Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Jun 23.
Published in final edited form as: Dev Cell. 2014 Jun 5;29(6):635–648. doi: 10.1016/j.devcel.2014.04.023

Homeotic function of Drosophila Bithorax-Complex miRNAs mediates fertility by restricting multiple Hox genes and TALE cofactors in the central nervous system

Daniel L Garaulet 1,2, Monica Castellanos 3, Fernando Bejarano 1, Piero Sanfilippo 1,4, David M Tyler 1, Douglas W Allan 3, Ernesto Sánchez-Herrero 2,5, Eric C Lai 1,5
PMCID: PMC4111139  NIHMSID: NIHMS603878  PMID: 24909902

Abstract

The Drosophila Bithorax-Complex (BX-C) Hox cluster contains a bidirectionally-transcribed miRNA locus, and a deletion mutant (∆mir) lays no eggs and is completely sterile. We show these miRNAs are expressed and active in distinct spatial registers along the anterior-posterior axis in the central nervous system. ∆mir larvae derepress a network of direct homeobox gene targets in the posterior ventral nerve cord (VNC), including BX-C genes and their TALE cofactors. These are phenotypically critical targets, since sterility of ∆mir mutants was substantially rescued by heterozygosity of these genes. The posterior VNC contains Ilp7+ oviduct motoneurons, whose innervation and morphology are defective in ∆mir females, and substantially rescued by heterozygosity of ∆mir targets, especially within the BX-C. Collectively, we reveal (1) critical roles for Hox miRNAs that determine segment-specific expression of homeotic genes, which are not masked by transcriptional regulation, and (2) that BX-C miRNAs are essential for neural patterning and reproductive behavior.

Introduction

Hox genes encode homeodomain proteins that confer positional identities along the antero-posterior axis of bilaterians. These sequence-specific DNA-binding proteins activate or repress particular cohorts of transcriptional targets, in concert with homeodomain cofactors of the TALE (three amino acid loop extension) class, to endow unique characteristics to different organs (Mann et al., 2009; Pearson et al., 2005). Hox genes are best-studied in Drosophila and mice, but principles regarding their regulation and function in these model organisms have generally proven broadly conserved.

Hox genes are almost always included in gene complexes, and the Drosophila cluster is split into the Antennapedia complex (ANT-C) and the Bithorax complex (BX-C). The latter contains three homeobox genes, Ultrabithorax (Ubx), abdominal-A (abd-A) and Abdominal-B (Abd-B) (Lewis, 1978; Sánchez-Herrero et al., 1985) (Figure 1A), which specify the posterior thoracic, abdominal and genital segments. In both Drosophila and mouse, the genomic position of genes within the complex correlates with their relative expression domain along the antero-posterior axis, a phenomenon known as colinearity (Lewis, 1978). Thus, in the BX-C, Ubx, abd-A and Abd-B are expressed in progressively more posterior domains that correlate with their 5′→3′ order along the genome.

Figure 1. Organization and expression of Bithorax-Complex miRNAs.

Figure 1

Open in a new tab

(A) Organization of Hox protein-coding and miRNA genes in the Bithorax Complex (BX-C) and the Antp gene of the Antennapedia Complex (ANT-C). Both miRNAs have conserved seed matches to the 3′ UTRs of genomically anterior Hox genes; the relative thickness of the repression lines reflects their targeting capacity. (B) BX-C aberrations used in this study. ∆mir is a 45 nt deletion of the mir-iab-4/8 hairpin (Bender, 2008). Df(3R)P109 deletes Ubx, abd-A and mir-iab-4/8. Df(3R)Fab7[R59] deletes only abd-A and mir-iab-4/8. Dp(3;2)P10 and Dp(3;3)bxd[100] are duplications of Ubx that are transposed elsewhere on chromosome 2 and 3, respectively. The breakpoints of Df(3R)P109 and Df(3R)Fab7[R59] have not been defined molecularly. (C) Dominant small RNAs produced by mir-iab-4/8. (D) PCR genotyping of the BX-C miRNA locus on different BX-C aberrations. (E) Northern blots of embryo RNA using probes antisense to miR-iab-4-5p and miR-iab-8-5p (C) demonstrate that ∆mir and Df(3R)P109 chromosomes do not express either miRNA.

Although Hox genes are deployed in specific spatial domains, their expression overlaps in some settings. However, two mechanisms preserve the major impact of a single Hox gene in a specific region: transcriptional down-regulation (Hafen et al., 1984; Struhl and White, 1985) and phenotypic suppression (Duboule and Morata, 1994; Gonzalez-Reyes and Morata, 1990; Gonzalez-Reyes et al., 1990). Transcriptional down-regulation involves repression of one Hox gene by another Hox gene expressed more posteriorly. Although observed in different tissues, the phenomenon has been most thoroughly studied in the Drosophila embryo. Thus, Ubx is transcribed in the embryonic ventral cord at high levels in the first abdominal segment and shows reduced expression in more posterior segments. However, when the two more posteriorly-expressed genes of the BX-C (abd-A and Abd-B) are mutated, Ubx is strongly expressed all along the ventral cord (Struhl and White, 1985).

The second mechanism, phenotypic suppression, allows a posteriorly-expressed Hox protein to dictate its own pattern of development even in the presence of more anteriorly-expressed Hox proteins. For instance, forced expression of high levels of Ubx in the posterior segments of the embryo do not compromise the formation of structures that are determined by the most posteriorly-expressed gene, Abd-B (Gonzalez-Reyes and Morata, 1990). Although observed in several Hox genes, there are some cases where this hierarchy is mildly broken (Heuer and Kaufman, 1992; Lamka et al., 1992). The molecular basis of phenotypic suppression seems to rely, at least for some genes, on the competition of different Hox proteins to bind DNA from common targets with the aid of TALE cofactors (Noro et al., 2011). Phenotypic suppression is also present within the mouse Hox genes, where it is known as posterior prevalence (Duboule and Morata, 1994), although the term posterior prevalence may also include transcriptional down-regulation in vertebrates.

The discovery of miRNAs within Hox complexes adds further complexity. Within the BX-C, two miRNA hairpins are generated from sense and antisense transcription across the same locus: mir-iab-4 and mir-iab-8 (Bender, 2008; Ronshaugen et al., 2005; Stark et al., 2008; Tyler et al., 2008) (Figure 1A). A precise deletion of mir-iab-4/mir-iab-8, which eliminates all BX-C miRNAs, only slightly alters Ubx expression and does not impact Abd-A or Abd-B in embryos (Bender, 2008; Lemons et al., 2012). Therefore, although miRNAs are inferred to contribute to posterior prevalence in Drosophila and vertebrates (Ronshaugen et al., 2005; Yekta et al., 2008), functional evidence for this is scant. Nevertheless, flies homozygous for the mir-iab-4/8 deletion are completely sterile (Bender, 2008), indicating critical functions for these miRNAs. This is particularly notable given that most miRNA mutants exhibit mild or no detectable phenotypes (Miska et al., 2007; Smibert and Lai, 2008). Although mir-iab-4 and mir-iab-8 are encoded by the same DNA, strand-specific mutant conditions achieved by placing BX-C breakpoint alleles in trans to the miRNA deletion revealed that mir-iab-8 is primarily required for fertility (Bender, 2008). However, the mechanisms that underlie this sterility have not been elucidated.

In this study, we provide evidence for specific, colinear expression of mir-iab-4 and mir-iab-8 within the posterior larval ventral nerve cord (VNC). Elimination of mir-iab-8 derepresses its known targets Ubx and abd-A in the posterior VNC, and we assign both Hox TALE cofactors Extradenticle and Homothorax as new functional targets of the BX-C miRNAs. In contrast to the demonstrated essential role of abd-A and Abd-B in the embryo to down-regulate the expression of anterior Hox genes in the epidermis, we find they lack a substantial role in the larval VNC. Instead, the BX-C miRNAs execute a critical role in the Hox regulatory hierarchy in this setting, since multiple Hox proteins and their cofactors accumulate inappropriately in the posterior VNC regions of BX-C miRNA mutant larvae. These changes in the spatial accumulation of BX-C miRNA targets are responsible for female sterility, since mild reduction in the expression of these genes substantially rescues fertility. Finally, we identify innervation and bouton defects in miRNA-mutant posterior VNC neurons that project to the oviduct. Altogether, we identify essential roles of miRNAs to restrict the spatial domains of Hox genes and cofactors, which determines normal CNS patterning and reproductive capacity.

