Template-assembled triple-helical peptide molecules: mimicry of collagen by molecular architecture and integrin-specific cell adhesion

Most proteins fold into specific structures to exert their biological functions, and therefore the creation of protein-like molecular architecture is a fundamental prerequisite toward realizing a novel biologically active protein-like biomaterial. To do this with an artificial collagen, we have engineered a peptide template characterized by its collagen-like primary structure composed of Gly-Phe-Gly-Glu-Glu-Gly sequence to assemble (Pro-Hyp-Gly)n (n = 3 and 5) into triple-helical conformations that resemble the native structure of collagen. The peptide template has three carboxyl groups connected to the N-termini of three collagen peptides. The coupling was accomplished by a simple and direct branching protocol without complex strategies. A series of biophysical studies, including melting curve analyses and CD and NMR spectroscopy, demonstrated the presence of stable triple-helical conformation in the template-assembled (Pro-Hyp-Gly)3 and (Pro-Hyp-Gly)5 solution. Conversely, nontemplated peptides showed no evidence of assembly of triple-helical structure. A cell binding sequence (Gly-Phe-Hyp-Gly-Glu-Arg) derived from the collagen alpha1(I) chain was incorporated to mimic the integrin-specific cell adhesion of collagen. Cell adhesion and inhibition assays and immunofluorescence staining revealed a correlation of triple-helical conformation with cellular recognition of collagen mimetics in an integrin-specific way. This study offers a robust strategy for engineering native-like peptide-based biomaterials, fully composed of only amino acids, by maintaining protein conformation integrity and biological activity.     

Conformational effects of Gly-X-Gly interruptions in the collagen triple helix

The collagen model peptide with sequence (Pro-Hyp-Gly)4-Pro-Gly-(Pro-Hyp-Gly)5 contains a central Gly-Pro-Gly interruption in the consensus collagen sequence. Its high-resolution crystal structure defines the molecular consequences of such an interruption for the collagen triple-helical conformation, and provides insight into possible structural and biological roles of similar interruptions in the -Gly-X-Y- repeating pattern found in non-fibrillar collagens. The peptide (denoted as the Hyp minus peptide or Hyp-) forms a rod-like triple helix structure without any bend or kink, and crystallizes in a quasi-hexagonal lattice. The two Pro-Hyp-Gly zones adopt the typical triple-helical collagen conformation with standard Rich and Crick II hydrogen bonding topology. Notably, the central zone containing the Gly-Pro-Gly interruption deviates from the standard structure in terms of hydrogen bonding topology, torsion angles, helical, and superhelical parameters. These deviations are highly localized, such that the standard features are regained within one to two residues on either side. Conformational variations and high temperature factors seen for the six chains of the asymmetric unit in the zone around the interruption point to the presence of a local region of considerable plasticity and flexibility embedded within two highly rigid and ordered standard triple-helical segments. The structure suggests a role for Gly-X-Gly interruptions as defining regions of flexibility and molecular recognition in the otherwise relatively uniform repeating collagen conformation.     

Transformation of the mechanism of triple-helix peptide folding in the absence of a C-terminal nucleation domain and its implications for mutations in collagen disorders

Folding abnormalities of the triple helix have been demonstrated in collagen diseases such as osteogenesis imperfecta in which the mutation leads to the substitution of a single Gly in the (Gly-X-Y)n sequence pattern by a larger residue. Model peptides can be used to clarify the details of normal collagen folding and the consequences of the interruption of that folding by a Gly substitution. NMR and CD studies show that placement of a (GPO)4 nucleation domain at the N terminus rather than the C terminus of a native collagen sequence allows the formation of a stable triple helix but alters the folding mechanism. Although C- to N-terminal directional folding occurs when the nucleation domain is at the C terminus, there is no preferential folding direction when the nucleation domain is at the N terminus. The lack of zipper-like directional folding does not interfere with triple-helix formation, and when a Gly residue is replaced by Ser to model an osteogenesis imperfecta mutation, the peptide with the N-terminal (GPO)4 domain can still form a good triple helix N-terminal to the mutation site. These peptide studies raise the possibility that mutant collagen could fold in a C to N direction in a zipper-like manner up to the mutation site and that completion of the triple helix N-terminal to the mutation would involve an alternative mechanism.     

