Article

• The EMBO Journal (1997) 16, 6590 - 6602
• doi:10.1093/emboj/16.21.6590

Functional differences between the human LINE retrotransposon and retroviral reverse transcriptases for in vivo mRNA reverse transcription

Olivier Dhellin¹, Joël Maestre¹ and Thierry Heidmann¹

1. Unité de Physicochimie et Pharmacologie des Macromolécules Biologiques, CNRS URA147, Institut Gustave Roussy, 39 rue Camille Desmoulins, 94805 Villejuif Cedex, France

Correspondence to:

Thierry Heidmann, E-mail: heidmann@igr.fr

Received 6 May 1997; Revised 14 August 1997

• Keywords:

  • LINE,
  • pseudogene,
  • retrovirus,
  • reverse transcriptase,
  • transposon

Introduction

Mammalian genomes contain two major types of reverse transcriptase (RT)-encoding elements: the retroviral-like long terminal repeat (LTR) retrotransposons and the non-LTR (or LINE) retrotransposons (reviewed in Gabriel and Boeke, 1993; Eickbush, 1994). These two classes of elements transpose in a replicative manner, via reverse transcription mediated by the transposon-encoded RT of a genomic RNA transcript from the element (Boeke et al., 1985; Heidmann and Heidmann, 1991; Jensen and Heidmann, 1991; Pélisson et al., 1991; Moran et al., 1996). Phylogenetic examination of the RT-encoding domains from both types of elements reveal strong similarities, with seven highly conserved boxes (Xiong and Eickbush, 1990; McClure, 1993) which are also found in other RT-containing sequences, including prokaryotic retrons (Inouye et al., 1989), mitochondrial plasmids (Nargang et al., 1984; Kuiper and Lambowitz, 1988) and group II introns (Michel and Lang, 1985; Lambowitz and Belfort, 1993). In the LINE elements (reviewed in Martin, 1991a; Eickbush, 1992) the RT domain is part of a large ORF. A sequence has been identified in its 5'-region, by database analysis, as most probably encoding an endonuclease (Barzilay and Hickson, 1995; Martin et al., 1995, 1996). This prediction was recently confirmed experimentally for the human LINE (L1) element (Feng et al., 1996). At the 3'-end a domain with homology to RNase H can be detected, as is also observed in retroviral RTs (McClure, 1991), as well as a zinc finger motif (Fanning and Singer, 1987; Schwarz-Sommer et al., 1987). Phylogenetic analyses have shown that the LINE RTs are distantly related to those of retroviruses and retroviral-like elements, but functional studies have not yet revealed any significant difference between these two classes of polymerases. However, there exist fundamental differences between the mechanisms involved in transposition of the two types of elements. LTR elements make reverse transcripts of their genomic RNA within retroviral-like particles, initiation of reverse transcription being mediated by a three-component association between a specific viral sequence (the primer binding site, PBS), a complementary tRNA and the RT (reviewed in Coffin, 1996). The extrachromosomal proviral DNA copies thus generated then integrate into the genome via an integrase-directed process. Conversely, LINE most probably undergo both processes simultaneously, through in situ reverse transcription of their genomic RNA intermediate, initiated at the level of a 3'-OH from a nick within the target genomic DNA. The nick is generated by a transposon-encoded endonuclease, as demonstrated for the R2Bm element (Luan et al., 1993) and strongly suggested for the human LINE (Feng et al., 1996). Such differences in the overall transposition process could be due to the nature of the structural intermediates for reverse transcription and replication. Retroviral-like elements are associated with particles of known content and organization (reviewed in Coffin, 1996), whereas LINE elements are associated with ribonucleoprotein particulate intermediates of still poorly defined structure (Martin, 1991b; Hohjoh and Singer, 1996). In fact, another possibility could be that they more directly result from intrinsic differences between the encoded RTs themselves. In this respect, previous assays in which the RT domain of the yeast Ty1 LTR retrotransposon had been replaced by the human LINE RT-containing ORF (Mathias et al., 1991; Dombroski et al., 1994; Teng et al., 1996) have resulted in transposed elements with unusual structures (see Discussion), suggesting that the two RTs are not exchangeable. To investigate this issue we have developed an in vivo assay for reverse transcription, whereby the human LINE RT is not forced into a retroviral-like particle but is assayed in a biologically more relevant situation within homologous human cells (as well as in murine or feline cells). Using this assay we show that a RT domain can be delineated within the functional human L1 element (Dombroski et al., 1991; Moran et al., 1996), which shows a high efficiency RT activity allowing in vivo reverse transcription of RNAs with no sequence specificity, a property which is not shared by the retroviral RTs (from the murine MoMLV and human HIV retroviruses) that we have similarly tested. Hence, LINE and retroviral RTs have distinct reverse transcription capacities and the 'wide spectrum' RT activity of the former is likely to be responsible for the endogenous activity that we previously revealed within mammalian cells through de novo formation of pseudogene-like structures (Tchénio et al., 1993; Maestre et al., 1995).

Rationale of the assay

The rationale of the assay (Figure 1) relies on the use of expression vectors for the human LINE ORFs and for retroviral RTs and on the capacity of these expression vectors to make a cDNA copy of an intron-containing reporter gene in mammalian cells in culture. In a standard experiment cells (human NTera2D1 or 293 cells, murine 3TDM1 cells or feline G355.5 cells) were co-transfected with both the expression vector and the reporter gene. Two to 4 days post-transfection the transfected cell DNA was extracted and PCR carried out to test for occurrence of cDNA copies of transcripts from the reporter gene, which can be unambigously identified as a result of splicing out of the intron (Figure 2).

Figure 1.

Structure of retroviral and LINE elements and rationale of the RT assay. (A) Structure of the MoMLV provirus with the GAG, POL and ENV ORFs and of the human LINE retrotransposon with ORF1 and ORF2. The boundaries of the cleavage products within the retroviral ORFs are indicated with dotted lines, with the matrix (Ma), capsid (Ca) and nucleocapsid (Nc) proteins within Gag and the protease (Pr), reverse transcriptase (RT) and integrase (In) within Pol. Transcription start sites are indicated by arrows. RT-containing expression vectors derived from both elements are schematized in Figures 4 and 5. Reporter genes for the RT assay contain an intron (from the previously devised neoRT indicator gene; Heidmann et al., 1988; Heidmann and Heidmann, 1991), the CMV promoter and the SV40 polyadenylation signal and are schematized in Figure 3. Spliced transcripts, if reverse transcribed by the products of the RT-containing expression vectors, should result in intronless cDNA molecules that can be identified upon PCR analysis using primers bracketing the splice junction (see also Figure 2). (B) Experimental procedure for the in vivo RT assay. Mammalian cells (of human, murine or feline origin; see text) are co-transfected with both intron-containing reporter genes and RT-containing expression vectors and DNA is extracted 2–4 days post-transfection for PCR analysis as indicated in (A) and in Figure 2.

View full figure (39 KB)

Figure 2.

