Normal ciliogenesis requires synergy between the cystic kidney disease genes MKS-3 and NPHP-4

Abstract

Cilia dysfunction contributes to renal cyst formation in multiple human syndromes including nephronophthisis (NPHP), Meckel-Gruber syndrome (MKS), Joubert syndrome (JBTS), and Bardet-Beidl syndrome (BBS). Although genetically heterogeneous, these diseases share several loci that affect cilia and/or basal body proteins, but the functions and interactions of these gene products are incompletely understood. Here, we report that the ciliated sensory neurons (CSNs) of C. elegans express the putative transmembrane protein MKS-3, which localized to the distal end of their dendrites and to the cilium base but not to the cilium itself. Localization of MKS-3 and other known MKS and NPHP proteins partially overlapped. By analyzing mks-3 mutants, we found that ciliogenesis did not require MKS-3; instead, cilia elongated and cilia-mediated chemoreception was abnormal. Genetic analysis indicated that mks-3 functions in a pathway with other mks genes. Furthermore, mks-1 and mks-3 genetically interacted with a separate pathway (involving nphp-1 and nphp-4) to influence proper positioning, orientation, and formation of cilia. Combined disruption of nphp and mks pathways had cell nonautonomous effects on C. elegans sensilla. Taken together, these data demonstrate the importance of mutational load on the presentation and severity of ciliopathies and expand the understanding of the interactions between ciliopathy genes.

Figure 1.

Expression and DAF-19 regulation of mks-3. (A through C) Confocal fluorescence images of worms coexpressing mks-3::GFP. (A) mks-3::GFP was strongly expressed in amphid (arrowhead) and labial (arrow) sets of neurons of the head, which both extend dendritic processes to the (double arrowhead) anterior of the worm where they project sensory cilia. (B) mks-3 transgene expression was almost completely abolished in worms crossed into daf-19(m86) mutant background. mks-3::GFP was faintly detected only in labial neurons (arrow). (C) Outcrossing transgenic worms from daf-19(m86) to daf-19(+) restored expression of mks-3::GFP.

Figure 2.

MKS-3 concentrates at the base of cilia in C. elegans. (A) An illustration depicting the anatomical positions of (left) amphid cilia bundles in the head and (right) phasmid cilia bundles in the tail with respect to (middle) a differential interference contrast image of the full body of an adult hermaphrodite worm. Red represents cilia axonemes, green represents bases of cilia, and blue represents the dendritic processes from which the cilia project. Short labial cilia surrounding the distal region of the amphid bundles and the degenerate PQR cilium near the phasmid bundles are not shown. (B through D) Confocal fluorescence images of wild-type transgenic worms coexpressing MKS-3::GFP (driven by its endogenous promoter) and XBX-1/dynein light intermediate chain::tdTomato (driven by the osm-5 promoter). With respect to XBX-1, which freely entered the cilia (arrowheads), MKS-3::GFP concentrated specifically at the base of cilia (arrows) on (B and D, left column) sensory neurons in the head and (C and D, right column) sensory neurons in the tail. (E) Confocal fluorescence images of wild-type transgenic worms coexpressing MKS-3::GFP and MKS-1::tdTomato (driven by its endogenous promoter). With respect to MKS-3::GFP, MKS-1::GFP localized only to the proximal end of cilia where MKS-3::GFP concentrated most intensely (arrows). Scale = 5 μm.

Figure 3.

Localization of MKS-3 in the dendritic tip resembles that of the membrane-associated protein RP2. Single-section confocal fluorescence images of wild-type transgenic worms expressing (A) RP2::GFP (driven by its endogenous promoter) or (B) MKS-3::GFP. One phasmid neuronal ciliated ending is shown in each panel. RP2::GFP accumulated around the edges of the dendritic tip (dt), indicative of membrane association, but was not present at the cilium base (cb). MKS-3::GFP also accumulated around the edges of the dendritic tip in addition to concentrating at the cilium base. Scale = 5 μm.

Figure 4.

mks-3(tm2547) mutants dye-fill normally. The ability to uptake DiI was examined in wild-type worms and mks-3(tm2547) mutants. Confocal fluorescence images of DiI staining are shown. Indicative of properly formed cilia, (left) wild-type and (right) mks-3(tm2547) worms consistently absorbed DiI into CSNs in the head (large panels) and tail (insets). The four phasmid cell bodies are outlined in the insets. Scale = 7 μm.

