Copyright 0 1992 by the Genetics Society of America

Analysis of Subtelomeric Heterochromatin in the Drosophila Minichromosome Dpll87 by Single P Element Insertional Mutagenesis

Gary H. Karpen’ and Allan C. Spradling

Howard Hughes Medical Institute Research Laboratories, Carnegie Institution of Washington, Department of Embryology, Baltimore, Maryland 21210 Manuscript received May 28, 1992 Accepted for publication July 28, 1992

ABSTRACT We investigatedwhether single P elementinsertional mutagenesis could beused to analyze heterochromatinwithin the Drosophila minichromosome Dpll87. Forty-fiveinsertions of the P[lacZ,rosy+] element onto Dpll87 (recovered among 7,825 transpositions) were highly clustered. None was recovered in centromeric heterochromatin, but 39 occurred about 40 kb from the distal telomere withina 4.7-kb hotspot containing tandem copies a of novel 1.8-kb repetitiveDNA sequence. The DNA within and distalto this region lacked essential genes and displayed severalother properties characteristic of heterochromatin. The rosy+ genes within the inserted transposons were inhibitedby position-effect variegation, and the subtelomeric region was underrepresented in polytene salivary gland cells. These experiments demonstrated that P elements preferentially transpose into a small subset of heterochromatic sites, providing a versatile method for studying the structure andfunction of these chromosome regions.This approach revealed thata Drosophila chromosome containsa large region of subtelomeric heterochromatinwith specific structural and genetic properties.

HROMOSOMES in multicellular eukaryotes due to the priming problem (WATSON1972; CAVA- C have been studied extensively at thegenetic and LIER-SMITH1974). In Drosophila, chromosomes bear- cytogenetic levels. However,much remains to be ingterminal deletions lose 50-75 bp of DNA per learned concerning the molecular structure of chro- generation (BIESSMANNand MASON 1988; LEVIS mosomes andthe mechanisms underlyingchromo- 1989), presumably for this reason (LEVIS1989; BIESS- some behavior during thecell cycleand development. MANN, CARTERand MASON 1990). The structure of The study of small natural andartificial chromosomes normal Drosophila telomeres has not been described, has greatly aided the analysis of chromosome function however, and the existence of terminal GT-rich re- in yeast (MURRAYand SZOSTAK1983; NEWLONet al. peats remains to be demonstrated. 1991). For this reason we previously initiated molec- Telomericregions frequently contain other re- ular studies of a Drosophila minichromosome, peated DNA sequences in addition to hexanucleotide Dp(l;f)1187, or Dp1187 (KARPEN and SPRADLINC or oligonucleotide repeats. For example, most yeast 1990). Dpl187 contains only 1000 kb of centromeric chromosomes contain tandem repeats of a sequence heterochromatin, and 300 kb of distal DNA, making called Y’, and an X sequence (CHANand TYE1983). it the smallest known functional chromosome in Dro- Large arrays of repetitive sequences are found at the sophila and other multicellular eukaryotes. We now telomeres of many other organisms, including Plas- report the use of P element insertional mutagenesis modium (PACE et al. 1987), Secale (BEDBROOKet al. to analyze the distal subtelomeric region of @I 187. 1980), Chironomus (SAIGAand EDSTROM1985), and Chromosomal telomeres carry out specialized func- humans (COOKE, BROWNand RAPPOLD1985; BROWN tions essential forchromosome maintenance [re- et al. 1990; DELANGEet al. 1990). A specific family of viewed by ZAKIAN (1 989) and BLACKBURN(1 991)]. repeats called HeT has been mapped at many Dro- The short GT-rich oligonucleotide repeats found at sophila telomeres (RUBIN 1978; YOUNGet al. 1983). the extreme termini of chromosomes in a wide range The functions of these“subtelomeric” repeats are of organisms serveas substrates for theenzyme telom- unclear, since they varywidely in amount between erase (GREIDERand BLACKBURN1985). The sequences chromosomes andundergo rapidevolutionary added by telomerase counteract the 5’ shortening of changes (YOUNG et al. 1983; CORCORANet al. 1988; chromosomal molecules that would otherwise occur ZAKIANand BLANTON 1988).Recently, BIESSMANNet al. (1990) reported that HeT repeats can be added The sequence data presented in this article have been submitted to the onto a broken chromosome end, possible by retro- EMBL/GenBank Data Libraries under the accession number L03284. ’ Current address: Molecular Biology and Virology Laboratory, The Salk transposition. Sequence additions, along with telom- Institute, 10010 North Torrey Pines Road, La Jolla, California 92037. ere-telomere recombination, might contribute to the

Genetics 132: 737-753 (November, 1992) 738 G. H. Karpen and A. C. Spradling rapid changes observed in subtelomeric regions. females. Among an average of 100 FP progeny from each

In Drosophila, telomeres have been associated with vial, y+ .ry+ Sb+ males were recognized as bearing new transpositions of P[lacZ, rosy+] (hereafter called PZ; see additional genetic and cytogenetic properties. Telom- MLODZIKand HIROMI1992) not linked to the X chromo- eresare generally thoughtto beheterochromatic some. These males were mated individually with twoy ;ryro6 (SCHULTZ 1947),and share common sequences with females, however, a maximum of only twoy+ ry+ Sb+ males centric regions and the Y chromosome (YOUNGet al. were used from any one F1 vial to avoid clusters, and males 1983). It has been suggested by cytological examina- from the same vial were recorded and kept together for further study. Approximately 8,300 such F2 crosses were tions,but not proven by molecular analyses, that established, including 950 pairs. In some F1 crosses the telomeric regions become underrepresented in poly- transpositions all segregated from Dpll87, or segregated tene DNA (ROBERTS1979). Rearrangements that jux- with the chromosome containing A2-3, hence the number taposeeuchromatic andheterochromatic chromo- of F:! vials established wasless than the total number of some regions frequently cause genes near the break- transpositions that occurred. Lines containing insertions onto Dpll87 were recognized point to display position-effect variegation (reviewed by co-segregation of they+ andry+ markers in progeny from in SPOFFORD1976; HENIKOFF 1990; SPRADLINGand the F2 outcrosses (at least some y+ ry+ progeny, but no y- KARPEN1990). The expression of a white+ gene relo- ry+). Linkage of the PZ element to the Y chromosome was cated near the 3R telomere variegated (HAZELRIGG, established by observing transmission of the ry+ marker only LEVISand RUBIN1984; LEVIS,HAZELRIGC and RUBIN from fathers to sons.Lines containing an insertion on Dp1187 and a second insertion on another chromosome 1985), however, it remains unclear if telomeric re- were recognized (despite the presence of y-ry+ animals) by gions are generally capable of causing variegated po- observing that all y+ progeny were ry+; the insertions were sition effects. then separated in a subsequent generation. A total of 42 Studies of Dpl187 telomeres, and of its centromeric independent lines with insertions on Dp1187 were eventu- region, were limited by a lack of specific molecular ally recovered (Table 2). Autosomal PZ insertions were balanced over CyO, TM? ryK,or (rarely) CxL) and recessive probes and of mutations mapping to these regions. (or rare dominant) phenotypes associated withthe insertion- We thereforeused insertional mutagenesis with single bearing chromosome classified. A small number oflines P elements (COOLEY,KELLEY and SPRADLING1988) with insertions on the 4th chromosome or lines that were to determine if genetically marked P elements bearing not balanced over either Cy0 or TM? for undetermined unique sequence tags could be recovered in reasons were discarded. In addition, ovaries were dissected Dp1187 from adults of each line, stained for &galactosidase activity, heterochromatin.Heterochromatin is generally and the expression pattern recorded. Subsequent molecular thoughtto be an infrequenttarget of P element and phenotypic study revealed that approximately 50% of insertion (ENGELS1989), but our experimentsre- the lines isolated as pairs contained identical insertions, so vealed that asignificant number of insertions occurred the total number of independent insertions recovered was nearthe distal telomere of this minichromosome. approximately 8,300 - 475 = 7,825. Some of these results will be presented in further detail elsewhere. These studies have provided new informationcon- Plasmidrescue: Sequences flanking an insertion were cerning the structure and function of Drosophila te- rescued by first preparing DNA from ten adult males as lomeres, andtheir relationship to transposable ele- described (BENDER,SPIERER and HOCNESS1983). Following ments. digestion with XbaI, or XbaI and SpeI, the reaction was diluted, ligated overnight at 15", and used to transform competent cells to kanamycin resistance, essentially as MATERIALS AND METHODS DH5 described (PIROTTA1986; COOLEY,KELLEY and SPRADLINC Drosophila stocks: Flies were grown on standard corn 1988). Approximately one colony was obtained for each fly meallagar media [see ASHBURNER(1990)], at 22". Unless equivalent of DNA used for transformation. stated otherwise, strains and mutations are as described in Restrictionmapping using pulsed-field and conven- LINDSLEYand ZIMM (1992). The original stock containing tional gel electrophoresis:The location ofthe PZ insertions Dp1187 was provided byD. LINDSLEY,and was balanced in Dpll87 was determined by pulsed-field Southern analysis, with fl-YL,Df(1)259 y w. and C(l)RA, l(1)Jl In(l)sc8.The X utilizing the NotI and Sf1 restriction map (see Figure 3) and chromosome was found to have broken down, andthe the presence of sites for these enzymes within the PZ tran- duplication was subsequently maintained in a y background, sposon. High molecular weight DNAwas isolated from hand by selection for y+. dissected larval brains and imaginal discs, or adult ovaries, Transposon mutagenesisof Dp1187: Prior to initiating using the agarose insert method described previously (KAR- the screen, sublines of the y ac PZ;cn ;ry ;Dp1187 stock, PEN and SPRADLINC1990). DNA from each Dp1187 PZ the cn ;ry Sb A2-3 ry+ /TM2,Ubx stock (see ROBERTSONand strain was digested with Not1 or SfI. Pulsed-field Southern ENCELS1989 for the derivation of the A2-3 transposase- blots (see KARPENand SPRADLINC 1990)were probed with encoding transformant), and they ;rySo6 stock were started a fragment (pBS 12.1BH9)corresponding to positions -80 from single pair matings. Two sublines of each stock that to -89 on the Dp1187 map (see Figure 3) for the NotI showed very high levelsof viability and fertility were used. digests, anda fragment from -40 to-51.5 (PBS The crossing scheme usedfor transposon mutagenesis of TGlBPll.5)for the Sf1 digests. The 1 00-kb1 proximal Not1 DpI187 is diagrammed in Figure 1, and the results of the fragment was reduced in size by transposon insertions prox- screen are summarized in Table1. Fo crosses were per- imal to position -100 on the Dp1187 map (see Figure 3), formed en masse, in bottles started with approximately 20 while insertions to the left of -40 reduced the size of the males and 20 females. Approximately 22,000 F1 dysgenic 250-kb distal S'I band. The orientation of the insertion males were individuallymated in vialswith twoy; ry506 virgin then was determined by reprobing the blots with probes Dpl187 Subtelomeric Heterochromatin 739

