ABSTRACT
A specific terminal structure of preintegrative DNA is required for transposition of retroviruses and LTR-retrotransposons. We have used an anchored PCR technique to map the 3' ends of DNA intermediates synthesized inside yeast Ty1 and Ty3 retrotransposon virus-like particles. We find that, unlike retroviruses, Ty1 replicated DNA does not have two extra base pairs at its 3' ends. In contrast some Ty3 preintegrative DNA molecules have two extra nucleotides at the 3' end of upstream and downstream long terminal repeats. Moreover we find that some molecules of replicated Ty3 DNA have more than two extra nucleotides at the 3' end of the upstream LTR. This observation could be accounted for by imprecise RNAse H cutting of the PPT sequence. The site of Ty1 and Ty3 plus-strand strong-stop DNA termination was also examined. Our results confirm that the prominent Ty1 and Ty3 plus-strand strong-stop molecules harbor 12 tRNA templated bases but also show that some Ty1 and Ty3 plus-strand strong-stop DNA molecules harbor less tRNA templated bases. We propose that these less than full length plus-strand molecules could be active intermediates in Ty retrotransposon replication.
Retroviruses and retrotransposons such as yeast Tys and Drosophila copia elements replicate through RNA intermediates and alternate their genetic material between RNA and DNA (1 -4 ). The genomic RNA of these mobile genetic elements is converted into double-stranded DNA by the process of reverse transcription. The replicated DNA is then integrated into the genomic DNA of the host cell where it can be transcribed to produce new molecules of genomic RNA. Synthesis of each strand of retrotransposon DNA begins with the synthesis of short DNA products called minus-strand and plus-strand strong-stop DNA. Minus-strand strong-stop DNA synthesis is initiated from the 3' hydroxyl group of a primer tRNA annealed at a primer binding site (PBS) located just downstream of the R-U5 sequence of the genomic RNA. Plus-strand strong-stop DNA synthesis commences from an RNAse H resistant oligoribonucleotide spanning a purine rich sequence (the PPT, polypurine tract) located just upstream of the 3' U3-R sequence of the RNA. Minus-strand and plus-strand strong-stop DNA are elongated after they have moved from their sites of synthesis at one end of the template to an acceptor region at the other end of the template in a process called strand transfer. As a consequence of the two strand transfers, the unique 3' U3 RNA sequence is duplicated at the 5' end of the preintegrative DNA to form the upstream LTR and the unique 5' U5 RNA sequence is duplicated at the 3' end of the DNA to form the downstream LTR. The final product of reverse transcription is a two LTRs linear double-stranded DNA molecule that is longer than the genomic RNA. For most retroviruses the replicated linear extrachromosomal DNA has 2 bp at each end which are not present at the end of integrated proviral DNA. The first step in retroviral integration is the cleavage of two nucleotides from the 3' end of preintegrative DNA. Thus, the 3' termini of retrovirus DNA intermediates represent a mixture of species with either two or no extra nucleotides. Recently Kirchner and Sandmeyer (5 ) have analyzed the ends of preintegrative Ty3 DNA and found that it has retrovirus-like two extra base pairs at each end. The terminal structure of Ty1 preintegrative DNA has not been directly examined. In the present study we have used an anchored PCR technique to map the 3' ends of DNA intermediates synthesized inside Ty1 and Ty3 VLPs. We find that Ty1 replicated DNA does not have two extra base pairs at its 3' ends whereas some Ty3 preintegrative DNA molecules have two extra nucleotides at the 3' ends of the upstream and downstream LTRs. Moreover we find that a few replicated extrachromosomal Ty3 molecules have more than two extra nucleotides at the 3' end of the upstream LTR. To explain this result we propose that the cleavage of the PPT primer by the RNase H activity of Ty reverse transcriptase is imprecise, leading to templated addition of extra nucleotides at the end of minus-strand after the second strand transfer.
