The Meckel-Gruber Syndrome Gene, MKS3, Is Mutated in Joubert Syndrome

Abstract

Joubert syndrome (JS) is an autosomal recessive disorder characterized by cerebellar vermis hypoplasia associated with hypotonia, developmental delay, abnormal respiratory patterns, and abnormal eye movements. The association of retinal dystrophy and renal anomalies defines JS type B. JS is a genetically heterogeneous condition with mutations in two genes, AHI1 and CEP290, identified to date. In addition, NPHP1 deletions identical to those that cause juvenile nephronophthisis have been identified in a subset of patients with a mild form of cerebellar and brainstem anomaly. Occipital encephalocele and/or polydactyly have occasionally been reported in some patients with JS, and these phenotypic features can also be observed in Meckel-Gruber syndrome (MKS). MKS is a rare, autosomal recessive lethal condition characterized by central nervous system malformations (typically, occipital meningoencephalocele), postaxial polydactyly, multicystic kidney dysplasia, and ductal proliferation in the portal area of the liver. Since there is obvious phenotypic overlap between JS and MKS, we hypothesized that mutations in the recently identified MKS genes, MKS1 on chromosome 17q and MKS3 on 8q, may be a cause of JS. After mutation analysis of MKS1 and MKS3 in a series of patients with JS (n=22), we identified MKS3 mutations in four patients with JS, thus defining MKS3 as the sixth JS locus (JBTS6). No MKS1 mutations were identified in this series, suggesting that the allelism is restricted to MKS3.

Joubert syndrome (JS [MIM 213300]) is an autosomal recessive, multisystem disease characterized by developmental delay, hypotonia, irregular breathing pattern, eye movement abnormalities,¹ and cerebellar vermis hypoplasia/dysplasia with accompanying brainstem abnormalities visualized on axial images, where the depth of the interpeduncular fossa is greater than that usually seen, with an associated narrowing of the midbrain tegmentum and a thickening of the superior cerebellar peduncles resulting in the “molar tooth sign” (MTS).² Other variable features include retinal dystrophy and renal anomalies that define type B of JS, also called “cerebello-oculo-renal” syndrome (CORS). JS is a genetically heterogeneous condition. To date, mutations in two genes have been identified as responsible for JS: AHI1 (JBTS3 on chromosome 6q23.3)³ and, most recently, CEP290 (JBTS5 on 12q21.3).^4,5 In addition, an NPHP1 deletion identical to the one that causes juvenile nephronophthisis (NPHP) has been identified in a subset of patients with NPHP and a mild form of MTS, thus defining NPHP1 as the fourth JS locus (JBTS4 on 2q13).⁶ Two more loci, JBTS1/CORS1⁷ and JBTS2/CORS2,^8,9 are mapped to chromosomes 9q34.3 and 11p12-11q13.3, respectively.

Occipital encephalocele and polydactyly have occasionally been reported in some patients with JS. These features are also found in Meckel-Gruber syndrome (MKS [MIM 249000]), a rare autosomal recessive lethal condition characterized by postaxial polydactyly, multicystic kidney dysplasia, liver bile–duct proliferation, hepatic developmental defects, and characteristic CNS malformations. In typical cases of MKS, these include posterior occipital encephalocele, prosencephalic dysgenesis, and rhombic roof dysgenesis, but Dandy-Walker (DW) malformation, hydrocephalus, and agenesis of the corpus callosum are also described as occasional features. MKS is genetically heterogeneous, and three loci have been mapped to 17q23 (MKS1), 11q13 (MKS2), and 8q24 (MKS3). Very recently, two genes have been identified: MKS1 on 17q in Finnish kindreds¹⁰ and MKS3 on 8q in families from Pakistan and Oman.¹¹ The phenotypic overlap between JS and MKS is further supported by the CNS malformations seen in the Wpk rat model of MKS, which include agenesis of the corpus callosum and hydrocephalus but not exencephaly.¹² The missense mutation, P394L, seen in the rat Mks3 gene is presumably a hypomorphic allele because of the mild phenotype and viability of the Wpk rat.¹¹

