Adh Nucleotide Variation in Drosophila willistoni: High ReplacementPolymorphism in an Electrophoretically Monomorphic Protein

Elizabeth C. Griffith, Jeffrey R. Powell

Department of Biology, Yale University, New Haven, CT 06520-8104, USA

Received: 5 August 1996 / Accepted: 12 April 1997

Abstract. Drosophila willistoniwas the subject of in-tensive allozyme studies and the locus coding for alcoholdehydrogenase (Adh) was found to be virtually mono-morphic. DNA sequence analysis of 18 alleles through-out the distribution of the species has revealed six re-placement polymorphisms. The ratio of replacement tosilent polymorphisms is higher inD. willistoni than inany otherDrosophilaspecies studied forAdhnucleotidevariation. Also in contrast to other species, the variationin introns and noncoding DNA is about the same as in thecoding region. We speculate that both these differencesindicateD. willistoni has historically had a small popu-lation size possibly related to Pleistocene refugia in theNeotropics.

Key words: Drosophila willistoni —DNA polymor-phism — Alcohol dehydrogenase

Introduction

The Alcohol dehydrogenase (Adh) locus in Drosophilahas served as a paradigm for molecular population ge-netics since Kreitman’s (1983) path-breaking work.DNA sequence variation at theAdh locus has been stud-ied extensively withinD. melanogaster(Kreitman 1983;Hudson et al. 1987; Kreitman and Hudson 1991; Mc-Donald and Kreitman 1991; Laurie et al. 1991) andD.pseudoobscura(Schaeffer and Miller 1991, 1992a,1993), and to a lesser extent in other species of theme-lanogastersubgroup (McDonald and Kreitman 1991).

These species represent two of the three major lineagesin subgenusSophophora.Here we add information onAdhnucleotide polymorphism in the third major lineage,the willistoni group. Previously, we showed thatAdh inthewillistoni group displayed some unusual features: theloss of an intron and a significant shift in codon usage(Anderson et al. 1993).

Species of theDrosophila willistonigroup have beenanalyzed extensively for electrophoretic variation insoluble enzymes in natural populations (reviewed inAyala 1975). Variation at theAdh locus was surveyed inseveral of these studies. ADH was one of the least poly-morphic enzymes analyzed in thewillistoni group, andwas electrophoretically monomorphic in several popula-tions of D. willistoni, D. paulistorum,andD. tropicalis.In D. willistoni, a total of four different alleles weredetected in a sample size of nearly 5000 alleles, althoughthe most common allele had a frequency of about 0.994.Thus, by virtually all criteria, this protein can be consid-ered electrophoretically monomorphic.

Here we present the nucleotide sequence variation in18 D. willistoni Adh alleles, sampled from strains thatrepresent 10 populations distributed throughout the geo-graphic range of this species. These data reveal unusualpatterns of nucleotide polymorphism in coding versusnoncoding regions, and a high level of replacement poly-morphism.

Materials and Methods

DrosophilaStocks. Strains of Drosophilawere obtained from severalsources. Collection sites and sources ofDrosophilaare listed in Table1. Each strain is an isofemale line. The strains sampled in this studyCorrespondence to:J.R. Powell

J Mol Evol (1997) 45:232–237

© Springer-Verlag New York Inc. 1997

basically represent the geographic range of this species, ranging fromMexico to Porto Alegre, Brazil (on the Uruguay border), and fromLima, Peru in western South America to eastern Brazil, and one Ca-ribbean strain. Strains ofD. willistoni from Lima, Peru have previouslybeen identified as a subspecies,D. willistoni quechua,based on sterilityin male hybrids (Ayala 1973). However, crosses of the Lima strain westudied to knownD. w. willistoni in our lab produced fertile F1 males(J. Powell and J. Gleason, unpublished observations).

