In defense of adiabatic ϕ/ψ mapping for cellobiose and other disaccharides

Abstract

Thorough conformational study of cellobiose requires consideration of numerous arrangements of the exocyclic groups. Therefore, it is customary to prepare a number of structures with different arrangements of hydroxyl and hydroxymethyl groups. These ''starting geometries'' are then given different values of the glycosidic linkage torsion angles / and w. At each increment of / and w, the energy is calculated. Usually, the final product is an ''adiabatic'' contour plot of the lowest energy at each //w point after considering all of the starting geometries. The present paper advocates for adiabatic maps despite the statement by Schnupf and Momany (preceding paper) that adiabatic maps are not of interest because they contain sparse details about the structures at each minimum. Similar information is computed by their method and adiabatic mapping, and comparable details can be provided from adiabatic studies. Although Schnupf and Momany presented maps from calculations in vacuum and in water that considered all of their calculated energies, they favored the presentation of two to four maps for each of 36 individual minima, each with its own zero of relative energy. However, previous work showed that more structures are needed to provide the lowest energies at each point in //w space. Following their preferred strategy would result in even more maps when the added structures are considered. The need to map individual minima can be avoided by starting calculations with the same exocyclic orientations at each //w point instead of using the preceding optimized structure to start the next energy minimization. Using the same orientations at each point allows periodic maps that depict barriers between minima.

Combining Computational Chemistry and Crystallography for a Better Understanding of the Structure of Cellulose

“…Instead, the literature on the same or similar molecules, such as cellobiose or lactose, describes work performed in various ways by different groups, employing different software. That approach is different 133 from the technology that is sometimes incorporated in modeling software for "scanning" a potential energy surface wherein which each calculation is started with the energy-minimized structure from the previous f/c point. 10, the same set of 155 starting geometries was used at each f/c point.…”

“…That concept was rebutted 133 with the consideration that some 23 different starting geometries had been needed to yield the lowest energy at each of 81 f/c map points in a quarter of f/c space, 146 and there are typically four minima for the entire f/c space. That concept was rebutted 133 with the consideration that some 23 different starting geometries had been needed to yield the lowest energy at each of 81 f/c map points in a quarter of f/c space, 146 and there are typically four minima for the entire f/c space.…”

“…133 Another problem with that Schnupf and Momany work was the limited number of starting geometries. That strategy used the structure from the previous minimization to start the next one.…”

“…As discussed in the previous section, such scanning of the conformations has the disadvantage that inelastic structural conversions occur on proceeding from one conformation to the next. 126 The new project was designed to allow maximum comparability of the two studies, while taking advantage of different mapping strategies 132,133 and a more complete, SMD 147 solvation model. 133 Another problem with that Schnupf and Momany work was the limited number of starting geometries.…”