Widespread regulatory activity of vertebrate microRNA* species

Abstract

An obligate intermediate during microRNA (miRNA) biogenesis is an ~22-nucleotide RNA duplex, from which the mature miRNA is preferentially incorporated into a silencing complex. Its partner miRNA* species is generally regarded as a passenger RNA, whose regulatory capacity has not been systematically examined in vertebrates. Our bioinformatic analyses demonstrate that a substantial fraction of miRNA* species are stringently conserved over vertebrate evolution, collectively exhibit greatest conservation in their seed regions, and define complementary motifs whose conservation across vertebrate 3'-UTR evolution is statistically significant. Functional tests of 22 miRNA expression constructs revealed that a majority could repress both miRNA and miRNA* perfect match reporters, and the ratio of miRNA:miRNA* sensor repression was correlated with the endogenous ratio of miRNA:miRNA* reads. Analysis of microarray data provided transcriptome-wide evidence for the regulation of seed-matched targets for both mature and star strand species of several miRNAs relevant to oncogenesis, including mir-17, mir-34a, and mir-19. Finally, 3'-UTR sensor assays and mutagenesis tests confirmed direct repression of five miR-19* targets via star seed sites. Overall, our data demonstrate that miRNA* species have demonstrable impact on vertebrate regulatory networks and should be taken into account in studies of miRNA functions and their contribution to disease states.

FIGURE 1.

Evolutionary profiles of well-conserved vertebrate miRNA genes. (A,B) Multispecies alignments of two miRNA genes that are conserved from humans to fish. (Green) Mature strands; (yellow) star strands; (red) nucleotides diverged with respect to human. (A) mir-20b is highly conserved; its mature product has sustained a few positions of divergence, but none involve its seed (nucleotides 1–8). On the other hand, miR-20b* has accumulated many more positions of divergence, including in seed nucleotides in its fish orthologs. (B) mir-18a is perfectly conserved along both miRNA and star arms among all vertebrates, from human to fish. Such extreme constraint is suggestive of conserved regulatory activities of both small RNAs produced by mir-18a. (C) Sequence divergence in 7-nt windows across 106 miRNA genes whose star arms sustained ≤3 diverged positions between human and chicken. We note three observations: (1) miRNA strands (green) are better conserved than star strands (yellow); (2) the ends of both miRNA and star strands are better conserved than their central regions; and (3) the 2–8 seed windows exhibit highest conservation along miRNA and star sequences (dotted reference line); the mature 1–7 and 2–8 windows had similar scores.

FIGURE 2.

Validation of the regulatory activity of mature strand and star strand targets of mir-142. (A) mir-142 has been exceptionally conserved along both its miRNA (green) and star (yellow) strands; positions of divergence (red) reside in the pre-miRNA flanks or in the terminal loop. (B) Schematics of artificial sensors for miR-142-5p or miR-142-3p, containing either two perfect target sites, four bulged sites, or four bulged sites with seed mismatches. (C) Target repression by both miR-142-5p and miR-142-3p is seed-dependent. Sensors were assayed and normalized as described in Figure 4. Introduction of the mir-142 expression plasmid yields robust repression of both perfect and bulged sensors, but mutation of seed nucleotides abolishes target repression.

FIGURE 3.

General tests of the dual regulatory capacity of miRNA genes. (A) Renilla luciferase sensors containing four antisense matches to either the mature miRNA or star species of a given miRNA gene were tested for their response to a cognate pri-miRNA expression plasmid in HeLa cells. Sensor values were normalized to an internal firefly luciferase transfection control and then represented as the fold repression relative to the sensor level in the presence of a non-cognate miRNA expression plasmid (usually mir-1-2; the control for mir-1-2 was mir-199a-2). We deemed a miRNA expression plasmid to be “dual function” if the mean repression value was at least two standard deviations above 1. Values are derived from two independent sets of quadruplicate transfection experiments performed on different batches of cells; standard deviations are shown. Some miRNA constructs were only capable of repressing the mature strand sensor, but a majority of constructs could repress both mature strand and star sensors. Three genes for which the inferred miRNA* species (based on meta-analysis of published library data) yielded slightly higher repression than their partner miRNA species are segregated to the right. (B) Correlation of ectopic miRNA sensor tests and endogenous cloning ratios. We collected small RNA reads from the 22 loci tested in Figure 2 and this figure and compared their miRNA:miRNA* cloning ratios with their miRNA:miRNA* sensor repression ratios. A rank analysis was performed to group genes with higher cloning ratio (i.e., more asymmetric accumulation of the duplex strands) or lower cloning ratio (i.e., more balanced accumulation of the two strands). The correlation with higher versus lower sensor repression ratios was statistically significant. (C) A linear regression was performed between the miRNA:miRNA* sensor repression ratio and the log₂(miRNA:miRNA*) cloning ratio. The correlation was significant according to both Pearson's tests.

