The Meckel syndrome protein meckelin (TMEM67) is a key regulator of cilia function but is not required for tissue planar polarity

Abstract

Meckel syndrome (MKS) is a lethal disorder associated with renal cystic disease, encephalocele, ductal plate malformation and polydactyly. MKS is genetically heterogeneous and part of a growing list of syndromes called ciliopathies, disorders resulting from defective cilia. TMEM67 mutation (MKS3) is a major cause of MKS and the related ciliopathy Joubert syndrome, although the complete etiology of the disease is not well understood. To further investigate MKS3, we analyzed phenotypes in the Tmem67 null mouse (bpck) and in zebrafish tmem67 morphants. Phenotypes similar to those in human MKS and other ciliopathy models were observed, with additional eye, skeletal and inner ear abnormalities characterized in the bpck mouse. The observed disorganized stereociliary bundles in the bpck inner ear and the convergent extension defects in zebrafish morphants are similar to those found in planar cell polarity (PCP) mutants, a pathway suggested to be defective in ciliopathies. However, analysis of classical vertebrate PCP readouts in the bpck mouse and ciliary organization analysis in tmem67 morphants did not support a global loss of planar polarity. Canonical Wnt signaling was upregulated in cyst linings and isolated fibroblasts from the bpck mouse, but was unchanged in the retina and cochlea tissue, suggesting that increased Wnt signaling may only be linked to MKS3 phenotypes associated with elevated proliferation. Together, these data suggest that defective cilia loading, but not a global loss of ciliogenesis, basal body docking or PCP signaling leads to dysfunctional cilia in MKS3 tissues.

INTRODUCTION

Cilia are highly conserved microtubule-based organelles that protrude from the apical surface of most cells. The core structure of cilia, the axoneme, is nucleated and extended out of a modified centriole called the basal body. The axoneme is a ring of nine microtubule doublets that is ensheathed by a membrane contiguous with the plasma membrane (1–4). Cilia can be either motile or non-motile. Specialized motile cilia are found at the embryonic node and generate a leftward fluid flow responsible for left–right body symmetry. Motile cilia are also present in respiratory epithelia, brain ependymal cells and epithelial cells lining the fallopian tubes, where they are responsible for generating flow (3). Primary cilia are specialized immotile cilia, present at a ratio of one per cell, and are thought to have mechanosensory functions. In addition, primary cilia may play a role in mediating signaling events involved in development, such as Sonic Hedgehog (Shh), Wnt signaling (canonical and non-canonical) and cell-cycle signaling events (1,2,5).

Primary cilium disruption can lead to the development of a set of disorders called ciliopathies. The ciliopathies encompass several syndromes including nephronophthisis (NPHP), Joubert syndrome (JBTS), Bardet Biedl syndrome (BBS) and Meckel syndrome (MKS) (6). Proteins associated with ciliopathies are primarily localized to the basal body or primary cilium, suggesting the proteins share common functions to preserve cilia and basal body structure, function and protein composition (2,7–10). The ciliopathies have a high degree of genetic and phenotypic overlap, especially NPHP, JBTS and MKS, suggesting that these syndromes represent variants of a related disease with differing phenotypic severity, dependent on allelic effects, genetic complexity and genetic background (1,4,6,11–13).

MKS is the most severe ciliopathy, being embryonic lethal by definition. It is relatively rare in outbred populations with the most common phenotypes being renal cystic dysplasia, central nervous system defects, usually occipital encephalocele, hepatic ductal plate malformations and polydactyly (4,14,15). At least 10 loci have been associated with typical MKS, including MKS1 (16), TMEM216 (17), TMEM67 (MKS3, Meckel syndrome type 3, OMIM 607361) (18), CEP290 (19), RPGRIP1L (20), CC2D2A (21), TCTN2 (22), B9D1 (23), B9D2 (24) and NPHP3 (25). Our studies focus on the TMEM67 locus, which is one of the most common causes of MKS and is also associated with JBTS and NPHP (26–29). The resulting product, meckelin, is a proposed 7-transmembrane receptor-like protein with few defined domains, but contains an extracellular cysteine-rich region (18). A number of protein complexes containing MKS, JBTS and NPHP proteins have recently been described (7,8,12,30–32). These complexes have been localized to a specialized region of the cilium called the transition zone and are proposed to regulate ciliary membrane protein composition (7,11). Meckelin has been localized to one of these complexes, the ‘MKS-B9’ complex. In addition, meckelin depletion was shown to regulate centrosome number and cilia length in animal and cell models (33,34).

Several signaling pathways, including the developmental Wnt signaling cascades, are suggested to require a functional cilium for normal levels of signaling; however, the pathway relationship with PKD and the ciliopathies remains controversial (35). Wnt signaling has two major arms, the canonical (Wnt) and non-canonical pathways, including the planar cell polarity (PCP) pathway. Canonical Wnt signaling is dependent on the transcription factor β-catenin, which functions as a transcription co-factor to regulate cell proliferation and differentiation events (36). Canonical Wnt signaling has been described as upregulated in PKD and transgenic mice overexpressing an activated form of β-catenin develop cystic disease, linking elevated canonical Wnt signaling to increased proliferation rates and cystogenesis (37–40). The primary cilium may modulate canonical Wnt signaling levels through sequestration of signaling components such as β-catenin at the basal body, causing decreased nuclear translocation and transcriptional activity (39).

The PCP pathway, in contrast, is suggested to be downregulated in PKD and the ciliopathies. PCP controls cellular polarization within a plane perpendicular to the apical-basal axis and employs cellular interactions to coordinate polarization among neighboring cells (41,42). A set of core PCP proteins were identified from Drosophilia studies including: frizzled (fz), dishevelled (dsh), prickle (pk), van gogh/strabismus (vang/stbm), flamingo (fmi) and diego (dgo) (43). These proteins become asymmetrically distributed to the distal and proximal cell membranes for planar polarization to occur (42,44). A set of homologous PCP proteins in vertebrates are involved in convergent extension (CE), which controls polarized cell movements leading to narrowing and lengthening of tissues, including elongation of the anterior–posterior axis during gastrulation and neural tube development (45,46). Besides CE, several vertebrate phenotypes have been used to assess disruption of PCP signaling, including uniform orientation of actin-based stereociliary bundles and the kinocilium, a specialized primary cilium in the organ of Corti (OC) of the inner ear (46). In addition, PCP has also been shown to play important roles in orienting the axis of mitotic spindles during cell division in developing tissues (47). Defective mitotic spindle orientation in developing renal tubules of rodent PKD models has been linked to defective PCP and cystogenesis; however, the importance for cyst initiation remains controversial (42,48–50).

In this study, we utilized zebrafish tmem67 ciliopathy morphants and the null Tmem67 (bpck) mouse model (51) to explore the role of meckelin in cilia maintenance and functionality in various organs, including examination of potential defects in canonical Wnt and PCP output.

