Proc. Nati. Acad. Sci. USAVol. 84, pp. 3946-3950, June 1987Population Biology

Ribosomal RNA-encoding DNA introgression across a narrowhybrid zone between two subspecies of grasshopper

(Caledia capiva/gene conversion/specation)

M. L. ARNOLD, D. D. SHAW, AND N. CONTRERASPopulation Genetics Group, Research School of Biological Sciences, The Australian National University, Canberra, ACT 2601 Australia

Communicated by R. 0. Slatyer, February 3, 1987 (received for review October 10, 1986)

ABSTRACT A ribosomal RNA-encoding DNA (rDNA)cloned sequence, consisting of a 0.8-kilobase fragment from the26S/nontranscribed spacer region, was used to identify diag-nostic restriction enzyme fragments that distinguish the More-ton and Torresian subspecies of the grasshopper Calediacaptiva. These restriction fragments were then used to studypatterns of rDNA variation across a narrow geographicalhybrid zone between the two subspecies. The pattern of rDNAvariation that emerged after the analysis ofover 250 individualsclearly demonstrates the asymmetrical introgression of theMoreton ribosomal RNA genes into the Torresian subspecies.This asymmetric movement of genetic material occurs eventhough there exists extreme postmating F2 and backcrossinviability between the two subspecies. From our data, as wellas those of previous chromosomal and allozymic studies, we areable to support the occurrence ofnonrandom processes such asbiased gene conversion and/or natural selection. Because therDNA loci in the Moreton and Torresian individuals are locatedin different regions on chromosomes 10 and 11, it should bepossible to determine the relative contributions of conversion,natural selection, and these sorts of processes to the pattern ofintrogression of the Moreton rDNA into the Torresian subspe-cies.

How reproductive isolation develops during the speciationprocess remains a fundamental and unanswered question inevolutionary biology. As a corollary of this question, one caninquire about the consequences of repeated hybridizationbetween partially reproductively isolated populations. Grant(1) has stated, in reference to the findings of Anderson (2),that long-term backcrossing of hybrid individuals to theparental types can result in ".... convergences betweenpreviously separate phyletic lines ...." However, it is alsoapparent that the introduction, maintenance, and persistenceof foreign genetic material, despite significant pre- and/orpostmating isolation, must be mediated by factors such asgenetic drift, migration, and/or natural selection. Thus, thedefinition of interacting processes that lead to introgression(2) of genetic material is of fundamental importance inunderstanding how genetic systems may be perturbed. In-deed, it has been advocated that a major stimulus to speci-ation involves hybridization (3). Furthermore, the quantita-tion of the effects of factors such as genetic drift, migration,and selection is equally fundamental in understanding theevolution of organisms.

In order to examine the above parameters, it is necessaryto identify hybridizing taxa that have the following charac-teristics: (i) significant levels of reproductive isolation; (ii)morphological, physiological, or genetic markers that can beused to identify genomic components in parental and hybridindividuals; and (iii) definable differences in the habitats

occupied by the taxa. Such a system has been identified in thegrasshopper Caledia captiva. In particular, two subspecies(Moreton and Torresian) of C. captiva meet and form a zoneof hybridization in southeast Queensland (4, 5). Thesesubspecies are differentiated at the level oftheir chromosomestructure, the patterns of allozyme variation and by thedistribution and abundance of highly repeated DNA se-quences (6-9). In laboratory hybridization experiments it hasbeen demonstrated that the F2 generation is completely in-viable, while the backcross generations show =50%o reduc-tion in viability (10). The Moreton and Torresian subspeciesoccupy different climatic regions for the majority of theirdistributions. They do, however, meet and form narrowzones of hybridization in which a major changeover (>50o)in chromosomal frequencies from the Moreton metacentricform to the Torresian acrocentric form occurs over a distanceof only 200 meters.The development of molecular techniques has made it

possible to assay genetic variation of natural populations ata very fine level. In relation to these analyses at the molecularlevel, the unique attributes of the Moreton/Torresian hybridzone now permits the assessment of those factors that affectthe exchange and introgression of specific DNA sequences.In this paper we report the pattern of ribosomal RNA-encoding DNA (rDNA) variation derived from over 250individual grasshoppers collected from within and outsidethis narrow hybrid zone. We are thereby able to test for theeffects from the action of the processes of natural selection,genetic drift, gene conversion and/or migration on this locusand to compare equivalent patterns ascertained from geno-mic analysis ofthe multiple chromosomal differences ofthesetwo subspecies.