Results

Characterization of alleles of Bithorax-Complex miRNAs

Welcome Bender extended Ed Lewis’ classic collection of BX-C aberrations (Lewis, 1978) by generating a 45 nt deletion of the mir-iab-4/8 hairpin (hereafter referred to as ∆mir) (Bender, 2008) (Figure 1A, B). Both arms of pre-mir-iab-4 and pre-mir-iab-8 are well-conserved and generate functional miRNAs (Okamura et al., 2008), but the dominantly-cloned species derive from their 5p arms (Figure 1C). Mature miR-iab-4-5p and miR-iab-8-5p are related but have distinct “seed” sequences; thus, most target sites are selective for one or the other miRNA (Tyler et al., 2008).

We characterized several BX-C aberrations with respect to the miRNA locus. The deletion chromosome Df(3R)P109 (Figure 1B) has classically been used as a Ubx/abd-A double mutant (Sánchez-Herrero et al., 1985), but its miRNA status was unknown. PCR tests indicated that Df(3R)P109 lacks mir-iab-4/8, and this aberration did not relocate the miRNA elsewhere (Figure 1D). We used Northern analysis to confirm that neither miR-iab-4 nor miR-iab-8 were expressed in Df(3R)P109/∆miR embryos (Figure 1E). We also analyzed Df(3R)Fab7[R59] (Gyurkovics et al., 1990), an abd-A deletion that removes BX-C sequences from bxd to iab-7 (Figure 1B), and for which expression of Ubx protein is maintained (data not shown). PCR tests demonstrated this chromosome also lacks the miRNA locus (Figure 1D). Therefore, Df(3R)P109 is deficient for Ubx, abd-A and mir-iab-4/8, while Df(3R)Fab7[R59] is deleted for abd-A and mir-iab-4/8, and affects some Ubx regulatory regions but not the Ubx transcription unit (Figure 1B).

Expression of BX-C miRNAs in the embryonic and larval ventral nerve cord

Although ∆mir animals lack patterning defects in embryo or adult cuticles, the phenotypic requirement of BX-C miRNAs was evident in the complete sterility of ∆mir flies (Bender, 2008). To obtain clues as to the mechanism of ∆mir sterility, we characterized the expression of BX-C miRNAs more fully. We initially used in situ hybridization to nascent primary transcripts of the mir-iab-4 and mir-iab-8. Similar to the homeobox genes of Hox clusters, BX-C miRNAs are restricted to specific domains along the anterior-posterior axis, but are transcribed around the dorsal-ventral axis in the blastoderm embryo (Ronshaugen et al., 2005) (Figure 2A). Notably, the expression of BX-C miRNAs follows the colinearity rule. Although the BX-C miRNAs are generated from the same genomic DNA, the promoter of mir-iab-4 is located proximally to that of mir-iab-8, which is transcribed from the opposite strand (Figure 1A). Accordingly, primary transcripts of mir-iab-4 are expressed more anteriorly than those of mir-iab-8 during early embryogenesis (Ronshaugen et al., 2005) (Figure 2A).

Figure 2. Spatial expression and in vivo activity of BX-C miRNAs.

Figure 2

Open in a new tab

(A, B) In situ hybridizations to primary transcripts of mir-iab-4 and mir-iab-8; this technique detects nuclear, nascent transcription. (A) Double staining of these miRNA transcripts at stage 5 (anterior part of the embryo is to the top) shows that iab-4 (red) is expressed throughout the dorsal-ventral axis, anteriorly to iab-8 (green). (B) The segmental registers of these miRNA transcripts is maintained at stage 12, but their expression is now confined to the ventral nerve cord (VNC). Anterior is to the top and dorsal is to the right. (C–E) Transcription of these miRNAs is actively maintained in the third instar larval CNS, in the same relative domains. (F–I) GFP expression in tub-GFP-mir-iab-4-5p (F′, G) and tub-GFP-mir-iab-8-5p (H′, I) transgenes, in the larval ventral cord of wildtype (F′, H′) or ∆mir/∆mir homozygotes (G, I). Lk expression (F, H) marks the extent of repression by miRNAs.

Following germband retraction, the expression of Hox homeobox genes is enhanced in the ventral nerve cord (VNC), and transcription of the BX-C miRNAs is in fact exclusive to the VNC (Bender, 2008; Tyler et al., 2008) (Figure 2B). Notably, we also detected transcription of both miRNA loci in the larval central nervous system (CNS). They maintained their A–P registers, with mir-iab-4 being expressed more anteriorly to mir-iab-8 (Figure 2C–E). Since we detected nascent nuclear transcripts, this reflects ongoing, active transcription in the larval CNS.

Since some miRNA loci are regulated post-transcriptionally, we sought direct evidence for functional miRNAs within these domains of pri-miRNA expression. To do so, we analyzed “sensor” transgenes consisting of ubiquitously expressed tub-GFP linked to miR-iab-4-5p or miR-iab-8-5p binding sites. Analysis of larval CNS revealed striking spatial patterns of GFP accumulation from these sensors, which corresponded with the relative expression domains of the cognate miRNAs (Figure 2F–I). We studied the precise segmental registers of the miRNA sensors by double staining for Leucokinin (Lk), which is expressed by a pair of lateral neurons in abdominal segments A1–7, but not in the terminal segment A8 (Figure 2F, H). Figure 2F′ shows that miR-iab-4-5p suppressed its sensor in A2–A7 segments, but not in A8, similar to its embryonic expression domain. Importantly, we demonstrated that this spatial pattern was directly induced by the miRNA, since tub-GFP-mir-iab-4-5p expressed GFP more uniformly throughout the VNC of ∆mir/∆mir homozygotes (Figure 2G). We also found that endogenous miR-iab-8-5p suppressed its GFP sensor in A8–A9 (Figure 2H′), and this spatial modulation was again fully dependent on the endogenous BX-C miRNA locus (Figure 2I).

In summary, there is ongoing expression and potent regulatory activity of top and bottom strand BX-C miRNAs in the larval VNC, in relative anterior-posterior domains that reflect their conformance to the colinearity principle.

BX-C miRNAs strongly repress the Hox heterodimeric cofactors hth and exd

BX-C miRNAs target the BX-C members Ubx and abd-A (Bender, 2008; Ronshaugen et al., 2005; Stark et al., 2008; Thomsen et al., 2010; Tyler et al., 2008) via unusually large arrays of binding sites (Figure 3A). In fact, both genes contain five canonical, conserved seed matches for miR-iab-8-5p, which is exceptionally rare amongst Drosophila genes (Ruby et al., 2007). To provide a broader foundation for understanding BX-C miRNA function, we tested additional targets using in vivo assays of tub-GFP-target 3′ UTR sensor transgenes.

Figure 3. Extensive targeting of a network of homeobox genes by BX-C miRNAs.

Figure 3

Open in a new tab

(A) The 3′ UTRs of four homeobox genes, two members of the BX-C (Ubx and abd-A) and their dimeric cofactors hth and exd, contain large numbers of seed matches to the mature products of mir-iab-4/8; see inset for site type and quality. Many of these genes are alternatively polyadenylated, with extensions in the CNS. (B) Northern blots using universal and extended hth probes, showing that the distal 3′ UTR isoform accumulates robustly in heads but not bodies. (C–L) Sensor tests. Shown are the central pouch regions of wing imaginal discs expressing a tub-GFP-3′ UTR sensor (as indicated and either UAS-DsRed-mir-iab-4 (C-G) or UAS-DsRed-mir-iab-8 (H-L) under the control of ptc-Gal4. GFP expression is shown in grayscale (C-L) and as a merge (in green) with DsRed (in red) (C′-L′). Sensors contain the 3′ UTRs shown as well as flanking downstream genomic DNA. mir-iab-4 weakly represses Ubx and hth, and strongly represses exd. mir-iab-8 strongly represses Ubx, abd-A, hth and exd. Neither miRNA detectably represses the homeobox gene Antp, even though its 3′ UTR contains conserved sites for these miRNAs. See also Figure S1–S3.

The 3′ UTR of the third BX-C member, Abd-B, lacks obvious target sites, and an Abd-B sensor was not responsive to ectopic mir-iab-4 or mir-iab-8 (Supplementary Figure 1). The 3′ UTR of the ANT-C Hox gene Antennapedia (Antp) contains multiple well-conserved seed matches for BX-C miRNAs (Stark et al., 2008; Tyler et al., 2008); however, the Antp 3′ UTR did not confer detectable repression in imaginal discs (Figure 3C, H). We infer that these sites are biologically meaningful since they are highly conserved; still, these tests placed regulation of Antp on a different quantitative level than that of Ubx and abd-A, which are particularly strongly repressed by mir-iab-8 (Figure 3D, E, I, J). Outside of the Hox clusters, the neural BTB-POZ transcription factor abrupt is responsive to mir-iab-4 (Okamura et al., 2008), and here we found it is also modestly repressed by ectopic mir-iab-8 (Supplementary Figure 1).

