Sergey Leikin, PhD, Head, Section on Physical Biochemistry
Edward L. Mertz, PhD, Staff Scientist
Se Jin Han, BS, Postbaccalaureate Fellow
Natalia V. Kuznetsova, PhD, Contractor
Elena N. Makareeva, PhD, Contractor
Mary Beth Sutter, BS, Technician

Recognition and self-assembly reactions are some of the most fundamental molecular processes in living organisms. They are responsible for, among other reactions, folding, interactions, and aggregation of proteins and nucleic acids. Our goal is to develop a better understanding of the common physical principles governing such reactions. We laid the foundation for our work with the discovery of symmetry features in the structural organization of helical biomolecules that govern assembly and stability of collagen fibers, sequence homology recognition and condensation of DNA, and a number of other reactions. We continue to build our understanding of DNA self-assembly. However, in recent years, we have shifted the focus of our research to collagen. We concentrate most of our efforts on understanding how disruption of the helical structure of collagen by osteogenesis imperfecta (OI) mutations leads to pathology of collagen matrix formation in what is a devastating "brittle bone" disease. Together with extramural and NICHD clinicians and scientists, we strive to gain better knowledge of the molecular origins of OI. We hope to use such knowledge for diagnostics, characterization, and treatment of this and related diseases, bringing our expertise in physical biochemistry and theory to clinical research and practice.

Structure and stability of the triple helix in function and pathology of collagen

Kuznetsova, Makareeva, Mertz, Leikin, Sutter; in collaboration with Barnes, Cabral, Marini

Type I collagen is a triple-helical protein that forms the stable matrix of bone, skin, and other tissues. We found, however, that the equilibrium state of collagen as well as its procollagen precursor at body temperature is a random coil rather than the triple helix. Even in a crowded environment mimicking the endoplasmic reticulum (ER), procollagen triple helix folding occurs only below 35°C. Apparently, cells have to use specialized chaperones to fold procollagen within the ER. Once procollagen is secreted from cells, it begins to unfold. Local weakening of collagen's least stable regions and concurrent cleavage of N- and C-terminal propeptides appear to trigger self-assembly of fibers. In fibers, collagen helices are protected from complete unfolding, but local unfolding/refolding of the triple helix still occurs and plays an important physiological role. Over 80 percent of moderately severe to lethal collagen mutations are substitutions of a single glycine in the (Gly-X-Y)n sequence of the triple helix. By disrupting the stability and consequently the folding of the triple helix, such mutations result in abnormal collagen production and secretion and in abnormal fibrillogenesis. We found that the effect of a glycine substitution on collagen folding and stability is determined by the structural domain within which the substitution is located rather than by the type of substituting residue. We found evidence for at least six such structural domains within the triple helix and observed that distinct changes in the folding and structure of the triple helix caused by mutations within the domains appear to correlate with variations in disease phenotype. For instance, we recently established that mutations within the N-terminal ("N-anchor") domain lead to a distinct combination of osteogenesis imperfecta and Ehlers-Danlos syndrome (EDS).

We are progressing rapidly in our understanding of the global and local structural stability and flexibility of the triple helix and how mutations affect it, but many crucial questions remain unanswered. For instance, we still do not know which chaperones fold the triple helix, how they fold it, or how they manage to fold OI collagens with stability reduced by as much as 5-6°C. Over the last year, our studies on collagens secreted by cultured fibroblasts from OI patients with non-collagen mutations indicated that a complex of cyclophilin B, prolyl-3-hydroxylase, and a cartilage-associated protein (CRTAP) might act as one such chaperone; we are currently testing this hypothesis. Our knowledge of the structural domains is also incomplete. In the last year, we determined that the region between amino acids 700 and 800 forms a large, flexible domain within which many proposed ligand-binding sites and the collagenase (MMP1) cleavage site are clustered. We know that virtually all α1(I) glycine substitutions within this domain are lethal, but we do not know the domain's exact boundaries, the role of its flexibility in ligand binding, and the molecular mechanism of the increased lethality of its mutations.

