Prospects and limitations of the rational engineering of fibrillar collagens

Recombinant collagens are attractive proteins for a number of biomedical applications. To date, significant progress was made in the large-scale production of nonmodified recombinant collagens; however, engineering of novel collagen-like proteins according to customized specifications has not been addressed. Herein we investigated the possibility of rational engineering of collagen-like proteins with specifically assigned characteristics. We have genetically engineered two DNA constructs encoding multi-D4 collagens defined as collagen-like proteins, consisting primarily of a tandem of the collagen II D4 periods that correspond to the biologically active region. We have also attempted to decrease enzymatic degradation of novel collagen by mutating a matrix metalloproteinase 1 cleavage site present in the D4 period. We demonstrated that the recombinant collagen alpha-chains consisting predominantly of the D4 period but lacking most of the other D periods found in native collagen fold into a typical collagen triple helix, and the novel procollagens are correctly processed by procollagen N-proteinase and procollagen C-proteinase. The nonmutated multi-D4 collagen had a normal melting point of 41 degrees C and a similar carbohydrate content as that of control. In contrast, the mutant multi-D4 collagen had a markedly lower thermostability of 36 degrees C and a significantly higher carbohydrate content. Both collagens were cleaved at multiple sites by matrix metalloproteinase 1, but the rate of hydrolysis of the mutant multi-D4 collagen was lower. These results provide a basis for the rational engineering of collagenous proteins and identifying any undesirable consequences of altering the collagenous amino acid sequences.     

The crucial role of trimerization domains in collagen folding

Collagens contain large numbers of Gly-Xaa-Yaa peptide repeats that form the characteristic triple helix, where the individual chains fold into a polyproline II helix and three of these helices form a right-handed triple helix. For the proper folding of the triple helix collagens contain trimerization domains. These domains ensure a single starting point for triple helix formation and are also responsible for the chain selection in heterotrimeric collagens. Trimerization domains are non-collagenous domains of very different structures. The size of trimerization domains varies from 35 residues in type IX collagen to around 250 residues for the fibrillar collagens. These domains are not only crucial for biological functions, but they are also attractive tools for generating recombinant collagen fragments of interest as well as for general use in protein engineering and biomaterial design. Here we review the current knowledge of the structure and function of these trimerization domains.     

Production of human type I collagen in yeast reveals unexpected new insights into the molecular assembly of collagen trimers

Substantial evidence supports the role of the procollagen C-propeptide in the initial association of procollagen polypeptides and for triple helix formation. To evaluate the role of the propeptide domains on triple helix formation, human recombinant type I procollagen, pN-collagen (procollagen without the C-propeptides), pC-collagen (procollagen without the N-propeptides), and collagen (minus both propeptide domains) heterotrimers were expressed in Saccharomyces cerevisiae. Deletion of the N- or C-propeptide, or both propeptide domains, from both proalpha-chains resulted in correctly aligned triple helical type I collagen. Protease digestion assays demonstrated folding of the triple helix in the absence of the N- and C-propeptides from both proalpha-chains. This result suggests that sequences required for folding of the triple helix are located in the helical/telopeptide domains of the collagen molecule. Using a strain that does not contain prolyl hydroxylase, the same folding mechanism was shown to be operative in the absence of prolyl hydroxylase. Normal collagen fibrils were generated showing the characteristic banding pattern using this recombinant collagen. This system offers new opportunities for the study of collagen expression and maturation.     

A peptide study of the relationship between the collagen triple-helix and amyloid

