Directed evolution strategies for enantiocomplementary haloalkane dehalogenases: from chemical waste to enantiopure building blocks

We used directed evolution to obtain enantiocomplementary haloalkane dehalogenase variants that convert the toxic waste compound 1,2,3‐trichloropropane (TCP) into highly enantioenriched (R)‐ or (S)‐2,3‐dichloropropan‐1‐ol, which can easily be converted into optically active epichlorohydrins—attractive intermediates for the synthesis of enantiopure fine chemicals. A dehalogenase with improved catalytic activity but very low enantioselectivity was used as the starting point. A strategy that made optimal use of the limited capacity of the screening assay, which was based on chiral gas chromatography, was developed. We used pair‐wise site‐saturation mutagenesis (SSM) of all 16 noncatalytic active‐site residues during the initial two rounds of evolution. The resulting best R‐ and S‐enantioselective variants were further improved in two rounds of site‐restricted mutagenesis (SRM), with incorporation of carefully selected sets of amino acids at a larger number of positions, including sites that are more distant from the active site. Finally, the most promising mutations and positions were promoted to a combinatorial library by using a multi‐site mutagenesis protocol with restricted codon sets. To guide the design of partly undefined (ambiguous) codon sets for these restricted libraries we employed structural information, the results of multiple sequence alignments, and knowledge from earlier rounds. After five rounds of evolution with screening of only 5500 clones, we obtained two strongly diverged haloalkane dehalogenase variants that give access to (R)‐epichlorohydrin with 90 % ee and to (S)‐epichlorohydrin with 97 % ee, containing 13 and 17 mutations, respectively, around their active sites.     

Stabilization of the Reductase Domain in the Catalytically Self‐Sufficient Cytochrome P450BM3 by Consensus‐Guided Mutagenesis

The multidomain, catalytically self‐sufficient cytochrome P450 BM‐3 from Bacillus megaterium (P450_BM3) constitutes a versatile enzyme for the oxyfunctionalization of organic molecules and natural products. However, the limited stability of the diflavin reductase domain limits the utility of this enzyme for synthetic applications. In this work, a consensus‐guided mutagenesis approach was applied to enhance the thermal stability of the reductase domain of P450_BM3. Upon phylogenetic analysis of a set of distantly related P450s (>38 % identity), a total of 14 amino acid substitutions were identified and evaluated in terms of their stabilizing effects relative to the wild‐type reductase domain. Recombination of the six most stabilizing mutations generated two thermostable variants featuring up to tenfold longer half‐lives at 50 °C and increased catalytic performance at elevated temperatures. Further characterization of the engineered P450_BM3 variants indicated that the introduced mutations increased the thermal stability of the FAD‐binding domain and that the optimal temperature (T_opt) of the enzyme had shifted from 25 to 40 °C. This work demonstrates the effectiveness of consensus mutagenesis for enhancing the stability of the reductase component of a multidomain P450. The stabilized P450_BM3 variants developed here could potentially provide more robust scaffolds for the engineering of oxidation biocatalysts.     

Improving the diastereoselectivity of penicillin G acylase for ampicillin synthesis from racemic substrates

