A genome-wide transgenic resource for conditional expression of Drosophila microRNAs

Abstract

microRNAs (miRNAs) are endogenous short RNAs that mediate vast networks of post-transcriptional gene regulation. Although computational searches and experimental profiling provide evidence for hundreds of functional targets for individual miRNAs, such data rarely provide clear insight into the phenotypic consequences of manipulating miRNAs in vivo. We describe a genome-wide collection of 165 Drosophila miRNA transgenes and find that a majority induced specific developmental defects, including phenocopies of mutants in myriad cell-signaling and patterning genes. Such connections allowed us to validate several likely targets for miRNA-induced phenotypes. Importantly, few of these phenotypes could be predicted from computationally predicted target lists, thus highlighting the value of whole-animal readouts of miRNA activities. Finally, we provide an example of the relevance of these data to miRNA loss-of-function conditions. Whereas misexpression of several K box miRNAs inhibited Notch pathway activity, reciprocal genetic interaction tests with miRNA sponges demonstrated endogenous roles of the K box miRNA family in restricting Notch signaling. In summary, we provide extensive evidence that misexpression of individual miRNAs often induces specific mutant phenotypes that can guide their functional study. By extension, these data suggest that the deregulation of individual miRNAs in other animals may frequently yield relatively specific phenotypes during disease conditions.

Fig. 1.

Use of the Gal4-UAS binary system for systematic in vivo screens of activated miRNA transgenes. We made two collections of UAS-miRNA transgenes, one consisting of random insertions in P-element backbones (165 different transgenes covering 149 different miRNA hairpins) and another consisting of defined insertions in attP landing sites (106 different transgenes covering 108 different miRNA hairpins), including stand-alone loci, miRNA clusters and subdivided clusters. The collections have complementary experimental strengths. We crossed these with a variety of ubiquitous and tissue-specific Gal4 drivers to assess systematically the consequences of ectopic miRNA activity. Examples of Gal4 driver patterns are shown on the left in wing imaginal discs (A-C) in which a lacZ or GFP reporter is driven by representative Gal4 insert; A′-C′ show reporter accumulation counterstained with DAPI. (A,A′) ptc-Gal4 is active at the border of the anterior (a) and posterior (p) compartments. (B,B′) bx-Gal4 is active throughout the wing pouch, but is elevated in the dorsal (d) relative to the ventral (v) compartment, and is not active along the presumptive wing margin (wm). (C,C′) sd-gal4 is active more uniformly throughout the wing pouch.

Fig. 2.

A diversity of patterning defects can be scored in the Drosophila wing. Shown are wild-type (A) and mutant (B-K) adult wings that illustrate major classes of mutant phenotypes. The directionality of pathway activity that yields these phenotypes is noted in the upper right corner of each panel. (A) The normal wing has a characteristic size and shape, and reproducible pattern elements such as the five longitudinal wing veins (L2-L5 are labeled) and sensory bristles that decorate the anterior margin of the wing. (B) Wing notching caused by mutant clones of the transcription factor in the Notch pathway, Su(H). (C) Loss of wing caused by knockdown of the Hedgehog ligand. (D,E) Examples of wing vein loss caused by misexpression of the Notch pathway target E(spl)mγ (D) or knockdown of the BMP ligand dpp (E). (F) Thick veins caused by high level expression of the Notch pathway component neuralized. (G) Expression of the activated MAP kinase rolled (Sevenmaker, Sem) causes ectopic wing veins. (H,I) Examples of wings that bear multiple mutant phenotypes. Asterisk indicates area of wing notching. (H) Knockdown of Su(H) induces thick veins and wing notching. (I) Misexpression of an activated BMP receptor tkv-QD causes both thick veins and ectopic veins. (J) Knockdown of the Hippo pathway component expanded in the central domain of the wing causes overgrowth (arrows). (K) Expression of a dominant-negative version of the Hedgehog receptor Smoothened (SmoDN) causes loss of the L3-L4 domain.

Fig. 3.

Selected examples of miRNA-induced wing phenotypes illustrate general properties of the UAS-DsRed-miRNA transgenes. (A,B) Wings of a heterozygous bx-Gal4 (‘bx’) female (A) and hemizygous male (B); males are smaller than females, which accounts for size difference. (C-F) Examples of dose effects. Vein thickening (arrows) induced by mir-263b is weaker in bx-Gal4 females (C) than males (D). Wing notching (asterisks) induced by mir-977 is weaker in sd-Gal4 (sd) females (E) than males (F). (G-I) Different miRNA misexpression phenotypes are evident in different Gal4 backgrounds. (G) bx/Y>mir-7 exhibits massive vein thickening (arrows), but the margin is continuous. (H) sd/Y>mir-7 exhibits massive wing notching (asterisks), and only mild vein thickening. (I) ptc-Gal4>mir-7 exhibits distal wing notching and reduction in the L3-L4 domain (arrows). (J,K) Dissection of a miRNA cluster. (J) Activation of the mir-11/mir-998 operon induces vein (arrow) and crossvein (arrowhead) loss; these phenotypes are recapitulated by ectopic mir-11 (K). (L) Similarity of seed families; mir-2 is in the same family as mir-11 and also induces vein loss (arrows).

