|Home | About | Journals | Submit | Contact Us | Français|
Ty1 retrotransposons of the yeast Saccharomyces cerevisiae are activated by different kinds of stress. Here we show that Ty1 transcription is stimulated under severe adenine starvation conditions. The Bas1 transcriptional activator, responsible for the induction of genes of the de novo AMP biosynthesis pathway (ADE) in the absence of adenine, is not involved in this response. Activation occurs mainly on Ty1 elements, whose expression is normally repressed by chromatin and is suppressed in a hta1-htb1Δ mutant that alters chromatin structure. Activation is also abolished in a snf2Δ mutant. Several regions of the Ty1 promoter are necessary to achieve full activation, suggesting that full integrity of the promoter sequences might be important for activation. Together, these observations are consistent with a model in which the activation mechanism involves chromatin remodeling at Ty1 promoters. The consequence of Ty1 transcriptional activation in response to adenine starvation is an increase in Ty1 cDNA levels and a relief of Ty1 dormancy. The retrotransposition of four native Ty1 elements increases in proportion to their increase in transcription. Implications for the regulation of Ty1 mobility by changes in Ty1 mRNA levels are discussed.
Retrotransposons are a class of mobile genetic elements that replicate through an RNA intermediate and resemble retroviruses from many points of view. Five different families of retrotransposons (Ty1 to Ty5) have been identified in the genome of the yeast Saccharomyces cerevisiae (reviewed in references 31, 48, and 57). They all share the same basic structure, consisting of two direct long terminal repeats (LTR) flanking the TYA and TYB open reading frames, which are analogous to the retroviral gag and pol genes. Ty1 is the most abundant of the retrotransposon families, with around 30 copies per haploid genome, and is a model for LTR-containing elements. Ty1 is transcribed from LTR to LTR by RNA polymerase II, with the resulting transcript serving as a template for both translation and reverse transcription. Translation leads to the synthesis of Ty1Ap, the structural component of the virus-like particle and of the Ty1A-Ty1B polyprotein containing protease (PR), integrase (IN), reverse transcriptase (RT), and RNase H (RH) catalytic domains, all of which are essential for retrotransposition. The linear double-stranded cDNA molecule, produced by reverse transcription within the virus-like particle, enters the genome either by integrase-mediated integration or, to a lesser extent, by homologous recombination with genomic elements. Ty1 preferentially integrates next to RNA polymerase III-dependent promoters, but less frequent insertions into, or upstream of, genes transcribed by RNA polymerase II have been reported. In these cases, Ty1 insertions have been shown to alter the expression of the neighboring genes by activating or inactivating them or by importing new regulatory mechanisms on their expression. Such insertions may provide a means of evolution to the yeast genome. The Ty1 element and its host have evolved numerous control systems that keep retrotransposition at a low level, yet allowing it to be activated under stress (10, 31, 57). In contrast to the host defense, which acts mainly at the posttranscriptional level, the Ty1 response to stress generally involves activation of transcription.
The promoter of Ty1 is complex, and its structure is very similar to that of higher eukaryotes. It extends over 1 kb, both upstream and downstream of two TATA boxes, and includes the 5′ LTR and part of the TY1A open reading frame. At least eight transcription factors (Gcr1, Ste12, Tec1, Mcm1, Tea1, Rap1, Gcn4, and Mot3), which bind to the Ty1 promoter (15, 16, 21, 22, 29, 33, 36, 37, 56), and three chromatin-remodeling complexes (Swi/Snf, SAGA, and ISWI) regulate Ty1 transcription in haploid cells (7, 20, 24, 28, 42). The activation of Ty1 transcription by the invasive/filamentous pathway in response to environmental signals, such as nitrogen starvation, occurs through the transcriptional activators Ste12 and Tec1 (9, 37). These proteins recognize a sequence called FRE (filamentous responsive element), located in TY1A sequences downstream of the TATA boxes of the Ty1 promoter (2). In addition to their role in Ty1 activation by the invasive/filamentous signaling pathway, Ste12 and Tec1 are important for basal levels of Ty1 transcription in haploid cells (29, 37). Exposure of yeast cells to DNA-damaging agents also increases Ty transcript levels and activates Ty retrotransposition (5, 35, 44, 47, 53). The mechanism of transcriptional activation by DNA damage has not been elucidated, however.
Insights into the transcription of individual Ty1 elements have recently been obtained using a set of haploid strains, each expressing lacZ from the transcription control signals of a specific Ty1 element at its native location. This set of 31 strains allows study of the transcription of all but one of the Ty1 elements present in the genome of S288C (36). The comparison of lacZ expression in these strains identified two basic classes of Ty1 elements according to their level of expression: weakly and highly expressed elements. Based on genetic data, it was proposed that repression of transcription of the highly expressed Ty1 elements by chromatin structures is antagonized by Swi/Snf and SAGA. In addition, several endogenous Ty1 elements, mostly those expressed at high levels, contain five potential Gcn4 binding sites in their 5′ LTR (36). Gcn4 is a transcriptional activator that binds to multiple sites upstream of amino acid biosynthetic genes (25). The transcription of the highly expressed Ty1 elements depends on GCN4 in the absence of amino acids and is activated when GCN4 is overexpressed.
The Bas1 transcriptional activator recognizes a DNA sequence (TGACTC) (13, 55) similar to the Gcn4 binding site [TGA(C/G)TCA] (39). The difference is that an A nucleotide is generally present at the 3′ extremity of Gcn4 binding sites but is never found in Bas1 binding sites. All but one of the Gcn4 binding sites located in Ty1 elements contain a T nucleotide at the last position and could be recognized by Bas1, suggesting that Bas1 could potentially activate Ty1 transcription. Bas1, together with Bas2, is required for the regulated activation of the ADE genes of the de novo AMP biosynthesis pathway (13). When adenine is provided in the environment, it enters the yeast cells and is converted to AMP and then to other adenine nucleotides. In the absence of purines, Bas1 and Bas2 interact to activate the ADE genes (40, 60). Since the Gcn4 binding sites located in Ty1 were potentially recognized by Bas1, we asked whether Ty1 transcription was stimulated under conditions of adenine starvation.
We discovered that although Bas1 does not activate Ty1 transcription, severe adenine starvation does. Our results indicate that activation occurs mainly on poorly expressed elements, whose transcription is repressed by chromatin. We also show that the activation of transcription of individual Ty1 elements under severe adenine starvation correlates with a proportional increase in their retrotransposition. Finally, we provide evidence that the activation mechanism requires chromatin remodeling at Ty1 promoters.
