Transcriptome-wide analysis for glucocorticoid receptor-mediated mRNA decay reveals various classes of target transcripts

Abstract

The glucocorticoid receptor (GR) can bind to DNA or RNA, eliciting transcriptional activation/repression or rapid messenger RNA (mRNA) degradation, respectively. Although GR-mediated transcriptional regulation has been well-characterized, the molecular details of rapid mRNA degradation induced by glucocorticoids are not yet fully understood. Here, we demonstrate that glucocorticoid-induced GR-mediated mRNA decay (GMD) takes place in the nucleus and the cytoplasm, acting on pre-mRNAs and mRNAs. We also performed cross-linking and immunoprecipitation coupled with high-throughput sequencing analysis for GMD factors (GR, YBX1, and HRSP12) and mRNA sequencing analysis to identify endogenous GMD substrates. Our comprehensive coupled with high-throughput sequencing and mRNA sequencing analyses reveal that a range of cellular transcripts containing a common binding site for GR, YBX1, and HRSP12 are preferential targets for GMD, suggesting possible new functions of GMD in various biological events.

Keywords: Glucocorticoid; Glucocorticoid receptor; Heat-responsive protein 12; Proline-rich nuclear receptor 2; messenger RNA decay.

Fig. 1

GMD targets both pre-mRNA and mRNA. (A) Schematic of C5′-RL-gGl, a GMD reporter construct containing 2 introns. Oligonucleotides for the amplification of pre-mRNA and its spliced mRNA are indicated by arrows. (B) GMD of pre-mRNA and mRNA. HeLa cells transiently expressing C5′-RL-gGl mRNAs, and as a control, firefly luciferase (FLuc) mRNA, were either untreated or treated with Dex for 12 hours before harvesting. The levels of C5′-RL-gGl pre-mRNA and mRNA were normalized to those of FLuc mRNAs. Normalized levels obtained from cells not treated with Dex were arbitrarily set to 100%. (C) Relative levels of endogenous pre-mRNA and mRNA of GMD target transcripts. HeLa cells were either treated or not treated with Dex for 1 hour before harvesting. The levels of endogenous pre-mRNA and mRNA of GMD target transcripts were normalized to the levels of GAPDH. Normalized levels obtained from cells not treated with Dex were arbitrarily set to 100%. (D and E) The effect of downregulation of GR or HRSP12 on GMD. (D) Specific downregulation of GR or HRSP12 using siRNA was confirmed via western blotting. For a quantitative comparison, 3-fold serial dilutions of total cell extracts were loaded in the 3 leftmost lanes. (E) The levels of endogenous pre-mRNA and mRNA of GMD target transcripts were normalized to those of endogenous GAPDH mRNAs. Endogenous CCL5 pre-mRNA and mRNA lacking a GR-binding site served as a negative control. The columns and bars in each panel represent the mean and standard deviation of 3 independent biological replicates (n = 3). Two-tailed, equal-sample variance Student’s t-tests were used to calculate the P values. **P < .01; *P < .05. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GMD, GR-mediated mRNA decay; GR, glucocorticoid receptor.

Fig. 2

GMD of both pre-mRNA and mRNA requires HRSP12 trimerization. (A and B) Effect of HRSP12 KO on the GMD of CCL2 mRNA. (A) Specific KO of HRSP12 in HAP1 cells was confirmed via western blotting. (B) GMD efficiency of endogenous CCL2 mRNA was analyzed via qRT-PCR. (C and D) Complementation experiments with FLAG-HRSP12-WT and -P105A/R107E. (C) Expression of FLAG-HRSP12-WT and its variant was comparable to that of endogenous HRSP12, as confirmed via western blotting. (D) GMD of pre-mRNA and mRNA of endogenous GMD target transcripts. n = 3; *P < .01; *P < .05. GMD, glucocorticoid receptor-mediated mRNA decay.