The site of Ty1 and Ty3 plus-strand strong-stop DNA termination was also examined using the anchored PCR technique. In the case of Ty1, hybridization experiments (6 ,7 ) and our previous mapping results (8 ) indicated that most plus-strand strong-stop DNA molecules are extended into the tRNA primer. It has been recently reported that the majority of Ty1 and Ty3 plus-strand strong-stop DNA molecules harbor 12 tRNA templated bases at the 3' end and may not be direct intermediates in retrotransposition (5 ,9 ). Our results confirm that the prominent Ty1 and Ty3 plus-strand strong-stop molecules harbor 12 tRNA templated bases but also show that some molecules harbor less tRNA templated bases. These less than full length plus-strand molecules could be active intermediates in Ty replication and could carry on a retrovirus-like second strand transfer involving complementarity between the tRNA inherited sequences and the PBS sequence at the end of the minus-strand DNA.
The yeast strain AGY9 (Mat a, his 4-539, lys 2-801, leu 2[Delta]1, trp 1[Delta]63, ura3-52, spt 3-202, GAL+) kindly provided by J.D.Boeke was used to minimize the amount of reverse transcript specified by endogenous Ty1 elements. Plasmid pJEF 1105 (10 ) kindly provided by J.D.Boeke is a high copy number (2 [mu]m) plasmid marked with URA3 containing a Ty1-neo element fused to the GAL1 promoter. Plasmid pEGTy3-1 (11 ) kindly provided by S.B.Sandmeyer is a high copy number plasmid marked with URA3 containing the Ty3 element fused to the GAL1 promoter.
AGY9 transformed with plasmid expressing Ty1 or Ty3 was used for VLPs production. Upon galactose induction, cells transformed by these plasmids produce large amounts of VLPs. Ty1-VLPs were isolated on a sucrose step gradient using a method described by Eichinger and Boeke (12 ) with minor modifications (7 ). Extraction of DNA from VLPs was as described (7 ).
The poly dA tailing and the anchored PCR was done as described by Charneau et al. (13). Terminal nucleotidyltransferase was used to add homomer tails to DNA fragments extracted from purified VLPs. The DNA was heat-denatured at 955C for 2 min and polydA tailing was achieved for 15 min at 375C in a final volume of 20 [mu]l in the presence of 50 U terminal nucleotidyltransferase EC2.7.7.31 (Boehringer Mannheim), 100 pmol of dATP, 0.75 mM CoCl2, 200 mM potassium cacodylate, 25 mM Tris-HCl pH 6.6, 0.25 mg/ml bovine serum albumin. The reaction was stopped by EDTA and the DNA was phenol/chloroform purified and precipitated by ethanol. The poly dA tailed DNA fragments were subjected to PCR amplification. The reaction conditions were 50 mM KCl, 10 mM Tris-HCl pH 8.4, 0.1 mg/ml gelatin, 1.5 mM MgCl2, 0.8 mM dNTPs, 16 ng/[mu]l each of primers, 1 [mu]l of DNA solution and Taq polymerase in a final volume of 100 [mu]l. PCR amplification was carried out for 35 cycles at 925C for 2 min, 505C for 2 min and 725C for 2 min. The resulting PCR amplification products were extracted with phenol/chloroform and digested by EcoRI and HindIII or KpnI and HindIII. The digested products were ligated to EcoRI and HindIII or KpnI and HindIII cut pSL1180 phagemid. Competent Escherichia coli cells were transformed. Small-scale phagemid DNA preparation of individual clones were used for double-stranded DNA sequencing.
Oligonucleotide primers were synthesized using an Applied Biosystem 381 A DNA synthesizer. The following primers were used to amplify plus-strand 3' ends, Ty1: oligo dT primer 5'-ATCGAAGCTTTTTTTTTTTT-3' and primer 275 (positions 7254-7273 of pJEF1105) 5'-ATCGAATTCAGAATTGTGTAGAATTGCAG-3' or primer 276 (positions 7000-7020 of pJEF1105) 5'-ATCGAATTCAACACTGGCAGAGCATTACGC-3'. Ty3: oligo dT primer and primer 2321 (positions 5274-5295 of Ty3) 5'-ATCGAATTCCAACTGGTTACTTCCCTAAGAC-3'.
