The Evolutionary Analysis of the Tnt1 Retrotransposon in Nicotiana Species Reveals the High Variability of its Regulatory Sequences

Transcription

1 The Evolutionary Analysis of the Tnt1 Retrotransposon in Nicotiana Species Reveals the High Variability of its Regulatory Sequences Samantha Vernhettes,* Marie-Angèle Grandbastien,* and Josep M. Casacuberta *Laboratoire de Biologie Cellulaire, Institut National de la Recherche Agronomique, Versailles, France; and Departament de Genètica Molecular, Centre d Investigació i Desenvolupament (Consejo Superior Investigaciones Cientificas), Barcelona, Spain We studied the evolution of the tobacco Tnt1 retrotransposon by analyzing Tnt1 partial sequences containing both coding domains and U3 regulatory sequences obtained from a number of Nicotiana species. We detected three different subfamilies of Tnt1 elements, Tnt1A, Tnt1B, and Tnt1C, that differ completely in their U3 regions but share conserved flanking coding and LTR regions. U3 divergence between the three subfamilies is found in the region that contains the regulatory sequences that control the expression of the well-characterized Tnt1 94 element. This suggests that expression of the three Tnt1 subfamilies might be differently regulated. The three Tnt1 subfamilies were present in the Nicotiana genome at the time of species divergence, but have evolved independently since then in the different genomes. Each Tnt1 subfamily seems to have conserved its ability to transpose in a limited and different number of Nicotiana species. Our results illustrate the high variability of Tnt1 regulatory sequences. We propose that this high sequence variability could allow these elements to evolve regulatory mechanisms in order to optimize their coexistence with their host genome. Introduction Retrotransposons are mobile genetic elements closely related to retroviruses. Both types of elements share structural characteristics and are functionally related (Boeke and Corces 1989). Replication of retroviruses is a very error-prone process because of the lack of proofreading activity of the enzymes involved. As a consequence, the replication of a single virus generates a population of closely related, but different, sequences (Domingo and Holland 1994). This dynamic population structure, which is referred to as quasispecies (Domingo et al. 1985), endows RNA genomes with extremely high adaptive capacity (Domingo and Holland 1994; Domingo et al. 1996). Retrotransposition seems to be as error prone as retroviral replication (Gabriel et al. 1996). We have already shown that the expression of the tobacco retrotransposon Tnt1, one of the very few plant retrotransposons that has been shown to be active (Grandbastien, Spielmann, and Caboche 1989), generates a population of RNA sequences that resembles RNA viral quasispecies (Casacuberta, Vernhettes, and Grandbastien 1995). However, the influence of this population structure on Tnt1 evolution has not been evaluated. Tnt1 expression is highly regulated, and its RNA is only detectable in a few tissues and under certain conditions (Pouteau et al. 1991). The expression of the Tnt1 94 element has been studied in detail and shown to be induced in association with plant defense mechanisms (Pouteau, Grandbastien, and Boccara 1994; Moreau-Mhiri et al. 1996). This restricts Tnt1 activity to stress situations, greatly reducing its mutagenic capacity. The control of transcription thus seems to play a key role in the control of Tnt1 activity. The cis-regulatory elements that control Tnt1 94 transcription are located Key words: Tnt1, retrotransposon, Nicotiana, sequence variability, transcriptional regulation, evolution. Address for correspondence and reprints: Josep M. Casacuberta, Departament de Genètica Molecular, CID (CSIC), Jordi Girona 18, Barcelona, Spain. jcsgmp@cid.csic.es. Mol. Biol. Evol. 15(7): by the Society for Molecular Biology and Evolution. ISSN: within the U3 region of the 5 long terminal repeat (LTR) (Casacuberta and Grandbastien 1993; Grandbastien et al. 1997; Mhiri et al. 1997; Vernhettes, Grandbastien, and Casacuberta 1997), and previous reports have shown that its transcriptional induction could be mediated by short repeated elements present within this region (Casacuberta and Grandbastien 1993; Casacuberta, Vernhettes, and Grandbastien 1995; Grandbastien et al. 1997; Vernhettes, Grandbastien, and Casacuberta 1997). In this report, we study the evolution of Tnt1 elements in the Nicotiana genus. Our results show that the U3 regulatory regions of Tnt1 have diverged to give rise to different Tnt1 subfamilies which, on the contrary, have conserved their flanking coding and LTR