Generation CrosS FO

select yellow+, rosy+, Stubble male progeny (dysgenic maleswith PZ & Dp1187) FIGURE1 .-Crossing scheme F1 y;rySwyy x yacP[lacZ.ry+]IY;cn;rySb42-3ry+;Dp1187,y+b used for single PZ mutagenesis of Dp1187. See MATERIALS AND METH- J. ODS for details. select yellow+, rosy+, Stubble+ male progeny (animalswith a new mnsposition of PZ) F2

score segregation of y+ and ry+: ligeindicates a new insertionof PZ on DpI 187

TABLE 1 appropriate regions were subcloned and additional primers were synthesized. A summary of the information used to Summary of PZ transposition screen assemble the sequence is given in Figure 2. Sequence assem- bly and analysis was accomplished with the Genetics Com- Class Number puterGroup software, version 7 (DEVEREAUX, HAEBERLI FI crosses (approximate) 22,000 and SMITHIES1984). In Transposition lines 8,300 situ hybridization: In situ hybridization to polytene Independent lines 7,825 chromosomes was carried out as described previously (KAR- PEN and SPRADLINC1990). Briefly, RNA probes were pre- Lethals: II + 111 958 pared by transcribing plasmidDNAs (subcloned in Blue- Female steriles: I1 + 111 190 script vectors) with polymerase in the presence of [‘%] Male steriles: I1 + 111 95 T7 UTP and [35S]CTP. Hybridization, washing and autoradi- Insertions on DpI 187 42 Insertions on Y 24 ography were then performed as described. DNA preparation and Southern blotting:DNA was pre- Lines with recessive phenotypes represent those maintained in pared from brains, imaginal discs, and salivary glands of stock. A small fraction were lost during balancing. Approximately third instar larvae, as described previously (KARPEN and 200 lines bearing an identical third chromosome background fe- SPRADLINC1990). Electrophoresis, blotting, autoradiogra- male sterile mutation are not included among the steriles. Approx- phy, and quantitation of bands was as before (KARPENand imately 200 lethal lines that were accidentally contaminated are also not included in the lethal category. SPRADLINC1990). homologous to the 5‘ (pSS 5’ Pend HR 0.55) or 3‘ (pSS 5’ RESULTS ryHR3.2)ends of the transposon. Yeast chromosomes (strain YPH 149, supplied by PHILHIETER), multimerized bacteri- Restrictionmapping of Dp1187: Dpl187 is the ophage X DNA, and a “5-kb ladder” (obtained from Bio- Rad), served as size markers. smallest of a series of free X chromosome duplications The map of the 9.9-kb SpeI-XbaI fragment was con- (minichromosomes) generated by deleting all but the structed using several sources of information. The genomic proximal and distal sequences within In(I)sc8, an X DNA present in the plasmid rescued from most of the 42 chromosome containing a largeinversion [see KARPEN insertion lines was mapped usingseveral restriction en- and SPRADLING(1 990) for adescription of the origin zymes. To independently verify the location of each inser- tion, genomicDNAs (isolated from adult flies; BENDER, of Dpl187]. Previous studies using pulsed-field gel SPIERERand HOCNESS1983) were subjected to conventional electrophoresis (KARPEN and SPRADLING1990) de- gel electrophoresis, and Southern blots were hybridized with fined the sc8 breakpoint as 0 kb on the Dp1187 map, the transposon-specific probes (see above), according to the and showed that the chromosome contained approxi- methods described in KARPENand SPRADLING1990. The mately 1000 kb of centromericheterochromatin sizes of the XbaI, XbaI-Spel,and EcoRI fragments were used to localize and orient the insertions within the distal hotspot proximal to this site, and about 300 kb of distal DNA (see Figure 5). When preliminary studies indicated that a (see Figure 3). Prior to initiating insertional mutagen- strain contained a second insertion on an autosome (see esis, we carried out additionalmapping to further Table Z), the insert-bearing duplication was isolated free of define possible target sites. High molecular weight other P elements by outcrossing males to y ;rySu6 females DNA was prepared from Drosophila tissues in agarose for at least two generations, then reanalyzed by Southern analysis. inserts, digested with restriction enzymes, and sepa- DNA sequencing: DNA sequencing reactions were car- rated on pulsed-field gels containing size markers. ried out essentially as described (LEVINEand SPRADLING Southern hybridization analysis revealed several facts 1985), using Sequenase 2.0 kits (U.S. Biochemical). Prepa- about Dp1187 structure not reported previously. The rations of the marker rescued clones were used directly as size of the portion of Dp1187 corresponding to the templates, with a P element-specific primer located just inside the 5’ inverted terminal repeat of the PZ element normal X tip, ie. sequences distal to theSC’ breakpoint, (sequence: 5’ GTATACTTCGGTAAGCTTCGGCTATA measured 290 f 5 kb (rather than 340 kb, as reported 3’). To complete the sequence of the SpeI-XbaI fragment, previously). A putative telomere was subsequently po- 740 G. H. Karpen and A. C. Spradling

TABLE 2 Properties of Dp1187 insertions

Site in Site in Of11187 hotspot Rosy d Line (Wa (bP)b Orientation‘ expression Comments 0367 -249 4549150 < +++ 040 1 -246 730516 > +++ 0517 -246 ND > ND Isolated in strain with second autosomal insert 0801 -248 493617 < ND Isolated in strain with second autosomal insert 0809 -246 674213 > ++ 1401 -247 586011 > ++ 1601 -247 ND > +++ 1630 -247 640617 < ++ 1801 -247 639718 > ++ 1803 -242 < ND Isolated in strain with second autosomal insert; isolated as part of a premeiotic cluster 2001 -249 4549150 < Wt 2 inserts on Dp1187 -85 < 2202 -247 ND < ND Isolated in strain with second autosomal insert 2622 -250 356718 < ++ 2801 -246 ND > ND lsolated in strain with second autosomal insert 300 1 -246 662819 > +++ 3401 -25 1 257516 < ++ 3402 -250 3 19819 < +++ 360 1 -250 ND < ++ Terminal deficiency 3901 -249 370617 > ++ 400 1 -246 ND > Wt 2 inserts in Dp1187 hotspot 4003 -246 730516 > ++ Isolated in strain with second autosomal insert 4201 -246 6569170 < +++ 440 1 -246 7339140 > Wt 4404 -249 ND < +++ 5402 -249 ND < ND Isolated in strain with second autosomal insert 5403 -248 530 112 > +++ Isolated as part of a premeiotic cluster 6201 -246 675112 > +++ 6231 -75 < Wt 6401 -248 ND > +++ 660 1 +goo ND + Unstable centromeric insert (see text) 720 1 ND ND ND ND 2 inserts in Dpl187 hotspot 740 1 -248 545112 < +++ 7403 -247 ND < wt 7606 -249 454011 > +++ 8002 -75 < Wt 860 1 -250 300617 < +++ 8602 -247 639718 > +++ 880 1 -80 < wt 921 1 -247 639718 > ND Isolated in strain with second autosomal insert 9901 -248 47 1314 > +++ 10240 -246 ND > ND Isolated in strain with second autosomal insert 10500 -246 730516 < ND Isolated in strain with second autosomal insert

a Approximate position of PZ insertions; kb = kilobases from the SC’ breakpoint (=0 kb). See Figures 3 and 5. ND = not determined. Position of the PZ insertion in the Dpll87 hotspot; bp = basepairs proximal to the SpeI site (see Figure 7). Orientation of the PZ element within Dpll87. Arrowhead indicates the 5’ P-ZacZ end of the PZ element; the Dpl187 centromere is to the right, the euchromatic telomere to the left (see Figure 5). Distinguishing weak ry* from ry- is difficult. Lines were therefore categorized according to the following scale, basedon counts of >ZOO flies per line: wt = 0-5% of the progeny ry-, +++ = 6-25% ry-, ++ = 26-49% ry-, + = 50-100% ry-. Note that the presence of a second insert affected the classification of some lines (ie., compare 0367 and 2001). wt, wlld type. sitioned at coordinate -290 by more precise methods site for each enzyme within the centromeric hetero- (see below). chromatin. The first 50 kb contained few restriction In addition, detailed mappingof the first 200 kbof sites (Figure 3), as expected for a region of simple the centric heterochromatinrevealed evidence of sub- sequence DNA. The 359bp (or 1.688) “complex” structure. Southernblots of the gels were probed with satellite DNA (HSIEHand BRUTLAG1979) was sus- a 3.7-kb XhoI-Hind111 fragment adjacent to the sc8 pected to lie at the sc8 breakpoint from cytological breakpoint (pBSscXR3.7) (KARPEN and SPRADLINC analyses (HILLIKERand APPELS1982); this has been 1990), to moreaccurately map the location of the first confirmed by cloning and sequencinga restriction DpI 187 HeterochromatinSubtelomeric 74 1

c - "- ""4 = f " uninterrupted across the 15-20-Mb centromere re- c - c c- c-- c "C " "= =- " "2r" r- L" c - c "" c " - gions of the major chromosomes. Two enzymes, SpeI I """ I """ I """I """ I """I """ I """I_ ""_ I """I """1 0 1000 2000 3000 4000 50006000 7000 8000 9000 10000 and MluI, did not cut within Dp1187 until positions bP +180 and +215. This may indicate the presence of FIGURE2.-Overview of the DNA sequencing strategy used for another region of simple sequence DNA followed by the .'+el-Xbal fragment. Arrows indicate the location, orientation and size of sequencedfragments used to assemble the 9,872-bp a second island. sequence shown in Figure 5. bp = base pairs, where 0 refers to the Insertional mutagenesis of Dpl187 heterochro- tlist;ll-most (Spel), and 10000 the centromere-proximal (Xbal),ends matin; a new approach to genomic structural analy- of the fragment. sis: Genomic regions rich in repetitive DNA sequences are difficult to analyze by standard cloning methods. -290 0 +IO00 I I I Approximately 120 kbof DNA distal to thesc' break- Not I - 100kb point was cloned from standard genomic libraries by SfiI -.-'I' ' chromosome walking (FLEMING,DESIMONE and WHITE 1989; KARPEN and SPRADLING1990). How- ".-.. ever, attempts to walk more distally (and within the -. centromeric region) were discouraged when regions