In vivo RT activity of LINE elements. The complementation assay in Figure 1 was carried out in human NTera2D1 cells using an intron-containing reporter gene (Figure 3, a) and the expression vector for the full-length LINE element (Figure 4, c). (A) Structure of the expected cDNA molecules and results. (Top) Primers bracketing the intronic domain of the reporter gene (primers neo1 and tk1, neo2 and tk2, d2 and tk2) are indicated, with the length of the PCR fragments expected upon splicing out of the intron; for reverse transcripts of the reporter gene the neo1–tk1 fragment obtained after PCR amplification (387 bp) should yield two fragments (of 193 and 194 bp) after SacI restriction (SacI generated at the splice junction; see Heidmann et al., 1988). (Bottom) PCR fragments were analysed by electrophoresis of 10 l of the PCR reactions in 1.5% agarose gels and ethidium bromide staining: lane a, PCR products (restricted or not by SacI) with both the intron-containing reporter gene and the LINE expression vector; lane b, reporter gene alone; lane –, control PCR without DNA; M, size marker. (B) Assay for double-strand DNA synthesis in the RT assay. (Top) Double-stranded DNA should be restricted by the indicated enzymes (XbaI and BglII). PCR amplification as in (A) using primers neo1 and tk1 should then result in a 387 bp amplified fragment in the case of single-strand cDNA synthesis and no amplified fragment in the case of double-strand synthesis, whereas control amplification using primers d2 and tk2 (both on the same side of the cut DNA) should yield a 180 bp fragment even after restriction. (Bottom) PCR fragments obtained with the primers indicated for each panel, with cellular DNA either uncut (nc) or cut (c) with XbaI and BglII prior to amplification. Lane –, control PCR without DNA. (C) Assay for second-strand DNA synthesis. Primer extension specific for the putative DNA plus strand was first achieved by a 45 cycle 'linear' PCR amplification with the tk1 primer, resulting in multiple single-strand copies as schematized. PCR amplification with two primers (neo2 and tk2) was then allowed to proceed for a limited number of cycles (23 cycles). PCR results obtained with or without preliminary primer extension are presented for transfection experiments as in (A), with both the expression vector and the reporter gene (a) or with the reporter gene alone (b). Controls for RNAs (for instance within putative RNA–DNA hybrids) as possible templates were performed by prior treatment of the extracted nucleic acids with RNase A (after denaturation at 95°C [+RNase A(1)] or with NaOH [RNase A(2)]) or upon treatment with RNase H. The expected 277 bp PCR fragments are indicated and their identity was ascertained upon restriction with SacI (not shown).

View full figure (57 KB)

A first series of expression vectors were constructed containing either the entire LINE element (i.e. both ORFs and the 5'- and 3'-untranslated domains) or only the LINE ORFs under control of the potent immediate early cytomegalovirus (CMV) promoter. Since it has been shown that LINE ORF2 is expressed within the bicistronic LINE at a level at least 100-fold lower than ORF1 (McMillan and Singer, 1993), we also constructed expression vectors harbouring only ORF2 (see vector structures in Figure 4). Similarly, several reporter genes were constructed harbouring the previously described intron-containing indicator gene for retrotransposition (see for example Heidmann et al., 1988; Heidmann and Heidmann, 1991). In a first series of reporter genes the indicator gene was inserted into a LINE element, with a small deletion to limit the overall size of the transcript, to take into account the possibility that some sequences might be essential in cis for reverse transcription. Other reporter genes were also constructed in which LINE sequences were deleted, encompassing domains 5' and/or 3' of the indicator gene (see reporter gene structures in Figure 3).

Figure 4.

LINE sequences required in trans for in vivo reverse transcription. (A) Co-transfection experiments were as in Figure 2, with the CMVLINEneoRT reporter gene and the LINE-derived expression vectors whose structures are indicated on the left, lanes a–d. Lane –, control PCR without DNA. (B) Experimental conditions as in (A) with the CMVORF2 expression vector and CMVLINEneoRT (lane 1), CMVneoRT (lane 2) or no reporter gene (lane 3). (C) Quantitation of in vivo RT activity in homologous and heterologous cells. Human (NTera2D1) or murine (3TDM1) cells were co-transfected with the CMVneoRT reporter gene and the LINE CMVORF2 expression vector and DNA was extracted 2 days post-transfection. Aliquots of 1.5 g cellular DNA and serial dilutions (by 10-fold steps) were used for PCR amplifications as in Figure 2.The indicated amount of pre-spliced plasmid copies (diluted into genomic DNA from non-transfected cells) were amplified in parallel as a reference. Lane –, control PCR without DNA.

View full figure (39 KB)

Figure 3.

Absence of effect of LINE sequences in the reporter gene for in vivo reverse transcription. Co-transfection experiments were as in Figure 2, with the CMVLINE expression vector (Figure 4, c) and the indicated LINE sequence-containing (or not containing) reporter genes (see reporter gene structures on the left); LINE sequences [including for (a) the 5' untranslated domain, ORF1, part of ORF2 and the 3' untranslated domain] are represented in grey and the intron-containing neoRT cassette with open boxes; the size of the expected PCR fragments before and after SacI restriction (-s, +s) are indicated.

View full figure (35 KB)

Evidence for in vivo reverse transcription by LINE-containing vectors

As illustrated in Figure 2A, co-transfection of cells with both a marked LINE element under control of the CMV promoter and an expression vector for the full-length LINE element resulted in cDNA copies that could be easily detected upon PCR amplification of the transfected cell DNA. PCR amplification of the DNA extracted 4 days post-transfection using primers bracketing the intronic domain of the indicator gene generated a 387 bp band, consistent with splicing out of the intron, which further contained the SacI site expected to be generated at the splice junction (Heidmann et al., 1988; Figure 2A). This band was not observed in the absence of the expression vector for the LINE element (Figure 2A) and no fragment could be detected after a second nested PCR amplification, which further increases PCR sensitivity at least 10-fold (data not shown). Amplification of larger fragments using primers located distantly from the reporter intron (namely at the 5'-end of the neo gene and at the polyadenylation signal at the 3'-end of the reporter gene) were also positive, although with a reduced intensity, again only in the presence of the expression vector for the LINE element (data not shown).

The nature of the cDNA molecules revealed by the PCR assay was further investigated to determine whether they corresponded to single- or double-stranded DNA (both would result in production of the 387 bp PCR fragment) and, more precisely, to determine whether the DNA plus strand was generated during reverse transcription. In a first series of experiments the extracted cellular DNA was first treated with a combination of restriction enzymes (XbaI and BglII) which do not cut single-stranded DNA and should disrupt the DNA region to be amplified (see scheme in Figure 2B). PCR amplification was then carried out as previously described. As illustrated in Figure 2B, under these conditions no PCR product was generated. As an internal positive control PCR amplification was also carried out before and after enzymatic digestion but using a pair of oligonucleotide primers located on the same side of the XbaI and BglII restriction sites, which should produce a PCR fragment of a distinct size (see Figure 2B, bottom): in that case, as expected, amplification was observed with both cut and uncut DNA. Altogether, these data suggest that a major fraction of the LINE-induced cDNA molecules are double-stranded (among which RNA–DNA hybrids cannot be excluded, but see below). In a second series of experiments (see scheme in Figure 2C), primer extension of the putative plus strand DNA was first performed by repeated cycles of PCR using a single primer (45 cycles) and occurrence of extended DNA molecules was then tested by standard PCR amplification using two primers, with a reduced number of cycles (23 cycles). As illustrated in the figure, PCR amplification under these conditions resulted in the expected fragment only upon preliminary primer extension, thus again strongly suggesting synthesis of plus strand DNA in the reverse transcription assay. As expected, amplification was only observed in the presence of the LINE expression vector and was insensitive to treatment of the extracted nucleic acids, prior to primer extension, by RNase A (after denaturation) or RNase H, which degrade RNAs and RNA strands within putative RNA–DNA hybrids (Figure 2C).

To measure the efficacy of in vivo reverse transcription mediated by the LINE expression vectors the amount of reverse transcript was estimated by semi-quantitative PCR analysis, using a previously described method (Ségal-Bendirdjian and Heidmann, 1991). Basically, PCR was allowed to proceed for a number of cycles adjusted so as to be in the range of linearity between the amount of DNA to be amplified and the amount of PCR products generated, as quantitated after gel electrophoresis and ethidium bromide staining (i.e. during the exponential phase of PCR). DNA plasmid molecules, diluted into genomic DNA from non-transfected cells, were amplified in parallel as standards (Figure 4C). Accordingly, the number of cDNA molecules could be systematically determined and was found to be in the range 10⁴–10⁵/g DNA, which corresponds to 0.01–0.1 molecules/cell, depending on the LINE expression vector tested (see below). The reverse transcripts found in relatively large amounts are most probably not integrated into the genome, but rather correspond to extrachromosomal cDNA molecules, as previously observed using rat LINE reporters (Ségal-Bendirdjian and Heidmann, 1991). This conclusion is consistent with the fact that despite the large number of cDNA copies generated in these transiently transfected cells and taking advantage of the indicator neo gene (see for example Heidmann et al., 1988; Heidmann and Heidmann, 1991) contained in the reporter plasmids, no G418-resistant clones could be isolated (unpublished results).