regions much more constant. This specific divergence of the transcriptional regulatory regions suggests that Tnt1 sequence can be molded by environmental influences. We discuss the importance of the high sequence variability of the regulatory regions of Tnt1 for the evolution of this element in the Nicotiana genus. Materials and Methods Plant Material All Nicotiana species were obtained from the SEI- TA Institut du Tabac in Bergerac, France. DNA Extraction, PCR Amplification, Cloning, and Sequencing DNA from the different Nicotiana species was extracted as described (Dellaporta, Wood, and Hicks 1984). PCR amplifications were performed with the same oligonucleotides and the same conditions used to amplify Tnt1 sequences from tobacco DNA, as published in Casacuberta, Vernhettes, and Grandbastien (1995). Cloning and sequencing of the amplified sequences was performed as previously described (Casacuberta, Vernhettes, and Grandbastien 1995). Phylogenetic Analysis Sequences were aligned using the CLUSTAL W multiple-alignment program (version 1.5; Thompson, 827

2 828 Vernhettes et al. FIG. 1. A, Diagram of the Tnt1 region analyzed. The approximate positions of the oligonucleotides used for PCR amplification are shown by closed arrows. The different regions into which we have divided the analyzed sequences are indicated at the top. B, Phylogenetic relationships between Nicotiana species, as reported by Goodspeed (1954). The species used in this work are shown in bold. Higgins, and Gibson 1994) with some minor refinements. DNADIST in Felsenstein s (1989) PHYLIP package was used to generate a distance matrix based on the Jukes-Cantor algorithm (Jukes and Cantor 1969). This was used to generate neighbor-joining trees (Saitou and Nei 1987). Bootstrap analyses were performed using the Seqboot and Consense programs from Felsenstein s (1989) PHYLIP package. Parsimony trees were obtained using the PAUP program (Swofford 1993). Sequence variability as measured by Nei s (1987) measure of nucleotide diversity,, was calculated using the DnaSP program (Rozas and Rozas 1997). Results Presence of Tnt1 Sequences Within Different Nicotiana Species We analyzed the Tnt1 elements in different Nicotiana species by cloning and sequencing PCR products obtained from different Nicotiana genomic DNA. The region analyzed was 400 nt divided into four different zones: RT, linker, box, and end (see fig. 1A). The RT region comprises 50 bp of the 3 end of the Tnt1 coding region. The linker region contains the 50 bp of the 3 untranslated region between the open reading frame (ORF) and the 3 LTR. The box region consists of the first 240 bp of the U3 region, from the beginning of the LTR up to the TATA box. This region is known to contain most of the elements responsible for the defense-associated expression shown for Tnt1 94 (Casacuberta and Grandbastien 1993; Mhiri et al. 1997; Vernhettes, Grandbastien, and Casacuberta 1997; Grandbastien et al. 1997). The end region comprises the 30 bp of the 3 end of the U3 region, the R region, and the first 20 bp of U5. In total, 140 sequences were analyzed, with 20 sequences obtained from each of 7 different Nicotiana species chosen as representatives of the major evolutionary branches of the Nicotiana genus (see fig. 1B). Two different phylogenetic methods (neighbor joining and parsimony; see Materials and Methods) were used to analyze the 140 sequences. Both methods showed that the sequences cluster into three major subfamilies, which we have named Tnt1A, Tnt1B, and Tnt1C. Figure 2 shows the results of the neighbor-joining analysis. Tnt1A and Tnt1C clusters are supported by high bootstrap values (100 and 88, respectively), and they can be considered monophyletic groups. On the contrary, cluster B, which groups the rest of the sequences, is not supported by a high bootstrap value (less

3 Tnt1 Evolution in Nicotiana Species 829 FIG. 2. Phylogenetic analysis of the Tnt1 partial sequences; neighbor-joining tree obtained comparing the 20 Tnt1 partial sequences obtained from each of seven different Nicotiana genomes. Nicotiana glauca sequences are indicated by glau, N. tomentosiformis sequences by tom, N. tabacum sequences by tab, N. sylvestris sequences by syl, N. plumbaginifolia sequences by plumb, N. benthamiana sequences by ben, and N. debneyi sequences by deb. A vertical bar with a capital letter on the right of the scheme indicates the division of the sequences into three different subfamilies. Bootstrap values above 50% supporting major clusters are shown. Horizontal distances are proportional to evolutionary divergence expressed in substitutions per hundred sites. Vertical lines are for clarity.

