A GENETIC ASSESSMENT OF BRITISH POPULATIONS OF THE SAND LIZARD (LA CERTA AGILIS)

Transcription

1 HERPETOLOGICAL JOURNAL, Vol 11, pp (2001) A GENETIC ASSESSMENT OF BRITISH POPULATIONS OF THE SAND LIZARD (LA CERTA AGILIS) TREVOR J C BEEBEE A D GRAHAM ROWE School of Biological Sciences, University of Sssex, Falmer, Brighton BN I 9QG We investigated sand lizard (Lacer/a agilis) poplations in Britain by genetic analysis across eight polymorphic microsatellite loci Genetic diversity as determined by mean expected heterozygosity was high in all three distinct regions where the species occrs (Dorset, Srrey and Merseyside), thogh allelic diversity was lower on Merseyside than in Srrey or Dorset There was significant genetic differentiation between poplations in all three of these widely separated zones, as jdged both by Fst and Rst estimators A genetic test for poplation bottlenecks confirmed that in at least two of the areas crrently inhabited, Srrey and Merseyside, L agilis has ndergone sbstantial recent declines The significance of these findings for sand lizard conservation is discssed Key words: sand lizards, Lacer/a, conservation, genetics, microsatellites INTRODUCTION The sand lizard Lacerta agilis is one of only six reptiles native to Britain It is also one of the two rarest, confined within recent historical times to three small areas of sitable habitat (Smith, I 95 1 ): the lowland heaths of Dorset and soth-west Hampshire, heathland in the Srrey Weald, and the coastal dnes of Merseyside and North Wales A major decline in the distribtion and abndance of L agilis has followed widespread losses of these critical habitats (eg Moore, 1962; Jackson, 1979; Corbett 1988; Webb 1990) Sand lizard poplations have also become increasingly fragmented within the three distribtion zones These conditions can lead to genetic depaperization and inbreeding depression, isses that have cased concern among conservation biologists in relation to a wide range of endangered species, inclding sand lizards (eg Frankham, 1996; Olsson, Gll berg & Tegelstrom, 1996; Gllberg, Olsson & Tegelstrom, 1999) Althogh sccessfl management practices have been developed for L agilis in Britain and the species is given maximm protection nder the law (Corbett & Tamarind 1979; Molton & Corbett, 1999), neither of these important developments will necessarily alleviate any conseqences of genetic impoverishment Polymorphic microsatellite loci are widely sed markers for the stdy of poplation genetics in the context of relatively short (ecological) time periods (eg Jame & Lagoda, 1996; Snncks, 2000) A site of L agilis microsatellites was recently developed by Gllberg, Tegelstrom & Olsson (1997) and sed to investigate the strctre of Swedish sand lizard poplations (Gllberg, Olsson & Tegelstrom, 1998) In this paper we describe a stdy of British sand lizards, sing those microsatellite markers to investigate genetic diversity and differentiation among animals living Correspondence: T J C Beebee, School of Biological Sciences, University of Sssex, Falmer, Brighton BN 1 9QG tjcbeebee@sssexack in the three areas where the species crrently exists in Britain MATERIALS AND METHODS SAMPLING Terminal digits from toes were obtained, nder licence, from lizards in all three areas of the British distribtion (Fig I) Toe clipping cases no significant damage to the animal and for many years has been sed as a marking procedre for L agilis Nevertheless, becase the species is both rare and endangered the sample sizes were small Seven Merseyside lizards, eight Dorset lizards and eleven Srrey lizards were toeclipped dring 1999 and the toes were stored in ethanol prior to DNA extraction The Merseyside lizards and some of the Srrey lizards were maintained in vivaria for captive breeding prposes at the time of sampling, bt all were wild-caght animals None were from sites where there had been previos releases of translocated lizards The Dorset animals came from two separate heathland sites (for from each) All the lizards were released immediately