Impaired glycosylation and cutis laxa caused by mutations in the ...

© 2008 Nature Publishing Group http://www.nature.com/naturegenetics

B R I E F C O M M U N I C AT I O N S

Impaired glycosylation and cutis laxa caused by mutations in the vesicular H+-ATPase subunit ATP6V0A2 Uwe Kornak1, Ellen Reynders2, Aikaterini Dimopoulou1, Jeroen van Reeuwijk3, Bjoern Fischer1, Anna Rajab4, Birgit Budde5, Peter Nu¨rnberg5, Francois Foulquier6, the ARCL Debre´-type Study Group12, Dirk Lefeber7, Zsolt Urban8, Stephanie Gruenewald9, Wim Annaert2, Han G Brunner3, Hans van Bokhoven3, Ron Wevers7, Eva Morava10,13, Gert Matthijs6,13, Lionel Van Maldergem11,13 & Stefan Mundlos1 We identified loss-of-function mutations in ATP6V0A2, encoding the a2 subunit of the V-type H+ ATPase, in several families with autosomal recessive cutis laxa type II or wrinkly skin syndrome. The mutations result in abnormal glycosylation of serum proteins (CDG-II) and cause an impairment of Golgi trafficking in fibroblasts from affected individuals. These results indicate that the a2 subunit of the proton pump has an important role in Golgi function. Congenital disorders of glycosylation (CDG) form a growing class of hereditary disorders caused by defective glycosylation at the level of the endoplasmic reticulum or the Golgi apparatus1. Although most CDG syndromes are caused by loss-of-function mutations in glycosyltransferases or sugar transporters, it was recently discovered that mutations in different subunits of the conserved oligomeric Golgi (COG) complex involved in Golgi membrane trafficking can also disturb glycosylation and cause disease2–4. We recently described several families and sporadic cases with autosomal recessive cutis laxa (ARCL) type II (Debre´ type) (MIM 219200) or wrinkly skin syndrome (WSS, MIM 278250) of varying severity5–7. Besides excessive congenital skin wrinkling, symptoms include a large fontanelle with delayed closure, a typical facial appearance with downslanting palpebral fissures, a general connective tissue weakness, and varying degrees of growth and developmental delay and neurological abnormalities (Fig. 1a,b). Some affected

individuals develop seizures and mental deterioration later in life, whereas the skin phenotype tends to become milder with age. An association of a cutis laxa phenotype with CDG has been previously described7, and wrinkly skin has been observed in an individual with a defect in the COG complex2. On the basis of these findings, we investigated glycosylation of serum proteins isolated from the individuals with ARCL type II described here, by either capillary zone electrophoresis, isofocusing of transferrin and apolipoprotein CIII, or mass spectrometry of glycans from total serum proteins. All affected individuals showed a CDG type 2 (CDG-II) pattern, which corresponds to a defect of N-glycosylation at the level of processing in the Golgi apparatus (Fig. 1c and Supplementary Fig. 1 online). In addition, apolipoprotein CIII isofocusing showed a typical ApoCIII1 profile characterized by elevated levels of the monosialo isoform, indicative of a disturbance of O-glycosylation (Table 1). This was confirmed by mass spectrometric analysis of O-glycans from total serum proteins (Supplementary Fig. 1). The reduced sialic acid content of the glycans from the affected individuals indicates that sialylation, a terminal step of glycan synthesis, is particularly impaired. A careful comparison of clinical data did not show a strict correlation between the severity of the phenotype and the degree of the glycan abnormality. We carried out homozygosity mapping in 15 consanguineous families using Affymetrix GeneChip Human Mapping 10K and 250K arrays and identified a homozygous region on chromosome 12q24 with a maximum lod score of 3.2 in 12 families (Supplementary Fig. 2 and Supplementary Methods online). Using polymorphic microsatellite markers and haplotype analysis, we found that the locus spans 5.7 Mb between D12S395 and D12S304 (Fig. 1d). Knowing that the genetic defect should lead to Golgi dysfunction, we prioritized the candidate genes in the region according to the localization of the gene product in the Golgi apparatus. The products of nine genes were part of the Golgi proteome, as previously described8. One of these genes, ATP6V0A2, turned out to harbor mutations in the consanguineous families and in a sporadic case with ARCL type II (Table 1 and Supplementary Table 1 online). It encodes the a2 subunit of the V-type H+ ATPase, which has been described to reside in endosomes and in a compartment overlapping with the trans-Golgi network (TGN)9,10. The a subunit, for which four isoforms exist, is embedded in the membrane, serves as an anchor for the large ATPase protein complex and is directly involved in proton transport11. The

