Cranio-lenticulo-sutural dysplasia is caused by a SEC23A mutation leading to abnormal endoplasmic-reticulum-to-Golgi trafficking

© 2006 Nature Publishing Group http://www.nature.com/naturegenetics

LETTERS

Cranio-lenticulo-sutural dysplasia is caused by a SEC23A mutation leading to abnormal endoplasmic-reticulumto-Golgi trafficking Simeon A Boyadjiev1,2, J Christopher Fromme3, Jin Ben4, Samuel S Chong4, Christopher Nauta2, David J Hur1, George Zhang1, Susan Hamamoto3, Randy Schekman3, Mariella Ravazzola5, Lelio Orci5 & Wafaa Eyaid6 Cranio-lenticulo-sutural dysplasia (CLSD) is an autosomal recessive syndrome characterized by late-closing fontanels, sutural cataracts, facial dysmorphisms and skeletal defects mapped to chromosome 14q13–q21 (ref. 1). Here we show, using a positional cloning approach, that an F382L amino acid substitution in SEC23A segregates with this syndrome. SEC23A is an essential component of the COPII-coated vesicles that transport secretory proteins from the endoplasmic reticulum to the Golgi complex. Electron microscopy and immunofluorescence show that there is gross dilatation of the endoplasmic reticulum in fibroblasts from individuals affected with CLSD. These cells also exhibit cytoplasmic mislocalization of SEC31. Cell-free vesicle budding assays show that the F382L substitution results in loss of SEC23A function. A phenotype reminiscent of CLSD is observed in zebrafish embryos injected with sec23a-blocking morpholinos. Our observations suggest that disrupted endoplasmic reticulum export of the secretory proteins required for normal morphogenesis accounts for CLSD.

Roughly a third of all cellular proteins are synthesized in association with endoplasmic reticulum membranes. These proteins are retained in the endoplasmic reticulum lumen for proper folding, oligomerization and post-translational modifications. COPII-coated transport vesicles represent the first step of the intracellular secretory pathway responsible for the transport of properly modified proteins from the endoplasmic reticulum to the endoplasmic reticulum–Golgi intermediate compartment and ultimately to the cis-Golgi complex (reviewed in refs. 2,3). The secretory pathway in eukaryotic cells has been extensively studied and well characterized, mostly by genetic and biochemical analyses of temperature-sensitive conditional mutants of Saccharomyces cerevisiae4–6. The COPII coat is a polymer complex formed by at

a

b

+/+

+/–

Figure 1 Craniofacial features of CLSD with depiction of the 1144T-C SEC23A mutation. (a) Top, disease progression in an affected male at 4 yr (left) and 9 yr (middle), and cranial radiogram demonstrating persistent ossification defect (right). Bottom, lateral view of an affected female showing marked forehead hyperpigmentation at 28 months (left), facial appearance of the same female at 7 yr (middle), and cranial radiogram showing wide-open anterior fontanel at 28 months (right). (b) All six affected individuals were homozygous for the 1144T-C transition in exon 10 of SEC23A, which was not present in 600 control chromosomes. (c) Protein sequence alignment shows that F382 is invariably conserved in at least ten species, including Arabidopsis thaliana.

–/–

F382L

c 356 358

S. cerevisiae S. pombe C. elegans A. thaliana D. rerio G. gallus C. familiaris R. norvegicus M. musculus P. troglodytes H. sapiens

415 361 358 358 358 358 349 190 358

1McKusick-Nathans Institute of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, USA. 2Section of Genetics, Department of Pediatrics, University of California, Davis, Sacramento, California 95817, USA. 3Department of Molecular and Cellular Biology, Howard Hughes Medical Institute University of California, Berkeley, California 94720, USA. 4Department of Pediatrics, Yong Loo Lin School of Medicine, National University of Singapore, 119074 Singapore. 5Department of Cell Physiology and Metabolism, University of Geneva Medical Center, 1211 Geneva 4, Switzerland. 6Department of Pediatrics, King Fahad Hospital, Khasim Alaan, 11426 Riyadh, Saudi Arabia. Correspondence should be addressed to S.A.B. ([email protected]).