Type XXV collagen, or collagen‐like amyloidogenic component, is a component of amyloid plaques, and recent studies suggest this collagen affects amyloid fibril elongation and has a genetic association with Alzheimer's disease. The relationship between the collagen triple helix and amyloid fibrils was investigated by studying peptide models, including a very stable triple helical peptide (Pro‐Hyp‐Gly)₁₀, an amyloidogenic peptide GNNQQNY, and a hybrid peptide where the GNNQQNY sequence was incorporated between (GPO)_n domains. Circular dichroism and nuclear magnetic resonance (NMR) spectroscopy showed the GNNQQNY peptide formed a random coil structure, whereas the hybrid peptide contained a central disordered GNNQQNY region transitioning to triple‐helical ends. Light scattering confirmed the GNNQQNY peptide had a high propensity to form amyloid fibrils, whereas amyloidogenesis was delayed in the hybrid peptide. NMR data suggested the triple‐helix constraints on the GNNQQNY sequence within the hybrid peptide may disfavor the conformational change necessary for aggregation. Independent addition of a triple‐helical peptide to the GNNQQNY peptide under aggregating conditions delayed nucleation and amyloid fibril growth. The inhibition of amyloid nucleation depended on the Gly‐Xaa‐Yaa sequence and required the triple‐helix conformation. The inhibitory effect of the collagen triple‐helix on an amyloidogenic sequence, when in the same molecule or when added separately, suggests Type XXV collagen, and possibly other collagens, may play a role in regulating amyloid fibril formation. © 2012 Wiley Periodicals, Inc. Biopolymers 97: 795–806, 2012.     

Presence of a thermostable domain in the helical part of the type I collagen molecule and its role in the mechanism of triple helix folding.

A study has been done of the effect of neutral salts (NaCl and CaCl2) on the mechanism of type I collagen triple helix folding and unfolding in concentrated acetic acid solutions (2-8.8 M). It is shown that in these conditions, thermoabsorption and secondary structure change in heated solutions proceed in two consecutive stages. Salts exert a different destabilizing effect on different sites of the macromolecule, promoting the detection of a thermostable domain. The presence of a thermostable domain permits one to carry out reversible denaturation of collagen and to study the mechanism of the triple helix folding. Proceeding from the mechanism of the triple helix folding, an assumption has been made on the localization of the thermostable domain and its biological role.     

Computation of 3D structure and distribution of helical parameters for D-period segments of human fibrillar collagen III

The structures of D-period segments of collagen (234 amino residues or ～1/4 of whole length) are established by methods of molecular mechanics and geometry analysis. Each D-period segment proves to have a unique spatial structure. The distributions of local helical parameters along the molecule are calculated. It is found that a second hydrogen bond is formed in every case when the second residue in the tripeptide G-X-Y is an amino acid. With such a combined H-bond network, all the peptide CO groups of glycines and of the third residues in tripeptides have quasi-equivalent positions on the surface of the collagen molecule. The local deformations of the polyproline II helix in the triple complex give rise to the observed modulation of structure at the macromolecular level, which may be important for the mutual orientation of collagen molecules during fibrillogenesis.     

Self‐association of streptococcus pyogenes collagen‐like constructs into higher order structures

A number of bacterial collagen‐like proteins with Gly as every third residue and a high Pro content have been observed to form stable triple‐helical structures despite the absence of hydroxyproline (Hyp). Here, the high yield cold‐shock expression system is used to obtain purified recombinant collagen‐like protein (V‐CL) from Streptococcus pyogenes containing an N‐terminal globular domain V followed by the collagen triple‐helix domain CL and the modified construct with two tandem collagen domains V‐CL‐CL. Both constructs and their isolated collagenous domains form stable triple‐helices characterized by very sharp thermal transitions at 35–37°C and by high values of calorimetric enthalpy. Procedures for the formation of collagen SLS crystallites lead to parallel arrays of in register V‐CL‐CL molecules, as well as centrosymmetric arrays of dimers joined at their globular domains. At neutral pH and high concentrations, the bacterial constructs all show a tendency towards aggregation. The isolated collagen domains, CL and CL‐CL, form units of diameter 4–5 nm which bundle together and twist to make larger fibrillar structures. Thus, although this S. pyogenes collagen‐like protein is a cell surface protein with no indication of participation in higher order structure, the triple‐helix domain has the potential of forming fibrillar structures even in the absence of hydroxyproline. The formation of fibrils suggests bacterial collagen proteins may be useful for biomaterials and tissue engineering applications.     

