The substance of the cotton fiber is the secondary cell wall that is nearly pure, microcrystalline cellulose. Normally there is about 5% moisture in cotton fiber, but variations of a few percent make differences as large as 40% in the strength, with more water resulting in higher strength. Different samples of cotton have different moisture capacities, but the location of the water in the fiber is not known. At 5%, there is about one water for every two glucose units. Improved understanding of how water interacts with cellulose at the molecular level could lead to improved products and would also be useful in other areas, including bioenergy.

To understand the role of water, it is also necessary to understand the detailed structure of the cell wall. Although the cell wall is complex, the crystal structure [1] of cellulose Iβ is an excellent start. Cotton crystallites are small enough that short lengths can be modeled, either with energy minimization or molecular dynamics, the latter being especially suited to studies with water present. Recent solvated dynamics simulations by Matthews et al. [2] found that half of the individual cellulose chains rotated somewhat from their positions in the crystal structure, along with reorientation of their O6 groups from tg to gg. In turn, that increased the unit cell a dimension by about 0.7Ĺ from the well-established experimental value. The model chains also underwent a right-handed twisting from the 2-fold screw shape, resulting in an overall twist of the crystallite. Matthews et al. quote similar work by Yui et al. [3] to reinforce these findings, as well as experimental findings of twist for the larger microfibrils [4].

So far, our work arrives at different conclusions. Crystalline β-(1,4)-linked saccharides generally prefer left-handed shapes [5]. Our HF/6-31G(d) quantum calculations on cellobiose suggest an energy penalty of about 0.4 kcal/mol for the untwisted, 2-fold screw axis shape, and B3LYP/6-31+G(d) calculations support the 2-fold shape as a minimum [6]. Energy minimization studies with MM3 led to some chain rotation, but MM4 minimizations did not. Further, 10 ns MD simulations with AMBER/GLYCAM show that increasing the number of cellooctaose chains in clusters from two to seven to 19 progressively decreases the RMS deviations from the crystal structure positions. The core region of a 19-chain cluster was mostly stable. We saw a momentary transition to gg for residue 4 of the central chain after 7 ns. The only “long term” transition of an O6 group occurred in a terminal residue after 4 ns and lasted about 0.7 ns.

[1] Y. Nishiyama P. Langan H. Chanzy. J. Am. Chem. Soc. 124, 9074 (2002). [2] J.F. Matthews C.E. Skopec P.E. Mason P. Zucccato R.W. Torget J. Sugiyama M.E. Himmel J.W. Brady. Carbohydr. Res., 341, 138(2006). [3] T. Yui S. Akiba S. Hayashi. Carbohydr. Res. 341, 2521-2530 (2006). [4] S.J. Hanley J.-F. Revol L. Godbout D. Gray. Cellulose 4, 209 (1997). [5] A.D. French G.P. Johnson. Cellulose 11, 5 (2004). [6] A.D. French G.P. Johnson. Can. J. Chem., 84, 603 (2006).