Molecular Dynamics Simulation of Solvated Cell Wall Polysaccharides: A Structural and Dynamic Stud

We investigate behavior of cell wall polysaccharides using computational approaches. Research on cell wall has become an increasingly emphasized topic due to the lignocellulosic biomass is a great candidate for biofuel production. Understanding the cell wall hierarchical structure is not only beneficial for fundamental understanding. It also provides a scientific basis for developing more effective methods for the conversion process. In this thesis, we focus on both the structural behavior of the polysaccharides and the dynamic properties of their surrounding water.;The load-bearing network of cell wall is the cellulose/xyloglucan network. In order to avoid harsh chemical extraction treatment and to simulate the polysaccharides in the length scale that close to that in plant cell walls, we developed a coarse-grained force field for both cellulose and (XXXG) xyloglucan, and combined these two force fields for studying the network structure. The force field for cellulose is built based on atomistic simulation of a 6x6x40 microfibril. The force sites are defined as the geometric average of the six member glucose ring. The force field is parameterized such that the chain configuration, intermolecular packing, and hydrogen bonding of the two levels of modeling are consistent. To retain the directionality of the interfibril interactions, we define pair-wised interactions between the interchain neighbors, and add the potentials sequentially until the crystal structure of the coarse-grained fibril matches that of the atomistic target. The coarse-grained simulation shows that microfibrils longer than 100nm tend to form kinks along their longitudinal direction. The kink structure may be linked to the periodic disorder of the microfibril observed based on small angle neutron scattering measurements.;The xyloglucan force field is build based on atomistic simulation of 15 (XXXG)3 segments, which is the shortest length of xyloglucan that shows significant interaction on cellulose microfibril surfaces. In order to make the two force fields compatible, we also define the xyloglucan force sites as the geometric average of the glucose rings or xylose rings. Thus, there are two types of beads in the xyloglucan chains. The coarse-grained simulation uses the atomistic chain configuration and intermolecular spacings as the target. Upon completion, we performed a simulation combining the two force fields. The simulation box in this case contains one 6x6x200 microfibril surrounded by 20 randomly placed (XXXG)50 xyloglucan chains. The coarse-grained simulation shows three types of xyloglucan based on their interaction with cellulose chains: the bridge chains (that interact with both the microfibril and its periodic image), the single chains (that only interacts with the microfibril), and the isolated chains (that do not directly interact with the microfibril). We also see that some of the isolated chains may bind to the other two types of xyloglucan and participating indirectly in bridging the microfibrils. Therefore, even though some xyloglucans may not directly interact with the microfibril, they may also contribute to the mechanical strength of the network structure. In addition, we observe that the interaction between xyloglucan and cellulose tends to extend along the fibril longitudinal direction. The above observations are very useful for revising the current cell wall models.;Water occupies up to 90% of the cell wall, and it has been shown that water may modify the mechanical properties of cellulose by varying the hydration level. Thus, in order to fully understand the properties of the cell wall network structure, one cannot ignore the role of water. We studied the dynamic properties of water at 5% and 20% hydration levels by fitting self-intermediate scattering function using a stretched exponential model. Atomistic simulation allows us to completely decouple the contributions of translation and rotation in the scattering functions. By applying jump model, we can determine the translational diffusion coefficient, jump length, and the residence time of water proton within the local cage. We observe that multiple types of translational motion exist in 20% hydrated system, but even the faster motion is still slower than bulk water. We further performed simulation of the fibril at these two hydration levels at various temperatures. From each simulation, the rotation and translation is analyzed separately. The activation energy of rotation and translation are obtained using Arrhenius plots of relaxation time of rotation and diffusion coefficient of translation, respectively. The translation activation energy is comparable with bulk water, but the rotational activation energy of the confined water is much higher comparing to the bulk water, indicating a difference in the water rotation mechanism. However, this mechanism seems to be independent on hydration level. By performing anisotropy analysis on water rotation, we determined that the difference in the values of rotation relaxation times of the confined water is due to the difference in their extent of anisotropy.;In summary, we study the structural properties of the cell wall load-bearing network by developing coarse-grained force fields by requiring consistency with atomistic simulation target. The surrounding of the network is examined based on atomistic simulation, which can provide useful insights on the mechanism of the water motions.