Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Human Molecular Genetics, 2003, Vol. 12, No. 18 DOI: 10.1093/hmg/ddg241 2269–2276 Allelic variation in normal human FBN1 expression in a family with Marfan syndrome: a potential modifier of phenotype? Sarah Hutchinson1, Andre Furger1, Dorothy Halliday1, Daniel P. Judge2, Andrew Jefferson3, Harry C. Dietz2, Helen Firth4 and Penny A. Handford1,* 1 Department of Biochemistry, University of Oxford, Oxford OX1 3QU, UK, 2Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA, 3Wellcome Trust Centre for Human Genetics, Roosevelt Drive, Headington, Oxford OX3 7BN, UK and 4Department of Medical Genetics, Addenbrooke’s Hospital, Cambridge CB2 2QQ, UK Received February 28, 2003; Revised and Accepted July 10, 2003 FBN1 mutations cause Marfan syndrome (MFS), an autosomal dominant disorder of connective tissue. One of the unexplained features of MFS is the pathogenic mechanism that leads to marked inter- and intra-familial clinical variability, despite complete disease penetrance. An FBN1 deletion patient [46,XXdel(15)(q15q22.1)] was identified whose fibrillin-1 protein and mRNA levels were significantly higher than expected for a single FBN1 allele. This suggested that allelic variation in normal FBN1 expression might occur in MFS families, and have potential clinical implications particularly for those with premature termination codon (PTC) mutations who usually display low levels of expression from the mutant allele due to nonsense-mediated decay (NMD). RNA analyses identified a variable reduction in total FBN1 transcript (78 2.2 to 27.3 2.3%) in three related individuals carrying PTC-causing mutation 932insT, compared with unaffected control individuals. Both pulse chase analysis of fibrillin-1 biosynthesis and RNase protection analyses demonstrated that these differences were due to variation in the expression of the normal FBN1 allele and not NMD of mutant RNA. We suggest that differences in normal FBN1 expression could contribute to the clinical variability seen in this family with MFS, and should be considered as a potential modifier of phenotype in other cases of MFS. INTRODUCTION Marfan syndrome (MFS; OMIM 154700) is one of the major inherited disorders of connective tissue and results in prominent skeletal, ocular and cardiovascular manifestations (1). Over 500 mutations have been identified throughout FBN1 (OMIM 134797) that give rise to MFS and a spectrum of related ‘fibrillinopathies’ (2). With the exception of a small number of recurrent mutations, the vast majority of mutations are unique to families. Premature termination codon (PTC) mutations resulting from nonsense mutations or insertions/ deletions causing a frameshift constitute 25% of FBN1 mutations so far detected (2). Inter- and intra-familial variability is a common feature of the disease and genotype– phenotype relationships have been difficult to deduce. The affected protein in MFS, fibrillin-1, is a major component of 10–12 nm connective tissue microfibrils that are widely distributed in both elastic and non-elastic tissues (3). Prior to, or during, secretion from the cell profibrillin-1 is proteolytically processed at its N- and C-termini and assembled in the extracellular matrix to form microfibrils (4– 9). The precise mechanism of normal assembly is poorly understood and, as a consequence, the pathogenic mechanism(s) that underlie the MFS phenotype remain unclear. Although direct evidence is lacking, it has been hypothesized that mutant fibrillin-1 monomers exert a dominant negative effect impairing the function of normal monomers (10). Haploinsufficiency has been thought unlikely to be a pathogenic mechanism in MFS, since PTC patients with low levels of mutant FBN1 due to nonsense-mediated decay (NMD) often exhibit milder phenotypes that fall outside the clinical criteria required for a MFS diagnosis (10). In addition, in an earlier study a kindred exhibiting isolated skeletal features was identified which harboured C-terminal *To whom correspondence should be addressed. Tel: þ44 1865 285347; Fax: þ44 1865 285327; Email: [email protected] Human Molecular Genetics, Vol. 12, No. 18 # Oxford University Press 2003; all rights reserved 2270 Human Molecular Genetics, 2003, Vol. 12, No. 18 Table 1. Clinical features of del(15) and MFS patients II.2, II.3, II.12 and III.1 Case Age Musculoskeletal Pectus carinatum Severe pectus excavatuma Disproportionate body ratiob Wrist and