Non-coding RNA derived from the region adjacent to the human HO-1 E2 enhancer selectively regulates HO-1 gene induction by modulating Pol II binding

Recent studies have disclosed the function of enhancer RNAs (eRNAs), which are long non-coding RNAs transcribed from gene enhancer regions, in transcriptional regulation. However, it remains unclear whether eRNAs are involved in the regulation of human heme oxygenase-1 gene (HO-1) induction. Here, we report that multiple nuclear-enriched eRNAs are transcribed from the regions adjacent to two human HO-1 enhancers (i.e. the distal E2 and proximal E1 enhancers), and some of these eRNAs are induced by the oxidative stress-causing reagent diethyl maleate (DEM). We demonstrated that the expression of one forward direction (5′ to 3′) eRNA transcribed from the human HO-1 E2 enhancer region (named human HO-1enhancer RNA E2-3; hereafter called eRNA E2-3) was induced by DEM in an NRF2-dependent manner in HeLa cells. Conversely, knockdown of BACH1, a repressor of HO-1 transcription, further increased DEM-inducible eRNA E2-3 transcription as well as HO-1 expression. In addition, we showed that knockdown of eRNA E2-3 selectively down-regulated DEM-induced HO-1 expression. Furthermore, eRNA E2-3 knockdown attenuated DEM-induced Pol II binding to the promoter and E2 enhancer regions of HO-1 without affecting NRF2 recruitment to the E2 enhancer. These findings indicate that eRNAE2-3 is functional and is required for HO-1 induction.


Transcript mapping of human HO-1 enhancer RNAs
To detect transcripts derived from HO-1 enhancer regions, we performed transcript mapping. Briefly, cDNA was synthesized using total RNA from DEM-treated HeLa cells as a template with random hexamers, and RT-PCR was performed with primer sets listed in Supplementary Table S1. In RT-PCR analysis, following samples were used as controls: HeLa genomic DNA was used as a positive control of PCR amplification; a reaction without reverse transcriptase was used as a negative control to rule out genome DNA contamination. RT-PCR products were electrophoresed on TAE-agarose gel, visualized by ethidium bromide staining, cloned using a Zero Blunt TOPO PCR Cloning kit (Life Technologies) and the cloned DNA was subjected to the DNA sequence analysis.

Determination of direction of hHO-1 eRNAs
To determine the direction of transcripts around the HO-1 enhancer regions, we performed a strand-specific RT reaction using internal primers followed by cDNA amplification using the primer sets. Briefly, cDNA was synthesized using total RNA from DEM-treated HeLa cells as a template with region specific forward or reverse primers, and subsequent PCR was performed with primer sets listed in Supplementary Table S1.
Amplified RT-PCR products were electrophoresed on TAE-agarose gel, visualized by ethidium bromide staining, cloned using a Zero Blunt TOPO PCR Cloning kit (Life Technologies) and the cloned DNA was subjected to the DNA sequence analysis.

5' Rapid Amplification of cDNA Ends (5' RACE)
5' RACE analysis was carried out using a Gene Racer Kit (Life Technologies) according to the manufacturer's protocol. Briefly, total RNA from DEM-treated HeLa cells was isolated, dephosphorylated and 5' Cap structure was removed. Then, the GeneRacer RNA oligo was ligated to de-capped 5' end of RNA and reverse transcription was done with random hexamers. To amplify 5' RACE product, RT-PCR was performed using the GeneRacer 5' primer and reverse region specific primers listed in Supplementary Table S1. Subsequently, nested PCR was performed using the GeneRacer 5' Nested primer and reverse region specific primers. Nested 5' RACE PCR products were electrophoresed on TAE-agarose gel, visualized by ethidium bromide staining and sub-cloned using a Zero Blunt TOPO PCR Cloning kit. The DNA sequence of the inserts in isolated clones was validated by DNA sequencing. To determine 5' ends, we analyzed DNA sequences of the junction where the GeneRacer RNA oligo was ligated.

