Gj. Salter et Db. Kell, SOLVENT SELECTION FOR WHOLE-CELL BIOTRANSFORMATIONS IN ORGANIC MEDIA, Critical reviews in biotechnology, 15(2), 1995, pp. 139-177
Although they were used historically as antimicrobial agents, there is
a modem requirement to devise organic solvent systems for exploitatio
n in the biotransformation by intact cells of substrates that are poor
ly soluble in water. Water-immiscible solvents are normally less cytot
oxic than are water-miscible ones. While a unitary mechanism is exclud
ed, damage to the membrane remains the likeliest major mechanism of cy
totoxicity, and may be conveniently assessed using an electronic bioma
ss probe. Studies designed to account for the mechanisms of action of
general anesthetics and of uncouplers parallel those designed to accou
nt for the cytotoxicity of organic solvents. Although there are hundre
ds of potential physical descriptors of solvent properties, many are b
roadly similar to each other, such that the intrinsic dimensionality o
f solvent space is relatively small (<10). This opens up the possibili
ty of providing a rational biophysical basis for the optimization of t
he solvents used for biotransformations. The widely used descriptor of
solvent behavior, log P (the octanol:water partition coefficient), is
a composite of more fundamental molecular descriptors; this explains
why there are rarely good correlations between cytotoxicity and log P
when a wide variety of solvents is studied. Although the intrinsic dim
ensionality of solvent space is relatively small, pure solvents still
populate it rather sparsely. Thus, mixtures of solvents can and do pro
vide the opportunity of obtaining a solvent optimal for a biotransform
ation of interest.