Aug 11, 1981 - ate was normally arched but she had had several teeth removed because of crowding. She had a mild pectus carinatum of the lower chest, mild ...
Proc. NatL Acad. Sci. USA
Vol. 78, No. 12, pp. 7745-7749, December 1981 Medical Sciences
Marfan syndrome: Abnormal a2 chain in type I collagen (procollagen/prepro a chains/collagen crosslinkdng/two-dimensional peptide mapping)
PETER H. BYERS*t, ROBERT C. SIEGELO§, KAREN E. PETERSON*, DAVID W. ROWE$, KAREN A. HOLBROOKiII, LYNNE T. SMITHS, YU-HUA CHANGt, AND JOSEPH C. C. Fu* Departments of *Pathology, IlBiological Structure, and tMedicine, University of Washington, Seattle, Washington 98195; tDepartments of Medicine and Orthopedic Surgery, University of California, San Francisco, California 94143; and IDepartment of Pediatrics, University of Connecticut, Farmington, Connecticut 06032
Communicated by Earl P. Benditt, August 11, 1981
ABSTRACT Cells in culture from a woman with a variety of the Marfan syndrome produce two species ofthe a2 chains of type I collagen. One a2 chain appears normal; the abnormal chain has a higher apparent molecular weight than normal and migrates more slowly during electrophoresis in sodium dodecyl sulfate/ polyacrylamide gels. A similar change in electrophoretic behavior is seen in the preproa2 chain and the pNa2 chain (which contains the amino-terminal extension). Asymmetric cleavage of the pepsin-treated procollagens with a fibroblast collagenase locates the abnormal segment amino terminal to the cleavage site, and analysis of cyanogen bromide peptides of collagenase cleavage peptides and of whole collagens indicates that the abnormal segment is in either the a2CB3 peptide or the short segment of c2CB5 amino terminal to the collagenase site of the altered a2 chain. The higher apparent molecular weight is consistent with the insertion of a small peptide fragment of approximately 20 amino acids. This alteration in chain size has marked effects on crosslinking because collagen from the patient's skin was 5-10 times more extractable in nondenaturing solvents than that from control skins. Although the abnormal chain was not found in several other individuals with the Marfan syndrome, these findings suggest that the phenotype may be the expression of a variety of primary structure alterations in the chains of type I collagen that interfere with normal crosslink formation.
The Marfan syndrome (1-4) is a group of inherited connective tissue disorders that share certain clinical features: severe cardiovascular abnormalities (valvular heart disease and aortic dissection), skeletal alterations (arachnodactyly, pectus deformities, and kyphoscoliosis), and ocular lens dislocation. Despite considerable clinical interest (1) and research activity (3, 4), a clear understanding of the syndrome at the biochemical level has not emerged. Investigations of glycosaminoglycan synthesis (5-7) and of collagen metabolism (8-13) have suggested that alterations in these macromolecules may be important in the pathogenesis of the syndrome. Dermal fibroblasts from some patients with the Marfan syndrome have been shown to have an increased rate of synthesis of hyaluronic acid (5, 6), and cellfree systems from these cells also synthesize hyaluronic acid at a higher than normal rate (7). It is not clear, however, that this is the primary disorder rather than a secondary change. Similarly, it is not clear that increased extractability ofcollagen from tissues or of that synthesized by cultured cells reflects the primary biochemical disturbance. Recently Siegel and Chang (11) and Scheck et al. (12) studied tissues from a woman with a variety of the Marfan syndrome. They found two species of the a2** chain of type I collagen in skin and aorta. We have extended their studies to an analysis of the type I collagen synthesis by dermal fibroblasts in culture
and have found that the abnormal a2 chain appears to contain a small insertion in the a2CB3 or a2CB5 peptide that affects collagen crosslink formation. This alteration appears to provide the pathogenetic basis for connective tissue disease in this patient and suggests that additional mechanisms that interfere with crosslink formation may underlie other forms ofthe Marfan
