A tubulin-MAPKKK pathway engages tubulin isotype interaction for neuroprotection

Abstract

The microtubule (MT) cytoskeleton is essential for neuronal morphology, neurite growth, synapse formation and maintenance, as well as regulation of signal transduction. Most cells express multiple isotypes of α- and β-tubulin that can coassemble into MTs. While a variety of signaling pathways regulate MT integrity and homeostasis, little is known about how tubulin isotypes interact in vivo. Here, we report a mechanism in which altered function of a neuronal β-tubulin in Caenorhabditis elegans activates the conserved kinase DLK-1 and its downstream signal transduction, which in turn upregulates expression of an α-tubulin isotype to ensure MT integrity. We find that alteration in the T7 loop of the β-tubulin/BEN-1 causes the formation of BEN-1-enriched islands along MTs in neurites. Combining genome editing with cellular imaging, we identified amino acid residues in α-tubulin/TBA-2 that are necessary for formation of BEN-1 islands. Activation of DLK-1 signaling in ben-1 mutants promotes TBA-2 transcription and protects axon and synapse morphology. These data uncover a positive feedback loop between DLK-1 and regulation of tubulin isotype interaction that maintains neuronal resilience.

Keywords: DLK kinase; bZip protein CEBP-1; microtubule plasticity; neuronal resilience; tubulin isotype.

Fig. 1.

ben-1(L246F) elevates DLK-1 signaling. (A) 3D structural model of α/β-tubulin heterodimer based on the structure of bovine brain tubulins (PDB accession code: 1TUB) (35). PDB DOI: https://doi.org/10.2210/pdb1TUB/pdb. GDP is bound to β-tubulin, GTP to α-tubulin; color highlights T7 loop (yellow), Leu246 (red), H7 α-helix (brown), and Thr238 (orange). (B) Compound fluorescence microscope images of CEBP-1::EGFP (wgIs563) in the ventral nerve cord (VNC) in L4-stage animals of indicated genotypes and treatment. Blue arrows mark neuronal nuclei showing visually detectable increased CEBP-1::EGFP, compared to WT animals. (Scale bar, 20 μm.) Null (0) alleles: ben-1(ju1813), dlk-1(tm4024). (C) Quantification of neuronal nuclei with increased CEBP-1::EGFP. n = 15 animals. Kruskal–Wallis test and Dunn’s posttest; **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. (D) Bright field images of animals of genotypes indicated cultured on control plates or plates containing 7 μM benomyl for 3 d. (Scale bar, 1 mm.) (E) Quantification of paralysis in animals of indicated genotypes cultured on plates containing benomyl of indicated concentration. 100 animals were scored per genotype/condition, from three experiments.

Fig. 2.

BEN-1(L246F) displays enrichment within stable MTs. (A) Kymographs of EBP-2::GFP in the ventral neurites of D-type GABAergic motor neurons in day 1 adults (D1A). Red line indicates a single EBP-2::GFP track. (B–D) Quantification of number (B), run length (C), and velocity (D) of EBP-2::GFP tracks in WT and ben-1(L246F) animals. n = number of animals in (B) and number of EBP-2::GFP tracks in (C and D). t test; NS (not significant), *P ≤ 0.05. (E) Schematics of Split GFP labeling of BEN-1(WT) and BEN-1(L246F) with representative confocal images of the DNC shown below. GFP11::BEN-1 (KI, ju1855) and GFP11::BEN-1(L246F) (KI, ju1959) are reconstituted with pan-neuronal expressed Prgef-1-GFP1-10 (juIs409) to generate RecGFP::BEN-1(WT) and RecGFP::BEN-1(L246F), respectively. Arrows indicate RecGFP::BEN-1(L246F) islands. (Scale bar, 10 μm.) (F) Quantification of RecGFP islands in DNC of animals at indicated stages. n = 18 animals. Kruskal–Wallis test and Dunn’s posttest; ****P ≤ 0.0001. (G and H) Quantification of the length (G) and intensity (H) of RecGFP::BEN-1(L246F) islands in animals at indicated stages. Relative intensity was calculated by dividing the intensity by the average intensity of RecGFP::BEN-1(WT). n = 96 islands. t test; NS (not significant), ****P ≤ 0.0001. (I) Quantification of RecGFP islands in DNC of animals cultured on plates containing 4 μM nocodazole (+) or buffer (−) for 2 d. n = 14 animals. Kruskal–Wallis test and Dunn’s posttest; ****P ≤ 0.0001. (J and K) Time-lapse confocal imaging of RecGFP::BEN-1(L246F) islands in commissures of D1A. (J) shows the formation of an island: The white arrowhead indicates the site of island formation, white arrows the island, yellow arrows the direction of island extension. (K) shows the dynamics of an island after it reached a steady length. (Scale bar, 5 μm.)