11 si I 11 Eag I of repeated DNAs began to be encountered. In prin- .-..&y I ciple, insertional mutagenesis with single P elements Nco I (COOLEY,KELLEY and SPRADLING1988) could solve IIIIIIIIII -120 -80 -40 0 +40 +SO +I20 +160 +ZOO +240 kb many of the problems associated with studying large FIGURE3.-Molecular structure of Dpl187. The restriction map chromosome segmentslacking unique sequences. The of Dpl I87 is shown above, with sites for the rare-cutting enzymes presence of genetically marked insertions would Not1 and Sf1 indicated by hashmarks. Values indicate distance in greatly facilitate structural analysis, since each inser- kilob;tses from the sc' breakpoint (= 0) into the euchromatin (solid tion would contain unique DNA sequences that could line, - values) and the heterochromatin (stippled bar, values). + be used forrestriction mapping. Furthermore,the 1)et;ded restriction map of the region surrounding the sc' break- point is shown below. Only the first restriction site proximal to the chromosome region could be mutagenized by impre- heakpoint is shown for the enzymes that digest the heterochro- cisely excising inserted transposons, facilitating sub- matin. For cl;wity, only the NotI, SjI and Xhol sites are shown in sequent functional studies. the euchromatin. P element insertional mutagenesis, however, has generally been thought to be of little use in analyzing fragmentthat spans the Dp1187 breakpoint (R. GLASER,G. KARPENand A. SPRADLING,unpublished). heterochromatic sequences. P elements located within heterochromaticregions have only rarely been ob- The two enzymes that cut close to theSC' breakpoint, served in natural populations, or following transfor- AseI and Sac1 (Figure 3), are known to have sites within the 359-bpsequence (HSIEH and BRUTLAG mation (ENGELS1989). The paucity of characterized 1979). Of the remaining enzymes tested, 12/15 first insertions into heterochromatin could be caused by cut between +55 and +90 suggesting that approxi- an inability of P elements to transpose into these mately 50 kb of simple sequence DNA is followed by regions, due to a lackof favorabletarget sites or an "island" of more complex DNA between +55 and because of the chromatin environment.Alternatively, +90. The presence of complex DNA within the cen- heterochromatic P insertions might be unable to ex- tromeric region encouraged the view that P element press theirmarker genes and become underrepre- insertions might occur within them. sented in polytene cells, making either genetic or A limited amount of data suggested that simple physical detection difficult. A few P element insertions sequence DNA might alternatewith "islands" of more within heterochromatin of the4th and Y chromo- complex sequences throughout Drosophila centrom- somes have been recovered previously, and lines bear- eric heterochromatin. When the pulse-field gel South- ing these insertions expressedthe marker genes pres- ern blots usedin thesemapping experiments were ent on the transposon at reduced levels (SPRADLING rehybridized with satellite DNA-specific probes (359- and RURIN1983; LEVIS 1989; BERGand SPRADLING bp complex satellite, 1.672 = AATAT and 1.705 = 199 1). AAGAG, LOHEand BRUTLAG1986) a heterogeneous We therefore designed an experiment to determine pattern of bands between 50-900 kb was observed if insertional mutagenesis could be used to analyze (data not shown). Most Drosophila satellite DNAs are Dpll87 heterochromatin. A large mutagenesis screen based on simple penta- or heptanucleotide repeats and was carried out in which a single PZ element (MLODZIK would not be expectedto contain sites for most of the and HIROMI1992), located initially on the X chro- enzymes tested (LOHEand BRUTLAG1986). The ob- mosome, was mobilized in a strain that also carried served sizes were therefore difficult to reconcile with Dpll87 (see MATERIALS AND METHODS). The Pz ele- a model in which simple sequence DNA stretches ment was selected as the insertional mutagen, since it 742 G.and H. Karpen A. C. Spradling carries an enhancer-sensitive P-lac2 fusion gene for Insert Lines enhancertrapping, as well assequences allowing markerrescue in Escherichia coli. Furthermore, we preferred using the rosy+ (ry+) gene as a marker be- cause very low levels of ry+ function are required to produce wild-type eye pigmentation. From a total of 8300 lines containing new insertions(representing -210 approximately 7825 independent insertions), 42lines were recovered that contained independent PZ inser- tionson Dp1187 (Table 1). Subsequentmolecular 49 analyses revealed that a total of 45 insertions were - - - 39 actually present in these lines (Table 2). Thisfre- quency(45/7825 = 0.58%) was similar tothat ex- pected assuming all Dp1187 genomic DNA was avail- able for insertion in the F1 males (1.3 Mb/l80 Mb = IXI -320 "io kb 0.72%), but higher than expectedif only euchromatin was a target (0.29 Mb/107 Mb = 0.27%). The location of the PZ insertion in each strain was -290 -40 o +loo0 determined initially by mapping new Not1 and SfiI Sfi I restriction sites caused by the presence of the PZ element in their resident minichromosomes (Figure 4). Subsequently, genomic DNA flanking the 5' end of each insertion was recovered by marker rescue, and mapped in greater detail with restriction enzymes. -290 -248 -40 0 +loo0 Flanking DNAs cloned by marker rescue were com- FIGURE 4."Mapping new PZ insertions on DpIIR7 by puked- field gel electrophoresis. Pulsed-field Southern hybridization analv- pared to the known genomic structure determined sis of Sfil-digested ovary DNA from DpII87 insertion lines was from Southern blots, to minimize the possibility that performed as described in MATERIAL5 AND METHODS, and a repre- clones had undergone rearrangements during isola- sentative autor;diogranl is shown. The genotypes of the lines are tion and growth in E. coli. y ;ry"" (= Dp-), y ;ly5'J6 ;Dp(1$1187, y+ (= Dp+), and y ; ;Dp

The distribution of insertionsalong the Dp1187 (with PZ insertion), y+ ry+ (= 0801-6231). Sizes of restriction fragments are indicated to the right; X= X chromosome-specific map was extremely nonuniform (Figure 5). A major fragment, Dp = DplI87-specific fragment. Diagrams of the struc- "hotspot" located near the distal telomere, 246-251 ture of this region of the X, Dpl I87 and DpOSOI chromosomes are kb from the SC' breakpoint(coordinate -246 to shown below, with the locations of the Sf1 sites indicated. Open -251), contained 39 of the 45 insertions. However, box denotes the probe used in this experiment (TGIBP11.5). Not1 digests, and probe 12. IBH9. were used to analyze the structure of within this 4.7-kb region, insertions occurredat many the derivatives. proximal to -100, in a similar manner, capitalizing different sites and in both orientations (Table2). The on thepresence of two Not1 sites in the PZ element. For a description insertion in strain 1803 mapped a few kilobases prox- of other symbols refer to the Figure 3 legend. See Figure 5 and imal to the hotspot, while four other insertions were legend for the structure of the PZ element. positioned between -75 and -85, a well studied re- allowed the subtelomericregion to be mapped in gion of euchromatin. Only oneline, 6601, contained greater detail.Southern blots with genomic DNA acandidate insertion within DpI187's 1000-kb seg- flanking the hotspot insertions revealed only the pres- ment of centromeric heterochromatin. Initially, this ence of repetitive DNA sequences (see below). Con- line expressed ry+ very weakly, but showed clear ge- sequently, we carried out additional mapping using netic linkage to y+ for several generations. However, line 3401, in which the single copy sequences within the ry+ marker subsequently became linked to chro- the PZ transposon were located closest to the telomere mosome 3. Mapping with Not1 revealed a structural (Figure 6A). These studies verified that both the Not1 change within DpIl87 at +900 but no evidence of P and SfiI distal "sites" mapped to the same location. element sequences (data not shown). We have been Furthermore, two others enzymes, XbaI and XhoI, unable to determine whetherPZ initially inserted onto also appeared to cleave at this same position, while no Dp1187 in strain 6601and then subsequently was enzyme was found with amore distal site. These transferred to chromosome 3 by a heterochromatic observations suggested that the minichromosome dis- exchange event, or if an original insertion on chro- tal telomere resides at this location. The existence of mosome 3 was associated with unstable re- a telomere at -290 is consistent with size measure- arrangements that caused ry+ and y+ to appear linked. ments of uncut DplI87, that suggested a total length The subtelomeric region distal to the hotspot: The of about 1300 kb (data not shown). presence of insertions near thedistal Dpl187 telomere The insertion in strain3601 differed from the DpI 187 Subtelomeric Heterochromatin 743 ciated with the genes uncovered by these deficiencies, and restored the flies to a fully wild-type state. We conclude that all the essential genes on the X, and on Dp1187, lie proximal to -250, and that the terminal 40 kb of the minichromosome lacks genes with easily +lo00 kb '. detectable phenotypes. Whether redundant functions +900 are encoded by repetitive sequenceswithin this region could not be addressed by these studies. DNA sequence of the hotspot region: The wide distribution of PZ insertions throughout the 4.7-kb *. hotspot greatly facilitated determination of its DNA 3402 08011 1630 sequence. A primer specific for the 5' P end was used to sequence flanking genomic DNA within most of the rescued plasmid templates, revealing much of the sequence of both strands. The sequences were then extended using oligonucleotides homologous to the ends of the initial runs, and templates from subclones of the marker-rescued plasmids (see MATERIALS AND METHODS). Since the insertion sites were unambigu- FIGURE5.-Location of PZ insertions on DplIR7. The approxi- ously positioned within the hotspot based on the ge- mate location and orientation of the PZ insertions, determined by nomic and plasmid maps, it was possible to join the pulsed-field Southern analysis (see MATERIALS AND METHODS and Figure 4). is shown above. The locations and orientations of 37 of sequences into a continuous IO-kb segment spanning the 40 insertions within the subtelomeric hotspot, determined by theentire SpeI-XbaI fragment(Figure 7). Further- conventional Southern analysis and/or analysis of plasmid rescued more, this method revealed the exact nucleotide at flanking DNA, are displayed below. Numbers indicate strain name. which most insertions occurred (Table 2). Arrow points toward the 5' P end of the PZ transposon. Black The hotspot region contained a series of different boxes = P ends: box with diagonal lines = 5'P-lacZ fusion plus the kanamycin" gene and the bacterial origin of replication; box with repetitive elements arranged in an organized, specific checkerboard is the rosy+ gene (ry+). Spe = Spel, Hind = HindllI, pattern. Starting at the distal SpeI site (Figure 7), the K = RcoRI. Xba = XbaI. See Table2 for nucleotidelocations within first 2 163 bpcomprised a member of the HeTfamily the SpeI-XbaI fragment. Refer to the Figure 3 legend for all other (RUBIN 1978;YOUNG et al. 1983; TRAVERSEand PAR- symbols. DUE 1989; BIESSMANNet al. 1990), followed by 3.1 others in the region distal to the hotspot. 3601 con- nearly perfect tandem repeats of a novel 1.8-kb se- tained an insertion at-250, and in addition,the quence (called 1.8A-1.8D), then 2.2 tandem repeats chromosome appeared to be broken -650 bp distal of a 0.9-kb sequence (called 0.9A and 0.9B). At the from the 5' P end. A series of restriction enzymes junction between the different types of repeated ele- showed the existence of apparent sites at this position ments, the 5'-most element in each group was trun- (Figure 6B), althoughsuch sites are absent in the other cated or internally deleted.These large repeats in lines (i.e.,Dp6201, Figure 6B).Furthermore, thesmall turn contain internally repeated, short sequence ele- terminal bands migrated heterogeneously ongels only ments (-80- 170 bp). in 3601 (Figure 6B, EcoRI digest), as previously ob- The HeTelement strongly resembled HeT repeats served for broken chromosomeends (BIESSMANNand sequenced previously (VALGEIRSD~TTIR,TRAVERSE MASON 1988; LEVIS 1989).Perhaps the insertion and PARDUE 1990),except a gapwas present between event that produced the 3601 strain simultaneously bases 910 and 1 120, and the first 450 bases of the caused a deletion of the distal 40 kb. Alternatively, a HeT DNA was present as a tandem, diverged repeat. secondary P element-catalyzed event may have caused Heterogeneity among membersof this family has been the deletion during germ-line cell divisions subsequent noted previously (TRAVERSEand PARDUE 1989). The to insertion, similar to the breakage events observed 3' terminus of the HeT element joined the first 1.8- by LEVIS(1 989)following experimental mobilization kb tandem repeat in a region of five A residues. This of a subterminal P insertion. junction was identical in structureto that recently We used the terminally deleted DpI187 derivative observed at the tip of two terminally deficient chro- in strain 3601 (Dp3601) to learn if any genes reside mosomes that had been "capped" by addition of an in the terminal 40 kb that lies distal to the hotspot. HeT element (BIESSMANNet al. 1990); both the pres- Appropriate crosses were made to aseries of mutated ence of poly(A) residues and the position of the junc- or deleted X chromosomes, including DXl)sc8 which tion within the HeT sequence were the same in the lacks all the DNA sequences distal to the sc8 break- Dpl187 hotspot and in the terminal additions. The 5 point (Figure 6C). Dp3601 rescued the lethality asso- A's may have been part of an HeT mRNA poly(A) 744 G. H. Karpen and A. C. Spradling EcoRI BamHISal1 Not1