after sampling, either at the site of captre or back into vivaria MICROSA TELLITE ANALYSIS DNA was obtained from each toe by a standard proteinase K digestion, phenol-chloroform extraction and ethanol precipitation procedre ng DNA were sed in PCR assays of final volme 20 µi, otherwise as described by Gllberg et al (1997), sing a33p- da TP as a radioactive label and with separate primers for each of the microsatellite loci The PCR prodcts were electrophoresed throgh 6% polyacrylamide gels, sbjected to atoradiography and alleles were scored by reference to M 13 seqence markers all as described elsewhere (Rowe, Beebee & Brke, 1997) DATA ANALYSIS Conformance to Hardy-Weinberg eqilibrim and linkage eqilibrim between loci were tested sing the compter programs BIOSYS- 1 and GENEPOP 3 1

2 24 T J C BEEBEE AND G ROWE MERSEYSIDE (Swoffo rd & Selander, 1981; Raymond & Rosset, 1995) Genetic diversity indices, notably mean nmber of alleles per locs, percentage of loci polymorphic at the 95% criterion (P95), observed and expected (nbiased) heterozygosities (H0 and H, respectively) and Cavalli-Sforza chord (D) genetic distances (Cavalli Sforza & Edwards, 1967) were also estimated sing BIOSYS- 1 Genetic differentiation between poplations was measred by pairwise Fst (Weir & Cockerham, 1984) and Rst (Slatkin, 1995) sing the programs FST AT 12 and RSTCALC 2 1 respectively (Godet, 1999; Goodman, 1997) Recent poplation trends were investigated with the BOTTLENECK program (Piry, Likart & Comet, 1999) Randomization tests were carried ot sing the program RT version 2 1 (Manly 1997) sing 5000 randomizations in two-sample comparisons Other statistical analyses were performed sing the ST A TISTIX compter package after testing data for normality as appropriate 200 km FIG I Distribtion of L agilis in Britain showing sampling sites RESULTS Of the 10 microsatellite loci available fo r stdy, eight (La-I, -2, -3, -4, -5, -8, -9 and - JO) demonstrated consistent polymorphic banding patterns in British L 03 c :::J 2 Cii ]! < ci z c: ::!!: f<' e o6 2l :,, 05 c i:l 04 B c: f! 01 -;----, , 4 08 i:5 07 "E 0 :!:! 06 iu > D , FIG2 dependency of genetic estimators A: Allele nmber per locs Data are averages of five randomly selected grops of each sample size except 7 (for which there was only one possible grop in Merseyside) B: Mean expected heterozygosity, H, Data are averages of five randomly selected grops of each sample size except 7 (for which there was only one possible grop in Merseyside) C: Mean Fst Data are averages of five random comparisons between some of the randomly chosen grops sed in A and B, excepting 7 where there was jst one possible comparison D: Cavalli-Sforza Chord Distance (D) Data are averages of five random comparisons between some of the randomly chosen grops sed in A and B, excepting 7 where there was jst one possible comparison Solid circles: Dorset poplation (A & B) or Dorset x Merseyside comparison (C & D); Open circles: Merseyside poplation (A & 8) or Merseyside x Srrey comparison (C & D); Open sqares: Srrey poplation (A & B) or Dorset x Srrey comparison (C & D)

3 SAND LIZARD GENETICS 25 TABLE I Genetic variation among Lagilis poplations?95, percentage of polymorphic loci at the 95% criterion H mean obs rved heterozygosity; H,, mean expecte heterozygos1ty; SO, standard deviation Location Total alleles Dorset 39 Srrey 37 Merseyside 24 Mean no alleles/ locs±sd 488± ± ±057 p H H agilis la-6 yielded no PCR prodcts from British sand lizards while la- 7 prodced no prodcts from Swedish lizards (Gllberg et al I 997) or from British ones Dorset and Srrey lizards were polymorph ic at all eight loci whereas Merseyside lizards were monomorphic at La-3 and la-10 Only la-10 in Srrey lizards fa iled to conform with Hardy-Weinberg expectations and no sets of loci showed significant linkage diseqilibrim in any poplation The markers were therefore considered appropriate for investigating genetic diversity in the British sand lizard poplations Becase the sample sizes were so small it was important to