Research ArticleNeuroscience

The lin-4 MicroRNA Targets the LIN-14 Transcription Factor to Inhibit Netrin-Mediated Axon Attraction

See allHide authors and affiliations

Science Signaling  12 Jun 2012:
Vol. 5, Issue 228, pp. ra43
DOI: 10.1126/scisignal.2002437

Abstract

miR-125 microRNAs, such as lin-4 in Caenorhabditis elegans, were among the first microRNAs discovered, are phylogenetically conserved, and have been implicated in regulating developmental timing. Here, we showed that loss-of-function mutations in lin-4 microRNA increased axon attraction mediated by the netrin homolog UNC-6. The absence of lin-4 microRNA suppressed the axon guidance defects of anterior ventral microtubule (AVM) neurons caused by loss-of-function mutations in slt-1, which encodes a repulsive guidance cue. Selective expression of lin-4 microRNA in AVM neurons of lin-4–null animals indicated that the effect of lin-4 on AVM axon guidance was cell-autonomous. Promoter reporter analysis suggested that lin-4 was likely expressed strongly in AVM neurons during the developmental time frame that the axons are guided to their targets. In contrast, the lin-4 reporter was barely detectable in anterior lateral microtubule (ALM) neurons, axon guidance of which is insensitive to netrin. In AVM neurons, the transcription factor LIN-14, a target of lin-4 microRNA, stimulated UNC-6–mediated ventral guidance of the AVM axon. LIN-14 promoted attraction of the AVM axon through the UNC-6 receptor UNC-40 [the worm homolog of vertebrate Deleted in Colorectal Cancer (DCC)] and its cofactor MADD-2, which signals through both the UNC-34 (Ena) and the CED-10 (Rac1) downstream pathways. LIN-14 stimulated UNC-6–mediated axon attraction in part by increasing UNC-40 abundance. Our study indicated that lin-4 microRNA reduced the activity of LIN-14 to terminate UNC-6–mediated axon guidance of AVM neurons.

Introduction

During development of the Caenorhabditis elegans nervous system, fate specification of the anterior ventral microtubule (AVM) neurons is a postembryonic event. The AVM cell migrates to its final destination along the anterior and posterior axis and starts sending out its axon at the first larval stage (L1). The axon of the AVM neuron is guided to the ventral midline by two cues, the UNC-6 (a member of the netrin family of secreted guidance cues) attractant, recognized by its receptor UNC-40 [a homolog of the vertebrate netrin receptor Deleted in Colorectal Cancer (DCC)], and the SLT-1 (a member of the slit family of secreted guidance cues) repellent, recognized by the receptors SAX-3 [a homolog of the vertebrate receptor roundabout (Robo)] and EVA-1 (a homolog of the vertebrate C21orf63 proteins identified from the Down syndrome project). (Throughout the text, the vertebrate homologs are provided after the C. elegans genes or proteins.) Upon reaching the ventral midline, the AVM axon moves anteriorly to the nerve ring (1), a neuropil that may be analogous to a vertebrate’s brain. Axonal growth cones are attracted to targets, but must switch their responsiveness upon arrival so that they are no longer sensitive to guidance cues and can proceed with synapse formation. Thus, AVM axons become less responsive to UNC-6 as their development progresses. Such a developmental loss of responsiveness to a guidance cue, in this case UNC-6, could potentially result from either extrinsic inhibitory signals or intrinsic neuronal aging signals.

MicroRNAs were first described in C. elegans (24). Since that time, these small, noncoding RNAs have been implicated in such diverse processes as development, metabolism, apoptosis, cell fate determination, and carcinogenesis. MicroRNAs decrease gene expression by interacting with partially complementary sequences in the 3′ untranslated regions (3′UTRs) of their target genes (2, 3). The C. elegans genome encodes hundreds of microRNAs, about 30% of which are phylogenetically conserved, suggesting ancient functions (5). Many genes that encode microRNAs in C. elegans, including lsy-6 and mir-273 (6, 7), are selectively expressed in specific regions of the nervous system, suggesting that microRNAs may contribute to spatial pattern formation in the nervous system. Other microRNA-encoding genes, including lin-4 (encoding the C. elegans homolog of miR-125) and let-7, are expressed at specific times in neurons during development (810), raising the possibility that these microRNAs may temporally regulate steps in neuronal differentiation. Although microRNAs have been shown to specify left-right neuronal asymmetry in C. elegans (6, 7) and the midbrain-hindbrain boundary in the zebrafish (11), their roles in other aspects of neuronal development remain largely unknown. Here, we used genetic analysis of growth of the AVM axon to the ventral midline in C. elegans to show that the lin-4 microRNA contributes to the regulation of axon guidance during formation of a neuronal circuit. Our results show that lin-4 microRNA acts as a potent and specific negative regulator of UNC-6–mediated attraction without affecting SLT-1–mediated repulsion, and we provide evidence that lin-4 microRNA suppresses production of the transcription factor LIN-14 in AVM neurons to inhibit attraction of their axons to UNC-6. Our data indicate that by suppressing LIN-14 production, lin-4 microRNA decreases signaling through the receptor UNC-40 to reduce responsiveness to UNC-6 in AVM neurons and terminate axon guidance at the appropriate time in development.

Results

A microRNA pathway inhibits ventral guidance of the AVM axon through a mechanism that may involve decreased UNC-6 signaling

The Dicer family of ribonuclease III (RNase III)–related enzymes is responsible for processing most, if not all, precursor microRNAs. The Argonaute protein is a core component of microRNA-induced silencing complex required for the activity of microRNAs. In C. elegans, two nearly identical Argonaute proteins, ALG-1 and ALG-2, copurify with DCR-1 (Dicer) (12). Like Dicer, ALG-1 and ALG-2 are involved in microRNA processing; however, ALG-1 and ALG-2 process only a subgroup of microRNAs (12, 13). To determine whether microRNAs are involved in axon guidance, we examined ventral guidance of the AVM axon in alg-1, alg-2, or dcr-1 mutants. The attractive guidance signaling initiates at the ventral side of the AVM cell body and the repulsive guidance signaling initiates at the dorsal side of the cell body, collectively resulting in ventral projection of the AVM axon (Fig. 1A). Because there are two parallel guidance pathways for the AVM axon, mediated by an attractive cue from the ventral nerve cord (UNC-6) and a repulsive cue from the dorsal body wall muscles (SLT-1) (Fig. 1A), genetic manipulations that potentiate signaling through one guidance pathway could potentially suppress defects caused by loss of the other. Mutations in either the unc-6 or the slt-1 gene result in a 30 to 50% penetrant defect in AVM ventral guidance (Fig. 1, B and C). To evaluate the specific contribution of dcr-1, alg-1, and alg-2 to either the UNC-6 or the SLT-1 guidance pathways, we introduced either of the dcr-1, the alg-1, and the alg-2 mutations into unc-6 or slt-1 mutants and examined AVM axon guidance phenotypes of the double mutants.

