Conformational analysis of natural products

Introduction

Carbohydrates have a number of different degrees of conformational freedom. The first step of any conformational analysis is to determine which degrees of freedom exist, if they are of importance and if they are independent. Since one may suspect that many degrees of freedom are correlated the analysis can become quite complex.
The results of conformational analyses for some carbohydrates are summarized here and a few tools and methods presented.

Degrees of freedom

Depending on the type of carbohydrate there are different degrees of conformational freedom. The monosaccharides can be divided into cyclic, e.g. furanoses, pyranoses, and acyclic compounds, e.g. alditols. Furthermore the cyclic compounds also have a number of exocyclic groups, e.g. hydroxy methyl groups and hydroxyl groups, that may also be of interest. Disaccharides and larger oligo- or polysaccharides add two more degrees of freedom, namely the orientation of the glycosidic linkage.

Conformational models

Depending on what degree of freedom is investigated different physical models are used to describe the conformation. The conformation of the furanose or pyranose ring may be described either in terms of pseudo-rotational phase and degree of puckering or as an equilibrium between idealized conformational states, e.g. 1C4 vs. 4C1.

Experimental methods

The most important experimental method for the study of carbohydrates in solution is NMR spectroscopy. NMR gives several important parameters that have been used, with varying degree of success, in conformational analyses:
Chemical shifts
Chemical shifts (mainly 1H or 13C) are sensitive probes of solvation and conformation. They are easy to record and abundant in litterature. Unfortunately our knowledge of the influence of conformation on the chemical shifts is still very limited.
Nuclear Overhauser effects
The nuclear Overhauser effects (nOe) between two nuclei is dependent on the distance between them. By measuring the build-up of the nOe as a function of time and determining the rate at t=0 it is possible to avoid complications that arise through three-spin effects (indirect nOe). This rate is then proportional to 1/<r3>2 or 1/<r6> depending on the correlation time of the molecule. Apart from practical problems that may arise during the measurement of the nOe, it is difficult to translate the distances to conformations since the latter often are defined in terms of torsion angles. An accurate model of the different conformations is therefore required for the interpretation.
NOe:s between 1H nuclei are most often used but it is possible to observe them for other nuclei as well.
Residual dipolar couplings
Residual dipolar couplings can be observed if a molecule is aligned in the magnetic field e.g. by interaction with an oriented phase such as a lyotropic liquid crystal. The observed couplings (D) depend on the distance between the nuclei and, the orientation of the vector connecting the nuclei and the external magnetic field.
The distance dependence is 1/<r3> (cf. nOe where it is 1/<r6>) and thus longer distances can be measured.
Scalar coupling constants
The three-bond scalar coupling between two nuclei depends on the torsion angle and is described by a truncated Fourier series. If the equation is in the form below it is refered to as a Karplus-equation.
3J(θ)=A*cos(θ)2+B*cos(θ)+C
Often this form is adequate but modified Karplus equations have been proposed where the effect of electronegative substituents is accounted for. Both the magnitude of the curve and the position of the maxima and minima are affected.
1J and 2J values are occationally used in conformational analyses but they depend on bond lengths and the orientation of susbstituents (since there is no torsion between the nuclei) and often more difficult to interpret.
See: Coupling constants