Interactions of collagen fibrils with other matrix proteins

Sutter, Makareeva, Leikin; in collaboration with Forlino, Marini, Nagase, Visse, Tenni

Interactions of collagen with ligands, particularly other extracellular matrix proteins and proteoglycans, are believed to be an important factor in connective tissue diseases. In tissue, ligands interact primarily with collagen fibrils rather than with triple helical monomers. During the past year, we completed the development of a new confocal microscopy assay with differential fluorescent labeling. It allows us to visualize and measure quantitatively the binding of various matrix proteins to individual collagen fibrils under physiological conditions. We found that MMP1 attacks damaged and poorly assembled rat-tail-tendon collagen fibers via preferential binding to microfibrils, which then become more exposed at fiber defects. Given that MMP1 is responsible for degradation and remodeling of type I collagen fibers, such preferential binding is of potential physiological importance. We also demonstrated high reproducibility of the assay for measurement of recombinant decorin binding to human collagen fibrils. We observed a four-fold decrease in binding to one of the studied OI mutants, which likely contributes to the patient phenotype. Now that we have demonstrated the concept and reliability of the assay in trial studies, we are beginning systematic measurements of interactions between collagen fibers and different matrix proteins potentially important in OI and other connective tissue disorders.

Type I collagen homotrimers and their role in osteoporosis, OI, and EDS

Han, Kuznetsova, Makareeva, Sutter, Leikin; in collaboration with Byers, McBride, Pace

While normal type I collagen is a heterotrimer of two α1(I) and one α2(I) chains, formation of α1(I) homotrimers was implicated in several cases of OI and EDS and might therefore play at least some role in age-related osteoporosis. A polymorphism in a promoter region resulting in increased production of the α1(I) chain was recently linked to increased predisposition to osteoporosis in the general population. It was proposed that such predisposition might be caused by α1(I) homotrimers, which comprise about 10 to 15 percent of all newly synthesized type I collagen in affected individuals. Nevertheless, the molecular mechanism of pathology associated with the homotrimers remains unclear. Our earlier studies revealed an approximately 2.5°C increase in thermal stability, an altered domain structure, and reduced intermolecular attraction, resulting in higher solubility and abnormal fibrillogenesis of murine α1(I) homotrimers. Initially, these abnormalities appeared to explain the severe OI observed in a patient with only homotrimeric type I collagen and in the skeletal deformities in oim mice with a similar genetic defect. But the literature later reported on several patients with only α1(I) homotrimers and no congenital osteoporosis, thus raising doubts about the earlier interpretation. Accordingly, over the last year, we returned to more detailed studies of murine and human homotrimers. We found, for example, that type I homo- and heterotrimers co-assemble into the same fibrils but that their distribution appears heterogeneous (with homotrimers observed primarily at fibril tips), suggesting a previously unappreciated effect of the homotrimers on nucleation and assembly of mixed fibers. Most importantly, we discovered that cleavage of homotrimers by MMP1 is much less efficient than that of heterotrimers. The reduced turnover and potential accumulation of "aged" homotrimers in mixed tissues offers a clue to their detrimental effect in the α1(I) promoter polymorphism. At present, we have more questions than answers, but the new approaches and results are promising.

Translational studies of patients with novel/unusual OI mutations

Kuznetsova, Makareeva, Sutter, Leikin; in collaboration with Barnes, Cabral, Marini

To gain further insight into the molecular mechanisms of bone pathology associated with abnormal folding, function, and interactions of collagens, we continued our collaboration with clinical researchers at the Bone and Extracellular Matrix Branch of NICHD on the characterization of pathology in patients with unusual OI mutations and/or phenotypes. Our goal is to find and characterize molecular abnormalities of collagen from such patients. Earlier studies revealed that the molecular mechanism of OI/EDS is attributable to mutations within the N-anchor domain of the triple helix. During the past year, we studied type I collagen from a patient with a familial R888C substitution in the α1(I) chain and a peculiar combination of OI and EDS symptoms. We found that approximately half the molecules containing two mutant chains form an intramolecular S-S dimer between the Cys-888 residues. Formation of the dimer results in a kink in the triple helix, altered stability, and a register shift between the chains, propagating all the way to the N-terminal end of the molecule and resulting in abnormal N-propeptide cleavage. Apparently, the dimer can form only while the chains are unfolded. If the dimer formations do not occur before the triple helix folding at the mutation site, further folding proceeds normally without the kink and the chain register shift. Such molecules are normally processed by the N-proteinase and do not thereafter form the S-S dimer. The abnormal N-propeptide cleavage might be partially responsible for the patient's EDS-like joint laxity. In addition to permitting us to develop both a better understanding of the clinical phenotype and an improved treatment strategy, the study of the patient's collagen might offer new insights into the potential role of aberrant S-S bonds in more common OI cases with Gly→Cys substitutions.

