Group 1: Ruti Parvari1, Eli Hershkovitz2, Nili Grossman3, Rafael Gorodischer2, Bart Loeys4, Alexandra Zecic5, Geert Mortier4, Simon Gregory6, Reuven Sharony7

Group 2: Marios Kambouris8, Nadia Sakati8, Brian F. Meyer8

Group 3: Aida I. Al Aqeel8,9, Abdul Karim Al Humaidan8, Fatma Al Zanhrani8, Abdulrahman Al Swaid9, Johara Al Othman9

Group 4: George A. Diaz10,11, Rory Weiner10, K. Tahseen S. Khan12, Ronald Gordon13 & Bruce D. Gelb10,11

Departments of 1Developmental Molecular Genetics and 2Pediatrics, 3Skin Bank and Investigative Dermatology Laboratory and Department of Microbiologyand Immunology, Soroka Medical Center and Faculty of Health Sciences, Ben Gurion University of the Negev, Beer Sheva 84105, Israel. 4Department of Med-ical Genetics and 5Department of Neonatology, Ghent University Hospital, Ghent, Belgium. 6The Wellcome Trust Sanger Institute, Hinxton, Cambridge, UK.7The Genetic Institute, Sapir Medical Center, Meir Hospital, Sackler School of Medicine, Tel Aviv University, Israel. 8King Faisal Specialist Hospital & ResearchCenter, Riyadh 11211, Kingdom of Saudi Arabia. 9Department of Pediatrics, Riyadh Armed Forces Hospital, Kingdom of Saudi Arabia. Departments of10Human Genetics and 11Pediatrics, Mount Sinai School of Medicine, One Gustave Levy Place, New York, New York 10029, USA. 12Department of Pediatrics,Al-Jahra Hospital, Safat, Kuwait. 13Department of Pathology, Mount Sinai School of Medicine, New York, New York, USA. Correspondence should beaddressed to G.A.D. (e-mail: george.diaz@mssm.edu).

letter

448 nature genetics • volume 32 • november 2002

Mutation of TBCE causes hypoparathyroidism–retardation–dysmorphism and autosomal recessiveKenny–Caffey syndrome

The HRD/Autosomal Recessive Kenny–Caffey Syndrome Consortium

The syndrome of congenital hypoparathyroidism, mentalretardation, facial dysmorphism and extreme growth failure(HRD or Sanjad–Sakati syndrome; OMIM 241410) is an autoso-mal recessive disorder reported almost exclusively in MiddleEastern populations1–3. A similar syndrome with the addi-tional features of osteosclerosis and recurrent bacterial infec-tions has been classified as autosomal recessive Kenny–Caffeysyndrome4 (AR-KCS; OMIM 244460). Both traits have previ-ously been mapped to chromosome 1q43–44 (refs 5,6) and,despite the observed clinical variability, share an ancestralhaplotype, suggesting a common founder mutation7. Wedescribe refinement of the critical region to an interval ofroughly 230 kb and identification of deletion and truncationmutations of TBCE in affected individuals. The gene TBCEencodes one of several chaperone proteins required for theproper folding of α-tubulin subunits and the formation ofα–β-tubulin heterodimers. Analysis of diseased fibroblastsand lymphoblastoid cells showed lower microtubule densityat the microtubule-organizing center (MTOC) and perturbedmicrotubule polarity in diseased cells. Immunofluorescenceand ultrastructural studies showed disturbances in subcellular

organelles that require microtubules for membrane traffick-ing, such as the Golgi and late endosomal compartments.These findings demonstrate that HRD and AR-KCS are chaper-one diseases caused by a genetic defect in the tubulin assem-bly pathway, and establish a potential connection betweentubulin physiology and the development of the parathyroid.Haplotype analysis of pedigrees of Israeli Bedouin and Palestin-ian families affected with HRD using new polymorphic markersflanking the critical region associated with HRD showed that allof the Middle Eastern pedigrees that we identified shared a com-mon founder (Fig. 1a). These studies also identified additionalrecombination events that refined the critical region between

