Copyright 0 1990 by the Genetics Society of America

Evidence for Horizontal Transmission of theP Transposable Element Between Drosophila Species

Stephen B. Daniels,* Kenneth R. Peterson: Linda D. Strausbaugh,* Margaret G. Kidwell+ and Arthur Chovnick*'

*Department of Molecular and Cell Biology, University of Connecticut, Stows, Connecticut 06268, and ?Departmentof Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721 Manuscript received September 2 1, 1989 Accepted for publication November 2, 1989

ABSTRACT Several studies have suggested that P elements have rapidly spread through natural populations of Drosophila melanogaster within the last four decades. This observation, together with the observation that P elements are absent in the otherspecies of the melanogaster subgroup, has lead to the suggestion that P elements may have entered the D. melanogaster genome by horizontal transmission from some more distantly related species. In an effort to identify the potential donor in the horizontal transfer event, we have undertaken an extensive survey of the genus Drosophila using Southern blot analysis. The results showed that P-homologous sequences are essentially confined to the subgenus Sophophora. The strongest P hybridization occurs in species from the closely related willistoni and saltans groups. The D. melanogaster P element is most similar to the elements from the willistoni group. A wild- derived strain of D. willistoni was subsequently selected for a more comprehensive molecular exami- nation. As part of the analysis, a complete P element was cloned and sequenced from this line. Its nucleotide sequence was found to be identical to the D. melanogaster canonical P, with the exception of a single basesubstitution at position 32. When the cloned element was injected into D. melanogaster embryos, it was able to both promote transposition of a coinjected marked transposon and induce singed-weak mutability, thus demonstrating its ability to function as an autonomous element. The results of' this study suggest that D. willistoni may have served as the donor species in the horizontal transfer of P elements to D.melanogaster.

ERY little is known about the origin and evolu- have been shown to be the causative agents of one of V tionaryhistory of mobile element systems.A the two well-documentedsystems of hybrid dysgenesis mobile element family might originate in a species de in D. melanogaster. Hybrid dysgenesis is a syndrome novo by mutational or recombination events that alter of correlated genetic abnormalities thatis induced in or reorder sequences already present within the ge- thegermline of progeny from certain intraspecific nome (e.g., FINNEGAN1985). Alternatively, mobile crosses(KIDWELL, KIDWELL and SVED1977). [For elements might be acquiredby horizontal transmission comprehensive reviews of P elements andP-M hybrid from some other species. Such transfer events could dysgenesis, see ENGELS (1 983, 1989).] be mediated by one or moreof a number of different P elements in D. melanogaster exhibitalmost no vectors (e.g., viruses). At present, information on the sequence polymorphism but are heterogeneous with mechanismsand relative frequencies of these two respect to size (O'HARE andRUBIN 1983). They can modes of origin is completely lacking. If horizontal bedivided into two types, complete (autonomous) transmission should proveto be other thana relatively elements and defective (nonautonomous) ones. Com- isolated, rare event, then the implications could have pleteelements are 2.9 kb in lengthand encode a a dramatic impact on many aspects of evolutionary transposaseenzyme (RIo, LASKI and RUBIN 1986). theory. Defective elements are deletion-derivatives that are The well-characterized P transposable element fam- smaller and variable in size. The deletion of internal ily represents an ideal system in which to examine a P element sequences occurs at a high frequency under variety of evolutionary issues. P elements were first conditions of activetransposition (VOELKERet al. detected in Drosophilamelanogaster andhave been 1984; DANIELSet al. 1985; SEARLESet al. 1986; Tsu- extensively studied at both the genetic and molecular BOTA and SCHEDL1986). Both types of elements have levels. They are particularly interesting because they 31-bp inverted terminal repeats and create an 8-bp target site duplication upon insertion. Defective ele- ' To whom reprint requests 5hould be addressed. ments can undergo transposition onlyin the presence The public;ktion costs of this article were partly defrayed by the payme~~t of active autonomous elements. of page charges. This articlen~ust therefore be hereby marked "advertisement" in accord;~ncewith 18 U.S.C. $1754 solrly to indicate this fact. One way tostudy the evolutionary origins and

Genettcc 124: :4:39-:4.5.5 (b'rb~n~a~.y,1990) 340 S. B. Daniels et al. histories of mobile element families is to undertake analysis, we isolated, cloned and sequenced an intact, surveys of their phylogenetic distributions.At present, functional D. willistoni P element. Our results support it is known that P-homologous sequences are distrib- the hypothesis that P elements have been transferred uted throughout thespecies groups that comprise the horizontally between Drosophila species. subgenus Sophophora (ANXOLAB~HERE, NOUAUDand P~RIQUET1985; LANSMANet al. 1985; DANIELSand MATERIALS AND METHODS STRAUSBAUGH1986), although they are absent from Fly stocks: All of the species used in this study are listed the species most closely related toD. melanogaster, i,e., in Table 1 or in the legend to Figure 3; each listing includes the species of the melanogaster subgroup (BROOKFIELD, one of the following designations to denote the laboratory MONTGOMERYand LANGLEY1984). This observation, from which the stock was obtained: BG, National Drosophila Species Resource Center, Bowling Green State University; together with theobservation that P elements are WH, W. HEED,University of Arizona; JC, J. COYNE,Uni- completelylacking in old laboratory strains of D. versity of Chicago, FA, F. AYALA,University of California melanogaster (BINGHAM,KIDWELL and RUBIN 1982; at Davis; LE, L. EHRMAN,State University of New York at ANXOLAB~H~RE,KIDWELL and P~RIQUET1988), sug- Purchase; MS, M. SEIGER,Wright State University. The D. willistoni P element was cloned from a line estab- gests that the transposon may entered the P have D. lished from a single female caught in Florida in 1983 by M. melanogaster genome by horizontal transmission, per- SEIGER.This isofemale line has been designated "willi #7." hapsas the result of some infectious process (e.g., Strains of D. melanogaster that wereemployed in this BINGHAM,KIDWELL, and RUBIN1982). Moreover, it study are as follows: Harwich and r2are strong P strains has been proposed that a distantly related memberof that contain many P element copies. Canton S is an M strain devoid of P elements (i.e., it is a true M strain). rySU6is a true the genusmay have served as the donorspecies (DAN- M strain with a small deletion in the coding region of the IELS et al. 1984; DANIELSand STRAUSBAUCH1986). rosy gene. C(I)DX,yf;ry5"6 is a compound-X, true M strain Whether P elements are of recent origin in D. mela- bearing the ry5"' allele. y sn"; ry50his an M strain containing nogaster or whether their introduction occurred at the hypermutable singed-weak allele (ENGELS1979, 1989) some more distant point in time has been the subject and a rosy eye color marker. ry5u6P[ry+ A2-3](99B), abbre- viated A2-3(99B), is a rySfJhline transformed with the engi- of some debate (e.g., ENGELS1986). neered P element, P[ry+A2-31, which contains a copy of a If the introduction weresufficiently recent to essen- wild type rosy gene inserted into afunctional P element with tially preclude any appreciable base sequence diver- a deleted third intron [for further details, see LASKI,RIO gence,then: (1) there must be at least one species and RUBIN(1 986) and ROBERTSONet al. (1988)]; A2-3(99B) otherthan D.melanogaster thatcontains an intact, contains two elements in reverse tandem at cytological functional P element, and (2) the elements (or atleast position 99B(H. ROBERTSON,personal communication). A2- 3(99B)/ryjOS individuals were obtained from the F, progeny a subset of them) within this species must have se- of a cross between homozygous A2-3(99B) and ry506 lines. quences that are virtually identical to those of the D. See LINDSLEYand GRELL(1 968) for a description of genetic melanogaster canonical P (O'HAREand RUBIN1983). symbols not described. The results of previous studies would suggest that the Plasmids: pr25.1 contains a genomic fragment from the D. melanogaster P strain, 82 (SPRADLING andRUBIN 1982). mostreasonable place to start a search for such a The cloned DNA consists of a complete P element flanked potential donor species is within the genus itself. Al- by genomic sequences from cytologicalregion 17C; the though there have beena number of studies that have nucleotide sequence of this P element has been determined examined the phylogenetic distributions of P-homol- (O'HAREand RUBIN1983). The 1.7-kb XhoIISalI fragment ogoussequences within Drosophila (BROOKFIELD, from the center of the P element was used as probe in some of our experiments. MONTGOMERY and LANGLEY1984; DANIELSet al. pr25.7BWC ("both wings clipped"), constructed by K. 1984; ANXOLAB~H~RE,NOUAUD and P~RIQUET1985; O'HARE,contains a P element that is lacking 39 bp from its LANSMANet al. 1985; DANIELSand STRAUSBAUCH 5' and 23bp from its3' end. This plasmid contains no 1986; STACEYet al. 1986), most of these have been flanking genomic DNA, and, for this reason, is an acceptable limited in scope to a fewspecies or to a particular probe for the purpose of screening different species for P element homologues. taxon within the genus. We have undertaken, there- The pry+4 plasmid carries a ry' transposon that was fore, a survey that is more comprehensive than any constructed by inserting a 7.2-kb Hind111 rosy locus frag- other performed to date both to screen for donor ment from ry'4 into the polylinker ofthe defective P element species candidates and tolook for other discontinuities borne by the Carnegie-4 plasmid vector (RUBIN and in the P elementdistribution that might represent SPRADLING 1983).The ry+ transposon is able to undergo transposition only in the presence of an element capable of potential cases of horizontal transfer. supplying trans-acting transposase. pry+4 wasused in the Although our survey did not reveal any previously transformation experiments and was provided by S. CLARK. unknown discontinuities, it did establish the willistoni The derivations of the D. willistoni plasmids, pDwP13- species group,and in particularthe willistoni X30 and pDwP13F-XE4, are diagrammed in Figure 1 and subgroup, as the only likely source of a donor species described below. The cloning of a P element from D. willistoni: Manip- within the genus. We subsequently selected one mem- ulations involving recombinant phages were performed by berfrom this subgroup, D.willistoni, for a more the methods described by DAVIS, BOTSTEINand ROTH extensive molecular examination. In the courseof this (1980). Adultgenomic DNA from the D. willistoni line, Interspecific Transfer of P Elements 34 1