Yeast strains were grown in rich yeast extract-peptone-dextrose (YPD), Hartwell's synthetic complete (HC), and synthetic minimum (SDc; SD minimal medium containing arginine, isoleucine, tryptophan, leucine, and valine) media, all supplemented with 2% glucose (1, 45). Adenine was added to a final concentration of 0.3 mM (high adenine) or 0.025 mM (low adenine) or not at all (no adenine), as indicated. The final concentration of 0.3 mM (high adenine) represses the de novo AMP biosynthesis pathway in wild-type strains, while the absence of adenine or 0.025 mM adenine does not. Yeast transformations were performed by the lithium acetate procedure.
Yeast strains used in this work are described in Table Table1.1. Strains carrying the different Ty1-lacZ fusions are FYBL1-23D derivatives. Ty1 elements are named according to the Ty1 sequence annotation by the Munich Information Center for Protein Sequences, i.e., TY1A(PR1)-lacZ corresponds to lacZ fused to the first Ty1 element of chromosome XVI on the right arm of the centromere (http://mips.gsf.de/genre/proj/yeast/). Strains containing a TY1A-lacZ fusion at the chromosomal locus of the different native Ty1 elements have already been described (36). Briefly, lacZ is fused in frame to TY1A (at coordinate 1571 of Ty1H3 (3), upstream of the URA3 gene and followed by sequences of TY1B (coordinates 2171 to 3726 of Ty1H3 (3), except for Ty1(ML2), in which URA3 is followed immediately by sequences downstream of that element. These fusions harbor all of the Ty1 transcription control sequences. The TY1Amir-lacZ fusion at Ty1(DR3) and Ty1(ML2) has the same structure as the TY1A-lacZ fusion, except that TY1A sequences between coordinates 818 and 926 of Ty1H3 (3), which contain the Mcm1, Tea1, and Rap1 binding sites, have been replaced by a DNA fragment containing three Gal4 binding sites. Consequently, the TY1A open reading frame is modified over 23 amino acids, but the frame is conserved. Strains containing the TY1AF-lacZ fusion at Ty1(DR3), Ty1(ML1), Ty1(ML2), and Ty1(PR1) loci were obtained upon transformation of FYBL1-23D with a Ty1up-TY1AF-lacZ-URA3-′ty1b′ DNA fragment. For each Ty1 element, the Ty1up-TY1AF portion was obtained by high-fidelity PCR amplification (Roche) of a region covering 100 bp upstream of the element and Ty1 sequence up to coordinate +425 of Ty1-H3 (just downstream of the Ste12 and Tec1 binding sites). Sequences were checked for errors introduced during the amplification process before subsequent cloning upstream of the lacZ-URA3-′ty1b′ portion. The Ty1up sequence allows targeting of the integration to the right Ty1 locus. In TY1A-lacZ, TY1Amir-lacZ, and TY1AF-lacZ, the lacZ-URA3-′ty1b′ organization is identical.
Strains containing a his3AI indicator of transposition at Ty1(DR5), Ty1(GR1), Ty1(NL2), and Ty1(PR1) were obtained by transforming FYBL1-23D with each of the four ′ty1b-his3AI-LTR-URA3-downTy1 constructs. In these constructs, the ty1b′-his3AI-LTR sequence was obtained by high-fidelity PCR amplification (Triple Master; Eppendorf), using the plasmid pGTy1-H3his3AI as a template, such that the his3AI indicator gene is located at the last codon of TY1B (12). In these constructs, TY1B sequences encompass coordinate 5455 to the end of Ty1H3 (3). Sequences were checked for errors introduced during the amplification process. The downTy1 portions correspond to a sequence of approximately 400 nucleotides immediately downstream of either Ty1(DR5), Ty1(GR1), Ty1(NL2), or Ty1(PR1) that was amplified by PCR from genomic DNA. Targeted integration at these four elements was obtained by homologous recombination through the ty1b and downTy1 sequences.
Null alleles of BAS1, ADE2, TEA1, SNF2, and GAL4 were obtained in these strains by one-step gene replacement, using a PCR fragment of the TRP1, HphMX, or KanMX cassettes flanked with 5′ and 3′ sequences of BAS1, ADE2, SNF2, and GAL4. The spt3-101 allele was introduced in FYBL1-23D by the pop-in/pop-out gene replacement method, using the integrative plasmid pFW33 (URA3 spt3-101), kindly provided by F. Winston. Gene replacement was confirmed by Southern blot analysis.
All constructs and gene replacements were checked by PCR analysis.
Plasmids pAM19 (GCN4 HIS3 CEN), Ycp50, and pRS313 have already been described (36, 46, 52). Plasmids p79 (BAS1 URA3 CEN), p115 (ADE1-lacZ URA3 2μm) and p473 (ADE1-lacZ LEU2 2μm) were kindly provided by B. Daignan-Fornier (13, 23, 41). Plasmid p11-4/HIS3 (STE11-4 HIS3 CEN) was a generous gift from G. Fink (32).
Precultures were grown to saturation under high-adenine conditions in liquid HC or SDc medium at 30°C, washed in sterile water, and diluted 100-fold in the same medium, in the presence or absence of adenine. Cultures were grown to mid-log phase at 22°C. The media used in each experiment are indicated in the corresponding figure legends. Total cellular RNA or proteins were extracted from 10-ml culture aliquots according to procedures already described (37).
For Northern blot analysis, 10 μg of each RNA sample was loaded onto a 1% agarose-1× Tris-borate-EDTA gel. The size-fractionated RNA was transferred to a Hybond-XL membrane (Amersham). Transfer and hybridization procedures were performed as recommended by the supplier. Probes against Ty1, Ty1-lacZ (lacZ probe), ADE1, and ACT1, were generated by random priming (Roche). The Ty1 probe was derived from a region of Ty1 with no homology to Ty2 (coordinates 3137 to 3682 in Ty1H3 (3). Results were quantified on a Molecular Dynamics PhosphorImager and using ImageQuant software.
β-Galactosidase assays were performed as already described (37). β-Galactosidase units are expressed in nanomoles of 2-nitrophenyl-β-d-galactopyranoside hydrolyzed per minute per milligram of protein. Values are averages of at least three independent measurements. Standard deviations were ≤10%.
Cells were grown at 30°C overnight in HC (high-adenine) medium. Precultures were washed in water and diluted 100-fold into 25 ml of HC medium in the presence or absence of adenine and grown for four generations at 22°C. Aliquots were plated on YEPD and HC media lacking histidine to determine the fraction of His+ prototrophs.