Sequence dependence of kinetics and morphology of collagen model peptide self-assembly into higher order structures

The process of self-assembly of the triple-helical peptide (Pro-Hyp-Gly)(10) into higher order structure resembles the nucleation-growth mechanism of collagen fibril formation in many features, but the irregular morphology of the self-assembled peptide contrasts with the ordered fibers and networks formed by collagen in vivo. The amino acid sequence in the central region of the (Pro-Hyp-Gly)(10) peptide was varied and found to affect the kinetics of self-assembly and nature of the higher order structure formed. Single amino acid changes in the central triplet produced irregular higher order structures similar to (Pro-Hyp-Gly)(10), but the rate of self-association was markedly delayed by a single change in one Pro to Ala or Leu. The introduction of a Hyp-rich hydrophobic sequence from type IV collagen resulted in a more regular suprastructure of extended fibers that sometimes showed supercoiling and branching features similar to those seen for type IV collagen in the basement membrane network. Several peptides, where central Pro-Hyp sequences were replaced by charged residues or a nine-residue hydrophobic region from type III collagen, lost the ability to self-associate under standard conditions. The inability to self-assemble likely results from loss of imino acids, and lack of an appropriate distribution of hydrophobic/electrostatic residues. The effect of replacement of a single Gly residue was also examined, as a model for collagen diseases such as osteogenesis imperfecta and Alport syndrome. Unexpectedly, the Gly to Ala replacement interfered with self-assembly of (Pro-Hyp-Gly)(10), while the peptide with a Gly to Ser substitution self-associated to form a fibrillar structure.     

Staggered molecular packing in crystals of a collagen-like peptide with a single charged pair

The crystal structure of the triple-helical peptide, (Pro-Hyp-Gly)(4)-Glu-Lys-Gly-(Pro-Hyp-Gly)(5) has been determined to 1.75 A resolution. This peptide was designed to examine the effect of a pair of adjacent, oppositely charged residues on collagen triple-helical conformation and intermolecular interactions. The molecular conformation (a 7(5) triple helix) and hydrogen bonding schemes are similar to those previously reported for collagen triple helices and provides a second instance of water mediated N--H . . . O==C interchain hydrogen bonds for the amide group of the residue following Gly. Although stereochemically capable of forming intramolecular or intermolecular ion pairs, the lysine and glutamic acid side-chains instead display direct interactions with carbonyl groups and hydroxyproline hydroxyl groups or interactions mediated by water molecules. Solution studies on the EKG peptide indicate stabilization at neutral pH values, where both Glu and Lys are ionized, but suggest that this occurs because of the effects of ionization on the individual residues, rather than ion pair formation. The EKG structure suggests a molecular mechanism for such stabilization through indirect hydrogen bonding. The molecular packing in the crystal includes an axial stagger between molecules, reminiscent of that observed in D-periodic collagen fibrils. The presence of a Glu-Lys-Gly triplet in the middle of the sequence appears to mediate this staggered molecular packing through its indirect water-mediated interactions with backbone C==O groups and side chains.     

Ferrocene-assisted stabilization of collagen mimetic triple helices: solid-phase synthesis and structure

A series of ferrocene-containing collagen models Fc-CO-(Pro-Hyp-Gly)n-Cys (n = 4 (1), 6 (2), 7 (3), 8 (4), 9 (5)) were synthesized by solid-phase synthesis. Biophysical studies using circular dichroism (CD) show that these collagen analogues form triple-helical conformations, and the peptides showed a range of thermal stabilities ((T(m)), 38-74 degrees C). Results also indicate that the ferrocene (Fc)-labeled collagen models possesses a higher triple-helical propensity than the unlabeled collagen models as demonstrated by the higher melting temperatures and thermodynamic parameters, and we conclude that the Fc group at the N-terminal position of the peptide strands increases the stability of the triple helix.     

Effects of the stereo-configuration of the hydroxyl group in 4-hydroxyproline on the triple-helical structures formed by homogenous peptides resembling collagen.