thumb signs Scoliosis of >20 Reduced extension at the elbowsc Pes planus Protrusio acetabulaed Moderate pectus excavatum Joint hypermobilitye Highly arched palate with crowding Typical facial appearance Cardiovascular Aortic root dissection Aortic root dilatationf Maximum value of the aortic rootg Mitral value prolapse Ocular Ectopia lentish Myopia Other Dural ectasiai Pneumothorax Striae Microcephalyj Profound learning difficultyj II.2 47 II.3 48 II.12 40 III.1 24 Del(15) 10 þ þ þ n/a þ þ þ þ þ n/a þ þ þ þ þ þ þ n/a þ þ þ þ n/a þ þ þ þ n/a þ þ þ þ 70 þ þ 43 þ 35 þ þ >50k þ þ þ n/a þ n/a þ þ þ þ n/a þ n/a þ þ The clinical features are taken from the Ghent criteria (29). a Pectus excavatum requiring surgery. b Upper segment /lower segment ratio <0.86 or arm span/height ratio >1.05. c <170 . d Demonstrated on X-ray. e Beighton index >4. f The maximum diameter of the aortic root at the sinuses of Valsalva exceeding published body surface area normograms (30). g In millimetres. h By slit lamp examination. i Presence determined using magnetic resonance imaging and a recommended classification (31). j Not a feature of MFS. k Aortic root replaced. n/a, Information not available. mutation R2726W that abolished processing of mutant profibrillin-1, resulting in its lack of incorporation into microfibrils (6). It was suggested that this kindred displayed the phenotype associated with the haploinsufficiency state and was essentially equivalent to having an FBN1 null allele. More recently, mouse models with reduced FBN1 expression have led to the suggestion that disease presentation in MFS is dependent on a critical threshold of functional microfibrils (11,12). A defective population or a reduced number of microfibrils produced by the pathogenic effect of an FBN1 mutation may therefore account for the change in connective tissue function. Since the level of normal fibrillin-1 protein has the potential to modulate the pathogenic effect of the mutant protein, it is possible that allelic variation in normal FBN1 expression could explain some of the clinical heterogeneity seen in MFS, particularly in PTC patients who usually have low levels of mutant transcript due to NMD. We have studied an isolated patient with an interstitial deletion of chromosome 15 including the FBN1 locus, together with three members of an MFS kindred carrying a PTC mutation 932insT, in order to gain insight into the variability of normal FBN1 expression. Using both RNA analyses and pulse chase assays, we assess normal FBN1 transcript and fibrillin-1 protein levels in these individuals and identify significant differences between them. The potential clinical implications of these data for MFS pathogenesis are discussed. RESULTS Identification of an FBN1 deletion patient [del(15)] An isolated patient with profound handicap, including a number of Marfanoid characteristics (Table 1), was investigated for a suspected chromosomal deletion. Initial cytogenetic analyses revealed a karyotype of 46,XXdel(15)(q15q22.1) (data not shown). Fluorescence in situ hybridization (FISH) of patient interphase cells showed that a single FBN1-specific cosmid Human Molecular Genetics, 2003, Vol. 12, No. 18 2271 probe signal (green) was detected in 94 out of 100 interphase cells in comparison to two red 15q telomere control signals (Fig. 1). This provided conclusive evidence of a single copy FBN1 deletion and exclusion of a translocation in this patient. Subsequent FISH analysis also revealed that the MFAP1 locus, encoding a microfibril associated protein, was deleted in this patient. A single MFAP1-specific cosmid probe signal was detected in 76 out of 100 interphase cells in comparison to two 15q telomere control signals (data not shown). Pulse chase analysis of fibrillin-1 biosynthesis in del(15) dermal fibroblasts Dermal fibroblasts from patient del(15) were established in order to measure fibrillin-1 protein and RNA levels associated with expression of a single FBN1 allele. Initially, pulse chase analyses were performed to investigate profibrillin-1 synthesis. Comparison of the incorporation of radioactivity by profibrillin1 (P) and internal standard X were used to ascertain levels of profibrillin-1 synthesis by control and patient fibroblasts as has previously been described (5,13). Unexpectedly, it was found that similar amounts of profibrillin-1 were synthesised by both patient and control fibroblasts at 0 h, despite patient del(15) having only