Fig. 1

Loss of alg-1 suppresses AVM axon ventral guidance defects in slt-1 but not unc-6 mutants. (A) Schematic drawing highlights two parallel guidance pathways, an attractive cue from the ventral nerve cord (UNC-6) and a repulsive cue from the dorsal body wall muscles (SLT-1), that influence ventral guidance of the AVM axon during development. (B) Schematic diagram of wild-type (WT) and mutant AVM axons at the L2 stage. AVM (red) has a cell body located in the anterior third of the animal, sending its axon ventrally in wild-type animals and laterally in some unc-6 and slt-1 mutants. Each ALM neuron (gray) sends an axon toward the anterior of the animal where the axon stops short of the nose. The dashed rectangle delineates the region around the AVM and ALM cell bodies where axon trajectory initiates. (C) alg-1 mutation specifically suppressed AVM axon ventral guidance defects in worms bearing slt-1, but not those with unc-6, mutations. The alg-1(gk214) and alg-2(ok304) mutants used in this study are null alleles. dcr-1(mg375) is a reduction-of-function allele. Error bars represent standard error of the proportion (SEP). **P < 0.001 by Z test for two proportions.

We found that dcr-1 mutation had no effect on AVM axon ventral guidance, whereas alg-2 mutation affected both UNC-6– and SLT-1–mediated axon guidance (Fig. 1C). The alg-1 mutation suppressed the slt-1 mutant phenotype in AVM axon ventral guidance without affecting the unc-6 mutant phenotype. This is consistent with a model in which the alg-1 mutation specifically enhanced UNC-6–mediated axon attraction, thereby compensating for slt-1 loss of function (Fig. 1C), although it is possible that other attractive forces are enhanced in the alg-1; slt-1 mutants.

The effect of alg-1 mutation on AVM axon guidance may result from loss of lin-4 microRNA in AVM neurons

After the embryonic (E) stage, C. elegans develops through four larval stages (L1 to L4) separated by molts before reaching the young adult (YA) stage. The fate of the AVM neuron is specified soon after the embryonic stage. The AVM cell migrates to its final destination along the anterior and posterior axis and starts sending out its axon at the early L1 stage and ends axon outgrowth at the late L1 stage (Fig. 2A). We postulated that ALG-1 is required for the maturation of one or more microRNAs that inhibit UNC-6–mediated attraction of the AVM axon and that the microRNAs are likely expressed in the first larval stage. We used stem-loop reverse transcription–polymerase chain reaction (RT-PCR), which can discriminate among related microRNAs that differ by as little as one nucleotide (14) (fig. S1A), to globally survey the presence and abundance of mature microRNAs at different stages of C. elegans development. Of the 83 microRNAs surveyed (fig. S1B), including all of those with vertebrate homologs, we identified 4 microRNAs (lin-4, mir-71, mir-83, and mir-90) that appeared abruptly in the first larval stage, at a time coinciding with the time at which the AVM axon is guided to the ventral midline (Fig. 2B). Of these four microRNAs, promoter reporters indicated that only lin-4 and mir-90 were present in AVM neurons (Fig. 2, C and D), and, of these two microRNAs, only maturation of lin-4 microRNA was affected by alg-1 mutation (Fig. 2E). The lin-4 promoter reporter was highly expressed in AVM neurons from the late first larval stage onward (Fig. 2D). Furthermore, expression of the lin-4 promoter reporter was stronger in AVM neurons, which depend on UNC-6 for guidance, than in anterior lateral microtubule (ALM) neurons, which do not depend on UNC-6 (Fig. 2D).

Fig. 2

Analysis of the presence and abundance of microRNAs during C. elegans development. (A) Schematic drawing depicting steps and timing of events in the development of the AVM neuron. The green cell represents AVM; the purple cell represents SDQR. (B) Stem-loop RT-PCR analysis of RNA isolated from populations of staged animals revealed the abrupt appearance of mature lin-4, mir-71, mir-83, and mir-90 microRNAs at the L1 stage of C. elegans development. E, L1, L2, L3, L4, and YA indicate the embryonic, first-larval, second-larval, third-larval, fourth-larval, and young adult stages, respectively. Equal amounts of the RNA preparation from staged animals were used for RT-PCR amplification of mature microRNAs and of actin transcripts. Each stem-loop RT-PCR was carried out in duplicate but only one result was shown. (C) Expression of a mir-90 reporter in neurons of the anterior body of an L3-stage animal. (D) Expression of a lin-4 reporter in neurons in the anterior body of animals at the indicated stages of development. The AVM neurons were labeled with a mec-4 promoter reporter (Pmec-4::rfp) composed of a 1-kb mec-4 promoter driving red fluorescent protein (RFP) expression. The lin-4 reporter (Plin-4::gfp) composed of a 2-kb lin-4 promoter driving GFP expression; the mir-90 reporter (Pmir-90::gfp) composed of a 1.2-kb mir-90 promoter driving GFP expression. The superimposed images identify the GFP-expressing cells as the AVM neurons. Asterisk indicates the ALM neuron; arrow indicates the AVM neuron. Anterior is to the left; dorsal is up. Scale bars, 20 μm. (E) Expression of mature mir-90 and lin-4 microRNAs in wild-type (N2) versus alg-1 mutants. Actin served as a control. Values shown below each lane indicate the amount of lin-4 normalized by actin. A representative of three experiments is shown.

