Functional microRNA targeting without seed pairing

MicroRNAs (miRNAs) associate with Argonaute (AGO) proteins to serve as guides, directing binding to partially complementary sites in mRNAs, ultimately causing post-transcriptional repression. Complementarity to the miRNA seed (miRNA nucleotides 2–7) is typically necessary and sufficient for repression. Here, we investigate unusual sites with extensive complementarity to the miRNA 3′ region (nucleotide 9 and onwards) but without complementarity to the seed. The top examples of these 3′-only sites bind as well as top canonical sites and impart similar repression, which can be further boosted by as few as 2–3 additional pairs to the miRNA seed. Despite these similarities, 3′-only sites have slower association and dissociation rates than seed-matched sites. They also impart different conformations to bound AGO–miRNA complexes than do seed-matched sites, and individual miRNAs differ substantially with respect to how well they bind their respective 3′-only sites. Thus, pairing to the seed is not always required for binding and repression, or for a target to gain access to the 3′ region of the guide. For miRNAs that recognize 3′-only sites, those sites are estimated to constitute <1% of the set of endogenous target sites, a proportion resembling that of other rare but functional site types such as 3′-compensatory sites. AGO-RBNS: AGO-RBNS was conducted with various concentrations of purified AGO-miR-155 and AGO-miR-124 and a 100nM RNA pool consisting of 50nM random-sequence library and 50nM programmed library. Bound RNAs were captured, and libraries were prepared for high-throughput sequencing as in "The biochemical basis of microRNA targeting efficacy" (McGeary, Lin et al., 2019 Science) from both AGO-miR-bound and input RNA samples, and then sequenced on an Illumina HiSeq 2500. Relative abundance of sequence motifs in bound versus input RNA-derived sequencing reads were used to determine the relative AGO-miR binding affinity of each motif. MPRA: HeLa or F9 cells were transfected with miR-155, miR-124, miR-1, or mock transfected, as well as a massively parallel reporter library in which each library variant was the same mEGFP expression plasmid with a different target site to miR-155 or miR-124 cloned into the EGFP 3ʹ UTR. After 24 or 48h, total RNA was extracted, reverse-transcribed, and cDNAs containing library contant regions amplified by PCR, for sequencing on an Illumina Novaseq SP. Relative abundance of individual library members in miRNA-transfected versus mock-transfected samples were used to determine the relative repression imparted to the reporter mRNA by the presence of a given miRNA target site in its 3ʹ UTR