Helix packing in the lactose permease of Escherichia coli determined by site-directed thiol cross-linking: helix I is close to helices V and XI

Coexpression of lacY gene fragments encoding the first two transmembrane domains and the remaining 10 transmembrane domains complement in the membrane and catalyze active lactose transport [Wrubel, W., Stochaj, U., et al. (1990) J. Bacteriol. 172, 5374-5381]. Accordingly, a plasmid encoding contiguous, nonoverlapping permease fragments with a discontinuity in the cytoplasmic loop between helices II and III (loop II/III) was constructed (N2C10 permease). When Phe27 (helix I) is replaced with Cys, cross-linking is observed with two native Cys residues, Cys148 (helix V) and Cys355 (helix XI). Cross-linking of a Cys residue at position 27 to Cys148 occurs with N,N'-o-phenylenedimaleimide (o-PDM; rigid 6 A), with N,N'-p-phenylenedimaleimide (p-PDM; rigid 10 A), or with 1,6-bis(maleimido)hexane (BMH; flexible 16 A). On the other hand, with the Phe27-->Cys/Cys355 pair, cross-linking is observed with p-PDM or BMH but not o-PDM. In neither case is cross-linking observed with iodine. It is suggested that a Cys residue at position 27 is within 6-10 A from Cys148 and about 10 A from Cys355. The results provide evidence for proximity between helix I and helices V or XI in the tertiary structure of the permease. In addition, the findings are consistent with other results [Venkatesan, P., Kaback, H. R. (1998) Proc. Natl. Acad. Sci. U.S.A. 95, 9802-9807] indicating that Glu126 (helix IV) and Arg144 (helix V) are within the membrane, rather than at the membrane-water interface on the cytoplasmic face.     

Mechanical implications of the domain structure of fiber-forming collagens: comparison of the molecular and fibrillar flexibilities of the alpha1-chains found in types I-III collagen

Fibrillar collagens store, transmit and dissipate elastic energy during tensile deformation. Results of previous studies suggest that the collagen molecule is made up of alternating rigid and flexible domains, and extension of the flexible domains is associated with elastic energy storage. In this study, we model the flexibility of the alpha1-chains found in types I-III collagen molecules and microfibrils in order to understand the molecular basis of elastic energy storage in collagen fibers by analysing the areas under conformational plots for dipeptide sequences. Results of stereochemical modeling suggest that the collagen triple helix is made up of rigid and flexible domains that alternate with periods that are multiples of three amino acid residues. The relative flexibility of dipeptide sequences found in the flexible regions is about a factor of five higher than that found for the flexibility of the rigid regions, and the flexibility of types II and III collagen molecules appears to be higher than that found for the type I collagen molecule. The different collagen alpha1-chains were compared by correlating the flexibilities. The results suggest that the flexibilities of the alpha1-chains of types I and III collagen are more closely related than the flexibilities of the alpha1-chains in types I and II and II and III collagen. The flexible domains found in the alpha1-chains of types I-III collagen were found to be conserved in the microfibril and had periods of about 15 amino acid residues and multiples thereof. The flexibility profiles of types I and II collagen microfibrils were found to be more highly correlated than those for types I and III and II and III. These results suggest that the domain structure of the alpha1-chains found in types I-III collagen is an efficient means for storage of elastic energy during stretching while preserving the triple helical structure of the overall molecule. It is proposed that all collagens that form fibers are designed to act as storage elements for elastic energy. The function of fibers rich in type I collagen is to store and then transmit this energy while fibers rich in types II and III collagen may store and then reflect elastic energy for dissipation through viscous fibrillar slippage. Impaired elastic energy storage by extracellular matrices may lead to cellular damage and changes in signaling by mechanochemical transduction at the extracellular matrix-cell interface.     