lin-4 microRNA is homologous to vertebrate miR-125a and miR-125b; its sequence differs from theirs only in the central region (Fig. 3A), which is not believed to participate in target mRNA recognition (15). miR-125a and miR-125b, which were first isolated from mouse brain tissue, have been implicated in neuronal development (15). The lin-4 microRNA has also been implicated in regulation of development and is expressed at specific stages of C. elegans development (16). Using a reporter with the lin-4 promoter, we detected limited expression of lin-4 in embryos (Fig. 3B) and abundant lin-4 expression in L1 larvae in many tissues, including muscles, hypodermal cells, and neurons in the head, the tail, the ventral nerve cord, the anterior body, and the mid-body (Fig. 3C). lin-4 microRNA binds to the 3′UTR of the target gene, lin-14, which encodes a transcription factor implicated in specification of the timing and sequence of stage-specific developmental events in C. elegans (2, 3, 17, 18). Indeed, overexpression of lin-4 (fig. S2) or miR-125 (fig. S3) repressed the expression in AVM neurons of a mec-4::gfp::lin-14 3′UTR sensor containing seven complementary lin-4 (mir-125) binding sites.

Fig. 3

Neuronal expression of lin-4. (A) Sequence alignment of C. elegans lin-4 microRNA with Drosophila melanogaster miR-125, Xenopus miR-125, and mouse miR-125. Letters in red highlight identical sequences at the 5′ seed region. (B) Expression of the lin-4::gfp reporter in a late embryonic stage animal. (C) Expression of the lin-4::gfp reporter in a late L1-stage animal; lateral view, whole body. Scale bars, 20 μm.

lin-4 microRNA inhibits UNC-6 signaling during AVM axon guidance

We used a lin-4 loss-of-function mutant [the lin-4(e912) mutation, in which lin-4 gene was deleted] to investigate the role of lin-4 microRNA in AVM axon ventral guidance and found that this mutant recapitulated the effects of the alg-1 mutant on ventral guidance of AVM. Worms with the lin-4(e912) mutation in isolation, like those with only the alg-1 loss-of-function mutation, did not display abnormal ventral guidance of the AVM axon. However, the lin-4(e912) mutation, like the alg-1 mutation (Fig. 1C), enhanced UNC-6–mediated attraction, suppressing the AVM axon guidance defect in slt-1–null mutants but not unc-6–null mutants (Fig. 4, A and B). EVA-1 is a SAX-3 (Robo) co-receptor that binds to SLT-1 and SAX-3 on the cell surface (19), and both EVA-1 and SAX-3 are required for SLT-1–mediated AVM axon repulsion. Whereas most of the effects of EVA-1 depend on SLT-1, signaling through SAX-3 can inhibit UNC-40 signaling in the absence of SLT-1. To determine the specific role of lin-4 in limiting UNC-6 signaling, we tested genetic interactions between lin-4 and eva-1 and between lin-4 and sax-3. The lin-4(e912) mutant significantly suppressed the eva-1–null mutant phenotype as expected from the interaction between EVA-1 and SLT-1 (Fig. 4B). However, the lin-4(e912) mutation did not suppress the sax-3 loss-of-function mutant phenotype (Fig. 4B). These results suggest that the function of lin-4 in UNC-6–mediated AVM guidance depends on sax-3 but not eva-1.

Fig. 4

Enhancing UNC-6–mediated axon attraction by a lin-4 microRNA mutation. (A) Representative images showing specific rescue of AVM axon guidance defects in slt-1 mutants, but not in unc-6 mutants, by a loss-of-function lin-4 microRNA mutation. Anterior is to the left; dorsal is up. Scale bar, 20 μm. (B) Mutation of lin-4 microRNA specifically suppressed AVM axon ventral guidance defects in worms with slt-1 (Slit) or eva-1 mutations but not in those with unc-6 (netrin) or unc-40 (DCC) mutations. Error bars represent SEP. **P < 0.01, comparisons between slt-1 and lin-4; slt-1 and between eva-1 and lin-4; eva-1 are significantly different by Z test for two proportions. (C) Cell-autonomous rescue of lin-4 mutant phenotypes in AVM axon guidance by a mec-4::lin-4 transgene that specifically drives lin-4 expression in AVM neurons. unc-54 and ajm-1 promoters were used to drive lin-4 expression in body wall muscles and hypodermal cells, respectively. Ex stands for extrachromosomal arrays. Asterisks indicate that comparison is significantly different between lin-4; slt-1 with or without mec-4::lin-4 transgene at P < 0.01 by Z test for two proportions.

Expression of lin-4 microRNA in AVM neurons, but not body wall muscles or hypodermal cells, significantly rescued the lin-4–null mutant phenotype [suppressing AVM axon guidance defects in slt-1 mutants] (Fig. 4C), indicating that the effect of the lin-4 mutation on AVM axon guidance was cell-autonomous. The mec-4::lin-4 transgene that specifically drives lin-4 expression in AVM neurons and was used in the rescue experiment also repressed the expression of a mec-4::gfp::lin-14 3′UTR sensor in AVM neurons, indicating that the mec-4::lin-4 transgene produced functional lin-4 microRNA (fig. S2). Because depletion of lin-4 microRNA in AVM neurons enhanced UNC-6–mediated axon attraction, suppressing the AVM axon guidance defect in slt-1 mutants, lin-4 microRNA may inhibit either the duration or the amplitude of UNC-6 signaling in AVM neurons.

The lin-4 loss-of-function mutation enhances PVM neuron specification (increases the number of PVM neurons by 40% compared to wild type) but does not affect AVM specification, as judged by immunocytochemical analysis of the MEC-7 proteins (20). Consistent with these previous findings, we did not observe any effect of lin-4 mutation on AVM cell fate. Here, we also did not see any effect of lin-4 mutation on the initial migration of the AVM cell body, its ventral axon guidance and anterior axon growth, or expression of the AVM markers mec-4 and mir-90 (table S1). Indeed, we detected lin-4 in AVM neurons only after the AVM cell had already migrated to its final position along the anterior and posterior axis and had started sending out its axon (Fig. 2D).

lin-14, a target gene of lin-4 microRNA, contributes to the ventral guidance of AVM axons