of this construct, which has been designated pDwP13-X30, is shown in Figure 1. The plasmid, pDwP13F-XE4, is de- rived from pDwP13-X30 and contains the 1.4-kb XbaI/ EcoRI D. wiltistoni genomic fragment that is located approx- imately 1.3 kb upstream from the 5’ P element terminus. X X DNA sequencing: P-containing DNA fragmentsfrom pDwP 13-X30 were subcloned into pEMBLl8 orpEMBL 19 -5.8 kb plasmid vectors and single-stranded template DNA was pre- pared as described by DENTE,CESARENI and CORTESE (1983). DNA sequences were determined by the dideoxy- chain termination method described by SANGER,NICKLEN and COULSON(1977) using T7 DNA polymerase (United States Biochemicals). Southern blotting: Genomic DNA samples were routinely prepared from approximately 0.12 g of adult flies by the method described by DANIELSand STRAUSBAUCH(1986). pEMBL18 DNA was quantitated by the fluorometric assay described 4.0 kb by KISSANE and ROBINS(1 958). Procedures for restriction pDwP13-X30 enzyme digestion, agarose-gel electrophoresis, gel blotting f 9.7 kb and preparation of nick-translated probes are described in rn RUSHLOW,BENDER and CHOVNICK (1984).Gels involved in the survey of the genus were blotted to Nytran (Schleicher & Schuell); those involved in all other experiments were blotted to nitrocellulose (Schleicher & Schuell). The same blotting conditions were used for both membranes. Probe DNAs were double labeled with [“‘PIdCTP and [”PIdTTP. Hybridizations were carried out overnight in a mixture containing 50% formamide (Fluka), 5 X SSPE t0.75 M NaCI, 0.05 M NaHP04, 0.01 M ethylenediamine-tetraacetate (EDTA), pH 7.01, 2 X Denhardt’s solution [0.04% (w/v) FIGURE1 .-The isolation and cloning of a full-si7ed P element Ficoll (type 400), 0.04% (w/v) polyvinylpyrrolidone, 0.04% from the D. willistoni genome. A lambda phage clone, designated (w/v) bovine serumalbumin], 1% sodium dodecyl sulfate PI:,, containing a potentially intact P element was isolated from a (SDS), 100 pg/ml sheared salmon sperm DNA, and 0.1 rg/ “willi #7” lambda EMBL3 phage library as described in the MATE- ml “‘P-labeled probe DNA. Filters were washed for 3 hr in RIALS AND METHODS. A 5.8-kb Xbal genomic fragment containing a series of solutions of increasing stringency, the final one the P element and someflanking DNA was subsequently subcloned of which contained 0.1 X SSPE and 0.5% SDS. Hybridiza- from PI:, into thepEMBL18 plasmid vector to create pDwPl3-X30, tions and subsequent washes were done at either 42 or 37”. a limited restriction map of which is shown. Sequencesderived Manipulations at 42” detect sequences with -78% identity fromlambda EMBLS are shown as fully shaded, those from to the probe DNA; manipulations at 37” detect those with pEMBLlX as unshaded and those from the P element as striped. -73% identity, assuming a 40% G/C content, which is the The genomic sequences flanking the P element are depicted by content of the canonical D. melanogaster P element (O’HARE thin bars. Constructs are drawn roughly to scale. The pDwP13- and RUBIN 1983). X30 restriction map is magnified threefold. Restriction enzyme Scanning densitometry: All autoradiograms of the P recognition sites are designated as follows: A = AvaI; E = BcoRI; H element survey of the genuswere subjected to scanning = HindIIl; K = KpnI; P = Pstl; S = Sal];T = SstI; V = PuuIl; X = densitometry(LKB UltroScan XL) in orderto roughly Xbal; Xh = XhoI; and XI11 = XmnI. The plasmid, pDwP13F-XE4, quantitate relative levels of P hybridization under defined is a derivative of pDwPIY-XSO, and contains the 1.4-kb genomic conditions. Each autoradiogram contained a positive (Har- fragment that is located approximately 1.3 kb upstream from the wich) and negative (Canton S) control lane and several lanes 5’ end of the P element. It was made by first digesting pDwPI3- containing DNA from different species. Identical amounts X30 DNA with EcoRl to remove all of the P element and some of of DNA were loaded in each lane with the exception of the the adjacent DNA and then recircularizing the plasmid by ligation. Harwich control, which had ‘hthe amount of DNA of the other samples. The entire length of each lane was scanned and the density of all bands measured; a single, composite “willi #7,” was digested to completion with BamHI and “raw score” was thencalculated by integratingthe area ligated to BamHI-digestedlambda EMBL3 phage DNA. under allof the peaks. The Canton S lane provided the The ligation mix was packaged invitro and plated onto baseline value. So that comparisons could be made between Escherichia coli KH802 cells. The resultingphage library different autoradiograms, afinal, standardized score was was screened by plaque hybridization using nick-translated, computed for each sample by expressing its “raw score” as “P-labeled BWC DNA. Phage DNAs containing P element a percentage of the Harwich score. All samples were then sequences were further characterizedby restriction enzyme placed on an arbitrarily defined, four pointscale to indicate and Southern blot analyses in order to identify those with relative strength of hybridization (see Table 1). potentially full-sized elements. One clone, designated PIS, Northern blotting: Total RNA was extracted from adult was selected as apossible candidate. A 5.8-kb XbaI genomic flies as described by KRAWETZ,STATES and DIXON(1 986), fragment containing theP element and some of its flanking andchromatographed on oligo(dT) cellulose (AVIV and DNA was subsequently isolated from PIS andsubcloned into LEDER1972). Polyadenylated RNA was size fractionated in thepEMBLl8 plasmid vector(DENTE, CESARENIand a 1% agarose-formaldehyde gel and transferred toa Nytran CORTESE1983) by themethods described by MANIATIS, membrane (Schleicher & Schuell) by gel blotting. Hybridi- FRITSCHand SAMBROOK(1982). A limited restriction map zation and washes were performed according to the rec- 342 S. B. Daniels et al. ommendations of Schleicher 8c Schuell for RNA transfer to guarani pallidipennis Nytran. cardini In situ hybridization: Larval salivary gland chromosomes tripunctata modifiedmouth Darts were prepared and hybridized to biotin-labeled BWC DNA picture-winged by a slight modification of the method describedby ENGELS histrio Dorsilopha' et al. (1986). Hybridizations were performed at 37 O. immigrans Transformation: Plasmid DNAs were introduced into Scaptomyza * the polarplasm of prepole cell embryos by the microinjection technique described by SPRADLINCand RUBIN(1 982). canalinea

RESULTS mesophragmatica robusta Thedistribution of P-homologoussequences melanica within the genus Drosophila: The primary goal in undertaking an extensive survey of the genus was to virilis-repleta identify possible candidates that might have served as radiation thedonor species in the horizontal transfer of P elements to D. melanogaster. A secondary goal was to obtain a more precise description of the phylogenetic Drosophila radiation distribution of P-homologous sequences within the Chymomyza genus, and, to a limited extent, among several of the 'melanogaster willistoni. saltans' closely-related, derivative genera. In all, 144 species *obscura~16 were examined, 136 from within the genus (- 10% of Sophophora radiation the presently described species), representing 26 spe- 'victoria Irlatifasciaeformis* cies groups and all the major radiations, and eight from outside the genus, representing five other gen- Scaptodrosophila radiation era. Information concerning the evolution and tax- onomy of the genus Drosophila, can be found in I THROCKMORTON(1975) and BEVERLEY andWILSON FIGURE2.-Depiction of the evolution of the genus Drosophila, (1 984).The phylogenetic relationships of the groups showing the major radiations and the present day species groups examined in this study are depicted in Figure 2; the that have resulted from each (after THROCKMORTON1975). Only arrangement of subgenera, species groups and subgenera and species groups screened in this study are included; subgroups is given in Table 1. several derivative genera are also shown. Those groups exhibiting evidence of P element hvbridization in our survey are indicated by To assay forP-homologous sequences, genomic asterisks (*). DNA samples from the species of interest were di- gested with PuuII and subjected toSouthern blot analysis. In some experiments, filters were hybridized subgenus Dorsilopha. None of the 50 species tested and washed under several sets of conditions, using from the subgenus Drosophila showed evidence of P different stringency and probe DNA combinations. hybridization under the conditions employed in this Stringency conditions are described in the MATERIALS study. Previously, STACEYet al. (1986) reported that AND METHODS; the BWC plasmid andthe internal D.nasuta, an immigrans group memberfrom the XhoIISalI P fragment were used as probes. All auto- Drosophila subgenus, exhibited hybridization to P ele- radiograms were analyzed by scanning densitometry ment probes under conditions of moderate stringency. as described in the MATERIALS AND METHODS, and We have tested the same D. nasuta stock under similar samples were placed on an arbitrarily defined, four conditions and have found it to be negative. More- point scale to indicate relative strength of hybridiza- over, we have examined 14 other species from the tion. The results of the survey are shown in Table 1. immigrans species group andhave found no detectable The points to be made from the data areas follows: evidence of P hybridization. 1. Within the genus Drosophila, P-homologous se- 2. By far,the strongest P element signals were quences are essentially confined to the subgenus SO- observed among the species from the closely related phophora. Of the80 sophophoran species tested in willistoni and sultans groups (Figure 3). Only minor this study, 61 showed discernible hybridization to P differences were noted in the strength and patterns element probes. Positive results were obtained in all of hybridization when different probesand stringency 10 of the willistoni group species, all 1 1 of the obscura conditions were employed. With the exception of D. group species, 6 of the 9 sultans group species and 34 melanogaster, the only other species that consistently of the 50 melanogaster group species tested. In addi- demonstrated relatively strong hybridization under tion, very weak hybridization signals were detected in all conditions was D.narragansett fromthe obscura 2 of the 5 species tested in the subgenus Scaptodroso- group. Most of the other P-bearing species from the phila and in D.busckii, the sole member of the melanogaster and obscura groups exhibited either no Interspecific Transfer of P Elements 343 TABLE 1 Distribution of P element sequences within the genus Drosophila and among several related genera

