Modulation of mRNA and lncRNA expression dynamics by the Set2-Rpd3S pathway

Abstract

H3K36 methylation by Set2 targets Rpd3S histone deacetylase to transcribed regions of mRNA genes, repressing internal cryptic promoters and slowing elongation. Here we explore the function of this pathway by analysing transcription in yeast undergoing a series of carbon source shifts. Approximately 80 mRNA genes show increased induction upon SET2 deletion. A majority of these promoters have overlapping lncRNA transcription that targets H3K36me3 and deacetylation by Rpd3S to the mRNA promoter. We previously reported a similar mechanism for H3K4me2-mediated repression via recruitment of the Set3C histone deacetylase. Here we show that the distance between an mRNA and overlapping lncRNA promoter determines whether Set2-Rpd3S or Set3C represses. This analysis also reveals many previously unreported cryptic ncRNAs induced by specific carbon sources, showing that cryptic promoters can be environmentally regulated. Therefore, in addition to repression of cryptic transcription and modulation of elongation, H3K36 methylation maintains optimal expression dynamics of many mRNAs and ncRNAs.

Figure 1. Set2 negatively regulates the kinetics of transcriptional induction.

(a) Schematic representation of the time course experiments to determine changes in transcript levels upon carbon source shifts. (b) RNA samples from the time course experiments in a were analysed by high-density tiling arrays. Normalized, log₂-transformed mRNA expression levels were visualized with the Multi Experiment Viewer. +and − indicate SET2 and set2Δ, respectively. Set2-repressed genes were identified as those showing at least twofold increase in transcript levels at one or more time points. (c) Averaged profile of expression signals of 60 genes from top set in b. (d) Averaged profile of expression signals of 18 genes from bottom set in b. Gal, galactose; Glu, glucose; Ra, raffinose.

Figure 2. Overlapping lncRNA transcription localizes H3K36me3 and Rpd3S to target promoters.

(a) AAD10 has a constitutively active distal promoter (blue arrow) and a proximal promoter (red arrow) induced during galactose incubation in set2Δ. SUL1 has a proximal promoter (red arrow) activated in galactose media in set2Δ and a constitutively active antisense promoter (blue arrow). Light blue boxes show open reading frame (ORF) positions and orange box shows lncRNA transcription. Time course as in Fig. 1a is arrayed from bottom to top for the indicated strains. Increased blue colour indicates more transcript hybridizing to the array. (b) Histone methylation patterns of AAD10 and SUL1 from cells grown in YPD were analysed using the data set from Pokholok et al.. Whereas the AAD10 distal promoter and SUL1 antisense promoter have high levels of H3K4me3, the core mRNA promoters have high levels of H3K36me3. White boxes show the ORF position. (c) Northern blot analysis of AAD10 transcript with a 3′-strand-specific DNA probe. The indicated cells were grown in synthetic complete (SC) medium containing raffinose (Raf) and shifted to SC-galactose media for 120 min (Gal120). Bottom panels show two transcripts of AAD10 detected by northern blot analysis, which are schematicized at top. Blue arrow is a distal promoter that produces a lncRNA and red arrow is a proximal promoter for AAD10 mRNA transcription. A bar underneath upper panel indicates position of probe used for northern blot analysis. (d) The Set2–Rpd3S pathway deacetylates histones at the AAD10 core promoter. Crosslinked chromatin from the indicated strains grown in YPD (where set2Δ also derepresses AAD10) was precipitated with anti-H3 or anti-acetyl H4 as indicated. PCR analysis of the precipitated DNA was carried out on both distal (blue bars) and proximal (red bars) promoters of AAD10. A non-transcribed region near the telomere of chromosome VI was used for an internal control. The signals for acetyl H4 were quantitated and normalized to the total H3 signal, and the ratios were graphed. Error bars show the s.d. calculated from two biological replicates, each with three technical replicates. Similar results were obtained with cells grown in galactose for 120 min. WT, wild type.

Figure 3. Distinct effects of Set2 and Set3 on gene repression.

(a) DCI1 induction is delayed by Set3 but not by Set2. Cells indicated were grown in synthetic complete (SC) medium containing raffinose and shifted to SC-galactose media for 120 min (Gal120). Bottom panels show two transcripts of DCI1 detected by northern blot analysis, which are schematicized at top. Blue arrow is a distal promoter that produces a lncRNA and red arrow is a proximal promoter for DCI1 mRNA transcription. The number between a distal and a lncRNA promoter of upper panel indicates the distance. The bar underneath indicates the probe position used for northern blot analysis. The number underneath of the gel images is the average expression value of DCI1 from northern blot analyses with two independent RNA samples. (b) Set3C does not affect AAD10 induction. Northern blot analysis of AAD10 was done as in a. (c) Northern blot analyses for YNL068C and DAL5 were carried out as in a. WT, wild type.

Figure 4. Shortening the distance between the two promoters at AAD10 alleviates Set2–Rpd3S-mediated repression.

