Some Directed Evolution Successes from the FHA Group

Biological systems have evolved over billions of years to perform very specific biological functions and to do so within the context of a living organism. Some of the features required for function in a complex chemical network are undesirable when the catalyst is lifted out of context. Conversely many of the properties we wish enzymes would have clash with the needs of the organism, or at least were never required. The Arnold group is developing and using methods of directed evolution to explore the vast space of novel enzyme functions never explored in nature.

One early example was the directed evolution of an enzyme to carry out the hydrolysis of a para-nitrobenzyl ester of an antibiotic (Moore & Arnold, Nature Biotechnology 14, 458-467 (1996) and Moore et al., J. Molecular Biol. 272, 336-347 (1997)). By applying sequential generations of random mutagenesis, recombination and screening, the enzyme's catalytic efficiency was increased more than 100-fold. More recently we evolved a cytochrome P450 monooxygenase to no longer require its ancillary electron-transfer proteins or any external cofactor (NADH) (Joo et al., Nature 399, 670-673 (1999)). The evolved enzyme hydroxylates its substrate using hydrogen peroxide rather than molecular oxygen, via the "peroxide shunt" pathway. While yielding powerful new catalysts for important synthetic reactions, this work clearly demonstrates that enzymes can acquire capabilities not found in naturally occurring organisms.

In another project, sponsored by Degussa AG, we dramatically improved the hydantoinase process for production of L-methionine (L-met) in E. coli by inverting the enantioselectivity and increasing the total activity of the hydantoinase. This process is being evaluated for commercial production.

Mutations can be accumulated through multiple cycles of random mutagenesis in an 'asexual' approach to evolution. Alternatively, they can be accumulated by gene recombination. Both approaches were used to evolve the pNB esterase for higher activity towards an antibiotic intermediate in the presence of organic solvent. Four generations of random mutagenesis followed by two cycles of recombination yielded an enzyme that has more than 100 times the activity of wild type.

The thermostability of this esterase has also been increased significantly using directed evolution [Giver et al., 1998; Gershenson, et al., 2000]. A simple screen was developed based on retention of catalytic activity after incubation at high temperature. Positives were then subjected differential scanning calorimetry (DSC) to verify an increase in melting temperature (T_m) relative to the parent enzyme. Accumulating mutations over eight generations of random mutagenesis and DNA shuffling yielded an increase in T_m of 18 ^oC. This very significant increase in the thermostability of the enzyme is equivalent to the difference between proteins from mesophiles and many thermophilic organisms. Furthermore, the increased thermostability was accompanied by a significant increase in enzyme activity at elevated temperatures. Reflecting the fact that the thermostability screen involved activity towards p-nitrophenyl acetate, the resulting enzymes are highly active towards this substrate and less active towards the antibiotic substrate and p-nitrobenzyl esters. The most thermophilic pNB esterase variant has 13 amino acid substitutions (out of 490). Contradicting the widely-held belief that high thermostability and high activity at low temperatures are incompatible (due to conflicting demands on enzyme flexibility), it is clearly possible to evolve enzymes that exhibit both properties at once.