Triple-helical peptides: An approach to collagen conformation, stability, and self-association

Abstract

Peptides have been an integral part of the collagen triple-helix structure story, and have continued to serve as useful models for biophysical studies and for establishing biologically important sequence-structure-function relationships. High resolution structures of triple-helical peptides have confirmed the basic Ramachandran triple-helix model and provided new insights into the hydration, hydrogen bonding, and sequence dependent helical parameters in collagen. The dependence of collagen triple-helix stability on the residues in its (Gly-X-Y)_n repeating sequence has been investigated by measuring melting temperatures of host-guest peptides and an on-line collagen stability calculator is now available. Although the presence of Gly as every third residue is essential for an undistorted structure, interruptions in the repeating (Gly-X-Y)_n amino acid sequence pattern are found in the triple-helical domains of all nonfibrillar collagens, and are likely to play a role in collagen binding and degradation. Peptide models indicate that small interruptions can be incorporated into a rod-like triple-helix with a highly localized effect, which perturbs hydrogen bonds and places the standard triple-helices on both ends out of register. In contrast to natural interruptions, missense mutations which replace one Gly in a triple-helix domain by a larger residue have pathological consequences, and studies on peptides containing such Gly substitutions clarify their effect on conformation, stability, and folding. Recent studies suggest peptides may also be useful in defining the basic principles of collagen self-association to the supramolecular structures found in tissues. © 2008 Wiley Periodicals, Inc. Biopolymers 89: 345–353, 2008.

This article was originally published online as an accepted preprint. The “Published Online” date corresponds to the preprint version. You can request a copy of the preprint by emailing the Biopolymers editorial office at [e-mail]

INTRODUCTION

Polypeptides have been an integral part in the history of the collagen triple-helix structure. In the 1950s, elucidation of the structural features of polyproline1 and polyglycine2 provided a conceptual framework for the proposal of the triple-helical structure of collagen and insight into its basic stabilizing interactions. The collagen triple-helix structure, as proposed by Ramachandran and Kartha (1955),3 Rich and Crick (1955),4 and Cowan et al. (1955),5 consists of three polypeptide chains, each in a polyproline II-like (PPII) conformation, supercoiled around a common axis. The triple-helix is stabilized by its high content of proline and hydroxyproline (Hyp), which promotes a PPII helix and by the formation of interchain backbone H bonds similar to those found in polyglycine II. The three polypeptide chains are staggered by one residue and their close packing near the central axis requires every third residue in each chain to be glycine, generating the characteristic Gly-X-Y repeating pattern of collagens.

In the 1960s, advances in peptide chemistry set the stage for the synthesis and characterization of polytripeptides as collagen models in various laboratories around the world (see Traub and Piez, 1971 for review6). Elkan Blout's laboratory became one center of this activity.7–11 His background in chemistry facilitated synthesis of oligopeptides and polytripeptides, while his pioneering research in spectroscopy provided the impetus to apply optical rotary dispersion, circular dichroism spectroscopy, and infrared spectroscopy to establish relationships between amino acid sequence and triple-helix stability, and conformation. Stabilizing solvents such as hexafluoroisopropanol and trifluoroethanol were used to induce helix formation, and comparisons were made of solid state and solution. These studies contributed to an understanding of the importance of the residues in the X and Y positions of the Gly-X-Y repeating sequence and of solvent for stabilization of the triple-helix.

In recent years, modern biophysical techniques have been effectively applied to triple-helical peptides but not to the triple-helix domains of collagen molecules themselves. The inability to crystallize collagen and its unsuitability for multidimensional NMR techniques have been offset in part by the success of X-ray crystallography and NMR studies on peptides. The extended nature of the triple-helix, which precludes long range interactions, allows relatively short peptides to be good models for biologically significant regions of collagen. Peptides of defined length and sequence are now easily made by solid state synthesis and allow precise correlations between sequence, structure, and function. Triple-helical peptides with specific sequences have been shown to have biological activity (e.g. Miles et al., 199512) and sets of overlapping peptides that cover the entire collagen sequence in the Farndale laboratory have led to the establishment of sequence requirements for collagen binding to integrin and other receptors.13 Several recent publications comprehensively review many aspects of collagen model peptides,14–16 and here the focus will be on recent studies of collagen-like peptides from our laboratory and the perturbation of the classic triple-helix by natural and pathological interruptions in the Gly-X-Y repeating sequence.

