Search

Prolyl 4-hydroxylase (P4H) catalyzes the posttranslational hydroxylation of (2 S)-proline (Pro) residues in procollagen strands. The resulting (2 S,4 R)-4-hydroxyproline (Hyp) residues are essential for the folding, secretion, and stability of the collagen triple helix. Even though its product (Hyp) differs from its substrate (Pro) by only a single oxygen atom, no product inhibition has been observed for P4H. Here, we examine the basis for the binding and turnover of substrates by human P4H. Synthetic peptides containing (2 S,4 R)-4-fluoroproline (Flp), (2 S,4 S)-4-fluoroproline (flp), (2 S)-4-ketoproline (Kep), (2 S)-4-thiaproline (Thp), and 3,5-methanoproline (Mtp) were evaluated as substrates for P4H. Peptides containing Pro, flp, and Thp were found to be excellent substrates for P4H, forming Hyp, Kep, and (2 S,4 R)-thiaoxoproline, respectively. Thus, P4H is tolerant to some substitutions on C-4 of the pyrrolidine ring. In contrast, peptides containing Flp, Kep, or Mtp did not even bind to the active site of P4H. Each proline analogue that does bind to P4H is also a substrate, indicating that discrimination occurs at the level of binding rather than turnover. As the iron(IV)-oxo species that forms in the active site of P4H is highly reactive, P4H has an imperative for forming a snug complex with its substrate and appears to do so. Most notably, those proline analogues with a greater preference for a C (gamma)- endo pucker and cis peptide bond were the ones recognized by P4H. As Hyp has a strong preference for C (gamma)- exo pucker and trans peptide bond, P4H appears to discriminate against the conformation of proline residues in a manner that diminishes product inhibition during collagen biosynthesis.

The procollagen C-proteinase (PCP) is a zinc peptidase of the astacin family and the metzincin superfamily. The enzyme removes the C-terminal propeptides of fibrillar procollagens and activates other matrix proteins. Besides its catalytic protease domain, the procollagen C-proteinase contains several C-terminal CUB modules (named after complement factors C1r and C1s, the sea urchin UEGF protein, and BMP-1) and EGF-like domains. The two major splice forms of the C-proteinase differ in their overall domain composition. The longer variant, termed mammalian tolloid (mTld, i.e., PCP-2), has the protease-CUB1-CUB2-EGF1-CUB3-EGF2-CUB4-CUB5 composition, whereas the shorter form termed bone morphogenetic protein 1 (BMP-1, i.e., PCP-1) ends after the CUB3 domain. Two related genes encode proteases similar to mTld in humans and have been termed mammalian tolloid like-1 and -2 (mTll-1 and mTll-2, respectively). For mTll-1, it has been shown that it has C-proteinase activity. We demonstrate that recombinant EGF1-CUB3, CUB3, CUB3-EGF2, EGF2-CUB4, and CUB4-CUB5 modules of the procollagen C-proteinase can be expressed in bacteria and adopt a functional antiparallel beta-sheet conformation. As shown by surface plasmon resonance analysis, the modules bind to procollagen I in a 1:1 stoichiometry with dissociation constants (K(D)) ranging from 622.0 to 1.0 nM. Their binding to mature collagen I is weaker by at least 1 order of magnitude. Constructs containing EGF domains bind more strongly than those consisting of CUB domains only. This suggests that a combination of CUB and EGF domains serves as the minimal functional unit. The binding affinities of the EGF-containing modules for procollagen increase in the order EGF1-CUB3 < CUB3-EGF2 < EGF2-CUB4. In the context of the full length PCP, this implies that a given module has an affinity that continues to increase the more C-terminally the module is located within the PCP. The tightest binding module, EGF2-CUB4 (K(D) = 1.0 nM), is only present in mTld, which might provide a quantitative explanation for the different efficiencies of BMP-1 and mTld in procollagen C-proteinase activity.

