Collagen composition of normal and myxomatous human mitral heart valves.

The collagens were studied in 13 normal and 19 myxomatous human mitral valves. The collagens of the valve were completely solubilized by using a method consisting of guanidinium chloride extraction, limited pepsin digestions and CNBr cleavage of the residue. The normal valves contained 74% type I, 24% type III and 2% type V collagen. The type I and type III collagens had similar solubility patterns, although only type I collagen was detected in the guanidinium chloride extract. Type V collagen was only detected in the first pepsin extract. The type I and III collagens had higher contents of hydroxylysine than did the same collagens from age-matched dermis. The two-dimensional electrophoretic 'maps' of CNBr-cleavage peptides showed low recoveries of the C-terminal alpha 1(I) CB6 and alpha 1(III) CB9 peptides, which are involved in forming intermolecular cross-linkages. Most of the reducible cross-linkages were present in large-Mr peptide complexes, and these complexes were shown by labelling with 125I to include the tyrosine-containing alpha 1(I) CB6 peptide. The myxomatous valves contained 67% type I, 31% type III and 2% type V collagens. There was a significant increase in the concentration of each type of collagen, which consisted of a 9% increase of type I collagen, a 53% increase of type III collagen and a 25% increase of type V collagen. The contents of hydroxylysine in type I and III collagens and the electrophoretic 'maps' of the CNBr-cleavage peptides involved in cross-linkages did not differ significantly from the results obtained from the normal valves. The biochemical findings suggest that there is an increased production of collagen, in particular type III collagen, and glycosaminoglycan as well as a proliferation of cells as part of a repair process in the myxomatous valves.     

Collagen defects in lethal perinatal osteogenesis imperfecta.

Quantitative and qualitative abnormalities of collagen were observed in tissues and fibroblast cultures from 17 consecutive cases of lethal perinatal osteogenesis imperfecta (OI). The content of type I collagen was reduced in OI dermis and bone and the content of type III collagen was also reduced in the dermis. Normal bone contained 99.3% type I and 0.7% type V collagen whereas OI bone contained a lower proportion of type I, a greater proportion of type V and a significant amount of type III collagen. The type III and V collagens appeared to be structurally normal. In contrast, abnormal type I collagen chains, which migrated slowly on electrophoresis, were observed in all babies with OI. Cultured fibroblasts from five babies produced a mixture of normal and abnormal type I collagens; the abnormal collagen was not secreted in two cases and was slowly secreted in the others. Fibroblasts from 12 babies produced only abnormal type I collagens and they were also secreted slowly. The slower electrophoretic migration of the abnormal chains was due to enzymic overmodification of the lysine residues. The distribution of the cyanogen bromide peptides containing the overmodified residues was used to localize the underlying structural abnormalities to three regions of the type I procollagen chains. These regions included the carboxy-propeptide of the pro alpha 1(I)-chain, the helical alpha 1(I) CB7 peptide and the helical alpha 1(I) CB8 and CB3 peptides. In one baby a basic charge mutation was observed in the alpha 1(I) CB7 peptide and in another baby a basic charge mutation was observed in the alpha 1(I) CB8 peptide. The primary defects in lethal perinatal OI appear to reside in the type I collagen chains. Type III and V collagens did not appear to compensate for the deficiency of type I collagen in the tissues.     

A structural mutation of the collagen alpha 1(I)CB7 peptide in lethal perinatal osteogenesis imperfecta