X-RAY CRYSTALLOGRAPHY OF TRIPLE-HELICAL PEPTIDES

The original models of the collagen triple-helix structure were based on fiber diffraction data, and high resolution structures of triple-helix peptides obtained by X-ray crystallography provided confirmation and new insights into the molecular conformation. The first crystal structure was obtained by Okuyama et al. (1981)17 for (Pro-Pro-Gly)₁₀, followed by structures for peptides with varying sequences from a number of laboratories (Table I) (for reviews, see Okuyama et al., 200616 and Brodsky and Persikov, 200515). In addition to the peptide structures, there are high resolution structures of complexes of a triple-helical peptide bound to the I domain of integrin19 and one bound to a bacterial cell surface receptor.20 The high-resolution structures of collagen-like peptides confirm the basic triple-helical model and show details of hydration, hydrogen bonding, and helical parameters.

Table I. List of Examples of Collagen Peptides with High Resolution Structures Solved by X-Ray Crystallographya

	Peptide Sequence	PDB ID	Properties
a For a complete listing of all solved structures as of 2006, see Okuyama et al., 2006.16 b Okuyama et al, 2007.18
Imino acid Polytripeptides	(Pro-Pro-Gly)n where n = 9,10	1A3I, 1K6F, 2CUO	7/2 symmetry, NH(Gly)…CO(X)
	(Pro-Hyp-Gly)10	1V7H
	(Gly-Hyp-Hyp)9	1YM8
Peptides with collagen sequences:			NH(Gly)…CO(X), NH(X)..W..CO(Gly), 7/2 symmetry in ends, more relaxed in central imino acid poor zone
LOG2	(Pro-Hyp-Gly)4-(Leu-Hyp-Gly)2-(Pro-Hyp-Gly)4b
Integrin binding peptide	(Gly-Pro-Hyp)2 -Gly-Phe-Hyp-Gly-Glu-Arg-(Gly-Pro-Hyp)3	1DZI, 1Q7D
T3-785 peptide	(Pro-Hyp-Gly)3-Ile-Thr-Gly-Ala-Arg-Gly-Leu-Ala-Gly-(Pro-Hyp-Gly)4	1BKV
Peptides with ImperfectGly-X-Y repeat:			Altered H bonding atinterruption site, good helix at both ends, but out of register
Peptide with natural interruption	(Pro–Hyp–Gly)3– Pro–Hyp–Gly– Pro–Gly–(Pro–Hyp–Gly)5	1EI8
Peptide with Gly→Ala mutation	(Pro-Hyp-Gly)4-Pro-Hyp-Ala-(Pro-Hyp-Gly)5	1CAG, 1CGD

The importance of water in collagen structure has been known from many techniques, and extensive hydration networks are seen in all crystal structures of collagen triple-helical peptides.21, 22 These hydrogen bonded networks often adopt pentagonal arrangements of water and are anchored to the peptide chain through backbone carbonyls and hydroxyl groups of Hyp (Figure 1a). The observation of Hyp involvement in a hydrogen-bonded water network is consistent with that proposed by Ramachandran's group on the basis of modeling and Privalov on the basis of calorimetry studies on collagens.23, 24 The mechanism of stabilization of the triple-helix by Hyp has been controversial, and studies on peptides with fluoroproline suggest stereoelectronic effects may lead to an exo-puckering of the imino acid ring, which is favorable for the triple-helix.14

Figure 1. Insights from high resolution X-ray crystal structures of collagen peptides. (a) Hydration network involving backbone carbonyl groups and hydroxyproline in the Gly→Ala peptide (reprinted from Structure 1995, 3, 893–906 with permission from Elsevier). (b) Structure of EKG peptide26 showing water mediated hydrogen bond NH(X)··· W···CO(Gly) where Glu residue is in the X position (Courtesy of Dr. Helen Berman). (c) Schematic diagram representing variable helical twist expected along the collagen sequence, with a tighter 7/2 helix expected for the imino acid rich (Gly-Pro-Hyp)_n region and a. looser, more relaxed helix in the imino poor region GITGARGLA found in peptide T3-785.29

The crystal structures of peptides clarified a long standing debate regarding hydrogen bonding within the triple-helix. The direct interchain Gly NH…CO (X position) has been a basic feature of all models,3, 4 but there had been proposal of a second H bond between NH (X position) to the CO of a Gly in a neighboring chain, either directly or mediated by water.25 The determination of several crystal structures of peptides with amino acids, rather than Pro, in the X position, showed a second interchain backbone hydrogen bond mediated by a water NH(X)…W…CO(Gly)26, 27, 18 (Figure 1b). NMR hydrogen exchange studies support its presence in solution as well.28

