refinement of the 7/2-and 10/3-helical structures based on the fiber diffraction pattern from native collagen

K. Okuyama, X. Xu, and K. Noguchi

Faculty of Technology, Tokyo University of Agriculture and Technology, Koganei, Tokyo 184-8588, Japan ([email protected])

After fifty years since the first proposal of a triple helical structure, there are still two models for collagen with different helical symmetry and different fiber repeating period. One is the Rich and Crick model in which three strands form a left-handed 10/3-helix with an axial repeat of 28.6Å [1]. The other is our model in which three strands form a left-handed 7/2-helix with an axial repeat of 20Å [2]. The latter model was strongly supported by single crystal analyses of several collagen model peptides reported so far. On the other hand, the former model has no such supporting data other than the fiber diffraction analyses of poly(Pro-Gly-Pro) [3] and native collagen [4]. Since the quality of fiber diffraction pattern from poly(Pro-Gly-Pro) was very poor compared with that from native collagen, the former analysis is not a powerful supporting evidence for the 10/3-helical model. On the other hand, in the latter analysis, only the 10/3-helical structure was refined by using continuous layer intensities. In this study, we refined both 7/2- and 10/3-helical structures by using continuous X-ray diffraction intensities on the layer lines corresponding to their axial repeat.

The fiber diffraction pattern from kangaroo tail tendon was taken by using synchrotron radiation (BL40B2, SPring-8) and recorded on an imaging plate (RAXISIV++, Rigaku). The background subtraction and acquisition of continuous X-ray intensities on the corresponding layer lines were performed by using CCP13 softwares. According to the recent single crystal analyses of collagen model peptides at high resolution, in most of the cases, there are water molecules bound to the carbonyl oxygen of Gly and Pro, and those to the hydroxyl oxygen of Hyp. Therefore, some of these water molecules were also included in the triple helical structures. Refinement calculation was performed by WinLALS, Windows version of Linked-Atom Least-Squares (LALS) program for fiber diffraction analyses.

In both models, triple-helical structures with no bound water molecules gave fairly high discrepancy (R-) factor (~0.40) compared with that (0.27) of the previous analysis [4]. Although Pro-Hyp-Gly triplet was taken as a helical asymmetric unit, occupancies of Cg, Cd and Od atoms of imino acids were reduced to 1/3 according to the imino acid occurrences in the amino acid sequence of tendon collagen, which decreased R-factors by several percent. Furthermore, addition of water molecules decreased R-factors to 0.25 (7/2-helix) and 0.30 (10/3-helix). This fairly large improvement of the discrepancy factor by addition of water molecules was also experienced in many single crystal analyses of collagen model peptides. The obtained result showed that the 7/2-helical model can explain fiber diffraction patterns from native collagen better than, or at least as well as, the 10/3-helical model can do.

References

1 Rich, A. and Crick, F.H.C. (1961) J. Mol. Biol. 3, 483-506.

2 Okuyama, K., Takayanagi, M., Ashida, T. and Kakudo, M. (1977) Polymer J. 9, 341-343.

3 Yonath, A. and Traub, W. (1969) J. Mol. Biol. 43, 461-477.

4 Fraser, R.D.B., MacRae, T.P. and Suzuki, E. (1979) J. Mol. Biol. 129, 463-481.