To investigate whether let-7-C miRNAs collectively regulate developmental timing in Drosophila, we generated two independent let-7-C knockout strains, let-7-C^KO1 and let-7-C^GKI (Fig. 1B,C; Supplemental Fig. S2). Both strains lack expression of the mature processed forms of miR-100, let-7, and miR-125 (Fig. 1B,C). To reduce the potentially complicating effects of genetic background on our study, we analyzed the let-7-C^KO1 and let-7-C^GKI knockout alleles in trans and refer to this trans-heterozygous let-7-C-null strain as let-7-C^KO1/GKI. We found that ~43% of let-7-C^KO1/GKI animals died prematurely during the course of development, with the majority (74%) of these arresting at the very end of metamorphosis. The remaining 57% of let-7-C^KO1/GKI mutants eclosed as adults, but displayed chronic defects in adult function, including severely reduced motility, flight, and fertility (Fig. 2). let-7-C^KO1/GKI mutants that carried a transgene that restored let-7-C miRNA expression were fully rescued for developmental viability and adult functions (Fig. 2; data not shown). Despite their developmental and behavioral defects, let-7-C^KO1/GKI mutant pupae and adults appeared morphologically normal (Supplemental Fig. S3), indicating that let-7-C miRNAs are not required for the morphogenesis of the adult exterior. These data indicated that let-7-C expression is predominantly required for adult behavior and are consistent with the hypothesis that let-7-C miRNAs play an essential role in regulating the developmental remodeling of internal tissues during metamorphosis.

To test whether the activity of each of the let-7-C miRNAs is required for let-7-C function, we analyzed the phenotypes of three different let-7-C derivative strains in which the expression of miR-100, let-7, or miR-125 had been eliminated individually (Fig. 1C; Supplemental Material). We refer to these singly mutant strains as miR-100^Δ, let-7^Δ, and miR-125^Δ, respectively. miR-100^Δ mutants functioned normally in all behavioral assays (Fig. 2), indicating that miR-100 was not solely responsible for any of the identified let-7-C functions. None of the single mutant strains displayed strong male fertility or climbing defects (Fig. 2D), suggesting that for normal male fertility and climbing behavior, the combinatorial action of any two let-7-C microRNAs could suffice. In contrast, let-7^Δ and miR-125^Δ mutants displayed severely reduced spontaneous locomotion as well as partial defects in flight (Fig. 2A–C). The normal climbing and nearly normal flight of let-7^Δ and miR-125^Δ mutants suggested that their severely impaired spontaneous locomotory activity was not simply the consequence of physically or metabolically impaired mobility, but rather likely reflected a behavioral deficit of neurological origin. Finally, let-7^Δ mutants alone displayed moderately severe defects in female fertility and oviposition (Fig. 2E,F), indicating that let-7 was required for an essential function to promote female reproduction.

To identify the specific place where let-7-C miRNAs may function to promote adult behavior, we examined the spatiotemporal expression pattern of the let-7-C locus. We used the let-7-C^GKI strain, in which the yeast transcriptional activator Gal4 had been inserted into the let-7-C locus (Fig. 1B), to drive expression of Gal4-dependent transgenes encoding membrane-bound or nuclear forms of GFP. A UAS-let-7-C transgene placed under the control of the let-7-CGal4 insertion restored miR-100, let-7, and miR-125 expression (Fig. 1D, lane 9) as well as climbing activity (Fig. 3A) to let-7-C^KO1/GKI mutants. Three characteristics of the let-7-CGal4 expression pattern are outlined below. First, let-7-CGal4 was expressed in neurons throughout the adult brain and ventral nerve cord, and this adult CNS expression was the culmination of a dramatic expansion in the spatial expression pattern of let-7-CGal4 that occurred in the CNS during the first half of metamorphosis (Fig. 3B; data not shown). Second, let-7-CGal4 was expressed in neurons that innervated structures throughout the adult (Fig. 3C), including sensory organs in the head, flight muscles in the thorax, and the alimentary tract, the male and female reproductive tracts, and the male and female genitalia in the abdomen (data not shown). We noted that let-7-CGal4 was very densely expressed in the posterior tip of the adult abdominal ganglion (Fig. 3B), as well as in motoneurons that projected posteriorly and innervated two distinct sets of abdominal muscles, the dorsal internal oblique muscles (DIOM) and the dorsal muscles (DM) (Fig. 3C). The DIOMs are remnants of the larval body wall that persist through metamorphosis (presumably to function in the process of eclosion) and in the wild type are fated to die within 12 h of eclosion (Crossley 1978; Kimura and Truman 1990). In contrast, the DMs are the adult body-wall muscles and are derived from larval myoblasts that undergo myogenesis during metamorphosis (Miller 1950; Currie and Bate 1991, 1995). Third, let-7-CGal4 was not only expressed in motoneurons but in muscle cells as well, including the DIOMs and DMs (Fig. 3E; data not shown). Taken together, the expression of let-7-CGal4 in pupal and adult neurons and muscles is consistent with the hypothesis that the behavioral phenotypes of let-7-C mutant adults are the consequence of defects in the metamorphosis of the neuromusculature.