To explore further the recent demonstration that hydroxyproline stabilizes the triple-helical structure of collagen, two peptides containing allohydroxyproline, (aHyp-Pro-Gly)10 and (Pro-aHyp-Gly)10, were synthesized by a modified Merrifield technique which yields products of defined molecular weight. Examination of the peptides by optical rotation and circular dichroism showed that neither of them formed triple-helical structures in aqueous solution. Since the peptides had less tendency than (Pro-Hyp-Gly)10 to become helical, the results demonstrated that the trans-4-hydroxyl group of hydroxyproline makes a specific contribution to stability of the triple helix formed by (Pro-Hyp-Gly)10. Since the peptides also had less tendency than (Pro-Pro-Gly10 to become helical, the results further demonstrated that the cis-4-hydroxyl group on allohydroxyproline decreases the stability of the triple helix. The observations provided direct support for previous data indicating that incorporation of proline analogues such as allohydroxyproline into pro-alpha chains during procollagen biosynthesis prevents the polypeptides from becoming triple helical.     

The specific recognition of a cell binding sequence derived from type I collagen by Hep3B and L929 cells

In this study, the affinity of two different cell types toward a specific cell binding sequence (Gly-Phe-Hyp-Gly-Glu-Arg or GFOGER) derived from type I collagen using peptide template (PT)-assembled collagen peptides of different triple helicity as a model for natural collagen is examined. A series of biophysical studies, including melting curve analysis and circular dichroism spectroscopy, demonstrated the presence of stable triple-helical conformation in the PT-assembled (GPO)3-GFOGER-(GPO)3, (GPO)-GFOGER-(GPO), and (Pro-Hyp-Gly)5 solution. Conversely, non-templated peptides, except (GPO)3-GFOGER-(GPO)3, showed no evidence of assembly into triple-helical structure. Biological assays, including cell adhesion, competitive inhibition, and immunofluorescence staining, revealed a correlation of triple-helical conformation with the cellular recognition of GFOGER in an integrin-specific manner. The triple helix was shown to be important, but not crucial for cell adhesion to native collagen. Hep3B and L929 cells displayed significant differences in the recognition of GFOGER, mainly because of the differences in their expression of specific integrin receptors for collagen. For example, PT-assembled (GPO)3-GFOGER-(GPO)3 was shown to perform comparably to collagen for L929, but not Hep3B, cell adhesion. The result showed that a specific cell binding motif may not fully mimic the extracellular matrix (ECM) microenvironment, suggesting the need to use a combination of two or more cell binding sequences for targeting a wide range of integrin receptors expressed by a specific cell type to better mimic the ECM.     

Predicting the clinical lethality of osteogenesis imperfecta from collagen glycine mutations

Osteogenesis imperfecta (OI), or brittle bone disease, often results from missense mutation of one of the conserved glycine residues present in the repeating Gly-X-Y sequence characterizing the triple-helical region of type I collagen. A composite model was developed for predicting the clinical lethality resulting from glycine mutations in the alpha1 chain of type I collagen. The lethality of mutations in which bulky amino acids are substituted for glycine is predicted by their position relative to the N-terminal end of the triple helix. The effect of a Gly --> Ser mutation is modeled by the relative thermostability of the Gly-X-Y triplet on the carboxy side of the triplet containing the substitution. This model also predicts the lethality of Gly --> Ser and Gly --> Cys mutations in the alpha2 chain of type I collagen. The model was validated with an independent test set of six novel Gly --> Ser mutations. The hypothesis derived from the model of an asymmetric interaction between a Gly --> Ser mutation and its neighboring residues was tested experimentally using collagen-like peptides. Consistent with the prediction, a significant decrease in stability, calorimetric enthalpy, and folding time was observed for a peptide with a low-stability triplet C-terminal to the mutation compared to a similar peptide with the low-stability triplet on the N-terminal side. The computational and experimental results together relate the position-specific effects of Gly --> Ser mutations to the local structural stability of collagen and lend insight into the etiology of OI.     

NMR and CD studies of triple-helical peptides.