Structure of the integrin alpha2beta1-binding collagen peptide

We have determined the 1.8 Å crystal structure of a triple helical integrin-binding collagen peptide (IBP) with sequence (Gly-Pro-Hyp)₂-Gly-Phe-Hyp-Gly-Glu-Arg-(Gly-Pro-Hyp)₃. The central GFOGER hexapeptide is recognised specifically by the integrins α2β1, α1β1, α10β1 and α11β1. These integrin/collagen interactions are implicated in a number of key physiological processes including cell adhesion, cell growth and differentiation, and pathological states such as thrombosis and tumour metastasis. Comparison of the IBP structure with the previously determined structure of an identical collagen peptide in complex with the integrin α2-I domain (IBP^c) allows the first detailed examination of collagen in a bound and an unbound state. The IBP structure shows a direct and a water-mediated electrostatic interaction between Glu and Arg side-chains from adjacent strands, but no intra-strand interactions. The interactions between IBP Glu and Arg side-chains are disrupted upon integrin binding. A comparison of IBP and IBP^c main-chain conformation reveals the flexible nature of the triple helix backbone in the imino-poor GFOGER region. This flexibility could be important to the integrin-collagen interaction and provides a possible explanation for the unique orientation of the three GFOGER strands observed in the integrin-IBP^c complex crystal structure.     

Interruptions in the collagen repeating tripeptide pattern can promote supramolecular association

The standard collagen triple‐helix requires a perfect (Gly‐Xaa‐Yaa)_n sequence, yet all nonfibrillar collagens contain interruptions in this tripeptide repeating pattern. Defining the structural consequences of disruptions in the sequence pattern may shed light on the biological role of sequence interruptions, which have been suggested to play a role in molecular flexibility, collagen degradation, and ligand binding. Previous studies on model peptides with 1‐ and 4‐residue interruptions showed a localized perturbation within the triple‐helix, and this work is extended to introduce natural collagen interruptions up to nine residue in length within a fixed (Gly‐Pro‐Hyp)_n peptide context. All peptides in this set show decreases in triple‐helix content and stability, with greater conformational perturbations for the interruptions longer than five residue. The most stable and least perturbed structure is seen for the 5‐residue interruption peptide, whose sequence corresponds to a Gly to Ala missense mutation, such as those leading to collagen genetic diseases. The triple‐helix peptides containing 8‐ and 9‐residue interruptions exhibit a strong propensity for self‐association to fibrous structures. In addition, a small peptide modeling only the 9‐residue sequence within the interruption aggregates to form amyloid‐like fibrils with antiparallel β‐sheet structure. The 8‐ and 9‐residue interruption sequences studied here are predicted to have significant cross‐β aggregation potential, and a similar propensity is reported for ～10% of other naturally occurring interruptions. The presence of amyloidogenic sequences within or between triple‐helix domains may play a role in molecular association to normal tissue structures and could participate in observed interactions between collagen and amyloid.     

Collagen’s primary structure determines collagen:HSP47 complex stoichiometry

Collagens play important roles in development and homeostasis in most higher organisms. In order to function, collagens require the specific chaperone HSP47 for proper folding and secretion. HSP47 is known to bind to the collagen triple helix, but the exact positions and numbers of binding sites are not clear. Here, we employed a collagen II peptide library to characterize high-affinity binding sites for HSP47. We show that many previously predicted binding sites have very low affinities due to the presence of a negatively charged amino acid in the binding motif. In contrast, large hydrophobic amino acids such as phenylalanine at certain positions in the collagen sequence increase binding strength. For further characterization, we determined two crystal structures of HSP47 bound to peptides containing phenylalanine or leucine. These structures deviate significantly from previously published ones in which different collagen sequences were used. They reveal local conformational rearrangements of HSP47 at the binding site to accommodate the large hydrophobic side chain from the middle strand of the collagen triple helix and, most surprisingly, possess an altered binding stoichiometry in the form of a 1:1 complex. This altered stoichiometry is explained by steric collisions with the second HSP47 molecule present in all structures determined thus far caused by the newly introduced large hydrophobic residue placed on the trailing strand. This exemplifies the importance of considering all three sites of homotrimeric collagen as independent interaction surfaces and may provide insight into the formation of higher oligomeric complexes at promiscuous collagen-binding sites.     

The C-propeptide domain of procollagen can be replaced with a transmembrane domain without affecting trimer formation or collagen triple helix folding during biosynthesis.