Structurally abnormal type I collagen was identified in the dermis, bone, and cultured fibroblasts obtained from a baby with lethal perinatal osteogenesis imperfecta. Two-dimensional gel electrophoresis of the CNBr peptides demonstrated that the alpha 1(I)CB7 peptide from the alpha 1(I)-chain of type I collagen existed in a normal form and a mutant form with a more basic charge distribution. This heterozygous peptide defect was not detected in the collagens from either parent. The defect was localized to a 224-residue region at the NH2 terminus of the alpha 1(I)CB7 peptide by mammalian collagenase digestion. Analysis of unhydroxylated collagens produced in cell culture indicated that the mutant alpha 1(I)CB7 migrated faster on electrophoresis suggesting that the abnormality may be a small deletion or a mutation that alters sodium dodecyl sulfate binding. The post-translational hydroxylation of lysine residues was increased in the CB7 peptide and also in peptides CB3 and CB8 which are toward the NH2 terminus of the alpha 1(I)-chain. The COOH-terminal CB6 peptide was normally hydroxylated. These findings support the proposal that the lysine overhydroxylation resulted from a perturbation of helix propagation from the COOH to NH2 terminus of the collagen trimer caused by the structural defect in alpha 1(I)CB7.     

Quantitation of type I and III collagens using electrophoresis of alpha chains and cyanogen bromide peptides

The methods of quantitating the relative amounts of type I and III collagens in samples containing crosslinked collagen chains were evaluated using electrophoresis of alpha chains and cyanogen bromide peptides. The densitometry areas of the alpha I(I) chains from type I collagen and the alpha I(III) chains from type III collagen were reduced because of the failure of the crosslinked chains to dissociate. However, the ratios of the unit densitometry areas of these chains (area of chain/micrograms type I or III collagen loaded) were constant for type I and III collagens prepared from the same samples of tissue. A calibration factor, which was the same for dermis and mitral valve, was derived to convert the densitometry area ratios to the weight ratios of type I to III collagens. In contrast, the densitometry areas of the alpha I(I) CB8 (type I collagen marker) and the alpha I(III) CB5 (type III collagen marker) were not reduced by crosslinked collagen chains. A calibration factor was also derived to convert the ratios of the densitometry areas of the marker peptides to weight ratios of type I to type III collagens. Almost identical results were obtained when electrophoresis of alpha chains and of cyanogen bromide peptides was used with these calibration factors to quantitate the relative amounts of type I and III collagens in tissue extracts which contained different amounts of crosslinked chains.     

Platelet-reactive sites in collagens type I and type III. Evidence for separate adhesion and aggregatory sites.

The adhesion of human and rabbit platelets to collagens and collagen-derived fragments immobilized on plastic was investigated. Adhesion appeared to be independent of collagen conformation, since similar attachment occurred to collagen (type I) in monomeric form, as fibres or in denatured state. The adhesion of human platelets was stimulated to a variable degree by Mg2+, but rabbit platelet adhesion showed little if any dependence on this cation. Collagens type I, III, V and VI were all able to support adhesion, although that to collagen type V (native) was lower than that to the other collagens. Adhesion to a series of peptides derived from collagens I and III was measured. Attachment did not require the presence of peptides in triple-helical configuration. The extent of adhesion ranged from relatively high, as good as to the intact parent collagen molecule, to little if any adhesive activity beyond the non-specific (background) level. The existence of very different degrees of activity suggests that platelet adhesion is associated with specific structural sites in the collagen molecule. Adhesion in many instances was essentially in accord with the known platelet-aggregatory activity of individual peptides. However, two peptides, alpha 1(I)CB3 and alpha 1(III)CB1,8,10,2, exhibited good adhesive activity although possessing little if any aggregatory activity. Of particular interest, despite its near-total lack of aggregatory activity, adhesion to peptide alpha 1(I)CB3 was as good as that to the structurally homologous peptide alpha 1(III)CB4, in which is located a highly reactive aggregatory site. This implies that platelet adhesion to collagen may involve sites in the collagen molecule distinct from those more directly associated with aggregation.     