4 830 Vernhettes et al. FIG. 3. Phylogenetic analysis of the Tnt1 partial sequences belonging to (A) Tnt1A, (B) Tnt1B and (C) Tnt1C subfamilies. The tree was constructed according to the neighbor-joining method. The different groups of each Tnt1 subfamily are shown. Sequences are designated as in figure 2. Bootstrap values above 50% supporting the clusters are shown. Horizontal distances are proportional to evolutionary divergence expressed in substitutions per hundred sites. Vertical lines are for clarity. than 50) and cannot be considered a monophyletic group. Nevertheless, for simplicity, we will refer to these sequences as the Tnt1B subfamily. The 140 sequences analyzed represent Tnt1 elements, as they all contain very conserved RT and end regions (see below). Nevertheless, sequences belonging to different subfamilies contain highly divergent box regions (see below) and, consequently, it is very difficult to properly align the 140 sequences within this region. Thus, we also compared the sequences belonging to each subfamily separately. New alignments for each subfamily were performed and the analysis by neighbor joining is presented in figure 3. Within each of the three subfamilies, the sequences obtained from a given ge-

5 Tnt1 Evolution in Nicotiana Species 831 FIG. 3 (Continued) nome tend to group in the tree. Moreover, sequences from phylogenetically related genomes tend to be closer than those from more phylogenetically distant genomes. As an example, the distances between the sequences obtained from N. benthamiana and N. debneyi, which are phylogenetically closely related (see fig. 1B), are in general lower than the distances between sequences from any of these genomes and those from N. glauca, which is more distant in the phylogeny (see fig. 3). Nicotiana tabacum is an amphidiploid species derived from an in-

6 832 Vernhettes et al. terspecific cross of two parental species, one of which is the ancestor of N. sylvestris, and the other of which is phylogenetically closely related to the ancestor of N. tomentosiformis (Goodspeed 1954). Nineteen out of 20 sequences obtained from N. sylvestris and 15 out of 20 sequences obtained from N. tabacum belong to the Tnt1A subfamily and plot together in the tree (see fig. 3A). In contrast, the tab4 sequence, which is the only Tnt1C sequence obtained from N. tabacum, falls into a group of Tnt1C sequences obtained from N. tomentosiformis (see fig. 3C). The only Tnt1B sequence obtained from N. sylvestris, syl16, plots together with tab2, a Tnt1B sequence obtained from N. tabacum (see fig. 3B). The high sequence similarity between glau8, an N. glauca sequence belonging to the C subfamily, and the rest of the sequences that cluster in the C2 group, which all belong to N. plumbaginifolia, is surprising. This could be interpreted as an example of a rare horizontal transmission event between N. glauca and N. plumbaginifolia. However, as glau8 is the only sequence belonging to the Tnt1C subfamily obtained from N. glauca, it is difficult to unambiguously interpret this result. Sequence Variability of the Three Tnt1 Subfamilies Within Different Nicotiana genomes The sequences belonging to the Tnt1A, Tnt1B, and Tnt1C subfamilies cluster into different groups (see fig. 3). We have calculated the sequence variability within each one of these groups using Nei s (1987) measure of nucleotide diversity,. The nucleotide diversity is very low for A1, B1, and C2 (values of of , , and , respectively), whereas it is much higher for other groups of sequences (for exampe, the value for A6 is , for B4 is , and for C1 is ). Interestingly, the variability of the sequences belonging to a given subfamily depends on the host genome. While most of the Tnt1A sequences obtained from N. tabacum (15 out of 16) and N. sylvestris (14 out of 19) appear in the highly homogeneous A1 group, all the Tnt1A sequences obtained from N. tomentosiformis fall into the A6 group, the most variable group of the Tnt1A subfamily. On the contrary, in N. plumbaginifolia, the Tnt1B subfamily seems to be very homogeneous (9 out of 12 sequences plot into the low variable B1 group), while it is highly variable in N. tabacum and N. glauca, whose Tnt1B sequences plot in three different highly variable groups, B2 (B value of ), B3 (B value of ), and B4. In N. plumbaginifolia, the Tnt1C subfamily is extremely homogeneous (5 out of 7 sequences cluster in the highly homogeneous C2 group), but it is much more variable in N. tomentosiformis (7 out of 10 sequences plot in the C4 group with a value of , and the remaining 3 sequences fall into the C1 group). The Three Tnt1 Subfamilies Are Highly