How primary microRNAs (pri-miRNAs) with complex secondary structures are recognized and processed in plant cells has up until now been unclear. A new study suggests that unlike canonical processing of pri-miRNAs, terminal loop–branched pri-miRNAs can be processed by Dicer-like 1 (DCL1) complexes bidirectionally, either from the lower stem to the terminal loop or vice versa, resulting in productive and abortive processing of miRNAs, respectively.
Multibranched terminal loops in pri-miR-166s substantially suppress miR-166 expression in vivo. In both cases, DCL1 complexes cut pri-miRNAs at a distance of 16–17 bp from a reference single-stranded (ss) loop region. DCL1 also adjusts processing sites toward an internal loop through its helicase domain.
Pri-miRNAs are first processed by 2 proteins: Drosha, and Pasha (a.k.a. “DGCR8”). Pasha associates with the base of the stem-loop structure, insodoing defining the catalytic cleavage site for the RNAse III enzyme Drosha precisely 11 bp away from the ssRNA-dsRNA junction. This complex, aptly named “Microprocessor”, produces a shorter, ‘precursor’ miRNA (pre-miRNA), which is then cleaved by an RNAse II enzyme, Dicer, to release ~21bp miRNA:miRNA* duplexes.
* indicates the less predominant form of the precursor where cloning studies have allowed researchers to determine which arm gives rise to the predominantly expressed miRNA.
Mature miRNAs are loaded into Argonaute (AGO)-containing RNA-induced silencing complexes, guiding AGO to repress their complementary targets. Arabidopsis rockcressare a model genus of plants, widely used in research. DCL1, one of four Dicer homologs, orchestrates this process, with 2 cofactors (serrate and HYL1) acting as a “molecular ruler” for each miRNA, much like Pasha.
Many plant pri-miRNAs have branched terminal loops (BTLs), or multiple bulges or internal loops in their lower and upper stems. The protein family in this study, miR-165 and miR-166, differ by just 1 nt and both target the HD-Zip III family of plant transcription factors, which combine a homeodomain (that forms a triple helix structure to bind RNA or DNA) with a leucine zipper dimerisation motif (that “zips up” with another Leu zipper molecule at hydrophobic regions).
As shown above, miR-166c assumes a terminally branched stem loop structure, in contrast to the unbranched miR-166f. As such, the concomitant productive and abortive processing of the pri-miRNA is observed only for the -166c transcript.
The finding of differential regulation of miRNA production via uni- or bidirectional processing of DCL1 complexes unveils a little more of the intricate relationship between pri-miRNA structure and miRNA biogenesis in plants.
What’s more, bidirectional processing was found beyond the pri-miR-166s:
We found that 77 out of 232 Arabidopsis hairpins from miRBase had BTLs. To examine whether bidirectional processing of pri-miRNAs occurs beyond the pri-miR-166 family, we first explored eight previously described degradome data sets for evidence of abortive processing. Degradome sequencing determines the RNAs that have 5′ monophosphates, which include remnants of DCL1-catalyzed pri-miRNA processing, among other types of processed RNAs. We defined abortive processing by the presence of degradome-inferred remnants ending between 15 and 17 nt from the loop-proximal end of the miRNA-containing helix. Of the 155 hairpins with normal, unbranched loops, 6 had evidence of abortive processing (3.9%). 7 of the 77 (9.1%) hairpins with BTLs had such evidence.
These products, they argue, are probably unstable, thereby escaping detection in degradome sequencing operations.
Ingeniously, the group thought to look for small RNA (sRNA) byproducts, formed from self-complementary intermediate abortive processing fragments as evidence of abortive processing in existing “degradome” sequencing data. Mining high-throughput sequence data produced matches with the degradome-inferred results.
In one instructive case, we found that pri-miR-825 harbors branched loops in its lower stem and a normal loop in the upper stem. Notably, the directions of productive and abortive processing are also switched for pri-miR-825. Similar studies in rice also recovered several pri-miRNAs, including pri-miR-166c, a result suggesting that the presence of bidirectional processing of the same pri-miRNAs extends beyond Arabidopsis.
Lastly, the authors reasoned the other pre-miRNAs produced from miR-166 such as -166c must have been downregulated through some other mechanism, as the low in vivo levels observed across the family couldn’t be explained wholly due to this bidirectional abortive processing. They found that
a large internal loop in the upper stem or multibranched terminal structures also affected pre-miRNA stability... indicating that the secondary structures of terminal regions in pre-miRNAs regulate their stability, further controlling homeostasis in vivo.
Many pri-miRNAs (e.g. pri-miR-166c) contain large terminal loops, which can ‘disguise’ a ssRNA-dsRNA junction to recruit DCL1 complexes for cleavage.
This outside-in mechanism has two outcomes: if DCL1 goes from base to loop, it produces mature miRNAs; if it goes in the opposite direction, it destroys them.
■ Zhu et al (2013) Bidirectional processing of pri-miRNAs with branched terminal loops by Arabidopsis Dicer-like1. Nature Structural and Molecular Biology 20(9):1106–1115
■ Han et al (2006) Molecular Basis for the Recognition of Primary microRNAs by the Drosha-DGCR8 Complex. Cell 125:887–901