1Institute for Medical Genetics, Charite ´ Universitaetsmedizin Berlin and Max Planck Institute for Molecular Genetics, Berlin, Germany. 2Laboratory for Membrane Trafficking, Center for Human Genetics, University of Leuven and Department for Molecular and Developmental Genetics, Flanders Institute for Biotechnology (VIB), B-3000 Leuven, Belgium. 3Department of Human Genetics, Radboud University Nijmegen Medical Centre, 6525GA Nijmegen, The Netherlands. 4Genetic Unit, Directorate General of Health Affairs, Ministry of Health, Muscat, 113 Sultanate of Oman. 5Cologne Center for Genomics (CCG) and Institute for Genetics, University of Cologne, 50674 Cologne, Germany. 6Laboratory for Molecular Diagnostics, Center for Human Genetics, University of Leuven, B-3000 Leuven, Belgium. 7Laboratory of Pediatrics and Neurology, University Medical Centre Nijmegen, 6525GA Nijmegen, The Netherlands. 8Department of Pediatrics and Department of Genetics, Washington University, St. Louis, Missouri, 63110 USA. 9Great Ormond Street Hospital for Children Trust NHS, London, WC1N 3JH UK. 10Department of Pediatrics, University Medical Centre Nijmegen, 6525GA Nijmegen, The Netherlands. 11Centre de Geńe´tique Humaine, Centre Hospitalier Universitaire du Sart-Tilman, Universite´ de Lie`ge, B-4000 Lie`ge, Belgium. 12A full list of authors appears at the end of this paper. 13These authors contributed equally to this work. Correspondence should be addressed to U.K. ([email protected]).

Received 15 August; accepted 9 October; published online 23 December 2007; doi:10.1038/ng.2007.45

32

VOLUME 40

[

NUMBER 1

[

JANUARY 2008 NATURE GENETICS


a

c


1,836

2,081 2,040

2,285 3,603 2,431 Affected individual

2,226

Control

1,960

3,480

m/z

d

b

AFM343XE5 AFM330YH9 D12S395 D12S2073 AFM337ZD9 D12S1611 D12S304

AFM343XE5 AFM330YH9 D12S395 D12S2073 AFM337ZD9 D12S1611 D12S304

1 3 2 2 1 3 3

2 3 3 1 1 1 2

1 3 2 2 1 3 3

1 3 2 2 1 3 3

1 3 2 2 1 3 3

1 3 2 2 1 3 3

1 3 2 2 1 3 3

1 3 2 2 1 3 3

2 2 4 2 2 2 0

1 3 2 2 1 3 3

1 3 1 2 1 3 2

1 3 1 2 1 3 2

2 1 2 1 2 1 1

1 2 2 2 1 3 2

1 2 2 2 1 3 2

2 1 2 1 2 1 1

2 3 1 2 1 1 1

1 2 2 2 1 1 1

a2 subunit has also been shown to recruit proteins that regulate vesicular trafficking10. In 12 families, we identified ten different mutations that are predicted to result in a loss of protein function. Four were splicesite mutations, of which three could be confirmed to lead to altered transcripts with either frameshifts or the deletion of an exon (Supplementary Figs. 3 and 4 online). Furthermore, we identified three nonsense and three frameshift mutations. Five mutations lead to a premature stop in the cytoplasmic N terminus, which is thought to mediate interaction with other ATPase subunits, and the other mutations lead to truncations in transmembrane segments III, VI and VIII (Supplementary Fig. 5 online). A complete loss of function is very likely, as a similar distribution of mutations was found in ATP6V0A3 (TCIRG1) in individuals with autosomal recessive osteopetrosis, known to be caused by a loss of the a3 subunit12. Using different markers for all Golgi subcompartments and for the Golgi matrix, we did not identify any major alteration in Golgi

Figure 1 Phenotypes of individuals with autosomal recessive cutis laxa type II or wrinkly skin syndrome. (a) Affected 9-month-old girl from family Ra with typical facial dysmorphism and depression in the frontal calvaria due to a large open fontanelle. Note skin wrinkling at neck and abdomen. (b) Affected individual 3 at 4 years of age showing facial and abdominal skin laxity and typical facial appearance. Informed consent was obtained to publish these photos in this journal. (c) N-glycosylation defect in affected individual CoFe compared to a healthy control determined by MALDI-TOF analysis. Values are normalized to the most abundant ion, m/z 2,792, corresponding to a disialylated biantennary glycan (shown in Supplementary Fig. 1). Only altered peaks are shown. Abnormal glycans m/z 2,081 and m/z 2,226 (green) are detected only in the affected individual. Relative amounts of glycans m/z 1,836 and m/z 2,431 (green) are increased, whereas the triantennary glycan m/z 3,603 (red) was less abundant in the affected individual. (d) Microsatellite analysis of the chromosome 12q24 locus in two consanguineous families from Oman with a common founder. The homozygous haplotype shared by affected individuals is indicated by boxes. Black square, N-acetylglucosamine; gray circle, mannose; gray triangle, fucose; white circle, galactose; black diamond, sialic acid.