rl3 333rlwrl Dp 3401 00 0000 00 r4UNWr4Wr4W kb \om wmwm wm - """I-

-290

1.2 -b

C) Dp 3601 ,, \I ...... ,, ~~ ...... :250 0 ...... +loo0 Dp 3601 k I) -650 bp BamHI 6 EcoRI Sal1 Not1 Xbal centromere 3 I(I)ECI (dmd)I(I)JlI(I)EC2 cin ewg arrh J uc sc. n ...... * * * mutations 0 PllacZ rosy + DJl) 05-22-1 - 0.6 kb Dfcl)259 3.5 Of(!)x* 4*5 7.0 FIGURE6."Structure of the subtelomeric region of DpI 187 elucidated by analyses of Pi! insertion lines. A) The restriction map distal to the insertional hotspot W;IS determined by pulsed-field Southern analysis, using the indicated enzymes, DNA from line Dp340l (the distal- nlost insert in the Dpl I87 hotspot), and a single-copy probe from within the transposon (open box). B) Dp360I contains a terminal deficiency. The restrictiotl map of the region distal to two different hotspot insertions, in lines Dp360I and Dp6201, was determined by conventional electrophoresis and Southern analysis. The restriction map for Dp3601 is shown below the autoradiogram; probe = the 5' P end (open box). The coincidence of sites for all 6 enzymes -650 bp distal to the endof the 3601 insertion (the Xbal digest yielded a fragment of 7.6 kb, data not shown), plus the heterogeneous migration of the short EcoRI fragments, indicate that 3601 is terminally deleted. C) Complementation of X chromosome mutations (asterisks)and deletions (bars with diagonal lines indicate the genes that are absent in the deficiencies) by Dp3601 localizes the distal-most X-linked genes to the region proximal to the 3601 insertion. Other symbols as in previous figures. tail prior to its reverse transcription and ligation to gous from the truncated 160 bp repeat through to the distal 1.8-kb repeat at some previous time in the theend of the80-bp element. Finally, thefourth evolution of the Dpl187 telomeric region. repeat(1.8D = 1574bp, positions 6102-7676) is The two complete 1.8-kb repeats (1.8B = 187 1 bp >95% homologous to 1.8B and C, except that -280 and 1.8C = 1857 bp, Figure 7) are >95% homologous bp are deleted near the 3' end of 1.8D, including the over positions 2372-4243and 4244-6101. These entire 160-bp repeat. sequences exhibit nosignificant homology to anything The SpeI-XbaI fragmentends with two repeats in the available databases, and have no large open (0.9A = 874 bp, positions 7809-8683, 0.9B = 910 reading frames. Each 1.8-kb repeat itself contains bp, positions 8685-9595) that also have no homology three tandem copies of a 173/173/161-bp subrepeat to current database entries, and lack significant open (dashed line, hollow arrow in Figure 7); the two 5' readingframes. These elements each contain two most repeats contain a12-bp insertion (box with copies of a167-bp repeat (jagged line, half-arrow, slashes) absent in the third repeat. The 1.8-kb repeats Figure 7), plus one complete and one truncated copy each contain single copies of 80-bp (double line) and of a 160-bp element thatis identical to that presentin 160-bp (solid line and arrow)elements. The first copy the 1.8-kb repeats (solid line, solid arrow Figure 7; of the 1.8-kb repeat (1.8A) is only 205 bp in length; see above). The truncated 160-bp repeats are joined it is truncated at its 5' end (the junctionwith HeT) in at their 3' ends to oneof the complete 167-bp repeats. comparison to 1.8B and 1.8C, but is >99% homolo- The region between 1.8D and 0.9A is in fact a deleted Subtelomeric HeterochromatinDpl187 Subtelomeric 745 version of a 0.9-kb repeat, containing a nearly com- elementinsertional hotspot in the singedlocus plete 167-bp repeat fused to a truncated 160-bp ele- (CACTGGAG; ROIHA, RUBIN andO’HARE 1988; ment. Thus, the region proximal to the 1.8A-D re- HAWLEYet al. 1988). We concludethat the 1.8-kb peats, from 7677 to 9595, contains a series of related repeats displayed the properties of a regional hotspot sequences whose distal-most element is rearranged in for P element insertion,with some sequence specificity comparison to the more proximal elements. Finally, in site choice within the region. The molecular mech- the last 276 bp of the hotspot region (9596-9872) anisms responsible for the unusual distribution of PZ contains sequences that are unrelated to the more insertions in this hotspot have yet to be elucidated(see distal elements. Further analysis will be necessary to DISCUSSION). determine if these sequences are repeated in the re- Multiple copiesof the 1.8-kb hotspot sequence are gion proximal to the XbaI site, and thus follow the located predominantly at salivary gland telomeres: pattern established for the organization of the HeT, A 1.8-kb EcoRI fragment containing a single 1.8-kb 1.8- and 0.9-kb elements. repeat and a 0.9-kb EcoRI fragment containing one Distribution of P element insertion sites within of the proximal repeats were subcloned fromthe thehotspot: The P elementinsertions within the plasmid rescued clones. Southern analysis of whole hotspotregion were distributednonrandomly, yet Drosophila genomic DNA indicatedthat sequences displayed a regional, rather than a site-specific, inser- within the 1.8-kbelement are moderatelyrepeated tional preference. PZ elements only inserted within a within the Drosophila melanogaster genome (50-1 00 4.7-kb portion of the 1.8-kb repeats;no elements were copies, data not shown). These sequences are present recovered in the distal HeT element or the proximal predominantly as 0.95 kb and 0.85 kb repeats. The DNA (see Figure 5, and Figure 7 flags). The only location of these sequences was determined by in situ exception was theinsertion in line 1803, which hybridization to salivary gland polytene chromo- mapped by restriction analysis to be less than 1 kb somes. Both fragments strongly labeled the telomere proximal to theXbaI site. The sequences at the5’ end regions of two chromosomes, 2R and 3R (Figure 9, of this insertion have not been determined, and could large arrows). The chromocenter region also was la- turnout to match sequences within the SpeI-XbaI beled weakly (Figure 9A, bracket), as was one basal fragment. There was a disparity in the distribution of euchromatic site (Figure9A, small arrow).Strong insertions among the different 1.8-kb elements, de- labeling of the X chromosome tip was not observed, spitetheir similarity in sequence.Surprisingly, the presumably because the X chromosome in the y ;ry506 internally deleted 1.8D repeat received more inser- strain used to preparechromosomes containedat most tions than the complete elements 1.8Band C (Figures a few copies of these sequences. The labeling at the 5 and 7; 7 PZ insertions in 1.8B, 11 in 1.8C, and 18 chromosome tips frequently was observed to extend in 1.8D. along ectopic fibers that emanate from the tips and Insertion sites were distributed throughout the in- extend to other chromosome regions (Figure 9, open dividual 1.8-kb repeats, and were rarely located at the arrow). In some cases, labeling completely spanned same nucleotide (only 3 clusters of 2, 3 and 3 inser- the ectopic fiber linking the 2R and 3R tips, which tions, Figure 7). However, there was some preference are rich in these sequences (Figure 9B, open arrow). forthe 173/161-bp subrepeats, since 56% of the Insertions in thehotspot region are subject to insertions (1 4/25 sequenced insertion sites) occurred position effect: The subtelomeric region of Dpll87 there, while these elements constituteonly 27% of the exhibited a deleterious position effect on the expres- DNA in the 1.8-kb element. Furthermore, a compi- sion of the inserted ry+ gene. Unlike the majority of lation of insertion sites for allof the173/161-bp PZ insertions, including those in the Dp1187 euchro- repeats revealed that most (1 2/14) of the insertions matin, alllines containinga single insertion in the inserted within an 1 l-bpregion of the repeat (Figure hotspotproduced progeny withweak and variable 8A). Two sites, at the 5’ and 3’ ends of this region, expression of ry. Afraction of the progeny were contained multiple insertions (8 and 4 insertions, re- phenotypically ry-; this number depended on the site spectively). Compiling data from the different 1.8-kb of insertion(Table 2). Among the8300 lines re- repeats demonstrated that four independentPZ inser- covered in the screen, the weak and variable ry phe- tions also occurred within another specific sequence notypes associated with hotspot insertions were similar located outside the 173/161-bp repeats (Figure 8B). only to those displayed by 24 insertions linked to the The flanking sequences at all three sites with multiple Y chromosome (data not shown). Regardless of the insertions are significantly diverged from each other, eye phenotype, all progeny transmitted the minichro- and from the P insertion8-bp consensus sequence mosome and expressed the yellow+ (y+) gene it con- derived by O’HARE andRUBIN (1983). However, the tained. Loss of ry+ expression was not due togerminal region 3’ to one of these clusters (CACTGGCG; Fig- mutation, since progeny of ry- individuals showed the ure 8A, stippled box) contains a 7/8 match to the P same range of phenotypes as their ry+ sibs. The re- 746 G. H. Karpen and A. C. Spradling

M- HeT ___) Sp.l 1 ACTAGTMTGCCGCGCTGCCAGACGGGATACACAAACA

121 CATCCMTGAAACAGACAG~GAAGGGCCCG-CGCAT-TCGCCAACTATGCGATTATAMCA~TTGACMTTTTGC~TGCCGTCCCCGACTCCTGATGCCA

241 CAGCATTGACMGGATCACTAG~GGAGCTGACATTACATTAAAAAGTCTGCATCCGTCCA-TTTCATATTTTTCCTCAGTATCGTATCTTCAAT~TTTTCCGACMCCTGA

361 M~GG-TTMTMGTTATA~CATATTMGGACACA~TAG-CCAGACAGGCAAACTMCAAATACAMTATGTGACTCCATCCTGCTGACGACACACM

481 GTAMTCCTTCMCGCMCCGCAACAG~CGGGCCTTGC~GC-C-TCGCCAMTTTTGCGATTATAMCA~TT~CAATTTTACMTGCCGCCTGCACCTCC

601 TGTTGCCACTGCATCAACMGGATCAGTAGCGCG~GCTGATGCCACTTT~GCT~TCCGTCCA~TGTATACTCTTCCTCAGTATCATACTGTCMCGMCTTCCACT

721 MCCTGCAAT~GGAAMT~TMGTATATGATAC~--TGACMTCT-TTAGCAAACCTGACAGGCATACA~CATACTAGCAGATGCTMTATG~CTCCATC

841 CTGCTGACGAGAACCACGCAACTCCTTTCTCC~CCGCAAATACTGAMCMGGMGCA~GCTAAAACTG~TTATTTATTTCMC-TACCTATCTMTTGCCMTTCGAC

961 GACTCATATCCGCGGCACACTGGCGGCGATGGCCCATAAATAAAAGGCCTCCTMTAAATTA~CGTACCT~CC~TTMCGCMCCAT~CAAAACAATCATMC Hlndlll 1081 ACTCACCTTCAGAT~CCACCTMGTACCT~TACAAAAACATTMTTMCGCAATTM~T-TMTACMGTATMCACTTACCTC~GCTMTGTACCT~C

