Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
1 1 Number 18 1983 Vofume 11Nme Voum Nucleic Acids Research 818ucecAisRsac Vectors for P element-{nediated gene transfer in Drosophila Gerald M.Rubin + and Allan C.Spradling Department of Embryology, Carnegie Institution of Washington, 115 West University Parkway, Baltimore, MD 21210, USA Received 23 June 1983; Revised and Accepted 30 August 1983 ABSTRACT We have constructed and tested several new vectors for P elementmediated gene transfer. These vectors contain restriction sites for cloning a wide variety of DNA fragments within a small, non-autonomous P element and can be used to efficiently transduce microinjected DNA sequences into the germ line chromosomes of D. melanogaster. The P element in one vector also carries the rosy gene which serves as an easily scored marker to facilitate the transfer of DNA fragments that do not themselves confer a recognizable phenotype. The failure of certain P element constructs to function as vectors suggests that P element sequences, in addition to the 31 bp inverse terminal repeats, are required in cis for transposition. Moreover, removal of the first 38 bp of the autonomous 2.9 kb P element appears to destroy its ability to provide a trans-acting factor(s) required for the transposition of non-autonomous P elements. Finally, we describe a genomic sequence arrangement that apparently arose by the transposition of a 54 kb composite P element from a tetramer plasmid. INTRODUCTION P element-mediated gene transfer permits the introduction of cloned DNA sequences into the germ line chromosomes of a complex metazoan organism, Drosophila melanogaster (1,2). Moreover, the transduced DNA sequences appear to be stably inherited and appropriately expressed in future generations (3,4,5). Thus, this genetic transformation method should allow the assessment of the functional properties of aenes that have been isolated and subjected to in vitro mutagenesis. The P element-mediated gene transfer method exploits the properties of the P family of transposable elements (6,7,8). A 2.9 kb P element has been isolated from the genome of a P strain and its DNA sequence determined (8). This transposable element appears to be autonomous in that it is able to transpose from plasmid sequences into the chromosomes of germ line cells after microinjection into Drosophila embryos lacking such elements (M embryos) (1). Smaller P elements are also present in the genome of P strains © I RL Press Limited, Oxford, England. 6341 Nucleic Acids Research (6,8) and the DNA sequences of these elements suggest that they arose by internal deletion from the 2.9 kb element (8). These smaller elements are non-autonomous; they cannot transpose when injected into M embryos (2). When such non-autonomous elements are coinjected with the 2.9 kb element however, they are able to transpose suggesting that the 2.9 kb element encodes a tras-acting factor required for transposition (2). Other DNA sequences can be caused to transpose into germ line chromosomes as well, by constructing a P transposon carrying the DNA segment of interest within a non-autonomous P element (2-5). We describe here several new non-autonomous P element vectors. These vectors are small plasmids which have many unique restriction sites for cloning DNA fragments within a P element but lack sites in the plasmid backbone outside the element. The results of testing several intermediates made in the course of constructing these plasmids reveal certain sequence requirements for P element transposition. Our results also suggest that P elements as large as 54 kb can