Proc. Nati. Acad. Sci. USAVol. 87, pp. 2951-2954, April 1990Genetics

The polydeoxyadenylate tract of Alu repetitive elements ispolymorphic in the human genome

(J3-globin gene cluster/adenosine dnse gene/polymerase chain reaction/DNA marker)

EFFROSINI P. ECONOMOU*, ANDREW W. BERGEN*, ANDREW C. WARRENt,AND STYLIANOS E. ANTONARAKIS*-Departments of *Pediatrics and tPsychiatry, Center for Medical Genetics, The Johns Hopkins University School of Medicine, Baltimore, MD 21205

Communicated by Victor A. McKusick, January 22, 1990

ABSTRACT To identify DNA polymorphisms that areabundant in the human genome and are detectable by poly-merase chain reaction amplification ofgenomic DNA, we testedthe hypothesis that the polydeoxyadenylate tract of the Alufamily of repetitive elements is polymorphic among humanchromosomes. We analyzed the 3' ends of three specific Alusequences and found that two (in the adenosine deaminase geneand the /3-globin pseudogene) were polymorphic. This novelclass ofpolymorphisms, termed AluVpA [Alu variable poly(A)Jmay represent one of the most useful and informative group ofDNA markers in the human genome.

The mapping of the human genome is greatly facilitated bythe enormous normal variability of the DNA sequencesbetween two randomly chosen homologous chromosomes.This DNA polymorphism provides the basis for the largenumber of markers currently being used to construct linkagemaps and search for the location ofunknown genes that causehereditary disorders (1). Several types of DNA polymor-phisms have been described: (i) single nucleotide substitu-tions that can be detected by restriction analysis (2); (ii)variable number of tandem repeats (VNTR), in which a 9- to45-nucleotide unit is repeated (3, 4); (iii) presence or absenceof retrotransposons-i.e., Alu and Li repetitive elements-and pseudogenes (5-7); (iv) variable number of dinucleotiderepeats (VNDR), in which a dinucleotide is repeated (8-12).The last class of DNA polymorphisms can only be detectedafter polymerase chain reaction (PCR) amplification ofDNA(13) and separation of the different alleles by polyacrylamidegel electrophoresis.

Despite all of these DNA polymorphisms, there are stillnumerous instances in which a certain DNA fragment cannotbe used as a marker because of the lack of such polymor-phisms or in which a given family with a specific disorder isnot informative for a known polymorphism and thereforecannot be used in a particular linkage analysis. In this studywe explored the possibility that the tracts of adenine residues[poly(A)] within the 3' ends of the Alu repetitive elements inthe human genome (14, 15) are polymorphic, i.e., that theycontain a variable number of adenine residues on differentchromosomes. We studied the poly(A) tracts in three Alusequences: in the f-globin gene cluster on the short arm ofchromosome 11 (lip); in the adenosine deaminase gene onchromosome 20; and in the factor VIII gene on the Xchromosome. We found that the poly(A) tracts ofthe first twoAlu sequences were polymorphic and termed them AluVpA[Alu variable poly(A)]. Because Alu repeats comprise about1% of the human DNA, a large number of AluVpA polymor-phisms may exist in the human genome and may serve as a

majorDNA marker system in almost any DNA fragment thatcontains Alu repetitive elements.

MATERIALS AND METHODSSubjects. The DNA from members of the 40 CEPH (Centre

d'Etude du Polymorphism Humain; ref. 16) families weretested for the presence and extent of the AluVpA polymor-phisms. These DNA samples have been extensively used ina major collaborative effort to construct initial linkage mapsof human chromosomes (17-19).