Procollagen C-proteinase-2 (pCP-2, mTld) is derived from the longest splicing variant of the gene encoding bone morphogenetic protein 1 (BMP-1). The variants have identical amino terminal signal peptides, prodomains and astacin-like protease domains. However, they differ in the length of their carboxy terminal part, which in pCP-2 has the composition CUB1, CUB2, EGF-like1, CUB3, EGF-like2, CUB4, CUB5, and C-tail. In the shorter form, pCP-1 (i.e., BMP-1), the sequence ends after the CUB3-domain. Using a combination of mutagenesis and structural approaches, we have investigated the structure and function of subfragments of pCP-2. The full-length latent recombinant enzyme and its N-terminally truncated form lacking the prodomain were tested for their enzymic activity. The intact protein showed only partial processing of procollagen type I, whereas the truncated form expressed enzymic activity indistinguishable from its native counterpart purified from chick embryo tendons. These results clearly demonstrated that the prodomain is required for the latency of the enzyme but not for its correct folding. Limited proteolysis of the recombinant protein with alpha-chymotrypsin produced four discrete fragments revealing the location of cleavage sites between the repetitive CUB/EGF domains. The results provide evidence that the CUB sequences form independently folded modules that are stabilized by two pairs of internal disulfide bridges. The modules are linked to each other by more flexible, hinge-like peptides. Solid-phase binding assays with isolated CUB domains and immobilized procollagen type I demonstrated that the first three but not the last two CUB domains specifically bound to the substrate. To define putative sites for CUB-CUB or CUB-substrate interactions, we generated molecular models for pCP-2 CUB domains. The models were obtained using as a template the structure of CUB domain in zona pellucida adhesion protein PSP-I/PSP-II from porcine sperm. The predicted conformations for homology models were, subsequently, confirmed by circular dichroism spectroscopy of polypeptide domains isolated following limited proteolysis with alpha-chymotrypsin.

Assembly of the heterotrimeric procollagen I molecule is initiated by interactions between the carboxyl propeptide domains of the completed nascent pro alpha chains. The [pro alpha 1(I)]2[pro alpha 2(I)] heterotrimer is the predominant molecule, with much smaller amounts of stable [pro alpha 1(I)]3 homotrimer also being formed. However, the [pro alpha 2(1)]3 homotrimer has not been detected, raising questions as to the mechanism of chain assembly and why [pro alpha2(1)]3 homotrimers are not formed. These questions have been examined here by expressing the intact and amino- or carboxyl-terminal truncated C-propeptides of the pro alpha chains recombinantly in bacteria and in a coupled transcription/translation reticulocyte lysate system. Their interactions were studied in vitro by binding analyses and in vivo by using the yeast two-hybrid system. The C-pro alpha 1(I) interacted with itself, and with C-pro alpha 2(I), as expected. Surprisingly, the C-pro alpha 2(I) also interacted with itself, both in vitro and in vivo. While the interaction of C-pro alpha 2(I) with itself and C-pro alpha 1(I) in vitro was strong, these interactions were weaker in vivo. Deletion of 36 amino acids from the C-terminal domain of C-pro alpha 1 had no effect on its binding to intact self or intact C-pro alpha 2, but the same deletion in C-pro alpha 2 completely abolished its binding to intact C-pro alpha 2 and to C-pro alpha 1. Comparable N-terminal deletions in C-pro alpha 1 or C-pro alpha 2 diminished, but did not abolish, their binding to intact C-pro alpha 1 and C-pro alpha 2. In the yeast two-hybrid system, C-pro alpha 2 interacted with itself more weakly than with C-pro alpha 1. Molecular modeling and circular dichroism analyses showed that C-pro alpha 1 and C-pro alpha 2 have different folded structures and stability. Studies with antibodies specific to the C-pro alpha1 and alpha2 peptides showed them to precipitate different, specific, and distinct cell proteins from fibroblast lysates. The C-pro alpha 2(I) antibody complexed with more cell proteins. We hypothesize that the lack of pro alpha 2(I) homotrimers is not due to the inability of the C-pro alpha 2(I) to interact with itself, but rather to the competing presence of other cell proteins. The specificity of these interactions may reside in conformational differences in N- and C-terminal sequences of the two propeptides or in their different folding patterns.

