The DLK-1 kinase promotes mRNA stability and local translation in C. elegans synapses and axon regeneration

Abstract

Growth cone guidance and synaptic plasticity involve dynamic local changes in proteins at axons and dendrites. The Dual-Leucine zipper Kinase MAPKKK (DLK) has been previously implicated in synaptogenesis and axon outgrowth in C. elegans and other animals. Here we show that in C. elegans DLK-1 regulates not only proper synapse formation and axon morphology but also axon regeneration by influencing mRNA stability. DLK-1 kinase signals via a MAPKAP kinase, MAK-2, to stabilize the mRNA encoding CEBP-1, a bZip protein related to CCAAT/enhancer-binding proteins, via its 3'UTR. Inappropriate upregulation of cebp-1 in adult neurons disrupts synapses and axon morphology. CEBP-1 and the DLK-1 pathway are essential for axon regeneration after laser axotomy in adult neurons, and axotomy induces translation of CEBP-1 in axons. Our findings identify the DLK-1 pathway as a regulator of mRNA stability in synapse formation and maintenance and also in adult axon regeneration.

Figure 1. Loss of function in mak-2 suppresses rpm-1(lf)

(A) MAK-2 is a member of the MAPKAPK family and has two isoforms that differ at their N-termini. Mouse MK2 is NP_032577. (B) mak-2(lf) suppresses motor neuron presynaptic morphology defects of rpm-1 cell-autonomously. Images are juIs1[Punc-25-SNB-1::GFP] in the adult dorsal cord. SNB-1::GFP puncta are uniformly-shaped and evenly distributed in wild type and mak-2, but are irregular and fewer in rpm-1 mutants. Right panel, quantification of SNB-1::GFP puncta # (n ≥ 12 animals per genotype; data shown as mean ± SEM). (C) mak-2 suppresses touch neuron axon termination defects of rpm-1 cell-autonomously. Images are muIs32[Pmec-7-GFP] in adults. The ALM axon (red arrows) ends before the tip of the nose in wild type, but overextends and frequently loops in rpm-1 mutants. The PLM axon (blue arrows) terminates posterior to the ALM cell body (*) in wild type, but overextends and turns ventrally in rpm-1. The PLM synaptic branch (purple arrows) is frequently missing in rpm-1. Axon termination defects of PLM and ALM neurons are quantified as % of total # of animals. Quantification of PLM synaptic branch shows similar pattern (not shown). (D) Expression of MAK-2(EE) induces gain of function phenotypes. (E) MAK-2 rescuing activity is abolished by mutation of Thr250 and Thr262 or of Lys103 in the ATP-binding site, but is not affected by deletion of the NLS or NES. (B–E) Unless noted, alleles are rpm-1(ju44) and mak-2(ok2394). Promoters: Prgef-1 (F25B3.3) for pan-neural, Pmyo-3 for body wall muscles; Pmec-4 for touch neurons; Punc-25 for GABA motor neurons. N, total # of animals; TL, total # of transgenic lines; RL, # of transgenic lines showing rescuing activities. Statistics in this and subsequent figures, ANOVA (panel B) or Fisher exact test (C–E). *, P < 0.05; **, P < 0.01; ns, not significant. Scales, 10 μm.

Figure 2. Loss of function in cebp-1 suppresses rpm-1(lf)

(A) CEBP-1 contains a bZip domain most similar to those of C/EBP proteins. (B) cebp-1(lf) suppresses motor neuron presynaptic morphology defects of rpm-1 cell-autonomously (juIs1 marker). Right panel, quantification of SNB-1::GFP puncta (n ≥ 10 animals per genotype; data shown as mean ± SEM). (C) cebp-1(lf) suppresses the touch axon overextension of rpm-1 cell-autonomously (muIs32 marker). (D) The bZip domain is required for cebp-1 function. (E) CEBP-1(R290C) and CEBP-1(S289L) have dominant-negative activity. (B–E) Unless noted, alleles are rpm-1(ju44) and cebp-1(ju659). Quantitation, labels, details of mutant phenotypes, promoters and statistics as in Figure 1. Scales, 10 μm.

Figure 3. rpm-1 and the MAP kinases regulate levels of cebp-1 mRNA via its 3′ UTR

(A) Quantitative RT-PCR analysis shows up-regulation of cebp-1 mRNA in rpm-1(lf), which is eliminated in dlk-1, pmk-3 or mak-2 mutants (n=3). The control transcript ama-1 encodes the large subunit of RNA polymerase II (Sanford et al., 1983). (B–D) Transgenic expression of mCherry reporters in GABA motor neurons and quantification of fluorescence (n = 35 per genotype). Blue lines and letters denote inclusion of the cebp-1 3′ UTR in the transgene. (B) CEBP-1 with its own 3′ UTR is more highly expressed in rpm-1(lf) and mak-2(EE) transgenic animals than in wild type, dlk-1, and MAK-2(AA) transgenic animals. (C) A CEBP-1 transgene with the unc-54 3′UTR is expressed at similar levels in rpm-1 and wild type. (D) The cebp-1 3′UTR confers up-regulation of mCherry in rpm-1. (E) Expression of cebp-1 with the unc-54 3′UTR causes abnormal motor synapses. Data in A–E shown as mean ± SEM; n = 30 animals per genotype; statistics, Anova. (F) Overexpression of cebp-1 with its own UTR causes PLM axon overextension; transgenes with the unc-54 3′ UTR cause further defects in ALM. Overexpression of cebp-1 from muscles does not affect neuronal morphology. Quantitation, labels, details of mutant phenotypes, promoters and statistics as in Figure 1; GL, # of transgenic lines exhibiting gain-of-function abnormalities. Scales, 10 μm.

