Germline loss-of-function mutations in SPRED1 cause a ... - Lirias

© 2007 Nature Publishing Group http://www.nature.com/naturegenetics

LETTERS

Germline loss-of-function mutations in SPRED1 cause a neurofibromatosis 1–like phenotype Hilde Brems1,9, Magdalena Chmara1,2,9, Mourad Sahbatou3,9, Ellen Denayer1, Koji Taniguchi4, Reiko Kato4, Riet Somers1,5, Ludwine Messiaen6, Sofie De Schepper7, Jean-Pierre Fryns1, Jan Cools1,5, Peter Marynen1,5, Gilles Thomas3,8, Akihiko Yoshimura4 & Eric Legius1 We report germline loss-of-function mutations in SPRED1 in a newly identified autosomal dominant human disorder. SPRED1 is a member of the SPROUTY/SPRED family1 of proteins that act as negative regulators of RAS-RAF interaction and mitogenactivated protein kinase (MAPK) signaling2. The clinical features of the reported disorder resemble those of neurofibromatosis type 1 and consist of multiple cafe´-au-lait spots, axillary freckling and macrocephaly. Melanocytes from a cafe´-au-lait spot showed, in addition to the germline SPRED1 mutation, an acquired somatic mutation in the wild-type SPRED1 allele, indicating that complete SPRED1 inactivation is needed to generate a cafe´-au-lait spot in this syndrome. This disorder is yet another member of the recently characterized group of phenotypically overlapping syndromes caused by mutations in the genes encoding key components of the RAS-MAPK pathway3,4. To our knowledge, this is the first report of mutations in the SPRY (SPROUTY)/SPRED family of genes in human disease. Neurofibromatosis type 1 (NF1), or von Recklinghausen disease, is an autosomal dominant condition characterized by multiple cafe´-au-lait spots, axillary freckling, Lisch nodules in the iris and tumors of the nervous system. Other frequently observed features are short stature, macrocephaly and learning and behavioral problems. NF1 is caused by inactivating mutations in the NF1 tumor suppressor gene5 encoding neurofibromin, a positive regulator of RAS inactivation6. NF1 was the first human disorder shown to originate from germline mutations in a gene encoding a component of the RAS-MAPK pathway. Subsequently, mutations in genes encoding other components of this pathway were identified in disorders showing some phenotypic overlap with NF1, such as PTPN11 (ref. 7), KRAS8 and SOS1 (refs. 9,10) in Noonan syndrome, PTPN11 (refs. 11,12) in LEOPARD syndrome, HRAS13 in Costello syndrome, and KRAS14, BRAF14,15, MEK1 (ref. 15)

and MEK2 (ref. 15) in cardio-facio-cutaneous syndrome. These disorders, now known as the ‘neuro-cardio-facial-cutaneous’ syndromes3,4, present with a variable degree of cognitive impairment, facial dysmorphism, congenital heart defects and skin abnormalities. In a multidisciplinary neurofibromatosis clinic, we identified five families with an autosomal dominant trait (Fig. 1a) consisting of multiple cafe´-au-lait spots (Fig. 1b), axillary freckling, macrocephaly and a Noonan-like dysmorphy in some individuals (5/37) (Table 1 and Fig. 1c). We were surprised not to find any NF1 gene abnormalities in the affected individuals after comprehensive mutation analysis16, and we excluded linkage to the NF1 region in the two largest families (data not shown). We then performed a genome-wide linkage scan on two pedigrees (families 1 and 2; Fig. 1a) to identify the location of the disease-causing gene. Genotypic data analysis uncovered a maximum multipoint parametric lod score on chromosome 15 of 4.8. We analyzed meiotic recombinations and narrowed the disease locus between markers D15S1040 and D15S659 (Supplementary Fig. 1 online), covering a 12.25-Mb region on chromosome 15 (position 31913669–44161483; Ensembl release 43). We considered SPRED1 a good candidate gene, as the protein it encodes negatively regulates MAPK signaling2 (as does neurofibromin), and the clinical phenotype in the five families resembles the NF1 phenotype. We performed mutational analysis of SPRED1 in the affected individuals from the first two families. In both families, all affected individuals showed a nonsense mutation (family 1: R117X; family 2: R24X) that was absent in unaffected relatives. We then extended mutation analysis to three additional families. In family 3, we found a splice mutation in the canonical GT splice donor site of intron 5 (423+1G4A), causing skipping of exon 5 and resulting in an out-of-frame deletion of 47 nucleotides at the mRNA level (Supplementary Fig. 2 online). Family 4 showed the same nonsense mutation (leading to amino acid change R117X) as in family 1, and in family 5, we identified a nonsense mutation (leading to Q215X) in exon 7

1Department of Human Genetics, Catholic University Leuven, 3000 Leuven, Belgium. 2Department of Biology and Genetics, Medical University of Gdansk, Gdansk, Poland. 3Fondation Jean Dausset–CEPH, 75010 Paris, France. 4Division of Molecular and Cellular Immunology, Medical Institute of Bioregulation, Kyushu University, Fukuoka 812-8582, Japan. 5Human Genome Laboratory, Department of Molecular and Developmental Genetics, Flanders Institute for Biotechnology (VIB), Leuven, Belgium. 6Medical Genomics Laboratory, Department of Genetics, University of Alabama at Birmingham, Birmingham, Alabama, USA. 7Department of Dermatology, Ghent University Hospital, Ghent, Belgium. 8Department of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, Maryland 20892, USA. 9These authors contributed equally to this work. Correspondence should be addressed to E.L. ([email protected]).