Published online 21 October 2002; doi:10.1038/ng1012

Fig. 1 Haplotype data for markers tightly linked to the critical region. a, Criticalrecombinant chromosomes defining the telomeric (individual PA) and cen-tromeric (individual SA) boundaries, with the consensus haplotype found inhomozygosity in non-recombinant chromosomes shown at right. b, Transcriptmap and clone coverage of the critical region. The relative positions ofmapped exons are shown at top, with orientation indicated by arrows. Theapproximate position of elements is indicated along the horizontal axis in kilo-bases (kb). Thin, vertical lines through the axis represent gaps in the sequencecontig. PAC and BAC clones that have been completely (dark bars) or partially(light bars) sequenced at the Wellcome Trust Sanger Institute or the MITWhitehead Center for Biomedical Research (293G6) are represented at bot-tom. BCAA, retinoblastoma binding protein-like 1 gene; GGPS1, geranylger-anylpyrophosphate synthase 1 gene; TBCE, tubulin-specific chaperone E gene;LOC148789, novel predicted gene with galactosyltransferase-like domain.

a

b

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

reg

enet

ics

letter

nature genetics • volume 32 • november 2002 449

markers 617f4CA and 860f22CA. We mapped the new criticalregion markers onto a sequence assembly of roughly 230 kb cre-ated manually from overlapping BAC and PAC clones (Fig. 1b).The derived contig contained the same number of gaps (nine) asthe estimate obtained from the most recent available HumanGenome Browser data freeze (April 2002) and was of similar size.

Sequence analysis of the critical region identified genes encod-ing geranylgeranyl pyrophosphate synthase 1 (GGPS1), tubulin-specific chaperone E (TBCE), the first two exons ofretinoblastoma-binding protein–like 1 (BCAA) and an alterna-tively spliced 12-exon gene predicted by GENSCAN8 that had aregion of homology to the family of β-1,3-galactosyltransferases(Fig. 2). All spliced expressed sequence tags (ESTs) in the criticalregion could be assigned to one of these four genes. Mutationanalysis of these candidate genes showed a deletion of 12 bp inthe second coding exon of TBCE in all of the affected MiddleEastern individuals that we evaluated (>50; Fig. 2a). The deletionwas not present in more than 350 control chromosomes fromArab individuals, excluding it as a common polymorphism inthis population. Analysis of a Belgian pedigree in which two sib-lings manifested features typical of HRD showed that the onesurviving sibling was a compound heterozygote with respect to adeletion of 2 bp in the first coding exon and a nonsense mutationin exon 12 of TBCE (Fig. 2a).

Tubulin is a heterodimer composed of α- and β-subunits thatrequire both the cytosolic chaperonin9 and a set of tubulin-specificchaperones (A–E)10,11 for proper folding. The gene TBCE encodesa chaperone required for the folding of α-tubulin and its het-erodimerization with β-tubulin. Assembled tubulin moleculessubsequently polymerize into microtubules in a polar fashion,with protofilaments nucleating at the MTOC and extensionoccurring at the rapidly growing (plus) end. The protein TBCEcontains two known functional motifs, a cytoskeleton-associatedprotein glycine-rich (CAP-Gly) α-tubulin–binding domain12

and a series of leucine-rich repeats (Fig. 2b), which can mediateprotein–protein interactions13. The del52–55 mutation deletedfour residues in the CAP-Gly domain, including a highly con-served glycine, and was adjacent to residues essential for α-tubulinbinding12,14 (Fig. 2c). The two mutations identified in the Belgiansiblings caused, respectively, a frameshift after residue 22 with ter-mination at residue 48 (Val23fs48X), and truncation of the proteinafter residue 370 (Cys371X), distal to the leucine-rich repeatmotifs. Epitope-tagged wildtype, del52–55 and Cys371X con-structs were stable when overexpressed (data not shown), raisingthe possibility that the mutant proteins could retain residual activ-ity. Further work will be necessary to explore the activity of themutants in tubulin folding.

To assess the effect of TBCE mutation on cellular tubulin assem-bly, we used AR-KCS lymphoblastoid and HRD fibroblast and ker-atinocyte cell lines derived from individuals who werehomozygous with respect to the del52–55 mutation. Tubulin pro-tein abundance and α-tubulin transcript levels (Fig. 3a,b) weresimilar in diseased and control lymphoblastoid samples, but the