While sequencing MKS1 and MKS3 genes in 31 unrelated fetuses presenting a cerebro-reno-digital syndrome, which was diagnosed as “Meckel-like” because of the absence of at least one of the MKS diagnostic criteria,¹³ we identified MKS3 mutations in a family with two siblings (fetuses JS-661 and JS-660). In both cases, pregnancies were terminated at 30 wk gestation (WG) and 28 WG, respectively, because of abnormal posterior fossa and hyperechogeneic, enlarged kidneys detected by ultrasound. Kidney histology showed a conserved corticomedullary organization, with microcysts mainly in the medulla (fig. 1A [panels a, b, d, and e]), and liver bile–duct proliferation in both fetuses (fig. 1A [panels c and f]). The full fetal autopsies revealed no other malformations. Detailed neuropathological analysis in fetus JS-661 found a cerebellum weighing 14 g (50th percentile for age: 8 g), with a transverse diameter of 33 mm (50th percentile for age: 32 mm). Sagital section of the cerebellum and the brainstem showed an enlarged, widely open fourth ventricle, hypoplastic/elevated vermis, and thick cerebellar peduncles. Histological study of the cerebellum confirmed vermian hypoplasia reduced to a few folia, fragmented dentate nuclei, and Purkinje cell heterotopias. The brainstem was reduced in size and contained aberrant longitudinal tracts. The inferior olive nuclei were hypoplastic, and the pyramids were asymmetric and reduced in size (data not shown). In fetus JS-660, the cerebellum weighed 10 g (50th percentile for age: 10 g) and had a transverse diameter of 33 mm (50th percentile for age: 35 mm). The transversal section showed a dilated, dysmorphic fourth ventricle flanked laterally by enlarged cerebellar peduncles (mimicking a molar tooth shape) (fig. 1B [panel g]). On histological examination, a remnant of vermian structure was found with fragmented dentate nuclei and Purkinje cell heterotopias. The brainstem was hypoplastic, the longitudinal tracts were aberrant, and the transversal tracts were reduced (fig. 1B [panel i]). We identified two MKS3 mutations in this family. A missense mutation in exon 15, Y513C, was inherited from the father, and a complex indel mutation (13-bp deletion encompassing the exon22/intron 22 boundary replaced by 2 bp) was inherited from the mother. This 2315_2323+4del13insGG mutation removes the donor splice site and was absent in 228 control chromosomes. The Y513C missense mutation is in the extracellular domain of the protein and affects an amino acid conserved throughout evolution.¹¹ This mutation was not identified in 214 chromosomes. In silico analysis of the effect of this amino acid substitution on the protein structure and function, with use of the PolyPhen program, predicts that this change is pathogenic, with a high position-specific independent counts (PSIC) score difference >2.¹⁴

Figure  1. .

Pathological features of fetal patients with MKS3 mutations. A, Histological pattern of kidney and liver (hematoxylin-eosin-safran staining) in fetuses JS-661 (at 28 WG) and JS-660 (at 30 WG). Kidney is shown at low (panels a and d) and high (panels b and e) magnification. Liver is shown at high magnification (panels c and f). Kidney histology shows conserved corticomedullary organization with 6 to 8 generations of mature glomerules. Microcysts are found in the deep cortex in JS-661. Tubular microcysts in medulla (black arrows) are observed in both. Liver histology shows a portal fibrosis with important and diffuse bile-duct proliferation (blue arrows) in both fetuses. B, Neuropathological findings of brainstem and cerebellum. Transversal section of the brainstem and cerebellum at the level of cerebellar peduncles in fetus JS-660 (panel g) compared with a control individual (panel h). In panel g, note the inverted molar tooth aspect due to a dysmorphic fourth ventricle (V4) flanked laterally by thick cerebellar peduncles (P). The roof of the fourth ventricle is formed by the remnant of vermian structure (V). Cerebellar folia contain heterotopic Purkinje cells (data not shown). The hypoplastic pons longitudinal tracts (LT) have a chaotic organization, and the transverse fibers are reduced (panel i), as compared with a control individual (panel j); this is better seen on a caudal level with higher magnification.

Despite the absence of neurological symptoms required for postnatal diagnosis in these fetuses, JS was highly suspected in this family. We, therefore, questioned whether MKS gene mutations could be found in patients with typical JS and sequenced 22 patients with JS with no NPHP1 deletion. Sequence analysis of MKS3 revealed mutations in three patients, confirming that MKS3 is indeed a gene for JS. Mutations and clinical data are summarized in table 1.

Table 1. .

