TMEM237 is mutated in individuals with a Joubert syndrome related disorder and expands the role of the TMEM family at the ciliary transition zone

Abstract

Joubert syndrome related disorders (JSRDs) have broad but variable phenotypic overlap with other ciliopathies. The molecular etiology of this overlap is unclear but probably arises from disrupting common functional module components within primary cilia. To identify additional module elements associated with JSRDs, we performed homozygosity mapping followed by next-generation sequencing (NGS) and uncovered mutations in TMEM237 (previously known as ALS2CR4). We show that loss of the mammalian TMEM237, which localizes to the ciliary transition zone (TZ), results in defective ciliogenesis and deregulation of Wnt signaling. Furthermore, disruption of Danio rerio (zebrafish) tmem237 expression produces gastrulation defects consistent with ciliary dysfunction, and Caenorhabditis elegans jbts-14 genetically interacts with nphp-4, encoding another TZ protein, to control basal body-TZ anchoring to the membrane and ciliogenesis. Both mammalian and C. elegans TMEM237/JBTS-14 require RPGRIP1L/MKS5 for proper TZ localization, and we demonstrate additional functional interactions between C. elegans JBTS-14 and MKS-2/TMEM216, MKSR-1/B9D1, and MKSR-2/B9D2. Collectively, our findings integrate TMEM237/JBTS-14 in a complex interaction network of TZ-associated proteins and reveal a growing contribution of a TZ functional module to the spectrum of ciliopathy phenotypes.

Figure 1

Mutations in TMEM237 Cause a Joubert Syndrome Related Disorder (A) Chromosomal location of the overlapping JSRD loci of the Hutterite families and the Austrian family. (B) Targeted sequence capture followed by massively parallel sequencing and data analysis identified a homozygous mutation in TMEM237 (ALS2CR4) in an affected Hutterite individual. The black box shows the G>A change at the genomic level. The red box shows the resultant amino acid change (R>X). The green highlights show that all the sequence reads from the affected child contain an A. (C) TMEM237 genomic organization. (D) Domain structure of TMEM237 protein. Yellow indicates the transmembrane domains. Grey indicates the extramembrane region. Red arrows and blue arrows indicate the highly conserved short motifs of basic and acidic amino acids, respectively. Black arrows indicate the positions of identified mutations. (E) Sequence traces of normal controls (top) and affected individuals (bottom) are shown.

Figure 2

TMEM237 Is Localized to the Ciliary Transition Zone Region and Mutation or Knockdown Results in Defective Ciliogenesis (A) Overlapping localization of endogenous N-terminal (N-term) or C-terminal (C-term) mouse Tmem237 (red) and polyglutamylated (polyglut.) tubulin (green) at the base of primary cilia and/or mother centriole (arrowheads in magnified insets) of IMCD-3 cells. The scale bar represents 5 μm. (B) Two different TMEM237-mutated fibroblasts lines (JSRD1 with homozygous splice-site mutation c.677+1G>T and JSRD2 with homozygous nonsense mutation c.52C>T as indicated) show defective ciliogenesis and pairing of centrioles (marked by γ-tubulin, red) compared to a normal control line. The scale bar represents 5 μm. Ciliogenesis is quantitated in the lower panel (n = 500 to 600 for each line; statistical significance of comparisons are shown (^∗∗p < 0.01, ^∗∗∗∗p < 0.0001 Student t test); error bars indicate standard error of the mean (SEM). (C) Tmem237 antisera against the C-terminal mouse protein detects a specific 45 kD band in immunoblot of control cells transfected with scrambled (scr.) siRNA that is lost after transfection with Tmem237 siRNA duplexes (dup.) 1 or 2. Equal loading is indicated by immunoblot analysis for β-actin. (D) Defective ciliogenesis in IMCD-3 cells after Tmem237 siRNA knockdown with duplexes (dup.) 1 or 2. Cells stained and ciliogenesis quantified as in (B); the scale bar represents 10 μm.