RESULTS

Renal phenotypes in the bpck mouse model

To better understand the tissue-specific requirement for meckelin in cilia maintenance and functioning, we first examined the extent and progression of kidney cystogenesis in the bpck mouse, a spontaneous mutant with all of Tmem67 removed (51). The bpck phenotypes are similar to human MKS3 and the wpk rat model of this disease (18,26,27,52). However, many MKS3 phenotypes, especially the extrarenal phenotypes, have not been characterized in detail. For these studies, we inbred the original bpck mouse from a B6C3FeF1/Ja/a mixed background into C57BL6/J and compared the severity of cystic disease to the original outbred background. In the outbred state, mild cystic tubule dilation was observed as early as embryonic day (E)16.5 in bpck mice (Fig. 1A). Cysts became more pronounced by postnatal day (P)1, with cystogenesis in the inbred animal more severe than the outbred state (Fig. 1B and D). Comparison of disease severity near the end of lifespan of outbred (P16) and inbred mice (P13; Fig. 1C and E) showed widespread cystogenesis that was more extensive in the younger inbred animal. Staining for cystic tubule origin at P12 showed that the largest cysts were derived from the collecting duct, with occasional dilation of the proximal tubule and loop of Henle (Fig. 1F and G; average cyst diameter: collecting ducts = 1018 μm, 1889 tubules; proximal tubules = 470 μm, 1113 tubules; loop of Henle = 84 μm, 1078 tubules). Additionally, bpck mice displayed a survival disadvantage in both outbred and inbred states. While the level of outbred bpck mice was normal at P1, it dropped well below the expected 25% by P4, while in the inbred state survival dropped below the expected level beginning at E18.5 with only rare liveborn mutants (Fig. 1H and I, Supplementary Material, Table S1).

Figure 1.

Depletion of meckelin in the bpck mouse leads to polycystic kidney disease (PKD). Mallory SS trichrome stained kidney sections at high (A–C) and low magnification (D, E) of (A) E16.5, (B, D) P1 and (C, E) P16 (outbred), P13 (inbred) comparing wild-type and inbred and outbred bpck. Arrowheads in (A) indicate dilated tubules. Scale bar in (A–C) is 20 μm and 100 μm in (D) and (E). (F) LTA (lotus tetragonolobus agglutinin; proximal tubules) and THP (Tamm-Horsfall protein; loop of Henle) staining, plus (G) DBA (Dolichos biflorus; collecting ducts) labeling in P12 outbred bpck kidney sections (scale bar = 100 μm). Survival analysis of (H) outbred and (I) inbred bpck mutants (Supplementary Material, Table S1). Insufficient numbers of inbred P13 bpck mice survived for accurate analysis. Statistics are based on a two-tailed Fisher's exact test. *P < 0.05.

Extrarenal phenotypes in bpck mice

Meckelin requirement for cilia maintenance was next examined in extrarenal tissues. Retinal sections from P16 outbred bpck animals displayed severe retinal degeneration, as seen by thinning of both the inner and outer retinal segments (Supplementary Material, Fig. S1A). Higher resolution analysis of the tissue by transmission electron microscopy (TEM) revealed a general disorganization of retinal layers in bpck mice with no structurally identifiable outer segment, as seen with the disc-shaped rod cells in the wild-type retina (Fig. 2A). However, the connecting cilium was present in bpck retinas and appeared structurally similar to the wild-type (Fig. 2A, inserts). Functionality of the connecting cilium was examined via transport of the G-coupled protein receptor rhodopsin in retinal segments. In wild-type tissue, rhodopsin was transported to the outer segment; however, in the bpck retina, rhodopsin accumulated in the inner segment and outer nuclear layer (Fig. 2B). Apoptotic rates were increased in the outer nuclear layer and inner and outer segments of the retina in P12 and P16 bpck mice (Supplementary Material, Fig. S1B and C), suggesting that the loss of protein transport along the connecting cilium leads to death of retinal cell layers, and underlies the degenerative phenotype.

Figure 2.

bpck mice display extrarenal abnormalities. (A) TEM images of P12 outbred retinal sections (scale bar = 2 μm). Inserts show the connecting cilium (scale bar = 0.5 μm). OS, outer segment. (B) Rhodopsin transport to the outer segment (arrowhead) in P12 mice. ONL, outer nuclear layer; IS, inner segment (double arrowheads). Scale bar = 20 μm. (C) Skeletal staining of E16.5 mice with lateral view of the hindlimbs with bowed fibula in mutant (arrowhead). Fe, femur; Ti, tibia; Fb, fibula. Scale bar = 2 mm. Alizarin red staining indicates mineralized bone and alcian blue shows cartilage. (D) Quantitation of bone lengths in the hindlimb (n = 3 wild-type, 2 bpck). (E) TEM images of P1 respiratory cilia in mouse trachea. Scale bar = 0.5 μm. (F) Quantitation of respiratory cilia density in P1 mouse trachea (n = 2 wild-type, 3 bpck). Statistics in (D, F) based on Student's t-test.

Skeletal abnormalities such as postaxial polydactyly, clubfeet and short limbs have been documented in MKS, although may be rare in MKS3 patients (14,26,27). In addition, mouse models of Mks1 display a range of skeletal deformations, prompting us to analyze meckelin functioning during skeletal development (53,54). Gross examination of E16.5 skeletons showed no overt malformations in bpck mice, including normal ossification of vertebrate bodies and no clear evidence of cleft palate (Supplementary Material, Fig. S1D–F). Forelimb development was also normal, with no evidence of polydactyly (5/5 bpck mice; Supplementary Material, Fig. S2G and H). However, fibula bones had a significant bowing in all mutants examined (5/5), and the fibula, tibia and femur were slightly shorter than the wild-type, although not reaching statistical significance (Fig. 2C and D). This verifies that although Mks1 and meckelin may be involved in similar complexes, they play divergent roles in different primary cilia-dependent developmental processes (26,27,33).

Lastly, the function of meckelin in cilia maintenance in multiciliated respiratory cells in P1 mouse trachea was examined. Meckelin removal did not affect docking of the basal body onto the apical cell surface in bpck mice, and the cilia appeared structurally normal compared with the wild-type (Fig. 2E). The average density of respiratory cilia was not significantly different between mutants and wild-type littermates, although cilia density was more variable in mutant mice (Fig. 2F; wild-type sd = 0.215, bpck sd = 1.65). Our results demonstrate that basal body trafficking and docking are largely unaffected by meckelin loss, at least in multiciliated airway epithelial cells. This is in contrast to previous in vitro studies where transient meckelin depletion led to apparent apical basal body docking defects (34,55).

Tmem67 deletion results in ciliogenesis and hair cell orientation defects in the OC

We next investigated the effects of meckelin depletion on formation and maintenance of another specialized primary cilium, the kinocilium, in the inner ear of bpck mice. Oriented alignment of the ‘V’-shaped stereociliary bundles in sensory hair cells in the cochlea of the inner ear is a robust demonstration of planar polarization of the underlying sensory epithelium. Thus, deregulation of hair cell alignment in other ciliopathy models has been interpreted as a PCP-dependent defect (56). Hair cell orientation was examined in cochleae from P1 littermates by scanning electron microscopy (SEM). Cell alignment was analyzed by measuring the degree of rotation of the stereociliary bundle from the lateral tissue edge in the apical, middle and basal regions of the OC, the sensory portion of the cochlea. We found the gross alignment of stereocilia across the tissue to be largely intact, although frequent morphological and orientation defects were observed within individual hair cells (Fig. 3A–D and F–H). Hair cells in apical regions of the OC were the most disorganized in bpck mice, with many stereociliary bundles displaying abnormal or flattened morphology (Fig. 3A and B), with statistically significant stereociliary alignment defects (Fig. 3F). Extra rows of hair cells were present in this region and throughout the OC, as observed in PCP mouse mutants, likely due to cochlear CE defects (Fig. 3B) (57,58). However, the percentage of hair cells present in extra rows in the bpck OC did not reach statistical significance compared with wild-type controls (Fig. 3E), showing that the bpck OC is not as disorganized as in PCP mutants with CE defects. Cells were also disorganized in the middle and basal OC, with some stereociliary bundles having an abnormal ‘U’-shaped morphology (Fig. 3C and D). Rotational defects also occurred in middle and basal OC regions, but were not as severe as in the apex (Fig. 3G and H).