The alpha 2 beta 1 integrin serves as a cell surface collagen or collagen/laminin receptor. Binding of the integrin to its ligands is largely mediated by the alpha 2 subunit I domain and requires the presence of divalent cations. Terbium ion (Tb3+), a fluorescent trivalent cation that often binds divalent cation-binding sites on proteins, supported binding of the I domain to collagen with half-maximal binding occurring at 5.2 +/- 1.7 microM Tb3+. By fluorescence resonance energy transfer spectroscopy, Tb3+ showed specific and saturable binding to the recombinant I domain with a Kd of 27 +/- 4 microM. Although both Mg2+ and Mn2+ were capable of quenching Tb3+ fluorescence, Mn2+ was much more effective than Mg2+. The alpha 2 beta 1 integrin also binds the pro-alpha 1(I) collagen carboxyl-terminal propeptide in a Mg2+-dependent manner via the I domain. Recombinant propeptide was used to examine the effect of ligand on the Tb3+ binding properties of the alpha 2 integrin I domain. As propeptide bound to the I domain, Tb3+ fluorescence progressively diminished suggesting that as ligand binds to the I domain, either Tb3+ is displaced or its fluorescence is quenched. Consistent with the former possibility, little dissociation of collagen-bound I domain occurred upon the addition of EDTA and subsequent incubation. These data support a model in which (1) the divalent cation is required for initial ligand-binding activity of the I domain and (2) ligand binding results in subsequent metal ion displacement to generate a metal-free I domain-ligand complex.

Type I procollagen was purified from cultured fibroblasts of a proband with a lethal variant of osteogenesis imperfecta. The protein was a mixture of normal procollagen and mutated procollagens containing a substitution of cysteine for glycine in either one pro alpha 1(I) chain or both pro alpha 1(I) chains, some or all of which were disulfide-linked through the cysteine at position alpha 1-748. The procollagen was then examined in a system for generating collagen fibrils de novo by cleavage of the pCcollagen to collagen with procollagen C-proteinase [Kadler et al. (1987) J. Biol. Chem. 262, 15696-15701]. The mutated collagens and normal collagens were found to form copolymers under a variety of experimental conditions. With two preparations of the protein that had a high content of alpha 1(I) chains disulfide-linked through the cysteine alpha 1-748, all the large structures formed had a distinctive, highly branched morphology that met one of the formal criteria for a fractal. Preparations with a lower content of disulfide-linked alpha 1(I) chains formed fibrils that were 4 times the diameter of control fibrils. The formation of copolymers was also demonstrated by the observation that the presence of mutated collagens decreased the rate of incorporation of normal collagen into fibrils. In addition, the solution-phase concentration at equilibrium of mixtures of mutated and normal collagens was 5-10-fold greater than that of normal collagen.(ABSTRACT TRUNCATED AT 250 WORDS)

The effect of 17 beta-estradiol (E2) on the biosynthesis of collagen in cultured bovine aortic smooth muscle cells was explored. Cells treated with various concentrations of the hormone for 14 days following subcultivation were subjected to growth studies. The cultures were also evaluated for [14C]hydroxyproline formation, the presence of collagenase-susceptible protein, prolyl hydroxylase activity, and procollagen types. There were no effect of E2 on the growth of these cells. At 10(-8) M E2, the hydroxylation of proline when compared to control cultures was reduced by 25-30%; however, little difference in extractable prolyl hydroxylase activity or total [14C]proline incorporation into protein was observed. The effect on collagen synthesis appears to be dose dependent over concentrations of E2 ranging from 10(-6) to 10(-12) M when measured by collagenase susceptibility. Procollagen typing on diethylaminoethylcellulose displayed reduced amounts of procollagen type I and type III fractions as well as other collagenous components. More importantly, however, the ratio of these two procollagen types were also altered. Similar results were obtained from the medium or cell layer. It is concluded that aortic smooth muscle cells cultured in the presence of 17 beta-estradiol display a decreased production of collagen in addition to altering the ratio of type I to type III procollagen fractions produced.

A procollagen C-proteinase which cleaves the C-propeptides from type I procollagen was purified about 125-fold from membranous bones of chick embryos. As estimated by gel filtration, the enzyme was about 80 000 daltons. When a reaction with modified procollagen was carried out, the enzyme preferentially cleaved the C-propeptides from the pro alpha chains in the order pro alpha 1, pro alpha 1, and then pro alpha 2. The enzyme was inhibited by several metal chelators and high concentrations of dithiothreitol. It was also inhibited by 5% fetal calf serum. A series of inhibitors of serine proteinases and sulfhydryl-containing proteinases were not inhibitory. Four oligopeptides were synthesized with amino acid sequences similar to the amino acid sequences around the sites at which the C-propeptides are cleaved during the conversion of procollagen top collagen in vivo. The peptide Tyr-Tyr-Arg-Ala-Asp-Asp-Ala inhibited the enzyme 35--60% in concentrations of 6--12 mM. Shorter peptides containing the Ala-Asp bond cleaved by the enzyme were less effective. The partially purified enzyme was also found to cleave the C-propeptides from type II and type III procollagens.

