|Home | About | Journals | Submit | Contact Us | Français|
The Ty1 long terminal repeat (LTR) retrotransposon of Saccharomyces cerevisiae is a powerful model to understand the activation of transposable elements by stress and their impact on genome expression. We previously discovered that Ty1 transcription is activated under conditions of severe adenine starvation. The mechanism of activation is independent of the Bas1 transcriptional activator of the de novo AMP biosynthesis pathway and probably involves chromatin remodeling at the Ty1 promoter. Here, we show that the 5′ LTR has a weak transcriptional activity and is sufficient for the activation by severe adenine starvation. Furthermore, we demonstrate that Ty1 insertions that bring Ty1 promoter sequences into the vicinity of a reporter gene confer adenine starvation regulation on it. We provide evidence that similar coactivation of genes adjacent to Ty1 sequences occurs naturally in the yeast genome, indicating that Ty1 insertions can mediate transcriptional control of yeast gene expression under conditions of severe adenine starvation. Finally, the transcription pattern of genes adjacent to Ty1 insertions suggests that severe adenine starvation facilitates the initiation of transcription at alternative sites, partly located in the 5′ LTR. We propose that Ty1-driven transcription of coding and noncoding sequences could regulate yeast gene expression in response to stress.
Long terminal repeat (LTR) retrotransposons are widespread transposable elements in eukaryotes. They resemble retroviruses in both their structure and their mode of replication, but they are not infectious (8). Their structure consists of two direct LTRs almost identical in sequence and two open reading frames (ORFs), analogs of the retroviral gag and pol genes. Retrotransposons propagate via reverse transcription (RT) of their mRNA and integration of the newly synthesized cDNA onto the chromosomes. In addition to full-length elements, genomes are scattered with solo LTRs, which originate from recombination between the two LTRs of an ancestral element. Full-length elements and solo LTRs constitute a significant percentage of genomes: 3% of the yeast Saccharomyces cerevisiae genome, up to 50% of the maize genome, 10% of the mouse genome, and 8% of the human genome (17, 23, 47, 68).
The mobility and the abundance of LTR retrotransposons can be harmful to the organisms. There is also increasing evidence that their recruitment as alternative exons or promoters has been a driving force in genome evolution in many organisms (30, 38, 46, 66). LTR sequences confer tissue-specific transcription of the neuronal apoptosis inhibitory gene (NAIP) family in humans, rats, and mice (55). The synchronous expression of a subset of genes expressed in oocytes and preimplantation embryos during mouse development depends on MalR LTR retrotransposons, which provide alternative promoters and 5′ exons to these genes (51). In the yeast Schizosaccharomyces pombe, the Tf2 LTR functions as an oxygen-dependent promoter and can also direct transcription of adjacent coding and noncoding sequences when oxygen is limiting (59). Although LTR retrotransposons are generally repressed at both transcriptional and posttranscriptional levels by a wide variety of mechanisms of gene silencing or by host factors that restrict different steps of their life cycle (32, 44, 60), they can be activated by stress and environmental challenges in several organisms (24, 40, 62).
The yeast Ty1 element is an interesting model to understand the activation of LTR retrotransposons by stress and the impact of such activation on gene expression. Ty1 is the most abundant LTR retrotransposon in S. cerevisiae, with 32 full-length elements and 185 solo LTRs in the sequenced genome of the S288C strain (34). These insertions are mainly located in a 1-kb window upstream of genes transcribed by RNA polymerase III (Pol III), where they have no or minor effects on tRNA gene expression (10, 34). Rare insertions in the promoter or 5′ noncoding regions of Pol II genes have also been recovered by genetic screens (for review, see references 7 and 40). These insertions were selected by their ability to inactivate or activate gene expression, indicating that Ty1 insertions are not neutral to Pol II gene expression. Solo LTR insertions inactivate gene expression probably via promoter competition (69). Full-length insertions that inactivate gene expression are mainly in the same transcriptional orientation as the target gene and separate the structural gene from its promoter, while activating insertions are in the divergent transcriptional orientation and bring Ty1 promoter sequences in proximity to the 5′ end of the target gene. The consequence is that the transcription of the adjacent gene depends on Ty1 promoter sequences. A well-documented example of this effect are Ty1 insertions upstream of the his3Δ4 promoterless allele that restore histidine prototrophy to yeast cells (57).
Different kinds of environmental stress, such as ionizing radiation, DNA damage, nitrogen starvation, and severe adenine starvation, activate Ty1 transcription and mobility (11, 45, 49, 54, 56, 61, 64). Studies to understand the impact of Ty1 insertions on adjacent gene transcription and to decipher the mechanisms of Ty1 activation by stress have provided several clues on Ty1 promoter structure and on the transcription factors involved in Ty1 transcription. The full Ty1 promoter extends over 1 kb both upstream and downstream of two TATA boxes and includes the 5′ LTR and part of the TYA (gag) ORF. Several transcription factors bind to the Ty1 promoter and regulate its transcription (19, 20, 26, 27, 36, 42, 48, 49, 65) (Fig. (Fig.1).1). The Swi/Snf, SAGA, and ISWI complexes that act on chromatin structures also regulate Ty1 transcription (14, 25, 33, 53).
Recently, we reported that Ty1 transcription is activated under conditions that strongly affect cellular adenine nucleotide synthesis: i.e., severe adenine starvation. The activation is followed by a proportional increase in retrotransposition events (64). The activation was observed in adenine-deprived cells, lacking the Bas1 transcriptional activator responsible for the induction of the ADE genes of the de novo AMP biosynthesis when adenine is limiting. This observation suggested a new mode of regulation of Ty1 transcription by adenine nucleotide levels. Since adenylic nucleotide availability probably reflects the energy status of the cell, severe adenine starvation could mimic conditions that impair the energetic metabolism in the yeast's natural environment.