syndrome. MATERIALS AND METHODS Clinical Summary. The patient, a 39-year-old woman, was the only affected child of three in the family; both parents were unaffected. She was noted to have equinovarus deformities of both feet at birth; arachnodactyly was first noted at age 9; lumbar scoliosis became apparent and a heart murmur was heard at age 10. Because ofpain and progressive deformity, the lumbar scoliosis was corrected by fusion at age 12 without serious problems. She remained relatively healthy until her mid-20s, when she began to experience frequent tachyarrhythmias and decreased exercise tolerance. By age 30 aortic dilatation, aortic valvular insufficiency, and mitral regurgitation were apparent and episodes of tachyarrhythmias accompanied by congestive heart failure became more frequent. Cardiac catheterization at age 37 revealed marked aortic dilation with aneurysms of the sinuses of Valsalva, and aortic and mitral valvular insufficiency. Because of cardiac decompensation the aortic valve and a portion of the ascending aorta were replaced with a porcine heterograft. Histologic study of the aorta and aortic valve revealed accumulation of lakes of mucopolysaccharide, consistent with cystic medial necrosis, and ultrastructural examination showed marked disorganization of collagen fibrils in those areas (12). When she was seen in early 1979 her height was 164.5 cm and her span was 178 cm with a lower segment of 91.5 cm (upper to lower segment ratio 0.80) Her sclerae were bluish-grey; there was no iridodonesis and previous careful ophthalmologic examination of the eyes with full pupillary dilatation had detected no evidence oflens dislocation. She had mild myopia. Her palate was normally arched but she had had several teeth removed because of crowding. She had a mild pectus carinatum of the lower chest, mild lower thoracic scoliosis with concavity to the left, long slender extremities with increased mobility ofalljoints except in the 4th and 5th fingers of both hands where she had marked camptodactyly. § To whom reprint requests should be addressed at: 101 S. San Mateo Drive, #211, San Mateo, CA 99401. **Type I collagen contains three a chains: the molecule is denoted [al(I)]2a2. The chains are synthesized as preproa chains, which contain "leader" sequences that are cleaved in the rough endoplasmic reticulum to produce proa chains. The proa chains differ from a chains by virtue of peptide extensions at both the amino and carboxy terminals. pN chains contain only the amino-terminal extension; pC chains contain only the carboxy-terminal extension.
The publication costs ofthis article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact. 7745
7746
Medical Sciences: Byers et aL
Methods. Skin samples were obtained from all subjects with their informed consent by biopsy of the inner aspect of the upper arm. Site-matched samples were obtained from the proband's sibs and parents, from several controls, and from 10 other patients with typical features of the Marfan syndrome. A portion of each was prepared for electron microscopic studies by standard techniques (14), a second portion was used to explant and establish cell strains, and a third was used to determine collagen extractability. To measure collagen extractability from tissues (15), small amounts of skin from the affected woman and from eight ageand sex-matched controls were homogenized, separately, in 1 M NaCl/50 mM Tris HCl (pH 7.5) and stirred for 48 hr. The insoluble residue was then extracted sequentially for 48 hr each in 0.5 M acetic acid and 4 M CaCl2. All extractions were done at 40C. The soluble collagens and insoluble residual collagen were measured as hydroxyproline by automated amino acid analysis after hydrolysis for 24 hr in 6 M HC1 at 108TC. The amino acid composition ofwhole skin was determined similarly. Fibroblasts from skin were grown in Dulbecco-Vogt modified Eagle's medium containing 10% fetal calf serum or 10% newborn calf serum. Cells were plated at high density (7.5 x 106 per 60-mm plate), allowed to attach overnight, and then labeled for 24 hr with [2,3-3H]proline at 50 ,Ci/ml (New England Nuclear, 25 Ci/mmol; 1 Ci = 3.7 x 1010 becquerels) in the modified Eagle's medium lacking fetal calf serum but supplemented with ascorbic acid at 50 Ag/ml. When proteins were labeled for peptide mapping, 2.5 X 105 cells were plated in 35mm culture dishes and then incubated with [3H]proline at a concentration of 200 ACi/ml in 0.4 ml. Medium and cell layer were harvested separately into 25 mM EDTA/10 mM phenylmethanesulfonyl fluoride (medium) or in 0.15 M NaCV50 mM Tris-HCI (pH 7.5)/25 mM EDTA/10 mM phenylmethanesulfonyl fluoride (cell layer), to inhibit proteolysis. For electrophoresis in sodium dodecyl sulfate/polyacrylamide slab gels (16) proteins were dialyzed against 1 mM ammonium bicarbonate/0. 