Cells were harvested following the same procedure as for transposition assays. Total genomic DNA was extracted from 10 ml of culture according to the methods described in reference 1 and digested with PvuII restriction endonuclease. DNA samples were subjected to electrophoresis on 1% agarose gel and transferred to a Hybond-XL membrane (Amersham). Transfer and hybridization procedures were performed as recommended by the supplier. A [32P]DNA probe for the TY1B sequence was generated by random priming on a PCR fragment whose sequence is included in the 2.0-kb PvuII fragment of the unincorporated Ty1 cDNA (coordinates 3992 to 5432 in Ty1H3 (3). Results were visualized with ImageQuant software and a Molecular Dynamics PhosphorImager.
The possibility that the Bas1 transcriptional activator might recognize Ty1 promoter sequences prompted us to study the role of Bas1 in Ty1 transcription. Bas1 interacts with Bas2 to activate the ADE genes in the absence of extracellular adenine (13, 40, 60). Deletion of either gene decreases the basal expression of ADE genes and impairs their activation. bas1Δ (and bas2Δ) cells produce low levels of adenine nucleotides, allowing growth in the absence of adenine, albeit at a slower rate than wild-type cells. To determine whether BAS1 activates Ty1 transcription, steady-state Ty1 mRNA levels were compared by Northern blot analysis in wild-type and bas1Δ haploid cells grown in the presence and in the absence of adenine (Fig. (Fig.1A).1A). The ADE1 gene was used as a positive control, since it was previously shown that an ADE1-lacZ reporter fusion carried on a 2μm plasmid is strongly activated by BAS1 in the absence of adenine (13). Although ADE1 mRNA was almost undetectable in wild-type cells grown in the presence of adenine, a faint signal was detected when wild-type cells were grown in the absence of adenine. As expected, this induction was dependent on BAS1, since no signal corresponding to ADE1 mRNA was detected in bas1Δ cells. Steady-state Ty1 mRNA levels were similar in wild-type cells grown in the presence and in the absence of adenine. There was also no difference in Ty1 mRNA levels between wild-type and bas1Δ cells grown in the presence of adenine. However, steady-state Ty1 mRNA levels were markedly elevated when bas1Δ cells were grown in the absence of adenine. This increase in Ty1 transcription suggests that in the absence of extracellular adenine, BAS1 might act as a repressor of Ty1 transcription, instead of as an activator. However, there is no previous evidence that BAS1 represses transcription, either directly or indirectly. In particular, an analysis of the yeast proteome by two-dimensional gels revealed no spot which increased in intensity in bas1Δ cells in the absence of adenine among 900 identifiable gene products (14). The increase in the amount of Ty1 mRNA in adenine-deprived bas1Δ cells could alternatively indicate that adenine starvation activates Ty1 transcription by a mechanism independent of BAS1. In this latter hypothesis, the deletion of BAS1 would exacerbate the deficiency in adenine nucleotides, since the ADE genes cannot be activated.
The GCN4 transcriptional activator was a potential candidate for the activation of Ty1 transcription under conditions of adenine starvation for the following reasons. First, although amino acid starvation is the major activation signal of GCN4 expression, other kinds of stress can activate GCN4, one of them being purine limitation (25, 45). Second, GCN4 overexpression increases the transcription of several Ty1 elements, in particular those containing five Gcn4 potential binding sites in their promoter sequences. The transcriptional activation of these elements by Gcn4 is accompanied by an increase in Ty1 mRNA levels (36). To test the involvement of GCN4 in the activation of Ty1 transcription in bas1Δ cells starved for adenine, we compared steady-state Ty1 mRNA levels in gcn4Δ bas1Δ double mutant cells grown in the presence and in the absence of adenine. We observed the same fivefold increase in Ty1 mRNA levels obtained with the bas1Δ single mutant (compare Fig. 1A and B). There was also no difference in Ty1 mRNA levels in gcn4Δ cells grown in the presence and in the absence of adenine (Fig. (Fig.1B).1B). Together, these results establish that GCN4 is not responsible for the activation Ty1 transcription under conditions of adenine starvation.
To determine whether starvation for adenine nucleotides per se activates Ty1 transcription by a BAS1-independent mechanism, we used a strain wild type for BAS1 but harboring a complete deletion of ADE2, which blocks the production of adenine nucleotides (Fig. (Fig.2A).2A). Since the ade2Δ strain is unable to grow in the complete absence of adenine, starvation conditions were created by growing the cells in limiting amounts of adenine that were sufficient to allow their growth for several hours but which were not high enough to repress the de novo AMP biosynthesis pathway in wild-type strains (see Materials and Methods). Derepression of the pathway under adenine-limiting conditions was monitored using the β-galactosidase activity of an ADE1-lacZ reporter fusion carried on a 2μm plasmid. This approach was chosen instead of Northern analysis of ADE1 mRNA levels, since this mRNA only gave faint signals on blots (Fig. (Fig.1A1A and data not shown). As expected, ADE1-lacZ expression increased severalfold in wild-type cells when adenine was limiting and was constitutively low in ade2Δ cells (Table (Table2).2). The poor expression of ADE1-lacZ in ade2Δ cells is explained by the absence of 5′-phosphoribosyl-4-succinocarboxamide-5-aminoimidazole (SAICAR). This metabolic intermediate is produced in the de novo AMP biosynthesis pathway downstream of ADE2 activity (Fig. (Fig.2A)2A) and has been shown to be critical for the activation of ADE genes by Bas1 and Bas2 (43). Hence, mutations affecting the steps of the pathway upstream of SAICAR production impair the induction of ADE gene expression under low-adenine concentrations, while mutations acting downstream of SAICAR synthesis lead to constitutive derepression of ADE gene expression. In Northern analysis, a 3.5-fold increase in Ty1 mRNA levels was observed in ade2Δ cells grown in limiting amounts of adenine, indicating that Ty1 transcription is activated when the de novo AMP biosynthesis pathway is blocked (Fig. (Fig.2B).2B). The increase in Ty1 mRNA levels was slightly higher in ade2Δ cells than in bas1Δ cells, probably because the de novo AMP biosynthesis pathway, which still functions at basal levels in bas1Δ cells, produces low levels of adenine nucleotides in these cells. These results demonstrate that Ty1 transcription is activated under conditions of adenine starvation, when the cells are unable to overcome the adenine deficiency. They also imply that BAS1 is not directly involved in the activation mechanism, although its absence exacerbates the conditions of adenine starvation, which in turn causes Ty1 derepression.