Structures of peptides with different sequences revealed a sequence-dependent variation in triple-helical parameters. Peptides containing only imino acids in the X and Y positions, e.g. Pro-Pro-Gly or Pro-Hyp-Gly triplets, have the 7/2 helical symmetry first seen by Okuyama for (Pro-Pro-Gly)_10.17 Peptide T3-785, which contains a 9-residue collagen sequence with no imino acids flanked by Pro-Hyp-Gly triplets, showed a sequence dependent variation in the superhelix twist: the repeating Pro-Hyp-Gly ends of the peptide adopted the 7/2 helix (3.5 residues/turn) while the imino acid free central zone was more relaxed, closer to a 10/3 helix (3.3 residues/turn).27, 29 Other imino acid poor sequences in peptides have also shown a looser triple-helix, and these parameters have been recently tabulated by Okuyama et al. (2006).16

These peptide structures indicate a high probability that the collagen molecule will have a subtle variation in its helical conformation along the chain. Regions which are very rich in imino acids, e.g. the (Gly-Pro-Hyp)₄ sequence at the C-terminus of the α1 chain of Type I collagen are likely to have a tighter helix, close to 7/2 symmetry (Figure 1c). Regions with sequences deficient in imino acids will have a second water mediated hydrogen bond, whenever Pro is not in the X position, and are likely to have a more relaxed helix (Figure 1c). These subtle variations in helix parameters may be important for collagen fibril assembly or binding of other molecules.

AMINO ACID SEQUENCE AND TRIPLE-HELIX STABILITY

The stability of collagen is an important biological property. The melting temperatures of collagens correlate with the upper temperature environment of the organism, while a decreased local stability is implicated in ligand binding and microunfolding regions important for fibril formation. Host-guest peptides of the form (Gly-Pro-Hyp)₃-Gly-X-Y-(Gly-Pro-Hyp)₄ have been used to clarify the relation between amino acid sequence and triple-helix stability.30 The melting temperature (T_m) values of host-guest peptides were measured by circular dichroism spectroscopy for all 20 amino acids in the X position in Gly-X-Hyp guest triplets and for 20 amino acids in the Y position in Gly-Pro-Y triplets. Interactions between X and Y residues in a Gly-X-Y guest triplet, e.g. Gly-Glu-Lys and Gly-Leu-Ala, were measured, as well as interactions between neighboring triplets Gly-X1-Y1-Gly-X2-Y2, e.g. Gly-Pro-Lys-Gly-Glu-Hyp. These thermal stability measurements were incorporated into an algorithm, designated the Collagen Stability Calculator (http://php.umdnj.edu/∼ccalcapp).30

Using the Collagen Stability Calculator, the amino acid sequence of a (Gly-X-Y)_n peptide is entered together with information about terminal group blocking, and the predicted T_m value is given. Such calculations support peptide design, since experimental T_m measurements are generally within 3–4°C of the predicted value. For collagen, the complete or partial amino acid sequence is entered online, and the calculator predicts relative stability variations along the sequence. The plot of relative stability allows visualization of biologically interesting regions, to see if they have abnormally high or low stability. For instance, Type III collagen is a homotrimer always found together with heterotrimeric Type I collagen in D periodic fibrils. In skin, Type III is a minor component (Type III:I ∼ 15%:85%), whereas it is present in much larger proportions in blood vessels and wound healing, where it has been suggested to confer flexibility on fibrils. A comparison of the stability profiles shows Type III has a larger range of stability than Type I collagen, with more regions of lower and higher stability (Figure 2). This greater modulation of stability along the Type III collagen could contribute to creation of more flexible regions and relate to its biological role.

Figure 2. Relative stability profiles for Type III (a) and Type I (b) collagens obtained from the Collagen Stability Calculator (http://php.umdnj.edu/∼ccalcapp). Bands are shown at 42°C (pink) and 35°C(green) to highlight the greater variability in stability seen along the Type III collagen molecule, homotrimeric α1(III), compared with the heterotrimer Type I collagen, [(α1(I))₂]α2(I). The cross linking sites are of low stability and are indicated by small arrows, whereas the large block arrow denotes the unique collagenase (Matrix metalloproteinase) cleavage site.