Triple-helix formation of the peptide (Pro-Hyp-Gly)10 was monitored by nmr and CD spectroscopy. The two-dimensional nmr spectra indicated that the Gly C alpha H and Pro C delta H proton resonances shift upfield in going from the nonhelical to helical form, while hydroxy-proline resonances are unchanged. The integrated areas of the helical and nonhelical resonances could be monitored in the one-dimensional nmr spectrum, and indicate that in the (Pro-Hyp-Gly)10 about 90% of the residues are in a defined triple-helical conformation. The introduction of a glycine to alanine substitution or the deletion of a single hydroxyproline residue in the stable triple-helical peptide (Pro-Hyp-Gly)10 still allows trimers to be formed, but the trimers show a substantial loss of triple helix and decreased thermal stability compared with (Pro-Hyp-Gly)10. Two computer models were generated for the Gly----Ala peptide, one with the Ala side chains packed inside the helix and the other with the region containing the alanines forming a beta-bend that loops out from the helix. The nmr data is more consistent with the latter model.     

Induction of protein-like molecular architecture by self-assembly processes

One of the most intriguing self-assembly processes is the folding of peptide chains into native protein structures. We have developed a method for building protein-like structural motifs that incorporate sequences of biological interest. A lipophilic moiety is attached onto an N(alpha)-amino group of a peptide chain, resulting in a 'peptide-amphiphile'. The alignment of amphiphilic compounds at the lipid solvent interface is used to facilitate peptide alignment and structure initiation and propagation. Peptide-amphiphiles containing potentially triple-helical structural motifs have been synthesized. The resultant head group structures have been characterized by circular dichroism and NMR spectroscopies. Evidence for a self-assembly process of peptide-amphiphiles has been obtained from: (a) circular dichroism spectra and melting curves characteristic of triple-helices, (b) one- and two-dimensional NMR spectra indicative of stable triple-helical structure at low temperatures and melted triple-helices at high temperatures, and (c) pulsed-field gradient NMR experiments demonstrating different self-diffusion coefficients between proposed triple-helical and non-triple-helical species. The peptide-amphiphiles described here provide a simple approach for building stable protein structural motifs using peptide head groups.     

Collagen triple helix formation can be nucleated at either end

The directional dependence of folding rates for rod-like macromolecules such as parallel alpha-helical coiled-coils, DNA double-helices, and collagen triple helices is largely unexplored. This is mainly due to technical difficulties in measuring rates in different directions. Folding of collagens is nucleated by trimeric non-collagenous domains. These are usually located at the COOH terminus, suggesting that triple helix folding proceeds from the COOH to the NH(2) terminus. Evidence is presented here that effective nucleation is possible at both ends of the collagen-like peptide (Gly-Pro-Pro)(10), using designed proteins in which this peptide is fused either NH(2)- or COOH-terminal to a nucleation domain, either T4-phage foldon or the disulfide knot of type III collagen. The location of the nucleation domain influences triple-helical stability, which might be explained by differences in the linker sequences and the presence or absence of repulsive charges at the carboxyl-terminal end of the triple helix. Triple helical folding rates are found to be independent of the site of nucleation and consistent with cis-trans isomerization being the rate-limiting step.     

The Recognition of Collagen and Triple-helical Toolkit Peptides by MMP-13: SEQUENCE SPECIFICITY FOR BINDING AND CLEAVAGE*

