Trans-acting antisense RNAs mediate transcriptional gene cosuppression in S. cerevisiae - PubMed (original) (raw)
Trans-acting antisense RNAs mediate transcriptional gene cosuppression in S. cerevisiae
Jurgi Camblong et al. Genes Dev. 2009.
Abstract
Homology-dependent gene silencing, a phenomenon described as cosuppression in plants, depends on siRNAs. We provide evidence that in Saccharomyces cerevisiae, which is missing the RNAi machinery, protein coding gene cosuppression exists. Indeed, introduction of an additional copy of PHO84 on a plasmid or within the genome results in the cosilencing of both the transgene and the endogenous gene. This repression is transcriptional and position-independent and requires trans-acting antisense RNAs. Antisense RNAs induce transcriptional gene silencing both in cis and in trans, and the two pathways differ by the implication of the Hda1/2/3 complex. We also show that trans-silencing is influenced by the Set1 histone methyltransferase, which promotes antisense RNA production. Finally we show that although antisense-mediated cis-silencing occurs in other genes, trans-silencing so far depends on features specific to PHO84. All together our data highlight the importance of noncoding RNAs in mediating RNAi-independent transcriptional gene silencing.
Figures
Figure 1.
A PHO84 extra copy leads to cosuppression of PHO84 mRNA expression. (A) Map of the PHO84 locus. Arrows indicate the orientation of PHO84 sense, TUB3 sense, and PHO84 antisense transcripts. YML122C described as a hypothetical ORF contains four Pho4-binding sites (black circles) and encompasses the PHO84 upstream activating sequence (UAS). The line below indicates the PHO84 sequence N°1 subcloned into the centromeric vector YCpLac111. The size of the insert is indicated on the right. Position 1 lies 105 bp upstream of the beginning of hypothetical ORF YML122C (725 bp upstream of PHO84 ATG), and position 2651 is located at +166 bp downstream from the PHO84 stop codon. Filled rectangles indicate promoter (P), 5′ and 3′ regions of PHO84 amplified in ChIP analyses (Fig. 2). The dotted line corresponds to the 3′ region spanned by the ssRNA probe used in the blots below. (B) Presence of an extra gene copy induces PHO84 cosuppression. Northern analysis of total RNA extracted from wild-type cells transformed with YCpLac111 empty vector (V) or YCpLac111 containing insert N°1 (N°1). After transformation cells were exponentially grown in −LEU media for 24 h. Total RNA was extracted and PHO84 sense and antisense RNAs detected with riboprobes mapping to PHO84 3′ end. Membranes were rehybridized with an ACT1 random labeled probe to control for equal loading. (C) PHO84 gene silencing is extra gene copy-dependent and reversible. Wild-type cells containing an empty vector (V) or plasmid N°1 were exponentially cultivated in −LEU for 24 h. A fraction of the culture was diluted into YEPD medium to lose vector (V) and N°1 and the rest collected for RNA extraction. Cured cells were retransformed with empty vector followed by cultivation in −LEU and RNA extraction. PHO84 sense and antisense RNA levels from cells containing vector (V) or N°1 (lanes 1,2) were compared with those of cured cells retransformed with the empty vector (lanes 3,4) as in B. (D) PHO84 gene silencing is position-independent. Wild-type and Δpho84 strains, each containing a PHO84 ectopic copy replacing either the LYS2 (lys∷PHO), the YML023 (yml∷PHO), or the GUD1 (gud∷PHO) loci, were transformed with an empty vector (V) or PHO84 construct N°1 and cultivated in −LEU medium followed by RNA extraction. Membranes were probed as in B. The chromosomal positions of endogenous (white square) and ectopic (black squares) PHO84 gene copies are diagrammed on the left. (E) An extra gene copy induces silencing in diploid cells. PHO84 sense and antisense RNA levels in haploid and diploid wild-type cells transformed with an empty vector (V) or plasmid N°1 (N°1) were analyzed as in B. In B–E, the numbers within the box indicate the number of PHO84 gene copies in the genetically modified and transformed strains.
Figure 2.
Silencing occurs at the level of transcription initiation and is Hda1/2/3-independent. (A) ChIP analysis of RNA Pol II over the PHO84 gene. Wild-type and Δpho4 strains transformed with an empty vector (V) or plasmid N°1 (N°1) were grown in −LEU medium. RNA Pol II immunoprecipitated DNA was quantified by real-time PCR with _PHO84_-specific primers (Fig. 1A; Supplemental Table 2). The relative enrichment of the gene segments was expressed as the _n_-fold increase with respect to a nontranscribed intergenic sequence. Values derive from three independent experiments. (B) ChIP analysis of TBP at the PHO84 promoter. The same chromatin extracts were immunoprecipitated with α-TBP antibodies. The DNA was quantified by real-time PCR with primers specific for PHO84 promoter and normalized as in A. ChIP is measuring the level of association of both endogenous and plasmidic PHO84 with these factors. (C) The Hda1/2/3 complex is not required for _trans_-silencing. PHO84 sense and antisense RNAs in wild-type, Δhda1, and Δhda2 strains transformed with an empty vector (V) or plasmid N°1. Cell cultures and Northern blotting were performed as in Figure 1B. PHO84 sense mRNAs were quantified and normalized to ACT1. For each strain, the levels of sense transcripts detected in presence of plasmid N°1 were expressed as a percentage of those detected in presence of empty vector (bottom line).