Upon evaluating other Hox-related genes, we realized that the TALE homeobox genes homothorax (hth) and extradenticle (exd) also exhibit well-conserved binding sites for these miRNAs (Figure 3A). Hth functions as a homeotic selector in the nervous system that is regulated by Ubx and abd-A, and is a nuclear escort for the homeotic protein cofactor Extradenticle (Exd) (Kurant et al., 1998; Pai et al., 1998; Rieckhof et al., 1997); Exd partners with Hox proteins to increase their binding specificity (Mann et al., 2009). Notably, their target sites were poorly accounted in genome-wide scans of conserved miRNA target predictions (i.e. TargetScan, http://www.targetscan.org), owing to incomplete 3′ UTR annotations and suboptimal genome alignments.

Recently, we reported that hundreds of Drosophila genes encode alternative polyadenylation (APA) variants with dramatic 3′ UTR extensions in the CNS (Smibert et al., 2012), and APA induces differential regulation of Hox targets by BX-C miRNAs in the CNS (Thomsen et al., 2010). The 3′ UTRs of canonical homeodomain-encoding hth transcripts exhibit evidence of transcript extension of 2240 nt past the 3′ end annotated in FlyBase (http://www.flybase.org), yielding a predicted 3.8 kb 3′ UTR that terminates in a highly conserved polyadenlyation signal (Figure 3A, Supplementary Figure 2). To provide direct evidence for a 3′ UTR extension, we performed Northern blots using hth universal and distal extension probes (Figure 3B). The universal probe primarily detected a ~4.3kb band in bodies, but detected an additional ~6.6kb in heads. Indeed, the distal extension probe specifically co-detected this large band in heads, and did not hybridize to body transcripts (Figure 3B). Therefore, hth expresses an unusually long 3′ UTR in the CNS.

Given that we observed complex relationships between sense/antisense and 5p/3p species of BX-C miRNAs and their targets, we catalogued various types of seed matches (8mer, 7mer, 6merA and 6mer) (Lewis et al., 2005). Interestingly, conserved matches to all four BX-C miRNA seeds exist in the hth-RA 3′ UTR, mostly within the unannotated extension (Figure 3A). Notably, only a few other genes in the entire Drosophila genome contain a comparable number of highly-conserved sites for a given miRNA (Ruby et al., 2007; Stark et al., 2005).

We also observed multiple BX-C miRNA target sites in the 3′ UTR of D. melanogaster exd. Since several Drosophilid genomes are broken within the exd-CG8939 interval, the exd 3′ region is not present in Multiz genomewide alignments available at the UCSC genome browser (http://genome.ucsc.edu). Consequently, no conserved miRNA target sites are predicted for exd by TargetScan. Nevertheless, simply by searching for miR-iab-4/8 seeds downstream of the exd stop codon, we identified multiple miR-iab sites in the sequenced Drosophilid species most distant to D. melanogaster; a manual alignment is presented in Supplementary Figure 3. Therefore, regulation of exd by BX-C miRNAs is under substantial constraint.

We observed substantial repression of the D. melanogaster hth-RA and exd 3′ UTRs by mir-iab-4 and mir-iab-8 in luciferase sensor assays in S2 cells, (3). Since exd was not predicted to have any conserved mir-iab-4/8 sites in genomewide scans, we tested a distant ortholog for miRNA regulation. Indeed, the D. mojavensis exd 3′ UTR was also repressed by both mir-iab-4 and mir-iab-8 (Supplementary Figure 3). To provide evidence for regulatory relationships in a more in vivo setting, we tested the capacity of the BX-C miRNAs to repress tub-GFP-hth/exd 3′UTR sensors in imaginal discs. mir-iab-4 was only weakly active on the hth-RA sensor, but this was strongly repressed by mir-iab-8 (Figure 3F, K). We note that miR-iab-8 was also a much stronger repressor of Ubx and abd-A than was miR-iab-4 (Tyler et al., 2008); both miRNAs strongly repressed the exd 3′ UTR (Figure 3G, L). Therefore, both sense and antisense BX-C miRNAs regulate multiple homeotic targets, but mir-iab-8 has generally (excepting exd) been selected for much stronger repressive interactions. These interactions correlate with the fact that mir-iab-8, but not mir-iab-4, is associated with a detectable in vivo mutant phenotype, namely sterility (Bender, 2008).

BX-C miRNAs control segmental expression of homeotic factors in the larval CNS

As many targets of BX-C miRNAs (Antp, Ubx, abd-A, hth, exd and abrupt) are expressed in the larval CNS, we investigated if their expression was affected in this setting in BX-C ∆mir animals. In contrast to subtle effects observed in the embryo, which included mildly ectopic expression of Ubx and abd-A in the embryonic VNC (Bender, 2008; Gummalla et al., 2012) we observed substantial derepression of Ubx within the mir-iab-4 domain in the larval VNC (Figure 4A, D). In addition, Ubx and Abd-A proteins expanded into the most posterior segments of the VNC, overlapping the expression domain of mir-iab-8 (Figure 4D, E, compare with 4A, B). Double labeling with Lk confirmed misexpression of Ubx and Abd-A in A8 (Figure 4D, E), where Abd-B and mir-iab-8 are normally expressed.

Figure 4. Defective segmental expression of multiple homeotic genes in the VNC of ∆mir homozygous animals.

Figure 4

Open in a new tab

Shown are posterior VNCs of third instar larvae, oriented with anterior to the top and posterior to the bottom. (A–F) Expression of Ubx (A, D), Abd-A (B, E) and Hth (C, F) in wildtype (A–C) or ∆mir mutants (D–F). Leucokinin (Lk) marks abdominal segments A1–A7. Yellow brackets in A–F indicate A2–A7 and A8–A9 regions of the VNC, corresponding to the mir-iab-4 and mir-iab-8 domains, respectively. Close-ups of the posterior regions (A8–A9) are shown to the right (A′–F′), showing expression of Ubx, Abd-A and Hth proteins in wildtype and ∆mir mutants. The yellow arrows indicate the Lk expression in A7. Samples were dissected, stained and imaged in parallel. Ubx and Abd-A are excluded from A8 in the wildtype (A, A′, B, B′) but they are present in ∆mir (D, D′, E, E′). Misexpression of Hth is more dramatic, as it is not expressed in the bulk of abdominal segments in wildtype, save for a few cells in A1 and A8/9 (C, C′), but is present throughout the abdominal VNC in ∆mir mutants (F, F′). (G, H) Double staining for Hth and Exd confirm that ectopic Hth in ∆mir VNC is a functional nuclear escort for Exd; TOPRO (blue) stains DNA. A single optical section of the whole VNC is shown in G; thus, only a few cells ectopically expressing these markers are stained. The region boxed in (G) is magnified in (H), and highlights nuclear coexpression of Hth and Exd in A8 of ∆mir. See also Figure S4.

The spatial derepression of Hth was more dramatic. While it is normally absent from the abdominal segments of the VNC, in ∆mir its expression invades throughout the VNC, overlapping both mir-iab-4 and mir-iab-8 domains, as confirmed by double labeling with Lk (Figure 4C, F). This was consistent with the demonstration that both of these miRNAs directly target hth. Moreover, we were able to perform a functional test of the de-repressed Hth by assaying whether ectopic Hth served as an Exd nuclear escort. Although Exd is also directly targeted by BX-C miRNAs, Exd is ubiquitous and excluded from the nucleus in the absence of Hth. We therefore stained for Exd in ∆mir larvae, and observed striking accumulation of nuclear Exd in the Hth-derepressed segments, precisely in nuclei exhibiting ectopic Hth (Figure 4G, H).

Finally, we note that Antp and Abrupt proteins were not derepressed in ∆mir (Supplementary Figure 4). Thus, only a subset of conserved targets of BX-C miRNAs are substantially altered in the mutant, in keeping with the general fact that most miRNA targets are only subtly regulated (Flynt and Lai, 2008). In summary, we define for the first time major defects in the segmental expression of homeotic factors in a Hox miRNA deletion mutant, including strong misexpression of more “anterior” Hox genes in more posterior segments.

abd-A and Abd-B mutations induce little or no derepression of anterior Hox genes or homothorax in the larval VNC

The defects we observed in the BX-C miRNA mutant background violate the expression hierarchy long-believed to be maintained by transcriptional regulatory activity of protein-coding Hox genes (Bender, 2008; Hafen et al., 1984; Karch et al., 1990; Macias et al., 1990; Struhl and White, 1985). Notably, the expression of Abd-B in the posterior VNC was maintained normally in ∆mir (not shown). We therefore compared the contribution of BX-C Hox genes and miRNAs to the normal spatial restriction of BX-C gene expression. We also examined expression of hth, which is similarly down-regulated by BX-C genes in abdominal segments (Kurant et al., 1998).

