Glutamylation Regulates Transport, Specializes Function, and Sculpts the Structure of Cilia

Abstract

Ciliary microtubules (MTs) are extensively decorated with post-translational modifications (PTMs), such as glutamylation of tubulin tails. PTMs and tubulin isotype diversity act as a "tubulin code" that regulates cytoskeletal stability and the activity of MT-associated proteins such as kinesins. We previously showed that, in C. elegans cilia, the deglutamylase CCPP-1 affects ciliary ultrastructure, localization of the TRP channel PKD-2 and the kinesin-3 KLP-6, and velocity of the kinesin-2 OSM-3/KIF17, whereas a cell-specific α-tubulin isotype regulates ciliary ultrastructure, intraflagellar transport, and ciliary functions of extracellular vesicle (EV)-releasing neurons. Here we examine the role of PTMs and the tubulin code in the ciliary specialization of EV-releasing neurons using genetics, fluorescence microscopy, kymography, electron microscopy, and sensory behavioral assays. Although the C. elegans genome encodes five tubulin tyrosine ligase-like (TTLL) glutamylases, only ttll-11 specifically regulates PKD-2 localization in EV-releasing neurons. In EV-releasing cephalic male (CEM) cilia, TTLL-11 and the deglutamylase CCPP-1 regulate remodeling of 9+0 MT doublets into 18 singlet MTs. Balanced TTLL-11 and CCPP-1 activity fine-tunes glutamylation to control the velocity of the kinesin-2 OSM-3/KIF17 and kinesin-3 KLP-6 without affecting the intraflagellar transport (IFT) kinesin-II. TTLL-11 is transported by ciliary motors. TTLL-11 and CCPP-1 are also required for the ciliary function of releasing bioactive EVs, and TTLL-11 is itself a novel EV cargo. Therefore, MT glutamylation, as part of the tubulin code, controls ciliary specialization, ciliary motor-based transport, and ciliary EV release in a living animal. We suggest that cell-specific control of MT glutamylation may be a conserved mechanism to specialize the form and function of cilia.

Keywords: C. elegans; cilia; extracellular vesicles; glutamylation; intraflagellar transport; kinesin-2; kinesin-3; microtubule; polycystin; post-translational modifications.

Figure 1. ttll-11 regulates PKD-2∷GFP ciliary localization and encodes two TTLL family glutamylase isoforms with non-overlapping expression patterns

(A) Diagrams (adapted from [23]) of male-specific extracellular vesicle-releasing neurons (“EVNs”: CEMs in head, and HOB and RnBs, where n=1–9, except 6 in tail), which express PKD-2∷GFP. In the diagram of the nose, only the left side CEM neurons of the two bilateral pairs is shown. Boxed area indicates region shown in micrographs. Abbreviations: d, dendrite; cb, ciliary base; c, cilium. Cell bodies are further posterior, out of view. In the tail, each ray is innervated by a single RnB neuron; the dendrite and cilium of R3B are shown as an example. Panels show PKD-2∷GFP localization in nose and tail for indicated genotypes. See also Figure S1 and Table S1. (B) Genomic structure of ttll-11 locus, which encodes two isoforms. Bars below the diagram show the gk482 and tm4059 deletion alleles used in this study. Both TTLL-11B and TTLL-11A contain a TTL domain of 364 amino acids (black region). MYR indicates a predicted myristoylation site [29] in TTLL-11B. By analogy with other TTLL proteins, the C-terminal gray area is expected to bind microtubules. (C, D) A ttll-11b promoter drove GFP expression in the EVNs, marked by klp-6 promoter-driven tdTomato. Expression of both transgenes was somewhat mosaic, and therefore did not completely overlap; for example, one of the CEM ventral neurons in the head was not labeled by tdTomato. (E, F) GFP expression driven by the ttll-11a promoter was not visible in the EVNs, marked by expression of Pklp-6∷tdTomato. (G) Behavioral response to hermaphrodites in male mating was scored. ccpp-1 and ccpp-1;ttll-11 mutant males responded significantly less frequently than wild-type. Data represents mean ± sem; N = 4 or 5 trials, n = 35 – 56 males tested for each genotype. ***indicates p<0.0001 by ANOVA and Tukey post-hoc tests. (H) Vulva location behavior was scored for 10 – 25 males for all genotypes. ** p<0.001, *** p<0.0001 by Kruskal-Wallis and Dunn’s Multiple Comparison Test.