Divergent in their Box Region As each Tnt1 subfamily is composed of groups of sequences of different variabilities, it is not easy to define a consensus sequence for each subfamily. We thus defined a consensus sequence from the most homogeneous group of each Tnt1 subfamily (which corresponds to Tnt1A1, Tnt1B1, and Tnt1C2) and we compared the three consensus sequences. As shown in figure 4, the RT regions are almost identical between the three consensus sequences (100% identity between Tnt1A1 and Tnt1B1 and 88% identity between these and Tnt1C2). The linker and the end regions also display a high degree of sequence similarity (the linker regions of Tnt1A1 and Tnt1B1 have 92% identity, those of Tnt1A1 and Tnt1C2 have 73% identity, and those of Tnt1B1 and Tnt1C2 have 75% identity; the end regions of Tnt1A1 and Tnt1B1 have 82% identity, those of Tnt1A1 and Tnt1C2 have 74% identity, and those of Tnt1B1 and Tnt1C2 have 67% identity). On the contrary, the box region of the three consensus sequences is highly divergent (less than 35% identity between any pair of sequences), although the sequences that belong to the same subfamily share the same box motifs (not shown). The three different Tnt1 subfamilies are thus defined by different box sequences flanked by much more conserved regions. It is interesting to note that all Tnt1A sequences contain motifs related to the tandemly repeated BII box, which has been characterized as a transcriptional regulator for Tnt1A (Casacuberta and Grandbastien 1993; Casacuberta, Vernhettes, and Grandbastien 1995, Vernhettes, Grandbastien, and Casacuberta 1997), although the number and conservation of such motifs differ between the different Tnt1A groups. Moreover, there is a clear relationship between the degree of homogeneity of a group and the number and conservation of the repeated BII motif within the box region of the sequences of that group. As an example, while 30% of the sequences that plot in the less variable A1 group have four BII boxes in their box regions, and no sequence has less than three BII boxes, 60% of the sequences in the highly variable A6 group contain only two BII tandemly repeated motifs, with the remaining 40% having only one such motif in their box region. In order to investigate the contribution of recombination events in the generation of the three box regions that define the three Tnt1 subfamilies, we also investigated the evolution of the different parts of the sequence. Figure 5 shows three examples of the phylogenetic trees obtained by comparing the sequences obtained from a given genome, analyzing the total amplified sequence (RT-linker-box-end), the 5 half (RT-linker), or the 3 half (box-end). The results obtained show that, although the variability of the RT-linker region is very low, the overall evolutionary patterns of this region are identical to those of the much more variable boxend region. Discussion Three Different Tnt1 Subfamilies Were Present in an Ancestral Nicotiana Genome. Our results show that the genomes of several Nicotiana species chosen as representatives of different branches of the genus (Goodspeed 1954) all contain se-

7 Tnt1 Evolution in Nicotiana Species 833 FIG. 4. Comparison of the consensus sequences for the low variable groups A1, B1, and C2. A scheme indicating the relative position of the region analyzed is shown above the corresponding sequence. Nucleotides that appear in at least two sequences are boxed. BI and BII motifs, known to control transcription of Tnt1A elements, are shown. The putative TATA-box is underlined. quences related to the Tnt1 retrotransposon. This illustrates the ubiquitous presence of Tnt1 in the Nicotiana genus and suggests that this element is an ancient component of this plant genus. This result is in accordance with previous observations that Tnt1-related sequences can be detected in other genera of the Solanaceae subfamily (Grandbastien, Spielmann, and Caboche 1989). The 140 Tnt1 partial sequences obtained plot into three different major subfamilies, which we have named Tnt1A, Tnt1B and Tnt1C. These three subfamilies of Tnt1 elements are present in the seven Nicotiana genomes analyzed. Although we have not cloned any Tnt1B sequence from N. tomentosiformis or any Tnt1C sequences from N. sylvestris, slot blot hybridizations performed with probes specific for each Tnt1 subfamily have revealed the presence of Tnt1B and Tnt1C sequences within these genomes (data not shown). Thus, since all the Nicotiana species analyzed contain sequences belonging to these three different subfamilies, the simplest model for their evolution is that they were