PCR Amplification.Total genomic DNA was prepared using stan-dard phenol–chloroform extraction followed by proteinase K digestion(Werman et al. 1990). A 1.3-kilobase (kb) region of theAdhgene wasamplified by polymerase chain reaction (PCR) from each strain usingoligonucleotide primers identical to highly conserved regions 58 and 38to the Adh coding region of theD. willistoni W4L genomic clone(Anderson et al. 1993). Oligonucleotide synthesis was done using anApplied Biosystems 391 DNA synthesizer. All of the oligonucleotideprimers used in PCR and sequencing reactions are listed in Table 2.PCR conditions were optimized for these primers, and the followingconditions were used. Onemg template DNA, 300 ng of each primer,1 unit Taq polymerase (Perkin Elmer Cetus), and 100mmol of eachdNTP were used per reaction, in a total volume of 100ml. The reactionbuffer was the same as that in the GeneAmp kit (Perkin Elmer Cetus)except for the MgCl2 concentration (50 mM KCl, 10 mM Tris-Cl, pH8.3, 30 mM MgCl2, 0.1% gelatin). Forty cycles of amplification werecarried out using a Perkin Elmer Cetus DNA Thermal Cycler under thefollowing conditions: 94°C, 1 min; 50° or 55°C, 1 min; 72°C, 1 min.PCR products were checked by electrophoresis on 2% agarose gels.

Cloning and Sequencing of PCR Products.Unpurified PCR prod-ucts were ligated into pCR1000 and pCRII vectors (Invitrogen) andtransformed into INVaF8 (Invitrogen) or DH5a (BRL) competent cells.Plasmid DNA for sequencing was obtained from either miniprepara-tions done with the Magic Minipreps DNA purification kit (Promega)or by large-scale plasmid preps, done by the polyethylene glycol/LiClprecipitation method described by Sambrook et al. (1989). DNA se-quencing was done by the dideoxy chain-termination method (Sangeret al. 1977) using Sequenase (USB). The sequences of primers used in

these experiments are shown in Table 2. M13 forward and reverseprimers were used as well, and were obtained from New EnglandBiolabs. The 5XINT primer was obtained from C. Anderson. Bothstrands of DNA were sequenced for all strains ofD. willistoni, but anindependent PCR product was sequenced only for those strains that hadamino acid replacement changes. Sequences were aligned by eye.

Results

A PCR amplified product of approximately 1.3 kb of theAdh region was obtained for each of the strains listed inTable 1. These PCR reactions always yielded a verystrong single band on an agarose gel. This region in-cluded the complete coding sequence of 759 bp, as wellas the 214-bp noncoding 58 DNA, a single 67 to 68-bpintron, and 127-bp 38 noncoding DNA. Both strands ofone PCR product were sequenced for each allele in thepresent study, unless there was an amino acid change, inwhich case an additional PCR product was cloned andsequenced. There were no sequence differences betweenindependent PCR products for these strains. The com-

Table 1. Strains ofD. willistoni used in this study, their original collection sites, and sources from which they were obtained

Strain Abbreviation Collection site SourceGenBankaccession no.

W4La wil4L Aguas do los Rios, Brazil J. Powell L08648W3La wil3L Aguas do los Rios, Brazil J. Powell U95253A57a wilA5 Aguas do los Rios, Brazil J. Powell U952540811.0 wil0 Santa Maria, Nicaragua Bowling Greenb U952510811.4 wil4 Cuernavaca, Mexico Bowling Green U95252Belize II wilB2 Belize F. Ayala U95256Belize VI wilB6 Belize F. Ayala U95257Cano Mora wilC Cano Mora, Costa Rica F. Ayala U95258Atlixco wilA Atlixco, Mexico F. Ayala U95255Guana wilG2 Guana Island P. Chabora U95259Lima wilL Lima, Peru F. Ayala U95260Manaus 1 wilM1 Manaus, Brazil V. Valente U95261Manaus 2 wilM2 Manaus, Brazil V. Valente U95262Manaus 3 wilM3 Manaus, Brazil V. Valente U95263Manaus 4 wilM4 Manaus, Brazil V. Valente U95264PA 1 wilPA1 Porto Alegre, Brazil V. Valente U95265PA 2 wilPA2 Porto Alegre, Brazil V. Valente U95266PA 3 wilPA3 Porto Alegre, Brazil V. Valente U95267

a Isofemale line; obtained as a genomic DNA preparation from C. Andersonb NationalDrosophilaSpecies Resource Center, Bowling Green State University, Bowling Green, Ohio, USA