Amplified Alu Repeats. Three Alu repeats and their corre-sponding 3' poly(A) tracts along with a small stretch ofnonrepetitive DNA were used for amplification to determinewhether these poly(A) tracts were polymorphic. The first Alurepeat studied is located within the P-globin gene clusterabout 500 nucleotides upstream from the /3-globin pseudo-gene (20) on the short arm of human chromosome 11. A158-nucleotide DNA fragment was amplified by using oligo-nucleotides OLl and OL2 as primers for the PCR. Oligonu-cleotide OLi (5'-AGAGATCGCGCCACTGCACA-3') is partof the "right" arm of the Alu repeat (14) and thereforehybridizes to a large number of Alu repeats in the genome;oligonucleotide OL2 (5'-CACAGCCTTTCTTGGTTTTC-3'), however, is derived from a single-copy sequence. PCRamplification resulted in a DNA fragment that contained inthe published sequence (21) a poly(A) tract of 45 nucleotidesinterrupted by four guanines. The second Alu repeat studiedis located about 1100 nucleotides 5' to the transcriptioninitiation site of the adenosine deaminase gene on humanchromosome 20 (21). A 188-nucleotide DNA fragment wasamplified by using oligonucleotides OL3 and OL4 as PCRprimers. Oligonucleotide OL3 (5'-CCAGATCGCGCCACT-TCACT-3') is similar, but not identical, to OL1 and is part ofthe "right" arm of the Alu repeat; oligonucleotide OL4(5'-AGATGAGCATAGATACGAGA-3') is derived from thenonrepetitive area 3' to the Alu sequence. The amplifiedDNA from this latter reaction contained in the publishedsequence (21) a poly(A) tract that consisted of the shortsequence TAAA repeated nine times. The third Alu repeatstudied is located in the first intron of the factor VIII gene onthe X chromosome. A 197-nucleotide fragment was amplifiedby using oligonucleotides OL5 and OL6 as PCR primers.Oligonucleotide OL5 (5'-GTGATTGTTCCACTGCACTG-3') is also similar but not identical to OL1 and OL3 and isagain part of the right arm of the Alu repeat; oligonucleotideOL6 (5'-GTGCCTTGGTGAAAAATAAAGC-3') is derivedfrom the nonrepetitive area 3' to the Alu sequence (22). The

Abbreviations: CEPH, Centre d'Etude du Polymorphism Humain;PCR, polymerase chain reaction; VNTR, variable number oftandemrepeats; AluVpA, Alu variable poly(A).tTo whom reprint requests should be addressed at: Department ofPediatrics, CMSC 1004, The Johns Hopkins University School ofMedicine, Baltimore, MD 21205.

2951

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

8, 2

020

Proc. Natl. Acad. Sci. USA 87 (1990)

end of this Alu repeat has a short poly(A) tract of sevenadenines.

Detection of Polymorphic Alleles. Amplification of genomicDNA (200-500 ng) isolated from lymphoblastoid cell lines ofthe CEPH collection was performed in a reaction mixturecontaining 25 ,.l of buffer (10 mM Tris HCI, pH 8.3/50 mMKCI/1.5 mM MgCl2/0.02% gelatin), 200 ,uM each dNTP, 0.5unit of Thermus aquaticus (Taq) DNA polymerase (23), 40nM repetitive primer (OL1, OL3, or OLS for the threedifferent amplifications, respectively), and 400 nM nonrepet-itive primer. For the amplification of specific DNA fragmentsit was important to use the repetitive and nonrepetitiveoligonucleotides at a concentration ratio of 1:10. The mixturewas heated for 6 min at 94°C, followed by 30 cycles ofamplification in aDNA thermal cycler (Perkin-Elmer/Cetus)using a step program (denaturation at 94°C for 30 sec,annealing at 60°C for 30 sec, extension at 72°C for 30 sec).Amplified DNA was initially analyzed by electrophoresis ina 3% Nusieve (FMC)/1% agarose gel.

In order to detect the different polymorphic alleles, theamplified DNA fragments were labeled with 32p. Usually, thenonrepetitive oligonucleotide (OL2, OL4, or OL6) was 5'-end-labeled by using [y-32P]ATP (3000 Ci/mmol, New En-gland Nuclear; 1 Ci = 37 GBq) and T4 polynucleotide kinase(BRL). Alternatively, 1-2 uCi of[a-32P]dCTP at 800 Ci/mmolwas added to the PCR mixture, with or without reduction ofunlabeled dCTP to 50 ,uM. Two microliters of the amplifiedproducts was mixed with an equal volume of formamidebuffer (95% formamide/20 mM EDTA/0.05% bromophenolblue/0.05% xylene cyanol FF), boiled for 2 min, and elec-trophoresed in 6% polyacrylamide/8 M urea sequencing gelfor 2-2.5 hr at 58 W. After electrophoresis, the gel was driedand exposed to Kodak XAR-5 film for 1-16 hr.DNA sequencing with modified T7 DNA polymerase (Se-