8 834 Vernhettes et al. FIG. 5. Phylogenetic relationships between the sequences of a single Nicotiana genome. The analyses were performed with the complete sequences (RT-linker-box-end) or only the 5 half (RT-linker) or the 3 half (box-end). The parsimony trees obtained with the sequences from N. tabacum (top), N. plumbaginifolia (middle), and N. benthamiana (bottom) are shown. present in the ancestral genome at the time of species divergence. Although retrotransposons are considered to be noninfectious, the possibility that they could be horizontally transmitted has been repeatedly suggested to explain incongruent retrotransposon and organismal phylogenies in previous studies (Flavell et al. 1992; Flavell, Smith, and Kumar 1992; Hirochika and Hirochika 1993; Flavell, Pearce, and Kumar 1994). Since plant copia-like retrotransposon populations are extremely heterogeneous and are usually partitioned into distinct lineages or families even within a single host, a limited random sampling can generate an incongruent phylogeny (Voytas et al. 1992; VanderWiel, Voytas, and Wendel 1993; Capy, Anxolabéhère, and Langin 1994). In this paper, we analyzed 140 sequences of very limited divergence (all the sequences represent Tnt1 elements) in closely related genomes. The phylogenetic relationships between the sequences obtained are, in general, compatible with the accepted phylogeny of the Nicotiana genus (Goodspeed 1954), indicating that vertical transmission is the major mechanism for Tnt1 persistence, in agreement with the results for copia-like elements in cotton (VanderWiel, Voytas, and Wendel 1993). High Variability of Tnt1 Regulatory Sequences The three different Tnt1 subfamilies, Tnt1A, Tnt1B and Tnt1C, are defined by completely different box regions. We have previously shown that the transcriptional regulatory elements that activate Tnt1 94 in association with plant defense mechanisms are located within this variable region of the U3 (Casacuberta and Grandbastien 1993; Mhiri et al. 1997). Two different transcriptional activator elements, BI and BII, have been characterized within the U3 region of Tnt1 94 (Casacuberta and Grandbastien 1993), and interactions of nuclear factors with one of them, BII, have been shown to be concomitant with the induction of Tnt1 94 in association with plant defense mechanisms (Vernhettes, Grandbastien, and Casacuberta 1997). The sequence of the box region, and especially those of the BI and BII elements, are very well conserved among Tnt1A elements expressed in tobacco protoplasts (Casacuberta, Vernhettes, and Grandbastien 1995). Moreover, tobacco Tnt1A populations expressed under two different situations (protoplasts and roots) have slightly different box regions (Casacuberta, Vernhettes, and Grandbastien 1995). Thus, the observation that the three Tnt1 subfamilies completely differ in their regulatory box regions, Tnt1B and Tnt1C sequences lacking BI and BII ele-

9 Tnt1 Evolution in Nicotiana Species 835 ments, strongly suggests that their expression is induced under different conditions. To our knowledge, this is the first description of different subfamilies of the same retrotransposon that differ in their transcriptional regulatory sequences. We have previously proposed that the activation of Tnt1A in association with plant defense responses could allow this element to transpose, and thus to persist, but to maintain its activity under a threshold in order to avoid compromising the viability of the host genome (Grandbastien et al. 1994, 1997). This could also be the reason why the activation of transposable elements by stress seems to be a common situation, and it would not be surprising if Tnt1B and Tnt1C elements were also activated by other types of local and transient stress situations. It is interesting to note that the appearance of the three different subfamilies of Tnt1 elements is the result of a long-lasting evolution in which the box region has diverged, while the flanking regions have been well conserved. Recombination does not seem to have played a major role in the generation of box-region diversity, as the overall evolutionary patterns of the different regions of the sequence are identical regardless of their respective degrees of sequence variability (see fig. 5). Retroviruses display an enormous degree of sequence variability, which is assumed to be the consequence of the high mutation rate associated with retroviral replication. Indeed, retroviral populations are made up of huge swarms of different but closely related sequences referred to as quasispecies (Domingo et al. 1985), and this endows these elements with extreme adaptability (Domingo and