Fibril-forming collagens in lamprey

Five types of collagen with triple-helical regions approximately 300 nm in length were found in lamprey tissues which show characteristic D-periodic collagen fibrils. These collagens are members of the fibril forming family of this primitive vertebrate. Lamprey collagens were characterized with respect to solubility, mobility on sodium dodecyl sulfate-polyacrylamide gel electrophoresis, carboxylmethyl-cellulose chromatography, peptide digestion patterns, composition, susceptibility to vertebrate collagenase, thermal stability, and segment long spacing-banding pattern. Comparison with fibril-forming collagens in higher vertebrates (types I, II, III, V, and XI) identified three lamprey collagens as types II, V, and XI. Both lamprey dermis and major body wall collagens had properties similar to type I but not the typical heterotrimer composition. Dermis molecules had only alpha 1(I)-like chains, while body wall molecules had alpha 2(I)-like chains combined with chains resembling lamprey type II. Neither collagen exhibited the interchain disulfide linkages or solubility properties of type III. The conservation of fibril organization in type II/type XI tissues in contrast to the major developments in type I and type III tissues after the divergence of lamprey and higher vertebrates is consistent with these results. The presence of type II and type I-like molecules as major collagens and types V and XI as minor collagens in the lamprey, and the differential susceptibility of these molecules to vertebrate collagenase is analogous to the findings in higher vertebrates.     

Lethal perinatal osteogenesis imperfecta due to the substitution of arginine for glycine at residue 391 of the alpha 1(I) chain of type I collagen

A baby with the lethal perinatal form of osteogenesis imperfecta was shown to have a structural defect in the alpha 1(I) chain of type I procollagen. Normal and mutant alpha 1(I) CB8 cyanogen bromide peptides, from the helical part of the alpha 1(I) chains, were purified from bone. Amino acid sequencing of tryptic peptides derived from the mutant alpha 1(I) CB8 peptide showed that the glycine residue at position 391 of the alpha 1(I) chain had been replaced by an arginine residue. This substitution accounted for the more basic charged form of this peptide that was observed on two-dimensional electrophoresis of the collagen peptides obtained from the tissues. The substitution was associated with increased enzymatic hydroxylation of lysine residues in the alpha 1(I) CB8 and the adjoining CB3 peptides but not in the carboxyl-terminal CB6 and CB7 peptides. This finding suggested that the sequence abnormality had interfered with the propagation of the triple helix across the mutant region. The abnormal collagen was not incorporated into the more insoluble fraction of bone collagen. The baby appeared to be heterozygous for the sequence abnormality and as the parents did not show any evidence of the defect it is likely that the baby had a new mutation of one allele of the pro-alpha 1(I) gene. The amino acid substitution could result from a single nucleotide mutation in the codon GGC (glycine) to produce the codon CGC (arginine).     

Partial covalent structure of the human alpha 2 type V collagen chain

Human cDNA libraries were screened with a cDNA fragment presumably encoding the 3' terminus of a procollagen carboxyl propeptide not identifiable as types I, II, III, or IV by protein sequence or Northern blot hybridization. One clone contained a 1350-base pair insert coding in part for 55 uninterrupted Gly-X-Y triplets. Comparison with the amino acid composition of the COOH-terminal cyanogen bromide (CB) peptides of the alpha 1 and alpha 2 type V collagen chains showed similarity only to the alpha 2(V)CB fragment. To identify the NH2 terminus of the peptide designated by methionine, an additional isolate was sequenced and found to contain a Gly-Met-Pro triplet. Thirty-one amino acids from the NH2 terminus of the alpha 2(V)CB9 fragment were then determined by Edman degradation and found to be identical to those derived from the cDNA clone. The DNA sequence encoding part of the triple helical region establishes for the first time the partial structure of a type V collagen chain. Although comparison of residues 796-1020 of the alpha 2(V) collagenous region with alpha 1 (III), alpha 1(I), and alpha 2(I) shows strong conservation of charged positions, the latter three chains appear considerably more similar to each other than to alpha 2(V). A striking feature of the alpha 2(V) sequence between 918-944 is the absence of proline residues. In the analogous region of alpha 1(I) where this amino acid is also lacking, a flexible site in the rigid triple helical structure of type I collagen has been observed (Hofmann, H., Voss, T., Kuhn, K. and Engel, J. (1984) J. Mol. Biol. 172, 325-343).     