Fig. 2 TBCE mutation analysis and expression studies. a, Sequence traces fromaffected individuals. All affected Middle Eastern subjects evaluated werehomozygous with respect to the 155–166del mutation, and a Belgian individualwith HRD was compoundly heterozygous with respect to 66–67delAG and1113T→A (mutation positions are relative to +1 at the initiation ATG). Themutated sequences are shown above the electropherograms, with the positionsof deletion or point mutations indicated by arrows and wildtype sequencesgiven above the arrows. b, Mutations in TBCE relative to the CAP-Gly andleucine-rich repeat domains. Positions of truncating mutations (Val23fs48X,Cys371X) are indicated by arrows, and the in-frame deletion (del52–55) is indi-cated by an arrowhead. Approximate amino-acid positions are indicated belowthe cartoon. c, ClustalW alignment of the CAP-Gly domains for TBCE, the bud-ding yeast ortholog Pac2p, TBCB (tubulin cofactor B), the TBCB fission yeastortholog Alf1p and tandem repeats of the microtubule-binding protein CLIP-170. A trio of substitutions introduced by site-directed mutagenesis thatabrogated microtubule binding by CLIP-170 and Alf1p are indicated abovethe alignment12,14. Residues deleted in the del52–55 mutant are indicated bythe black bar. d, Expression pattern of TBCE. A multiple-tissue northern blot(Clontech) was probed with a radiolabeled amplicon corresponding to the 3′end of the TBCE transcript, then stripped and reprobed with a radiolabeledamplicon specific for β-actin. Approximate sizes of transcripts are indicatedat right. H, heart; B, brain; Pl, placenta; Lu, lung; Li, liver; M, skeletal muscle;K, kidney; Pa, pancreas.

Fig. 3 Tubulin expression in cells homozygous with respect to the del42–55 mutation and in control lymphoblastoid cells. a, Western blot of protein lysates fromlymphoblastoid cells was probed with an antibody specific to α-tubulin (top), then stripped and reprobed with antibodies specific to either β-tubulin (middle) oractin (bottom). 1, del42–55 sample; 2, control sample. b, Northern blot of total cellular RNA from lymphoblastoid cells probed with a radiolabeled α-tubulinRT–PCR amplicon (top), then stripped and reprobed with a probe specific to β-actin. 1, del52–55 homozygote; 2, del52–55 heterozygote; 3, unrelated controlsample. c, Fractionation of unpolymerized soluble tubulin and precipitable microtubule-incorporated tubulin from lymphoblastoid cell protein lysates by ultra-centrifugation at 100,000g. The proportion of precipitable α-tubulin was lower in diseased cells than in controls. Total, soluble and precipitable α-tubulin isshown for control (lanes 1–3) and del52–55 homozygous (lanes 4–6) lymphoblastoid cells.

a

b

c

d

a b c

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

reg

enet

ics

letter

450 nature genetics • volume 32 • november 2002

proportion of precipitable, microtubule-incorporated α-tubulinwas lower in diseased cells (Fig. 3c). Immunofluorescence studiesshowed lower α-tubulin staining at the MTOC in diseased lym-phoblastoid cells with aberrant polarity of the microtubule net-work (Fig. 4a,b, insets). We also observed lower MTOC-associatedtubulin staining in diseased fibroblasts (Fig. 4c,d) and slightly, butreproducibly, lower peripheral staining intensity in keratinocytes(Fig. 4e,f). The fibroblasts and keratinocytes shown are from oneindividual with HRD, suggesting that the cellular phenotypemay be tissue-specific despite ubiquitous transcriptional TBCEexpression (Fig. 2d). Phenotypic differences (that is, presence orabsence of osteosclerosis) between individuals who are homozy-gous with respect to the del52–55 mutation might therefore berelated to inter-pedigree variability in tissue-specific micro-tubule perturbation.

The absence of the MTOC and the microtubule polarity defect infibroblasts and lymphoblastoid cells that were homozygous withrespect to the del52–55 mutation was notable, given the role of theMTOC in the organization of the Golgi complex15 and in the estab-lishment of cellular polarity in some cell types16,17. The distributionof MTOC-associated organelles was therefore assessed in diseasedfibroblasts. Organization of the Golgi complex in control cells,visualized with a fluorescently labeled lectin (GSII), was compactand juxtanuclear, whereas in diseased fibroblasts the Golgi complexsurrounded the nuclear envelope diffusely (Fig. 5a–d). We assessedthe distribution of late endosomes positive for Rab7 (ref. 18), whichis microtubule-dependent, by indirect immunofluorescence. Incontrast with the predominantly perinuclear pattern observed incontrol cells, staining in diseased cells was more diffuse, typicallyencircling the nucleus completely and extending farther towardsthe plasma membrane (Fig. 5e,f). Ultrastructural studies of dermalfibroblasts from affected individuals confirmed the observed