Collagen synthesis and deposition in the predentine of the continuously erupting rat incisor was analyzed in vivo following a single intraperitoneal injection of [14C]glycine. Newly synthesized collagen was extracted from the dissected predentine with a solution of 1.0 m sodium chloride containing proteolytic enzyme inhibitors. Mature, cross-linked collagen was solubilized by limited pepsin digestion following extractions with 0.5 M acetic acid or 0.1 M penicillamine. Analysis of the radiolabeled collagens by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and fluorography showed that over 97% of collagen synthesized migrated in the alpha 1 (I) and alpha 2 positions following pepsin digestion; the remaining 3% migrated in the positions of alpha 1 (V) and alpha 3 (V) collagens. However, from the ratio of the alpha 1 (I): alpha 2 it was estimated that approximately 30% of the salt-extractable collagen was alpha 1 (I) trimer. The presence of this collagen was confirmed by salt-fractionation and cyanogen bromide digestion patterns. Type I, alpha 1 (I) trimer, and type V collagens were also found in the salt-insoluble tissue residue. In this fraction the alpha 1 (I) trimer comprised 10-15% of the collagen measured as radioactivity but was difficult to discern colorimetrically. Type III collagen could not be detected in any of the fractions analyzed. From the profiles of isotope incorporation into collagens and collagen precursors, it was evident that collagen synthesis and processing was rapid. Processing of type I collagen and probably also alpha 1 (I) trimer proceeded almost entirely through procollagen intermediates. Rapid maturation of the types I and V collagens in the salt-insoluble fraction and the appearance of beta and gamma chains as early as 30 min after isotope administration. Radiolabeled procollagens were also extracted with acetic acid and penicillamine, indicating that cross-linking of collagen precursors may be involved in fiber formation.

[14C]Proline-labeled protocollagen, the unhydroxylated form of procollagen, was isolated from cartilage cells incubated with alpha, alpha'-dipyridyl. For examination of the initial steps in the hydroxylation of the protein, it was incubated in vitro with prolyl hydroxylase so that an average of 1.3-2.7 prolyl residues per chain was hydroxylated. The partially hydroxylated alpha chain were cleaved with cyanogen bromide, and the fragments were separated by polyacrylamide gel electrophoresis or column chromatography. The cyanogen bromide fragments were hydroxylated to the same degree. The results indicated, therefore, that in the initial hydroxylation of alpha chains in vitro, there was no preferential hydroxylation of any specific regions of the protein. In a second series of experiments, cartilage cells were incubated with [14C]proline and alpha, alpha'-dipyridyl so that prolyl hydroxylase in the cells was extensively, but not completely, inhibited. Partially hydroxylated alpha chains were isolated, and cyanogen bromide fragments of the alpha chains from the cells were assayed for hydroxy[14C]proline. The alpha chains contained an average of two residues of hydroxyproline per chain, and the cyanogen bromide fragments were hydroxylated to about the same degree. The results indicated, therefore, that when prolyl hydroxylase activity in cells is low relative to the rate at which pro alpha chains are synthesized, hydroxylation of prolyl residues occurs as it does in vitro, and there is no preferential hydroxylation of a specific region of the protein.

Procollagen type III amino-terminal protease was detected in cultures of smooth muscle cells of fetal calf aorta, and this protease was purified about 400-fold. Only about half of the enzyme activity was consistently attached to concanavalin A-agarose (Con A-agarose). After affinity chromatography on type III pN-collagen-Sepharose, the Con A bound fraction showed only one major band with a molecular weight of about 72000, this value corresponding well with the elution position of enzyme activity in gel filtration. The enzyme did not cleave procollagens type I or type IV, and denatured type III pN-collagen also remained uncleaved. The Km of the enzyme activity for iodo[14C]acetamide-labeled type III pN-collagen was 0.76 microM. Neutral pH and Ca2+ wer required for maximal enzymic activity. The metal chelators ethylenediaminetetraacetic acid and ethylene glycol bis(beta-amino-ethyl ether)-N,N,N',N'-tetraacetic acid inhibited the activity as well as did dithiothreitol, but there was only little if any inhibition by several other proteinase inhibitors tested.