quenase, United States Biochemical) was performed on thePCR products as described (24). As sequencing primers weused either OL2 or OL4 or a third internal primer [OL7(5'-TCTCTAGCGCGGTGACGTGT-3') for the OL1/OL2amplification; OL8 (5'-AGTGGTTATCTCAGGTGAAAG-3') for the OL3/OL4 amplification]. Sequencing primerswere end-labeled by using [-32P]ATP and T4 polynucleotidekinase. Approximately 100 ng of PCR product, purified byspin dialysis with Centricon-30 (Amicon), was annealed to 10ng of the sequencing primer in 11 ,ul on ice for 3 min. Afterheat denaturation for 5 min at 95°C, 2.5 ,ul of the sample wasadded to each of four tubes with 3 ,ul of the sequencingmixture containing 62 ,M nonradioactive dNTPs, 6.2 ,uMdideoxynucleoside triphosphates, and 2 units of T7 DNApolymerase in buffer (25 mM Tris-HCl, pH 7.5/10 mMMgCl2/70 mM NaCl/7 mM dithiothreitol). The mixtureswere incubated for 10 min at 40°C and then loaded on apolyacrylamide gel.Two-point linkage analyses were performed with the com-

puter program packages LINKAGE version 4.9 (25) and LIPED(26).

RESULTSAfter amplification of the 3' end of the Alu repeat within the5' flanking sequence of the adenosine deaminase gene inmembers of the CEPH families and electrophoresis of theamplified DNAs, seven alleles were identified. Fig. LA showsthe analysis of DNA from several unrelated individuals.Mendelian inheritance of the alleles was demonstrated in all40 CEPH families. Fig. 1B shows an example of the inheri-tance of several alleles in one family. The frequency of theseven alleles observed, and the informativeness of the poly-morphic system, is shown in Table 1. The observed heterozy-gosity in the unrelated individuals of the CEPH families is0.74 and therefore this polymorphic system is highly infor-mative for chromosome 20. To study the molecular basis of

A

Alleles

54321

6

B

Alleles

4321

6

Cl)

V- VI 04 C1 4 7, Y, CN ° 0 C0

csN q Itt Ln LO rcTAb it) C'

P:A

Family CEPH 1333

4

4s

FIG. 1. (A) Autoradiogram of PCR-amplified products from the3' end of the Alu sequence in the adenosine deaminase gene. Thenumbers above each lane identify unrelated individuals from theCEPH families. The numbers of the various alleles are shown at left.These alleles differ from one another by 4 nucleotides. For example,allele 1 contains 9 TAAA repeats, allele 2 contains 10, and allele 3contains 11 such repeats. Oligonucleotides OL3 and OL4 were usedfor the PCR amplification (see Materials and Methods). The size ofallele 1 detected by using these oligonucleotides is 188 nucleotides.(B) AluVpA polymorphism in the adenosine deaminase gene. Themembers of CEPH family 1333 are represented above the lanesshowing their corresponding alleles. In this family, five alleles can berecognized and are numbered at left. Oligonucleotides OL3 and OLAwere used for the PCR amplification.

the polymorphic alleles, the nucleotide sequences of PCRamplification products from individuals homozygous for agiven allele were determined. The difference among alleles isdue to different numbers ofthe TAAA repeated units (Fig. 2).No new alleles were observed in 750 meioses studied in the40 CEPH families. Linkage analyses using data from the V3CEPH data base of chromosome 20 DNA markers and theAluVpA polymorphism in the adenosine deaminase genedescribed here showed that this polymorphic marker is

Table 1. Number of polymorphic alleles and informativeness ofthe AluVpA polymorphisms of the adenosine deaminase gene(ADA) and /3-globin pseudogene (HBBPI)

AluVpA ObservedLocus allele(s) n % PIC* heterozygosityADA 1 68 25.3 0.74 0.74 (98/132)

2 34 12.63 73 27.14 66 24.55 26 9.76 1 0.47 1 0.4

HBBPI Upper 48 19.1 0.26 0.33 (39/118)Lower 203 80.9

*Polymorphism information content.