test the effe cts of this factor on the varios genetic estimators To do th is, random samples of 3-7 individals were selected from each poplation and a range of genetic parameters estimated as a fnction of sample size (Fig 2) Mean allele nmbers per locs and genetic distance (D) both showed linear samplesize effects within the range available for analysis The absolte vales of these parameters were therefore meaningless in the present stdy, bt relative comparisons were nevertheless informative Ths Merseyside allele nmbers were consistently lower than those of Dorset and Srrey, whereas genetic distances were consistently similar between all three localities Randomization tests indicated that Merseyside allele nmbers were significantly lower than those in Srrey (P=0030) or in Dorset (P=0025) bt that there were no differences in this parameter between Srrey and Dorset (P=0356) By contrast, sample-size dependent trends were weak or non-existent for heterozygosity and differentiation estimators (H, and Fst respectively), a sitation which also held for Rst (data not shown) Estimates of the partitioning of genetic variation (pooling across all loci) indicated that inter regional diffe rentia- TABLE 2 Genetic differentiation of L agilis poplations The probability (P) that Fst or Rst is not significantly different from 0 Comparison Dorset x Srrey Dorset x Merseyside Srrey x Merseyside Fst (P) 0133 (<000 1) (<000 1) (<000 1) Rst (P) (002) 0295 (0009) 0288 (000 I) tion (mean Fst=O 191) somewhat exceeded variation within regions (mean Fis=0 1 16) Estimates of the sample-size independent parameters for the fll data set are shown in Tables I and 2 Taken toget er the data imply that genetic diversity in the Merseys 1de sand l zards was lower than in Srrey a d Dorset h ards, which were broadly similar, thogh differences m He were not statistically significant (Krskal -W llis statistic = , P=03437) Random1zatton tests of heterozygosity also failed to show any significant differences between reoions Genetic differentiation, however, was sbstanti;l between all three regions with both Fst and Rst vales significantly diffe rent from zero in all pairwise comparisons Application of a bottleneck test based on excess heterozygosity relative to allele nmbers (Cornet & Likart, 1996) spported the inference that there have been sbstantial recent declines in sand lizard nmbers Despite the fact that sample size was lower than that recommended for adeqate statistical power in this test, two of the three areas (Merseyside and Srrey) demonstrated significant heterozygote excess - indicative of bottlenecking - with P=0023 and P=0 027 respectively DISCUSSION Althogh sample sizes were small, microsatellite analysis across eight loci has provided sefl insights into the British sand lizard poplations Larger samples wold have permitted statistically more robst analyses, bt in or estimation these wold probably not have altered the main conclsions Genetic diversity was related to poplation size in the expected way, with the smallest and most isolated poplation, in Merseyside, demonstrating the lowest diversity of the three British regions Recent estimates of adlt sand lizard poplation sizes in Merseyside, Srrey and Dorset are in the region of , <I OOO and adlts respectively (Corbett, 1994; Wheeler, Simpson & Hoston, 1993) This relationship was nlike the sitation in Sweden where, srprisingly, no sch correlation between genetic diversity at microsatellite loci and poplation size was evident (Madsen et al, 2000) Reasons for th is difference are nknown, bt the Swedish stdy was at a finer level of scale than ors and there is clearly scope for more detailed analysis of British sand lizard genetics within each region One conseqence of heathland fragmentation, in particlar, may be a redction in the genetic diversity of sand lizards at a local level However, assming or data are representative of regional patterns, lizards in all three areas maintained sbstantial diversity at microsatellite loci Indeed, British l agilis compare favorably in terms of heterozygosity with Hngarian animals tested across the same loci (mean H,=070) and proved sbstantially more diverse than Swedish sand lizards with a mean H of 045 (Gllberg, Olsson & Tegelstrom, 1998) Sand lizards also make an interesting comparison with natterjack toads (Bfo ca/arnita), a species with a