Received 23 April; accepted 5 July; published online 19 August 2007; doi:10.1038/ng2113

NATURE GENETICS ADVANCE ONLINE PUBLICATION

1

LETTERS the duration of pathway activation was prolonged in the Spred1/ MEFs (SupplemenI:1 I:2 I:1 I:2 tary Fig. 3 online). The nature of the identified SPRED1 II:9 II:10 II:1 II:2 II:3 II:4 II:5 II:6 II:7 II:8 II:3 II:4 II:6 II:8 II:1 II:2 II:7 II:8 II:8 II:10 mutations suggested that they were loss-offunction mutations. We compared their abilIII:1 III:2 III:3 III:4 III:5 III:6 III:7 III:8 III:9 ity to inhibit MEK and ERK activation with III:8 III:7 III:3 III:8 III:10 III:11 III:12 III:13 III:14 III:1 III:2 III:3 III:4 III:6 * that of wild-type SPRED1. We cotransfected * b sequence variants (Fig. 2) with an Erk2 IV:3 IV:4 IV:6 IV:8 IV:7 IV:1 IV:2 IV:3 IV:8 IV:10 expression vector in HEK293T cells (hereafter ‘293T cells’). Some constructs produced an UZL3 UZL4 UZL5 insufficient amount of protein to assess their I:1 I:2 I:1 I:2 I:1 I:2 c effect and were not further analyzed. Results from the constructs producing a sufficient amount of protein showed that the mutant II:1 II:2 II:3 II:1 II:2 II:1 II:2 II:3 constructs encoding the I81_V85del, M266fsX4 and R325X SPRED1 proteins III:1 III:2 III:3 III:4 were unable to reduce FGF- and EGFinduced ERK activation, contrary to the Figure 1 Pedigrees and clinical photographs. (a) Pedigrees of families 1, 2, 3, 4 and 5. Squares and constructs encoding the wild-type protein circles indicate males and females, respectively. Open symbols indicate unaffected individuals, filled and the S149N missense change (Supplesymbols indicate affected individuals and symbols with a slash indicate deceased family members. mentary Fig. 4 online). (b) Clinical phenotype with multiple cafe´-au-lait spots (individual III4* from family 2). (c) Face of individual III9 from family 2. Note the mild hypertelorism, epicanthic folds, broad nasal tip, full lips As expression of some mutant proteins and cafe´-au-lait spot on the left upper arm. This child does not show the Noonan-like facial was not as high as that of wild-type characteristics seen in some individuals with SPRED1 mutations. Written parental consent was SPRED1, we performed more quantitative obtained to publish the photograph in c. assays to assess ERK activation. Growth factor–induced activation of ERK can be (Fig. 2 and Supplementary Table 1 online). This mutation occurred monitored by measuring the rate of Elk1-dependent transcription18. de novo on the paternal allele of individual II2 (data not shown). In all We transfected 293T cells with various amounts of wild-type and individuals, we detected only a single SPRED1 mutation, in agreement mutant SPRED1 constructs and cotransfected the cells with Elk1 with a dominant inheritance pattern. reporter plasmids. Wild-type SPRED1 and S149N dose-dependently We screened an additional set of 86 anonymized unrelated (mostly suppressed FGF-induced Elk1 activation (Fig. 3b). The mutants young) individuals who underwent clinical NF1 testing because of the I81_V85del, M266fsX4 and R325X had no effect on luciferase activity presence of multiple cafe´-au-lait spots but who did not have detectable (Fig. 3b). We obtained similar findings for EGF-induced Elk activaNF1 mutations. We identified seven different mutations in SPRED1: tion (data not shown). The I81_V85del variant protein was unstable three nonsense mutations (leading to R64X, K322X and R325X), two and was degraded by the ubiquitin-proteasome pathway (Supplemenframeshift mutations (leading to M266fsX4 and S383fsX21), one tary Fig. 5 online). However, even when expressed at levels as high as missense mutation (leading to S149N) and one in-frame deletion those of wild-type SPRED1 (Fig. 3b; 0.6 mg per lane), this mutant did (leading to I81_V85del) (Table 1, Fig. 2 and Supplementary Table 1). not suppress FGF-induced Elk activation. It seems that both the Ena/ To study the potential role of SPRED1 in the pathogenesis of cafe´- vasodilator-stimulated phosphoprotein (VASP) homology (EVH-1) au-lait spots, we established melanocyte cultures17 from normal skin domain and the SPROUTY-like cysteine-rich (SPR) domain are and from a cafe´-au-lait spot of individual II6 from family 2. The necessary for the suppression of Elk1 activation, as both I81_V85del germline mutation leading to R24X was present in melanocytes from (with normal SPR and mutant EVH-1) as well as M266fsX4 and normal skin and the cafe´-au-lait spot, but in melanocytes from the R325X (with normal EVH-1 and mutant SPR) are defective. We have shown that mouse Spred1 inhibits ERK activation by cafe´-au-lait spot, we identified an additional somatic SPRED1 frameshift mutation (304_305insA, leading to T102fsX6). The two SPRED1 suppressing Raf kinase activation2. To confirm that the observed mutations were located on different alleles, suggesting that SPRED1 SPRED1 variants were unable to suppress activation of ERK, we function was completely absent in these cells. To further investigate the investigated the effect of the different SPRED1 constructs on Raf functional effects of the SPRED1/ status on the RAS-MAPK path- kinase activation in response to growth factors (Fig. 3c). We transiway, we compared three different melanocyte cultures (SPRED1+/+, ently transfected 293T cells with SPRED1 constructs together with a SPRED1+/ and SPRED1/). We observed higher MAP-kinase Flag-Raf1 construct and stimulated the cells with FGF and EGF. We kinase 1 and 2 (MEK1/2) phosphorylation and higher extracellular monitored Raf1 kinase activity by phosphorylation of Ser338 and by signal-regulated kinase 1 and 2 (ERK1/2) phosphorylation in an in vitro kinase assay using a recombinant inactive form of MEK as a SPRED1/ melanocytes compared with SPRED1+/ melanocytes, substrate2. In comparison with cells transfected with empty vector, with the lowest levels of MEK1/2 and ERK1/2 phosphorylation in Raf1 protein from cells overexpressing wild-type SPRED1 showed SPRED1+/+ melanocytes (Fig. 3a). lower phosphorylation of Ser338 Raf1 and MEK1 (Fig. 3c), confirmComplementary to the melanocyte experiment, we studied the effect ing that the wild-type SPRED1 protein inhibits signaling through of Spred1 on the Ras-MAPK pathway in Spred1+/+ and Spred1/ suppression of Raf1 kinase activation. We observed a similar level of mouse embryonic fibroblasts (MEFs) stimulated with basic fibroblast inhibition in cells expressing the S149N construct, indicating that the growth factor (FGF). FGF stimulation resulted in Raf1, MEK1/2 and encoded variant protein functions normally in these assays. For ERK1/2 activation in Spred1+/+ as well in Spred1/ MEFs; however, constructs encoding the I81_V85del, M266fsX4 and R325X SPRED1


