H3K4 trimethylation by Set1 promotes efficient termination by the Nrd1-Nab3-Sen1 pathway

Abstract

In Saccharomyces cerevisiae, the Nrd1-Nab3-Sen1 pathway mediates the termination of snoRNAs and cryptic unstable transcripts (CUTs). Both Nrd1 and the Set1 histone H3K4 methyltransferase complex interact with RNA polymerase II (Pol II) during early elongation, leading us to test whether these two processes are functionally linked. The deletion of SET1 exacerbates the growth rate and termination defects of nrd1 mutants. Set1 is important for the appropriate recruitment of Nrd1. Additionally, Set1 modulates histone acetylation levels in the promoter-proximal region via the Rpd3L deacetylase and NuA3 acetyltransferase complexes, both of which contain PHD finger proteins that bind methylated H3K4. Increased levels of histone acetylation reduce the efficiency of Nrd1-dependent termination. We speculate that Set1 promotes proper early termination by the Nrd1-Nab3-Sen1 complex by affecting the kinetics of Pol II transcription in early elongation.

Fig. 1.

SET1 interacts genetically with NRD1. (A) Schematic representation of the Nrd1 domain structure and deletion mutants. Annotated are the CTD interaction domain (CID), the Nab3 interaction domain (NID), and the RNA recognition motif (RRM). (B) Growth of the WT (BY4741), set1Δ (YSB2253), pho23Δ (YF688), nrd1(1-164Δ) (YSB2094), nrd1(1-164Δ) set1Δ (YSB2153), and nrd1(1-164Δ) pho23Δ (YSB2409) strains. Strains were plated with 10-fold serial dilutions on synthetic complete (SC) medium and incubated for 2 days at 30°C. (C) Growth curves of WT (BY4741), set1Δ, nrd1(151-214Δ), nrd1(6-150Δ), nrd1(151-214Δ) set1Δ, and nrd1(6-150Δ) set1Δ strains assayed in liquid synthetic complete medium at 30°C starting from an A₆₀₀ of 0.1. For strain information see Table 1. The graph key shows the calculated doubling time (Td) of each strain in parentheses.

Fig. 2.

Deletion of SET1 enhances the termination defect of nrd1 and nab3 mutants. (A, top) Schematic representation of the SNR13 region. SNR13 is upstream of the infrequently transcribed TRS31 gene. The probe used for hybridization is indicated by the horizontal bar below the schematic. (Bottom) Total RNA (5 μg) from WT (BY4741), set1Δ (YSB2253), nrd1(1-164Δ) (YSB2094), nrd1(1-164Δ) set1Δ (YSB2153), nrd1(151-214Δ) (YSB2086), and nrd1(151-214Δ) set1Δ (YSB2152) strains was analyzed by Northern blot using the SNR13 probe. The positions of mature full-length snR13 and stable snR13-TRS31 read-through RNAs are marked on the left. The bottom shows methylene blue staining of 25S and 18S rRNAs as a loading control. Transcripts were quantitated with a phosphorimager. snR13 read-through transcripts were calculated as a percentage of full-length snR13. These numbers were divided by the value for the nrd1(151-214Δ) strain to give the “fold over nrd1(151-214Δ)” values listed below each lane. (B) Effect of set1Δ on the snR13 read-through transcript levels in nrd1(1-164Δ) and nrd1(151-214Δ) mutants quantified as the ratio of the nrd1 set1 double mutant to the nrd1 mutant alone using values from at least four Northern blot experiments. Error bars show standard deviations. (C) RT-PCR of snR13 read-through (26 cycles, using primers downstream of the normal termination site) and total (full-length plus read-through) transcripts (20 cycles, using primers within the mature snR13) were performed on RNA from three independent cultures of nab3-11 (YF1471) and nab3-11 set1Δ (YSB2559) strains grown at the semipermissive temperature of 30°C. Quantitation by phosphorimager analysis is expressed by dividing the read-through by the total signal. Because these ratios were not corrected for numbers of PCR cycles, they are a relative, not absolute, measure of the read-through. This experiment was repeated by using RT-qPCR, which gave similar results (see Fig. S1B in the supplemental material). (D, top) Schematic representation of the rDNA region. 35S rRNA genes are interspersed with nontranscribed spacer 1 (NTS1), the 5S rRNA gene (black arrow), and nontranscribed spacer 2 (NTS2). The position of the probe used to detect the NTS1 CUT is shown below the diagram as a black line. (Bottom) RNA samples as in panel A were analyzed by Northern blotting using the NTS1 CUT probe. Levels of the NTS1 CUT were normalized to the methylene blue 25S rRNA signal. These values for each strain were then divided by the value of the nrd1(151-214Δ) strain to give the “fold over nrd1(151-214Δ)” values that are listed below each lane.

Fig. 3.