(a) Schematic representation of AAD10 locus showing positions of 500 (Δ500) and 941 bp (Δ941) deletions. The upstream and downstream promoters are shown in blue and red, respectively, and the long transcript is shown as blue arrow below. Numbers in parentheses indicate size of the associated open reading frame and black bar indicates the probe position used for northern blot analysis. (b) Northern blot analysis of AAD10 transcripts as in Fig. 3a. Strains from a were grown in synthetic complete (SC) medium containing raffinose (Ra) and shifted to SC-galactose media for 120 min (Gal120). SCR1 is used for a loading control (bottom). (c) H3K36me3 at the AAD10 promoter is significantly decreased in Δ941cells. Crosslinked chromatin from the indicated strains grown in YPD was precipitated with anti-H3 or anti-H3K36me3 as indicated. PCR analysis of the precipitated DNA was carried out on the core promoter of AAD10. A non-transcribed region near the telomere of chromosome VI was used for an internal control. The signals for anti-H3K36me3 were quantitated and normalized to the total H3 signal, and the ratios were graphed. Error bars show the s.d. calculated from two biological replicates, each with three technical replicates. (d) Set2-targeted histone deacetylation is abrogated in Δ941cells. ChIP assay for anti-acetyl H4 was done as in Fig. 2d. Error bars show the s.d. calculated from two biological replicates, each with three technical replicates. (e) Set2 repression of AAD10 is reduced in Δ941 cells. Northern blot analysis of AAD10 was done as in b. SCR1 is used as a loading control (bottom). (f) Insertion of a 941 bp of DNA from chromosome 10 restores AAD10 repression by Set2. Northern blot analysis of AAD10 was done as in b. SCR1 is used as a loading control (bottom). WT, wild type.

Figure 5. Set3C represses AAD10 transcription in Δ941 cells.

(a) Shortening the distance between the two promoters of AAD10 increases H3K4me2 at the core promoter. ChIP assay for H3K4me2 was done as in Fig. 4c. Error bars show the s.d. calculated from two biological replicates, each with three technical replicates. (b) Set3 deacetylates histones at AAD10 promoter in Δ941 cells. ChIP assay for anti-acetyl H4 was done as in Fig. 2d. Error bars show the s.d. calculated from two biological replicates, each with three technical replicates. (c) Loss of Set3 derepresses AAD10 transcription in Δ941 cells. Figure shows northern blot analysis of the upstream and AAD10 transcripts. The indicated strains were grown in synthetic complete (SC) medium containing raffinose (Ra) and shifted to SC-galactose media for 120 min (Gal120). SCR1 is used for a loading control (bottom). (d) Set3 delays induction of AAD10 in Δ941 cells. Northern blot analysis of AAD10 was done as in c. SCR1 is used as a loading control (bottom). (e) Set3C represses AAD10 transcription in Δ941 cells. Wild-type (WT) and a mutant for Set3C (hst1Δ) were analysed by northern blot as in c. (f) The average distance between mRNA and lncRNA promoters for Set3- or Set2-repressed genes. The distance between the two promoters of individual genes was measured using the tiling array data sets. Circles represent measured distance values, with x showing the mean value. The two boxes and two whiskers represent the four quartiles, with the line between the two boxes showing the median distance value.

Figure 6. Set2 represses inducible cryptic promoters.

(a) Tiling array data identified 757 cryptic transcripts increased in set2Δ. Venn diagram shows the number of genes with internal cryptic promoters repressed by Set2: 639 antisense and 118 sense direction. Among these, 55% (416 promoters) are more active during carbon source shifts in the absence of SET2, indicating that they are inducible cryptic promoters responding to environmental changes. In all, 44 cryptic promoters repressed by Set2 produce both sense and antisense divergent transcripts. (b) Examples of distinct cryptic promoter types. Red arrows and blue arrows show the core promoters and cryptic promoters, respectively. Light blue boxes show the open reading frame position. SLD3 or TMA108 has an internal cryptic promoter or an antisense promoter that is activated during galactose incubation. RAD28 gene has internal cryptic promoters that produce short sense and antisense transcripts. (c) Northern blot analysis of PCA1 cryptic transcripts. SET2 or set2Δ cells were grown in synthetic complete (SC) medium containing raffinose (Ra) and shifted to SC-galactose media for 120 min (Gal120). WT, wild type.

Figure 7. Models for regulation of gene expression by Set2–Rpd3S pathway and overlapping lncRNA transcription.

(a) At Set2-repressed promoters, transcription from a distal or antisense promoter targets H3K36me3 to the core promoter of mRNA target genes. Rpd3S deacetylates histones in the mRNA promoter region, resulting in delayed or reduced induction. For suppression of internal cryptic promoters, transcription from the mRNA promoter of mRNA gene targets H3K36me3 and Rpd3S deacetylation in 3′-transcribed regions. These cryptic promoters may also respond to environmental changes. (b) Position of the lncRNA promoter specifies the repressive effects of two distinct HDACs, Set3C and Rpd3S. H3K4me2 and H3K36me3 are targeted to 5′- and 3′-regions of lncRNA expression, respectively. In 5′-regions, Set3C binds to H3K4me2 to deacetylate histones and repress the mRNA promoter. In contrast, in 3′-regions of lncRNA transcription, Rpd3S binds to H3K36me3 to deacetylate histones and repress mRNA promoters. These mechanisms are also applicable to genes with antisense transcription.