Holland 1994). We have recently shown that when Tnt1 is expressed, its RNA displays a population structure that resembles retroviral quasispecies (Casacuberta, Vernhettes, and Grandbastien 1995), suggesting that different types of retroelements, such as retrotransposons and retroviruses, could use similar adaptive mechanisms regardless of their infective capacity. Thus, the high mutation rate associated with the retrotransposition process could be a major player in Tnt1 U3 diversity, and we propose that Tnt1 quasispecies-like structure could give this element a high adaptive capacity. Differential Maintenance of Active Tnt1 Subfamilies During Evolution Within each of the three different Tnt1 subfamilies, the sequences obtained cluster into different groups of different nucleotide diversities. A limited nucleotide diversity may indicate that the sequences obtained represent recent insertions that have had no time to diverge, while more variable groups of sequences probably represent ancient insertions of elements that might have become defective. We have previously shown that the transcribed Tnt1A population is much more homogeneous than the genomic population of Tnt1 elements (Casacuberta, Vernhettes, and Grandbastien 1995). In addition, we have also shown that deletions of a tandemly repeated sequence, BII, probably accompanies the inactivation of Tnt1A elements (Casacuberta, Vernhettes, and Grandbastien 1995). We think, therefore, that the A1 group of the Tnt1A subfamily contains sequences that represent active elements, while the A6 group probably represents elements that are no longer active. As most Tnt1A sequences obtained from N. tabacum and N. sylvestris belong to the A1 group, it may be that this subfamily is still actively transposing in these genomes. This is in agreement with the fact that the three mobile Tnt1 elements cloned to date, from tobacco, belong to the Tnt1A subfamily (Grandbastien, Spielmann, and Caboche 1989; unpublished data), and their sequences also closely resemble those of the Tnt1A1 group (not shown). On the contrary, the Tnt1A subfamily has probably lost its ability to transpose in other genomes such as N. tomentosiformis, as all the Tnt1A sequences obtained from this genome are in the highly variable A6 group. Interestingly, Tnt1B sequences that are very heterogeneous in N. tabacum are, like Tnt1C sequences, extremely homogeneous in N. plumbaginifolia (see figs. 4 and 5). This may indicate that these subfamilies are still active in N. plumbaginifolia, while they have lost this ability in N. tabacum. This would be consistent with the fact that the two Tnt1 elements cloned to date after transposition in N. plumbaginifolia (C. Meyer, personal communication) both belong to the Tnt1B subfamily (not shown). It seems, therefore, that while at least three different subfamilies of Tnt1 elements were present in the genome of Nicotiana at the time of species divergence, they have evolved independently since then, leading to the amplification of different Tnt1 subfamilies in different species. As retrotransposons are highly mutagenic agents, a given host genome can probably support a limited number of active elements without compromising its viability. The success of a particular subfamily of Tnt1 elements may lead to the inactivation of other subfamilies, as if they were competing for a common ecological space, perhaps as predicted by the competitive exclusion principle of population biology (Gause 1971). As the three Tnt1 subfamilies are probably differently regulated, the success of different subfamilies within different hosts could simply be a consequence of a different activation story in each Nicotiana genome. Alternatively, a particular subfamily could be better expressed in a particular host genome, thus having more chances to persist. We are analyzing the regulation of representatives of the three Tnt1 subfamilies in different Nicotiana species in order to clarify this point. In summary, our results show that Tnt1 regulatory elements display a high sequence variability that could allow this retrotransposon to evolve and acquire different regulatory mechanisms, improving its coexistence with the genomes which it inhabits, and this could be the major reason for its persistence in Nicotiana genomes. Sequence Availability The nucleotide sequence data reported in this paper will appear in the EMBL, GenBank and DDBJ nucleotide sequence databases under the following accession numbers: AJ AJ for sequences tab1 tab20, AJ AJ for syl1 syl20, AJ AJ for plumb1 plumb20, AJ AJ for ben1 ben20, AJ AJ for deb1 deb20, AJ AJ for glau1 glau20, and AJ AJ for tom1 tom20.