Ty1 elements can be divided into two basic classes according to their level of expression: the weakly transcribed elements and the highly transcribed elements. Based on genetic data, it was proposed that Ty1 transcription is repressed by chromatin structure that is alleviated at the highly expressed elements by chromatin remodeling complexes (48). The activation by severe adenine starvation occurs mainly on weakly transcribed Ty1 elements. The activation is abolished in an hta1-htb1Δ mutant that opens chromatin structure at these elements by depleting histones H2A and H2B or in an snf2Δ mutant that inactivates Swi/Snf, suggesting that the activation involves chromatin remodeling at the Ty1 promoter (64). To obtain more information about the mechanism of activation of Ty1 by severe adenine starvation, we have characterized the regions of the Ty1 promoter important for the activation. Although previous work suggested that important Ty1 regulatory sequences reside mostly in Ty1 internal sequences (21, 72), we demonstrate here that the 5′ LTR possesses a weak transcriptional activity that is stimulated by severe adenine starvation. By investigating the impact of several Ty1 insertions on the transcription of different reporter genes, we also discovered that severe adenine starvation can modify the transcription pattern of genes located up to 300 bp from Ty1 insertions.
All strains used in this study are S288C derivatives. FYBL1-23D (MATα flo8-1 ura3Δ851 trp1Δ63 his3Δ200), LV426 (FYBL1-23D bas1Δ::TRP1) (64), LV69 (FYBL1-23D his3Δ4), and LV150 (FYBL1-23D Ty1-his3Δ4) (49) have already been described. Null alleles of BAS1 and STE12 were obtained in LV150 by one-step gene replacement using a PCR fragment of the KanMX and TRP1 cassettes flanked with 5′ and 3′ sequences of BAS1 and STE12, respectively. The resulting strains are LV922 (LV150 bas1Δ::KanMX), LV993 (LV150 ste12Δ::TRP1), and LV926 (LV150 bas1Δ::KanMX, ste12Δ::TRP1). The strain LV1107 (FYBL1-23D LTR-his3Δ4 bas1Δ::KanMX) was obtained in two steps. We first introduced in LV922 the URA3 gene at position 5560 of the Ty1 element, which is adjacent to his3Δ4. We then selected uracil auxotrophic cells, on 5-fluorouracil plates (6). We confirmed by PCR (Eppendorf PCR extender system) that these cells lost URA3 and Ty1 internal regions, by homologous recombination between the two LTRs of the Ty1 element, leaving a solo LTR.
Strains containing a TYA-lacZ fusion at the chromosomal locus of a native Ty1 element, with lacZ expressed from all of the Ty1 transcription control sequences, have already been described (48). Briefly, lacZ was fused in frame to TYA at an MfeI restriction site (at coordinate 1569 of Ty1-H3) (5), and followed by the URA3 gene and TYB sequences (coordinates 2171 to 3726 of Ty1-H3). The Munich Information Center for Protein Sequences (MIPS) annotation is used to name Ty1 elements (http://mips.gsf.de/genre/proj/yeast/). Strains containing an LTR-lacZ fusion at Ty1-DR3, Ty1-ML1, and Ty1-PR1 loci were obtained upon transformation of FYBL1-23D with a Ty1up-LTR-lacZ-URA3-“ty1b” DNA fragment. For each Ty1 element, The Ty1up-LTR portion was obtained by high-fidelity PCR amplification (Eppendorf PCR extender system) of a region covering approximately 100 bp upstream of the element and the 5′ LTR. The Ty1up sequence allows targeting of the integration to the right Ty1 locus. The “ty1b” sequences come from Ty1-H3 (coordinates 2171 to 3726). LTR sequences were checked for errors introduced during the amplification process. Null alleles of BAS1 and SNF2 were obtained in these strains by one-step gene replacement, using a PCR fragment of the KanMX and TRP1 cassettes flanked with 5′ and 3′ sequences of BAS1 and SNF2, respectively (Eppendorf PCR extender system). All constructs and gene replacements were checked by PCR analysis. Yeast transformations were performed by the lithium acetate procedure.
To construct pPL143, pPL144, and pPL147 plasmids, which express lacZ from the complete promoters of Ty1-DR3, Ty1-ML1, and Ty1-PR1, respectively (TYA-lacZ fusions), sequences extending from the first nucleotide and the MfeI restriction site of each element (at coordinate 1569 of Ty1-H3) (5) were amplified by high-fidelity PCR (Roche) and cloned into Yep356R (2μm and URA3) (50). A similar strategy was used to obtain pGS7, pGS8, pGS15, which are Yep356 (2μm and URA3) derivatives carrying LTR-lacZ fusions, with 5′ LTR sequences of Ty1-DR3, Ty1-PR1, and Ty1-ML1, respectively.
Plasmids pIL10, pIL11, and pIL12 carry Ty1 portions upstream of the UAS-less TDH3-lacZ reporter gene, and pIL5 carries a UAS-less TDH3-lacZ fusion. These plasmids were kindly provided by E. Dubois (37). A PCR-amplified DNA fragment corresponding to Ty1-DR3 5′ LTR (Eppendorf PCR extender system) was introduced at the HindIII site of pIL5 to create pGS17.
For all of the constructs, we checked that the amplification process did not introduce detrimental errors into the sequences. Details can be obtained upon request.
