The human EDN1 chromatin state was analyzed using the UCSC Genome Browser¹ (Karolchik et al., 2014). The EDN1-AS predicted promoter was identified using the Genome Segments and Broad Chromatin HMM tracks with HUVEC cell information selected. The Transcription Factor ChIP track and DNase Clusters track was also used to analyze transcriptional regulation of the predicted promoter site.

HMEC cells were cultured in MEGM^TM Mammary Epithelial Cell Growth Medium with BulletKit^TM (Lonza) and 10% charcoal stripped FBS. S9 cells were cultured in F12 Ham Kaighn’s modification (F12K) supplemented with 25 mM NaHCO₃, 4 mM glutamine, 1% Penicillin/Streptomycin and 10% FBS. HK-2 cells were cultured in DMEM/Hamm’s F12 media supplemented with 10% FBS and 1% Penicillin/Streptomycin. HEK293 cells were cultured in DMEM containing 4.5 g/L glucose supplemented with 1% Penicillin/Streptomycin and 10% FBS. All cells were grown in a 37°C incubator, humidified at 5% CO₂.

RNA was isolated from cells using TRIzol (Ambion) per manufacturer instructions. In general, 1 ml TRIzol was used per well in a 6-well plate. Total RNA was treated with DNase (Ambion) per manufacturer instructions to remove genomic DNA. RNA from an adult human female kidney was purchased from Life Technologies.

Human strand-specific EDN1-AS primers (SS1-6; Table 1) were designed to lay down at locations progressively closer to the 5′ end of the EDN1 gene for use in reverse transcriptase reactions. Reverse Transcriptase (RT) from Thermo Fisher was used as per manufacturer instructions. Oligo-dT primers were used as a positive control for any poly-adenylated tailed mRNA. All samples were used in –RT and + RT reactions. Primers have a complementary sequence to the EDN1 sense strand so they will only anneal to antisense RNA. PCR primers (PCR1 and PCR2) were designed to amplify the same region of cDNA regardless of the strand-specific primer used.

HK-2 cells were grown to ∼80% confluence in 6-well cell culture plates. To synchronize the circadian clock in all cells, 100 nM Dexamethasone was added for 30 min and then media was changed. After 24 h, cells were trypsinized and RNA was isolated every 2 h for 24 h.

Guide RNA (gRNA) design, cloning and cell engineering was done as previously outlined (Dükel et al., 2016). Briefly, gRNAs were designed using the online tool CC Top (² Stemmer et al., 2015) using a ∼150 bp sequence of from the predicted human EDN1-AS promoter as input. The region was selected based on identified transcription factor binding sites hypothesized to be important to transcription regulation.

gRNA sequences were synthesized as complementary oligonucleotides and included overhangs for insertion into BsmB1-digested LentiCRISPRv2 plasmid (Addgene catalog # 52961). Oligonucleotides were 5′ phosphorylated and annealed as outlined by the Zhang lab³ and ligated to linearized LentiCRISPR v2 with Quick Ligase (NEB). Ligation reactions were transformed into Stbl3 bacteria following manufacturer instructions. Using this protocol, four (4) recombinant LentiCRISPR v2 plasmids (designated gRNA 1–4) were cloned and verified by automated Sanger sequencing.

HEK293 cells were co-transfected with the four LentiCRISPR + gRNA constructs (sgRNA 1, 2, 3, and 4) along with psPAX2 (Addgene #11260) and pMD2.G (Addgene #11259) for generation of pseudotyped lentiviral particles. Media containing virus particles was collected 72 h post-transfection and cleared viral supernatant was then added to HK-2 cells singly and pairwise (e.g., gRNA 1 + 2, 1 + 3, 1 + 4 etc.). 2 μg/ml puromycin was used for 2 weeks to select for cells transduced with virus.

After puromycin selection, cells were plated into the first of 4 columns of a 96 well plate in 200 μl of media. Cells were then serial diluted into the following columns. After growth time, wells were inspected for the presence of one to three colonies of cells. These wells were then trypsinized with 1 drop of 0.25% Trypsin per well and plated into 10 cm dishes. Cells were allowed to grow undisturbed for 1 week until colony formation. Cloning rings were dipped in sterile vacuum grease to create a seal and pressed down over the colony. Media was aspirated and 1 drop of Trypsin was added. After ∼5 min, trypsinized cells were removed and plated in 6-well plate and allowed to grow. Genomic DNA was isolated from these cells and used for genotyping PCR.