We also determined Ty1 mRNA levels in the presence of the ade13-52 mutation by Northern analysis. This defective allele has only 1% of adenylosuccinate activity compared to wild type and impairs growth in the complete absence of adenine (43). The Ade13 protein acts in two steps of the de novo AMP biosynthesis pathway. First, it metabolizes SAICAR to AICAR (5′-phosphoribosyl-4-carboxamide-5-aminoimidazole). Thus, the ade13-52 allele leads to SAICAR accumulation in the cells and constitutive activation of ADE genes by Bas1p and Bas2p. Thus, as expected, ADE1-lacZ expression was strongly elevated in ade13-52 compared to wild type, independently of adenine availability (Table (Table2).2). Second, it acts in one of the two steps that transform IMP to AMP. These steps are essential for extracellular adenine to be transformed to adenine nucleotides (Fig. (Fig.2A).2A). Therefore, in ade13-52 cells, adenine nucleotide pools remain low, even in the presence of adenine in the environment. Consistent with our hypothesis, Ty1 mRNA levels were constitutively higher in ade13-52 cells than in wild-type cells (Fig. (Fig.2C).2C). This result, together with the increase of Ty1 mRNA levels observed in ade2Δ and bas1Δ adenine-depleted cells, demonstrates that Ty1 transcription is activated under conditions that severely decrease cellular adenine nucleotide pools. For simplicity, we will refer to this situation as “severe adenine starvation.”
To learn more about this newly discovered mode of regulation of Ty1 transcription by adenine levels, we decided to use a set of S288C-derived strains in which each expresses a lacZ chromosomal fusion from the transcriptional signal of a different endogenous Ty1 element. In these strains, lacZ is fused to TY1A, downstream of the Ty1 promoter sequences (36). These strains allowed us previously to identify two classes of endogenous Ty1 elements, according to their expression level, and to understand certain aspects of Ty1 transcription regulation (36, 37). Based on our previous findings that the highly expressed and poorly expressed Ty1 elements are differently regulated, we asked whether severe adenine starvation activated all elements or only a specific class. For this purpose, the expression of 11 TY1A-lacZ fusions representative of the two classes was monitored in wild-type and bas1Δ strains grown in the presence or in the absence of adenine. Consistently with the Northern data described above, adenine starvation did not modify the β-galactosidase activity of the fusions in wild-type strains but significantly increased the activity of all except one of the fusions in bas1Δ cells (Fig. (Fig.3A).3A). Interestingly, the level of activation of the TY1A-lacZ fusions in bas1Δ cells in the absence of adenine varied in most cases with their level of expression in the presence of adenine. Poorly expressed fusions were activated between 3- and 20-fold, while highly expressed fusions were activated less than 2-fold (Fig. (Fig.3A).3A). This difference in activation suggests that adenine starvation preferentially activates Ty1 elements expressed at low levels. That most of these elements lack Gcn4 binding sites is consistent with the finding that Gcn4 does not contribute to the regulation. In the presence of adenine, the β-galactosidase activity of several lacZ fusions [at Ty1(ML1), Ty1(DR1), Ty1(PR2), and Ty1(OR)] was lower in bas1Δ cells than in wild-type cells. This difference in expression was observed in other experiments (see Fig. Fig.5B5B and 6B and D, below) but was not systematic and is probably not significant.
We also compared the expression of two fusions, one expressed at low levels, TY1A(ML1)-lacZ, and one expressed at high levels, TY1A(ML2)-lacZ, in wild-type, ade2Δ, bas1Δ, and bas1Δ ade2Δ cells. Since the ade2Δ and the ade2Δ bas1Δ strains are auxotrophic for adenine, starvation conditions were created by cultivating the cells in limiting amounts of adenine. The β-galactosidase activity of the weakly expressed TY1A(ML1)-lacZ fusion increased by a factor of 6.9, 4.5, and 9.8 in bas1Δ, ade2Δ, and bas1Δ ade2Δ cells, respectively, while the expression of the TY1A(ML2)-lacZ fusion did not respond to adenine limitation in these three mutants (Fig. (Fig.3B).3B). Failure to activate TY1A(ML2)-lacZ during adenine limitation provided further confirmation that activation preferentially affects weakly expressed Ty1 elements.
We also analyzed the expression level of the weakly expressed TY1A(ML1)-lacZ fusion in response to decreasing levels of adenine in a bas1Δ strain. Although the expression was similar in the presence of 0.3 mM, 0.15 mM, and 0.075 mM adenine (0.5 U of β-galactosidase), it progressively increased 2-fold, 6-fold, and 10-fold in the presence of 0.05 mM, 0.025 mM, and without adenine, respectively (1, 2.9, and 5.8 U of β-galactosidase, respectively). This result confirms that Ty1 transcription is activated in response to reduced pools of adenine nucleotides, independently of Bas1p. It also explains why Ty1 mRNA levels are more abundant in bas1Δ cells grown in the complete absence of adenine than those grown in a limiting amount of adenine (compare Fig. Fig.1A1A and and2B2B).
Lastly, Northern blot analysis was performed on wild-type and bas1Δ strains containing either the weakly expressed TY1A(ML1)-lacZ fusion or the highly expressed TY1A(ML2)-lacZ fusion. The presence of the fusions did not alter the activation of Ty1 transcription under conditions of adenine starvation: in the absence of adenine, total Ty1 mRNA levels increased six- to eightfold in bas1Δ strains carrying TY1A(ML1)-lacZ and TY1A(ML2)-lacZ, respectively (Fig. (Fig.3C).3C). Due to the relative weakness of the Ty1 promoter (36), the signals corresponding to the TY1A-lacZ transcripts were very faint and did not allow a precise quantification of their level under the different conditions. Nevertheless, TY1A-lacZ transcripts derived from TY1A(ML2)-lacZ could be detected in wild-type strains, whereas those derived from TY1A(ML1)-lacZ could not, consistent with the difference in β-galactosidase levels expressed from the fusions (Fig. (Fig.3A).3A). Furthermore, in bas1Δ cells in the absence of adenine, TY1A-lacZ mRNA levels derived from both fusions increased, confirming that individual Ty1 elements are activated under conditions of adenine starvation (Fig. (Fig.3C3C).