Semi-synthetic β-lactam antibiotics are synthesized enzymatically with the use of penicillin G acylase (PGA). Currently, PGA only exhibits weak diastereoselectivity with respect to the alpha amino group of rac-phenylglycine methyl ester (rac-PGME) when it is coupled with 6-aminopenicillanic acid to synthesize ampicillin. Therefore, we sought to improve the diastereoselectivity of PGA by targeting residues for site-saturation based on the proximity to the substrate's chiral center. Four variants with improved selectivity for (R)-ampicillin synthesis were identified, all resulting from a mutation at the β24 position. βPhe24Ala, while not identified from our library screening, was obtained with site-directed mutagenesis because it has been previously shown to be selective for (R)-enantiomers with substituents other than an amino group. The diastereomeric excess (d.e.(R)) value of 37% for the wild-type enzyme was improved to a d.e.(R) value of 98% for our most selective mutant, βPhe24Ala. Also, four mutations at the α146 position that resulted in (S)-selective PGA variants were identified. βPhe24 and αPhe146 are on opposite sides of the alpha carbon of the substrate, and we have shown that altering these residues results in enhanced selectivity in opposite directions. All variants that showed selectivity for (S)-ampicillin synthesis showed decreased synthetic activity for pure substrates and a decreased synthesis-to-hydrolysis ratio. In contrast, the mutants that were selective for (R)-ampicillin showed significantly decreased primary and secondary hydrolysis when synthesizing ampicillin from pure (R)-PGME, resulting in up to 4-fold decrease in the synthesis to hydrolysis ratio and up to 2-fold increase in the yield achieved. Finally, it was discovered that the selective PGA variants have racemase or epimerase activity, a fascinating phenomenon that has never been reported.     

Changing the substrate specificity of P450cam towards diphenylmethane by semi-rational enzyme engineering

A focused library comprising nine residues of the active site of P450cam monooxygenase resulting in ～ 300,000 protein variants was screened for activity on diphenylmethane (DPM). The assay was based on the depletion of NADH by an in vitro reconstituted P450cam system in a 96-well scale. The throughput was increased by the parallel cultivation, purification and analysis of 20 variants per well (cluster screening). Thus ～ 20,000 protein variants could be screened in summary of which five were found to transform DPM with a specific activity of up to 75% of the wild-type activity on d-camphor and a coupling rate of 7-18%. One variant converting DPM to 4-hydroxydiphenylmethane (4HDPM) was subjected to site-directed mutagenesis and saturation mutagenesis, which revealed the particular importance of positions F87, Y96 and L244 for substrate selectivity and the possibility for further improvements of this variant. Moreover, a reduction in size of the amino acid at position 396 decreased specific activity dramatically but increased coupling and switched the main product formation from 4HDPM towards diphenylmethanol.     

Engineering Cytochrome P450 BM3 for Terminal Alkane Hydroxylation

Enzymes that catalyze the terminal hydroxylation of alkanes could be used to produce more valuable chemicals from hydrocarbons. Cytochrome P450 BM3 from Bacillus megaterium hydroxylates medium‐chain fatty acids at subterminal positions at high rates. To engineer BM3 for terminal alkane hydroxylation, we performed saturation mutagenesis at selected active‐site residues of a BM3 variant that hydroxylates alkanes. Recombination of beneficial mutations generated a library of BM3 mutants that hydroxylate linear alkanes with a wide range of regioselectivities. Mutant 77‐9H exhibits 52% selectivity for the terminal position of octane. This regioselectivity is octane‐specific and does not transfer to other substrates, including shorter and longer hydrocarbons or fatty acids. These results show that BM3 can be readily molded for regioselective oxidation.     

Comprehensive mutagenesis on yeast cytosine deaminase yields improvements in 5-fluorocytosine toxicity in HT1080 cells

Improved prodrug-activating enzymes have the potential to increase the therapeutic efficacy of gene-directed enzyme prodrug therapy (GDEPT). Yeast cytosine deaminase (yCD) is commonly used to convert the prodrug 5-fluorocytosine (5-FC) to the chemotherapeutic 5-fluorouracil for GDEPT. Mutagenesis studies on yCD aimed at improving its application in GDEPT have been limited to subsets of residues or have sought to improve a single property of the enzyme. We performed comprehensive site-saturation mutagenesis (CSM) on yCD designed to create all 2,983 possible unique protein mutants with a single amino acid substitution. We identified active variants through Escherichia coli genetic complementation and screened these mutants, and combinations thereof, for increased ability to sensitize E. coli and HT1080 fibrosarcoma cells to 5-FC. Several mutants identified in this study showed increased sensitization ability for both E. coli and HT1080 cells indicating that CSM is an effective directed evolution tool for identifying unexpectedly beneficial mutations.     