4 26 T J C BEEBEE AND G ROWE similarly restricted distribtion in Britain (Beebee, 1977) Natterjacks on the Merseyside coast were also assayed across eight microsatellite loci and exhibited lower genetic variation than sand lizards in the same area, despite the fact that the crrent censs poplation size of natterjacks is at least tenfold larger than that of sand lizards (Corbett, 1994; Rowe, Beebee & Brke, 1998; 1999) Ths Merseyside natterjacks (with a sample size of 200) exhibited a mean of only 235 alleles per locs, a P95 of 625% and a mean He of 0295 These differences may stem from the very diffe rent poplation dynam ics of lizards and toads, with the latter ndergoing larger poplation flctations over short time periods (Beebee, Denton & Bckley, 1996) Unlike lizards, toads have a breeding system in which many individals in any particlar generation probably fail to reprodce sccessflly (Scribner, Arntzen & Brke, 1997) Effective poplation sizes (ie nmbers of animals reprodcing sccessflly averaged over mltiple generations) are therefore likely to be mch smaller in these amphibians, relative to censs sizes, than is the case with lizards Both of these featres are likely to impact on genetic diversity, althogh other reasons (sch as different mtation rates in toads and lizards) cold also accont for the interspecific differences observed The estimators of genetic differentiation in sand lizards revealed significant diffe rences between all three sample areas that are consistent with separation of the three regions at roghly the same time, presmably soon after postglacial colonization when forest development eliminated intervening open habitats (Vinc nt, 1990) These reslts also sggest that for conservation prposes poplations in the three regions sho ld e considered as distinct clades worthy of protect10n m their own right The genetic bottleneck tests indicated that there have been sbstantial recent declines of sand lizard effective poplation sizes in at least two of the three geographical regions This independent genetic assessment of the fate of British sand lizards accords with conclsions derived from direct field srvey (Corbett, 1994; Molton & Corbett, 1999) There can be little dobt that this species has responded dramatically to the extensive losses of, and damage to, its sensitive heathland and dne habitats, and that conservation measres for it are flly jstified ACKNOWLEDGEMENTS We thank the Herpetological Conservation Trst, especially Chris Davies, Nick Molton and Mike Preston for assistance with sampling, Madryn Lake and Inga Zeisset for assistance with the microsatellite analyses, Bryan Manly for advice abot randomization tests, and Tony Gent, Peter Rothery and an anonymos referee for helpfl sggestions The work was carried ot nder Home Office licence PPL 70/3950 REFERENCES Beebee, T J C (1977) Environmental change as a case of natterjack toad (Bfo calamita) declines in Britain Biological Conservation 11, Beebee, T J C, Denton, J S & Bckley, J (1996) Factors affecting poplation densities of adlt natterjack toads Bfo calamita in Britain Jornal of Applied Ecology 33, Cavalli-Sforza, L L & Edwards, A W F (1967) Phylogenetic analysis: models and estimation procedres Evoltion 21, Corbett, K F (1988) Distribtion and stats of the sand lizard Lacer/a agilis agilis in Britain Mertensiella 1, Corbett, K F ( 1994 ) Pilot stdy for sand lizard UK recovery programme English Natre, Peterborogh Corbett, K F & Tamarind, DL ( 1979) Conservation of the sand lizard, Lacer/a agilis, by habitat management British Jornal of Herpetology 5, Cornet, J M & Likart, G ( 1996) Description and power analysis of two tests for detecting recent poplation bottlenecks from allele freqency