3’UTR of the CYP20A1_Alu-LT isoform modulates the expression of selected genes
To test the potential sponge function of the 3’UTR of CYP20A1_Alu-LT, we cloned fragments of these Alu-rich sequences downstream of the firefly luciferase gene. Four segments encompassing the whole UTR region were separately amplified, cloned, and transfected in the SK-N-SH cell line. The introduction of these fragments led to significant downregulation of firefly luciferase signals with a maximum reduction of 91% in clone 1 (Supplementary figure S1). Clone 1 (hereafter named Clone1-UTR-CYP20A1), the proximal 1/3rd of the UTR, with 48% of the total miRNA sites (56 out of 116) and 9 out of the 23 Alu repeats was selected for further experiments.
In our previous study, mRNA seq analysis in primary neurons showed upregulation and downregulation in the expression of 380 genes (including CYP20A1) after Tat treatment and heat shock, respectively. To ascertain if the overexpression of Clone1-UTR-CYP20A1 in the SK-N-SH cell line could also increase the expression of these genes, a few genes namely, GAP43, URB1, SLC20A2, NEIL2, EIF4H, MCAM, and ORAI2 were selected based on their RNA seq expression levels (Supplementary figure S2). Their expression was checked via qRT-PCR after the transfection of Clone1-UTR-CYP20A1. A ~ 21-fold increase in the expression of the proximal 3’UTR was achieved by Clone1-UTR-CYP20A1 transfection (Fig. 1a). Interestingly, a pattern of upregulation in gene expression was observed for all the seven genes that were tested and GAP43, URB1, and SLC20A2 were significantly upregulated (Supplementary Figure S3a, Fig. 1b).
In order to strengthen our findings, we performed a knockdown of CYP20A1_Alu-LT. siRNAs targeting the 3’UTR of CYP20A1_Alu-LT significantly downregulated CYP20A1_Alu-LT levels (30%) and also suppressed GAP43 and MCAM expression (Fig. 1c). Though the knockdown efficiency was not high it was intriguing to find that all seven genes showed an approximate 20% decrease in the expression after the knockdown of the transcript (Supplementary Figure S3b). This implies that differentially expressed CYP20A1_Alu-LT can alter the target gene expression even in the absence of stress triggers.
To ascertain if the Clone1-UTR-CYP20A1 is mediating these effects via binding to miRNAs, we tested its association with RNA-induced silencing complex (RISC). When RISC was enriched using Ago2 immunoprecipitation, there was a 4-fold enrichment of CYP20A1_Alu-LT compared to the negative control IgG, confirming the association of the transcript with the miRNA machinery (Fig. 1d).
Interestingly, GAP43 (Growth Associated Protein 43) was significantly upregulated as well as downregulated after overexpression and knockdown of CYP20A1_Alu-LT, respectively. We earlier predicted miRNA binding sites on GAP43 using miRanda (version 3.3a). miRanda predicted two potential miRNA binding sites one each for miR-619-5p and miR-3677-3p on the 3’UTR of GAP43. However, the 3’UTR of CYP20A1_Alu-LT has far more predicted sites i.e. 26 and 12 sites for miR-619-5p and miR-3677-3p, respectively [14]. The region of Clone1-UTR-CYP20A1 alone had 12 and 8 sites for each of the miRNA all on exonized Alus (Fig. 2a). Both these miRNAs were expressed in SK-N-SH cell lines at comparable levels (Supplementary Figure S4). Therefore, we hypothesize that CYP20A1_Alu-LT might be acting as a molecular sponge for these miRNAs to alter the expression of GAP43. When the Alu-rich region is overexpressed, miRNAs would bind more to the CYP20A1_Alu-LT than GAP43, leading to increased expression of GAP43. However, when the expression of CYP20A1_Alu-LT is decreased the miRNAs are free and bind to target sites on GAP43 thus decreasing its expression (Fig. 2b).