Fig. 3.

BEN-1(L246F) islands within MT shafts are enriched for GTP tubulin-binding protein. (A–F) Confocal images (Top) and line scan (Bottom) of Prgef-1-mKate2::MAPH-1.1 with RecGFP::BEN-1(WT) (A) and RecGFP::BEN-1(L246F) (B) in DNC, Prgef-1-PTRN-1::Crimson with RecGFP::BEN-1(WT) (C) and RecGFP::BEN-1(L246F) (D) in commissures, Prgef-1-EBP-2::mKate2 with RecGFP::BEN-1(WT) (E) and RecGFP::BEN-1(L246F) (F) in commissures of D2A. Dashed lines denote regions used for line scan analysis. (Scale bar, 10 μm.) (G) Percentage of RecGFP::BEN-1(L246F) islands coenriched for the indicated protein in D2A. n = number of islands. Quantification was performed on images of neurites coexpressing RecGFP::BEN-1(L246F) and the indicated protein.

Fig. 4.

TBA-2 pairs with BEN-1(L246F) to form MT islands, dependent on specific residues. (A and B) Confocal images (Top) and line scan (Bottom) of RecGFP::BEN-1(L246F) with Prgef-1-mKate2::TBA-1 (A) and Prgef-1-mKate2::TBA-2 (B) in DNC of D2A. Dashed lines denote regions used for line scan analysis. (C) Percentage of RecGFP::BEN-1(L246F) islands coenriched for mKate2-tagged tubulin isotypes in D2A. n = number of islands. Quantification was performed on images of neurites coexpressing RecGFP::BEN-1(L246F) and the indicated tubulin isotype. (D and E) Confocal images (D) and island quantification (E) of RecGFP::BEN-1(L246F) in DNC of animals of indicated genotypes. Ex(TBA-2), Prgef-1-TBA-2. Arrows indicate RecGFP::BEN-1(L246F) islands. (Scale bar in A, B, and D, 10 μm.) n = 18-23 animals. One-way ANOVA and Tukey’s posttest; NS (not significant), ****P ≤ 0.0001. (F and G) Models of BEN-1(L246F)/TBA-1 (confidence scores: pLDDT = 90.9 pTM = 0.905 ipTM = 0.909) and BEN-1(L246F)/TBA-2 (confidence scores: pLDDT = 89.9 pTM = 0.906 ipTM = 0.904) heterodimers using ColabFold. (H) Quantification of RecGFP::BEN-1(L246F) islands in DNC. n = 20 animals. One-way ANOVA and Tukey’s posttest; NS (not significant), ****P ≤ 0.0001. (I) Bright field images of animals of indicated genotypes on NGM plates. (Scale bar, 1 mm.) tba-2(ΔC/EBP motif), tba-2(ju2008). Null (0) allele: tba-1(ok1135), tba-2(tm6948).

Fig. 5.