The following primers were used to amplify the minus-strand 3' ends, Ty1: oligo dT primer and primer 7680 (positions 1074-1093 of pJEF1105) 5'-TACGGTACCTTGGTTTTGGGTCATCATGC-3'. Ty3: oligo dT primer and primer 1845 (positions 817-843 of Ty3) 5'-TATGAATTCGGTGGCATTCTGTCCCAAATCTTTCTG-3'.
The overall flow of the LTR retroelement replication process is presented in Figure 1 . Question marks in step E and G indicate the 3' ends of DNA intermediates which were analyzed using the anchored PCR technique described in Figure 2 . In step E the 3' end of plus-strand strong-stop DNA was mapped to determine how far in the tRNA molecule plus-strand synthesis is proceeding. In step G the question mark indicates that the anchored PCR technique was used to examine the terminal structure of Ty1 and Ty3 DNA after completion of full length DNA synthesis.
Specific Ty1 or Ty3 primers complementary to sequences in the U3 region and an oligo dT primer complementary to the 3' poly dA tail were used to amplify the 3' end of plus-strand DNA. This combination of primers (U3 + dT) allows amplification of the ends of plus-strand strong-stop DNA as well as full length plus-strand DNA and plus-strand DNA initiated at the central PPT2 sequence in the case of Ty1 (7 ,8 ). The 3' ends of Ty1 plus-strand DNA were also amplified with a combination of primers (a primer located upstream of U3 and the poly dT primer) which allows separate analysis of the ends of full length plus-strand DNA. 3' ends of Ty1 plus-strand. Figure 3 shows the distribution of 3' ends of Ty1 plus-strand DNA analyzed with the U3 + dT primers. About 75% of the molecules terminate at the end of the LTR. The remaining 25% plus-strand molecules harbor tRNA-templated bases presumably generated by reverse transcription of the primer tRNA molecule attached to the minus-strand template.
3' ends of Ty1 minus-strand. A specific Ty1 primer located downstream of the PBS sequence and the oligo dT primer were used to amplify the 3' ends of Ty1 minus-strand DNA. The 3' end of 20 Ty1 full length minus-strand DNA molecules were analyzed. We find that all molecules terminate at the end of the LTR (data not shown). No molecules terminate with two extra nucleotides at the end of the LTR. Thus the 3' termini of full length minus-strand DNA coincides with the end of the LTR. This result is in agreement with the prediction from the sequence of Ty1 retrotransposon that preintegrative Ty1 DNA should have the same terminus as that of the integrated element.3' ends of Ty3 minus-strand. The distribution of Ty3 minus-strand 3' ends is illustrated in Figure 4 . The majority of molecules terminate at the end of the LTR. A few molecules have two extra nucleotides CT at the end of the LTR. Surprisingly molecules with 4, 5 and 14 extra nucleotides complementary to the PPT sequence have also been characterized. The 3' end of minus-strand is determined by the site of initiation of plus-strand synthesis at the 3' end of the PPT plus-strand primer which is generated by RNase H cleavage of genomic plus-strand RNA after minus-strand has proceeded beyond the PPT sequence (Fig. 5 ). If cleavage of the 3' end of the PPT by RNase H is imprecise and cuts within the PPT sequence, plus-strand strong-stop DNA initiated from the 3' end of the PPT would be heterogeneous and slightly longer. After strand-transfer, extra nucleotides would be templated to the 3' complementary end of minus-strand DNA. This mechanism could explain how 4 or 5 nucleotides can be added at the 3' end of minus-strand DNA. To explain that some minus-strand 3' ends have a 14 extra nucleotide extension spanning the entire PPT we propose that the strong-stop plus-strand DNA still attached to the PPT primer could be prematurely dissociated and transferred at the end of minus-strand DNA. Reverse transcription of the PPT sequence would template addition of the 14 extra nucleotides at the end of minus-strand DNA. All Ty3 preintegrative DNA molecules with more than two extra nucleotides at the end of the upstream LTR are likely dead-end products in the Ty3 life cycle.