Table 2. Sequences of oligonucleotide primers used for PCR ampli-fication and DNA sequencing

Primer Sequence (58–38) Use

ADH 1 TTAGTTGAGAAGAGAAGAGCC PCR amplificationADH 2 CGATTATCAAATCAGCCTTC PCR amplificationB3 TCTGTGACCGGTTTCAATGC sequencingB4 GGCCTTAACAAAGTTCTGGG sequencingB5 AACATCTTGGTGGCCGG sequencing5XINT ACAGCAATGGTAGCTC sequencing

233

plete nucleotide sequence and deduced amino acid se-quence of the wil4Adhgene was presented in Andersonet al. (1993) and will not be repeated here. Comparisonof these strains to other species of thewillistoni group aswell as the GenBank accession numbers are in Gleasonet al. (submitted).

Figure 1 illustrates the polymorphic sites detected inthe 18D. willistoni Adhalleles. A total of 35 nucleotidesubstitutions and four single nucleotide indels werefound. In the noncoding DNA, introns had the highestpercentage of polymorphic sites (7.46%), with the 58region having 3.27%, and 38 region 2.36%. In the codingregion, there were 20 nucleotide polymorphisms, 14 ofwhich were synonymous. Tables 3 and 4 summarize thenature of these polymorphisms. Twelve of 14 synony-mous polymorphisms are transitions, a not unexpectedresult given the structure of the genetic code. The fre-quency of synonymous polymorphism for each codonredundancy class is not significantly different from ran-dom expectations (G4 4.48, 3 d.f.,p > 0.1). Surpris-ingly, many of the synonymous polymorphisms areshared across strains from widely separated localities.

Table 4 summarizes the replacement polymorphism.The predicted charge at pH 7.0 of each polymorphicprotein was determined, and none differed from that ofwil4 ADH. Hydrophilicity profiles and predictions of

Table 3. Summary of polymorphic synonymous codons in the 18D.willistoni Adh alleles. Sites are labeled as in Figure 1

SiteChange in NTsequence

Amino acidencoded

Codonredundancygroup

Transition ortransversion

54 GGT-GGC Gly fourfold Transition63 ACC-ACA Thr fourfold Transversion

367 ATC-ATT Ile threefold Transition403 GTC-GTT Val fourfold Transition436 GAT-GAC Asp twofold Transition460 GGC-GGT Gly fourfold Transition469 TGC-TGT Cys twofold Transition571 TTA-TTG Leu sixfold Transition629 TTG-CTG Leu sixfold Transition637 CAT-CAC His twofold Transition678 GCT-GCC Ala fourfold Transition718 TGT-TGC Cys twofold Transition772 AAA-AAG Lys twofold Transition823 TCG-TCT Ser fourfolda Transversion

a Serine cannot be analyzed as a sixfold redundant codon, becausesubstitutions in two codon positions must occur between some synony-mous codons. Serine codons are thus divided between the two- andfourfold redundant groups

Fig. 1. Polymorphic sites in the 18D. willistoni Adh alleles. The wil4 sequence is used as the reference sequence.n/, refer to indels.Replacement polymorphisms are indicated in bold type. Strain abbreviations are as listed in Table 1.

234

secondary structure were obtained for each of the poly-morphic proteins as well as for wil4 ADH, using theMacVector program (IBI) (data not shown). None of theamino acid replacements caused a major shift in theseproperties. Thus our results are consistent with an inabil-ity of electrophoresis to detect these polymorphisms.