Received 23 March; accepted 4 August; published online 17 September 2006; doi:10.1038/ng1876

1192

VOLUME 38

[

NUMBER 10

[

OCTOBER 2006 NATURE GENETICS

LETTERS

© 2006 Nature Publishing Group http://www.nature.com/naturegenetics

a

SEC23A+/+ PDI

b SEC23A

+/+

COL1A1

c

Merge

d

SEC23A–/– PDI

e SEC23A

–/–

COL1A1

f

Merge

5 µM

5 µM

g

SEC23A+/+ SEC31

h

10 µM

least five well-characterized proteins: SAR1, SEC23, SEC24, SEC13 and SEC31. Several isoforms of these proteins have been identified, and a distinct coat structure is responsible for the endoplasmic reticulum export of specific sets of secretory proteins7–10. Formation of the COPII complex is initiated when the small GTPase SAR1 becomes anchored to the cytosolic surface of the endoplasmic reticulum, which requires GDP/GTP exchange by SEC12, an integral membrane glycoprotein11,12. Activated SAR1 directly binds the SEC23-SEC24 complex, a heterodimer protein responsible for recognition of cargo proteins in the membrane. Cargo molecules tethered to SAR1 and SEC23-SEC24 are coated by the SEC13-SEC31 complex, forming buds and vesicles destined for the Golgi apparatus. In addition to its role in coating and stabilizing COPII vesicles and selective recruitment of cargo proteins in the budding vesicle13, SEC23A functions as a SAR1-specific GTPase-activating protein that hydrolyzes the GTP bound to SAR1, which in turn triggers uncoating of the vesicles and exposure of the SNARE proteins needed for vesicle fusion to an acceptor compartment7. The SEC24 component of the COPII coat selects cargo proteins for export from the endoplasmic reticulum9,14. The precise sites of interaction of SAR1, SEC23 and SEC24 in the COPII complex have been determined, but it remains unclear where exactly SEC13 and SEC31 bind15,16. Conditional yeast sec23 mutants show distention of the terminal portions of the endoplasmic reticulum at restrictive temperatures owing to a block of the secretory pathway at the endoplasmic reticulum exit sites4. Similar cellular changes are present in Caenorhabditis elegans sec-23 mutants, which show mislocalization of the collagen DPY-7 and defects in morphogenesis and exoskeleton formation17. CLSD was originally described in five males and one female from a large consanguineous Saudi Arabian family of Bedouin descent1 (Fig. 1a, Supplementary Note, Supplementary Figs. 1–3 and Supplementary Table 1 online). We previously showed linkage to chromosome 14q13–q21 by genome-wide scan, with a maximum lod score of 4.58 at GATA126A04 (ref. 1). We identified 47 known and predicted genes in the candidate region. PSMA6, GARNL1, BRMS1L and MBIP were screened by direct sequencing of cDNA generated by RT-PCR, and no variations were identified. We observed a 1144T-C transition in SEC23A that segregated in a homozygous form in all affected individuals and was

NATURE GENETICS VOLUME 38

[

NUMBER 10

[

OCTOBER 2006

SEC23A–/– SEC31

Figure 2 Immunofluorescence analysis. (a–c,g) Wild-type fibroblasts; (d–f,h) Mutant fibroblasts. PDI immunofluorescence shows the fine reticular appearance of the endoplasmic reticulum in wild-type fibroblasts (a), as opposed to the marked dilatation of the endoplasmic reticulum in mutant fibroblasts (d). Immunofluorescence of wild-type (b) and homozygous mutant (e) cells visualizing procollagen COL1A1 shows similar structures with clear colocalization of PDI and COL1A1 (c and f, respectively). Immunofluorescence with SEC31 antibody produces punctate staining mostly in the perinuclear region in wild-type (g) and diffuse cytoplasmic mislocalization in mutant (h) fibroblasts.

not present in 600 control chromosomes (Fig. 1b). The F382L substitution involved a residue that is invariably conserved in at least ten species (Fig. 1c). 10 µM On the basis of the known biological function of SEC23A, we predicted excessive accumulation of secretory proteins in the rough endoplasmic reticulum of the mutant fibroblasts. Using an antibody to the intralumenal endoplasmic reticulum chaperone PDI18, we detected abundant vacuolar structures in homozygous mutant cells that we interpreted as distended endoplasmic reticulum (Fig. 2a–f). Immunofluorescence with an antibody to procollagen COL1A1 showed substantial accumulation

a

SEC23A+/+

0.5 µm

b

c

SEC23A+/–

SEC23A–/–

0.5 µm

0.5 µm

Figure 3 Electron microscopy. Wild-type (a), SEC23A heterozygous (b) and SEC23A homozygous mutant (c) fibroblasts shows the appearance of normal, moderately dilated and grossly dilated endoplasmic reticulum, respectively.

