Human SWI/SNF-associated PRMT5 methylates histone H3 arginine 8 and negatively regulates expression of ST7 and NM23 tumor suppressor genes

Abstract

Protein arginine methyltransferases (PRMTs) have been implicated in transcriptional activation and repression, but their role in controlling cell growth and proliferation remains obscure. We have recently shown that PRMT5 can interact with flag-tagged BRG1- and hBRM-based hSWI/SNF chromatin remodelers and that both complexes can specifically methylate histones H3 and H4. Here we report that PRMT5 can be found in association with endogenous hSWI/SNF complexes, which can methylate H3 and H4 N-terminal tails, and show that H3 arginine 8 and H4 arginine 3 are preferred sites of methylation by recombinant and hSWI/SNF-associated PRMT5. To elucidate the role played by PRMT5 in gene regulation, we have established a PRMT5 antisense cell line and determined by microarray analysis that more genes are derepressed when PRMT5 levels are reduced. Among the affected genes, we show that suppressor of tumorigenicity 7 (ST7) and nonmetastatic 23 (NM23) are direct targets of PRMT5-containing BRG1 and hBRM complexes. Furthermore, we demonstrate that expression of ST7 and NM23 is reduced in a cell line that overexpresses PRMT5 and that this decrease in expression correlates with H3R8 methylation, H3K9 deacetylation, and increased transformation of NIH 3T3 cells. These findings suggest that the BRG1- and hBRM-associated PRMT5 regulates cell growth and proliferation by controlling expression of genes involved in tumor suppression.

FIG. 1.

PRMT5 coelutes with endogenous BRG1 and hBRM complexes. HeLa nuclear extracts (300 mg) were loaded on a 25-ml BioRex 70 column. After the flowthrough (FT) was collected, the column was washed with increasing salt concentrations as indicated. Fractions were collected and analyzed by Western blotting with either anti-PRMT5, anti-BRG1, or anti-hBRM antibodies. Peak fractions from the BioRex 70 column (fractions 4 to 28 for the 0.5 M KCl elution and fractions 4 to 12 for the 0.8 M KCl elution) were pooled and dialyzed against BC100 buffer. Dialyzed proteins were loaded on two DE 52 columns and eluted as described for the BioRex 70 column. Peak fractions (0.3 M KCl) were pooled and immunoprecipitated using either preimmune (PI) or immune (I) anti-PRMT5 antibodies. After extensive washing, proteins were analyzed by Western blotting with the indicated antibodies.

FIG. 2.

PRMT5 coelutes with flag-tagged INI1 complexes. Either HeLa S3 or Fl-INI1 nuclear extract (180 mg) was incubated with anti-flag M2 affinity gel, and hSWI/SNF complexes were eluted with a 20-fold molar excess of flag peptide. Eluted complexes were analyzed by silver staining (A) or by Western blotting (B) with specific antibodies. Western blot assays were performed using 20 μg of Fl-INI1 nuclear extract (Input, lane 1), 15 μl of eluted fraction with HeLa nuclear extract (Ctrl, lane 2), and 15 μl of eluted flag-tagged hSWI/SNF (Fl-hSWI/SNF) complexes (lane 3). (C) PRMT5 does not cross-react with anti-flag antibodies. In vitro-translated and ³⁵S-labeled PRMT5 was incubated with either agarose beads or anti-flag M2 antibody cross-linked to agarose beads. As a control, preimmune (PI) and immune (I) anti-PRMT5 antibodies were also included. The immunoprecipitated complexes were washed as described for panel A and visualized by autoradiography. (D) Recombinant and hSWI/SNF-associated PRMT5 can methylate histones H3 and H4. H1-depleted HeLa core histones were incubated with either Fl-hSWI/SNF complexes, affinity-purified wild-type (WT) Fl-PRMT5, or mutant (Mut) Fl-PRMT5/G367A-R368A in the presence of [³H]SAM. Histones were visualized by Coomassie blue staining, and methylated products were detected by autoradiography. (E) Levels of PRMT5 in Fl-hSWI/SNF complexes were quantitated by Western blot analysis with increasing amounts of either immunopurified Fl-hSWI/SNF fractions (100, 200, and 400 ng) or Sf9-expressed and affinity-purified Fl-PRMT5 (6.25, 12.5, and 25 ng). The asterisk indicates a long exposure of the anti-PRMT5 Western blot. (F) Methylation of histone H3 and H4 N-terminal tails was carried out as described for panel D. Reaction mixtures were spotted onto Whatman P-81 filter paper, and methylated peptides were quantitated by liquid scintillation counting as described in Materials and Methods. As controls, methylation of BSA and H3 peptide (aa 60 to 84) is shown.