abnormalities of the vesicular compartments; we observed largeelectrolucent vacuoles, lamellated membranous inclusions anddilated endoplasmic reticulum in diseased cells (Fig. 6). It has yet tobe established whether the vesicular abnormalities result directlyfrom disorganization of MTOC-associated organelles or fromsome other mechanism, such as a disruption of microtubule-dependent trafficking related to defects in microtubule polarity,and whether the observed changes will also be present in cells thatare functionally hemizygous with respect to the Cys371X mutation,in which the CAP-Gly domain is intact.

The findings presented here identify mutations in TBCE inindividuals affected with HRD/AR-KCS and provide a new exam-ple of a chaperone-related human disease. To our knowledge, thisis the first report of a disease caused by a defect in the tubulinfolding and assembly pathway. Notably, recent work reported inan accompanying paper19 demonstrates that the autosomal reces-sive progressive motor neuronopathy (pmn) phenotype in mice is

caused by a missense mutation in Tbce(Trp524Gly) that seems to destabilize theprotein and reduce steady-state levels. Neu-rodegeneration in pmn mice affects descend-ing motor neurons but spares sensory

Fig. 4 Immunostaining for α-tubulin in lymphoblastoid, dermal fibroblast andepidermal keratinocyte cells from affected del52–55 homozygotes. Represen-tative control (a) and diseased (b) lymphoblastoid cells incubated with mouseantibody against α-tubulin and FITC-labeled rabbit secondary antibody againstmouse. White arrows indicate cells shown in inverted image mode in insets, inwhich a radial pattern of microtubules emanating from the MTOC wasobserved in control cells, whereas microtubule-network polarity was lost in dis-eased cells. Immunostaining of α-tubulin in control (c) and diseased (d) dermalfibroblasts showed lower microtubule density at the perinuclear MTOC in dis-eased cells. Immunostaining of α-tubulin in control (e) and diseased (f) epider-mal keratinocytes showed a modest decrease in α-tubulin staining (mostnotable at cell periphery) in diseased cells, but no difference in microtubuledistribution was apparent.

Fig. 5 Subcellular localization of Golgi and late endo-somal compartments. Control (a) and del52–55homozygous (b) dermal fibroblasts were staineddirectly with the GC-specific Griffonia simplicifolialectin GS-II Alexa Fluor 488 conjugate (MolecularProbes) and DAPI nuclear stain. c,d, GC-specific stain-ing was compact in control cells (c) but dispersedaround the nucleus and cytoplasm in diseased cells(d). Late endosomes positive for Rab7 were visualizedby indirect immunofluorescence in representativecontrol (e) and diseased (f) fibroblasts. The distribu-tion of Rab7-positive vesicles in control cells was gen-erally perinuclear with an asymmetric distributionpredominantly to one side of the nucleus (whitearrowheads), whereas diseased cells had a more uni-form and more extensive distribution.

a b

c d

e f

a b

c d

e f

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

reg

enet

ics

letter

nature genetics • volume 32 • november 2002 451

neurons, and leads to death at six weeks from degeneration of thephrenic nerve. The divergent human and mouse phenotypes mayreflect mutation-specific genotype–phenotype variability,mouse-specific sensitivity to microtubule perturbation in motorneurons, lethality in mice before features in common with thehuman phenotype could be observed or some combination ofthese factors. We note that the original description of pmn micereported that brain mass and body weight were 40% and 60%lower, respectively, in pmn mice than in wildtype mice at birth,and that spermatogenesis was defective in six-week-old mice20.These findings are reminiscent of the microcephaly, growth fail-ure and hypogonadism seen in humans with TBCE mutations,and suggest that, despite the differences in postnatal course, simi-larities between the two syndromes may exist in other tissues.