We performed two statistical tests to determine howwell the data conform to expectations of the neutraltheory. One is that of McDonald and Kreitman (1991)which tests whether the ratio of silent to replacementpolymorphisms are significantly different from that ratiobetween species. We usedD. nebulosaas the comparisonspecies because it has sufficient divergence to provide areasonable sample of changes, yet is sufficiently closelyrelated that saturation has not occurred (Gleason et al.submitted). The silent:replacement ratio for polymor-phisms is 14:6 and for interspecific differences is 54:9. AG-test yielded a value of 2.19, 1 d.f., p > 0.1. Thus by thistest we cannot reject the hypothesis of neutrality.

The other test is the Tajima (1989) test which testshow well the average observed number of nucleotidesdifferences between alleles (p) corresponds to the aver-age number of segregating nucleotides per site predictedby a panmictic neutral model at equilibrium (u). For thecoding region, the test statistic D4 −1.21, p4 0.1. Forthe noncoding region, D4 −1.69, 0.1 >p > 0.05. Thusin each case the D is negative, but not quite statisticallysignificant.

Discussion

As mentioned in the Introduction, virtually no electro-phoretically detectable variation was observed in allo-zyme studies ofD. willistoni Adh, despite a very largesample size. Thus it was a surprise to find six replace-ment polymorphisms in a sample of only 18 alleles.None of the detected polymorphisms cause a change incharge, although several of these changes involve polar/nonpolar R groups, which potentially could have affectedelectrophoretic mobility. We performed electrophoresison the strains with amino acid substitutions using cellu-

lose acetate strips, and detected no mobility differences(G. Allegrucci and J.R. Powell, unpublished data). Ayala(1975) and his coworkers used starch gels. It is possiblethat a method that can detect conformational changesmay detect mobility differences. It is unlikely that sizedifferences could be detected as the predicted molecularweights of the ADH molecule for these alleles differ atmost by 0.16%

Table 5 compares our results with previous nucleotidepolymorphism studies onDrosophila Adh.Most notableis the uncharacteristically high ratio of replacement tosilent polymorphism inD. willistoni. This implies theremay be reduced selective constraints on this gene inD.willistoni compared to the previously studied species.Consistent with this is the fact that theD. willistoni Adhgene has lower codon usage bias, having an ‘‘effectivenumber of codons’’ (Wright 1990) of 45.1 compared to31.4 for D. melanogaster Adh.However, the level ofADH enzyme activity in thewillistoni group species isabout the same as for the Slow allele inD. melanogasterandD. simulansand is about the mean for species breed-ing in rotting fruit (Mercot et al. 1994). Unless there is asignificant difference in specific activity of the enzyme,it would seem that the level of expression is not the causeof the relaxation of selection. Considering the similarecology ofwillistoni andmelanogastergroups (tropicalfruit breeders) it would be surprising if the level of ex-pression ofAdh would be greatly different.

Generally,D. melanogasterhas proportionately morereplacement polymorphisms than doesD. simulans(Moriyama and Powell, 1996). This has been attributedto the possibility that the effective population size ofD.melanogasteris smaller than that ofD. simulans.Thismakes selection against mildly deleterious mutations lesseffective, and implies that most replacement polymor-phisms are slightly deleterious. If this is the explanationfor differences in the proportion of replacement to silentpolymorphisms among species, then this implies that theeffective population size ofD. willistoni is even less thanthat of D. melanogaster.This may seem improbablegiven the high densities that have been measured forD.willistoni (Burla et al. 1950) and the extremely largerange of the species. Doubtless today, the census size ofD. willistoni is extremely large. This suggests that per-hapsD. willistoni had a smaller population size in itsrecent history and has only recently expanded to its pres-ent size. One indication of an expanding population is anegative Tajima’s D statistic because expansion of asmall population is analogous to a selective sweep(Tajima 1993; Braverman et al. 1995). For these data,this statistic is negative but not quite significantly so (seeResults). The only other study of nucleotide polymor-phism inD. willistoni is for theperiodlocus where all theTajima’s Ds were also negative, but not significant(Gleason and Powell, 1997). Another possible indicationof a small effective population size is the unusually

