Crosstalk between histone modifications indicates that inhibition of arginine methyltransferase CARM1 activity reverses HIV latency

Abstract

In eukaryotic cells, the gene expression status is strictly controlled by epigenetic modifications on chromatin. The repressive status of chromatin largely contributes to HIV latency. Studies have shown that modification of histone H3K27 acts as a key molecular switch for activation or suppression of many cellular genes. In this study, we found that K27-acetylated histone H3 specifically recruited Super Elongation Complex (SEC), the transcriptional elongation complex essential for HIV-1 long terminal repeat (LTR)-mediated and general cellular transcription. Interestingly, H3K27 acetylation further stimulates H3R26 methylation, which subsequently abrogates the recruitment of SEC, forming a negative feedback regulatory loop. Importantly, by inhibiting methyltransferase activity of CARM1, the enzyme responsible for H3R26 methylation, HIV-1 transcription is reactivated in several HIV latency cell models, including a primary resting CD4+ T cell model. When combined with other latency disrupting compounds such as JQ1 or vorinostat/SAHA, the CARM1 inhibitor achieved synergistic effects on HIV-1 activation. This study suggests that coordinated and dynamic modifications at histone H3K27 and H3R26 orchestrate HIV-1 LTR-mediated transcription, and potentially opens a new avenue to disrupt latent HIV-1 infection by targeting specific epigenetic enzymes.

Figure 1.

Histone H3 with K27 acetylation is a better substrate for CARM1 methylation. (A) In vitro histone methylation assay catalyzed by CARM1. Unmodified and K27-acetylated recombinant histone H3 proteins were used as substrates, and the methylation activity was determined either by tritium-labeled assay (top panel), or western blot analysis (bottom panel). The relative intensity of signals was determined by ImageJ. (B) Unmodified and K27-acetylated histone peptides (amino acids 23–42) were used as substrates for tritium-labeled in vitro methylation assay catalyzed by CARM1. (C) A sequence alignment of known CARM1 substrates. The ‘R’ residues in red are CARM1 methylation sites. (D) In vitro methylation assay using wildtype histone H3 or histone H3 with a K27M mutation. The names of the recombinant methyltransferases/acetyltransferase were shown in blue. (E) Unmodified and R26-methylated histone H3 peptides (amino acids 18–35) harboring the K27M mutation were used in a methylation assay catalyzed by CARM1 in vitro. (F) Western blot analysis of histones extracted from 293T cells stably expressing vector, wildtype histone H3, or K27M mutated histone H3. The relative intensity of signals was determined by ImageJ. (G) Histone methylation assay catalyzed by CARM1. Unmodified, K27-acetylated, and K27-trimethylated histone H3 proteins were used as substrates. The total histone proteins were shown by Blue stain reagent, and H3R26me2a level was shown by Western blot analysis. The relative intensity of H3R26me2a was normalized by H3 and shown at the bottom. (H) A cartoon model of H3K27ac-regulated H3R26 methylation.

Figure 2.

Super Elongation Complex (SEC) are H3K27ac ‘readers’. (A) Key components of SEC were highly enriched on H3K27ac peptide by pull-down analysis. The fold changes over unmodified histone H3 peptide are shown as red (enriched) or blue (reduced). A complete list of all proteins identified by mass spectrometry is shown in Table 1. (B) Peptide pull-down followed by silver stain. (C) Peptide pull-down followed by Western blot analysis to confirm the mass spectrum results in (A). The peptides shown on the bottom were in proportion to the amount of peptides used in the pull-down. (D) Immunoprecipitation (IP) of SEC subunits from HeLa nuclear extract by recombinant full-length histone H3 proteins (unmodified or K27-acetylated). Histone H3 antibody and protein G beads were added to the IP samples following 2 h of incubation. The mock sample contains all the reagents except for recombinant histone H3. The amount of histones used in the IP was shown by Blue stain.

Figure 3.

CARM1 methyltransferase activity is repressive for HIV-1 LTR-mediated transcription. (A) Chromatin IP on HIV-1 LTR in 2D10 cells treated with or without TNFα for 2 h. The experiment has been repeated two times and the most representative results were shown. *P < 0.05 determined by two-tailed t-test with duplicated ChIP samples per treatment. The error bars were shown as standard error of mean (SEM). DHS: DNase I hypersensitive site. (B) Wildtype or CARM1 methyltransferase-dead mutants Y154R and Y154A were transfected in TZM-bl cells in the absence or presence of Tat, and the relative luciferase activity was measured two days after transfection. Western blot analysis showed the expression levels of CARM1 proteins and H3R26me2a. *P < 0.05 determined by two-tailed t-test. NS, not significant. All transfections were in quadruplicates, and the error bars were shown as SEM. (C) TZM-bl cells were transfected with Tat expressing vector and siRNA for 24 h, followed by treatment of DMSO, JQ1 (0.5 μM), or SAHA (0.5 μM) for another 24 hours before harvesting cells for luciferase assay. RT-qPCR analysis showed the knockdown efficiency of CARM1 siRNA in TZM-bl cells. *P < 0.05 determined by two-tailed t-test. The experiment was performed in duplicates. The error bars were shown as SEM.