Yeast strains were grown in rich yeast extract-peptone-dextrose, Hartwell's synthetic complete (HC), and synthetic minimum (SDc, which is SD minimal medium containing arginine, isoleucine, tryptophan, leucine and valine) media, all supplemented with 2% glucose (1). Adenine was added to a final concentration of 0.3 mM.
For β-galactosidase assays, Northern blotting, and primer extension analyses, precultures were grown to saturation in liquid HC or SDc medium supplemented with adenine at 30°C and diluted 100-fold in the same medium, in the presence or absence of adenine. Ten-milliliter cultures were grown to mid-log phase between three and five generations at 22°C, which is the permissive temperature for Ty1 retrotransposition, or at 30°C. The media used in each experiment are given in the corresponding figure legends.
Total RNA was extracted by the hot acid phenol procedure as previously described (58). For Northern blotting, 5 to 15 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-N membrane (Amersham). Transfer and hybridization procedures were performed as recommended by the supplier. ACT1, ESF1, and Ty1 probes 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 Ty1-H3). An anti-HIS3 RNA probe corresponding to HIS3 618 to 213 with a GG sequence at the 5′ end was synthesized by T7 RNA polymerase (Promega), yielding uniformly labeled RNA with [α-32P]UTP as a tracer, from a DNA fragment obtained by PCR amplification of genomic DNA, using the T7 RNA polymerase promoter containing oligonucleotide 5′-TAATACGACTCACTATAGGGCGAGGTGGCTTCTCTTATGG as a forward primer and 5′-GCATTCCGGCTGGTCGCTAATC as a reverse primer.
Primer extension experiments were performed with 0.1 pmol of 5′-labeled (32P) primer and 100 μg of RNA, denatured at 65°C for 15 min in 10 μl (final volume) reverse transcriptase buffer (50 mM Tris-HCl [pH 8.3], 10 mM MgCl2, 80 mM KCl). The denatured mixture was cooled slowly to room temperature for 10 min. A 10.4-μl mix of 2 mM each deoxynucleoside triphosphate (dNTP), 8 mM dithiothreitol, and 8 U avian myeloblastosis virus (Finnzymes) reverse transcriptase in reverse transcriptase buffer was then added. Reaction mixtures were incubated for 60 min at 42°C. RNA templates were degraded with 20 μmol of NaOH. Samples were ethanol precipitated, and the pellets were resuspended in 8 μl 95% formamide, 20 mM EDTA, 0.05% bromophenol blue, and 0.05% xylene cyanol. Samples were loaded on 9% sequencing gels. The primers used in this study were O-GS71 (ESF1, coordinates −190 to −211 relative to the initiation codon) and O-GS73 (HIS3, coordinates +102 to +81 relative to the initiation codon). Sequences were performed in parallel, using the same primers (Thermosequenase; USB).
Northern blotting and primer extension analyses were reproduced at least three times. Results were quantified on a Molecular Dynamics PhosphorImager with ImageQuant software.
β-Galactosidase assays were performed as already described (49). 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.
The promoter of Ty1 contains three regions important for its transcription: the 5′ LTR and two regions called FRE (filamentous response element) and MIR (Mcm1, Tea1 [previously Ibf1], and Rap1), which are located in the TYA ORF (Fig. (Fig.1).1). We previously showed that severe adenine starvation activates the transcription of lacZ fusions, carrying all Ty1 promoter sequences. The activation was more efficient on weakly expressed elements than on highly expressed elements (64). These results are exemplified in the right panel of Fig. Fig.2A,2A, with TYA-lacZ fusions introduced at three Ty1 elements located on different chromosomes, either Ty1-DR3 and Ty1-ML1, which are weakly expressed, or Ty1-PR1, which is highly expressed. The increase in expression was independent of the Ste12 transcriptional activator, which binds to the FRE site (64). Furthermore, the activation was conserved with lacZ fusions containing the 5′ portion of Ty1 promoter, including the 5′ LTR and the FRE site (64). These observations suggested that the 5′ LTR might be sufficient to promote the activation. To test this idea, we assayed the expression of LTR-lacZ fusions driven from Ty1-DR3, Ty1-ML1, and Ty1-PR1. The 5′ LTRs of Ty1-PR1 and Ty1-DR3 are very similar in their sequences, except that Ty1-DR3 lacks the Gcn4E binding site, which is a signature of highly expressed Ty1 elements (Fig. (Fig.1)1) (48). The sequence of the Ty1-ML1 5′ LTR differs at several positions, mostly in the five potential Gcn4 binding sites and TATA box 1, which are far from consensus. In the three fusions, lacZ is translated from the Ty1 initiation codon, which is located in the 5′ LTR. Conditions of severe adenine starvation were obtained in bas1Δ cells grown in the absence of adenine, as described previously (64). For each LTR-lacZ fusion, there was no significant difference in the β-galactosidase activities of wild-type cells grown with or without adenine, and their activity was similar to that observed in bas1Δ cells grown in the presence of adenine (data not shown) (Fig. (Fig.2A,2A, left panel). When the bas1Δ cells were grown in the absence of adenine, we observed 10-, 3.5-, and 1.6-fold increases in β-galactosidase activity with LTR-lacZ fusions at Ty1-DR3, Ty1-ML1, and TY1-PR1, respectively. Except for Ty1-ML1, these activation factors were almost identical to those observed with TYA-lacZ fusions. In addition, the LTR-lacZ fusions were expressed at lower levels at Ty1-DR3 and Ty1-ML1 than at Ty1-PR1, indicating that the hierarchy in the level of expression observed with TYA-lacZ fusions was conserved. Taken together, these observations suggest that the 5′ LTR contains regulatory elements which are determinant for both the expression level of Ty1 elements and for their level of activation by severe adenine starvation.
