α-Helical Coiled-coil Oligomerization Domains Are Almost Ubiquitous in the Collagen Superfamily
2003; Elsevier BV; Volume: 278; Issue: 43 Linguagem: Inglês
10.1074/jbc.m302429200
ISSN1083-351X
AutoresAudrey McAlinden, Thomasin A. Smith, Linda J. Sandell, D. Ficheux, David Parry, David Hulmes,
Tópico(s)Wnt/β-catenin signaling in development and cancer
Resumoα-Helical coiled coils are widely occurring protein oligomerization motifs. Here we show that most members of the collagen superfamily contain short, repeating heptad sequences typical of coiled coils. Such sequences are found at the N-terminal ends of the C-propeptide domains in all fibrillar procollagens. When fused C-terminal to a reporter molecule containing a collagen-like sequence that does not spontaneously trimerize, the C-propeptide heptad repeats induced trimerization. C-terminal heptad repeats were also found in the oligomerization domains of the multiplexins (collagens XV and XVIII). N-terminal heptad repeats are known to drive trimerization in transmembrane collagens, whereas fibril-associated collagens with interrupted triple helices, as well as collagens VII, XIII, XXIII, and XXV, were found to contain heptad repeats between collagen domains. Finally, heptad repeats were found in the von Willebrand factor A domains known to be involved in trimerization of collagen VI, as well as in collagen VII. These observations suggest that coiled-coil oligomerization domains are widely used in the assembly of collagens and collagen-like proteins. α-Helical coiled coils are widely occurring protein oligomerization motifs. Here we show that most members of the collagen superfamily contain short, repeating heptad sequences typical of coiled coils. Such sequences are found at the N-terminal ends of the C-propeptide domains in all fibrillar procollagens. When fused C-terminal to a reporter molecule containing a collagen-like sequence that does not spontaneously trimerize, the C-propeptide heptad repeats induced trimerization. C-terminal heptad repeats were also found in the oligomerization domains of the multiplexins (collagens XV and XVIII). N-terminal heptad repeats are known to drive trimerization in transmembrane collagens, whereas fibril-associated collagens with interrupted triple helices, as well as collagens VII, XIII, XXIII, and XXV, were found to contain heptad repeats between collagen domains. Finally, heptad repeats were found in the von Willebrand factor A domains known to be involved in trimerization of collagen VI, as well as in collagen VII. These observations suggest that coiled-coil oligomerization domains are widely used in the assembly of collagens and collagen-like proteins. The mechanisms controlling chain oligomerization in extracellular matrix and related proteins are currently topics of active research (1Engel J. Kammerer R.A. Matrix Biol. 2000; 19: 283-288Crossref PubMed Scopus (36) Google Scholar, 2Frank S. Boudko S. Mizuno K. Schulthess T. Engel J. Bachinger H.P. J. Biol. Chem. 2003; 278: 7747-7750Abstract Full Text Full Text PDF PubMed Scopus (50) Google Scholar). In the case of the collagen superfamily, which includes both collagens and collagen-like proteins (3Myllyharju J. Kivirikko K.I. Ann. Med. 2001; 33: 7-21Crossref PubMed Scopus (568) Google Scholar, 4Kadler K.E. Protein Profile. 