Increased Length of Long Terminal Repeats

Copyright 0 1997 hy the Genetics Society of America

Increased Length of Long Terminal Repeats Inhibits Tyl Transposition and Leads to the Formation of Tandem Multimers Vit Lauermann,’ Monika Hermankova and Jef D. Boeke Department qf Molerular Biology and Genetics, Johns Hopkins University School of Medicine, Baltimore, M a y l a n d 21205

Manuscript received April 16, 1996 Accepted for publication December 23, 1996 ABSTRACT The Tyl retrotransposon of Saccharomyces cerevisiae is bounded by long-terminal repeats ( LTRs) . We have constructed a variety of Tyl elementsin which the LTR length has been increased fromthe normal length of 334 bp to > 2 kb. Although small insertions in the LTR have minimal effects on transposition frequency, larger insertions dramatically reduce it. Nevertheless, elements with long LTRs are incorporated into the genome at a low frequency. Most of these rare insertion events represent Tyl tandem (head to tail ) multimers.

T

HE Ty elements of yeast are characterized by longterminal repeats (LTRs) that flank their coding regions and contain critical cis-acting signals.These signals include transcriptional promoters and 3’ end formation signals as well as the sequences that define the tips of the element and are recognized by the transposon-encoded IN protein to mediate insertion of Ty DNA into host target DNA. The Ty elements structurally resemble retroviral proviruses, which are also flanked by LTRs. Furthermore, numerous studies have confirmed that the transposition mechanisms of Tyl-Ty3 and Ty5 functionally resemble the replication / integration processes used by retroviruses. Tyl, the element studied here, produces a transcript whose 5 ‘ end is specified in part by promoter sequences in the 5 ’ LTR ( LIAOet (11. 1987; HIRSCHMAN et al. 1988 ) and whose 3 ’ terminus is specified by 3 ’ LTR sequences (Yu and ELDER1989; HOC et al. 1994). This transcript is packaged into virus-like particles (VLPs) , where it is converted into a linear Tyl DNA copy by theTyl-encoded reverse transcriptase (RT) ( BOEKEet al. 1985; GARFINKEIet al. 1985; MELLOKet al. 1985; ADAMS et nl. 1987; EICHINGER and BOEKE1988). This DNA copy contains intact LTR sequences at both termini. The termini of the Tyl LTRs define the Tyl DNA as a substrate for Tyl IN (integrase) protein (Er(:HINGEK and BOEKE 1990; DEVINEand BOEKE 1994; MOORE and GNINKEI. 1994; MOOREet al. 1995). The five families ofTy elements from Saccharomyces cermzsiue, designated Tyl-Ty5, vary dramatically in primary sequence,but interestingly, thelength of the LTRs in these families is relatively well conserved, from Corrrsponding author:Jef D. Boeke, Department of Molecular Biology and Genetics, Hunterian 617, Johns Hopkins University School of Medicine, 725 N. Wolfe St., Baltimore, MD 21205. E-mail: [email protected] Prfsmt addrpss: Frederick Cancer Researchand Development Center, Frederick, MD 21702.

’

C;enrtich 145: 9Il-WY (April, 1997)

251 (Ty5) to 371 bp (Ty4) (Table 1 ) . Retrotransposons from Schizosaccharomyces, Candida, and Pichia fall in the same range. In nonyeast elements, LTR lengths range from as short as 77 [Bombyx mori mag element ( MICHAILLE et al. 1990) ] to 2136 bp (Drosophila Lrlysseselement (EVGEN’EV et al. 1992) ) . Investigations of Moloney murine leukemia virus have indicated that LTR length can be dramatically increased in this virus with onlymodest effects on replication efficiency (REDDY et al. 1991; FAUSTINELLA et al. 1994). Mie have begun investigations of the effects of altering LTR lengthon Tyl retrotransposition. The “marked” LTRs generated alsoserveasuseful substrates for evaluating the inheritance patterns ofsequences in the LTRs. We have examined the effect of lengthening the Tyl LTR from 334 bp upward and find that transposition is dramatically inhibited when the LTR length exceeds -500 bp. Nevertheless, these long LTR elements are mobilized into the genome at a low frequency. These rare events represent Tyl multimers rather than normal Tyl simple insertion events. Tyl multimers have previously been reported as arising at a relatively low frequency in cells undergoing transposition of a normal Tyl element ( ~ ’ E I N S T O C K et al. 1990; MASTIWNGELO et al. 1992), andat a relatively high frequency in cells in which normal integration is blocked by either mutations in IN or mutations in LTR termini (SHARON et al. 1994). MATERIALS AND METHODS Yeast strains, media and measurement of transposition frequencies: A I I yeast strains used are described in Table 1. Media used were as described by ROSEet al. ( 1990). SC Foa and