The less-efficient activation by severe adenine starvation of LTR-lacZ fusions than of TYA-lacZ fusions observed at Ty1-ML1 might suggest that promoter sequences located in TYA are necessary for complete activation in the context of degenerated TATA boxes. Likewise, in bas1Δ cells grown in the presence of adenine, the β-galactosidase activity of the LTR-lacZ fusion at Ty1-PR1 was markedly lower than the activity of the TYA-lacZ fusion. Together, these differences confirm that TYA sequences are also important for the full expression and possibly the activation of these elements.
Since the β-galactosidase activities produced by the LTR-lacZ fusions were low, especially at Ty1-DR3 and Ty1-ML1, we wished to confirm the activation by severe adenine starvation with equivalent LTR-lacZ fusions carried on high-copy-number plasmids. In bas1Δ cells transformed with these plasmids and grown in the presence of adenine, the β-galactosidase activities were between 70- and 300-fold higher than the activities obtained with chromosomal fusions (compare Fig. Fig.2A,2A, left panel, and B). Interestingly, LTR-lacZ and TYA-lacZ fusions remained activated by severe adenine starvation when carried on plasmids, although the level of activation varied less between weakly and highly expressed elements (Fig. (Fig.2B).2B). The hierarchy of expression between Ty1 elements was also maintained on plasmids, although not as clearly. From this experiment, we conclude that the chromosomal environment is not essential for the activation of Ty1 by severe adenine starvation or the hierarchy of expression between elements.
We previously reported that the activation by severe adenine starvation of several TYA-lacZ fusions is abolished when the Swi/Snf chromatin remodeling complex is inactivated, suggesting that Swi/Snf could be involved in the mechanism of activation (64). Having established that the 5′ LTR is sufficient to promote the activation by severe adenine starvation, we asked whether Swi/Snf contributes to the activation through the 5′ LTR. Since SNF2 encodes an ATPase essential for Swi/Snf remodeling activity (39), the levels of expression of LTR-lacZ fusions at Ty1-DR3 and Ty1-PR1 were compared in bas1Δ and bas1Δ snf2Δ cells grown with and without adenine. At the weakly expressed Ty1-DR3 element, the SNF2 deletion led to a significant decrease in the activation of LTR-lacZ by severe adenine starvation, indicating that Swi/Snf is necessary for the activation to occur (Fig. (Fig.2C,2C, compare the 10-fold activation in bas1Δ to the 2.2-fold activation in bas1Δ snf2Δ cells). At the highly expressed Ty1-PR1 element, the activation of LTR-lacZ, which was already low in bas1Δ cells, was completely abolished in bas1Δ snf2Δ cells (Fig. (Fig.2C,2C, compare the 1.6-fold activation in bas1Δ cells to the 0.9-fold activation in bas1Δ snf2Δ cells). The loss of activation by severe adenine starvation of LTR-lacZ fusions at Ty1-DR3 and Ty1-PR1, in the absence of SNF2, is consistent with a role of the Swi/Snf chromatin remodeling complex in the mechanism of activation.
It is noteworthy that the SNF2 single deletion led to an increase, although less than twofold, in the expression of LTR-lacZ at Ty1-PR1, independently of the adenine availability. This observation is surprising since previous data clearly established that Ty1 mRNA levels synthesized from all Ty1 elements strongly decrease in snf2Δ cells (14, 28). In addition, the expression of a TYA-lacZ fusion at Ty1-PR1 is abolished in snf2Δ cells, indicating that Swi/Snf plays a major role in Ty1 transcription (48). One explanation is that Swi/Snf plays antagonistic roles at Ty1-PR1: a minor repressive effect at the 5′ LTR and a major positive effect at downstream Ty1 promoter sequences.
Ty1 insertions can activate the expression of adjacent genes (reviewed in references 7 and 40). In all cases, Ty1 elements are inserted within a window of 175 bp upstream of the target gene, in such a way that Ty1 and the adjacent gene are divergently transcribed. These insertions have been extremely useful in the identification of Ty1 transcription factors, such as Ste12, which binds to the FRE site in TYA and is important for basal Ty1 transcription in haploid cells (3, 14, 36, 49). Thus, we speculated that such Ty1 insertions might also be helpful to understand the regulation of Ty1 transcription by severe adenine starvation. To investigate whether conditions of severe adenine starvation could activate the expression of a gene adjacent to a Ty1 element, we used a reporter Ty1-his3Δ4 allele, which consists of a Ty1 insertion 56 bp upstream of the promoterless his3Δ4 allele (Fig. (Fig.3).3). Cells containing the Ty1-his3Δ4 allele were able to grow in the absence of histidine, while cells containing the his3Δ4 allele alone could not, indicating that his3Δ4 expression depends on the presence of the Ty1 insertion (Fig. (Fig.3,3, rows 1 and 2). Furthermore, the ability of Ty1-his3Δ4 cells to grow in the absence of histidine required the Ste12 transcriptional activator (Fig. (Fig.3,3, row 5). We showed previously that STE12 is not required for the activation of TYA-lacZ fusions by severe adenine starvation (64). Therefore, we analyzed whether the activation of Ty1 by severe adenine starvation could rescue the histidine auxotrophy of ste12Δ cells, by stimulating his3Δ4 transcription from the adjacent Ty1 sequences. The growth of ste12Δ and ste12Δ bas1Δ cells was compared on plates lacking either histidine only or both histidine and adenine. Although ste12Δ cells grew poorly on both plates, the growth of ste12Δ bas1Δ cells was greatly improved in the absence of adenine (Fig. (Fig.3,3, compare rows 4 and 5). The suppression of the growth defect of an ste12Δ mutant on plates lacking histidine by severe adenine starvation was strictly dependent on the presence of Ty1 sequences upstream of his3Δ4, since severe adenine starvation did not rescue the histidine auxotrophy of bas1Δ cells containing the his3Δ4 allele alone (data not shown). These results indicate that Ty1 can activate the transcription of an adjacent his3Δ4 allele under conditions of severe adenine starvation. The impact of severe adenine starvation on Ty1-his3Δ4 could not be detected in a bas1Δ single mutant (Fig. (Fig.3,3, row 3), probably because in these cells, STE12 activates Ty1-his3Δ4 transcription at sufficient levels to allow the growth in the absence of histidine, independently of adenine availability (see below).