1995; 2: 491-619PubMed Google Scholar, 5Ricard-Blum S. Dublet B. van der Rest M. Unconventional Collagens. Oxford University Press, Oxford, UK2000Google Scholar), a number of structural features have been identified that are essential for bringing together component polypeptide chains with the correct stoichiometry, thus leading to folding of the triple helix. Early studies on the procollagen precursors of the fibrillar collagens (types I–III, V, and XI) (6Mclaughlin S.H. Bulleid N.J. Matrix Biol. 1998; 16: 369-377Crossref PubMed Scopus (100) Google Scholar) showed that the C-propeptide regions are necessary to direct correct chain association, which is followed by zipper-like folding of the triple helix in the C- to N-terminal direction (7Engel J. Prockop D.J. Annu. Rev. Biophys. Biophys. Chem. 1991; 20: 137-152Crossref PubMed Scopus (284) Google Scholar). This concept was subsequently extended to basement membrane collagen type IV (8Boutaud A. Borza D.B. Bondar O. Gunwar S. Netzer K.O. Singh N. Ninomiya Y. Sado Y. Noelken M.E. Hudson B.G. J. Biol. Chem. 2000; 275: 30716-30724Abstract Full Text Full Text PDF PubMed Scopus (183) Google Scholar, 9Than M.E. Henrich S. Huber R. Ries A. Mann K. Kuhn K. Timpl R. Bourenkov G.P. Bartunik H.D. Bode W. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 6607-6612Crossref PubMed Scopus (105) Google Scholar, 10Sundaramoorthy M. Meiyappan M. Todd P. Hudson B.G. J. Biol. Chem. 2002; 277: 31142-31153Abstract Full Text Full Text PDF PubMed Scopus (137) Google Scholar), microfibril-forming collagen VI (11Ball S.G. Baldock C. Kielty C.M. Shuttleworth C.A. J. Biol. Chem. 2001; 276: 7422-7430Abstract Full Text Full Text PDF PubMed Scopus (36) Google Scholar, 12Lamande S.R. Morgelin M. Selan C. Jobsis G.J. Baas F. Bateman J.F. J. Biol. Chem. 2002; 277: 1949-1956Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar), members of the C1q family (13Kishore U. Reid K.B. Immunopharmacology. 2000; 49: 159-170Crossref PubMed Scopus (405) Google Scholar) including collagens VIII and X (14Illidge C. Kielty C. Shuttleworth A. Int. J. Biochem. Cell Biol. 2001; 33: 521-529Crossref PubMed Scopus (25) Google Scholar, 15Marks D.S. Gregory C.A. Wallis G.A. Brass A. Kadler R.E. Boot-Handford R.P. J. Biol. Chem. 1999; 274: 3632-3641Abstract Full Text Full Text PDF PubMed Scopus (35) Google Scholar, 16Dublet B. Vernet T. van der Rest M. J. Biol. Chem. 1999; 274: 18909-18915Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar, 17Bogin O. Kvansakul M. Rom E. Singer J. Yayon A. Hohenester E. Structure. 2002; 10: 165-173Abstract Full Text Full Text PDF PubMed Scopus (97) Google Scholar) and the emilins (18Mongiat M. Mungiguerra G. Bot S. Mucignat M.T. Giacomello E. Doliana R. Colombatti A. J. Biol. Chem. 2000; 275: 25471-25480Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar), and the FACITs 1The abbreviations used are: FACITs, fibril-associated collagens with interrupted triple helices; CHO, Chinese hamster ovary; CPII, C-propeptide region of procollagen II; CRD, carbohydrate recognition domain; SP-D, surfactant protein-D; BS3, bis-(sulfosuccinimidyl)suberate; MBP, mannan-binding protein; TFE, trifluoroethanol. (19Mazzorana M. Cogne S. Goldschmidt D. Aubert-Foucher E. J. Biol. Chem. 2001; 276: 27989-27998Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar, 20Mechling D.E. Gambee J.E. Morris N.P. Sakai L.Y. Keene D.R. Mayne R. Bachinger H.P. J. Biol. Chem. 1996; 271: 13781-13785Abstract Full Text Full Text PDF PubMed Scopus (30) Google Scholar, 21Lesage A. Penin F. Geourjon C. Marion D. van der Rest M. Biochemistry. 1996; 35: 9647-9660Crossref PubMed Scopus (33) Google Scholar). α-Helical coiled coils have been shown to be oligomerization domains in both collagens and collagen-like proteins. The paradigm here is the collectin family (22Hakansson K. Reid K.B. Protein Sci. 2000; 9: 1607-1617Crossref PubMed Scopus (105) Google Scholar), which includes the lung surfactant proteins (SP-A and SP-D), mannan-binding proteins (MBP-A and MBP-C), conglutinin, and collectin-43. Collectins are trimeric molecules in which each chain contains a collagen-like domain followed by an α-helical coiled-coil region and