Remodeling of collagen by matrix metalloproteinases (MMPs) is crucial to tissue homeostasis and repair. MMP-13 is a collagenase with a substrate preference for collagen II over collagens I and III. It recognizes a specific, well-known site in the tropocollagen molecule where its binding locally perturbs the triple helix, allowing the catalytic domain of the active enzyme to cleave the collagen α chains sequentially, at Gly⁷⁷⁵–Leu⁷⁷⁶ in collagen II. However, the specific residues upon which collagen recognition depends within and surrounding this locus have not been systematically mapped. Using our triple-helical peptide Collagen Toolkit libraries in solid-phase binding assays, we found that MMP-13 shows little affinity for Collagen Toolkit III, but binds selectively to two triple-helical peptides of Toolkit II. We have identified the residues required for the adhesion of both proMMP-13 and MMP-13 to one of these, Toolkit peptide II-44, which contains the canonical collagenase cleavage site. MMP-13 was unable to bind to a linear peptide of the same sequence as II-44. We also discovered a second binding site near the N terminus of collagen II (starting at helix residue 127) in Toolkit peptide II-8. The pattern of binding of the free hemopexin domain of MMP-13 was similar to that of the full-length enzyme, but the free catalytic subunit bound none of our peptides. The susceptibility of Toolkit peptides to proteolysis in solution was independent of the very specific recognition of immobilized peptides by MMP-13; the enzyme proved able to cleave a range of dissolved collagen peptides.     

Free energy simulation to investigate the effect of amino acid sequence environment on the severity of osteogenesis imperfecta by glycine mutations in collagen

Molecular dynamics simulations were carried out to calculate the free energy change difference of two collagen-like peptide models for Gly --> Ser mutations causing two different osteogenesis imperfecta phenotypes. These simulations were performed to investigate the impact of local amino acid sequence environment adjacent to a mutation site on the stability of the collagen. The average free energy differences for a Gly --> Ser mutant relative to a wild type are 3.4 kcal/mol and 8.2 kcal/mol for a nonlethal site and a lethal site, respectively. The free energy change differences of mutant containing two Ser residues relative to the wild type at the nonlethal and lethal mutation sites are 4.6 and 9.8 kcal/mol, respectively. Although electrostatic interactions stabilize mutants containing one or two Ser residues at both mutation sites, van der Waals interactions are of sufficient magnitude to cause a net destabilization. The presence of Gln and Arg near the mutation site, which contain large and polar side chains, provide more destabilization than amino acids containing small and nonpolar side chains.     

Location of glycine mutations within a bacterial collagen protein affects degree of disruption of triple-helix folding and conformation

The hereditary bone disorder osteogenesis imperfecta is often caused by missense mutations in type I collagen that change one Gly residue to a larger residue and that break the typical (Gly-Xaa-Yaa)(n) sequence pattern. Site-directed mutagenesis in a recombinant bacterial collagen system was used to explore the effects of the Gly mutation position and of the identity of the residue replacing Gly in a homogeneous collagen molecular population. Homotrimeric bacterial collagen proteins with a Gly-to-Arg or Gly-to-Ser replacement formed stable triple-helix molecules with a reproducible 2 °C decrease in stability. All Gly replacements led to a significant delay in triple-helix folding, but a more dramatic delay was observed when the mutation was located near the N terminus of the triple-helix domain. This highly disruptive mutation, close to the globular N-terminal trimerization domain where folding is initiated, is likely to interfere with triple-helix nucleation. A positional effect of mutations was also suggested by trypsin sensitivity for a Gly-to-Arg replacement close to the triple-helix N terminus but not for the same replacement near the center of the molecule. The significant impact of the location of a mutation on triple-helix folding and conformation could relate to the severe consequences of mutations located near the C terminus of type I and type III collagens, where trimerization occurs and triple-helix folding is initiated.     

CD and NMR investigation of collagen peptides mimicking a pathological Gly–Ser mutation and a natural interruption in a similar highly charged sequence context

Even a single Gly substitution in the triple helix domain of collagen leads to pathological conditions while natural interruptions are suggested to play important functional roles. Two peptides—one mimicking a pathological Gly–Ser substitution (ERSEQ) and the other one modeling a similar natural interruption sequence (DRSER)—are designed to facilitate the comparison for elucidating the molecular basis of their different biological roles. CD and NMR investigation of peptide ERSEQ indicates a reduction of the thermal stability and disruption of hydrogen bonding at the Ser mutation site, providing a structural basis of the OI disease resulting from the Gly–Ser mutation in the highly charged RGE environment. Both CD and NMR real‐time folding results indicate that peptide ERSEQ displays a comparatively slower folding rate than peptide DRSER, suggesting that the Gly–Ser mutation may lead to a larger interference in folding than the natural interruption in a similar RSE context. Our studies suggest that unlike the rigid GPO environment, the abundant R(K)GE(D) motif may provide a more flexible sequence environment that better accommodates mutations as well as interruptions, while the electrostatic interactions contribute to its stability. These results shed insight into the molecular features of the highly charged motif and may aid the design of collagen biomimetic peptides containing important biological sites.     

