Section on Physical Biochemistry, OD/NICHD conducts experimental and theoretical studies of structure and function of biomolecules. In recent years we invested most of our efforts into investigation of molecular mechanisms of connective tissue disorders (Osteogenesis Imperfecta, Ehlers-Danlos Syndrome, osteoporosis, etc) and fundamental studies of underlying processes of type I collagen folding, interactions and fiber assembly. To better understand these mechanisms, we continue systematic studies of mutant collagens with substitutions of obligatory glycine residues (responsible for over 80% of moderate to lethal OI cases), studies of interesting and unusual individual OI cases, and studies of existing murine OI models (Brtl, oim and G610C). ? In the last year, in collaboration with BEMB/NICHD researchers we completed the study of collagen from a patient with a combination of OI and Ehlers-Danlos-Syndrome (EDS) symptoms resulting from an atypical point mutation (Arg888Cys in alpha1(I) chain) not involving any of the obligatory glycine residues. We confirmed formation of aberrant S-S bonds between a1(I) chains in a small fraction of molecules and found that these bonds lead to kinking and chain register shift in the molecule. The resulting structural alterations in the triple helix appear to be responsible for the observed changes in the thermal stability and N-propeptide cleavage and are likely to be the cause of the OI and EDS symptoms in the patient. ? We expanded our analysis of structural domains in collagen triple helix and their role in OI, focusing on mutations within the most flexible domain (~aa 700-800), which overlaps with one of major ligand binding regions. Glycine substitutions within this domain cause little change in the thermal stability of the triple helix, yet such mutations in the a1(I) chain are predominantly lethal. Normal thermal stability of collagen with such mutations appears to contribute to the lethality by allowing efficient folding, secretion and incorporation of mutant molecules into tissues. We also found that the structure of this domain was particularly strongly affected by the chain register shift caused by a dowstream Gly-Ala-Hyp tripled insertion in two siblings with lethal OI. In addition to destabilization of this domain, the chain register shift resulted in abnormal cleavage of collagen by matrix metalloproteases, e.g., cleavage by the catalytic domain of MMP-1 and cleavage by MMP-2 (normally incapable of cleaving type I triple helix). ? Although abnormal procollagen folding and stability are at least partially responsible for pathology, studies of our group and other researchers increasingly point to the incorporation of mutant collagen into fibrils and the resulting abnormal interactions of the fibrils with other extracellular matrix molecules as an important factor in pathology. To understand how incorporation of mutant molecules affects such interactions, we developed a new assay for visualization and measurement of binding of extracellular matrix proteins to collagen fibrils based on differential fluorescent labeling and quantitative confocal microscopy. Preliminary experiments revealed, e.g., strongly heterogeneous binding of MMP-1, decorin and biglycan, which accumulate at fibril tips, kinks, branching points and structural defects. The binding constants of decorin and biglycan at physiological conditions estimated from these measurements are highly reproducible, allowing us to begin characterization of the changes caused by collagen mutations. ? In addition to studies of folding and interactions of mutant collagens, we continued the study of type I collagen homotrimers, normally present only in fetal tissues. A polymorphism in a promoter region resulting in overproduction of the a1(I) chain and potential synthesis and incorporation of homotrimers into adult tissues was recently linked to predisposition to osteoporosis. Yet, several homozygous and compound heterozygous patients, which produce only the homotrimer form of type I collagen do not have significant bone phenotype. We, therefore, initiated a study of homotrimeric collagen from these patients. We confirmed normal posttranslational modification and processing of homotrimers, slightly increased thermal stability and slightly altered microunfolding pattern, previously reported for murine homotrimers. At the same time we discovered significantly slower cleavage of human homotrimers by MMP-1, potentially resulting in abnormal matrix remodeling and accumulation of type I homotrimers in tissues even when they comprise only a small fraction of secreted molecules. Since this pathology might offer the key to understanding the role of homotrimers in osteoporosis, its further investigation is currently under way. ? ? Another important direction of our research is closely related recognition and assembly reactions involving DNA, which play an important role in packaging of genetic material inside cells and viruses and in many other fundamentally important biological processes. In particular, we uncovered several common physical principles of interactions which govern formation, structure and physical properties of collagen and DNA aggregates. Our theory of these interactions provided 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 multiple quasi-crystalline phases of even more densely packed DNA aggregates. It also suggested that electrostatic interactions might contribute to sequence homology recognition and pairing of intact DNA double helices observed prior to genetic recombination. One of the most important and yet controversial predictions of this theory, distinguishing it from other models, was that strong azimuthally-dependent interaction should align strands and grooves on opposing surfaces of adjacent molecules. Such alignment was observed in crystals of DNA and nucleosomal particles, but in contrast to our predictions it was not traditionally expected for highly hydrated, liquid-crystalline aggregates. As proposed by Franklin and Gosling, it was expected that in such aggregates DNA should be ?relatively free from the influence of neighboring molecules, each unit being shielded by a sheath of water?. The latter assumption was used, e.g., by Watson & Crick, Wilkins et al., and Franklin & Gosling for interpretation of DNA diffraction patterns in their celebrated set of back-to-back papers on DNA structure. We revisited these classical patterns using a more detailed set reported by Zimmerman & Pheiffer in 1979. We adapted the classical Cochran-Crick-Vand diffraction theory to account for possible short-range azimuthal order in the aggregates and analyzed the observed changes in the diffraction patterns caused by varying aggregate hydration. We found that the observed changes do not affect the classical interpretation of DNA structure. However, they unequivocally reveal strong azimuthally-dependent interactions between adjacent molecules up to ~ 20 ? surface-to-surface separations in good qualitative and quantitative agreement with our predictions, lending strong support to our theory. Building on these observations, in the last year we further developed the diffraction theory to allow for sequence-dependent variation in the twist angle between adjacent base pairs and other nonidealities in the double helix structure. Based on the new theory, we extended the analysis of the Franklin &Gosling and Zimmerman & Pheiffer diffraction patterns. This analysis confirmed significant torsional deformation of the double helices in hydrated fibers predicted by the theory.
Showing the most recent 10 out of 31 publications