a

2

UZL1

UZL2

ADVANCE ONLINE PUBLICATION NATURE GENETICS

LETTERS Table 1 Clinical features of individuals with SPRED1 mutations Subject

Age (years)

CAL spots

Freckling

Head circumference (percentile)

Length (percentile)

Other

II1 II4

66 62

46 2

Mild –

497 97

75 10

Three lipomas Multiple lipomas

II6 II8

56 60

6 1

– –

97 50–75

10–25 25

Four lipomas Multiple lipomas

II9 III1

61 38

46 5–6

+ –

497 50

50–75 U

Loose integument, two lipomas One depigmentation spot

III3 III5

39 37

7–8 46

– –

497 U

U U

III7 III9

33 32

5 46

– –

25–50 50–75

75 50–75

Noonan-like face, pseudotumor orbitae

III10 III11

28 30

46 2

– –

U 75

U 50–75

One lipoma Two lipomas

III13 IV1

33 11

46 3

Mild –

75–97 497

50–75 U

Three depigmentation spots

IV2

9

7

–

75–97

U

Two depigmentation spots

IV3 IV5

11 9

46 12

– –

U 75

U 90

Noonan-like face, mild pectus excavatum

IV7 IV8

5 17

14 12

– –

90 75–97

90 50–75

IV9 IV10

14 9

7 2–3

+ –

75–97 97

3 3–25

I1 II2

60 Died at 44

– 15

– Mild

97 50–75

75 75–97

II6 II9

41 36

46 15

Mild +

3 97

75 3–25

One lipoma Multiple lipomas

III4 III8

5 1

46 410

– –

25–50 3–25

3–25 3–25

ADHD, eczema ADHD, learning problems

III9

3

46

U

U

U

UZL3 I2

41

6–7

+

50–75

3–25

II1 II2

16 15

8 6–7

– +

497 497

25–50 50–75

II3

13

7

Mild

75–97

3

Mild pectus excavatum, previously diagnosed as Noonan syndrome, learning problems

I1

39

7

1

75–97

50–75

Noonan-like face, loose integument, two lipomas, appendicular skin tumor

II1

11

Some

–

97

3

UZL5 II2

46

3

–

U

3–25


UZL1

Multiple lipomas

Tuberous hemangioma, naevus flammeus Hemihyperplasia, one lipoma, accessory nipple Noonan-like face

UZL2 Cataract, multiple lipomas Non–small cell lung cancer, learning problems

Slow psychomotor development

Learning problems Mild supravalvular pulmonic stenosis, learning problems Mild pectus excavatum, learning problems

UZL4

VUJ-stenosis, nephrolithiasis

Wilms tumor, tubular colonadenoma, paroxysmal atrial tachycardia, two lipomas

III1

17

Some

–

U

3–25

III4

5

7–8

–

97

75

UAB15 UAB31

4 5

4 +

– +

U U

U U

Familial history of CAL spots U

UAB43 UAB48

5 3

46 46

– –

U U

U U

U Familial history of CAL spots

UAB74 UAB84

10 4

46 +

– Mild

U U

U U

No familial history of CAL spots U

UAB88

37

46

Mild

U

U

Familial history of CAL spots

UAB

UZL1–UZL5 and UAB represent families. CAL, cafe´-au-lait. Percentiles are according to growth charts of Flanders 2004 (http://www.vub.ac.be/groeicurven). –, negative; +, positive; U, unknown; ADHD, attention deficit hyperactivity disorder; VUJ, vesico-ureteric junction.