Effect of Set1 on Nrd1-dependent termination is dependent on H3K4me3. (A) Schematic representation of the Set1 domain structure and mutants used in this study. Set1 has two predicted RNA recognition motifs (RRM1 and RRM2) in its N-terminal region. A SET domain in the C-terminal region is responsible for the methyltransferase activity of Set1. The single asterisk represents the position of the H422A point mutant that abolishes the in vitro RNA binding activity of Set1. The double asterisk shows the H1017K mutation in the methyltransferase active site of Set1. (B) An nrd1(1-164Δ) set1Δ strain (YSB2153) was transformed with the indicated plasmids expressing WT or mutant Set1 (the SET1 allele on the plasmid is shown in square brackets) (for strain names and full genotypes see YSB2507 to YSB2511 in Table 1). The growth of the strains was assayed by 10-fold serial dilutions on synthetic complete medium plates lacking uracil at 30°C for 2 days. (C) Northern blot analysis of total RNA (5 μg) extracted from strains BY4741, YSB2253, YSB2086, YSB2502, YSB2152, YSB2505, YSB2506, and YSB2503. The plasmid-borne SET1 alleles transformed into nrd1(1-164Δ) set1Δ cells are shown in square brackets. The positions of mature full-length snR13 and snR13-TRS31 read-through transcripts are indicated on the left. The methylene blue staining of 25S and 18S rRNAs as a loading control is shown at the bottom. The numbers below the gels show the quantitation of the snR13 read-through transcript level as fold over nrd1(151-214Δ) as described in the legend of Fig. 2A. (D) Northern blot analysis as in C except with the blot probed for the NTS1 CUT. (E) Northern blot analysis as in panel C using total RNA (5 μg) extracted from strains YF534, YSB2094, YSB2415, and YSB2153. (F) RT-PCR of snR13 read-through transcripts (31 cycles) and total (full-length plus read-through) transcripts (20 cycles) performed by using three independent cultures of WT (YSB2256) and H3K4R (YSB2257) strains grown at 30°C.

Fig. 4.

Cells lacking Set1 have reduced Nrd1 recruitment to snoRNAs. (A) Schematic representation of the chromatin immunoprecipitation (ChIP) PCR probes spanning the SNR13 region. See Table 2 for primer names and sequences. (B) ChIP analysis of SNR13 using anti-Nrd1 antibody was carried out with nrd1(151-214Δ) (YSB2086), nrd1(151-214Δ) set1Δ (YSB2152), and nrd1(151-214Δ) set1(230-335Δ) (YSB2503) cells. (Left) Representative ChIP gel. The asterisk denotes a PCR product from a nontranscribed region used as a background control for normalization. (Right) Quantitation of data from three independent experiments normalized to anti-Rpb3 ChIPs (see panel D) using the same chromatin samples. Error bars indicate standard errors. (C) Immunoblot analysis of Nrd1 (antihemagglutinin) and Rpb3 levels in nrd1(151-214Δ) (YSB2086) and nrd1(151-214Δ) set1Δ (YSB2152) cells. Twenty micrograms of whole-cell extract from each strain was resolved on 12% SDS-PAGE gels. The quantitation of Nrd1 protein levels is shown below the blots. (D) Quantitation of three independent anti-Rpb3 ChIP experiments with nrd1(151-214Δ) (YSB2086), nrd1(151-214Δ) set1Δ (YSB2152), and nrd1(151-214Δ) set1(230-335Δ) (YSB2503) cells.

Fig. 5.

Increased histone acetylation levels in pho23Δ cells enhance the termination defects of nrd1 mutants. (A) Northern blot analysis of total RNA (5 μg) extracted from set1Δ (YSB2253), pho23Δ (YF688), nrd1(151-214Δ) (YSB2086), nrd1(151-214Δ) set1Δ (YSB2152), nrd1(151-214Δ) pho23Δ (YSB2412), and nrd1(151-214Δ) pho23Δ set1Δ (YSB2420) strains was carried out and quantitated as described in the legend of Fig. 2. (B) ChIP analysis of SNR13 using anti-tetra-acetyl H4 antibody was carried out with WT (BY4741) and pho23Δ (YF688) cells. (Left) Representative ChIP gel, with the asterisk denoting a PCR product from a nontranscribed control region. (Right) Quantitation of three independent experiments normalized to anti-H3 ChIP data. Error bars indicate standard errors. (C) ChIP-chip trace for Rpd3-Myc for the snR13 region. Data were reported previously by Bumgarner et al. (7) and are accessible through ArrayExpress accession number E-MEXP-2269. (D) Northern blot analysis of total RNA extracted from yng1Δ (YF1495), nrd1(151-214Δ) (YSB2086), nrd1(151-214Δ) pho23Δ (YSB2412), nrd1(151-214Δ) pho23 yng1Δ (YSB2492), and nrd1(151-214Δ) yng1Δ (YSB2489) strains and probed for snR13. The positions of mature full-length snR13 and snR13-TRS31 read-through transcripts are indicated on the left. The quantitation of the snR13 read-through transcript level is expressed as described in the legend of Fig. 2. (E) Quantitation of three independent anti-tetra-acetyl H4 ChIP experiments of snR13 in WT (BY4741), set1Δ (YSB2253), and yng1Δ (YF1495) cells. Error bars represent standard errors. (F) RT-PCR of snR13 read-through (26 cycles) and total (full-length plus read-through) transcripts (20 cycles) performed by using three independent cultures of nrd1(151-214Δ) (YSB2086), nrd1(151-214Δ) set1Δ (YSB2152), and nrd1(151-214Δ) pho23Δ (YSB2412) strains. The quantitation was done by the normalization of the read-through signal to the total signal. This result was also verified by qPCR (see Fig. S1C in the supplemental material). (G) Northern blot analysis of RNA as in panel A except with the blot probed for the NTS1 CUT (position indicated on the left). The quantitation of the NTS1 CUT level as “fold over nrd1(151-214Δ)” is shown below the gels.