1 mM phenylmethanesulfonyl fluoride/0.5 mM N-ethylmaleimide and then lyophilized. Dry samples were dissolved in 50 ,ul of electrophoresis sample buffer and denatured in boiling water for 3 min. Radioactive proteins in slab gels were detected by autoradiofluorography (17). Some radiolabeled collagenous proteins were dialyzed against 0.5 M acetic acid and then subjected to limited proteolysis with pepsin (18) to produce collagen-size molecules. After pepsin treatment some samples were analyzed by slab gel electrophoresis and others were cleaved with purified fibroblast collagenase (19, 20), a generous gift from Eugene Bauer. The fibroblast collagenase cleavage products were separated by electrophoresis in sodium dodecyl sulfate/8% polyacrylamide gels and then digested, in the gels, with cyanogen bromide (21). Gel strips that contained the separated collagenase peptides were cut out immediately after electrophoresis, without prior fixation, equilibrated with 70% (vol/vol) formic acid, and incubated in 70% formic acid that contained cyanogen bromide at 50 mg/ml for 6 hr at 250C. At the end of the digestion, each strip was equilibrated with 100 mM Tris HCl (pH 6.8)/30% (vol/vol) glycerol and placed horizontally over a 12.5% polyacrylamide separating gel with a 5% stacking gel. A 100 mM Tris buffer, pH 6.8, which contained 5% sodium dodecyl sulfate and 20% (vol/vol) glycerol was layered over the gel strip and the cyanogen bromide-derived peptides were separated by electrophoresis at a current of 20 mA. mRNA from confluent cells in two 100-mm dishes was prepared and translated as described (22). An aliquot of the translation mixture that contained [3H]proline-labeled preproa chains (23) was mixed with concentrated sample buffer and applied to a slab gel for electrophoresis.
Proc. Nad Acad. Sci. USA 78 (1981) Table 1. Collagen extractability from skin % of collagen extracted Controls Solvent Patient (n = 8) 1 M sodium chloride 8.6 0-2 0.5 M acetic acid 17.3 2-6 4 M calcium chloride 11.7 5-15 Insoluble
62.4
80-95
Two-dimensional electrophoresis of cyanogen bromide peptides of pepsin-digested [3H]proline-labeled procollagens was performed as described by Benya (24) except that the second dimension was run in a 12.5% polyacrylamide gel. Lysyl oxidase was assayed as described (25). RESULTS Collagen from skin of the affected woman was considerably more extractable in nondenaturing solvents than that from normals. More than 25% of total skin collagen was soluble in these solvents, compared to less than 8% for controls (Table 1). The amino acid composition of the affected skin was normal; the degree of lysyl and prolyl hydroxylation was similar in affected and control skin. The increased extractability was not due to abnormalities in lysyl oxidase because activities measured in affected skin, in affected aorta, or in medium from affected cultured dermal fibroblasts were similar to their respective controls (not shown). When collagenous proteins synthesized in culture by cells obtained from skin were examined by sodium dodecyl sulfate/ polyacrylamide gel electrophoresis, there appeared to be two species of a2 (after treatment of procollagens with pepsin) and pNa2 (the precursor chain that contains the amino-terminal extension but lacks the carboxy-terminal extension) (Fig. 1, lanes a-d). This abnormal chain was not found in the parents or sibs ofthe patient or in 10 additional patients with the Marfan syndrome. To determine whether the altered migration of the a
b
c
d
e
f
A-*
-prepro c I
preprooa2
cXl(l)- so
so
cx2 M C medium pepsin
M C M C medium cell-free translation
FIG. 1. Collagenous chains synthesized by cells in culture (lanes a-d) and synthesized in cell-free translation systems (lanes e and 0 by
Marfan (M) or control (C) cells. Medium proteins for lanes a and b were treatedwith pepsin prior to electrophoresis. In the cell-free translation, two preproa2 chains are clearly visible from the Marfan medium (lane e); in lanes a and c the a2 and pNa2 chains are represented by a broad band with a major component that migrates more slowly than the normal. This difference in band clarity is probably due to extensive posttranslational modification of the chains made by cells in culture that is absent from the chains translated in a cell-free synthetic system. Only the regions of the gel that contain the collagenous proteins are shown in order to demonstrate more clearly the significant differences between the Marfan and control proteins. The identity of the band between preproal and preproa2 in lane f is unknown. When lanes c and d are exposed for a shorter time there is a complex band pattern surrounding proa2, and a proa2 doublet cannot be distinguished from the pattern produced by pNa(I) and proa2.