Since severe adenine starvation can activate Ty1-his3Δ4 and LTR-lacZ expression, we asked whether the 5′ LTR would also be sufficient to promote activation of his3Δ4 under these conditions of starvation. We obtained an LTR-his3Δ4 allele in bas1Δ cells, by eviction of the Ty1 internal region in Ty1-his3Δ4, through homologous recombination between the two LTRs of the element. In the resulting LTR-his3Δ4 allele, the LTR is at the same distance from his3Δ4 and in the same orientation as it is in Ty1-his3Δ4. Unlike the Ty1-his3Δ4 allele, the LTR-his3Δ4 allele did not confer histidine prototrophy to bas1Δ cells, and severe adenine starvation did not suppress the histidine deficiency of these cells (Fig. (Fig.3,3, row 6). Together, these results indicate that severe adenine starvation can activate the expression of a gene adjacent to a Ty1 element and that the activation requires Ty1 internal sequences.
To further analyze the requirement for Ty1 promoter sequences in the activation of adjacent gene expression under conditions of severe adenine starvation, we used centromeric plasmids carrying different portions of the Ty1 promoter, inserted 140 bp upstream of a TDH3-lacZ reporter construct, with the same divergent organization as for the Ty1-his3Δ4 allele (Fig. (Fig.4A).4A). These plasmids have been previously used to demonstrate the involvement of the Ty1 FRE site in the activation of Ty1-adjacent genes (37). TDH3 encodes a glyceraldehyde-3-phosphate dehydrogenase involved in glycolysis and neoglucogenesis. In these plasmids, TDH3-lacZ expression is dependent upon Ty1 promoter sequences. The β-galactosidase activity of bas1Δ cells grown in the presence of adenine was weak with all of the constructs (Fig. (Fig.4B).4B). When TDH3-lacZ expression was driven from the complete Ty1 promoter, a 10-fold increase in β-galactosidase activity was observed in adenine-starved cells (Fig. (Fig.4B,4B, plasmid pIL11). The activation decreased to 4.9- and 6.2-fold, in the absence of the 3′ portion of Ty1 promoter including MIR, suggesting that MIR contributes to achieving complete activation of TDH3-lacZ under conditions of severe adenine starvation (Fig. (Fig.4B,4B, plasmids pIL10 and pIL12). This observation is consistent with our previous report that severe adenine starvation does not fully activate TYA-lacZ fusions lacking MIR (64) and with the observation that two and three tandem copies of MIR inserted upstream of a CYC1-lacZ reporter gene confer the activation by severe adenine starvation to the construct (data not shown). Furthermore, the activation by severe adenine starvation was abolished when TDH3-lacZ was expressed from a solo LTR (Fig. (Fig.4B,4B, plasmid pGS17). These results are consistent with those obtained with Ty1-his3Δ4 and LTR-his3Δ4 alleles. Together, they indicate that Ty1 internal sequences are important for the activation of adjacent gene expression under conditions of severe adenine starvation.
To understand the impact of severe adenine starvation on the transcription of genes adjacent to Ty1, Northern blotting analysis was undertaken with cells carrying Ty1-his3Δ4 and LTR-his3Δ4 alleles. As a control, Ty1 mRNA levels increased severalfold in bas1Δ cells grown in the absence of adenine and were extremely low in ste12Δ cells (Fig. (Fig.5A,5A, lanes 7, 8, and 9). In bas1Δ ste12Δ cells grown in the absence of adenine, the amounts of Ty1 mRNA increased to reach a level similar to that observed in wild-type cells (Fig. (Fig.5A,5A, compare lane 13 with lanes 4 and 5). This increase is consistent with the activation of TYA-lacZ by severe adenine starvation in the absence of STE12 (64). It is worth noting that the decrease in Ty1 transcript observed in ste12Δ cells was accompanied by the appearance of a transcript of smaller size (Fig. (Fig.5A,5A, lanes 8 and 9). A similar short species has been described in spt3 mutants that affect Ty1 transcription and corresponds to an initiation of transcription in the TYA ORF (70).
Two mRNA species were detected in Ty1-his3Δ4 wild-type cells, using a HIS3 riboprobe specific for HIS3 mRNA (Fig. (Fig.5A,5A, lanes 4 and 5). Using an RNA ladder, we estimated an approximately 180-nucleotide difference in lengths between these two mRNA species (data not shown). Their synthesis was dependent on the presence of the full-length Ty1 element upstream of his3Δ4 since they were not detected in wild-type cells carrying the Ty1-less his3Δ4 or LTR-his3Δ4 alleles (Fig. (Fig.5A,5A, lanes 2, 3, 14, and 15). These species were only detected in cells able to grow in the absence of histidine: i.e., in wild-type and bas1Δ cells and in adenine-deprived ste12Δ bas1Δ cells (compare Fig. Fig.33 and and5A,5A, lanes 4 to 7 and 10 to 13). The mRNA species of lower molecular weight comigrated with functional HIS3 mRNA (Fig. (Fig.5A,5A, lane 1). Furthermore, its amount increased under severe adenine starvation: i.e., in bas1Δ and ste12Δ bas1Δ cells grown in the absence of adenine (Fig. (Fig.5A,5A, compare lanes 6 and 7 and lanes 12 and 13). Altogether, these observations indicate that the mRNA species of lower molecular weight allows the synthesis of a functional His3 protein and that its expression is stimulated by severe adenine starvation. Although the amount of the mRNA species of higher molecular weight showed no obvious correlation with the adenine availability in wild-type and bas1Δ cells, its expression was stimulated by the absence of adenine in ste12Δ bas1Δ cells (Fig. (Fig.5A,5A, compare lanes 12 and 13). We conclude that severe adenine starvation activates the synthesis of both HIS3 mRNA species.