then a carbohydrate recognition domain (CRD). The amino acid sequence of the coiled-coil domain is characterized by up to four heptad repeats (a-b-c-d-e-f-g) in which hydrophobic amino acid residues occur at positions a and d (23Kammerer R.A. Matrix Biol. 1997; 15: 555-565Crossref PubMed Scopus (68) Google Scholar). Three-dimensional structures of the coiled-coil and CRD regions in MBP and SP-D show that chains within the trimer are associated via a three-stranded coiled coil, with the three CRDs arranged as distinct lobes (24Sheriff S. Chang C.Y. Ezekowitz R.A. Nat. Struct. Biol. 1994; 1: 789-794Crossref PubMed Scopus (219) Google Scholar, 25Weis W.I. Drickamer K. Structure. 1994; 2: 1227-1240Abstract Full Text Full Text PDF PubMed Scopus (297) Google Scholar, 26Hakansson K. Lim N.K. Hoppe H.J. Reid K.B. Structure Fold. Des. 1999; 7: 255-264Abstract Full Text Full Text PDF Scopus (127) Google Scholar). Several studies have shown that the coiled-coil region is both necessary and sufficient for trimerization of SP-D (27Hoppe H.J. Barlow P.N. Reid K.B.M. FEBS Lett. 1994; 344: 191-195Crossref PubMed Scopus (114) Google Scholar, 28Zhang P. McAlinden A. Li S. Schumacher T. Wang H. Hu S. Sandell L. Crouch E. J. Biol. Chem. 2001; 276: 19862-19870Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar). In a recent study (29McAlinden A. Crouch E. Bann J.G. Zhang P. Sandell L. J. Biol. Chem. 2002; 277: 41274-41281Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar), the first two heptad repeats of SP-D were shown to be sufficient to drive trimerization and folding when fused C-terminal to the N-propeptide region of procollagen IIA, which also contains a collagen triple helix but which alone does not spontaneously trimerize. Coiled-coil oligomerization domains have also been shown to be essential for trimerization and folding of the membrane-associated collagens XIII (30Snellman A. Tu H.M. Visnen T. Kvist A.P. Huhtala P. Pihlajaniemi T. EMBO J. 2000; 19: 5051-5059Crossref PubMed Scopus (81) Google Scholar, 31Latvanlehto A. Snellman A. Tu H. Pihlajaniemi T. J. Biol. Chem. 2003; 278: 37590-37599Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar) and XVII (32Areida S.K. Reinhardt D.P. Muller P.K. Fietzek P.P. Kwitz J. Marinkovich M.P. Notbohm H. J. Biol. Chem. 2001; 276: 1594-1601Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). Similar domains occur in collagen XXIII (33Banyard J. Bao L. Zetter B.R. J. Biol. Chem. 2003; 278: 20989-20994Abstract Full Text Full Text PDF PubMed Scopus (77) Google Scholar), the Alzheimer amyloid plaque component precursor CLAC-P/collagen XXV (34Hashimoto T. Wakabayashi T. Watanabe A. Kowa H. Hosoda R. Nakamura A. Kanazawa I. Arai T. Takio K. Mann D.M. Iwatsubo T. EMBO J. 2002; 21: 1524-1534Crossref PubMed Scopus (179) Google Scholar) and the collagen-like membrane proteins MARCO and ectodysplasin-A1 (31Latvanlehto A. Snellman A. Tu H. Pihlajaniemi T. J. Biol. Chem. 2003; 278: 37590-37599Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar). In all of these proteins, the transmembrane domain is found close to the N terminus and is followed immediately by a heptad repeat region at the beginning of the large ectodomain. Once trimerized via the coiled-coil domains, folding of the C-terminal collagen helix then proceeds in the N- to C-terminal direction. N-terminal coiled-coil oligomerization domains are thought to play similar roles in the assembly of macrophage scavenger receptor proteins (35Frank S. Lustig A. Schulthess T. Engel J. Kammerer R.A. J. Biol. Chem. 2000; 275: 11672-11677Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar), which also contain collagen-like regions. Several coiled-coil domains, as well as collagen-like regions, are also present in