1193

LETTERS

–

–

–

–

+ –

+

+ –

+

+ –

+

+ –

+

SEC23A-SEC24D F382L SEC23A-SEC24D

D

D

T

T

D

D

T

T

GDPβS or GTPγS

did not reveal abnormal cellular phenotypes. There was no difference in the abundance of SEC23A mRNA or protein among wild-type, heterozygous and homozygous mutant fibroblasts, as assessed by northern and western blotting (data not shown). Wild-type and F382L SEC23A heterozygous and homozygous mutant fibroblasts were examined by thin-section electron microscopy. Most wild-type cells showed a typical organization of the rough endoplasmic reticulum with narrow cisternae (Fig. 3a), and about 10% of cells showed mild focal dilatation of the endoplasmic reticulum. Roughly 35% of heterozygous fibroblast sections showed a moderate generalized dilatation of endoplasmic reticulum (Fig. 3b). More than 80% of the homozygous mutant cells had endoplasmic reticulum cisternae that were greatly distended by an accumulation of secretory material (Fig. 3c). The skeletal phenotype of CLSD predominantly involves the membranous bones of the calvaria, but vertebral and pelvic bone defects and generalized joint hypermobility also occur (Supplementary Note online). These clinical features suggest abnormal formation of the bone and connective tissue, potentially as a result of a defect in the secretion of collagen and/or other related extracellular matrix proteins. A defect in collagen secretion was suggested by immunofluorescence (Fig. 2e). We further analyzed collagen secretion by immunoblot analysis of secreted and intracellular COL1A1, but we did not observe a reproducible quantitative defect in secretion of the COL1A1 protein from skin fibroblasts under cell culturing conditions. To assess the functional consequence of the F382L substitution, we purified recombinant F382L SEC23A in complex with histidine-tagged SEC24D from baculovirus-infected insect cells. In the presence of activated SAR1B, the F382L SEC23A-SEC24D complex bound to synthetic membranes at a level similar to that of the wild-type protein (Fig. 4a). We next tested the function of F382L SEC23A in an in vitro vesicle formation assay20. Under conditions requiring the addition of purified SEC23-SEC24 complex for the formation of cargo-containing

SEC24D SEC23A (WT or F382L) SAR1B

b

20% input

– – – – –

Floats (membrane bound) – 4 – – –

+ 4 – – –

+ 1 – – –

+ 1 + – –

+ 1 + + –

+ 1 + – +

ATPr + GTP Rat liver cytosol (mg/ml) SEC13-SEC31A + SAR1B SEC23A-SEC24D F382L SEC23A-SEC24D Ribophorin-I p58 Sec22b

Figure 4 In vitro studies of F382L SEC23A. (a) Liposome-binding assay shows that, when SAR1B is activated (triphosphate bound), the mutant protein binds to synthetic membranes to an extent similar to that of the wild-type protein. D, GDPbS; T, GTPgS. (b) Vesicle formation assay shows that F382L SEC23A has markedly lower activity for generating cargocontaining vesicles as compared with the wild-type protein. The endoplasmic reticulum resident protein ribophorin-I was used as a negative control; p58 and Sec22b are COPII cargo proteins. ATPr, ATP regeneration system.

a

14-som

of this protein in endoplasmic reticulum cisternae identical in morphology to those marked by PDI (Fig. 2d–f). Immunofluorescence with an antibody to SEC31 showed diffuse cytoplasmic mislocalization of this protein in the mutant fibroblasts, suggestive of abnormal formation of the COPII complex (Fig. 2g,h). Additional immunofluorescence experiments with antibodies to SAR1, SEC24A, SEC24B, SEC24C, SEC22B and GM130, a cis-Golgi matrix protein19,

j

e

1 d.p.f.