one FBN1 allele (Fig. 2A). This is in contrast to profibrillin-1 synthesis by MFS PTC patient fibroblasts which has previously been shown to be significantly reduced due to NMD of mutant FBN1 transcript (14,15). These results demonstrate that patient del(15) fibroblasts express fibrillin-1 protein at a level significantly higher than that expected for a single FBN1 allele. Figure 1. Dual-colour fluorescence in situ hybridization of patient del(15) interphase cells using FBN1 specific (green) and 15q telomere (red) control probes showing the deletion of one FBN1 allele. Comparative northern blot analysis of patient del(15) and control fibroblast mRNA verified the pulse chase result (Fig. 2B). The increased profibrillin-1 protein synthesis observed in patient del(15) correlated with mRNA levels of 78.4 3.2% from the remaining FBN1 allele, compared with an unaffected control (100%). Parallel quantitation of mRNA from an MFS patient control with PTC-causing mutation 5358delT (unpublished data) revealed a reduction in total fibrillin-1 mRNA expression to 26.8 1.9% of control levels. Collectively, the del(15) data suggested that differences in levels of normal FBN1 transcript may occur within MFS families. If correct, this would be most easily visualized in MFS patients with PTC mutations where the mutant FBN1 transcript is usually degraded by NMD. Mutation 932insT was confirmed in individuals II.3, II.12 and III.1 by direct sequencing of exon 8 of their respective genomic DNA (data not shown). Initially RNA extracted from patient dermal fibroblasts was analysed by northern blot and RNase protection. In each case, after normalization against GAPDH, fibrillin-1 RNA levels were expressed relative to an unaffected control (see Materials and Methods). In both assays, a reduction in fibrillin-1 RNA was observed for each MFS patient (Fig. 4). However, the amount of fibrillin-1 RNA from individual III.1 was significantly reduced compared with II.12 and II.3. Comparative northern blot analysis of patient and control RNA showed that fibrillin-1 RNA levels from patient II.12 were reduced to 78 2.2% of the control value compared with 27.3 2.3% for patient III.1 and 26.8 1.9% for an isolated MFS patient with PTC-causing mutation 5358delT (unpublished data). Quantitation of RNase protection data demonstrated that the amount of fibrillin-1 RNA detected from III.1 was 42.4 3.3% compared to levels of 69.7 4.9% and 64.1 2.5% from patients II.12 and II.3 respectively. Analysis of FBN1 RNA in a kindred with PTC mutation 932insT RNase protection assays and pulse chase analysis exclude NMD as a cause of the variable expression In a previous study, mutation 932insT resulting in a frameshift and PTC in exon 9 of FBN1, and NMD, was detected in an MFS kindred (15) (Fig. 3). As in many MFS kindreds, intrafamilial phenotypic variability was observed despite complete disease penetrance. To investigate levels of FBN1 expression in this family, three related individuals (II.12, II.3 and III.1) who differ in their severity of MFS clinical manifestations were studied (Table 1). Haplotype analysis confirmed that the mutant allele co-segregated with the MFS phenotype (Fig. 3). Since it could be argued that the differences in RNA levels within the MFS 932insT kindred were due to variations in NMD, further experiments were performed to specifically analyse RNA and protein derived from the normal FBN1 allele. An RNase protection probe (MUT) was designed in order to produce a 447 nucleotide protected fragment when added to RNA transcribed from the 932insT mutant allele (Fig. 5A). In contrast, RNA transcribed from the normal allele should mismatch with the probe at the site of the T nucleotide insertion Northern blot analysis 2272 Human Molecular Genetics, 2003, Vol. 12, No. 18 Figure 3. Pedigree and FBN1 haplotype segregation of the MFS kindred with FBN1 mutation 932insT demonstrating co-segregation of the mutant allele (red) with the MFS phenotype. Extended markers (D15S119, D15S126, D15S1028, D15S143) were used that occur within 1 Mb of the gene as well as intragenic markers (MTS1, MTS2, MTS3, MTS4) for complete informativeness (25). Genotypes that are inferred by pedigree analysis are in brackets. Symbols corresponding to affected individuals are shaded. mutant FBN1 transcript from all three individuals was present at similarly reduced levels (see asterisk, Fig. 5B), therefore excluding variable rates of NMD in these