the emilins (36Colombatti A. Doliana E. Bot S. Canton A. Mongiat M. Mungiguerra G. Paron-Cilli S. Spessotto P. Matrix Biol. 2000; 19: 289-301Crossref PubMed Scopus (94) Google Scholar). Here we show that putative α-helical coiled-coil oligomerization domains are present in most members of the collagen superfamily, located either before, after, or between collagen-like regions, suggesting a general role in triple-helical assembly. Sequence Analysis—All sequences were obtained from Swiss-Prot, SPTREMBL, or REMTREMBL data bases. Potential coiled-coil sequences in all polypeptide chains of the collagen superfamily were searched initially using the program COILS2 (37Lupas A. Methods Enzymol. 1996; 266: 513-525Crossref PubMed Google Scholar). Since the reliability of this computational approach is very limited, particularly for short (14-residue) sequences, all "hits" (regardless of score) were also examined by informed eye in order to assess the likelihood of coiled-coil formation. Experience has shown that this approach can more easily recognize heptad discontinuities and individual or small groups of residues that are generally disruptive to coiled-coil formation, hence allowing elimination of some regions from further consideration. Although sequences were analyzed for their potential to form either two- or three-stranded helices using SCORER (38Woolfson D.N. Alber T. Protein Sci. 1995; 4: 1596-1607Crossref PubMed Scopus (201) Google Scholar), their very short lengths (2–3 heptads in general) rendered the results statistically uncertain. The presence of three-chain structures in the proteins reported here would nonetheless strongly favor the formation of three-stranded rather two-stranded coiled coils. To test the plausibility of a three-stranded coiled-coil geometry for the heptad containing region of type I procollagen C-propeptides, a model three-stranded coiled coil was built based on the homotrimeric GCN4-pII leucine zipper structure (Protein Data Bank code 1gcm). Appropriate residue replacements were made, placing the pro-α1 chains onto chains A and B and the pro-α2 chain onto chain C of GCN4-pII, in such a way as to ensure the heptad repeat patterns were in phase. The packing of the core residues was then checked manually using Turbo-Frodo 5.5 (afmb.cnrs-mrs.fr/TURBO_FRODO), noting particularly the positions of the a and d residues relative to those in the GCN4-pII structure. Automated energy minimization was done in a multistep process using conjugate gradients followed by simulated annealing with CNSsolve 0.5 (39Brunger A.T. Adams P.D. Clore G.M. DeLano W.L. Gros P. Grosse-Kunstleve R.W. Jiang J.S. Kuszewski J. Nilges M. Pannu N.S. Read R.J. Rice L.M. Simonson T. Warren G.L. Acta Crystallogr. D Biol. Crystallogr. 1998; 54: 905-921Crossref PubMed Scopus (17024) Google Scholar). In the conjugate gradient minimization, the main chain atoms were first held fixed followed by harmonically restraining them to their initial positions with an energy constant of 10 kcal/mol·A2. Annealing using torsion dynamics was employed with the main chain atoms again harmonically restrained. Synthesis of Chimeric IIA N-pro/CPII cDNA Fusion Constructs— cDNA fusion constructs encoding the full-length human type IIA procollagen N-propeptide linked to 18, 25, or 31 amino acids of the coiled-coil domain present at the beginning of the type II procollagen C-propeptide (IIA N-pro/CPII; Fig. 3) were synthesized by overlap extension PCR. Oligonucleotide primers used for PCR and overlap extension PCR procedures were from Invitrogen. One PCR was carried out to amplify type IIA N-propeptide cDNA using primers 1 and 2 (Table I) for 30 cycles (95 °C, 30 s; 55 °C, 30 s; 72 °C, 50 s). Here IIA N-pro/SP-D cDNA in pGEM3Z was