LINE sequences required in trans or in cis for in vivo reverse transcription

The demonstration that a full-length LINE expression vector can generate double-stranded reverse transcripts of an intron-containing, LINE-derived reporter gene and the simple in vivo assay described above allow an investigation of both the LINE-coding sequences in the expression vector which are required in trans for this process and the LINE sequences in the reporter gene which might be required in cis.

A series of deletion derivatives of the initial LINE-derived reporter gene were therefore constructed (Figure 3) and assayed as described above to test for the possible role in reverse transcription of LINE sequences in cis. As illustrated in Figure 3, similar levels of reverse transcript were obtained using reporter genes with complete deletion of the 5'- and/or 3'-domains of the LINE elements (including the LINE 3'-untranslated domain), ending in the minimal CMVneoRT reporter gene completely devoid of LINE sequences (Figure 3, lane d): clearly, no LINE sequence is specifically required in cis for reverse transcription (see also Figure 6 and the related section).

Figure 6.

LINE-mediated reverse transcription of mRNA from cellular genes. Cells (293) were transfected with expression vectors for the LINE ORF2, the MoMLV RT or no vector, either alone (B and C) or together with an expression vector for genomic TNF (A). Reverse transcripts of mRNAs from the TNF (A), E1A (B) or c-myc (C) genes were assayed by PCR using the pairs of primers indicated in the figure for the corresponding genes. Introns are indicated together with the expected size of the PCR products. For the TNF gene the second intron has a splicing efficiency close to unity, whereas the third intron is not fully processed (Neel et al., 1995). For the E1A gene two alternative splices have been described (Stephens and Harlow, 1987).

View full figure (69 KB)

Deletion derivatives of the initial LINE expression vector were then constructed and assayed to delineate the role of the LINE ORFs in trans for reverse transcription. The assays were performed either with the minimal CMVneoRT reporter gene mentioned above or with the LINE-containing reporter, in which case a deletion within ORF1 was introduced to prevent expression of this ORF from the reporter gene itself (see Materials and methods). As illustrated in Figure 4A for the LINE reporter gene and in Figure 4B for the CMVneoRT reporter, ORF1 is dispensable for reverse transcription activity of the LINE expression vector. Reverse transcription efficiency was even higher (up to 10-fold) with the CMVORF2 vector than with CMVLINE, as expected, since ORF2 expression in the former vector does not require translational re-initiation. An alternative interpretation involving a possible negative effect of ORF1 expression was ruled out by an experiment using a LINE expression vector with an in-phase deletion within ORF1, which gave the same result as the full-length LINE vector (not shown). Finally, a vector with a deletion within ORF2 (CMVORF2^*, Figure 4), assayed as a control, actually lacked reverse transcription activity in the assay. The LINE ORF2 is therefore necessary and sufficient to generate reverse transcripts in vivo. The number of cDNA molecules synthesized under these conditions was determined as indicated in the previous section and was found to be close to 10⁵/g DNA, i.e. close to 0.1 molecules/cell, taking into account the transfection efficiency measured by X-gal staining of cells transfected with a lacZ-containing plasmid. Figure 4C also demonstrates, interestingly, that similar levels of reverse transcript were obtained in murine and human cells (as well as in feline cells; data not shown), strongly suggesting that the reverse transcription activity mediated by the LINE ORF2 is not dependent on species-specific factors, but is rather a property of the protein per se.

In vivo reverse transcription of the reporter gene cannot be induced by a provirus or by retroviral RTs

Since the LINE-mediated in vivo reverse transcription activity revealed in the present assay does not require specific sequences in cis, we tested whether similar results would be obtained with retroviral RTs. We previously demonstrated that a cloned MoMLV provirus was competent for both viral particle formation and intracellular transposition in murine as well as human cells (Heidmann et al., 1988; Tchénio and Heidmann, 1991, 1992) and this proviral construct was tested. As illustrated in Figure 5A (lane b) for murine cells (similar results were obtained for human cells; not shown), this element was negative in the in vivo reverse transcription assay, whatever the reporter gene used. We checked that the transfected plasmid was 'functional' as a provirus by measuring RT activity in the viral particles released in the supernatant of transfected cells, using a classical in vitro RT assay with synthetic templates and primers (Table I). Since particle formation could impair reverse transcription from a non-viral RNA template, we then derived an expression vector for the retroviral RT, as described in Jean-Jean et al. (1989). Rather unexpectedly, this retroviral RT (Figure 5A, lane c) was similarly unable to generate cDNA copies of transcripts from the reporter gene (whatever the reporter gene and the cells tested) and these remained undetectable even after a second nested PCR amplification. Quantitation of the PCR assay using 10-fold serial dilutions of the DNA from the transfected cells demonstrated an at least 1000-fold lower in vivo reverse transcription activity, if any, of the MoMLV RT as compared with the LINE ORF2 (Figure 5B). In that case also we could demonstrate that the MoMLV RT was actually produced in the transient transfection assay. This was tested upon protein extraction from the transfected cells and assaying in vitro RT activity as described above, using synthetic templates and primers: as indicated in Table II, the MoMLV RT was active and its activity was even higher than that of the LINE RT (extracted under identical conditions), as determined under two assay conditions for divalent cations (Mg²⁺ or Mn²⁺). The inefficiency of the MoMLV RT to achieve reverse transcription of the reporter gene was also observed with another retroviral RT. Indeed, using expression vectors encoding the two subunits of the HIV RT described in Ansari-Lari and Gibbs (1994), no spliced reverse transcripts could be detected by the in vivo assay in either murine (Figure 5A, lane d) or human cells (not shown); in that case the in vitro RT activity of the HIV RT was not assayed, but its functionality has been previously documented (Ansari-Lari and Gibbs, 1994).

Figure 5.

Assay for in vivo reverse transcription using retroviral elements. (A) Co-transfection assays were as in Figure 2, with the CMVneoRT reporter gene and the indicated expression vectors (structure on the left): lane a, LINE ORF2; lane b, MoMLV GAG–POL; lane c, MoMLV RT; lane d, HIV RT (using the CMV p66 and SV p51 expression vectors for the full-length and RNase H-truncated subunits constituting the dimeric HIV RT). PCR and controls were as in Figure 2. (B) Assay for quantitation of the relative efficiency of the LINE RT versus the MoMLV RT for in vivo reverse transcription. Co-transfection was as in (A) with CMVORF2 (lane a) or MoMLV CMVRT (lane c). Cellular DNAs (1 g) were PCR amplified either directly or after serial dilutions (10-fold steps, the total amount of DNA before amplification being maintained constant by complementation with cellular DNA from untransfected cells). Lane –, control PCR without DNA.