Figure 4. MAK-2 and CEBP-1 are present at motor neuron synapses

(A) Functional mCherry::MAK-2 in GABA motor neurons. Top panels are motor neuron soma in animals coexpressing Punc-25-mCherry::MAK-2 (red) and Histone::GFP (zuIs178). MAK-2 is present both in nucleus and cytoplasm. Bottom panels are motor neuron synapses in the dorsal cord coexpressing mCherry::MAK-2 with Punc-25-SNB-1::GFP (juIs1, left), or with RPM-1::GFP (juIs77, right). MAK-2 partially colocalizes with SNB-1 but not with RPM-1 (quantitation in Figure S7). (B) Functional mCherry::CEBP-1 in motor neuron cell bodies (top panels) and synapses (bottom panels) using the same markers as in (A). CEBP-1 shows nuclear and cytoplasmic localization in soma, and co-localizes with SNB-1 but is excluded from areas of RPM-1 at synapses (Figure S7). Synapse morphology (SNB-1::GFP) is unaltered in these transgenic animals. (C) mCherry::PHR::MAK-2 is confined to motor neuron synapses (left panels), and is excluded from the nucleus (N) in the soma (right panel). (D) Synaptic targeting of constitutively active MAK-2(EE), but not wild type MAK-2(+) or inactive MAK-2(AA), alters synapse morphology and reduces SNB-1::GFP puncta number (juIs1). (E) Quantification of the SNB-1::GFP phenotypes in D (n ≥ 30 per genotype). (F) Induction of un-regulated cebp-1 in adult neurons alters synapse morphology and number (juIs1). Quantification is on the right (n ≥ 30 for each group). 15°C–15°C indicates animals cultured continuously at 15°C; 15°C–25°C indicates animals grown at 15°C until late L4, then shifted to 25°C for 24 h. Data shown as mean ± SEM; statistics, Anova. Scales, 10 μm.

Figure 5. cebp-1 protein and mRNA are localized in touch neuron axon and synapses

(A) Expression of functional GFP::CEBP-1 in touch neurons. GFP is present along the axon (left panel), in the cell body (middle) and synapse (right). (B) GFP::NLS::MS2 reporter detects mRNA of cebp-1 in touch neuron axons. Illustrations of the constructs are shown above each image. Axonal fluorescence intensity as a fraction of peak cell body intensity is plotted versus distance along the axon from the cell body. GFP::NLS::MS2 alone (juEx1982) is concentrated in the cell body, and faintly seen along the axon. Coexpression of juEx1982 with cebp-1 transgenes containing six MS2 binding sites in the cebp-1 3′ UTR (juEx1984), but not in unc-54 3′ UTR (juEx2013), results in increased punctate GFP in axons. Axonal expression of GFP::NLS::MS2 + CEBP-1-cebp-1 3′ UTR is decreased in dlk-1(lf) animals, but significantly elevated in rpm-1(lf). (C) Average GFP::NLS::MS2 fluorescence intensity in axons when co-expressed with juEx1984 as % of cell body intensity (n = 15 per genotype). Data shown as mean ± SEM; statistics, Anova. Scale, 10 μm.

Figure 6. The DLK-1 pathway and CEBP-1 are required for touch axon regeneration in adults

(A) Regeneration of PLM axon (muIs32 marker) is reduced in loss of function mutants in each MAP kinase and cebp-1. mak-2(lf) has modest effects on regeneration compared to dlk-1(lf); the defect is rescued by wild type MAK-2 but not by MAK-2(K103R). Overexpression of DLK-1 enhances regrowth in a CEBP-1-dependent manner. Images are at 0 h and 24 h; quantitation at 24 h. Alleles are dlk-1(ju476), mkk-4(ju91), pmk-3(ok169), cebp-1(ju659), juEx1301[Prgef-1-dlk-1(+)] as dlk-1(++), juEx2111[Pmec-4-GFP::MAK-2(+)], and juEx2262[Pmec-4::MAK-2(K103R)]. n, # of animals shown below genotype or in column. (B) Heat-shock induced expression of cebp-1(R290C) transgene in adults blocks PLM regeneration. n, # of animals, in columns. Data shown as mean ± SEM; statistics, Anova. Scales, 10 μm.

Figure 7. Axotomy induces translation of CEBP-1 in a MAP kinase dependent manner and involves the 3′ UTR of cebp-1

Shown are images of Dendra reporter transgenes expressed in touch neurons, with illustration of constructs above. Green images are unconverted or newly synthesized Dendra; red images show photoconverted Dendra (Dendra*). (A) Under normal conditions synthesis of new Dendra::CEBP-1 from transgenes containing the cebp-1 3′ UTR (juEx1915) is slow. Quantification data are collected from soma, and axons of 60 μm away from soma (n = 10 for each time point). (B) Axotomy of PLM axons (middle), but not laser wounding of nearby epidermis (top panels), increases synthesis of Dendra::CEBP-1 (juEx1915). The increase is blocked by dlk-1(lf) or cycloheximide, and reduced by mak-2(lf). White arrows mark the laser cutting sites. (C) Axotomy of PLM axons does not induce new protein synthesis from a Dendra::CEBP-1 transgene lacking cebp-1 3′ UTR. (D) Quantitation of B and C (n = 12 for each group). (E, F) cebp-1 3′ UTR is sufficient to confer axotomy-induced protein synthesis (n = 12 for each group). Data shown as mean ± SEM. Statistics, Anova or t-test. Scales, 10 μm.