used as a substrate (28Zhang P. McAlinden A. Li S. Schumacher T. Wang H. Hu S. Sandell L. Crouch E. J. Biol. Chem. 2001; 276: 19862-19870Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar). Another set of PCRs was done to amplify cDNA encoding the 18-, 25-, or 31-amino acid sequence of the C-propeptide (plus the terminal alanine residue of the telopeptide at the procollagen C-proteinase cleavage site) corresponding to amino acid numbers 1172–1189, 1172–1196, or 1172–1202, respectively (GenBank™ accession number X16468). Table I shows the sequence of the PCR primer pairs used to amplify the cDNA fragments encoding each of the three C-propeptide regions. Each of these PCRs was carried out for 30 cycles (95 °C, 30 s; 55 °C, 30 s; 72 °C, 25 s), and oligonucleotides encoding amino acid residues 1172–1189, 1172–1196, or 1172–1202 of the C-propeptide were used as substrates. Overlap extension PCRs were then prepared by mixing cDNA encoding the IIA N-propeptide with cDNA encoding one of the three C-propeptide cDNA fragments. To promote overlap extension, a preliminary PCR was carried out for 5 cycles in the absence of primers (95 °C, 30 s; 52 °C, 30 s; 72 °C, 45 s). Primer pairs were then added to each PCR (Table I), and IIA N-pro/CPII cDNA fusion constructs were amplified for 30 cycles (95 °C, 30 s; 55 °C, 30 s; 72 °C, 55 s). cDNA fusion constructs were purified from 1% agarose gels, digested with EcoRI, and subcloned into pcDNA3 (Invitrogen). DNA sequencing using the T7 and Sp-6 promoter primers confirmed correct orientation of the inserted cDNA fragments.Table IOligonucleotide primers used for amplification of cDNA encoding the type IIA N-propeptide (IIA N-pro) or three regions of the type II procollagen C-propeptide domain (CPII)cDNAPrimer pairs for PCR (5′-3′)IIA N-propeptideF 1, ggtacgaattcatgattcgcctcgggR 2, tgcctggtcggccattggtccttgcatCPII (aa 1172-1189)F 3, aggaccaatggccgaccaggcagcagccR 4, catgggaattctcatcatgtggcatccacctcCPII (aa 1172-1196)F 3, aggaccaatggccgaccaggcagcagccR 5, catgggaattctcatcactggttgttgagggaCPII (aa 1172-1202)F 3, aggaccaatggccgaccaggcagcagccR 6, catgggaattctcatcagctgcggatgctctcFusion cDNA constructPrimer pairs for overlap extension PCRIIA N-pro/CPII (aa 1172-1189)1 + 4IIA N-pro/CPII (aa 1172-1196)1 + 5IIA N-pro/CPII (aa 1172-1202)1 + 6 Open table in a new tab Transient Transfections—Chinese hamster ovary (CHO) cells were plated overnight in 6-well tissue culture dishes at a density of 3 × 105 cells/ml. Cells were transiently transfected with pcDNA3 alone or pcDNA3 containing each of the three cDNA fusion constructs (IIA N-pro/CPII amino acids 1172–1189/1196/1202) using FuGENE 6 reagent (Roche Applied Science), in the presence of 50 μg/ml ascorbate. Cells transfected to synthesize full-length IIA N-pro/SP-D fusion protein (28Zhang P. McAlinden A. Li S. Schumacher T. Wang H. Hu S. Sandell L. Crouch E. J. Biol. Chem. 2001; 276: 19862-19870Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar) or just the type IIA N-propeptide were included as controls; IIA N-pro/SP-D forms trimers in solution, whereas IIA N-propeptide alone remains monomeric. Conditioned medium (12–36 ml per fusion construct) was harvested after 2 days in culture and clarified by centrifugation. Phenylmethylsulfonyl fluoride (1 mm) was added to the medium, and proteins were precipitated overnight at 4 °C with 33% ammonium sulfate. Precipitated proteins were collected by centrifugation and washed three times in saturated ammonium sulfate. Protein pellets were resuspended in phosphate-buffered saline (1–3 ml) and dialyzed overnight against phosphate-buffered saline at 4 °C. Chemical Cross-linking and Detection of Chimeric Fusion Proteins— Covalent cross-linking of fusion proteins was done using