1870

Current Organic Chemistry, 2010, 14, 1870-1882

Engineering of Enzymes for Selective Catalysis Nikhil U. Nair1,#, Carl A. Denard1,#, and Huimin Zhao1,2,* 1

Department of Chemical and Biomolecular Engineering, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA

2

Department of Chemistry, Department of Biochemistry, Department of Bioengineering, and Institute for Genomic Biology, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA Abstract: Naturally occurring enzymes are truly remarkable catalysts. However, many are not suitable for practical synthetic chemistry because of poor substrate or product selectivity. This review highlights recent advances in engineering enzymes for selective catalysis. Selected topics include altering substrate specificity, altering substrate and product selectivity, engineering enzymes that catalyze carboncarbon bond formation or carbon-oxygen bond cleavage and formation, and engineering multi-function enzymes such as polyketide synthases.

Key Words: Directed evolution, enzyme engineering, biocatalysis, selectivity. 1. INTRODUCTION Enzymes are biocatalysts with numerous potential applications in industry and medicine [1-3]. Compared to chemical catalysts, one of the most important advantages of a biocatalyst is its high selectivity, namely stereoselectivity, regioselectivity, and chemoselectivity. Such high selectivity is desirable in chemical synthesis as it may reduce or eliminate the use of protecting groups, minimize side reactions, simplify separation, and reduce environmental problems. Other advantages of a biocatalyst include high catalytic efficiency and mild operational conditions (pH, temperature, and pressure). Unfortunately, many naturally occurring enzymes are not suitable for industrial applications because of their poor selectivity, low stability in organic solvents, slow reaction rates, and substrate or product inhibition [4]. To overcome these limitations, two complementary enzyme engineering approaches, rational design and directed evolution, have been developed over the past decades. Rational design involves site-specific alterations of selected residues in a protein to cause predicted changes in function, whereas directed evolution mimics the natural evolution process in the laboratory and involves repeated cycles of generating a library of protein variants and selecting the variants with desired properties. Largely due to its high success rate and general applicability, directed evolution has become the preferred engineering approach to generate tailor-made enzymes [5, 6]. The two key steps of directed evolution are generating molecular diversity and identifying improved variants through screening or selection. In order to generate molecular diversity, numerous methods of random mutagenesis and in vitro gene recombination have been developed [2, 7]. For random mutagenesis, error-prone PCR (epPCR) is the most convenient method, whereas DNA shuffling [8] and staggered extension process (StEP) [9] are the two most widely used in vitro gene recombination methods. In parallel to the development of these library creation methods, numerous high throughput screening or selection methods have been reported [4, 10, 11]. Screening involves examining every variant individually for the desired enzyme

*Address correspondence to this author at the Department of Chemical and Biomolecular Engineering, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA; Tel: (217) 333-2631; Fax: (217) 333-5052; E-mail: [email protected] #

These two authors contributed equally 1385-2728/10 $55.00+.00

property and usually relies on colorimetric or fluorogenic substrates or products. In contrast, selection links the survival of the host with the desired enzyme property. Hundreds of articles on enzyme engineering have been published between 2003 and 2009, and cannot be covered here. In this review, we focus on protein engineering of enzymes for selective catalysis, including altering substrate specificity, altering substrate and product selectivity (enantioselectivity and regioselectivity), engineering enzymes that catalyze carbon-carbon bond formation (aldolases) and carbon-oxygen bond cleavage and formation (glycosidases and glycosyltranferases), and engineering of multifunction enzymes such as polyketide synthases (PKSs). The latter two topics are included as separate categories because unlike the enzymes in the first two topics, these enzymes utilize multiple catalytic reactions or more than one substrate. 2. ALTERING SUBSTRATE SPECIFICITY Altering substrate specificity is one of the most frequently required property changes in protein engineering of enzymes for selective catalysis, as enzymes in nature usually do not exhibit a wide substrate range or high catalytic activity towards various substrates. Since enzymes are highly specific both in the nature of the substrate they utilize and the type of reaction they catalyze, the need to modify the enzymes to show broad substrate specificity is industrially important and falls into three categories: completely switching the specificity towards novel substrates, expanding, and narrowing substrate specificity. Directed evolution and rational design approaches have been widely successful in such endeavors. We will highlight the most recent examples in the following sections. Switching Substrate Specificity Cytochrome P450 monooxygenases are versatile biocatalysts that introduce oxygen into a vast range of molecules. They are truly remarkable “workhorses”, capable of introducing atomic oxygen into allylic positions, double bonds, or even into non-activated C–H bonds in a regio- and stereo- selective manner (Fig. 1A) [12-14]. The cytochrome P450 BM-3 from Bacillus megaterium catalyzes the sub-terminal hydroxylation of long-chain (C12–C20) fatty acids. Due to its high activity and catalytic self-sufficiency (heme and diflavin reductase domains are fused in a single polypeptide chain) © 2010 Bentham Science Publishers Ltd.

Engineering of Enzymes for Selective Catalysis

[12, 13], the P450 BM-3 makes an excellent platform for biocatalysis. Arnold and coworkers successfully engineered the P450 BM-3 enzyme, through successive rounds of laboratory evolution, to catalyze the selective hydroxylation of the unnatural short alkane propane with native-like activity and coupling efficiency [15], a reaction for which no practical catalysts are available [16]. An initial round of mutagenesis afforded the P450 BM-3 variant 139-3 that showed low activity towards propane [17]. A domain-based protein engineering strategy was used, in which the heme, flavin mononucleotide (FMN), and flavin adenine dinucleotide (FAD) domains were evolved separately in the context of the holoenzyme, and beneficial mutations were recombined in a final step. Mutant libraries were created by random mutagenesis and site-directed mutagenesis. Two mutants, P450PMO R1 and R2, containing mutations in both the reductase and heme domains, could catalyze propane oxidation with high total turnovers (35,600 and 45,800 for R1 and R2, respectively), while maintaining close to 98% coupling efficiency. This work highlights a complete re-specialization of the P450 BM-3 to a propane monooxygenase P450PMO achieved by a profound reshaping and partitioning of the substrate access pathway (Fig. 1B) [18]. In P450PMO, little trace of the original P450 BM-3 activity remained. Through saturation mutagenesis, Arnold and coworkers altered the chemoselectivity of P450 BM-3 by engineering of the heme domain to create two mutants, RH-47 and SH-44, for the selective epoxidation of simple terminal alkenes (Fig. 1B) [19]. The two mutants show inverted enantiospecificity (toward an enantiomer of product), with the RH-47 variant forming the (R)-epoxide, while the (S)-epoxide is formed by the SH-44 variant. Up to 83% enantiometric excess (ee), high catalytic turnovers (up to 1,370) and high epoxidation selectivities (up to 95%) were obtained. The authors performed saturation mutagenesis on 11 residues located in a radius of 5Å around the substrate active site of the P450 BM-3 wild-type crystal structure in complex with N-palmitoylglycine and combined beneficial mutations to yield the positive mutants [20]. Furthermore, using a biocatalytic system with E. coli lysates containing P450 variants as the epoxidation catalysts and an in vitro NADPH regeneration system, each of the epoxide enantiomers was generated in high yields. This exemplifies an alteration of enzyme chemoselectivity, in addition to switching substrate specificity from alkanes and alkenes, since minimal hydroxylation reaction products were observed. In separate works, ethane conversion to ethanol was achieved using two different P450s, P450 BM-3 [21] and P450CAM [22]. Both variants, however, show total turnover numbers towards ethane and coupling efficiencies (the percentage of NADH consumed that lead to product formation) that were still too low for practical purposes. Using epPCR, Kim and coworkers altered the substrate specificity of the Thermus caldophilus GK-24 ADP-glucose pyrophosphorylase, an important enzyme for the enzymatic synthesis of activated sugars, towards enantiomeric substrates [glucose-1-phosphate (G-1-P) + uridine triphosphate (UTP)] and [N-acetylglucosamine-1phosphate (GlcNAc) + UTP] to produce uridine diphosphate (UDP)-glucose and UDP-N-acetylglucosamine, respectively [23]. ADP-glucose-1-phosphate adenyltransferase is a major regulatory enzyme in the biosynthesis of
-glucans in bacteria and plants [24]. From 656 colonies screened, two colonies showed UDP-glucose pyrophosphorylase (UGPase) activity, while three had UDP-Nacetylglucosamine pyrophosphorylase (UNGPase) activity.

