HIF-1-dependent induction of Jumonji domain-containing protein (JMJD) 3 under hypoxic conditions

Abstract

Jumonji domain-containing proteins (JMJD) catalyze the oxidative demethylation of a methylated lysine residue of histones by using O2, α-ketoglutarate, vitamin C, and Fe(II). Several JMJDs are induced by hypoxic stress to compensate their presumed reduction in catalytic activity under hypoxia. In this study, we showed that an H3K27me3 specific histone demethylase, JMJD3 was induced by hypoxia-inducible factor (HIF)-1α/β under hypoxia and that treatment with Clioquinol, a HIF-1α activator, increased JMJD3 expression even under normoxia. Chromatin immunoprecipitation (ChIP) analyses showed that both HIF-1α and its dimerization partner HIF-1β/Arnt occupied the first intron region of the mouse JMJD3 gene, whereas the HIF-1α/β heterodimer bound to the upstream region of the human JMJD3, indicating that human and mouse JMJD3 have hypoxia-responsive regulatory regions in different locations. This study shows that both mouse and human JMJD3 are induced by HIF-1.

Fig. 1.

HIF-1α-dependent hypoxic induction of the mouse JMJD3 gene. (A–C) NIH-3T3 cells, and MEF were incubated under hypoxia (< 0.5% O₂) for 48 h. 3T3-L1 cells were incubated under hypoxia (< 0.5% O₂) for 36 h. The relative mRNA levels of BNIP3 and JMJD3 were analyzed by qRTPCR and normalized by 18S ribosomal RNA. The mean value of qRT-PCR from normoxic cells was set to 1. The data represent the average and standard deviations of triplicate measurements from qPCR. (D, E) HIF-1α knockdown 3T3-L1 cells and control 3T3-L1 cells were treated with hypoxia for 16 h; (D) qRT-PCR was performed using the primer sets for the indicated mouse genes; (E) Western blot analyses were performed by using the indicated antibodies and lysates from hypoxia treated mouse 3T3-L1 cells. HSP70 and β-Actin protein were used as the loading control. The results of two independent experiments were shown. The numbers indicated the relative intensity of each band which was measured using LAS-3000 luminescent image analyzer and Multi Gauge ver.3.0 imaging software. (F, G) 3T3-L1 cells were treated with DFN (100 μM), CoCl₂ (200 μM), or hypoxia for 48 h; (F) Western blot analyses were performed using the indicated antibodies; (G) The RT-PCR analyses of JMJD3 and BNIP3 are shown. The 18S ribosomal RNA was used as the loading control. N, normoxia; H, hypoxia (< 0.5% O₂). The p-values were obtained using the Student’s t-test and the significance between the groups is indicated (^**p < 0.05, ^*p < 0.1).

Fig. 2.

Arnt-dependent hypoxic induction of the mouse JMJD3 gene. (A, B) Hepa1c1c7 cells and BpRc1 cells were incubated under hypoxia (< 0.5% O₂) for 16 h; (A) Western blot analyses were performed by using the indicated antibodies. The 14-3-3γ protein was used as the loading control; (B) The relative mRNA levels of mouse JMJD3 were analyzed by qRT-PCR. (C) The putative HREs in the mouse JMJD3 gene are identified by using the MatInspector program (www.genomatix.de) (Cartharius, Frech et al., 2005). The bars indicate the positions of the primer sets used for the ChIP analyses. The nucleotide sequences of putative HRE sites (−0.2 Kb, +3.9 Kb, +4.2 Kb) and flanking regions were shown (lower panel). (D) 3T3-L1 cells were incubated under hypoxia (< 0.5% O₂) for 16 h. ChIP analyses were performed with the indicated antibodies and primer sets of mouse JMJD3 and mouse Stra13 genes. Input values were obtained from samples treated in the same way as the experimental method, except that no immunoprecipitation steps were performed. The values on the y-axis are presented as the percentage input of the average and standard deviation of triplicate determinations of qPCR. (E) MEF were incubated under < 0.5% O₂, 3% O₂, and 21% O₂ (normoxia) for 48 h. ChIP analyses (upper panel) and Western blot analyses (lower panel) were performed using the indicated antibodies. N, normoxia; H, hypoxia (< 0.5% O₂). The p-values were obtained using the Student’s t-test and the significance between the groups is indicated (^**p < 0.05, ^*p < 0.1).

Fig. 3.

Hypoxic induction of human JMJD3 in hADSC. (A, B) hADSC were treated with hypoxia (<0.5% O₂, 48 h) or Clioquinol (50 μM, 16 h); (A) The relative mRNA levels of BNIP3 and JMJD3 were analyzed by qRT-PCR. The data represent the average and standard deviations of two independent experiments; (B) Western blot analyses were performed using the indicated antibodies. The results of two independent experiments were shown. The relative intensity of each band was measured as described in Fig. 1E. (C) The putative HREs in the human JMJD3 gene are identified as described in Fig. 2C. The bars indicate the positions of the primer sets used for the ChIP analyses. The nucleotide sequences of putative HRE sites (−4.1 kb, −4.0 kb, −3.3 kb) and flanking regions were shown (lower panel). The p-values were obtained using the Student’s t-test and the significance between the groups is indicated (^**p < 0.05).

Fig. 4.

ChIP analyses of the human JMJD3 gene. hADSC were treated with hypoxia (< 0.5% O₂, 48 h) or Clioquinol (50 μM, 16 h). (A–F) ChIP analyses were performed using the indicated antibodies and primer sets for the human JMJD3 gene; (A, C, and E) ChIP-eluted DNA was analyzed by PCR and visualized on 2% agarose gel electrophoresis. (B, D, and F) ChIP-eluted DNA was analyzed by qPCR. The values on the y-axis are presented as the percentage input of the average and standard deviation of triplicate determinations of qPCR. The primer set for the human BNIP3. promoter was used as the positive control. The p-values were obtained using the Student’s t-test and the significance between the groups is indicated (^**p < 0.05, ^*p < 0.1).