bis-(sulfosuccinimidyl)suberate (BS3; Pierce). Increasing amounts of BS3 (0, 0.1, 0.5, 1, or 2 mm final concentration) prepared in 5 mm sodium citrate, pH 5, were added to dialyzed conditioned medium for 1 h at room temperature. Addition of SDS-PAGE loading buffer containing Tris-HCl (0.5 m) inhibited the reaction. Samples were boiled for 5 min prior to SDS-PAGE, which was carried out without sulfhydryl reduction. Proteins were transferred to nitrocellulose membranes by Western blotting, and the presence of cross-linked dimers or trimers was identified by immunolocalization using anti-IIA polyclonal antisera (40Oganesian A. Zhu Y. Sandell L.J. J. Histochem. Cytochem. 1997; 45: 1469-1480Crossref PubMed Scopus (95) Google Scholar). Peptide Analysis—Peptides corresponding to GCN4-pII (RMKQIEDKIEEILSKIYHIENEIARIKKLIGER-NH2) and residues 1172–1189 (ADQAAGGLRQHDAEVDAT-NH2), 1172–1196 (ADQAAGGLRQHDAEVDATLKSLNNQ-NH2), and 1172–1202 (ADQAAGGLRQHDAEVDATLKSLNNQIESIRS-NH2) from human procollagen II were synthesized on a Milligan 9050 instrument with Fmoc/DIC/HOBt chemistry. They were then cleaved using trifluoroacetic acid with classical scavengers and precipitated in diethyl ether, followed by centrifugation. Pellets were then dissolved in water and lyophilized. The crude peptides were then re-dissolved in water and purified on a Vydac column (C18, 5 μm, 25 × 1 cm) with an appropriate gradient of acetonitrile in 0.1% trifluoroacetic acid. They were then characterized by electrospray mass spectrometry (SCIEX API 165) and by high pressure liquid chromatography (HP1100) using an analytical Vydac C18 column with a 30-min gradient from 7 to 63% acetonitrile. Far-UV (180–260 nm) circular dichroism measurements were carried out using thermostated 0.2-mm path length quartz cells in a Jobin-Yvon CD6 instrument, calibrated with aqueous d 10-camphorsulfonic acid. Peptides (200-400 μm) were analyzed at 25 °C in 20 mm KH2PO4/NaOH, 150 mm NaF, pH 7.2, in the presence or absence of 50% TFE (41Pan O.H. Beck K. J. Biol. Chem. 1998; 273: 14205-14209Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar). Spectra were measured with wavelength increment 0.2 nm, integration time 1 s, and bandpass 2 nm. For cross-linking analysis with BS3, conditions were the same as those used for the fusion proteins (see above), except that peptide concentrations were as for the circular dichroism. Coiled Coils in Fibrillar Procollagen C-propeptides— Fig. 1A shows an amino acid sequence alignment of the C-propeptide domains of the human fibrillar procollagens. From the pattern of repeating hydrophobic residues at positions a and d, and the frequent occurrence of charged residues at positions e and g, it can be seen that there are up to four heptad repeats, indicative of coiled coils, beginning soon after the procollagen C-proteinase cleavage site. In all chains, the g position at the end of the fourth heptad repeat is a proline which therefore defines the end of the α-helical region. The extent of the heptad repeat pattern is similar to that found in the collectins (Fig. 1B), again C-terminal to the collagenous region. This suggests the existence of a coiled-coil structure near the N terminus of all fibrillar C-propeptide trimers. Putative coiled-coil regions were also found C-terminal to the triple helix in the fibrillar procollagens XXIV (42Gordon M.K. Hahn R.A. Zhou P. Bhatt P. Song R. Kistler A. Gerecke D.R. Koch M. FASEB J. 2002; 16: 359Google Scholar) and XXVII (43Pace J.M. Corrado M. Missero C. Byers P.H. Matrix Biol. 2003; 22: 3-14Crossref PubMed Scopus (117) Google Scholar, 44Boot-Handford R.P. Tuckwell D.S. Plumb D.A. Farrington R.C. Poulsom R. J. Biol. Chem. 2003; 278: 31067-31077Abstract Full Text Full Text PDF PubMed Scopus (126) Google