Current Organic Chemistry, 2010, Vol. 14, No. 17 1871

A few rational design approaches have been reported in recent years. Edmiston and coworkers used a rational design approach to change the substrate specificity of the rabbit muscle creatine kinase (CK) from creatine to glycocyamine [25]. Although creatine and glycocyamine have similar structures that differ only by a single methyl at the -N of the guanidine group, wild-type creatine and glycocyamine kinases (GK) exhibit high levels of substrate specificity. As glycocyamine is a metabolic precursor for creatine, the authors wanted to investigate the binding determinants that have evolved to yield such substrate discrimination. Structural information obtained from sequence alignment of various CK and arginine kinases provided insight into the conserved residues of the two flexible loops that sequester the bound substrates from bulk solvent at the active site. A CK mutant created by site-directed mutagenesis exhibited a 2,000-fold change in substrate specificity for glycocyamine versus creatine; however, a large drop in enzymatic activity ensued in all cases. To switch the substrate specificity of the eukaryotic holoenzyme protein phosphatase 2A using site-directed mutagenesis, the catalytic subunit C of the enzyme was modified by the removal of leucine 309 at the carboxy-terminus of the C subunit [26]. This abolished the binding of the catalytic subunit C towards subunits B/PR55 and modified the holoenzyme composition and subsequent substrate specificity. The abolishment of B/PR55 binding decreased the holoenzyme activity towards cdc-2-phosphorylated histone H1, while still being active towards the general substrate phosphoylase A. This is one of the few examples where protein engineering has been employed to study multi-subunit enzymes in mammalian cells. One of the key problems in utilizing xylose for ethanol production is the cofactor imbalance created by the coenzyme specificities of xylose reductase (XR; NADPH) and xylitol dehydrogenase (XDH; NAD+) from the xylose utilization pathway [27]. Lin and coworkers achieved complete reversal of coenzyme specificity of the XR from Pichia stipitis by site-directed mutagenesis from NADPH to NADH [28]. In a similar work, Watanabe and coworkers performed a complete reversal of coenzyme specificity of the NAD+-dependent XDH from P. stipitis [29]. Although the double mutants obtained still retained NAD+ activity, the triple and quadruple mutants showed 4,500-fold higher activity with NADP+ than the wild-type enzyme, comparable to the wild-type activity with NAD+. Expanding Substrate Specificity Using an in vivo selection method, Arnold and coworkers evolved the medium chain alkane hydroxylase AlkB from Pseudomonas putida GPo1 and P450 enzyme CYP153A6 from Mycobacterium sp. strain HXN-1500 for the terminal hydroxylation of butane to 1-butanol [30]. Previous work on alkane hydroxylases, P450CAM [22] and P450 BM-3 [15], has yielded variants that hydroxylate propane and higher alkanes at the more energetically favorable subterminal positions; highly selective terminal hydroxylation is difficult to achieve by engineering a subterminal hydroxylase. Terminal alkane hydroxylase activity of propane- and butaneoxidizing enzymes was selected based on enhanced growth complementation of an adapted P. putida GPo12 (pGEc47_B) strain. In addition, E. coli mutator strains, rather than epPCR, were used to generate randomly mutated plasmids encoding AlkB or CYP153A6 in order to afford a larger library for the selection. Several mutants were obtained that showed high rates of 1-butanol production from butane and maintained their preference for terminal hydroxylation.

1872 Current Organic Chemistry, 2010, Vol. 14, No. 17

A

Nair et al.

Carbon hydroxylation [FeO]3+

HC

[FeOH]3+

Fe3+

C

HOC

Heteroatom release [FeO]3+

[FeO]2+

N CH2R

OH Fe3+

N CHR

[FeOH]3+

N CH2R

N CHR

O NH + CHR

N CHR

Heteroatom oxygenation [FeO]3+

X

[FeO]2+

Fe3+

X

O X

Epoxidation and group migration R

R [FeO]3+

O

R

R

O R [FeO]2+

HO

R

(OH)

(OH)

O N N

N Fe4+ N

R HO

N

N

N N

B

O OH

O

WT P450 BM3

OH (OH)

NADP+

NADPH + O2 + H+

+ H2O

random mutagenesis recombination

(2%) (66%) (OH) (OH) 139-3 (OH)

DNA shuffling site saturation recombination

(32%)

35E11

9-10A site saturation recombination

site-directed mutagenesis

O

random mutagenesis site saturation recombination site-directed mutagenesis P450PMO

(10%) (HO)

SH-44

71% ee (S)

RH-47

60% ee (R)

(90%) (OH)

Fig. (1). A) Major P450 reactions [94, 95], B) Examples of evolved P450 BM-3 mutants obtained from the wild-type fatty acid hydroxylase by directed evolution [15, 17].