FIGURE 4.

Transcriptome-wide evidence for target repression directed by partner miRNA and star species. (A) Target repression by mature (green) and star (yellow) products of mir-17. Linsley et al. (2007) reported microarray profiles for cells transfected with mimics for either miR-17-5p (mature) or miR-17-3p (star), but only analyzed the mature strand response. Using miREDUCE, we observe that the motifs that are most highly correlated with down-regulated transcripts in these data sets are the miR-17-5p seed (2–8 or 3–8) and the miR-17-3p seed (2–8 or 2–7seed + t1A), respectively, all with P-values = 0. Other highly statistically enriched motifs in down-regulated transcripts include various seed variants (see also Supplemental Table 2). (B) Mendell and colleagues reported microarray profiles following infection with a retroviral mir-34a construct (Chang et al. 2007) and reported that the miR-34a seed was enriched among the 100 most down-regulated transcripts. miREDUCE on this transcript set yielded statistically significant enrichment of miR-34a 2-7 seed (green) among the top 200 down-regulated transcripts, but this enrichment increased when larger sets of down-regulated transcripts were analyzed. These same transcript subsets yielded even stronger enrichment for miR-34a* seeds (yellow) across all comparable cohorts of down-regulated genes, peaking at 1.23E-06 in the top 500 most down-regulated transcripts.

FIGURE 5.

Evidence for targeting by endogenous miR-19 and miR-19*. (A) Dominant read counts of mmu-mir-19a and mmu-mir-19b-1 analyzed from six data sets reported in GSE11724 (Marson et al. 2008; see also Supplemental Fig. 2). Both genes generate a mature species with a shared seed (red box); miR-19a* generates two 5′ isomiRs, one of which is shared with miR-19b-1* (green box); note that the GUUUUGC miR-19* seed is by far the dominant star seed generated by the mir-19 genes. (B) Cumulative distribution function of gene expression in lymphoma cell lines deleted for the mir-17→92 cluster compared to those re-expressing mir-19a/b under retroviral control. Transcripts whose 3′ UTRs contained exclusively miR-19 (GUGCAAA) or miR-19* (GUUUUGC) seed sites are segregated for analysis; transcripts containing both types of seed matches were discarded to avoid ambiguity in assigning the targeting species. Transcripts with miR-19 seed sites (red) or miR-19* seed sites (green) were down-regulated in the presence of mir-19 relative to background gene expression of all other transcripts (black); this trend was more substantial for the mature strand miR-19 but still statistically significant for star strand targets. The region boxed in the main graph is expanded to illustrate this trend more clearly. (C) Fold change of transcript down-regulation for selected targets bearing conserved miR-19* seed matches. Following multiple hypothesis correction by the false discovery rate (FDR), log fold changes with adjusted P-values (FDR <%5) were considered significant; all of these gene expression changes were highly significant. (D) Confirmation of direct repression by miR-19*. 3′-UTR target sensors and matched mutants bearing specific point changes within the miR-19* seed matches were tested for their response to transfection of mir-19a/b in HeLa cells; sensor activities were normalized to their level in the presence of functional mir-1-2 construct. All five were repressed in a manner that was completely dependent on the integrity of the miR-19* seed matches (cf. wild-type sensors in green with mutant sensors in tan).