1201 AAAAACAAAACAAAAATTMTT~TAMTAAAT~TAAATACAAATATMTACTTACCTCCMTTTnCCTCCCAGCCMCGTACCTA~n~TAAAAATTTGATACMTCTT-G

1321 CAAATMCAAATGTMTACTTACCAAATTTT~T~TGTATTCATTTCCATGACCCCAACGCTGCGATTGTCCTCGGCMCMTTCCTGTTCCGGCGGCTCCATGCTGCCMTCCGGACG

1441 CACTGGCCACMGACGCGGCGCTACTGGCMTCC~CGATG

1561 AMCAATACMCGACAAAATTCCCTCTTTCTGACGCCCGGCGMGTGCCCTGGMTTATCMTGTATTMTTG-CATCTACCMTGAGGGCAGAAGAGATACTCACCAMT~CTGC

1681 nnCGCGGGMCAATCTACCTGCMCGCCGGCCGGACATACATGTTGCAAGTGGCGCGCATCCAGCCGCTGTMCATAGCCCCMGTCAAGTACMCMCTACTTACCTGTMCGTCGCCA

1801 GAGGCTCCCAGCGAATCGGTGCTTCCGTCC~CTGGCGGGGGTACCTGAAAA~CAMTTAMCMTATTAATCCTAAATTTCMTGTTTTTTGT~TTMTTTAAATTGTTAMTG

1921 T~CMGCCTTGCMTATGTTMTGTTACCAGTCCATGTTACTGTCT-GCCAAGMTA~TACTMTTATAAACTMCTCACCACGCCCAAGCCCCAAACTCACCCCATG

2041 CMTGTTAAACCTATAAATTC~TMTTGTCCTATATATTGCACATACTGTMTCAAAGGC-TAMTCGT~ATGCGGAACAGAATTTACTCTGTCTCCGTACCTCCACCAGCAAA

4 HeT- 1.8 A ___) 2161

2281 2401 ATTTTGTMGATTATATTCGATAAAAATT~GATTTCCA~GGATMTGTTTT~TATMTTTTTTTTATACACCGGCAAGAACACGATTATTMTATTGTTATTTM?""""""- 2521 """""""""""""""""""""""""""-~~~""""""- TGCGGCAGCAAGTACGGATCAACTCGACCAGGTCAMCTGA TGAAGGATCTTCTTACATTTCC 2641 CTTCTTCAACATATTCTCTMTATCTATACACAC~CGTTCACTACATTTTCATCTCCCTCTMCACACTATTCTTGA~CCCGATCGGACCT~CGCAGCTGCA~GGT~GTTCGTGC """"""""~""""""""""""""""""""""""""""- 2761 CCTCTGCGCAGCGCTCGACAGAATT TACTGAAGGAATCTTCATACATTTCCCTTCTTCAACATATTCTCTMTATCTATACACACTGGCGTTCACTACATTTTCATCTCCCTCTMCAC """"""_ qrr;rTT1"""""""""""""" ~"""""""""""" 2881 ACTATTCTTGACMCCCGATCG~CCTCACGCAGCTGCAGAGGTTGTTCGTGCCTCTGCGCAGCGCTCGACAGMTTTTACTGAAGGAATCTTCTTACATTTCCCTTCTTCAACATATTC ...... 0 3001 TCTM TCTATACACACTGGCGTTCACTACATTGTCATCTCACTCTMCACACTATTCATGACMCCCGATCGCTGCAGAGGCTGTTCGTGCCTCTGCCGCAGAGCTCGMCG~TTTT EcoRl 3121 TGTGTACGCATCGTTACATTTTTGCAATGCTTTTACACCA

3241

3361

3481

3601

3721

3841

3961

4081 TTGGAAATTGAGTGGCCATTAAAAGACA TTATTGAAATTCGACTACTATT-TT-TTGATMTTMG~TAACMCATATTTCATGGM 4 1.8 C ___) 4201 AAAAGAATAAAGACTCATATGTACMCTTTGCC~TTAGGTTCGGCACTGTGTTMTTTTTGTTGGCTGATTTTGTAAGATTATATTCGAT~TTGAAAARAGATTTC~GMG

4321 GATMTGTTTT~TATMTTTTTTTTATACACCGGCTTGTGTCTCATCCATTTCGCGTATTCAGTTTG

4441 """""""""""""""P~~"""""""""""""" p.. - - - - - 4561 ACTACATTTTCATCTCCCTCTMCACACTATTCTTGACMCCCGATCGGACCTCACGCAGCTGCAGAGGTTGTTCGTGCCTCTGCCGCAGCGCTCGACAGMTTTTACTGMGGMTCTT """"""""""-~"""""""""""""""""""" +==+- - - - 4681 CATACATTTCCCTTCTTCAACATATTCTCTAA TCTATACACACTGGCGTTCACTACATTTTCATCTCCCTCTMCA~CTATTCTTGA~CCCGATCGGACCTCACGCAGCTGCAGA """"""""""""[n----""""""""""""""""""""""""" 4801 GGTTGTTCGTGCCTCTGCCGCAGCGCTCGACAGAATTTTACTW\AGGAATCTTCTTACATTTCCCTTCTT~CATATTCTCTMTATCTATACACACTGGCGTTCACTA~TTGTACTC """".4.""""""""""""""""""" 0 4921 TCACTCTMCACACTA TCATGACAACCCGATCGCTGCAGAGGCTGTTCGTGCCTCTGCCGCAGCGCTC~CGAATTTTACCGCMTATTACMTTTTATAA~TGTT~TTACAGA EcoRl 5041 GMTTCCTAW\ATGCGAACAGTCACGCTGCGACATATGTGTACGCATCGTTACATTTTTGCAATGCTTTTACAC~CAGAG~T~TATAMTATTTATATATT

5161 TTATGCAAAARATATMTMTAAATATAAAAATATMTTTCCTTTATACACCGGCAAGMC-CGATTATTMTAT~CCCAGCAAATTTATG~T~CAACATTTGCAC

5281

5401 Dpll87 Subtelomeric Heterochromatin 7 47

5521 GTGGCCGCGTTATCGATGTGCGCGCGATAT~TATTGCTCMCCTTTAGTGTCTCTGATATGGCGGCGGGATTGAGAGTGACCGTACCGCATTTTCAACMGATTGTGCCCTTGCTCTA

5881 AAMTGGCCGCCGCCUJIOPGA~GCAATATCTAGGCCACTCCTCTCTCT~AGGGCGATTGGAMTTGAGTGGCCATTAAMGA~

6001 TTCGACTACTATTAAMTTAAMTTGATMTTM~TMCMCATATTTCAT~~TAAAGACTCATATGTACMCTTTGC~TTAGGTTCGGCACTGTGT~TTT

6121 TTGTTGGCTGATTTTGTMGATTATATTCGAT~TT~GATTTCCAGAAGGATMTGTTTTAAAAATATMTTTTTTTTATACACCGGCAAGAC~CGMTMTMTATT 0""- 6241 """""""""""-GTTATTTMTATATTATTMTATMTTGTGTCTCATCCATTTCGCGTATTCAGTTTGTMTTTGGTGCGGCAGC~GCAAGTACCGATCAACACGACCAGT~CTTGACTGMGGGA ~"-~"""""""""""""""""" "- 6361 """"""""""""-TCTTCTTACATTTCCCTTCTTCAACATATTCTCTM TCTATACACACTGGCGTTCACTACATTTTCATCTCC~CTMCACACTATTCTTGACAACCCGATCGGACCTCACGCAGCTC"""""""""_ 6481 TCTATACACACTGGCGTTCACTACATTTTC """"""""_ "-v-~"""""""""""""" W""""""" 6601

6721 """"_ Q EcoRl 6841 CGCTCGAACGAATTTTACCG~TATTACAATTTTATM~TGTTAAATTACA~G~CTAGMTGC~CAGTCACGCTGCGACATATGTGTACGCATCGTTACATTTTTGCA

6961 ATGCTTTTACACCAAAMCAG~T~TAT~TATTTATATATTATTAT~TATMTMT~TATAAAAATATMTTTCCTTTATACACCGGCMGAACA

7081 ~CGATTATTMTAT~CCCAG~TTTATGGAT~CAACATTTGCAGCAACATATGT~TTTCGTAGTMCTAGATMTACATTTTATTTATAT~TATAGATACGC

7201 GCATACGGAACTGGOCTAMTTCGCGTGCAATATGACGCACMGC~TTTCAGTCCACGCGCAGTTTACCATAGTAGCTCT~GC~CGACTGCCTCTA~GACTGAGCTATCA~* 7321 TGTAGCGGCGGCGCTTAT PTAGTGCCCGMCAATCGATAGTGACTGTCGCTACTGGTTATCGTGGCCGCGTTATCGATGTGCGCGCGATATCGATATTGCTCMCCTTTAGTGT~CTGA 7441

7561

7681 0.9 A d 7801 GCGTATGCGAGAGGAGTGACMTATMTGTACTCTTC~TTTTTTGGCMTCCAAMTAGGAGA~CCAGnTTTCCTCTTTG~CCATTTTTGAnTTAAAAATG~G~TM 7921 TGTGGAGCMTAAAMTACCTTCATAGCTGTTATCATTCACCCTTTCAGCTGn~TAGGAACAGATTACAGTTTTTAAAAATTTGTCTCATTTTATTCCCCMTAT~cTCATCAT EcoRl 8041 GGCAACTGTTGACGAGGGCTG~~GGGGCAACGGT~TGTTTCATC~TGGACGCT~CAGGCTT~TAT-G"CTGCCTCTCATTCTCTGTCTT~TTACC 81 61 GCAAATCCAACATGACAATGCT CAGATATTTAMTTGCCTCTCGTTTTCTCTCCCTTATATAGGGACClUMTGATCGCGTATnGAnGAGTGCCAACATATTGTACTCTTAGATTTTm 8281 TTnGCAACCGAAAATAGCGGCGGTCGAAGTGAGACCMTATCTAGGCCACTCCTCCCTCT CGACCACTTTGGGCCTACCTATTTTTTTTnCCTCCGCCATTTTTAGATGCGGACC 8401 \ 8521 CAAAMTAGCAACAGATT~TTTTT~TTTGTTTMTTTGTTCTCATTTTTCCCCCGATATCTGTCGATATGCT~~TAGTAGTAATTTCGGTTCGCATTTCCGCTTGCTGA 0.9 6 d +0.9 A ,-W e)r 8641 CTAAATCCTTTTCTATTTTTTTCCCMTTTTATTACCAGTATTTCAGAGAGGAGTGACATATACTGTACTCTTCGATTTTTTGGCAATCCMGATA~GA~CCGnTTTCTTCTTTG

8761 bACCCCArrTTTC EcoRl 8881 CCATCTCATTTTATTCCCCAATATTCCTCGATATGGCA~C~GTATAGMTTTCGCGTTnGCATC~~AGCAGCTGTTGACGAGGGCTGGA~GGGGCMCGGTMGTGTTT

9001 CATCGAAAAATGGACGCTCCAGGCTTGMTAT~GACTGCCTCTCATTCTCTGTCTT~TTACCGCMnCCCMCAAGACMTGCACGACAGATATTAAMTTGccTcTcGT

9121

9241

9361

9481

9601 MTTTAAAAAAGTTAAMGTT~GTTAAMTTTACCTCCTGACTTGCCAGnG~ATTCC~T~GCGGGTTTCATTTTTATGCTCCC~CGCTGCTACAAAGTT~~CTTTGACT

9721 TCGCCAMGAClUMGGnAAGCGATTTT~TCTACTM~TMTTGCTGGGACGTGATATTG~GGGAC~GCTMTT~CAATCMTT~GGCCTTATTGTTAAMGT~ffi Xbal 9841 AAAGTTTTMGCMTTAAAGCGMCGTCTAGA FIGURE7.-DNA sequence and structural features of theSpeI-XbaI hotspot region. The 9872 bpDNA sequence starts with the sequences closest to the distal telomere (the SpeI site, positions 1-6; see Figure 5 for the position of this region within Dp1187), and ends with the centrotnere-proximal XbaI site (positions 9867-9872). Structural features of this region are noted above eachline. These include restriction sites (Spel, Hindlll, BcoRI and XbaI) and major repeats (HeT, 1.8A-D; 0.9A and B). Subrepeats located within the major repeats are designated by double lines (80-bp repeat), dashed lines/open arrowheads (1 73/173/161-bp tandem repeats), solid lines/solid arrowheads (160-bp repeat) and jagged lines/half arrowheads (167-bp repeat). Boxes with diagonal lines are 12-bp insertions that constitute the only difference between the 173- and 161-bp repeats. // indicates subrepeats that are truncated at their 5' and/or 3' ends. Flags indicate the positions of the sequenced PZ insertions, and the flag direction denotes the transposon's orientation (tip = 5' P end; see Figure 5). 1803 is inserted 9' to the XbaI site, and is not included in this sequence. Numbers inside flags represent the number of independent insertions at that site, ifgreater than one. See Figure5 and Table 2 for insertion line designations. 748 G. H. Karpen and A. C. Spradling

