Origin of a novel CYP20A1 lncRNA through 23 Alu exaptations in the human lineage creates a potential multi-miRNA sponge

Alu repeats contribute to lineage specific novelties in conserved transcriptional regulatory networks. We report for the first time the origin of a multi-miRNA human specific sponge through exaptation of 23 Alu repeats that forms a novel principal isoform of CYP20A1 gene with a 9kb 3'UTR. This 3'UTR, confirmed by RACE, is an outlier in terms of its length with expression in multiple cell lines, including brain as evidenced from single nucleus RNA-seq data of 16000 human cortical neurons. It has diverged from its parent gene through exon skipping into a novel lncRNA. Its uniqueness in humans was validated by its presence in rosehip neurons and absence in closely related primate species through RNA-seq datasets. Strikingly, prediction by miRanda revealed ~4700 MREs for ~1000 different miRNAs, majorly in Alu repeats, in this 3'UTR. Permutations on 1000 random sets suggest their creation is non-random and post Alu exaptation. We hypothesise this lncRNA is an miRNA sponge as it has cytosolic localisation and harbours =10 MREs for 140 miRNAs (threshold > - 25kcal/mol). Under experimental conditions where CYP20A1 displays differential expression, we probed the expression of the miRNAs that map to this 3'UTR and their cognate targets through small RNA and mRNA-seq, respectively. We observed correlated expression of our lncRNA and a set of 380 genes with downregulation in heat shock and upregulation in HIV1-Tat treatment in primary neurons. GO analyses suggest the involvement of this sponge lncRNA in modulation of processes linked to neuronal development and hemostasis pathways. Overall design: A total of 18 human primary neuron samples (Control + 5 different treatment conditions; 3 biological replicates for each) were used for stranded paired-end RNA seq and small RNA seq. Since data for RNA as well as small RNA seq have been generated from the same sample, these are paired. Additionally, small RNA seq data have been generated from untreated MCF-7 cells (6 biological replicates).