Because reporter gene expression indicated that the transcription factor LIN-14 was present in many neurons, including AVM neurons (Fig. 5A), and the lin-14 3′UTR contains multiple lin-4 binding sites (2, 3), we investigated whether lin-14 in AVM neurons was inhibited by lin-4 microRNA. If lin-14 is the pertinent target of lin-4 microRNA, a gain-of-function lin-14 mutant allele (n355gf) lacking all seven lin-4 binding sites (2, 3) should mimic the loss-of-function lin-4 mutant allele in enhancing ventral guidance of the AVM axon. The lin-14(n355gf) mutation overcame AVM axon guidance defects displayed in slt-1 loss-of-function mutants similar to the effect of the lin-4(e912lf) mutation (Figs. 4B and 5B). In contrast, lin-14(n355gf) mutation failed to suppress AVM axon guidance defects displayed in an unc-40 (DCC) loss-of-function mutant, suggesting that the function of LIN-14 depends on UNC-40 (Fig. 5B). Decreasing lin-14 activity with the reduction-of-function allele lin-14(n360) suppressed the lin-4 loss-of-function AVM axon guidance phenotype (Fig. 5C). Indeed, lin-4; lin-14 slt-1 triple mutants were phenotypically indistinguishable from those bearing lin-14 slt-1 double mutations, suggesting that lin-14 acts downstream of lin-4 microRNA in the same genetic pathway (Fig. 5C). Similar to the lin-4 loss-of-function phenotype, overexpression of lin-14 in AVM neurons under the control of the cell type–specific promoter mec-4 resulted in enhanced UNC-6–mediated attraction of AVM axons and suppressed the defects in AVM axon guidance seen with slt-1 mutants (Fig. 5, D and E). This result indicates that LIN-14 likely functions within AVM neurons to regulate their axonal guidance.

Fig. 5

The transcription factor LIN-14 enhances netrin-mediated axon attraction. (A) Representative images showing expression of lin-14 in neurons using a lin-14 reporter (Plin-14::gfp). Left: Late embryonic stage. Red arrowheads indicate ventral cord motor neurons. Right: A late L1–stage animal. Scale bars, 20 μm. SDQR, the sister neuron of AVM. (B) A gain-of-function lin-14 mutant allele (n355gf) recapitulated the lin-4 loss-of-function phenotype in enhancing AVM axon ventral guidance. (C) A reduction-of-function lin-14 mutant allele (n360) suppressed the lin-4 loss-of-function phenotype in AVM axon ventral guidance. (D) Representative images showing specific rescue of AVM axon guidance defects in slt-1 mutants by overexpression of lin-14 in AVM neurons. Arrow indicates AVM and arrowhead points to ALM. Anterior is to the left; dorsal is up. Scale bar, 20 μm. (E) Cell type–specific promoter-driven lin-14 expression in AVM neurons enhances UNC-6–dependent attraction, suppressing ventral guidance defects in slt-1 mutants. In all figures, error bars represent SEP. **P < 0.01 by Z test for two proportions. (F) Model of developmental switch of responsiveness to UNC-6 (netrin) by lin-4 microRNA. Netrin responsiveness in early AVM neurons is mediated by the transcription factor LIN-14. lin-4 microRNA expressed in AVM neurons during the later stages of development inhibits netrin-mediated axon attraction by decreasing lin-14 expression.

To test the effect of lin-4 microRNA on lin-14 expression in AVM neurons, we expressed two sensor constructs in AVM neurons, one consisting of the lin-14 3′UTR fused to a green fluorescent protein (GFP) reporter (Fig. 6A), and the other consisting of a control 3′UTR (that of unc-54), containing no known lin-4 binding sites, fused to an mCherry reporter. We detected a significantly higher GFP signal (normalized to the mCherry signal) in AVM neurons of the lin-4 loss-of-function mutants (Fig. 6, B and C) and a significantly lower signal in AVM neurons overexpressing lin-4 (fig. S2) compared to that in wild-type AVM neurons, at the L4 stage, suggesting that lin-4 microRNA decreased the lin-14 3′UTR sensor expression in AVM neurons. Even though the unc-54 3′UTR does not contain any known lin-4 binding sites, the expression of the control reporter (mCherry::unc-54 3′UTR) was increased in the AVM neurons in the lin-4 mutants and was modestly decreased in response to lin-4 overexpression compared to that in wild-type animals. However, these effects were less than those detected with the lin-14 3′UTR sensor.

Fig. 6

Endogenous lin-4 microRNA binds to sites in the lin-14 3′UTR to decrease expression of a reporter in AVM neurons. (A) Organization of the Pmec-4::gfp::lin-14 3′UTR reporter construct. Seven lin-4 microRNA binding sites were computationally predicted in the lin-14 3′UTR, which was used to create a GFP reporter construct. (B) Representative images showing WT animals or lin-4 mutants expressing reporter constructs fused to either the lin-14 3′UTR or a control (unc-54) 3′UTR. Reporter constructs were injected at 1 ng/μl. (C) Quantification of reporter abundance in AVM neurons. The fluorescence intensity of the Pmec-4::gfp::lin-14 3′UTR reporter was normalized to that of the Pmec-4::mCherry::unc-54 3′UTR reporter. Substantially higher GFP (normalized by mCherry) signal was detected in AVM neurons harboring a mutation in lin-4 gene than in neurons with WT lin-4. Error bars indicate the SEM. **P < 0.001, Student’s t test.

Reporter gene analysis indicated that the expression of lin-4 increased at the L1 stage (Fig. 7A). To understand the dynamic regulation of lin-14 3′UTR activity during AVM axon guidance, we analyzed changes in the expression of the gfp::lin-14 3′UTR sensor in AVM neurons during the L1 stage. Expression of the lin-14 3′UTR sensor decreased in AVM neurons as wild-type animals progressed through L1 (Fig. 7B). In contrast, the expression of the lin-14 3′UTR sensor in AVM neurons was stable or tended to increase slightly as L1 progressed in animals bearing a lin-4 loss-of-function mutation (Fig. 7C). These results suggested that by binding to the lin-14 3′UTR, the lin-4 microRNA reduces the abundance of LIN-14 in AVM neurons as the axons are guided to their final destination.