Clinical Data of Patients with JS with MKS3 Mutations

						Phenotypic Featuresa
Patient and Nucleotide Change	Exon	Predicted Effect on Protein	Family	Age	Origin	Kidney	Liver	Ocular	CNS
JS-661:			1	28 WG	France	Microcysts	HF, BDP	−	CVH, open V4, heterotopias, brainstem anomalies
1538A→G	15	Y513C
2315_2323+4del13insGG	22	Splice
JS-660:			1	30 WG	France	Microcysts	HF, BDP	−	CVH, open V4, heterotopias, brainstem anomalies
1538A→G	15	Y513C
2315_2323+4del13insGG	22	Splice
JS-05:			2	14 years	Algeria	−	−	−	HT, MR, BA, CVH, MTS
IVS23+5G→C homozygous	23	I775_A813del
JS-09:			3	7 years	France	−	−	OMA	HT, AT, CVH, enlarged V4, MTS
IVS6+2T→G	6	Splice
1634G→A	16	G545E
2341G→A	21	Q747fsX761
NPH-786:			4	7 years	France	Microcysts	Liver disease	−	AT, MR, BA, CVH, no MRI
637C→T	6	R213C
2132A→C	21	D711A

^aAT = ataxia; BDP = bile-duct proliferation of liver; BA = breathing anomalies; CVH = cerebellar hypoplasia; HF = hepatic fibrosis; HT = hypotonia; MR = mental retardation; OMA = oculomotor apraxia; V4 = fourth ventricle.

JS-05 is a 14-year-old Algerian girl who had consanguineous parents and had already been reported by Romano et al. (designated as “case 2” in their article).¹⁵ She was severely handicapped and presented with hypotonia, severe mental retardation, stereotypic movements, and no independent walking. She had breathing abnormalities but no oculomotor apraxia or abnormal eye movements. Electroretinography and optic disc imaging results were normal. She had no renal or hepatic involvement. Brain imaging showed a cerebellar dysplasia and MTS with no supratentorial anomalies (fig. 2). A homozygous MKS3 mutation was identified near the donor splice site of intron 23 (IVS23+5G→C). This mutation was absent in 402 chromosomes. Both parents and a healthy sister were heterozygous for the mutation, whereas a healthy brother inherited both normal alleles. RNA was extracted from lymphocytes from the patient and from his father, and RT-PCR with primers located in exons 20 (forward: GTAAGCATATGGAGAACATATTT) and 26 (reverse: CATTGGTTCCATGAATTCCATT) showed a transcript smaller than the expected normal size (754 bp) (fig. 3A). Sequencing confirmed an exon skipping from exons 22 to 24 (fig. 3B). This inframe deletion predicts a protein lacking amino acids 775 to 813, which compose most of the putative coiled-coil domain (amino acids 756 to 912) of the protein.

Figure  2. .

Brain MRI of a normal control individual, an affected girl aged 7 years (JS-09), and an affected girl aged 14 years (JS-05) carrying MKS3 mutations. A, Sagittal MRI in JS-09 shows a dilated fourth ventricle slightly superiorly displaced (white arrows). The fourth ventricle in JS-05 was normally sized but superiorly displaced. Both patients JS-09 and JS-05 show a superior vermian dysplasia. The middle and inferior segments of the vermis are hypoplastic. B, Axial images at the level of the cerebellar hemispheres show that the middle and inferior segments of the vermis are hypoplastic in patients JS-09 and JS-05; the two hemispheres are not separated by any vermian structure. Patient JS-09 had also an hemispheric cerebellar dysplasia. C, Axial images showing the MTS (white arrow) at the level of the superior cerebellar peduncles in patients JS-09 and JS-05.

Figure  3. .

Analysis of two MKS3 splice-site mutations at the RNA level. A, RT-PCR with primers located in MKS3 exons 20 (forward) and 26 (reverse) in the father of patient JS-05, JS-05, and a control individual. Expected wild-type size (754 bp) is seen in the control individual, whereas in patient JS-05 the smaller band (617 bp) corresponds to the transcript without exon 23. The father of patient JS-05 shows both normal and mutated transcripts. B, Direct sequencing of the RT-PCR of patient JS-05 shows a skipping from exons 22 to 24. C, Direct sequencing of the RT-PCR of patient JS-09 showing the wild-type (upper line) and abnormal (lower line) transcripts corresponding to the insertion of 16 intronic bp (boxed).