TCACTAC

B)

GCGAACGACTGCCTCTAGCk ATCAGTTGTAG FIGURE8.-Sequence-specific PZ insertion within the 1.8-kb re- peats. Homologous subrepeats from the 1.8-kb elements were aligned to identify sequence-specific insertional hotspots not appar- ent from the dispersion of PZ insertions throughout 1.8B. C and 1). Only three sequences, which correspond to the three clusters noted in Figure 7, were found to contain more than one insertion when summed in this manner; two (A) were located 9 bp apart in the 173/161-bp repeats and one(B) in the 3' portion of the 1.8-kb FIGURE9.-ln situ hybridization of 1.8-kb repeat sequences to repeat. Dp line numbers are next to flags that indicate the orienta- Canton-S salivary gland polytene chromosomes. The distribution of tion of each insertion (see Figures 5 and 7). Boxed sequences 1.8-kb element sequences within the genome was determined to be indicate the 8-bp genomic DNA duplicated upon P element inser- primarily telomeric (2R and 3R, large solid arrows)and centromeric tion (O'HAREand RURIN1983). Stippled box highlights a 7/8-bp heterochromatin(chromocenter, bracketed area). One basal eu- Inatch to a P element insertion hotspot identified at thesinged locus chromatic site was observed (small arrow). Ectopic fibers joining (KOIHA, RURINand O'HARE1988; HAWLEYet al. 1988). chromosomal telomeres (open arrows) were frequently labeled. Sometimes (A) only one of the two tips contained significant amounts of the 1.8-kb sequences, whereas in other cases (B) the ductions were due to position effects, since transpo- 1.8-kb element was well represented on both tips (2R and 3R). sons remobilized to other locations subsequently ex- pressed ry+ in a normal, stable manner(J. TOWER, G. cloned, repeatedDNAs from the hotspot regioncould KARPEN,N. CRAIG andA. SPRADLING, unpublished). not be used as probes in these experiments, since it Genes juxtaposed with heterochromatic regionsfre- was difficult to recognize the bands specific for the quently display variegated position effects [reviewed duplication amongthe many complementaryfrag- by SPOFFORD(1 976), HENIKOFF (1990) and SPRA- ments on a Southern blot. This is a general problem DLING and KARPEN (1990)l. We tested whether ry+ encountered when analyzing the copy number of sub- expression from hotspot insertionswas affected by the telomeric repeats, since all characterized to date in dosage of the Y chromosome, a strong modifier of Drosophila are present in both centromeric and sub- position-effect variegation. All the hotspot insertion telomericregions, andcannot be distinguished by lines tested failed to express ry+ in X/O males, but restrictionfragment polymorphisms. However, expressed fully wild-type eye color in X"Y/Y males. unique sequences within transposons inserted in the Thus, theposition effect on ry+ expression responded DpZZ87 hotspot provided a tag to specifically follow as expectedfor a variegated position effect. Previ- the behavior of the minichromosome telomere. ously, the R401.1 insertion within repetitive se- DNA from Dp34OZ males with X/O and X/Y sex quences on the 4th chromosome was shown to exhibit chromosome constitutionswas examined to determine variegated ry+ expression with similar characteristics if telomeric sequences are underrepresented in poly- (SPRADLINCand RUBIN 1983;DANIELS et al. 1986; A. tene nuclei, and if DNA copy number in telomeric SPRADLINGand D. THOMPSON,unpublished). regions is sensitive to a strong modifier of position- The Dpll87 telomere becomes underrepresented effect variegation. High molecular weight DNA iso- in the DNA of polytene cells: Centromeric hetero- lated from late third instar larval salivary glands and chromatin is severely underrepresented in chromo- imaginal discs was digested with BgZII or XbaI, sepa- somes of many polytene tissues (reviewed by SPRA- rated on a pulsed-field gel, and hybridized with a ZacZ DLING andORR-WEAVER 1987). Cytological studies probe specific for the insertion (Figure 10). The 48- have led to suggestions that telomeric regions also kb XbaI fragment that extended all the way from the may be underrepresented (ROBERTS1979). To deter- hotspot region to the telomerewas strongly underrep- mine if the relative copy number of the DpZ 187telom- resented in X/O male salivary gland polytenized DNA, ere decreased as a result of salivary gland polyteniza- compared to thedisc diploid DNA (1 3-fold, Table 3). tion, we compared larval salivary gland DNA to DNA The amount of underrepresentation of the XbaI frag- from predominantly diploid imaginal disc tissue. Sub- ment was reduced to 2.6-fold when a Y chromosome Subtelomeric HeterochromatinDpl187 Subtelomeric 749

RglII XbaI TABLE 3 D SG D SG -kb "" Underrepresentation of Dp1187 subtelomeric DNA in salivary gland polytene nuclei

48 - Fold-underrepresented' 23 Genotvpe RglII fragmentb Xbal fragmentb

X,y/O;Dp 340I ,y+ ND I3

X,y/Y;Dp 340I J+ I .6 2.6 Xbal RglllRglll Xbal ' lues indicate the amount of underrepresentation of Dpl187 -*1"2b750-1 /H#Okb +1000 subtelomeric DNA in polytene salivary gland nuclei (folddifference in the amount of DNA), when compared to the Same region in the Dp 3401 preclon~in;~ntlydiploid imaginal discs(see MATERIALS AND METHODS. and Figure IO). The values include normalization for differences FIGURE IO.-Underrepresentation of the subtelomeric region in in the total anlount of DNA loaded in the lanes, determined by sillivary gl;uld DNA. Pulsed-field Southern hybridization analysis of reprobing with a ry+ probe and measuring the representation of the y/Y ;ry "" ;Dp3401,y+ ry+ larval imaginal disc (D) diploid cells and endogenous ry+ fragment (see MATERIALS AND METHODS). salivary gland (SG)polytene nuclei was performed as described in See Figures 6.4 and IO for thesize and position of the restriction MATERIALS AND METHODS. The structureof the subtelomeric region fragments. ND = not determined. in this line is shown helow. Symbols are the Sameas in previous figures. The autoradiogram shown was produced by probing with Notch (KELLEY et al. 1987). In these cases, however,