Fig. 7

Dynamic regulation of lin-14 during AVM axon guidance at the L1 stage. (A) Expression of Plin-14::gfp and Plin-4::gfp reporters in AVM neurons at different times during L1. The Pmec-4::mCherry reporter was used to label AVM and ALM neurons, which are located at different positions in the animal. The merged images identify the GFP-expressing cells as AVM neurons. Closed arrowhead indicates the AVM and the open arrowhead marks the ALM. Anterior is to the left; dorsal is up. Scale bar, 20 μm. (B and C) Expression of the gfp::lin-14 3′UTR reporter in the AVM neurons was assessed every hour during L1, starting at 4 hours after hatching and ending at 12 hours after hatching, in WT animals (B) and lin-4 mutants (C). Eight animals each were measured for WT and lin-4 mutants. Bars represent the average intensity of the gfp::lin-14 3′UTR reporter measured at each hour. The comparison of average fluorescence intensity between WT and lin-4 mutants at 10, 11, or 12 hours is significantly different at **P < 0.01 or ***P < 0.001. P values were calculated with a Student’s t test.

There are other targets of lin-4 microRNA in addition to lin-14. Two lin-4 complementary elements are present in the 3′UTR of hbl-1, which is the C. elegans ortholog of Drosophila hunchback, and lin-4 microRNA mediates the down-regulation of hbl-1 in ventral nerve cord neurons (9, 10). The gene lin-41, encoding a member of a large family of RBCC (ring finger, B box, coiled coil) proteins, contains one lin-4 microRNA binding site in the 3′UTR of the transcript (21). However, overexpressing hbl-1 or lin-41 in AVM neurons failed to recapitulate the lin-4 loss-of-function phenotype of suppressing AVM axon guidance defects in slt-1 mutants, suggesting that hbl-1 and lin-41 are not lin-4 microRNA targets involved in mediating AVM axon ventral guidance (Fig. 5B).

The lin-4 and lin-14 mutations do not affect the timing of AVM axon outgrowth

Defective axon guidance can arise through altered signaling without any change in the timing of differentiation and axon outgrowth or can arise through mutations that cause temporal decoupling of axon outgrowth and axon guidance (22). However, analysis of AVM axon outgrowth in wild-type animals or lin-4 loss-of-function or lin-14 gain-of-function mutants indicated that timing of AVM axon outgrowth was not affected by either lin-4 loss-of-function or lin-14 gain-of-function mutations (fig. S4). Five hours into the L1 stage, the axonal marker (Pmec-4::gfp) was not visible in AVM axons in wild-type animals or worms harboring lin-4 or lin-14 mutations. Seven hours into L1 stage, AVM axon outgrowth (both ventral and anterior projections) was detected in most wild-type worms and those with lin-4 or lin-14 mutations, and, by 10 hours into L1 stage, all strains (wild type and lin-4 and lin-14 mutants) showed 100% AVM axon outgrowth (fig. S4). Thus, the timing of AVM axon outgrowth was not altered by lin-4 or lin-14 mutations, suggesting that the effect of lin-4 and lin-14 mutations on AVM axon guidance is likely not due to the temporal decoupling of axon outgrowth and axon guidance.

LIN-14 promotes UNC-6 attraction through the receptor UNC-40 and its cofactor MADD-2 and a combination of downstream UNC-34 and CED-10 pathways

MADD-2, a cofactor for the receptor UNC-40, is required for UNC-40 signaling in AVM and other neurons (23, 24) (Fig. 8). UNC-40 signaling in AVM is mediated by two parallel downstream pathways, one involving UNC-34, an Enabled homolog [a member of the VASP (vasodilator-stimulated phosphoprotein) family of actin regulators], and the other involving CED-10, a Rac guanosine triphosphatase (25, 26) (Fig. 8). MIG-10, a homolog of the adaptor protein lamellipodin, functions in both UNC-34 and CED-10 pathways in AVM neurons (2729). To determine whether the role of LIN-14 in UNC-6 attraction was mediated by the receptor UNC-40 and its cofactor MADD-2, we investigated whether unc-40 or madd-2 loss-of-function mutations blocked the effects of lin-14 overexpression in the suppression of AVM axon guidance defects in slt-1 loss-of-function mutants. Indeed, lin-14 overexpression failed to suppress AVM guidance defects in worms bearing slt-1; unc-40 double mutations or slt-1; madd-2 double mutations (Fig. 8), suggesting that the effects of LIN-14 depended on UNC-40 and its cofactor MADD-2. Next, we selectively disrupted either of the downstream pathways in a slt-1 mutant background, where only unc-40 guidance signaling is active in AVM neurons, and determined the effects of lin-14 overexpression on the AVM phenotype. Overexpression of lin-14 significantly suppressed AVM defects in worms with double mutations of slt-1 and unc-34, slt-1 and mig-10, or slt-1 and ced-10 (Fig. 8), indicating that these double mutants retained a downstream pathway that was stimulated by LIN-14. Animals overexpressing lin-14 in the presence of slt-1; ced-10; unc-34 and slt-1; mig-10; unc-34 triple mutations were not viable. Therefore, to determine whether the UNC-34 and CED-10 pathways combined to mediate the effects of lin-14 overexpression on the AVM phenotype, we used Pmec-7::FP4::mito transgenes to specifically disrupt unc-34 function only in AVM neurons in the slt-1; ced-10; Ex(Pmec-4::lin-14) and slt-1; mig-10; Ex(Pmec-4::lin-14) strains. FP4-MITO depletes enabled (VASP) proteins from their normal subcellular locations, sequesters them on mitochondria, and neutralizes their function, thus selectively blocking UNC-34 signaling (30). We also used RNA interference (RNAi) to knock down ced-10 in the slt-1; unc-34; Ex(Pmec-4::lin-14) strain. FP4-MITO and ced-10 RNAi reagents generated AVM axon guidance defects that were less severe than those observed in the loss-of-function mutations (26) (Fig. 8), suggesting that these reagents only caused reduction of function effects. We found that lin-14 overexpression significantly suppressed AVM guidance defects in worms bearing the slt-1; mig-10; unc-34 triple mutations but not in those with the slt-1; unc-34; ced-10 triple mutations (Fig. 8), suggesting that whereas MIG-10 is not involved in the promotion of UNC-6–mediated attraction by LIN-14, the effects of LIN-14 are mediated by the receptor UNC-40 and its cofactor MADD-2 through a combination of the downstream UNC-34– and CED-10–dependent pathways.