patients. Pulse chase analyses verified the RNase protection result, since a significantly increased level of profibrillin-1 synthesis was observed at 0 h in II.12, compared with III.1 (Fig. 5C). The profibrillin-1 band observed in the pulse chase can only have been produced from the normal allele since the mutant protein is truncated as a consequence of the PTC mutation. Taken together these data prove that allelic variation in normal FBN1 expression occurs in dermal fibroblasts from this family with MFS. Figure 2. (A) Profibrillin-1 synthesis and secretion by del(15) fibroblasts in comparison to an unaffected control. Cells were pulse-labelled for 30 min with 35 35 L-[ S]methionine/L-[ S]cysteine (0 h) and then chased in the absence of label. Total cell lysates chased to 2 h are shown. The positions of profibrillin-1 (P) and protein X are indicated. (B) Northern blot analysis of mRNA extracted from unaffected control fibroblasts (lane 1), an MFS patient control with PTC mutation 5358delT (unpublished data) (lane 2) and del(15) fibroblasts (lane 3). Hybridization was performed with random primed [a-32P]dCTP-labelled cDNA probes complementary to fibrillin-1 and GAPDH transcripts. Quantitation values after normalization using GAPDH mRNA signals are indicated below the figure expressed as mean SEM from six independent experiments. and after RNase digestion result in two fragments 293 and 154 nucleotides in length. RNase protection of fibroblast RNA, using the MUT probe produced fragments of the expected size (Fig. 5A and B). Quantitation of these bands (see Materials and Methods) demonstrated that normal FBN1 allele expression in II.12 and II.3 (78 2.7 and 71.9 7.2%, respectively) was significantly greater than that of III.1 (45.5 5.4%; Fig. 5A and B). However DISCUSSION In this study we have demonstrated that levels of normal FBN1 transcript are higher than expected for a single allele in isolated patient del(15) fibroblasts and variable in an MFS kindred with a previously described PTC mutation. Fibroblast RNA obtained from affected individuals was analysed by both northern blotting and allele-specific RNase protection assays and showed III.1 to have a significantly lower level of normal transcript than either II.12 or II.3. NMD of mutant transcript was similar in all three samples. This is the first study to quantitate normal FBN1 allele expression in clinically affected patients and complements a recent study highlighting differences in normal allelic expression of six genes in individuals from the Centre d’Etude du Polymorphisme Humain (CEPH) genotype database, including FBN1 (16). In this study 2/19 normal individuals showed enhanced expression of FBN1. Previous attempts to quantify FBN1 transcription in MFS patients have focused on estimation of mutant FBN1 transcript levels, particularly in PTC patients where it has been proposed Human Molecular Genetics, 2003, Vol. 12, No. 18 2273 Figure 4. (A) Northern blot analysis of dermal fibroblast mRNA extracted from individuals II.12 (lane 1) and III.1 (lane 2) in parallel with a MFS patient with PTC mutation 5358delT (lane 3) and unaffected control fibroblast mRNA (lane 4). Hybridization was performed using random primed [a-32P]dCTP-labelled cDNA probes complementary to fibrillin-1 and GAPDH transcripts. Quantitation values after GAPDH normalization are indicated below the figure expressed as mean SEM from five independent experiments. (B) RNase protection analysis of total RNA extracted from unaffected control cells (lane 1), individual II.12 (lane 2), individual II.3 (lane 3) and individual III.1 (lane 4). Fibrillin-1 (FIB) and GAPDH pBluescript II constructs were used for this experiment (see Materials and Methods). The resulting protected fragments are indicated above the figure. Lane () shows a digestion control. Lane (þ) shows a fifth of the undigested riboprobes used for the RNase protection. Lane M, size markers indicated in nucleotides. Quantitation values after normalization using protected GAPDH mRNA are indicated below the figure expressed as mean SEM from four to 10 independent experiments. that there is a correlation between the level of mutant transcript and the severity of disease (10,17–19). In these studies the mutant transcript was expressed as a percentage of normal transcript, without normalization against a housekeeping gene such as GAPDH. The demonstration of allelic variation in FBN1 in normal and clinically affected