3

LETTERS

423+1G>A UZL3

7

8

TGA EVH-1

KBD

SPR

R24X R64X l81_V85del R117X 423+1G>A S149N Q215X M266fsX4 K322X R325X S383fsX21

proteins, we observed higher ratios of phosphorylated MEK1 to total MEK1 and phosphorylated Raf1 to total Raf1 compared with the wildtype construct (Fig. 3c), indicating that these mutant SPRED1 proteins have lost their ability to inhibit Raf-MEK-ERK signaling.

MEK1/2 pERK1/2 ERK2

b

FGF

c

140 120 100

WT l18_V85del M266fsX4 R325X S149N

80 60 40

– + – + – + – + – + – +

IP-Flag

0.03 0.06

0.1

0.3

0.6 STAT5 pRaf-1/Raf-1

0.2

0.8 0.6 0.4 0

l8

y pt

W T 1_ V8 5d el M 26 6f sX 4

– + – + – + – + – + – + FGF or EGF

Em

N

5X

49 S1

32

R

26

6f s

X4

T

el 5d

V8 1_ l8

–/–

1.0

– + – + – + – + – + – +

Em

+/–

*

1.2

0.2

0 FGF or EGF

+/+

p-Raf1/Raf1 ratio

0.4

M

STAT5

R325X M266fsX4

0.6

y

Flag-SPRED-1

1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0

*

1.4

0.8

W

–/–

(µg) WT, S149N l81_V85del

**

1.0

pt

V8 5 26 de l 6 R fsX 32 4 5X

1_

M

S149N

p-MEK1/MEK1 ratio

+/–

**

1.2

0. 03 0. 0 0. 6 1 0. 3 0. 6 0. 1 0. 6 0. 1 0. 6 0. 1 0. 6 0. 03 0. 06 0. 1 0. 3 0. 6

WT

l8

pt

y

pMEK1/MEK1

Em

pMEK/MEK ratio

pS338

Flag-Raf-1

0

+/+

Raf-1 kinase activity

TCL Flag-Spred-1

Plasmid (µg) 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0

– + – + – + – + – + – +

pMEK-1 MEK-1 pFlag-Raf-1 Flag-Raf-1

20 0

Actin

EGF

N

pMEK1/2

–/–

49

+/–

5X

+/+

Relative luciferase activity (%)

a

pERK/ERK ratio


ATG

In the biochemical assays, the construct containing the S149N change functioned similarly to the wild-type SPRED1 construct. Moreover, Ser149 is not conserved in mouse and rat. Thus, we assume that S149N is encoded by a rare sequence variant without a functional effect on Raf suppression. The question arises whether the observed SPRED1 mutations result in haploinsufficiency or whether they have a dominant-negative effect. To address this, we transfected 293T cells with an enhanced yellow fluorescent protein (EYFP)-Erk2 vector together with a SPRED1 wildtype vector in the presence of the expression vectors encoding mutants I81_V85del, M266fsX4, R325X and S149N. As expected, overexpression of S149N further suppressed ERK activation. Transfection of equimolar quantities of wild-type and the other mutant constructs was not associated with any dominant negative effect (Supplementary Fig. 6 online). We observed only a partial dominant-negative effect with mutants M266fsX4 and R325X, and only when they were transfected in a 5:1 ratio with the wild-type construct (Supplementary Fig. 6). Several mutants, including I81_V85del, were degraded very

32

6

S1

5

R

4

pt W y T l8 1_ V M 85 26 de l 6 R fsX 32 4 5X S1 49 N

3

Em

1 2

S383fsX21 UAB84

pt W y T l8 1_ V M 85 26 de l 6 R fsX 32 4 S1 5X 49 N

R325X UAB88

Q215X UZL5

R64X UAB74

Em

K322X UAB48

S149N UAB43

l81_V85del UAB31

R24X UZL2

Figure 2 Exon-intron structure of human SPRED1 showing noncoding sequences as open boxes and protein-coding exons as filled black boxes. Exons are numbered from 1 to 8. The filled gray boxes represent the structure of the SPRED protein, showing the Ena/vasodilator-stimulated phosphoprotein (VASP) homology (EVH-1) domain, KIT-binding (KBD) domain and the SPROUTY-like (SPR) domain. The different SPRED1 mutations in functional and nonfunctional domains are indicated.