morphology in fibroblasts from affected individuals (data not shown). However, we observed a significant delay in the retrograde translocation of Golgi membranes to the endoplasmic reticulum (ER) in fibroblasts of all affected individuals after treatment with brefeldin A (Fig. 2). This delay in Golgi trafficking can best be explained by an impairment of vesicle fusion with the ER. It has been shown that the V0 part of the V-type H+ ATPase is crucial for the fusion of vesicles with the yeast vacuole13. It remains unclear whether this impairment of Golgi-to-ER membrane trafficking is the only cellular abnormality leading to the glycosylation defect. As the main function of the ATPase is proton transport, it is very likely that a defect in pH regulation also results in impaired enzymatic and sorting processes in the Golgi compartment. Golgi pH decreases from the cis to trans cisternae and is estimated to range around 6.2 in the TGN14. Notably, the observed glycosylation defect strongly affects the last steps of glycan synthesis, which take place in the trans Golgi. Acute neutralization of the Golgi pH has been shown to impair glycosylation and lead to an abnormal distribution of glycosylation enzymes, whereas Golgi morphology remains normal15. Given that this experimental disruption of pH is harsh and affects all compartments, it is to be expected that the effects seen in cells from individuals with ARCL type II are rather subtle. Furthermore, the a1 subunit, which has also been detected in the Golgi proteome, is likely to partially compensate for the loss of the

Table 1 Mutations in ATP6V0A2 and their functional effects CDG Family Am, Ra, Ri

Origin

Number affected

Phenotype

N

ATP6V0A2 mutation O

CNS

cDNA

Effect

Oman

9

WSS

+

+

+

c.294+1G4A

V66fsX107

Oman Hispanic

2 1

WSS ARCL type 2

+ +

n.d. +

+ +

c.1929delA c.2293C4T

T643fsX683 Q765X

Yi Aff. Ind. 3a

Turkey Turkey

1 1

ARCL type 2 ARCL type 2

+ +

n.d. +

++ ++

c.187C4T c.187C4T

R63X R63X

Aff. Ind. 2a Aff. Ind. 1a

Germany Turkey

1 1


+ +

+ +

++ ++

c.353_354delTG c.2176-2_3delCA

K117fsX144 n.d.

France

1

ARCL type 2

+ +

–b

++

c.732-2A4G c.1327-2A4C

D243fsX258 E442fsX506

Portugal Italy

1 1


+ +

+b –b

+++ +++

c.839delC c.1324G4T

T280fsX285 E442X

Su Pat. 4

Me CoFe Sa

Data on glycosylation anomalies (CDG) are included. Columns N and O describe N- and O-glycosylation defects, respectively. Severity of CNS involvement (mental retardation, dementia, seizures, abnormal MRI) is given in column ‘CNS’. Aff. Ind., affected individual. aThese

cases have been previously described7. bThese cases were investigated by mass spectrometry, whereas all others were assessed by isofocusing.

NATURE GENETICS VOLUME 40

[

NUMBER 1

[

JANUARY 2008

33


Percentage of cells with Golgi remnants


45

Note: Supplementary information is available on the Nature Genetics website.

** *

40

**

35 30 25

**

**

20 15 10

ACKNOWLEDGMENTS We thank W. Morelle for the MALDI-TOF glycan analysis, J. Jaeken and H. Carchon for the capillary zone electrophoresis, and C. Schlack for her brilliant technical assistance. This work was supported by Bundesministerium fu¨r Bildung und Forschung (BMBF) grant 01GM0623 and a grant from the Deutsche Forschungsgemeinschaft to UK, by the VIB (to E.R. and W.A.), by the Fonds Wetenschappelijk Onderzoek-Vlaanderen (grants G.0504.06 and G.0173.04 to W.A. and G.M., respectively) and by the Sixth Framework program of the European Union (Euroglycanet: LSHM-2005-512131). E.R. holds an Instituut voor de aanmoediging van innovatie door wetenschap en technologie in Vlaanderen (IWT) doctoral fellowship.