FIG. 3.

Recombinant and hSWI/SNF-associated PRMT5 can specifically methylate H3R8 and H4R3. Wild-type and mutant H3 peptides containing an arginine-to-alanine mutation at a single position (2, 8, or 17) or at all three positions were incubated with either Fl-hSWI/SNF complexes (A) or flag-tagged wild-type PRMT5 (B) in the presence of [³H]SAM. Similarly, wild-type and mutant H4 peptides with either a single point mutation (R3A, R17A, or R19A) or a triple point mutation (R3A/R17A/R19A) were incubated with either Fl-hSWI/SNF complexes (C) or Fl-PRMT5 (D), and samples were processed as described for Fig. 2E. (E) H3 peptides acetylated at K9 or K14 were incubated with Fl-hSWI/SNF as described above. As controls, BSA and wild-type H3 peptide are shown.

FIG. 4.

Characterization of sense and antisense (AS) PRMT5 cell lines. (A) Expression of endogenous PRMT5 was assessed in NIH 3T3 and AS-PRMT5 cell lines. RT-PCR was performed using 20 μg of total RNA from either NIH 3T3 cells (lanes 1 to 3) or AS-PRMT5 cells (lanes 4 to 6) with primers specific for either PRMT5 or GAPDH. PCRs were performed using 2 μl (lanes 1, 3, 4, and 6) or 0.2 μl (lanes 2 and 5) of the respective 20-μl RT reaction mixtures. Control (Ctrl) represents PCRs lacking the 5′ primer (lanes 1 and 4). Primers used to determine the PRMT5 transcript levels were placed downstream (+1093 to +1306) from the sequences used to knock down PRMT5 (+39 to +1165). (B) Increasing amounts (10, 20, and 40 μg) of whole-cell extracts from NIH 3T3 and AS-PRMT5 cells were analyzed by Western blotting with anti-PRMT5 and anti-MAD antibodies. For quantitation purposes, increasing amounts (6.25, 12.5, and 25 ng) of Sf9-expressed and affinity-purified Fl-PRMT5 were analyzed on the same blot. The asterisks indicate long exposures of anti-PRMT5 and anti-MAD Western blots. (C and D) RT-PCR was conducted as described for panel A with primers specific for the indicated up-regulated genes (C) and down-regulated genes (D). As a control, GAPDH levels were also analyzed.

FIG. 5.

PRMT5 induces cell growth and proliferation. (A) Proliferation of NIH 3T3, puromycin-resistant stable cell lines that express AS-PRMT5 or Fl-PRMT5, or hygromycin-resistant MYC/RAS-transformed NIH 3T3 cells was measured using 2 × 10⁵ cells as described in Materials and Methods. The experiment was repeated three times, and the data points represent the average count from six plates. Standard deviations are included but are too small for the error bars to appear on the graph. (B) Approximately 40 μg of whole-cell extract from the indicated cell lines was analyzed by Western blotting with anti-flag antibodies to detect expression of Fl-PRMT5 in the flag-tagged PRMT5 cell line. The same blot was stripped and probed with either anti-PRMT5 or anti-MAD antibodies. (C) BrdU incorporation in NIH 3T3, AS-PRMT5, and Fl-PRMT5 cells was determined after 4.5 or 9 h of incubation with BrdU as described in Materials and Methods. The percentage of BrdU-positive cells was determined by FACS analysis. (D) NIH 3T3, AS-PRMT5, and Fl-PRMT5 cells were grown for 4 days and stained with propidium iodide, and the DNA content of each cell line was analyzed by FACS analysis. The percentage of cells in each stage of the cell cycle including cells undergoing apoptosis (A) is shown.