M266fsX4 UAB15

R117X UZL1, UZL4

Figure 3 Effect of SPRED1 on the RAS-MAPK pathway. (a) Comparison of SPRED1+/+ (+/+), SPRED1+/ (+/–) and SPRED1/ (–/–) melanocytes stimulated with stem cell factor. Top: cell extracts immunoblotted with the indicated antibodies. Bottom: pMEK/MEK and pERK/ERK ratios normalized to 1 for SPRED1+/+ melanocytes. Error bars represent s.d. (n ¼ 2). Average pMEK/MEK and pERK/ERK ratios of the three groups (SPRED1+/+, SPRED1+/ and SPRED1/) were significantly different (analysis of variance (ANOVA), P ¼ 5.5 10–4 for pMEK/MEK, P ¼ 1.4 10–3 for pERK/ERK). The Tukey post hoc test showed that, for both ratios, the average of each group was significantly different from the other two groups. (b) Quantitative luciferase assay for ERK activation. Top: different concentrations of empty vector or vectors carrying wild-type or mutant SPRED1 were transfected into 293T cells along with Elk1 reporter plasmids. Cells were treated with 50 ng/ml basic FGF or were left untreated for 6 h, and luciferase activity was then measured and normalized to b-galactosidase. Relative luciferase activity was normalized to 100% for the wild-type construct. Error bars represent s.d. SPRED1 expression was determined by immunoblotting (bottom). (c) Suppression of Raf1 kinase activity by SPRED1. Top: 293T cells were transfected with plasmids encoding Flag-Raf1 together with wild-type, I81_V85del, M266fsX4, R325X or S149N SPRED1 constructs. After 24 h, cells were treated with 100 ng/ml basic FGF or EGF for 2 min; then, total cell lysates (TCLs) were immunoprecipitated with anti-Flag, and immunoprecipitates were subjected to the in vitro Raf1 kinase assay, using an inactive form of MEK1 as substrate (see Methods). Reactions were immunoblotted with anti-pMEK1 and anti-pRaf1. Bottom: pMEK1/MEK1 and pRaf1/Raf1 ratios were normalized to 1 for the empty vector (stimulated). Student’s t test was used for statistical analysis. * P o 0.05, ** P o 0.01. Error bars represent s.d. (n ¼ 3).

4



LETTERS rapidly, and thus expression was very low (Supplementary Fig. 5 and data not shown). Therefore, we believe it is unlikely that these mutants exert a dominant-negative effect under physiological conditions, and thus haploinsufficiency would be the most likely mechanism. Spred proteins are membrane-associated suppressors of growth factor–induced Ras-MAPK activation, and they seem to function downstream of Ras2, although a target upstream of Ras has also been suggested19. Spred1 and Spred2 have three known functional domains: an N-terminal EVH-1 domain, a central c-KIT–binding domain (KBD) and a C-terminal SPR domain (reviewed in ref. 1). Spred3 lacks a functional KBD. The SPR domain is responsible for the translocation of Spred proteins to the plasma membrane and the interaction with Raf. It also mediates heterodimer formation of Spred proteins19. EVH1 domains recognize and bind proline-rich sequences, but the specific ligands of the Spred proteins have not yet been identified. The physiological function of the KBD is still unknown. Spred1 has been shown to regulate not only the Ras-ERK pathway but also Rho and downstream actin organization and cell motility20. In mice, Spred1 is expressed predominantly in adult brain but also in adult lung and colon and in some fetal tissues such as brain, liver and heart21. Spred1/ mice are fertile and have a lower body weight and a shortened face. Spred1 negatively regulates interleukin-5–mediated eosinophil proliferation, and Spred1/ mice show allergen-induced airway hyperresponsiveness caused by eosinophilia22. Facial abnormalities (Supplementary Fig. 7a online) as well as melanin deposits in the spleen (Supplementary Fig. 7b online) of Spred1/ mice may relate to the abnormalities seen in affected individuals with SPRED1 mutations. The individuals discussed here with mutations in SPRED1 presented with an autosomal dominant trait consisting of multiple cafe´-au-lait spots, axillary freckling and macrocephaly. Noonan-like dysmorphic features were present in a minority of individuals. Here, we have presented detailed clinical and familial data for the five families ascertained in a multidisciplinary neurofibromatosis clinic. Several affected individuals from these five families fulfilled the NIH diagnostic criteria for NF1 (ref. 23), although some typical features of NF1 were systematically absent, such as Lisch nodules in the iris, neurofibromas and central nervous system tumors. Similar to individuals with NF1, some children were diagnosed with attention deficit and/or learning difficulties. We also found that the melanocytes from a cafe´-au-lait spot harbored a mutation in the wild-type SPRED1 allele and that SPRED1 function is completely inactivated in these cells. These findings are in agreement with recent findings in melanocytes from cafe´-au-lait spots in individuals with NF1, where a second mutation was demonstrated in NF1 in melanocytes derived from cafe´-au-lait spots24. This means that the benign congenital proliferations of melanocytes causing cafe´-au-lait spots in NF1 (ref. 17) as well as in this new syndrome may be due to an early somatic mutation in a melanocyte precursor cell (melanoblast). Subcutaneous lipomas were present in several affected adults, but a second mutation in SPRED1 was not detected in a lipoma (individual II9 family 2). As lipomas are commonly found in the general population, it is not clear if they are part of the phenotype. We also do not know if other observed tumors are causally related to the SPRED1 mutation; these include the lung cancer in individual II2 (42 years old, non-smoker) from family 2 and the childhood malignant renal cancer (possibly Wilms tumor) and colonadenoma in individual II2 (45 years old) from family 5. In conclusion, we report the identification of germline mutations in a member of the SPRED/SPRY gene family in individuals with an


