Drosophila primary microRNA-8 encodes a microRNA-encoded peptide acting in parallel of miR-8

Abstract

Background: Recent genome-wide studies of many species reveal the existence of a myriad of RNAs differing in size, coding potential and function. Among these are the long non-coding RNAs, some of them producing functional small peptides via the translation of short ORFs. It now appears that any kind of RNA presumably has a potential to encode small peptides. Accordingly, our team recently discovered that plant primary transcripts of microRNAs (pri-miRs) produce small regulatory peptides (miPEPs) involved in auto-regulatory feedback loops enhancing their cognate microRNA expression which in turn controls plant development. Here we investigate whether this regulatory feedback loop is present in Drosophila melanogaster.

Results: We perform a survey of ribosome profiling data and reveal that many pri-miRNAs exhibit ribosome translation marks. Focusing on miR-8, we show that pri-miR-8 can produce a miPEP-8. Functional assays performed in Drosophila reveal that miPEP-8 affects development when overexpressed or knocked down. Combining genetic and molecular approaches as well as genome-wide transcriptomic analyses, we show that miR-8 expression is independent of miPEP-8 activity and that miPEP-8 acts in parallel to miR-8 to regulate the expression of hundreds of genes.

Conclusion: Taken together, these results reveal that several Drosophila pri-miRs exhibit translation potential. Contrasting with the mechanism described in plants, these data shed light on the function of yet undescribed primary-microRNA-encoded peptides in Drosophila and their regulatory potential on genome expression.

Keywords: Drosophila; Small peptides; lncRNA; miPEP; miR-8; sORF.

Fig. 1

Translatability of pri-miR-8. a Model of miPEP regulation in plants. b Box plot representation of the number of ORFs in different classes of RNAs. 3′UTR, 5′UTR and CDS represent coding RNAs, whereas lncRNAs and pri-miRs represent non-coding RNAs. An ORF was defined as starting with an ATG and coding for a minimum of 10 amino acids. Pri-miRs reveal comparable numbers of ORFs/kb as lncRNAs. c GWIPS-vis [42] genome viewer of the Drosophila miR-8 locus. Top: genomic positions and ORFs in the three reading frames. Green bars define ATGs and red bars stop codons. Below, RNA-seq profile is shown in green and ribosome profiling is shown in red. The blue horizontal lines represent the two CR43650 non-coding RNA transcripts and potential miR-8 pri-miRs. In black is schematized the transcript we identified as detectable pri-miR-8. Bottom: miPEP-8 amino acid sequence is shown. d Western blot experiments using the anti-miPEP-8 antibody. Left panel: in vitro synthetized miPEP-8 corresponding to the constructs indicated on top. The asterisk indicates the upstream initiated peptide. The arrow indicates the miPEP-8 initiated at ATG1. Middle panel: detection of miPEP-8 in S2 cells over-expressing miPEP-8 placed in different translational contexts; Kozak (optimal); K mt (ATG mutated into TGA); mt (ATG mutated into AGT). Note that the pri-miR is translated and endogenous miPEP-8 expression is undetectable in S2 cells. Right panel: anti miPEP-8 western blot of adult Drosophila extracts and in the miR-8 deleted line Δ2/Δ2 [26] in which no miPEP-8 is detected. We noticed the presence of non-specific bands as well as additional specific bands representing possibly miPEP-8 multimeric forms or PTM modifications. Ctrl corresponds to cell extracts transfected with an empty vector

Fig. 2

Drosophila miPEP-8 is biologically active during development. a Lethality assay on flies over-expressing miR-8, miPEP-8 or miPEP-8mt (ATG mutated) using the miR-8 GAL4 driver. Left: details of the genetic cross and expected percentage depending on the effect (neutral, deleterious or advantageous) on Drosophila development. Right: graph indicating the percentage of hatched flies over-expressing the different constructs. White flies (w) crossed with the driver line were used as a control. Expressing miR-8 resulted in developmental lethality since less than 20% of flies hatched (expected value 50%). A significant decrease occurred following miPEP-8 over-expression but not with the untranslatable miPEP-8mt construct. Number of independent crosses: for w and miR-8 n = 23; for miPEP-8 wt and mt n = 24. b Same as in a except the constructs were expressed in wings using the MS1096 driver and the phenotypes scored on wing size. Number of wings analyzed: for w n = 20; for miR-8 n = 27; for miPEP-8 wt and mt n = 27. * or ns: Significant differences are indicated relative to white recipient flies. AU arbitrary units

