- option of REAP was used to test for heterogeneity of fied (Table 1). Although screening with the Pma2 haplotype numbers via x2usinga Monte Carlo primer pair was restricted, variation within the procedure with 10000 randomizations, following the limited sample screened was also high, with 36 method of Roff & Bentzen (1989) to avoid problems composite haplotypes identified (Table 2). Screening associated with empty cells and small sample sizes. was more extensive for the Pmal fragment, and data Chi-squared tests were performed over all data, in from this fragment are concentrated on. However, pairwise comparisons of populations and on a single the limited quantity of data for the Pma2 fragment enzyme haplotype basis. 0, a haploid analogue of did not alter the conclusions from these data. F1, was calculated using FSTAT (Goudet, 1995). Although there is variation in the frequency of Pmal composite haplotypes (Table 1), for example thefrequency of the common haplotype Results Tworegions of P maximus mtDNA were successfully cloned, a 2 kb EcoRI—HindIII fragment and a 3.85 kb Hindlil—Hindlil fragment, from which sequence was obtained and primers designed to facilitate PCR amplification. The location of the 3.85 kb Hindlil— Hindlil fragment is evident from the restriction map of Rigaa et al. (1993). Using NsiI, the recognition sequence of which is within the sequence identified from forward primed sequencing of the 2 kb EcoRI— b Hindill fragment, and which cleaves P maximus mtDNA only once (Wilding, 1996), the location and orientation of this fragment was also confirmed (Fig. 2). Neither of the fragments, which proved to be contiguous, extends into the length-variable domain noted by Rigaa et al. (1993). These primers were applied under optimized conditions to both pure mtDNA and crude genomic DNA isolated from scallops from all sampling sites. There was no evidence of length variation in ampli- fied DNA. Amplification with Pmal was consistently successful. However, after initial success, amplifica- tion using Pma2 later failed and could not be resur- Fig. 2 Circular restriction map of Pecten maximus mtDNA as published by Rigaa et al. (1993) with additional infor- rected, in spite of further optimization, purchase of mation on the location of the 2 kb and 3.85 kb fragments new primers or change of manufacturers of Taq or isolated in this study. Ac, AccI; Av, AvaI; Bg, BglII; E, dNTPs. EcoRI (only one site mapped); H, Hindill; N, NsiI. Open region: 3.85 kb HindIII—HindIII fragment. Heavy shading: Restriction 2 kb EcoRI—HindIIT fragment. F, location of Pmal analysis of amplified products primer designed from forward-primed sequence. R, loca- Sixof the nine enzymes screened against the Pmal tion of Pmal primer designed from reverse-primed product identified polymorphisms at the 0.95 level sequence. Inner arc indicates length-variable domain.

The Genetical Society of Great Britain, Heredity, 79, 178—189. 182 C.S. WILDING ETAL.

(AAAAAA) varies widely between sample sites, between populations using x2analysis,either table- accounting for between 25 per cent and 53 per cent wise (37.3 per cent of pseudo 2-values from the of all composite haplotypes, no significant differ- randomized data were higher than those generated ences in haplotype numbers could be detected from the real data), in pairwise comparisons of

Table 1 Pmal-amplified mitochondrial DNA haplotypes in 12 populations of Pecten maximus

Haplotype 'a;r U STE KU.. 1VIIY CHI DOU PEE STO MUL LYM POL AAAAAA 8 8 17 14 15 8 13 11 9 12 16 8 AAAAAR 1 AAAABA 2 2 1 1 3 1 1 AAAACA 2 2 2 3 1 1 1 4 AAAACB AAAADA I AAAABA I I I AAAAFA AAAAGA I AAAABB I AABAAI3 8 4 5 7 2 6 3 5 4 7 9 8 AABAAA 3 4 1 2 4 3 2 2 1 1 AAABAA I I I I 2 I I AAABCB AAACHA I I AAACCA AAAGEA AAAGCA 1 2 AAAGAA AAAGBA I AAAFCC I AAACAA 1 AAACBA AABDCA AABABB I I 1 1 2 AABFAB I AABACB 1 I I AABAAD AADW I I I I I I I I AABBAE I I I 2 1 1 2 AACGCA I AABABA I MBLA I AABEAB I AADBAB I AABGAB I AAEAAA 1 AAFAAA I AAFACA I ABAAAA I AIIBACA I AFBAAA I ACAAAA I I I ACABEA I ACI3AAB I ACAGCA I ACACCA 1 ADBAAB 1 2 AEAAAA I