Probeb

Soul-ce" SubgenusGenus SpeciesSubgroup group Species X/S(42") X/S(37") BWC(42") BWC(37") __ BG 10 10-00 1 1.O DrosophilaScaptodro- victoria lebanonensis casteeli - - sophila BG 1010-0021.0 lebanonensis lebanonensis - - BG 1010-0031.0 pattersoni - - BG 1010-0041.0 stoni - F BG 1020-005 1.O coracina dimorpha - - BG 1020-006 1.0 latijasciaeformis latijasciaeformis - F Sophophorawillistoniwillistoni equinoxialis (2)' ++++ ++++ ++++ ++++ insularis (5) - + F + paulistorum (20) ++++ ++++ ++++ ++++ Paulovskiana (1) ++++ ++++ ++++ ++++ tropicalis (1) ++ ++ ++ ++ willistoni (12) ++++ ++++ ++++ ++++ bocainensis capricorni (2) + + + + fumipennis (2) ++++ ++++ ++++ ++++ nebulosa (7) ++++ ++++ ++++ ++++ sucinea (3) + + + + saltans cordata neocordata (1) - - - - elliptica emarginata (2) - - - - neoelliptica (1) - - - - sturtevanti milleri (1) + + + + sturtevanti (2) ++++ ++++ ++++ ++++ saltans austrosaltans (1) ++++ ++++ ++++ ++++ lusaltans (1) ++ ++ ++ ++ prosaltans (2) +++ +++ +++ +++ saltans (1) ++++ ++++ ++++ ++++ BG 141 0 12-01141.0 obscura affnis afjnis - F - + BG 14012-0151.0 algonquin - F - ++ BG 14012-0171.0 azteca - + - + BG 14012-0181.0 bijasciata + ++ + ++ BG 14012-0191.0 narragansett +++ ++++ +++ ++++ BG 140 12-020.O 1 tolteca - F - + BG 14011-0091.0 obscura ambigua - F - + BG 14011-0101.0 miranda F + F + BG 1401 1-0111.0 persirnilis - F - + U' H pseudoobscura - F - + BG 14011-0131.0 subobscura - F - + BG 14024-0371.0 melanogasterananassae ananassae - - - - BG 14024-038 1.O bipectinata - F - F BG 14024-0391.0 malerkotliana - F - F BG 14024-0401.O parabipectinata - F - F BG 14024-0421.0 pseudoananassae - F - F BG 14024-043 1 .O uarians - - - - BG 14027-0461.0 elegans elegans - - - - BG 14026-0451.O eugracilis eugracilis - - - - BG 14025-0441.O ,.ficusbhila jcusphila F + F + BG 14021-0241.0 melanogastermauritiana - - - - Various melanogasterd J <: sechellia Various simulans BG 14021-0261.0 yakuba BG 14028-047 1.1 montium auraria BG 14028-048 1.O baimaii BG 14028-049 1 .O barbarae BG 14028-0501.O biauraria BG 14028-051 1 .O bicornuta BG 14028-052 1.O birchii BG 14028-0586.0 diplacantha BG 14028-053 1.O jambulina BG 14028-0541.0 kanapiae BG 14028-055 1 .O khaoyana 344 S. B. Daniels et al.

Table 1-Continued Probeb

Soulre" SpeciesSubgenus Genus group Subgroup Species X/S(42") X/S(37") BWC(37")BWC(42")

BG 14028-0561.0 kikkawai F BG 14028-057 1.0 lacteicornis t+ BG 14028-0581.0 lini F BG 14028-0591.0 mayrz F BG 14028-0601.0 nikananu + BG 14028-061 1.0 orosa F BG 14028-0621.0 pamula F BG 14028-063 1.0 pennae F BG 14028-0641 .0 punjabiensis F BG 14028-0651 .0 quadraria + BG 14028-066 1 .0 rufa + BG 14028-067 1 .0 seguyi - BG 14028-068 1 .0 serrata F BG 14028-0691.0 triauraria f+ B(; 14028-070 1 .O tsacasz + BG 14028-07 I 1 .0 vulcana + B(; 14023-033 1 .o suzukii luripennis + BG 14023-034 1.0 mimetica + BG 14023-0351 .0 pulchrella + BG 14023-036 1 .0 rajasekari - BG 14022-027 1 .0 takahashii lutescens F BG 14022-0281 .0 paralutea - BG 14022-0291.0 prostipennis F BG 14022-0301.1 pseudotakahashii - - BG 14022-03 1 1.0 takahashii BG 14022-0321.0 trilutea F BG 1 5 1 20- 19 1 1.O Drosophila funebris filnebris BG 15120-1921.0 macrospina limpiensis BC; 15120-1931.2 macrospina macrospinu BG 1.5120-1941.0 multispina BG 15120-1951.1 subfunebris BG 1.5040-1 171.1 annulimana aracatacas BCr 1.5040-1 181.0 gibberosa BC, 15040-1 191.0 talamancana WH melanica melanzca BC; 1.5100-1711.0 polychaeta polychaeta BG 15020-1 1 1 1.1 robusta robusta B(; 15250-1451.0 tumiditarsus repletoides LE virilis virilis BG 35050-1201.0 canalinea canalinea B(; 15060-1221.2 dreyfusi ramargoa BG 15070-12:11.3 gaucha BG 15070-1241.0 mesophragma- pavani tira BG 15083-1551.0 repleta fasciola ellisoni BG 15085-1641.4 hydei hydei BG 15082-1511.0 mercatorum mercatorum "H mulleri arizonensis w H mojauensis "H mulleri 1%' H peninsularis BG 15081-1431.1 serido BG 15270-246 1.0 histrio sternopleuralis B(; 1511.5-1871.0 immigrans hypocausta hypocausta BG 1511.5-1881.0 neohypocausta .5-1891.0 pararubida 5- 190 1 .0 rubida 1-1721.0 immigrans formosana 1-1731.0 immigrans 1-1741.0 signata 4-1861.0 lineosa lineosa 2-1751.0 nasuta albomicans 2-1761.0 kepulauana Interspecific Transfer of P Elements 345

Probeb

Source" Genus Subgenus Species group Subgroup Species X/S(42") X/S(37") BWC(42")BWC(37")

BG 15112-1771.0 kohkoa BG 15112-1781.0 nasuta BG 15112-1801.0 pulaua BG 15112-1811.0 sulfuragaster albostrigata BG 15112-1821.0 sulfurigaster bilimbata BG 15112-1831.0 sulfurigaster sulfurigaster BG 15112-1841.0 tongpua BG 15181-2181.9 cardini cardini cardini BG 15181-2201.0 neocardini BG 15 182-2291.O dunni dunni dunni BG 15171-2141.2 guaramunu guarani guaramunu BG 15210-2331.0 pallidipennis pallidipennis pallidipen- nis BG 15220-24 1 1.2 tripunctata unipunctata BG 15291-2551.0 mod.mth.parts hystricosa biseriata BG 15287-2541.0 grimshawipicture-winged grimshawi BG 15284-150 1 .0 hawaiiensis gymnobasis BG 1528.5-2521 .O pilimana aglaia BG 13000-0081.O Dorsilopha busckii BG 20000-2621 .O Chymomyza amoena BG 20000-263 1.1 procnemis BG 60000-275 1 .0 Liodrosophila aerea BG 80000-276 1 .0 Samoaia leonensis BG 3 1000-2641 .0 Scaptomyza Parascapto- adusta myra BG 3 1000-2651.O elmoi +++ BG 50000-2746.0 Zaprionus inermis - BG 50000-276 1 .O tuberculatus -

All genomic blots were analyzed by scanning densitometry (see MATERIALS AND METHODS). A plus (+) indicates that hybridization was observed; a minus (-) indicates that no hybridization was seen at the stringency employed. The number of pluses indicates the relative strength of the hybridization signal. An F indicates that one or more bands were faintly discernible on the blot but that the strength of the signal was not significantly above the negative control baseline level. * Letter designations are defined in MATERIALS AND METHODS and denote the laboratories from which the stocks were obtained. The sources for the willistoni and sultans group species are given in DANIELSand STRAUSBAUCH(1986). * X/S = 1.7-kb Xhol/Sall internal fragment from ~~25.1;BWC = pn25.7BWC (see MATERIALS AND METHODS). The hybridization and wash temperatures are in parentheses. ' More than one isolate was tested for many of the species in the willistoni and sultans groups; the number tested is in parentheses. P element-free D.melanogaster strains werelast isolated from the wild in the early 1970s, since then all wild-derived strains have contained multiple P element copies (ANXOLAB~H~RE,KIDWELL and P~RIQUET1988).

discernible or only weak hybridization when the in- Underthe simplest set of assumptions, thedonor ternal P probe was used atthe higherstringency. species in a recent transfer of P elements to D. mela- Signal intensity usually increased when eitherthe nogaster must fulfill one basic criterion: it must possess lower stringency or the BWC probe was employed. at least one complete P element with virtually the Sometimes the changes seen under the different hy- same base sequence, and hence restriction map, as the bridization conditions were quite striking, as is dem- D. melanogaster canonical P. When a complete P ele- onstrated in Figure 4. ment is digested with PvuII, the restriction enzyme 3. Of the five other genera surveyed, two, Liodro- employed in our survey, it producesan internally sophila and Scaptomyza, showed evidence of P element derived, 0.9-kb fragment. This fragment shares ap- hybridization. The single Liodrosophila species tested, proximately 750 bases with the XhoI/SalI P fragment L. aerea, exhibited veryweak signals tothe BWC and produces a strong hybridization signal when the probe. Significantly stronger signals were observed in latter is used as probe. All candidates for the donor the two Scaptomyza species examined, with one, S. species, therefore, should show the 0.9-kb band when elmoi, displaying relatively strong P hybridization. P probed with the XhoI/SalI fragment, even under con- homologues have also been reported in another Scap- ditions of high stringency. (It should be noted that tomyza species, S. pallida (ANXOLAB~H~RE,NOUAUD defective elements with deletions outside the PvuII/ and P~RIQUET1985). PvuII interval will also produce the 0.9-kb band.) Identifying potential donor species in the hori- Of all the species tested, only five, D. equinoxialis, zontal transfer of P elements to D. melanogaster: D. paulistorum, D. pavlovskiana, D. willistoni and D. 346 S. B. Daniels et al.