SeSaM-Tv-II generates a protein sequence space that is unobtainable by epPCR

Generating high‐quality mutant libraries in which each amino acid is equally targeted and substituted in a chemically diverse manner is crucial to obtain improved variants in small mutant libraries. The sequence saturation mutagenesis method (SeSaM‐Tv⁺) offers the opportunity to generate such high‐quality mutant libraries by introducing consecutive mutations and by enriching transversions. In this study, automated gel electrophoresis, real‐time quantitative PCR, and a phosphorimager quantification system were developed and employed to optimize each step of previously reported SeSaM‐Tv⁺ method. Advancements of the SeSaM‐Tv⁺ protocol and the use of a novel DNA polymerase quadrupled the number of transversions, by doubling the fraction of consecutive mutations (from 16.7 to 37.1 %). About 33 % of all amino acid substitutions observed in a model library are rarely introduced by epPCR methods, and around 10 % of all clones carried amino acid substitutions that are unobtainable by epPCR.     

Enhancing the Efficiency and Regioselectivity of P450 Oxidation Catalysts by Unnatural Amino Acid Mutagenesis

The development of effective strategies for modulating the reactivity and selectivity of cytochrome P450 enzymes represents a key step toward expediting the use of these biocatalysts for synthetic applications. We have investigated the potential of unnatural amino acid mutagenesis to aid efforts in this direction. Four unnatural amino acids with diverse aromatic side chains were incorporated at 11 active‐site positions of a substrate‐promiscuous CYP102A1 variant. The resulting “uP450s” were then tested for their catalytic activity and regioselectivity in the oxidation of two representative substrates: a small‐molecule drug and a natural product. Large shifts in regioselectivity resulted from these single mutations, and in particular, for para‐acetyl‐Phe substitutions at positions close to the heme cofactor. Screening this mini library of uP450s enabled us to identify P450 catalysts for the selective hydroxylation of four aliphatic positions in the target substrates, including a C(sp³)?H site not oxidized by the parent enzyme. Furthermore, we discovered a general activity‐enhancing effect of active‐site substitutions involving the unnatural amino acid para‐amino‐Phe, which resulted in P450 catalysts capable of supporting the highest total turnover number reported to date on a complex molecule (34 650). The functional changes induced by the unnatural amino acids could not be reproduced by any of the 20 natural amino acids. This study thus demonstrates that unnatural amino acid mutagenesis constitutes a promising new strategy for improving the catalytic activity and regioselectivity of P450 oxidation catalysts.     

Beating Bias in the Directed Evolution of Proteins: Combining High‐Fidelity on‐Chip Solid‐Phase Gene Synthesis with Efficient Gene Assembly for Combinatorial Library Construction

Saturation mutagenesis (SM) constitutes a widely used technique in the directed evolution of selective enzymes as catalysts in organic chemistry and in the manipulation of metabolic paths and genomes, but the quality of the libraries is far from optimal due to the inherent amino acid bias. Herein, it is shown how this fundamental problem can be solved by applying high‐fidelity solid‐phase chemical gene synthesis on silicon chips followed by efficient gene assembly. Limonene epoxide hydrolase was chosen as the catalyst in the model desymmetrization of cyclohexene oxide with the stereoselective formation of (R,R)‐ and (S,S)‐cyclohexane‐1,2‐diol. A traditional combinatorial PCR‐based SM library, produced by simultaneous randomization at several residues by using a reduced amino acid alphabet, and the respective synthetic library were constructed and compared. Statistical analysis at the DNA level with massive sequencing demonstrates that, in the synthetic approach, 97 % of the theoretically possible DNA mutants are formed, whereas the traditional SM library contained only about 50 %. Screening at the protein level also showed the superiority of the synthetic library; many highly (R,R)‐ and (S,S)‐selective variants being discovered are not found in the traditional SM library. With the prices of synthetic genes decreasing, this approach may point the way to future directed evolution.     