Scholar) (Fig. 1A). When the C-propeptide heptad repeats were further analyzed using SCORER (38Woolfson D.N. Alber T. Protein Sci. 1995; 4: 1596-1607Crossref PubMed Scopus (201) Google Scholar), a preference for three-stranded, as opposed to two-stranded, coiled coils was evident (data not shown). Modeling of the three-dimensional structure of the type I procollagen C-propeptide di-heptad sequence spanning the d positions of heptads two and four (Fig. 1A), based on that adopted by the mutated form of the GCN4 leucine zipper (GCN4-pII, in which residues at the a and d positions are isoleucines) (45Harbury P.B. Kim P.S. Alber T. Nature. 1994; 371: 80-83Crossref PubMed Scopus (427) Google Scholar) was consistent with a three-stranded coiled-coil conformation free of steric clashes (Fig. 2). Possible ionic interactions stabilizing the structure are listed in Table II. Based on this analysis, and in view of the recent observation that a di-heptad repeat in SP-D is sufficient for trimerization (29McAlinden A. Crouch E. Bann J.G. Zhang P. Sandell L. J. Biol. Chem. 2002; 277: 41274-41281Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar), this suggests that the heptad repeats in the fibrillar procollagen C-propeptides could also act as oligomerization domains.Table IIPossible interactions between charged amino acid residues in the putative coiled regions of the human fibrillar procollagen C-propeptidesi → i + 3i → i + 4g → ePro-α1(I)Glu15-Arg18Asp4-Lys8Pro-α2(I)Asp4-Lys8Pro-α1(II)Glu15-Arg18Asp4-Lys8Pro-α1(III)Pro-α1(V)Glu15-Lys18Lys11-Glu15Glu13-Lys18E15-Arg19Pro-α2(V)Glu15-Arg18Pro-α1(XI)Glu15-Lys18Lys11-Glu15Asp13-Lys18Pro-α2(XI)Asp8-Arg11Arg12-Glu16Glu13-Arg18Glu15-Arg18Glu15-Arg19 Open table in a new tab Coiled Coils in Procollagens as Oligomerization Domains—In order to test whether the heptad repeats in the procollagen C-propeptides can act as oligomerization domains, we adopted the approach described by McAlinden et al. (29McAlinden A. Crouch E. Bann J.G. Zhang P. Sandell L. J. Biol. Chem. 2002; 277: 41274-41281Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar) in which putative trimerization domains were fused C-terminal to the type IIA procollagen N-propeptide and expressed in CHO cells. The IIA N-propeptide contains a short, interrupted (GXY)n region (25 triplets), which alone does not spontaneously trimerize. When fused to at least the first two heptad repeats of the neck region of SP-D, however, trimerization ensues, as revealed by electrophoretic analysis after stabilization of trimers by cross-linking. By using the same IIA N-propeptide as a trimerization reporter molecule, we made fusion constructs using sequences from the C-propeptide of human type II procollagen, beginning at the alanine residue in the procollagen C-proteinase cleavage site, up to the end of the second, third, or fourth heptad repeats (Fig. 3). As shown in Fig. 4A, positive controls with the IIA N-propeptide fused to the entire C-terminal region of SP-D (beginning at the first heptad repeat) readily formed dimers and trimers, as expected (29McAlinden A. Crouch E. Bann J.G. Zhang P. Sandell L. J. Biol. Chem. 2002; 277: 41274-41281Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar), whereas in the absence of an oligomerization domain only monomers were observed (Fig. 4B). In contrast, when the IIA N-propeptide was fused to the N-terminal sequence of the type II procollagen C-propeptide, up to position f in the fourth heptad repeat, cross-linkable dimers and trimers were again observed (Fig. 4C). Similar results were obtained with fusion constructs containing only the first three (Fig. 4D) or two (Fig. 4E) heptad repeats. The data obtained with the fusion constructs show that sequences from the N-terminal