DLK-1 promotes BEN-1(L246F) island formation in mature neurons. (A and B) Confocal images (A) and island quantification (B) of RecGFP::BEN-1(WT) and RecGFP::BEN-1(L246F) in DNC of animals of indicated genotypes. Arrows indicate RecGFP::BEN-1(L246F) islands. Null (0) alleles: dlk-1(tm4024), rpm-1(ju44). Si(Pn-dlk-1), Prgef-1-GFP::DLK-1L (juSi163). n = 20-29 animals. One-way ANOVA and Tukey’s posttest; ****P ≤ 0.0001. (C) Schematic of the HS assay. (D) Confocal images of Phsp-16.2-GFP::DLK-1L [juSi277, Si(Phsp-dlk-1)] in VNC of L4-stage animals with no HS or at the indicated time after HS. (E and F) Confocal images (E) and island quantification (F) of RecGFP::BEN-1(L246F) in VNC of animals of indicated genotypes and treatment. Arrows indicate RecGFP::BEN-1(L246F) islands. n = 9-20 animals. t test; NS (not significant), ***P ≤ 0.001. (Scale bar in A, D, and E, 10 μm.)

Fig. 6.

DLK-1 promotes TBA-2 expression for BEN-1(L246F) island formation. (A) Confocal images of GFP::TBA-2 (KI, tj38) in VNC of day 2 adults of indicated genotypes. Dashed ovals mark GABAergic DD motor neurons labeled by Pflp-13::mCherry. dlk-1(0), dlk-1(tm4024); tba-2(ΔC/EBP motif), tba-2(ju2008). (Scale bar, 10 μm.) (B) Quantification of relative intensity of GFP::TBA-2 in the soma of VNC motor neurons in animals of indicated genotypes. n = 18-57 neurons. One-way ANOVA and Tukey’s posttest; ***P ≤ 0.001. (C and D) Confocal images (C) and island quantification (D) of RecGFP::BEN-1(L246F) in DNC of animals of indicated genotypes. Ex(TBA-1), Prgef-1-TBA-1; Ex(TBA-2), Prgef-1-TBA-2. Arrows indicate RecGFP::BEN-1(L246F) islands. (Scale bar, 10 μm.) n = 20-23 animals. Kruskal–Wallis test and Dunn’s posttest; NS (not significant), **P ≤ 0.01, ****P ≤ 0.0001. (E) Graphic illustration of DLK-1-mediated neuronal resilience to ben-1(L246F)-induced MT perturbation.

Fig. 7.

DLK-1 ensures synapse formation and axon morphology in ben-1(L246F) mutants. (A and B) Airyscan images (A) and quantification (B) of ELKS-1::mNG (KI, syb1710) in DNC of indicated genotypes. (Scale bar, 10 μm.) n = 13-14 animals. One-way ANOVA and Tukey’s posttest; NS (not significant), ****P ≤ 0.0001. (C) Airyscan images of UNC-17::mKate2 (KI, ot907) in D2A of indicated genotypes. Boxed DNC regions are shown enlarged. Arrowheads indicate commissures accumulating UNC-17::mKate2; dashed oval marks a neuronal soma accumulating UNC-17::mKate2. (Scale bar, 10 μm.) (D) Quantification of animals showing ectopic accumulation of UNC-17::mKate2 in the commissures (Commi). 60 animals were scored per genotype. n = 3 experiments. Kruskal–Wallis test and Dunn’s posttest; **P ≤ 0.01. (E) Confocal images of RecGFP::BEN-1(WT)_DD (Left) or RecGFP::BEN-1(L246F)_DD (Right) with Punc-25::mCherry::RAB-3 (juIs236) in DD neuron dorsal axons in adults of indicated genotypes. GFP1-10 is expressed in DD neurons using Pflp-13-GFP1-10 (juIs463). White arrows, RecGFP::BEN-1(L246F)_DD islands; Yellow brackets, regions where RecGFP::BEN-1(L246F)_DD is undetectable. (Scale bar, 10 μm.) Null (0) alleles are dlk-1(tm4024) and ben-1(ju1813).