Structural stability of the solvated cellulose IIII crystal models: a molecular dynamics study

Abstract

Molecular dynamics (MD) simulations of cellulose III I crystal models have been carried out. The crystal models were composed by either 24 or 48 cellooligomers consisting of either 20 or 40 residues and were surrounded by waters in a periodic boundary box. Two base plane types differing in a constituent crystal lattice plane, (0 -1 0) 9 (0 1 0) and (1 0 0) 9 (0 1 0), were additionally considered. Among the resulting eight crystal models, an overall structure conversion was observed for the seven models. The final structures had a triclinic-like chain arrangement involving one-quarter staggering chains with respect to its axis. The successive, local transformation involving cooperative bends in cellooligomers was observed during the structure conversion. Only the 48 9 20-mer model having the (0 -1 0) lattice plane retained the original crystal structure throughout a 2.5-ns simulation. The MD simulations with an implicit solvent system and a vacuum system were also performed to asses a solvent effect on the structure conversion.

Structure conversions of cellulose IIII crystal models in solution state: a molecular dynamics study

“…Since the final chain packing structure could not be indexed with any set of unit cell parameters of known cellulose allomorphs, it is uncertain whether the conversion scheme could, at least partly, describe the actual crystalline transition from III I to Ib. The most serious limitation for the finite crystal model was the limited fiber lengths, with the degree of polymerization (DP) ranging from 20 to 40 in the previous MD studies (Yui and Hayashi 2007;Yui and Hayashi 2009). This may have decreased the three-dimensional stability of the crystal models in solution state.…”

“…The labeling of each crystal model and their corresponding ab projection dimensions and DP are listed in Table 1. Further details, such as the number of water molecules and the size of the periodic box, are provided in Table 1 of our previous paper (Yui and Hayashi 2009). Fig.…”

“…This allowed thermal vibrations of the atoms at the original positions during the heating and equilibration dynamics. It should be noted that the minimization and equilibration procedures had already been performed in our previous MD study (Yui and Hayashi 2009). Therefore, the current MD calculations started from the equilibrated coordinate set of the solvated crystal model systems.…”

“…Therefore, the current MD calculations started from the equilibrated coordinate set of the solvated crystal model systems. In our previous MD calculations (Yui and Hayashi 2009), the SHAKE option (Ryckaert et al 1977) coupled with a 2 fs time step for integration was adopted to constrain the stretching motions of hydrogen atoms. In this study, two independent sets of production runs of NTP dynamics were carried out for each crystal model without any constraints.…”

“…The 1-4 scaling factors for the electrostatic (SCEE) and nonbonded (SCNB) interactions were set to default values of SCEE = 1.2 and SCNB = 2.0. The preliminary results of MD calculations with the SHAKE option off were briefly discussed in our previous report (Yui and Hayashi 2009), where preliminary 1 ns simulations were carried out for the four 24-chain models. In the other set, 1-4 scaling was not introduced (SCEE = 1.0 and SCNB = 1.0), and the SHAKE option was adopted with a 2 fs time step.…”