10 836 Vernhettes et al. Acknowledgments We are grateful to Pere Puigdomènech (CID-CSIC) for his support in this work and for critical reading of the manuscript. We thank Jordi Gomez and Esteban Domingo for helpful discussions. We are indebted to José García- Martínez for his help in the pylogenetic analyses of Tnt1 sequences. We also thank Ian Drouau, Colette Audéon and Véronique Tanty for help in sequencing. S.V. was supported by a fellowship of the Ministère de la Recherche et de l Enseignement (MRE). Work done at the INRA (Versailles) was partly funded by a MERS Action Concertée Coordonnée (ACC-SV3, 1995) and by the EEC BIOTECH program BIO4-CT Work done at the CID-CSIC (Barcelona) was funded by Plan Nacional de Investigación Científica y Técnica (grant BIO94/0734 to Professor Puigdomènech). LITERATURE CITED BOEKE, J. D., and V. CORCES Transcription and reverse transcription of retrotransposons. Annu. Rev. Microbiol. 45: CAPY, P., D. ANXOLABÉHÈRE, and T. LANGIN The strange phylogenies of transposable elements: is horizontal transfer the only explanation? Trends Genet. 10:7 12. CASACUBERTA, J. M., and M.-A. GRANDBASTIEN Characterisation of LTR sequences involved in the protoplast specific expression of the tobacco Tnt1 retrotransposon. Nucleic Acids Res. 21: CASACUBERTA, J. M., S. VERNHETTES, and M.-A. GRANDBASTIEN Sequence variability within the tobacco retrotransposon Tnt1 population. EMBO J. 14: DELLAPORTA, S. L., J. WOOD, and J. HICKS Molecular biology of plants: a laboratory course manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. DOMINGO, E., C. ESCARMIS, N. SEVILLA, A. MOYA, S. F. ELENA, J. QUER, I. S. NOVELLA, and J. HOLLAND Basic concepts in RNA virus evolution. FASEB J. 10: DOMINGO, E., and J. J. HOLLAND Mutation rates and rapid evolution of RNA viruses. Pp in S. S. MORSE eds. Evolutionary biology of viruses. Raven Press, New York. DOMINGO, E., E. MARTINA-SALAS, F. SOBRINO et al. (14 co-authors) The quasispecies (extremely heterogeneous) nature of viral RNA genome populations: biological relevance a review. Gene 40:1 8. FELSENSTEIN, J., PHYLIP phylogeny inference package (version 3.56). Cladistics 5: FLAVELL, A. J., E. DUMBAR, R. ANDERSON, S. R. PEARCE, R. HARTLEY, and A. KUMAR Ty1-copia group retrotransposons are ubiquitous and heterogeneous in plants. Nucleic Acids Res. 20: FLAVELL, A. J., S. R. PEARCE, and A. KUMAR Plant transposable elements and the genome. Curr. Opin. Genet. Dev. 4: FLAVELL, A. J., D. R. SMITH, and A. KUMAR Extreme heterogeneity of Ty1-copia group of retrotransposons in plants. Mol. Gen. Genet. 231: GABRIEL, A., M. WILLEMS, E.H.MULES, and J. D. BOEKE Replication infidelity during a single cycle of Ty1 retrotransposition. Proc. Natl. Acad. Sci. USA 93: GAUSE, G. F The struggle for existence. Dover Publications, New York. GOODSPEED, T. H The genus Nicotiana. Chronica Botanica Company, Waltham, Mass. GRANDBASTIEN, M.