Fig. 3

Pri-miR-8 expression is independent of miPEP-8 control/activity. a schematic representation of constructs tested on miR-8 expression and activity levels. Arrows locate the primers used in the qPCR experiments determining miPEP and pri-miR relative expression levels. b the characterized pri-miR-8 produces a mature miR-8. S2 cells were transfected with a vector expressing pri-miR-8. Left: detection of the over-expression level of pri-miR-8 by qPCR. Right: detection of the over-expression level of mature miR-8 using the same RNA samples, n = 11 c miPEP-8 lacks repressive activity towards miR-8 expression. Left: level of miPEP-8 over-expression. Middle and right panels: quantification of pri-miR-8 and mature miR-8 in miPEP-8 over-expressing cells compared to control transfected cells (ctrl). n = 13 for the ctrl and 14 for miPEP-8. d, e Insensitivity of miR-8 sensor to miPEP-8 over-expression in S2 transfected cells (n = 16) (d) or in wing imaginal discs when miPEP-8 is expressed under the ptc-GAL4 promoter (e). In d, a miR-8 construct (n = 12) [17] was used as a positive control repressing the miR-8 luciferase sensor [20]. Of note, pri-miR-8 (n = 21) also repressed the miR-8 luciferase sensor. In e, first panel to the left: ptc GAL4 crossed with a UAS mCherry. Second panel to the left: expression pattern of the GFP miR-8 sensor alone. Scale bars (white) indicate 100 μm. A repressive activity is observed with miR-8 expressed in the ptc domain but not with miPEP-8. A representative disc is shown out of ten analyzed

Fig. 4

targeting miPEP-8 in vivo in Drosophila induces a wing phenotype. a Strategy for endogenous miPEP-8 edition. The pri-miR-8 gene region was deleted by CRISPR and a P landing site was created. Wild type and miPEP-8 ATG mutated pri-miR-8 in pattB were inserted at the P landing site. b Similar rescue efficiency was observed in at least three independent transgenic lines (left panel). qPCR on mature miR-8 in wild type and mutant (mt) pri-miR-8 Knock In (KI) lines showed similar miR-8 levels (n = 4), (right panel). c Wing phenotype in miR-8 deletion edited line. The pri-miR-8 miPEP-8 mutated (mt) shows a reduced wing size compared to the wild type pri-miR-8. (n = 15 and 28 respectively). d–f Analyses in miPEP-8 mutant identified in DGRP polymorphism. d miR-8 level determined by qPCR in white recipient flies (w) and in white flies carrying the miPEP-8 truncated form (miPEP-8alt), (n = 6 and 8, respectively). e, f Wing size determination in different genetic contexts. miPEP-8alt homozygotes or over miR-8 deficiencies revealed significant reduced wing size relative to the white recipient flies (w, n = 19; miPEP-8alt, n = 21; miPEP-8alt/miR-8 deletions, n = 40). Expressing miPEP-8 rescued the wing phenotype of miPEP-8alt flies relative to sibling flies not expressing miPEP-8 (n = 18 and 28, respectively). Significant (*) or nonsignificant (ns) differences are indicated either relative to white recipient flies or between the two groups

Fig. 5

Uncoupled activity of miR-8 and miPEP-8. a Rescue assay of miR-8- or miPEP-8-induced wing phenotype in flies co-over-expressing miR-8 or miPEP-8 along with a miR-8 sponge (miR-8sp) or a miR-8 scramble (miR-8scr). Only miR-8-sp (and not miR-8scr) compensates for miR-8-induced wing size reduction, hence efficiently titrating miR-8, while it has no effect on miPEP-8-induced wing phenotype. b Quantification of a. “ctrl” (MS1096/+) n = 19; “mir-8; mir-8scr” n = 20; “mir-8; mir-8sp” n = 21; “miPEP-8; mir-8scr” n = 23; “miPEP-8; mir-8sp” n = 19. * p < 0.05

Fig. 6

miR-8 and miPEP-8 control distinct set of genes. a Heatmap representing the RNA-seq results obtained from S2 cells over-expressing either miR-8 or miPEP-8. Significant sets of genes are modulated in response to mirR-8 or miPEP-8 over-expression, when compared to control transfected cells (ctrl). N = 5. b Venn diagram representing the miR-8 versus miPEP-8 modulated genes. c–e Different subgroups are distinguished; miPEP-8 specific (c), miR-8 specific (d) and co-regulated by miPEP-8 and miR-8 (e)