An efficient route to selective bio-oxidation catalysts: an iterative approach comprising modeling, diversification, and screening, based on CYP102A1

Perillyl alcohol is the terminal hydroxylation product of the cheap and readily available terpene, limonene. It has high potential as an anti‐tumor substance, but is of limited availability. In principle, cytochrome P450 monooxygenases, such as the self‐sufficient CYP102A1, are promising catalysts for the oxidation of limonene or other inert hydrocarbons. The wild‐type enzyme converts (4R)‐limonene to four different oxidation products; however, terminal hydroxylation at the allylic C7 is not observed. Here we describe a generic strategy to engineer this widely used enzyme to hydroxylate exclusively the exposed, but chemically less reactive, primary C7 in the presence of other reactive positions. The approach presented here turns CYP102A1 into a highly selective catalyst with a shifted product spectra by successive rounds of modeling, the design of small focused libraries, and screening. In the first round a minimal CYP102A1 mutant library was rationally designed. It contained variants with improved or strongly shifted regio‐, stereo‐ and chemoselectivity, compared to wild‐type. From this library the variant with the highest perillyl alcohol ratio was fine‐tuned by two additional rounds of molecular modeling, diversification, and screening. In total only 29 variants needed to be screened to identify the triple mutant A264V/A238V/L437F that converts (4R)‐limonene to perillyl alcohol with a selectivity of 97 %. Focusing mutagenesis on a small number of relevant positions identified by computational approaches is the key for efficient screening for enzyme selectivity.     

Mapping of residues involved in the interaction between the Bacillus subtilis xylanase A and proteinaceous wheat xylanase inhibitors

The Bacillus subtilis xylanase A was subjected to site-directed mutagenesis, aimed at changing the interaction with Triticum aestivum xylanase inhibitor, the only wheat endogenous proteinaceous xylanase inhibitor interacting with this xylanase. The published structure of Bacillus circulans XynA was used to target amino acids surrounding the active site cleft of B.subtilis XynA for mutation. Twenty-two residues were mutated, resulting in 62 different variants. The catalytic activity of active mutants ranged from 563 to 5635 XU/mg and the interaction with T.aestivum xylanase inhibitor showed a similar variation. The results indicate that T.aestivum xylanase inhibitor interacts with several amino acid residues surrounding the active site of the enzyme. Three different amino acid substitutions in one particular residue (D11) completely abolished the interaction between T.aestivum xylanase inhibitor and B.subtilis xylanase A.     

Prediction of the Candida antarctica lipase A protein structure by comparative modeling and site-directed mutagenesis

A number of model structures of the CalA suggested by comparative modeling were tested by site-directed mutagenesis. Enzyme variants were created where amino acids predicted to play key roles for the lipase activity in the different models were replaced by an inert amino acid (alanine). The results from activity measurements of the overproduced and purified mutant enzymes indicate a structure where the active site consists of amino acid residues Ser184, His366, and Asp334 and in which there is no lid. This model can be used for future targeted modifications of the enzyme to obtain new substrate acceptance, better thermostability, and higher enantioselectivity.     

Directed evolution of (βα)(8)-barrel enzymes: establishing phosphoribosylanthranilate isomerisation activity on the scaffold of the tryptophan synthase α-subunit

Phosphoribosylanthranilate (PRA) isomerase (TrpF) and tryptophan synthase α-subunit (TrpA) are (βα)(8)-barrel enzymes that are involved in the biosynthesis of tryptophan. They contain a conserved phosphate binding site, which indicates a common evolutionary origin. In order to experimentally back this hypothesis, we have established TrpF activity on the scaffold of TrpA from Salmonella typhimurium using protein engineering. Based on the superposition of crystal structures with bound ligands, two residues in the active site of TrpA were replaced with catalytic residues from TrpF using site-directed mutagenesis. This TrpA variant as well as wild-type TrpA were each subjected to random mutagenesis using error-prone PCR. The two resulting trpA gene libraries were used to transform an auxotrophic Escherichia coli trpF deletion strain, and TrpA variants with PRA isomerisation activity were isolated by in vivo complementation. The amino acid substitutions of the selected TrpA variants were recombined by DNA shuffling, again followed by complementation in vivo. Several TrpA variants were produced in E. coli and purified, and their catalytic TrpF activities were determined in vitro by steady-state enzyme kinetics. Our results support that TrpA and TrpF have evolved by gene duplication and diversification from each other or a common predecessor, and provide insights into the minimum requirements for the catalysis of PRA isomerisation.     