Human fibroblasts when induced to make nonhelical , defective collagen have mechanisms for degrading up to 30% of their newly synthesized collagen intracellularly prior to secretion. To determine if at least a portion of the degradation of defective collagen occurs by lysosomes, extracts of cultured HFL-1 fibroblasts were examined for proteinases capable of degrading denatured type I [3H]procollagen. The majority of the proteolytic activity against denatured [3H]-procollagen had a pH optimum of 3.5-4; it was stimulated by dithiothreitol and inhibited 95% by leupeptin, 10% by pepstatin, and 98% by leupeptin and pepstatin together. Extracts of purified lysosomes from the fibroblasts were active in degrading denatured [3H]procollagen and were completely inhibited by leupeptin and pepstatin. To demonstrate directly that human lung fibroblasts can translocate a portion of their defective collagen to lysosomes, cultured cells were incubated with cis-4-hydroxyproline and labeled with [14C]proline to cause the cells to make nonhelical [14C]procollagen. About 3% of the total intracellular hydroxy[14C]proline was found in lysosomes. If, however, the cells were also treated with NH4Cl, an inhibitor of lysosomal function, 18% of the intracellular hydroxy[14C]proline was found in lysosomes. These results demonstrate that cultured human lung fibroblasts induced to make defective collagen are capable of shunting a portion of such collagen to their lysosomes for intracellular degradation.

The number of peptide binding sites of prolyl 4-hydroxylase was manipulated with the peptide photoaffinity label N-(4-azido-2-nitrophenyl)glycyl-(Pro-Pro-Gly)5, and the effect on hydroxylation of the relatively short peptide substrate (Pro-Pro-Gly)5 and of the long natural substrate procollagen was studied. With (Pro-Pro-Gly)5 as a substrate, a linear relation was found between enzyme activity and the amount of covalently bound photoaffinity label, approximately 50% inactivation being reached at 1 mol of label/mol of enzyme. No difference in Km value for (Pro-Pro-Gly)5 was detected between unlabeled and partially labeled enzyme preparations. These results indicate that enzyme molecules with only one free active site hydroxylated the synthetic substrate (Pro-Pro-Gly)5 with the same Km and at half the rate of native enzyme. In contrast, with procollagen as a substrate a 5-10-fold increase in Km was found with the fraction of enzyme containing only one free active site, as compared to the Km for procollagen with nonlabeled enzyme. This finding is explained by an enzyme-kinetic model based on a processive action of the two peptide substrate binding sites of prolyl 4-hydroxylase, preventing dissociation of the enzyme-substrate complex between successive hydroxylations of a long peptide with multiple substrate sites. Such a mechanism leads to a low Km for a long peptide by overcoming the diffusional constraints on the rate of association between the enzyme and the individual substrate sites.

The conversion of type I procollagen to type I collagen was studied by cleaving the protein with partically purified type I procollagen N-proteinase from chick embryos. Examination of the reaction products after incubation for varying times at 30 degrees C indicated that, during the initial stages of the reaction, pro alpha 1(I) and pro alpha 2(I) chains were cleaved at about the same rate. As a result, all the pro alpha 2(I) chains were converted to pC alpha 2(I) chains well before all the pro alpha 1 chains were cleaved. When the reaction products were examined by gel electrophoresis without reduction of interchain disulfide bonds, a distinct band of an intermediate was detected. The same intermediate was seen when the reaction was carried out at 35, 37, and 40 degrees C. The data established that over two-thirds of the type I procollagen was converted to the intermediate and that this intermediate was then slowly converted to the final product of pCcollagen. The kinetics for the reaction, however, did not fit a simple model for precursor-product relationship among substrate, intermediate, and product. Examination of the reaction products with a two-step gel procedure demonstrated that the intermediate consisted of three polypeptide chains in which the N propeptide was cleaved from one pro alpha 1 chain and one pro alpha 2(I) chain but the N propeptide was still present on one of the pro alpha 1(I) chains. In further experiments it was demonstrated that a similar intermediate was seen when a homotrimer of pro alpha 1(I) chains was partially cleaved by the enzyme.(ABSTRACT TRUNCATED AT 250 WORDS)