Fig. 3

GMD occurs both in the nucleus and cytoplasm. Extracts of HeLa cells treated or not treated with Dex for 1 hour were fractionated into the nuclear (N) and cytoplasmic (C) extracts. (A) The distribution of GMD factors was determined via western blotting. All results are representative of 3 biological replicates. (B) The GMD of pre-mRNA and mRNA among GMD target transcripts in the nucleus and cytoplasm was analyzed by qRT-PCR. n = 3; **P < .01; *P < .05. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GMD, GR-mediated mRNA decay; GR, glucocorticoid receptor; qRT-PCR, quantitative real-time PCR.

Fig. 4

CLIP-seq analysis of endogenous GR, MYC-YBX1, and MYC-HRSP12 in HeLa cells. CLIP-seq experiments were conducted as described in the Materials and Methods section. (A-C) Scatter plots of reads obtained for individual genes. All CLIP-seq experiments with endogenous GR (A), MYC-YBX1 (B), and MYC-HRSP12 (C) were performed on 2 biological replicates. Each dot represents a gene. The Pearson’s correlation coefficients (r) are indicated. (D-F) Relative peak distributions of GR (D), YBX1 (E), and HRSP12 (F). The number of peaks shown in Figures S2D-F was divided by the length of nucleotides of each region in the genome. 3′UTR, 3′-untranslated region; GR, glucocorticoid receptor; HRSP12, heat-responsive protein 12; YBX1, Y-box-binding protein 1.

Fig. 5

A common binding site for GR, YBX1, and HRSP12 is essential for efficient GMD. (A-C) In the transcriptome-wide analysis of CLIP-seq data on GR, YBX1, and HRSP12, only the peaks that overlapped from 2 biological replicates of each CLIP-seq experiment were selected. The Venn diagrams display the number of peaks (A), the number of genes (B), which have the corresponding peaks, and the number of genes (C), which have the corresponding peaks and are downregulated after Dex treatment. (D-F) Cumulative distribution function plots for GMD factor dependency. Fold-change values in transcript levels following Dex treatment were normalized to those without Dex treatment. Next, the relative fold changes after downregulation of GR (siGR, D), YBX1 (siYBX1, E), or HRSP12 (siHRSP12, F) were further normalized to those before downregulation. The relative changes were compared among group I (black), group II (blue), and group III (red). P values were calculated using the 2-sided Mann-Whitney U test. (G and H) Box plot analysis shows that GMD is independent of the position (G) and the number of (H) the common peaks. 3′UTR, 3′-untranslated region; 5′UTR, 5′-untranslated region; CDS, coding sequence; GR, glucocorticoid receptor; HRSP12, heat-responsive protein 12; YBX1, Y-box-binding protein 1.

Fig. 6

GMD requires an intact CBS. (A) Schematic of a GMD reporter construct containing a CBS split by an intron. The full-length CCL2 5′UTR and CBS are indicated in red and yellow, respectively. Oligonucleotides for the amplification of pre-mRNA and its spliced mRNA are indicated by the arrows. (B) qRT-PCR of pre-mRNA and mRNA generated from the GMD reporter construct. (C and D) Read maps for transcripts harboring a CBS split by an intron. The reads obtained from each CLIP-seq experiment were mapped to the UCSC reference genome (hg19). Read maps for the HYI (C) and CCBE1 (D) transcripts are shown. Genomic regions corresponding to the identified peaks are underlined in green. (E) Validation for the GMD of newly identified transcripts containing a CBS split by an intron. Cellular transcripts containing a CBS split by an intron were identified in the CLIP-seq data. The efficiency of GMD of pre-mRNA and mRNA among the identified transcripts was analyzed via qRT-PCR. n = 3; **P < .01; *P < .05. 5′UTR, 5′-untranslated region; CBS, common binding site; GMD, glucocorticoid receptor-mediated mRNA decay; HRSP12, heat-responsive protein 12; qRT-PCR, quantitative real-time PCR; YBX1, Y-box-binding protein 1.