Insights into OI mechanisms from murine models

Mertz, Han, Kuznetsova, Leikin; in collaboration with Forlino, Marini, McBride, Uveges

Murine OI models offer unique opportunities for more systematic studies of the molecular mechanism of pathology in animals with the same mutation but different genetic background, age, sex, and so forth. We currently work with all three existing murine models: the oim mouse with nonfunctional α2(I) chains, the Brtl IV mouse with a knockin G349C substitution in the α1(I) chain, and a new mouse model with a knockin G610C substitution in the α2(I) chain. Our earlier studies revealed several changes in the physical and chemical properties of collagen from oim and G610C animals, including altered thermal stability (increased in oim and decreased in G610C) and abnormal collagen-collagen interactions. However, the Brtl IV animals with no such abnormalities in type I collagen exhibit a more severe bone phenotype. The most important molecular abnormality that we discovered in these animals is selective retention and intracellular degradation of molecules with a single mutant chain. The retention of the mutant collagen within ER produces visible swelling of ER cisternae, potentially resulting in ER stress and malfunction of collagen-producing cells, including osteoblasts. To understand more fully how such retention might result in bone failure in Brtl IV animals, we performed a more detailed characterization of the structure and composition of the animals' bone by infrared microspectroscopy. Preliminary results indicate that the mutation has no significant effect on the collagen and mineral composition and alignment in lamellar bone. However, non-lamellar layers produced during initial, faster growth of long bones in mutant animals have abnormal collagen and mineral organization, which might be responsible for bone brittleness. These findings appear to be consistent with the hypothesis that more severe cellular malfunction is associated with faster collagen production and with the observed gradual normalization of bone strength in post-pubertal animals. However, more studies are needed to verify our hypothesis.

Infra-red microspectroscopy of connective tissues

Mertz, Leikin; in collaboration with Forlino, Marini, McBride, Uveges

After investing several years into pushing the precision limits of analytical micro-Fourier Transform InfraRed spectroscopy (FTIR), we completed the development and testing of a novel temperature-controlled, flow-through optical cell for micro-FTIR, which, for the first time, also permits analysis of tissue samples at physiological solvation. We tested the apparatus by measuring interactions of phosphate and sulfate ions with collagen fibrils, assessing collagen backbone dynamics at different hydrations, characterizing the structure and orientation of bound water within collagen fibrils, and determining hydroxyl content in bone mineral. These and other measurements indicate that the sensitivity and reproducibility of our micro-FTIR setup appear to be at least 100 times better than those of any commercially available instrument and thus represent a significant expansion of the capabilities of IR technology in chemistry and biology (a U.S. patent application is pending). In addition to the study of anomalies in the composition and structural properties of bone from murine OI models discussed above, we began to study cartilage and bone disorders caused by disruptive mutations in a transporter responsible for sulfate uptake by connective tissue cells. We demonstrated that infra-red microspectroscopy can measure the extent, spatial distribution, and even type of sulfation of connective tissue glycosaminoglycans across tissue sections, thereby permitting us to determine, for example, otherwise undetectable differences between the wild-type and heterozygous carriers of a recessive sulfate transporter mutation in a murine model. The ability of the technique to characterize intracellular composition of glycosaminoglycans at different stages of cell differentiation appears to be particularly promising for future studies of sulfate transporter deficiency and other connective tissue disorders.