Ty1 RNA serves as a template for reverse transcription to synthesize Ty1 cDNA (4). Thus, we wondered whether Ty1 cDNA levels would increase under conditions of severe adenine starvation, as a consequence of the activation of Ty1 transcription. Ty1 cDNA levels were measured in wild-type and bas1Δ cells grown in the presence and in the absence of adenine, using a Southern blot assay developed to detect unintegrated Ty1 cDNA (49) (Fig. (Fig.4).4). In this assay, the TY1B radiolabeled probe hybridizes to a 2.0-kb fragment of unintegrated Ty1 cDNA and numerous other larger fragments, corresponding to various Ty1 elements joined to genomic DNA (Fig. (Fig.4A).4A). A fus3Δ mutant accumulates Ty1 cDNA and served as positive control in this experiment, while a spt3-101 mutant that impairs Ty1 transcription served as negative control (Fig. (Fig.4B).4B). We could barely detect Ty1 cDNA in wild-type cells, regardless of the growth conditions, or in bas1Δ cells grown in the presence of adenine. In contrast, unincorporated Ty1 cDNA was detected in bas1Δ cells grown in the absence of adenine. Ty1 cDNA levels were approximately 15-fold higher in bas1Δ cells grown in the absence of adenine compared to those grown in the presence of adenine. Therefore, the increase in Ty1 mRNA levels under conditions of severe adenine starvation is followed by a significant increase in Ty1 cDNA levels.
An increase in Ty1 cDNA levels is generally associated with a hyper-retrotransposition phenotype (reviewed in references 31 and 57). We therefore asked whether the concomitant increase in Ty1 mRNA and Ty1 cDNA levels in bas1Δ cells starved for adenine resulted in more retrotransposition events. As the activation is more efficient on weakly expressed Ty1 elements, we decided to compare the retrotransposition frequency of endogenous elements expressed at different levels. For this purpose, we chose two endogenous elements expressed at low levels, Ty1(DR5) and Ty1(NL2), and two elements expressed at high levels, Ty1(PR1) and Ty1(GR1). In bas1Δ cells, transcription of these elements fused to lacZ was activated by the absence of adenine by 7.4-, 2.9-, 1.9-, and 0.8-fold, respectively (Fig. (Fig.3A).3A). To determine the retrotransposition frequency of these elements, the his3AI indicator gene was introduced into each of them (see Materials and Methods). Cells bearing a Ty1his3AI-tagged element that undergoes retrotransposition are detected as His+ prototrophs (12). This assay is specific for retrotransposition, as the formation of a functional HIS3 gene requires both splicing and reverse transcription of the Ty1his3AI transcript.
In wild-type cells, Ty1(NL2), which is expressed at low levels, had a retrotransposition frequency 1 order of magnitude lower than Ty1(GR1) and Ty1(PR1), which are expressed at high levels (Table (Table3).3). In contrast, the weakly expressed Ty1(DR5) element retrotransposed at a frequency similar to Ty1(PR1) and Ty1(GR1). Thus, the frequency of His+ formation was not systematically proportional to the relative transcriptional level of each element. This could be the consequence of differences in the sequences of these elements that affect steps in their life cycle downstream of transcription.
There was no change in the retrotransposition frequency of any of the four elements when wild-type cells were grown in the absence of adenine. In bas1Δ cells, however, the absence of adenine increased the retrotransposition frequency of Ty1(DR5)his3AI 30-fold, while it did not substantially affect the retrotransposition frequency of Ty1(NL2)his3AI, Ty1(PR1)his3AI, or Ty1(GR1)his3AI. Therefore, conditions of severe adenine starvation preferentially activate the mobility of Ty1 elements whose expression is strongly activated in the absence of adenine. In conclusion, the activation of Ty1 transcription in bas1Δ cells under conditions of adenine starvation is correlated with a relatively proportional increase in Ty1 retrotransposition.
To get further insights into the mechanism of regulation of Ty1 transcription by adenine, we searched for Ty1 sequences necessary for the control. The cis-acting region of Ty1 transcription regulation extends over 1 kb, upstream and downstream of the TATA box, and is present in the TY1A-lacZ fusions used in the preceding experiments (Fig. (Fig.5A).5A). To test the involvement of the 5′ portion of the Ty1 promoter, lacZ was fused just downstream of the FRE site in the weakly expressed elements Ty1(DR3) and Ty1(ML1) and the highly expressed elements Ty1(ML2) and Ty1(PR1) (Fig. 5A and B). In wild-type cells, the β-galactosidase activities produced by the four TY1AF-lacZ fusions were markedly higher than the activities observed with the corresponding TY1A-lacZ fusions containing all the regulatory regions (compare Fig. Fig.5B5B with 3A). The differences in activity between the two sets of fusions could be due to different protein stabilities or activities, since the hybrid Ty1A-β-galactosidase proteins expressed from the two sets of fusions differ in the Ty1A portion. It might alternatively indicate that regions downstream of the FRE site act negatively on Ty1 transcription. Nevertheless, the hierarchy of expression between the fusions was conserved, indicating that these fusions still reflect the relative transcription level of the endogenous Ty1 elements.
In bas1Δ cells, the absence of adenine activated the expression of TY1A(DR3)F-lacZ and TY1A(ML1)F-lacZ threefold and had no significant effect on TY1A(ML2)F-lacZ and TY1A(PR1)F-lacZ. The persistence of regulation of the weakly expressed elements suggests that sequences in the 5′ portion of Ty1 promoter might be involved in the activation of Ty1 transcription by severe adenine starvation. Among these sequences, the FRE site recognized by Ste12 and Tec1 was a potential candidate, since this site is involved in the activation of Ty1 transcription in response to environmental stress that induces the invasive/filamentous pathway (37). To determine whether STE12 was necessary for the activation of Ty1 transcription by adenine starvation, the expression of TY1A(DR3)-lacZ and TY1A(ML2)-lacZ fusions containing the whole regulatory region was monitored in bas1Δ cells and in bas1Δ ste12Δ cells grown in the absence and in the presence of adenine. Consistent with the major role of STE12 in Ty1 basal-level transcription (37), the STE12 deletion led to a strong decrease in the activity of the highly expressed TY1A(ML2)-lacZ fusion (Fig. (Fig.5C).5C). The decrease in expression was less obvious with the TY1A(DR3)-lacZ fusion, which is already expressed at low levels in wild-type cells. The absence of adenine resulted in a similar increase in expression of TY1A(DR3)-lacZ in ste12Δ bas1Δ and in bas1Δ cells, while the expression of TY1A(ML2)-lacZ actually showed a greater increase in ste12Δ bas1Δ cells (Fig. (Fig.5C).5C). This indicates that the activation mechanism does not require STE12 and a fortiori does not act through the FRE site. Thus, sequences located in the 5′ portion of the Ty1 promoter, other than the FRE site, might be involved in the activation of Ty1 transcription by severe adenine starvation.