Engineering of Enzymes for Selective Catalysis

Similar to the above studies, Arnold and coworkers used directed evolution to expand the substrate scope of the P450 BM-3 from B. megaterium [17]. The resulting 139-9 mutant not only supported higher turnover rates on C3-C8 alkane chains than the wildtype, but also showed higher turnovers on the wild-type substrates lauric acid and palmitic acid. Of the 11 residue changes that resulted from this engineering work in the heme domain, surprisingly, only residue 87 was in the active site of the enzyme. Furthermore, starting with the P450 139-3 mutant, Peters and coworkers used a combination of epPCR, DNA shuffling, and site-directed mutagenesis to evolve a P450 mutant capable of hydroxylating linear alkanes regio- and enantio- selectively using atmospheric dioxygen as an oxidant [31]. Two variants, 9-10A-A328V and 1-12G, hydroxylate alkanes larger than hexane primarily at the 2-position to form S-2-octanol (40% ee) and R-2-octanol (40-55% ee), respectively. Enantiomerically pure chiral amines are valuable synthetic intermediates. Various methods have been developed for the purification of chiral amines, including resolution-based procedures [32-34] and asymmetric approaches [35] that can, in theory, deliver the product with 100% yield and 100% ee. Alexeeva and coworkers expanded the substrate range of the Type II monoamine oxidase from Aspergillus niger (MAO-N) by increasing its activity and enantioselectivity to effectively deracemize
-methylbenzylamine [36]. This deracemization method involves the stereoinversion of L
-methylbenzylamine to D-
-methylbenzylamine by repeated cycles of enzyme-catalyzed oxidation to the imine. This is followed by the nonselective reduction back to the amino acid using reductants such as sodium cyanoborohydride or ammonium formate with Pd/C. MAO-N, whose natural substrates are simple aliphatic amines such as amylamine and butylamine, showed barely detectable activity towards L-
-methylbenzylamine. Using a colorimetric 96-well plate based screening assay for amine oxidase activity by capture of the hydrogen peroxide produced using 3,3’-diaminobenzidine with peroxidase, the enzyme mutant N336S showed 47-fold higher activity towards
-methylbenzylamine. Furthermore, the selectivity towards L-
-methylbenzylamine versus D-
-methylbenzylamine was increased 5.8 fold (from 17:1 for the wild-type enzyme to 100:1 for the evolved enzyme). Reetz and coworkers recently provided a remarkable example of expanding the substrate specificity of an enzyme [37]. The authors applied CASTing (combinatorial active site testing, see also Increase in Enantioselectivity) to expand the substrate acceptance of the lipase from Pseudomonas aeruginosa as a catalyst in the hydrolysis of carboxylic acid esters, so that it would include the wild-type triglycerides or fatty acid esters such as palmitic acid pnitrophenyl ester and more sterically demanding substrates. Three thousand colonies from each of five libraries were subjected to a multisubstrate screening (11 different esters simultaneously). The use of an in vitro cofactor regeneration system provides an economic incentive in enzyme reactions that use equimolar amounts of expensive cofactors [38, 39]. Of the enzymatic NADH regeneration systems, the most widely used enzyme is formate dehydrogenase (FDH) from Candida boidinii [40]. In order to expand the toolset of regeneration systems, Zhao and coworkers engineered the phosphite dehydrogenase (PTDH) from Pseudomonas stutzeri, an enzyme which may have kinetic and practical advantages over FDH in certain applications [41], to relax its cofactor specificity [42]. Site-directed mutagenesis was performed to create mutations E175A and A176R, both separately and in combination to yield three mutants, all exhibiting significantly better catalytic efficiency for both cofactors. The double mutant showed the highest kinetic

Current Organic Chemistry, 2010, Vol. 14, No. 17 1873

parameters, namely a 3.6-fold higher catalytic efficiency for NAD + and a 1,000-fold higher efficiency for NADP+. This work highlights the creation of a dual cofactor enzyme, with efficiencies higher than the wild-type enzyme for both cofactors. Narrowing Substrate Specificity In literature, there are very few examples of enzymes engineered to have a narrowed substrate range. Recently, Nair and Zhao used a combination of semi-rational design approaches, namely targeted site-saturation mutagenesis (TSSM), and epPCR to engineer a xylose-specific xylose reductase (XR) [43]. XRs catalyze the conversion of D-xylose to xylitol, a sweetener and a platform chemical for the production of industrially important chemicals [44]. However, XRs are promiscuous and can reduce a number of pentoses and hexoses, including L -arabinose to L-arabinitol, an undesired by-product. In their work, the authors presented an engineered XR with increased preference for D-xylose over L-arabinose. The mutagenic libraries were subjected to positive selection pressure linking cell growth on D-xylose to the assimilation of D-xylose through the presence of an active XR in a constructed selection strain. A negative selective pressure by L-arabinose against promiscuous XR resulted in the accumulation of toxic arabinitol phosphate and led to growth inhibition. The epPCR library yielded several mutants, including mutant Q (L109Q), which had 8.9-fold preference for D-xylose. Iterative rounds of TSSM identified additional mutations and finally yielded mutant VMQCI, which showed a near complete loss of activity towards L-arabinose (K M >2000 mM) compared to the wild-type enzyme (KM = 40 mM), along with a slight decrease in affinity towards D-xylose (430 mM compared to 34 mM for the wild-type XR). However, this loss of affinity was compensated by a higher catalytic activity. Most recently, Gupta and Farinas converted the laccase CotA from Bacillus subtilis from a generalist, an enzyme with broad specificity, to a specialist enzyme with narrowed specificity [45]. 3. ALTERING SUBSTRATE AND PRODUCT SELECTIVITY Reaction steps requiring >95% selectivities are quite difficult to obtain by any chemocatalysis. Such high selectivities are of particular interest to pharmaceuticals to ensure minimal production of potentially toxic or antagonistic isomeric byproducts. High product purities are also mandated by regulatory authorities like FDA (Food and Drug Administration) to ensure product safety and minimal drug side effects. Therefore, use of biocatalysts in the pharmaceutical industry has been steadily growing with increasing ability to create a large and diverse set of chiral molecules due to the ability to engineer and discover new enzymes more readily. Many indirect methods of controlling the enantioselectivity of a biocatalyst have been described in the past [46-48]. These primarily involve altering either reaction conditions (pH, pressure, temperature, etc.) or solvent properties (polarity, hydrophobicity, ionic strength, etc.) to control enzyme-substrate interactions. While these techniques have had success, the circuitous nature of the enantiomeric control makes them less than ideal methods. Enzyme engineering, particularly directed evolution, is a more direct and powerful method to alter enantioselectivity. It has been used to not only increase the natural selectivity of certain enzymes, but also completely invert the selectivity of other enzymes. Increase in Enantioselectivity The moderate enantiomeric excess produced by naturally occurring enzymes is usually insufficient for industrial application. Pre-

