MKS and NPHP modules cooperate to establish basal body/transition zone membrane associations and ciliary gate function during ciliogenesis

Abstract

Meckel-Gruber syndrome (MKS), nephronophthisis (NPHP), and related ciliopathies present with overlapping phenotypes and display considerable allelism between at least twelve different genes of largely unexplained function. We demonstrate that the conserved C. elegans B9 domain (MKS-1, MKSR-1, and MKSR-2), MKS-3/TMEM67, MKS-5/RPGRIP1L, MKS-6/CC2D2A, NPHP-1, and NPHP-4 proteins exhibit essential, collective functions at the transition zone (TZ), an underappreciated region at the base of all cilia characterized by Y-shaped assemblages that link axoneme microtubules to surrounding membrane. These TZ proteins functionally interact as members of two distinct modules, which together contribute to an early ciliogenic event. Specifically, MKS/MKSR/NPHP proteins establish basal body/TZ membrane attachments before or coinciding with intraflagellar transport-dependent axoneme extension and subsequently restrict accumulation of nonciliary components within the ciliary compartment. Together, our findings uncover a unified role for eight TZ-localized proteins in basal body anchoring and establishing a ciliary gate during ciliogenesis, and suggest that disrupting ciliary gate function contributes to phenotypic features of the MKS/NPHP disease spectrum.

Figure 1.

C. elegans B9 and C2 domain ciliopathy proteins are found at the transition zone, adjacent to the basal body/transition fiber region. (A) Schematic of a prototypical basal body (BB)/transition zone (TZ)/cilium, highlighting the microtubule (MT) backbone of the organelle, the docking of vesicles at the base of a “ciliary gate,” and its intraflagellar transport (IFT) trafficking machinery. In C. elegans, the BB lacks discernable MTs (shown dashed) and consists mainly of transition fibers (TFs). The TZ is demarcated by Y-links spanning the axoneme to the membrane and likely organizes the “ciliary necklace” on the ciliary membrane. The TFs and TZ are proposed to form a gate that restricts entry of vesicles and potentially also nonciliary proteins. TEM cross sections from C. elegans show relevant substructures (TFs, TZ, middle and distal cilia segments; Bars, 100 nm). IFT particles carry cargo from the BB into cilia; kinesins transport two multi-protein complexes (IFT subcomplexes A and B) and a BBS protein complex (S) along with cargo, and dynein recycles components back to the BB. (B) MKS-1::YFP localization to the TZ in relation to the CHE-13 IFT protein, which concentrates at BB/TFs and is present along the axoneme. Two phasmid (tail neuron) cilia are shown, as in C and D; all tagged proteins are expressed under endogenous promoters unless specified otherwise; MS, middle segment; DS, distal segment. Bar, 2 µm. (C and D) Similar to MKS-1, MKS-5::tdTomato and MKS-6::GFP (both osm-5 promoter driven) localize to the TZ, which largely does not overlap with the peak intensities of tagged IFT proteins (DYF-11 and XBX-1, respectively) at the adjacent BB/TFs. Bars, 2 µm. (E) B9 domains of MKS-1, MKSR-1, and MKSR-2 may be structurally related to C2 domains of RGRIP1L/MKS-5 and CC2D2A/MKS-6 (see F and G). A representative structure of synaptotagmin I C2 domain (PDB code 1byn), with bound Ca²⁺, is shown. (F) A Hidden Markov Model profile was created using B9 domains from C. elegans, C. briggsae, and C. remanei to search the C. elegans proteome for related domains in evolutionarily conserved proteins. Only four are retrieved: the B9 input proteins and a C2 domain protein, synaptotagmin-4 (SNT-4). (G) The top hits from the structure prediction algorithm GenTHREADER reveal that all three human B9 domains can be modeled onto known C2 domain NMR/crystal structures from different proteins (synaptotagmins, E3 ubiquitin-protein ligase NEDD4-like, phospholipase C-delta1, and rabphilin-3A; PDB codes in parentheses).

Figure 2.