FIGURE J.”Southernblots of the willistoni and saltans species groups. A and B. Blots were hybridized and washed at 42”; BWC was used as probe. The first two lanes in each panel represent positive and negative controls from, respectively, Harwich and Canton S, both strains of D. mefanogasfer.The source of each stock is given in parentheses; abbreviations are defined in the MATERIALS AND METHODS. An arrow indicates the position of the 0.9-kb Puull fragment. The approximate positions of the fragments generated from a Hind111 digest of lambda DNA are indicated on the left. A, The wilfisfonispecies group. Samples are as follows: (3) D. equinoxiafis (BG14030-0741.1); (4) D. insufaris (FA); (5) D. paulisforum (C-2 isolate from LE); (6) D.paufouskiana (FA); (7) D. fropicafis(BG14030-0801 .O); (8) D.wiflistoni (“willi #7”); (9) D. capricorni (BG14030-0721.1); (10) D.fumipennis (BG14030-0751.1); (1 1) D. nebulosa (MS);and (12)D. sucinea (BG14030-0791.1). Note that P signals in the D. insufaris sample (lane 4) are only weakly visible at this exposure. B, The saftans species group. Samples are as follows: (3) D. neocordafa (BG1404 1-083 1 .O);(4) D.emarginata (BGI 4042-084 1 .O); (5) D. neoelfipfica(BG 14042-085 1 .O); (6) D.mifleri (BGI 4043-086 1 .O); (7)D. sturfevanfi(BG14043-0871.0);(8)D.avstrosalfans(BG14045-0881.0);(9)D.fusaftans(BG14045-0891.0);(lO)D.~rosaltans(BG14045- 0901.0); and (1 1) D.salfans (BG1404.5-091 1.0). nebulosa, possessed the diagnostic PvuII fragment (see 6 in DANIELSand STRAUSBAUCH1986), indicating the Figure 3). All of these reside within the willistoni possible presence of active elements. Based on these species group; all but D. nebulosa reside within the considerations, D. willistoni was chosen for the com- willistoni subgroup (see Table 1). The field of candi- prehensive analysis described below. Intact, active P dates was further narrowed by focusing on those that elements were notfound in theone strain of D. have ranges that extend into the southern portion of nebulosa examined by LANSMANet al. (1987). North America, where the earliest P-activity was de- Molecular characterizationof the P elements of a tected in D. melanogaster (ANXOLAB~H~RE,KIDWELL D. willistoni strain: The strain of D.willistoni selected and P~RIQUET1988). Of the five candidates, only D. for molecular analysis has been given the designation, willistoni and D.nebulosa have ranges that overlapthis “willi #7” (see MATERIALS AND METHODS). This line region. A more extensive screen of these two species, has been inbred for approximately five years. The which included seven D. nebulosa and 12 D. willistoni analysis consisted of estimating the total number of P lines, representing both recently caughtand older elements and determining whether any were intact laboratorystrains, was subsequently undertaken and transcriptionally active. (DANIELSand STRAUSBAUCH1986). Although all of An estimate of the total number of elements in the lines contained the 0.9-kb PvuII fragment, its “willi #7” was obtained by Southern blot analysis. intensity was always stronger in D. willistoni. More- Genomic DNA samples were simultaneously digested over, P element hybridization patterns were usually with BamHI and BglII, restriction enzymes that lack more varied in D. willistoni than in D. nebulosa, even sites in the P element. Such a digest yields a single among lines collected from thesame locale (see Figure band for each hemizygous and homozygous element Interspecific Transfer of P Elements 347

FIGURE4.-The influence of hy- bridizationconditions on P element patterns. Within the obscura and mela- nogaster groups,P-bearing species oftenshowed marked differences in the degree of hybridization depending on probe and hybridization conditions. A-C, Hybridizationintensities exhib- ited by three species from the melano- gaster group under three different sets of conditions. In each panel, the first twolanes contain positive (Harwich) and negative (Canton S) control sam- ples.Lanes 3-5 contain,respectively, DNAsamples from D. auraria, D. biauraria and D. facteicornis. The anlount of DNAin the Harwich lane is only ‘/5 that of the other samples.The same filter was used in (A) and (E); a different one was used in (C). A, Hy- bridi7ation and subsequentwashes per- formed at 42“; the internal Xhol/SafI fragment was used as probe.E, Hybrid- ization and washes performed at 37”. using the same probe as in (A). C, Hy- bridilationand washes performedat 42”. using the BWC plasmid as probe.

in the genome. BWC was used as probe. DNA samples 31 transposon yields a fragment that is about 200 bp from males and females were analyzed separately. By smaller because of the deleted third intron (see MA- this method, it was estimated that “willi #7” females TERIAL.~ AND METHODS). Dilutions of digested DNA have as many as 12, and “willi #7” males as many as from the “willi #7” samples were thencompared 13, prominent P elements per haploid DNA comple- against the A2-3(99B) reference samples. The A2- ment (Figure 5A), although other, much weaker sig- 3(99B) line contains two tandem P element transpo- nals were also noted upon longer exposure. The ad- sons (H. ROBERTSON,personal communication), with ditional signal in males suggests that there is a P on hemizygous [(A2-3(99B)/rySo6)]and homozygous [A2- the Y chromosome in this line. In situ hybridization of 3(99B)/A2-3(99B)]lines containing, respectively, two biotinylated-BWC DNA to larval salivary gland chro- and fourP transposons per diploid DNA complement. mosomes showed sites of hybridization in pericentric The approximate number of complete elements is and euchromatic regions (data notshown). determined by selecting the 2.2-kb band in the dilu- By comparing male and female DNA, it is also tion series that corresponds in intensity to the 2.0-kb possible to estimate the numberof X-linked sites, since band in the A2-3(99B) controls. From this analysis, bands representing them are roughly twice as intense we estimated that the“willi #7” line has twopotentially in female DNA as they are in male DNA. The number complete elementsper haploid DNA complement of X-linked sites in “willi #7” was determined by sub- (Figure 5B). jecting autoradiograms of male and female DNA to A Northern analysis of polyadenylated RNA from scanning densitometry. The results revealed that five “willi #7” revealed a P transcript of 2.5 kb (Figure 6). of the 12 prominent bands in this line are X-linked This is the sizeof the transcript produced by an (see Figure 5A). autonomous P element. An assessment of the number of potentially com- Isolating a full-sized P element from D. w’llistoni: plete elements was made by an analysis similar to the A lambda library of genomic DNA from “willi #7” one described by DANIEL.^ et al. (1987b). Samples adults was prepared and screened for P-hybridizing containing equal amountsof genomic DNA from A2- plaques by one of us (L.S.) as described in the MATE- 3(99B)/rySo6, A2-3(99B)/A2-3(99B),“willi #7” males RIALS AND METHODS. Positive clones were further and “willi #7” females were digested with DdeI. Com- screened by restriction analysis for those containing plete elements produce a large, 2.2-kb internal frag- potentially complete elements. One such clone, PIS, ment when digested with this enzyme; the P[ry+ A2- was identified. A 5.8-kb XbaI fragment containingthe 348 S. B. Daniels et al.

3 4 5 6

FIGURE&--Northern analysis of P element transcripts from "willi #7." Polyadenylated RNA was isolated, size fractionated in a 1 % agarose-formaldehyde gel and hybridized with a "'P-labeled BWC probe ias described in the MATERIALS AND METHODS. Lanes 1-3 contain control samples from, respectively, up, Harwich and Canton S; each lane contains approximately 2 pg of poly(A+)RNA. Lanes 4-6 contain. respectively, 2, 4 and 6 gg of poly(A+) RNA from "willi #7." The up and Harwich P strains display prominent 2.5-kb P element transcripts (arrow), whichis the predominant message produced by autonomous elements. No P element tran- scripts are seen in the Canton S negative control. The "willi #7" line exhibits a single P element transcript of 2.5 kb, which is much less abundant than in the strong P strains of D. melanogaster. The low molecular weight band seen in every lane is presumed to be an artifact, perhaps resulting from nonspecifichybridization to the abundant 9s globin message, with which it comigrates.