Fig. 8

LIN-14 functions through UNC-40, MADD-2, UNC-34, and CED-10 to promote UNC-6–mediated ventral guidance of the AVM axon. LIN-14 requires UNC-40, MADD-2, UNC-34, and CED-10 to stimulate UNC-6 signaling. AVM cells were visualized with zdIs5[mec-4::gfp]. **P < 0.001, significantly different from Ex[lin-14](-) controls by Z test for two proportions. ns, not significant.

LIN-14 promotes UNC-6–mediated axon attraction in part by increasing the abundance of UNC-40

Given that MADD-2, UNC-34, and CED-10 all transduce signals downstream of UNC-40, we postulated that unc-40 might represent a direct target of LIN-14. Therefore, we analyzed the effect of LIN-14 on unc-40 promoter activity and UNC-40 abundance in the anterior touch neurons (AVM and ALM). Overexpression of lin-14 did not affect the expression of either a 3.5- or a 5-kb unc-40 promoter::gfp reporter in AVM neurons (fig. S5). Indeed, LIN-14 did not affect the activity of gene reporters driven by the promoters of various components of the unc-40 pathway (madd-2, mig-10, and ced-10) (fig. S5). To determine whether UNC-40 abundance might be regulated by LIN-14 through a mechanism independent of transcription, we generated three transgenic lines that expressed different amounts of UNC-40::GFP fusion proteins in AVM and ALM neurons. In all neurons expressing small amounts of UNC-40::GFP, from the Mos single-copy insertion line, the GFP signal was localized to vesicle-like compartments in the perinuclear region (Fig. 9, A and B). Expression of a medium amount of UNC-40::GFP, from the extrachromosomal transgenic line that produced 1.2-fold more protein relative to the low-expression Mos line, reduced the percentage of neurons showing this restricted UNC-40::GFP perinuclear localization and increased the percentage of neurons with broader UNC-40::GFP distribution (Fig. 9C). All of the AVM and ALM neurons expressing a large amount of UNC-40::GFP, from the extrachromosomal transgenic line that produced 9.5-fold more protein relative to the low-expression Mos line, showed UNC-40::GFP distribution throughout the whole cell (Fig. 9D). Overexpression of lin-14 expanded the distribution of UNC-40::GFP from the perinuclear region to the whole cell in the low and medium unc-40::gfp–expressing lines (Fig. 9, A to C), similar to the effect caused by the expression of a larger amount of UNC-40::GFP (Fig. 9D). These results suggest that lin-14 stimulates UNC-6–mediated axon attraction in part by expanding the distribution of UNC-40 protein or increasing the abundance of UNC-40 protein or both. Consistent with this interpretation, overexpression of unc-40::gfp in AVM neurons using the high-expression line recapitulated the effect of lin-14 overexpression, suppressing ventral guidance defects of the AVM axon in slt-1, but not unc-6, mutants (Fig. 9E).

Fig. 9

Effects of LIN-14 on UNC-40 distribution in anterior touch neurons. (A) Distribution of UNC-40::GFP and LIN-14::mCherry fusion proteins in ALM and AVM neurons. The Mos[Pmec-4::unc-40::gfp] single-copy insertion line was used to analyze the distribution of UNC-40::GFP in ALM and AVM with or without overexpression of LIN-14::mCherry. (B to D) Distribution of UNC-40 in either the perinuclear region alone or throughout the cell depends on the abundance of UNC-40 and that of LIN-14 in anterior touch neurons. (E) The high-expression line was used to overexpress unc-40::gfp in AVM neurons. Overexpression of unc-40::gfp in AVM suppressed ventral guidance defects of the AVM axon in animals harboring slt-1, but not unc-6, mutations. In all figures, error bars represent SEP. **P < 0.01, ***P < 0.001 by Z test for two proportions.

The lin-4 and lin-14 relationship may extend beyond AVM neurons

A GFP reporter under the control of the lin-4 promoter was broadly expressed in C. elegans neurons during development, and the lin-4 promoter reporter expression was broader and generally stronger during the L1 stage than in the embryonic stage (Fig. 3, B and C). The lin-4 promoter reporter was expressed in lateral ganglion neurons (ADL, AWB, AWC, AFD), ventral nerve cord neurons (DA, DB, DD, VD), anterior-body region neurons (SDQR), tail neurons (PQR), touch neurons (AVM, PVM, ALM, PLM), and HSN neurons. A GFP reporter under the control of the lin-14 promoter was also broadly expressed in C. elegans neurons, with strong expression in neurons starting in the late embryonic stage (Fig. 5A). Analysis of the expression patterns of these two reporters revealed overlapping expression of the two genes in several neurons, including AVM, ALM, PVM, PLM, DD, VD, DA, DB, SDQR, HSN, and PQR. This suggests that the lin-4 and lin-14 relationship may extend beyond AVM neurons. Some of these neurons are guided by UNC-40 signaling alone (AVM, PVM, HSN, and ADL), whereas others are guided by both UNC-40 and UNC-5 signaling (SDQR, DD, VD, DA and DB) (1).

Discussion

Observation of defects in asymmetric neuronal development in C. elegans bearing mutation in specific microRNAs, such as lsy-6 microRNA and mir-273 microRNA, and in vivo studies of neuronal morphogenesis in vertebrates harboring microRNA mutations have led to the suggestion that microRNAs may play a role in neural development (6, 7, 11). Local translational regulation of axon guidance molecules at the developing growth cones by microRNAs has been long speculated. Here, we showed that a microRNA implicated in developmental timing represses the expression of a transcription factor and consequently inhibits netrin-mediated axon attraction. We propose that this microRNA signals the neuron to undergo a developmental switch in responsiveness to netrin as growth cones complete pathfinding.

