Enhancer-mediated enrichment of interacting JMJD3-DDX21 to ENPP2 locus prevents R-loop formation and promotes transcription

Abstract

ENPP2, which encodes for the enzyme autotaxin (ATX), is overexpressed during chronic inflammatory diseases and various cancers. However, the molecular mechanism involved in the ENPP2 transcription remains elusive. Here, in HEK 293T cells, we demonstrated that lipopolysaccharide (LPS) increased the transcription process at ENPP2 locus through a NF-кB pathway and a reduction of H3K27me3 level, a histone repressive mark, by the demethylase UTX. Simultaneously, the H3K27me3 demethylase JMJD3/KDM6B was recruited to the transcription start site (TSS), within the gene body and controlled the expression of ENPP2 in a non-enzymatic manner. Mass spectrometry data revealed a novel interaction for JMJD3 with DDX21, a RNA helicase that unwinds R-loops created by nascent transcript and DNA template. Upon LPS treatment, JMJD3 is necessary for DDX21 recruitment at ENPP2 locus allowing the resolution of aberrant R-loops. CRISPR-Cas9-mediated deletion of a distant-acting enhancer decreased the expression of ENPP2 and lowered the recruitment of JMJD3-DDX21 complex at TSS and its progression through the gene body. Taken together, these findings revealed that enhancer-mediated enrichment of novel JMJD3-DDX21 interaction at ENPP2 locus is necessary for nascent transcript synthesis via the resolution of aberrant R-loops formation in response to inflammatory stimulus.

Figure 1.

NF-κB response elements mediate the expression of ENPP2. (A) Luciferase assay demonstrating increased transcriptional activity at the ENPP2 promoter following LPS treatment for the indicated time. (B and C) Endogenous ENPP2 expression in HEK 293T cells treated with LPS at the indicated time points (B) qPCR analysis (C) representative western blots and quantifications. (D) Effect of a 24 h treatment with BAY11-7085 on LPS-mediated ENPP2 expression, qPCR analysis. (E) Effect of p65 overexpression on ENPP2 promoter activity as shown by luciferase reporter assay. (F) Expression of endogenous ENPP2 following p65 overexpression in HEK 293T, qPCR analysis. (G) Scheme of the two кB consensus sites localization in the ENPP2 promoter. (H) ChIP-qPCR measurement of p65 enrichment at the ENPP2 locus in HEK 293T cells treated with LPS for 6 h. A set of primer targeting an intergenic region downstream of SNRPN was used a negative control (NEG). (I) Scheme of the ENPP2 luciferase reporter. (J) Luciferase reporter assay showing the effect of the кB sites deletion on ENPP2 promoter activity. LPS 100ng/ml, BAY11-7085 1 μM.

Figure 2.

ENPP2 promoter has bivalent histone marks. (A) ChIP-qPCR measurement of H3K27me3 enrichment at the ENPP2 locus and a negative locus (NEG; downstream of SNRPN) in HEK 293T cells following 6 h of LPS exposure (B and C) ChIP-qPCR assays showing (B) enrichment of the H3K4me3 histone mark and (C) RNA pol II recruitment at the ENPP2 TSS. (D–F) Effect of siRNA against (D) EZH2, (E) UTX and (F) JMJD3 on LPS mediated expression of ENPP2, qPCR analyses. (G and H) H3K27me3 enrichment at the ENPP2 promoter, determined by ChIP-qPCR, in response to siRNA targeting (G) UTX and (H) JMJD3. (I) qPCR measurements of ENPP2 expression in response to JMJD3 siRNA transfection followed by rescue with constructs encoding for JMJD3 wild-type or JMJD3 H1390A (catalytic dead mutant). LPS 100 ng/ml.

Figure 3.