Experiments
Primer names 5'-3' sequences Purposes Tanscription mapping, RT-PCR  map_E2_fwd_1  CGG ATG TGG TAT AAT TAC AGC TGT  transcript mapping E2-1, strand specific cDNA synthesis and PCR  map_E2_rev_1  CTA GTA GTT GAT ACT CAC CGG GT  transcript mapping E2-1, strand specific cDNA synthesis and PCR  map_E2_fwd_2 CAG TTC TCA CTT CTG CTC ACT TCT transcript mapping E2-2, strand specific cDNA synthesis map_E2_rev_2 GAT TAA ACC TGG AGC AGC TGG AAC TCT transcript mapping E2-2, strand specific cDNA synthesis E2 Forward_1 GTC TGG GGC CTG AAT CCT A transcript mapping E2-3, strand specific cDNA synthesis, E2-2, E2-3 overlap detaction, UPL_realtime PCR for eRNA E2-3 with universal probe library probe #17 5' RACE E2 Reverse_1  TGC ATT GCA TTC ACT TGT CCT GCC TCT  transcript mapping E2-3, strand specific   We performed RT-PCR using the primer sets indicated in (A) and (C), and the following templates: g: genomic DNA of HeLa cells; 1: RT reaction with random hexamers and DEM-untreated HeLa total RNA; 2: RT reaction with random hexamers and DEM-treated HeLa total RNA; 3: RTase minus reaction with random hexamers and DEM-untreated HeLa total RNA; 4: RTase minus reaction with random hexamers and DEM-treated HeLa total RNA. Amplified RT-PCR products (asterisks) were subcloned using a ZeroBlunt TOPO PCR cloning Kit and the DNA sequence was analyzed. Note that the E2-4 PCR product was detected using genomic DNA as a template, but not using the RT reaction performed with random hexamer and DEM-treated HeLa total RNA (data not shown). On the other hand, E1-2 PCR product was not amplified even when using genomic DNA as a template (data not shown). To determine the direction of transcripts that were detected around the HO-1 enhancers, strand-specific RT-PCR was performed using strand-specific gene internal primers and total RNA from DEM-treated HeLa cells. We performed RT-PCR using the following templates: g: genomic DNA of HeLa cells; 1: RTase minus reaction with random hexamers and DEM-treated HeLa total RNA; 2: RT reaction with random hexamers and DEM-treated HeLa total RNA; 3: RT reaction with a 3'end-directed cDNA synthesis primer and DEM-treated HeLa total RNA ; 4: RT reaction with a 5'end-directed cDNA synthesis primer and DEM-treated HeLa total RNA . Amplified RT-PCR products (asterisks) were subcloned using a ZeroBlunt TOPO PCR cloning Kit and the DNA sequence was analyzed.  We performed RT-PCR using the following templates: g: genomic DNA of HeLa cells; 1: RT reaction with random hexamers and DEM-untreated HeLa total RNA; 2: RT reaction with random hexamers and DEM-treated HeLa total RNA; 3: RTase minus with random hexamers and DEM-untreated HeLa total RNA; 4 : RTase minus reaction with random hexamers and DEM-treated HeLa total RNA. Amplified PT-PCR products (asterisks) were subcloned using a ZeroBlunt TOPO PCR cloning Kit and the DNA sequences were analyzed. Map E2_fwd_2(Nest) and eRNA E2_UPL_Reverse primers were used to amplify the overlapping region between E2-2 and E2-3, and hHO-1 E1_Forward and 5' RACE E1 Reverse_1 primers were used to amplify the overlapping region between E1-3 and E1-4.