MATERIALS AND METHODSDNA Preparation. The individuals used in this study

included 189 grasshoppers from six populations along atransect across the Moreton/Torresian hybrid zone that hadall been previously analyzed for their chromosomal andallozymic characters (11). Total DNA was isolated fromindividual grasshoppers using the technique of Appels andDvofdk (12). In addition, DNA was isolated from individualsoriginating from six other Queensland populations. Theseincluded three Moreton populations (Peregian Beach, n = 16;Mary Smokes Creek, n = 5; and Scrubby Creek, n = 16) andthree Torresian populations (Bongmuller, n = 16; NearaCreek, n = 5; and Insulator Creek, n = 16).

Restriction Enzyme Analysis and Gel Electrophoresis. Twomicrograms ofDNA was digested with four to six units of ClaI restriction endonuclease for 1 hr at 37°C. The restrictedDNA was electrophoresed in a 1% agarose gel in 0.4 MTris/0.2 M NaOAc/0.02 M EDTA at 100 mA for 4-5 hr.

Abbreviation: rDNA, ribosomal RNA-encoding DNA.

3946

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

Janu

ary

8, 2

020

Proc. Natl. Acad. Sci. USA 84 (1987) 3947

5' ...TTCGCCAAGCGTTGGGATTGTTCACCCACTAATAGGGAACGTGAGCTGGOTTTAGACCGTCGTGAGACAG

GTTAGTTTTACCCTACTGATGACTGTGTCGTTGCGATAGTAATCCTGCTCAGTACGAGAGGAACCGCAGTTCGGA

CATTTGGTTCACGCACTCGGCCGAGCGGCGTGGTGTGAAGTACATCTGGATTAAGCTAGCTTAGCATGTTAGCAT

TGCA.. .3'

FIG. 1. Partial nucleotide sequence for the 0.8-kb clone from the C. captiva rDNA locus. The 0.8-kb sequence was first cloned into the phageM13mpl8 and then sequenced using the dideoxy chain-termination sequencing protocol (14). This cloned sequence demonstrates significanthomology (bold-faced type) to the 3' end of previously sequenced 26S rRNA genes (15). The portion of this clone not yet sequenced extends3' to the sequence shown and, therefore, would include the remainder of the 3' end of the 26S gene as well as a portion of the nontranscribedspacer region.

Phage X, digested completely with HindIII, was used for sizedeterminations on each gel.Southern Blotting and Hybridization of rDNA Probes to

Filter-Bound DNA. DNA was transferred from agarose gels toGeneScreen (New England Nuclear) (13). Nick-translatedprobes were synthesized using standard techniques. Probeswere made using a 0.8-kb cloned sequence from the 26S/nontranscribed spacer region from the C. captiva rDNAcistron.

Quantification of the Relative Amounts of rDNA. Therelative amounts of the rDNA in an individual were estimatedusing a Quick Scan Jr. scanning densitometer (Helena Lab-oratories, Beaumont, TX). This quantification facilitated asensitive assay of the proportions of the restriction fragmentscharacteristic for the Moreton and Torresian subspecies forboth individuals and populations.

RESULTSrDNA Variation in the Moreton, Torresian, and Hybrid

Zone Individuals. Cla I-digestion of individual DNA samples,followed by hybridization to the 26S/nontranscribed spacerprobe (Fig. 1), produces a number of fragments from -2.1kilobases (kb) to 11.5 kb in size (Fig. 2). Hybridization to an18S-specific probe from rye (12) revealed that none of thefragments, with the possible exception of the 11.5-kb band,have homology to the 18S gene (data not shown). OftheDNAfragments that show homology to the 26S/spacer sequence,the 2.1-kb and 2.9-kb fragments are diagnostic for the

11.510.0

2.1 _

A BFIG. 2. Autoradiograph of Torresian (A) and Moreton (B) indi-

vidual DNA that has been digested with Cla I, subjected toelectrophoresis, blotted, and then hybridized to the 26S/nontran-scribed spacer probe. Note the diagnostic fragments correspondingto sizes of 2.1 kb (Torresian) and 2.9 kb (Moreton).