Fig. 6.

Deletion of SET1 or PHO23 affects termination of multiple snoRNAs and CUTs. (A) RT-PCRs for snR128 read-through transcripts (26 cycles) and total (full-length plus read-through) transcripts (20 cycles) were performed by using WT (BY4741), set1Δ (YSB2253) pho23Δ (YF688), nrd1(151-214Δ) (YSB2086), nrd1(151-214Δ) set1Δ (YSB2152), and nrd1(151-214Δ) pho23Δ (YSB2412) strains. The schematic above the gel represents the position of read-through and total primer pairs used for the PCRs of snoRNAs. The graph on the right shows the quantitation of the snR128 read-through signal relative to the total snR128 transcript signal. The x axis indicates the strains used, designated below the gels. Error bars indicate standard errors from three biological replicates. (B) RT-PCRs for the NEL025C CUT (30 cycles) and snR190 transcripts (20 cycles) performed as described above for panel A. The schematic above the gel shows the position of the NEL025C initiation site (arrow) and the primer pair used for the PCRs. The snR190 signal shows total transcript levels that do not change with defects in the Nrd1 termination pathway and was therefore used for normalization. The graph on the right shows the quantitation of the NEL025C CUT signal relative to snR190. x axis numbers refer to the numbered lanes at the left. (C) RT-PCRs for snR3, snR71, and snR9 read-through transcripts (30 cycles) and total (full-length plus read-through) transcripts (20 cycles) were performed with WT (BY4741), set1Δ (YSB2253), and pho23Δ (YF688) strains. The numbers below each panel represent the quantitation of read-through signal relative to total transcript signal levels. (D) RT-PCRs for the FMP40 CUT (30 cycles), the RPR2 CUT (30 cycles), and snR190 total transcripts (20 cycles) were performed as in panel C. The schematic above the gels shows the positions of the CUT initiation site and the primer pairs used for the PCRs.

Fig. 7.

NET-Seq analysis of set1Δ cells. (A) Sequenced 3′ ends of nascent transcripts from WT (BY4741) and set1Δ cells were mapped to the chromosomal (chr) locations of snoRNAs. The x axis shows the positions of nucleotides along the corresponding chromosome. The y axis shows the number of sequence reads for each nucleotide position in WT (blue, above zero) or set1Δ (orange, below zero) cells. The positions of snR9 (top left), snR71 (top right), and snR13 (bottom) are shown above each graph. (B) NET-Seq results for the NEL025C CUT (left) and FMP40-proximal CUT883 (right).

Fig. 8.

Pol II kinetics can affect Nrd1-dependent termination. (A) RT-PCRs for snR13 read-through (26 cycles) and total (full-length plus read-through) transcripts (20 cycles) were performed by using three independent cultures of nrd1(151-214Δ) set1Δ (YSB2582) and nrd1(151-214Δ) set1Δ rpb2-10 (YSB2583) (left) or nrd1(151-214Δ) pho23Δ (YSB2574) and nrd1(151-214Δ) pho23Δ rpb2-10 (YSB2595) (right) strains. The graphs show the quantitation as read-through signal relative to the total signal. These results were also confirmed by qPCR (see Fig. S1D in the supplemental material). (B) Growth rates of nrd1(151-214Δ) set1Δ (YSB2582), nrd1(151-214Δ) set1Δ rpb2-10 (YSB2583), nrd1(151-214Δ) pho23Δ (YSB2574), and nrd1(151-214Δ) pho23Δ rpb2-10 (YSB2595) strains were assayed in liquid synthetic complete medium at 30°C starting from an A₆₀₀ of 0.1. The graph key shows the calculated doubling time (Td) of each strain in parentheses. (C) Proposed model for how H3K4me3 in the promoter-proximal region affects histone acetylation and the early termination pathway. The NuA3 acetyltransferase complex may recognize H3K4me3 via the Yng1 PHD finger. Similarly, the Pho23 PHD finger may bring the Rpd3L deacetylase complex to H3K4me3-modified nucleosomes at active promoters. The actions of these two complexes maintain an optimum level of histone acetylation (black circles), which promotes proper Nrd1-dependent termination by affecting the kinetics of early Pol II elongation.