View full figure (42 KB)

Table 1: Assay for the MoMLV GAG–POL expression vector

View full table

Table 2: RT activity in cellular extracts from cells transfected with expression vectors for the MoMLV RT, LINE ORF2 and deletion derivatives

View full table

LINE-mediated in vivo reverse transcription of mRNA from cellular genes

The specific reverse transcription activity of the LINE ORF2, not shared by retroviral RTs, as revealed by its capacity to in vivo reverse transcribe transcripts from the CMVneoRT reporter gene, was further assayed on transcripts from cellular genes. In a first series of experiments a cellular gene, the murine TNF gene, was assayed as in the previous sections in a transient transfection experiment within human cells, using a genomic copy of the gene under control of the strong CMV promoter (see Figure 6A). As illustrated in the figure, co-transfection of the TNF and LINE ORF2 expression vectors followed by PCR using primers bracketing introns 2 and 3 disclosed the bands expected for splicing out of the TNF introns: a 608 bp fragment associated with splicing out of both introns and an 831 bp fragment associated with splicing of only intron 2 (splicing efficiency of intron 2 is close to unity, whereas intron 3 is not fully processed; Neel et al., 1995). As expected, reverse transcripts of the spliced mRNAs were not observed in the absence of LINE ORF2 nor upon addition of the MoMLV RT. In a second series of experiments we assayed for in vivo reverse transcription of transcripts from integrated genes. Two genes were selected that do not possess processed pseudogenes (such pseudogenes would also result in intronless PCR fragments), namely the adenovirus E1A gene, which has integrated into human 293 cells (Graham et al., 1977), and the endogenous c-myc gene. As illustrated in Figure 6B and C, reverse transcripts could be detected in both cases, again exclusively upon transfection of the cells with the LINE ORF2 expression vector and not with the MoMLV RT. For the E1A gene the two expected cDNAs corresponding to the previously described alternative splicing of the E1A transcripts (Stephens and Harlow, 1987) were detected upon PCR amplification. For the c-myc gene, PCR fragments of the size expected for splicing out of intron 2 were similarly obtained after either a first or a second nested PCR amplification. For both genes (as well as the TNF gene mentioned above) identity of the amplified fragments was ascertained by the presence of the restriction sites indicated in the figure and, in addition, for c-myc, by nucleotide sequencing of the fragment, which confirmed precise splicing out of the intron (not shown).

Altogether, these results demonstrate that the in vivo RT activity of the LINE ORF2 is not restricted to transcripts from neoRT-containing genes but extends to transcripts from cellular genes, whereas the MoMLV RT is unable to do so for any of the gene transcripts that we have tested.

Delineation of the minimal LINE protein domain active for in vivo reverse transcription

The product of the LINE ORF2 is a large polypeptide (1275 amino acids) in which the RT consensus sequence occupies only a limited central domain (see Figures 7 and 8). To delineate the minimal sequence responsible for the specific in vivo RT activity of the LINE RT and to compare it with that previously defined for retroviral RTs we generated a series of derivatives with deletions on either side of ORF2. In these various constructs ORF2 translation was under control of a common AUG initiation codon (from a His-tag; see Materials and methods). They were tested as described above, after re-insertion into the CMV expression vector. The results are shown in Figure 7 and disclose that the 5'-end of the LINE ORF2 can be deleted without loss of in vivo RT activity, up to a domain 720 bp from the first box of homology between retroviral and LINE RTs. Loss of activity upon further deletion within the remaining 5'-domain, called the 'Z domain' (Doolittle et al., 1989; McClure, 1991), is complete and takes place rather abruptly (within 100–200 bp). This loss of in vivo activity is paralleled by a loss of RT activity measured in vitro, using the assay described in the previous section (ORF2 5'; see Table II). Deletions at the 3'-end of the LINE ORF2, within the putative zinc finger and RNase H domains, similarly did not affect the specific RT activity of the LINE element. However, more extensive deletions (e.g. 650 bp and 1 kb) resulted in a significant decrease in activity (up to 10-fold), but the RT still remained active. Accordingly, a minimal RT domain capable of in vivo reverse transcription can be delineated within ORF2 as shown in Figure 8.

Figure 7.

Delineation of the RT domain of the LINE ORF2. Assays were as in Figure 2, using the CMVneoRT intron-containing reporter gene and expression vectors for LINE ORF2 and the indicated deletion derivatives. RT activities, indicated on the right relative to that of the full-length ORF2, are the means of at least three independent co-transfection assays. For the symbolic representation of the domains within LINE ORF2 see the legend to Figure 8; the His-tag including the ATG initiation codon common to all constructs is indicated by a bold line at the N-terminus.

View full figure (90 KB)

Figure 8.

Comparison of the structure of the RT domain within the MoMLV retrovirus (POL, top) and the human LINE element (ORF2, bottom). The RT domains, as delineated in the present study for the LINE element and in Tanese and Goff (1988) for MoMLV, are indicated within bracketts (RT+). The seven highly conserved domains within both RTs are indicated in black (Xiong and Eickbush, 1990) and domains with lower conservation in grey (McClure, 1993). The MoMLV RNase H is indicated, as well as the putative homologous domain within LINE ORF2 (with a zinc finger indicated by a hatched box). A domain recently identified within LINE ORF2 as showing endonuclease activity (Feng et al., 1996) is positioned at the N-terminus. The protease and integrase domains in MoMLV are part of POL, but are cleaved by the MoMLV-encoded protease itself.

View full figure (41 KB)

A minimal RT domain can be defined within the human LINE retrotransposon which closely resembles the consensus sequence found in most RTs but still discloses a specific reverse transcription activity not shared by retroviral RTs. This difference is manifested in the ability of the former to generate very efficiently reverse transcripts in vivo from RNAs with no sequence specificity. This specific property could be responsible for the endogenous reverse transcription activity that we previously identified within human and murine cells through the formation of pseudogene-like structures (Tchénio et al., 1993; Maestre et al., 1995).

Delineation of a minimal LINE domain with specific RT activity

The present experiments have shown that both the 5'- and 3'-ends of the human L1 ORF2 could be deleted without loss of the RT activity revealed in vivo. These domains are most probably not complemented by some endogenous elements since identical results were obtained with murine, feline and human cells. The 5'-domain, also found in several non-LTR retrotransposons (reviewed in Feng et al., 1996; Martin et al., 1996), contains a sequence with close similarity to endonucleases and was recently demonstrated to possess the expected enzymatic activity (Feng et al., 1996). Deletion of this domain clearly does not impair RT activity of the protein, thus showing that the product of ORF2 is a multifunctional protein with distinct domains associated with distinct functions. Similarly, 3' deletions within ORF2 did not result in impairment of reverse transcription activity. This result is surprising, as this domain contains a putative zinc finger motif that was hypothesized to be a domain of interaction with nucleic acids (Fanning and Singer, 1987), possibly involved in the specific ability of the L1-encoded RT to initiate reverse transcription from the 3'-end of the LINE transcripts. Actually, this domain also discloses similarities with retroviral RNase H, in both its sequence and its position relative to the RT domain. In this respect, the lack of a significant effect of its deletion is reminiscent of the similarly observed lack of effect of RNase H deletion on the MoMLV RT (Kotewicz et al., 1987; Levin et al., 1988; Tanese and Goff, 1988).

This newly defined 'minimal' functional unit within the LINE ORF2 shows strong sequence similarities with retroviral RTs. As illustrated in Figure 8, both RTs have a closely related overall organization and harbour the seven highly conserved subdomains which were highlighted in Xiong and Eickbush (1990) and McClure (1991). The only significant difference that can be illustrated from a comparison of the corresponding domains is that the LINE RT domain is less compact than that of its retroviral counterparts: for instance, the seven subdomains are interspersed by longer interdomains in the former case and, similarly, the domain 5' of these highly conserved sequences is slightly longer. These structural features, including the large 5'-domain (called the Z domain in Doolittle et al., 1989; McClure, 1991), are actually also found in other phylogenetically related and/or ancestral RTs (reviewed in McClure, 1993) and might be relevant to evolutionary constraints imposed on retroviruses for compactness.