Torresian and Moreton taxa, respectively. With respect tothis.polymorphism, DNA sequencing analyses are currentlyunderway that will demonstrate whether the size differencebetween these two fragments is due to an insertion/deletionevent or the loss/gain of a Cia I restriction site.'The diagnostic nature of the above two fragments has been

demonstrated by assaying Moreton and Torresian popula-tions that occur at varying distances from the hybrid zone(Fig. 3). The Insulator Creek population is located =1200 kmfrom the hybrid zone. Of the 16 individuals analyzed in thisstudy, 14 possessed the 2.1-kb fragment, whereas two indi-viduals were characterized by a'3.2-kb Cla I restrictionfragment (Table 1). This latter fragment pattern has not beendetected in any other individual analyzed thus far. None ofthe Insulator Creek individuals possessed the 2;9-kb restric-tion fragment. Each of the remaining five populations fromoutside the hybrid zone, as well as the hybrid zone samples

1'

_ 1100 Km L .

FIG. 3. Populations sampled for rDNA variation. 1, InsulatorCreek (pure Torresian); 2, Bongmuller (introgressed Torresian); 3,Scrubby Creek (pure Moreton); 4, Peregian Beach (pure Moreton);5, Neara Creek (introgressed Torresian); 6, hybrid zooe populations(see text and ref. 11 for details); and 7, Mary Smokes Creek (pureMoreton).

Population Biology: Arnold et al.

I

Dow

nloa

ded

by g

uest

on

Janu

ary

8, 2

020

3948 Population Biology: Arnold et al.

Table 1. Frequencies of the 2.1-kb and 2.9-kb restrictionfragments for six C. captiva populations

Population Sample size 2.1-kb 2.9-kb

Insulator Creek* 16 0.88Bongmuller 16 0.92 0.08Neara Creek 5 0.83 0.17Scrubby Creek 16 1.00Peregian Beach 16 1.00Mary Smokes Creek 5 - 1.00

*Two individuals from the Insulator Creek population possessed a3.2-kb length variant.

themselves, possessed a proportion Qf individuals character-ized by the Moreton 2.9-kb fragment (Table 1). Significantly,all of the 37 individuals from the three populations (ScrubbyCreek, Peregian Beach, and Mary Smokes Creek) locatedwithin the Moreton subspecies range, possessed only theMoreton 2.9-kb fragment. In contrast, the two populations(Bongmuller and Neara Creek) from the Torresian subspeciesthat were within 10 km of the hybrid zone contained both theTorresian and Moreton restriction fragments.The Distribution ofrDNA Variants Across the Hybrid Zone.

In the present study we have utilized the same individualsanalyzed by Shaw et al. (11). These individuals, collectedfrom the hybrid zone in southeast Queensland (Fig. 3), havebeen previously analyzed with respect to chromosomal andallozymic markers. We have also examined individuals fromsix additional populations (see Fig. 3 and Table 1) fromoutside this zone.

Fig. 4 represents an autoradiograph of 13 individuals fromthe Moreton side of the hybrid zone along with a pureTorresian (lane N) and a pure Moreton (lane 0) controlsample. Although there is variation in the presence/absenceand the relative amounts of the-larger fragments seen in thisfigure (e.g., lanes C, D, and G), these fragments did not proveto be diagnostic for the two subspecies and were, therefore,not included in our analysis; however, it is important to notethat this variation may be due to the gain or loss of Cla Irestriction sites or to insertion/deletion events involving therDNA cistron. In terms of the 2.1-kb Torresian fragment andthe 2.9-kb Moreton fragment, there are differences in theirrelative amounts in the heterozygous individuals located inlanes F, I, and M (Fig. 4). Densitometric readings from thisautoradiograph demonstrated that the proportion of theMoreton 2.9-kb fragment relative to the Torresian 2.1-kbfragment was 0.56, 0.54, and 0.60 in individuals F, I, and M,respectively.

Sixty-three individuals from the six hybrid zone popula-tions were heterozygous for the 2.9-kb and 2.1-kb fragments.

3.12.9

Table 2. Frequency of the 2.9-kb fragment relative to the 2.1-kbfragment for 20 heterozygous TA-1 individuals

Frequency of the Frequency of theIndividual 2.9-kb fragment Individual 2.9-kb fragment

1 0.56 11 0.082 0.61 12 0.913 0.32 13 0.574 0.64 14 0.635 0.65 15 0.276 0.58 16 0.547 0.48 17 0.628 0.51 18 0.189 0.22 19 0.4910 0.72 20 0.66

Table 2 contains the relative proportion of the 2.9-kb to the2.1-kb Cla I fragment for 20 randomly chosen, heterozygousindividuals from the TA-1 hybrid zone population (Fig. 5).The 2.9-kb fragment ranges in relative proportion from 0.08in individual 11 to 0.91 in individual 12. The relative propor-tions of the 2.9-kb fragment in each individual analyzed werethen averaged to give an estimate of the populational 'fre-quency of this variant. Fig. 5 illustrates the frequency of theMoreton (2.9-kb) rDNA variant across the hybrid zone, aswell as in populations located within 10 km on either side ofthe zone. In addition, the frequency of the Moreton chro-mosome 1, which is representative of the variation foundwithin the zone for chromosomes 1-10 (11), is also included.

