Inositol pyrophosphates regulate JMJD2C-dependent histone demethylation

Abstract

Epigenetic modifications of chromatin represent a fundamental mechanism by which eukaryotic cells adapt their transcriptional response to developmental and environmental cues. Although an increasing number of molecules have been linked to epigenetic changes, the intracellular pathways that lead to their activation/repression have just begun to be characterized. Here, we demonstrate that inositol hexakisphosphate kinase 1 (IP6K1), the enzyme responsible for the synthesis of the high-energy inositol pyrophosphates (IP7), is associated with chromatin and interacts with Jumonji domain containing 2C (JMJD2C), a recently identified histone lysine demethylase. Reducing IP6K1 levels by RNAi or using mouse embryonic fibroblasts derived from ip6k1(-/-) knockout mice results in a decreased IP7 concentration that epigenetically translates to reduced levels of trimethyl-histone H3 lysine 9 (H3K9me3) and increased levels of acetyl-H3K9. Conversely, expression of IP6K1 induces JMJD2C dissociation from chromatin and increases H3K9me3 levels, which depend on IP6K1 catalytic activity. Importantly, these effects lead to changes in JMJD2C-target gene transcription. Our findings demonstrate that inositol pyrophosphate signaling influences nuclear functions by regulating histone modifications.

Keywords: inositides; metabolism; phosphorylation.

Fig. 1.

IP₆K1 and JMJD2C are binding partners in vitro and in vivo. (A) IP₆K1 (Diff) interacts with JMJD2C (847–1,056) by yeast two-hybrid. Yeast reporter strain AH109 was cotransformed either with empty pGBKT7, pGBKT7-Diff1, or pGBKT7-Diff2 and pACT2-JMJD2C (847–1,056). Serial dilution of yeast grown on selective media to assess interaction strength (Upper) or permissive media (Lower) and grown at 30 °C for 3 d (n = 3). (B) Immunoblot analysis of HEK293T cells transfected with JMJD2C-Flag and either GST-IP₆K1, GST-Diff, or GST empty vector control and subjected to GST pull-down assays. The Ponceau staining confirmed the pull-down of the GST-tagged proteins (n = 3). (C) Immunoblot analysis of HEK293T cells transfected with JMJD2C-Flag and subjected to coimmunoprecipitation with anti-IP₆K1 antibody. The blot was probed with anti-FLAG antibody (n = 3). (D) MEF cell extracts from either wild-type (WT) or IP₆K1 null (ip6k1^−/−) mice were subjected to coimmunoprecipitaton with Sepharose-conjugated anti-IP₆K1 antibody and probed with either anti-JMJD2C or anti-IP₆K1 antibodies (n = 3). (E) Recombinant GST-JMJD2C (847–1,056) or GST were subjected to in vitro binding assays either with recombinant His-Diff1 or His-IP₆K1 in the presence of 10 μM IP₆ or IP₇. Pull-downs and 5% of the inputs were immunoblotted with anti-His antibody and Ponceau staining demonstrates pull-down of the GST-tagged proteins (n = 3).

Fig. 2.

Nuclear localization and chromatin association of IP₆K1. (A) IP₆K1 associates with chromatin. HEK293T nuclear pellets were subjected to fractionation into soluble (nucleoplasm) and chromatin fractions. Immunoblotting was subsequently performed with antibodies to actin and histone H3 as quality controls and anti-IP₆K1. (B) IP₆K1 associates with histone H3. MEF cells from wild-type (WT) and IP₆K1 knockout (ip6k1^−/−) mice were subjected to coimmunoprecipitaton with anti-IP₆K1 antibody after DNase treatment of lysates. Pull-downs and 5% inputs were separated by SDS/PAGE and membranes were subsequently blotted with anti-histone H3 antibody. (C) IP₆K1 associates with chromatin bearing the H3K9me3 mark. Coimmunoprecipitation with anti-IP₆K1 was performed on DNase-treated HEK293T lysates. Inputs and immunoprecipitated samples were separated by SDS/PAGE and blotted with anti-H3K9me3 antibody. (D) IP₆K1 associates with histone H3. In vitro binding assays were performed with recombinant GST-IP₆K1 and purified H3.1 protein.

Fig. 3.