Figure 3.

Meckelin disruption leads to ciliogenesis and ‘PCP-like’ phenotypes in the OC of the inner ear. Hair cells in the three outer hair cell (OHC) rows were analyzed by SEM in all regions of the OC in P1 mice. Rows of OHCs in (A) wild-type and (B) bpck apical regions of the OC. L, lateral edge of the tissue. Arrows indicate abnormally shaped stereociliary bundles and cells with rotational defects from the lateral edge, and the boxed areas illustrate additional rows of stereociliary bundles. Hair cells in the middle region of the (C) wild-type and (D) bpck OC (scale bar in A–D = 10 μm). (E) Quantitation of percentage of hair cells in P1 OC in extra rows. (n = 2 wild-type, 2 bpck mice, ∼200 cells/genotype). Statistics based on two-tailed Student's t-test. Quantitation of stereociliary bundle rotation from the lateral edge in (F) apical, (G) middle and (H) basal regions of the OC (n = 2 wild-type, 2 bpck mice). Approximately 90–200 cells were analyzed in each region of the OC. (I) High-resolution images of hair cells with or without kinocilia in wild-type or bpck OC (scale bar = 1 μm). (J) Quantitation of kinocilia absence in OC rows (n = 2 wild-type, 2 bpck mice). Between 100 and 200 cells were analyzed in each region of the OC. ***P ≤ 0.0001, **P < 0.01 based on a two-tailed Fisher's exact test (in F–H, J).

Furthermore, we found a significant proportion of bpck hair cells lacked a kinocilium in all regions of the OC, while these defects were rarely observed in wild-type mice. The highest percentage of ciliogenesis defects occurred in the apical region with somewhat milder defects in the middle and basal regions of the OC (Fig. 3I and J). The mild orientation defects in the bpck mice resemble those seen in the ciliopathy BBS or ciliary Kif3a models (56,59), but differ markedly in degree from the much greater misorientation found in core PCP mouse mutants, such as in Vangl2^Lp/Lp animals where the stereocilia orientation can reach 120° (57,58,60). These data, along with data from other ciliopathy models, suggest that input may be required from both a functional kinocilium and the PCP proteins for complete polarization in this tissue.

PCP is intact in the bpck mice

To clarify the potential involvement of meckelin with PCP signaling, PCP-mediated cellular polarization was analyzed in the bpck model by examining polarity of basal feet in P1 mouse tracheal tissue. Basal feet are dense appendages of the basal body of motile cilia that form connections to the apical cytoskeleton (61). PCP regulates synchronous orientation of basal bodies of motile cilia in multiciliated cells, such as those in the airway epithelium, to ensure coordinated ciliary beating to propel fluid in a uniform direction (62–64). We found that the organization of basal feet was maintained in both wild-type and bpck littermates (Fig. 4A and B), suggesting that the airway epithelium is properly polarized in the bpck mutant.

Figure 4.

Classical PCP signaling is maintained in bpck mice. TEM imaging of P1 (A) wild-type and (B) bpck tracheal tissue showing orientation of basal feet (arrows). Scale bar = 0.5 μm. Localization of the core PCP protein Vangl1 in P1 (C) wild-type and (D) bpck OHCs (M, medial edge). Scale bar = 5 μm. Stereociliary bundles are stained with phallodin.

To further confirm that PCP signaling is intact in the bpck mouse, we analyzed a different tissue and assayed the asymmetric segregation of core PCP proteins to the apical adherens junctions. Here, we examined the asymmetric localization of the core PCP protein Vangl1 to the medial edge of P1 bpck OC hair cells as a robust demonstration of planar polarization of these sensory cells. In agreement with our observation of largely uniform sensory sterocilia alignment in the OC of the bpck mouse inner ear, medial localization of Vangl1 was retained in bpck OC hair cells and was indistinguishable from wild-type littermates (Fig. 4C and D). This demonstrated that tissue planar polarization is intact in the bpck mouse and argues against a clear role for meckelin in PCP signaling. While the loss of meckelin causes the ‘PCP-like’ phenotype of disorganized stereociliary bundles, the defect is more likely due to the role of the kinocilium in this process.

Meckelin loss leads to upregulation of canonical Wnt signaling in bpck cystic tubules

As a relationship between dysregulation of canonical Wnt and PCP signaling in PKD and the ciliopathies has been proposed, we investigated the levels of canonical Wnt signaling in the bpck mouse to determine whether this pathway was defective (38,65–71). Transcriptional dysregulation of Wnt signaling was first examined in P1 whole kidney tissue using a RT² Profiler real-time PCR array (SABiosciences) that assays 84 Wnt targets. Most targets were upregulated, with 12 transcripts significantly upregulated >2-fold in mutant compared with wild-type kidneys (Fig. 5A, Table 1). Four of these genes are direct targets of Wnt transcription factors: Ccnd1 (cyclin D1) (72,73), Fn (fibronectin) (74), Myc (74,75) and Ppard (peroxisome proliferation-activated receptor delta) (73). These results indicate that Wnt signaling is globally upregulated in meckelin negative kidneys.

Figure 5.

Wnt transcription is upregulated in P1 bpck kidneys. (A) Fold change of Wnt responsive transcripts in bpck kidney compared with the wild-type based on RT² Profiler Array analysis. Image of (B) TOPGAL;wild-type and (C) TOPGAL;bpck kidney after X-gal staining indicating regions of Wnt transcriptional activation (scale bar = 100 μm). Higher resolution images of (D) P1 and (E) P14 TOPGAL;bpck kidney cysts (scale bar = 20 μm). (F) P1 outbred wild-type and bpck kidney tissue labeled with biotinylated-DBA (collecting ducts) and PCNA to determine levels of proliferation in the collecting ducts. Arrowheads indicate PCNA positive proliferating cells. Scale bar = 20 μm. (G) Quantitation of (F). Statistics was based on a two-tailed Fisher's exact test. (***P ≤ 0.0001; n = 1108 wild-type, 1019 bpck cells; n = 167 wild-type, 102 bpck collecting duct tubules).

Table 1.

Wnt transcript fold changes in P1 bpck kidneys

Gene	Description	Fold change	P-value
Abcb1a	ATP-binding cassette	2.31	0.0246
Ccnd1	Cyclin D1	2.22	0.000468
Cd44	CD44 antigen	2.39	0.0137
Ctgf	Connective tissue factor	3.98	0.000211
Egfr	Epidermal growth factor receptor	2.28	0.0141
Fn1	Fibronectin 1	2.1	0.0383
Irs1	Insulin receptor substrate 1	2.4	0.00386
Klf5	Kruppel-like factor 5	2.7	0.0206
Myc	Myelocytomatosis oncogene	2.11	0.0186
Pdgfra	Platelet-derived growth factor receptor alpha	2.48	0.0118
Ppard	Peroxisome proliferator activator receptor delta	2.09	0.000015
Tgfb3	Transforming growth factor beta 3	2.16	0.00211

Statistics are based on Student's t-test. n = 3 wild-type, 3 bpck mice.