Asymmetry in the triple helix of collagen-like heterotrimers confirms that external bonds stabilize collagen structure

Heating and subsequent cooling mixtures of (Pro-Pro-Gly)(10) and (Pro-Hyp-Gly)(10) peptides leads to formation of model heterotrimeric collagen helices that can be isolated by HPLC. These heterotrimeric collagen peptide helices are shown to be fundamentally unstable as denaturing then renaturing experiments result in heterotrimeric/homotrimeric mixtures.As the proportion of hydroxyproline-containing chains in the trimers increases, differential scanning calorimetry shows that the helix melting temperatures and denaturation enthalpies increasing non-linearly. Three types of Rich-Crick hydrogen bonds observed by NMR allow modelling of heterotrimeric structures based on published homotrimeric X-ray data. This revealed a small axial movement of (Pro-Hyp-Gly)(10) chains towards the C-terminal of the helix, demonstrating heterotrimeric asymmetry.     

Mapping of SPARC/BM-40/osteonectin-binding sites on fibrillar collagens

The 33-kDa matrix protein SPARC (BM-40, osteonectin) binds several collagen types with moderate affinity. The collagen-binding site resides in helix alphaA of the extracellular calcium-binding domain of SPARC and is partially masked by helix alphaC. Previously, we found that the removal of helix alphaC caused a 10-fold increase in the affinity of SPARC for collagen, and we identified amino acids crucial for binding by site-directed mutagenesis. In this study, we used rotary shadowing, CNBr peptides, and synthetic peptides to map binding sites of SPARC onto collagens I, II, and III. Rotary shadowing and electron microscopy of SPARC-collagen complexes identified a major binding site approximately 180 nm from the C terminus of collagen. SPARC binding was also detected with lower frequency near the matrix metalloproteinase cleavage site. These data fit well with our analysis of SPARC binding to CNBr peptides, denaturation of which abolished binding, indicating triple-helical conformation of collagen to be essential. SPARC binding was substantially decreased in two of seven alpha2(I) mutant procollagen I samples and after N-acetylation of Lys/Hyl side chains in wild-type collagen. Synthetic peptides of collagen III were used to locate the binding sites, and we found SPARC binding activity in a synthetic triple-helical peptide containing the sequence GPOGPSGPRGQOGVMGFOGPKGNDGAO (where O indicates 4-hydroxyproline), with affinity for SPARC comparable with that of procollagen III. This sequence is conserved among alpha chains of collagens I, II, III, and V. In vitro collagen fibrillogenesis was delayed in the presence of SPARC, suggesting that SPARC might modulate collagen fibril assembly in vivo.     

Substitution of cysteine for glycine within the carboxyl-terminal telopeptide of the alpha 1 chain of type I collagen produces mild osteogenesis imperfecta

We have characterized a mutation that produces mild, dominantly inherited osteogenesis imperfecta. Half of the alpha 1 (I) chains of type I collagen synthesized by cells from an affected individual contain a cysteine residue in the 196-residue carboxyl-terminal cyanogen bromide peptide of the triple-helical domain (Steinmann, B., Nicholls, A., and Pope, F. M. (1986) J. Biol. Chem. 261, 8958-8964). Unexpectedly, sequence determined from a proteolytic fragment of the alpha 1 (I) chain derived from procollagen molecules synthesized in the presence of both [3H]proline and [35S]cysteine indicated that the cysteine is located at the third residue carboxyl-terminal to the triple-helical domain, normally a glycine. The nucleotide sequence of a fragment amplified from genomic DNA confirmed the location of the cysteine residue and showed that the mutation was a single nucleotide change in one COL1A1 allele. This represents a new class of mutations, point mutations outside the triple-helical domain of the chains of type I collagen, that produce the osteogenesis imperfecta phenotype.