-A., C. AUDEON, J. M. CASACUBERTA, P. GRAPPIN, H. LUCAS, and S. POUTEAU Functional analysis of the tobacco Tnt1 retrotransposon. Genetica 93: GRANDBASTIEN, M.-A., H. LUCAS, J.-B. MOREL, C. MHIRI, S. VERNHETTES, and J. M. CASACUBERTA The expression of the tobacco Tnt1 retrotransposon is linked to the plant defense responses. Genetica 100: GRANDBASTIEN, M.-A., A. SPIELMANN, and M. CABOCHE Tnt1, a mobile retroviral-like transposable element of tobacco isolated by plant cell genetics. Nature 37: HIROCHIKA, H., and R. HIROCHIKA Ty1-copia group retrotransposons as ubiquitous components of plant genomes. Jpn. J. Genet. 68: JUKES, T. H., and C. R. CANTOR Evolution of protein molecules. Pp in H. N. MUNRO, ed. Mammalian protein metabolism. Academic Press, New York. MHIRI, C., J.-B. MOREL, S.VERNHETTES,J.M.CASACUBERTA,H. LUCAS, and M.-A. GRANDBASTIEN The promoter of the tobacco Tnt1 retrotransposon is induced by wounding and abiotic stress. Plant Mol. Biol. 33: MOREAU-MHIRI, C., J.-B. MOREL, C. AUDEON, M. FERAULT, M.- A. GRANDBASTIEN, and H. LUCAS Regulation of the expression of the tobacco Tnt1 retrotransposon in heterologous species following pathogen-related infection. Plant J. 9: NEI, M Molecular evolutionary genetics. Columbia University Press, New York. POUTEAU, S., M.-A. GRANDBASTIEN, and M. BOCCARA Microbial elicitors of plant defence responses activate transcription of a retrotransposon. Plant J. 5: POUTEAU, S., E. HUTTNER, M.-A. GRANDBASTIEN, and M. CA- BOCHE Specific expression of the tobacco Tnt1 retrotransposon in protoplasts. EMBO J. 10: ROZAS, J., and R. ROZAS DnaSP version 2.0: a novel software package for extensive molecular population genetic analysis. Comput. Appl. Biosci. 13: SAITOU, N., and N. NEI The neighbour-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4: SWOFFORD, D. L PAUP: phylogenetic analysis using parsimony. Version 3.0. Illinois Natural History Survey, Champaign. THOMPSON, J. D., D. G. HIGGINS, and T. J. GIBSON CLUS- TAL W: improving the sensitivity of progressive multiple sequence alignement through sequence weighting, populationspecific gap penalties and weight matrix choice. Nucleic Acids Res. 22: VANDERWIEL, P. L., D. F. VOYTAS, and J. F. WENDEL Copia-like retrotransposable element evolution in diploid and polyploid cotton (Gossypium L.). J. Mol. Evol. 36: VERNHETTES, S., M.-A. GRANDBASTIEN, and J. M. CASACUBERTA In vivo characterisation of transcriptional regulatory sequences involved in the defence-associated expression of the tobacco retrotransposon Tnt1. Plant Mol. Biol. 36: VOYTAS, D.F.,M.P.CUMMINGS, A.KONIECZNY,F.M.AUSUBEL, and S. R. RODERMEL Copia-like retrotransposons are ubiquitous among plants. Proc. Natl. Acad. Sci. USA 89: PIERRE CAPY, reviewing editor Accepted March 16, 1998