Leukocyte-derived proteases may contribute to the destruction of basement membranes during inflammation. We have, therefore, examined the degradation of human type IV procollagen (PC) by purified human neutrophil elastase (HLE). Native [14C]proline-labeled type IV PC was isolated from cultures of human HT-1080 cells and incubated with HLE for various times at 25 or 37 degrees C. Cleavage products were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and identified by CNBr peptide mapping. Incubation of type IV PC with HLE (less than 1:10 HLE:type IV weight ratio) resulted in cleavage of the pro alpha 1 (IV) and pro alpha 2 chains (Mr 180,000 and 175,000) to discrete components of Mr greater than 140,000. Peptide mapping indicated that the carboxy-terminal collagenase-resistant domains of both chains were rapidly and preferentially degraded. Longer incubations or incubations at higher enzyme:substrate ratios resulted in extensive and asymmetric internal cleavage with the generation of fragments similar in size distribution to the major pepsin-resistant fragments of type IV collagen. Our findings indicate that soluble, native human type IV PC is a substrate for HLE and is preferentially cleaved within the globular carboxy-terminal domains of the pro alpha 1 and pro alpha 2 chains. We suggest that even limited cleavage of type IV PC by HLE may disrupt intermolecular carboxy-terminal interactions believed to be important for basement membrane assembly and for maintaining basement membrane structure in vivo.

Derivatives of ferrocene (dicyclopentadienyliron) (Fc) were examined as active site directed inhibitors of type I procollagen N-proteinase, the enzyme that cleaves the NH2-terminal propeptides from type I procollagen. The compounds were shown here to be reversible, competitive inhibitors of the enzyme. The effectiveness of the Fc inhibitors varied with modification of the cyclopentadienyl (cp) rings. The monocarboxylic acid (I) and the 1,1'-dicarboxylic acid (II) derivatives of Fc inhibited 50% of the enzymic activity (I50) at concentrations of 1.0 and 0.5 mM, respectively. The Ki values were 0.3 mM for both I and II. Derivatization of the carbonyl alpha to the cp ring of compound I (FcCOCH2CH2COOH, III) increased the inhibitory activity (I50 = 0.100 mM; Ki = 0.065 mM). Removal of the carbonyl alpha to the cp ring of III did not improve inhibitory activity: FcCH2CH2COOH, I50 = 2 mM; FcCH = CHCOOH, I50 = 1.5 mM. The active inhibitory species apparently contained iron in the 3+ valence state since two ferrocenium derivatives were very effective inhibitors: ferrocenium tetrachloroferrate, IV (I50 = 0.030 mM; Ki = 0.004 mM), and carboxyferrocenium hexafluorophosphate, V (I50 less than 0.1 mM; Ki less than 0.05 mM). In addition, reduction of III with ascorbic acid abolished its inhibitory activity. Compounds I and III stabilized the enzyme to heat denaturation in the absence of exogenous calcium; compound IV did not stabilize the enzyme. Further observations indicated that Fc derivatives were specific inhibitors of procollagen N-proteinase.(ABSTRACT TRUNCATED AT 250 WORDS)

Type I/II procollagen N-proteinase was partially purified from chick embryos and used to examine the rate of cleavage of a series of purified type I procollagens synthesized by fibroblasts from probands with heritable disorders of connective tissue. The rate of cleavage was normal with procollagen from a proband with osteogenesis imperfecta that was overmodified by posttranslational enzymes. Therefore, posttranslational overmodification of the protein does not in itself alter the rate of cleavage under the conditions of the assay employed. Cleavage of the procollagen, however, was altered in several procollagens with known mutations in primary structure. Two of the procollagens had in-frame deletions of 18 amino acids encoded by exons 11 and 33 of the pro alpha 2(I) gene. In both procollagens, both the pro alpha 1(I) and the pro alpha 2(I) chains were totally resistant to cleavage. With a procollagen in which glycine-907 of the alpha 2(I) chain domain was substituted with aspartate, both pro alpha chains were cleaved but at a markedly decreased rate. The results, therefore, establish that mutations that alter the primary structure of the pro alpha chains of procollagen at sites far removed from the N-proteinase cleavage site can make the protein resistant to cleavage by the enzyme. The long-range effects of in-frame deletions or other changes in amino acid sequence are probably explained by their disruption of the hairpin structure that is formed by each of the three pro alpha chains in the region containing the cleavage site and that is essential for cleavage of the procollagen molecule by N-proteinase.