individuals suggests a potential role for normal FBN1 expression in the pathogenesis of MFS. Since fibrillin-1 is a structural protein of connective tissue, and transgenic mouse data have demonstrated a critical threshold of fibrillin-1 microfibrils for tissue function (11,12), it is reasonable to suggest that allelic variation in FBN1 expression could be a potential modifier of phenotype in MFS, particularly in PTC families where the population of FBN1 transcripts is predominantly derived from the normal allele due to NMD. For example, when expression of a normal FBN1 allele is below a critical threshold, it is plausible that this could result in more profound disease. Conversely a higher level of normal FBN1 expression could moderate any pathogenic effect of the mutant allele resulting in a milder phenotype. A low expressing normal FBN1 allele could explain the previous observation of two affected MFS patients with mutations R1541X and R529X that were shown to have severe disease, despite very low mutant RNA levels of 2 and 7%, respectively (17,20). Quantitation of normal FBN1 transcript levels was not performed in these patients. The explanation for allelic variation seen in the patient fibroblasts is unknown. Variation in normal transcript levels is known to arise from a number of factors including promoter polymorphisms that affect the binding of transcription factors and mRNA stability (21–24). In this kindred haplotype analysis demonstrates that individuals II.12 and II.3 who are moderately affected, with MFS manifestations in two of the three main organ systems, and individual III.1 who is more severely 2274 Human Molecular Genetics, 2003, Vol. 12, No. 18 affected, with manifestations in the skeletal, ocular and cardiovascular systems, have different normal alleles. Individual II.2 (Fig. 3, Table 1) who has inherited the identical FBN1 alleles as II.12 and II.3, exhibits the same system involvement as II.12 and II.3, although has more severe cardiovascular manifestations. Unfortunately cells were not available from II.2 to test RNA and protein levels. It is interesting to speculate that complete system involvement (skeletal, cardiovascular, ocular) in this kindred is dependent on the inheritance of the lower-expressing normal allele as in III.1. However other modifying factors may also contibute to the phenotypic spectrum displayed. Patient del(15), although profoundly affected by the extent of the chromosomal deletion exhibits a number of Marfanoid characteristics predominantly affecting the skeletal system (Table 1). However it is not possible to draw any association between fibrillin-1 levels in this patient and phenotype exhibited due to the young age of the patient and the deletion of the MFAP1 locus which may affect microfibril function, independently of fibrillin-1. Establishment of allelic variation in normal FBN1 expression in fibroblast samples from a single family with MFS does not by itself prove the influence of the normal allele on intrafamilial variability. Although an attractive hypothesis, it remains possible that a different modifier is responsible and it is simply by chance that, in this family, allelic variation is observed. It is interesting to note that previous pulse chase studies of PTC patients have not previously revealed raised levels of normal fibrillin-1 expression. This may be because the patients usually studied by this method are index cases with clinically diagnosed MFS, rather than the more mildly affected family members, or alternatively it may indicate that allelic variation in FBN1 is rare and unlikely to be a major factor moderating phenotype. Proof will require analysis of larger numbers of MFS families, particularly those with PTC mutations where allelic variation can be most easily visualized. However our data on individuals from a single family and patient del(15) suggest that, in future, consideration be given to the role of the normal FBN1 allele in MFS pathogenesis. MATERIALS AND METHODS Cell culture and pulse chase analysis of fibrillin-1 Figure 5. (A) RNase protection probe and resulting protected fragments after RNase digestion using the MUT probe. (B) RNase protection analysis of fibroblast nuclear RNA from two unaffected control individuals (lanes 1 and 2), individual II.12 (lane 3), individual II.3 (lane 4) and individual III.1 (lane 5) using MUT and GAPDH probes. The GAPDH probe was as used previously (Fig. 4B). II.12, II.3 and