Bovine type I collagen: A study of cross-linking in various mature tissues

The cyanogen bromide peptides from insoluble and pepsin solubilised type I collagen of bovine bone, dentine, meniscus, tendon, skin and cornea were compared by SDS-polyacrylamide gel electrophoresis. In each case alpha 1CB6 was shown to be the only peptide of molecular weight greater than 10 000 involved in cross-linking. The major helical peptides alpha 1CB3, alpha 1CB8, alpha 1CB7 and alpha 2CB4 were not implicated in cross-linking in any tissue either by end overlap or helix-helix interaction. The C-terminal alpha 2 chain peptide alpha 2CB3,5, which contains a large helical region, was not involved in cross-linking to any large peptides, although a slight increase in molecular weight in all tissues examined did suggest a possible interaction(s) with a very small peptide of molecular weight 4--5000.     

Characterization of pepsin-solubilized bovine heart-valve collagen.

Collagens extracted from heart valves by using limited pepsin digestion were fractionated by differential salt precipitation. Collagen types were identified by sodium dodecyl sulphate/polyacrylamide-gel electrophoresis, amino acid analysis and cleavage with CNBr. Heart-valve collagen was heterogeneous in nature, consisting of a mixture of type-I and type-III collagens. The identity of type-III collagen was established on the basis of (a) insolubility in 1.7 M-NaC1 at neutral pH, (b) behaviour of this collagen fraction on gel electrophoresis under reducing and non-reducing conditions, (c) amino acid analysis showing a hydroxyproline/proline ratio greater than 1, and (d) profile of CNBr peptides on sodium dodecyl sulphate/polyacrylamide-gel electrophoresis showing a peak characteristic for type-III collagen containing peptides alpha1(III)CB8 and alpha1(III)CB3. In addition to types-I and -III collagen, a collagen polypeptide not previously described in heart valves was identified. This polypeptide represented approx. 30% of the collagen fraction precipitated at 4.0 M-NaCl, it migrated between beta- and alpha1-collagen chains on sodium dodecyl sulphate/polyacrylamide-gel electrophoresis and its electrophoretic behaviour was not affected by disulphide-bond reduction. All collagen fractions from the heart valves contained increased amounts of hydroxylysine when compared with type-I and -III collagens from other tissues. The presence of beta- and gamma-chains and higher aggregates in pepsin-solubilized collagen indicated that these collagens were highly cross-linked and suggested that some of these cross-links involved the triple-helical regions of the molecule. It is likely that the higher hydroxylysine content of heart-valve collagen is responsible for the high degree of intermolecular cross-linking and may be the result of an adaptive mechanism for the specialized function of these tissues.     

A novel binding site in collagen type III for integrins alpha1beta1 and alpha2beta1

Previously identified high affinity integrin-binding motifs in collagens, GFOGER and GLOGER, are not present in type III collagen. Here, we first characterized the binding of recombinant I domains from integrins alpha(1) and alpha(2) (alpha(1)I and alpha(2)I) to fibrillar collagen types I-III and showed that each I domain bound to the three types of collagens with similar affinities. Using rotary shadowing followed by electron microscopy, we identified a high affinity binding region in human type III collagen recognized by alpha(1)I and alpha(2)I. Examination of the region revealed the presence of two sequences that contain the critical GER motif, GROGER and GAOGER. Collagen-like peptides containing these two motifs were synthesized, and their triple helical nature was confirmed by circular dichroism spectroscopy. Experiments show that the GROGER-containing peptide was able to bind both alpha(1)I and alpha(2)I with high affinity and effectively inhibit the binding of alpha(1)I and alpha(2)I to type III and I collagens, whereas the GAOGER-containing peptide was considerably less effective. Furthermore, the GROGER-containing peptide supported adhesion of human lung fibroblast cells when coated on a culture dish. Thus, we have identified a novel high affinity binding sequence for the collagen-binding integrin I domains.     