The phenotype of pmn mice is not unexpected in light of thepredominantly neurological manifestations of mutations affect-ing human microtubule-associated proteins such as tau21 (fron-totemporal dementia with parkinsonism), LIS1 (ref. 22) andDCX23,24 (lissencephaly), the microtubule motor KIF1Bβ (ref. 25;Charcot–Marie–Tooth disease type 2A) and MID1 (ref. 26; Opitzsyndrome). X-linked retinitis pigmentosa is caused by mutationof RP2 (ref. 27), a protein with homology to tubulin-specificchaperone C but no involvement in tubulin heterodimerization28.The pleiotropic disease manifestations of HRD/AR-KCS affect tis-sues with abundant microtubules, such as brain and testis. Thespecific absence of parathyroid glands, with accompanying nor-mal development of the thyroid and other branchial-pouch deriv-atives (R.P. and E.H., unpublished observations), and theoccurrence of osteosclerosis in a subset of affected individuals4

are, however, intriguing and unexpected aspects of a derangementin tubulin physiology. Aberrant microtubule polarity as observedin diseased lymphoblastoid cells might affect numerous func-tions, including intracellular transport, signal transduction andcellular migration. Additional studies will be necessary to deter-mine the mechanisms operative in the pathophysiology ofHRD/AR-KCS. Analysis of existing and additional animal modelsmay be of value in this regard.

MethodsSubjects. We diagnosed and recruited Middle Eastern subjects withHRD/AR-KCS for the positional cloning project as described elsewhere1,3–7.Individuals with AR-KCS (8 pedigrees, 13 affected individuals) differedfrom those with HRD/SSS (17 Saudi pedigrees, 27 affected individuals; 9Israeli pedigrees, 25 affected individuals) owing to the additional presenceof medullary stenosis of the long bones, calvarial osteosclerosis and suscep-

tibility to bacterial infection. The presence of patchy osteosclerosis in thelong bones of some Saudi subjects with HRD (M.K., unpublished observa-tions) and deaths secondary to sepsis in some Israeli Bedouin individualswith HRD (E.H., unpublished observations) suggested variable expressionof these phenotypic features in a pedigree-specific fashion.

We identified a non-consanguineous Belgian pedigree in which twomale siblings had features of HRD including facial dysmorphism, rela-tively short limbs, small hands and feet, small genitalia, hypoparathy-roidism and severe pre- and postnatal growth retardation. An oldersibling who died from bronchiolitis with sepsis had aortic coarctation,which has not been described in previous cases. Analysis by fluorescencein situ hybridization excluded interstitial deletion of chromosome 22q11.We enrolled subjects with informed consent approved by local institu-tional review boards (King Faisal Specialist Hospital & Research Center,Mount Sinai School of Medicine, Soroka Medical Center) and in accor-dance with the Helsinki agreement. We extracted DNA from peripheralblood samples using the Puregene kit (Gentra Systems) and establishedlymphoblastoid cell lines immortalized with Epstein–Barr virus. We alsoderived dermal fibroblast and epidermal keratinocyte primary cell linesfrom punch skin-biopsy samples.

Critical region refinement. We developed short tandem repeat markersfrom sequenced BAC and PAC clones in the critical interval using TandemRepeats Finder29, carried out standard genotyping PCR reactions andresolved products on 7% sequencing gels. To determine marker heterozy-gosity, we typed candidate short tandem repeats on a panel of parental car-rier DNAs and used potentially informative markers for genotyping.Primer sequences for the markers used are available upon request.

Sequence analysis. We used the NIX software package (UK-HGMP), theHuman Genome Browser and the MacVector software package for genom-ic sequence analysis. To generate sequence contigs, we manually alignedunassembled sequence fragments using physical mapping data forsequence-tagged site markers (data not shown) and the exonic structure ofknown genes in the interval. We prioritized predicted genes in the intervalon the basis of the number of algorithms identifying a candidate exon andthe identification of EST clones that aligned to the predicted exons. Weidentified a 12-exon gene candidate predicted by the GENSCAN programthat corresponded largely to the National Center for Biotechnology Infor-mation locus 148789 (Acembly locus Hs1_31876_28_1_2015), amplified itby RT–PCR from diseased lymphoblastoid cDNA and sequenced it (datanot shown). In addition to the publicly available genomic sequence data-base, we also evaluated the critical region in the proprietary Celera humangenome sequence database. No additional candidate genes were identified.