We used genetic analysis of a well-characterized axon guidance event, the growth of the AVM axon to the ventral midline in C. elegans, to uncover a role for lin-4 microRNA in axon guidance. We found that lin-4 microRNA functions as a potent and specific negative regulator of UNC-6 (netrin)–mediated attraction, without affecting SLT-1 (Slit)–mediated repulsion, and provide evidence that lin-4 represses the expression of the gene encoding the LIN-14 transcription factor in AVM neurons to inhibit its attraction to UNC-6. The promotion of UNC-6–mediated attraction by LIN-14 involved the receptor UNC-40 and its cofactor MADD-2 and occurred through a combination of downstream UNC-34 and CED-10 pathways. Our data showed that overexpression of lin-14 altered the cellular distribution of UNC-40, which we interpreted to mean that LIN-14 increased the abundance of UNC-40, resulting in a broader UNC-40 distribution throughout the whole cell. We propose that as growth cones complete pathfinding, LIN-14 abundance decreases through increased production of the lin-4 microRNA and UNC-40 would become restricted to the perinuclear area, where its efficacy in signal transduction could be reduced, leading to decreased UNC-6 signaling. Here, we have identified the mechanisms by which lin-4 regulates UNC-6–mediated attraction of the AVM axon; the identity of the signal that turns on lin-4 expression in AVM neurons, however, remains unknown.

There is an emerging view that microRNAs contribute to specifying the transition of cells from totipotency to commitment. Our results suggest that lin-4 microRNA may signal the transition of AVM neurons from an early phase of differentiation characterized by responsiveness to axon guidance cues to later phase of differentiation characterized by loss of responsiveness to axon guidance cues (Fig. 5F). The mRNA encoding the LIN-14 transcription factor is a lin-4 microRNA target, and LIN-14 may stimulate UNC-6–mediated AVM axon guidance in part by enhancing UNC-40 abundance. Previous studies showed that by preventing lin-14 translation, lin-4 microRNA allows L2 to occur (8). Mutations that decrease lin-4 expression or enhance that of lin-14 delay developmental timing of cells at the L1-L2 transition (2, 3). These findings suggest that lin-4 mutation may prolong the state of AVM neuron differentiation characteristic of the L1 stage, thereby extending the duration of axon guidance signaling in AVM neurons. The fact that lin-4 microRNA acts in the AVM neurons to decrease UNC-6–mediated axon attraction suggests that lin-4 microRNA either decreases the amplitude or restricts the duration of UNC-40 signaling in AVM neurons. Given that developmental mechanisms are frequently conserved across animal phylogeny, the general rules of microRNA regulation of axon guidance in C. elegans are likely to apply to higher organisms.

Materials and Methods

Genetics and strains

C. elegans strains were cultured by means of standard methods (31). All strains were grown at 20°C unless otherwise specified. A strain list appears as table S2.

Transgenic animals

Germline transformation of C. elegans was performed with standard techniques (32). For example, the Pmec-4::lin-14 transgene was injected at 10 ng/μl along with the coinjection marker Podr-1::rfp at 40 ng/μl. Transgenic lines were maintained by following Podr-1::rfp fluorescence. All transgenic lines are high-transmitting extrachromosomal array lines with higher than 80% transmission rate to minimize the mosaic nature of the extrachromosomal arrays.

PCR fusion reaction

mec-4::lin-4 was generated by PCR fusion, in which two separate primary PCR products were fused by a secondary PCR. The template used was genomic DNA.

The PCR primers are as follows: lin-4_G1, CTATCAAGTTATAGAGGGATCATGCTTCCGGCCTGTTCCCTG; lin-4_G1_R, CAGGGAACAGGCCGGAAGCATGATCCCTCTATAACTTGATAG; lin-4_G2, AGATCTGCTCAAACCGTCCTG; and mec-4_P1, CCAAGCTTCAATACAAGCTC.

Constructs

Pmec-4::lin-14 and Pmec-4::hbl-1 were constructed by PCR amplification of a genomic fragment of lin-14 or hbl-1 followed by digestion with Nhe I and Kpn I and ligation to Pmec-4-pSM vector. Pmec-4::gfp::lin-14 3'UTR was constructed by PCR amplification of a 1.8-kb genomic fragment of lin-14 3'UTR followed by digestion with Eco RI and Spe I and ligation to Pmec-4::gfp-pSM vector. Plin-4::gfp was constructed by PCR amplification of a 1.9-kb lin-4 promoter fragment from genomic DNA followed by digestion with Fse I and Xma I and ligation to gfp-pSM vector. Plin-14::gfp was constructed by PCR amplification of a 4-kb lin-14 promoter fragment from genomic DNA followed by digestion with Fse I and Asc I and ligation to gfp-pSM vector. Punc-40::gfp was constructed by PCR amplification of either a 3.5- or a 5-kb unc-40 promoter fragment from genomic DNA followed by digestion with Fse I and Asc I and ligation to gfp-pSM vector. Pmadd-2::gfp was constructed by PCR amplification of a 3-kb madd-2 promoter fragment from genomic DNA followed by digestion with Fse I and Asc I and ligation to gfp-pSM vector. Pced-10::gfp was constructed by PCR amplification of a 3-kb ced-10 promoter fragment from genomic DNA followed by digestion with Fse I and Asc I and ligation to gfp-pSM vector. Pmig-10::gfp was constructed by PCR amplification of a 3.5-kb mig-10 promoter fragment from genomic DNA followed by digestion with Fse I and Asc I and ligation to gfp-pSM vector. For the generation of Pmec-4::unc-40::gfp, a DNA fragment containing the unc-40 complementary DNA (cDNA) was inserted at the 5′ end of gfp in a gfp-pSM vector containing a 1-kb fragment of the mec-4 promoter between Nhe I and Kpn I sites. For MosSCI (Mos-mediated single copy insertion) experiments, Pmec-4::unc-40::gfp sequence was cut out of pSM by use of the Fse I and Spec I restriction sites and was cloned into a pCFJ151 MosSCI insertion vector (33) that had been modified to include an Fse I site in the multiple cloning site (34).