To determine in which part of the kidney Wnt signaling was upregulated, we performed X-gal (5-bromo-4-chloro-indolyl-β-d-galactopyranoside) staining in TOPGAL;bpck and wild-type kidneys. The TOPGAL transgenic mouse contains three multimerized TCF/LEF consensus-binding sites with a minimal c-Fos promoter controlling expression of LacZ in a β-catenin responsive manner (76). P1 wild-type kidneys showed a low level of staining with the majority of signal in the nephrogenic zone (Fig. 5B), consistent with a role of Wnt in normal nephron development. In bpck kidneys, the overall staining intensity was greater than in wild-types (Fig. 5C) with the strongest signal occurring in the epithelial cells surrounding the dilated tubules (Fig. 5D). Elevated levels of Wnt signaling were maintained in P14 bpck cyst linings compared with the surrounding interstitum (Fig. 5E), indicating that upregulation of Wnt is specifically associated with cystogenesis in these animals.

To further investigate the consequences of elevated Wnt activity to cyst development, we examined levels of proliferation in P1 kidney tissue in collecting ducts, the most affected kidney segment in the bpck mouse model (Fig. 1F and G). Elevated levels of proliferation were detected in bpck collecting duct cells compared with wild-type littermate controls (Fig. 5F–G). This elevated proliferation is observed at the same time point where elevated levels of Wnt signaling were present (Fig. 5F–G), indicating that increased Wnt signaling activity may be linked to an increase in proliferation in bpck kidney tubules.

Meckelin levels can influence normal canonical Wnt signaling

To further investigate the level of Wnt deregulation with meckelin removal, we examined Wnt upregulation in affected bpck retinal and cochlea tissue by examining levels of nuclear β-catenin. However, in P17 retinal sections from wild-type or bpck mice, we did not find a clear difference in β-catenin levels, and limited nuclear β-catenin (data not shown). Also, while nuclear β-catenin was observed in OHCs and supporting cells in the OC, there was not a significant difference between bpck and wild-type mice (wild-type = 16.7%, bpck = 15.0%, P = 0.63), suggesting that abnormal Wnt levels do not contribute to these phenotypes (data not shown).

We further investigated the general requirement of meckelin for Wnt signaling by examining Wnt levels in isolated bpck cells and cells with exogenous TMEM67 expression. Immunofluorescence (IF) labeling of bpck mouse embryonic fibroblasts (MEFs) showed increased levels of nuclear β-catenin compared with heterozygous littermate controls (Fig. 6A and B). This was confirmed and shown to affect both nuclear and membrane fractions by western analysis of subcellular fractions (Fig. 6C). To further elucidate the role of TMEM67 in canonical Wnt signaling, we performed reporter assays employing a β-catenin-responsive TCF/LEF-1 luciferase reporter plasmid in HEK293T cells. Upon induction of Wnt activity with exogenous expression of the canonical pathway components DVL2 (Dishevelled-2) or CTNNB1 (β-catenin), Wnt signaling was induced in these cells. However, in the presence of exogenously expressed TMEM67, Wnt activity was significantly decreased when stimulated with β-catenin (Fig. 6D). TMEM67 expression also decreased Wnt activity after stimulation with the upstream modulator of Wnt signaling, DVL2, although this did not reach statistical significance (Fig. 6D). The overall expression of the individual components, and the reduction in Wnt signaling by stimulation with both DVL2 and β-catenin is consistent with a role for meckelin as a negative regulator of Wnt activity and suggests a direct role for the protein in modulating Wnt signaling in kidney epithelial cells as well as in MEFs. In addition, enhanced Wnt signaling does not appear to be associated with some other phenotypes observed in the bpck mouse where proliferation may be less evident.

Figure 6.

Meckelin regulation of canonical Wnt signaling. (A) IF analysis of β-catenin localization in heterozygous (+/−) and homozygous (−/−), bpck, MEFs. Scale bar = 10 μm. (B) Quantitation of nuclear β-catenin levels in MEFs. Data are combined from two representative experiments. (n = 3336 +/− cells, 2802 bpck cells). Statistics are based on Chi-squared analysis (***P ≤ 0.0001). (C) Western blot displaying levels of β-catenin from subcellular MEF fractionation, showing increased levels in bpck membrane and nuclear fractions. Lamin and N-cadherin were used as loading controls. WCL, whole cell lysate; Memb., membrane fraction; Nuc., nuclear fraction. (D) β-Catenin responsive luciferase reporter assay (TOPflash) data from HEK293T cells transiently transfected (48 h) with various cDNA constructs: Wnt signaling reporter (firefly luciferase, TOPflash), transfection control (renilla luciferase, pGL4.73), control plasmid (V5 epitope tag, V5-pcDNA3.1(+)), epitope tagged human TMEM67 cDNA plasmid (V5-TMEM67[pcDNA3.1(+)]), control plasmid (HA epitope tag, HA-pcDNA3.1(+)), epitope-tagged human DVL2 cDNA (HA-DVL2 [pcDNA3.1(+)]) and epitope-tagged CTNNB1 cDNA (HA-β-cat [pcDNA3.1(+)]). Data were collected from two independent experiments. Each condition tested was prepared in triplicate for each experiment, and luciferase activity measured independently. Luciferase activity is reported as a ratio, TOPflash/renilla luciferase RLUs. Statistics were based on Student's t-test (**P < 0.01).

Zebrafish tmem67 depletion leads to ciliopathy and CE phenotypes

To gain further insight into conserved roles of meckelin in cilia functionality and cilia-related signaling events, we turned to zebrafish as a model system. We identified one zebrafish tmem67 ortholog (ENSDARG00000076752). The resultant protein has 982 amino acids and is 58% identical and 74.9% similar to the human protein, based on ClustalW Alignment (Supplementary Material, Fig. S2). To examine potential functions of meckelin during zebrafish development, phenotypes were examined after transient protein depletion by morpholino injection at ∼48 hpf. Morphants developed ciliopathy phenotypes, such as hydrocephalus, body curvature and bilateral pronephric cysts, consistent with other ciliopathy zebrafish models (Fig. 7A–D) (17,69–71,77,78). Two different morpholinos were used to target tmem67, a splice blocking morpholino (tmem67 sp-MO) and a translation blocking morpholino (tmem67 ATG-MO). Both produced dose-dependent phenotypes, indicating morpholino specificity, with the tmem67 sp-MO being more potent (Fig. 7E and F, Supplementary Material, Table S2). For this reason, the sp-MO was used for all other experiments, unless indicated. The sp-MO led to complete degradation of mRNA, likely through nonsense mediated decay (Supplementary Material, Fig. S3A). Rescue of these ciliopathy phenotypes by overexpression of human TMEM67 was attempted; however, injection of even very small amounts of RNA led to the development of similar phenotypes (data not shown), making phenotypic rescue challenging and suggesting that these zebrafish ciliopathy phenotypes are highly sensitive to the amount of functional meckelin present.

Figure 7.