Gel-purified PCR products from PCR reactions using genotyping primers GT1 and GT2 were cloned into the TA TOPO vector according to the manufacturer’s instructions (Thermo Fisher). Cells with insert were identified by blue/white screening. Plasmids containing inserts were sent to GENEWIZ for sequencing using M13F and M13R primers. The resulting sequences were analyzed with Serial Cloner and sequences from NCBI and UCSC Genome Browser.

HK-2 cells were plated into 6-well Transwell plates (Corning) and grown to confluence. Media from HK-2 cell lines was spun down to remove cell debris and then frozen at −80°C until all samples were ready to use. ELISA for ET-1 was performed using the Human Endothelin-1 QuantiGlo ELISA Kit (R&D Systems) according to manufacturer instructions.

HK-2 cells were counted and 20,000 cells were plated into 6-well Transwell dishes. Cells were grown under normal conditions for 2, 4, 6, and 8 days. After the specified amount of growth time, cells were treated with Trypsin, diluted, and counted using the BioRad TC20 Automated Cell Counter.

The chromatin state of the human endothlien-1 gene (EDN1) gene was analyzed using the UCSC Genome Browser. The Genome Segments track and ChromHMM revealed a predicted promoter region with high confidence immediately downstream of the exon encoding the 3′ untranslated region (UTR) of EDN1 (about 800 bp long) in human umbilical vein endothelial cells (HUVEC) (Figure 1). This region also contained large clusters of DNase sensitivity sites, suggesting that the chromatin is in an open state and transcription factors could bind to this region (Figure 2). Using the UCSC Genome Browser transcription factor binding site database created using data from chromatin immunoprecipitation (ChIP) assays (ENCODE) (Karolchik et al., 2014), multiple transcription factors including C-JUN, CCAAT/enhancer-binding protein β (CEBPβ), and Forkhead Box A1 (FOXA1) were shown to bind at the location of the predicted promoter region. Additionally, Transcription Factor II D (TFIID), Upstream Transcription Factor 1 (USF1), and Glucocorticoid Receptor (GR) are predicted to bind to this region, and these transcription factors are associated with sites of active transcription. RNA polymerase II was also shown to bind to the predicted promoter region. Based on this information, we hypothesized that the predicted downstream promoter contained a transcription start site and potentially could be a promoter for an antisense RNA.

To determine if an antisense RNA is being transcribed from the EDN1 downstream promoter, a strand-specific reverse transcriptase PCR (ssRT-PCR) strategy was designed (Figure 3). Primers for the reverse transcriptase reactions (RT) were designed to have a complementary sequence to the EDN1 sense strand, so the primers will only anneal to RNA transcribed from the antisense strand (Table 1). Multiple strand-specific primers were designed at varying distances from the predicted promoter toward the 5′ end of the EDN1 gene (SS1-6). Oligo dT primers were used as a positive control for the RT reactions. For each RT reaction, a –RT reaction was performed without the reverse transcriptase enzyme as a negative control to ensure the absence of genomic DNA contamination in our samples.

Total RNA was isolated from primary human mammary endothelial cells (HMEC), human kidney proximal tubule epithelial cells (HK-2), human bronchoendothelial cells (S9), and human embryonic kidney cells (HEK293). Additionally, human RNA from kidney tissue was obtained from Life Technologies. Each of the RNA samples was treated with DNase and then converted to cDNA using a strand-specific primer or an oligo dT primer. For every sample, both + RT and –RT reactions were performed. The resulting cDNA products were amplified in a PCR using the primers PCR1 and PCR2, which amplified the same region of cDNA regardless of the SS or oligo dT primer used. An antisense RNA transcript was detected in HMEC, HUVEC, HK-2, and S9 cell cultures (Figures 4A,B). The antisense transcript was also detected in human kidney tissue RNA samples (Figure 4C). Interestingly, this antisense RNA was not detected in HEK293 cells, but EDN1 mRNA was also not detected in these cells (Figure 4D). These results demonstrate that there is an RNA transcribed in an anti-sense direction within the EDN1 gene locus, and we have designated this transcript EDN1-AS. SS primers 4–6 (Figure 3) were designed to detect if the length of EDN1-AS spanned the entire EDN1 coding region (Figures 4B,C). The results indicate that transcription of EDN1-AS proceeds at least as far as the first exon of EDN1.