DISCUSSIONrDNA Introgression Across the Hybrid Zone. The individ-

uals collected from the hybrid zone in southeast Queenslandhave been previously analyzed with respect to chromosomaland allozymic markers. As a result ofthis analysis, significantlevels of gametic disequilibrium (D) in the chromosomalgenotypic frequencies on the Moreton side ofthe zone but noton the Torresian side, have been detected (11). This resultwas in direct contrast to an earlier study of the hybrid zonein which the gametic disequilibrium (D) was also highlysignificant, but on the Torresian and not theMoreton side ofthe zone (4). This complete reversal of D was shown tocorrelate with an extreme fluctuation (drought) in yearly

L..Q)

cJ

0Q)

LL

0

0

0)

a)

LL

*2.9

2.1

A B C D E F G H J K L M N 0

90 H

70 K

50D

30H

lo - ChromosomeL,Ii INeara-5km-< n.T mcN-1IOkmMSKCreek < < H

FIG. 4. Autoradiograph of 13 individual DNA samples from thehybrid zone population TB (lanes A-M) and the pure Torresian (laneN) and pure Moreton (lane 0) markers. Each DNA sample was

digested with Cla I and, following electrophoresis and blotting, was

hybridized to the 26S/nontranscribed spacer probe. Note thatindividuals F, I, and M contain both Torresian and Moreton rDNAtypes.

Population

FIG. 5. Frequency of the Moreton rD)NA and chromosomalmarkers for the six hybrid zone populations (TA, TA-1, TA-2, TA-3,TA-4, TA-5, and TB), as well as an introgressed Torresian (NearaCreek) and a pure Moreton (MSK, Mary Smokes Creek) population.Note that the distance from TA to TB is 1 km.

Proc. Natl. Acad Sci. USA 84 (1987)

An

la a* no 40 a* db 4b AN -

Dow

nloa

ded

by g

uest

on

Janu

ary

8, 2

020

Proc. Natl. Acad. Sci. USA 84 (1987) 3949

rainfall. On the basis of the more recent survey it wasproposed that the disequilibrium was a result of directionalselection favoring the Moreton metacentric karyotype duringmesic conditions and a Torresian acrocentric type duringxeric periods. In each of the above studies, the majorchangeover (>50% for each member of the genome) from theMoreton to the Torresian karyotype occurred at the samepoint over a distance of 200 meters. One of the nucleolarorganizing chromosomes (chromosome 11) was not assayedin this analysis (11) due to the lack of diagnostic markers;however, no introgression was detected for either chromo-somes 1-10 or the allozyme variants.

Several observations indicate differences in the patterns ofvariation across the hybrid zone with respect to the rDNAand chromosomal markers, even though the large-scalefrequency changeover occurs in precisely the same 200-meterinterval (see Fig. 5). On the Moreton side of the zone theMoreton rDNA variant and the Moreton chromosomesrapidly attain fixation. Likewise, the Torresian chromosomesare fixed on the Torresian side of the zone within 5 km of theTA population. This changeover can be explained by thehybrid breakdown in the F2 and backcross generations (10).Therefore, there is an intense postmating reproductive bar-rier against the introduction of chromosomes from onesubspecies into the other. Thus, the significantly higherfrequency of the Moreton rDNA variant in the hybrid zonepopulations TA-3, TA-2, TA-i, and TA (Fig. 5), as well as itsoccurrence in Torresian populations outside the zone, indi-cate the involvement of factors that are sufficiently strong tooverride the severe postmating inviability.Random and Nonrandom Evolutionary Processes and the