transpose. MATERIALS AND METHODS Restriction enzymes and BAL-31 nuclease were purchased from New England Biolabs. T4 DNA ligase as well as Hpa I and Kpn I DNA linkers were purchased from Collaborative Research. M13mplO DNA was purchased from P-L Biochemicals. Microinjection of Drosophila embryos and other aspects of the transformation procedure are described in detail in references 1 and 2. DNA blotting and hybridization were as previously described (9). BAL-31 digestions were carried out at room temperature in 600 mM NaCl, 12 mM CaC12, 12 mM MgC92, 1 mM EDTA and 20 mM Tris-HCl pH 8 for 5 or 10 min using 6 units of BAL-31 and 20 ag of plasmid DNA. Growth of recombinant M13 phage, preparation of DNA templates and DNA sequencing using di-deoxynucleotides were as described by Sanger et al. (10) except that the sequence reactions were not in glass capillaries, but in 1.5 ml tubes as described by Heidecker et al. (11). RESULTS AND DISCUSSION Vector Construction Plasmid vectors were constructed according to the protocol diagrammed in Figure 1. First, the Sal I fragment of p6.1 (2), which contains a nonautonomous P element inserted into white locus DNA (6), was ligated into the 6342 Nucleic Acids Research HindMi Sol P EcoRI Smol 5.3 kb u dM HindMU Xhol Po.ub Digest with BamHl Digest with exonuclease Bal 31 Ligate with T4 DNA ligase Tronsform E. co// Figure 1. Construction of vectors. The indicated steps were performed. DNA sequences derived from the pUC8 plasmid are shown stippled, those from the P element are fully shaded and those from the D. melcmogaster white locus are unshaded. HindUm Ps Il ((pCIW2~ \\i4.6 kb, indm HFlnd Digest with Digest with Ligate with Transform Sall exonuclease Bal 31 T4 DNA igose E co/i 3.9kbg HOndM Hindm Insert EcoRI linker into Pvu0 site to create pCIW4 Insert polylinker into pCIW4 between EcoRl and left HindIm site Caonegie ind M Hind-1 Pstl Sol1 BamHl*Smal EcoRI Carnegie 2 HindM Pstl Sal -BamHl Hpa EcoRI Carnegie 3 HindlMPstl-Sall-BomHl.Kpnl EcoRI' Carnegie 4 'HindE-Pstl Soll-Xbol-BomHt.Smol-Sstl-EcoRl Sal I site of the plasmid vector pUC8 (12) to create the plasmid pCIW1. Then two successive treatments with BAL-31 nuclease were used to eliminate undesirable restriction sites and reduce the size of the plasmid. Finally, the indicated linkers were inserted to create the plasmid vector Carnegie 1. The Carnegie 1 vector is 3616 bp long and its nucleotide sequence is shown in Figure 2. This vector carries the Hind III - Pst I - Sal I - Bam HI - Sma I - Eco RI polylinker derived from pUC8. The Carnegie 2 and Carnegie 3 vectors were constructed by inserting a Hpa I (GTTMC) or Kpn I (CGGTACCG) linker into the Sma I site of Carnegie 1, respectively. The Carnegie 4 vector was made by substituting the polylinker of M13mplO for the Hind IIIEco RI polylinker contained in Carnegie 1. The vector Carnegie 20, which is diagrammned in Figure 3, was made by inserting a 7.2 kb Hind III fragment containing the rosy gene (2) into the Hind III site of the Carnegie 2 polylinker. This rosy DNA fragment 6343 Nucleic Acids Research HindM 120 90 110 70 100 s0 00 s0 30 20 TAACA?AAGG ?GYCGCGAThOCCAA OCT?vACCGAA GUTAACAC? hhATTAG?G CAC0W"GCY 9T000-AOGW OAAMG7TCY?GTCGGACGAA ?7YP? 240 230 220 210 190 200 130 170 1S0 160 140 130 0AA~AACTAA rC~C0AATAAAAAA AAA?GAAAYA ?90MAATrY 90CYGAAAO CTO9GAC900 AOYAAAMTPA A?TCAC0T0C CGAAQ00TC TAIT7AAG AA TG?CGGG 3600 340 330 350 310 320 290 300 200 270 200 290 TG&O0c?0G70CAOC CTT?GTGAAA &CTCCCAAT ?TTGTAT0TCC CACTnTAA20. 