MKS/MKSR and NPHP proteins localize at the ciliary TZ. (A) Fluorophore-tagged MKS/MKSR and NPHP proteins localize at the TZ in tail cilia. Each TZ protein signal is overlaid with IFT or other TZ protein signal, as indicated. Bar, 10 µm. (B) MKS-5::tdTomato and MKS-6::GFP localize at the TZ in head cilia. MKS-5 and MKS-6 are overlaid with IFT proteins DYF-11::GFP and XBX-1::tdTomato, respectively. Bar, 10 µm. (C) GFP-tagged MKS-1, MKSR-1, and MKS-6 localize at the base of ciliary axonemes (consistent with the TZ), which are marked by staining with an antibody against polyglutamylated tubulin. Bar, 10 µm. (D) Schematic of MKS/MKSR/NPHP protein localization at the TZ.

Figure 3.

Functional interactions between MKS-5/MKS-6 and other TZ proteins are required for environmental exposure of cilia to a fluorescent dye, consistent with a role in ciliogenesis. (A) Gene structures of C. elegans mks-5 (C09G5.8) and mks-6 (K07G5.3) and the nature of three deletion alleles, tm3100, gk674, and nx105. (B and C) Representative images of fluorescence staining of environmentally exposed sensory neurons via DiI uptake through cilia in head (amphid) neurons (left panels) and tail (phasmid) neurons (right panels). Dye filling is noted as normal (+++), reduced (+), or absent (−). (B) Wild-type (N2) and mks-6 strains exhibit normal dye filling; nphp-1 and nphp-4 show normal or slightly reduced (++) dye filling (see D; Jauregui et al., 2008). mks-6;mks-5, mks-6;nphp-1, and mks-6;nphp-4 double-mutant combinations show little or no dye filling. Expressing GFP-tagged MKS-6 rescues the mks-6 (double) mutant phenotype. (C) mks-5(tm3100) mutants have reduced dye filling, whereas mks-5;mksr2, mks-5;mks-6, and mks-5;nphp-4 double mutants show little to no dye filling. Bars, 17.5 µm. (D) Summary of dye-filling phenotypes in single and double mks/mksr/nphp mutants. NT, not tested; red, present study; black, data from Jauregui et al. (2008) and Williams et al. (2008, 2010).

Figure 4.

Functional interactions between TZ proteins are required for correct cilium length and positioning of basal bodies/TZs, as well as chemosensation. (A and B) Disruption of mks-5 (A) or mks-6 (B) has no major effect on the gross morphology/presence of amphid (head) and phasmid (tail) cilia, as revealed by GFP-tagged IFT markers (OSM-6 and CHE-2, respectively). BB, basal body. Bars, 5 µm. (C) Analyses of amphid and phasmid cilia length (white and light gray bars), spatial distribution of BB/TZ in amphid channel neurons (black bars), and position of the BB with respect to the cell body in tail neurons (dark gray bars) are summarized based on phenotypes and statistical analyses shown in Fig. 5. Mean values obtained for wild-type (N2) are shown; *, statistically significant difference (P < 0.01) vs. N2; §, statistically significant variance (P < 0.01) vs. N2. Schematics illustrate how measurements were made for amphid and phasmid cilia lengths, the distribution of BBs within amphid neurons (some BBs in mutants do not form a tight cluster as in wild type), and distance of the cilia from the phasmid neuron cell body (some mutants have abnormally short dendrites, indicating lack of anchoring of upon retrograde extension of the cell body during development). (D) Graph of cilium-dependent osmotic avoidance of N2 and TZ mutant strains. osm-5 (control IFT mutant) fails to avoid high osmolarity solution. Two mks-6 alleles show osmosensation defects in an nphp-4 mutant background, as do mksr-2 and mks-5 mutants. *, significant difference compared with N2 (P < 0.05); **, significant difference compared with either single mutant for the indicated double mutant.

Figure 5.

Detailed analysis of cilia length and positioning in strains with disrupted TZ proteins. (A–D) Example of correct (wild type; top panels) or defective (for TZ gene mutants; bottom two panels) amphid (A) and phasmid (B) cilia length, distribution of BBs in amphid ciliated neurons (C), and positioning of the BB with respect phasmid neuron cell body (D), using a CHE-11::GFP IFT marker (with the exception of the mks-5 data, which was obtained using OSM-6::GFP). CB, cell body; cil, cilia; BB, basal body. Bar, 5 µm. (A′–D′) Individual data points (n > 40) of the measurements in A–D are shown with a bar graph showing average and standard error. Statistically significant differences (P < 0.01) in the mean (*) or variance (§) are indicated for all combinations.