BAAAAA 1 1 1 1 BABAAA I BAI3AAB I I BCAMA I BCAACA I BDMCA I CAAAAA I I i CA13AAA I CAHAAB I N 29 30 32 30 30 30 30 31 24 30 30 32 Hap. Div 0.8522 0.9011 0.7056 0.7425 0.7471 0.8851 0.8 0.8538 0.8442 0.7931 0.6414 0.871 ND. Frag 0.0148 0.01590.0132 0.0128 0.0107 0.0153 0.0109 0.01380.0168 0.01460.0092 0.0153 N.D. Site 0.0238 0.0256 0.021 0.02 0.0184 0.0243 0.01 84 0.023 0.0265 0.0235 0.0145 0.0249 Population abbreviations as in Fig. 1. Order of restriction enzymes in composite haplotypes: Cfol, Hinfl, MspI, RsaI, Sau3AI and TaqI.

The Genetical Society of Great Britain, Heredity, 79, 178—189. MITOCHONDRIAL DNA VARIATION IN PECTENMAXIMUS 183 populationsor from single-enzyme haplotype data, Bonferroni procedure. Some haplotypes are when calculated values were adjusted via the confined to particular populations ('private haplo-

Table 2 Pma2-amplified mitochondrial DNA haplotypes in six populations of Pecten maximus Haplotype CIII POL PEE KIL STB MUL AAAAAA 3 6 4 6 3 AAAACA 2 1 1 6 4

AABAAA 1 1 1 BFAABA 1 1 ACABAA 1 BBAABA 3 4 1 2 3 4 AAAEAC 1 BEAABA 1 1 AAAAFA 1 1 AAAGCA 1 AAABAA 1 1 DAAACA 1 AEAHAA I AACAAA 1 AABGAA 1 AAAJAA 1 AAAAEA 1 AAABCA 1 CBAABA 1 1 BBBADA 1 AACACA 1 AGAAAA 1 AAACBA 1 AAAFEA 1 BEAABC 1 AAAACB 1 AABACA 2 1 1 AAADHA 1 BBAAAA 1

ABAACA 1 1

BMBAAA 1 ANEACA 1 ADAACA 1 BHAABA I ADAAAA 1 BBAADA N 18 14 6 16 23 20 Hap. Div. 0.9477 0,7692 1 0.9417 0.8696 0.9211 N.D. 0.01089 0.01348 0.01348 0.01118 0.00802 0,00992 Population abbreviations as in Fig. 1. Order of restriction enzymes in composite haplotypes: DraI,HaeIII,Hinfl, MspI, RsaI and Sau3AI.

The Genetical Society of Great Britain, Heredity, 79, 178—189. 184 C. S. WILDING ETAL. types') but cannot be taken as evidence of popula- tion) and extensive nucleotide diversity within popu- tion structure, only as an indication of considerable lations (Tables 1 and 2). Nucleotide diversities diversity. A UPGMA dendrogram of haplotype rela- (corrected for within-population diversity below the tedness did not provide any evidence of geographi- diagonal) are shown in Tables 3 and 4 for the two cal structure (Fig. 3). fragments. A number of (corrected) pairwise nucleo- Variation was extensive for the Pmal-amplified tide divergence values are negative, indicating that fragment with high values of haplotype diversity (the within-population diversity is greater than between- probability of encountering different haplotypes population differences, and thus there is little separ- when two individuals are sampled from a popula- ation of populations on the basis of sequence

El4EUtOX, 2 • .. S•SS S SI • S. S . S. S I • S. I . S S II. I S S

I • I S S S IS .I...I•.•.. SI

.. S •IS S .IS. S S ISSI...... •...•.... • I..•.. . S S.SSISS••1 I I. S S Fig. 3 UPGMA dendrogram generated I. from sequence divergence estimates . for all Pecten maximus haplotypes. S Presence of the haplotype in each I population is indicated. For details of population abbreviations, see Fig. I.