IP₆K1 catalytic activity regulates epigenetic modifications of histone H3 at the K9/S10 hot-spot. (A) IP₆K1 knockout MEFs (ip6k1^−/−) display reduced levels of H3K9me3 and increased levels of both acetyl-H3K9 and phospho-H3S10, which can be rescued by expression of IP₆K1. In the case of the acetyl-H3K9, a dividing line is shown to indicate that nonessential lanes were removed from the single original blot. Immunoblots of cell extracts obtained either from ip6k1^−/− or WT MEFs and probed with the corresponding antibodies. The experiments were performed at least three times with two independent lines of WT and ip6k1^−/− MEFs. (B) Immunohistochemistry of ip6k1^−/− MEFs confirms reduced levels of H3K9me3. MEFs obtained from ip6k1^−/− mice were labeled with cell tracker (green), mixed with WT MEFs, and immunostained with anti-H3K9me3 antibody (red). Images were taken on a Leica TCS SPE confocal microscope and quantified using LAS AF software (Leica). Left graph represents the mean ±SE of the average H3K9me3 fluorescence intensity of 97 KO and 68 WT cells (n = 3). Right graph represents the mean ±SE of the number of H3K9me3 foci per nuclei of 77 KO and 74 WT cells. Statistical analysis was performed using the Mann–Whitney u test. (Scale bar, 30 μm.) (n = 3). (C) Immunoblot analysis of H3K36me3 shows no difference between WT and ip6k1^−/− MEFs (n = 4). (D) Immunoblot of HEK293T cells expressing GST-IP₆K1 and analyzed for H3K36me3 mark (n = 3). (E) Immunoblot analysis of HEK293T cells expressing either GST-IP₆K1 or the catalytically inactive mutant GST-IP₆K1 (DF/AA) and analyzed for H3K9me3 mark (n = 4). (F) H3K9me3 immunoblot of HEK293T cells depleted of IP₆K1 for 48 h by using two short hairpin RNAs (shRNA-A and shRNA-B) (n = 3).

Fig. 4.

Inositol pyrophosphates regulate histone modifications through JMJD2C chromatin association. (A) JMJD2C protein levels were reduced in human HEK293T cells with two different specific siRNAs for 96 h. Mus musculus (Mm) JMJD2C was expressed 16 h before cells were lysated and immunoblotting performed. (B) The effect of IP₆K1 kinase activity on H3K9me3 levels depends on JMJD2C. Cells were transfected with JMJD2C siRNAs as above and subsequently transfected either with GST-IP₆K1 or GST-IP₆K1 DF/AA, 16 h before analysis. Total proteins were extracted, separated by SDS/PAGE, and blots were probed with the indicated antibodies. Experiments were performed two times in duplicate. The increase in H3K9me3 levels relative to total histone H3 observed between lanes 3 and 4 over lanes 1 and 2 is no longer apparent in the context of JMJD2C knockdown, now comparing lanes 7 and 8 to lanes 5 and 6. (C) Quantitative analysis of histone H3K9me3 levels normalized to total histone H3. Average ± SEM of the four experiments is shown. The Mann–Whitney u test was used to determine statistical significance. (D) IP6K1 induces dissociation of JMJD2C from chromatin. HEK293T cells were transfected with either empty GST or GST-IP₆K1 and JMJD2C-Flag, and nuclear extracts were separated into chromatin or nucleoplasmic fractions. Western blot analyses were performed with anti-Flag, anti-GST, anti-histone H3, and anti-actin antibodies. (E) Quantitative analysis of JMJD2C levels normalized to actin or histone H3 for nucleoplasm or chromatin, respectively. NS, nonsignificant. Average ± SD of two experiments both run in duplicate is shown. The Mann–Whitney u test was used to determine statistical significance.

Fig. 5.

Inositol pyrophosphates control the expression of JMJD2C-regulated genes. (A) Chromatin immunoprecipitation analysis of MEFs obtained either from WT or ip6k^−/− mice. IP₆K1, H3K9me3, JMJD2C, H3S10ph, and H3K9ac immunoprecipitation followed by qPCR of either P1 or P2 promoters of the JMJD2C-regulated gene Mdm2. PI, preimmune serum. Data are represented as percentage of total input. Shown are the averages ±SEM. Parried Student t test was applied to calculate the statistical significance of the value of the ip6k1^−/− against WT (n = 3; *P < 0.01, **P < 0.001). (B) IP₆K1 regulates levels of Mdm2 expression from exon 2 but not exon 1. qRT-PCR of mRNA extracted from either WT or ip6k^−/− MEF. Mdm2 exon 1 and exon 2 mRNA was normalized to 18S ribosomal RNA (n = 3; *P < 0.01). (C and D) IP₆K1 regulates expression of JMJD2C-dependent genes in embryonic stem (ES) cells. (C) Specific siRNA to IP₆K1 or scrambled siRNA were transfected in embryonic stem cells in two rounds of transfection. RNA was extracted 24 h after the second transfection and analyzed by qRT-PCR (n = 3; **P < 0.001). (D) Myc-IP₆K1 or empty vector controls were transfected into ES cells and after 48 h RNA was extracted and subjected to qRT-PCR. Levels of Nanog, Sox2, and Cdx2 cDNA were normalized to actin and GAPDH and expressed as fold induction over control samples (n = 3; **P < 0.001).