Figure 6.

MKS-6 and NPHP-4 are collectively required for BB/TZ attachments to membrane. Shown are low and high magnification images from TEM serial cross sections of amphid channel cilia from wild-type (N2), mks-6, and mks-6;nphp-4 worms. Top row, distal segments (DS); middle row, middle segments (MS); bottom row, transition zone (TZ) and distal dendrites (DD). Below bottom row are magnified images of representative TZs. Boxed number denotes proximal positioning of section relative to top section. Schematics (longitudinal, transverse views) summarize the major ultrastructural observations. BB, basal body; TFs, transition fibers. Bars, 100 nm. (A) N2 worms showing 10 singlet microtubule (MT)-containing axonemes in DS, 10 doublets in MS (+3 µm), and TZs (constriction of MT doublets surrounding apical ring) with intact Y-links (connecting doublets with ciliary membrane) anchored at distal dendrite tips (+7 µm). TFs are also observed below this region. (B) Apart from rare axonemes lacking MTs in DS, mks-6 mutant cilia possess normal DS, MS (+2 µm), and TZ (+5 µm) regions. (C) In DS and MS (+4 µm) regions of mks-6;nphp-4 worms, many axonemes are missing and amphid pore size is reduced. Open/unzipped B-tubules are also observed, as in nphp-4 mutants (Jauregui et al., 2008). In TZ/DD region, TZs are not anchored at the membrane of the DD tips, but instead are often mispositioned in more proximal regions of the dendrites (see also schematic). Y-links at the TZ are missing and TFs are not observed.

Figure 7.

Functional interactions between five TZ-associated proteins (MKS-1, MKSR-1, MKS-3, MKS-5, and MKS-6) and NPHP-4 contribute to anchoring the BB/TZ to membrane. (A) TEM images of amphid channel TZ regions obtained from longitudinal sections of wild-type (N2) and mks-6;nphp-4 double mutants. Compared with N2 worms in which membrane (green outlines) associates tightly with the TZ (demarcated by dashed lines), mks-6;nphp-4 double-mutant TZs fail to anchor to the surrounding membrane (red outlines). Bars, 500 nm; 200 nm (insets). (B) TEM images of amphid channel ciliary TZ regions obtained from cross sections of mks-3;nphp-4, mks-5;nphp-4, mks-1;nphp-4, mksr-1;nphp-4 and mks-6;nphp-4 double mutants. Boxed number denotes the proximal positioning of the imaged section relative to the most distal sections of the amphid pore. In all worms, the MT doublet-containing TZ ring is not anchored at distal dendrite tips (which are typically positioned at +5 to +6 µm relative to distal amphid pore); instead, TZs are found at abnormal proximal positions in the dendrites (+8 to +16 µm). Y-links are frequently not observed. Bars, 100 nm.

Figure 8.

MKS/MKSR and NPHP proteins are organized hierarchically as modules, and independently of IFT/BBS proteins, at the TZ. (A) Schematic representing the region of phasmid (tail) cilia analyzed in B–E. Entire cilia/cilia regions are shown in F–K. (B–E) Co-dependent localization of fluorophore-tagged TZ proteins. TZs lacking normal localization of a particular protein are circled. (B) MKS-6::GFP localizes normally in wild-type (N2) and mks-1, mks-3, and nphp-4 mutants, but mislocalizes in mks-5, mksr-1, and mksr-2 mutants. (C) mks-5 mutants fail to localize MKS-1, MKSR-1, MKSR-2, and MKS-3, the latter dispersing along the cilium axoneme (see also Fig. 9 L). NPHP-1 and NPHP-4 localize to the TZ in mks-5 mutants, but to a subregion smaller than that occupied in N2 (compare brackets). (D) MKS-5::tdtomato localization at the TZ is not perturbed in mksr-2, nphp-4, or mks-6;nphp-4 double mutants. (E) MKS-3::GFP localizes correctly to the TZ in nphp-4 and mks-6 single mutants but is mislocalized in the double mutant. (F) MKS-6::GFP localizes normally in bbs mutants (bbs-7, bbs-8) and ift mutants (osm-5, che-11), indicating the presence of an intact TZ. (G) tdTomato-tagged NPHP-1 (top) and MKSR-1 (bottom) localize normally at the TZ in mks-6(gk674) mutants. (H–K) CFP- or YFP-tagged MKS-1, MKSR-2, NPHP-1, and NPHP-4 localize normally at the TZ in mks-3 mutants. Bars, 2.5 µm.