Bud

Shield

64-cell

1k-cell

Figure 5 Developmental expression of sec23a and loss-of-function phenotype in zebrafish. (a) RT-PCR analysis detected sec23a transcript sec23a MO at the one-cell stage, suggesting that sec23a is present as a maternal transcript. (b,c) WholeActin mount in situ hybridization analysis confirmed WT the presence of maternal transcript at the onecell stage (b) until the 1,000-cell mid-blastula transition stage (c). (d) At the 12-somite stage, weak but distinct expression is detected in the nc developing notochord. (e) Notochord expression is pf strongest in the 1-d.p.f. embryo, especially in the WT MO tail bud region and the ventral tail edge (inset: aud sco sco 25-somite stage). In 2-d.p.f. embryos, expression act act ep is no longer detectable in the notochord, but ed ed begins to be observed in the developing head m cartilages (not shown). (f–i) Strong expression is ch cb: 1 2 3 4 5 observed in the neurocranial and viscerocranial pop pop MO WT cartilages of the head in 3-d.p.f. embryos cnp (f, ventral view; g, lateral view), the cranial act sco project and bulge of the otic vesicle (h), and the ed cnb scapulocoracoid and postcoracoid processes and distal edge of the endoskeletal disc of the pectoral fin (i). (j–q) Loss-of-function 5-d.p.f. pop WT MO morphants show reduced body length and dorsal curvature (j) as compared with wild-type larvae (k), kinked pectoral fins (l,m) owing to a larger non-cartilaginous fin segment at the distal edge (n,o), and malformation or dysgenesis of the head cartilages (p,q). 1k-cell, 1,000-cell stage; 14-som, 14-somite; MO, morphant; WT, wild type; act, actinotrichs; aud, auditory capsule; cb1–cb5, ceratobranchial arches 1–5; ch, ceratohyal arch; cnb, cranial bulge; cnp, cranial project; ed, endoskeletal disc; ep, ethmoid plate; m, Meckel’s cartilage; nc, notochord; pf, pectoral fin; pop, postcoracoid process; sco, scapulocoracoid process. 1-cell

© 2006 Nature Publishing Group http://www.nature.com/naturegenetics

Inputs

No RNA

a

k

b

c

d

l

f

m

g

n

h

1194

i

o

p

VOLUME 38

[

NUMBER 10

q

[

OCTOBER 2006 NATURE GENETICS

© 2006 Nature Publishing Group http://www.nature.com/naturegenetics

LETTERS vesicles from donor endoplasmic reticulum membranes, F382L SEC23A-SEC24D showed inefficient vesicle formation in vitro (Fig. 4b). Activity was not restored even if a threefold excess of the mutant protein complex was added to the assay (data not shown). These biochemical data thus show that F382L SEC23A can bind SAR1B and SEC24D, but has lost an important aspect of its function in vesicle budding. On the basis of these biochemical data and SEC31 immunofluorescence in homozygous mutant cells (Fig. 2g,h), the mutant is probably defective in its ability to form a competent complex with SEC13-SEC31. We next determined the developmental expression pattern of the zebrafish (Danio rerio) ortholog of SEC23A using reverse transcription PCR (RT-PCR) and whole-mount in situ hybridization analyses, and carried out a morpholino-mediated knockdown analysis to assess the phenotypic consequences of sec23a loss of function. RT-PCR analysis of 16 embryonic stages detected sec23a transcript as early as the onecell stage, indicating the presence of maternal transcript (Fig. 5a). This expression was confirmed by whole-mount in situ hybridization analysis (Fig. 5b,c). Distinct expression in the developing notochord was observed at the 12-somite stage (Fig. 5d). Notochord expression was strongest at roughly the 25-somite to 1 day post fertilization (d.p.f.) stage (Fig. 5e), especially in the tail bud region and ventral tail edge. Notochord and tail bud expression was undetectable by 2 d.p.f. (data not shown). At this stage, however, some expression in the developing head cartilages was observed (data not shown). In 3-d.p.f. embryos, there was strong expression in all main neurocranial and viscerocranial cartilages of the head (Fig. 5f,g). Weak expression was also detected in the cranial project and bulge of the otic vesicle (Fig. 5h). In addition, distinct sec23a expression was observed both in the scapulocoracoid and postcoracoid processes of the pectoral fin and in the distal edge of the endoskeletal disc, but not in the surrounding actinotrichs (Fig. 5i). To try to recreate the CLSD phenotype in zebrafish, we designed two morpholinos targeting different regions immediately upstream of the translation start site, and complementary to sec23a but not sec23b. To verify that both morpholinos specifically inhibited the translation of sec23a, we designed expression vectors containing the 5¢ end of sec23a or sec23b fused in-frame with the enhanced green fluorescent protein gene (EGFP). The transcripts from these constructs were injected into zebrafish embryos at the one- to four-cell stage in the presence of negative control morpholino or either of the two sec23a morpholinos, enabling us to verify the effectiveness and specificity of sec23a translation inhibition (Supplementary Fig. 4 online). Both morpholinos produced identical phenotypes. Morphant 5d.p.f. hatchlings had a reduced body length and a dorsally oriented curvature as compared with uninjected hatchlings (Fig. 5j,k), and showed reduced and upward swimming motion, suggesting the possibility of an effect on vertebral development analogous to the scoliosis observed in CLSD. Unfortunately, the morphants did not survive beyond 9 d.p.f., which precluded analysis of the bone defects because hatchlings at 5–7 d.p.f. lack an ossified skeleton. Kinking of the pectoral fins was also observed in morpholino-injected hatchlings from 4 d.p.f. onwards (Fig. 5l,m). Alcian blue staining of pectoral fins of 5-d.p.f. morphants revealed an elongated actinotrich region as compared with wild-type zebrafish, in which the whole endoskeletal disc appeared intact (Fig. 5n,o). Given the absence of sec23a expression in the actinotrichs, these results suggest that actinotrich outgrowth is regulated by sec23a expression at the distal edge of the endoskeletal disc. As expected, sec23a morphants showed morpholino dosage– dependent malformations of all principal neurocranial and viscero-