SC -Trp Foa plates contained 0.1 % 5-fluoro-orotic acid (Foa) ( B o E ~et al. 1984). Transposition frequency measurements were made asfollows. Transformantsbearing the 7RPlmarked GAL-Ty1 plasmid of interest were maintained as patches on SC -Ura glucose plates at 30” (these plasmids

V. Lauermann, M. Hermankova and J. D. Boeke

912

TABLE 1 Yeast retrotransposon LTR lengths Element

Host

LTR length (bp)

Reference

TYl TY2 Ty3 Ty4 Ty5 Tcal Tfl Tf2 Pm-T 1

S. rereuisiae S. cereuisiae S. cmeuisiae S. rereuisiae S. paradoxus C. albicans Sch. pombe Sch. pombe P. membranaefacians

334 332 340 371 25 1 388 358 349 289

CLARE and FARABAUCt1 (1985) WARMIN(:TON et al. (1985) HANSEN et al. (1988) JANETZIW and LEHI.E(1992); STUCKA et al. ( 1992) ZOL~rt al. (1996) CI-IENand FONZI(1992) LEVINet al. (1990) WEAVER et al. (1993) PEIU~SON et al. (1995)

contain URA? as the selection marker on the plasmid backbone). The patches were then replica-plated to SC -Ura galactose plates and incubated for 5 days at 22”. The patches were then sequentially replica-plated to SC -Ura glucose (overnight at 30”) and YPD medium (overnight at 30”) to allow plasmid loss. The patches were then treatedin one of two ways to determine percentage transposition, which is defined as [number of Trp+ Ura- cells] / [total number of Ura- cells] X 100. Method 1: Cells were scraped from the W D patch into water and then diluted and plated on YPD plates at -200-400 colonies per plate at 30”; the colonies were then replica-plated to SC -Ura glucose and SC -Trp glucose plates and thefrequencies of the two types of colonies were tallied. This method was used when transposition frequencies were relatively high ( >1% ) . For those constructs that gaverise to lower transposition frequencies, method 2 was used. Method 2: The cells were scraped from the YPD patches as above and appropriate dilutions were plated onto SC Foa plates (these colony counts give the population of total Ura- cells) and SC -Trp Fwd plates (these colony counts give the population of Trp Ura- cells) . Methods 1 and 2 gave similar results when performed in parallel with strain VL3 (data not shown). For qualitative transposition assays, the W D patches were simply replica-plated to SC. -Trp Foa plates (Figure 2 ) . Oligonucleotides: Oligonucleotides used were asfollows: JB14,5’ CTAGTTAGTAGAGGATCCCGGGAGCTCTGATAG TTGATT 3’; JB746,5’ CAGTGAGCGCGCGTAATACGACTCACTATAG 3 ’ ; JB772, 5 ’ CCCCGGATCCGCTGTAAAAG CAGTAAAATCAATC 3 ‘ ; JB868, 5‘ CCCCGGATCCCTATTTGTATGTTCATCCATGCC 3’ ; JB873, 5’ CCCCGAGCTCACAATGAGTAAAGGAGAAGAACTTTTC 3 ’ ; JB878, 5 ’ CCCCGGATCCAGTATATTATCATATACGGTG 3 ’ ; JB879, 5 ’ CCCCGGATCCCGGGAGCTCGTAATGTAAATACTAG TTAG 3 ‘ ;JB880, 5 ’ CCCCGGATCCGAAGATGACGCAAATGATG 3 ’ ;JBSSI, 5 ‘ CCCCGGATCCCGGGAGCTCTAACACCGTATATGATAATATAC 3 ’ ; JB882, 5 ’ CCCCGGATCCTAGTCATCTAAATTAGTGGAAG 3 ’ ;JB883,5’ CCCCGGATCCCGGGAGCTCTTTCTCATCATTTGCGTCATC3 ’. Construction of GALTyl plasmids with altered LTR length: All marked plasmids used were derived from pX3, which is a GAL1-Tyl-TRPI 2-p plasmid containing UxA3 as the vector backbone marker (XU andBOEKE1987). A small polylinker, referred to as “SSB” and containing theSad, SmaI and BamHI sites, was introduced into the 3‘ Tyl LTR by sitedirected mutagenesis as described by KUNKEL( 1985) using oligonucleotide JB14 ( 5‘ CTAGTTAGTAGAGGATCCCGG GAGCTCTGATAGTTGATT 3 ’ ; underlined bases represent the SSB polylinker). A 1.8-kb HindIII-Sal1 fragment containing the 3’ LTR from pGTyl-H3 (pJEF724) was subcloned into similarly digested M13mp18, creating recombinant phage fV IT25. Mutagenesis was carried out in single-stranded +