The activation levels observed with the four TY1AF-lacZ fusions in adenine-deprived bas1Δ cells were significantly lower than those observed with the corresponding TY1A-lacZ fusions containing the complete Ty1 promoter sequences (compare Fig. Fig.3A3A and and5B).5B). This difference in activation raised the possibility that regulatory sequences located between the FRE site and the end of TY1A were also necessary for the response to adenine. Two sequences located in this segment of the Ty1 promoter regulate Ty1 transcription (Fig. (Fig.5A).5A). The first sequence is bound by the a1/α2 repressor complex in diploid cells to repress Ty1 transcription (17). It was unlikely that the a1/α2 complex plays a role in the regulation by adenine, since a1/α2 is formed in diploid cells only. The second sequence (referred as MIR) is recognized by Mcm1p, Tea1p (Ibf1p), and Rap1 and is known to stimulate Ty1 transcription (17, 22). To determine whether the MIR sequence was necessary for the regulation of Ty1 transcription by adenine, we replaced it with an unrelated sequence in the weakly and highly expressed fusions TY1A(DR3)-lacZ and TY1A(ML2)-lacZ, respectively (Fig. (Fig.5A,5A, TY1Amir-lacZ) (see Materials and Methods). In adenine-deprived bas1Δ cells, the activation of TY1Amir(DR3)-lacZ and TY1Amir(ML2)-lacZ was lower than the activation of the TY1A(DR3)-lacZ and TY1A(ML2)-lacZ fusions harboring the wild-type MIR sequence, indicating that an intact MIR sequence is necessary for complete activation (Fig. (Fig.5D;5D; 3.4- and 1.5-fold versus 10.6- and 2.5-fold, respectively). Thus, these results suggest that several regions of the Ty1 promoter are necessary for the Ty1 response to severe adenine starvation. The dispersion of regulatory sequences might also highlight the need to keep the integrity of the promoter sequences instead of specific sequences to achieve full activation.
Among the three proteins that bind to the MIR sequence, Tea1p belongs to the family of zinc cluster transcriptional activators, which are often activated by environmental challenges. Although TEA1 has been shown to activate Ty1 transcription, the conditions of TEA1 activation are not known (21). We checked whether TEA1 could mediate the activation of Ty1 transcription under conditions of adenine depletion by comparing the activation of the TY1A(DR3)-lacZ fusion in bas1Δ and bas1Δ tea1Δ cells. The response of this fusion to adenine depletion was similar in both strains, indicating that TEA1 is dispensable for the activation (Fig. (Fig.5E).5E). Deletion of TEA1 also had no impact on TY1A(DR3)-lacZ expression in cells grown in the presence of adenine. This result is consistent with the modest contribution of TEA1 to normal Ty1 mRNA levels (21).
It was striking that conditions of severe adenine starvation activated poorly expressed Ty1 elements more efficiently. The difference in expression between weakly and highly expressed copies can be explained by repressive chromatin at Ty1 elements that is exclusively antagonized by Swi/Snf and SAGA chromatin remodeling complexes at the highly expressed elements (36). Thus, one attractive hypothesis was that severe adenine starvation decreased chromatin repression at Ty1 promoters. This would explain why the activation was mostly detected with weakly expressed Ty1 elements, the transcription of which is repressed by chromatin. To explore this possibility, we tested whether adenine starvation still caused an increase in Ty1 mRNA levels in bas1Δ hta1-htb1Δ cells. Deletion of the HTA1-HTB1 gene pair, one of the two sets of genes encoding H2A and H2B, leads to an unbalanced production of histones. The resulting alteration in chromatin structure is correlated with an increase in Ty1 mRNA levels and, more specifically, with an increase in the expression of Ty1 elements expressed at low levels (26, 36). Northern blot assays indeed revealed a twofold increase in Ty1 mRNA levels in hta1-htb1Δ cells compared to wild-type cells (Fig. (Fig.6A).6A). However, the increase in Ty1 mRNA levels observed during adenine starvation in bas1Δ cells no longer occurred in hta1-htb1Δ bas1Δ cells (Fig. (Fig.6A).6A). Therefore, alteration in chromatin structure caused by the HTA1-HTB1 deletion abolishes the activation of Ty1 transcription observed in adenine-starved bas1Δ cells.
To confirm these data, we also tested whether adenine starvation still activated transcription of the weakly expressed TY1A(DR1)-lacZ fusion in hta1-htb1Δ cells. In agreement with the effect of the HTA1-HTB1 deletion on Ty1 transcription, expression of TY1A(DR1)-lacZ was higher in hta1-htb1Δ cells than in wild-type HTA1-HTB1 cells (Fig. (Fig.6B).6B). In the absence of adenine, the increase in β-galactosidase activity observed in bas1Δ cells did not occur in the double hta1-htb1Δ bas1Δ mutant (Fig. (Fig.6B).6B). It was possible that the lack of increase in β-galactosidase activity in the hta1Δ-htb1Δ bas1Δ mutant was due to an inability to increase TY1A(DR1)-lacZ expression further when histones are depleted. However, we could rule out this possibility since the constitutive STE11-4 allele, known to activate Ty1 transcription (37), was able to activate TY1A(DR1)-lacZ expression severalfold in both hta1-htb1Δ mutant cells and wild-type cells (Fig. (Fig.6C).6C). These data indicate that the activation of transcription of a weakly expressed Ty1 element under severe adenine starvation conditions is suppressed by a mutation that alters chromatin repression at Ty1 promoters.
The Swi/Snf and SAGA chromatin remodeling complexes are necessary for the activation of Ty1 transcription in haploid cells. Although a previous study showed that both complexes activate the highly expressed elements only (36), Swi/Snf or SAGA could be necessary for the activation of weakly expressed Ty1 elements under conditions of severe adenine starvation. We did not analyze the role of SAGA, because the pattern of Ty1 transcription is complex in mutants that affect SAGA and we suspected that the results would be difficult to understand (58, 59). Since SNF2 encodes an ATPase essential for Swi/Snf activity (30), we analyzed the involvement of Swi/Snf by comparing the β-galactosidase activities of the weakly expressed TY1A(DR1)-lacZ fusion in bas1Δ and bas1Δ snf2Δ cells grown in the presence and in the absence of adenine. The activation of Ty1(DR1)-lacZ observed in the absence of adenine was abolished in bas1Δ snf2Δ cells (Fig. (Fig.6D).6D). Similarly, TY1A(ML1)-lacZ and TY1A(PR1)-lacZ fusions, expressed at low and high levels, respectively, were not activated under severe adenine starvation conditions when SNF2 was deleted (data not shown). Together, these results indicate that Swi/Snf is necessary for the activation of Ty1 transcription under severe adenine starvation conditions.