Introgression of the 2.9-kb Variant. Shaw et al. (11) haveindicated the lack of evidence for either population sizefluctuations, which would favor genetic drift, or significantmigration of individuals across the zone in either direction. Inthe present analysis, we have sampled populations that arelocated on both the Torresian and the Moreton side of thehybrid zone. Furthermore, we have sampled populationsproximal to both the Northern (Bongmuller and ScrubbyCreek populations) and Southern (Neara Creek, MarySmokes Creek, and Peregian Beach populations) portions ofthe zone (5). In both of these regions, which are =80 kmapart, the rDNA variants demonstrate the same distribution.Thus, the 2.9-kb and the 2.1-kb fragments are found on theTorresian side of the hybrid zone, but only the 2.9-kbfragment is present in the Moreton populations. It seemsunlikely that random population size fluctuations and/ormigration would result in such an identical pattern. Inaddition, the exact correlation in the position of the hybridzone over a period of six generations, which included a majorclimatic perturbation (11), would strongly suggest that thepresence of the Moreton rDNA variants on the Torresian sideof the zone is not a result of a shift in the position of the zone.Therefore, it would appear most likely that the factors thatresulted in the widespread introgression of the MoretonrDNA variant into the formerly pure Torresian populationswere nonrandom. Three such processes are biased geneconversion, natural selection favoring the Moreton rDNAvariant or, similarly, natural selection favoring the rDNAchromosomes as a unit.

In its classical sense, gene conversion occurs when oneallele at a locus converts a second allele at a homologouslocus into its own allelic state (16). Such intragenic conver-sion results in an atypical ratio of the segregants produced bymeiosis (17). A similar process has seemingly been respon-sible for intergenic conversion events in the chorion andglobin multigene families (18, 19) and in a duplicated leucinegene (20). For a new sequence variant to spread throughouta family of repeats, as the result of conversion events, theconversion must be biased (i.e., it must favor the new

variant). This type of biased conversion has been observedfor intragenic recombination involving spore color loci inAscobolus immersus (21) and Sordaria brevicollis (22). In thepresent case, it is possible that biased gene conversion hasbeen one of the mediating factors in the introgression of theMoreton rDNA into the Torresian subspecies, and we willdiscuss below how this process may be detected in theMoreton/Torresian hybrid zone. The level of conversionevents necessary to account for the pattern of rDNA varia-tion in this hybrid zone would be high. However, it isconceivable that the hybrid nature of the individuals from thiszone facilitates such an elevated level of gene conversion.With respect to this and similar processes, it is important tonote the amount of variation in the relative proportions of the2.9-kb and 2.1-kb fragments in heterozygous individuals.Data for 20 heterozygous individuals were presented in Table2. If there are equal numbers ofrRNA genes on chromosomes10 and 11, as well as between Moreton and Torresianindividuals, and if there are no gene conversion-type eventsoccurring, then one would expect any variation to be duesimply to meiotic recombination and/or segregation. Fur-thermore, because very little, if any, recombination withinthe rDNA locus is expected (ref. 23, and see below), thevariation in the relative proportion of the two variants shouldbe directly attributable to the number of Torresian andMoreton rDNA loci present in the hybrid individual. Al-though we do not yet have an estimate ofthe relative numbersof rRNA genes on chromosome 10 compared with chromo-some 11, preliminary data do suggest that overall there areequivalent numbers of these genes in Torresian and Moretonindividuals. Thus, individuals that diverge greatly fromintegral proportions for the 2.9-kb and 2.1-kb fragments (forexample, compare individuals 11 and 12 of Table 2) mayreflect the action ofgene conversion or some form of unequalexchange. With respect to the latter process, Dvorak andAppels (23), by combining classical genetic and moleculartechniques, suggested that an unequal sister chromatid ex-change occurred within the wheat rDNA locus. This eventresulted in an unexpected proportion in the copy numbers oftwo of the length variants.

Saghai-Maroof et al. (24) have argued that frequencychanges in rDNA variants in a population of cultivated barleywere mediated by natural selection. However, these authorsconcluded that the mixture of selfing and random matingdictated that their measure of selective effects included notonly the effects generated from the rDNA loci, but also allother loci in the genome. In contrast, the unique analysis ofthe Moreton/Torresian hybrid zone, in terms of total geno-mic assessment, provides the opportunity to test for selectiveadvantage for the Moreton chromosomes 10 and 11 and,indeed, the rDNA loci themselves. Thus, it has been argued(11) that the Moreton metacentric karyotype was favored onthe Moreton side of the zone under specific environmentalconditions. However, an analysis of the same individualsindicates that the rDNA loci, or chromosomes 10 and 11 asa unit, may be selectively advantageous in a largely Torresiangenomic background. A feature of the Moreton and TorresianrDNA loci will facilitate the determination of whether or notthe rDNA loci themselves (or closely associated loci), or thecumulative effects of the loci on the nucleolar organizingchromosomes, may be responsible for the asymmetricalintrogression. This is due to the fact that the ribosomal RNAgenes occur in different locations on chromosomes 10 and 11in the Moreton and Torresian taxa (N.C., unpublished data).Thus, the Moreton rDNA loci are located on the short armsof chromosomes 10 and 11. In contrast, the Torresian rDNAgenes are present on the long arms of these same chromo-somes. If the rDNA from the Moreton type is found tointrogress by recombination with the Torresian chromo-somes rather than in an unrecombined form, then selection

