Two major approaches have been developed in protein engineering. One is directed evolution. A library of variants with random mutations in the protein-encoding gene is created through sequential error-prone PCR (epPCR) or in vitro DNA recombination, and then screening methods are used to isolate functionally improved mutant proteins for further study or the next round of improvement. Directed evolution does not rely on the knowledge of the enzyme structure and catalytic mechanism, but large libraries generated by directed evolution need to be screened for the best mutants with desired properties, which is considered to be time-consuming and laborious. The development of new efficient and convenient screening methods is the key to obtaining novel biocatalysts by directed evolution
[37][38]. The second approach is the rational design which generates beneficial mutants based on protein sequences, structural information, and the relationship between enzyme structure and function. Advanced computational methods are often adopted to introduce specific mutagenesis in conserved functional domains, catalytic active sites, and certain amino acid positions for yielding remarkable desired properties (
Figure 2). The combination of directed evolution and rational design strategies, named semi-rational design, with the benefits of both strategies has also been developed for improving substrate specificity, enantioselectivity, and stability.
Many members of α/β hydrolase family show “catalytic promiscuity” and catalyze diverse chemical transformations. Enzyme promiscuity plays a key role in protein engineering because enzyme activities can be changed by only altering a few amino acid residues
[39][40]. SABP2-G12T-M239K, the variant of plant esterase salicylic acid binding protein 2 (SABP2) can be converted into a hydroxynitrile lyase only with the replacement of two amino acid residues. This mutant can also catalyze reverse reactions to form mandelonitrile with low enantio-specificity
[28]. Nedrud et al. substituted the active site residues of hydroxynitrile lyase
HbHNL with the corresponding residues in esterase SABP2 and found key residues to generate the esterase activity
[35]. On the contrary, Jochens et al. attempted to convert the esterase of
Pseudomonas into epoxide hydrolase by replacing the residues of active site loops in the esterase with the residues in the epoxide hydrolases. After residues in multiple sites were carefully chosen and replaced, the mutant enzyme showed an epoxide hydrolase activity with enantioselectivity for R-type enantiomers, but it was widely inhibited by substrates and had a limited turnover number
[25]. These examples provide evidence for specificity engineering by utilizing the catalytic promiscuity of enzymes, although their performance, compared with natural enzymes, still needs further improvement.
Beier et al. reported a study on haloalkane dehalogenase which provides a new way to expand the application of α/β hydrolases in chiral preparations
[29]. By mutating the catalytic residue histidine and blocking the hydrolysis process, the variant DbjA H280F could catalyze a transhalogenation reaction and realize the replacement of halogen substitutes. This strategy was further demonstrated in several different haloalkane dehalogenases
[29].
4. Catalytic Activity Enhancement of α/β Hydrolases
High catalytic activities are required for enzymes in commercial applications. It is possible to increase the activities of natural enzymes through rational design and directed evolution approaches. Kotik et al. significantly improved the activity of
Aspergillus niger M200 epoxide hydrolase through single site mutagenesis of Ala217 at the substrate entrance. For example, A217C enhanced the activity of the enzyme by three times towards allyl glycidyl ether and styrene oxide, while A217V increased the activity of the enzyme to allyl glycidyl ether by six times
[13]. A completely different approach was used by Yu et al. to improve the activity of
Candida antarctica lipase B. A mutant cp283 obtained through circular permutation showed higher catalytic activity, compared with the wild-type enzyme in the transesterification reaction with different substrates using 1-butanol and ethyl acetate as acyl receptors
[36]. Langermann et al. replaced the active site residue in hydroxynitrile lyase
HbHNL with the corresponding site in esterase SABP2, thereby improving the catalytic activity of
HbHNL on an unnatural substrate mandelonitrile
[34]. More recently, Marek et al. transplanted a nine-residue-long extension of L9 loop and α4 helix from DbjA into the corresponding site of another haloalkane dehalogenase DbeA. The mutation not only altered its enantio-preference with several linear β-bromoalkanes but also enhanced the catalytic activity of DbeA towards various halogenated substrates
[30].
5. Regio- and Stereo-Selectivity Engineering of α/β Hydrolases
Since the resolution process of racemates is costly and often difficult, enzymatic catalytic synthesis of enantiopure compounds has gradually become a useful alternative to various chemical preparation routes. While enantio-convergent bioconversions by natural enzymes are usually not able to meet industrial requirements, it is necessary to improve or alter the enantioselectivity of biocatalysts by protein engineering.
Substrate binding sites are frequently engineered for the improvement of regio- and stereo-selectivity because they contain the catalytic residues and determine the location of substrates during the catalysis. By mutating the residues of the atypical oxirane oxygen hole (Trp186/Trp187/Tyr247) around the active site in epoxide hydrolase Alp1U, Zhang et al. obtained mutants Y247F and W187F/Y247F with high regioselectivity, respectively, towards C-2 and C-3 of fluostatin C
[11]. Recently, Wen et al. reported the improvement of the enantio-specificity of
AuEH2 towards racemic
ortho-trifluoromethyl styrene oxide by tuning the substrate binding pocket
[17]. With a similar method, Hu et al. performed the molecular docking of
AuEH2 with (
R)-styrene oxide by AutoDock Vina and screened suitable sites for site-directed saturation mutagenesis and combinatorial mutagenesis. A mutant A214C/A250I exhibited a 12.6-times increase of the enantiomeric ratio towards the substate
rac-styrene oxide
[16]. Accompanied by the deep understanding of the relationship between the structure and activity of enzymes, more successful studies have been reported to enhance the enantioselectivity by the design and engineering of active site residues
[14][18][20][41].
In addition to the active site, the entrance tunnel of the enzyme has also been targeted to modulate the selectivity. For example, Kotik et al. reported that a single amino acid mutation at the entrance site of the
AnEH tunnel can notably increase its enantioselectivity
[13].
6. Stability Enhancement of α/β Hydrolases
Enzymes with high thermostabilities are required in many industrial production processes because the elevated temperature has positive effects on many bioconversion processes such as improving the solubility of substrates, increasing the catalytic rate, and reducing microbial contaminants
[42]. Through a combination of computational design and experimental screening, Floor et al. established a method named FRESCO, which based on the calculation of folding energies for all possible substitutions and the integration of conformational sampling in disulfide-bond designs, utilized the molecular dynamics as a fast-screening tool. Using this method, a multisite mutant of a large monomeric protein, haloalkane dehalogenase LinB, was obtained with drastically improved thermostability (a 23 °C increase in apparent melting temperature and an over 200-fold longer half-life at 60 °C) and retained moderate enantioselectivity
[33]. Similarly, Bednar et al. established another method, FireProt, which is a computational strategy aimed at the prediction of highly stable multipoint mutations and significantly increased the thermostability of haloalkane dehalogenase DhaA
[31]. Utilizing the circular permutation strategy to analyze the three characteristic regions of typical esterase
RhEst1, Li et al. obtained three mutants CP-20, CP-22, and CP-24 with ameliorated thermostability
[26]. A systematic study by Jones et al. used and compared five protein engineering strategies (random mutagenesis, two computational methods Rosetta and FoldX, consensus mutation, and homoproline mutation) to enhance the stability of an α/β-hydrolase fold enzyme, salicylic acid binding protein 2. The results showed that all five methods could obtain mutants with enhanced stability, but consensus mutation and homoproline mutation seemed to be the best choices.