Functional differences between LINE and retroviral RTs

The present assay has revealed that the LINE RT could generate reverse transcripts with high efficiency from RNAs with no sequence specificity, including RNAs with no LINE sequences, whereas retroviral RTs were unable to do so under strictly identical in vivo experimental conditions. Interpretations of such differences should take into account the recently recognized fact that reverse transcription is a complex process involving, at least under in vivo conditions, several distinct steps, including initiation of reverse transcription and elongation, which are mediated by template–enzyme complexes with distinct enzymatic and even distinct structural characteristics (Isel et al., 1996; Lanchy et al., 1996). These considerations have not always been taken into account under in vitro experimental conditions in which artificial primers and templates are used, in many cases at concentrations of enzymes and substrates much higher than those encountered in vivo. Yet, it is noteworthy that the in vitro assays for LINE RT activity as reported in this paper, as well as those previously reported for the same LINE RT purified from yeast Ty particles, where they had been generated via chimeric Ty–LINE constructs (Mathias et al., 1991), did not reveal significant differences in polymerase activity when compared with 'retroviral' RTs, namely the MoMLV and Ty1 RTs respectively. The most likely interpretation could be that in these in vitro assays the measured polymerase activities essentially report on the RT activity in the elongation step as being the rate limiting step and, therefore, estimate incorrectly the first step in reverse transcription, i.e. the initiation step. The latter has recently been characterized in an extensive analysis of the HIV RT as disclosing very specific properties, including kinetic rate constants for interaction between the RT and the nucleic acids and the processivity of polymerization, several orders of magnitude different from those for elongation (Isel et al., 1996; Lanchy et al., 1996). These basic kinetic considerations could provide a hint as to the differences between LINE and retroviral RTs as presently observed in vivo, as these might rely on the ability of the former efficiently to prime reverse transcription from any RNA, whereas retrovirus or LTR retrotransposons most probably require more specific interactions (for instance in the formation of ternary complexes between RT, tRNA molecules and viral PBS sequences). Along these lines, it is noteworthy that the priming process for several RTs, not necessarily associated with transposable elements but phylogenetically close to those of the LINEs, has been demonstrated to be significantly different from that for retroviruses (reviewed in Levin 1997). For instance, in the case of group II introns (Zimmerly et al., 1995) and the non-LTR R2Bm element (Luan et al., 1993) priming is at a 3'-OH within the target DNA; for the bacterial retrons (Inouye and Inouye, 1993) priming is at a 2'-OH internal to RNA and in the case of the Mauriceville plasmid priming can take place either from short non-complementary RNA molecules or without any primer at all (Wang and Lambowitz, 1993; Kennell et al., 1994). Although in all these cases refined kinetic analyses have not been carried out, these distinct, and in some cases dual, modes of initiation of reverse transcription are likely to result in distinct rate limiting steps and overall reverse transcription activities under in vivo conditions. The present investigation, leading to delineation of a LINE RT domain, should permit a physico-chemical analysis as in Isel et al. (1996) and Lanchy et al. (1996) of the various steps of LINE reverse transcription and unravelling of the molecular basis of the observed functional differences between the LINE and retroviral RTs.

A source of endogenous RT activity within mammalian cells possibly associated with processed pseudogene formation

The specific activity of the LINE RT as revealed in this in vivo analysis (not observed with the retroviral enzymes) also strongly suggests that LINEs might be the source of the endogenous RT activity that we previously unravelled within both murine and human cells and found to be responsible for formation of processed pseudogene-like structures in those cells (Tchénio et al., 1993; Maestre et al., 1995). This conclusion would be consistent with previous experiments disclosing that retroviral infection, as well as forced expression of retroviral-like retrotransposons, in all cases resulted in cDNA genes which lacked the hallmarks of the processed pseudogenes naturally found in mammalian genomes (Dornburg and Temin, 1990; Levine et al., 1990; Derr et al., 1991). Similarly, forced expression of a human L1 RT inserted within the Ty LTR-containing retrotransposon as a chimeric construct did not result in canonical processed pseudogenes, but in cDNA genes which always included retroviral sequences (i.e. Ty sequences) (Dombroski et al., 1994; Teng et al., 1996). This might be the consequence of 'ectopic' expression of the LINE RT within a retroviral particle and it emphasizes the importance of in vivo analyses for a correct appraisal of naturally occurring processes. The present results showing LINE-specific reverse transcription of mRNA from cellular genes lend support to the plausible involvement of these elements in the generation of processed pseudogenes, whose formation in mammalian cells can now be analysed using the presently derived expression vectors together with integrated reporter genes as in Maestre et al. (1995).

DNA constructs

Expression vectors for LINE ORFs. Nucleotide positions refer to the L1.2A element sequence (Dombroski et al., 1991). pCMVL1 was constructed by insertion of the cloned LINE L1.2A (a generous gift from H.Kazazian), as a NotI–NsiI blunt-ended fragment, between the CMV promoter and the SV40 polyadenylation sequence of pCMV (Clontech) restricted with BspEI and NotI and blunt ended. pCMVORFs was constructed as above with a smaller PvuII–AccI Klenow-treated LINE fragment (nt 796–5964) containing both ORFs. An expression vector for the ORF2 protein, pCMVORF2, was constructed by a three-fragment ligation including a 199 bp HphI–BglII fragment and a 4010 bp BglII–AccI fragment, both derived from pCMVORFs, and the pCMV vector cleaved by NotI and AccI, after Klenow treatment of the HphI and NotI ends. In pCMVORF2^* the RT domain was inactivated by removing the fragment between the two KpnI sites (nt 3454 and 4659).

Deletion derivatives of the LINE ORF2. A His-tagged pCMVHisORF2 plasmid was first constructed which allowed 5'-deletion within ORF2 while keeping the same ATG (from the His-tag) for all expression vectors. It was constructed from pBacHisORF2 (a gift from M.I.Thoulouze, constructed by inserting a BstUI–AflIII fragment from L1.2A into pBlueBacHisA; Invitrogen) by inserting a NdeI–AflII fragment from pBacHisORF2, containing part of ORF2 with the histidine tag, into pCMVORF2 opened by EagI and AflII, after Klenow treatment of the NdeI and EagI ends.

ORF2 N-terminal deletions. N-terminal deletions within ORF2 were generated from pCMVHisORF2 by in-frame deletions between unique restriction sites at the 3'-end of the His tag (ClaI or EagI) and restriction sites in ORF2 (nucleotide positions refer to L1.2A sequence). pCMVHisORF2[ClaI–PmlI] and pCMVHisORF2[ClaI–BglII] were derived from pCMVHisORF2 by excision of the indicated fragments (PmlI, nt 2128; BglII, nt 2172), followed by religation of the blunt-ended vectors. pCMVHisORF2[EagI–AseI], pCMVHisORF2[EagI–BstXI], pCMVHisORF2[EagI–EcoRI] and pCMVHisORF2[EagI–KpnI] (AseI, nt 2410; BstXI, nt 2605; EcoRI, nt 3425; KpnI, nt 3454) were constructed by re-inserting respectively the AseI–AflII, BstXI–AflII, EcoRI–AflII and KpnI–AflII fragments from the N-terminal ORF2 domain into pCMVHisORF2 opened by EagI and AflII (after Klenow treatment of the non-AflII ends). pCMVHisORF2[ClaI–BspMI] (BspMI, nt 2751) was constructed by inserting a BspMI–AflII fragment into pCMVHisORF2 opened by ClaI and AflII (after Klenow treatment of ClaI and BspMI). pCMVHisORF2[EagI–NheI] and pCMVHisORF[EagI–EcoRV] (NheI, nt 3118; EcoRV, nt 3185) were constructed using pCMVHisORF2, which was first modified upon restriction with ClaI, Klenow treatment and religation, to have insertions of NheI–AflII or EcoRV–AflII fragments from ORF2 in-frame; these fragments were inserted into the modified vector (opened by EagI and AflII) after Klenow treatment of the NheI and EagI ends.