In conclusion, the loss of activation of Ty1 transcription observed in a hta1-htb1Δ mutant, which alters chromatin structure, and in a snf2Δ mutant, which impairs the chromatin remodeling activity of Swi/Snf, is consistent with a model in which the mechanism of activation of Ty1 transcription by adenine starvation involves chromatin remodeling at Ty1 promoters.
Here we show that Ty1 mRNA levels increase under three conditions that all strongly affect cellular adenine nucleotide levels. Under the first condition, the BAS1 gene, encoding an essential transcriptional activator of de novo AMP biosynthesis genes (ADE genes), is deleted. In this mutant context, the basal activity of the de novo AMP biosynthesis pathway is not sufficient to ensure adequate levels of adenine nucleotides in the absence of adenine. Under the second condition, the de novo AMP biosynthesis pathway is not functional because the ADE2 gene, which encodes an enzyme of the pathway, is deleted. In both bas1Δ and ade2Δ cells, Ty1 transcription significantly increases when extracellular adenine is limiting. Under the third condition, the ade13-52 bradytrophic allele was used. In the ade13-52 mutant, adenine nucleotide pools remained low, even in the presence of extracellular adenine, as the Ade13 protein acts both in the de novo AMP biosynthesis pathway and in the pathway that metabolizes extracellular adenine to AMP. In this mutant, Ty1 mRNA levels increase independently of adenine availability in the growth medium.
The regulation of Ty1 transcription by adenine differs from the regulation of ADE gene expression. In contrast to the ADE genes, Ty1 transcription does not respond to transient adenine starvation, since activation occurs only when adenine nucleotide pools become extremely limiting, i.e., in adenine-deprived cells defective in de novo AMP biosynthesis synthesis (referred to in the text above as “severe adenine starvation”). The second difference with the regulation of ADE genes is that the Bas1 transcriptional activator is not responsible for the activation of Ty1 transcription under severe adenine starvation conditions. Instead, the absence of Bas1 increases Ty1 transcription in adenine-deprived cells by exacerbating the starvation conditions. Therefore, prolonged growth of a wild-type strain in the absence of adenine should have no effect on Ty1 transcription, since Bas1p would permanently activate the de novo AMP biosynthesis pathway. This activation should provide sufficient AMP pools to avoid severe adenine starvation and activation of Ty1 transcription.
We do not yet know whether the complete absence of adenine for cells defective in the de novo AMP biosynthesis pathway is physiological or not. It remains possible that yeast cells have evolved a specific regulatory mechanism to deal with adenine starvation, under natural conditions lacking purines and with a temperature or a pH, at which one of the enzymes of the de novo AMP biosynthesis pathway is inactive. Alternatively, extreme adenine starvation could mimic much more physiological but as-yet-unknown conditions that induce Ty1 transcription.
The exceptional stability of Ty1 RNA (its half-life is at least 3 hours, comparable to that of rRNA ) makes its unlikely that the main consequence of adenine deprivation is increased Ty1 RNA stability. Furthermore, we have previously shown that the β-galactosidase activity produced by the TY1A-lacZ fusions reflects the promoter activity of each native Ty1 element (36). Thus, the increase in expression of several TY1A-lacZ fusions observed in response to severe adenine starvation is likely to be the result of an increase in promoter activity, and we believe that the mechanism of activation occurs at the level of Ty1 RNA synthesis. Although a partial proteome analysis did not reveal any protein whose production increased independently of BAS1 in adenine-starved cells (14), Rolfes and Hinnebusch mentioned an activation of GCN4 transcription in ade1 cells deficient in the de novo AMP pathway and cultivated in the absence of adenine (45). Thus, it is possible that the new regulatory pathway that activates Ty1 transcription in response to conditions of severe adenine starvation might also regulate the transcription of additional genes.
The absence of adenine could lower the production of dATP in the cell and create an imbalance in deoxynucleoside triphosphate pools, which would generate DNA damage during replication. Indeed, a link between imbalance of deoxynucleoside triphosphate pools and DNA damage has been observed in eukaryotic tumor cells (27). Since Ty1 transcript levels increase in the cells in response to several DNA-damaging agents (5, 35, 44, 47, 53), it is possible that conditions of severe adenine starvation activate Ty1 transcription in the cell as part of a general DNA damage response. However, the lack of activation of RNR2 in bas1Δ cells grown under conditions of adenine starvation makes this hypothesis unlikely (data not shown). RNR2 encodes a ribonucleotide reductase subunit whose expression is induced by DNA damage.
The mechanism of Ty1 activation in response to severe adenine starvation is able to overcome the absence of Ste12, which is normally required for basal levels of Ty1 transcription. Nevertheless, the expression of the highly expressed fusion TY1A(ML2)-lacZ was lower in the absence of adenine in ste12Δ bas1Δ cells than in bas1Δ cells, reaching the expression level of the poorly expressed fusion TY1A(DR3)-lacZ in bas1Δ cells (Fig. (Fig.5C).5C). This observation indicates that Ste12 is still necessary to achieve full expression of highly expressed elements in adenine-starved bas1Δ cells. Ste12-independent transcriptional activation has also been recently observed for the activation of Ty1 transcription by γ-irradiation (47). These two examples of Ty1 transcription activation under stress conditions reveal for the first time that Ste12 is dispensable for Ty1 transcription under certain circumstances. Our results also indicate that Gcn4 and Tea1 are not responsible for the activation of Ty1 transcription under severe adenine starvation conditions.