The Genetical Society of Great Britain, Heredity, 79, 178—189. MITOCHONDRIAL DNA VARIATION IN PEC TEN MA XIM US 185

divergence. However, figures generated from Pmal o ' .-4 00 fl 0 fl 0 N N fl fl f) 00 fl data in pairwise comparisons involving Muiroy Bay 000 0 C') C') C') C') 0 C')0 C') C') N N 0 0 are consistently positive, indicating that between- NNC')NN N N N J N N ooc 00000000 population differences are greater than within-popu- 00000000000 lationdiversity. A UPGMA dendrogram was 0(')00000000 C')CO constructed using the Pmal-derived nucleotide 0000N000000 divergence data, correcting negative (artefactual) 0000000000 0 values to zero but, as a consequence of the multiple 00000 zero values in the matrix, this produced a number of 0000NOC')ONfl000)D NO W equally likely trees. The consensus tree (Fig. 4) indi- 000 0 N)-)00NN C')0N 0 C') 000 ,-1 C )NN 00 cates that only the Muiroy Bay sample is separable, 000000000 00 is en but that this separation clearcut.Trees Ce 000 constructed from Pma2-derived values do not add N'-)ONO)OON 00C') 00')000,')0 N0-10r) further resolution (figure not shown). From the C) 0)C')NN00N0C')0 000 0CCJNN 00000 000 000 Pmal-derived data, an estimate of FST (0)of0.005 was calculated. o0o0c0 000 C" Ce fl C') C C Or- C') 0)0) C'), C') 00 r- C C) Ce r)00).C') 0 cm0 C') 00 ,-,-iC') 0 0000+ Discussion 0000000NN N N N <') N 09000000 C.) -C 0000 Thepresence of considerable length variation in P C) maximus mtDNA causes substantial difficulties for 0 N 000 N 0 N0ON00O 0 N C .-1COO C')N '-4 N CC)(0' N the interpretation of RFLP patterns from whole S N .-)00(0 -4 N 00000 Cd, 000000N 00000 mtDNA (Rigaa et a!., 1993; Wilding, 1996). The C 000 0000C.C application of PCR—RFLP circumvents these diffi- C.) N0NN Nr-C ' CC)CCN N0CC) 0) N fl C') CO culties and, in addition, potentially permits the 0 , N 0 fl N -')N ii C.) cC)0'N C')N 000000 analysis of DNA extracted from minute quantities of NN N N (N 000000 Cd) 00000o00c0•0, tissue, perhaps even larvae. Here, primers are -c 000000000I I presented that permit reliable amplification of a -C I) 0 400 40 C')CC) 00 N (0 C') 0)N 40 N highly polymorphic section of mtDNA. MtDNA of P Ce C')0(00 0 CC) NC') flN V., r')N CC 0 00 0 0 C') maximus studied by this method proved to be C NC')'C N 0000000 C.) oooooo Ce 00.0, considerably polymorphic, with variation within C.) doooodooooo populations often as great as or greater than that OC') C') N CO 0)400 NC') C').-4 '1 C) ,-4 C' between sampling sites and numerous private haplo- 0)C') ,-4 0"1 0C')') 0 ,-40000000 types (haplotypes confined to single populations). ocON N N oooOOcC00000000 coo0000000 The high values of haplotype diversity reflect this III I I I r. variation and, although these figures may not be 0.0 oNor-' c'j fl 0 C) N N N strictly comparable with values based on data from NNNfl 000000000000000000 whole mtDNA digests, the diversity levels in this 00 000000000 study are of a similar level and distribution to other 0 o000000I III 0 0 0 0 1 0 0) marine species. The observed pattern of two . 00 N (I '-1 N 0 fl NN fl00000000000 fl ,-NN common haplotypes dominating the distribution and 0N 0000000CC0000000000 multiple, rarer or private haplotypes has been recog- g d 0000000000I'll nized in both Pacific (Reeb & Avise, 1990; Ce fl 0 N N r-00 N 000 Boom et a!., 1994) and pollock (Shields & Gust, 0C) N0N00O0O 1995) and been attributed to a variety of possible C 000N0000000 Ce C.) 00000000000 C C causal mechanisms (Shields & Gust, 1995). No diag- C) Co 00000000000 Ce I I I II nostic haplotypes or high-frequency haplotypes -C confined to any particular location were seen, and i2 C) C.) C analysis failed to reveal any significant structure, nor -C CC was there any evidence of clustering of haplotypes 0C) with respect to area, suggestive of little phylogeo- C.) C- graphical structure. When allied to the low 0(FsT) C- 0C estimate, the mtDNA data appear to be consistent C with the conclusion from allozyme data (Beaumont

The Genetical Society of Great Britain, Heredity, 79, 178—189. 186 C. S. WILDING ETAL.