lari! sequences (open box). Reprobing the Same blots with ry+ insertions were spread over only a few hundred base sequences (5' ryffIi3.2)showed that equivalent amounts of DNA pairs, significantly smaller than the 4.7-kb hotspot were loaded in the D and SG lanes for each digest (data notshown). near the Dpll87 telomere. Some structural or func- Note that the underrepresentation of the terminal X6aI fragment tional property of the hotspot region, for example in polytene DNA is greater than for the subterminal BgLlI fragment (see Table 3). transcriptional activity or chromatin structure in male germ-line cells, might explain the generally high fre- was present. However, the 23-kb BgZII fragment con- quency of insertion, and the apparent preference for taining sequences from the transposoninsertion to the proximal 1.8 kb repeat. However, some role for position -267 was underrepresented only 1.6-fold in sequence specificity in site choice within regional hot- salivary gland DNA compared to diploid DNA. Both spots is suggested by the striking homology between restriction fragments migrated more heterogeneously one of the Dp1187 multiple insertion sites and the when prepared from salivary gland DNA than diploid singed hotspot (ROIHA, RURIN and O'HARE1988; DNA, possibly as a consequence of anelimination HAWLEYet al. 1988). mechanism that was proposed to be responsible for The screen employed would have detected any in- sequenceunderrepresentation (KARPEN and SPRA- sertion on Dpll87 that was still capable of expressing DLING 1990; R. GLASER, G. KARPEN and A. SPRA- the ry+ gene, but nonewere recovered in centromeric DLING, unpublished). The reduced intensity and het- heterochromatin or other subtelomeric regions. The erogeneous migration were not simply the result of simple sequences that constitute much of centromeric nonspecific DNA degradation, since restriction frag- heterochromatinmight lack P elementtarget se- ments of similar or greatersize that lay within euchro- quences. However, the identification of an island of matin (the endogenous rosy locus) were unaffected. complex sequences within Dpl187 heterochromatin makes it more likely that potential P target sequences DISCUSSION reside in the centromeric heterochromatin. Since all the hotspot insertions tested suffered strong position A P element hotspot within subtelomeric, repeti- effects on ry+, and similar reductions were seen with tive DNA: Our results demonstrated that P element insertionson the Y chromosome (BERGand SPRA- insertionscompatible with markergene expression DLING 199 1; G. KARPEN and A. SPRADLINGunpub- occurred selectively but frequently within repetitive, lished), 4thchromosome (SPRADLINGand RURIN subtelomeric DNA. The fraction of insertions intothe 1983), and near the 3R telomere (HAZELRIGG, LEVIS Dp1187 hotspot, 39/7,825 (0.5%), was several fold and RUBIN 1984; LEVIS, HAZELRIGGand RUBIN greater than any other hotspot observedin our screen, 1985), marker geneexpression is likely to be a limiting such as lethal stripe insertions (19/7,825). The inser- factor. Our results suggest that this problem can be tion frequency into the XbaI-SpeI fragment appears mitigated by scoring insertions in the presence of an similar to the average frequency inferred for individ- extra Y chromosome, or in a genotypecontaining ual elements from data on previously described hot- suppressors of variegation. spots such as singed (ENGELS 1989). The recovery of Our choice of the Dp1187 minichromosome and insertions at more than one site within a hotspot was insertional mutagenesis greatly assisted the subse- reported previously for insertions at singed (ROIHA, quent structural analysis of the repetitive, subtelom- RURINand O'HARE1988; HAWLEYet al. 1988) and ericregion. Direct cloning of repetitive DNA into 750 G. andH. Karpen A. C. Spradling plasmids may have helped avoid sequencere- the small loss of DNA due to the priming problem. arrangements. Equally valuable was the ability to map The absence of a strict mechanism to balance the rate each insertion site in the genomic DNA using pulsed of addition and loss would provide an explanation for field and standard gel electrophoresis. It would have both the existence and evolutionary instability of te- been much more difficult to correctly position the lomeric heterochromatin. The structure of the SpeI- transposon insertion sites, and to delineate the struc- XbaI fragment provided some support for this model. ture of each 1.8 kb repeat, without this information. The HeT element was linked with the first 1.8-kb The introduced single-copy sequences within the tran- repeat via a short stretch of A residues, like the HeT sposon allowed the copy number of the tip region to addition events observedby BIESSMANNet al. (1 990). be studied in polytene cells. Furthermore, the DNA Possibly, at some previous time the telomere of the X in this region can now be manipulated genetically by chromosome progenitor of Dpll87 terminated in 1.8- remobilizing the insertions u. TOWER,G. KARPEN,N. kb repeats, when this HeT was added. Subsequently, CRAIGand A. SPRADLING,unpublished). Manyof the equilibrium may have favored addition events so these advantages would hold for P insertions within that 40 kb of additional sequences accumulated. Se- other heterochromatic sites. quencing of DNA lying distal to the SpeI site would Structure and evolution of subtelomeric hetero- test some of the predictions of this model. chromatin: The subtelomeric region of Dpll87 that The frequent insertion of a transposable element we analyzed displayed several properties characteristic into repetitive target sequences, as observed in our of heterochromatin. First, sequences from this region studies, provides an additional mechanism that could were repetitive, butwere found primarily at telomeres contribute to the maintenance andevolution of telom- and in centromeric regions. Second, no known or eres, as well as other heterochromaticregions not essential genes were present in the 40-kb region distal susceptible to terminal additions. The acquisition of to the3601 insertion. Third,the euchromatic ry+ 1.8-kb elements in a subtelomeric region might be gene suffered variegated position effects upon inser- followed closely by the accumulation of P elements. If tion within the Dp1187 subtelomeric region. Finally, transposons themselves served as targetsfor other the copy number of restriction fragments from the transposable elements, once startedthis process would subtelomericregion was underrepresented in DNA lead to the rapid growth in the size and complexity of from polytene salivary gland cells, relative to euchro- this region. The structure of the junctions between matic fragments. All these properties are shared by members of the 1.8- and 0.9-kb repeatsdo not provide centromeric heterochromatin in Drosophila chromo- sufficient information to support or discount this somes. model. Further studies of the structure and behavior Sequence analysis of the 10-kb subtelomeric XbaI- of these elements are needed to resolve these ques- SpeI fragment revealed a high degree of sequence tions. However, some other mechanism(s) may have structure. Distal HeT sequences were joined to tan- generated the tandem repetitionof major (ie., 1.8A- dem repeats of the 1.8-kb element, followed by im- D) and minor (ie., 173/173/161-bp) repeats. Perhaps perfect tandem repeats of the 0.9-kb sequence. HeT the poorly understood mechanism that is involved in elements are located near other Drosophila telomeres maintaining tandem repetitions within other hetero- (RUBIN 1978;YOUNG et al. 1983), and some of these chromatic regions, including satellite DNA and ribo- chromosome tips also contain 1.8-kb repeats (AJIOKA somal DNA repeats, also acts on these sequences. 1987).A P elementinsertion in the subtelomeric Sequence underrepresentation of telomeric het- region of chromosome 3R (HAZELRIGG,LEVIS and erochromatin: Our experimentsprovided the first RUBIN 1984)has recently been shown to reside in a direct demonstration that subtelomericsequences be- sequencerelated tothe 1.8-kbelement (R. LEVIS, come underrepresented in a Drosophila polytene cell personal communication).Further analyses of normal type, thesalivary gland. The existence of copynumber and newly generatedtelomeres will beneeded to changes in telomere regions has previously been pos- determine if theorganization of HeT, 1.8kb and tulated on cytological grounds (ROBERTS1979). Both possibly other repeatedsequences is conserved in Dro- heterochromaticsequence underrepresentation and sophila, like the X and Y’ elements of yeast. copy number decreases within “constrictions” in po- The addition and loss of DNA sequences may be a lytene chromosomes have invariably been assumed to regular feature of subtelomeric chromosome regions. result from differential DNA replication (reviewed in BIESSMANNet al. (1 990) recently demonstrated that SPRADLINGand ORR-WEAVER 1987). However, recent HeT elements can be added onto terminally deleted studies of Dp1187 led us to propose that a process of chromosomeends. R. LEVIS, R. GANESAN,K. DNA elimination similar to that observed in some HOUTCHENSand F. SHEEN(in preparation) have pro- nematodes and in ciliate macronuclei was actually posed that sequence additionsto normal chromosome responsible for the decreases in heterochromatic se- termini by retrotranspositionmight counterbalance quences in polyploid and polytene cells (KARPENand Of1187Heterochromatin Subtelomeric 75 1 SPRADLING1990; SPRADLING and KARPEN 1990; in the case of the hotspotinsertions, the v+gene SPRADLINGet al. 1992). Telomeresof nonhomologous within PZ was reduced less than twofold, while ry+ polytene chromosomes display a strong tendency to expression must have been severely reduced, perhaps be joined together by thin strands termed “ectopic as much as 1 00-fold,to explain the effect on eye color fibers”(HINTON and ATWOOD1941). The labeling (CHOVNICK, GELBARTand MCCARRON1977). Conse- along ectopic fibers spanning the two 1.8-kb-rich te- quently, the heterochromatic position effects on these lomeres that was observed in polytene chromosomes subtelomeric insertions predominantly suppress gene can readily be explained by the elimination model. expression, rather than DNA copy number. Religation of individual DNA strands following elim- Hotspot insertion andP element regulation: Pre- ination of more distal heterochromatin at both tips vious studies revealed that some wild strains of D. couldjoin strands from separate chromosomes. melanogaster frequentlycarry P element inserts in Stretching of such covalently joined chromatids dur- region 1A (AJIOKAand EANES1989). The chromo- ing chromosome preparation would produce ectopic somal location of theseinsertions are likely to lie strands rich in the 1 .8-kb repeats.Other subtelomeric within the hotspot sequencewe analyzed, since several heterochromatic sequences, such as HeT repeats, also were mapped to a “1.9”-kbEcoRI repeat that hybrid- would be expected to become located in some ectopic ized to the tips of2R and 3R (AJIOKA1987). The fibers (depending on thesite of breakage and ligation), identity of these sites remains to be directly verified. and labeling of ectopic fibers with HeT probes has Nonetheless, the presence of an insertion hotspot in been observed previously (RUBIN 1978;YOUNG et al. our experimentsdiffers from the observed lackof 1983). Another puzzling attribute of polytene chro- elevated P insertion into region 1A in a chromosome mosomes, the occasional linkage of telomeric and tested by these authors (AJIOKAand EANES 1989;also chromocentral regions via ectopic fibers, also can be see RONSSERAY,LEHMANN and ANXOLABEHERE explained if the elimination of sequences in both 1991). The existence of such ahotspot may vary heterochromatic regions is occasionally followed by between strains. The fact that 1.9-kb repeatswere not centromere-telomere, rather than telomere-telomere observed on all X chromosome tips examined might or centromere-centromere, religation. provide the basis for these differences(AJIOKA 1987). Underrepresentation of DNA in telomeric regions Recent studies suggested that P insertions within can also explain the cytological appearance of polytene the 1A hotspot might exhibit an unusual ability to chromosome tips. The morphologies of X chromo- repress P element movement. Intact P elements pro- some tips are altered significantly by the addition or duce a 66-kD proteinthat can repress P element removal of Y chromosomes (SCHULTZ1947). These activity (MISRAand RIO 1990), but only low levels of morphological differences may be caused by altera- repression that varied depending on the insertionsite tions in telomeric sequence representation, which we were associated with any single element, even if it had observed to be influenced in trans by the number of been engineered to produce large amounts of 66-kD Y chromosomes. In addition, the morphology of te- protein [reviewed by RIO (1990)l. A strain derived lomeres has been observed to vary among salivary from the wild in which all the P elements were re- gland cells from the same larva (DOBZHANSKY1944; moved except those located at position 1A appeared LEFEVRE 1976;ROBERTS 1979). Recent studies show to be an exception (RONSSERAY,LEHMANN and ANX- that the level of heterochromatin-associated underre- OLABEHERE 1991).Although containing only two presentation differs among polytene nuclei (KARPEN complete P elements, it strongly repressed germ-line and SPRADLING1990); thus, variations in tipmor- P element activity, and repression was maternally phology could result from cell-to-cell differences in transmitted to progeny (RONSSERAY,LEHMANN and the underrepresentation process. ANXOLABEHERE 199P insertions 1). within the hotspot Sequence underrepresentation was observed in the may be subject to position effects that enhance their subtelomeric region of Dp1187 DNA isolated from effectiveness as regulators of transposition. salivary glands; a terminal 48-kb fragment was under- The preferentialinsertion of P elementsinto a represented more than a subterminal 27-kb fragment. heterochromatic site where they may play a role in A copy number gradient in salivary gland DNA was regulating P element activity has potential implica- found previously within the first 103 kb of Dp1187 tions for the evolutionary role of transposable ele- euchromatin that resides adjacent to centromeric het- ments. Previously, it has been argued that Drosophila erochromatin (KARPENand SPRADLING1990). Genes transposons are unlikely to confer a selective advan- within both regions were subject to variegated posi- tage because individual sites detectable by in situ hy- tion effects. We argued previously that the reductions bridization are usually occupied only at low frequency in DNA copy number might be sufficient to explain within a population (see AJIOKAand HARTL 1989). the reduction in yf expression in variegatingchro- However, many transposable element family members mosomes (KARPENand SPRADLING 1990). However, are located in centromeric heterochromatin, where 752 G. andH. Karpen A. C. Spradling they frequently occupy conserved sites (BUCHETONet FINNEGAN, 1984The molecular basis of IR hybrid dysgenesis al. 1984; SHEVELYOV,BALAKIREVA and GVOZDEV in D. melanogaster: identification, cloning and properties of the I factor. Cell 38: 153- 163. 