NF1-like phenotype. Heterozygous SPRED1 mutations are associated with dominant inherited multiple cafe´-au-lait spots, axillary freckling, macrocrania and a Noonan-like dysmorphy in some of the individuals. However, it remains important to further document the spectrum and frequency of symptoms and complications associated with this new disorder in a larger set of affected individuals. METHODS Subjects. All blood samples for DNA extraction were collected after informed consent. Affected and unaffected individuals from the five families were identified in the Leuven neurofibromatosis clinic. This study was approved by the local institutional review board (IRB). We performed Epstein-Barr virus transformation on B lymphocytes from peripheral blood of individual II2 from family 3 to study the familial splice mutation (423+1G4A) at the mRNA level. The set of 86 anonymous samples of individuals having multiple cafe´-au-lait spots but lacking NF1 mutations was provided by the Medical Genomics Laboratory at the University of Alabama at Birmingham. Only limited clinical data were available on these 86 samples, and IRB approval was obtained for further analysis of these samples to determine the genetic cause of the phenotype. Linkage analysis. A genome-wide linkage analysis was performed using a set of 382 microsatellite markers (ABI Linkage Mapping Set 1 (LMS1)) in two families on a total of 14 unaffected individuals and 26 affected individuals. Analysis of the genotypic data was performed with GeneHunter2. As the initial cosegregation analysis clearly strengthened the hypothesis of a fully penetrant dominant mode of inheritance, we performed an analysis under the following restrictive assumptions: penetrance vector ¼ 0, 1, 1; population frequency of the disease allele ¼ 104. The marker allele frequencies in the population were computed from a group of unrelated individuals composed of the unaffected founder members of the two pedigrees and the grandparents of ten CEPH reference families. We used the ABI LMS2 panel and seven additional markers on chromosome 15 to increase the lod score. Mutation detection. Mutational screening of NF1 was carried out on cDNA as previously described16. Genomic DNA extracted from peripheral blood leukocytes was used for mutation analysis of SPRED1. Each exon was amplified by PCR, and purified PCR products were sequenced bidirectionally using the ABI BigDye Terminator Sequencing Kit (Applied Biosystems) and an ABI3100 Capillary Array Sequencer (Applied Biosystems). Primers for PCR amplification of SPRED1 exons 1–8 are listed in Supplementary Table 1. The parents of child UAB 43 (with the S149N missense change) were not available for further analysis. We did not find this sequence variant in 1,000 NF1 mutation-negative controls, excluding it as a frequent polymorphism. Study of the splice site mutation. RNA was extracted from an Epstein-Barr virus–transformed cell line derived from blood of individual II2 from family 3 using the RNeasy Mini Kit (Qiagen). cDNA was synthesized with 3–5 mg total RNA using oligo(dT)12-18 (Invitrogen) and SuperScript III Reverse Transcriptase (Invitrogen). Exons 4–7 of SPRED1 were amplified by PCR using SPRED1 cDNA as a template, with primers cDNA_F2 and cDNA_R2 (Supplementary Table 1). Agarose gel electrophoresis showed two different fragments, which we purified using the Gel Extraction Kit (Qiagen). Purified fragments were cloned in a pGEM-T Easy vector (Promega) and transformed in competent cells. Plasmid DNA from seven transformed clones was extracted and sequenced as described above. Skin biopsy and melanocyte culture. After we obtained informed consent, we obtained biopsies from normal skin of a control and normal skin and a cafe´-aulait spot of individual II6 from family 2. Primary epidermal melanocytes were cultured in Ham’s F10 medium (Invitrogen) supplemented with 2.5% FBS (Hyclone), 1% Ultroser G (BioSepra), 5 ng/ml basic FGF (Sigma), 10 ng/ml endothelin-1 (Sigma), 0.33 nM cholera toxin (Axelle), 33 mM isobutyl-methylxanthine (Sigma), 5.3 nM 12-tetradecanoyl phorbol 13-acetate (Sigma) and 20 ng/ml stem cell factor (Sigma). RNA isolation, cDNA synthesis, amplification and cloning. Before isolating RNA, we grew melanocytes to confluence and treated cells with puromycin