DNA interactions: new insights from classical diffraction patterns

Leikin; in collaboration with Kornyshev, Lee, Wynveen, Zimmerman

Interactions within the DNA molecule play an important role in the packaging of genetic material inside cells and viruses and in many other fundamentally important biological processes. Earlier, we found that such interactions are intimately related to the finer details of the molecular structure of DNA, particularly the helical nature of DNA's surface charge pattern. Our theory of these interactions suggested explanations for the observed counter-ion specificity of DNA condensation, DNA overwinding from 10.5 base pairs (bp) per helical turn in solution to 10.0 bp/turn in aggregates, nontrivial cholesteric pitch behavior upon compression of DNA aggregates, subsequent transition from the cholesteric to hexagonal (hexatic) phase, and several quasi-crystalline phases of even more densely packed DNA aggregates. In addition, we found that electrostatic interactions might contribute to sequence homology recognition and pairing of intact DNA double helices observed before genetic recombination. One of the most important and yet controversial predictions of our theory, which is what distinguishes it from other models, was that strong azimuthally dependent interactions should align strands and grooves on opposing surfaces of adjacent molecules, resulting in significant torsional deformation of DNA in aggregates. To test this prediction, we revisited the classical diffraction patterns from hydrated DNA aggregates reported by Zimmerman and Pheiffer in 1979. First, we adapted the classical Cochran-Crick-Vand (CCV) diffraction theory to account for possible short-range azimuthal order and reanalyzed the changes in the positions of the diffraction peaks observed upon aggregate hydration. In good qualitative and quantitative agreement with our predictions, the analysis unequivocally revealed strong azimuthally dependent interactions between adjacent molecules with up to about 20Å surface-to-surface separations. Next, we developed a theory of X-ray diffraction on non-ideal helices that incorporates accumulating effects of sequence-dependent DNA structure and thermal disorder beyond the Debye-Waller factors. Comparison of the corresponding, expected changes in the intensity and width of the diffraction peaks with the observed diffraction patterns confirmed the predicted DNA deformation upon aggregation, further supporting the theory of structure-dependent interactions between DNA.

Supercoiling of biological helices

Leikin; in collaboration with Kornyshev, Lee, Wynveen

Most biological helices tend to wind around each other, forming coiled coils or supercoils. The three left-handed helical chains of the collagen triple helix are wound together in a right-handed helical supercoil. The two right-handed α-helical chains in myosin filaments coil around each other, forming an extended, left-handed coiled coil. Circular DNA fragments form a variety of supercoiled structures with the help of specialized enzymes (topoisomerases). Researchers have paid surprisingly little attention to the physics of interactions governing the structural hierarchy in these coils. For instance, we still do not know how the pitch of the molecular helix (e.g., DNA) affects the energy and structure of the supercoil formed by two such helices tightly wound around each other. We do not know how the supercoil structure (e.g., in collagen) affects its stability and mechanical properties. In 1953, Francis Crick took the first step toward a rigorous theory of supercoils. He derived general expressions for the corresponding structure factors. However, the Crick theory found application only in the context of X-ray diffraction and did not extend to investigating relationships between the structural factors and physical properties of coiled coils (supercoils). Over the last year, we found how the Crick structure factors can be used to calculate supercoiling energy as well as interactions between several coiled coils. The initial results for the simplest model of electrostatic interactions in a tight, two-stranded, zwitterionic coiled coil with a large supercoil pitch revealed some new, qualitative physics. For instance, the model suggested, just as a first step, the origin of the generally observed opposite handedness of a "relaxed" supercoil and its helical strands and indicated when and why the handedness might become the same. The development of a complete theory is a more challenging task that will require significant additional effort.

COLLABORATORS

Aileen M. Barnes, MS, Bone and Extracellular Matrix Branch, NICHD, Bethesda, MD
Peter H. Byers, MD, University of Washington, Seattle, WA
Wayne A. Cabral, AB, Bone and Extracellular Matrix Branch, NICHD, Bethesda, MD
Alain Colige, PhD, Université de Liège, Liège, Belgium
Antonella Forlino, PhD, Università degli studi di Pavia, Pavia, Italy
Alexei A. Kornyshev, PhD, Imperial College, London, UK
Lev I. Krishtallik, PhD, Frumkin Institute of Electrochemistry, Moscow, Russia
Dominic J. Lee, PhD, Imperial College, London, UK
Wolfgang Losert, PhD, University of Maryland, College Park, MD
Joan C. Marini, MD, PhD, Bone and Extracellular Matrix Branch, NICHD, Bethesda, MD
Daniel J. McBride, Jr., MD, PhD, University of Maryland School of Medicine, Baltimore, MD
Hideaki Nagase, PhD, Kennedy Institute of Rheumatology Division, Imperial College, London, UK
James M. Pace, PhD, University of Washington, Seattle, WA
John M. Seddon, PhD, Imperial College, London, UK
Ruggero Tenni, PhD, Università degli Studi di Pavia, Pavia, Italy
Thomas E. Uveges, PhD, Bone and Extracellular Matrix Branch, NICHD, Bethesda, MD
Robert Visse, PhD, Kennedy Institute of Rheumatology Division, Imperial College, London, UK
Aaron Wynveen, PhD, Imperial College, London, UK
Steven B. Zimmerman, PhD, Laboratory of Molecular Biology, NIDDK, Bethesda, MD

For further information, contact leikins@mail.nih.gov.