Figure 3.
A PHO84 sequence containing both the UAS and ORF is required to trigger silencing. (A) Map of the PHO84 gene. The lines below indicate all PHO84 portions (N°1–N°6 and UAS) subcloned into the centromeric vector YCpLac111 or other expression vectors (see below). The size of each insert is indicated on the right. Position 1 lies 105 bp upstream of YML122C (725 bp upstream of PHO84 ATG), and position 2651 is located at +166 bp downstream from the PHO84 stop codon. Dotted lines correspond to 5′ and 3′ regions spanned by the ssRNA probes. PHO84 sense and antisense RNA expression in Δpho84 (B) or in wild-type cells (C) transformed with PHO84 plasmids N°1–N°6. In both cases, sense RNAs were detected with a probe mapping to PHO84 3′end. Antisense RNAs were revealed with a probe mapping to PHO84 3′ end in B, or specific for PHO84 3′ end or 5′ end (common to all six constructs) in C. (D) Deletion of 120 bp within the PHO84 terminator (Δter strain) leads to the production of a long PHO84-TUB3 read-through transcript. Wild-type and Δter strains transformed with empty vector (V) or plasmid N°1 were analyzed as in Figure 1B. (E) Expression of the PHO84-TUB3 read-through transcript in a Δter strain transformed with empty vector (V) or plasmids N°1–N°6.
Figure 4.
GAL-driven PHO84 antisense RNAs silence endogenous PHO84 gene expression in trans. (A) _PHO84 cis_-silencing is recapitulated on plasmid. PHO84 sense and antisense RNA levels in Δpho84, Δpho84Δrrp6, and Δpho84Δhda2 strains transformed with a PHO84 plasmid in which antisense RNAs are driven by a galactose-inducible promoter (GAL-N°1) as drawn at the top. Cells were exponentially cultivated 24 h in −LEU medium containing 2% glucose (GLU) or 2% galactose (GAL) and PHO84 sense and antisense RNAs produced from the plasmid were detected as in Figure 1B. (B) GAL-inducible antisense RNAs promote TGS in trans. (Top) Map of the PHO84 + HIS5 endogenous gene. The S. pombe HIS5 gene was inserted within PHO84 in reverse orientation and is flanked by TEF promoter and terminator sequences (black triangles). Arrows indicate the orientation of the sense transcripts expressed from this locus. Dotted lines correspond to the position of PHO84 and HIS5 probes used below. (Bottom) Northern analysis of RNA from the _PHO84+HIS_5 strain transformed with empty vector (V) or plasmid GAL-N°1. Transformants were exponentially grown for 24 h in −LEU media containing 2% glucose (GLU) or 2% galactose (GAL). (Top panels) PHO84 + HIS5 and HIS5 sense transcripts were detected with HIS5 antisense- and sense-specific probes. (Middle panels) PHO84 + HIS and PHO84 plasmid transcripts were also detected with a PHO84 sense-specific probe. GAL1 and ACT1 mRNAs were detected with random primed probes and served, respectively, as galactose induction and loading controls. (C) TGS by GAL-induced PHO84 antisense RNAs is rapid. The PHO84 + HIS5 cells transformed with plasmid GAL-N°1 were grown in −LEU medium containing 2% glucose. After 24 h, cells were spun and resuspended in −LEU media containing either 2% glucose (GLU) or 2% galactose (GAL) and grown for an additional 20 min.
Figure 5.
PHO84 antisense RNA self-cleavage blocks silencing in cis and trans. (A) Map of the _PHO84_-ribozyme chimeric gene. A PHO84 gene containing a 51-mer wild-type (Rz) or mutant (Rzm) hammerhead ribozyme sequence was integrated in place of the endogenous PHO84 gene (strains PHO84Rz and PHO84Rzm) or inserted within the plasmid PHO84 N°1 (N°1Rz and N°1Rzm). Arrows indicate PHO84 sense and antisense transcripts. (B) Cleavage of endogenous PHO84 antisense RNA blocks PHO84 silencing in aged cells. Northern analysis of PHO84 sense and antisense RNAs in the progeny 25-d-old wild-type, PHO84Rz, and PHO84Rzm strains and the progeny of 2-d-old wild-type cells. (C) Cleavage of plasmid-encoded antisense RNA blocks silencing of endogenous PHO84. Δter cells (Fig. 3D) transformed with empty vector (V), plasmid N°1 (N°1) as such or containing wild-type (N°1Rz) or mutant ribozyme (N°1Rzm) were analyzed as in Figure 1B. (D) Cleavage of endogenous PHO84 antisense RNAs does not affect silencing in trans. The PHO84Rz strain (see A,B) transformed with empty vector or plasmid N°1 (N°1) was analyzed as in Figure 1B.