1874 Current Organic Chemistry, 2010, Vol. 14, No. 17

dicting the exact mutations required to increase the enantioselectivity of an enzyme can be difficult considering the complex nature of enzymes. Not only would this require a high resolution crystal structure to predict free energies of molecular interactions, but also entropic calculation, a factor demonstrated to play a significant role in determining an enzyme’s enantioselectivity [49]. Unfortunately, these calculations cannot be made if high resolution crystal structures are unavailable, as is the case in many situations. Consequently, most improvements in enantioselectivity of enzymes have resulted from directed evolution experiments. In this section, we will outline some of the recent successes in this area. The Acinetobacter sp. NCIMB 9871 cyclohexanone monooxygenase is only moderately enantioselective (14% ee) in creating a chiral (R)-sulfoxide center by oxidizing methyl-p-methylbenzyl thioether. Using epPCR, Reetz and coworkers created a library of random mutants and screened ~10,000 clones, a number large enough to provide near complete coverage of the entire library [50]. With the help of automation, they implemented a mediumthroughput screen able to handle microtiter-plate format for cell growth, product extraction, and analysis by chiral HPLC. They were able to isolate three mutants with >98% ee for the (R)sulfoxide in just a single round of screening. In a similar example, Reetz and coworkers improved the enantioselectivity of Aspergillus niger epoxide hydrolase from E = 4.6 to 10.8 [51]. Again, they used epPCR to introduce mutations in the wild-type gene and screened ~20,000 mutants for hydrolytic kinetic resolution of glycidyl phenyl ether. The best mutant carried three mutations, two of which were distant from the active site. Although they were able to solve the crystal structure of this mutant, they were unable to rationalize the mode of action of the distant mutations. Such results, however, are not uncommon in directed evolution experiments, further highlighting the difficulty in trying to rationally design altered function into an enzyme. Using gene site saturation mutagenesis (GSSM™) for library creation, DeSantis and coworkers engineered a nitrilase for enantioselective hydrolysis of 3-hydroxyglutaronitrile to (R)-4-cyano-3hydroxybutyric acid, a precursor to the cholesterol-lowering drug Lipitor® [52]. The enantioselectivity of parent enzyme (ee = 87.6%) was insufficient for industrial application, and therefore an improved mutant with an ee of 98.5% was isolated after screening >30,000 clones. A high-throughput screen was implemented such that an (R)-15N labeled substrate was hydrolyzed, resulting in the formation of the desired (R)-product and radiolabeled ammonia. An alternative to GSSM™ is to first perform epPCR to identify “hot-spots” that improve the desired property and then perform saturation mutagenesis on those sites. While not as thorough as GSSM, fewer clones need to be screened due to the smaller library size. Using this strategy, Reetz and coworkers evolved a lipase from Pseudomonas aeruginosa to catalyze enantioselective hydrolysis of 2-methyldecanoic acid p-nitrophenyl ester. The wildtype enzyme exhibits a selectivity of E = 1.1, a value far too low for application in the fine chemicals industry where lipases are routinely used for kinetic resolutions [53]. After six iterative rounds of library generation by epPCR and screening, they were able to increase E to 13.5 in favor of the (S)-acid. Dissatisfied with the improvement over six rounds of engineering, they switched to saturation mutagenesis for further rounds. Each site identified by epPCR was thereafter randomized using saturation mutagenesis and screened for improved enantioselectivity. The best mutant identified had an E = 25.8, a significant improvement over the final mutant identified by epPCR.

Nair et al.

From their previous work, Reetz and coworkers realized that epPCR may not be the ideal mutagenesis technique to improve enantioselectivity. Revisiting the same Pseudomonas aeruginosa lipase, they applied combinatorial saturation mutagenesis to a few “hot-spot” residues [54]. After screening only 5,000 clones, they found a mutant with E = 30, which is superior to the best mutant they identified previously. From these findings, Reetz and coworkers discerned that focused mutagenesis at these “hot-spots” is more likely to result in improved mutants. Channeling these insights into a generalized scheme, they developed a technique called CASTing (combinatorial active site testing) for rapid creation of enantioselective enzymes [55]. To demonstrate the power of this methodology, Reetz and coworkers revisited the Aspergillus niger epoxide hydrolase, an enzyme that they had previously engineered by epPCR (above) [51]. Remarkably, they were able to increase the enzyme’s selectivity from E = 4.6 to 115 in favor of the (S)-diol after screening a relatively small number of mutants [55]. Compared to the best mutant identified by epPCR (E = 10.8), this newly identified mutant is far superior, and yet required screening of the same number of clones (20,000). Reversal of Enantioselectivity While it would be generally ideal to start with an enzyme having at least a marginal desired enantioselectivity, there may be situations where such an enzyme does not exist. Reversing the enzyme selectivity may then be one of the only available options. Optically pure amino acids are in demand for synthesis of many antibiotics and drugs. L-5-(2-Methylthioethyl)hydantoin can be converted to L-methionine using a hydantoinase. However, all known hydantoinases are selective for D-5-(2-methylthioethyl) hydantoin rather than the L -enantiomer. To improve the productivity of L -methionine, Arnold and coworkers inverted the enantioselectivity of the hydantoinase from Arthrobacter sp. DSM9771 [56]. After isolating a slightly L-selective mutant (ee = 7% compared to 40% for the wild type enzyme) from a library created by epPCR, all attempts to further improve the enantioselectivity by a second round of random mutagenesis failed. Subsequently, they switched to saturation mutagenesis and focused on a residue identified in their first round of mutagenesis. The mutant identified had an ee = 20%, which albeit low, was accompanied by a 5-fold increase in specific activity. Reetz and coworkers were more successful in isolating an enzyme with high and reversed enantioselectivity in a single round of epPCR. While screening for (R)-specific mutants of cyclohexanone monooxygenases from Acinetobacter sp. NCIMB 9871 (see Increase in enantioselectivity), Reetz and coworkers also identified (S)-specific mutants with ee > 95% [50]. These mutants had a larger number of substitutions than the highly (R)-specific enzymes, indicating that a larger change in structure may be required to reverse enantioselectivity than what is required to simply increase it. This is also demonstrated in another work, where Reetz and coworkers used epPCR with a relatively high mutagenic rate to find enzymes with inverted selectivity [57]. Another excellent example was demonstrated by Boersma and coworkers [58]. By altering the length of the active-site-lining loop in Bacillus subtilis 168 lipase A, they hoped to invert its enantioselectivity. In one of the rare examples of rational design, they modeled the effect of grafting loops from other similar proteins, and followed through with experimental evidence demonstrating the

Engineering of Enzymes for Selective Catalysis

Current Organic Chemistry, 2010, Vol. 14, No. 17 1875

O O

OH

O

wt (+)-IPG

O

O

O

O

R

(-)-IPG ester

O

O

O

R

mutant O

O

O

O

rac-IPG ester O

OH (-)-IPG

O

R

(+)-IPG ester

Fig. (2). The wild-type (wt) Bacillus subtilis 168 lipase A resolves the racemic ester in favor of the (R)-(+)-enantiomer (ee = 12.9%), whereas the mutant forms primarily the 1,2-O-isopropylidene-(S)-(-)-glycerol (IPG) enantiomer [58]. The mutant was created by first altering the active site loop in contact with the substrate to affect steric interactions, followed by saturation mutagenesis.

inversion of enantioselectivity (Fig. 2). To further improve the enantioselectivity, they switched over to iterative saturation mutagenesis and randomized two residues predicted to interact directly with the substrate. The best mutant exhibited an ee of 57.4% for 1,2-O-isopropylidene-(S)-glycerol, compared to the wildtype enzyme’s preference for the (R)-enantiomer (ee = 12.9%).

called GeneReassembly™, Diversa Corporation engineered a cytochrome P450 oxidase to perform regioselective oxidation of avermectin [60]. By incorporating the genetic diversity present in 17 Streptomyces cytochrome P450 enzymes in a combinatorial manner, and screening for improved regioselectivity, they were able to eliminate non-selective oxidative activity in their chimeric mutant (Fig. 3).