We first investigated the accumulation of Ubx, Abd-A and Hth proteins under Abd-B mutant conditions. Abd-B codes for two isoforms, Abd-B[M] and Abd-B[R], both of which are detected by the antibody we used (Delorenzi and Bienz, 1990). As the complete absence of Abd-B is lethal prior to the third instar, we investigated Abd-B function using three approaches: (1) Flp-out clones expressing an Abd-B-RNAi construct that eliminates both Abd-B proteins (Figure 5A–C; Supplementary Figure 5), (2) MARCM clones of Abd-BM5 (Supplementary Figure 5), and (3) the Abd-B[M] mutant combination (Abd-BM5/Abd-BD14) (Figure 5D–F). We were particularly interested to study the effect of removing Abd-B in the A8 larval segment, since the contribution of Abd-B and miRNAs in BX-C gene regulation has been studied most closely in this metamere of the embryonic VNC (Bender, 2008; Gummalla et al., 2012; Struhl and White, 1985). Accordingly, we analyzed Abd-BM5/Abd-BD14 mutants, which are null for Abd-B protein within embryonic A5–A8 segments (Delorenzi and Bienz, 1990; Sánchez-Herrero, 1991). We observed a similar elimination of Abd-B within the larval VNC (Figure 5D–F).

Figure 5. abd-A and Abd-B are mostly dispensable for exclusion of Ubx, abd-A and Hth in the larval VNC.

Figure 5

Open in a new tab

(A–C) Expression of Ubx, Abd-A and Hth in Abd-B-RNAi clones in the larval VNC. Clones are marked by absence of GFP (green) and Abd-B (blue). There is no derepression of Ubx (A, in red; insets in A′–A′″ magnify the region marked in A) or Abd-A (B, in red; insets in B′–B′″ magnify the region marked in B) in these mutant clones. Some of clones show variable expression of Hth (C, red). Insets of the region labeled in C (C′–C′″) show a few nuclei expressing Hth (arrows), but most of the clone lacks Hth. (D–F) Expression of Ubx, Abd-A and Hth (all in white) in Abd-BD14/AbdBM5. In these mutants, Abd-B (in red) is absent in segments A5–A8 but remains in A9. Lk (green) is shown to the left of each panel as a marker of segments A1–A7 in wildtype. Note that Abd-B mutants exhibit ectopic Lk in A8 (yellow arrows), indicating transformation of this segment into an anterior one; Abd-B mutants also exhibit extended VNCs. Insets of Ubx (D), Abd-A (E) and Hth (F) expression in A8 are shown below each panel (D′–F′). Note that Abd-A is not derepressed in the absence of Abd-B, but that Ubx and Hth show ectopic expression in some cells of A8 (between the yellow lines). (G) abd-AM1 MARCM clones marked by GFP expression (green). Expression of Ubx (red) in these clones does not differ from the background; the inset (G′–G′″) of the region marked in G shows variable Ubx expression within and outside the two clones. TOPRO (blue) marks nuclei. (H) Expression of Hth in abd-AM1 clones induced in the larval ventral cord, marked by the absence of the GFP marker (green) and of Abd-A (blue). These clones do not derepress Hth (red; insets in H′–H′″ magnify the region marked in H). See also Figure S5, S6.

We recovered Abd-B-RNAi and Abd-BM5 clones in which neither Ubx nor Abd-A were derepressed (Figure 5A, B and Supplementary Figure 5). Similarly, there was no ectopic Abd-A protein in the VNC of Abd-B mutant larvae (Figure 5E), as was the case in the embryo (Gummalla et al., 2012), and only a few cells of A8 exhibited ectopic Ubx in these mutants, albeit with very low signals (Figure 5D). Moreover, Abd-B mutant clones and Abd-B-RNAi clones accumulated Hth protein in only a few cells at most (Figure 5C; Supplementary Figure 5). Hth was derepressed in the posterior VNC of Abd-B transheterozygous mutant larvae, but this was not as intense or consistent as in ∆mir homozygous larvae (Figure 5F, compare with Figure 4F). Comparison of Ubx, Abd-A and Hth expression in the A8 segment of wildtype, ∆mir and Abd-B− larvae demonstrates a substantially higher ectopic expression of these proteins when the larvae lack the miRNAs, relative to the effect of lacking Abd-B (Supplementary Figure 6).

We also examined Ubx and Hth expression in the absence of abd-A. Since all abd-A mutants are embryonic lethal, we analyzed mitotic clones of the null mutant abd-AM1 (Figure 5G, H), as well as clones that express an abd-A-RNAi construct (Supplementary Figure 5). Ubx is widely expressed within the abd-A domain of wildtype larvae (Figure 4A), so determining its potential derepression in abd-A mutant clones was not as straightforward as when eliminating Abd-B. However, in contrast to the expectation that abd-A loss-of-function should result in uniform de-repression of Ubx (Struhl and White, 1985), we observed many abd-A-mutant and knockdown clones that lack Ubx protein, or express Ubx in only a few cells (Figure 5G), indicating that abd-A is at best only mildly required for correct expression of Ubx in the VNC. Moreover, neither type of abd-A mutant clone exhibited any ectopic Hth protein within the abd-A domain (Figure 5H; Supplementary Figure 5).

Taken together, our results unexpectedly show that posterior Hox genes of the BX-C do not (or only weakly) repress more anterior BX-C genes or hth in the larval VNC, in contrast to their established roles in the embryo.

Derepression of its homeotic targets causes female sterility of mir-iab-8 mutants

We wished to establish whether derepression of BX-C miRNA targets was causal to ∆mir sterility. As is the case with most conserved miRNAs, >100 transcripts bear highly conserved seed matches to BX-C miRNAs (e.g. TargetScan predictions, http://www.targetscan.org). It could be that the cumulative de-repression of the large cohort of BX-C miRNA targets, in addition to the ones that we studied, underlies the fertility defect. On the other hand, the phenotypes of some miRNA mutants are due to derepression of individual targets (Flynt and Lai, 2008; Sun and Lai, 2013), for which compelling genetic evidence has come with the observation that certain miRNA mutants are rescued by heterozygosity of specific targets (Dai et al., 2012).

We performed fertility and egg-laying tests using single female crosses. The vast majority (97.5%, n=121) of individual wild-type females were fertile and produced progeny (Figure 6A). In contrast, ∆mir females were 100% sterile and failed to lay any eggs (n=148 flies). We first tested for rescue by introducing abrupt mutations into the miRNA deletion background. Although abrupt is expressed in the VNC, contains multiple highly conserved target sites for BX-C miRNAs, and is repressed by these miRNAs in sensor assays, neither of two abrupt mutants tested restored any fertility to ∆mir (Figure 6A).

Figure 6. Derepression of several homeotic targets causes sterility in ∆mir females.

Figure 6

Open in a new tab

(A) Fertility and egg-laying defects of ∆mir homozygotes are rescued by heterozygosity of homeotic target genes. Substantial rescue was obtained in Ubx abd-A double heterozygotes (in the Df(3R)P109 background). Restoration of a genomic copy of Ubx to the Df(3R)P109 background [provided by Dp(3;3)bxd[100] or Dp(3;2)P10] strongly obviated its ability to rescue ∆mir, whereas heterozygosity of abd-A alone (in Df(3R)Fab7[R59]) did not rescue. Heterozygosity for hth (using the P2 null allele or a deficiency of the locus) provided strong fertility rescue of ∆mir; this was enhanced further in the Df(3R)P109 background. Heterozygosity for exd provided slight rescue of egg-laying but not fertility, while heterozygosity for either of two alleles of abrupt (ab) did not rescue at all. (B) Neuronal knockdown of Ubx using an RNAi transgene (UAS-iUbx) driven by elav-Gal4 provided demonstrable rescue of ∆mir egg-laying and fertility, relative to control ∆mir homozygotes carrying either Gal4 or UAS transgene.

We next tested BX-C targets of these miRNAs, by taking advantage of BX-C deletions that remove the miRNAs and their genomically adjacent target loci (Figure 1B). Importantly, ∆mir/Df(3R)P109 females exhibited demonstrable fertility (20.2%), and an additional 11.8% of sterile flies exhibited rescue of egg-laying (as defined by >3 eggs; n=381). As Df(3R)P109 removes both Ubx and abd-A, we attempted to distinguish their relative contributions to ∆mir sterility. As discussed above, Df(3R)Fab7[R59] lacks the mir-iab-4/8 locus and the abd-A gene (Figure 1B), but maintains Ubx expression. Df(3R)Fab7[R59]/∆mir females were completely sterile and failed to lay eggs (Figure 6A, n=103), indicating that heterozygosity of abd-A alone was insufficient to rescue ∆mir.