2952 Genetics: Economou et al.

..Wmbo. 4040

40 4001:1!,l4low 4b

_O*

Dow

nloa

ded

by g

uest

on

Mar

ch 2

8, 2

020

Proc. Natl. Acad. Sci. USA 87 (1990) 2953

Allele 1 Allele 4 Allele 5

A C G T A C G T A C G T

23-01 1340-12 1332-03

FIG. 2. Nucleotide sequence analysis of the poly(A) tract of theAlu repeat in the adenosine deaminase gene. Alleles 1, 4, and 5 weresequenced from individuals homozygous for these alleles. The iden-tification numbers of these CEPH individuals are shown. Oligonucle-otides 0L3 and 0L4 were used for PCR amplification and oligonu-cleotide 0L8 was used as a sequencing primer. Note that allele 4contains three more, and allele 5 four more, TA3 repeats than allele 1.

closely linked to locus D20S17 recognized by probe CRI-L127 (18) [J = 0.019; Z' = 17.07].The PCR amplification of the 3' end of the Alu repeat in the

/3-globin pseudogen'e region in members ofthe CEPH familiesrevealed two categories of polymorphic alleles that wetermed upper and lower. Fig. 3 shows the analysis of onefamily in which the parents were homozygous for the upperand lower alleles and all the offspring were heterozygous.The inheritance of the alleles was Mendelian in all CEPHfamilies examined, and no new alleles were observed in 730meioses examined. Nucleotide sequence analysis of ampli-fied DNAs from individuals homozygous for the upper orlower alleles (Fig. 4) indicated that the upper allele containsfour more adenine residues than the lower. However, differ-ences in one nucleotide between alleles are not easily distin-guished and therefore the "upper" or "lower" allele may notbe identical in different families. The appearance of the allelicfragments in the autoradiogram is not a perfect sharp band

Family CEPH 37

Alleles

UpperE[ *LowerE ftde w

FIG. 3. AluVpA polymorphism in the ,B-globin pseudogene. Themembers of CEPH family 37 are represented above the lanesshowing their corresponding alleles. Oligonucleotides OL1 and OL2were used for the PCR amplification.

Upper AlleleA C G T

'.

16 I'

Lower AlleleA C G T

FIG. 4. Nucleotide sequenceof the poly(A) tract of the Alurepeat in the (-globin pseudo-gene. The upper allele containsfour more adenines than thelower allele. The sequenceswere obtained from PCR ampli-fication products from individu-als homozygous for the upperand lower alleles. Oligonucleo-tide OL7 was used as sequencingprimer.

but rather a group ofthree or more bands of variable intensityof which the most prominent band represents the actualallelic size. This is presumably due to errors by Taq poly-merase (i.e., the polymerase incorporates one more or oneless adenine in the newly growing nucleotide chain) when itreads through the poly(A) stretch of DNA. When the twoends of the PCR amplification product were cleaved by arestriction endonuclease, only the fragment with the repeatsshowed the pattern of multiple bands (data not shown),suggesting that there is no polymerase error at the beginningor end of the amplification reaction. Table 1 shows thefrequency of the "two" alleles of this polymorphic systemand the observed heterozygosity in the unrelated individualsof the CEPH families. Linkage analyses using data from theV3 CEPH data base ofchromosome 11 DNA markers and theAluVpA polymorphism in the f3-globin pseudogene showed(as expected) no recombination with other markers within the/3-globin gene cluster. For example, there was no recombi-nation between the pseudoj3-globin AluVpA and the HindIIIsite in the y-globin gene (6 = 0.00; z = 11.3).The amplified 3' end of the Alu repeat from intron 1 of the

factor VIII gene was not polymorphic in the DNA of 20unrelated females (40 X chromosomes examined) from theCEPH families.

DISCUSSIONPCR amplification of specific DNA fragments (23) greatlyfacilitates the study ofthe human genome because of its manyadvantages over techniques that require nucleic acid hybrid-ization (27). It is therefore desirable to detect DNA polymor-phic markers by using PCR amplifications that eliminate theuse of molecular probes. We describe here a novel class ofDNA polymorphic markers due to normal variation of thepoly(A) tracts ofAlu repeats among individuals. This class ofDNA polymorphisms, which we term AluVpA, may repre-

1W.-..