NATURE GENETICS VOLUME 38

[

NUMBER 10

[

OCTOBER 2006

cranial cartilage structures, including the ethmoid plate, parachordal and Meckel’s cartilage, the auditory capsules, ceratohyal and ceratobranchial arches (Fig. 5p,q). Although detailed characterization of these cranial defects is difficult in the absence of an adult morphant or germline mutant, the observed changes and weak Alcian blue staining suggest generalized cartilage hypoplasia, most probably caused by aberrant or delayed chondrogenesis and/or delayed deposition of extracellular matrix molecules. These observations are consistent with the idea that sec23a has a role in craniofacial development and reflect the observed craniofacial defects in CLSD. We suspect that the specificity of the CLSD phenotype is determined by the tissue-specific expression of SEC23A versus SEC23B, as observed in zebrafish (S.S.C., unpublished data) and/or by disrupted endoplasmic reticulum export of secretory proteins required for normal morphogenesis. Abnormal secretion of collagen could explain the skeletal features observed in individuals affected with CLSD; however, the mechanisms leading to cataract formation remain unclear. Further studies of the CLSD developmental phenotype in animal models are needed to address these important issues. In conclusion, we have delineated CLSD as a dysmorphic genetic syndrome with characteristics of a skeletal dysplasia and have identified its genetic cause. Our experiments show that CLSD occurs as a result of defective COPII-mediated endoplasmic reticulum export owing to loss of function of SEC23A. As a result, collagen and (probably) other secretory proteins accumulate and distend the endoplasmic reticulum, ultimately leading to the clinical manifestations of CLSD. The relatively mild phenotype of affected individuals suggests that the 1144T-C SEC23A mutation is a hypomorph and that the mutant protein retains some residual functional activity. Further studies of the mutant cells and/or a SEC23A animal model will allow more precise identification of the cargo proteins retained in the endoplasmic reticulum as a result of mutations in SEC23A. The characteristic phenotype of the SEC23A mutant cells suggests that screening methods could be developed to facilitate the identification of other human disorders caused by defects in endoplasmic-reticulum-to-Golgi trafficking. Analysis of the orthologous sec23a gene in zebrafish revealed an anatomical and morphological correlation with the human CLSD phenotype, providing further evidence that loss of SEC23A function is responsible for this genetic syndrome. Although the endoplasmic-reticulum-to-Golgi trafficking has been extremely well characterized by both genetic and biochemical methods, very few human disorders21–23 have been attributed to defects in its individual components. The functional redundancy of the COPII pathway is likely to lead to non-lethal phenotypes that have escaped classification. A systematic survey of tissues from similar bone morphogenesis diseases may uncover other previously unknown mutant alleles of the COPII machinery. METHODS Subjects. Blood samples, fibroblast cells and photos were obtained, with informed consent, in accordance with protocols approved by the institutional review boards of the Johns Hopkins University and the King Khaled Eye Specialist Hospital of the Kingdom of Saudi Arabia. RT-PCR, candidate gene screening and TaqMan assays. Primary fibroblast cell lines from a heterozygous carrier and a homozygous affected individual were established. Total RNA was extracted with TRIzol reagent (Invitrogen) and purified with an RNeasy micro kit (Qiagen) in accordance with the manufacturers’ instructions. cDNA was prepared with Superscript II reverse transcriptase (Invitrogen). All known genes and putative transcripts in the region between markers D14S1014 and D14S301 on chromosome 14 were identified by using the Human Genome Browser (http://genome.ucsc.edu). Candidate