Candidate-gene mutation analysis. We did initial mutation analysis bygenerating overlapping RT–PCR amplicons using Superscript II reversetranscriptase (Roche) followed by sequencing on ABI377 or ABI3700 auto-mated sequencers and sequence trace analysis using the AutoAssemblerprogram (ABI). We amplified the exonic and flanking intronic regions ofTBCE and sequenced them using primer pairs developed from availablegenomic and cDNA sequence data (available upon request). We then useda PCR primer pair that amplified a 138-bp product from wildtype genom-ic DNA to screen for the presence of a 126-bp mutation-specific product byelectrophoresis in 1.5% agarose gels (data not shown).

Cell culture. We cultured lymphoblastoid cell lines immortalized byEpstein–Barr virus from subjects diagnosed with AR-KCS, their parents

Fig. 6 Ultrastructural studies of diseased dermal fibroblasts. Control and diseasedfibroblasts homozygous with respect to del52–55 were grown to confluence in60-mm culture dishes, fixed in 3% glutaraldehyde, secondarily fixed in 1%osmium tetroxide and mounted for transmission electron microscopy. a, Typicaldiseased cell at low magnification (3,000×) showing large electrolucent vacuoles(asterisks), scroll-like lamellated inclusions (arrowheads) and dilated endoplas-mic reticulum (arrows). b, Higher-magnification view (20,000×) of typical vesicu-lar abnormalities. c, Some diseased cells showed less severe pathology withsmaller vacuoles and lamellations. d, Rare disease cells had recognizable Golgicomplex, though developing vacuoles and inclusions were still present. e, Mor-phologically normal control cell shown for comparison. N, nucleus; GC, Golgicomplex; ER, endoplasmic reticulum.

a b c

d e

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

reg

enet

ics

letter

452 nature genetics • volume 32 • november 2002

and unaffected controls in RPMI medium (Cellgro) with 10% fetal bovineserum (Gibco), and we cultured dermal fibroblasts or keratinocytesderived from Israeli Bedouin or Saudi individuals with HRD in DMEMwith 10% fetal bovine serum and specific supplements30. We stored low-passage cultures and control human dermal fibroblasts purchased from theHuman Genetic Cell Repository (Coriell Institute for Medical Research)frozen for use in subsequent microscopy studies.

Protein analysis. For protein extraction, we pelleted AR-KCS or controllymphoblastoid cells (106 per assay), washed them in PBS and resuspend-ed them in 100 µl lysis buffer (50 mM Tris (pH 8.0), 150 mM NaCl,0.02% sodium azide, 1% (v/v) Triton X-100) containing a cocktail ofprotease inhibitors (Complete Protease Inhibitor Tablets; Roche). Wediluted protein extracts in 2× Laemmli sample buffer (LSB), resolvedthem by SDS–PAGE and transferred them to nylon membranes. Weprobed blots with monoclonal mouse primary antibodies specific for α-tubulin or β-tubulin (Sigma) and peroxidase-labeled sheep secondaryantibody against mouse (Sigma), and visualized bands using the SuperSignal kit (Pierce). Some blots were stripped and reprobed with an anti-body against actin (Sigma). To assess the polymerization status of tubu-lin in diseased and control lymphoblastoid cells, we pelleted 106 cells atroom temperature, washed them with prewarmed PBS and then lysedthem by sonication in 100 µl microtubule-stabilization buffer (MSB; 0.1 MPIPES (pH 6.75), 1 mM EGTA, 1 mM MgSO4, 30% glycerol, 5 mM GTP,5% DMSO, 1 mM DTT) containing a protease inhibitor cocktail. We col-lected one sample immediately into 2× LSB (total protein) and a parallelsample of the supernatant after ultracentrifugation at 100,000g for 30 minat room temperature into 2× LSB (soluble fraction), washed pellets oncein MSB, resuspended them in an equal volume of MSB and collectedthem into 2× LSB (precipitable fraction). We resolved equal volumes ofthe final protein fractions by SDS–PAGE and determined protein contentof individual fractions by the method of Bradford to ensure that recoverywas approximately equal from sample to sample.