5

LETTERS


(200 mg/ml) (Sigma) for 3 h. Total RNA was extracted with the RNeasy Plus Mini Kit (Qiagen) according to the manufacturer’s instructions. cDNA was synthesized as described previously. Exons 2 to 4 of SPRED1 were amplified by PCR using cDNA as template, with primers cDNA_F and cDNA_R1 (Supplementary Table 1). The resulting PCR fragments were cloned into a pGEM-T Easy vector (Promega) and were used to transform competent cells. We sequenced 16 transformed clones with inserts. Expression constructs. Wild-type SPRED1 cDNA was cloned into a pcDNA3.1 (Invitrogen) construct encoding a Flag tag at the N terminus. SPRED1 mutants were generated by PCR-directed mutagenesis and were verified by sequencing. Wild-type and mutant SPRED1 constructs were subcloned in a pmax vector (Amaxa Biosystems). cDNA encoding mouse Erk2 was amplified by PCR and subcloned into the multiple cloning site of pEYFP-C1 (Clontech). We analyzed the stability of some mutant proteins used for the biochemical experiments. Transfected 293T cells were pretreated with MG132 (Calbiochem) or with dimethylsulfoxide and then incubated with cycloheximide (Nacalai Tesque) (Supplementary Fig. 5). We showed that I81_V85del was unstable and degraded by the proteasome. Pretreatment with a proteasome inhibitor resulted in stability of the protein. We hypothesize that the truncated proteins, which are unstable in vitro, are also unstable in vivo. Biochemical analysis of melanocytes. Melanocyte cultures were grown to confluence, starved overnight in serum-free medium and stimulated with stem cell factor (30 ng/ml) for 10 min. The cells were lysed and used for protein blot analysis. Antibodies used for immunoblotting included anti-phospho-MEK1/2 (Cell Signaling), anti-MEK1/2 (Santa Cruz), anti-phospho-ERK1/2 (Cell Signaling), anti-ERK2 (Santa Cruz), anti-b-actin (Sigma) and anti-Flag (Sigma). Quantitative analysis of protein blot images was performed using Scion software. Results are the average of two experiments and are presented as normalized ratios ± s.d. Biochemical analysis of MEFs. Spred1+/+ and Spred1/ MEFs were prepared as described previously25. MEFs were stimulated with basic FGF (100 ng/ml) (PeproTech) for 2, 5, 15, 30 and 60 min. Cell extracts were immunoblotted with anti-phospho-Raf1 (Cell Signaling), anti-Raf1 (BD Transduction Laboratories), anti-phospho-MEK1/2, anti-MEK1 (Santa Cruz), anti-phospho-ERK1/2, antiERK2 (Santa Cruz) and anti-Spred1 as described2,26. The anti-Spred1 was prepared by immunizing rabbits and was affinity purified before use2. Biochemical analysis of 293T cells. Wild-type or mutated SPRED1 constructs were transiently transfected and cotransfected with an enhanced yellow fluorescent protein (EYFP)-Erk2 expression construct into 293T cells, using the polyethylenimine method or Fugene HD (Roche). Twenty-four hours after transfection, the cells were stimulated with basic FGF (100 ng/ml) or EGF (100 ng/ml) (PeproTech) for 5 min. Cell extracts were immunoblotted with antiphospho-ERK1/2, anti-ERK2 and anti-Flag antibodies. Three independent transfection experiments were carried out. Luciferase assay. FGF- or EGF-induced ERK activation was measured by a luciferase assay using an Elk1-responsive reporter, as described previously25. In all reporter assays, 2 105 293T cells were plated on six-well dishes, and plasmids were transfected with the polyethylenimine method. We defined the percentage of relative luciferase activity as follows: 100 ((luciferase activity in the presence of FGF / b-galactosidase activity in the presence of wild-type SPRED1 or mutant vectors) – (luciferase activity in the absence of FGF / b-galactosidase activity in the presence of wild-type SPRED1 or mutant vectors)) / ((luciferase activity in the presence of FGF / bgalactosidase activity in the presence of the empty vector) – (luciferase activity in the absence of FGF / b-galactosidase activity in the presence of the empty vector)). Cell lysates were blotted with and hybridized with anti-Flag and anti-STAT5. In vitro Raf1 kinase assay. 293T cells were transiently transfected with 0.5 mg Flag-Raf1 and SPRED1 constructs, stimulated with 100 ng/ml basic FGF or EGF for 2 min and lysed. Cleared lysates were immunoprecipitated with antiFlag. Immunoprecipitates with protein-A Sepharose were washed three times with buffer A (50 mM Tris-HCl (pH 7.6), 150 mM NaCl, 1% Nonidet P-40, 1

6

mM sodium vanadate) and once with buffer B (20 mM MOPS (pH 7.2), 25 mM glycerol phosphate, 5 mM EGTA, 1 mM sodium vanadate and 1 mM dithiothreitol). The Raf1 kinase assay was carried out in buffer B containing 250 mM ATP, 37.5 mM MgCl2 and 0.5 mg inactive recombinant MEK1 (Upstate Biotechnology) for 30 min at 30 1C, according to the manufacturer’s instructions (Upstate). Reaction products were resolved, separated by SDS-PAGE, followed by immunoblotting with anti-phospho-MEK and anti-phospho-Raf1 (S338). Results were obtained from three independent experiments. Dominant-negative effect. EYFP-Erk2 was transfected into 293T cells with or without wild-type SPRED1 vector (0.3 mg) in the presence or absence of expression vector (0.3 mg) (I81_V85del, M266fsX4, R325X or S149N). After 24 h, cells were treated without or with 100 ng/ml basic FGF for 5 min. Then, cell extracts were separated on an SDS-polyacrylamide gel, followed by immunoblotting with anti-phospho-ERK1/2 and anti-ERK2. The experiment was repeated with five times as much mutant expression vector (1.5 mg) as wildtype construct. Note: Supplementary information is available on the Nature Genetics website. ACKNOWLEDGMENTS The authors thank T. de Ravel for critically reading the manuscript, M. De Mil for technical assistance in melanocyte cell culture and F. Okamoto for technical assistance in cell culture and plasmid preparation. H.B. is supported by the Institute for the Promotion of Innovation through Science and Technology in Flanders (IWT-Vlaanderen). M.C. was supported by the Marie Curie European Community fellowship (contract HPMT-CT2001-00273). E.D. is a predoctoral researcher (Aspirant of the Fonds voor Wetenschappelijk Onderzoek-Vlaanderen), J.C. is a postdoctoral researcher and E.L. is a part-time clinical researcher of the Fonds voor Wetenschappelijk Onderzoek-Vlaanderen (FWO). This work is also supported by research grants from the Fonds voor Wetenschappelijk Onderzoek Vlaanderen (G.0096.02, E.L.) and (G.0507.04, P.M.); by the Interuniversity Attraction Poles (IAP) granted by the Federal Office for Scientific, Technical and Cultural Affairs, Belgium (2002–2006; P5/25) (P.M. and E.L.); by a Concerted Action Grant from the Catholic University Leuven; by the Federation des Maladies Genetiques Orphelines (G.T.) and by grants-in-aid from the Ministry of Education, Science, Technology, Sports, and Culture of Japan (A.Y.). This work was made possible by the Centre of Excellence SymBioSys (Research Council, Catholic University Leuven EF/05/007) (E.L.). AUTHOR CONTRIBUTIONS The study was coordinated by H.B., G.T., P.M., A.Y. and E.L.; patient phenotyping was performed by J.-P.F. and E.L. and clinical data collected by E.D.; linkage analysis was done by M.S. and G.T.; NF1 mutation analysis was conducted by L.M. and SPRED1 mutation analysis by H.B., M.C. and E.D.; skin biopsies were carried out by S.D.S. and melanocytes cultured by H.B. and S.D.S.; biochemical analysis was performed by H.B., M.C., E.D., K.T., R.S. and J.C.; mouse characterization was performed by R.K.; the manuscript was written by H.B., M.C., E.D., L.M., S.D.S., A.Y., J.-P.F., J.C., P.M., G.T. and E.L. COMPETING INTERESTS STATEMENT The authors declare 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. Bundschu, K., Walter, U. & Schuh, K. The VASP-Spred-Sprouty domain puzzle. J. Biol. Chem. 281, 36477–36481 (2006). 2. Wakioka, T. et al. Spred is a Sprouty-related suppressor of Ras signalling. Nature 412, 647–651 (2001). 3. Shannon, K. & Bollag, G. Sending out an SOS. Nat. Genet. 39, 8–9 (2007). 4. Bentires-Alj, M., Kontaridis, M.I. & Neel, B.G. Stops along the RAS pathway in human genetic disease. Nat. Med. 12, 283–285 (2006). 5. Legius, E., Marchuk, D.A., Collins, F.S. & Glover, T.W. Somatic deletion of the neurofibromatosis type 1 gene in a neurofibrosarcoma supports a tumour suppressor gene hypothesis. Nat. Genet. 3, 122–126 (1993). 6. Cichowski, K. & Jacks, T. NF1 tumor suppressor gene function: narrowing the GAP. Cell 104, 593–604 (2001). 7. Tartaglia, M. et al. Mutations in PTPN11, encoding the protein tyrosine phosphatase SHP-2, cause Noonan syndrome. Nat. Genet. 29, 465–468 (2001). 8. Schubbert, S. et al. Germline KRAS mutations cause Noonan syndrome. Nat. Genet. 38, 331–336 (2006).



