How Does a Skin Fibroblast Make Type I Collagen Fibers?

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

022-202X/82/7903-003s$02.00/0 THE JOURNAL OF INVESTIGATIVE DERMATOLOGY, 79:38-68, 1982 Vol. 79, Supplement I Copyright © 1982 by The Williams & Wilkins Co. Printed in U.S.A.

REPORTS

How Does a Skin Fibroblast Make Type I Collagen Fibers?

DARWIN J. PROCKOP, M.D., PH.D.

Department of Biochemistry, College of Medicine and Dentistry, of New Jersey-Rutgers Medical School, Piscataway, New Jersey, U.S.A.

We now have an impressive amount of information proa2(I) chain [9]. By using a :l2P-Iabeled cloned cDNA and the about several steps in the synthesis of type I collagen technique known as "Southern blotting," it was Rhown that fibers by skin fibroblasts: (a) the transcription of the DNA fragments arising from the human proa2(I) gene were genes for proal(I) and proa2(I) polypeptide chains; (b) obtained only from mouse-human cell hybrids and hamster­ the processing of the initial RNA transcripts of the genes human cell hybrids which contained the human chromosome to mature mRNAs; (c) translation of the mRNAs to pre­ #7. proal(I) and preproa2(I) chains; (d) processing of the In the second step, the genes for proal (I) and proa2(I) chains preproal(I) and preproa2(I) chains to procollagen; and are transcribed into RNA. Experiments carried out first with a (e) the further processing of procollagen to collagen proa2(I) gene from chick embryos [10-12], then with a proa2(I) fibers. The speed with which we are currently develop­ gene from sheep [13] and finally a pro(�2(l) gene from man ing this information about collagen and comparable in­ [14,15], demonstrated that both the gene and the initial RNA formation about other constituents of skin suggests that transcript of the collagen gene are about 40,000 bases, or eight skin may well be the first human organ which we fully times longer than the nucleotides necessary to code for the understand. protein. The large size of the gene and the initial RNA transcript is accounted for by the presence of about 50 separate interven­ ing sequences (Fig 1). Most of the intervening sequences divide As the initial speaker at this conference, I thought it might the central portion of the gene, the part which codes for the a be helpful to set the stage by reviewing the information we now chain domain, into short coding sequences or exons of 54 or 108 have about the question, "How does a fibroblast make type I bases. The 54 or 108 base exons code for 18 or 36 amino acid collagen, the most abundant constituent of skin?" (Table). sequences with the structure (Gly-X-Y)6 or (Gly-X-Yh2 (10-12). The synthesis of any extracellular protein occurs in at least Occasionally the exons are 49 and 95 bases [12]. These obser­ 6 steps: (a) The structural gene for the protein is "turned on" vations have led to the suggestion that the collagen gene may in the sense of being prepared for transcription; (b) the DNA of have originated by extensive duplication of a primitive collagen the structural gene is transcribed into an initial RNA transcript; of 54 bases [10-12]. Alternatively, it is possible that the large (c) the initial RNA transcript is processed by a series of modi­ number of intervening sequences arose during evolution to fications to a mature mRNA; (d) the mRNA is translated into prevent recombination events (crossover mutations), which fre­ a polypeptide chain; (e) the polypeptide chain is processed by quently occur in highly repetitive DNA sequences. one or more modifying enzymes; (f) the polypeptide chain folds In the third step of collagen synthesis, the hnRNAs are into the correct three-dimensional conformation and is further processed to functional mRNAs. Because there are 50 interven­ moditied so that its biological activity is expressed. ing sequences, over 50 enzymic steps are required. In fact, In considering the synthesis of type I collagen (for more examination of one part of this process has suggested that the detailed reviews, see references 1-7), we know relatively little excision of a single large intervening sequence may occur in about the first step, the "turning on" of the structural genes for more than one step [16]. The RNA must also be processed proal (I) and proa2(l) polypeptide chains which comprise the through steps that add a "cap" containing 7-methylguanosine molecule. In fact, although we have considerable information to its 5' -end, and a stretch of 20-100 adenine nucleotides to the about regulation of the operon and similar gene systems in lac :1'-end, much as occurs with most other messengers of eukar­ E. coli, we have almost no information about the regulatory yotic cells. Recent observations on several types of protein mechanisms which turn on genes in eukaryotic cells. What we mRNAs have indicated that the length of the untranslated know about collagen genes is that most cells, particularly cells nucleotides at the 3' -end of the mRNAs for a given protein may in culture, either synthesize collagen or contain some of the vary [17]. Therefore, transcription of a single gene can give rise necessary components for collagen synthesis such as active to 2 or more different length mRNAs coding for the same prolylhydroxylase [8]. We also know that the gene system for protein. In the case of type I procollagen, mRNAs of several collagen synthesis is among the oldest gene systems in the different sizes have been identified for the proa2(I) chain [15]. multi-cellular organisms, and it dates as far back in evolution The significance of this variability in the size of the mRNAs as the origin of sponges [2]. Therefore, it seems that synthesis coding for a given protein is unclear. of extracellular collagen fibers is a major mechanism which In the fourth step, the mRNAs are translated into preproal(I) nature has developed for holding together cells to form a multi­ and preproa2(I) chains. As appears to be true of all protein cellular organism and is an intrinsic property of many cell types. synthesized for export from cells, the amino terminal ends of the We have one further point of information about collagen at the newly-synthesized polypeptides contain a stretch of hydropho­ level of regulation of the gene in that we know the chromosomal bic amino acids which direct binding of the ribosome to the location of the proa2(I) chain. The proa2(I) gene was assigned membranes of the rough endoplasmic reticulum. This region of to chromosome #7 by use of a cloned cDNA for the human the protein, called a "signal sequence" or "pre-sequence," is rapidly cleaved after the newly synthesized polypeptide chain This work was supported by in part by Grant AM-16516 from the National Institutes of Health. passes through the membrane of the rough endoplasmic retic­ Reprint requests to: Darwin J. Prockop, M.D., Ph.D., Department ulum into the cisternae of the organelle. In the case of collagen of Biochemistry, College of Medicine and Dentistry of New .Jersey, synthesis, both the proal(I) [18,19] and the proa2(I) [19] chains Rutgers Medical School, Piscataway, NJ 08854. have been shown to have signal peptides, and the signal pep-