Zebrafish meckelin depletion leads to ciliopathy phenotypes. Studies of (A) control and (B–D) morphant zebrafish at ∼48 hpf. (Bi–ii) Representative images of zebrafish morphants with bilateral pronephric cysts, (C) body curvature and (C, D) hydrocephalus. Arrows indicate affected region (B–D). Quantitation of (E) tmem67 splice blocking and (F) translation blocking morpholino phenotypes. Between 30 and 80 embryos were examined for each condition (Supplementary Material, Table S2). Statistics are based on Chi-squared analysis and compared with control injected embryos (***P ≤ 0.0001, **P < 0.01, *P < 0.05).

We next used the zebrafish model to examine the role of meckelin in cilia maintenance by examining olfactory and lateral line cilia. Cilia organization and length in the olfactory placode and anterior and posterior lateral lines were normal in ∼48 hpf morphants (Supplementary Material, Fig. S3B–D). These data suggest that depletion of zebrafish meckelin levels leads to mild ciliopathy phenotypes; however, the structure of cilia was intact in the examined tissues, similar to the data from the bpck mouse.

CE has previously been used as a read-out of functional levels of PCP signaling during zebrafish gastrulation and segmentation, a directional tissue remodeling process that is severely disrupted in many PCP mutants (67–69,77,79–83). Several ciliopathy zebrafish models also display defective CE, suggesting a potential role for these genes in PCP events (67,69,77), although our rodent data did not indicate that meckelin was necessary for PCP-mediated planar polarization (Fig. 4). We investigated if meckelin removal in zebrafish also leads to ‘PCP-like’ phenotypes by examining CE events. For these studies, meckelin was depleted using the ATG-MO in 7–8 somite-staged embryos. The angle of embryonic tissue elongation (TE) spread around the yolk was measured to assess CE defects. Morphants with a TE angle greater than 65 degrees were defined as having defective TE (Fig. 8A and B). Injection of increasing amounts of morpholino led to a dose-dependent increase in affected TE morphants that could be rescued by overexpression of TMEM67 RNA, suggesting that tmem67 is involved in TE during zebrafish development (Fig. 8C and D and Supplementary Material, Table S3). This rescue is in contrast to the effect observed with overexpression of TMEM67 for the ciliopathy phenotypes (Fig. 7), suggesting that the latter phenotype may be more sensitive to absolute levels of functional meckelin and that the CE and ciliopathy phenotypes may be associated with different functions of meckelin. As an additional assessment of CE, somite length and width measurements were taken after staining the same morphants with myoD and krox20 riboprobes to label the somites and hindbrain, respectively (Fig. 8E and F) (67,83). This analysis confirmed that targeting tmem67 leads to defective CE events.

Figure 8.

‘PCP-like’ CE phenotypes occur with meckelin depletion in zebrafish morphants. All TE analysis was performed in 7–8 somite-staged embryos. (A) Zebrafish embryo with TE angle indicated; affected embryos had TE angles greater than 65°. (B) Representative images of TE unaffected and affected morphants. (C) Quantitation of TE defects with dose escalation of tmem67 ATG-MO and (D) rescue of TE phenotype with human TMEM67 RNA injection. Approximately 25–95 embryos were examined for each condition (Supplementary Material, Table S3). Statistics are based on Chi-Squared analysis compared with control-MO or MO + RNA injected embryos, as indicated (***P ≤ 0.0001, **P < 0.01, *P < 0.05). (E) Representative 7–8 somite-staged embryos stained with myoD (somites) and krox20 (hindbrain) riboprobes to analyze somite length (SL) and somite width (SW) ratios. Images are dorsal views of flatten embryos with the yolk removed. (F) Quantitation of somite length to width defects with tmem67 ATG-MO. Control morphants n = 19; tmem67 morphants (3 ng) n = 45; tmem67 morphant (6 ng) n = 44. Statistics are based on Student's t-test (**P < 0.01).

Ciliary defects in zebrafish tmem67 and vangl2 morphants

We describe above the ‘PCP-like’ phenotypes of hair cell mis-orientation in the bpck OC and CE defects in tmem67 morphants. However, we found intact planar polarity in the bpck mouse, suggesting that PCP signaling is functional with meckelin removal. Consequently, we first examined maintenance of planar polarity in the tmem67 morphants by comparing to phenotypes induced by targeting the core PCP protein Vangl2. Depletion of zebrafish Vangl2 levels using a translation blocking morpholino (83) led to the development of the ciliary phenotypes of body curvature, cysts and hydrocephalus in a similar manner to tmem67 morphants (Fig. 9A). The pronephric duct was also dilated in maternal zygotic (MZ) vangl2 mutants (84), supporting the specificity of our vangl2 morphant data. We attempted to examine planar polarity in tmem67 morphants by orientation of basal feet in the pronephric duct, as performed in bpck tracheal tissues; however, this proved to be technically challenging. As a surrogate for this analysis, we examined a downstream consequence of basal feet orientation in the pronephric duct: pronephric cilia beat frequency. MZvangl2 mutants have disorganized multiciliated pronephric duct cilia (84), and our vangl2 morphants had a significantly lower pronephric cilia beat frequency than injected control morphants (Fig. 9B and C, Supplementary Material, Movies S1 and S3). Cilia beat frequency was also decreased in tmem67 morphants compared with controls (Fig. 9B and C, Supplementary Material, Movie S2). However, vangl2 morphants had a significantly lower beat frequency than tmem67 morphants and displayed disorganized cilia with no clear coordinated movements (Supplementary Material, Movie S3), which was not observed in tmem67 morphants (Fig. 9C, Supplementary Material, Movie S2). Consequently, while tmem67 morphants display cilia beat frequency defects, they are significantly milder than that observed in the vangl2 morphants and pronephric cilia do not display a dramatic organizational defect as seen in the pronephric duct of vangl2 morphants. This effect is similar to that observed in the OC of the bpck mouse, where the ‘PCP-like’ defect is much milder than that observed in core PCP mutants. Together, these data provide evidence that although the phenotypes in the tmem67 and vangl2 morphants are similar, they appear to result from different mechanisms further showing that planar polarity is intact with meckelin depletion.

Figure 9.

Comparison of ciliary phenotypes in zebrafish morphants of tmem67 and the core PCP gene vangl2. (A) Quantitation of ciliopathy phenotypes after vangl2 depletion in ∼48 hpf embryos. Approximately 50–80 embryos were examined for each condition. Statistics are based on Chi-squared analysis (***P ≤ 0.0001, **P < 0.01, *P < 0.05). (B) Quantitation of pronephric cilia beat frequency in ∼48 hpf morphants. Data are from two representative experiments, with 4–6 embryos examined for each condition. Statistics are based on Student's t-test (***P ≤ 0.0001, **P < 0.01). (C) Still frames of ciliary waves from Supplementary Material, Movies S1–S3. Arrowheads indicate the progression of a single ciliary wave throughout the examined time span. The pronephric duct is outlined.