data Genetics 144, Frankham, R ( 1996) Relationship of genetic variation to poplation size in wildlife Conservation Biology 10, Goodman, S J ( 1997) RSTCALC: A collection of compter programs for calclating nbiased estimates of genetic differentiation and gene flow from microsatellite data and determining their significance Moleclar Ecology 6, Godet, J (1999) FSTA T, a program for IBM PC compatibles to calclate We ir & Cockerham 's (1984) estimators of F-statistics (version 12) Institte of Zoology and Animal Ecology, Lasanne University, Switzerland Gll berg, A, Olsson, M & Tegelstrom, H ( 1998) Colonisation, genetic diversity, and evoltion in the Swedish sand lizard, Lacer/a agilis (Reptilia, Sqamata) Biological Jornal of the Linnean Society 65, Gllberg, A, Olsson, M & Tegelstrom, H (1999) Evoltion in poplations of Swedish sand lizards: genetic differentiation and loss of variability revealed by mltilocs DNA fingerprinting Jornal of Evoltionary Biology 12, Gll berg, A, Tegelstrom, H & Olsson, M ( 1997) Microsatellites in the sand lizard (Lacer/a agilis): description, variation, inheritance, and applicabilty Biochemical Genetics 35, Jackson, H (1979) The decline of the sand lizard Lace rt a agilis L poplation on the sand dnes of the Merseyside coast, England Biological Conservation 16, Jame, P & Lagoda, P J L (1996) Microsatellites, from molecles to poplations and back Trends in Ecology & Evoltion 11,

5 SAND LIZARD GENETICS 27 Madsen, T, Olsson, M, Wittzell, H, Stille, B, Gllberg A, Shine, R, Andersson, S & Tegelstrom, H (2000) Poplation size and genetic diversity in sand lizards (Lacer/a agilis) and adders (Vipera bers) Biological Conservation 94, Manly, B F J ( 1997) RT: A program for randomisation testing Version 21 Western Ecosystem Technolog, Inc, Wyoming, USA Moore, N W (1962) The heaths of Dorset and their conservation Jornal of Ecology 60, Molton, N & Corbett, K F (1999) The Sand Lizard Conservation Handbook English Natre, Peterborogh Olsson, M, Gll berg, A & Tegelstrom, H ( 1996) Malformed offspring, sibling matings, and selection against inbreeding in the sand lizard (Lacer/a agilis) Jornal of Evoltionary Biology 9, Piry, S, Likart, G & Cornet, J M (1999) BOTTLENECK: a compter program for detecting recent redctions in the effective poplation size sing allele freqency data Jornal of Heredity 90, Raymond, M & Rosset, F (1995) GENEPOP (Version 12): poplation genetics software for exact tests and ecmenicism Jornal of Heredity 86, Rowe, G, Beebee, T J C & Brke, T ( 1997) PCR primers for polymorphic microsatellite loci in the anran amphibian Bfo calamita Moleclar Ecology 6, Rowe, G, Beebee, T J C & Brke, T ( 1998) Phylogeography of the natterjack toad Bfo calamita in Britain: genetic differentiation of native and translocated poplations Moleclar Ecology 7, Rowe, G, Beebee, T J C & Brke, T ( 1999) Microsatellite heterozygosity, fitness and demography in natterjack toads Bfo calamita Animal Conservation 2, cribner, K T, Arntzen, J W & Brke, T (1997) Effective nmber of breeding adlts in Bfo bfo estimated from age-specific variation at minisatellite loci Moleclar Ecology 6, Slatkin, M (1995) A measre of poplation sbdivision based on microsatellite allele freqencies Genetics mith, M A ( 1951 ) The British Amphibians and Reptiles Collins, London Snncks P (2000) Efficient genetic markers for poplation biology Trends in Ecology & Evoltion Swofford, D L & Selander, R B ( 1981 ) BIOSYS-1: A FORTRAN program for the comprehensive analysis of electrophoretic data in poplation genetics and systematics Jornal of Heredity 72, Vincent, P ( 1990) The Biogeography of the British Isles Rotledge, London Webb, N R ( 1990) Changes on the heath lands of Dorset, England, between 1978 and 1987 Biological Conservation 51, Weir, B S & Cockerham, C C (1984) Estimating F statistics for the analysis of poplation strctre Evoltion 38, Wheeler, D J, Simpson, D E & Hoston, J A (1993) In: The Sa nd Dnes of the Seft on Coast Ed D Atkinson & J Hoston National Msems and Galleries, Merseyside Accepted: I I