Figure 5.

mks-3 mutants exhibit reduced chemotaxis toward benzaldehyde. Compared with wild type, mks-3(tm2547) and nphp-4(tm925) single mutants were mildly but significantly defective in response to a volatile attractant. Severity of mks-3;mks-1;nphp-4 triple mutant chemotaxis deficiency was similar to retrograde IFT che-11(e1810) mutants, which were used as a negative control. Each group was assayed at least four times with 50 to 120 worms per run. Error bars represent SD. No significant difference was observed in groups represented by bars of matching color. P < 0.05 was deemed significant.

Figure 6.

mks-3;nphp-4 double mutants have short and incorrectly positioned cilia. Cilia morphology was analyzed in confocal fluorescence images of wild-type, mks-3(tm2547), mks-3;mksr-2, and mks-3;nphp-4 worms expressing the transgenic IFT particle B protein CHE-13::YFP. (A and B) Cilia morphology appeared overtly normal in wild-type (top), mks-3(tm2547) (middle), and mks-3;mksr-2 (bottom) strains. Arrowheads, distal ends of cilia; double arrowheads, cilia bases. (A′ and C′) 1.66× magnification of amphid bundles in A and C, respectively. Arrows indicate the shorter labial cilia that are not part of the amphid channel organs. (B) Compared with wild type (top), mutant strains mks-3(tm2547) and mks-3;mksr-2 exhibited no abnormalities in the extension of dendrites (dotted lines) from the phasmid cell bodies (arrows) nor in the projection of full-length cilia (double-headed arrow) from the distal tips of the dendrites. (C) Amphid and (D) phasmid cilia arrangement and morphology was abnormal in mks-3;nphp-4 double mutants. Compared with wild type, some amphid cilia were positioned posterior with the respect to the rest of the sensory organ (C′, arrowhead), and other amphid cilia often altogether lacked axonemes (arrow). (D) Phasmid dendrites often failed to properly extend from the cell bodies and projected stunted cilia. Scale = 5 μm.

Figure 7.

Partial restoration of phasmid dye-filling in mks-3;nphp-4 double mutants by transgene rescue. The ability to uptake DiI was examined in wild-type worms, mks-3(tm2547);nphp-4(tm925) double mutants, and transgenic mks-3(tm2547);nphp-4(tm925) double mutants coexpressing MKS-3::tdTomato and the coelomocyte marker UNC-122::GFP from a nonintegrated extrachromosomal array (Ex1). Phasmid dye-filling was regularly observed in wild-type worms but rarely seen in mks-3(tm2547);nphp-4(tm925) double mutants. Progeny of mks-3(tm2547);nphp-4(tm925) double mutants coexpressing MKS-3::tdTomato and UNC-122::GFP were exposed to DiI and then siblings were segregated as transgenic versus nontransgenic on the basis of the presence or absence of UNC-122::GFP signal, respectively. Partial rescue of phasmid dye-filling was observed in transgenic double mutants compared with nontransgenic siblings. At least 100 worms in each category were analyzed. Error bars represent SEM.

Figure 8.

Cell nonautonomous perturbation of phasmid sensilla morphology in mks-3;nphp-4 double mutants. Illustrations of (A) amphid and (D) phasmid sensilla showing attachment of cuticle to socket channels, connections between socket channels and sheath cells (double arrowheads), and connections between sheath cells and dendritic tips (den) (arrowheads). Scale = 1 μm. Modified with permission from Perkins et al. (B, C, E, and F) Confocal fluorescence images of worms coexpressing f16f9.3::GFP to mark sheath cells and XBX-1::tdTomato to mark cilia. Compared with (B) wild type, amphid sensory neurons in (C) mks-3;nphp-4 double mutants projected malformed cilia that remained in association with the surrounding sheath cells. Despite apparent detachment from the phasmid channels (dotted lines mark predicted location of channels), phasmid sheath cells remained in association with the misplaced ciliated endings of phasmid neurons (arrowheads) in (F) mks-3;nphp-4 double mutants. (G, H) Confocal fluorescence images of worms coexpressing MAGI-1::GFP to mark socket channels and XBX-1::tdTomato to mark cilia. In (G) wild-type worms, socket cell-expressed MAGI-1::GFP localizes around the distal-most segments of phasmid cilia (arrows), consistent with the location of the socket-derived portion of the phasmid channels. In (H) mks-3;nphp-4 double mutants, socket channels were present. (Left) A pair of short phasmid cilia positioned normally with respect to the socket channel. (Right) Socket channels were most often present (arrow) despite misplacement of cilia, although on rare occasions no socket-expressed MAGI-1::GFP was observed (arrowhead). Scale of all fluorescence images = 5 μm.