Modification of Regioselectivity In addition to controlling the enantioselectivity of a biocatalyst, many enzyme engineering efforts are focused on the modification of regioselectivity. For example, C-H bonds are usually quite unreactive, which makes catalysts capable of activating this bond very useful in organic syntheses. Traditionally, organometallic catalysts containing precious metal centers are required to initiate this reaction by inserting between the two atoms. Monooxygenases and oxidases are able to perform similar C-H bond activation, quite often in a regioselective manner. In addition, they are able to perform these reactions using common transition metals like iron or copper. Realizing that these classes of enzymes can be very useful in difficult syntheses, several groups have used protein engineering tools to make them more amenable for industrial use. Alkanes are quintessential nonreactive compounds, yet cytochrome P450 are able to hydroxylate them at various positions. To perform controlled oxidation of linear alkanes, Lentz and coworkers engineered the cytochrome P450 CYP102A3 from Bacillus subtilis for terminal hydroxylation [59]. By first using epPCR to generate genetic diversity, they isolated a mutant with the ability to produce 1-octanol from octane. The authors then backcrossed the mutations with the parent in order to find the essential mutations, as well as reduce instabilities caused by disruptive mutations. Their final mutant primarily oxidized the terminal position, although residual activity toward the internal position could not be completely eliminated. Synthesis of the potent semisynthetic insecticide Emamectin benzoate requires regioselective oxidation of the 4’’ alcohol of the natural product avermectin, followed by the reductive amination of the resulting key intermediate 4’’-oxo-avermectin. Unfortunately, the large number of groups that can be activated results in formation of many unwanted byproducts. Using a proprietary method

O

O

O

O O O

O 4" HO

O O

O

H

H R

OH

O H OH

mutant P450

O

O

O

O O O

O 4"

O

O O

O

H

H R

OH

O H OH

Fig. (3). Using GeneReassembly™, Diversa Corporation incorporated the genetic diversity present in 17 Streptomyces cytochrome P450 enzymes to create a regioselective hybrid that was able to oxidize the 4”-hydroxyl to produce the oxo-avermectin intermediate for avermectin synthesis [60].

1876 Current Organic Chemistry, 2010, Vol. 14, No. 17

Cytochrome P450s in humans, among other functions, play an important role in drug metabolism and clearance, as well as in hormone synthesis. Due to their clinical relevance, several highresolution crystal structures have been solved, which enables rational design of these proteins by comparative analysis. Following this strategy, Kumar and coworkers engineered an enzyme with high activity and 80% regioselectivity for position 21 on progesterone [61]. Starting with cytochrome P450 2B1, a progesterone 16
hydroxylase, they introduced mutations to model its active site after cytochrome P450 2C5, a progesterone 21-hydroxylase. Using the best mutant for the next round of site-directed mutagenesis, they gradually converted the low-affinity hydroxylase to a highly regioselective 2C5-like enzyme over four rounds. A final example is the engineering of tetrachlorobenzene dioxygenase from Ralstonia sp. PS12 for hydroxylation of sites not oxidized by the wild-type enzyme [62]. Halogenated aromatics are usually toxic pollutants and certain microbes, including Ralstonia sp. PS12, are able to metabolize them, making them good candidates for bioremediation. Tetrachlorobenzene dioxygenase is the first step in the aerobic biodegradation of chlorinated toluenes and supports the growth of Ralstonia sp. PS12 on these substrates. Using the crystal structure of the homologous enzyme naphthalene dioxygenase, and mechanistic insights from studies in nitrotoluene dioxygenase, the authors changed the regioselectivity of tetrachlorobenzene dioxygenase using a single substitution. The mutant was able to produce various novel dichloromethylcatechols. 4. ENGINEERING ENZYMES CATALYZING CARBONCARBON BOND FORMATION Carbon-carbon (C-C) bond formation is one of the most powerful synthetic reactions for the industrial synthesis of chiral and asymmetric compounds [63, 64]. Known as a cornerstone of modern organic chemistry, the reactions lead to the formation of a new carbon-carbon bond and up to two new stereogenic centers under mild conditions [63]. C-C bond forming reactions have been widely used in the stereo-controlled synthesis of natural products and bioactive small molecules using various metal enolates, as well as metal-complex-catalyzed and organocatalytic methods [65]. In addition, several enzyme classes have been shown to catalyze C-C bond forming reactions, including the transketolases, oxynitrilases, and aldolases [63], as well as other thiamine pyrophosphate (TPP) independent enzymes [66]. By modifying or expanding the substrate repertoire and by modifying the stereochemical properties, alternative stereoisomeric products can be synthesized. The most recent directed evolution examples have mainly been geared towards aldolases and transketolases. Few successes have been achieved with oxynitrilases and TPP dependent enzymes in C-C bond formation. Directed evolution of aldolases with new and improved properties has been achieved by two methods. First, promiscuous aldolases have been created that are able to accept more than one stereoisomeric substrate in order to prepare more than one stereoisomeric product through the choice of the substrate provided. Second, a genuinely selective enzyme has been generated that is capable of catalyzing the formation of a diastereoisomeric product, using the same substrates as the wild-type enzyme [64]. Among the aldolases shown to be useful in selective catalysis is the 2-deoxy-D-ribose-5-phosphate aldolase (DERA) from E. coli. In vivo, DERA catalyzes the reversible retro-aldol degradation reaction of 2-deoxy-D-ribose-5-phosphate to D-glyceraldehyde-3phosphate and acetaldehyde, although the equilibrium lies on the 2-

Nair et al.

deoxy-D-ribose-5-phosphate side (Fig. 4A) [67]. Unlike other aldolases that catalyze the aldol addition of an aldehyde and a ketone, DERA catalyzes the asymmetric aldol addition of two aldehydes. In light of this novel property, Wong and coworkers used rational design to expand the substrate scope of DERA towards the unnatural nonphosphorylated substrate D-2-deoxyribose [68]. Of five mutants created by site-directed mutagenesis, the S238D mutant showed a 2.5-fold improvement over the wild-type enzyme in the retro-aldol reaction of D-2-deoxyribose. Moreover, this mutant accepted 3-azidopropinaldehyde as a substrate to form deoxyazidoethyl pyranose, which is a precursor for LipitorTM (Fig. 4B) [68]. DERA is also a potential biocatalyst for the industrial synthesis of (3R, 5S)-6-chloro-2,4,6-trideoxyhexapyranoside, a versatile precursor for statin drugs like LipitorTM [63]. However, low affinity towards chloroacetaldehyde and enzyme deactivation by high aldehyde concentration limit the enzyme’s usefulness for industrial applications. Using independent high throughput screens for chloroacetaldehyde resistance and activity, several mutants from a random mutagenesis library were identified [63]. By combining the beneficial mutations leading to improved stability or catalytic properties of DERA, a triple mutant showed the highest chloroacetaldehyde tolerance, while being an efficient catalyst of the precursor (Fig. 4C) [63]. Directed evolution has also proven to be successful in the engineering of the 2-keto-3-deoxy-6-phosphogalactonate (KDPGal) aldolase from the shikimate pathway [69, 70]. Ran and Frost devised a new strategy to synthesize 3-deoxy-D-arabino-heptylosonic acid 7-phosphate (DAHP), originally catalyzed by DAHP synthase in the first step of the shikimate pathway, by using the KDPGal. Yields of microbially synthesized chemicals are limited due to the competition between DAHP synthase and the sugar-transporting phosphoenolpyruvate: carbohydrate phosphotransferase system (PTS) for cytoplasmic supplies of phosphoenolpyruvate. In this new strategy, the E. coli KDPGal, which normally catalyzes the cleavage of KDPGal to pyruvate and D-glyceraldehyde-3-phosphate, was evolved by directed evolution to catalyze the reversible condensation of pyruvate and D-erythrose-4-phosphate to form DAHP. By utilizing pyruvate (a PTS byproduct), the KDPGal synthesis of DAHP completely avoided competition for phosphoenolpyruvate from the PTS system [70]. A combination of epPCR, DNA shuffling and multiple site-directed mutagenesis yielded KDPGal variant NR8.276-2, which showed a 60-fold improvement in catalytic efficiency relative to the wild-type KDPGal aldolase in catalyzing the addition of pyruvate and D-erythrose-4-phosphate to form DAHP [69]. This study shows high applicability in the industrial production of shikimic acid [71], a relevant starting material for the manufacture of a plethora of commodity chemicals. In a similar study, using a combination of epPCR and DNA shuffling, Wong and coworkers evolved the E. coli D-2-keto-3deoxy-6-phosphogluconate (KDPG) aldolase into one with improved catalytic efficiency, altered substrate specificity and stereoselectivity [72]. By screening a small library (10,000), some aldolases evolved the capability to accept both the D- and L glyceraldehyde in the non-phosphorylated forms as substrates for the reversible aldol reaction. This provides a new direction in the enzymatic synthesis of D- and L -sugars. In another related study, N-acetylneuraminic acid aldolase (Neu5Ac aldolase) catalyzes the aldol condensation of N-acetyl-Dmannosamine and pyruvate to give N-acetyl-D-neuraminic acid (Dsialic acid). By using epPCR, Wong and coworkers generated mu-