DISCUSSION

In these studies, we have more fully characterized the phenotypes associated with meckelin loss/depletion using the bpck mouse model and zebrafish tmem67 morphants. In the bpck mouse, we detected mild tubule dilations as early as E16.5, with mortality and kidney disease severity increased in C57BL6/J inbred animals, similar to the recently described Tmem67^−/− mouse (7). These mice also exhibited defects in cilia function, as seen by retinal degeneration and tissue disorganization in the eye, due to a non-functional connecting cilium [Fig. 2, Supplementary Material, Fig. S1; (33,85)]. Novel cochlea defects were also described (Fig. 3) with mild rotational defects of the stereociliary bundles and absence of kinocilia. In tmem67 zebrafish morphants, pronephric cysts and hydrocephalus were found at a higher level in tmem67 zebrafish morphants than previously described (55), but in contrast to the previous study, only these phenotypes and body curvature (but not otic, eye defects and cardiac edema) were found to be significant. However, less than half of the animals were affected suggesting that in zebrafish, as in C. elegans, depletion of additional ciliopathy genes may be necessary to cause strong, highly penetrant cilia-related phenotypes (9,10,86).

The data described here and in previous studies of rodent models and MKS3 human and mouse cells reveal a complex set of ciliary defects that vary between cell types and tissues. Previous data in kidney cysts showed that cilia formed but were longer than the wild-type and had numerical defects, in contrast to the absence of flagella in the rat sperm (7,33,51). Connecting cilia in MKS3 rodent retinas appeared structurally normal, but in the mouse were unable to transport rhodopsin, transducin or arrestin to the outer segment [Fig. 2, Supplementary Material, Fig. S1; (33,85)]. A similar ciliary trafficking defect was observed in isolated MEFs, where Arl13b and adenylyl cyclase III (AC3) transport to the primary cilium was reduced but Smoothened trafficking unperturbed (7). In our study, a high percentage of sensory hair cells in the cochlea were detected without a kinocilium in bpck animals at P1, which we propose occurs due to earlier re-absorption of the kinocilium, which occurs normally at approximately P10 after patterning of the stereocilia bundles is complete, likely from being non-functional (Fig. 3) (87). Evaluation of basal body docking in multiciliated respiratory cells in the bpck trachea showed no defective docking (Fig. 2), despite earlier studies suggesting that this process was defective in meckelin-depleted cells (34,55). Clear structural and numerical defects in cilia were also not detected in several ciliated tissues in tmem67 zebrafish morphants (Supplementary Material, Fig. S3), although these animals developed ciliopathy phenotypes. These data are consistent with the defect in MKS3 only rarely being associated with failure to form cilia (33), but more typically due to abnormal trafficking of a subset proteins on to the cilium, linked with a role for meckelin as part of a membrane protein filter at the transition zone (7).

The role of Wnt signaling in ciliopathies and PKD has been controversial. Canonical Wnt is involved in metanepehric mesenchyme induction and cell proliferation during branching morphogenesis in early kidney development and may become reactivated in PKD to repair damaged tissue (88–91). In human NPHP2, a switch from PCP signaling to canonical Wnt activation was suggested as the pathogenic principle (78). This model remains a strong paradigm for the contribution of canonical Wnt signaling to renal cystogenesis. Here, we demonstrate that the canonical pathway is upregulated in P1 and P14 bpck kidneys with canonical Wnt activity concentrated in cyst linings, with enhanced proliferation also detected in P1 bpck kidney tissue. Wnt activity was not elevated in cochlear or retinal bpck cells, where the phenotype is unlikely to be caused by heightened proliferation levels. This is consistent with canonical Wnt signaling playing an active role in promoting or accelerating cystogenesis by enhancing proliferation, similar to other PKD models (36,55,67,69,92,93). Our data suggest that meckelin restrains canonical Wnt activity, which could not be rescued by overexpression of β-catenin. The identification of meckelin in the MKS-B9 protein module at the transition zone (TZ) suggests that meckelin may exert this observed negative regulatory function by a general role as a gating factor at the TZ to control trafficking of a specific set of signaling factors to and from the ciliary compartment. In support of this, lack of a functional cilium, such as with meckelin removal, may disrupt the sequestration of canonical Wnt components, leading to increased transcriptional activity and cellular proliferation (36,39). More specifically, meckelin may act as a membrane anchor to spatially organize the β-catenin destruction complex at the ciliary base (38). Lack of meckelin may disrupt formation of this complex, allowing β-catenin to escape degradation and remain transcriptionally active (39,94). This is supported by our finding of elevated levels of β-catenin in bpck kidney epithelial cells and MEFs.

Conversely, the non-canonical Wnt pathway, PCP, is proposed to be downregulated in PKD (2) and to directly contribute to cystogenesis (95) and other ciliopathy phenotypes; however, this involvement is controversial. Loss of meckelin in the bpck mouse caused the ‘PCP-like’ phenotype of stereocilia misalignment in the OC with hair cells displaying significant rotational defects away from the lateral edge of the OC (Fig. 3). However, hair cells in core PCP mouse mutants, such as Vangl2/Stbm (Looptail), rotate up to 100 to 210 degrees from the lateral edge of the OC (57), while bpck hair cells rotated a maximum of 51 to 60 degrees. The phenotype of extra rows of hair cells seen in PCP mutants was also not significantly elevated in the bpck mice. A milder rotational phenotype, similar to that in the bpck, was observed in BBS (56) and structural cilia protein mutant models, such as in conditional Ift88 and Kif3a mice (43,59). It is thought that input from both the core PCP pathway and the kinocilium and associated proteins are necessary for complete polarization of hair cells, implying interaction between these pathways (43,59). Given the milder stereociliary phenotype detected in the bpck mouse, the stereocilia misalignment seems more likely associated with loss of kinocilium function than defective PCP signaling. This interpretation is supported by the failure to detect direct planar polarization defects in bpck tissues when well-established vertebrate PCP-dependent processes, such as coordinated polarization of motile cilia in airway epithelial cells and asymmetric protein localization of Vang1 in the OC, were examined. Although we cannot rule out PCP defects in other tissues or stringently conclude that all of PCP signaling is functional by examining a few phenotypes, these data support a functional PCP pathway in the bpck mouse model. Furthermore, no renal cysts have been documented in mouse Vangl2/Stbm mutants and only Fat4 and Nphp2/Inversin PCP-related mouse mutants develop PKD (48,96), perhaps because of their additional roles in the development of functional cilia or in directional tissue morphogenesis, although it is unclear how closely renal cystogenesis has been examined in many core PCP mutants.

‘PCP-like’, CE defects were also observed with meckelin depletion in zebrafish morphants, although our analysis also suggested that planar polarity, and consequently PCP signaling, was intact in tmem67 morphants due to coordinated, synchronized ciliary beating compared with that in vangl2 morphants. Defective CE has been used as a functional readout of PCP signaling in ciliopathy models and is observed in core PCP zebrafish mutant models (67,78,82,83). However, CE is a complex process that involves many pathways and, therefore, CE phenotypes do not necessarily equate with disrupted PCP signaling (35,97–99). Although the PCP pathway is a major player in CE, its role appears to be more permissive than instructive and involves input from multiple different signaling pathways and molecules including: Wnt, Nodal, BMP (bone morphogenic protein) and FGF (fibroblast growth factor), Stat3 and G protein coupled-receptor pathways (97–99). Additionally, transcription factors such as Brachyury and Snail, adhesion molecules, extracellular matrix components, cytoskeleton factors, various receptors, axon guidance proteins and components involved in endocytosis have all been implicated in maintaining proper CE, demonstrating the complexity of events involved in this process (99).