We performed a reciprocal test using Dp (3;2)P10 (Figure 1B), a duplication that restores the function of Ubx but not of abd-A (Lewis, 1978). The fertility of Dp(3;2)P10/+; ∆mir/Df(3R)P109 females was reduced by >4.5-fold compared to ∆mir/Df(3R)P109 (Figure 6A, n=90), indicating that the restoration of a genomic copy of Ubx to the Df(3R)P109 background obviates its capacity to suppress ∆mir. We confirmed this using Dp(3;3)bxd[100], a rearrangement that duplicates Ubx (except for some bxd regulatory sequences) (Lewis, 1978). Indeed, ∆mir/Dp(3;3)bxd[100], Df(3R)P109 females exhibited even stronger infertility and egg-laying defects (Figure 6A, n=137). Thus, while abd-A is substantially derepressed in ∆mir, Ubx may contribute more substantially to ∆mir sterility.

We next tested the ability of hth mutations to rescue ∆mir. We initially analyzed the strong loss-of-function allele hth[P2] (Rieckhof et al., 1997). Strikingly, heterozygosity for hth[P2] produced stronger fertility rescues than did heterozygosity of the BX-C, since over half of ∆mir/hth[P2] ∆mir females were fertile (Figure 6A, n=71). We confirmed this result using a chromosomal deficiency that uncovers hth; this aberration also conferred fertility to half of ∆mir females, with an additional ~10% sterile flies exhibiting rescue of egg-laying (Figure 6A, n=82). We also tested exd heterozygotes; this did not rescue fertility but conferred some rescue of egg-laying. Finally, we tested the effect of double heterozygosity of the BX-C and hth, using hth[P2] ∆mir/Df(3R)P109 flies. These females exhibited the most robust rescue (Figure 6A, n= 84), with 2/3 of females being fertile and an additional ~10% of sterile flies now able to lay eggs.

These data explicitly demonstrate that deletion of non-coding RNAs causes failure of the normal transcriptional repression hierarchy (as evidenced by coexpression of Abd-B and other BX-C proteins), as well as of posterior prevalence (because the misexpression of Ubx and other homeotic factors in the posterior VNC has a phenotypic consequence, namely sterility).

Evidence for a neural basis for ∆mir sterility

The misexpression of BX-C miRNA targets was manifest in the VNC, but this did not rule out a spatially broader contribution of target deregulation. We therefore sought a neural basis for the sterility phenotype. We first asked if neural knockdown of any target genes could rescue ∆mir. We combined RNAi transgenes against abd-A, hth, and exd with the pan-neuronal driver elav-Gal4 into the ∆mir background. Unfortunately, all of these genotypes were inviable, precluding assessments of female fertility. We attempted to restrict RNAi transgene expression by using tub-Gal80ts to temporally restrict elav-Gal4 driven UAS-RNAi constructs in the ∆mir background. We shifted flies from restrictive (17°C) to permissive (29°C) temperature at different timepoints, but all regimens tested were lethal except for shifts to 29°C in late pupal stages, for which all genotypes maintained sterility. These tests did not distinguish if the knockdowns were insufficient, or involved an earlier phenocritical period. However, ∆mir animals carrying elav-Gal4>UAS-Ubx[RNAi] transgenes were viable, which allowed us to assess its potential modification of ∆mir sterility. Neuronal knockdown of Ubx yielded rescue of egg-laying in 16.7% of sterile ∆mir females, and an additional 7.1% were actually fertile (Figure 6B, n=127). As elav-Gal4 is expressed postmitotically, this implied a post-developmental role of Hox miRNAs in neurons to regulate fertility.

With this in mind, we sought a physical basis for ∆mir fertility defects. Bender described that ∆mir mutants have normal ovaries and can mate (Bender, 2008), indicating that failure of egg-laying is not due to the absence of eggs. We confirmed that ∆mir females exhibit normal ovary morphology (data not shown). We also obtained evidence that ∆mir females can be fertilized by males bearing Don Juan-GFP, which allow direct visualization of fluorescent transferred sperm (not shown).

On the other hand, ∆mir females exhibited defective passage of eggs from the ovary through the reproductive tract. This led us to search for potential defects in the innervation of the genitalia. We attempted to trace the total innervation of the uterus using pan-neural elav-Gal4 to drive UAS-CD8-GFP; however, this revealed a dense axon patterning that precluded quantitative analysis (data not shown). Moreover, the identity and function of neurons that innervate the uterus are currently poorly-defined. We subsequently focused on the oviduct, whose innervation is simpler and functionally understood. The oviduct is innervated by two efferent subsets required for oviduct function and egg-laying; a dedicated excitatory motoneuron subset that expresses insulin-like peptide 7 (Ilp7) that terminates on the radial muscles of the oviduct (Castellanos et al., 2013; Yang et al., 2008), and an inhibitory neuromodulator subset of octopaminergic neurons that ramify over the oviduct muscle and inner epithelial lining, as well as along other reproductive tract structures (Rodriguez-Valentin et al., 2006). Both of these neuronal populations are located in the posterior VNC, where target gene expression is altered the most in ∆mir.

We quantified the numbers of adult Ilp7-motoneurons (Ilp7-Gal4, UAS-CD8::GFP activity) and octopaminergic neurons (tyrosine decarboxylase 2 (TDC2)-Gal4, UAS-CD8::GFP activity). We observed no change in the number of Ilp7+ neurons in ∆mir homozygotes, and only a mild (although statistically significant) reduction in TDC2+ neurons (Supplementary Figure 7). These neurons retained their transmitter identity; ∆mir Ilp7-motoneurons expressed Ilp7 protein and the vesicular glutamate transporter (VGluT) (Castellanos et al., 2013) (Supplementary Figure 7) and ∆mir octopaminergic neurons expressed the octopamine biosynthetic enzyme tyramine β-hydroxylase (TβH, not shown).

We next examined the capacity of these neuronal populations to innervate the oviduct. To do so, we confocal imaged Ilp7 axons (Ilp7-GAL4, UAS-CD8::GFP) and octopaminergic axons (TDC2-GAL4, UAS-CD8::GFP) within the anterior 210μm of the common oviduct (the oviduct is evenly innervated through its 450–500μm length in all genotypes examined herein), and quantified the extent of innervation using Simple Neurite Tracer (Longair et al., 2011). Octopaminergic neurons showed no change in innervation (Supplementary Figure 7). However, total Ilp7 innervation was reduced by 50% in ∆mir (Figure 7A, B and quantified in Figure 7K), indicating a specific defect in this neuronal population. This was paralleled by a 50% reduction of anti-Discs large (Dlg)-labeled neuromuscular junction (NMJ) boutons made by Ilp7-motoneurons onto oviduct muscle (Castellanos et al., 2013), in ∆mir compared to controls (Figure 7F, G and quantified in Figure 7L). Therefore, motoneuron innervation of the oviduct is compromised in ∆mir females. Despite this morphological defect, ∆mir oviduct neuromuscular junctions (NMJs) still clustered post-synaptic glutamate receptors (GluR), GluRIIB and GluRIIC (Supplementary Figure 7).

Figure 7. Reduced innervation of oviduct by Ilp7-motoneurons in ∆mir.

Figure 7

Open in a new tab

(A–J) Innervation by Ilp7-motoneurons and NMJ numbers on the oviduct are reduced in ∆mir, rescued by heterozygosity of its targets, and phenocopied by ectopic Ubx. (A–E) Whole reproductive tract innervation by Ilp7-Gal4>UAS-CD8-GFP neurons in the indicated genotypes. (A′–E) 3-D renderings of Ilp7 innervation used to quantify total Ilp7 neuronal branch lengths on the common oviduct. Ov, ovary; LO, lateral oviduct; CO, common oviduct, Ut, uterus. (F–J) Maximum sum projection through the common oviduct showing Dlg-labeled Ilp7-motoneuron NMJs. (K) Quantification of total innervation by Ilp7-axonal arbors shows a 50% reduction in ∆mir homozygotes. This was not altered in hth[P2], ∆mir/∆mir, but was significantly rescued in Df(3R)P109/∆mir Misexpression of Ubx also induced loss of Ilp7 oviduct innervation. (L) Quantification of discrete Dlg+ synaptic contacts made between Ilp7-neurons and the oviduct muscle. NMJ reduction in ∆mir homozygotes was restored in Df(3R)P109/∆mir. The hth[P2], ∆mir/∆mir oviducts exhibited partial restoration of NMJ bouton numbers. Misexpression of Ubx also induced loss of Dlg+ contacts. Quantifications are shown ±S.E.M.; Tukey tests: **, p<0.01, ***, p>0.001, N.S.=not significant. (M) Directed misexpression of BX-C miRNA targets in Ilp7+ cells induced marginal sterility, but misexpression of Ubx and Hth using fru-Gal4 induces high frequency sterility. (N) fru-Gal4 is active in Dlg+ Ilp7 neurons that innervate the oviduct (top panels), but further labels a large set of neurons that innervate the uterus. ov, oviduct; sp, spermathecae, sr, seminal receptacle, ut, uterus. See also Figure S7.