We and others have shown that the EDN1 gene is under the regulation of the circadian clock proteins (Hanai et al., 2005; Stow et al., 2012; Kozakai et al., 2014; Richards et al., 2014). Additionally, GR was shown to bind to the EDN1-AS promoter region and glucocorticoids have been shown to modulate the circadian clock (Balsalobre et al., 2000; Kino and Chrousos, 2010; Surjit et al., 2011). E-box response elements, to which circadian clock transcription factors bind, were also present in the predicted EDN1-AS promoter region. These data suggest that, similarly to the EDN1 gene, EDN1-AS may be under control of the circadian clock and may display a circadian rhythm of expression. To determine if EDN1-AS exhibits a circadian rhythm of expression, a 24-h expression profile of the antisense transcript was generated (Figure 5). The HK-2 cells were treated with dexamethasone for 30 min to synchronize their circadian clocks (Feillet et al., 2014). Twenty-four hours after the dexamethasone treatment, RNA was isolated from the HK-2 cells every 2 h for 24 h and the ssRT-PCR protocol was performed as described above. RT reactions with random hexamers were also performed on the same samples in order to detect the levels of GAPDH expression as a control. As seen in Figure 5, EDN1-AS expression varies over a 24- h period.

Since the expression of EDN1-AS seems to be under circadian regulation, we wanted to investigate the effect of removing the E-box response element and GR binding site in the predicted promoter region of EDN1-AS. In order to create a KO of this regulatory region in HK-2 cells, we utilized CRISPR/Cas9 technology. Guide RNAs (gRNAs) were designed using the online tool CC Top (Stemmer et al., 2015; Table 2). gRNAs were selected based on location and predicted off-target sites. If possible, only gRNAs with no off-target sites or those with sites only in regions between genes were chosen. No gRNAs were used that had predicted off-target sites in exonic or intronic regions of any gene to minimize possible off-target effects. Various combinations of all the gRNAs were used in order to create a biallelic KO of the EDN1-AS regulatory region. The gRNA combination of gRNA 2 and 3 resulted in two HK-2 cell lines that had a homozygous deletion of an approximately 350 bp region within the predicted EDN1-AS promoter by genotyping PCR (Figure 6). Sanger sequencing using genomic DNA from these cell lines was used to confirm the KO region. In addition to the two homozygous KO cell lines, control (CNTL) cell lines that had undergone the CRISPR process, but did not have KO of the EDN1-AS regulatory region, were identified to use as control cells for future experiments.

To determine how the KO of the EDN1-AS regulatory region containing the predicted GR and circadian clock binding sites affected the expression of EDN1-AS, we performed our ssRT-PCR protocol on the KO cell lines in addition to the CNTL cells. The CRISPR KO cells had higher expression of EDN1-AS compared to CNTL cells (Figure 7A). Secreted endothelin-1 peptide (ET-1) was measured to determine if the increase in expression of the EDN1-AS gene led to an increase in ET-1 peptide level. Briefly, KO and CNTL cells were plated into Transwell plates and media was collected from confluent cells. The media was then analyzed via an ET-1 ELISA to measure total ET-1 peptide levels. ET-1 peptide levels were normalized to total protein concentration measured by BCA assay. The CRISPR KO cells had a significant increase in total secreted ET-1 peptide levels compared to the CNTL cells (Figure 7B).

High expression of ET-1 in human proximal tubule cells, like HK-2 cells, is associated with increased cell proliferation (Zoja et al., 1995; Seccia et al., 2016). To determine if the increased ET-1 levels caused by KO of the EDN1-AS regulatory region affected cell proliferation, growth rates of the KO and CNTL cells were measured. Briefly, 20,000 cells were plated into 6-well dishes and counted after 2, 4, 6, and 8 days of growth. Both CRISPR KO cell lines grew significantly faster than the CNTL cells (Figure 8). Significant differences between individual KO cells and the CNTL cells total cell count were detected on day 6 and day 8.