III.1 were overloaded by 1.5, 1.5 and 5, respectively, to enable signal enhancement from the normal allele and detection of the protected mutant allele. The protected mutant allele (asterisk) can be seen in a longer exposure of lanes 3–5 on the right. M: size markers indicated in nucleotides. Quantitation values after normalization using protected GAPDH mRNA are indicated below the figure expressed as mean SEM from six to 10 independent experiments. (C) Pulse chase analysis of individuals II.12 and III.1 with mutation 932insT in comparison to an unaffected control. Cells were pulse-labelled for 30 min with L-[35S]methionine/L-[35S]cysteine (0 h) and then chased in the absence of label. Total cell lysates chased to 2 h are shown. The positions of profibrillin-1 (P) and protein X, used as an internal standard (5,13), are indicated. Deletion patient, control and Marfan patient fibroblasts were established from forearm skin biopsies. The MFS family was selected for study due to its large size, intra-familial clinical variability and previously detected PTC mutation (15). Individual II.2 was unavailable for biopsy. The primary cultures were maintained in Dulbecco’s modified Eagle medium supplemented with 10% fetal calf serum, 50 U/ml penicillin, 50 mg/ml streptomycin and 2 mM L-glutamine. For all experiments where only one unaffected control was used it corresponds to unaffected control 2 in Figure 5B. Pulse chase analysis of fibrillin-1 from patient and control fibroblasts was carried out as described previously (13,15) and profibrillin-1 was confirmed by immunoprecipitation using a prorich region antibody that has been previously characterized (13). Human Molecular Genetics, 2003, Vol. 12, No. 18 Probe isolation and dual colour FISH A chromosome 15 cosmid library was constructed at Los Alamos National laboratory (LANL) and the cosmid library filter obtained from UK HGMP Resource Centre, Cambridge, UK. The filter was screened for FBN1 specific cosmids and FBN1 cosmid clone (LA1537-H4 from the LA15NC01 library) obtained from UK HGMP Resource Centre, Cambridge, UK. Chromosome preparations were made by treating patient fibroblasts with hypotonic solution (7.5 mM KCl) for 20 min at 37 C followed by fixing in 3:1 methanol:acetic acid. Slides were prepared following standard procedures and aged at 20 C. Probe DNA was biotinylated and ethanol precipitated with salmon testis DNA (Gibco BRL), E. coli tRNA (Boehringer) and resuspended in 1 TE. A commercial digoxigenin labelled chromosome 15q sub-telomeric probe (Appligene Oncor) was used as a marker for chromosome 15. The cosmid probe and a 50-fold excess of Cot-1 DNA (Gibco BRL) were dried and resuspended in hybridization solution (50% formamide, 2 SSC, 10% dextran sulfate) to a final concentration of 20 ng/ml. Probe DNA was denatured at 72 C for 5 min and reannealed at 37 C for 15 min. The 15q sub-telomeric probe was used according to the manufacturer’s protocol. The slides were dehydrated in an ethanol series at room temperature and air-dried, then denatured in 70% formamide at 70 C for 2 min, quenched in 70% ethanol at 20 C then dehydrated in an ethanol series. Probe mix (FBN1 cosmid and 15q probe) was hybridized at 37 C for 20 h followed by washes in 50% formamide at 42 C for 10 min and 2 SSC at 42 C for 5 min. The biotinylated probes were detected with Texas red-conjugated streptavidin, biotinylated anti-streptavidin (Vector Laboratories) and a final layer of Texas Red-conjugated streptavidin. The digoxigenin probe was detected using mouse anti-digoxigenin antibody (Roche) and goat anti-mouse Alexa-488 (Molecular Probes). The slides were counterstained and mounted with Vectashield (Vector Laboratories) containing 40 ,6-diamidino-2phenylindole (DAPI). Slides were examined using an Olympus BX-51 epifluorescence microscope coupled to a Sensys CCD camera (Photometrics). One hundred randomly selected interphase nuclei were analysed for each hybridisation experiment. Texas red, Alexa-488 and DAPI fluorescence images were taken separately then pseudocoloured and merged using the software package Genus (Applied Imaging International). 