Instead of being directly involved in PCP-mediated CE, we propose that meckelin plays a role in CE through cilia and centrosome functioning. The centrosome is proposed to detect polarizing information in a three-dimensional sense in the cell and consequently may function as a mediator of polarity on the cellular level (35). We have previously shown that meckelin removal causes a multiple centrosome phenotype, which could cause defects in polarized cell migration and oriented cell division, both essential processes for CE at the level of the centrosome and thus may be responsible for CE defects mediated by meckelin removal rather than a direct involvement of meckelin in PCP-mediated CE (33,97,99).

This study indicates that no strong defects in cilia formation, basal body docking or PCP dysfunction occur with meckelin removal. Instead, our data support the hypothesis that defects associated with TMEM67 mutation are likely due to only partially functional cilia in a range of different tissues. Aberrations were observed in canonical Wnt signaling, which may be related to cilia-dependence of this pathway. Therefore, we reaffirm a role of meckelin at the cilia/centrosome compartment and propose that defects in protein functioning at this level are largely responsible for the phenotypes observed in MKS3.

MATERIALS AND METHODS

Mice

Mice were housed at the Mayo Clinic Mouse Facility according to NIH and Institutional Animal Care and Use Committee (IACUC, Mayo Clinic, Rochester, MN, USA) guidelines. bpck mice were obtained from Jackson Laboratories (Bar Harbor, ME, USA) and were maintained on a C57BL6/J background. Inbred animals were backcrossed 10 generations and outbred mice were in mixed C57BL6/J;B6C3FeF1/Ja/a or C57BL6/J;LacZ backgrounds. TOPGAL mice were kindly provided by John Hawse at the Mayo Clinic. bpck and LacZ mice were genotyped using the following primers: 5′-GCACTATGTTAAAAGATCCAG-3′ and 5′-ATTCTGCTGAACACCCCACAG-3′ (wild-type allele), 5′-TGTCAAACCACCCTCTAAGCCTG-3′ and 5′-CAGCCTCACATCACTTTACCTAACG-3′ (bpck allele), 5′-ATCCTCTGCATGGTCAGGTC-3′, 5′-CGTGGCCTGATTCATTCC-3′, 5′-CAAATGTTGCTTGTCTGGTG-3′ and 5′-GTCAGTCGAGTGCACAGTTT-3′ (LacZ alleles). Cystic tubule measurements were performed using the ImageJ software.

Zebrafish

Zebrafish were housed at the Mayo Clinic Zebrafish Core Facility according to the NIH and IACUC guidelines. Wild-type Tu or Segrest lines were utilized. Staging was performed by counting somites or by other standard protocols (79).

X-gal staining

Whole mouse kidneys were dissected and fixed in whole mount fix (0.2% glutaraldehyde, 5 mm EGTA and 2 mm MgCl₂ in 0.1 m phosphate buffer) on ice for 1 h. Kidneys were washed 3×30 min in detergent rinse (0.02% Igepal, 0.01% sodium deoxycholate and 2 mm MgCl₂ in 0.1 m phosphate buffer). Tissue was stained overnight at 37°C in the dark in X-gal staining solution (1 mg/ml X-gal 0.02% Igepal, 0.01% sodium deoxycholate, 5 mm potassium ferricyanide, 5 mm potassium ferrocyanide and 2 mm MgCl₂ in 0.1 m phosphate buffer). Kidneys were then washed in PBS and stored in 4% paraformaldehyde (PFA; Electron Microscopy Sciences; Hatfield, PA, USA). Kidneys were embedded in paraffin and sectioned. Slides were de-waxed, rehydrated, counter-stained with Nuclear Fast red and preserved with Vectashield (Vector Labs; Burlingame, CA, USA).

Electron microscopy

For SEM, inner ears were dissected from mice and fixed in 4% PFA for at least 24 h. Cochleae were dissected from fixed tissue, and the OC was isolated and fixed at least overnight in 4% PFA and post-fixed in Trump's fixative (4% PFA, 1% glutaraldehyde). The remaining process was carried out using standard procedures by the EM core facility at the Mayo Clinic, as previously described (33). For TEM, the samples were fixed in Trump's fixative for 2–3 h, washed in 0.1 m phosphate buffer 3 × 15 min, post-fixed in 1% osmium tetroxide in 0.1 m phosphate buffer for 1–2 h at room temperature, washed three to five times (10 min each) in distilled water, then En bloc stained with 2% aqueous uranyl acetate for 1 h at room temperature, washed two to three times (5 min each) in distilled water, dehydrated through an ethanol series, embedded in Spurrs resin, sectioned at 80 nm, stained with lead citrate for 3–15 min and examined with a JEOL 1400 electron microscope (JEOL USA, Peabody, MA, USA).

Zebrafish microinjection

All morpholino and RNA solutions were injected into 1–4 cell embryos in volumes of 3–9 nl at the indicated concentrations. RNA was synthesized from a linearized pcDNA3.1 vector template using the T7 mMESSAGE mMACHINE High Yield Capped RNA Transcription Kit (Ambion, Life Technologies). Morpholinos were manufactured by Gene Tools (Philomath, OR, USA). Unless otherwise noted, morpholinos were designed by Gene Tools using the Oligo Design service. Two morpholinos were used against tmem67: a translation blocking morpholino (5′-GATCAGATGACGCTTGATGACGTAT-3′) and a morpholino targeting the intron 8/exon 9 splice acceptor site (5′-AGAACAACTACAGAAGAACAATAAC-3′). The splice blocking morpholino efficacy was tested with primers in exon 7 (5′- AATGTGCTTCCCTGCGAACAGTCTGC-3′) and exon 10 (5′- CCAAGGTAGGTTTGTCCTCCAGTAG-3′). Additional morpholinos were previously tested in published literature against the translation start site of vangl2 (83). The control morpholino used targets mutated human β-globin and has no target in zebrafish (Gene Tools). TE measurements were performed on printed images using a protractor and compass. Ciliopathy phenotypes were analyzed in a blinded manner. All experiments were repeated three times and with three different dosages of morpholino. Level of control morpholino to use for each experiment was determined by matching death rates between the control and experimental morpholino. RNA levels for the CE rescue experiments were chosen based on levels that did not produce a phenotype when injected alone. All levels of TMEM67 RNA injected produced ciliopathy phenotypes.

RNA extraction and cDNA analysis

RNA for morpholino analysis was extracted from ∼5–10 pooled embryos. Embryos were homogenized with needles in Trizol and RNA was extracted by successive chloroform, isopropyl alcohol extractions. cDNA was synthesized using the Superscript III First Strand Synthesis System with random primers according to the manufacturer instruction (Life Technologies; Grand Island, NY, USA). RNA was extracted from whole mouse kidneys for transcript analysis on the RT² WNT Signaling Target PCR Array (Mouse) chip from SABiosciences (Valencia, CA, USA). RNA was isolated using a Qiagen RNeasy Mini Kit and cDNA synthesis performed using the RT² First Strand Kit (SABiosciences). Transcript analysis was performed employing the RT² WNT Signaling Target PCR Array (Mouse) chip (SABiosciences). Chips were run on a 7900HT Real-Time 384 well PCR machine (Applied Biosystem). Data were analyzed by the ΔΔCt method using the PCR Array Data Analysis Web Portal software (SABiosciences).