Proc. Natd Acad. Sci. USA 78 (1981)
Medical Sciences: Byers et al abnormal a2 was the result of posttranslational modifications or of a mutation in the chain itself, mRNA was isolated from cells in culture and translated in a modified reticulocyte lysate, and the collagenous proteins were examined by electrophoresis. In the affected individual, but not in four controls, there were two chains that migrated in the region of preproa2;. one migrated in the position of normal preproa2, the other had a slightly slower migration (Fig. 1, lanes e.and f). The chains labeled preproal and preproa2 in Fig. 1, lanes e and f, including the preproa2 chain with the higher apparent molecular weight from the affected cells, could be digested with purified bacterial collagenase and were precipitated with an antibody directed toward type I procollagen (not shown). Because the preproa2 chain, which is not modified by hydroxylation or glycosylation, had an altered electrophoretic mobility, the apparent alteration in molecular weight is probably due to inserted peptide material. When the radiolabeled collagenous proteins secreted into culture medium by skin fibroblasts were treated with pepsin
b
a
I.,
.2.
4
3+
-.
1
54
to remove the non-triple-helical domains of procollagen and then digested with mammalian collagenase, only the TCA fragment of the abnormal a2 chain had a slower than normal migration (Fig. 2). Mammalian collagenase cleaves the a2 chain in the carboxy-terminal a2CB5 peptide (see Fig. 3). However, the methionine residue between a2CB3 and a2CB5 is not well cleaved, so that much of the a2CB3 peptide exists in a form that includes the amino-terminal portion of a2CB5. We designate that peptide a2CB3-5c. The cyanogen bromide peptides from the TCA fragments of the al(I) and normal a2 chains from the affected cells were identical to the controls. The a2CB3-5c peptide in the abnormal a2 chain had a higher apparent molecular
i ziss.
aI2
.: . .
5
FIG. 3. Diagrammatic representation of the a2 chain for control and affected tissues. The normal a2 chain is above, the two possible abnormal chains are below. The insertion, not drawn to scale, is represented by the stippled region. The vertical lines represent the methionine residues in the a2 chain and the peptides obtained by cleavage with cyanogen bromide are numbered. The arrowhead indicates the site of cleavage by fibroblast collagenase.
4
TC A
,
3
7747
I_
0
I
I
I
-
-cx2CB3-5c
0
-- 2CB 4
od(I)CB8al(I)CB7c-
v
TCB C FIG. 2. Autoradiofluorogram of [3Hlproline-labeledfibroblast collagenase cleavage products from pepsin-treated medium collagen. Lane a, control; lane b, Marfan. The TCA segments of a2 from the Marfan medium contain an abnormal peptide (arrow). TCA fiagments are located amino terminal to the cleavage site while TB fragments are located carboxy terminal to the site (see Fig. 3). Some pepsin preparations produce heterogeneity in the a chain size fragments. However, because the location of the abnormality in the Marfan a2 chain is internal to the helical domain, heterogeneity of terminal cleavage does not introduce problems in interpretation of data (see Figs. 3 and 4).