Figure 3

Deregulation of Wnt Signaling Following Mutation or Loss of TMEM237 (A) Immunoblots showing the relative levels of the indicated endogenous mediators of both canonical and noncanonical-PCP Wnt signaling in normal control fibroblasts compared to the TMEM237-mutated fibroblast line (JSRD2 c.52C>T line). Identical results were seen for the JSRD1 c.677+1G>T line (data not shown). (B) Increase in levels of activated RhoA-GTP after Tmem237 siRNA knockdown with duplexes (dup.) 1 or 2, compared to scrambled (scr.) siRNA. Levels of β-actin are shown as loading controls. Positive control for the assay (loading with nonhydrolyzable GTPγS) and a negative control (loading with GDP) are shown on the left. (C) Pericentriolar and basolateral subcellular localization of RhoA (top panels) and increase of actin stress fibers (bottom panels; visualized by phalloidin staining, green) in TMEM237-mutated fibroblast lines; the scale bars represent 10 μm. (D) TOP Flash assays of canonical Wnt signaling, showing deregulation in affected fibroblasts (JSRD2 c.52C>T, red) compared to normal control cells (green) after treatment with 0.5 × L cell control-conditioned media (control), and conditioned media containing expressed Wnt3A and/or Wnt5A, as indicated. Activity is expressed as ratios of luciferase reporter construct expression, normalized for loading by measurement of a Renilla construct expression. Values shown are means of three independent replicates; error bars indicate SEM. (E) TOP Flash assays of canonical Wnt signaling after cotransfection of Mks3^+/+ and Mks3^−/− MEFs with reporter constructs and empty vector control, wild-type GFP-TMEM237, HA-MKS3, or GFP-TMEM216 as indicated. The empty vector results combine the data from cotransfections with pCMV-HA and pGFPN1. Responses are shown to L cell control-conditioned media (control, blue bars, values are means of three independent replicates) and conditioned media containing expressed Wnt3A (0.5 × Wnt3A, brown bars, values are means of four independent replicates). Error bars indicate SEM. Statistical significance of comparisons are shown (^∗p < 0.05, ^∗∗p < 0.01, Student t test).

Figure 4

Zebrafish Morphant Phenotypes and Functional Complementation of TMEM Proteins (A) Representative live embryo images of tmem237 morphants at 8–10 somites (top, lateral view; bottom, dorsal view). Class I and II embryos were scored on the basis of previously established objective criteria., , Class I, short anterior-posterior axis, small anterior structures, and mild somitic defects; Class II, severe defects in head and eye development, shortening of the body axis, kinked notochords, and broadening of the somites; Class III, severe thinning and shortening of the body axis, curved notochord, broadened and thin somites, and an ovoid embryonic shape suggestive of dorsalization—embryos were severely delayed in development and typically did not survive past the ten somite stage. (B) Zebrafish embryo flatmounts at the six somite stage hybridized with krox20 and myoD riboprobe cocktails. Representative embryos are shown: control MO (Ctr) injected (left); tmem237 MOs (MO) injected (right). Gastrulation defects were measured in two dimensions (medial-lateral and anterior-posterior) by the length of the 5th rhombomere (horizontal lines) and notochord (vertical lines), respectively. (C) Quantification of the mean length of the 5th rhombomere or notochord of morphants (n = 29) are shown as the percentage of the mean length of the controls (n = 22). t test, ^∗p < 0.01; ^∗∗p < 0.001; error bars indicate SEM. (D) Coinjection of tmem237 MOs with human TMEM237 mRNA, MKS3 mRNA, or TMEM216 mRNA significantly rescued morphant phenotypes at eight to ten somites. (E) TMEM237 can partially ameliorate mks3 suppression defects of midsomitic embryos. Comparison of rescue efficiency of human MKS3 versus TMEM237 mRNA is shown. (F) TMEM237 mRNA rescues tmem216 suppression phenotypes in a manner indistinguishable from TMEM216 mRNA coinjection in midsomitic embryos (^∗∗∗p < 0.0001; NS, not significant for D and F).

Figure 5

Ciliopathy TMEM Proteins Are Anchored at the Transition Zone by MKS5/RPGRIP1L (A) In scrambled (scr.) siRNA-treated IMCD-3 cells, Tmem237 (red channel) is at proximal regions of the cilium and partially colocalizes with γ-tubulin (green channel) at the basal body, consistent with Tmem237 localization to the TZ. Colocalization with γ-tubulin is disrupted after Rpgrip1l siRNA knockdown with duplexes (dup.) 1 or 2. Insets show magnified regions (indicated by arrowheads on the main image) for the red and green channels only; the scale bars represent 5 μm. (B) Top panels: Dvl-1 (green channel) is at the basal body that colocalizes with γ-tubulin (red channel) in IMCD-3 cells, as described previously.Rpgrip1l siRNA knockdown does not disrupt overall basal body architecture. Bottom panels: in scrambled (scr.) siRNA-treated IMCD-3 cells, Tmem237 partially colocalizes with Dvl-1, but this is disrupted after Rpgrip1l siRNA knockdown. (C) As above, but with RhoA staining; magnified insets are indicated by white frames.