0?YC AGCA" 0400aOCTA CCT=AAAAG GC0AaCAT TAAIAGG00GG CGACTCAACG 470 400 450 430 400 040 010 020 000 390 300 370 C0AG00CCGT ACC?AM?AAA GYGAThGAO CYGAACCAG.A AAAGAThAAAA G&AAGCTATA CCA0?GGGA0 TACACAAACA GAGTAAGTT GAATAGTAAA AAAAA?CATT TA?GTAAACA 10 MUMA0AA 340 590 330 910 520 S00 9000MGTCCYGTYVA T?GrYAAYG AAAATAAGAG CTTGAGOGAA AAAA?PGYA HindM 000 070 090 640 630 610 TO0CGC9CAOCCAAG CTIr00CGTC TCGCA&ATPAL VYAAAAATAA AAC?T0AAAA A?AA?TTCGY 790 700 770 730 700 740 730 jCA~ACAA-f ATATCOCY COCACTCAGA CTCAATACGA CAC?CAGAT? ACTATTCC?T TCAC?CGCA 910 900 330 390 39 070 600 YGZAr.r:r A?-GrA.&T TZ:G..TG7 ZG':AThTZrG 3CTCTAA7AC rZ AYGPATTvYc AYA?t7 1030 1020 1010 1000 990 930 970 TTTrT'r'rTT TACcArTAIT ACCATCGr;T T-rACGTr!TA TTGCCZCCT1 k9AAAAGCThA rGrkArTATA 490 A?AAC0T 970 500 ClPrGAGWA CGATOA0COT 6000 EcoRi So. AGA7CC 00AGcCCGAA ?MCCG0GGA 720 710 700 090 000 CTAA?FAATA TTAICOA ATYCAAJk(C CAC0G-ACA C?AAGGGYA 340 030 020 300 I10 TTATTOCAAG CAThCGITAA GTGGAYOTCT CrtGCCGACG GGACCACCrY 9600 940 930 920 990 TTrCGGGCCCG AGSCAGGAG TAGC--GA-ZAT ArAr1-:C7AAA T9.Acr:"Z-rG 1060 1070 1000 1040 1090 TT~rGTGCCAA T&AAAAAZ9A7 ATATGACCTA TNSAATA~ZAA~;rA rTTZ.CZ: 1200 1190 1170 1110 1140 1130 1100 1130 1100 1120 1110 1090 TT-GA,ACATC CCCACAAGTA SACT1TGGAT TrGTZTtrCT kCC99AAGAC TTA7cACAcr G'-AThccrTA :ArC1AA%A9A T0GTTTATZG CTACATAAAA ZACCGGGATA rA rtrrrTAT 1310 1326 1290 1270 1300 1200 1240 1200 1230 1230 1220 1210 C rr :ATkATTCAZ CCACCA:A wA7JCGrrTCG TAkGTTGC'rCT rTCGCTG rCT :CCACCZGCT CTCCGCAAZA CA?1CACCTr ?TrGrTTGACG ~:TC ATACATACTr T?CA.AA?CGC 1440 1430 1420 1410 1400 1390 1330 1370 1300 1330 1300 1330 ACCTTrGGAGC GACTGTCGTT AkGT'tCCGCGC GA?TT7GTGC OGYTIYTTCAC ACCGCATA'r G?GCACICT ATACAATCT GCYCYGATC CGCA?A09T AGCCAGCCO GACACCCOCC A 39bp 19600 1940 1920 1930 1910 1300 A728bp 14940 1470 1490 1460 YCCOOGAOCT 0CA%GTCYGA GhaG?rrCA CCIACAC CGAAACGCGC o&ocotGACCOCCCCY GACOGOCTYG TCTGCYCCCG OCAYCCOCY ACAGACAAGC T&ccGACG 1990 1000 1010 1970 1300 1930 1940 1990 1900 1970 2170 2130 2190 2200 2210 1090 1000 1070 1600 1630 1040 1900 1990 2000 2010 2020 2030 2040 2220 2230 2240 2290 2200 2270 2230 1020 G0AGaGo CCwGw rACGccYATYr m"ATAGGT AATGTAAGA YAATAA0GOT TVCTPAGMM TCA0TOCA CTTlTCGGOG AA?CG=GO GGAACCCCTA TrrG?rrAFT A 2bp 1730 1700 5[tort P-lactamnose gene 1740 1720 1730 18000 1710 1700 rTC-TAA&TA C&AIAAATrA 90TAICGC?C CATGOAC-AA TAACCCYGAT MAAACTMA ATAAT&ATA AAAAGGAAGA OTAYGAGTAT 1AACAWAV CGTGTCG=C ?FA~TFCC 1920 1910 1900 1300 1090 1090 1370 1000 1310 1040 1330 1320 ?WYOGCOOCA ?PrCC-M CTGOT?TfMCACCCAGAA ACGCTOGYGA AAGTAAAAGA TGCYGOAGAT CAGTT0G¶ CACCAGT0G TTA?CrA CGGAFCICA ACAGCOGYAA GAcTvCGwA harYTYcGcc ccGAAGAAcG. TrecAATG ATG.AGCAc?r TYAAAG?rT OCTAYGYGoC OCOGYTTAT ccTGTYA CGCCOGocAA GAGAACTM GTCOCcOAT 2140 2190 21600 2130 2110 2120 2100 2090 2000 2070 2090 2000 ACACTATYCYT cAcAAYGACT YGTPG?cAGTrA CYcAccwG ACAGAAAAGC ATCTTACGGA 9OGCAYGAA GTOAGAGAAT ravwCAGvC 0cccArAAC ATGOAaGYGA AcAcTc00C CAACrrACTrr C90ACAACGA YCGGAGG1M GAAGGAGC?CA ACCGCYOTr"T OCACAACAT oooooatva oTAAToCC ?rGATCOr OGGoACCG CT'AAC-M CCATXCAAA 24000 2390 2380 2360 2370 2390 C0AhAC0TQ GACACCACGA TOCCTGTAGC AAYOGCAACA ACGTYGCGCA AAC?AYTrAC TO0CGAACTA CTrACrCTAG CTTCOCG0CA ACAATrrAAYA GACrgGAY A00MGA?AA 2920 2900 2910 2490 2470 2430 2400 2490 2440 2430 2420 2410 GOaIOCOOGA ~cACrtcvCY acrCGcCY= Ycc0GcT0 TooYrYrArrG Cw?GAAArYc !oGGaGcG GA0cOa0T C1CGGaYwa cAivcAGA CYGGOCCAG AYGYAAGcC 2290 2300 OAOT!AY C-2933 GTAYC 2940 2310 2320 2330 2340 end a-IoCtomoSe gene] 2600 2603 2990 2930 2900 2370 2990 WACGhCOAGO GAGYCAGOCA ACTAYGGAYG AACG.AAAAG ACAGAicYc GawATAGOYG cCYCcAcmrA YAACAYFOS YAACYGYCAG AcAAGT"A 2720 2720 2300 2320 2330 2300 2010 2790 2700 2773 gWCAGh0= G?AGAAAAGA ICAAAOOAYc TYCrTArGAY CCTrrYYYYC YOGcGGAAY CTGCTGCTYG CAAACAAAAA 2990 2960 2940 2930 2920 2910 2000 2390 OCWA cY CrWrcCGA