M
FIG. 4. Autoradiofluorogram of [5Hlproline-labeled cyanogen bromide-derived peptides from TCA fragments of al(I) and a2 chains. The peptides from al(I) and the smaller a2 fragment in the Marfan type I collagen (M) are identical to control (C); However, the higher molecular weight TCAa2 fragment from the Marfan culture contains a peptide (a2CB3-5c-the fragment that contains the collagenase cleavage site) that is higher in apparent molecular weight than that from the control TCAa2 fragment (see arrow). The first-dimension gel is placed above the second dimension; large arrows indicate direction of migration in each dimension. Identities of peptide spots were confirmed by comparison with similar maps from proa and a chains.
7748
Medical Sciences: Byers et alPProc. Nad Acad., Sci. USA 78 (1981) acidic
-.-
basic
a
cx2CB3
a2CB4 A.* k2CB5 Fs 6.0 lw
a
CXICB8
b Qc2CB3
cx2CB5 CM'7
a2-CB4
odICB8*
FIG. 5. Autoradiofluorogram of two-dimensional electrophoresis of [5Hlproline-labeled cyanogen bromide peptides of pepsin-treated medium collagens. (a) Control; (b) Marfan. Nonequilibrium isoelectric focusing was performed in the first dimension; electrophoresis in sodium dodecyl sulfate/12.5% polyacrylamide was performed in the second dimension. In b there is an additional peptide (arrow) that is more acidic than control a2CB3 and that has a slightly higher molecular weight than the normal. This suggests that the abnormal peptide in a2 resides in the a2CB3 domain. The identity of each control spot was previously established by electrophoresis of individual isolated peptides. Only a small segment of the second-dimension gel is shown to demonstrate the region with the additional peptide.
weight than the control (Fig. 4, arrow), indicating that the abnormal segment of the chain was located in this domain. To determine whether the altered segment was located in a2CB3 and a2CB5 we separated the cyanogen bromide fragments from pepsin-treated medium procollagens by two-dimensional electrophoresis. The patterns for a2CB4 and a2CB5 from affected medium are similar to those of the control (Fig. 5a). However, in the region of the a2CB3 peptide from the affected cells there is an additional, more acidic, spot that has a slightly higher apparent molecular weight (Fig. Sb). This suggests that the altered segment is located in a2CB3, although we cannot exclude the possibility that it is in a2CB5 amino terminal to the collagenase cleavage site (Fig. 3). The apparent increase in molecular weight in the preproa2, pepsin-treated a2, the TCAa2 fragment, and the cleaved a2CB3-5c peptide is about 2000 to 2500 or 20-25 amino acids. This alteration in a2 does not appear to have major consequences for collagen fibril formation because fibril diameter, integration, and banding patterns appear normal in skin, as judged by light and electron microscopy, although the density of collagen in the papillary and reticular dermis was increased compared to controls. Thus the alterations in aorta seen by Scheck et al. (12) probably represent nonspecific changes. DISCUSSION The collagens are a family of molecules that share several structural features but have organ- and tissue-specific distributions (see refs. 3, 4, 26, and 27 for reviews). Type I collagen is the most abundant species, has a virtually ubiquitous distribution, and is a major component of blood vessels, heart valves, and skin, all of which are affected in the Marfan syndrome. This collagen molecule consists of two al(I) chains and one a2 chain-[al(I)]2a2-which are products of different genes. Like other proteins destined for export, the chains are initially synthesized as preproa chains in the rough endoplasmic reticulum.