Contribution of Ubx and hth to defective oviduct innervation in ∆mir females

In light of these neural phenotypes, we assessed whether the rescues of ∆mir infertility obtained in Df(3R)P109 and hth[P2] heterozygotes were associated with any alterations of Ilp7-neuron innervation or NMJ number. Heterozygosity for hth[P2] did not improve Ilp7 innervation in ∆mir homozygotes (Figure 7C, K), but partially rescued NMJ bouton numbers (Figure 7H, L). More notably, fertile Df(3R)P109/∆mir females exhibited restoration of oviduct innervation by Ilp7 neurons (Figure 7D, K) and Dlg-labeled NMJs (Figure 7I, L). These data directly connect the derepression of these specific BX-C miRNA targets to functional defects in a specific population of posterior VNC neurons that are essential for oviposition and fertility.

In reciprocal tests, we attempted to phenocopy ∆mir mutants by forcing the expression of Ubx or hth in Ilp7 neurons. Heterozygous ∆mir females that carried Ilp7-Gal4 and UAS-Ubx transgenes exhibited strongly decreased oviduct innervation by Ilp7+ neurons and of Dlg+ boutons (Figure 7E, J, quantified in K, L), similar to the effect observed in ∆mir homozygotes; this was accompanied by ~12% sterility (Figure 7M). On the other hand, ectopic Hth did not substantially affect oviduct innervation and only weakly induced sterility (5%). These findings are consistent with the rescue of Ilp7-motoneuron morphology in Df(3R)P109/∆mir that reduces Ubx, but not in hth[P2], ∆mir/∆mir. These data also support the notion that Ilp7-motoneurons comprise a subset of a larger neuronal population whose function is disrupted in ∆mir and that contribute combinatorially to the sterility phenotype. We attempted to test if more general neural misexpression of these Hox/homeotic targets could affect fertility. However, all combinations of UAS-Ubx, -abd-A, -hth, or -exd crossed to elav-Gal4 were inviable, attesting to adverse consequences of broadly ectopic Hox genes and their cofactors in neurons.

We subsequently assessed fru-Gal4, which directs restricted CNS expression to neurons with sexually dimorphic function (Stockinger et al., 2005). Within the posterior nerve cord, its expression domain includes Ilp7 oviduct motoneurons (Castellanos et al., 2013), within a larger set of mostly undefined Fru+ posterior nerve cord interneurons and efferents (Stockinger et al., 2005). Within the reproductive tract, fru-Gal4>UAS-CD8-GFP identified extensive innervation of Fru+ neurons on the uterus, and on the oviduct (Figure 7N). The identity of the Fru+ neurons innervating the uterus is unknown, but we confirmed that Fru+ innervation of the oviduct only comes from Ilp7-motoneurons, in part shown by the perfect congruence of Fru+ innervation and of Dlg+ staining that only marks Ilp7-motoneuron NMJs (Figure 7N) (Castellanos et al., 2013; Kapelnikov et al., 2008).

Interestingly, 22% of fru-Gal4>UAS-hth females (n=50) and nearly 90% of fru-Gal4>UAS-Ubx females (n=50) were sterile (Figure 7M). This was not merely the consequence of expressing a homeotic factor in fru-Gal4 neurons, since ectopic abd-A induced only minor effects on sterility and oviposition by comparison (Figure 7M). Currently, specific Fru+ neurons in the posterior nerve cord that impact female fertility and egg-laying, aside from Ilp7 neurons, are not defined. However, our observations motivate future efforts to identify such neurons and dissect their wildtype functions, and potentially examine how they are affected in the miRNA mutant.

Discussion

Essential, not fail-safe, function of Hox miRNAs in the Hox regulatory hierarchy

Although Hox miRNAs are documented to target other Hox genes, such regulation was suggested to be mostly secondary to transcriptional mechanisms (Bender, 2008; Hornstein et al., 2005; Lemons et al., 2012). The present work documents broadly ectopic expression of multiple homeotic factors in ∆mir nerve cords. In particular, the terminal VNC segments, which express the most posterior Hox protein Abd-B, misexpress multiple anterior Hox genes in ∆mir. Moreover, their deregulation is causal to ∆mir sterility. These data explicitly document that lack of Hox miRNAs causes failure not only of the Hox transcriptional hierarchy, but also of the posterior prevalence rule.

Consideration of the literature alongside our present work identifies an unexpected temporal transition in the mechanisms involved in the Hox regulatory hierarchy. Studies of a P-element insertion (HCJ199) that specifically disrupts Abd-B (Bender and Hudson, 2000) showed that loss of Abd-B derepresses Ubx in A8 of the embryonic VNC into a pattern similar to that of anterior abdominal segments (Bender, 2008). By contrast, embryos homozygous for ∆mir show only very weak Ubx derepression (Bender, 2008). Our results in the larval VNC show the opposite effect: in Abd-B mutants there is weak ectopic Ubx expression in the larval VNC, whereas ∆mir mutants exhibit substantial Ubx derepression. Therefore, there is a temporal switch from Abd-B to mir-iab-8 as the main Ubx repressor in the posterior ventral cord of embryos and larvae, respectively. The situation is slightly different for abd-A, since neither Abd-B nor BX-C miRNAs alone repress this gene in the posterior embryonic VNC (Gummalla et al., 2012), whereas we now show that ∆mir mutants exhibit substantial ectopic expression of Abd-A protein in the larval VNC.

BX-C miRNAs are critical for CNS patterning and reproductive behavior

The essential functions of BX-C miRNAs in neural segmental patterning and organismal fertility are striking, given that most miRNA mutants seem to have subtle phenotypes. Moreover, even though BX-C miRNAs are typical in having hundreds of conserved targets, mere heterozygosity of specific critical targets could rescue female egg-laying and fertility defects. The critical BX-C miRNA targets form protein complexes (i.e., Abd-A/Hth/Exd and Ubx/Hth/Exd) (Slattery et al., 2011), such that their coordinate deregulation may amplify the effects of miRNA loss.

The CNS region most profoundly affected by BX-C miRNAs comprises the posterior VNC, including segments A8–A9 that are the domain of mir-iab-8. Ilp7+ motoneurons in these segments are functionally aberrant in ∆mir mutants, since they exhibited substantially reduced oviduct innervation and bouton numbers. This was at least partly due to derepression of specific BX-C miRNA targets, since these defects were partially rescued by heterozygosity for hth and more substantially rescued by heterozygosity for a deficiency of Ubx and abd-A. However, as homeotic genes are derepressed broadly throughout ∆mir segments, we expect other neurons may potentially be affected. In fact, whereas the forced expression of Ubx or hth results in a modest percentage of sterile females, these numbers raise dramatically when they are misexpressed in the Fru+ domain, encompassing a broader set of neurons that includes populations that innervate the uterus. These data imply that other populations of Fru+ posterior VNC neurons contribute to the egg-laying program and thus to ∆mir sterility. It will be prerequisite in the future to first refine the identity and functional properties of such neurons in wildtype, before we can understand how they may be affected by loss of BX-C miRNAs.

Finally, we highlight the importance of genetics in deciphering miRNA functions. We not only identified phenotypically critical targets of the BX-C miRNAs, we also distinguished their relative dose-sensitivity for the ∆mir phenotype. For key targets such as hth and exd, their target sites were masked by incomplete 3′ UTR annotations and poor genome-wide alignments, yet we identified them due to their functional relationship to Hox genes. Moreover, while expression profiling indicated that BX-C miRNAs are expressed in segmental patterns in the early embryo, this proved not to be the location of their most substantial function. Instead, it was necessary to chase their functions inside the CNS, a location that in retrospect was hinted at by the behavioral phenotypes of ∆mir flies. As with the first recognized miRNAs, then, evidence from genetics and mutant phenotypes can be crucial for interpreting the biological roles of miRNAs.

Methods Summary

Drosophila genetics

The deletion of the BX-C miRNA locus (∆mir) was described (Bender, 2008). We generated new sensor transgenes by amplifying 3′ UTR segments of exd and hth-RA and inserting into tub-GFP, followed by P-element transgenesis using ∆2–3 helper transposase. A detailed description of other published mutants, transgenes, and stocks for generating clones, and primary antibodies used for immunostaining experiments are provided in the Supplemental methods.