Engineering of Enzymes for Selective Catalysis

A

Current Organic Chemistry, 2010, Vol. 14, No. 17 1877

O

O

O

OH

WT DERA H

OPO3

CH3

OPO3 OH

OH

HO

O O F

B

HO

O

O

O

1. S238D mutant

N

O

H

H

N3

O

2. Br2, BaCO3 35% for 2 steps O

NH N3

Atorvastatin (LipitorTM) O

C O

OH

O

O

OH

Cl

Cl

(S)

DERA triple mutant

(S)

DERA triple mutant

OH

O

(R)

O

O Cl

OH

Cl

O OH

(3R,5S)-6-chloro-2,4,6trideoxyhexapyranoside

Cl

LipitorTM OH

(3R,5S)-6-chloro-2,4,6trideoxyhexonolactone

Fig. (4). A) DERA catalyzed aldol reaction between the natural donor acetaldehyde and acceptor D-glyceraldehyde-3-phosphate to generate D-2-deoxyribose5-phosphate [67]. B) The S238D DERA variant is able to accept nonphosphorylated substrates as acceptors and catalyzes a novel sequential aldol reaction using 3-azidopropinaldehyde as the first acceptor and two molecules of acetaldehyde as donor to form an azidoethyl pyranose, a key intermediate useful for the synthesis of the cholesterol lowering agent LipitorTM [68]. C) Using chloroacetaldehyde and two equivalents of acetaldehyde, the triple DERA variant catalyzes the synthesis of (3R,5S)-6-chloro-2,4,6-trideoxyhexapyranoside, a versatile chiral building block for the synthesis of vastatin drugs like LipitorTM , under industrially relevant conditions [63].

tants with improved activity towards L-configured substrates including N-acetyl-L-mannosamine and L-arabinose [73]. Ten thousand colonies in each round were screened for their ability to accept 3-deoxy-L-manno-octulosonic acid (L-KDO) using a coupled assay based on the reduction of pyruvate to lactate by lactate dehydrogenase. The evolved aldolase, containing three mutations, showed an inverted enantioselectivity towards the formation of L-sialic acid and L-KDO. In particular, a >1,000-fold improvement in the ratio of the specificity constants [kcat/K M (L-KDO)]/[kcat/K M (N-acetyl Dneuraminic acid)] was observed for this mutant. In addition, Williams and coworkers showed that the configuration of the product of an aldolase-catalyzed reaction can be varied by altering the stereochemical course of carbon-carbon bond formation through directed evolution of the E. coli N-acetylneuraminic acid lyase (NAL) [74]. Starting from the E192N NAL mutant, the stereochemical control of an aldolase-catalyzed reaction was altered by epPCR, followed by an intense structure-guided program of saturation and site-directed mutagenesis. A complementary pair of variants were identified that were approximately 50-fold more selective towards the cleavage of the alternative 4S- and 4Rconfigured condensation products, respectively, from the same starting material as used for the wild-type enzyme [74].

As a final example, directed evolution was used to create a variant of the thiamine diphosphate-dependent enzyme benzoylformate decarboxylase (BFD) from Pseudomonas putida that accepts ortho-substituted benzaldehyde as a donor substrate [75]. Two identified mutants could selectively catalyze the formation of enantiopure (S)-2-hydroxy-1-(2-methylphenyl)propan-1-one (>93% ee) with high yields (99%) - a reaction not catalyzed by the wild-type enzyme. Different ortho-substituted derivatives such as 2-chloro-, 2-methoxy-, or 2-bromo-benzaldehyde were also accepted as donor substrates to form the corresponding (S)-2-hydroxy-1-(2methylphenyl)propan-1-one derivatives [75]. 5. ENGINEERING ENZYMES CATALYZING CARBONOXYGEN BOND CLEAVAGE OR FORMATION Glycosidases (also called glycoside hydrolases or saccharidases) catalyze the selective hydrolysis of glycosidic bonds between sugar molecules (Fig. 5). While they have found extensive use primarily in the food, textile, and paper industries, recent trends for biofuels production has made them attractive targets for lignocellulose disassembly. There have been many examples of glycoside engineering for improved thermal stability, stability at extreme

1878 Current Organic Chemistry, 2010, Vol. 14, No. 17

glycosidase

O

O

Nair et al.

O

O

OH +

O

HO

+ H2O O

O-pNP

O

transglycosidase

O

+

HO

O

O

+ pNP-OH

O

O

glycosynthase

O

+ HO

O

O

+ H+ + F-

F

O

O-NDP

+

glycosyltransferase

O HO OH

O

O O

+ H+ + O-NDP

OH

Fig. (5). Reactions catalyzed by various enzymes involved in breaking and forming bonds between sugars. Glycosidases catalyze the hydrolysis of glycosidic bonds between sugar molecules. Transglycosidases are a minor and thermodynamically unfavorable activity of glycosidase, whereby activated glycosides (like para-nitrophenyl sugars) can be linked to an acceptor sugar. Glycosynthases are glycosidases with a mutated active site residue that renders them unable to perform hydrolysis. They are however, able to transfer a glycosyl fluoride with the opposite anomeric configuration onto an acceptor sugar molecule. Glycosyltransferases are naturally occurring enzymes that catalyze the formation of glycosidic bonds between glycosylnucleotides.