The "signal" sequence is cleaved as the chain is inserted through the cisternal membrane, certain lysine and proline residues are hydroxylated, some hydroxylysine residues and other residues are glycosylated, the chains are assembled into trimers, disulfide bonds are formed, a triple helix configuration is assumed, and the procollagen molecules are packaged in the Golgi apparatus and then secreted. Subsequently, terminal non-triplehelical extensions are removed by proteolysis and the collagen molecules pack into fibrils that are stabilized by enzymatic crosslinking. Our findings indicate that a structural alteration in one allele for the a2 gene of type I collagen results in a clinical picture with features of the Marfan syndrome. In skin from this patient there are two species of a2 that are separated by electrophoresis in sodium dodecyl sulfate/polyacrylamide gels (11, 12). One chain comigrates with the normal a2, while the other migrates more slowly. The precursor chains to the abnormal a2 that are synthesized by cells in culture from two affected tissues also migrate more slowly than their. normal counterparts. The abnormal and normal chains are apparently incorporated equally well into procollagen molecules and appear to form equally stable molecules, as demonstrated by resistance of the collagen helix to protease (pepsin) digestion. The increase in molecular weight appears to be due to the insertion of a short collagenous segment in the domain of the a2CB3 peptide or in the small domain of a2CB5 amino terminal to the mammalian collagenase cleavage site (Fig. 3). That there is an insertion is indicated by the existence of two species of preproa chains that migrate in the region of preproa2 during electrophoresis in sodium dodecyl sulfate/polyacrylamide gels. Because these chains do not undergo posttranslational modification in the translation system (22), the abnormal migration is most likely due to an insertion. Although a hybrid al/a2 chain could be considered, this is an unlikely possibility, because cyanogen bromide cleavage of isolated a2 chains indicates no peptides are present that represent other collagen a chains. Furthermore, it does not appear to be due to abnormal procollagen-to-collagen conversion, because the abnormal segment is located within an internal helical fragment of a2 and, as indicated above, is seen in the initial translation product. This is in contrast to the abnormal a2 chain from a patient with Ehlers-Danlos syndrome type VII, recently described by Steinmann et aL (28), that appears to have a structural alteration, at or near the amino-terminal procollagen protease cleavage site, that interferes with proteolytic processing. It is possible that the altered migration in sodium dodecyl sulfate/polyacrylamide gels of the Marfan a2 is due to a single amino acid substitution, but the marked alterations in collagen solubility seem inconsistent with such a change. The size of the inserted fragment appears to be about 20-25 amino acids. Recently, Boyd et aL (29), Vogeli et ad (30), and Wozney et aL (31) demonstrated that the coding segments in the a2 gene in sheep and chicken are small (50-100 bases) and are separated -by extensive intervening sequences. Because the additional sequence in the abnormal a2 chain appears to have a helical structure (it is resistant to proteolysis), it may have been derived by duplication of an entire coding segment. The data from tissues (11, 12), from cells in culture, and from cell-free translation of isolated mRNA demonstrate the two chains are synthesized in equal amounts and so suggest that, in these tissues, only a single cr2 locus is expressed. In terms of function, the affected region of the c2 chain appears to be relatively dormant. It does not contain major crosslink sites and is not a site for collagenase cleavage. Nonetheless, there is a striking effect on collagen crosslink formation, reflected in the increased extractability of the collagen from tis-