Additional N-terminal deletion derivatives were constructed using PCR products generated with 5' primers (Z1, Z2 and Z3, at position 2908, 2972 and 3030 respectively, all including a ClaI site at their 5'-end) positioned within domains lacking appropriate restriction sites and a 3' primer in ORF2 (RT4). PCR fragments were restricted with ClaI and BclI and introducted into pCMVHisORF2 opened by ClaI at the homologous BclI site, to create pCMVZ1, pCMVZ2 and pCMVZ3. Primers used were: Z1, 5'-ATGATCGATTGAGAAAGCAGGAAAGATCC; Z2, 5'-AACATCGATTGCAAGAGCAAACACATTC; Z3, 5'-GAGATCGATTGAAGGAAATAGAGA; RT4, 5'-ATTGAACCAGCCTTGCATCCC. Amplified fragments were sequenced with an Applied Biosystems apparatus.

ORF2 C-terminal deletions. A first C-terminal deletion was generated from pCMVHisORF2 upon restriction with SpeI and EcoRV and religation after blunt ending. Other mutants were generated by bidirectional exonuclease III digestion of pCMVHisORF2 restricted with SpeI and religation of the linearized plasmids after blunt ending with S1 nuclease, using standard procedures. Deletions (extending from 360 to 1000 bp) resulted in plasmids pCMVHisORF2360, pCMVHisORF2650 and pCMVHisORF21000, which were characterized by restriction mapping. Expression vectors for MoMLV proteins and for HIV RT. Construct pCMV-GAG-POL, a 5430 bp fragment encoding the GAG and POL proteins from MoMLV, was excised from plasmid pCRIPenv^- (Danos and Mulligan, 1988) after complete EcoRI and partial HindIII digestion. The fragment was blunt ended by Klenow treatment and inserted into plasmid pCMV opened with XhoI and NotI and blunt ended by Klenow treatment.

Plasmid pCMV-MoMLV-RT contains a SacI–HindIII fragment of the MoMLV pol gene, encoding an active RT (Jean-Jean et al., 1989) under control of the CMV early promoter. It was derived from pMoRT1 (Jean-Jean et al., 1989) by substitution of the EcoRI–SacI adenovirus promoter fragment with the EcoRI–XhoI fragment from pCMV containing the CMV promoter.

Expression vectors for the HIV RT were pCMV66 and pSV51 (Ansari-Lari and Gibbs, 1994), which encode the two subunits, p66 and p51 respectively, of the heterodimeric HIV RT. They were a gift from Drs M.A.Ansari-Lari and R.A.Gibbs.

Construction of intron-marked reporter elements. The CMVneoRT reporter gene was constructed by inserting the neoRT indicator gene (Heidmann et al., 1988; Heidmann and Heidmann, 1991) as a SalI blunt-ended fragment between the CMV promoter and the SV40 polyadenylation sequence of pCMV restricted by XhoI and NotI and Klenow-treated (Maestre et al., 1995). CMVneoRTLINE3' was constructed by inserting a 1850 bp BsmI fragment from pL1.2B (Dombroski et al., 1991), containing the 3'-part of the cloned L1.2B, including the poly(A) tail, into the unique SacII site (3' of neoRT) of pCMVneoRT after Klenow treatment of both insert and vector. CMVneoRTLINE3'EE was constructed by digestion of CMVneoRTLINE3' with EcoNI and EcoRV and self-ligation after blunt ending, thus eliminating a large part of the ORF2 sequence. CMVLINEneoRT was constructed by inserting a 2 kb NotI–DraI fragment from pL1.2A, which contains the 5'-untranslated region and the entire ORF1 sequence (NotI is located in the polylinker 5' of L1.2A) into the ClaI site of CMVneoRT. CMVLINEORF1neoRT was derived from CMVLINEneoRT by deletion of the XhoI–BglII fragment within ORF1 and self-ligation.

The TNF reporter gene (Neel et al., 1995) contains genomic DNA from the murine TNF gene extending from exons 1 to 4 under control of the CMV promoter and ended by a reduced polyadenylation sequence from the rabbit -globin gene.

Cells and transfections

Cells were grown in 5% CO₂ at 37°C in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS; Gibco-BRL) for NTera2D1 (Andrews, 1988), 293 (Graham et al., 1977) and 3T3 murine cells and in DMEM supplemented with 5% FCS for 3TDM1 murine cells (Nicolas et al., 1981) and G355-5 feline cells (a gift from Dr J.F.Nicolas). For PCR detection 510⁵ cells (210⁶ for NTera2D1) per 60 mm dish were transfected by the calcium phosphate method with 5 g neoRT-marked construct and 5 g RT expression vector. For the in vitro RT activity assay 510⁵ 3T3 murine cells were transfected by the lipofectamine method (Gibco-BRL) with 4 g RT expression vector and 16 l lipofectamine.

Nucleic acid purification and PCR amplification

Two to 4 days after transfection genomic cellular DNAs were extracted using procedures described in Heidmann and Heidmann (1991). To prevent amplification of the numerous transfected and unspliced copies of neoRT-marked elements genomic DNAs were first digested with BamHI, which cuts in the intronic sequence of the indicator gene. PCR amplification of the intronic region was performed in 50 l containing 50 mM KCl, 1.5 mM MgCl₂, 0.001% gelatin (w/w), 0.2 mM each dNTP, 1 M each primer, 1 U Taq polymerase (Amersham) and 1.5 g DNA (equivalent to 310⁵ cells). After an initial step at 94°C (3 min), 30 cycles of amplification (60 s at 68°C, 90 s at 72°C, 60 s at 93°C) were carried out with primers neo1 (5'-CGGCATCAGAGCAGCCGATTGTCTG) and tk1 (5'-GGCGCGGGGGGTCCGAGGTCCACTTC). To determine whether the cDNA copies were double-stranded, 100 ng genomic DNA were digested with a set of two restriction enzymes, BglII and XbaI (8 U each), which cut between the two neo1 and tk1 primers, and PCR was performed as above or with primers d2 (5'-TGCCCAGGGACCACCGACCCAC) and tk2 (5'-GGCCTCGAACACCGAGCGACCCTG) as a positive control. The products were analysed by electrophoresis in 1.8% agarose gels. Primer extension experiments were performed with the tk1 primer alone for the indicated number of cycles (45 cycles) and 5 l of this reaction product was thereafter PCR amplified for 23 cycles using primers neo2 (5'-CCAGTCATAGCCGAATAGCCTCTCC) and tk2; controls for RNA (within putative RNA–DNA hybrids) as possible templates were performed by prior treatment of the extracted nucleic acids with RNase A (Sigma; 10 g at 30°C for 1 h) following denaturation at 95°C for 10 min or with 0.3 M NaOH and neutralization or upon treatment with Escherichia coli RNase H (USB; 5 U at 37°C for 1 h). In the assays for reverse transcription of cellular genes PCR amplification was performed as above, except that the number of cycles was increased to 35. Primers used for TNF, E1A and c-myc were respectively: TNF1 (5'-CTAGACCCCGCACAGCAGGTCTCCACATGACACTGCTC) and TNF2 (5'-AGGCTCCAAAGAATACACTGCT); E1A1 (5'-TCCGGTTTCTATGCCAAACCTTGTAC) and E1A2 (5'-GTTAAGCAAGTCCTCGATACATTC); myc1 (5'-CCCTACCCTCTCAACGACAG), myc2 (5'-CCAACTCCGGGATCTGGTCA), myc3 (5'-CTGGTGCTCCATGAGGAGAC) and myc4 (5'-TCCTCTGGCGCTCCAAGAC) (two nested PCRs were performed in the latter case).