Figure 4.

The CARM1 inhibitor 7g strongly induces HIV-1 transcription in latent cell models. (A) Chemical structure of 1-Benzyl-3,5-bis-(3-bromo-4-hydroxybenzylidene) piperidin-4-one, the CARM1 inhibitor 7g. (B) 293T cells were treated with 7g for 24 h and Western blot analysis was performed. (C) TZM-bl cells were first transfected with a Tat expressing vector, and then treated with 7g (5 μM) with or without the bromodomain inhibitor JQ1 (0.1 or 1 μM). Luciferase assay was performed after 24 h of drug treatment. *P < 0.05 determined by two-tailed t-test. The experiment was performed in duplicates, and error bars were shown as standard deviation (SD). (D) 2D10 cells were treated with increasing doses of 7g for 16 h and GFP level was determined by FACS analysis. Each treatment was duplicated and averages were shown. Error bars were shown as SD. (E) 2D10 cells were treated with different doses of 7g and SAHA for 16 h followed by FACS analysis to measure the GFP level. The percentage of activated cells and their geometric mean (GEO) were both shown. The GEO Mean was calculated specifically for GFP+ cells, and was done by gating GFP+ cells and displaying the geometric mean from that gate. (F) 2D10 cells were treated with different doses of 7g and JQ1 for 16 h followed by FACS analysis to measure the GFP level. (G) PBMCs were treated with various concentrations of 7g for 3 days, and the cell viability was determined by CellTiter-Glo luminescent assay. The experiment was performed in triplicates, and error bars were shown as SD. (H) PBMCs were treated with various concentrations of 7g for 24 h, and the levels of full-length (fl) PARP-1 and cleaved form of PARP-1 were determined by Western blot analysis. (I) CCL19-treated primary resting CD4+ T cells were infected by SF162 virus. Five days after infection, cells with latent HIV-1 were stimulated by increasing doses of 7g. The mRNA level of gag was measured by RT-qPCR to determine the activity of 7g on HIV-1 activation at the transcriptional level. The assay was performed in triplicates and SEM was shown. *P < 0.05 determined by two-tailed t-test in comparison with DMSO-treated sample. (J) Evaluation of 7g with a combinatory treatment of SAHA and JQ1 using CCL19 primary T cell model. 7g, 2 μM; SAHA, 0.5 μM; JQ1, 0.5 μM. The assay was performed in triplicates and SEM was shown. *P < 0.05.

Figure 5.

The CARM1 inhibitor 7g promotes transcriptional elongation by increasing the recruitment of SEC subunits on HIV-1 LTR. (A) Immuno-dot-blot assay to determine the specificity of the antibody. Various amounts of histone H3.1 (aa 23–42) peptides with/without methylated R26 or acetylated K27 were blotted on nitrocellulose membrane followed by western blot analysis. The H3R26me2aK27ac antibody was affinity-purified and used at 0.2 μg/ml. (B) Western blot assay to further determine the specificity of H3R26me2aK27ac antibody. Recombinant histone H3 and H3K27ac proteins were pre-methylated by CARM1 protein in vitro before being analyzed by Western blot analysis. PVDF membrane was used for protein transfer. (C) Western blot analysis to determine the protein levels of SEC subunits and histone modification enzymes in 2D10 cells treated with DMSO, 0.5 μM of SAHA, or 5 μM of 7g. Cells were treated for 24 h before harvest. (D) ChIP-qPCR analysis of histone modifications and the recruitment of transcriptional coregulators on HIV-1 LTR. The PCR primers amplified the DNA sequence on Nucleosome 1 (Nuc-1). 2D10 cells were treated with DMSO, 0.5 μM of SAHA, or 5 μM of 7g for 24 h before fixation for ChIP. The assay was performed in duplicates and SEM was shown. *P < 0.05 determined by two-tailed t-test in comparison with DMSO treated samples.

Figure 6.

Working model. Histone H3 K27 acetylation (H3K27ac) is an epigenetic mark that recruits SEC to HIV-1 LTR. K27ac also promotes R26 methylation (H3R26me2a) catalyzed by CARM1, which subsequently causes dissociation of SEC and attenuation of transcription, resulting in a negative feedback regulatory loop.