1989). Transposons may be able to efficiently inte- CAVALIER-SMITH,T., 1974 Palindromic base sequences and rep- grateinto certainheterochromatic regions, where lication of eukaryote chromosome ends. Nature250 467-470. they can effectively regulatethe activity and copy CHAN,C. S. M., and B.-K. TYE,1983 Organization ofDNA number of otherelement copies, an ability that is sequences and replication origins at yeast telomeres. Cell 33: likely to be of selective value. 563-573. CHOVNICK, A., W. GELBARTand M. MCCARRON, We thankthe following peoplefor their participation in the 1977Organization of the rosy locus in Drosophilamelano- mutagenesis screen, carried outin the SPRADLING laboratoryin the gaster. Cell 11: 1-10, sunimer of 1989: C. BERG,L. COOLEY,R. GLASER,B. HARKINS,M. COOKE, H.J., W. R. A.BROWN and G. A. RAPPOLD, HECK,G. KARPEN,M. KUHN,L. LEE, D. MCKEARIN,C. MONTELL, 1985 Hypervariable telomeric sequences from the humansex D. MONTELL,T. OYEBODE,J. RIESGO,D. SOMMERVILLE,A. SPRA- chromosomes are pseudoautosomal. Nature 317: 687. DLING, D. STERN,D. THOMPSON,J. TOWER, E. VERHEYEN,S. COOLEY,L., R. KELLEYand A.C. SPRADLING,1988 Insertional WASSERMAN, AND L. YUE. We are grateful to YASH HIROMIfor mutagenesis of the Drosophila genome with single P elements. sending the strain with a PZ insertion on theX. DIANNETHOMPSON Science 239 1121-1 128. provided skillful technicalassistance in DNA sequencing.G.K. CORCORAN,L. M.,J. K. THOMPSON,D. WALLIKER and D. J. KEMP, thanks L. COOLEY,R. GLASER,D. MCKEARIN,M. MCKEOWN,D. 1988 Homologous recombination within subtelomeric repeat MONTELL and J. TOWERfor helpful discussions andsupport sequencesgenerates chromosome size polymorphisms in P. throughout the course of these studies. Special thanks to R. LEVIS falciparum. Cell 53: 807-813. for providing critical comments on the manuscript, helpful discus- DANIELS,S. B., M. MCCARRON,C. LOVE, S. H.CLARK and A. sions, and sequence data prior to publication. This work was sup- CHOVNICK, 1986The underlying bases of gene expression in ported by the Howard Hughes Medical Institute and U.S. Public stable transformants of the rosy locus in Drosophila melanogaster. Health Service grant GM27875. G.K. was supported by a National Genetics 113: 265-285. Research Service Award of the U.S. Public Health Service, and by DELANGE, T.,L. SHIUE,R. M. MYERS,D. R. Cox, S. L. NAYLOR, the Howard Hughes Medical Institute. A. M. KILLERYand H. E. VARMUS, 1990 Structure and vari- ability of human chromosome ends. Mol. Cell. Biol. 10: 518- LITERATURE CITED 527. DEVEREAUX,J., P. HAEBERLIand 0. SMITHIES,1984 A compre- AJIOKA,J. W., 1987 The molecular natural historyof the P hensive set ofsequence analysis programs for theVAX. Nucleic element family in Drosophilamelanogaster. Ph.D.Thesis, Acids Res. 12: 387-395. SUNY, Stony Brook, N.Y. DOBZHANSKY,T., 1944 Distribution ofheterochromatin in the AJIOKA,J. W., and W. F. EANES,1989 The accumulation of P chromosomes of Drosophila pallidipennis. Am. Nat. 78: 193- elements on the tip of the X chromosome in populations of 213. Drosophila melanogaster. Gen. Res. 53: 1-6. ENGELS,W. R., 1989 P elements in Drosophila, pp. 437-484 in AJIOKA,J. W., and D. L. HARTL,1989 Population dynamics of Mobile DNA, edited by D. BERGand M. HOWE.ASM Publica- transposable elements, pp. 939-58 in Mobile DNA, edited by tions, Washington, D.C. D. BERGand M. HOWE.ASM Publications, Washington, D.C. FLEMING, R.J., S. M. DESIMONEand K. WHITE, 1989 Molecular ASHBURNER,M., 1990 Drosophila:A Laboratory Handbook. Cold isolation and analysis of theerect wing locus in Drosophila Spring Harbor Laboratory, Cold Spring Harbor, N.Y. melanogaster. Mol. Cell. Biol. 9 719-725. BEDBROOK,J. R., J.JONES, M. O’DELL,R. D. THOMPSONand R. B. GREIDER,C. W., and E. H. BLACKBURN,1985 Identification of a FLAVELL, 1980 A molecular description of telomeric hetero- specific telomere terminal transferase activity in Tetrahymena chromatin in Secale species. Cell 19: 545-560. extracts. Cell 43: 405-413. BENDER, W.,P. SPIERERand D. S. HOGNESS,1983 Chromosome HAWLEY, R.S., R. A. STEUBER, C.H. MARCUS,R. SOHN,D. M. walking and jumping toisolate DNA from theAce and rosy loci BARONAS,M. L. CAMERON,A. E. ZITRON and J. W. CHASE, and the bithorax complex in Drosophilamelanogaster. J. Mol. 1988 Molecular analysis of an unstable P element insertion Biol. 168: 17-33. at the singed locus of Drosophilamelanogaster: evidence for BERG,C. A., and A. C. SPRADLING, 1991 Studies on the rate and intracistronic transposition of a P element. Genetics 119: 85- site-specificity of P element transposition. Genetics 127: 5 15- 94. 524. HAZELRIGG,T., R. LEVIS^^^ G.M. RUBIN, 1984 Transformation BIESSMANN, H., S. B. CARTER and J. M. MASON, of white locus DNA in Drosophila: dosage compensation, zeste 1990 Chromosome ends without telomeric DNAsequences. interaction, and position effects. Cell 36 469-481. Proc. Natl. Acad. Sci. USA 87: 1758-1761. HENIKOFF,S., 1990 Position-effect variegation after60 years. BIESSMANN, H., and M.J. MASON,1988 Progressive loss of DNA Trends Genet. 6: 422-426. sequences from terminal chromosome deficiencies in Drosoph- HILLIKER,A. J., and R. APPELS,1982 Pleiotropiceffects associ- ila melanogaster. EMBO J. 7: 1081-1086. ated with the deletion of heterochromatin surrounding rDNA BIESSMANN,H., J. M. MASON, K. FERRY,M. D’HULST,K. VALGEIRS- on the X chromosome of Drosophila. Chromosoma 86: 469- DOTTIR, K. L. TRAVERSEand M. L. PARDUE, 1990 Addition 490. oftelomere-associated HeT DNAsequences “heals” broken HINTON,T., and K. C.ATWOOD, 1941Terminal adhesions of chromosome ends in Drosophila. Cell 61: 663-673. salivary gland chromosomes in Drosophila. Proc.Nat. Acad. BLACKBURN,E. H., 1991Structure and function of telomeres. Sci. USA 27: 491-496. Nature 350: 569-573. HSIEH,T., and D. BRUTLAG,1979 Sequence and sequence varia- BROWN,W. R. A., P. J. MACKINNON, A.VILLASANTE, N. SPURR, tion within the 1.688 g/cm5 satellite DNA of Drosophila mela- V. J. BUCKLEand M. J. DOBSON,1990 Structure and poly- nogaster. J. Mol. Biol. 135 465-48 1. morphism of human telomere-associated DNA. Cell 63: 119- KARPEN,G. H., and A. C. SPRADLING, 1990 Reduced DNApo- 132. lytenization of a minichromosome region undergoingposition- BUCHETON,A,, R. PARO,H. M. SANG,A. PELISSONand D. J. effect variegation in Drosophila. Cell 63: 97-107. S ubtelom eric HeterochromatinDpll87 Subtelomeric 753 KELLEY,M. R., S. KIDD, R.L. BERG and M. W. YOUNG, ROIHA,H., G. M. RUBINand K. O'HARE, 1988 P element inser- 1987 Restriction of P element insertions at theNotch locus of tionsand rearrangements at the singed locus of Drosophila Drosophila melanogaster. Mol. Cell. Biol. 7: 1545-1548. melanogaster. Genetics 119: 75-83. LEFEVRE JR.,G., 1976 A photographic representation and inter- RONSSERAY,S., M. LEHMANN andD. ANXOLABEHERE,1991 The pretation of the polytene chromosomes of Drosophila melano- maternally inherited regulation of P elements in Drosophila gaster salivary glands, pp. 31-66, in The Genetics and Biology of melanogaster can be elicited by two P copies at cytological site Drosophila, edited by M. ASHBURNERand E. NOVITISKI.Aca- 1 A on theX chromosome. Genetics129 50 1-5 12. demic Press, London. RUBIN,G. M., 1978 Isolation of a telomeric DNA sequence from LEVINE,J., and A. C. SPRADLING,1985 DNA sequence of a 3.8 Drosophilamelanogaster. ColdSpring Harbor Symp. Quant. kilobase pair region controlling Drosophila chorion gene am- Biol. 42: 104 1- 1046. plification. Chromosoma 92: 136-142. SAIGA,H., and J. E.EDSTROM, 1985 Longtandem arrays of LEVIS, R.,1989 Viable deletions of a telomere from a Drosophila complex repeat units in Chironomus telomeres. EMBO J. 4 chromosome. Cell 58: 791-801. 799-804. LEVIS,R., T. HAZELRIGGand G. M. RUBIN, 1985 Effects of SCHULTZ,J., 1947 The nature of heterochromatin. Cold Spring genomic position on the expression of transduced copies of the Harbor Symp. Quant. Biol. 12: 179-191. white gene of Drosophila. Science229 558-561. SHEVELYOV,M., D. BALAKIREVAand V. A. GVOZDEV, 1989 Heterochromaticregions in different Drosophilamela- LINDSLEY, D. L., and G. G. ZIMM, 1992 The Genome of Drosophila melanogaster. Academic Press, San Diego. nogaster stocks contain similar arrangements of moderate re- peats with inserted copia-like elements (MDGl). Chromosoma LOHE, A. R.,and D. L. BRUTLAG,1986 Multiplicity of satellite DNA sequences in Drosophila melanogaster. Proc. Natl. Acad. 98: 117-122. SPOFFORD,J. B., 1976 Position-effectvariegation in Drosophila, Sci. USA 83: 696-700. pp. 955- IO18, in The Genetics and Biology Drosophila, edited MISRA,S., and D. C. RIO, 1990 Cytotype control of Drosophila P of by M. ASHBURNERand E. NOVITISKI.Academic Press, London. elementtransposition: the 66 kdprotein is arepressor of SPRADLING, A. C., G.and H. KARPEN,1990 Sixty yearsof mystery. transposase activity. Cell 62: 269-284. Genetics 126: 779-784. MLODZIK, M., and HIROMI,1992 The enhancer trap method Y. SPRADLING,A. C., and T. ORR-WEAVER,1987 Regulationof in Drosophila: its application to neurobiology. Methods Neu- DNA replication during Drosophila development. Annu. Rev. rosci. 9: 397-414. Genet. 21: 373-403. MURRAY,A. W., and J. W. SZOSTAK,1983 Construction of artifi- SPRADLING,A. C., and G. M. RUBIN,1983 The effectof chro- cial chromosomes in yeast. Nature 305: 189-193. mosomal positionon the expression of the Drosophila xanthine NEWLON,C. S., L. R. LIPCHITZ, I. COLLINS,A. DESHPANDE,R. J. dehydrogenase gene.Cell 34: 47-57. DEVENISH,R. P. GREEN, H. L. KLEIN,T. G. PALZKILL,R. B. SPRADLING,A. C., G. H. KARPEN,R. GLASERand P.ZHANG, REN,S. SYNN andS. T. WOODY, 1991 Analysis of a circular 1992 Evolutionaryconservation of developmental mecha- derivative of Saccharomyces cereuisiae chromosome a physical ZIZ: nisms: DNAelimination in Drosophila,pp. 39-53, in 50th map and identification and location of ARS elements. Genetics Annual Symposium, Society for Developmental Biology. 129 343-357. TRAVERSE,K. L., and M. L. PARDUE,1989 Studies of He-T DNA ()'HARE,K., and G. M. RUBIN,1983 Structures off' transposable sequences in thepericentric regions of Drosophila chromo- elements and their sites of insertion and excision in the Dro- somes. Chromosoma97: 26 1-27 1. sophila melanogaster genome. Cell 34: 25-35. VALGEIRSD~TTIR,K., K. L. TRAVERSEand M. L.PARDUE, PACE,T., M. PoNZI, E. DOREand C. FRONTAL],1987 Telomeric 1990 HeT DNA: a family of mosaic repeated sequences spe- motifs are present in a highly repetitive element in the Plas- cific forheterochromatin in Drosophilamelanogaster. Proc. modium gerghei genome. Mol. Biochem. Parsitol. 24: 193-202. Natl. Acad. Sci. USA 87: 7998-8002. PIROTTA, V.,1986 Cloning Drosophila genes, pp.83-110 in Dro- WATSON,J. D., 1972 Originof concatameric T7 DNA.Nature sophila:A Practical Approach, edited by D. B. ROBERTS.IRL New Biol. 239 197-201. Press, Oxford. YOUNG, B. S., A. PESSION,K. L. TRAVERSE,C. FRENCH andM. L. RI~,D. C., 1990 Molecular mechanisms regulating Drosophila P PARDUE,1983 Telomereregions in Drosophila share com- element transposition. Annu. Rev. Genet. 24: 543-578. plexDNA sequences with pericentric heterochromatin. Cell ROBERTS,P. A,, 1979 Rapid change of chromomeric and pairing 3485-94. patterns of polytene chromosome tips in D. melanogaster: mi- ZAKIAN, V. A., 1989 Structure and function of telomeres. Annu. gration of polytene-nonpolytene transition zone? Genetics 92: Rev. Genet. 23: 579-604. 861-878. ZAKIAN, V. A., and H.M. BLANTON,1988 Distribution of telom- ROBERTSON,H. M., and W. R. ENGELS,1989 ModifiedPelements ere-associated sequences on natural chromosomes in Saccharo- that mimic the P cytotype in Drosophila melanogaster. Genetics myces cerevisiae. Mol. Cell Biol. 8: 2257-2269. 123: 815-824. Communicating editor: T. SCHUPBACH