NATURAL INTERRUPTIONS IN THE GLY-X-Y REPEATING PATTERN IN NONFIBRILLAR COLLAGENS

There are now 28 distinct types of collagens, defined as structural molecules in the extracellular matrix with a triple-helix domain.31, 32 The most abundant are the fibril forming collagens (Types I, II, III, V, XI) and these molecules have Gly as every 3rd residue throughout their ∼1000 residue triple-helix domain. Although the presence of Gly as every 3rd residue is a defining feature necessary to have a classic triple helix without distortion, it is now clear that the (Gly-X-Y)_n repeat is perfect only in fibril forming collagens. Nonfibrillar collagens form networks, hexagonal arrays, anchoring fibrils, or other supramolecular structures and contain triple-helical domains of varying lengths. In all nonfibrillar collagens, the triple-helix domain has interruptions in the Gly-X-Y repeating pattern. For example, the chains of Type IV collagen in basement membranes have a 1350 residue long triple-helix domain with >20 sites at which the (Gly-X-Y)_n pattern is interrupted. Some interruptions are known to play a role in binding to tumor cell integrins12 or as specific sites of matrix metalloproteinase cleavage,33 and they have been suggested to represent flexible or kink sites involved in the network structure of Type IV collagen.34–36

A total of 354 interruptions are found in all human nonfibrillar collagens and these can been classified according to number of residues within the interruption which are flanked by (Gly-X-Y)_n triplets37, 38 (Figure 3). The repeating (Gly-X-Y)_n sequence generates a pattern with two amino acids between Gly residues. The most common interruptions in all nonfibrillar collagens are cases where there is only one residue, rather than the usual two residues between Gly residues (e.g. Gly-Pro-Hyp-Gly-Phe-Gly-Pro-Hyp) or where there are four amino acids between Gly residues (e.g. Gly-Pro-Hyp-Gly-Ala-Ala-Val-Met-Gly-Pro-Hyp) (Figure 3). A hydrophobic residue is found 70% of the time when there is only one residue between Gly residues, whereas a consensus sequence, involving a small residue followed by a hydrophobic residue, is found for interruptions with four amino acids between Gly residues.

Figure 3. Plot of the frequency of occurrence of interruptions in human nonfibrillar collagens (total of 354 interruptions) as a function of the length of residues between the (Gly-X-Y)_n repeating sequences. Examples of interruptions found in Type IV collagen chains of 1, 4, 6, and 12 residue length are shown in single amino acid code (where O represents Hyp), flanked by their adjacent Gly-X-Y triplets.

Peptides are being used to investigate the structural consequences of interruptions in the Gly-X-Y repeating sequence.37–39 Peptides can form stable triple-helices in presence of interruptions which have 1, 4, or 6 amino acids between Gly residues, but the break leads to decreased stability, decreased triple-helix content, and decreased hydrogen bonding. The identity of the residue in the middle and the nature of the surrounding Gly-X-Y sequence affects the degree of destabilization for the smallest interruptions. Interruptions with one or four amino acids between Gly residues show similar effects on the thermal stability and calorimetric enthalpy, and NMR and X-ray structures indicate that both can be incorporated into a straight triple-helix with a highly localized distortion that disrupts the register of the helix.38–40 However, NMR studies show that in a peptide with a Gly-Pro-Hyp-Gly-Phe-Gly-Pro-Hyp interruption, the hydrophobic residue Phe is located on the outside of the helix38 (Figure 4a), while a peptide with the interruption Gly-Pro-Hyp-Gly-Ala-Ala-Val-Met-Gly-Pro-Hyp has the hydrophobic Val residue located on the inside of the triple-helix, forming a small hydrophobic core in the center of triple-helix usually occupied by Gly (Figure 4b).39 The disruption of the axial register of the triple-helix by such interruptions may account for their absence in fibril forming collagens, since this would interfere with the alignment of overlapping triple-helices.

Figure 4. Cross-section view of the models obtained by NMR of a peptide with (a) Gly-Pro-Hyp-Gly-Phe-Gly-Pro-Hyp interruption and (b) a peptide with a Gly-Pro-Hyp-Gly-Ala-Ala-Val-Met-Gly-Pro-Hyp interruption.38, 39 In (a), the cross-section view shows the Gly residues are closely packed at the center with the Phe residues on the outside, whereas in (b) the Val packs in the center at the position usually occupied by Gly, creating a small hydrophobic core. (c) View of the distribution of the eight interruptions present in human Type X collagen. Based on the structures of peptide models, it is expected these interruptions will alter the register of the superhelix and will adopt structures similar to those shown in Figures 4a and 4b. Sites of matrix metalloproteinase cleavage are shown, as determined by Welgus et al. (1990).33