JMJD3 and DDX21 are recruited to the promoter and gene body of ENPP2. (A) JMJD3 enrichment on promoter and gene body of ENPP2 in HEK 293T cells treated with LPS during 6 h compared to a negative locus (NEG; downstream of SNRPN), ChIP-qPCR analysis. (B) GO analysis of JMJD3 interactome identified by mass spectrometry analysis (LCMS/MS). (C) Table of the first ten JMJD3-associated proteins that were enriched after LPS stimulation, as determined by mass spectrometry analysis. (D) DDX21 and RNA pol II co-immunoprecipitated with HA-JMJD3 in HEK 293T cell extracts. (E) ChIP-qPCR experiment demonstrating DDX21 enrichment at ENPP2 promoter and gene body following 6 h of LPS treatment. Two set of primers targeting an intergenic region downstream of SNRPN (NEG; downstream of SNRPN) and promoter region of DDX23 (DDX23) were used as negative and positive control, respectively. (F) qPCR quantification of ENPP2 expression in response to DDX21 siRNA transfection and LPS exposure in HEK 293T cells. LPS 100 ng/ml.

Figure 4.

The complex formed by JMJD3 and DDX21 is required for R-loops resolution at ENPP2 locus. (A) ChIP-qPCR indicating increased R-loops formation at ENPP2 TSS, gene body and a negative locus (NEG; downstream of SNRPN) following LPS treatment (6 h) in HEK 293T cells; treatment of lysates with RNase H before immunoprecipitation abolished R-loops. (B) Effect of DDX21 siRNAs on R-loop formation, as measured by ChIP-qPCR, in presence of LPS. (C) JMJD3 siRNAs treatment hinders on DDX21 recruitment at ENPP2 TSS and gene body in presence of LPS, ChIP-qPCR analysis. (D) ChIP-qPCR showing the effect of JMJD3 siRNA on R-loop formation at ENPP2 TSS and gene body in the presence of LPS. (E) Scheme of the ChIP-ReChIP main steps. (F) ChIP-ReChIP-qPCR analysis indicating simultaneous binding of DDX21 and HA-JMJD3 at ENPP2 TSS, enhanced upon addition of LPS. LPS 100 ng/ml, RNaseH 60U/ml.

Figure 5.

R-loops resolution at ENPP2 locus promotes the synthesis of nascent transcript. (A and B) Endogenous ENPP2 expression after transfection of siDDX21 and rescue with a vector encoding (A) RNaseH1 or in (B) DDX21 WT or a DDX21 S375L/A376E (catalytic dead mutant), qPCR analyses. (C) Schematic representation of the experimental protocol for transcriptional run-on assay using the Click-it technology. HEK 293T cells were pulse labeled with 5-ethynyl uridine (EU) for 6 h, and the EU-incorporated RNAs were biotinylated and captured with streptavidin conjugated beads, followed by RT-PCR. (D–F) Transcriptional run-on assay indicates that depletion of DDX21 alters ENPP2 nascent transcript formation. LPS 100 ng/ml, EU 0.2 mM.

Figure 6.

A distant-acting enhancer contacts the promoter region of ENPP2 and controls its expression. (A) Circos plot showing interactions between ENPP2 TSS and a putative enhancer located at 82 kb, dataset from CD34+ cells. (B) 3C experiments replicating this interaction in HEK 293T cells. LPS treatment did not modulate this interaction. (C) H3K4me1 enrichment at the enhancer region, as measured by ChIP-qPCR, using six sets of primers covering the putative enhancer region. (D) Luciferase reporter assay demonstrating that the region has enhancer activity, independently of its orientation. (E) Scheme depicting the deletion of the enhancer region by using the CRISPR-Cas9 technology. Cas9 enzyme was transfected simultaneously with two gRNA targeting 5′ end and 3′ end of the enhancer region. (F) Endogenous expression of ENPP2 after CRISPR-Cas9-mediated deletion of the enhancer, as measured by q-RT-PCR. LPS 100 ng/ml.

Figure 7.

Enhancer-mediated enrichment of JMJD3–DDX21 at ENPP2. (A) Schematic of the ChIP-reChIP protocol. (B) Binding of JMJD3–DDX21 complex on ENPP2 enhancer region, as measured in ChIP-reChIP experiments and (C) after deletion of enhancer region by CRISPR-Cas9 technology in presence of LPS. LPS 100 ng/ml.

Figure 8.

Working model of the role of JMJD3 in the regulation of ENPP2 transcription. LPS increases transcription at the ENPP2 locus through a NF-кB pathway coinciding with a reduction of the H3K27me3 histone mark, which is promoted by the recruitment of UTX. Concomitantly, enhancer-mediated enrichment of JMJD3–DDX21 complex at the TSS stimulates transcription elongation through R-loops resolution.