IF labeling

Cochleae

OCs were isolated as above and fixed in 4% PFA. OCs were washed three times in buffer T [0.1 m Tris (pH 7.5), 0.15 m NaCl and 0.1%Triton X-100] and then incubated in blocking buffer (buffer T with 10% fetal bovine serum) for 1 h. Samples were incubated overnight in primary antibody diluted in blocking buffer. OCs were then washed with buffer T, incubated in secondary antibody for 1 h and washed again. Whole OCs were mounted in Vectashield (Vector Labs) on slides for analysis. Antibodies used include: rabbit anti-human Vangl1 pAb (Sigma-Aldrich; St Louis, MO, USA), Alexa Fluor-conjugated phallodin [Life Technologies (Molecular Probes)] and Alexa Fluor 488 goat anti-rabbit IgG [Life Technologies (Molecular Probes)]. All samples were visualized on a Zeiss AxioObserver microscope at ×40 magnification (Carl Zeiss).

Mouse embryonic fibroblasts

Fibroblasts were prepared from E13.5 embryos. Pregnant females and embryos were euthanized according to IACUC standards and embryos removed and washed briefly in PBS. Embyos were decapitated, eviscerated, diced and trypsinized. Fully trypsinized cells were plated on tissue culture plastic as passage 1. For IF, cells were grown to 90% confluence in DMEM/F12 10% FBS on glass cover slips. Cells were then washed twice with PBS, fixed and permeabilized 5 min with ice-cold methanol. Slides were washed twice in PBS and incubated in 5% BSA in PBS for 1 h. Primary anti-β-catenin antibody (Cell Signaling, 9562) was applied overnight in 1% BSA in PBS. Secondary Alexa Fluor 488 goat anti-rabbit IgG [Life Technologies (Molecular Probes)] was used for 1 h followed by DAPI. Cover slips were washed in PBS and mounted on charged slides using Vectashield H-1000 mounting medium (Vector Labs). Nuclear β-catenin/DAPI co-localization was scored and quantified with ImageJ.

Mouse kidney tissue

IF was performed on kidney tissue sections as previously described (33). Primary antibodies used included: biotinylated –LTA and –DBA (1:250, Vector Laboratories), THP (1:200, Santa Cruz Biotechnology Inc.), and PCNA (1:100, Santa Cruz Biotechnology Inc.). Secondary antibodies included: Alexa Fluor conjugated streptavidin, goat anti-rabbit IgG and goat anti-mouse IgG2a [Life Technologies (Molecular Probes)]. Nuclei were counterstained with DAPI (Life Technologies).

All IF samples were imaged using an AxioObserver and Axiovision 4.8 software (Carl Zeiss).

Statistics

Two-tailed Fisher's exact test was performed as previously described (33). Chi-squared distributions and Student's t-test were calculated using Microsoft Excel.

Whole mount zebrafish in situ

Regions of myoD and krox20 were amplified for probe synthesis using the following primer sets: myoD (5′-CTTGGACCCCAGGCTTGTTCAC-3′, 5′-GGTAATACGACTCACTATAGGGTTCCGTCTTCTCGTCTGACACG-3′) and krox20 (5′-TCGCACAGACACAACACATTC-3′, 5′-GGTAATACGACTCACTATAGGTGGGGTATTTCCTTGGACGC-3′). A T7 promoter was added to the reverse primer for amplification. Riboprobes were created using the Ampliscribe T7 kit with DIG RNA labeling mix (Roche). Zebrafish embryos for analysis were collected at the indicated times, the chorions were removed and embryos were fixed overnight at 4 degrees in 4% PFA. Whole-mount stainings were performed as described (100). Embryos were cleared in 70% glycerol/PBS for imaging.

Mouse skeletal whole-mount staining

Pregnant females and embryos were euthanized according to IACUC standards. Embryos were dissected and stained as previously described (101).

Cellular fractionations

For MEF subcellular fractionation, cells were resuspended in a hypotonic solution (10 mm HEPES pH 7.5, 10 mm KCl, 2.5 mm MgCl₂) and homogenized. Cells were spun at 600g for supernatant (SN1) and pellet (P1) fractionation. SN1 was spun at 22 000g for 30 min (SN2). The pellet was resuspended in a solubilization buffer (10 mm Tris–HCl pH 7.5, 500 mm NaCl, 1 mm EDTA, 1% Triton X-100) and centrifuged at 16 000g for 30 min. The resulting supernatant (SN3) represented the membrane fraction. Pellet from the original centrifugation (P1) was resuspended in 100 μl of nuclear lysis buffer (50 mm HEPES pH 7.9, 250 mm KCl, 0.1% NP40, 0.1 mm EDTA) and centrifuged at 16 000g for 10 min. Supernatant from the final centrifugation represented the nuclear fraction (N1). All buffers were supplemented with protease inhibitors (Roche) and fractionation performed on ice.

Western blotting

Antibodies for western blotting were diluted in 5% Non-fat dry milk in TBST: β-catenin (Cell signaling 9562S 1:2000), Lamin A/C (Santa Cruz N-18 1:2000), Alpha Tubulin (Sigma T6199, mouse IgG1 1:5000), N-cadherin (Life Technologies 33–3900, and mouse IgG1 1:1000).

Luciferase assays

HEK293T cells (ATCC) were grown according to the manufacturer's instructions. Cells were plated in six-well plates, and transfected with Polyfect (Qiagen) according to the manufacturer's protocol. The following constructs were transfected into cells: Wnt signaling reporter (firefly luciferase, TOPflash, Millipore), transfection control (renilla luciferase, pGL4.73, Promega), control plasmid (V5 epitope tag, pcDNA3.1(+), Life Technologies), epitope tagged human TMEM67 cDNA plasmid (N-terminal V5 epitope inserted after the signal peptide, pcDNA3.1(+), Life Technologies), control plasmid (HA epitope tag, pcDNA3.1(+), Life Technologies), epitope-tagged human DVL2 cDNA (N-terminal HA epitope, pcDNA3.1(+), Life Technologies) and epitope-tagged CTNNB1 cDNA (N-terminal HA epitope tag, pcDNA3.1(+), Life Technologies). The cell media was replaced 24 h post-transfection, and cells were lysed and prepared for luminescence measurements 48 h post-transfection. The Dual-Luciferase Reporter Assay system (Promega) was used for sample preparation according to the manufacturer's protocol. Sample luminescence was measured using dual automatic injection (Luminometer 20/20ⁿ, Turner Biosystems).

cDNA cloning

pcDNA3.1(+) vectors were modified by insertion of coding sequences for V5- and HA-epitope tags between the NheI- and BamHI-restriction site. Human cDNAs for DVL2, CTNNB1 and TMEM67 were amplified by RT–PCR from 293T cell mRNA using gene-specific primers. DVL2 and CTNNB1 were cloned into the respective plasmids using BamHI and NotI restriction enzymes to generate tagged constructs. For TMEM67, a V5-tag was inserted by fusion PCR between the signal peptide sequence and the putative first amino acid of the N-terminal domain and later subcloned into pcDNA3.1(+). The exact primer sequences and cloning strategies are available from the authors upon request.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.

FUNDING

This work was supported by the National Institute of Diabetics and Digestive and Kidney Disease grant (DK059597). A.C.L.'s predoctoral studentship is supported by the Zell Family Foundation. The Mayo PKD Translational Center (DK090728) provided support for the zebrafish and mouse studies.