In the present study we have directly examined the 3' ends of minus- and plus-strand Ty1 and Ty3 extrachromosomal DNA. We confirm the recent observation that some Ty3 replicated DNA molecules have two extra nucleotides at the 3' ends of the upstream and downstream LTR (5 ). In contrast, no Ty1 molecule with two extra nucleotides at the end of the LTR were found in our study. Recent results of A.Gabriel and E.Mules provide evidence that one non-templated base can be added at the 3' end of some full-length minus-strand molecules (A.Gabriel and E.Mules, personal communication).
Mapping of the 3' end of Ty3 preintegrative DNA reveals that some molecules have more than two extra nucleotides at the 3' end of the upstream LTR. We propose that incorrect cleavage of the Ty3 PPT sequence by RNase H could explain this heterogeneity (Fig. 5 ). It is interesting to note that we have not observed heterogeneity generated by imprecise cleavage of the PPT sequence by RNase H at the 3' end of Ty1 upstream LTR. One possible explanation for the difference observed between Ty1 and Ty3 could lie in the fact that the PPT sequence of Ty1 is shorter (8 or 9 nucleotides) and less purine rich than the 12 nucleotides Ty3 PPT sequence; imprecise cutting of Ty1 PPT sequence at its 3' end would generate RNA fragments too short to serve as primers for plus-strand DNA synthesis, thus only the PPT fragment with the correct size would be functional and would therefore generate the correct 3' ends. Imprecise primer processing has also been observed in vitro with retroviral MoMLV RNase H (15 ,16 ) and recent studies have provided direct evidence of imprecise cleavage by Ty1 RNase H at the U5-PBS border (A.Gabriel and E.Mules, personal communication).
Mapping of Ty1 and Ty3 plus-strand strong-stop DNA revealed that most molecules harbor 12 tRNA templated bases. In a recent report Lauermann and Boeke (9 ) suggested that plus-strand strong-stop DNA harboring 12 tRNA templated bases may not be a direct intermediate in Ty1 retrotransposition and proposed a type of plus-strand transfer in which less than full length plus-strand DNA with no tRNA templated bases could transfer and anneal to a minus-strand DNA which has U5 sequence on its 5' end. The observation that plus-strand molecules harboring less tRNA templated bases are present inside Ty1 and Ty3 VLPs suggests that a retrovirus like second strand-transfer could also be possible for yeast Ty1 and Ty3 elements. These molecules would be elongated at once to the full length plus-strand and therefore would not accumulate. This would not be very different from the mechanism used by retroviruses since it has been reported that in 80% of cases strand transfer of strong-stop plus-strand DNA takes place before reaching the end of the PBS sequence in the tRNA template attached to the minus-strand DNA (3 ). For Ty retroelements it is possible that copying of only part of the PBS could also be sufficient to allow the second strand transfer. It would now be interesting to know whether the retrovirus like plus-strand transfer and the type of plus-strand transfer suggested by Lauermann and Boeke (9 ) both occur during replication of yeast Ty retroelements.
We thank A.Gabriel and E.Mules for helpful discussion and communication of data before publication. We thank J.D.Boeke for providing the pJEF1105 plasmid and yeast strain AGY9. We thank S.B.Sandmeyer for providing the pEGTy3-1 plasmid. This work was supported in part by grants from the Association pour la Recherche contre le Cancer (ARC) and from la Ligue Nationale Franaise contre le Cancer, Comité Départemental du Haut-Rhin.
REFERENCES