Table4 Nucleotide divergence among Pecten mdximus populations calculated by site methods from the Pma2-amplified fragment

1 2 3 4 5 6 iChickenRock 0.009509 0.0116090.0108470.009500 0.010348 2Polpeiro 0.000081 0.0099070.0091370.0080020.008592 3Peel —0.000575—0.000820 0.0115430.0100100.010833 4KilkieranBay —0.000186—0.000438—0.000789 0.0097170.010177 5St.Brieuc 0.0000470.000005 —0.0007430.000115 0.008885 6Mull —0.000053—0.000352—0.000867—0.000372—0.000085

Uncorrected above diagonal (d) and below diagonal corrected for within-population diversity [a—(a+a)/2]. et al., 1993) that P maximus conform to panmixia. the individuals within this population and genetic Indeed, the low FST estimate is typical of species drift acting on haplotype frequencies. MtDNA has with no observed genetic heterogeneity and far increased susceptibility to genetic drift as a result of lower than figures at which restrictions to gene flow the lowered effective population size (half that of are detected, for example an FSTvalueof 0.027 nuclear genes for such a hermaphroditic species), calculated from allozyme data in the which could explain why mtDNA reveals population vulgata was taken as evidence of genetic homo- structure not uncovered by allozyme (nuclear) loci. geneity, whereas 0 of 0.427 from mtDNA was indica- Muiroy Bay is a semi-enclosed sea lough with two tive of substantial population differentiation (Hurst entrance sills acting to reduce water movement in & Skibinski, 1995). Thus, at the haplotype level, and out of the basin. This is probably an effective there is little evidence of genetic heterogeneity containment mechanism for larvae, which will among P maximus populations. However, when the promote genetic differentiation. The shell colour of data are analysed at the nucleotide level, then for I? maximus juveniles has previously been suggested the Muiroy Bay sample the corrected nucleotide as indicating that Mulroy Bay scallops may indeed divergence was consistently positive in all pairwise be variant (Minchin, 1991). Although the mtDNA comparisons (Table 3). Such a finding is indicative divergence values add support to this hypothesis, the of genetic divergence (Ovenden, 1990). The Mulroy Bay sample in this study, which came from a suggested differentiation of the Muiroy Bay popula- single spatfall ongrown elsewhere, may not be repre- tion could be explained by reproductive isolation of sentative of the population as a whole, whether because of temporal effects (variation in genetic composition of recruits between settlements) or postcollection mortality effects. Temporal variation Muiroy Bay — in the composition of recruits has been documented —Polperro in other marine species (Watts et al., 1990; Hedge- LymeBay cock, 1994). Convincing evidence that Muiroy Bay —Mull scallops are self-recruiting will only be possible Stonehaven through the analysis of a sample of native adults. —Peel The possible self-recruiting nature of the Mulroy — Bay population has implications for the management Douglas —ChickenRock of the population because, although Mulroy Bay P — maximus are no longer heavily fished (Minchin, pers. KilkieranBay comm., 1983), the population has been proposed as St. Brieuc a suitable site for the collection of scallop spat for — Radede Brest reseeding efforts (Slater, pers. comm., 1995). —LaTrinité The demonstration of substantial sequence diverg- Fig. 4 Consensus dendrogram for Pecten maximus popula- ence between populations of I? maximus represents tions calculated from all equally likely UPGMA dendro- the first genetic evidence that some populationsmay grams using clustered nucleotide divergence values be, to some extent, reproductively isolated. Muiroy calculated from PCR—RFLP data on the Pmal-amplified Bay P maximus have not previously been analysed at fragment. The average sequence divergence of the Muiroy the molecular level, but our own studies using allo- Bay sample from all other samples is 0.107 per cent. zyme electrophoresis (C. S. Wilding et al., in