pH conditions, improved catalytic activity, etc., making them more suitable for the aforementioned applications. Biocatalytic hydrolysis of oligosaccharides, while industrially relevant, is somewhat uninteresting for synthetic chemistry. However, synthesis of bioactive glycosides with either medicinal or microbicidal properties is of great interest to pharmaceuticals and biotech companies. Glycosidases have a minor transglycosidase activity whereby they can link activated sugars to form di- or oligosaccharides in a regioselective manner (Fig. 5). Unfortunately, transglycosylation is a thermodynamically unfavorable reaction and products are hydrolyzed quickly back to monosaccharides. To address this limitation, Dion and coworkers engineered the Thermus thermophilus glycosidase for improved -transglycosidase activity. The wild-type enzyme has 50% efficiency for self-condensation, but only 8% efficiency for transglycosylation on unactivated sugar. Using epPCR and combinatorial saturation mutagenesis on identified sites, they were able to significantly reduce the hydrolytic activity, thereby improving the overall yield of transglycosylation. For some of the best mutants, self-condensation was nearly quantitative, whereas transglycosylation to cellobiose or maltose reached 6075% yield. Similar work was subsequently performed by Osanjo and coworkers, who engineered
-l-fucosidase from Thermotoga maritima with improved transferase/hydrolysis kinetic ratio [76]. The wildtype
-l-fucosidase can catalyze the regioselective
-(12) transglycosylation of pNP-fucoside (para-nitrophenyl-fucoside) donor to pNP-galactoside with 7% yield. Using epPCR followed by combinatorial site-directed mutagenesis on their most promising mutants, they were able to isolate a variant with improved transglycosidase activity. It was able to catalyze formation of pNP-Gal-
-(12)-Fuc with 60% yield starting with equimolar amounts of glycosides. Even though the success with engineering transglycosidase activity was promising, Withers and coworkers were not content with just reducing hydrolytic activity. By mutating the active site nucleophile (glutamate) to alanine, they were able to completely eliminate hydrolytic activity, creating a new family of glycosidase mutants called glycosynthases (Fig. 5) [77]. These mutants are able to transfer a glycosyl fluoride with the opposite anomeric configu-

ration to their natural substrate onto a suitable acceptor molecule. The first glycosynthase created was from Agrobacterium sp. glucosidase/galactosidase with the hydrolytic glutamate replaced with alanine. Using this variant, they were able to synthesize Gal(1,4)-Glc-(1,4)-pNP-Glc starting from
-galactosyl fluoride and pNP--cellobioside with 92% yield. Cornish and coworkers took this approach a step further by looking for a more suitable substitution for the hydrolytic glutamate [78]. Using the glycosidase from Humicola insolens rather than the Agrobacterium sp., they randomized the active site residue responsible for hydrolysis and selected for an appropriate mutant using a high-throughput selection. By linking the glycosynthase activity to the transcription of an essential gene in yeast, they developed a three-hybrid assay. The best variant with serine substitution had a 5-fold higher synthase activity than the alanine substituent. While glycosynthases and transglycosidases are variants of glycosidases that have been engineered to catalyze the “reverse” of their natural reaction, nature’s catalysts for these reactions are glycosyltransferases (GTs) (Fig. 5). Much effort has been put into engineering the substrate range of GTs in hope of creating nonnatural glycosides with novel bioactive properties – a process called glycorandomization. Unfortunately, the inability to screen for variants with promiscuous activity in a high-throughput manner has limited the success of most such undertakings. Thorson and coworkers developed a high-throughput screen in order to find these elusive promiscuous GTs. In their screen, they searched for mutants able to transfer a sugar onto a surrogate substrate (a fluorescent coumarin). Successful GT activity with this substrate would result in quenching of fluorescence, providing a detectable output. Starting with the GT OleD from Streptomyces antibioticus involved in decorating the antibiotic oleandomycin, they performed epPCR to generate a library of mutants in hope of finding variants with promiscuous activity. By combining mutations from positive hits in a combinatorial manner, they identified a mutant with three substitutions that could accept 15 sugar nucleotides with improved or novel activity compared to the parent enzyme. Realizing that the three mutated residues may be “hot-spots” for promiscuous activity, Thorson and coworkers also tried to find mutants with the ability to

Engineering of Enzymes for Selective Catalysis

Current Organic Chemistry, 2010, Vol. 14, No. 17 1879

Fig. (6). The modular nature of type I polyketide synthases is demonstrated by 6-deoxyerythronolide B synthase (DEBS) from Saccharopolyspora erythraea. The wild type enzyme accepts propionyl-CoA (coenzyme A) as a starter unit and extends the polyketide with acetyl groups. The replacement of the loading domain of DEBS with that from the avermectin-producing polyketide synthase of Streptomyces avermitilis, allows for a wider range of starter units such as acetyl-CoA and isobutyryl-CoA [81]. AT = acyltransferase, ACP = acyl-carrier protein, KS = ketosynthase, KR = ketoreductase, DH = dehydratase, ER = enoyl reductase, and TE = thioesterase.

decorate the nonnatural acceptor novobiocic acid with nonnatural sugars [79]. Using saturation mutagenesis on the three “hot-spot” sites, they created a library of mutants and screened through them using their previously developed assay for promiscuity. Surprisingly, even though they screened for transferase activity with nonnatural donors, their final mutant was promiscuous even for acceptor molecules and decorated novobiocic acid with several hundred fold higher efficiency. Realizing the modular nature of certain GTs, Park and coworkers tried to combine fragments from different proteins to create chimeras with novel activity [80]. Since the N-terminus domain recognizes the acceptor while the C-terminus domain recognizes the donor sugar nucleotide, they created chimeras of kanamycin GT (KanF) and C-terminal fragments of vancomycin GT (GtfE). They found that the most active mutant was able to transfer the glycone substrate of GtfE onto the aglycone acceptor of KanF. However, in a fortuitous turn of events, they found that the hybrid was also able to accept substrates not natural to either parent protein. 6. ENGINEERING SINGLE-CHAIN POLYKEKETIDE SYNTHASES Polyketides are a class of secondary metabolites that encompass a wide range of biological functions, including antibiotic, antitumor, antifungal, immunosuppressive, and roles in defense against predators, amongst others. Among the most well known polyketides are lovastatin, doxorubicin, rapamycin S, erythromycin A, and various tetracyclines. As a result, this group of molecules constitutes a vast resource for the discovery of novel bioactive and therapeutic compounds. Like for most bioactive compounds, chemical

modification of polyketides has been used to diversify them in order to find variants with improved activity. However, certain functional groups are resistant to functionalization due to inability to activate them selectively. In such cases, modification may have to be introduced during their biosynthesis. Within cells, polyketides are produced by polyketide synthases (PKSs), a family of proteins closely related to fatty acid synthases. Although they have been discovered in a variety of organisms, those from plants, fungi, and bacteria are most well known and closely studied. PKSs are divided into three groups based on their structure and function. Type I PKSs are large polypeptides with multiple domains, each with its own catalytic site and function. Type II PKSs are complexes of smaller associated proteins, each with its own distinct function. Type III PKSs, unlike the other two, are not modular in nature, but consist of one polypeptide with a single active site. In all three cases, polyketide synthesis begins with a starter unit onto which elongation units are added in stepwise condensation reactions. Other domains are responsible for modifying the carbonyl groups, and in certain cases, prior to product release the polyketide may be circularized. With the aim of creating novel molecules with biological activity, several groups have undertaken efforts to understand and engineer PKSs. Taking advantage of the modular nature of Type I PKSs, Leadlay and colleagues grafted the loading module of the avermectin-producing PKS onto that of the erythromycin-producing protein [81]. Since the module from the avermectin PKS has a much more relaxed substrate preference, with the ability to accept more than 40 starter carboxylic acid units, their hybrid was able to produce various erythromycin derivatives (Fig. 6). Amazingly, they