Immunofluorescence and electron microscopy. We cultured dermalfibroblast and keratinocyte cells in chamber slides for 48 h, washed them inPBS, fixed them for 15 min in 3% paraformaldehyde in PBS or in 50%methanol–50% acetone on ice, washed them again and permeabilizedthem for 10 min in 1% Nonidet P-40 (NP-40) at room temperature. Weincubated slides in 2% bovine serum albumin (BSA) in PBS to block non-specific antibody binding, then in 2% BSA in PBS with primary antibodiesfor 60 min at room temperature, and finally in 2% BSA in PBS with sec-ondary antibodies for 30 min, and then washed them in 0.1% NP-40 inPBS at room temperature and mounted them in VectaShield Antifade(Vector Laboratories) with or without added DAPI nuclear stain. Primaryantibodies included mouse monoclonal antibody against α-tubulin B7(Santa Cruz Biotechnology), mouse monoclonal antibody against α-tubu-lin MCS78S (Serotec), mouse monoclonal antibody against β-tubulin D10(Santa Cruz Biotechnology) and goat antibody against Rab7. Secondaryantibodies included FITC-conjugated rabbit antibody against mouse,FITC-conjugated swine antibody against goat and rhodamine-conjugatedgoat antibody against mouse. Alexa Fluor 488–conjugated GSII (MolecularProbes) was diluted in 1 mM CaCl2. For fluorescence microscopy, weviewed samples using a Nikon Eclipse epifluorescence microscope with acooled CCD camera, and collected and processed images using the Photo-shop 6 software package (Adobe).

For ultrastructural analysis, we fixed cultured dermal fibroblasts(roughly 106) in 3% glutaraldehyde in PBS for 3 h, washed them in PBSbuffer, removed them from the culture dish with a rubber policeman, cen-trifuged them, treated them for 1 h with 1% osmium tetroxide, dehy-drated them in graded steps of ethanol through propylene oxide andembedded them in Embed 812. Representative areas for ultrathin sectionswere chosen by light microscopy from 1-µm plastic sections stained withmethylene blue and azure II. We stained ultrathin sections with uranylacetate and lead citrate.

URL. Tandem Repeats Finder is available at http://c3.biomath.mssm.edu/trf.html.

Accession numbers. Homo sapiens TBCE mRNA, NM_003193.

AcknowledgmentsThis research was supported in part by the Israel Science foundation (to R.P.and R.G.), the US National Institutes of Health (the Mount Sinai ChildHealth Research Center, Mount Sinai Microscopy Shared Instrument Facility;to B.D.G. and G.A.D.) and the March of Dimes (to G.A.D.). The authorswould like to thank the families that participated in the study, S. Zhang and P. Hernandez for excellent technical assistance, T. Volberg for help with thetubulin immunofluoresence staining studies and Y. Ioannou for helpfuldiscussions and critical reading of the manuscript. We gratefully acknowledgethe contributions of the Sanger Institute mapping and sequencing groups.

Competing interests statementThe authors declare that they have no competing financial interests.

Received 10 April; accepted 18 September 2002.

1. Sanjad, S.A., Sakati, N.A., Abu-Osba, Y.K., Kaddoura, R. & Milner, R.D. A newsyndrome of congenital hypoparathyroidism, severe growth failure, anddysmorphic features. Arch. Dis. Child. 66, 193–196 (1991).

2. Richardson, R.J. & Kirk, J.M. Short stature, mental retardation, andhypoparathyroidism: a new syndrome. Arch. Dis. Child. 65, 1113–1117 (1990).

3. Hershkovitz, E. et al. The new syndrome of congenital hypoparathyroidismassociated with dysmorphism, growth retardation, and developmental delay—areport of six patients. Isr. J. Med. Sci. 31, 293–297 (1995).

4. Khan, T.S.K. et al. Kenny–Caffey syndrome in six Bedouin sibships: autosomalrecessive inheritance is confirmed. Am. J. Med. Genet. 69, 126–132 (1997).

5. Parvari, R. et al. Homozygosity and linkage-disequilibrium mapping of thesyndrome of congenital hypoparathyroidism, growth and mental retardation,and dysmorphism to a 1-cM interval on chromosome 1q42-43. Am. J. Hum. Genet.63, 163–169 (1998).

6. Diaz, G.A., Khan, K.T. & Gelb, B.D. The autosomal recessive Kenny–Caffeysyndrome locus maps to chromosome 1q42–q43. Genomics 54, 13–18 (1998).

7. Diaz, G.A. et al. Sanjad–Sakati and autosomal recessive Kenny–Caffey syndromesare allelic: evidence for an ancestral founder mutation and locus refinement. Am.J. Med. Genet. 85, 48–52 (1999).