Maruyama et al.
Template:    Figure S6 3' primer walking analysis of hHO-1 eRNA E2-3. To estimate the full-length of eRNA E2-3, cDNAs synthesized using random hexamers and total RNA of DEMtreated HeLa cells were examined by 3' primer walking analysis. (A) The location of the primers used in 3' primer walking analysis are shown. (A) A downward E2 Forward_1 primer (red arrow) and a series of upward primers ((i) to (viii)) were used. The primer sequences are listed in Supplementary Table S1. The sizes of the PCR products are indicated to the right of the primer name. (B) Ethidium bromide staining of PCR products in 3' primer walking analysis. The PCR primers used for analysis are indicated in the figure. We performed RT-PCR using the following templates: 1: genomic DNA of HeLa cells; 2: RTase minus with random hexamer and DEM-treated HeLa total RNA; 3: RT reaction with random hexamers and DEM-treated HeLa total RNA. Amplified RT-PCR products (asterisks) were subcloned using a ZeroBlunt TOPO PCR cloning Kit and the DNA sequence was analyzed. A PCR fragment of (v)-lane 3 contained hHO-1 eRNA E2_L. We also confirmed by DNA sequencing that the PCR fragment of (i)-lane 3 contained DNA sequence corresponding to eRNA E2-3, whereas the PCR fragment of (ii)-lane 3 was a non-specific PCR amplification.  We performed a 3' primer walking analysis as shown in Figure S6 and the DNA sequences of the plasmid clones were determined. (A) A schematic representation of the putative transcript (hHO-1 eRNA E2_L) is shown by bold black arrows. (B) The sequence obtained from the subclone of PCR using E2 Forward_1 and (v) E2 Reverse_5 is underlined against the HO-1 E2 enhancer region sequences. Putative intron is indicated by hatching. We identified CAGC sequences at the both sides of the intron/exon junction of genomic DNA (shown in red). The hHO-1 eRNA E2_L contains only one of these. The DNA sequence that encode the putative ORF is indicated by green letters. The 5' end of hHO-1 eRNA E2-3 is indicated by a magenta letter. The primers used in this experiment are shown by arrows below the sequence.   We analyzed the expression of hHO-1 eRNAs in HaCaT and SH-SY5Y cells. HaCaT (A) and SH-SY5Y (B) cells were either untreated (-) or treated with 100 mM DEM for 6 hours (+) and total RNA samples were isolated. The cDNAs were synthesized using random hexamers with total RNA as a template. The arbitrary RNA levels of eRNA E2-1, eRNA E2-3 and eRNA E1-4 were measured by real-time RT-PCR using specific primers and Universal Probe Library Probes. The value was normalized to the expression of the cyclophilin A gene, and the arbitrary RNA level was expressed as the mean ± SEM of three independent assays. *: P<0.05; **: P<0.01 (two-tailed unpaired Student's t-test). HeLa cells were transfected with control siRNA (Ctrl) or two siRNAs against eRNA E2-2 (E2-2(1) or E2-2(2)). Then the cells were either untreated (-) or treated with100 mM DEM for 6 hours (+). The RNA level was analyzed by real-time RT-PCR using specific primers and Universal Probe Library Probes. The effect of eRNA E2-2-KD on the expression of HO-1 (D) and TXNRD1 (E). The expression of HO-1 and TXNRD1 was analyzed by real-time RT-PCR. Each value was normalized to cyclophilin A gene expression and the arbitrary RNA level was expressed as the mean ± SEM of three independent assays. *: P<0.05; **: P<0.01 (twotailed unpaired Student's t-test) compared to control siRNA without DEM (Ctrl (-)). #: P<0.05, ##: P<0.01; NS: no significance compared to the value of control siRNA with 100 mM DEM for 6 hours (Ctrl (+)) (one-way ANOVA followed by a Dunnett's post-hoc test for multiple parameter comparisons).   (2)). The cells were untreated (-) or treated with100 mM DEM for 6 hours (+). The eRNA E2-3 level was analyzed by real-time RT-PCR using specific primers and Universal Probe Library Probe #17. (B) The effect of eRNA E2-3-KD on the expression of HO-1. The expression of HO-1 was analyzed by real-time RT-PCR. Each value was normalized to cyclophilin A gene expression, and the arbitrary RNA level was expressed as the mean ± SEM of four independent assays. *: P<0.05; **: P<0.01 (two-tailed unpaired Student's t-test) compared to control siRNA without DEM (Ctrl (-)). #: P<0.05; ##: P<0.01; NS: not significant compared to the value of control siRNA with 100 mM DEM for 6 hours (Ctrl (+)) (one-way ANOVA followed by a Dunnett's post-hoc test for multiple parameter comparisons). (C) The effect of eRNA E2-3-KD on HO-1 protein expression in HaCaT cells. Wholecell lysates were separated by SDS-PAGE, and HO-1 protein expression was analyzed by immunoblotting using a HO-1-specific antibody (Abcam, ab68477). Actin was used as a loading control (Sigma-Aldrich, A1978).