Figure 6.
Antisense RNAs are stimulated by Set1 and require both 5′ and 3′ sequences to mediate silencing. (A) SET1 is implicated in PHO84 gene cosuppression. Δset1 cells containing an empty YCpLac33 vector (V), or complemented with plasmids encoding wild-type SET1 or the rrm1 and rrm2 mutant forms, were transformed with a YCpLac111 vector (V) or the PHO84 plasmid N°1 and analyzed as in Figure 1B. (B) PHO84 antisense RNAs driven by a GAL promoter bypass the requirement for Set1 to trigger PHO84 cosuppression. Wild-type and Δset1 cells transformed with empty vector or plasmid GAL-N°1 were grown for 24 h in −LEU medium containing 2% glucose (GLU) or 2% galactose (GAL). PHO84 sense and antisense RNA levels were measured as in Figure 1B. (C) Efficient silencing requires long antisense RNAs containing both PHO84 3′end and UAS sequences. The W303 wild-type strain was transformed with the TetO7 pCM252 vector alone (V) or containing the PHO84 portions UAS, N°3, N°5, N°2, N°1 (Fig. 3A). Cells were exponentially grown for 16 h in −TRP media without (−) or with (+) doxycycline (20 μg/mL) to induce antisense RNAs. PHO84 sense and antisense RNAs were detected with single-stranded probes specific for PHO84 5′ end (Fig. 3A). ACT1 mRNA served as loading control. (D) Diagram of the chromosomal pho84∷URA3 gene. The PHO84 ORF (codons 1 to stop) was replaced by the K. lactis URA3 ORF (codons 1 to stop), so that URA3 transcription is driven by the PHO84 UAS promoter region. Total RNA extracted from the Δ_pho84∷URA3_ strain containing an empty vector (V) or plasmid N°1 was probed with probes specific for URA3 sense and PHO84 antisense transcripts, and ACT1 for loading control.
Figure 7.
Analysis of candidate genes regulated by antisense transcripts like PHO84. (A) RT-qPCR analysis of PHO84, VTC3, YJR129c, and GYP5 sense and antisense RNA levels in wild-type, Δrrp6, Δhda2, and Δhda2Δrrp6 strains exponentially grown in SC medium + 2% glucose. Values are represented as fold change of sense or antisense RNA levels in Δrrp6 versus wild type and in Δhda2Δrrp6 versus Δhda2. (B) VTC3, YJR129c, and GYP5 antisense RNAs do not trigger silencing in trans. The W303 wild-type strain was transformed with empty vector (V), plasmids containing the three candidate genes with their own regulatory sequences (gene) or plasmids containing the three candidate genes behind a GAL promoter such that antisense RNAs are induced in galactose (GAL antisense). Cells were exponentially grown in −LEU + 2% glucose (vector and gene) or in −LEU + 2% galactose (vector and GAL antisense). Total RNA was subjected to RT-qPCR analyses with gene-specific probes. Values are expressed as fold change versus the vector control in the two different conditions. (C) Model shows antisense RNAs mediate cis and trans transcriptional gene silencing. Set1 H3K4 methylation promotes antisense RNA production. In cis, antisense RNAs escaping degradation by the nuclear exosome can trigger transcriptional gene silencing in an Hda1/2/3-dependent (PHO84 and VTC3) or Hda1/2/3-independent (GYP5 and YJR129C) way. In trans, PHO84 antisense RNAs silence the homologous target sequence at the transcriptional level due to a homology region with the UAS promoter sequence. We propose that this requires an unidentified silencing factor coating the PHO84 gene (named A). Importantly, while high amounts of short _trans_-acting antisense RNAs (Tet-N°3) are weak in inducing silencing by A, small amounts of long antisense RNAs originating from the 3′end of PHO84 gene are very effective (plasmid N°1). We speculate that a silencing factor (named B) might be recruited by these long RNAs and would interact with the prebound repressor A enhancing its silencing activity (A*). The same hypothesis might explain why other antisense RNAs are ineffective in trans (VTC3, GYP5, and YJR129C).
References
- Andrulis ED, Werner J, Nazarian A, Erdjument-Bromage H, Tempst P, Lis JT. The RNA processing exosome is linked to elongating RNA polymerase II in Drosophila . Nature. 2002;420:837–841. - PubMed
Publication types
MeSH terms
Substances
LinkOut - more resources
Full Text Sources
Other Literature Sources
Molecular Biology Databases