"Wom,

Genetics: Economou et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

8, 2

020

Proc. Natl. Acad. Sci. USA 87 (1990)

Table 2. Types of poly(A) tracts of the human Alu repeats

Type Characteristic n S

1 More than 10 adenines 154 79.82 Patterned repeat* 11 5.71 and 2 14 7.33 Less than 10 adenines 7 3.64 None of the above 7 3.6

One hundred ninety-three Alu repeats from the GenBank data base(release 62.0; December 15, 1989) were examined.*For example, (TA3)", (CA4)n, (GA3),, (GA)", or (TA2)n, where n 2 3.

sent one of the most abundant polymorphic markers in thehuman genome since >105 Alu repetitive elements exist in thehaploid human genome. The detection of these polymor-phisms is based on electrophoresis of radioactive PCR am-plified products. We found that the best results were obtainedby using a 1:10 concentration ratio of repetitive primer tononrepetitive primer. In addition to the specific PCR prod-ucts, other amplification products are also seen in the poly-acrylamide gels and may represent Alu-Alu amplificationproducts. These products are numerous when radioactivedNTPs are incorporated into the amplified DNA and are nottotally absent when the single-copy primer is end-labeled.Different alleles may be defined by differences in the lengthof the amplified products due to the different number ofadenine residues in the poly(A) tracts. Similarly, poly(A)tracts from Li repeats, processed pseudogenes, or otherretrotransposons, or any poly(A) tract or other polynucle-otide tract, have the potential of being polymorphic amonghuman chromosomes. The AluVpA polymorphisms can beused as markers in linkage analyses to create meiotic maps,to mark normal and defective genes, to search for the locationof unknown genes after linkage analyses in families withspecific phenotypes, and to follow the inheritance of specificchromosomes. The AluVpA polymorphisms examined in thisstudy are stable within a given family, since no new alleleswere observed in a large number of meioses examined in theCEPH families. These results are similar to those observedfor the VNTR polymorphisms (28). In the Human GenomeMapping/Sequencing Project AluVpA polymorphisms willbe of particular interest because they can serve both assequence-tagged sites (STS; ref. 29) for identification andfingerprinting of cosmid or yeast-artificial-chromosomeclones and as polymorphic markers with known location inthe genetic maps. Potential mechanisms of the genesis ofAluVpA polymorphisms during evolution may include poly-merase errors or slippage during replication (30, 31), sisterchromatid exchange with mispairing at the poly(A) tracts (32,33), or unequal homologous crossing-over.To estimate the extent ofAluVpA polymorphisms in the Alu

repeats, we have categorized the 3' ends ofa sample of 193 Alusequences included in the GenBank sequence data base. Asshown in Table 2, almost 80Wo ofAlu repeats have >10 adenineresidues at their 3' ends (type 1) and 5.7% contain a patternedshort A-rich unit such as TAAA, CAAAA, GAAA, or TAArepeated at least three times (type 2). Furthermore, 7.3% havecharacteristics of both types 1 and 2, for they contain morethan 10 adenines and a patterned short repeated unit. Weanticipate that a large number of those Alu types may bepolymorphic. The Alu sequences with a short poly(A) tract(type 3) or no poly(A) tract at all (type 4), which comprise 7.2%of the sequences examined in Table 2, may not be polymor-phic. Recently, Orita et al. (34) have reported that singlenucleotide changes in Alu repetitive sequences can be easilydetected by electrophoresis of single-stranded PCR amplifi-cation products under nondenaturing conditions. Therefore,

Alu repetitive elements may represent one of the most abun-dant and useful polymorphic systems in the human genome.

Note Added in Proof. Since the submission of the manuscript, wehave found AluVpA polymorphisms in the a-globin gene cluster onchromosome 16p, the HMG-14 gene on chromosome 21q, and theargininosuccinate synthetase gene on chromosome 9q.

We thank Drs. G. L. Semenza, A. F. Scott, and M. B. Petersenfor valuable suggestions and J. Strayer for assistance in the artwork.E.P.E. is partially supported by the Cooley's Anemia Foundation ofMaryland; S.E.A. is supported by grants from the National Institutesof Health and the U.S. Department of Energy. A.W.B. is a studentin the National Institutes of Health-supported Predoctoral Programin Human Genetics.

1. Botstein, D., White, R. L., Scolnick, M. & Davis, R. W. (1980) Am.J. Hum. Genet. 32, 314-331.

2. Kan, Y. W. & Dozy, A. M. (1978) Proc. Nat!. Acad. Sci. USA 75,5631-5635.

3. Jeffreys, A. J., Wilson, V. & Thein, S. L. (1985) Nature (London)314, 67-73.

4. Nakamura, Y., Leppert, M., O'Connell, P., Wolff, R., Holm, T.,Culver, M., Martin, C., Fujimoto, E., Hoff, M., Kumlin, E. &White, R. L. (1987) Science 235, 1616-1622.