JS-09 is a mildly affected 7-year-old girl who was also reported elsewhere (as “case 11”).¹⁵ She presented with hypotonia, ataxia, oculomotor apraxia, and abnormal eye movements. She had no breathing anomalies or retinal, renal, or hepatic involvement. She had mild motor delay (walked at age 2.5 years) and moderate mental retardation. Brain magnetic resonance imaging (MRI) showed a cerebellar dysplasia with an MTS, an enlarged fourth ventricle, and no supratentorial anomaly (fig. 2). We identified three MKS3 variations: a donor–splice-site mutation in intron 6 (IVS6+2T→G) inherited from her mother, a missense mutation located in exon 16 (G545E), and a mutation located at the last base of exon 21 (2341G→A). The last two changes were inherited from the father, and both could be deleterious. The G545E mutation concerns a conserved amino acid in vertebrates and is predicted to be pathogenic by the PolyPhen program. The 2341G→A mutation was predicted by the SSF program to decrease the donor–splice-site score. RNA extracted from lymphocyte cell lines of the patients and RT-PCR analysis (performed with the same primers as those for the patient JS-05) did not show an abnormal fragment on electrophoresis. However, the sequencing identified a 16-bp insertion in the transcript, corresponding to the first 16 bp of intron 21 (fig. 3C). An alternative intronic splice site is used when the mutated allele is present. This frameshift mutation predicts a truncated protein with a premature stop codon 14 aa downstream. Whereas the maternal IVS6+2T→G and the paternal G545E mutations were not identified in 380 and 392 chromosomes, respectively, the 2341G→A variation was identified once (1/332 control chromosomes). Since 2341G→A was found in the heterozygote state in the patient, we do not know if this mutation retains some normal splicing. It is, therefore, unclear whether the pathogenic paternal allele results from the G545E missense mutation, the 2341G→A splice mutation, or a combination of the two variations.

Patient NPH-786 is a 7-year-old boy presenting developmental delay, cerebellar ataxia, abnormal breathing, and vermis agenesis. No MRI was performed, and the presence of an MTS or vermis dysgenesis could not be assessed. Since he did not show cystic dilatation of the posterior fossa, a DW malformation was ruled out. The association of hyperechogenic kidneys with cysts and severe hepatic involvement with the vermis agenesis led to the diagnosis of JS in this patient. We identified two missense mutations, in exons 6 (R213C) and 21 (D711A). Both substitutions are conserved in mammals¹¹ and were absent in 380 and 332 chromosomes, respectively. Parental DNA samples were not available for mutation analysis. They are both predicted to be pathogenic by the PolyPhen program, with a PSIC score difference of 1.8 for both changes. In addition, D711 is the first amino acid after the third putative transmembrane domain of meckelin,¹¹ and the substitution of a hydrophobic amino acid, A711, may cause the aberrant inclusion of this alanine in the transmembrane helix, as predicted by the TMHMM Server.

The identification of MKS3 mutations in three patients with JS defines MKS3 as the sixth JS locus (JBTS6). Despite a relatively small JS series, mutations in the MKS3 gene could account for ∼10% (3/22) of patients with JS. No MKS1 mutation was identified, suggesting that allelism between JS and MKS is specific to the MKS3 locus.

MKS and JS are disorders featuring a hindbrain malformation and renal and hepatic involvement and are generally included in the “congenital hepatorenal fibrocystic” group of diseases.¹⁶ Although a severe and lethal cystic kidney dysplasia is a consistent feature of MKS, nephronophthisis or cystic dysplastic kidneys are associated with JS in some patients, defining the CORS group of JS-related disorders. Concerning the brain malformation, cerebellar vermis agenesis is the cardinal feature of JS, and occipital encephalocele is the most-frequent MKS brain anomaly. Interestingly, some patients with a cerebellar vermis agenesis also present an occipital encephalocele (reviewed by Satran et al.¹⁷), and intrafamilial variability has been reported in discordant monozygous twins presenting JS with or without encephalocele.¹⁸ DW malformation has been described in both disorders. In MKS, sibs were reported with either occipital encephalocele or DW malformation¹⁹ or both.²⁰ Patients with JS and DW malformations or variants have also been discussed in many cases.^21–23 However, these patients did not undergo brain MRI, and the state of the brainstem is unknown. Also, the diagnosis of DW malformation was based on either ultrasound or neuropathological examination results. The neuropathological definition of DW malformation differs from the radiological definition and includes partial or complete vermis agenesis with cystic dilatation of posterior fossa, with or without hydrocephalus, but does not take into account the imaging definition—in particular, on brainstem and high tentorium—that would distinguish among JS, a DW malformation, or a variant. Confusion still exists nowadays, especially for prenatal cases in which MRI was not performed, and many fetuses with a DW malformation diagnosed at neuropathological examination could be affected with JS.