3s 4s PROCKOP Vol. 79, Supplement 1

How does a skin fibroblast make type Collagen? In the sixth step, the secreted procollagen is converted into I ----- collagen fibers (Fig 3). Here at least 2 separate procollagen 1. Genes for prOa I (I) and proa2 (I) are "turned on." 2. DNA is transcribed into hnRNAs. proteinases are involved [20-25]. One, procollagen amino-pro­ -about 50 intervening sequences. teinase, removes the amino-propeptides from the 3 proa chains. -hnRNA about 8 times longer than coding sequences. Another procollagen carboxy-proteinase, removes the carboxy­ 3. hnRNAs are processed to mRNAs. propeptides. The amino- and carboxy-propeptides constitute -over 50 steps each for proal (I) and proa2 (I) about one-third of the procollagen molecule. With the removal -2 size mRNAs for proal (I); proa2 3 for (I) of the propeptides, the biological activity of the protein is 4. mRNAs are translated into preproal and preproa2 chains. (1) (I) released and the resulting collagen self-assembles into fibers 5. Preproal (I) and preproa2 (ll chains are processed to procoUagen. -over 11 enzymes. [2;3). Finally, in steps not discussed here, the collagen molecules -over 120 steps. which have self-assembled into fibers are cross-linked by a 6. Procollagen is converted to collagen fibers. series of enzymic and nonenzymatic steps so that large sections -at least 3 enzymes of the fiber become a single, covalently-linked structure. -one-third of the protein discarded At least 3 important generalizations emerge from this infor­ -the collagen self-assembles into fibers mation about the synthesis of type I collagen fibers by skin -the fibers are cross-linked. fibroblasts. One of these generalizations is that the structural and functional relationships of the collagen molecule are re­ markably simple. Each of the 3 a chains contain a large number DNA of repeating sequences of -Gly-Pro-Hyp-. These triplets in the polypeptide chains direct folding of the chains into the unique rope-like, triple-helical conformation which is characteristic of Transcription of 40,000 bose gene and addition of the collagen molecule. The remainder of the a chains consist of poly A 01 3' end. j triplets of -Gly-X-Y- in which the side chains of the amino acids in the X- and Y-positions point away from the center of the triple-helix and thereby provide precisely positioned reactive RNA ------(A)n groups that direct lateral and longitudinal assembly of the molecules into fibers. The second generalization is that the structure of the gene is remarkably complex because of the large number of intervening sequences. As indicated above, it is not clear as yet whether this complex structure reflects the evolution of the gene from a simple 54-base unit or whether it j reflects an arrangement which is necessary to prevent muta­ tions in the gene. A third generalization is that intracellular Splicing out of 50 inlrons processing of the protein after the mRN A is translated is (AI, of varying size. remarkably complex. There appear to be 2 reasons for this � complexity. One is that it is probably necessary to tailor the hydroxy/proline content of the polypeptide chains so that the triple helix is stable but not overly rigid. The second reason for j the complexity of the processing is probably that the protein must be synthesized as a precursor which does not form fibrils spontaneously. Otherwise, fibril formation might occur prema­ m RNA f ab ut 6000 ------o o (Aln turely and before the protein is either secreted or before it bases used for translation. reaches the appropriate extracellular site for self-assembly. FIG 1. Scheme for the transcription of a gene for a proa chain of collagen and processing of the initial RNA transcript to a mature mRNA. (Reproduced with permission [26]).