Enzymatic Thioxyloside Synthesis: Characterization of Thioglycoligase Variants Identified from A Site‐Saturation Mutagenesis Library of Bacillus Circulans Xylanase

Thioglycoligases are engineered enzymes for the synthesis of thioglycosides that are derived from retaining glycosidases by replacing the acid/base catalyst. The optimal choice of substitution for the acid/base mutant is currently unknown, so to investigate this question a complete acid/base library of the model glycosidase Bacillus circulans xylanase (Bcx) was generated by using site‐saturation mutagenesis. A novel screening approach combining active site titration with semiquantitative product analysis by thin layer chromatography was established and used to evaluate specific activities of each mutant enzyme within crude cell lysates. The six most active Bcx variants were analyzed in more detail, a pH optimum of 8.5 was established and the identity of reaction products was confirmed. Optimal choices for substitution were small, preferably polar amino acids such as threonine, cysteine, and serine. We discuss the resultant data in the context of previously published studies on thioglycoligases.     

A Promiscuous Bacterial P450: The Unparalleled Diversity of BM3 in Pharmaceutical Metabolism

CYP102A1 (BM3) is a catalytically self-sufficient flavocytochrome fusion protein isolated from Bacillus megaterium, which displays similar metabolic capabilities to many drug-metabolizing human P450 isoforms. BM3′s high catalytic efficiency, ease of production and malleable active site makes the enzyme a desirable tool in the production of small molecule metabolites, especially for compounds that exhibit drug-like chemical properties. The engineering of select key residues within the BM3 active site vastly expands the catalytic repertoire, generating variants which can perform a range of modifications. This provides an attractive alternative route to the production of valuable compounds that are often laborious to synthesize via traditional organic means. Extensive studies have been conducted with the aim of engineering BM3 to expand metabolite production towards a comprehensive range of drug-like compounds, with many key examples found both in the literature and in the wider industrial bioproduction setting of desirable oxy-metabolite production by both wild-type BM3 and related variants. This review covers the past and current research on the engineering of BM3 to produce drug metabolites and highlights its crucial role in the future of biosynthetic pharmaceutical production.     

Synthetic single-framework antibody library integrated with rapid affinity maturation by VL shuffling

Affinity maturation is often applied to improve the properties of antibodies isolated from universal antibody libraries in vitro. A synthetic human scFv antibody library was constructed in single immunoglobulin framework to enable rapid affinity maturation by updated Kunkel's mutagenesis. The initial diversity was generated predominantly in the V(H) domain combined with only 36 V(L) domain variants yielding 3 × 10(10) unique members in the phage-displayed library. After three rounds of panning the enriched V(H) genes from the primary library selections against lysozyme were incorporated into a ready-made circular single-stranded affinity maturation library containing 7 × 10(8) V(L) gene variants. Several unique antibodies with 0.8-10 nM (K(d), dissociation constant) affinities against lysozyme were found after panning from the affinity maturation library, contrasted by only one anti-lysozyme scFv clone with K(d) <20 nM among the clones panned from the primary universal library. The presented single-framework strategy provides a way to convey significant amount of functional V(H) domain diversity to affinity maturation without bimolecular ligation leading to a diverse set of antibodies with binding affinities in the low nanomolar range.     