Medical Sciences: Byers et aL sues. We think that the effect on crosslink formation occurs because the important sites for crosslinking (32) in the nonhelical telopeptide extensions at one or both ends and in the helical region of a2CB4 are shifted with respect to adjacent type I collagen molecules in tissues so that appropriate alignment is lost between relevant lysine residues. Because initial crosslink formation appears to involve preferential union of al(I) and a2, this shift in a2 chain register may profoundly alter connective tissue tensile strength (25). In animals, disturbances in collagen crosslink formation produce skeletal changes and dilation and rupture ofthe aorta similar to those seen in the Marfan syndrome (33-35). These changes are seen in animals treated with 13-aminopropionitrile, which inhibits lysyl oxidase, or penicillamine, which blocks reactive aldehyde groups produced by lysyl oxidase, and in mice that bear certain alleles at the mottled locus on the X chromosome (34). The disorder in mice, inherited in an X-linked recessive fashion, is due to marked decrease in lysyl oxidase activity in the relevant connective tissues (34, 35). Although it has been proposed that lysyl oxidase deficiency could result in the clinical features of the Marfan syndrome, lysyl oxidase deficiency in humans is an X-linked disorder with the clinical features of cutis laxa, accompanied by internal organ and bony involvement (36). Furthermore, direct assay of lysyl oxidase activity in our patient and in cultured fibroblasts from other patients with the Marfan syndrome (37) have shown it to be normal. Thus the genetic evidence and the biochemical data from this patient now favor abnormalities in the collagen a chains themselves. Others have suggested that abnormalities in the regulation of hyaluronic acid synthesis (5-7) or that alterations in the relative content of different types of collagen in affected vessels (13) may be the basis ofthe Marfan syndrome in some affected individuals. Taken together, these studies suggest that there is significant biochemical heterogeneity within the Marfan syndrome. It is likely that additional studies of individuals with the Marfan syndrome will reveal a spectrum of biochemical alterations that, in common, affect collagen crosslink formation. These could include substitutions for lysines involved in major crosslinks, alterations at lysyl oxidase binding sites on the substrate, or interference by noncollagen molecules with normal packing. The investigation of the different individuals and families should help to understand and predict differences in the natural history of the disease, give genetic counseling a more rational basis, and provide a more complete understanding of the function of certain connective tissue macromolecules. We thank the patient and her family for their interest, Eugene Bauer for the generous gift of purified fibroblast collagenase, Richard Palmiter for his help with translation of procollagen mRNA, Paul Benya for providing prepublication details of the method for two-dimensional separation of collagen cyanogen bromide peptides, Sue Linkhart for excellent assistance with cell culture, and Virginia Wejak for preparing the manuscript. This work was supported by grants from the National Institutes of Health (AM-21557, AM-18237, AM-21897, and GM-07266), a research grant from the National Foundation-March of Dimes, a Basil O'Connor starter grant from the National Foundation-March of Dimes, a Clinical Research Grant (6-298) from the March of Dimes-Birth Defects Foundation, and a supply grant from the Poncin Fund. R.C. S. was supported in part by Research Career Development Award AM-00114 from the National Institutes of Health; P.H.B. is an Established Investigator of the American Heart Association. 1.