tides on the proal chain differ in amino acid sequences from those of the proa2 chain [19]. In the fifth step, as defined here, the preproal(I) and pre­ proa2(l) chains are processed to proa chains by removal of the signal peptides, and then the pro a chains are processed into triple helical procollagen (Fig 2). The processing of the pro a chains involves at least 8 separate enzymes (see references 1 and 3). Three of these enzymes are hydroxylases which synthesize the 4-hydroxyproline, 3-hydroxyproline, and hydroxylysine found in collagen. Together these enzymes are involved in about 120 enzymic reactions in which 100 prolyl residues are hydroxylated to hydroxyproline, and about 20 lysyl residues are hydroxylated in hydroxylysine. In addition, 2 other enzymes add sugars to the hydroxylysyl residues and a separate system of enzymes, involving dolichol intermediates, adds sugars to the carboxy-terminal propeptide. The processing on the proa: chain also involves association of the chains through the large car­ boxy-terminal propeptides, formation of interchain disulfide links among the propeptides, and folding of the three a chain domains into the unique, triple-helical conformation found in FIG 2. Scheme for the intracellular biosynthesis of procollagen. He­ collagen. After these processing steps are completed, the pro­ produced with permission from the New England Journal of Medicine collagen is secreted. [4]. July 1982 TYPE I COLLAGEN FIBERS 5s

COLLAGEN MOLECULE

•

�: = I I I r� �(;!\L� L6:SL6:SL�

PACKING OF �

� i�I I t L�CYZ � :� = I Overlap zone 0.4 D--I--...J-, Hole zone 0.6D------'---. •

FIG :l. Scheme for the extra�ellular assembly of collagen fibers. Reproduced with permission from the New England Journal of lvIedicine [41·