To test whether let-7-C miRNAs play a role in specifying the configuration of the adult neuromusculature, we examined the abdominal muscle system of let-7-C^KO1/GKI mutants since, as shown above, let-7-C is expressed in abdominal motoneurons and muscles. We found two very clear and highly penetrant defects (Fig. 4A,B). First, the DIOMs that ordinarily decay during post-eclosion maturation of wild-type flies failed to disappear in older let-7-C^KO1/GKI mutants (Fig. 4A,B). We quantified this phenotype by scoring the presence of six DIOMs in aged wild-type and let-7-C^KO1/GKI mutant flies. Two-day-old wild-type males (n = 10) retained none of these DIOMs, whereas 2-d-old let-7-C^KO1/GKI males (n = 10) retained 89.9% ± 11.7% of these DIOMs. Nine-day-old male let-7-C^KO1/GKI mutants (n = 6), the oldest cohort of let-7-C^KO1/GKI mutants examined, retained 91.5% ± 9.3% of these DIOMs. Second, the DMs of let-7-C^KO1/GKI mutant adults were clearly smaller than those in age-matched wild-type controls (Fig. 4A–D). We quantified this phenotype by measuring the width and number of nuclei present in a set of approximately six to eight distinct DMs close to the dorsal vessel per A4 hemisegment per 2-d-old wild-type (n = 7) or let-7-C^KO1/GKI (n = 7) male. Wild-type DMs were 20.2 ± 2.7 μm in width and contained 13.9 ± 1.2 nuclei, whereas let-7-C^KO1/GKI DMs were 12 ± 1.6 μm in width and contained 10.6 ± 0.8 nuclei (Fig. 4C,D). Restoration of let-7-C expression rescued both the DIOM and DM phenotype; let-7-C^KO1/GKI mutants that carried the let-7-C⁺ transgene (n = 6) retained none of the six DIOMs scored above, and their DMs were 16.5 ± 1.4 μm in width and contained 12.8 ± 0.6 nuclei. To test whether let-7-C^KO1/GKI mutant muscle phenotypes were the consequence of defects apparent prior to the onset of metamorphosis, we examined the musculature and myoblasts of let-7-C^KO1/GKI larvae and found that both appeared normal (Supplemental Fig. S4). We therefore concluded that the abdominal muscle system of let-7-C^KO1/GKI mutant adults failed to complete its larval-to-adult remodeling, displaying both persistent pupal as well as immature adult characteristics. We interpreted this as a heterochronic phenotype, since let-7-C^KO1/GKI mutant adults exhibited both juvenile features (e.g., muscle system morphology) as well as mature adult traits (e.g., external appearance) at the same time.