Information obtained from triple-helical peptides with interruptions in the Gly-X-Y repeating sequence provides a useful framework for considering the structure of nonfibrillar collagens. Type X collagen, which forms hexagonal networks at the mineralizing front of cartilage, contains eight interruptions along its 463 residue triple-helix domain (Figure 4c), and model peptide studies predict highly localized structural perturbations at these short interruptions sites. Only two of these interruption sites are cleaved by matrix metalloproteinases, Gly-Pro-Ala-Gly-Ile-Ser-Val-Pro-Gly-Lys-Pro and Gly-Pro-Ala-Gly-Ile-Ala-Thr-Lys-Gly-Leu-Asn³³, suggesting the nature of the structural perturbation influences enzyme specificity. An understanding of the effects of these interruptions on stability and structure may shed light on their biological role, and clarify how pathological interruptions differ from natural ones.

COLLAGEN DISEASES: GLY SUBSTITUTION MUTATIONS

Many common diseases, such as arthritis, diabetes, and cancer, involve abnormal degradation, cross-linking, or reactivity of collagen. There are also rare hereditary disorders that result from specific mutations in collagen genes. The clinical manifestation of these genetic diseases depends on the type, location, and function of the collagen that has the mutation: a mutation in Type I collagen, the major fibril forming collagen in bone, leads to Osteogenesis Imperfecta (OI), a disease characterized by fragile bones,41 whereas a mutation in Type IV collagen in basement membranes, the filtration barrier in the kidney glomerulus, leads to Alport Syndrome with progressive kidney failure.42 Most of the mutations are missense mutations which replace one Gly residue in the Gly-X-Y repeating pattern by a larger residue. Recently, the OI Consortium reported 682 Gly missense mutations in the α1 and α2 chains of Type I collagen, which cause bone fragility of varying degrees of severity.43 Factors that appear to influence the clinical severity include the identity of the residue replacing Gly, the chain in which the mutation occurs, the location of the mutation with respect to the C-terminus of the protein, the local environment of the mutation, and its position relative to collagen interaction sites with other matrix molecules.41, 43, 44

To isolate some of these proposed factors influencing OI clinical phenotype, peptides were designed to investigate the consequences of Gly substitutions on the stability, conformation and folding of the triple-helix. Studies on host-guest peptides (Gly-Pro-Hyp)₃-Gly-Pro-Hyp-(Gly-Pro-Hyp)₄ with the central Gly replaced by residues obtained by a single base substitution indicated that all Gly replacements were highly destabilizing and the degree of triple-helix destabilization depends on the identity of the residue replacing Gly, with the order Ala, Ser<Cys<Arg<Val<Glu, Asp.45 The crystal structure of a (Pro-Hyp-Gly)₁₀ peptide containing one Gly to Ala replacement shows a loss of direct hydrogen bonds at the Ala site, a local untwisting of the helix, and a loss of register between the two triple-helical ends.21 OI collagens with a Gly missense mutation have a delay in folding46 and model peptides indicated that such a Gly replacement will arrest C- to N-terminal triple-helix folding unless there is a strong renucleation site downstream of the mutation.47 It is interesting to note that Gly replacement mutations in nonfibrillar collagens also lead to pathological conditions even though a perfect (Gly-X-Y)_n repeating pattern is not found in their natural triple-helix.42 For instance, a Gly replacement by a Val at one site in Type IV collagen of α5 chain leads to Alport Syndrome even though the chain already contains more than 20 interruptions in the Gly-X-Y repeating pattern and the mechanism of such pathology is not well understood.

The results from peptides have provided experimental stability data and conceptual information about the effect of Gly mutations in OI collagens and permit assessment of the effects of single amino acid mutations found in diseases on local stability. Recently, activation energies of local helix unfolding calculated from the host-guest triple-helix peptide data were found to be correlated with the decrease in T_m values observed for 41 different Gly substitutions in OI collagens.48

SELF-ASSOCIATION OF PEPTIDES AS MODELS FOR HIGHER ORDER STRUCTURE

Peptides, which have been useful in clarifying features of the triple-helix at the molecular level, may also contribute to an understanding of the biologically important higher order structures of collagen. In many of the crystal structures, triple-helical peptides show quasi-hexagonal packing and intermolecular distances similar to that of collagen molecules in fibrils.21, 26 Water-mediated interactions and Hyp-Hyp intermolecular hydrogen bonding are involved in the lateral packing of peptides. Such triple-helix to triple-helix interactions are consistent with the conclusions of the Leikin laboratory, based on X-ray diffraction of tendon under different osmotic pressures, that hydration provides a major force for collagen fibrillogenesis.49