In vitro assay for RT activity using synthetic templates and primers

Two days post-transfection 210⁶ 3T3 cells were washed with phosphate-buffered saline solution. From this step all operations were carried out at 4°C. Cells were scraped from their support and suspended in 300 l TKCM buffer (50 mM Tris–HCl, pH 7.6, 10 mM KCl, 3 mM MgCl₂, 5 mM dithiothreitol, 1 mg/ml Pefablock; Boehringer). Cells were homogenized with a Dounce B homogenizer and nuclei pelleted by centrifugation at 2000 r.p.m.. The supernatant (cytosol) was clarified by centrifugation at 200 000 g for 15 min in a TL 100 Beckman rotor and used as a crude cellular extract; at this step 200 l crude extract were collected. The RT assay was performed as previously described (Mathias et al., 1991) with 15 l crude extract in 50 l reaction mixture (with 50 mM KCl). Micrococcal nuclease pretreatment was performed prior to the RT assay, except for the assay in 0.6 mM Mn²⁺. The assay was carried out at 37°C and incorporation of radiolabelled nucleoside triphosphate ([³H]dTTP) on poly(rA).oligo(dT) template was determined after 15 and 30 min. Aliquots of 20 l of the reactions were spotted onto cellulose filters (DE81 Whatman), washed as described and radioactivity determined by liquid scintillation counting.

We wish to acknowledge especially Dr H.H.Kazazian for the gift of the pL1.2A and pL1.2B plasmids containing the human LINE elements, Dr J.-M.Rossignol for the gift of the pMoRT1 plasmid encoding the MoMLV RT, Drs Ansari-Lari and Gibbs for the gift of the pCMV66 and pSV51 plasmids encoding the two subunits of the HIV RT, Drs D.Weil and F.Dautry for the TNF plasmid and primers and Dr E.Prost for the E1A primers. We thank Dr C.Lavialle for comments and critical reading of the manuscript. This work was supported by grants from the CNRS (ACC-SV3), the Association pour la Recherche sur le Cancer (contract 6552) and Rhone-Poulenc Rorer (Bioavenir contract).

Andrews PW (1988) Human teratocarcinomas. Biochim Biophys Acta, 948, 17–36. | Article | PubMed | ChemPort |

Ansari-Lari A and Gibbs R (1994) Analysis of HIV type 1 reverse transcriptase expression in a human cell line. AIDS Res Hum Retroviruses, 10, 1117–1124. | PubMed |

Barzilay G and Hickson D (1995) Structure and function of apurinic/apyrimidinic endonucleases. BioEssays, 17, 713–719. | Article | PubMed | ISI | ChemPort |

Boeke JD, Garfinkel DJ, Styles CA and Fink GR (1985) Ty elements transpose through an RNA intermediate. Cell, 40, 491–500. | Article | PubMed | ISI | ChemPort |

Coffin JM (1996) Retroviridae: the viruses and their replication. In Fields,B.N. et al. (eds), Fields Virology. Raven Publishers, Philadelphia, PA, pp. 1767–1847.

Danos O and Mulligan RC (1988) Safe and efficient generation of recombinant retroviruses with amphotropic and ecotropic host ranges. Proc Natl Acad Sci USA, 85, 6460–6464. | Article | PubMed | ChemPort |

Derr LK, Strathern JN and Garfinkel DJ (1991) RNA-mediated recombination in S. cerevisiae. Cell, 67, 355–364. | Article | PubMed | ISI | ChemPort |

Dombroski BA, Mathias SL, Nanthakumar E, Scott AF and Kazazian HH (1991) Isolation of an active human transposable element. Science, 254, 1805–1808. | Article | PubMed | ISI | ChemPort |

Dombroski BA, Feng Q, Mathias SL, Sassaman DM, Scott AF, Kazazian HH and Boeke JD (1994) An in vivo assay for the reverse transcriptase of human retrotransposon L1 in Saccharomyces cerevisiae. Mol Cell Biol, 14, 4485–4492. | PubMed | ISI | ChemPort |

Doolittle RF, Feng DF, Johnson MS and McClure MA (1989) Origins and evolutionary relationships of retroviruses. Q Rev Biol, 64, 1–30. | Article | PubMed | ISI | ChemPort |

Dornburg R and Temin HM (1990) cDNA genes formed after infection with retroviral vector particles lack the hallmarks of natural processed pseudogenes. Mol Cell Biol, 10, 68–74. | PubMed | ISI | ChemPort |

Eickbush TH (1992) Transposing without ends: the non-LTR retrotransposable elements. New Biologist, 4, 430–440. | PubMed | ISI | ChemPort |

Eickbush TH (1994) Origin and evolutionary relationships of retroelements. In Morse,S. (eds), The Evolutionary Biology of Viruses. Raven Press, New York, NY, pp. 121–157.

Fanning T and Singer M (1987) The Line-1 DNA sequences in four mammalian orders predict proteins that conserve homologies to retrovirus proteins. Nucleic Acids Res, 15, 2251–2260. | PubMed | ISI | ChemPort |

Feng Q, Moran JV, Kazazian HH and Boeke JD (1996) Human L1 retrotransposon encodes a conserved endonuclease required for retrotranposition. Cell, 87, 905–916. | Article | PubMed | ISI | ChemPort |

Gabriel A and Boeke JD (1993) Retrotransposon reverse transcription. In Skalka,A.M. and Goff,S.P. (eds), Reverse Transcriptase. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 275–328.

Graham FL, Smiley J, Russel WC and Nairn R (1977) Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J Gen Virol, 36, 59–72. | Article | PubMed | ISI | ChemPort |

Heidmann O and Heidmann T (1991) Retrotransposition of a mouse IAP sequence tagged with an indicator gene. Cell, 64, 159–170. | Article | PubMed | ISI | ChemPort |

Heidmann T, Heidmann O and Nicolas JF (1988) An indicator gene to demonstrate intracellular transposition of defective retroviruses. Proc Natl Acad Sci USA, 85, 2219–2223. | Article | PubMed | ChemPort |

Hohjoh H and Singer M (1996) Cytolasmic ribonucleoprotein complexes containing human LINE-1 protein and RNA. EMBO J, 15, 630–639. | PubMed | ISI | ChemPort |

Inouye S and Inouye M (1993) Bacterial reverse transcription. In Skalka,A.M. and Goff,S.P. (eds), Reverse Transcriptase. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 391–410.

Inouye S, Hsu MY, Eagle S and Inouye M (1989) Reverse transcriptase associated with the biosynthesis of the branched RNA-linked msDNA in Myxococcus xanthus. Cell, 56, 709–717. | Article | PubMed | ChemPort |

Isel C, Lanchy J-M, Le Grice SFJ, Ehresmann C, Ehresmann B and Marquet R (1996) Specific initiation and switch to elongation of human immunodeficiency virus type 1 reverse transcription require the post-transcriptional modifications of primer tRNA₃^Lys. EMBO J, 15, 917–924. | PubMed | ISI | ChemPort |

Jean-Jean O, Moyret C, Bernard D, Recondo A and Rossignol J (1989) Expression of an enzymatically active murine retroviral reverse transcriptase in human cells. Biochem Biophys Res Commun, 158, 595–602. | PubMed | ChemPort |

Jensen S and Heidmann T (1991) An indicator gene for detection of germline retrotransposition in transgenic Drosophila demonstrates RNA-mediated transposition of the LINE I element. EMBO J, 10, 1927–1937. | PubMed | ISI | ChemPort |

Kennell J, Wang H and Lambowitz AM (1994) The Mauriceville plasmid of Neurospora spp. uses novel mechanisms for initiating reverse transcription in vivo. Mol Cell Biol, 14, 3094–3107. | PubMed | ISI | ChemPort |

Kotewicz ML, Sampson CM, D'Alessio J and Gerard GF (1987) Isolation of cloned Moloney murine leukemia virus reverse transcriptase lacking ribonuclease H activity. Nucleic Acids Res, 16, 265–277.