McKusick, V. A. (1972) Heritable Disorders of Connective Tissue (Mosby, St. Louis, MO), 4th Ed.
Proc. NatL Acad. Sci. USA 78 (1981)
7749
2. Pyeritz, R. E. -& McKusick, V. A. (1979) N. Engl J. Med. 300, 772-777. 3. Bornstein, P. & Byers, P. H. (1980) in Metabolic Control and Disease, eds. Bondy, P. & Rosenberg, L. (Saunders, Philadelphia), 8th Ed., pp. 1089-1153. 4. Pinnell, S. R. (1978) in The Metabolic Basis of Inherited Disease, eds. Stanbury, J. B., Wyngaarden, J. B. & Fredrickson, D. S. (McGraw-Hill, New York), 4th Ed., pp. 1366-1394. 5. Matalon, R. & Dorfman, A. (1968) Biochem. Biophys. Res. Commun. 32, 150-154. 6. Lamberg, S. I. & Dorfman, A. (1973) J. Clin. Invest. 52, 2428-2433. 7. Appel, A., Horwitz, A. L. & Dorfman, A. (1979) J. Biol Chem. 254, 12199-12203. 8. Macek, M., Hurych, J., Chvapil, M. & Kadlecova, V. (1966) Humangenetik 3, 87-97. 9. Laitinen, O., Uitto, J., fivanainen, M., Hannuksela, M. & Kivirikko, K. I. (1968) Clin. Chim. Acta 21, 321-326. 10. Priest, R. E., Moinuddin, J. R. & Priest, J. H. (1973) Nature (London) 245, 264-266. 11. Siegel, R. C. & Chang, Y.-H. (1978) Clin. Res. 26, 501A (abstr.). 12. Scheck, M., Siegel, R. C., Parker, J., Chang, Y.-H. & Fu, J. C. C. (1979) J. Anat. 129, 645-657. 13. Krieg, T. & Muller, P. K. (1977) Exp. Cell Biol 45, 207-221. 14. Vogel, A., Holbrook, K. A., Steinmann, B., Gitzelmann, R. & Byers, P. H. (1979) Lab. Invest. 40, 201-206. 15. Pinnell, S. R., Krane, S. M., Kenzora, J. E. & Glimcher, M. J. (1972) N. EngL J. Med. 286, 1013-1020. 16. Laemmli, U. K. (1970) Nature (London) 227, 680-685. 17. Bonner, W. M. & Laskey, R. A. (1974) Eur. J. Biochem. 46, 83-88. 18. Layman, D. L., McGoodwin, E. B. & Martin, G. R. (1971) Proc. Natl Acad. Sci. USA 68, 454-458. 19. Stricklin, G. P., Bauer, E. A., Jeffrey, J. J. & Eisen, A. Z. (1977) Biochemistry 16, 1607-1615. 20. Stricklin, G. P., Eisen, A. Z., Bauer, E. A. & Jeffrey, J. J. (1978) Biochemistry 17, 2331-2337. 21. Barsh, G. S. & Byers, P. H. (1981) Proc. Natl Acad. Sci. USA 78, 5142-5146. 22. Rowe, D. W., Moen, R. C., Davidson, J. M., Byers, P. H., Bornstein, P. & Palmiter, R. D. (1978) Biochemistry 17,1581-1590. 23. Palmiter, R. D., Davidson, J. M., Gagnon, J., Rowe, D. W. & Bornstein, P. (1979) J. Biol Chem. 254, 1433-1436. 24. Benya, P. D. (1981) Collagen Relat. Res. 1, 17-26. 25. Siegel, R. C. (1979) Int. Rev. Connect. Tissue Res. 8, 73-118. 26. Prockop, D. J., Kivirikko, K. I., Tuderman, L. & Guzman, N. A. .(1979) N. EngL J. Med.. 301, 13-23; 77-85. 27. Bornstein, P. & Sage, H. (1980) Annu. Rev. Biochem. 48, 957-1004. 28. Steinmann, B., Tuderman, L., Peltonen, L., Martin, G. R., McKusick, V. A. & Prockop, D. J. (1980) J. Biol Chem. 255, 8887-8893. 29. Boyd, C. D., Tolstoshev, P., Schafer, M. P., Trapnell, B. C., Coon, H. C., Kretschmer, P. J., Nienhuis, A. W. & Crystal, R. G. (1980)J. Biol Chem. 255, 3212-3220. 30. Vogeli, G., Avvedimento, E. V., Sullivan, M., Maizel, J. V., Lozano, G., Adams, S. L., Pastan, I. & deCrombrugghe, B. (1980) Nucleic Acids Res. 8, 1823-1837. 31. Wozney, J., Hanahan, D., Morimoto, R., Boedtker, H. & Doty, P. (1981) Proc. Natl Acad. Sci. USA 78, 712-716. 32. Fietzek, P. & Kuhn, K. (1976) Int. Rev. Connect. Tissue Res. 7, 1-60. 33. Ponseti, I. V. & Baird, W. A. (1952) Am. J. Pathol. 28, 1059-1078. 34. Rowe, D. W., McGoodwin, E. B., Martin, G. R., Sussman, M. D., Grahn, D., Faris, B. & Franzblau, C. (1974) J. Exp. Med. 139, 180-192. 35. Rowe, D. W., McGoodwin, E. B., Martin, G. R. & Grahn, D. (1977) J. Biol Chem. 252, 939-942. 36. Byers, P. H., Siegel, R. C., Holbrook, K. A., Narayanan, A. S., Bornstein, P. & Hall, J. G. (1980) N. Engl J. Med. 303, 61-65. 37. Layman, D. L., Narayanan, A. S. & Martin, G. R. (1972) Arch. Biochem. Biophys. 149, 97-101.
