Rational assignment of key motifs for function guides in silico enzyme identification.

Biocatalysis has emerged as a powerful alternative to traditional chemistry, especially for asymmetric synthesis. One key requirement during process development is the discovery of a biocatalyst with an appropriate enantiopreference and enantioselectivity, which can be achieved, for instance, by protein engineering or screening of metagenome libraries. We have developed an in silico strategy for a sequence-based prediction of substrate specificity and enantiopreference. First, we used rational protein design to predict key amino acid substitutions that indicate the desired activity. Then, we searched protein databases for proteins already carrying these mutations instead of constructing the corresponding mutants in the laboratory. This methodology exploits the fact that naturally evolved proteins have undergone selection over millions of years, which has resulted in highly optimized catalysts. Using this in silico approach, we have discovered 17 (R)-selective amine transaminases, which catalyzed the synthesis of several (R)-amines with excellent optical purity up to >99% enantiomeric excess.

Protein engineering of α/ß-hydrolase fold enzymes.

The superfamily of α/ß-hydrolase fold enzymes is one of the largest known protein families, including a broad range of synthetically useful enzymes such as lipases, esterases, amidases, hydroxynitrile lyases, epoxide hydrolases and dehalogenases. This minireview covers methods developed for efficient protein engineering of these enzymes. Special emphasis is placed on the alteration of enzyme properties such as substrate range, thermostability and enantioselectivity for their application in biocatalysis. In addition, concepts for the investigation of the evolutionary relationship between the different members of this protein superfamily are covered, together with successful examples.

3DM: systematic analysis of heterogeneous superfamily data to discover protein functionalities.

Ten years of experience with molecular class-specific information systems (MCSIS) such as with the hand-curated G protein-coupled receptor database (GPCRDB) or the semiautomatically generated nuclear receptor database has made clear that a wide variety of questions can be answered when protein-related data from many different origins can be flexibly combined. MCSISes revolve around a multiple sequence alignment (MSA) that includes "all" available sequences from the entire superfamily, and it has been shown at many occasions that the quality of these alignments is the most crucial aspect of the MCSIS approach. We describe here a system called 3DM that can automatically build an entire MCSIS. 3DM bases the MSA on a multiple structure alignment, which implies that the availability of a large number of superfamily members with a known three-dimensional structure is a requirement for 3DM to succeed well. Thirteen MCSISes were constructed and placed on the Internet for examination. These systems have been instrumental in a large series of research projects related to enzyme activity or the understanding and engineering of specificity, protein stability engineering, DNA-diagnostics, drug design, and so forth.

Natural diversity to guide focused directed evolution.

Simultaneous multiple site-saturation mutagenesis was performed at four active-site positions of an esterase from Pseudomonas fluorescens to improve its ability to convert 3-phenylbutyric acid esters (3-PBA) in an enantioselective manner. Based on an appropriate codon choice derived from a structural alignment of 1751 sequences of alpha/beta-hydrolase fold enzymes, only those amino acids were considered for library creation that appeared frequently in structurally equivalent positions. Thus, the number of mutants to be screened could be substantially reduced while the number of functionally intact variants was increased. Whereas the wild-type esterase showed only marginal activity and poor enantioselectivity (E(true)=3.2) towards 3-PBA-ethyl ester, a significant number of hits with improved rates (up to 240-fold) and enantioselectivities (up to E(true)=80) were identified in these "smart" libraries.

Converting an esterase into an epoxide hydrolase.

Entering the fold: A common structural motif in hydrolytic enzymes is the alpha,beta-hydrolase fold. The interconversion of one enzyme into another by introduction of mechanistically important residues is not enough; only substitution of a loop allows epoxide hydrolase activity in the esterase scaffold to be formed (see picture; structure comparison of epoxide hydrolases (green) with the esterase (orange)). The result is an enantioselective chimeric enzyme.

Thermostabilization of an esterase by alignment-guided focussed directed evolution.

Site-saturation libraries of the Pseudomonas fluorescens esterase were created targeting three surface positions to increase its thermostability on the basis of the B-factor iterative test principle. All three positions were saturated simultaneously using our recently developed protocol for the design of 'small, but smart' mutant libraries bearing only consensus-like mutations. Hence, the library size could be significantly reduced while ensuring a high hit rate. Variants could be identified that showed significantly improved stability (8° C higher compared with the wild type) without compromising specific activity. Subsequent iterative saturation mutagenesis gave an esterase mutant with a 9° C increased melting point, but unchanged catalytic properties.