The self-association of soluble collagen-like peptides to form aggregates can also serve as models for the formation of collagen higher order structure. The triple helical peptide (Pro-Hyp-Gly)₁₀ undergoes a temperature-induced self-assembly to form an irregular, branched higher order structure (Figure 5a).50 Despite the very different morphology of the aggregated form, the properties of self-association of (Pro-Hyp-Gly)₁₀ are very similar to collagen fibril formation (Table II). Collagen fibrillogenesis and (Pro-Hyp-Gly)₁₀ self-association both follow a nucleation-growth mechanism, which is fastest at high temperatures just below their T_m. Both self-assemble at neutral, but not acid pH, are inhibited by sugars, and show a requirement for hydroxyproline. Nonspecific lateral self-association of triple-helices involving water, backbone carbonyls, and hydroxyproline may underlie the similar properties of peptide self-association and collagen fibril formation.

Figure 5. Electron micrographs of (a) self-assembled structures formed by the triple helical peptide (Pro-Hyp-Gly)₁₀; (b) collagen fibrils with an axial D period; (c) self-assembled structure of peptide containing the Type IV sequence Gly-Gln-Hyp-Gly-Leu-Hyp-Gly-Leu-Hyp flanked by Gly-Pro-Hyp units; (d) supercoiled network like structures of Type IV collagen in the basement membrane of the EHS tumor (reprinted from Am J Pathol 1988, 132, 278–291 with permission from the American Society for Investigative Pathology).

Table II. Comparison of Properties of Self-Association of (Pro-Hyp-Gly)₁₀ with Fibril Formation of Type I Collagena

Self-Association Properties	(Pro-Hyp-Gly)10	Type I Collagen
a Adapted from Kar et al., 2006.50
Mechanism	Nucleation growth	Nucleationgrowth
Seeding effect	Yes	Yes
Temperature (T)	↑T ↑ rate	↑T ↑ rate
pH effect
Acid	–	–
Neutral	+	+
Stability of associated form	Tm + 25°C	Tm + 23°C
Activation energy	137 kJ/mol	113 kJ/mol
Critical concentration	1 mM	50 μM
Reversibility	Yes	Yes
Requirement for Hyp	Yes	Yes

The disordered aggregate formed by (Pro-Hyp-Gly)₁₀ contrasts with the highly ordered supramolecular structures formed by collagens, such as the D-periodic fibrils formed by Type I collagen or the network structures of Type IV collagen in basement membranes (Figure 5b,d). It is likely that specific hydrophobic and/or electrostatic interactions are required for the specificity to determine such regular structures.51 The incorporation of a hydrophobic sequence from Type IV collagen into a peptide resulted in temperature-induced self-association to form a fiber-like structure with supercoiling and branching which resembles the networks of Type IV collagen in basement membranes (Figure 5c) (Kar et al., unpublished observations). Recently, the introduction of positively charged residues at the N-terminus and negatively charged residues at the C-terminus of a triple-helical peptide was shown to lead to formation of banded periodic fibrils which resemble collagen fibrils.52 Thus, peptides may be very useful in deciphering the basic principles of collagen self-assembly to supramolecular structures.

CONCLUSIONS

In recent years, the biomedical importance of collagen has led to increasing interest in the structure of this protein and in establishing relationships between sequence, structure, and function. Scientists like Elkan Blout pioneered the strategy of applying spectroscopic techniques to peptides to focus on basic questions that are not easily accessible in larger and more complex intact proteins. Such a peptide approach is still important today and has proved productive in defining the fundamental principles governing the stability and function of the collagen triple-helix.

Acknowledgements

We thank Teresita Silva for technical support in all work reported here. The results reported in this review represent work from former members of the laboratory, including Anton Persikov (host-guest peptides) and Angela Mohs (peptides with interruptions). NMR studies of triple-helical peptides were done by Dr. Jean Baum and Yingjie Li (Rutgers University). The X-ray crystal structures discussed here were carried out in the laboratory of Dr. Helen Berman (Rutgers University) and Dr. Jordi Bella (University of Manchester). All host-guest peptide studies were done in collaboration with Dr. John Ramshaw (CSIRO).