To test whether let-7-C affects the remodeling of other internal tissues, we examined the morphogenesis of the CNS during metamorphosis in let-7-C^KO1/GKI mutants and found that at a gross level, CNS development appeared to have proceeded normally (data not shown). To examine the results of nervous system remodeling in finer detail, we focused on the morphology of motoneurons that innervate the DIOMs or the DMs. DIOMs are innervated by DIOM motoneurons, which also degenerate after eclosion. The DIOMs and their DIOM motoneurons, however, are triggered to die at different times and therefore may be controlled by independent signals (Kimura and Truman 1990). Interestingly, we found that the neuromuscular junctions (NMJs) connecting DIOMs and their innervating motoneurons failed to decay in let-7-C^KO1/GKI mutant adults (Fig. 4 E,F), indicating that the DIOMs and DIOM motoneurons persisted together. These data suggested that let-7-C functioned to coordinate the fates of DIOMs and DIOM motoneurons. Similarly, the reduced size of let-7-C^KO1/GKI mutant DMs was reflected in clear defects in let-7-C^KO1/GKI mutant DM NMJs, which were either completely absent, shorter in length than wild-type NMJs, or devoid of boutons, appearing as long, thin processes along the length of the DM (Fig. 4C,D). To quantify this phenotype, we measured the length of DM NMJs containing boutons in the same set of approximately six to eight DM muscles of wild-type (n = 7), let-7-C^KO1/GKI (n = 7), and rescued let-7-C^KO1/GKI (n = 6) flies; wild-type NMJs were 55 ± 7.8 μm in length, let-7-C^KO1/GKI mutant NMJs were 11.2 ± 1.6 μm in length, and rescued let-7-C^KO1/GKI mutant NMJs were 60.5 ± 6.4 μm in length. We concluded that the heterochronic abdominal muscle defect was reflected in a corollary nervous system defect, supporting the hypothesis that disruption of neuromusculature remodeling could underlie at least some of the let-7-C^KO1/GKI mutant behavioral phenotypes.

We note the striking similarity between let-7-C^KO1/GKI mutant phenotypes and the phenotypes associated with manual denervation of abdominal muscles prior to metamorphosis, reported by Currie and Bate in 1995. In both cases, adult DM muscles fail to grow to wild-type width, contain fewer nuclei, and display aberrant NMJs. However, the let-7-C mutation and denervation differ in at least one respect: their effect on the male-specific muscle of Lawrence (MOL) (Lawrence and Johnston 1986). MOLs are present in let-7-C^KO1/GKI adult males but absent in manually denervated adult males (Supplemental Fig. S5; Currie and Bate 1995). Interestingly, Currie and Bate did not report the persistence of DIOMs in denervated adults, which could mean either that DIOM degeneration is unaffected by denervation or that DIOMs degenerate precociously when denervated and were therefore not observed. In either case, the overall similarity between the effects of genetic depletion of let-7-C and muscle denervation during metamorphosis supports the hypothesis that let-7-C is required to regulate an interaction between muscles and motoneurons during neuromusculature remodeling.

To test whether the activities of miR-100, let-7, or miR-125 are required individually for neuromusculature remodeling, we examined the abdominal muscle pattern as well as DM NMJs in miR-100^Δ, let-7^Δ, and miR-125^Δ single mutants (Fig. 5). We found that 2-d-old miR-100^Δ (n = 6) and miR-125^Δ (n = 7) males retained none of the six DIOMs, while let-7^Δ males (n = 7) retained 61% ± 28.5% of DIOMs. Although the frequency of complete DIOM retention is lower in let-7^Δ mutants compared to let-7-C^KO1/GKI mutants, we noted that 83% ± 25% of let-7^Δ mutant DIOMs had arrested at some stage in the process of degeneration. With respect to both the DM and DM NMJ phenotype, we similarly found that miR-100^Δ (n = 5) and miR-125^Δ (n = 6) mutants appeared normal, whereas let-7^Δ mutants (n = 6) phenocopied let-7-C^KO1/GKI mutants. miR-100^Δ and miR-125^Δ DMs were 18.3 ± 0.8 μm and 16.8 ± 0.8 μm in width, respectively, while let-7^Δ DMs were 12 ± 1.8 μm in length (Fig. 5A–C). Similarly, miR-100^Δ and miR-125^Δ NMJs were 45.9 ± 10.2 μm and 55.1 ± 13.3 μm in length, respectively, while let-7^Δ NMJs were 12.5 ± 5 μm in length (Fig. 5D,E). For the sake of consistency, all the morphological data quantified in this study were collected from adult males. However, let-7-C^KO1/GKI and let-7^Δ mutant females exhibited DM and DIOM phenotypes identical to their male siblings (data not shown), suggesting that the reduced egg-laying displayed by let-7^Δ mutant females (Fig. 2F) might be a consequence of defects in their abdominal neuromusculature. From these data, we concluded that the activity of let-7 alone was predominantly responsible for let-7-C-dependent remodeling of the abdominal neuromusculature, and therefore that a heterochronic let-7 role in regulating developmental transitions had been evolutionarily conserved from worms to flies.