AGOtOACYG CTIcAGcAGA GcGCArGAAc cAhATYAci* cc?TmCaraTG AOCCGww 3330 3000 3370 3090 3040 3030 3013 3020 CLICGCiCvc CTMICCIt TACCOTOW YGcYGCCAOn OOGAAAGY COYGYCTYAC COOGYFOGAC YCAAGACGAY 3200 3130 3190 3170 3130 3100 3140 3133 fCACA CAGCCCAGCY GOGAQGAGAC CACCTACACC CGOALCGAGAY ACCY.ACAOCO TGAGCA?OG GAAAGCOCCA 2390 2090 2000 2070 2030 2090 2700 2710 2700 2790 2700 CYCrAvATAT CYTwAGATY AtTrAhAAcYr YCArrrrAa YTrAA&AAGA YcTAaGOYAA GAYCCYTYTr GAYAAYTCWA YGACCAAAAYr CCCYTrAACOY GAa rfGTY TCACYGAGC AACCA~CCC 2970 YhAOCCACC 3090 23300 2373 23060 GYGAYCAAGA ACCAGGYGY G?TTrr 2930 2990 3000 3100 3110 3120 c?CAAGAhC YCYGYAGCAC cOCCTAATA GAAC00GOG 3230 32400 3223 3210 COCTrTCCCGA AGOGAGAAAG OCOGACAGGY ATCGYAAG AGYYACCG YAOAOCOC.AGC00C 3300 3393 3300 3333 3310 3320 3300 3230 3290 3270 3290 3260 CG0CAG0GT WMACA0GAG AGCGCACGAG. 0GAGCTICCA 000GGAAA CCYGOYAY?CY YTAYAGICC @YCO0GM OCCACCYCYG ACTcGAaGGOY =&OA Y GIaYcYCm 34000 ORI CZ~7 3390 3470 3400 3490 3030 3000 3010 3020 34403 3370 AGGOOGOCOG AGOTAYGG A AAACC9Tr CT?Crt'GAAC TCGGGCrCGG TOGCCAGYATA CCTCIAAATGG rTGTCGTACC ?CTCATGGTT CCGTTACGCC AACGAkGGG?IC TGCTGAkTTAA 3993 36000 3090 3903 A695bp 3920 3930 3S40 3993 3900 CCAATGGGCG GACG?GGAGC CGGGCGAAAT TAGCTGCACA TCGTCG%ACA CCACG?GCCC CAGTTCGGGC AAGGTCATCC 3S70 3900 rGGAGAC-GCT TAACT]1CTCC GC-CGCCGATC TGCCGCTGGA 3610 CTACGTGGGT CTGGCC Figure 2. Predicted DNA sequence of Carnegie 1. The sequence is numbered starting with the first nucleotide of the P element and proceeding clockwise around the map as shown in Figure 1. P element DNA sequences (nucleotides 1-856) are shown both boldface and underlined. These are the sequences that are transferred to Drosophila chromosomes during transformation. Sequences derived from the Drosophila white locus (nucleotides 857-1358 and 3389-3616) are shown in normal typeface and those derived from pUC8 (1359-3388) are shown in boldface. The 31 bp inverse terminal repeats of the P element are indicated by the arrows. The positions of cleavage sites for Hind III and Eco RI, the start and end of the ~-Iactamase gene, and the plasmid origin of 6344 Nucleic Acids Research replication are shown. The positions of the two BAL-31 deletions (728 and 695 bp) made during the construction of Carnegie 1 (see Figure 1) as well as two smaller deletions of 39 bp and 12 bp are also indicated. The 12 bp deletion apparently resulted from the Si nuclease treatment used by Vieira and Messing (12) during the construction of pUC8 to remove an Eco RI site. The origin of the 39 bp deletion is unclear; these nucleotides were deleted from Carnegie 1 as compared to the corresponding region of pBR322 (13). The Carnegie 1 DNA sequence shown was assembled as follows: The DNA sequence of pCIW1 was predicted from the known sequences of pUC8 (12) and of the Sal I fragment of p6.1 (8; K. O'Hare and G. Rubin, unpublished results). The extent of the BAL-31 deletions made during construction of Carnegie 1 were determined by direct DNA sequence analysis of the Taq I - Sau 3A fragment (nucleotides 1315 - 1871) and the Taq I-Taq I fragment (nucleotides 3340-3527) which span the deletion endpoints. The sequence of Carnegie 1 was then derived by modifying this sequence to account for the replacement of the Eco RI-Hind III fragment of the P element in pCIW4 by polylinker fragments. does not contain cleavage sites for Hpa I or Sal I thereby permitting the convenient cloning of additional Sal I, Xho I or blunt-end fragments into this vector. In such constructs Carnegie 20 has been used successfully to transfer DNA fragments that do not themselves confer a recognizable phenotype by selecting for rosy+ transformants (see below). All of the vectors grow well in standard bacterial hosts and confer ampicillin resistance. A number of useful restriction sites found in Carnegie 1 are listed in Table 1. Tests of vectors and other constructs Figure 3 summarizes the results of testing various vector constructs by the Drosophila transformation assay. For these assays (Figure 3, experiments 1-8), a 7.2 kb Hind III or 8.1 kb Sal I fragment containing the rosy gene was inserted into the P element of each vector construct. The resultant plasmids were microinjected into rosy- Drosophila embryos. DNA of a plasmid carrying the 2.9 kb P element (either pff25.1 or pnr25.7) was coinjected to provide tras-acting functions required for transposition of the non-autonomous rosy transposons (2). Adults developing from injected embryos (GO adults) were mated to rosy- flies. The presence of rosy+ progeny in the next generation indicated that the vector being tested was competent for germ line transformation. Some of the GO adults themselves displayed a rosy+ eye color which we believe to be due to expression of the rosy gene on the injected plasmid DNA (2). While not directly correlated with germ line transformation (see below) this transient expression does provide a convenient indication of the ability of the rosy gene to function in a given construct. We first had to verify that the plasmid created by the BAL-31 treatments, 6345 Nucleic Acids Research TABLE I Restriction Sites in Carnegie 1 Location Enzyme Recognition Sequence 39 588 593 593 598 604 610 618 860 914 1219 1375 2178 2325 3194 3400 Hind III Eco RI Ava I Sma I Bam HI Sal I Pst I Hind III Bal I Apa I BssH II Nde I Pvu I Mst I Hae II Ava I AAGCTT GAATTC CCCGGG CCCGGG GGATCC GTCGAC CTGCAG AAGCTT TGGCCA GGGCCC GCGCGC CATATG CGATCG TGCGCA AGCGCC CTCGGG No cleavage sites for Bcl Bgl Cla Eco Hpa Kpn I II I RV I I Mlu I Nae Nar Nco Pvu Sac I I I II I Sac Sph Stu Xba Xho Xma II I I I I I The location numbers refer to the sequence shown in Figure 2 and correspond to the first nucleotide of the recognition site. All of the enzymes shown cleave Carnegie 1 either 0 or 1 time except for Hind III and Ava I which cleave twice. All of the enzymes shown in this table are commercially available. and from which all the vectors were derived, was still capable of acting as an efficient transformation vector. We inserted the 7.2 kb rosy Hind III fragment into the left Hind III site of plasmid pCIW4 (see Figure 1) to create the plasmids pVlO and pVll. These plasmids differ only in the relative orientations of rosy and P element sequences and both were able to transfer the rosy gene at frequencies in excess of 50% (Figure 3, experiments 1 and 2). We also inserted the same rosy fragment into the right Hind III site of pCIW4 to create the plasmid pV9. We did not observe transfer of the rosy gene into the germ line when this plasmid vector was used, however, suggesting that insertion of foreign DNA sequences at this site in the P 6346 Nucleic Acids Research Experiment Injected (No) DNAs injected HR~Xtl R.R X Fertile adults expression GO r, + progeny 4 (67%) H +pfI25.1 123 6( 5% 4 (67%) i+pn25.1 121 9(7%) 6(67%) 6(67%) 133 5 (4%) 4 (80%) 0 pV1O: HI~~~~X~~~I RH 2 pVII: 3 pV9: 4 5 -..CarnegelI, X , HR X5_H HRF/s4 pVI.2: f ""I)iitl,71 >-.'.vx'._:-.7l 6 Cornegie20: ,,,X,)+R X R ......... . H If R HRSH 7 HRXH . 