Population Biology: Arnold et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

8, 2

020

3950 Population Biology: Arnold et al.

for the rDNA locus will be strongly supported. However, ifthe entire rDNA chromosome(s) is found in heterozygousindividuals, then one would suppose that selection was actingupon other loci besides or in addition to those involving theribosomal RNA genes. Alternatively, if the chromosomalmarkers in individuals that possess the Moreton rDNAvariant are of the Torresian type, a mechanism such as geneconversion would need to be invoked in order to explain thefindings.

We thank A. Marchant for reviewing a draft of this report and D.Coates for his help with the field collections. We are grateful to R.Appels for providing laboratory space for a part of this study and forhis help in cloning the 0.8-kb rDNA sequence. M.L.A. was support-ed by an Australian National University Postdoctoral Fellowship.

1. Grant, V. (1963) The Origin of Adaptations (Columbia Univ.Press, New York), p. 487.

2. Anderson, E. (1953) Biol. Rev. Cambridge Philos. Soc. 28,280-307.

3. Anderson, E. & Stebbins, G. L. (1954) Evolution 8, 378-388.4. Moran, C. (1979) Heredity 42, 13-32.5. Shaw, D. D., Moran, C. & Wilkinson, P. (1980) Symp. R.

Entomol. Soc. London 10, 171-194.6. Shaw, D. D., Wilkinson, P. & Moran, C. (1979) Chromosoma

75, 333-351.7. Arnold, M. L. & Shaw, D. D. (1985) Chromosoma 93, 183-

190.

Proc. Nati. Acad. Sci. USA 84 (1987)

8. Arnold, M. L., Appels, R. & Shaw, D. D. (1985) Cytobios 43,149-157.

9. Arnold, M. L., Appels, R. & Shaw, D. D. (1986) Mol. Biol.Evol. 3, 29-43.

10. Shaw, D. D. & Wilkinson, P. (1980) Chromosoma 80, 1-31.11. Shaw, D. D., Coates, D. J., Arnold, M. L. & Wilkinson, P.

(1985) Heredity 55, 293-306.12. Appels, R. & DvotAk, J. (1982) Theor. Appl. Genet. 63,

337-348.13. Southern, E. M. (1975) J, Mol. Biol. 98, 503-517.14. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl.

Acad. Sci. USA 74, 5463-5467.15. Appels, R., Moran, L. B. & Gustafson, J. P. (1986) Can. J.

Genet. Cytol. 28, 669-685.16. Roman, H. (1963) in Methodology in Basic Genetics, ed.

Burdette, W. J. (Holden-Day, San Francisco),, pp. 209-227.17. Case, M. E. & Giles, N. H. (1964) Genetics 49, 529-540.18. Iatrou, K., Tsitilou, S. G. & Kafatos, F. C. (1984) Proc. Natl.

Acad. Sci. USA 81, 4452-4456.19. Slightom, J. L., Blechl, A. E. & Smithies, 0. (1980) Cell 21,

627-638.20. Klein, H. L. & Petes, T. D. (1981) Nature (London) 289,

144-148.21. Leblon, G. (1972) Mol. Gen. Genet. 115, 36-48.22. Yu-Sun, C. C., Wickramaratne, M. R. T. & Whitehouse,

H. L. K. (1977) Genet. Res. 29, 65-81.23. DvofAk, J. & Appels, R. (1986) Genetics 113, 1037-1056.24. Saghai-Maroof, M. A., Soliman, K. M., Jorgensen, R. A. &

Allard, R. W. (1984) Proc. Natl. Acad. Sci. USA 81, 8014-8018.

Dow

nloa

ded

by g

uest

on

Janu

ary

8, 2

020