Figure 2. TTLL-11 was required for ciliary MT glutamylation

Fixed young adult males expressing PKD-2∷GFP (green) stained with monoclonal antibody GT335 (red), which detects the branch point of glutamylation side-chains. Cartoon depicts cilia observed in the nose and tail. In the tail, each ray B-type neuron is equipped with a sensory cilium, but cartoon only shows cilium for the left R3B neuron. For each genotype, left panel shows ciliated endings in the tip of the nose; middle panel shows male tail fan, and right panel shows enlargements of the boxed areas in tails to show ray cilia. In wild type panels, GT335 staining of amphid ciliary MS (middle segments, which contain doublet MTs), CEP cilia, and IL cilia in the head, and phasmid cilia in the tail, are indicated. NS indicates puncta of non-specific staining, where antibody stuck to the cuticle or cellular debris on some animals. See also Figure S2 and Table S1.

Figure 3. Loss of TTLL-11 function suppressed GFP∷KLP-6 kinesin-3 ciliary accumulation of ccpp-1 mutants

(A) Localization of the kinesin-3 motor GFP∷KLP-6 was diffuse in wild type and ttll-11 single mutants. In ccpp-1 mutants, GFP∷KLP-6 accumulated in cilia of EVNs in the head (CEMs and IL2s) and tail (HOB and RnBs). Mutation of ttll-11 suppressed the abnormal ciliary enrichment of GFP∷KLP-6 in ccpp-1 mutants. (Tail images were normalized so that brightness of autofluorescent posterior tip of acellular tail fan was similar across genotypes. c, cilium; d, dendrite.) (B) The enrichment of GFP∷KLP-6 was quantified across genotypes by calculating the ratio of the maximum pixel value in cilia over the maximum pixel value in the distalmost 10 μm of dendrites for the head neurons only. Data represents Mean ± sem for N = 3 animals for each genotype; ** indicates significantly different from wild type with p= 0.002 by ANOVA and post-hoc Tukey test. See also Table S1.

Figure 4. The MT glutamylase TTLL-11 and deglutamylase CCPP-1 were required for release of PKD-2∷GFP-labeled extracellular vesicles (EVs)

(A) Cartoon (modified from [7] shows CEM EV “shedding,” which produces the EVs surrounding the CEM ciliary base inside a lumen formed by the sheath and socket glial cells; and EV “release” from CEM ciliated neurons to the environment. (B) Diagram shows PKD-2∷GFP-labeled EVs released from CEMs float and accumulate at the cover slip. (C) Abundant PKD-2∷GFP-labeled EVs were released outside from CEM neurons in wild-type adult males (several EVs indicated by arrowheads). Few EVs were seen in ttll-11 mutants. (D) Quantification of EVs released to the local environment by sensory ciliated CEM neurons in adult male head. N animals scored in parentheses for each genotype. See Figure S4 for images of additional genotypes, as well as images of PKD-2∷GFP-labeled EVs released by neurons in the male tail in adults and L4 larvae. (E) EVs shed into the glial lumen were reconstructed and counted from serial TEM sections. Representative sections are shown in Fig. S5. See also Table S1. In parentheses, N = cilia scored. (Mean ± sem;* p=0.0153; ** p=0.0065; ***p<0.001 by ANOVA and post-hoc Tukey tests.)

Figure 5. TTLL-11∷GFP was enriched in cilia and a cargo of environmentally released EVs

(A) TTLL-11∷GFP expressed under the pkd-2 promoter was enriched in CEM cilia and ciliary bases. TTLL-11∷GFP appeared in puncta in dendrites and cell bodies of C. elegans males. CEM cilia and ciliary base are indicated by a bracket and arrow, respectively. (B, C) TTLL-11∷GFP was observed in EVs released from CEM neurons in the head (B) and RnB neurons in the tail (C). Yellow arrowheads point to TTLL-11∷GFP-labeled EVs. (D) Kymograph and histograms of TTLL-11∷GFP movement in CEM dendrites. Mean ± sem dendritic velocities shown for 15 animals, 153 anterograde particles; 98 retrograde particles. (E) Mean ± sem anterograde velocity of TTLL-11∷GFP in CEM cilia for genotypes shown, and representative kymographs created using Kymograph Direct [43]. See also Table S1.

Figure 6. CCPP-1 and TTLL-11 regulators of MT glutamylation were required for normal CEM neuronal ciliary ultrastructure

(A) Representative TEM sections through middle regions of CEM cilia in genotypes indicated, characterized by singlets in the wild type. Illustrations indicate position of CEM cilium and MTs, as well as the CEP cilium. Scale bar = 250nm. (B) A series of CEM middle sections from ccpp-1 mutants. Open “C-shaped” tubules in section 2 are indicated by arrows. Approximate locations of numbered sections indicated by red dotted lines in panel C. Scale bar = 50nm. (C) Cartoon model of a single outer doublet microtubule along the ciliary length for wild type (WT), ttll-11, ccpp-1, and ccpp-1; ttll-11 double mutant genotypes. See also Table S1.