1195

© 2006 Nature Publishing Group http://www.nature.com/naturegenetics

LETTERS genes in the area were prioritized in accordance with their function and expression pattern with relevance to CLSD. Primers were designed to amplify and sequence the whole coding regions of these genes from cDNA. The whole SEC23A coding region was amplified in four overlapping segments (with primer sequences based on GenBank accession code NM_006364 in Supplementary Table 2 online). We carried out PCR with Platinum Taq (Invitrogen) and purified the PCR products from solution using shrimp alkaline phosphatase and Exonuclease I (US Biochemical) or from gel bands using a QIAquick gel extraction kit (Qiagen). A Sequenase v. 2.0 DNA sequencing kit (US Biochemical) and the ABI Prism 3700 fluorescent DNA analyzer (Applied Biosystems) were used for direct sequencing of PCR products. A custom TaqMan assay (Applied Biosystems) for the 1144T-C mutation was designed, and 300 individuals were tested. Cell culture and transfection. Primary fibroblast cell lines were derived from skin biopsies from affected and carrier individuals. Control fibroblasts were obtained from the American Type Culture Collection (ATCC 2091). Fibroblasts were maintained in MEM media with 20% fetal bovine serum. Immunofluorescence microscopy. Fibroblasts (2.5  105) were plated on glass coverslips in six-well plates 24 h before staining. Cells were washed briefly with PBS, fixed in 3% paraformaldehyde for 10 min at room temperature (22 1C) and then permeabilized in 0.5% Triton-X 100 for 3 min at room temperature. We used the following primary antibodies for indirect Immunofluorescence: polyclonal rabbit antibodies to SAR1 (a gift from C.M. Mansbach, University of Tennessee, Memphis), SEC23 (a gift from F. Gorelick, Yale University), SEC24A, SEC24B and SEC24C (a gift from J.-P. Paccaud, Athelas, Switzerland), Sec22b (a gift from J. Hay, University of Montana) and SEC31 (a gift from A. Hubbard, Johns Hopkins University); LF-67 anti-COL1A1 rabbit serum24 (a gift from L.W. Fisher, National Institutes of Health (NIH)); mouse monoclonal antibody to GM130 (a gift from C. Machamer, Johns Hopkins University); and monoclonal antibody to PDI (a gift from S. Fuller, EMBL-Heidelberg). We used the secondary antibodies Alexa Fluor 488–conjugated goat antibody to rabbit IgG, Alexa Fluor 546–conjugated goat antibody to mouse IgG, and Texas red–conjugated goat antibody to mouse IgG. After staining the cells with the appropriate primary and secondary antibodies, we visualized and imaged them with an LSM 510 confocal microscope (Zeiss). Electron microscopy. Confluent fibroblast cell cultures were fixed in 2% glutaraldehyde in 0.1 M phosphate buffer, gently scraped, centrifuged, post-fixed in osmium tetroxide, dehydrated and embedded in Epon. Thin sections were sequentially stained with uranyl acetate and lead citrate, and examined by a Philips electron microscope. In vitro assays. A baculovirus expression vector of human SEC23A was created by inserting a PCR product from a cDNA template (IMAGE accession code 4821858) into the pFastBac1 plasmid (Invitrogen). The 1144T-C mutation was subsequently introduced by means of primer-directed mutagenesis. Complexes of F382L SEC23A and His-tagged SEC24D, wild-type SEC23A and His-tagged SEC24D, and human SEC13 and His-tagged SEC31A were purified from baculovirus-infected SF9 cells as described20. Human SAR1B cDNA (SARA2 gene) was subcloned into the pGEX-2T vector (GE Healthcare) and purified as described for hamster Sar1a (ref. 20). Liposome-binding assays using major-minor mix liposomes were done essentially as described in ref. 24 (a gift from L.W. Fisher, NIH);25, with incubation at 37 1C for 20 min before centrifugation. NIH-3T3 cells were permeabilized with digitonin26 and used as donor endoplasmic reticulum membranes for in vitro vesicle formation reactions as described20. Zebrafish expression analysis. Total RNA was extracted from Singapore wildtype zebrafish embryos at different stages and analyzed for sec23a expression by using the SuperScript III One-Step RT-PCR System (Invitrogen). A 1,262-bp sec23a cDNA fragment was amplified from 1 mg of total RNA template over 35 step cycles. A 561-bp actin cDNA internal positive control fragment was coamplified. A digoxigenin-labeled antisense or sense RNA probe corresponding to the 3¢ region of sec23a cDNA (2341–2893 of GenBank accession code BC097063.1) was generated by cloning the 553-bp cDNA fragment, after