Figure 6

C. elegans JBTS-14 Functionally Interacts with NPHP-4 to Promote Ciliogenesis and Forms Part of a Functional Assembly of Proteins at the Transition Zone that Includes MKSR-1/B9D1, MKSR-2/B9D2, MKS-2, and MKS-5/RPGRIP1L (A) C. elegans JBTS-14 and MKS-2 (GFP-tagged) localize specifically to the transition zone (TZ) in amphid (head) and phasmid (tail) cilia. The TZ is immediately distal to the basal body (BB) on the ciliary axoneme, both of which are marked by the tdTomato-tagged XBX-1 intraflagellar transport protein. (B) Disruption of jbts-14 or mks-2 alone does not cause overt ciliogenesis defects, as judged by the ability of worms to take up fluorescent dye through environmentally-exposed cilia. In contrast, the jbts-14;nphp-4, or mks-2;nphp-4 double-mutant animals are dye-filling defective, indicative of cilia structure defects. The osm-5 ciliary mutant, which lacks cilia altogether, is included as a control. (C) jbts-14 or mks-2 single mutants can avoid a high osmolarity solution, which depends on functional cilia, whereas the jbts-14;nphp-4 or mks-2;nphp-4 double mutants are defective in this chemosensory behavior. Chi-square test, ^∗p<0.0001; error bars indicate SEM. (D) The localization of JBTS-14 (left panels) and MKS-2 (right panels) depends on the presence of several other TZ proteins but not others, as indicated by yes (properly localized) or no and a dotted ellipse for a mislocalized protein. The following abbreviations are used: BB, basal body; cil, cilium; TZ, transition zone; den, dendrite. The scale bar represents 5 μm.

Figure 7

Basal Body-Transition Zone Attachment to the Ciliary Membrane Is Disrupted in Mutant Strain Lacking JBTS-14 and NPHP-4 (A) Low and high-magnification images of C. elegans amphid channel sensory cilia from TEM serial cross sections of the amphidal pore. Top row: images from the distal pore region. Boxed numbers denote proximal positioning of sections relative to the top row sections. In wild-type (N2) worms, the amphid pore contains ten axonemes, each consisting of a distal segment (top row: nine singlet microtubules), a middle segment (second row: doublet microtubules) and a transition zone (TZ) (third row: ring of doublet microtubules drawn together by the apical ring and connected to the ciliary membrane via Y-links). At the base of the cilia, within the far distal dendrite region, are transition fibers and occasional vesicles. In jbts-14;nphp-4 mutants, two to three axonemes are typically missing from the distal pore (top row, left image) and occasionally from more proximal regions (+4; left image), microtubule number is sometimes elevated (+4; right image), and Y-links are consistently missing or reduced from the transition zone compartment (+7; middle and right images). In distal dendrite regions, vesicles and membranes are often abnormally accumulated (+8; middle image) and incomplete, ectopically positioned, ciliary axonemes are occasionally observed (+8; right image). The open arrowhead indicates a transition fiber. The scale bars represent 200 nm. (B) High-magnification images of transition zone phenotypes in the jbts-14;nphp-4 double mutant presenting multiple examples of the phenotypes described above, as well as an image showing an undocked TZ in the distal dendrite region (+10). The arrow indicates an apical ring; the closed arrowhead, Y-links; and the open arrowhead, inner microtubules. Images taken from three independently sectioned worms. Boxed numbers denote proximal positioning of sections relative to the distal end of the amphidal pore. Schematics to the right of each row depict normal and abrogated ciliary structures in wild-type and jbts-14;nphp-4 mutant worms, respectively. The scale bars represent 200 nm. (C) Schematics of amphid channel cilia (longitudinal and transverse views) showing the major ultrastructural defects observed, including the loss of TZ connections to the ciliary membrane via Y-links. TZ subcompartment (microtubules and Y-links) is highlighted in green, dendritic membrane is highlighted in red.