phage fVIT25DNA, creating mutant phage fV IT26. To construct GAL-Tyl plasmids in which the BamHI site in the SSB polylinker would be unique, the preexisting site downstream of Tyl in the pJEF724 plasmid was first removed by cutting pJEF724withBnmHI, filling in withKlenow fragment, and religating, generating plasmid pJEF1706. The full-length GAL-Tyl-containing plasmid pVIT41, containing LTR:: SSB, was then constructed by three-piece ligation of 8.8-kb BstEIISulI and 2.5-kb BstEII-Hind111 fragments from pJEFl706 and the HindIII-Sal1 fragment bearing LTR:: SSB from fVIT’L6. The GAL-Tyl-TRPI-marked plasmid pVIT43 that was the progenitor for allof the other constructs (Figure 1) was constructed by three-piece ligation of3-kb BstEII-BamHI and 8.8kb BstEII-Sal1 fragments from pX3 and the 0.85-kb SalI-BglII piece from pVIT41. Thus all Tyl codingsequences came from pX3 and were not exposed to PCR amplification. The LTR sequence was confirmed by DNA sequencing. Derivatives of the pGTyl-TRPI-L,TR:: SSB plasmid pVIT43 (Figure 1 ) were made by inserting various DNA fragments into Sad / BamHI-digested pVIT43 as follows: LTR: : lacZcy, a 698-bp Sad-BglII fragment from M13mp18; LTR:: g u ~1.8-kb , Sad-BamHI fragment frompVIT5l (see below) ; LTR:: ZMf’4, 182-bp Sad-BanzHI fragment from pVITlll(see below) ; LTR:: g,fp, 980-bp SacI-BamHI fragment constructed by PCR. The g f p fragment in pVITP18 was PCR amplified using prirners JB868 and JB873 and template plasmid pGFP1O.l (provided byM. CHAI.FIE) . We also constructed a frameshift version of the SSB polylinker by filling in the BamHI site with Klenow and religating, resulting in plasmid pVIT45. pVITl11, containing the 182-bp IMT4 gene fragment, was constructed asfollow~s: the 182-bp HindIII-BamHI fragment containing ZMT4 from plasmid pKC35 (CHAPMAN et al. 1992) was subcloned into cloning vector pKP59 digested with those enzymes. pKP59 contains a polylinker with the following restriction sites: XbaI, HpaI, MluI, BglII, SalI, Xhol, S a d , N d d , HindIII, PstI, SalI, BamHI, SmaI, EcoRI (the kind gift of -IT1 I PELEN).pVIT51, containing the Sad-BamHI p ,s fragment, was constructed by ligating the HindIII-EmRI fragment containing the g u s openreading frame (ORF) from plasmid pRAJ260 (JEFFERSON et al. 1986) into the equivalent sites i n pKP59. Construction of GALTyl plasmids with different marker insertion positions: A set of LTR:: g f p plasmids bearing the g f p marker at various positions in the 3’ LTR was constructed as follows. The SSB polylinker was introduced at three aclditional positions within the 3’ LTR within the context of intermediate plasmid pVIT219. pVIT219 is the HindIII-Sal1 fragment of pJEF1706 subcloned into pBLUESCRIPT SK’ (Stratagene).The original SSB polylinker (present in pVIT43 and its derivatives) was inserted 24 bp from the left end of the 3’ LTR element; SSBs were also inserted 48, 72 and 96 bp into the 3 ’ LTR; these are designated LTK:: SSB-