The Genetical Society of Great Britain, Hered/ly, 79, 178—189. MITOCHONDRIAL DNA VARIATION IN PECTEN MAXIMUS 187 preparation) provide little evidence for the distinct- underlying differentiation does exist and that further ness of this population on the basis of allozyme study of St Brieuc P maximus may prove productive. frequencies. Discrepancies between results based on protein electrophoresis and DNA-based markers, either in the ability to separate populations or in the Acknowledgements magnitude of differentiation, have previously been Weare grateful to Dr A. Brand (Port Erin Marine recognized in a number of species, for example Laboratory), Dr G. Burnell (University College, Drosophila melanogaster (Begun & Aquadro, 1993), Cork), Dr P. Dare (MAFF, Lowestoft), Dr P. G. Mytilus galloprovincialis (Karakousis & Fleury (IFREMER, Brest) and T. Howell (SOAFD, Skibinski, 1992) and the American Aberdeen) for provision of scallops. C.S.W. was virginica. In this species, Atlantic and Gulf of supported by NERC grant GT4/92/275 A. Mexico populations appear genetically homogeneous for allozyme markers but exhibit a pronounced genetic break for mtDNA and anonymous single- References copy nuclear DNA markers (Karl & Avise, 1992). A. D., DAO, J.-C. AND MASON, i. 1991. Three Euro- This was attributed to balancing selection maintain- ANSELL, pean scallops: Pecten maximus, Chiamys (Aequipecten) ing similar allele frequencies in populations experi- opercularis and C. (Chiamys) varia. In: Shumway, S. E. encing little gene flow. Such a mechanism may be an (ed.) Scallops: Biology, Ecology and Aquaculture, pp. alternative explanation for why P maximus appear 715—751. Elsevier, Amsterdam. panmictic on the basis of allozyme data, despite BARTLETT, S. E. AND DAVIDSON, W. s. 1995. Genetic analy- possible reproductive isolation or restrictions to sis of Salmo using mtDNA AMPFLPs. Aquaculture, gene flow. 137,47—48. The lack of significant differentiation of the St BEAUMONT, A. R. 1982. Geographic variation in allele Brieuc scallops at the mtDNA level is surprising, if frequencies at three loci in Chiamys opercularis from this population is indeed reproductively isolated as Norway to the Brittany coast. .1. Mar Biol. Ass. U.K., 62, 243—261. has previously been suggested (Paulet et al., 1988; BEAUMONT, A. R. AND BARNES, D. A. 1992. Aspects of Cochard & Devauchelle, 1993; Mackie & Ansell, veliger larval growth and byssus drifting of the spat of 1993). If population differentiation is present, then Pecten maximus and Aequipecten (Chiamys) opercularis. mtDNA is expected to reveal more differences than ICES J. mar Sci., 49,417—423. allozymes owing to high variability and lowered BEAUMONT, A. R., MORVAN, C., HUELVAN, S., LUCAS, A. AND effective population size, and failure to establish that AN5ELL, A. D. 1993. Genetics of indigenous and trans- differences are present questions whether the popu- planted populations of Pecten maximus: no evidence for lation is indeed genetically different. If the repro- the existence of separate stocks. J. Exp. Mar Biol. Ecol., ductive data are truly an indication of genetic 169,77—88. isolation, then the suitability of the marker is BEGUN, D. J. AND AQUADRO, C. F. 1993. African and North American populations of Drosophila melanogaster are brought into question and analysis with alternative very different at the DNA level. Nature, 365, 548—550. systems, e.g. microsatellites, may prove fruitful. It is BOOM, J. D. G., BOULDING, E. G. AND BECKENBACH, A. T. noteworthy that variation in the frequency of haplo- 1994. Mitochondrial DNA variation in introduced type AAAAAA for Pmal is similar to that of PT-A populations of , Crassostrea gigas, in allele frequencies in A. opercularis, in which substan- British Columbia. Can.J.Fish. Aquat. Sci., 51, tial population subdivision, including separation of 1608—1614. the St Brieuc population, is observed (Beaumont, BOULDING, E. G., BOOM, J. D. G. AND BECKENBACH, A. T. 1982). Thus, the haplotype frequency at St Brieuc 1993. Genetic variation in one bottlenecked and two appears to be more similar to the Irish Sea popula- wild populations of the Japanese scallop (Patinopecten tions than to the neighbouring Rade de Brest popu- yessoensis): empirical parameter estimates from coding lation, which is itself similar to the sample from La regions of mitochondrial DNA. Can. .1. Fish. Aquat. Sci., 50,1147—1157. Trinité (for queen scallops, a population from BRAND, A. R. 1991. Scallop ecology: distributions and nearby Belle Isle was studied). All Irish Sea samples behaviour. In: Shumway, S.E. (ed.)Scallops: Biology, are similar (although Chicken Rock appears to have Ecology and Aquaculture, pp. 517—584. Elsevier, a lower frequency of this haplotype). There are, Amsterdam. however, no great differences between Irish Sea ci-jow, s. AND INOUE, s. 1993. Intra- and interspecific populations and those from the west coasts of restriction fragment length polymorphism in mitochon- and Scotland. This similarity in pattern is by drial genes of Thunnus tuna species. Bull. Nati. Res. no means convincing, but it may be indicative that Inst. Far Seas Fish., 30, 207—225.