111.Z I '.1 lineor ros S H X R HRoS 'i1 R fragment: 8 frogment: t 9 pII25.7: II X UR R S BS H XH R S BS HXH k_ _ u pn25.7A1: H 777777777M,")il 156 10(6%) 4(40%/) 5(50%) 171 21 (12%) 7(33°%) 0 +prl 25.7 116 20117%/) 11(55%) [ RHB _ +prl25.7 83 +p[125.7 188 14 (70%) 9 5) 0 0 (14%) 0 0 S +pry 26 +pry 3 121 17 (14%) 6 (35%) 5 (29%) +pry3 208 21 (Y10) 3 (114%) 0 +pry 3 114 6 ( 5) 2 (33%/6) 0 RH _ +_'. +pry) H XH R S RHB k -.rD, m -l plI25.7A2: +pry r +pfl25.7 +pl125.7 H / *.5 linear pryl S HF X,/Sp 10 X +pI125.1 Figure 3. Testing of vector constructs. The structures of the microinjected DNAs are diagrammed. Rosy sequences are shown cross-hatched and P element, white and pUC8 sequences are indicated as in Figure 1. Cleavage sites for Hind III, Eco RI, Xho I, Sal I, and Bam HI are indicated by H, R, X, S, and B, respectively. X/S indicates the product of ligation of Xho I and Sal I ends. All DNAs were microinjected as supercoiled plasmids (except for experiments 7 and 8) at the following concentrations: p7r25.1, pir25.7 and their deletion derivatives, 50 pg/ml; rosy-containing DNAs in experiments 1-8, 300 pg/ml; pryl and pry3 in experiments 9-11, 150 pjg/ml each. The number of embryos microinjected and the number and percentage surviving to become fertile adults are given. Also shown are the number and percent of fertile adults showing GO expression of the rosy gene, and the number and percentage of fertile adults giving transformed rosy+ progeny. element destroyed its ability to transpose (Figure 3, experiment 3). We cannot rule out the formal possibility that it is not the transposition of this rosy transposon which is affected, but rather the expression of the rosy gene it carries. This seems unlikely, however, as the percentage of embryos showing GO expression was similar to that obtained with pVl0 and pVll. We inserted the pUC8 polylinker into pCIW4 in two different positions between the Eco RI site and the Hind III site located to its left to create the Carnegie 1 vector, and between the Eco RI site and the Hind III site to its right to create the plasmid pVl. The Carnegie 1 vector was shown to be capable of promoting efficient germ line transformation when tested with an inserted 8.1 kb rosy Sal I fragment (plasmid Carnegie 1.1, Figure 3, experiment 4). In contrast, no rosy+ transformants were recovered when pVl carrying the rosy gene (plasmid pV1.2) was injected, although normal GO 6347 Nucleic Acids Research expression was observed (Figure 3, experiment 5). In the pV1.2 construct, sequences from position 45 to 588 of the P element have been replaced by the polyl inker. The apparent failure of the rosy transposons of pV1.2 and pV9 to transpose suggests that sequences, in addition to the 31 bp inverse terminal repeats of the P element, are required in cis for transposition. Previous studies of the structures of non-autonomous, but transposition competent, P elements indicate that no more than the first 139 or last 216 nucleotides of the 2.9 kb P element are required in cis for transposition (8). Taken together with those findings, the results presented here suggest that, in addition to the 31 bp inverse terminal repeats, a cis-acting sequence essential for transposition is present in the first 140 bp of the P element. Carnegie 20 was shown to be effective in transferring the rosy gene (Figure 3, experiment 6) and has subsequently been used in our laboratories (unpublished results) to transfer additional DNA fragments inserted into its Hpa I site. In these latter experiments, transformants were identified by their acquisition of a rosy+ phenotype. In this way, DNA segments that do not themselves confer a recognizable phenotype can be routinely transferred into the germ line. Attempted improvements in P element-mediated gene transfer P element insertions into host chromosomes occur at a wide variety of locations (2,3); no evidence for