Our data suggest that the mechanism of Ty1 activation might involve chromatin remodeling at Ty1 promoters. The first line of evidence comes from the suppressing effect of the hta1-htb1Δ deletion on the activation of Ty1 transcription in adenine-starved bas1Δ cells, particularly on the activation of the weakly expressed TY1A(DR1)-lacZ fusion. Since this deletion creates a histone imbalance that opens chromatin at weakly expressed Ty1 elements (26, 36), the absence of activation in the hta1-htb1Δ mutant suggests that the activation mechanism implicates a removal of repressive chromatin. The second line of evidence relies on the comparison of the activation of the weakly and highly expressed TY1A-lacZ fusions in bas1Δ cells grown in the absence of adenine. Activation is higher on weakly expressed TY1A-lacZ fusions (3.1- to 20-fold) than on highly expressed TY1A-lacZ fusions (2- to 0.8-fold). The difference in expression between Ty1 elements has been correlated to repressive chromatin at the promoters of the weakly expressed Ty1 elements that is opened by Swi/Snf and SAGA at the promoters of the highly expressed elements (36). Since regulation under conditions of severe adenine starvation occurs mainly on weakly expressed elements, we believe that the activation mechanism involves an opening of the chromatin that induces the expression of these Ty1 elements. The loss of activation of two weakly expressed fusions, Ty1(DR1)-lacZ and Ty1(ML1)-lacZ, and one highly expressed fusion, Ty1(PR1)-lacZ, in a snf2Δ mutant suggests that the Swi/Snf chromatin remodeling complex participates in this mechanism. Finally, our finding that the two mutant TY1A-lacZ constructs we made that disrupted the integrity of the Ty1 promoter affected the regulatory mechanism could be explained if multiple cis-acting sequences, recognized by different transcription factors, were implicated in the control. However, it might also highlight the existence of a particular spatial nucleosome organization over the Ty1 promoter that is important for regulation to occur.
Conditions that severely reduce cellular adenine nucleotide pools could perturb intracellular levels of a metabolic intermediate that would in turn influence chromatin remodeling. Such a mechanism has been proposed in the case of the ATPase-dependent Swi/Snf and INO80 chromatin remodeling complexes, whose activities are modulated by inositol polyphosphates (51, 54). Alternatively, chromatin-modifying complexes might be recruited to Ty1 promoters by DNA-bound transcription factors under conditions of severe adenine starvation. Finally, our results might also indicate that an intact adenine biosynthesis pathway is necessary to organize a repressive chromatin structure at Ty1 elements and that repression is abolished when adenine pools become limiting. Further studies will be necessary to unravel the mechanism of activation of Ty1 transcription by adenine deficiency.
Since severe adenine starvation had a greater impact on the expression of weakly transcribed elements, we introduced the his3AI indicator of retrotransposition in four native Ty1 elements expressed at different levels to analyze the effect of adenine deficiency on Ty1 mobility. Interestingly, the mobility of these Ty1 elements was increased in relative proportion to the activation of expression in bas1Δ cells grown in the absence of adenine. This suggests that for a given Ty1 element, the increase in mobility follows the increase in its mRNA levels. A similar observation has already been made for de novo Ty1 insertions at the RDN1 locus, the mobility of which increases in mutants that disrupt RDN1 repressive chromatin structures and release transcriptional silencing (6). However, this observation was only made for Ty1 elements present at that particular chromosome locus that inhibits their expression. It has been generally assumed from other studies that Ty1 transposition is tightly controlled at a posttranscriptional level and that an increase in the transcriptional level of an endogenous Ty1 element cannot overcome this control (11). In addition, the fact that the increase in retrotransposition of native Ty1 elements observed under conditions of severe adenine starvation can be correlated to an increase in expression suggests that the increase in Ty1 cDNA levels is a direct consequence of Ty1 transcriptional activation rather than the consequence of an activation of Ty1 cDNA synthesis per se. There are other examples of activation of Ty1 retrotransposition under stress-like conditions that have been correlated with increased transcription. It is noteworthy that in each case, the Ty1 response is different. In the case of the Ty1 response to external signals that induce invasive/filamentous differentiation, Ty1 mobility is activated at both transcriptional and posttranscriptional levels (8, 9, 37). Likewise, γ-irradiation stimulates multiple steps of the Ty1 life cycle (47). Other DNA-damaging agents might have the same effect as γ-irradiation on the Ty1 life cycle (5, 35, 44, 53). In contrast, telomere erosion increases Ty1 mobility posttranscriptionally by stimulating Ty1 cDNA synthesis (50). The diversity of Ty1 responses to stress highlights the complex relationship between Ty1 and its host. In particular, it suggests that for each kind of stress, a specific response has evolved to fulfill specific yeast needs. Indeed, the activation of Ty1 cDNA synthesis and mobility in telomerase-negative survivors is associated with the synthesis of subtelomeric Y′ cDNA, which can recombine with Y′ elements at telomeres and play a role in telomere maintenance (34, 50). Likewise, it has been proposed that activation of Ty1 transcription and retrotransposition by the invasive filamentous pathway might lead to cis-activation of cellular genes because Ty1 insertions in the opposite orientation relative to a target gene have been shown to confer regulation by the invasive/filamentous pathway on adjacent genes, through its FRE site (37). More insights into the regulation mechanism of adenine deficiency will be required to determine whether the activation of Ty1 under these stress-like conditions confers an advantage to the cell.
Finally, this is the first study that provides insights into the mobility of native Ty1 elements expressed from their chromosomal location, although the mobilities of de novo Ty1 insertions at different locations in the yeast genome have already been studied (6, 11). It is noteworthy that the four elements studied here are able to retrotranspose at a rate similar to de novo Ty1 insertions (6, 11). It was assumed from different studies that RNA levels of individual Ty1 elements determine their rate of retrotransposition. However, it is striking that the weakly expressed Ty1(DR5) element does not follow this rule, since it retrotransposes at the same frequency as the two highly expressed elements, Ty1(PR1) and Ty1(GR1). In comparison, the weakly expressed elements Ty1(NL2) and Ty1(DR1) retrotranspose at 1 and 2 orders of magnitude less than Ty1(PR1) and Ty1(GR1), respectively (Table (Table33 and data not shown). These differences suggest that Ty1(DR5) contains particular features in its sequence that optimize a posttranscriptional step of its life cycle. It would be interesting to determine whether other native Ty1 elements expressed at low levels have the capacity to retrotranspose at high levels, similar to Ty1(DR5).
We are grateful to B. Daignan-Fornier, B. Pinson, L. Benard, and C. Condon for stimulating discussions and helpful comments on the manuscript. We are grateful to J. Curcio, G. Fink, and B. Daignan-Fornier for providing strains and plasmids.
This work was supported by grants from the CNRS (UPR 9073), the Université Paris VII, and the Association pour la Recherche contre le Cancer (grant number 4663 to P.L.). A.-L.T. was a recipient of a fellowship from the Ministère pour la Recherche et la Technologie and of a fellowship from the Association pour la Recherche contre le Cancer.