entire element and some flanking DNA was isolated from PISand subcloned into the pEMBLl8 plasmid vector to make pDwP13-X30 (see MATERIALS AND METHODS), a restriction map of which is shownin FIGURE5.-Analysis of the "willi #7" P element complement. A, Figure 1. Estimate of the total number of elements per genome. Genomic To verify that this element was derived from the DNA was digested with both BamHI and BgllI, enzymes that do not cut within the P element, and probed with BWC. The lane on "willi #7" genome, we performeda Southern blot the left contains DNA extracted from "willi #7" females; the lane analysis to determinewhether genomic digests of on the right contains DNA from "willi #7" males.Each band "willi #7" DNA produce fragments of the size pre- represents a single element. It is estimated that females and males dicted from the pDwP13-X30 restriction map. The have, respectively, 12 and 13 elements per haploid DNA comple- strategy is outlined in Figure 7A and involves restric- ment. Autosomal sites are indicated by dots, X-linked sites by stars. The male-specific site is indicated by an arrow. B, Estimate of the tion fragments that span both P element and flanking number of potentially complete elements in the "willi #7" genome. DNA sequences. The pDwP13F-XE4 plasmid, which Genomic DNA was digested with Ddel, which produces an internal contains the non-P DNA segment indicated in Figure 2.2-kb fragment from a complete P element and a 2.0-kb fragment 7A, was used as probe. Eight different digests were from the A2-3(YYB) transposon (see MATERIALS AND METHODS). The performed. In every case, a fragment of the expected BWC plasmid was used as probe. The two lanes on the left (C) contain equal amounts of DNA from the two reference samples, size was generatedfrom "willi #7" genomic DNA A2-3(YYB)/A2-3(YYB) and A2-3(9YB)/rysu6,respectively. The for- (Figure 7B), thus confirming that the cloned DNA mer contains four copies, and the lattertwo copies, of the P[ry+A2- fragment was derived from this line. 31 transposon per diploid DNA complement. The reference lanes are followed by two groups, each composed of three lanes. The groups represent dilution series of DNA from "willi #7" females and "willi #7" males. The first lane in each group contains the same amount ofDNA as the reference lanes (C); the next two lanes ing the dilution series to the reference samples, we estimate that contain, respectively, one-half and one-fourth the amount of DNA. males and females of the "willi #7" line contain two potentially full- Only the relevant Ddel fragment is shown in each lane. By compar- sized P elements per haploid DNA complement. Interspecific Transfer of P Elements 349

probe The D. willistoni P element and the D. melano- gaster canonical P differ by only a single base: The xs- EE V AXh KTP S X base sequence of the P element in pDwP13-X30 was determined by one of us (K.P.) as described in the

I 13.2 MATERIALS AND METHODS. The D. willistoni P element I 13.3 was found to be identical to the complete (and auton- I 13.3 omous) D. melanogaster P element described by I 13.7 I 13.8 O’HARE andRUBIN (1983), with the single exception I 13.9 of an A to G transition at position 32. As has been I 14.9 found in D. melanogaster, there was an 8-bp duplica- -,I.” tion of genomic DNA (in this case GCACAACT) at the site of insertion. A The single base difference in the cloned D. willistoni 12345678 P element results in the obliteration of a TaqI recog- - nition site. This site is absent in all of the “willi #7” P elements that could be monitored by the method described in Figure SA. The TaqI site was also found to be missing in the P elements of other members of the willistoni subgroup (Figure SB), suggesting that the guanine base at position 32 may be common or ubiquitous in P elements of this taxon. The D. w’llistoni element can promote the trans- position of other P element transposons: The North- ern and sequencing data together provide strong cir- cumstantial evidence that the D. willistoni element is capable of functioning as an autonomous P element, since there is no existing evidence to suggest that activity is affected by a base change at position 32. To confirm its ability to function, we have used the D. willistoni element as a “helper” plasmid in a transfor- mation experiment to see whether it is capable of promoting P-element-mediated genetransfer. The pry+4 plasmid (500 pg/ml), which carriesa ry+ transposon, was microinjected into the polarplasm of y sn’’; ry5”6 embryos, along with thepDwPl3-X30 “helper” (150 pg/ml), by the procedure described in the MATERIALS AND METHODS. Individual Go male and female survivors were mated to, respectively, C( ])DX, yf; ryro6females and ry5O6 males. AI1 GI progeny were screened for the presence of ry+ exceptions; GI males were also checked for mutations of the singed-weak FIGURE7.-Confirmation that the cloned P element is derived allele, which is known to contain two defective P from the D. willisfoni line, “willi #7.” A, Partial restriction map of elements in tandem (NITASAKAand YAMAZAKI1988; pDwPl 3-XSO. This plasmid contains a full-sized P element (thick bar) flanked on both sides by genomic sequences (thin bars). The ROIHA, RUBIN and O’HARE1988). Excision of one or 5.8-kb genomic fragment was cloned into theXbal site found in the the other of these elements produces two phenotypi- pEMBLI8 polvlinker. Restriction site designations are given in the cally distinguishable classes, sn(+) and sn‘ (ENCELS legend to Figure I. The fragments expected when this plasmid is 1979; ROIHA,RUBIN and O’HARE1988). Of the 35 digested as described in (B)are shown below the map. The size of fertile Go adults, eight (23%) had genetically trans- each (in kb) is given on the right.B. Southern blot of eight different restriction digests of “willi #7” genomic DNA. The pDvvPl 3F-XE4 formed ry+ individuals amongtheir progeny, and pl;tsnlid. which contains the piece of non-P DNA shown above the three (9%) had sn(+)or sn‘ exceptional progeny. All map in (A), \vas used as probe. DNA was digested as follows: (1) exceptions were true-breeding. In a separate experi- Xbal and Puull; (2) XbaI ant1 Aual; (3) Xbal and XhoI; (4) Xbal and ment, only pDwP13-X30 plasmid DNA (150 pg/ml) Kpnl: (5) Xbal and .%fl; (6) Xbal and Psfl; (7) Sall: and (8) XbaI. was injected intoy snw;ry5‘”j embryos, and Go survivors The second, weaker band in the Sall digest (lane 7) occurs because the probe DNA overlaps the genomic fragment to the left of the were mated as above. GI males were screened for Sal1 fr;tgnlent depicted in (A). The approximate positions of the evidence of snw mutability. Of the eight fertile Go fragments generated from a Hindlll digest of lambda DNA are survivors, two (25%) produced progeny with excep- indicated on the left. tional singed phenotypes. Theseexperiments dem- 350 S. B. Daniels et al. 20 3220 D. willistoni T ~CGA (~a9I onstrate that the P element is a functional element capable of providing the trans-acting product necessary for the mobilization of nonautonomous P AGC; (Hindlll) element transposons.

DISCUSSION

733 When did P elements enter the genus Drosophila? b28c+ (Xhol) The genus Drosophila arose during the Eocene age from one of several early radiations within the family Drosophilidae. It subsequently underwent extensive diversification in the form of five major radiations, ‘0 689bp 728 the earliest of which gave rise to the subgenus Scap-

30 700bp729 todrosophila. This was followed by the Sophophora - radiation, which produced the melanogaster, obscura, and closely related willistoni and sultans lineages. The A complex Drosophila radiation ensued. This radiation is composed of three main branches, represented by the virilis-repleta, immigrans and Hirtodrosophila radia- tions [see THROCKMORTON(1975) forfurther de- tails]. To identify potential donor species in the horizontal transfer of P elements to D. melanogaster, we first undertook an extensive survey of the distribution of P-homologous sequences within the genus Drosophila. Five other closely-related genera were also sampled, although much less extensively. The results showed that P sequences are essentially confined tothe subgenus Sophophora, although weak hybridization was also observed in a few species from the subgenus Scaptodrosophila and in D. busckii, the only species in the subgenus Dorsilopha. No P element hybridization was detected in species from the subgenus Drosophila. Relatively weak P-hybridizing sequences were found in species from the genera Liodrosophila and Scapto- myza. In other surveys, P homologues have been re- ported in D. phalerata and D. kuntzei, both assigned to thequinaria species group of the subgenus Drosoph- FIGURE8.-The base difference at position 32 may be ubiquitous ila (ANXOLAB~H~RE,NOUAUD and P~RIQUET1987), in P elements from the willistoni subgroup. A, Diagram of the D. and, interestingly, in two species from the nondroso- melanogaster P element, showing the restriction sites pertinent to philid families, Trixoscelididae and Opomyzidae this analysis. The TaqI recognition sequence is a subset of the Xhol (ANXOLAB~H~REand P~RIQUET1987). The occur- sequence. In P elements from D. melanogaster, the HindllI/Xhol and Taql/Taql fragments are nearly identical, spanning essentially rence of P-hybridizing sequences in species from the the same piece of DNAand differing in size by only 1 1 bases. When subgenus Scaptodrosophila, one of the earliest Drosoph- the 0.7-kb HindlII/Xhol fragment is present, it confirms the integ- ila lineages, suggests that P elements may have been ritv of both the DNA segment and the constituent Tag1 recognition present at thetime of the inception of the genus about sequence at the Xhol site. If the TagI recognition sequence between 60 million years ago. bases 29 and 32 is intact in all elements that yield a Hindlll/XhoI fragment. then a Tag1 digest should produce a fragment of about the same size and of equal intensity. B. Southern blot analysis of the Tuql polymorphism. Five pairs of lanes are shown. In each pair, comigrating, 0.7-kb bands that are of equal intensity, indicating the lane on the left contains DNA digested with Taql, and the lane that the Tag1 site spanning base 32 is intact in most or all of the on the right contains the same DNA digested withHind111 and elements with intact Hindlll/XhoI segments. On the other hand, P Xhol. BWC wvas used as probe. Samples are as follows:(1) A2-3(99B): elements of three willistoni subgroup species produce a prominent, (2) Harwich; (3) D. willistoni(“willi #7”); (4) D. equinoxialis 0.7-kb Hindlll/Xhol band, but lack the corresponding Taql/TaqI (BG 14030-074 1.1); and (5) D. paulistorum (C-2 strain). The arrow band. Since a functional Xhol site verifies the integrity ofthe indicates the position of the 698-bp HindIII/XhoI P element frag- constituent Tap1 recognition sequences at its 3’ end, the absence of ment. The approximate positions of the fragments generated from the Taql/TaqI band in willistoni subgroup species must be the result a Hind111 digest of lambda DNA are indicated on the left. In D. of an alteration within the Tag1 sequence that spans bases 29 to 32 melanogasfer (lanes 1 and 2). the two different digests produce at its 5’ end. InterspecificTransfer of P Elements 35 1