In situ hybridization to primary miRNA transcripts

We tested a panel of iab-4 and iab-8 probes (see Methods for sequences) and combined the strongest probes in a cocktail. Fluorescent secondary detection was performed using preadsorbed peroxidase-conjugated Fab fragments (Roche). Tyramide reactions were performed using Perkin Elmer tyramide reagents. Biotin was detected first; then antibodies were stripped off using PBS-Tween + 0.1 M HCl for 10 min, followed by detection of digoxigenin.

Molecular methods

We genotyped BX-C miRNA alleles using PCR primers that amplify a 246 bp product in wildtype and a 223 bp product from the ∆mir allele. We used published methods for small RNA Northerns (Tyler et al., 2008), agarose Northerns (Smibert et al., 2012), and luciferase sensor assays in S2 cells (Okamura et al., 2007). Additional details on probes and primer sets are listed in the Supplemental methods.

Fertility assays

Newly eclosed females were mated individually with three Canton-S males in a single vial for three days at 25 °C, and then checked for the number and viability of eggs. Flies laying viable eggs were considered fully fertile. Those flies unable to lay any egg were considered sterile and egg-laying defective. Flies that produced at least 3 inviable eggs were considered sterile, but capable of egg-laying.

Immunostaining and image analysis

Ventral nerve cords and oviducts from larvae or mated females were dissected in cold 1XPBS and fixed in 4% PFA with 0.1% Triton. Images showing target derepression in third instar larvae VNCs correspond to Z-projections of the ventral half of nerve cords, except those images used in clonal analysis or when specified. To analyze oviduct innervation, we imaged a 210μm span of the oviduct posterior of the point of lateral oviduct fusion to the common oviduct. Total neuronal arbor length within this region, expressing ilp7-GAL4, UAS-mCD8::GFP or TDC2-GAL4, UAS-mCD8::GFP, were quantified using the Simple Neurite Tracer plugin of FIJI (Longair et al 2011). In this software, we 3-D rendered oviduct innervation and report the total length of all arbors in contact with the oviduct. Dlg bouton number: Total number of Dlg-positive boutons was counted in a 200μm span of the common oviduct using mouse anti-Dlg (1:50). A maximum-intensity Z-stack projection was generated in FIJI (ImageJ) and selections were created for each individual bouton in this image. Total number of boutons was quantified using the Analyze Particles plugin of FIJI; boutons were defined as any selection that measured between 2 to infinity mm with a circularity of 0 to 1.

Statistical analysis

Image data were subjected to D’Agostino and Pearson as well as a Shapiro-Wilk Normality tests. Normally distributed data sets were compared using a parametric unpaired t-test and non-normally distributed groups were compared using a non-parametric Mann Whitney test. One-way ANOVA and Tukey’s tests were conducted for multiple comparisons. All statistical analysis and graph data were performed using GraphPad Prism 5. Data were presented as mean±SD.

Supplementary Material

01
02

Highlights.

  • Deletion of the Drosophila Hox miRNA (mir-iab-4/8) locus causes complete sterility.

  • Key targets of the Hox miRNAs include other Hox genes and their TALE cofactors.

  • Hox miRNAs mediate correct segmental expression of homeotic genes in the CNS.

  • Lack of Hox miRNA control impacts the innervation of oviduct motoneurons.