Quantitation of type I, III, and V collagens in human tissue samples by high-performance liquid chromatography of selected cyanogen bromide peptides

A method to determine the proportions of the major fiber-forming collagens (types I, III, and V) in noncartilaginous human tissues is presented. The procedure relies on direct solubilization of tissue collagen as cyanogen bromide peptides. The peptides are subjected to cation exchange chromatography followed by gel permeation chromatography in a manner consistent with the rapid resolution and quantitation of relatively low-molecular-weight marker peptides for each collagen. The marker peptides utilized for type I, III, and V collagens are alpha 1 (I)-CB2, alpha 1 (III)-CB2, and alpha 1 (V)-CB1, respectively. Quantitation of the peptides is attained as a function of ultraviolet absorbance during gel permeation chromatography. The nature of the marker peptides, the use of high-performance liquid chromatography techniques, and quantitation of the peptides by ultraviolet absorbance renders the method suitably rapid, sensitive, and accurate for routine evaluations of collagen composition. The utility of the method is illustrated in the presentation of analyses on specimens of placental membranes and blood vessel walls.     

Incorporation of type I collagen molecules that contain a mutant alpha 2(I) chain (Gly580-->Asp) into bone matrix in a lethal case of osteogenesis imperfecta.

To understand more directly the tissue defect in osteogenesis imperfecta (OI), bone matrix was analyzed from an infant with lethal OI (type II) of defined mutation (collagen alpha 2(I)Gly580-->Asp). Pepsin-solubilized alpha 1(I) and alpha 2(I) chains and derived CNBr-peptides migrated more slowly on sodium dodecyl sulfate-polyacrylamide gel electrophoresis compared with normal human controls. The peptide alpha 2(I)CB3,5, predicted to contain the mutation site, ran as a retarded doublet band and was purified by high performance liquid chromatography and digested with V8 protease. Two peptides with amino-terminal sequences beginning at residue 576 of the alpha 2(I) chain were isolated. One had the normal sequence. The other differed in that aspartic acid replaced glycine at residue 580 as predicted from cDNA analysis, and in having an unhydroxylated proline at residue 579. From yields on microsequencing and the relative intensities of the two forms of alpha 2(I)CB3,5 on SDS-polyacrylamide gel electrophoresis, the ratio of mutant to normal alpha 2(I) chains in the infant's bone matrix was 0.7/1. Although the effects of an efficient incorporation of mutant chains on the properties of the bone matrix are unknown, it may be that in this OI case the tissue abnormalities result more from the presence of mutant protein than from an underexpression of matrix.     

Collagen XXIV,a vertebrate fibrillar collagen with structural features of invertebrate collagens: selective expression in developing cornea and bone

Tissue-specific assembly of fibers composed of the major collagen types I and II depends in part on the formation of heterotypic fibrils, using the quantitatively minor collagens V and XI. Here we report the identification of a new fibrillar-like collagen chain that is related to the fibrillar alpha1(V), alpha1(XI), and alpha2(XI) collagen polypeptides and which is coexpressed with type I collagen in the developing bone and eye. The new collagen was designated the alpha1(XXIV) chain and consists of a long triple helical domain flanked by typical propeptide-like sequences. The carboxyl propeptide is classic, with 8 conserved cysteine residues. The amino-terminal peptide contains a thrombospodin-N-terminal-like (TSP) motif and a highly charged segment interspersed with several tyrosine residues, like the fibril diameter-regulating collagen chains alpha1(V) and alpha1(XI). However, a short imperfection in the triple helix makes alpha1(XXIV) unique from other chains of the vertebrate fibrillar collagen family. The triple helical interruption and additional select features in both terminal peptides are common to the fibrillar chains of invertebrate organisms. Based on these data, we propose that collagen XXIV is an ancient molecule that may contribute to the regulation of type I collagen fibrillogenesis at specific anatomical locations during fetal development.     