FIG. 6.

Overexpression of PRMT5 stimulates anchorage-dependent and -independent growth. (A) Approximately 4 × 10³ cells of either puromycin-resistant NIH 3T3, AS-PRMT5, Fl-PRMT5, or hygromycin-resistant MYC/RAS-transformed cell lines were grown in medium containing 10% fetal bovine serum, and after 7 days colonies were stained with crystal violet. (B) Equal numbers (2 × 10²) of cells of drug-resistant cell lines containing either vector alone, AS-PRMT5, Fl-PRMT5, or MYC and RAS were grown in soft agar for 10 days. Representative pictures showing the morphology and size of transformed cells are shown at an approximately 35× magnification. Colony formation assays were performed in triplicate and repeated three times. The number shown below each figure represents the average number of colonies from nine plates.

FIG. 7.

BRG1- and hBRM-associated PRMT5 is directly involved in transcriptional repression of ST7 and NM23. (A) RT-PCR was performed on 10 μg of total RNA from either Fl-PRMT5, NIH 3T3, or AS-PRMT5 cell lines with primers specific for MYT1l, NM23, ST7, and GAPDH. PCR for each gene was carried out using either 2 μl (lanes 1, 3, 4, 6, 7, and 9) or 0.2 μl (lanes 2, 5, and 8) of the RT reaction mixture. Ctrl represents PCRs without 5′ primer (lanes 1, 4, and 7). (B, C, and E) ChIP assays were conducted using cross-linked chromatin from either Fl-PRMT5, NIH 3T3, or AS-PRMT5 cells as described in Materials and Methods with either preimmune (PI) or immune (I) anti-PRMT5 or anti-flag antibodies (B), anti-H3(Me₂)R8 antibodies (C), or anti-H3AcK9 antibodies (E). As controls, mock (reaction mixture without chromatin) and no-antibody (Ab) (reaction mixture with chromatin but without antibody) reactions are shown. For mock, no-Ab, PI, and I reactions, 10 μl of eluted DNA was amplified and 15 μl of each PCR mixture was analyzed. For the input lane, 0.6 μl of eluted DNA was PCR amplified and 10 μl was loaded on the gel. Specific primer pairs were used to amplify MYT1l (−258 to +214), NM23 (−211 to +254), and ST7 (−228 to +209) sequences. (D) Specificity of anti-H3(Me₂)R8 was determined by Western blot analysis with 1 and 2 μg of symmetrically methylated H3R8 peptide, unmethylated N-terminal and internal H3 peptides, or BSA.

FIG. 8.

BRG1 and hBRM are differentially recruited to methylated ST7 and NM23 promoters. Cross-linked chromatin was prepared from asynchronous Fl-PRMT5, NIH 3T3, and AS-PRMT5 cell lines, and ChIP assays were conducted as described for Fig. 7 with either preimmune (PI) or immune (I) anti-BRG1 (A) and anti-hBRM (B) antibodies. As controls, mock and no-antibody (Ab) reactions are shown. (C) Levels of catalytically inactive BRG1 and hBRM were measured by Western blotting with 30 μg of nuclear extract from either mutant (Mut) BRG1 or hBRM cell lines. HeLa S3 nuclear extract was used as a control, and proteins were detected using the indicated antibodies. (D) MYT1l, NM23, ST7, and GAPDH transcript levels were analyzed by RT-PCR as described for Fig. 7A with RNA from either HeLa cells or HeLa cells that express either mutant BRG1 or hBRM. PCRs were carried out using either 2 μl (lanes 1, 3, 5, and 7) or 0.2 μl (lanes 2, 4, and 6) of the RT reaction mixture. Ctrl represents PCRs without 5′ primer (lane 1).