Semirational Protein Engineering of CYP153AM.aq.‐CPRBM3 for Efficient Terminal Hydroxylation of Short‐ to Long‐Chain Fatty Acids

The regioselective terminal hydroxylation of alkanes and fatty acids is of great interest in a variety of industrial applications, such as in cosmetics, in fine chemicals, and in the fragrance industry. The chemically challenging activation and oxidation of non‐activated C?H bonds can be achieved with cytochrome P450 enzymes. CYP153A_M.aq.‐CPR_BM3 is an artificial fusion construct consisting of the heme domain from Marinobacter aquaeolei and the reductase domain of CYP102A1 from Bacillus megaterium. It has the ability to hydroxylate medium‐ and long‐chain fatty acids selectively at their terminal positions. However, the activity of this interesting P450 construct needs to be improved for applications in industrial processes. For this purpose, the design of mutant libraries including two consecutive steps of mutagenesis is demonstrated. Targeted positions and residues chosen for substitution were based on semi‐rational protein design after creation of a homology model of the heme domain of CYP153A_M.aq., sequence alignments, and docking studies. Site‐directed mutagenesis was the preferred method employed to address positions within the binding pocket, whereas diversity was created with the aid of a degenerate codon for amino acids located at the substrate entrance channel. Combining the successful variants led to the identification of a double variant—G307A/S233G—that showed alterations of one position within the binding pocket and one position located in the substrate access channel. This double variant showed twofold increased activity relative to the wild type for the terminal hydroxylation of medium‐chain‐length fatty acids. This variant furthermore showed improved activity towards short‐ and long‐chain fatty acids and enhanced stability in the presence of higher concentrations of fatty acids.     

Automated design of degenerate codon libraries

Degenerate codon libraries are frequently used in protein engineering and evolution studies but are often limited to targeting a small number of positions to adequately limit the search space. To mitigate this, codon degeneracy can be limited using heuristics or previous knowledge of the targeted positions. To automate design of libraries given a set of amino acid sequences, an algorithm (LibDesign) was developed that generates a set of possible degenerate codon libraries, their resulting size, and their score relative to a user-defined scoring function. A gene library of a specified size can then be constructed that is representative of the given amino acid distribution or that includes specific sequences or combinations thereof. LibDesign provides a new tool for automated design of high-quality protein libraries that more effectively harness existing sequence-structure information derived from multiple sequence alignment or computational protein design data.     

Antibody affinity maturation using bacterial surface display

A quantitative system for screening combinatorial single-chain Fv (scFv) antibody libraries was developed utilizing surface display on Escherichia coli and fluorescence-activated cell sorting (FACS). This system was employed to isolate clones with high-affinity to a fluorescently-labeled hapten from libraries constructed by randomizing heavy and light-chain residues in the anti-digoxin 26-10 derived antibody, scFv(dig). The use of flow cytometry enabled the detection of rare library members directly in heterogeneous populations and the optimization of selection conditions prior to sorting. A heavy-chain mutant having wild-type affinity (KD = 0.91+/-0.22 nM) and an expected representation frequency of less than 1 x 10(6), was selected to homogeneity after three rounds utilizing increasingly stringent selection conditions. The isolated clone possessed two distinct point mutations relative to the wild-type DNA sequence, yet still coded for the wild-type amino acid sequence, suggesting that the wild-type residues may be optimal at the randomized positions. An affinity improved clone (KD = 0.30+/-0.05 nM), having a dissociation constant approximately threefold lower than the wild-type antibody, was isolated from a smaller light-chain library in a single sorting step. Flow cytometry was shown to be a simple and rapid method for the determination of the relative hapten dissociation rate constants of selected clones without requiring subcloning. The relative rate constants estimated by FACS were confirmed by producing the scFv antibodies in soluble form and measuring hapten binding kinetics by surface plasmon resonance (SPR). These results demonstrate that E.coli surface display, coupled with quantitative selection and analysis using FACS, has the potential to become a powerful tool for rapid isolation and characterization of desirable mutants from large polypeptide libraries.      

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 α/β‐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.