Quantitation of types I and III collagens in human tissue samples and cell culture by cyanogen bromide peptide analysis

A procedure for the quantitation of types I and III collagens by cyanogen bromide peptide analysis was developed with the aim of eliminating certain problems associated with this method. Ion-exchange chromatography reduced high background levels on gel scans used to quantitate the peptides; reduction with beta-mercaptoethanol substantially increased the efficiency of the cyanogen bromide cleavage; use of a concave gradient in acrylamide from 8 to 20% improved the resolution of cyanogen bromide peptides separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis; and a normalization procedure eliminated variations due to differences in the amount of material loaded on the gel system. This method of quantitation was applied to human aorta samples and to collagen secreted by human skin fibroblasts. Metachromasy of type I and type III collagen cyanogen bromide peptides stained with Coomassie blue R-250 was established and this was used as an index of the purity of the cyanogen bromide peptide preparations. Type I and III collagens were prepared from human placental tissue, and these purified collagens were used to construct calibration curves to determine the relationship between the quantity of diagnostic cyanogen bromide peptides present and the composition of the sample in terms of types I and III collagens.     

Collagen cross-linking. Isolation of cross-linked peptides from collagen of chicken bone

Cross-linked peptides were isolated from chicken bone collagen that had been digested with CNBr or with bacterial collagenase. Analyses of (3)H radioactivity in disc electrophoretic profiles of the CNBr peptides from bone collagens that had been treated with NaB(3)H indicated that a major site of intermolecular cross-linking in chicken bone collagen is located between the carboxy-terminal region of an alpha1 chain and a small CNBr peptide, probably situated near the amino-terminus of an alpha1 or alpha2 chain in an adjacent collagen molecule. A small amount of this cross-linked CNBr peptide was isolated from a CNBr digest of chicken bone collagen by column chromatography. Amino acid analysis showed that the CNBr peptide, alpha1CB6B, the carboxy-terminal peptide of the alpha1 chain, was the major CNBr peptide in the preparation, and the reduced cross-linking components were identified as hydroxylysinohydroxynorleucine (HylOHNle), with a smaller amount of hydroxylysinonorleucine (HylNle). However, the composition and the low recovery of the cross-linking amino acids suggested that the preparation was a mixture of CNBr peptides alpha1CB6B and alpha1CB6B cross-linked to a small CNBr peptide whose identity could not be determined. A small cross-linked peptide was isolated from chicken bone collagen that had been reduced with NaB(3)H(4) and digested with bacterial collagenase. This peptide was the major cross-linked peptide in the digest and contained a stoicheiometric amount of the reduced cross-linking compounds. A peptide which had the same amino acid composition, but contained the cross-linking compounds in their reducible forms, was isolated from a collagenase digest of chicken bone collagen that had not been treated with NaBH(4). The absence of the reduced cross-links from this peptide indicates that, at least for the cross-linking site from which the peptide derives, natural reduction is not a significant pathway for biosynthesis of stable cross-links. However, most of the reducible cross-linking component in the peptide appeared to stabilize in the bone collagen by rearrangement from aldimine to ketoamine form.     

Isolation and characterization of CNBr peptides of human (alpha 1 (III) )3 collagen and tissue distribution of (alpha 1 (I) )2 alpha 2 and (alpha 1 (III) )3 collagens.