Primer extension experiments were also performed to map the 5′ end of the different mRNA species produced by HIS3, Ty1-his3Δ4, and LTR-his3Δ4, respectively (Fig. (Fig.5B5B and and5C).5C). In HIS3 wild-type cells, two major products were identified, which correspond well to the two already characterized HIS3 transcripts (Fig. (Fig.5B,5B, lane 1: RT stops at positions −11 and −24 relative to HIS3 initiation ATG codon ). These products were also detected in Ty1-his3Δ4 cells that were His+: i.e., in wild-type and bas1Δ cells and in adenine-deprived ste12Δ bas1Δ cells (Fig. (Fig.5B,5B, lanes 4, 5, 6, 7, and 9). Another major product, which was specific for Ty1-his3Δ4, was identified in these cells. This product corresponds to a 5′ end located in the LTR sequence, approximately at position −71 relative to the HIS3 ATG codon (Fig. (Fig.5B,5B, position +15 relative to the +1 nucleotide of the LTR). In addition to these products, minor products were also revealed, which may correspond to other 5′ ends or may be the consequence of secondary structures that block RT (Fig. (Fig.5B:5B: RT stops at positions −5 and −27 relative to the HIS3 ATG and positions +5, +45, and +71 relative to the +1 nucleotide of the LTR). Interestingly, all of these products were absent in Ty1-less his3Δ4 cells, LTR-his3Δ4 cells, Ty1-his3Δ4 ste12Δ cells and Ty1-his3Δ4 ste12Δ bas1Δ cells grown in the presence of adenine (Fig. (Fig.5B,5B, lanes 2, 3, 8, and 10 to 13). Together, these observations suggest that the corresponding mRNA species are under the control of Ty1 internal sequences and respond to severe adenine starvation. Among these species, only the HIS3 −11 and −24 species should be functional, since species with 5′ ends in the LTR contain two short ORFs that might impair HIS3 translation (Fig. (Fig.5C5C).
The difference in sizes between the two HIS3 mRNA species detected by Northern blot analysis (around 180 nucleotides) suggests that the 5′ end of the mRNA species of higher molecular weight must be in the LTR sequence. However, this difference was bigger than the difference in sizes of the products obtained by primer extension analysis (65 to 120 nucleotides). We made several unsuccessful attempts to detect 5′-end products that could correspond to this mRNA species by using other primers or longer migration of electrophoresis gels. In conclusion, severe adenine starvation activates the synthesis of several mRNA species in Ty1-his3Δ4 cells. Their 5′ ends are located in the LTR and at normal HIS3 start sites. Their synthesis requires Ty1 internal sequences, since these are absent in LTR-his3Δ4 cells. Among these transcripts, the mature HIS3 mRNA might be the only one to produce a functional His3 protein.
Our findings that severe adenine starvation can activate the transcription of his3Δ4 and TDH3-lacZ, when they are 56 and 140 bp downstream of Ty1 promoter sequences, respectively, prompted us to analyze whether this activation could also happen with yeast genes naturally located in proximity to Ty1 insertions. On chromosome IV, the full-length Ty1-DR6 (YDRWTy1-5) element is located 320 bp upstream of the ESF1 gene that encodes an essential nucleolar protein involved in pre-rRNA processing (52). Two species of ESF1 mRNA were identified by Northern blot analysis in wild-type and bas1Δ cells (Fig. (Fig.6A,6A, lanes 1 to 4). Interestingly, the levels of the higher-molecular-weight species increased in bas1Δ cells grown in the absence of adenine (Fig. (Fig.6A,6A, lane 4). This species was also detected in bas1Δ ste12Δ cells grown in the absence of adenine, whereas it could not be detected in ste12Δ cells, independently of the adenine availability or in ste12Δ bas1Δ cells grown in the presence of adenine (Fig. (Fig.6A,6A, compare lane 8 with lanes 5, 6, and 7). This profile of expression is similar to that of Ty1 and suggests that the transcription of this ESF1 mRNA species is under the control of Ty1 transcription regulatory signals and is stimulated by severe adenine starvation. The situation was different for the ESF1 mRNA of lower molecular weight, as its amounts remained relatively constant in the wild-type and bas1Δ cells grown with and without adenine (Fig. (Fig.6A,6A, lanes 1 to 4) and increased in ste12Δ cells and in ste12Δ bas1Δ cells grown in the presence of adenine (Fig. (Fig.6A,6A, lanes 5, 6, and 7). Moreover, in ste12Δ bas1Δ cells, the amounts of this species decreased concomitantly with the increase of the higher-molecular-weight species, in the absence of adenine. (Fig. (Fig.6A,6A, compare lanes 7 and 8). This balance between the two species suggests a competition in their synthesis.