5 0 C

Ri

Su

Me

CoFe

Sa

Figure 2 Golgi trafficking defect in fibroblasts from affected individuals after brefeldin A treatment. The percentage of cells retaining Golgi remnants is markedly increased in cells from affected individuals. Results are given as means of three independent experiments. C, results obtained from fibroblasts from four healthy controls. Errors are given as standard error of the mean (s.e.m.). *P o 0.05; **P o 0.01.

a2 subunit8. The exact reason for the prominent developmental abnormalities of connective tissue is unclear. One likely explanation is a defective processing of proteins involved in elastic fiber formation (for example, fibulins). Furthermore, defects in the synthesis of dermatan sulfate and chondroitin sulfate might underlie the observed facial dysmorphism. In conclusion, we identified a new mechanism leading to a congenital glycosylation defect, demonstrated a pivotal role of the V-type H+ ATPase a2 subunit for Golgi function and showed that wrinkly skin syndrome and autosomal recessive cutis laxa type II are variable manifestations of the same genetic defect. Further analysis is necessary to dissect which aspects of the phenotypes observed here are caused by abnormal glycosylation or by impairment of other cellular pathways.

AUTHOR CONTRIBUTIONS U.K. coordinated the genetic analysis and wrote the paper, E.R. performed Golgi trafficking assays, A.D. and B.F. performed fine mapping and sequencing, B.B., P.N., J.v.R. and H.v.B. performed genetic mapping, A.R., S.G., Z.U., H.v.B., E.M. and L.V.M. provided patient material and clinical information, F.F., G.M. and W.A. coordinated cell biological and glycosylation analyses, D.L. and R.W. performed glycosylation analysis, L.V.M. and S.M. initiated and coordinated the study. Published online at http://www.nature.com/naturegenetics Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions 1. Freeze, H.H. Nat. Rev. Genet. 7, 537–551 (2006). 2. Wu, X. et al. Nat. Med. 10, 518–523 (2004). 3. Foulquier, F. et al. Hum. Mol. Genet. 16, 717–730 (2007). 4. Foulquier, F. et al. Proc. Natl. Acad. Sci. USA 103, 3764–3769 (2006). 5. Rajab, A. et al. Am. J. Med. Genet. (in the press). 6. Morava, E. et al. Eur. J. Hum. Genet. (in the press). 7. Morava, E. et al. Eur. J. Hum. Genet. 13, 414–421 (2005). 8. Gilchrist, A. et al. Cell 127, 1265–1281 (2006). 9. Pietrement, C. et al. Biol. Reprod. 74, 185–194 (2006). 10. Hurtado-Lorenzo, A. et al. Nat. Cell Biol. 8, 124–136 (2006). 11. Nishi, T. & Forgac, M. Nat. Rev. Mol. Cell Biol. 3, 94–103 (2002). 12. Frattini, A. et al. Nat. Genet. 25, 343–346 (2000). 13. Peters, C. et al. Nature 409, 581–588 (2001). 14. Miesenbock, G., De Angelis, D.A. & Rothman, J.E. Nature 394, 192–195 (1998). 15. Axelsson, M.A. et al. Glycobiology 11, 633–644 (2001).

The ARCL Debre´-type Study Group includes the following members: William B Dobyns1, Dulce Quelhas2, Laura Vilarinho2, Elisa Leao-Teles3, Marie Greally4, Eva Seemanova5, Martina Simandlova5, Mustafa Salih6, Arti Nanda7, Lina Basel-Vanagaite8, Hulya Kayserili9, Memmune Yuksel-Apak9, Marc Larregue10,14, Jacqueline Vigneron11, Sanda Giurgea12, Uwe Kornak13 & Stefan Mundlos13 1Department of Human Genetics, University of Chicago, Chicago, Illinois 60637, USA. 2Institute of Medical Genetics Jacinto de Magalha ˜ es, 4099-028 Porto, Portugal. 3Metabolic Unit, Department of Pediatrics, San Joao Hospital, 4200-465 Porto, Portugal. 4College of Medicine and Medical Sciences, Arabian Gulf 5 University, Manama, POB 26671, Bahrain. Department of Clinical Genetics, Motol Hospital, Charles University, 150 06, Prague 05, Czech Republic. 6Division of Pediatric Neurology, Department of Pediatrics, College of Medicine, King Saud University, 11451 Riyadh, Saudi Arabia. 7As’ad Al-Hamad Dermatology Center, Al-Sabah Hospital, 22016 Salmiya, Kuwait. 8Department of Medical Genetics, Schneider Children’s Medical Center, 49202, Petah Tikva, Israel. 9Department of Medical Genetics, Faculty of Medicine, University of Istanbul, 34390 Istanbul, Turkey. 10Consultation de dermatologie pe´diatrique, CHU Armand-Trousseau, 75012 Paris, France. 11Unite´ de Geńe´tique Me´dicale, Service de me´decine neónatale, CHU Nancy, 54042 Nancy, France. 12Department of Neurology, CHU Tivoli, 7100 La Louviere, Belgium. 13Institute of Medical Genetics, Charite´ Universitaetsmedizin and Max Planck Institute for Molecular Genetics, Berlin, Germany. 14Deceased.

34

VOLUME 40

[

NUMBER 1

[

JANUARY 2008 NATURE GENETICS