LTR 913 Length Transposition and Tyl 48, -72, and -96. These SSBs were constructed by ligating together two PCR products. In each case a “left” PCR product, extending from the Hind111 siteinpVIT43 (position 5145, Figure 1) to the new SSB, was amplified usingoligonucleotide 56772 todefine its leftboundary and one of three oligonucleotides JB879, 881 or 883 that define the SSB end. The left PCR product was combined with a “right” PCR product, extending from the new SSB to the Sal1 site in pVIT43, and was amplified usingoligonucleotide56746 to define the right boundary and one of three oligonucleotides JB878, 880 or 882 to define the SSB end. The left end products were digested with Hind111 and BamHI, the right end products were digested with Sal1 and BamHI, and the left and right products were ligated into pVIT219 digested withHind111 and SaZI. BgZlI-Sal1 fragments containing each altered LTR were then ligated into pX3 that had been digested with BamHI and SaZI, generating plasmids isostructural to pVIT43. The resulting pGTyl-TRPl-LTR::SSB plasmids weredubbed pVIT481,483, and 482, respectively. The gfp fragment was subloned into these three vectors as described for pVIT218 above. The resulting pGTyl-TRPl-LTR:: gf p plasmidswere dubbed pVIT484,485, and 486, respectively.The 1-kb neo BamHI fragment from plasmid pGH54 (JOYCE and GRINDLEY 1984) was inserted into the BamHI site of pVIT482 to generate pVIT487. Construction of mini-Tyl element with long LTRs: The pGTy1-LTR::gus plasmid pVIT56 was digested with HpaI and BglII, and the large fragment was ligated to an EmRV-BamHI fragment containing HIS3 previously used to construct miniTyl elements (from plasmid pX130) ( X u and BOEKE1990), resulting in plasmid pVIT2 1 1 . Southern blot analysis: Yeast genomic DNA was prepared as described (HOFFMAN and WINSTON1987). The DNA was digested overnight with appropriate restriction enzymesin the supplier specified bufferat 37” and run on a 1.2%agarose gel at 40 V overnight. The gelwas treated with HCl for 15 min, denatured for 30 min, renatured for 30min and the DNA was transferred to a nitrocellulose membrane overnight using 10X SSC. DNA wasUV-crosslinked to the membrane using a Stratalinker ( Stratagene). Themembrane was prehybridized at 65” for 2 hr and then hybridized overnight at 65” with the probe. The filter was then washed three times for 15 min in 0.1X SSC, 0.1% SDS at 55”. The probes usedwere and generated using the random priming method ( FEINBERG VOGEISTEIN1983) using the following input DNAs. The g u s probe was a PstI restriction fragment from pRAJ260. The larZ probe was a 698-bp SacI-BglII fragment from M13mp18. RESULTS

A family of GALTyl plasmids

with alteredLTRs:

GAL-Tyl plasmids are a family of plasmids ( pGTy plasmids) that can be used to study the stucture and function ofyeast transposons. In these plasmids, the controllable GALl promoter replaces the US region of the 5 ’ LTR, which normally provides the promoter sequences for Tyl. Furthermore, marker sequences of 1-2 kb can a nonessentialregion be convenientlyinsertedinto near the 3’ e n d of the transposon and downstream of the POL OW, butupstream of the 3 ’ LTR. These marked Tyls can be launched at high efficiency into either genomic or plasmid targets by induction of the GALl promoter, and the properties of the progeny elements can be studied (BOEKEet ul. 1985, 1988, 1991; Xu and BOEKE1987). We have constructed a new set of pGTyl-TRPI plasmids that bear selectable markersand other DNA frag-

ments in theU3 region of the 3 ’ LTR (Figure 1, Table 2 ) . Accordingtopreviousexperimentsandthe accepted models for reverse transcription of retroviruses as well as Tyl, the U3 regions of both progeny LTRs are expected to be inherited from the 3’ LTR. Transpositionfrequencies: The transpositionfrequencies of the pGTyl-TRPI elements with LTR alterations were measured in bothwild-type strains and spt3 strains(Figure 2, Table 3 ) . In wild-type strains, the analysis of T y l transposition is complicated by the fact Tyl elements(whose LTRs areunthatgenomic marked) are expressed at high levels. Previous studies have shown that the progeny Tyl elements that arise in such strains can be hybrids between the marked element and an unmarked chromosomallyderivedelement ( BOEKEet al. 1986). However, in spt? strains, the chromosomal elements are not transcribed (WINSTON et ul. 1984), and thetransposition of the marked GALTyl-ZRPI element can be studied in isolation, because GALI-driven transcription is unaffected by the spt3 mutation ( BOEKEet al. 1986). W‘e introduced a variety of pGTyl plasmids with LTRmarked elements into isogenic SZ’73+ and s p 3 strains and observed that transposition frequency was inversely related toLTR length (Table 3 ) . Elements with the short SSB polylinker insertion transposed at essentially wildtype efficiency, whereas elementswith LTR insertions > 1 kb in lengthhad extremely low transposition frequencies. element with the LTR: : lucZa marker had The Tyl-T’l an intermediate phenotype, namely -10% of wild-type LTR in the SpT3+ strain background, but this element transposed at -0.3% of the wild-typeLTR in the spt3 strain background. Thus it appears that transposition of at least this lucZa-marked element can be “helped” in some currently unknown way by genomic Tyl elements. LTR length and not position of marker insertioninhibitstransposition: The above experiments were carried out with elements that carried various markers inserted in the SSB polylinker inserted at position 24 of the 3 ’ LTR. It remained possible that an important element or coding region might be interrupted by these insertions. To address this point, we created a frameshift mutation in the SSB polylinker that had n o effect on transposition frequency (data not shown). In addition, we built a set of additional SSBcontaining plasmids with the SSB inserted at positions 48, 72, and 96 bp in the 3’ LTR (see MATERIALS AND METHODS). A set of transformants containing eitherSSB or g fp inserted at the various positions in the 3’ LTR U3 region were qualitatively assayed for transposition. All of the SSB plasmids transposed as well as the wild type, (data not shown). However, none of the LTR: : gfp plasmids derived from these LTR: : SSB plasmids transposed at high efficiency in spt3 strains, indicating either that the gfp marker was incompatible with LTRs, regardless of the position transposition or that long of insertion of the marker, interferedwith efficient transwe conposition(Figure 2 ) . To rule out the former, structed pVIT487, which contains the neomarker inserted

!)

1'.L.m~w1nann.M. Hrrmankova ancI.1. I). I%ot-kr

1 .I

I'I(;I'RI'. l.-T\l (.I('Inrnts w i t h a l w r c d 1,Tk. Thc diagram indicatrs. ~ ~ ~ ) ~ ~ r o tso i scdc*. ~~l~~t~~l t h r strllclllrrs o f t h r a l tc~.rtl 1,TRs prrsrnt i n 1l1cpl;lslnitls stutlicd i n t h i s p p c r . Thr cil-clc thr slrllclllrr r~'pr~'s~'l1ts o f ' plxmitl p\'IT43 t h a t is the parrI1tal vrctor li)r most o f the. plxsnlitls tlcscrilwtl here*. w i t h t l w positions o f inlportant rrstrictio11 sitrs intlicntctl.