The PCR primers are as follows: lin-14_F_NheI, GCTAGCTAGCTTGTCCTTCTGCCCAATCAAG; lin-14_R_KpnI, CGGGGTACCTCTATTGTGGACCTTGAAGAG; hbl-1_F_NheI, GCTAGCTAGCGAAAAAGGATTAGTGGTCCTG; hbl-1_R_KpnI, GGGGTACCTTATTGGTGTCTGGCTTGGTAC; lin-14 3′UTR_F_EcoRI, CGGAATTCACATCAGTCTCTTCACCCATC; lin-14 3′UTR_R_SpeI, GGACTAGTGTAAGGTTCAGAGATGCATC; lin-4_P_F_FseI, GTCTGGCCGGCCCATGTCCTGCCCATTCCGTAG; lin-4_P_R_XmaI, TCCCCCCGGGCTCATAAACCAACCAAAACTC; mig-10_P_F_FseI, GTCTGGCCGGCCGACAAGCAATCCATCATCATC; mig-10_P_R_AscI, CAGCGGCGCGCCGCATAAAAGAGCACAAATCAG; unc-34_P_F_FseI, GTCTGGCCGGCCAGCCCACCAAAATTACAGTAC; unc-34_P_R_AscI, CAGCGGCGCGCCCTGGCTCAAAAAAGTGCAATC; unc-40_P_F_FseI, CAGCGGCGCGCCCTTCTGTCGAATTATCGCATTC; unc-40_P_R_AscI, GTCTGGCCGGCCCATGGGCATTTATCATCACTG; madd-2_P_F_FseI, GTCTGGCCGGCCTTTCCGTGCGGTTGCTTAGAG; madd-2_P_R_AscI, CAGCGGCGCGCCTTGAGTATTGTCAGGTGGAAG; ced-10_P_F_FseI, GTCTGGCCGGCCACCGTATCCAATGGGAGCTTC; and ced-10_P_R_AscI, CAGCGGCGCGCCGAGCTCCGCGAGCACAAGCAC.

Stem-loop RT-PCR

We quantified mature microRNAs by modifying a microRNA assay developed previously (14). RT reactions contained purified total RNA, 50 nM stem-loop RT primer, 1× RT first-strand buffer, 0.25 mM each deoxynucleotide triphosphates (dNTPs), 10 mM MgCl2, 0.1 M dithiothreitol (DTT), 200 U of SuperScript III reverse transcriptase, and 40 U of RNase inhibitor. The mixture of RNA template, dNTPs, and RT primer was incubated for 5 min at 65°C. The mixture was then placed on ice for at least 1 min before the addition of RT buffer, DTT, MgCl2, SuperScript III, and RNase inhibitor. The reaction was incubated for 50 min at 50°C before heat inactivation at 85°C for 5 min. PCR was conducted with 0.25 μl of RT products as template in 20 μl of PCR for 17 cycles. Information about stem-loop RT primers and PCR primers is available in table S3.

RNAi assay

For RNAi experiments, single-stranded ced-10 RNA was transcribed from T7- and SP6-flanked PCR templates. Single-stranded RNAs were annealed and injected into slt-1; unc-34; Ex(Pmec-4::lin-14) animals.

MosSCI integrations

Mos single-copy integrants were generated by the direct insertion protocol described previously (33). Fifty EG4322 ttTi5605; unc-119(ed3) worms were injected with rab-3::mCherry, myo-2::mCherry, myo-3::mCherry, pJL43.1 (a vector containing the Mos1 transposase under the control of the germline promoter glh-2), and a vector containing the mec-4 promoter::unc-40::gfp sequence to be inserted flanked by sequences homologous to the insertion site. Animals that were rescued for the unc-119 phenotype (array-positive) were allowed to starve out twice, and then unc-119 rescued animals that lacked the three mCherry coinjection markers (integrant-positive, array-negative) were cloned out from separate plates to find independent integrated lines. These lines were outcrossed twice to wild-type animals, and the presence of the intact insertion was verified by PCR.

Microscopy

Axonal processes of AVM neurons were visualized with the integrated mec-4::gfp transgene zdIs5 at L4 stage. The observer was not blind to the genotype. Animals were placed on 5% Noble agar pads in M9 buffer containing 10 mM sodium azide and examined with a Plan-Neofluar 40×, 1.4 numerical aperture (NA) oil-immersion objective on a Zeiss AxioImager. Images for axons and body morphology were captured with a Hamamatsu ORCA-ER cooled charge-coupled device (CCD) camera and 60× oil-immersion objective (NA 1.4). Fluorescence intensity was measured with the Zeiss AxioVision Rel. 4.7 image analysis software.

Statistics

Average data of reporter expression intensity are presented as means ± SEM. Data of % axons with defective ventral guidance are presented as proportions ± standard error of the proportion (SEP). Statistical analyses were carried out with the Primer of Biostatistics software for the Student’s t test or the two-proportion Z test. P < 0.01 or P < 0.001 was considered statistically significant.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/5/228/ra43/DC1

Fig. S1. Global survey of microRNA presence and abundance during C. elegans development.

Fig. S2. Responsiveness of a lin-14 3′UTR reporter to ectopic expression of lin-4 microRNA in AVM neurons.

Fig. S3. Responsiveness of a lin-14 3′UTR reporter to ectopic expression of miR-125a in AVM neurons.

Fig. S4. lin-4 and lin-14 mutations do not affect the timing of AVM axon outgrowth.

Fig. S5. The effect of the LIN-14 transcription factor on activity of the promoters of various unc-40 pathway genes in AVM neurons.

Table S1. lin-4 mutation does not affect AVM cell fate.

Table S2. C. elegans strains used in this study.

Table S3. Stem-loop RT primers and forward PCR primers.

References and Notes

Acknowledgments: We thank F. Ciamacco, B. Bayne, and K. Campbell for technical assistance; B. Lesch for the MosSCI technique; A. Fire for C. elegans vectors, F. Gertler for the FP4-mito construct, and the C. elegans Genetic Center for C. elegans strains; the WormBase for readily accessible information; V. Cleghon for critical reading of the manuscript; and C. Bargmann, M. Tessier-Lavigne, V. Ambros, and members of the Chang and Chuang labs for helpful discussions. Funding: This work was funded by grants from the following: the Whitehall Foundation Research Awards (C.C. and C.-F.C.), the March of Dimes Foundation Research Award (C.C.), the Alfred P. Sloan Research Fellowship (C.-F.C.), and the Canadian Institutes of Health Research grant RMF-82501 (C.C.). Author contributions: Y.Z. and H.C. performed the experiments and analyzed the data; H.C. created the schematic diagrams; D.D. analyzed axonal phenotypes in lin-4; slt-1 and lin-4; unc-6 strains; and C.C. and C.-F.C. designed the experiments, analyzed the data, and wrote the paper. Competing interests: The authors declare that they do not have any competing financial, personal, or professional interest.
View Abstract

Stay Connected to Science Signaling

Navigate This Article