Hepatic involvement has been reported in JS, mainly in the form of congenital hepatic fibrosis,^17,24 but bile-duct proliferation of liver (a consistent feature of MKS¹³) was also described at the postmortem examination of a boy with JS who died at age 12 mo.²⁵ To support this observation, we describe in this report two fetuses that also had a bile-duct proliferation of liver. It is not clear if the MKS3 missense and splice-site mutations (table 1) that we have identified are hypomorphic alleles, but such a suggestion would be reasonable given the viability and marked phenotypic overlap of the Wpk rat, the model animal for MKS, with human JS. The Wpk rat carries a missense mutation in Mks3 and does not develop the full range of CNS malformations that are typical of human MKS.¹¹

Allelism between JS and MKS has been suggested elsewhere,^26,27 and whether patients with JS with occipital encephalocele compose a distinct entity or are a part of the MKS phenotypic spectrum has been debated.²⁸ The very recent identification of CEP290 mutations in two patients with JS with occipital meningoencephalocele⁴ provides support that occipital encephalocele belongs to the spectrum of JS. These results suggest that the rare patients reported to have a long-survival MKS because of an occipital meningocele were more likely to have JS.^26,29,30 The phenotypic overlap between these two disorders is also illustrated by the considerable phenotypic variability observed in JBTS2-linked patients with multiorgan involvement, including occipital encephalocele, polydactyly, microphthalmia, and kidney disease.^9,31 In view of the present study, at least one gene, MKS3, can lead to both phenotypes. It is possible that the allelic nature of these conditions may be extended to the MKS2 locus, which has been mapped to chromosome 11q13³² and may therefore be allelic to JBTS2, which has been mapped to 11p12-11q13.3.⁹ However, Valente et al. has refined the JBTS2 interval to the region between D11S4191 and D11S1344 at 11p11.2-11q12.1, which would argue against this interpretation.³¹

No MKS1 mutations were identified in the patients with JS, which suggests that the allelism is specific to the MKS3 locus. In a recent study, we analyzed MKS1 and MKS3 genes in 54 fetuses with MKS and found a phenotype/genotype correlation according to the mutated gene. Whereas the occipital encephalocele was constant in fetuses with MKS1 mutations, DW malformations were observed in some fetuses with MKS3, and polydactyly was significantly less frequent in fetuses with MKS3 mutations (R.K., unpublished data). The identification of MKS3 mutations in patients with JS adds to the phenotypic variability associated with MKS3 mutations, ranging from a severe and lethal cystic kidney dysplasia with exencephalocele to a mild form of MTS with normal kidneys and moderate mental retardation. Since no obvious genotype/phenotype correlation could be observed concerning the type or localization of MKS3 mutations in patients with MKS or JS, one can hypothesize that oligogenic inheritance and/or modifier genes are responsible for the highly variable phenotypes associated with MKS3 mutations. MKS1 is unlikely to be involved, since we did not identify MKS1 mutations in any patients with the MKS3 mutation.