homologous recombination has been obtained. In yeast, the frequency of recombination between introduced plasmids and homologous host sequences can sometimes be substantially increased by using linear rather than circular DNA molecules (14). Consequently we attempted to transfer rosy genes carried on linear DNA fragments. However, our attempts were unsuccessful regardless of whether the rosy gene was within a P element or not (Figure 3, experiments 7 and 8). Moreover, we did not observe any GO expression in these experiments suggesting either that the microinjected DNA was degraded or that linear fragments are poor templates for RNA transcription. Although the relative amounts of autonomous and nonautonomous P elements injected into embryos are designed to favor introduction of the non-autonomous element, occasionally transformed strains are obtained which have acquired an autonomous element as well. This problem could potentially be eliminated by the use of a transposition-defective P element which was still able to provide transposition functions in trans to non-autonomous elements. 6348 Nucleic Acids Research .. .. .s. ... ........ .... . ................ .... .... ...... Figure 4, Analysis of the structure of transferred rosy genes in line R310.1. The line R310.1 was generated by transformation of the ry42 host strain with pryl, a rosy transposon plasmid constructed by inserting the 8.1 kb rosy Sal I fragment into the Xho I site of p6.1 (2). DNA was isolated from the host strain ry42, which carries an apparent point mutation in the rosy gene (W. Bender and A. Chovnick, personal communication), and from two individual sublines of line R310.1. Sal I and Hiind III digests of these DNAs were electrophoresed on 0.5% agarose gels and transferred to nitro- cellulose filters to make thiree sets of identical filters. Each set of filters was hybridized with a different 32p-labeled plasmid probe as indicated: pDm2844S8.5 contains th .1k roy SlIfamn lndi pBR322. pm12.8 is a pBR322 clone of the 1.5 kb white locus Sal I fragment into which the P element is inserted in p6.1. The third probe was pBR322 alone. Autoradiograms of the resultant hybridizations are shown. The sizes of labelled bands are indicated in kb. A diagram of the genomic arrangement of pryl sequences deduced from the data in the autoradiograms is also shown. P element sequences are shown fully shaded, pBR322 sequences stippled, rosy sequences cross-hatched and white sequences unshaded. However, we were unsuccessful in our initial attempts to construct such a mutated 2.9 kb P element. Removal of either the first 39 (pTr25.7A1) or last 492 bp (pir25.7A&2) of the 2.9 kb P element carried on the plasmid pw25.7, destroyed its ability to provide the required trans-acting functions (Figure 3, experiments 9-11). Evidence for transposition of a 54 kb P element For many purposes it would be desirable to introduce large segments of DNA into Drosophila chromosomes. However, we have observed that 15 kb 6349 Nucleic Acids Research transposons containing the rosy gene are transferred at only about onethird the frequency of 9 kb rosy transposons, suggesting that transformation frequency may be highly size-dependent. We have carried out a series of experiments, reported elsewhere (3), to assay the level and tissue specificity of rosy gene expression in transformant lines. In the course of those experiments, we recovered one transformant line which produced approximately four times the expected level of the product of the rosy gene, xanthine dehydrogenase. DNA blotting experiments indicate that this line, R310.1, contains a tandem array