The evolutionary history of P elements in Dro- situation in which sequences are gradually drifting to sophila: To fully understand the evolutionary history extinction within the lineage. This may also be the of a particular transposable element within a phylo- case in the other lineages in which P hybridization is genetic lineage, it is necessary to determine its initial weak. A theoretical treatment of the evolution of a point of entry, its subsequent distributionand its mode hypothetical transposable element has suggested that of transmission between species (STACEYet al. 1986). a drift to extinction can occur under certain condi- If transmission has been strictly vertical (i.e., mating- tions (KAPLAN,DARDEN and LANGLEY 1985). dependent), then descendants of an ancestral species 2. Alternatively, it can be proposed that P elements bearing the element should also possess homologues may have entered Drosophila on more than one occa- of the element. If during evolution the element has sion and that certain features of the present distribu- been lost from a species, then all of its descendants tion may be the result of horizontal transmission (e.g., should be element-free.This mode of transmission LANSMANet al. 1985; STACEYet al. 1986). This form results in distribution patterns that are virtually con- of acquisition would perhaps explain the isolated oc- tinuous. Alternatively, if transmission has occurred currences of P sequences within the subgenus Dro- horizontally between reproductively isolated species, sophila (ANXOLAB~H~RE,NOUAUD and P~RIQUET then distribution patternsmay not follow phylogenetic 1987) and in the nondrosophilid dipterans (ANXOLA- groupings, i.e., they may be discontinuous. The fact BEH~REand P~RIQUET1987), as well as the conspicu- that vertical and horizontal modes of transmission are ous anomalies observed in the willistoni and sultans not mutually exclusive makes it difficult to decipher groups (Figure 3). All of the willistoni group species the exact chain of events leading to the present day have prominent P sequences, with the exception of D. distribution of P sequences within Drosophila, since insularis, which has elements that more closely resem- very complex patterns can result from the combina- ble the weak P homologues seen in the melanogaster tion of occasional horizontal transfer and subsequent and obscura groups. In the sultans species group, P vertical transmission. Moreover, the analysis can be sequences are absent in the moreprimitive subgroups, even further complicated by the possibility that P but are present in those that are more derived, the elements ]nay have been introduced into some line- reverse of what might be predicted. ages at more than one point during their evolution. The following account of the evolutionary histories Previous studies have uncovered two major discon- of the willistoni and sultans groups, taken from tinuities in the distribution of P-homologous se- THROCKMORTON(1975), may provide an understand- quences in the genus. With the exception of D.mela- ing of these anomalies and shed some light on the nogaster, P sequences were not found in species from possible time and location of P element introductions the melanogaster subgroup (BROOKFIELD,MONTGOM- into these lineages. During the Eocene age the pri- ERY and LANGLEY 1984), norwere they found in the mary tropical disjunction separated the New and Old more primitive subgroups of the sultans species group World tropics. As a result, the common ancestor of (DANIELS and STRAUSBAUGH 1986).present The study the willistoni-sultans lineage become sequestered in confirms the results of these earlier reports. No other tropical North America where the divergence of the major discontinuities were uncovered duringthe two groups is believed to have occurred. Thereafter, course of our survey. [In an earlier study, DANIELS it is proposed that an ancestor of the willistoni lineage and STRAUSBAUGH(1986) reported that P sequences invaded South America by crossing the water gap that were conspicuously absent from one (D. insularis) of then existed between the North and South American the species of the willistoni group; in this study, weak continents. The development and diversification of P hybridization was detected under conditions of re- the willistoni lineage is thought to have taken place duced stringency.] exclusivelyin South America. Today,the willistoni There aretwo basic ways to account for the current species group is divided into the willistoni and bocai- distribution of P-homologous sequences: nensis subgroups. The more primitive willistoni 1. If horizontal transmission is assumed to be very subgroup is composed of a cluster of sixsibling species, rare, then themost parsimonious interpretation of the one of which is D. insularis, a narrow island endemic data favors a single introduction of P elements into confined to the Lesser Antilles (DOBZHANSKY,EHR- the genus, with homologs being transmitted vertically MAN and PAVLOVSKY 1957).Analysis of allelic varia- between species. Under this set of conditions,the tion shows that it is the most genetically differentiated discontinuities in distribution can only be explained of the six sibling species, suggesting that it may have by proposing that elementshave been lost in a number been the earliest to diverge from the cluster (AYALA of species or lineages during diversification or upon et al. 1974). subsequent evolution. There is some evidence to sug- The sultans species group, which is composed of gest that elements can be lost over time. The patchy five subgroups, shows an orderly progression in evo- distribution of predominantly weak P homologues in lution from the more primitive cordata and elliptica the melanogaster species group probably represents a subgroupsthrough the sturtevanti and parasaltans 352 S. B. Daniels et al.