1196

amplification, into pCRII-TOPO (Invitrogen), followed by either linearization with EcoRV and in vitro transcription with SP6 RNA polymerase, or linearization with BamHI and in vitro transcription with T7 RNA polymerase, respectively. Whole-mount in situ hybridization was carried out as described27. All primer sequences are listed in Supplementary Table 2 online. Zebrafish morpholino knockdown analysis. Two different translation-blocking antisense morpholinos (MO-P and MO-Q; Supplementary Table 2 online) specific to sec23a mRNA were designed to target the 5¢ UTR sequence of zebrafish sec23a transcript (Gene Tools). To verify their in vitro effectiveness and specificity, the first 140 bp of coding sequence and adjoining 64 bp of the 5¢ UTR of sec23a cDNA (GenBank accession code BC052768), and the first 44 bp of coding sequence and adjoining 90 bp of the 5¢ UTR of sec23b cDNA (NM_199777), were cloned upstream of the EGFP gene in the pT7TS vector (a gift from P. Krieg, University of Texas, Austin). In vitro transcription of BamHI-linearized plasmids with T7 RNA polymerase produced in-frame fusion transcripts encoding 46 amino acids from SEC23a followed by EGFP peptide and 14 amino acids from SEC23b upstream of the EGFP peptide, respectively. Roughly 140 pg of capped sec23aEGFP or sec23b-EGFP mRNA and 8 ng of a negative control morpholino (Supplementary Table 2 online), MO-P or MO-Q, were co-injected into embryos at the one- to four-cell stage. After 8 h, injected embryos were observed under a fluorescence stereomicroscope (Leica). Singapore wild-type zebrafish embryos between the one- and four-cell stage were injected with 8 ng of either MO-P or MO-Q. Live and Alcian blue–stained 4- and 5-d.p.f. hatchlings were observed under a bright-field microscope. Images were taken with a DP50 digital camera (Olympus) and processed by Adobe Photoshop Version 7 software, or with a MicroPublisher color digital camera (QImaging) and processed by Image-Pro Plus Version 5 software (Media Cybernetics). Note: Supplementary information is available on the Nature Genetics website. ACKNOWLEDGMENTS We thank all members of the family that participated in this project; C. Machamer and A. Hubbard for discussions and help with Immunofluorescence; J. Mendell for help with expression vectors, D. Murphy and C. Cooke for assistance with electron microscopy; A. Fischer for help with cell culture; J. Kim and B. Kleizen for assistance with in vitro assays; the Pole Facultaire de Microscopie Ultrastructurale (PFMU) at the University of Geneva Medical School for access to electron microscopy equipment and L. Liu for assistance with zebrafish imaging. This work was supported by grants from the National Institute of Dental and Craniofacial Research–US National Institutes of Health (DE16342 and DE00462 to S.A.B.) and the Swiss National Science Foundation (to L.O.); J.C.F. is supported by a fellowship from the Miller Institute for Basic Research; R.S. is supported by funds from the Howard Hughes Medical Institute and S.S.C. is supported by grants from the NUS (R-178-000-080-112 and R-178-000-104-112). AUTHOR CONTRIBUTIONS This study was initiated by S.A.B., who interpreted clinical and radiologic data, designed and performed experiments, and wrote the manuscript with contributions from J.C.F., R.S., L.O. and S.S.C.; W.E. performed the clinical assessment and provided fibroblast cell lines for these studies; S.A.B., D.J.H. and G.Z. performed the cloning experiments; R.S. and J.C.F. designed, performed and interpreted in vitro liposome-binding and vesicle-formation assays and contributed to immunofluorescence experiments. L.O., M.R. and S.H. performed electron microscopy and immunofluorescence analysis with contributions from C.N. and S.A.B.; L.O. critically interpreted microscopy data and suggested experiments; J.B. and S.S.C. performed the zebrafish expression analysis, the morpholino knockdown experiments and the analysis of the morphants. COMPETING INTERESTS STATEMENT The authors declare that they have no competing financial interests. Published online at http://www.nature.com/naturegenetics Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions/ 1. Boyadjiev, S.A. et al. A novel dysmorphic syndrome with open calvarial sutures and sutural cataracts maps to chromosome 14q13–q21. Hum. Genet. 113, 1–9 (2003).