The folding and assembly of procollagen occurs within the cell through a series of discrete steps leading to the formation of a stable trimer consisting of three distinct domains: the N-propeptide, the C-propeptide and the collagen triple helix flanked at either end by short telopeptides. We have established a semi-permeabilized cell system which allows us to study the initial stages in the folding and assembly of procollagen as they would occur in the intact cell. By studying the folding and assembly of the C-propeptide domain in isolation, and a procollagen molecule which lacks the C-propeptide, we have shown that this domain directs the initial association event and is required to allow triple helix formation. However, the essential function of this domain does not include triple helix nucleation or alignment, since this can occur when the C-propeptide is substituted with a single transmembrane domain. Also the telopeptide region is not involved in triple helix nucleation; however, a minimum of two hydroxyproline-containing Gly-X-Y triplets at the C-terminal end of the triple helix are required for nucleation to occur. Thus, the C-propeptide is required solely to ensure association of the monomeric chains; once these are brought together, the triple helix is able to nucleate and fold to form a correctly aligned triple helix.     

A microcalorimetric and electron spin resonance study of the influence of UV radiation on collagen

It was demonstrated microcalorimetrically that UV radiation cardinally changed the behavioral pattern of the temperature dependence of heat capacity of a collagen solution and decreased the enthalpy of collagen denaturation. It was discovered by electron spin resonance (ESR) that the primary products of UV-irradiated acid-soluble collagen were atomic hydrogen and the anion radical of acetic acid. The latter, when exposed to long-wavelength UV light, converted into the methyl radical, which interacted with acetic acid to produce the acetic acid radical. These free radicals interacted with the collagen molecule, leading to the appearance of seven superfine components with the split ΔH = 1.13 mT in the ESR spectrum. It was assumed that this spectrum is related to the free radical that occurred in the proline residue of the collagen molecule, which, in this particular case, was the major defect in the triple helix of collagen that caused instability of the macromolecule.     

The Role of Cross-Chain Ionic Interactions for the Stability of Collagen Model Peptides

The contribution of ionic interactions to the stability of the collagen triple helix was studied using molecular dynamics (MD) simulations and biophysical methods. To this end, we examined the stability of a host-guest collagen model peptide, Ac-GPOGPOGPYGXOGPOGPO-NH₂, substituting KGE, KGD, EGK, and DGK for the YGX sequence. All-atom, implicit solvent MD simulations show that the fraction of cross-chain ionic interactions formed is different, with the most pronounced in the KGE and KGD sequences, and the least in the DGK sequence. To test whether the fraction of cross-chain ionic interactions correlates with the stability, experimental measurements of thermostability were done using differential scanning calorimetry and circular dichroism spectroscopy. It was found that the melting temperature is very similar for KGE and KGD peptides, whereas the EGK peptide has lower thermostability and the DGK peptide is the least thermostable. A novel, to our knowledge, computational protocol termed temperature-scan MD was applied to estimate the relative stabilities of the peptides from MD simulations. We found an excellent correlation between transition temperatures obtained from temperature-scan MD and those measured experimentally. These results suggest the importance of cross-chain ionic interactions for the stability of collagen triple helix and confirm the utility of MD simulations in predicting interactions and stability in this system.     

The Role of Cross-Chain Ionic Interactions for the Stability of Collagen Model Peptides

The contribution of ionic interactions to the stability of the collagen triple helix was studied using molecular dynamics (MD) simulations and biophysical methods. To this end, we examined the stability of a host-guest collagen model peptide, Ac-GPOGPOGPYGXOGPOGPO-NH₂, substituting KGE, KGD, EGK, and DGK for the YGX sequence. All-atom, implicit solvent MD simulations show that the fraction of cross-chain ionic interactions formed is different, with the most pronounced in the KGE and KGD sequences, and the least in the DGK sequence. To test whether the fraction of cross-chain ionic interactions correlates with the stability, experimental measurements of thermostability were done using differential scanning calorimetry and circular dichroism spectroscopy. It was found that the melting temperature is very similar for KGE and KGD peptides, whereas the EGK peptide has lower thermostability and the DGK peptide is the least thermostable. A novel, to our knowledge, computational protocol termed temperature-scan MD was applied to estimate the relative stabilities of the peptides from MD simulations. We found an excellent correlation between transition temperatures obtained from temperature-scan MD and those measured experimentally. These results suggest the importance of cross-chain ionic interactions for the stability of collagen triple helix and confirm the utility of MD simulations in predicting interactions and stability in this system.     