The Genetical Society of Great Britain, Heredity, 79, 178—189. 188 C.S. WILD1NG ETAL.

COCHARD, J. C. AND DEVAUCHELLE, N. 1993. Spawning, lations of Pecten maximus: evidence for the existence of fecundity and larval survival and growth in relation to separate stocks. J. Exp. Mar Biol. Ecol., 169, 57—75. controlled conditioning in native and transplanted MASON, i. 1983. Scallop and Queen Fisheries in the British populations of Pecten maximus (L.): evidence for the Isles. Fishing News Books, Farnham, England. existence of separate stocks, .1. Exp.MarBiol. Ecol., MINCHIN, D. 1991. Observations on shell colour in the 169, 41—56. scallop, Pecten maximus (L.). J. Conchol., 34, 41—46. CRONIN, M. A., SPEARMAN, W. J., WILMOT, R. L., PATTON, J. MICrON, J. B. 1994. Molecular approaches to population C. AND BICKHAM, j.w.1993. Mitochondrial DNA varia- biology. Ann. Rev. Ecol. Syst., 25, 45—69. tion in chinook (Oncorhynchus tshawytscha) and chum MORITZ, C., DOWLING, T. E. AND BROWN, W. M. 1987. Evolu- salmon (0. keta) detected by restriction enzyme analysis tion of animal mitochondrial DNA: relevance for popu- of polymerase chain reaction (PCR) products. Can. J. lation biology and systematics. Ann. Rev. Ecol. Syst., 18, Fish. Aquat. Sci., 50, 708—715. 269—292. CRONIN, M. A., HILLS, S., BORN, E. W. AND PA1TON, J. C. NE!,M. 1987.Molecular Evolutiona,y Genetics. Columbia 1994. Mitochondrial DNA variation in Atlantic and University Press, New York. Pacific walruses. Can. J. Zoo!., 72, 1035—1043. OVENDEN, j.R.1990. Mitochondrial DNA and marine DARE, P. J., DARBY, C. D., DURANCE, J. A., AND PALMER, 0. stock assessment: a review. Aust. J. Mar Freshw. Res., w. 1994. The distribution of scallops (Pecten maximus) 41, 835—853. in the English Channel and Celtic Sea in relation to PAVLET, Y. M., LUCAS, A. AND GERARD, A. 1988. Reproduc- hydrographic and substrate features affecting larval tion and larval development in two Pecten maximus (L.) dispersal and settlement. In: Bourne, N. F., Bunting, B. populations from Brittany. .1. Exp. Mar Biol. Ecol., 119, L. and Townsend L. D. (eds) Proceedings of the 9th 145—156. InternationalPectinid Workshop, Nanaimo, B. C., REEB, C. A. AND AVISE, .i. c. 1990. A genetic discontinuity Canada, 22—27 April 1993. Can. Tech. Rep. Fish. Aquat. in a continuously distributed species: mitochondrial Sci., 2, 20—27. DNA in the American oyster, Crassostrea virginica. FISHER, C. AND SKIBINSKI, D. 0. F. 1990. Sex-biased mito- Genetics, 124, 397—406. chondrial DNA heteroplasmy in the marine REPIN, P. V. AND BRYKOV, V. A. 1992. Mitochondrial DNA Mytilus. Proc. R. Soc. B, 242, 149—156. from sea scallops (Pectinidae) with unusual properties. GJETVAJ, B., COOK, D. I. AND ZOUROS, E. 1992. Repeated Genetika, 29, 460—467. sequences and large-scale size variation of mitochon- RIGAA, A., LE GAL, Y., SELLOS, D. AND MONNEROT, M. 1993. drial DNA: a common feature among scallops (Bival- Mapping and repeated sequence organization of mito- via: Pectinidae). Mol. Biol. Evol., 9, 106—124. chondrial DNA in scallop, Pecten maximus. Mol. Mar GOUDET, j.1995.FSTAT (Version 1.2). A computer Biol. Biotechnol., 2, 218—224. program to calculate F-statistics. J. Hered., 