of rosy transposons (Figure 4). The relative intensities of the hybridizations to various bands indicate that there are four copies of the ryl transposon in this tandem array. This can be most clearly seen in the hybridization of pDM2844S8.5 to Sal I digests (Figure 4, leftmost panel). The 8.1 kb fragment derives from the two mutant genomic copies of the rosy gene (one on each homologue) present in the ry42 host strain. R310.1 DNA was prepared from flies heterozygous for the chromosome carrying the transferred ryl transposon DNA. Thus, the equivalent intensities of hybridization to the 8.1 and 10.8 kb fragments indicate that the 10.8 kb fragment is present twice in the tandem array. The 18 and 22 kb Sal I fragments both contain a single rosy transposon as well as genomic sequences adjacent to the tandem array (see diagram in Figure 4). The structure of this array suggests that it arose by the transposition of a large composite P element from a tetramer plasmid. The tetramer plasmid could have been generated by homologous recombination of monomer plasmids either in E. coli, during growth of the plasmid-containing strain, or in the Drosophila embryos following microinjection. Such recombination between plasmids to produce multimers is known to occur both in recA- E.coli (15) and in mammalian cells (16). Alternatively, such a tandem array could be generated by a normal P transposon integration followed by successive events of homologous recombination between the integrated chromosomal sequences and free circular plasmid DNA as has been observed in yeast (17). Given the failure to observe homologous recombination between injected DNA sequences and their chromosomal homologues in Drosophila (2), we feel that this latter mechanism is highly unlikely in this case. Thus, although the frequency of transfer may be greatly reduced, these results suggest that P elements as large as 54 kb can transpose from plasmids into chromosomal sequences. 6350 Nucleic Acids Research ACKNOWLEDGEMENTS This work was supported by grant from the National Institutes of Health to G.M.R. and A.C.S. We are grateful to Kevin O'Hare and Christine Murphy for invaluable help with the DNA sequence determination of the Carnegie I vector. +Present address: Department of Biochemistry, University of California, Berkeley, CA 94720, USA REFERENCES 1. Spradling, A.C. and Rubin, G.M. (1982) Science 218, 341-347. 2. Rubin, G.M. and Spradling, A.C. (1982) Science 218, 348-353. 3. Spradling, A.C. and Rubin, G.M. (1983) Cell, in press. 4. Scholnick, S.B., Morgan, G.A. and Hirsh, J. (1983) Cell, in press. 5. Goldberg, D.A., Posakony, J.M. and M1aniatis, T. (1983) Cell, in press. 6. Rubin, G.M., Kidwell, M.G. and Bingham, P.M. (1982) Cell 29, 987-994. 7. Bingham, P.M., Kidwell, M.G. and Rubin, G.M. (1982) Cell 29, 995-1004. 8. O'Hare, K. and Rubin, G.14. (1983) Cell, in press. 9. Levis, R., Bingham, P.M. and Rubin, G.M. (1982) Proc. Natl. Acad. Sci. USA 79, 564-568. 10. Sanger, F., Coulson, A.R., Barrell, B.G., Smith, A.J.H. and Roe, B.A. (1980) J. Mol. Biol. 143, 161-178. 11. Heidecker, G., Messing, J. and Gronenborn, B. (1980) Gene 10, 69-73. 12. Vieira, J. and Miessing, J. (1982) Gene 19, 259-268. 13. Sutcliffe, J.G. (1978) Cold Spring Harb. Symp. Quant. Biol. 43, 77-90. 14. Orr-Weaver, T.L., Szostak, J.W. and Rothstein, R.J. (1981) Proc. Natl. Acad. Sci. USA 78, 6354-6358. 15. Doherty, M.J., Morrison, P.T. and Kolodner, R. (1983) J. Mol. Biol. 166, in press. 16. Foler, K., Wong, E., Wahl, G. and Capecchi, M. (1982) Mol. Cell Biol. 2, 1372-1387. 17. Orr-Weaver, T.L. and Szostak, J.W. (1983) Mol. Cell Biol. 3, 747-749. 6351