2275 were [a-32P]dCTP-labelled by random priming (Roche). The membrane was hybridized with the fibrillin-1 probe for 16 h at 65 C, washed to a final stringency of 0.2 SSC/0.1% SDS then autoradiographed. Hybridisation of the membrane to the GAPDH probe was repeated without stripping and quantitation carried out by PhosphorImager. RNase protection Total RNA was isolated using the hot-phenol method. Volumes of 450 ml NTE buffer (0.1 M NaCl, 10 mM Tris pH 8, 1 mM EDTA) and 50 ml 10% sodium dodecyl sulfate were added to 500 ml phenol and heated to 90 C. Cell pellets were added to the hot phenol mix, phenol–chloroform extracted, and precipitated. Total RNA pellets were resuspended in 100 ml R loop buffer (27). Nuclear RNA was extracted by hot phenol after cytoplasmic/nuclear separation (28). For total fibrillin-1 analysis fragments of fibrillin-1 cDNA (nucleotides 1086–1555: FIB) and GAPDH cDNA (nucleotides 84–256: GAPDH) were cloned into pBluescript II (Stratagene). For II.12, II.3 and III.1 single allele analysis a fragment of fibrillin-1 cDNA containing the kindred mutation 932insT was cloned into pBluescript II (nucleotides 639–1086: MUT). Riboprobes were made from linearized pBluescript II with T7 (GAPDH, MUT) or T3 (FIB) RNA polymerase according to the manufacturer’s instructions (Promega). Full-length riboprobe was purified from a denaturing polyacrylamide gel in 300 ml buffer (0.5 M ammonium acetate, 0.1% SDS, 1 mM EDTA) for 2 h at 37 C. Eluted RNA was ethanol precipitated, resuspended in water and 50 cps of probe hybridized to total RNA. RNase protection analysis was performed as described previously (27), except that the MUT probe experiment (Fig. 5A) was RNasedigested at 37 C. ACKNOWLEDGEMENTS The authors thank Dr Sally Rose, Dr Mick Dye, Dr Rob Davies and Dr Garry Brown for experimental advice and guidance, Professor Paul Wordsworth for continued support and Dr Lucy Raymond for taking the del(15) skin biopsy. P.H. and S.H. thank the British Heart Foundation and MRC for their support. REFERENCES Microsatellite marker analysis Haplotype analysis of markers D15S119, D15S126, D15S1028, MTS1, MTS2, MTS3, MTS4, D15S1028 was carried out on available pedigree DNA as described previously (25). Northern analysis Messenger RNA was isolated from primary human fibroblasts using the Oligotex direct mRNA kit (Qiagen). RNA was separated on a 0.8% formaldehyde–agarose gel, transferred to Hybond Nþ (Amersham Pharmacia) and pre-hybridized in Church buffer (26) at 65 C. Human cDNA probes encoding fibrillin-1 (nucleotides 1168–1527) and GAPDH (Clontech) 1. Pyeritz, R.E. (2000) The Marfan syndrome. A. Rev. Med., 51, 481–510. 2. Beroud, C., Collod-Beroud, G., Boileau, C., Soussi, T. and Junien, C. (2000) UMD (Universal Mutation Database): a generic software to build and analyze locus specific databases. Hum. Mut., 15, 86–94. 3. Sakai, L.Y., Keene, D.R. and Engvall, E. (1986) Fibrillin, a new 350 kD glycoprotein, is a component of extracellular microfibrils. J. Cell Biol., 103, 2499–2509. 4. Kielty, C.M., Baldock, C., Lee, D., Rock, M.J., Ashworth, J.L. and Shuttleworth, C.A. (2002) Fibrillin: from microfibril assembly to biomechanical function. Phil. Trans. R. Soc. Lond. B, 357, 207–217. 5. Milewicz, D.M., Pyeritz, R.E., Crawford, E.S. and Byers, P.H. (1992) Marfan syndrome: defective synthesis, secretion, and extracellular matrix formation of fibrillin by cultured dermal fibroblasts. J. Clin. Invest., 89, 79–86. 6. Milewicz, D.M., Grossfield, J., Cao, S.N., Kielty, C., Covitz, W. and Jewett, T. (1995) A mutation in FBN1 disrupts profibrillin processing and results in isolated skeletal features of the Marfan syndrome. J. Clin. Invest., 95, 2373–2378. 2276 Human Molecular Genetics, 2003, Vol. 12, No. 18 7. Raghunath, M., Putnam, E.A., Ritty, T., Hamstra, D., Park, E.S., TschodrichRotter, M., Peters, R., Rehemtulla, A. and Milewicz, D.M. (1999) Carboxy-terminal conversion of profibrillin to fibrillin at a basic site by PACE/furin-like activity required for incorporation in the matrix. J. Cell Sci., 112, 1093–1100. 8. Ritty, T.M., Broekelmann, T., Tisdale, C., Milewicz, D.M. and Mecham, R.P. (1999) Processing of the fibrillin-1 carboxyl-terminal domain. J. Biol. Chem., 274, 8933–8940. 9. Kettle, S., Cardy, C.M., Hutchinson, S., Sykes, B. and Handford, P.A. (2000) Characterisation of fibrillin-1 cDNA clones in a human fibroblast cell line that assembles microfibrils. Int. J. Biochem. Cell Biol., 32, 201–214. 10. Dietz, H.C., McIntosh, I., Sakai, L.Y., Corson, G.M., Chalberg, S.C., Pyeritz, R.E. and Francomano, C.A. (1993) Four novel FBN1 mutations— significance for mutant transcript level and EGF-like domain calcium binding in the pathogenesis of Marfan syndrome. Genomics, 17, 468–475. 