86, 485—486. RIGAA, A., MONNEROT, M. AND SELLOS, D. 1995. Molecular HALL, H. J. AND NAWROCKI, L. w. 1995. A rapid method for cloning and complete nucleotide sequence of the detecting mitochondrial DNA variation in the brown repeated unit and flanking gene of the scallop Pecten trout, Salmo trutta. J. Fish. Biol., 46, 360—364. maximus mitochondrial DNA: putative replication HEDGECOCK, D. 1994. Temporal and spatial genetic struc- origin features. J. Mol. Evol., 41, 189—195. ture of marine animal populations in the California ROFF, ii A. AND BENTZEN, i'. 1989. The statistical analysis current. CALCOFI Rep., 35, 73—81. of mitochondrial DNA polymorphisms: x2andthe HOLMES, D. S. AND QUIGLEY, M. 1981. A rapid boiling problem of small samples. Mo!. Biol. Evol., 6, 539—545. method for the preparation of bacterial plasmids. RO1-ILF, F. j.1990.NTSYS-PC: Numerical and Analyt. Biochem., 114,193—197. Multivariate Analysis System. Exeter Publishing, Setau- HURST, C. D. AND SKIBINSKI, 13. 0. F. 1995. Comparison of ket, NY. allozyme and mitochondrial DNA spatial differentiation SAMBROOK, J., FRITSCH, E. F. AND MANIATIS, T. 1989. Molec- in the limpet Patella vulgata. Mar Biol., 122, 257—263. ular Cloning: A Laboratoiy Manual, 2nd edn. Cold KARAKOUSIS, Y. AND SKJBINSKI, D. 0. F. 1992. An analysis Spring Harbor Laboratory Press, New York. of allozyme, mitochondrial DNA and morphological SHIELDS, G. F. AND GUST, J. R. 1995, Lack of geographic variation in mussel (Mytilus galloprovincialis) popula- structure in mitochondrial DNA sequences of Bering tions from Greece. Experientia, 48, 878—881. Sea walleye pollock, Theragra chacogramma. Mol. Mar KARL, S. A. AND AVISE, J. c. 1992. Balancing selection at Biol. Biotechnol., 4, 69—82. allozyme loci in oysters: implications from nuclear SINCLAIR, M., MOHN, R. K., ROBERT, G. AND RODDICK, D. L. RFLPs. Science, 256, 100—102. 1985. Considerations for the effective management of MCELROY, D., MORAN, P., BERMINGHAM, E. AND KORNFIELD, Atlantic scallop populations. Can. Tech. Rep. Fish. i. 1992. REAP: An integrated environment for the Aquat. Sci., 1382,1—113. manipulation and phylogenetic analysis of restriction SNYDER, M., FRASER, A. R., LAROCHE, J., GARTNER-KEPICAY, data. J. Hered., 83, 157—158. K. E. AND ZouRos, E.1987.Atypical mitochondrial MACKIE, L. A. AND ANSELL, A. D. 1993. Differences in DNA from the deep-sea scallop Placopecten magellani- reproductive ecology in natural and transplanted popu- cus. Proc. Nati. Acad. Sci. USA., 84, 7595—7599.

The Genetical Society of Great Britain, Heredity, 79, 178—189. MITOCHONDRIAL DNA VARIATION IN PEG TEN MA XIM US 189

WATTS, J., JOHNSON, M. S. AND BLACK, R. 1990. Effects of and Mitochondrial DNA Approaches. PhD thesis, recruitment on genetic patchiness in the urchin Echino- University of , Bangor. metra mathaei in Western Australia. Mar Biol., 105, ZHANG, D.-X. AND HEWITT, G. M. 1996. Nuclear integra- 145—161. tions: challenges for mitochondrial DNA markers. WILDING,C. S. 1996. Population Genetics of the Scallop Trends Ecol. Evol, 11,247—251. Pecten maximus: Morphologica4 Allozyme Electophoresis

The Genetical Society of Great Britain, Heredily, 79, 178—189.