Kuiper MTR and Lambowitz AM (1988) A novel reverse transcriptase activity associated with mitochondrial plasmids of Neurospora. Cell, 55, 693–704. | Article | PubMed | ChemPort |

Lambowitz AM and Belfort M (1993) Introns as mobile elements. Annu Rev Biochem, 62, 587–622. | Article | PubMed | ISI | ChemPort |

Lanchy J-M, Ehresmann C, Le Grice SFJ, Ehresmann B and Marquet R (1996) Binding and kinetic properties of HIV-1 reverse transcriptase markedly differ during initiation and elongation of reverse transcription. EMBO J, 15, 7178–7187. | PubMed | ISI | ChemPort |

Levin H (1997) It's prime time for reverse transcriptase. Cell, 88, 5–8. | Article | PubMed | ISI | ChemPort |

Levin JG, Crouch RJ, Post K, Hu SC, McKelvin D, Zweig M, Court DL and Gerwin BI (1988) Functional organization of the murine leukemia virus reverse transcriptase: characterization of a bacterially expressed AKR DNA polymerase deficient in RNase H activity. J Virol, 62, 4376–4380. | PubMed | ChemPort |

Levine KL, Steiner B, Johnson K, Aronoff R, Quinton TJ and Linial ML (1990) Unusual features of integrated cDNAs generated by infection with genome-free retroviruses. Mol Cell Biol, 10, 1891–1900. | PubMed | ISI | ChemPort |

Luan DD, Korman MH, Jakubczak JL and Eickbush TH (1993) Reverse transcription of R2Bm RNA is primed by a nick at the chromosomal target site: a mechanism for non-LTR retrotransposition. Cell, 72, 595–605. | Article | PubMed | ISI | ChemPort |

Maestre J, Tchénio T, Dhellin O and Heidmann T (1995) mRNA retroposition in human cells: processed pseudogene formation. EMBO J, 14, 6333–6338. | PubMed | ISI | ChemPort |

Martin F, Maranon C, Olivares M, Alonso C and Lopez M (1995) Characterisation of a non-long terminal repeat retrotransposon cDNA (L1Tc) from Trypanosoma cruzi: homology of the first ORF with the APE family of DNA repair enzymes. J Mol Biol, 247, 49–59. | Article | PubMed | ISI | ChemPort |

Martin F, Olivares M and Lopez MC (1996) Do non-long terminal repeat retrotransposons have nuclease activity? Trends Biochem Sci, 21, 283–285. | Article | PubMed | ChemPort |

Martin SL (1991a) LINEs. Curr Opin Genet Dev, 1, 505–508. | ChemPort |

Martin SL (1991b) Ribonucleoprotein particles with LINE-1 RNA in mouse embryonal carcinoma cells. Mol Cell Biol, 11, 4804–4807. | PubMed | ISI | ChemPort |

Mathias SL, Scott AF, Kazazian HH, Boeke JD and Gabriel A (1991) Reverse transcriptase encoded by a human transposable element. Science, 254, 1808–1810. | Article | PubMed | ISI | ChemPort |

McClure MA (1991) Evolution of retroposon by acquisition or deletion of retrovirus-like genes. Mol Biol Evol, 8, 835–856. | PubMed | ISI | ChemPort |

McClure MA (1993) Evolutionary history of reverse transcriptase. In Skalka,A.M. and Goff,S.P. (eds), Reverse Transcriptase. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 425–444.

McMillan J and Singer M (1993) Translation of the human LINE-1 element, L1Hs. Proc Natl Acad Sci USA, 90, 11533–11537. | PubMed | ChemPort |

Michel F and Lang B (1985) Mitochondrial class II introns encode proteins related to the reverse transcriptases of retroviruses. Nature, 316, 641–643. | Article | PubMed | ISI | ChemPort |

Moran JV, Holmes SE, Nass TP, DeBerardinis RJ, Boeke JD and Kazazian HH (1996) High frequency retroposition in cultured mammalian cells. Cell, 87, 917–927. | Article | PubMed | ISI | ChemPort |

Nargang F, Bell J, Stohl L and Lambowitz AM (1984) The DNA sequence and genetic organization of a Neurospora mitochondrial plasmid suggest a relationship to introns and mobile elements. Cell, 38, 441–453. | Article | PubMed | ISI | ChemPort |

Neel H, Gondran P, Weil D and Dautry F (1995) Regulation of pre-mRNA processing by src. Curr Biol, 5, 413–422. | PubMed | ChemPort |

Nicolas J-F, Jakob H and Jacob F (1981) Teratocarcinoma-derived cell lines and their use in the study of differentiation. In Sato,G. (ed.), Functionally Differentiated Cell Lines. A.R.Liss, New York, pp. 185–210.

Pélisson A, Finnegan DJ and Bucheton A (1991) Evidence for retrotransposition of the I factor, a LINE element of Drosophila melanogaster. Proc Natl Acad Sci USA, 88, 4907–4910. | PubMed | ISI | ChemPort |

Schwarz-Sommer Z, Leclercq L, Göbel E and Saedler H (1987) Cin4, an insert altering the structure of the A1 gene in Zea mays, exhibits properties of nonviral retrotransposons. EMBO J, 6, 3873–3880. | PubMed | ChemPort |

Ségal-Bendirdjian E and Heidmann T (1991) Evidence for a reverse transcription intermediate for a marked LINE transposon in tumoral rat cells. Biochem Biophys Res Commun, 181, 863–870. | PubMed |

Stephens C and Harlow E (1987) Differential splicing yields novel adenovirus 5 E1A mRNAs that encode 30 kDa and 35 kDa proteins. EMBO J, 6, 2027–2035. | PubMed | ISI | ChemPort |

Tanese NT and Goff SP (1988) Domain structure of the Moloney MuLV reverse transcriptase: mutational analysis and separate expression of the polymerase and RNase H activities. Proc Natl Acad Sci USA, 85, 1777–1781. | Article | PubMed | ChemPort |

Tchénio T and Heidmann T (1991) Defective retroviruses can disperse in the human genome by intracellular transposition. J Virol, 65, 2113–2118. | PubMed | ISI | ChemPort |

Tchénio T and Heidmann T (1992) High-frequency intracellular transposition of a defective mammalian provirus detected by an in situ colorimetric assay. J Virol, 66, 1571–1578. | PubMed |

Tchénio T, Ségal-Bendirdjian E and Heidmann T (1993) Generation of processed pseudogenes in murine cells. EMBO J, 12, 1487–1497. | PubMed | ISI |

Teng S-C, Kim B and Gabriel A (1996) Retrotransposon reverse-transcriptase-mediated repair of chromosomal breaks. Nature, 383, 641–644. | Article | PubMed | ISI

Wang H and Lambowitz AM (1993) The Mauriceville plasmid reverse transcriptase can initiate cDNA synthesis de novo and may be related to reverse transcriptase and DNA polymerase progenitor. Cell, 75, 1071–1081. | Article | PubMed | ISI | ChemPort |

Xiong Y and Eickbush TH (1990) Origin and evolution of retroelements based upon their reverse transcriptase sequences. EMBO J, 9, 3353–3362. | PubMed | ISI | ChemPort |

Zimmerly S, Guo H, Perlman P and Lambowitz AM (1995) Group II intron mobility occurs by target DNA-primed reverse transcription. Cell, 82, 545–554. | Article | PubMed | ISI | ChemPort |

Main navigation

• Search EMBO Journal

  Search

• Advanced search

• The EMBO Journal

  • Home
  • Aims and scope
  • Current issue
  • Advance online publication
  • Web focuses
  • Archive:
  • Browse by issue
  • Browse by subject
  • Browse by category
  • Free online sample issue
  • Press releases

• Authors and Referees

  • Editorial process
  • Guide for authors
  • Submit an article
  • Guide for referees
  • Editors and Editorial Board
  • Contact editorial office

• Customer Service

  • Subscribe
  • Order sample copy
  • Purchase articles
  • Reprints and permissions
  • Contact NPG
  • Advertising