Primer extension experiments were performed in parallel (Fig. 6B and C). Using a primer complementary to sequences in the ESF1 gene (coordinates +49 and +28 relative to ESF1 +1), we characterized four products corresponding to 5′ ends upstream of ESF1 at coordinates −306, −63, −70, and −88 relative to the ESF1 initiation codon (data not shown). The ESF1 −306 product was also identified with primer O-GS71, which is complementary to sequences upstream of ESF1 (Fig. 6B and C). Among these products, the ESF1 −70 and −88 products probably correspond to transcription start sites previously observed in a global mapping of yeast transcription start sites (73). The level of these products did not change in response to severe adenine starvation.
The O-GS71 primer also revealed four products corresponding to 5′ ends at positions −326, −336, −366, and −392 upstream of ESF1, in the Ty1 LTR (Fig. 6B and C, LTR +5, +15, +45, and +71). Remarkably, these 5′ ends were identical to those observed in the LTR sequence of Ty1-his3Δ4 mRNA (Fig. 5B and C). Moreover, the presence of these products in the different samples fit perfectly with the presence of the higher-molecular-weight RNA species detected by Northern blotting in the different cell cultures, as they were not detected in ste12Δ cells and in ste12Δ bas1Δ cells grown in the presence of adenine (Fig. (Fig.6B,6B, lanes 5, 7, and 8). These results demonstrate for the first time that a full-length Ty1 element controls the transcription of a yeast gene naturally located in proximity to a Ty1 promoter. It is noteworthy that Ty1-DR6 promoter sequences are located more than 300 bp upstream of ESF1, as this suggests a relatively long-distance effect of Ty1 on adjacent gene transcription.
The first striking observation is that the 5′ LTR can activate the transcription of lacZ reporter fusions in the absence of other activating sequences. Previous work has shown that while Ty1 transcription starts in the 5′ LTR, there is no upstream activating sequence (UAS) in the LTR and important regulatory sequences reside mostly in the TYA ORF (21, 72). Yet, an upstream activating activity (UAS) has been described with the his4-912δ composite promoter, consisting of the juxtaposition of HIS4 UAS and a Ty1 LTR (18). In this case, Gcn4 acting at the HIS4 UAS and Gcr1 acting at the Ty1 LTR are the primary activators required for his4-912δ expression. Here, we show that the LTRs of two weakly expressed elements, Ty1-DR3 and Ty1-ML1, and one highly expressed element, Ty1-PR1, activate the transcription of lacZ fusions. The activation is effective with LTR-lacZ fusions, either introduced at endogenous Ty1 locations or carried on multicopy plasmids. Interestingly, the hierarchy of expression previously observed with endogenous TYA-lacZ fusions carrying the full Ty1 promoter sequences is conserved with LTR-lacZ fusions, although the expression of LTR-lacZ fusions is weak compared to that of TYA-lacZ. This confirms the important role of Ty1 internal sequences for full transcription. Collectively, these data demonstrate that Ty1 LTRs possess a weak transcription activity and contribute to the variability in expression between Ty1 elements. A binding site for the Gcr1 transcriptional activator of genes encoding glycolytic enzymes, ribosomal proteins, and cyclins has been found in the Ty1 LTR, and Gcr1 activates Ty1 transcription (2, 41, 65). Likewise, five potential Gcn4 binding sites are located in the 5′ LTRs of several Ty1 elements (mostly those expressed at high levels) and Gcn4 overproduction stimulates the transcription of these elements (48). Thus, Gcr1 and Gcn4 might contribute to Ty1 LTR transcription activity. Since the yeast genome contains over 200 LTR sequences, our results raise the possibility that the intrinsic transcription activity of the LTRs contributes to intergenic transcription in yeast.
Our previous work has shown that severe adenine starvation activates TYA-lacZ fusions containing the full Ty1 promoter sequences (64). We now show that the activation also occurs on LTR-lacZ fusions. Interestingly, the results obtained with LTR-lacZ fusions recapitulate several previous observations made with TYA-lacZ fusions, such as the more efficient activation of weakly expressed elements and the loss of activation in a snf2Δ mutant that affects Swi/Snf chromatin remodeling activity. These results emphasize the important role of the 5′ LTR in the regulation of Ty1 transcription. Since Ty1 transcription starts in the 5′ LTR, our results also suggest that severe adenine starvation could facilitate the initiation of Ty1 transcription by Swi/Snf-dependent chromatin remodeling at Ty1 promoters. Alternatively, chromatin remodeling at the Ty1 promoter could be a prerequisite to the activation.
The discovery that severe adenine starvation can activate the transcription of genes adjacent to Ty1 insertions is also of interest. This conclusion is exemplified by the Ty1-his3Δ4 allele, the TDH3-lacZ reporter construct expressed from different Ty1 sequences, and the ESF1 gene, which is naturally located in the vicinity of Ty1-DR6 in the yeast genome. In all of these cases, Ty1 is transcribed divergently from its associated gene. Previous studies on the impact of Ty1 insertions on adjacent gene expression suggested that the activation strongly depends on the Ste12 transcriptional activator (reviewed in references 7 and 40). Thus, it is noteworthy that severe adenine starvation can activate the expression of these genes in the absence of STE12. Yet, it is consistent with our previous observation that the activation of Ty1 transcription by severe adenine starvation occurs in ste12Δ cells (64). Furthermore, although the 5′ LTR is sufficient to promote activation by severe adenine starvation when fused to lacZ in the same orientation, it is unable to activate the transcription of his3Δ4 or TDH3-lacZ reporter genes when it is inserted in the opposite orientation. An explanation for the different impact of the LTR on lacZ, his3Δ4, and TDH3-lacZ transcription is that Ty1 internal sequences could be essential for adjacent gene transcription. This hypothesis is supported by the activation of TDH3-lacZ, observed with constructs that contain some Ty1 internal sequences, and by the characterization of several HIS3 mRNA species in Ty1-his3Δ4 cells that are absent in LTR-his3Δ4 cells. Since chromatin structures play an important role in Ty1 transcription, one idea is that Ty1 internal sequences help to organize a chromatin structure permissive for transcription over the Ty1 LTR and adjacent regions, allowing activation by severe adenine starvation (48, 69). Also of note is the identification of ESF1 and his3Δ4 mRNA, with common 5′ends in the 5′ LTR of Ty1-DR6 in the case of ESF1 and in the 5′ LTR of the Ty1-his3Δ4 allele. Although we cannot rule out that some of these mRNA species correspond to the maturation of a precursor RNA by RNases, their increase under severe adenine starvation suggests that severe adenine starvation could stimulate the initiation of transcription at cryptic sites, especially in the LTR. It is noteworthy that AT-rich sequences, which might function as TATA boxes, are present in the LTR at a canonical distance from the +15 and +45 cryptic initiation sites (Fig. (Fig.5C5C and and6C6C).