Table 4. Summary of amino acid replacement polymorphisms in theD. willistoni Adhalleles

StrainAmino acidchange

Change innucleotidesequence

Position innucleotidesequence

Position inprimary aminoacid sequence

wil4 Val to Ile GTT to ATT 236 57wil0 Thr to Ala ACC to GCC 296 77WilLwilM1 Ser to Pro TCT to CCT 479 138wilM3 Ser to Pro TCC to CCC 647 194wilC Thr to Ala ACA to GCA 704 213wilA5 Ile to Thr ATT to ACT 741 225

235

low polymorphism in adjacent noncoding DNA (Table5). Nucleotide variation in noncoding regions in all otherspecies is two to five times that of coding regions, but inD. willistoni it is about equal. Reduced selection in thecoding region and insufficient time since the expansionfor mutation to generate equilibrium amounts of(nearly?) neutral variation in the noncoding region couldaccount for this pattern. There are also several polymor-phic sites shared among strains from throughout therange of this species, which is also consistent with anexpansion of a recent small population.

Is it reasonable to invoke a small effective populationin a species so numerous today? It is generally thoughtthat during times of glaciation, the Neotropics experi-enced a cooling and drying period (Haffer 1969; Colin-vaux 1993). The appropriate habitat forD. willistonicould have decreased considerably. Spassky et al. (1970)felt that the distributions of the semispecies of the veryclosely relatedD. paulistorum complex could be ac-counted for by Pleistocene refugia.

Seemingly inconsistent with a recent small effectivepopulation is the high allozyme polymorphism inD. will-istoni (Ayala 1975), among the 25% highest in species ofDrosophila (Powell 1975). However, a decoupling be-tween nucleotide and allozyme polymorphism has beenobserved inDrosophila (Begun and Aquadro 1993) aswell as in other organisms (Karl and Avise 1992). Thecase ofD. melanogastermay be analogous toD. willis-toni. In D. melanogaster,the ancestral African popula-tions have considerably more nucleotide variation ascompared to human commensal populations throughoutthe world, but the allozyme variation is very similar forall populations examined (Begun and Aquadro 1993). Itis possible that only a few individual colonists were theorigin of the human commensal populations. This historyhas apparently reduced nucleotide polymorphism whilenot affecting appreciably the allozyme polymorphism.

A potential alternative explanation is related to sup-pression of recombination due to the high inversion poly-morphism inD. willistoni. If the level of recombinationis very low, then natural selection is less effective atdetecting single nucleotide variants which may beslightly deleterious. However,Adh in D. willistoni is onthe IIR chromosome (Rohde et al. 1995) which is the

least polymorphic arm for inversions. In a sample of 57populations from throughout the range, an average of60% of all individuals are karyotypically homozygousfor the IIR (Dobzhansky 1957; da Cuhna et al. 1959).The two inversions which would containAdh withintheir breakpoints (C and D) are quite rare with hetero-karyotypes for these inversions averaging only 9% in the57 populations. Thus there is little reason to suspect thatinversion polymorphism is greatly affecting recombina-tion at this particular locus.

Whatever the explanation for the pattern ofAdhnucleotide polymorphism inD. willistoni, it is instructiveto note that theD. willistoni group has consistentlyyielded different molecular evolutionary patterns, espe-cially when compared toD. melanogaster.It has a sig-nificantly different codon usage (Anderson et al. 1993;Powell and Gleason 1996), itsAdh gene has lost an in-tron (Anderson et al. 1993), and in theperiod locus aregion that is highly polymorphic inD. melanogaster(Costa et al. 1991, 1992) is virtually monomorphic in thewillistoni group (Gleason and Powell 1997). Thus gen-eralizing about molecular evolutionary patterns and pro-cesses (in as much as processes can be deduced frompatterns) from a single species (D. melanogaster) to Dro-sophila is risky at best.