Acknowledgments

We thank Raquel Martin, Hong Duan, Salvador C. Herrera and Paloma Martín for technical support. We are grateful to N. Azpiazu, Welcome Bender, Aaron DiAntonio, Ernst Hafen, Yuh Nung Jan, Richard Mann, Maria Monastirioti, Michael Gordon, the Vienna Drosophila RNAi Center and the Bloomington Stock Center for fly stocks and antibodies, and Francois Karch for discussion. Work in D.W.A.’s lab was supported by the EJLB Foundation and the Canadian Institutes of Health Research. Work in E.S.’s laboratory was supported by grants from the Spanish Ministerio de Economía y Competitividad (BFU 2008-00632, BFU2011-26075, CSD2007-00008) and an institutional grant from the Fundación Ramón Areces. Work in E.C.L.’s group was supported by the Burroughs Wellcome Fund and the National Institutes of Health (R01-NS074037 and R01-NS083833).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Bender W. MicroRNAs in the Drosophila bithorax complex. Genes & development. 2008;22:14–19. doi: 10.1101/gad.1614208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bender W, Hudson A. P element homing to the Drosophila bithorax complex. Development. 2000;127:3981–3992. doi: 10.1242/dev.127.18.3981. [DOI] [PubMed] [Google Scholar]
  3. Castellanos MC, Tang JC, Allan DW. Female-biased dimorphism underlies a female-specific role for post-embryonic Ilp7 neurons in Drosophila fertility. Development. 2013;140:3915–3926. doi: 10.1242/dev.094714. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Dai Q, Smibert P, Lai EC. Exploiting Drosophila genetics to understand microRNA function and regulation. Current topics in developmental biology. 2012;99:201–235. doi: 10.1016/B978-0-12-387038-4.00008-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Delorenzi M, Bienz M. Expression of Abdominal-B homeoproteins in Drosophila embryos. Development. 1990;108:323–329. doi: 10.1242/dev.108.2.323. [DOI] [PubMed] [Google Scholar]
  6. Duboule D, Morata G. Colinearity and functional hierarchy among genes of the homeotic complexes. Trends Genet. 1994;10:358–364. doi: 10.1016/0168-9525(94)90132-5. [DOI] [PubMed] [Google Scholar]
  7. Flynt AS, Lai EC. Biological principles of microRNA-mediated regulation: shared themes amid diversity. Nature reviews Genetics. 2008;9:831–842. doi: 10.1038/nrg2455. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Gonzalez-Reyes A, Morata G. The developmental effect of overexpressing a Ubx product in Drosophila embryos is dependent on its interactions with other homeotic products. Cell. 1990;61:515–522. doi: 10.1016/0092-8674(90)90533-k. [DOI] [PubMed] [Google Scholar]
  9. Gonzalez-Reyes A, Urquia N, Gehring WJ, Struhl G, Morata G. Are cross-regulatory interactions between homoeotic genes functionally significant? Nature. 1990;344:78–80. doi: 10.1038/344078a0. [DOI] [PubMed] [Google Scholar]
  10. Gummalla M, Maeda RK, Castro Alvarez JJ, Gyurkovics H, Singari S, Edwards KA, Karch F, Bender W. abd-A Regulation by the iab-8 Noncoding RNA. PLoS genetics. 2012;8:e1002720. doi: 10.1371/journal.pgen.1002720. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gyurkovics H, Gausz J, Kummer J, Karch F. A new homeotic mutation in the Drosophila bithorax complex removes a boundary separating two domains of regulation. The EMBO journal. 1990;9:2579–2585. doi: 10.1002/j.1460-2075.1990.tb07439.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Hafen E, Levine M, Gehring WJ. Regulation of Antennapedia transcript distribution by the bithorax complex in Drosophila. Nature. 1984;307:287–289. doi: 10.1038/307287a0. [DOI] [PubMed] [Google Scholar]
  13. Heuer JG, Kaufman TC. Homeotic genes have specific functional roles in the establishment of the Drosophila embryonic peripheral nervous system. Development. 1992;115:35–47. doi: 10.1242/dev.115.1.35. [DOI] [PubMed] [Google Scholar]
  14. Hornstein E, Mansfield JH, Yekta S, Hu JK, Harfe BD, McManus MT, Baskerville S, Bartel DP, Tabin CJ. The microRNA miR-196 acts upstream of Hoxb8 and Shh in limb development. Nature. 2005;438:671–674. doi: 10.1038/nature04138. [DOI] [PubMed] [Google Scholar]
  15. Kapelnikov A, Zelinger E, Gottlieb Y, Rhrissorrakrai K, Gunsalus KC, Heifetz Y. Mating induces an immune response and developmental switch in the Drosophila oviduct. Proceedings of the National Academy of Sciences of the United States of America. 2008;105:13912–13917. doi: 10.1073/pnas.0710997105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Karch F, Bender W, Weiffenbach B. abdA expression in Drosophila embryos. Genes & development. 1990;4:1573–1587. doi: 10.1101/gad.4.9.1573. [DOI] [PubMed] [Google Scholar]
  17. Kurant E, Pai CY, Sharf R, Halachmi N, Sun YH, Salzberg A. Dorsotonals/homothorax, the Drosophila homologue of meis1, interacts with extradenticle in patterning of the embryonic PNS. Development. 1998;125:1037–1048. doi: 10.1242/dev.125.6.1037. [DOI] [PubMed] [Google Scholar]
  18. Lamka ML, Boulet AM, Sakonju S. Ectopic expression of UBX and ABD-B proteins during Drosophila embryogenesis: competition, not a functional hierarchy, explains phenotypic suppression. Development. 1992;116:841–854. doi: 10.1242/dev.116.4.841. [DOI] [PubMed] [Google Scholar]
  19. Lemons D, Pare A, McGinnis W. Three Drosophila Hox complex microRNAs do not have major effects on expression of evolutionarily conserved Hox gene targets during embryogenesis. PloS one. 2012;7:e31365. doi: 10.1371/journal.pone.0031365. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lewis BP, Burge CB, Bartel DP. Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell. 2005;120:15–20. doi: 10.1016/j.cell.2004.12.035. [DOI] [PubMed] [Google Scholar]
  21. Lewis EB. A gene complex controlling segmentation in Drosophila. Nature. 1978;276:565–570. doi: 10.1038/276565a0. [DOI] [PubMed] [Google Scholar]
  22. Longair MH, Baker DA, Armstrong JD. Simple Neurite Tracer: open source software for reconstruction, visualization and analysis of neuronal processes. Bioinformatics. 2011;27:2453–2454. doi: 10.1093/bioinformatics/btr390. [DOI] [PubMed] [Google Scholar]
  23. Macias A, Casanova J, Morata G. Expression and regulation of the abd-A gene of Drosophila. Development. 1990;110:1197–1207. doi: 10.1242/dev.110.4.1197. [DOI] [PubMed] [Google Scholar]
  24. Mann RS, Lelli KM, Joshi R. Hox specificity unique roles for cofactors and collaborators. Current topics in developmental biology. 2009;88:63–101. doi: 10.1016/S0070-2153(09)88003-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Miska EA, Alvarez-Saavedra E, Abbott AL, Lau NC, Hellman AB, McGonagle SM, Bartel DP, Ambros VR, Horvitz HR. Most Caenorhabditis elegans microRNAs Are Individually Not Essential for Development or Viability. PLoS genetics. 2007;3:e215. doi: 10.1371/journal.pgen.0030215. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Noro B, Lelli K, Sun L, Mann RS. Competition for cofactor-dependent DNA binding underlies Hox phenotypic suppression. Genes & development. 2011;25:2327–2332. doi: 10.1101/gad.175539.111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Okamura K, Hagen JW, Duan H, Tyler DM, Lai EC. The mirtron pathway generates microRNA-class regulatory RNAs in Drosophila. Cell. 2007;130:89–100. doi: 10.1016/j.cell.2007.06.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Okamura K, Phillips MD, Tyler DM, Duan H, Chou YT, Lai EC. The regulatory activity of microRNA* species has substantial influence on microRNA and 3′ UTR evolution. Nature structural & molecular biology. 2008;15:354–363. doi: 10.1038/nsmb.1409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Pai CY, Kuo TS, Jaw TJ, Kurant E, Chen CT, Bessarab DA, Salzberg A, Sun YH. The Homothorax homeoprotein activates the nuclear localization of another homeoprotein, extradenticle, and suppresses eye development in Drosophila. Genes & development. 1998;12:435–446. doi: 10.1101/gad.12.3.435. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Pearson JC, Lemons D, McGinnis W. Modulating Hox gene functions during animal body patterning. Nature reviews Genetics. 2005;6:893–904. doi: 10.1038/nrg1726. [DOI] [PubMed] [Google Scholar]
  31. Rieckhof GE, Casares F, Ryoo HD, Abu-Shaar M, Mann RS. Nuclear translocation of extradenticle requires homothorax, which encodes an extradenticle-related homeodomain protein. Cell. 1997;91:171–183. doi: 10.1016/s0092-8674(00)80400-6. [DOI] [PubMed] [Google Scholar]
  32. Rodriguez-Valentin R, Lopez-Gonzalez I, Jorquera R, Labarca P, Zurita M, Reynaud E. Oviduct contraction in Drosophila is modulated by a neural network that is both, octopaminergic and glutamatergic. J Cell Physiol. 2006;209:183–198. doi: 10.1002/jcp.20722. [DOI] [PubMed] [Google Scholar]
  33. Ronshaugen M, Biemar F, Piel J, Levine M, Lai EC. The Drosophila microRNA iab-4 causes a dominant homeotic transformation of halteres to wings. Genes & development. 2005;19:2947–2952. doi: 10.1101/gad.1372505. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Ruby JG, Stark A, Johnston WK, Kellis M, Bartel DP, Lai EC. Evolution, biogenesis, expression, and target predictions of a substantially expanded set of Drosophila microRNAs. Genome research. 2007;17:1850–1864. doi: 10.1101/gr.6597907. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Sánchez-Herrero E. Control of the expression of the bithorax complex genes abdominal-A and abdominal-B by cis-regulatory regions in Drosophila embryos. Development. 1991;111:437–449. doi: 10.1242/dev.111.2.437. [DOI] [PubMed] [Google Scholar]
  36. Sánchez-Herrero E, Vernos I, Marco R, Morata G. Genetic organization of Drosophila bithorax complex. Nature. 1985;313:108–113. doi: 10.1038/313108a0. [DOI] [PubMed] [Google Scholar]
  37. Slattery M, Riley T, Liu P, Abe N, Gomez-Alcala P, Dror I, Zhou T, Rohs R, Honig B, Bussemaker HJ, et al. Cofactor Binding Evokes Latent Differences in DNA Binding Specificity between Hox Proteins. Cell. 2011;147:1270–1282. doi: 10.1016/j.cell.2011.10.053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Smibert P, Lai EC. Lessons from microRNA mutants in worms, flies and mice. Cell cycle. 2008;7:2500–2508. doi: 10.4161/cc.7.16.6454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Smibert P, Miura P, Westholm JO, Shenker S, May G, Duff MO, Zhang D, Eads B, Carlson J, Brown JB, et al. Global patterns of tissue-specific alternative polyadenylation in Drosophila. Cell reports. 2012;1:277–289. doi: 10.1016/j.celrep.2012.01.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Stark A, Brennecke J, Bushati N, Russell RB, Cohen SM. Animal MicroRNAs confer robustness to gene expression and have a significant impact on 3′UTR evolution. Cell. 2005;123:1133–1146. doi: 10.1016/j.cell.2005.11.023. [DOI] [PubMed] [Google Scholar]
  41. Stark A, Bushati N, Jan CH, Kheradpour P, Hodges E, Brennecke J, Bartel DP, Cohen SM, Kellis M. A single Hox locus in Drosophila produces functional microRNAs from opposite DNA strands. Genes & development. 2008;22:8–13. doi: 10.1101/gad.1613108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Stockinger P, Kvitsiani D, Rotkopf S, Tirian L, Dickson BJ. Neural circuitry that governs Drosophila male courtship behavior. Cell. 2005;121:795–807. doi: 10.1016/j.cell.2005.04.026. [DOI] [PubMed] [Google Scholar]
  43. Struhl G, White RA. Regulation of the Ultrabithorax gene of Drosophila by other bithorax complex genes. Cell. 1985;43:507–519. doi: 10.1016/0092-8674(85)90180-1. [DOI] [PubMed] [Google Scholar]
  44. Sun K, Lai EC. Adult-specific functions of animal microRNAs. Nature reviews Genetics. 2013;14:535–548. doi: 10.1038/nrg3471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Thomsen S, Azzam G, Kaschula R, Williams LS, Alonso CR. Developmental RNA processing of 3′UTRs in Hox mRNAs as a context-dependent mechanism modulating visibility to microRNAs. Development. 2010;137:2951–2960. doi: 10.1242/dev.047324. [DOI] [PubMed] [Google Scholar]
  46. Tyler DM, Okamura K, Chung WJ, Hagen JW, Berezikov E, Hannon GJ, Lai EC. Functionally distinct regulatory RNAs generated by bidirectional transcription and processing of microRNA loci. Genes & development. 2008;22:26–36. doi: 10.1101/gad.1615208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Yang CH, Belawat P, Hafen E, Jan LY, Jan YN. Drosophila egg-laying site selection as a system to study simple decision-making processes. Science. 2008;319:1679–1683. doi: 10.1126/science.1151842. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Yekta S, Tabin CJ, Bartel DP. MicroRNAs in the Hox network: an apparent link to posterior prevalence. Nature reviews Genetics. 2008;9:789–796. doi: 10.1038/nrg2400. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS603878-supplement-01.pdf (74.7MB, pdf)
02
NIHMS603878-supplement-02.pdf (138.4KB, pdf)

RESOURCES