8. Burge, C. & Karlin, S. Prediction of complete gene structures in human genomicDNA. J. Mol. Biol. 268, 78–94 (1997).

9. Lewis, S.A., Tian, G., Vainberg, I.E. & Cowan, N.J. Chaperonin-mediated folding ofactin and tubulin. J. Cell Biol. 132, 1–4 (1996).

10. Tian, G. et al. Pathway leading to correctly folded β-tubulin. Cell 86, 287–296 (1996).11. Tian, G. et al. Tubulin subunits exist in an activated conformational state

generated and maintained by protein cofactors. J. Cell Biol. 138, 821–832 (1997).12. Pierre, P., Pepperkok, R. & Kreis, T.E. Molecular characterization of two functional

domains of CLIP-170 in vivo. J. Cell Sci. 107, 1909–1920 (1994).13. Kobe, B. & Deisenhofer, J. The leucine-rich repeat: a versatile binding motif.

Trends Biochem. Sci. 19, 415–421 (1994).14. Radcliffe, P.A., Hirata, D., Vardy, L. & Toda, T. Functional dissection and hierarchy

of tubulin-folding cofactor homologues in fission yeast. Mol. Biol. Cell 10,2987–3001 (1999).

15. Thyberg, J. & Moskalewski, S. Role of microtubules in the organization of theGolgi complex. Exp. Cell Res. 246, 263–279 (1999).

16. Kupfer, A., Dennert, G. & Singer, S.J. Polarization of the Golgi apparatus and themicrotubule-organizing center within cloned natural killer cells bound to theirtargets. Proc. Natl Acad. Sci. USA 80, 7224–7228 (1983).

17. Kupfer, A., Louvard, D. & Singer, S.J. Polarization of the Golgi apparatus and themicrotubule-organizing center in cultured fibroblasts at the edge of anexperimental wound. Proc. Natl Acad. Sci. USA 79, 2603–2607 (1982).

18. Meresse, S., Gorvel, J.P. & Chavrier, P. The rab7 GTPase resides on a vesicularcompartment connected to lysosomes. J. Cell Sci. 108, 3349–3358 (1995).

19. Martin, N. et al. A missense mutation in Tbce causes progressive neuronopathy inmice. Nat. Genet. 32, 443–447 (2002).

20. Schmalbruch, H., Jensen, H.J., Bjaerg, M., Kamieniecka, Z. & Kurland, L. A newmouse mutant with progressive motor neuronopathy. J. Neuropathol. Exp.Neurol. 50, 192–204 (1991).

21. Hutton, M. et al. Association of missense and 5′-splice-site mutations in tau withthe inherited dementia FTDP-17. Nature 393, 702–705 (1998).

22. Reiner, O. et al. Isolation of a Miller–Dieker lissencephaly gene containing Gprotein β-subunit-like repeats. Nature 364, 717–721 (1993).

23. Gleeson, J.G. et al. Doublecortin, a brain-specific gene mutated in human X-linked lissencephaly and double cortex syndrome, encodes a putative signalingprotein. Cell 92, 63–72 (1998).

24. des Portes, V. et al. A novel CNS gene required for neuronal migration andinvolved in X-linked subcortical laminar heterotopia and lissencephaly syndrome.Cell 92, 51–61 (1998).

25. Zhao, C. et al. Charcot–Marie–Tooth disease type 2A caused by mutation in amicrotubule motor KIF1Bβ. Cell 105, 587–597 (2001).

26. Quaderi, N.A. et al. Opitz G/BBB syndrome, a defect of midline development, isdue to mutations in a new RING finger gene on Xp22. Nat. Genet. 17, 285–291(1997).

27. Schwahn, U. et al. Positional cloning of the gene for X-linked retinitis pigmentosa2. Nat. Genet. 19, 327–332 (1998).

28. Bartolini, F. et al. Functional overlap between retinitis pigmentosa 2 protein andthe tubulin-specific chaperone cofactor C. J. Biol. Chem. 14, 14 (2002).

29. Benson, G. Tandem repeats finder: a program to analyze DNA sequences. NucleicAcids Res. 27, 573–580 (1999).

30. Rheinwald, J.G. Methods for clonal growth and serial cultivation of normalhuman epidermal keratinoctyes and mesothelial cells. in Cell Growth andDivision: A Practical Approach (ed. Baserga, R.) 81–93 (IRL, Oxford, 1989).

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

reg

enet

ics