[Alpha 1(III)]3 collagen was solubilized by pepsin digestion of normal human placental membranes and was purified by differential salt precipitation and carboxymethylcellulose chromatography. This collagen was digested with CNBr, and the resultant nine peptides were isolated and characterized. The chains are cross-linked by cysteinyl residues in the COOH-terminal peptide. Isolation of peptides derived from CNBr digestion of insoluble tissues was used as an assay for the presence of [alpha 1(I)]2alpha 2 and [alpha 1(III)]3 collagens. Both types are present in human skin, intestine, liver, spleen, kidney, lung, aorta, umbilical cord, placental membranes, and myocardium. Bone and tendon contain [alpha 1(I)]2alpha 2 collagen but, unlike the other tissues, lack [alpha 1(III)]3 collagen. Both [alpha 1(I)]2alpha 2 and[alpha 1(III)]3 collagens are present in scars of human skin, myocardium, tendon, and liver and of rabbit skin. The degree of hydroxylation of proline was 4 to 5% lower in the same peptides in skin, bone, and tendon than in the other tissues. The degree of hydroxylation of lysine in the same peptides derived from different tissues varied more widely.     

The covalent structure of calf skin type III collagen. IV. The amino acid sequence of the cyanogen bromide peptide alpha 1(III)CB5 (positions 552--788).

The cyanogen bromide peptide alpha 1(III)CB5 is 237 amino acid residues in length and occupies position 552--788 along the alpha 1(III) chain. For sequence analysis alpha 1(III)CB5 was fragmented with hydroxylamine, protease from Staphylococcus aureus V8, trypsin and the arginine-specific enzyme from mouse submaxillary gland. The peptides obtained were separated using molecular and ion exchange chromatography and sequenced with the automated Edman degradation procedure.     

Identification of cyanogen bromide peptides involved in intermolecular cross-linking of bovine type III collagen.

Cyanogn bromide peptides derived from bovine type III collagen and containing reducible cross-links were isolated and identified. Two peptides, alpha 1 (III)CB7 and alpha 1 (III)CB9B, from within the helical portion of the molecule were shown to contain the 'amino donor' residues cross-linked to non-helical 'aldehyde donor' residues in the formation of cross-links. This information, in conjunction with previously published data for the order of the cyanogen bromide peptides [Fietzek, Allman, Rauterberg & Wachter (1977) Proc. Natl. Acad. Sci. U.S.A. 74, 84-86], suggests that in type III collagen intermolecular cross-links are located in the end-overlap regions, so as to stabilize a quarter-stagger arrangement of molecules within the fibre in a similar manner to that proposed for type I and type II collagens.     

The human immune response to reconstituted bovine collagen

The human immune response to bovine dermal collagen was characterized through histologic, serologic, and immunoblotting methods. Collagen-sensitive patients were identified by hypersensitivity to intradermal exposure to ZYDERM Collagen Implant--a pepsin-solubilized, reconstituted, bovine dermal collagen. Biopsies of test sites in the forearm were obtained from several collagen-sensitive patients. Histologic examination revealed an implant-associated palisading foreign body granuloma. The lesion also contained a mixed cell infiltrate of histiocytes, lymphocytes, and eosinophils. Sera were collected from patients who developed erythema or induration at intradermal test or treatment sites, and were evaluated for antibodies to bovine dermal collagen by an enzyme-linked immunosorbent assay (ELISA). Sera with anti-collagen antibodies were further characterized in this study. The circulating antibodies were reactive with both native and heat-denatured bovine dermal collagen. By using purified alpha 1(I) and alpha 2(I) polypeptides, these sera were found to have antibodies reactive with both alpha-chains. Each alpha-chain was fragmented by using cyanogen bromide (CB). The CB peptides were electrophoretically separated, and these sera were evaluated for antibodies to the major fragments by using an immunoblotting technique. Of the sera evaluated by this method, 89% (23/26) had antibodies to alpha 1-CB6; 77% (20/26) had antibodies to alpha 2-CB4; and 65% (17/26) had antibodies reactive with both CB fragments. In addition, most sera (77%) contained antibodies reactive with two or more (up to five) of the major CB peptides. The least antigenic fragment was alpha 2-CB3,5 (8%). In addition, these sera had antibody activity against both native and heat-denaturated bovine types III and II collagens. Little or no interspecies (rat or guinea pig) cross-reactivity (types I and II) was detected. Furthermore, these sera did not have antibodies against human types I, II, and III collagens.