5. Frossard, P. M., Coleman, R. T., Protter, A. A., Seilhamer, J. J.,Funke, H. & Assmann, G. (1986) Nucleic Acids Res. 14, 8694.

6. Anagnou, N. P., O'Brien, S. J., Shimada, T., Nash, W. G., Chen,M. J. & Nienhuis, A. W. (1984) Proc. Nat!. Acad. Sci. USA 81,5170-5174.

7. Woods-Samuels, P., Wong, C., Mathias, S. L., Scott, A. F., Ka-zazian, H. H., Jr., & Antonarakis, S. E. (1989) Genomics 4, 290-296.

8. Weber, J. L. & May, P. E. (1989) Am. J. Hum. Genet. 44, 388-396.9. Litt, M. & Luty, J. A. (1989) Am. J. Hum. Genet. 44, 397-401.

10. Semenza, G. L., Malladi, P., Surrey, S., Delgrosso, K., Poncz, M.& Schwartz, E. (1984) J. Biol. Chem. 259, 6045-6048.

11. Chebloune, Y., Pagnier, J., Trabuchet, G., Faure, C., Verdier, G.,Labie, D. & Nigon, V. (1988) Proc. Natl. Acad. Sci. USA 85,4431-4435.

12. Tautz, D. (1989) Nucleic Acids Res. 17, 6463-6471.13. Saiki, R. K., Scharf, S., Faloona, F., Mullis, K. B., Horn, G. T.,

Erlich, H. A. & Arnheim, N. (1985) Science 230, 1350-1354.14. Deininger, P. L., Jolly, D. J., Rubin, C. M., Friedmann, T. &

Schmid, C. W. (1981) J. Mol. Biol. 151, 17-33.15. Schmid, C. W. & Jelinek, W. R. (1982) Science 216, 1065-1070.16. Dausset, J. (1986) Presse Med. 15, 1801-1802.17. Drayna, D. & White, R. L. (1985) Science 230, 753-758.18. Donis-Keller, H., Green, P., Helms, C. et al. (1987) Cell 51,

319-337.19. Warren, A. C., Slaugenhaupt, S. A., Lewis, J. G., Chakravarti, A.

& Antonarakis, S. E. (1989) Genomics 4, 579-591.20. Miyamoto, M. M., Slightom, J. & Goodman, M. (1987) Science 238,

369-373.21. Wiginton, D. A., Kaplan, D. J., States, J. C., Akeson, A. L.,

Perme, C. M., Bilyk, I. J., Vaughn, A. J., Lattier, D. L. & Hutton,J. J. (1986) Biochemistry 25, 8234-8244.

22. Woods-Samuels, P., Kazazian, H. H., Jr., & Antonarakis, S. E.(1989) Am. J. Hum. Genet. 45, 230 (abstr.).

23. Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R.,Horn, G. T., Mullis, K. B. & Erlich, H. A. (1988) Science 239,487-491.

24. Wong, C., Dowling, C. D., Saiki, R. K., Higuchi, R. G., Erlich,H. A. & Kazazian, H. H., Jr. (1987) Nature (London) 330, 384-386.

25. Lathrop, G. M. & Lalouel, J. M. (1988) Am. J. Hum. Genet. 42,498-505.

26. Ott, J. (1976) Am. J. Hum. Genet. 28, 528-529.27. Southern, E. M. (1975) J. Mol. Biol. 98, 503-517.28. Jeffreys, A. J., Royle, N. J., Wilson, V. & Wong, Z. (1988) Nature

(London) 332, 278-281.29. Olson, M., Hood, L., Cantor, C. & Botstein, D. (1989) Science 245,

1434-1435.30. Levinson, G. & Gutman, G. A. (1979) Mol. Biol. Evol. 4, 203-221.31. Sargentini, N. J. & Smith, K. C. (1985) Mutat. Res. 154, 1-27.32. Williams, S. M. & Strobeck, C. (1985) J. Theor. Biol. 116, 625-636.33. Wabl, M., Meyer, J., Beck-Engeser, G., Tenkhoff, M. & Burrows,

P. D. (1985) Nature (London) 313, 687-689.34. Orita, M., Suzuki, Y., Sekiya, T. & Hayashi, K. (1989) Genomics

5, 874-879.

2954 Genetics: Economou et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

8, 2

020