Acknowledgments. We thank E.N. Moriyama for statistical assis-tance, A. Calderon-Urrea for oligonucleotide synthesis, C. Andersonfor DNA samples, and the individuals listed in Table 1 for kindlysharing theirD. willistoni stocks. This study was supported by a NSFDissertation Improvement Grant (DEB9122899) to ECG and NSFgrant DEB 9318836 to JRP. ECG was supported by a PHS TrainingGrant.

References

Anderson C, Carew EA, Powell JR (1993) Evolution of theAdh locusin theDrosophila willistonigroup: the loss of an intron, and shift incodon usage. Mol Biol Evol 10:605–618

Ayala FJ (1973). Two new subspecies of theDrosophila willistonigroup (Diptera: Drosophilidae). Pan-Pacific Entomologist 49:273–279

Ayala FJ (1975) Genetic differentiation during the speciation process.Evol Biol 8:1–78

Table 5. Comparison of level ofAdh polymorphism amongDrosophilaspecies

SpeciesNumber ofalleles

Coding DNA Noncoding

ReferenceSilent Replacement pa ub p u

willistoni 18 14 6 0.0053 0.0076 0.0057 0.0104 This papermelanogaster 15 16 3 0.0081 0.0076 0.0191 0.0170 Laurie et al. 1991simulans 5 11 0 0.0068 0.0069 0.0271 0.0272 McDonald and Kreitman 1991pseudoobscura 107 38 1 0.0040 0.0098 0.0202 0.0315 Schaeffer and Miller 1992b, 1993

a Observed average proportion of nucleotide differences between allelesb Predicted average number of nucleotides segregating per site under a panmictic neutral model at equilibrium (Tajima 1993)Note: Values are taken from Moriyama and Powell (1996)

236

Begun DJ, Aquadro CF (1993) African and North American popula-tions of Drosophila melanogasterare very different at the DNAlevel. Nature 365:548–550

Braverman JM, Hudson RR, Kaplan NL, Langley CH, Stephan W(1995) The hitchhiking effect on the site frequency spectrum ofDNA polymorphisms. Genetics 140:783–796

Burla H, da Cuhna AB, Cavalcanti AGL, Dobzhansky Th, Pavan C(1950) Population density and dispersal rates in BrazilianDro-sophila willistoni.Ecology 31:393–404

Colinvaux P (1993) Pleistocene biogeography and diversity in tropicalforests of South America. In: Goldblatt P (ed) Biological relation-ships between Africa and South America. Yale University Press,New Haven, pp 473–499

Costa R, Peixoto AA, Thackeray JT, Dalgleish R, Kyriacou CP (1991)Length polymorphism in the threonine-glycine-encoding repeat re-gion of theperiod gene inDrosophila.J Mol Evol 32:238–246

Costa R, Peixoto AA, Barbujani G, Kyriacou CP (1992) A latitudinalcline in aDrosophilaclock gene. Proc Roy Soc Lond B 250:43–49

daCunha AB, Dobzhansky Th, Pavlovsky O, Spassky B (1959) Genet-ics of natural populations. XXVIII. Supplementary data on thechromosomal polymorphism inDrosophila willistoniand its rela-tion to the environment. Evolution 13:389–404

Dobzhansky Th (1957) Genetics of natural populations. XXVI. Chro-mosomal variability in island and continental populations ofDro-sophila willistonifrom Central America and The West Indies. Evo-lution 11:280–293

Gleason JM, Powell JR (1997) Interspecific and intraspecific compari-sons of theperiodlocus in theDrosophila willistonisibling species.Mol Biol Evol 14:741–753

Gleason JM, Griffith EC, Powell JR (submitted) A molecular phylog-eny of theDrosophila willistonigroup.