TABLE 2 Yeast strains and plasmids

Strain

\'I \'I

,Y'

..I"

1'1.1 I "

11.I "' \'l,lX' \'l,!!-l' \T,23" \'I, W ' \'!,ii' \'1,?(;4'!

\'I,Yi"' \'I.?i.5''

\'I .?i(i" \'I ,!!ii" \'I,?i8"

\'I ,!!i!V \'l,!!X"'

\ ' I 2 8 1" \'!,!X!?' \'I ,:$?X" \'I 33!V \'I %O"

\'I ,3(i1 SI .JIYdi!)'

1'1.1 ' S.1111.d

pS3 p\'IT4-1 pM; I 403 ll\'lT-l3

ps

? l

p\'IT4.5 pI'IT4.5 p\'IT43 pI'IT3.5 p]IlWX p\'IT43 p\'!T48 1 p\'!T4X3 p\'!T482 ps3 p\'ITI! I X p\'!T484 1>\'IT4X.5 p\'lT4X6 p\'IT4Xi p\w-1 I 3 p\'IT44 p\7T.',3

-

Pl;lsnlid drscription"

Krf'cw~ncc.

Native LTR. TIVI marked I,ody I .TR::IrcZCr Native LTR; protcasc f'rmncshif't m u t a n t ' 1.TR::SSR Native LTR, 7W'I m;lrkctl l,otly 1,TR::SSRframcshilt I~TR::SSRf~~meshilt 1,TII::SSR 1,TR::grt.s A3'LTRX 1,TR::SSB I,TR::SSRM'' LTR::SSRT~" LTR:SSR96" Native LTR, TRPI m;lrkctl body I.TR::&I I.TK::g/j1-48" LTR:&-7?" LTR:g/p96" I.TR:: w d ) 6 " LTR::Il\W4 1,TK::lnch 1,TR::~qt.s

St. and Ro~:ta,: ( 1 9 X i ) This str~tlv This str~tl!, This srr~tly SI. ; t n d ROI. LEVIN and,J. D. BOEKE,1993 Sequence analysis of closely related retrotransposon families from fission yeast. Gene 131: 135-139. WEINSTOCK, K., M. MASTRAN(;EI.O,T. BURKFTI',D. GARFINKEL. and J. STRAIHERN, 1990 Multimeric arrays of the yeast retrotransposon Ty, Mol. Cell. Biol. 10: 2882-2892. F., K.J. DURBINand G. R. FINK,1984 The SF73 gene is WINSTON, required for normal transcription of Ty elements in S. cereuisiae. Cell 39: 675-6382, XCJ, H.,and .J. D. BOEKE, 1987 Highfrequencydeletion between homologous sequences during retrotransposition of Ty elements in Saccharomycrscerm~isiae. Proc. Natl. Acad. Sci. USA 84: 85538557. Xu, H., andJ. D. BOEKF, 1990 Localization of sequences required in cis for yeast Tyl element transposition near the long terminal repeats: analysis of-mini-Tyl elements. Mol. Cell. Biol. 10: 26952702. YOUNGKEN, S. D., J. D. BOEKE,N. J. SANDI~.RS and D.J. GARFINKKI., 1988 Functionalorganization o f t h e retrotransposon Ty from Saccharomycrc cereuisiae: the Ty protease is required for transposition. Mol. Cell. Biol. 8: 1421-1431. Yu, K., and R. T. EI.DE:R,1989 Some of the signals for 3' end formation in transcription of the Saccharomyces cereuzsine Ty-Dl5 element arc immediately downstream of the initiation site. Mol. Cell. Biol. 9: 2431 -2444. Zorl, S., N. KE, J. M. KIM and D. F. VOWAS, 1996 The SacchuromycP,T retrotransposon Ty5 integrates preferentially into regions of silent chromatin at thetelomeres and mating loci. Genes Dev. 10: 634-645. Communicating editor: F. WINSTON