LETTERS 9. Roberts, A.E. et al. Germline gain-of-function mutations in SOS1 cause Noonan syndrome. Nat. Genet. 39, 70–74 (2007). 10. Tartaglia, M. et al. Gain-of-function SOS1 mutations cause a distinctive form of Noonan syndrome. Nat. Genet. 39, 75–79 (2007). 11. Digilio, M.C. et al. Grouping of multiple-lentigines/LEOPARD and Noonan syndromes on the PTPN11 gene. Am. J. Hum. Genet. 71, 389–394 (2002). 12. Legius, E. et al. PTPN11 mutations in LEOPARD syndrome. J. Med. Genet. 39, 571–574 (2002). 13. Aoki, Y. et al. Germline mutations in HRAS proto-oncogene cause Costello syndrome. Nat. Genet. 37, 1038–1040 (2005). 14. Niihori, T. et al. Germline KRAS and BRAF mutations in cardio-facio-cutaneous syndrome. Nat. Genet. 38, 294–296 (2006). 15. Rodriguez-Viciana, P. et al. Germline mutations in genes within the MAPK pathway cause cardio-facio-cutaneous syndrome. Science 311, 1287–1290 (2006). 16. Messiaen, L.M. et al. Exhaustive mutation analysis of the NF1 gene allows identification of 95% of mutations and reveals a high frequency of unusual splicing defects. Hum. Mutat. 15, 541–555 (2000). 17. De Schepper, S. et al. Cafe-au-lait spots in neurofibromatosis type 1 and in healthy control individuals: hyperpigmentation of a different kind? Arch. Dermatol. Res. 297, 439–449 (2006).


18. Gille, H. et al. ERK phosphorylation potentiates Elk-1-mediated ternary complex formation and transactivation. EMBO J. 14, 951–962 (1995). 19. King, J.A.J. et al. Distinct requirements for the Sprouty domain for functional activity of Spred proteins. Biochem. J. 388, 445–454 (2005). 20. Miyoshi, K. et al. The Sprouty-related protein, Spred, inhibits cell motility, metastasis, and Rho-mediated actin reorganization. Oncogene 23, 5567–5576 (2004). 21. Engelhardt, C.M. et al. Expression and subcellular localization of Spred proteins in mouse and human tissues. Histochem. Cell Biol. 122, 527–538 (2004). 22. Inoue, H. et al. Spred-1 negatively regulates allergen-induced airway eosinophilia and hyperresponsiveness. J. Exp. Med. 201, 73–82 (2005). 23. Stumpf, D. et al. Neurofibromatosis conference statement. National Institutes of Health Consensus Development Conference. Arch. Neurol. 45, 575–578 (1988). 24. Maertens, O. et al. Molecular dissection of isolated disease features in mosaic neurofibromatosis type 1. Am. J. Hum. Genet. 81, 243–251 (2007). 25. Taniguchi, K. et al. Spreds are essential for embryonic lymphangiogenesis by regulating vascular endothelial growth factor receptor 3 signaling. Mol. Cell. Biol. 27, 4541–4550 (2007). 26. Sasaki, A. et al. Identification of a dominant negative mutant of Sprouty that potentiates fibroblast growth factor- but not epidermal growth factor-induced ERK activation. J. Biol. Chem. 276, 36804–36808 (2001).

7