VOLUME 38

[

NUMBER 10

[

OCTOBER 2006 NATURE GENETICS

© 2006 Nature Publishing Group http://www.nature.com/naturegenetics

LETTERS 2. Duden, R. ER-to-Golgi transport: COP I and COP II function (review). Mol. Membr. Biol. 20, 197–207 (2003). 3. Schekman, R. & Orci, L. Coat proteins and vesicle budding. Science 271, 1526–1533 (1996). 4. Novick, P., Field, C. & Schekman, R. Identification of 23 complementation groups required for post-translational events in the yeast secretory pathway. Cell 21, 205–215 (1980). 5. Novick, P. & Schekman, R. Secretion and cell-surface growth are blocked in a temperature-sensitive mutant of Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 76, 1858–1862 (1979). 6. Novick, P. & Schekman, R. Export of major cell surface proteins is blocked in yeast secretory mutants. J. Cell Biol. 96, 541–547 (1983). 7. Barlowe, C. et al. COPII: a membrane coat formed by Sec proteins that drive vesicle budding from the endoplasmic reticulum. Cell 77, 895–907 (1994). 8. Antonny, B. & Schekman, R. ER export: public transportation by the COPII coach. Curr. Opin. Cell Biol. 13, 438–443 (2001). 9. Miller, E., Antonny, B., Hamamoto, S. & Schekman, R. Cargo selection into COPII vesicles is driven by the Sec24p subunit. EMBO J. 21, 6105–6113 (2002). 10. Miller, E.A. et al. Multiple cargo binding sites on the COPII subunit Sec24p ensure capture of diverse membrane proteins into transport vesicles. Cell 114, 497–509 (2003). 11. Nakano, A., Brada, D. & Schekman, R. A membrane glycoprotein, Sec12p, required for protein transport from the endoplasmic reticulum to the Golgi apparatus in yeast. J. Cell Biol. 107, 851–863 (1988). 12. Barlowe, C. & Schekman, R. SEC12 encodes a guanine-nucleotide-exchange factor essential for transport vesicle budding from the ER. Nature 365, 347–349 (1993). 13. Aridor, M., Weissman, J., Bannykh, S., Nuoffer, C. & Balch, W.E. Cargo selection by the COPII budding machinery during export from the ER. J. Cell Biol. 141, 61–70 (1998). 14. Pagano, A. et al. Sec24 proteins and sorting at the endoplasmic reticulum. J. Biol. Chem. 274, 7833–7840 (1999).

NATURE GENETICS VOLUME 38

[

NUMBER 10

[

OCTOBER 2006

15. Antonny, B., Madden, D., Hamamoto, S., Orci, L. & Schekman, R. Dynamics of the COPII coat with GTP and stable analogues. Nat. Cell Biol. 3, 531–537 (2001). 16. Bi, X., Corpina, R.A. & Goldberg, J. Structure of the Sec23/24-Sar1 pre-budding complex of the COPII vesicle coat. Nature 419, 271–277 (2002). 17. Roberts, B., Clucas, C. & Johnstone, I.L. Loss of SEC-23 in Caenorhabditis elegans causes defects in oogenesis, morphogenesis, and extracellular matrix secretion. Mol. Biol. Cell 14, 4414–4426 (2003). 18. Bottomley, M.J., Batten, M.R., Lumb, R.A. & Bulleid, N.J. Quality control in the endoplasmic reticulum: PDI mediates the ER retention of unassembled procollagen C-propeptides. Curr. Biol. 11, 1114–1118 (2001). 19. Nakamura, N. et al. Characterization of a cis-Golgi matrix protein, GM130. J. Cell Biol. 131, 1715–1726 (1995). 20. Kim, J., Hamamoto, S., Ravazzola, M., Orci, L. & Schekman, R. Uncoupled packaging of amyloid precursor protein and presenilin 1 into coat protein complex II vesicles. J. Biol. Chem. 280, 7758–7768 (2005). 21. Jones, B. et al. Mutations in a Sar1 GTPase of COPII vesicles are associated with lipid absorption disorders. Nat. Genet. 34, 29–31 (2003). 22. Tiller, G.E. et al. A recurrent RNA-splicing mutation in the SEDL gene causes X-linked spondyloepiphyseal dysplasia tarda. Am. J. Hum. Genet. 68, 1398–1407 (2001). 23. Zhang, B. et al. Bleeding due to disruption of a cargo-specific ER-to-Golgi transport complex. Nat. Genet. 34, 220–225 (2003). 24. Fisher, L.W., Lindner, W., Young, M.F. & Termine, J.D. Synthetic peptide antisera: their production and use in the cloning of matrix proteins. Connect. Tissue. Res. 21, 43–48 (1989). 25. Matsuoka, K. & Schekman, R. The use of liposomes to study COPII- and COPI-coated vesicle formation and membrane protein sorting. Methods 20, 417–428 (2000). 26. Wilson, R. et al. The translocation, folding, assembly and redox-dependent degradation of secretory and membrane proteins in semi-permeabilized mammalian cells. Biochem. J. 307, 679–687 (1995). 27. Ben, J., Jabs, E.W. & Chong, S.S. Genomic, cDNA and embryonic expression analysis of zebrafish IRF6, the gene mutated in the human oral clefting disorders Van der Woude and popliteal pterygium syndromes. Gene Expr. Patterns 5, 629–638 (2005).

1197