Figure 9.

MKS/NPHP proteins are required for ciliary gate function but not for ODR-10 trafficking. (A and B) GFP-tagged ODR-10 odorant receptor concentrates specifically at the AWB cilium in wild-type (N2) animals, showing a typical branched ciliary structure. Bars: (A) 25 µm; (B–K) 8 µm. (C–G) Disruption of the indicated TZ genes does not overtly affect the structure of the AWB cilium or presence of ODR-10::GFP. (H–K) Various double TZ mutants reveal abnormal AWB ciliary structures but otherwise normal localization of ODR-10::GFP to the ciliary membrane. (L) MKS-3::GFP localizes normally at the TZ in nphp-1 and mks-1 mutants but accumulates abnormally inside cilia (cil) and at dendritic tips (DT) in mks-5, mksr-1, and mksr-2 mutants. Bar, 2.5 µm. (M) GFP-tagged RPI-2 is found at dendritic tips but not inside cilia, marked by XBX-1::tdTomato in wild-type worms. RPI-2 localizes normally in mks-1, mks-3, and nphp-1 single mutants but accumulates within cilia of mks-3;mks-1 double mutants and mksr-1, mksr-2, mks-5, mks-6, and nphp-4 single mutants. Bar, 2.5 µm. (N) tdTomato-tagged TRAM-1a is found at dendritic tips but not inside cilia (ciliary TZ marked by MKS-3::GFP) in wild-type worms. TRAM-1a localizes normally in mks-1 and mks-3 mutants but accumulates within cilia of mksr-1, mksr-2, mks-5, mks-6, nphp-1, and nphp-4 mutants. Bar, 2.5 µm.

Figure 10.

MKS/MKSR and NPHP proteins form part of a functional interaction network required for an early stage of ciliogenesis and formation of an intact ciliary gate. (A) Previously identified physical interactions between C. elegans or mammalian TZ proteins (see also Table S1 A). Interactions between MKS1 and MKS3 and between NPHP1, NPHP4, and MKS5 were identified in mammals. The interaction between MKSR1 and MKSR2 was identified in C. elegans. (B) Summary of hierarchy analysis uncovering the requirement of some TZ proteins for normal localization of others. Black arrows represent previously known requirements (arrows point away from the protein required for localization of the other). Red arrows represent novel requirements uncovered in this study. (C) Summary of functional (genetic) interactions between C. elegans TZ components influencing ciliogenesis. Black lines represent previously known functional interactions, and red lines represent genetic interactions uncovered in this study (dye-filling/ciliogenesis and/or BB docking/anchoring and ciliary structures, as revealed by TEM). Modular assignments based on these functional interactions and hierarchy analysis data are represented by stars. (D) Model depicting early and late events in ciliogenesis, first involving transition zone (TZ) proteins and then IFT-associated proteins. Top pathway: in some mammalian cell types, the earliest step of ciliogenesis involves binding of a ciliary vesicle (CV) to the mother centriole, followed by migration of the centriole-CV toward the plasma membrane (PM). At some point a ciliary bud (likely to be a maturing TZ) emerges, invaginating the CV. Bottom pathway: in other cells, the CV may be expendable, and docking of the centriole/developing TZ occurs directly with the PM. During these early steps, interactions between the BB and TZ with the membrane depend on MKS/MKSR/NPHP TZ proteins. Assembly of the TZ in early ciliogenesis likely coincides with formation (of at least part) of the ciliary gate, and MKS/MKSR/NPHP proteins are functional/structural elements of this mechanism. IFT and BBS proteins are not necessary for formation of the TZ or for BB/TZ anchoring to the membrane during early ciliogenesis, but after these events IFT proteins participate in building (and maintaining) the rest of the ciliary axoneme, and BBS proteins participate in the delivery of ciliary cargo.