Proximity between Glu126 and Arg144 in the lactose permease of Escherichia coli.

Evidence has been presented [Venkatesan, P., and Kaback, H. R. (1998) Proc. Natl. Acad. Sci. U.S.A. 95, 9802-9807] that Glu126 (helix IV) and Arg144 (helix V) which are critical for substrate binding in the lactose permease of Escherichia coli are charge paired and therefore in close proximity. To test this conclusion more directly, three different site-directed spectroscopic techniques were applied to permease mutants in which Glu126 and/or Arg144 were replaced with either His or Cys residues. (1) Glu126-->His/Arg144-->His permease containing a biotin acceptor domain was purified by monomeric avidin affinity chromatography, and Mn(II) binding was assessed by electron paramagnetic resonance spectroscopy. The mutant protein binds Mn(II) with a KD of about 40 microM at pH 7.5, while no binding is observed at pH 5.5. In addition, no binding is detected with Glu126-->His or Arg144-->His permease. (2) Permease with Glu126-->Cys/Arg144-->Cys and a biotin acceptor domain was purified, labeled with a thiol-specific nitroxide spin-label, and shown to exhibit spin-spin interactions in the frozen state after reconstitution into proteoliposomes. (3) Glu126-->Cys/Arg144-->Cys permease with a biotin acceptor domain was purified and labeled with a thiol-specific pyrene derivative, and fluorescence spectra were obtained after reconstitution into lipid bilayers. An excimer band is observed with the reconstituted E126C/R144C mutant, but not with either single-Cys mutant or when the single-Cys mutants are mixed prior to reconstitution. The results provide strong support for the conclusion that Glu126 (helix IV) and Arg144 (helix V) are in close physical proximity.     

Chondrocalcin is identical with the C-propeptide of type II procollagen.

The primary structure of the cartilage matrix molecule chondrocalcin has been found to be identical with that of the C-propeptide of type II procollagen by comparing sequence analyses of the N-terminal regions and of tryptic peptides derived from chondrocalcin. This implies that in type II collagen the C-propeptide of type II collagen is employed not only in the assembly of the triple helix of type II collagen, as demonstrated previously, but in calcifying cartilage it may also be involved in those events leading to cartilage calcification, as earlier indicated.     

The presence of a thermally stable domain in the spiral part of a collagen molecule

It is shown that in 0,5 M NaCl 8 M CH3COOH heat absorption and the second structure alteration in a heated solution proceed between two stages following one another, and besides, salts not only decrease the macromolecule denaturation temperature in total, but produce different destabilization effect on different regions. The presence of the thermostable domain in the macromolecule helical part permits investigation of the folding mechanism of the triple collagen helix under partial denaturation. The localization and biological role of the stable domain in the triple helix formation are suggested.     

Thiol cross-linking of transmembrane domains IV and V in the lactose permease of Escherichia coli

Glu126 (helix IV) and Arg144 (helix V) in the lactose permease of Escherichia coli are critical for substrate binding and transport, and the two residues are in close proximity and charge-paired. By using a functional permease construct with two tandem factor Xa protease sites in the cytoplasmic loop between helices IV and V, it is shown here that Cys residues in place of Glu126 and Arg144, as well as Ala122 and Val149, spontaneously form disulfide bonds in situ, indicating that this region of transmembrane domains IV and V is in the alpha-helical conformation. To determine if the local structure or environment is perturbed by the presence of an unpaired charge, either Glu126 or Arg144 or both were replaced with Ala, and cross-linking between the Cys pair Ala122-->Cys/Val149-->Cys was studied. Ala replacement for Arg144 causes a marked decrease in cross-linking, while Ala replacement for Glu126 alone or for both Glu126 and Arg144 has little effect. The data provide strong support for the argument that Glu126 and Arg144 are within close proximity and suggest that an unpaired carboxylate at position 126 causes a structural change at the interface between helices IV and V.