Fig. 4.

Summary of wing phenotypes caused by misexpression of different Drosophila miRNAs. Shown are adult female wings, except as noted for X-linked Gal4 drivers (/X= female; /Yμmale). (A-E) Except for the wild-type [w1118] wing (A), all other flies contain a single copy of Gal4 and UAS-DsRed-miRNA transgene. (B) bx-Gal4/X heterozygous females exhibit a normal wing, as do sd-Gal4/X females (not shown). (C) Gal4 activity in sd-Gal4/Y males results in a minor loss of posterior wing margin, especially near the wing hinge (boxed regions in A and C are enlarged in D and E, respectively). (F-J) Examples of vein loss induced by different miRNAs. (K-T) Examples of vein thickening induced by different miRNAs. (U-AA) Examples of wing notching induced by different miRNAs. Note that P-S and U-X highlight miRNAs that induce vein thickening or margin loss, respectively, depending on the driver. Other combinations of phenotypes are evident by inspection. (BB-DD) Examples of ectopic wing veins (arrows) induced by different miRNAs. (EE-OO) Examples of miRNAs that have a selective effect on wing margin bristles; close-ups of the anterior margin are shown in JJ-OO. (PP) ptc-Gal4 heterozygous female; the ptc+ domain includes the L3-L4 region marked by the double arrow. (QQ-UU) Examples of miRNAs that induced overgrowth of the L3-L4 domain. (VV-ZZ) Examples of miRNAs that induced undergrowth or loss of the L3-L4 domain. (AAA-EEE) Examples of miRNAs that induced wing blisters. (FFF) miRNA that induces a potential defect along the anterior-posterior axis. (GGG) miRNA that induces a potential proximal-distal defect. (HHH-OOO) Examples of other severe wing deformities or wing loss induced by different miRNAs.

Fig. 5.

Validation of miRNAs that directly target abrupt. (A) Targetscan predictions of conserved miRNA-binding sites in the abrupt 3′ UTR. (B-D) Adult female wings. (B) The viable abrupt[1] mutant exhibits loss of distal L5 wing vein (arrow). (C) Misexpression of let-7 induces wing deformity even when cultured at low temperature to limit Gal4 activity; in addition, loss of the distal region of L5 is seen (arrow). (D) Misexpression of mir-275/mir-305 also induces loss of L5. (E-G″) Transgenic sensor assays in wing imaginal discs that carry tub-GFP-abrupt 3′ UTR, dpp-Gal4 and UAS-DsRed (linked to a miRNA in F-G″); the central domain of the wing pouch is shown. (E-E″) Control staining shows that expression of DsRed does not repress the abrupt sensor. (F-F″) Ectopic let-7 strongly represses the abrupt sensor. (G-G″) Ectopic miR-275 mildly represses the abrupt sensor. (H) Renilla-abrupt 3′ UTR sensor assays in S2 cells. Consistent with the in vivo results, mir-275 weakly repressed the abrupt sensor, while let-7 strongly repressed it; mir-iab-4 has previously been validated to repress the abrupt 3′ UTR (Okamura et al., 2008). Data are mean±s.e.m.

Fig. 6.

K box miRNA sponges enhance Notch signaling during wing development. (A,C-H) Notch[55e11]/+ (N/+) heterozygous females that carry ptc-Gal4 and two copies of the indicated miRNA sponges (SP); scr, scrambled sponge control. (A) N/+ females expressing control sponges exhibit a notch (asterisk) at the distal tip and mild vein thickening; the regions outlined are magnified to highlight these phenotypes. (B) Misexpression of any of the sponges used in this figure did not alter wing development; ptc-Gal4>2xmir-13aSP is shown as an example. The magnified insets exhibit normal vein thickness and can be used to judge N/+ haploinsufficiency. (C) Misexpression of the miR-13aSP rescued N/+ notching, but not vein thickening. (D) Misexpression of miR-7SP did not rescue either N/+ phenotype. Asterisk indicates area of wing notching. (E-H) Magnifications of the distal wing tips to highlight the status of wing notching in other sponge backgrounds. Asterisk indicates area of wing notching. (E) miR-2bSP, (F) miR-2cSP and (G) miR-13bSP all rescued N/+ notching, but (H) miR-6SP could not (asterisk). (I) Quantification of rescue of wing notching in various genotypes. N/+ in various ptc-Gal4>UAS-mir-SP backgrounds exhibit notching in about two-thirds of animals, this is reduced to less than 20% in the presence of miR-2cSP, to 10% in miR-2bSP and miR-13aSP, and to less than 2% in miR-13bSP. Inset shows the sequence relationship of these K box miRNAs.