region of the procollagen II C-propeptide can trimerize when linked to the IIA N-propeptide. To determine whether these sequences are capable of trimerizing independently, the corresponding peptides were synthesized along with GCN4-pII as a positive control. When analyzed by circular dichroism (Fig. 5A), GCN4-pII gave a spectrum characteristic of an almost fully α-helical coiled coil, as shown by the positions and heights of the peaks at 192, 208, and 222 nm, with a ratio of mean residue ellipticities θ220/θ208 of 1.039. In the presence of 50% TFE (Fig. 5B), which disrupts coiled coils and stabilizes α-helices (46Litowski J.R. Hodges R.S. J. Pept. Res. 2001; 58: 477-492Crossref PubMed Scopus (63) Google Scholar), this ratio decreased to 0.84, due mainly to changes in the peak at 208 nm, as expected. In contrast, all three peptides from the N-terminal region of the procollagen II C-propeptide yielded spectra characteristic of random coils, in the absence of TFE. Therefore these peptides do not form stable coiled coils independently. In the presence of 50% TFE, spectra clearly characteristic of α-helices were obtained with the two longer peptides (residues 1172–1196 and 1172–1202), whereas the shorter peptide (residues 1172–1189) showed some signs of α-helical structure. Therefore, all three peptides are capable of adopting an α-helical fold. The results obtained by circular dichroism were confirmed by cross-linking analysis using BS3, followed by SDS-PAGE, which showed trimerization of GCN4-pII but not of a similar length peptide (residues 1172–1202) from the human procollagen II C-propeptide (data not shown). C-terminal Coiled Coils in Other Members of the Collagen Superfamily—α-Helical coiled-coil heptad repeats were also found in the C-propeptides of known invertebrate fibrillar procollagens, including the sponge Ephydatia muelleri, hydra, annelids (alvinella, arenicola, and riftia), mollusc (abalone), and sea urchin (Fig. 1C) (47Exposito J.Y. Cluzel C. Garrone R. Lethias C. Anat. Rec. 2002; 268: 302-316Crossref PubMed Scopus (141) Google Scholar, 48Boot-Handford R.P. Tuckwell D.S. BioEssays. 2003; 25: 142-151Crossref PubMed Scopus (155) Google Scholar). Both the length and distance of the heptad repeats from the end of the collagen triple-helical region were similar to those in the vertebrate fibrillar collagens (Fig. 1A), although in hydra and annelid fibrillar procollagens the intervening sequence was relatively long (Fig. 1C). Similar C-terminally located heptad sequences were also found in some bacteriophage tail fiber collagen-like proteins (data not shown). As pointed out by Beck and Brodsky (49Beck K. Brodsky B. J. Struct. Biol. 1998; 122: 17-29Crossref PubMed Scopus (277) Google Scholar), a possible coiled-coil oligomerization domain C-terminal to the triple helix is also found in the collagen-like tail subunit of acetylcholinesterase (ColQ) (50Ohno K. Brengman J. Tsujino A. Engel A.G. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 9654-9659Crossref PubMed Scopus (233) Google Scholar). The sequence (Fig. 1D) consists of 2.5 heptad repeats and is found about 30 residues C-terminal to the last GPP triplet, similar to that in the procollagen C-propeptides. Probable trimerization domains have also been identified in the multiplexins (collagens XV and XVIII) (51Sasaki T. Fukai N. Mann K. Gohring W. Olsen B.R. Timpl R. EMBO J. 1998; 17: 4249-4256Crossref PubMed Scopus (328) Google Scholar, 52Sasaki T. Larsson H. Tisi D. Claesson-Welsh L. Hohenester E. Timpl R. J. Mol. Biol. 2000; 301: 1179-1190Crossref PubMed Scopus (197) Google Scholar), immediately after the most C-terminal triple-helical domain. Examination of the corresponding sequences showed that these also contain
Referência(s)