To date, the final step in the synthesis of collagen fibers has controlled manner and under physiological conditions of tem­ been among the most difficult to study. The problems here, perature and pH. When relatively small amounts of enzyme are however, may well be soluble with the development of a new in employed in the system, the removal of the propeptide occurs vitro system in which purified procollagen is treated with a slowly over one or two days. The fibers formed are more regular, mixture of partially purified procollagen amino- and carboxy­ and five to seven times thicker than any type I collagen fibers proteinases [23]. As the propeptides are cleaved from the pro­ seen in vivo. In addition, branch points are frequently seen in tein, the resulting collagen assembles into collagen fibers in a the fibers. Because the components are relatively pure and the PROCKOP 6s Vol. 79, Supplement 1 conditions for fiber formation are relatively controllable, the demic Press, in press system appears to provide a means of examining the process of 9. Junien C, Weil D, Myers JC, Cong MV, Chu M-L, Foubert C, Gross M-S, Prockop DJ, Kaplan J-C, Ramirez F: Assignment of the fibril formation in considerable detail. human proa2(l) collagen structural gene to chromosome 7 by The large amount of information we now have about the molecular hybridization. Am J Hum Genetics, in press processes by which fibroblasts, and other collagen synthesizing 10. Yamada Y, Avvedimento VE, Mudryj M, Ohkubo H, Vogeli G, cells, produce type collagen fibers has several consequences. Irani M, Pastan I, de Crombrugghe B: The collagen gene contains I evidence for its evolutionary assembly by duplication of a DNA One is that the information provides a basis with which to segment containing an exon on 54 bp. Cell 22:887-892, 1980 explore important biological questions such as the factors which 11. Wozney J, Hanahan D, Morimoto R, Boedtker H, Doty P: Fine regulate the growth of development of connective tissues. An­ structural analysis of the chick proa2 collagen gene. Proc Nat! other consequence is that the information provides a basis for Acad Sci 78:712-716, 1981 12. Wozney J, Hanahan D, Tate V, Boedtker H, Doty P: Structure of determining the molecular basis of a broad range of human the proa2(l) collagen gene. Nature 294:129-135, 1981 diseases. As will certainly become apparent at this meeting, 13. Boyd CD, Toistoshev P, Schafer M, Trapnell BC, Coon HC, genetic diseases involving collagen, such as Ehlers-Danlos syn­ Kretchmer PJ, Nienhuis AW, Crystal RG: Isolation and charac­ drome, have now become the target for highly productive terization of a 15 kb genomic sequence coding for part of the proa2 chain of sheep type collagen. Bioi Chern 255:3212-3220, research. Acquired diseases involving collagen are still relatively I J 1980 intractable. However, the basic information we now have about 14. Myers JC, Sangiorgi FO, Chu M-L, Ding G, Prockop DJ, Ramirez the structure and biosynthesis of collagen, together with ex­ F: Isolation and characterization of part of the human proa2 tremely powerful new tools provided by recombinant DNA, collagen gene. Federa Proc 40:1650, 1981 15. Myers JC, Chu M-L, Faro SH, Clark WJ, Prockop DJ, Ramirez F: offer new hope that we will be able to resolve many intriguing Cloning a cDNA for the proa2 gene of human type I collagen. questions, such as why collagen synthesis is increased in scle­ Proc Nat! Acad Sci 78:3516-3520, 1981 roderma. 16. Avvedimento VE, Vogeli G, Yamada Y, Maizel JV Jr, Pastan I, de Finally, it should be noted that the information we have Crombrugghe B: Correlation between splicing sites within an intron and their sequence complementarity with RNA. Cell about how a skin fibroblast synthesizes type I collagen fibers Ul 21:689-696, 1980 does not stand alone. We are also rapidly developing detailed, 17. Hagenbiichle 0, Tosi M, Schibler U, Bovey R, Wellauer PK, Young definitive information about proteoglycans, glycoproteins and RA: Mouse liver and salivary gland a-amylase mRNAs differ essentially all the other components of skin. The speed with only in 5' non-translated sequences. Nature 289:643-646, 1981 which this information is being acquired suggests, to me at 18. Palmiter RD, Davidson JM, Gagnon J, Rowe DW, Bornstein P: NH"-terminal sequence of the chick proal(l) chain synthesized least, that skin may well be the first tissue or organ of man in the reticulocyte lysate system. J Bioi Chern 254:1433-1436, which we fully understand. 1979 19. Graves PN, Olsen BR, Fietzek PP, Prockop DJ, Monson JM: Comparison of the NH"-terminal sequence of chick type I pre­ REFERENCES procollagen chains synthesized in an mRNA-dependent reticu­ locyte lysate. Eur J Biochem, 118:363-369, 1981 1. Kivirikko KI: Post-translational modifications of collagen, Gene 20. Tuderman L, Kivirikko K, Prockop DJ: Partial purification and Families of Collagen and Other Proteins. Edited by DJ Prockop, characterization of a neutral protease which cleaves the N-ter­ PC Champe. North-Holland, Elsevier, New York, 1980, pp minal propeptides from pro collagen. Biochemistry 17:2948-2954, 107-120 1978 2. Gross J: Some aspects of the biology of the extracellular matrix, 21. Lueng MKK, Fessler LI, Greenberg DB, Fessler JH: Separate Gene Families of Collagen and Other Proteins. Edited by DJ amino and carboxyl pro collagen peptidases in chick embryo Prockop, PC Champe, North-Holland, New York, Elsevier, 1980, tendon. J Bioi Chern 254:224-232, 1979 pp 5-36 22. Njieha FK, Morikawa T, Tuderman L, Prockop DJ: Partial purifi­ :1. Piez KA: Structure and function of collagen, Gene Families of cation of a procollagen C-proteinase. Inhibition by synthetic Collagen and Other Proteins. Edited by DJ Prockop, PC Champe. peptides and sequential cleavage of type I procollagen. Biochem­ North-Holland, New York, Elsevier, 1980, pp 143-160 istry, 21:757-764,1982 4. Prockop DJ, Kivirikko KI, Tuderman L, Guzman NA: The biosyn­ 23. Miyahara M, Njieha FK, Prockop DJ: Formation of collagen fibers thesis of collagen. N Engl J Med 301:13-23, 77-85, 1979 in vivo by cleavage of procollagen with procollagen proteinases. 5. Bornstein P, Traub W: The chemistry and biology of collagen. The Effect of kinetics on fiber morphology. J Bioi Chern, in press Proteins IV:412-632, 1979 24. Nusgens BV, Goebels Y, Shinkai H, Lapiere CM: Procollagen type 6. Bornstein P, Sage H: Structurally distinct collagen types. Ann Rev III N-terminal endopeptidase in fibroblast culture. Biochem J Biochem 49:957-1004, 1980 191:699-706, 1980 7. Fessler JH, Fessler LI: Biosynthesis of procollagen. Ann Rev Bio- 25. Goldberg B, Taubman MB, Radin A: Procollagen peptidase: Its chern 47:129-162, 1978 , mode of action on the native substrate. Cell 4:45-50, 1975 8. Kivirikko KI, Myllyla R: The hydroxylation of prolyl and lysyl 26. Prockop DJ, Williams CJ: Structure of the organic matrix: Collagen residues, The Enzymology of Post-translational Modification of structure, Biological Mineralization and Demineralization. Ed­ Proteins. Edited by RB Freedman, H Hawkins, London, Aca- ited by GH Nancollas. Heidelberg, Springer-Verlag, 1982, in press