Haffer J (1969) Speciation in Amazonian forest birds. Science 165:131–147

Hudson RR, Kreitman M, Aguade´ M (1987) A test of neutral molecularevolution based on nucleotide data. Genetics 116:153–159

Karl SA, Avise JC (1992) Balancing selection at allozyme loci inoysters: implications from nuclear RFLPs. Science 256:100–102

Kreitman M (1983) Nucleotide polymorphism at the alcohol dehydro-genase locus ofDrosophila melanogaster.Nature 304:412–417

Kreitman M, Hudson RR (1991) Inferring the evolutionary histories ofthe Adh and Adh-duploci in Drosophila melanogasterfrom pat-terns of polymorphism and divergence. Genetics 127:565–582

Laurie CC, Bridgham JT, Choudhary M (1991) Association betweenDNA sequence variation and variation in expression of theAdhgene in natural populations ofDrosophila melanogaster.Genetics129:489–499

McDonald JH, Kreitman M (1991) Adaptive protein evolution at theAdh locus inDrosophila.Nature 351:652–654

Mercot H, Defaye D, Capy P, Pla E, David JR (1994) Alcohol toler-ance, ADH activity, and ecological niche ofDrosophila species.Evolution 48:746–757

Moriyama EN, Powell JR (1996) Intraspecific nuclear DNA variationin Drosophila.Mol Biol Evol 13:261–277

Powell JR (1975) Protein variation in natural populations of animals.Evol Biol 8:79–119

Powell JR, Gleason JM (1996) Codon usage and the origin of P ele-ments. Mol Biol Evol 13:278–279

Rohde C, Abdelhay E, Pinto H, Schrank A, Valente VLS (1995) Analy-sis andin situ mapping of theAdh locus in species of thewillistonigroup ofDrosophila.Cytobios 81:37–47

Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a labo-ratory manual. Cold Spring Harbor Laboratory Press, Cold SpringHarbor, NY

Sanger F, Nickler S, Coulson AR (1977) DNA sequencing with chainterminating inhibitors. Proc Natl Acad Sci USA 74:5463–5467

Schaeffer SW, Miller EL (1991) Nucleotide sequence analysis ofAdhgenes estimates the time of geographic isolation of the Bogota`population ofDrosophila pseudoobscura.Proc Natl Acad Sci USA88:6097–6101

Schaeffer SW, Miller EL (1992a) Molecular population genetics of anelectrophoretically monomorphic protein in thealcohol dehydro-genaseregion of Drosophila pseudoobscura.Genetics 132:163–178

Schaeffer SW, Miller EL (1992b). Estimates of gene flow inDro-sophila pseudoobscuradetermined from nucleotide sequenceanalysis of thealcohol dehydrogenaseregion. Genetics 132:471–480

Schaeffer SW, Miller EL (1993) Estimates of linkage disequilibriumand the recombination parameter determined from segregatingnucleotide sites in the alcohol dehydrogenase region ofDrosophilapseudoobscura.Genetics 135:541–552

Spassky B, Richmond RC, Pe´rez-Salas S, Pavlovsky O, Moura˜o CA,Hunter AS, Hoenigsberg H, Dobzhansky Th, Ayala FJ (1971) Ge-ography of the sibling species related toDrosophila willistoni,andof the semispecies of theDrosophila paulistorumcomplex. Evolu-tion 25:129–143

Tajima F (1989) Statistical method for testing the neutral mutationhypothesis by DNA polymorphism. Genetics 123:585–595

Tajima F (1993) Measurements of DNA polymorphism. In: TakahataN, Clark AG (eds) Mechanisms of molecular evolution. SinauerPress, Sunderland, MA, pp 37–60

Werman SD, Davidson EH, Britten RJ (1990) Rapid evolution in afraction of theDrosophilanuclear genome. J Mol Evol 30:281–289

Wright F (1990) The ‘effective number of codons’ used in a gene. Gene87:23–29

237