11. Pereira, L., Andrikopoulos, K., Tian, J., Lee, S.Y., Keene, D.R., Ono, R., Reinhardt, D.P., Sakai, L.Y., Biery, N.J., Bunton, T. et al. (1997) Targetting of the gene encoding fibrillin-1 recapitulates the vascular aspect of Marfan syndrome. Nat. Genet., 17, 218–222. 12. Pereira, L., Lee, S.Y., Gayraud, B., Andrikopoulos, K., Shapiro, S.D., Bunton, T., Biery, N.J., Dietz, H.C., Sakai, L.Y. and Ramirez, F. (1999) Pathogenetic sequence for aneurysm revealed in mice underexpressing fibrillin-l. Proc. Natl Acad. Sci. USA, 96, 3819–3823. 13. Kettle, S., Yuan, X., Grundy, G., Knott, V., Downing, A.K. and Handford, P. (1999) Defective calcium binding to fibrillin-1: consequence of an N2144S change for fibrillin-1 structure and function. J. Mol. Biol., 285, 1277–1287. 14. Aoyama, T., Francke, U., Dietz, H.C. and Furthmayr, H. (1994) Quantitative differences in biosynthesis and extracellular deposition of fibrillin in cultured fibroblasts distinguish 5 groups of Marfan syndrome patients and suggest distinct pathogenetic mechanisms. J. Clin. Invest., 94, 130–137. 15. Halliday, D., Hutchinson, S., Kettle, S., Firth, H., Wordsworth, P. and Handford, P.A. (1999) Molecular analysis of eight mutations in FBN1. Hum. Genet., 105, 587–597. 16. Yan, H., Yuan, W., Velculescu, V.E., Vogelstein, B. and Kinzler, K.W. (2002) Allelic variation in human gene expression. Science, 297, 1143. 17. Hewett, D., Lynch, J., Child, A., Firth, H. and Sykes, B. (1994) Differential allelic expression of a fibrillin gene (FBN1 ) in patients with Marfan syndrome. Am. J. Hum. Genet., 55, 447–452. 18. Schrijver, I., Liu, W., Odom, R., Brenn, T., Oefner, P., Furthmayr, H. and Francke, U. (2002) Premature termination mutations in FBN1: Distinct effects on differential allelic expression and on protein and clinical phenotypes. Am. J. Hum. Genet., 71, 223–237. 19. Mátyás, G., Giunta, C., Steinmann, B., Hossle, J.P. and Hellwig, R. (2002) Quantification of single nucleotide polymorphisms: a novel method that combines primer extension assay and capillary electrophoresis. Hum. Mut., 19, 58–68. 20. Montgomery, R.A., Geraghty, M.T., Bull, E., Gelb, B.D., Johnson, M., McIntosh, I., Francomano, C.A. and Dietz, H.C. (1998) Multiple molecular mechanisms underlying subdiagnostic variants of Marfan syndrome. Am. J. Hum. Genet., 63, 1703–1711. 21. Wilusz, C.J., Wormington, M. and Peltz, S.W. (2001) The cap-to-tail guide to mRNA turnover. Nat. Rev. Mol. Cell. Biol., 2, 237–246. 22. Akai, J., Kimura, A. and Hata, R.I. (1999) Transcriptional regulation of the human type I collagen a2 (COL1A2) gene by the combination of two dinucleotide repeats. Gene, 239, 65–73. 23. van ‘t Hooft, F.M., von Bahr, S.J.F., Silveira, A., Iliadou, A., Eriksson, P. and Hamsten, A. (1999) Two common functional polymorphisms in the promoter region of the b-fibrinogen gene contribute to regulation of plasma fibrinogen concentration. Aterioscler. Thromb. Vasc. Biol., 19, 3063–3070. 24. Jacquelin, B., Rozenshteyn, D., Kanaji, S., Koziol, J.A., Nurden, A.T. and Kunick, T.J. (2001) Characterisation of inherited differences in transcription of the human integrin a2 gene. J. Biol. Chem., 276, 23518–23524. 25. Judge, D.P., Biery N.J. and Dietz H.C. (2001) Characterisation of microsatellite markers flanking FBN1: utility in the diagnostic evaluation for Marfan syndrome. Am. J. Med. Genet., 99, 39–47. 26. Church, G.M. and Gilbert, W. (1984) Genomic sequencing. Proc. Natl Acad. Sci. USA, 81, 1991–1995. 27. Ashe, M.P., Griffin, P., James, W. and Proudfoot, N.J. (1995) Poly(A) site selection in the HIV-1 provirus: inhibition of promoter-proximal polyadenylation by the downstream major splice donor site. Genes Dev., 9, 3008–3025. 28. Eggermont, J. and Proudfoot, N.J. (1993) Poly(A) signals and transcriptional pause sites combine to prevent interference between RNA polymerase II promoters. EMBO J., 12, 2539–2548. 29. De Paepe, A., Devereux, R.B., Dietz, H.C., Hennekam, R.C.M. and Pyeritz, R.E. (1996) Revised diagnostic criteria for the Marfan Syndrome. Am J. Med. Genet., 62, 417–426. 30. Roman, M.J., Devereux, R.B., Kramer-Fox, R. and O’Ranghlin, J. (1989) Two-dimensional echocardiographic aortic root dimensions in normal children and adults. Am. J. Cardiol., 64, 507–512. 31. Ahn, N.U., Sponseller, P.D., Ahn, U.M., Nallamshetty, L., Rose, P.S., Buchowski, J.M., Garrett, E.S., Kuszyk, B.S., Fishman, E.K. and Zinreich, S.J. (2000) Dural ectasia in the Marfan syndrome: MR and CT findings and criteria. Genet. Med., 2, 173–179.