Genes involved in NPHP/JS and MKS encode proteins involved in ciliary function, and the conditions are collectively known as “ciliopathies.”³³ Little is known about the function of AHI1, whereas NPHP genes leading to NPHP with (NPHP1 and CEP290^4,34) or without (NPHP2,³⁵ NPHP4,³⁶ and NPHP5³⁷) cerebellar vermis hypoplasia encode proteins located at the primary cilia and centrosomes. In addition, some of these proteins have been shown to interact, suggesting that they belong to the same signaling pathway.^35,36,38 Comparative genomics also suggests that MKS1 and MKS3 encode ciliary proteins.^10,11 Bardet-Biedl syndrome (BBS [MIM 209900]) is another multisystemic genetic disorder that has phenotypic overlap with MKS³⁹ and JS and is also thought to be a ciliopathy.⁴⁰ In BBS, the genetic basis of phenotypic variability is provided by oligogenism, since three mutated BBS alleles are necessary for the expression of the disease in some families,⁴¹ whereas a third mutated allele modulates the phenotype in other instances.⁴² Epistatic interactions between proteins interacting and colocalizing with the BBS proteins were also demonstrated in BBS.⁴³ Oligogenism is also suggested in NPHP and JS. Indeed, in several patients, only one mutation was identified at NPHP3³⁸ or CEP290⁴ loci. In our study, a single MKS3 heterozygous mutation, F29L, was also identified in a fetus with vermis agenesis, a small occipital bone defect, cystic kidney dysplasia, and no bile-duct proliferation of liver. Either we failed to find the second molecular event or this mutation, lying in the putative peptide signal of meckelin, plays an epistatic role, along with mutations in another gene.

The existence of a distinct phenotype/genotype correlation in JS with some rare cases of phenotypic variability argues for such an oligogenic model. Indeed, whereas AHI1 mutations lead to JS with retinal involvement and sometimes polymicrogyria,⁴⁴ AHI1 mutations were found only in two patients with kidney involvement.⁴⁵ Also, NPHP1 is responsible for isolated NPHP in the vast majority of cases; of patients with JS, only several with a mild form of the brain malformation^6,46 and one with retinal involvement⁴⁷ were found with the NPHP1 deletion. One can hypothesize that oligogenic or epistatic inheritance is responsible for the additional signs found in these rare patients. The probability of such an additional molecular event is high, in view of the large number of genes involved in this group of ciliopathies and retinal dystrophies.³³ In contrast with AHI1 and NPHP1, extensive phenotypic variability is observed in patients with CEP290 mutations who have both renal and ocular variable involvement and in two patients with occipital encephalocele. CEP290 mutations were even shown to be a frequent cause of isolated Leber congenital amaurosis.⁴⁸ This large variability is also observed in patients with MKS3 mutations. A phenotype/genotype correlation is also found at JBTS1 and JBTS2 loci: whereas no organ involvement is described in JBTS1-linked patients, a large variability is associated with JBTS2-linked patients with multiorgan involvement, including occipital encephalocele, polydactyly, microphthalmia, and kidney involvement.³¹

The identification of genes leading to vermis agenesis, DW malformation, or occipital encephalocele supports the existence of an embryologic mechanism common to these brain malformations. The encephalocele may be secondary to the brainstem and cerebellum herniation and not to a primary neural tube defect. Indeed, in most patients with MKS who underwent detailed neuropathological examination,^49,50 occipital encephalocele consists of extruded rhombic roof elements and third ventricle and a distended fourth ventricle through a widened posterior fontanelle. However, a second bony defect was frequently observed below the meningocele or meningoencephalocele and cannot be due to the herniation. In addition, a wide spectrum of CNS anomalies are usually observed in MKS, including heterotopia, micropolygyria, optic nerve and chiasma hypoplasia, and midline defects ranging from corpus callosum agenesis to arhinencephaly, hypothalamic fusion, and complete holoprosencephaly. Whether these prosencephalic and rhombencephalic dysgeneses could also be mechanical consequences of the exencephalocele is questionable. On the other hand, malformations of pontine and medullary structures, as well as cerebellar neuron heterotopias, are described in patients with JS,⁵¹ and mutations in AHI1 lead to JS with cortical polymicrogyria,⁴⁴ which suggests a more general defect in axon guidance in JS. Primary cilia are found in many brain cell types (reviewed by Fuchs et al.⁵²). Whether they act as mechanosensory organelles of the cerebrospinal fluid or are involved in the planar cell polarity pathway⁵³ or both is still unknown. It has recently been shown that intraflagellar transport proteins are necessary for neuronal cilia biogenesis and for Hedgehog signal transduction,⁵⁴ providing a potential molecular explanation for the midline defects/holoprosencephaly that can be observed in MKS.

In conclusion, the identification of MKS3 gene mutations in four patients with JS defines the MKS3 gene as the sixth locus for JS (JBTS6). Although the malformative spectrum observed in MKS is more severe, the identification of MKS3 mutations in patients with JS suggests that JS and MKS are allelic disorders and that other and still-unknown molecular defects could be responsible for the variable phenotypic expression of MKS3 mutations.