There are two yeast genes transcribed by RNA polymerase II, ESF1 and NDJ1, which are naturally located in proximity to full-length endogenous Ty1 elements, in a configuration similar to that of the Ty1-his3Δ4 allele and the Ty1-TDH3-lacZ construct (22). With ESF1, this study provides the first example of a yeast gene naturally under the control of an endogenous Ty1 insertion. Here, we identify two classes of ESF1 mRNA species. The first class contains several mRNA species with 5′ ends in the LTR of Ty1-DR6. All of these transcripts are likely not functional since they contain several short ORFs upstream of the ESF1 initiation codon. The second class corresponds to mRNA species with 5′ ends in the region comprised between the 5′ LTR and ESF1 and includes functional ESF1 mRNA. Ty1-DR6 affects the expression of these two classes by different means. Transcripts with 5′ ends in the LTR of Ty1-DR6 are under the control of Ty1-DR6 regulatory sequences and follow the regulation of Ty1 transcription, as they are absent in ste12Δ cells and their level increases in adenine-deprived cells. In contrast, conditions that inhibit Ty1 transcription (i.e., ste12Δ) increase the amounts of the second class of ESF1 mRNA and conversely conditions that stimulate Ty1 transcription (i.e., severe adenine starvation in ste12Δ cells) decrease their amounts. This regulation could be explained by a competition for the recruitment of transcription factors at the ESF1 promoter region and the 5′ LTR or by transcription interference with transcription from the Ty1-LTR masking the ESF1 promoter region.
Remarkably, NDJ1, which encodes a meiosis-specific telomere protein, is another gene potentially under the control of an adjacent endogenous Ty1 insertion since it is located 250 bp downstream of Ty1-OL promoter sequences (22). Although we were unable to detect NDJ1 mRNA by Northern blotting, microarray studies performed to characterize yeast genes upregulated under conditions of severe adenine starvation identified NDJ1 (Servant et al. unpublished data). Strikingly, the distance between ESF1 and Ty1-DR6 or between NDJ1 and Ty1-OL (320 and 250 bp, respectively) is wider than the classical distance of Ty1 insertions reported to activate his3Δ4 expression (within a 175-bp window ). Hence, Ty1 can influence the expression of genes located at a distance larger than predicted from previous studies. However, the distance of these genes from Ty1 sequences probably influences the way Ty1 will affect their expression. Effectively, in the case of Ty1-his3Δ4, the activation of Ty1 transcription allows the transcription of functional HIS3 mRNA, whereas with ESF1, conditions that inactivate Ty1 transcription favor the transcription of functional ESF1 mRNA. It has already been reported that an endogenous Ty2 insertion inhibits the transcription of the high-affinity copper transporter gene CTR3 (35). With ESF1 and NDJ1, we provide the first evidence that endogenous Ty1 insertions can remodel yeast gene transcription under certain growth conditions.
Genome-wide studies in cells deficient in RNA degradation reveal that the yeast transcriptome includes numerous antisense transcripts from coding regions and cryptic unstable transcripts from intergenic regions (15, 16, 71). In addition, important regulatory functions have been assigned to antisense transcripts at IME4, PHO84, and Ty1; to noncoding interfering transcripts at SER3; and to cryptic transcripts in the ribosomal DNA arrays (4, 13, 29, 31, 43, 67). Our study indicates that full-length Ty1 insertions can drive the transcription of adjacent regions. We also provide evidence with the LTR-lacZ fusions that transcription can be initiated from solo LTRs. Since 32 full-length Ty1 elements and 185 solo LTRs are scattered in the yeast genome (34), an attractive hypothesis is that Ty1-driven transcription could contribute to the production of noncoding RNAs. Since Ty1 transcription is activated by different kinds of environmental stress conditions, Ty1-driven transcription of noncoding sequences could perhaps influence the reprogramming of yeast gene expression in response to stress. This hypothesis makes particular sense with the nonrandom clustering of stress-related genes, tRNA genes, and Ty1 sequences described in the S. cerevisiae genome (12).
We are grateful L. Bénard and M. Springer for stimulating discussions and comments on the manuscript. Thanks also go to M. Bétermier and C. Condon for helpful comments on the manuscript. We are grateful to E. Dubois for providing plasmids and to J. Lederout for technical help with riboprobe synthesis.
This work was supported by grants from the CNRS (UPR 9073) and Université Paris-Diderot Paris 7. G.S. was a recipient of a fellowship from the CNRS and of a fellowship from the Association pour la Recherche contre le Cancer.
Published ahead of print on 30 June 2008.