subgroups to the sultans subgroup, which is consid- the ancestor of whichmay have become separated ered to be the most derived. Following the willistoni- from South America on an islandmass destined to sultans divergence in tropical NorthAmerica, the become the Lesser Antilles. This second introduction sultans lineage continued to diversify in this region, most likely would have taken place before the com- giving rise tothe ancestors of themore primitive plete blossoming of sibling clusters and ata time prior cordata and elliptica subgroups. However,at least once to the formation of the Isthmus of Panama about 5 prior to the closing of the water gap by the Isthmus million years ago. The independent introduction into of Panama, it is proposedthat a progenitor from D. melanogaster probably occurred at some later time, North America crossed the water gap and colonized following its divergence from D. simulans. We empha- the South American continent, where further diver- size that the above scenario is hypothetical and does sification occurred to produce the sturtevanti, para- notrule out other plausible interpretations of the sultans and sultans subgroups. data. Based on the Southern blot data and the biogeo- Did D. willistoni serve as the donor speciesin the graphical inferences of THROCKMORTON(1975), we proposed horizontal transfer of P elements to D. have formulated the following hypothetical scenario melanogaster? The results of our survey showed that toaccount for the present day P distributions: P D. melanogaster P elementsmore closely resemble elements initially entered the genus early in its radia- those of thewillistoni species group, and,in particular, tion and were transmitted vertically through one or the willistoni subgroup, than those of the melanogaster more of the early Scaptodrosophila lineages. During and obscura species groups. Within the willistoni thesophophoran radiation, they were passed to at subgroup,four species were initially considered as least the protomelanogaster lineage, which later gave potential donor species. From these, D. willistoni was rise to the melanogaster and obscura groups. Whether singled out for further analysis because: (1) different P elements were transmitted to the ancestors of the isolates exhibit very different restriction patterns, a willistoni-sultans lineage during this period is more possible indication of active transposition; and (2) it is problematic, but the presence of weak P homologues the only species inthe subgroupwhose range overlaps in D. insularis, similar to thoseof the melanogaster and the geographic region where the earliest P element obscura groups, suggests that P elements may have activity was detected in D. melanogaster. enteredat least the willistoni lineage at this time. Subsequently, an intact P element was cloned and Whether they also entered the sultans lineage is more sequenced from a D. willistoni iso-female line. It was difficult to determine, since there is no evidence of found that this element differs from the D. melano- their occurrence in the primitive subgroups as they gaster canonical P by only a single base, an A to G exist today. It is possible that elements once existed transition at position 32. Thisbase substitution creates in these subgroups, but subsequently have drifted to 32-bp terminal inverted repeats instead of the 31-bp extinction, as is probably the case for many of the repeats typical of D. melanogaster elements. Whether melanogaster group species that now completely lack this difference has any influence on the transposability P-hybridizing sequences. During this phase, P se- of the D. willistoni P element is not known. We further quences may also have been transmitted vertically to note that a guaninebase was also found at position 32 a very limited number of lineages that arose during in the two P elements cloned from D. nebulosa (LANS- the Drosophila radiation. The remnants of the initial MAN et al. 1987) andin a defective P element, referred invasion are perhaps manifested today by the very to as KP’, from D. melanogaster (ITOH and HORI 1988). weak P homologs seen in the Scaptodrosophila species, This element is identical to the KP element described in the melanogaster and obscura groups,and in D. by BLACKet al. (1987), with the exceptionof the single insularis. base substitution. This raises the possibility that there At some point following both the establishment of may be a subset of intact P elements within D. mela- the sultans lineage on the South American continent nogaster that are completely identical to the cloned D. and the physical isolation of the D. insularis ancestor, willistoni P element. a second invasion occurred, with P elements being Northern blot analysis of poly(A+) RNA from the introduced, or perhaps reintroduced, independently D. willistoni line from which the cloned P element was intothe extant willistoni and sultans lineages then derived revealed a P transcript of the size expected present on the South American continent. This intro- from an autonomous element,suggesting the presence duction could have been mediated by an infectious of at least one potentially active element. The func- process thatoccurred exclusively onthe continent tionality of this element was subsequently confirmed during this time. Such an event would explain the by the demonstrations of its ability to promote ry+ conspicuous absence of P homologs in the more prim- transformation when co-injected with a y+transposon itive sultans subgroups, which at the time were geo- and its ability to induce snw mutability. graphically isolated in tropical North America, and The similarity of the D. willistoni element to the D. the absence of prominent P sequences in D. insularis, melanogaster canonical P, together with the proof of Interspecific TransferInterspecific of P Elements 353 its functionality,strongly suggests that D.willistoni sumes that initial P element entry was by horizontal may have served as the donor species in the interspe- transmission. cific transfer event, although other species from the It is importantto note that the recent-invasion willistoni subgroup, namely D. equinoxialis, D. paulis- hypothesis does not directly address the issue of time torum, D. pavlovskiana and even D. tropicalis, of which of P element introduction into the D.melanogaster only a single isolate was tested, cannot be completely genome; it only addresses the time of P element spread ruled out at this time. All have broad distributions through natural populations. KIDWELL(1 986) has cau- within Central and SouthAmerica, with the exception tioned that “recent invasion does notnecessarily imply of D. pavlovskiana, which is a rare, narrow endemic recent origin.” P elements may have been acquired at (SPASSKYet al. 197 1). a time much prior to the global invasion, persisting at A comparison of P element dynamics in D. mela- low frequencies in isolated subpopulations until very nogaster and D. willistoni: Whether P elements be- recently. The rapid spread of P elements within the have in D. willistoni as they do in D. melanogaster is yet last half century could have been triggered by man’s to be determined. There is accumulating evidence to recent global activities, which have produced drastic suggest thatcertain host propertiesinfluence the changes in D.melanogaster demography (KIDWELL expression of P elements (DANIELSet al. 1987a; 1986; ENCELS1989). Supportfor such ascenario O’BROCHTAand HANDLER1988; RIO et al. 1988; RIO comes from simulation studies (UYENOYAMA1985) and RUBIN 1988). It is not yet known whether D. that show that a transposable element, such as P, can willistoni possesses the host functions that permit the exist within a population for a long period of time at phenotypic manifestations of P-M hybrid dysgenesis very low frequencies, thereby avoiding detection by in D.melanogaster. If P elements are of relatively limited sampling. However,once the element has recent origin in D. melanogaster, then this species may reacheda certain critical frequency, it can quickly not have had sufficient time to adapt to the effects of sweep to fixation.Under this setof conditions, P this element system. It seems possible, then, that the elements could have entered the D. melanogaster ge- expression of hybrid dysgenesis in D. melanogastermay nome as long ago as the time immediately following show interestingdifferences when comparedto D. the divergence from D. simulans, which is estimated willistoni, which apparently has harbored active P ele- to have occurred anywhere from 1 to 7 million years ments for a longer periodof time. Such a comparison ago, depending on the method used to measure the may shed light on the question of whether hybrid rate of evolutionary separation(SCAVARDA and HARTL dysgenesis is an early, transient phase of a P element 1984). Thereare at leasttwo pieces of evidence, invasion. however, that would suggest that P elements may have Certain differences have been noted between the entered D. melanogaster more recently than this. First, two species. We have examined 12 lines ofD. willistoni, all D. melanogasterP elements thus far examinedshow representing older laboratory strains and recent iso- almost complete sequence homogeneity, even when lates from thewild. P element copy number was found elements from geographically diverse strains are com- to be consistently lower in these lines than in most P- pared (e.g., O’HAREand RUBIN 1983; SAKOYAMAet bearingstrains of D.melanogaster. Additional lines al. 1985; BLACKet al. 1987). Second, the D. willistoni collected from different locales have been screened element is nearly identical to the D. melanogaster ca- by L. STRAUSBAUGHand J. POWELL(personal com- nonical P (this report). This observation is especially munication) with the same results. In the D. willistoni relevant if D.willistoni served as the donor in the line from which the cloned element was derived, low postulatedhorizontal transfer event. If P elements copy number was accompanied by far fewer P element have existed in D. melanogaster for a very long time transcripts than are found in strong P strains of D. (i.e., millions of years), then we would expect more melanogaster. Limiting the number of elements within sequence divergence among its elements, as wellas the genome could constitute one way of minimizing between them and the D. willistoni P. In fact, this is the deleterious phenotypiceffects of hybrid dysgenesis what was observed in a study of two P elements from at the level of the organism. D. nebulosa (LANSMANet al. 1987), a willistoni group The recent-invasion hypothesis: The recent-inva- species that has presumably harbored elements for a sion hypothesis, sometimes referred to as the rapid- relatively long period of time. These elements had invasion hypothesis, proposes that P transposable ele- diverged 10% from each other and 4-6% from the ments have spread rapidly through global populations D. melanogaster canonical P. of D. melanogaster within the last half century (KID- Since we know almost nothing about the rates of WELL 1979, 1983). This model was inspired by tern- sequence divergence for multicopy, mobile DNAs in poral and geographic surveys that showed that P Drosophila, it is impossible to pinpoint precisely the strains were absent in collections made before 1950, time of P element entry into D. melanogaster. All that but increasingly common in those madeafter that may be said currently is that P elements have not been time (KIDWELL 1983). The hypothesis generally as- present in this species long enoughto accumulate - 354 S. B. Daniels et al. appreciable sequence divergence. fiased upon theo- ANXOLAR~HERE,D., D. NOUAUD and G. PERIQUE?., 1987 retical considerations of the rate at which new trans- Evolutionary genetics of the P transposable elements in Ilro- sophila melanogaster and in the Drosophilidae family. Life Sci. posable elements are likely to originate within the D. Adv. Ser. C 1: 37-46. melanogaster genome, ENCELS(1986, 1989) has ar- AVIV,H., and P. LEDER,1972 Purification of biologically active gued that it is improbable that P elements have arisen globin messenger RNA by chrotnatography on oligothymidylic in D. melanogaster within the last half century. These acid-cellulose. Proc. Natl. Acad. Sci. USA 69: 1408-1412. arguments are not in conflict with the present data. AYALA,F. J., M. L. TRACEY, D. HEDGECOCKand R. C. RICHMOND, 1974 Genetic differentiation during the speciation process in Whether P elements arosein D. melanogasterjust prior Ihosophila. Evolution 28: 576-592. to their worldwide spread during the last few decades BEVCRLEY,S. IM., and A.C. WILSON, 1984 Molecularevolution or whether they were extant in the organism for some in Drosophila and the higher Diptera. 11. A time scale for fly period of timebefore the global invasion is still a evolution. J. Mol. Evol. 21: 1-13. matter of conjecture. BINGHAM,P. M., 5.1. G. KIDWELLand G. M. RURIN, 1982 ?'he tnoleculat- ba5is of P-M hybrid dysgenesis: the role of the P The role of horizontal transmission in eukaryotic element. 21 1'-stain-specific tr;~nsposon family. Cell 29: 99.5- evolution: Cross-species gene transfer is well known 1004. among prokaryotes, especially in environments where BLACK, D.M.. M. S. JACKSON,M. G.KIDWELL, and G. A. DOVER, selective pressures have been intensified (e.g., by an- 1987 K1' elementsrepress P-inducedhybrid dysgenesis in Drosophtla melanogaster. EMBO J. 6: 4125-4135. tibiotics). However, the extent to which such events BROOKFIk:l.D, J. F. Y., E. MONTGOMERYand C. H.LANGLEY, occur in higher organisms remains unknown. Thus I984 Apptrent absence of tlansposable elen~entsrelated to far, only a very few solid candidates for interspecific the P elernents ofD. melanogaster in other species ofllrosophila. gene transfer involving eukaryotic species have been Naturr 310: 330-392. described (LEWIN 1982, 1985), although there is in- DANIELS,S. B., md 1.. D. STRAUSRAUGH,1986 The distribution of 1'-element sequences in Drosophila: the willistoni and sultans creasing speculation that this form of inheritance may species groupr.J. Mol. Evol. 23: 138-148. play a more significant role in eukaryotic evolution DANIELS,S. B., L. D. STRAUSRAUGH,L. EHRMANand R. ARM- than previously thought (e.g., SYVANEN1986). Of the WRONG, 1984 Sequences homologous to P elements occur in possible mechanisms to shuttle DNA between species, Drosophilapaulzslorum. Proc. Natl. Acad. Sci USA 81: 6794- viruses are perhapsconsidered the most promising 6797. DANIELS,S. B., hl. TLICCARRON,C. LOVE and A. CHOVNICK, candidates. It is known that certain insect viruses are 198.5 Dysgenesis-induced instability of rosy locus transfor- able to incorporate foreign DNA into their genomes mation in Drosophila melanogasfer: analysis of excision events without affecting their ability to replicate or form :urd the selective recovery of control element deletions.Ge- infectious particles (FRASER1986). A significant pro- netics 109: 9.5- I 17. portion of these DNA insertion events involve host DANIELS,S. B., I. A. BOVSSY,A. TUKEY,M. CARRILLO;tnd M. G. KIIIWFI.I~,1987a Variability among"true M" lines in P-M transposable elements (FRASER 1986; MILLER and gonadal dysgenesis potential. Drosophila Inform. Serv.66: 37- MILLER1982). If the transferof transposable elements :39. between species should prove to be other than a rare I~NIELS,S. B.. S. H. CLARK,%I. G. KIDWELI.and .4.C:HOVNICK, event,then the case forthe involvement of these 1987b Genetic transfortnation of Drosophilamelanogaster genetic entities in eukaryotic evolution becomes even with an ~~utonon~ousP element: phenotypic and molecular more compelling. analyses of long-established transformed lines. Genetics 115: 71 1-723. We ;IIY nruch indehted to IAN BOUSSYand KIYOSHI KIMURAfor DAVIS, R. W., 1). BOISI'EIN and J. R. ROTH. 1980 Advanced n1any valuable discussions and for their comments on the penuiti- Bacterial Genetics:A Manualfor Genetic Engineering. Cold Spring mate draft of this manuscript. We also thank RITA VADNAISfor her Harbor Laboratory, Cold Spring Harbor, N.Y. diligent cvork in the fly facility and ATANU DUTTA ROYfor assisting DENTE,L. G,, G.CFSARENI and K. CORTESE,1983 pEMB1.: a new with the in situ hybridization analysis. ANDREW REAUMEpatiently fanlily of single-stranded plasmids. Nucleic Acids Res. 11: served as ;In unending resource, providing invaluable help at many 1645-1655. s~cgcsof this study. ?'his work was supported by U.S. Public Health DORZHANSKY,T., L. EHRMAN and 0. PAVLOVSKY, 1957 Service grants GM-37241 and GM-09886 to A.C, GM-28680 to Drosophila insularis,a new sibling species ofthe willistoni group. I..I).S. ;tnd GM-36715 to M.G.K. Univ. Texas Publ. 5721: 39-47. KNGELS, W. R., 1979 Extrachromosomal control of mutability in LITERATURE CITED Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 79: 401 1- 4015. ANXOI.AR~H~RE,D.. and G. P~RIQUET,1987 P-homologous se- ENGELS,W. K., I983 The P farnily of transposable elements in quences in Diptera arenot restricted to the Drosophilidae Drosophila. Annu. Rev. Genet. 17: 315-344. fittnily. Genet. Iber. 39: 21 1-222. ENGEIS,UT. K., 1986 Onthe evolution and populationgenetics ANXOLAR~H~RE,11,. M. G. KIDWELLand G. P~RIQUET,1988 of hybrid-dysgenesis-causing transposable elements in Drosoph- Molecular characteristics of diverse populations are consistent ila. t'llilos. Trans. K. Soc. Lond. Ser. B 312: 205-215. with the hypothesis of a recent invasion of Drosophila melano- ENGELS.W. R., 1989 P elements in Drosophilamelanogaster, pp. gasfer by mobile P elements. Mol. Bid. Evol. 5: 252-269. 437-484 in Mobile DNA, edited by D. BERGand M. HOWL ANXOt.ARkHk.RE, D., 1). NOVAUD atld G. PkRIQUET, 1985 An~ericanSociety for Microbiology, Washington, D.C. S&quenceshon~ologues i I'blbnlent P chez des espkces de Dro- ENGELS,W. R.,<:. R. PRESTON,P. THOMPSONand W. B. EGGLES- sophila du groupe obscura et chez Scaptomyza pallida (Droso- TON, 1986 In situ hybridization to Drosophila salivary chro- philidae). (;met. Sel. Evol. 17: 579-584. n~oso~neswith biotinylated probes and alkaline phosphatase. Interspecific Transfer of P Elements 355

FOCUS8: 6-8. in mammalian cells and yeast. J. Mol. Biol. 200: 4 11-4 15. FINNEGAN,D. J.. 1985 Transposable elements in eukaryotes. Int. RIO, D. C.. F. A. LASKI andG. M. RUBIN, I986 Identification and Rev. Cytol. 93: 281-326. it~un~ut~ochemicalanalysis of biologically active Drobophila P FRASER,M. J., 1985 Transposon-mediated mutagenesisof bacu- element transposase. Cell 44: 21-32. loviruses: transposon shuttling and implications for speciation. Rro, D. C., and G. M.RUBIN, 1988 Identification and purification Annu. Entomol. Soc. Am. 79: 773-783. of a Drosophila protein that binds to the terminal 31-base-pair ITOH, M., and S. H.HORI, 1988 Structure of the insertion se- invertedrepeats of the P transposableelement. Proc.Natl. quence involved in positive regulation of D. melanogaster G6PD Acad. Sci. USA 85: 8929-8933. gene. Jpn.J. Genet. 63: 563. ROBERTSON,H. M.,C. R. PRESTON,R. W. PHILLIS,D. M. JOHNSON- KAPLAN,N.. T. DARDEN and C. H. LANGLEY, 1985Evolution and SCHL~IL,W. K. BENZ and W. K. ENGELS,1988 A stable extinction of transposable elements in Mendelian populations. genomic source of P element transposase in Drosophila mela- Genetics 109: 4.59-480. nagaster. Genetics 118: 461-470. KIDWELL, M. G.,1979 Hybrid dysgenesis in Drosophilamelano- ROIHA,H.. G. hl. KURIN and K. O'HARE, 1988 P element inser- gaster: the relationship between the P-M and I-R interaction tions andre;lrrangements at the singed locus of Drosophila systems. Genet. Res. 33: 203-21 7. melanogaster. Genetics 119: 75-83. KIDWELL, M. G., 1983 Evolutionof hybrid dysgenesis determi- RUBIN,G. M., and A. C. SPRADLING,I983 Vectors for P element- nants in Drosophilamelanogaster. Proc. Natl.Acad. Sci. USA mediated gene transfer in Drosophila. Nucleic Acids Res. 11: 80: 16.5.5-16.59. 6341-635 1. KIDWELL, M. G., 1986 Molecular and phenotypic aspectsof the RUSHLOW,C. A., W. BENDER andA. CHOVNICK, 1984 Studieson evolution of hybrid dysgenesis systems, pp. 169-198 in Evolu- the mechanism of heterochromatic position effect at the rosy tionary Processes and Theory, edited by S. KARLIN and E. NEVO. locus of Drosophzla melanogaster. Genetics 108: 603-6 15. Academic Press, New Yorl. SAKOYAMA,Y., T. TODO,S. ISHIWA-CHICUSA,T. HONJO and S. KIDWELL,M. G., J. F. KIDWELLand J. A. SVED, 1977Hybrid KONDO, 1985 Structures of defective P transposable elements dvsgenesis in Drosophila melanogaster: a syndrome of aberrant prevalent in natural Q andQ-derived strains of Drosophila traits including mutation, sterility, and male recombination. melanogaster. Proc. Natl. Acad. Sci. USA 82: 81 19-8123. Genetic\ 36: 8 13-833. SANGER,F., S. NICKLEN andA. R. COULSON, 1977 DNA sequenc- KISSANE,H. M., and E. ROBINS,1958 The fluorometric tneasure- ing with chain-ter.~nin;ttit~~inhibitors. Proc. Natl. Acad. Sci. ment of DNA in animal tissues with special reference to CNS. USA 74: 5463-5467. J. Biol. Chem. 233: 184-188. SCAVARDA,N. J.,and D. L. HARTL,1984 Interspecific DNA KRAWETZ,S. A,, J. C. STATESand G. H. DIXON, 1986 Isolation tl-;lnsformation in Drosophila. Proc. Natl. Acad. Sci. USA 81: and fractionation of total nucleic acids from tissues and cells. 7.515-7519. Cell J. Biochetn. Biophys. Methods 12: 29-36. SEARLES,I.. L., A. L. GREENLEAF,U'. E. KEMP and R. A. VOELKER, LANSMAN, R. A., S. N. STACEY,T. A. GRICLIATTI andH. W. 1986 Sites of P element insertion and structures of P element BROCK,1985 Sequences homologous to the P mobile element deletions in the 5' region of Drosophila melanogaster RpII215. of Drosophilamelanogaster are widely distributed in the Mol. Cell Biol. 6: 33 12-33 19. subgenus Sophophora. Nature 318: 561-563. SPASSKY, B.,R. C. RICHMOND,S. PEREZ-SALAS,0. PAVLOVSKY,C. LANSMAN, R. A,,R. 0. SHADE,T. A. GRICLIATTI and H.W. BROCK, A. MOUR~O,A. S. HUNTER, H.HOENICSBERG, T. DOBZHANSKY 1987 Evolution of P transposableelements: Sequences of and F. J.AYALA, 1971 Geography of the sibling species Drosophila nebulosa P elements. Proc. Natl. Acad. Sci. USA 84: related to Drosophila willistoni, and of the semispecies of the 649 1-6495. Drosophila paulistorum complex. Evolution 25: 129-1 43. LASKI,F. A,. D. C. RIO and G. M. RUBIN, 1986 Tissue specificity SPRADLING,A. C., and G. M. RUBIN 1982 Transposition ofcloned of Drosophila P element transposition is regulated at the level P elernents into Drosophila germ line chromosomes. Science of mRNA splicing. Cell 44: 7-19. 218: 341-347. LEWIN,R., 1982 Can genes jump between eukaryotic species? STACEY,S. N., R. A. LANSMAN, Mr.H. BRocKand T. A. GRIGLIATTI, Science 217: 42-43. 1986 Distribution and conservation of mobile elements in the LEWIN,R.. 1985 Fish tobacterium gene transfer. Science 227: genus Drosophila. Mol. Biol. Evol. 3: 522-534. 1020. SYVANEN,M., 1986 Cross-species gene transfer: a major factor in LINDSLEY,D. L., and E. H.GRELL, 1968 GeneticVariations of evolution? Trends Genet. 2: 63-66. Drosophila melanogaster. Carnegie Inst. Wash. Publ. 627. THROCKMORTON,L. H., 1975 The phylogeny, ecology and ge- MANIATIS,T., E. F. FRITSCH and J. SAMBROOK,1982 Molecuhr ography of Drosophila, pp. 421-469 in Handbook of Genetics Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, (Vol. 3). edited by R. C. KING.Plenum Press, New York. Cold Spring Harbor, N.Y. TSUROTA,S., and P. SCHEDL, 1986 Hybrid dysgenesis-induced MILLER,D. W., and L. K. MILLER, 1982 A virus mutant with an revertants of insertions at the .5' end of the rudimentary gene insertion of a copia-like element. Nature 299: 562-564. in Drosophilamelanogaster: transposon-induced control muta- NITASAKA,E., and T. YAMAZAKI,1988 Excision of one oftwo tions. Genetics 114: 165-182. defective P elements as thecause of alternate mutational events UYENOYAMA,M. K., 1985Quantitative modelsof hybrid dysge- (sn' and sn') of the singedbristle allele sn"' in Drosophila nesis: rapid evolution under transposition, extrachromosomal melanogaster. Jpn. J. Genet. 63: 303-3 12. inheritance and fertility selection, Theor. Pop. Bioi. 27: 176- O'BROCHTA,D. A.,and A. M. HANDLER,1988 Mobility of P 201. elements in drosophilidsand nondrosophilids. Proc. Natl. VOELKER,R. A., A. L. GREENLEAF, H. GYURKOVICS,G. B. WISELY, Acad. Sci. USA 85: 6052-6056. S. HUANGand L. L. SEARLES,1984 Frequent impreciseex- O'HARK.K., and G. M. RUBIN, 1983 Structures of P transposable cision among reversions of a P element-caused lethal mutation elenleuts and their sites of insertion and excision in the Dro- in Drosophila. Genetics 107: 279-294. sophila melanogaster genome. Cell 34: 25-35. RIO, D. C., G. BARNES,F. LASKI,J. RINEand G. M. RUBIN, 1988 Evidence for Drosophila P element transposase activity Communicating editor: C. C. LAURIE