1880 Current Organic Chemistry, 2010, Vol. 14, No. 17

did not need to perform any linker engineering considering communication via interdomain sequences is essential for proper chain elongation. Patel and coworkers took this approach a step further and made even more extensive grafts to a Type I PKS [82]. To make derivates of the anticancer drug geldanamycin with substitutions at nonreactive functional groups, they exchanged six out of the seven acyltransferase domains of the corresponding PKS. Since acyltransferases are responsible for loading starter and elongation units, they reasoned this would create derivatized geldanamycin. Four of these substitutions resulted in fully functional proteins that produced geldanamycin derivates. Of these one had a higher affinity for the drug target (Hsp90), thus promising an even more potent anticancer activity. With a similar aim, Tang and coworkers aimed to create a more potent epothilone anticancer agent. Instead of swapping acyltransferase domains, they inactivated certain ketoreductase domains. Since these domains are responsible for reducing carbonyl groups to hydroxyl groups, inactivation of their active site would result in oxo-derivates. In addition to finding these expected compounds, they also found some unexpected variants suggesting alternate activities for domains based on the identity of bound substrate. Frost, Zhao and coworkers used a similar strategy to produce triacetic acid lactone (TAL), an intermediate in the production of the energetic compound 2,4,6-trinitrobenzene [83]. They inactivated the ketoreductase module of 6-methyl sialic acid synthase (6-MSAS), a type I PKS, using a single point mutation. The loss of reducing activity led to the premature lactonization of linear triketide with the formation of their desired product. Since type I PKSs are modular proteins, altering product specificity is conceptually simpler thanks to the ability to inactivate or modify a single domain whilst leaving the others unperturbed. However, as exemplified by 6-MSAS, where loss of ketoreductase activity resulted in premature lactonization, mutations can have unexpected results. Such cases are even more common for type III PKSs that have a single active site. Abe and coworkers demonstrated this by creating various mutations at a single position and found that it had a profound effect on the product [84]. Using the type III PKS from aloe, they found that mutation of the small amino acid glycine at position 207 to larger residues like tryptophan, methionine, leucine, or phenylalanine resulted in change of product profile from octaketides to tri- through hepta-ketides. Finally, they also found that changing a large residue of a pentaketide synthase, also from aloe, to small amino acids resulted in the formation of octaketides [85]. Furthermore, additional mutations to increase active site volume changed the PKS to a nonaketide synthase. FUTURE PROSPECTS The age of genomics has ushered in an exciting time for biocatalysis. The availability of metagenomic data and bioinformatic resources has significantly accelerated the pace of new enzyme and pathway discovery. This is particularly helpful since laboratory cultivation is no longer a necessity for such identifications, which would have otherwise limited scientists and engineers to culturable organisms only. Armed with the ability to find new pathways and enzymes, the challenge has now shifted on how to use this data to create industrially and medically relevant proteins and compounds. To that end, directed evolution has played a significant role in expediting the timeline from discovery to application. As described in the preceding sections, protein engineering has been used to engineer enzymes with a variety of properties to make them more ame-

Nair et al.

nable for catalysis. Substrate specificity, enantioselectivity, regioselectivity, and the range of products have all been altered to varying degrees using directed evolution. Although rational design has been used in some cases, there are far fewer reports of its successful application compared to directed evolution primarily because it requires structural or mechanistic data that are unavailable for most proteins. Directed evolution’s bottom-up approach has minimized the requirements to engineer a protein significantly – the only caveat is the requirement for a medium-to-high throughput screen. To some degree, this issue has been addressed by limiting sequence space explored using semi-rational design techniques like combinatorial saturation mutagenesis [86] or CASTing [55]. In addition, newer screens and selections are constantly being developed, as are inventive ways to adapt assays to high-throughput methods like FACS (fluorescence activated cell sorting) to a wide variety of enzymatic assays [86]. Not content with stepwise evolution of naturally occurring enzymes, many groups have been looking into creating synthetic enzymes. Schultz and coworkers have shown that noncanonical amino acids can be efficiently incorporated into proteins using designer organisms [87]. This can be used to extract new chemistry not available through the standard twenty amino acids. Others like Reetz and Ward have created hybrid enzymes, i.e., metalloenzymes with nonnatural metal centers [88, 89], thereby trying to combine the interesting chemistry afforded by organometallic catalysts with the high specificity offered by enzymes. Lu and coworkers have combined the two aforementioned ideas to generate proteins with unnatural amino acids as well as heavy metal centers [90]. Catalytic antibodies are another avenue that has been explored to impart catalytic activity to the innately highly selective antibodies [91]. Baker and coworkers have used computational techniques to create active sites designed around a specified reaction and grafted them onto naturally occurring scaffolds to create proteins with novel activities [92, 93]. Yet current research has barely scratched the surface of these fields. Their incorporation into protein engineering for industrial use is still in its nascent stages. Future prospects for protein engineering, and particularly its synergy with synthetic enzymes, should lead to an expansive set of catalysts for use in chemistry and industry. ACKNOWLEDGEMENTS We thank the National Institutes of Health (GM077596), the Biotechnology Research and Development Consortium (BRDC) (Project 2-4-121), the British Petroleum Energy Biosciences Institute, the National Science Foundation as part of the Center for Enabling New Technologies through Catalysis (CENTC), CHE0650456 for supporting our studies related to biocatalysis and biosynthesis. N.U.N. also acknowledges Drickamer Fellowship support from the Department of Chemical and Biomolecular Engineering at the University of Illinois. REFERENCES [1] [2] [3]

[4]

Schoemaker, H. E.; Mink, D.; Wubbolts, M. G. Dispelling the myths-biocatalysis in industrial synthesis. Science, 2003, 299, 1694-1697. Rubin-Pitel, S. B.; Zhao, H. Recent advances in biocatalysis by directed enzyme evolution. Comb. Chem. High Throughput Screen, 2006, 9, 247-257. Zhao, H.; Chockalingam, K.; Chen, Z. Directed evolution of enzymes and pathways for industrial biocatalysis. Curr. Opin. Biotechnol., 2002, 13, 104110. Otten, L. G.; Quax, W. J. Directed evolution: selecting today's biocatalysts. Biomol. Eng., 2005, 22, 1-9.

Engineering of Enzymes for Selective Catalysis [5] [6] [7] [8] [9]

[10] [11]

[12] [13]

[14]

[15]

[16] [17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

Arnold, F. H. Combinatorial and computational challenges for biocatalyst design. Nature, 2001, 409, 253-257. Johannes, T. W.; Zhao, H. Directed evolution of enzymes and biosynthetic pathways. Curr. Opin. Microbiol., 2006, 9, 261-267. Kaur, J.; Sharma, R. Directed evolution: an approach to engineer enzymes. Crit. Rev. Biotechnol., 2006, 26, 165-199. Stemmer, W. P. Rapid evolution of a protein in vitro by DNA shuffling. Nature, 1994, 370, 389-391. Zhao, H.; Giver, L.; Shao, Z.; Affholter, J. A.; Arnold, F. H. Molecular evolution by staggered extension process (StEP) in vitro recombination. Nat. Biotechnol., 1998, 16, 258-261. Zhao, H.; Arnold, F. H. Combinatorial protein design: strategies for screening protein libraries. Curr. Opin. Struct. Biol., 1997, 7, 480-485. Wen, F.; MacLachlan, M.; Zhao, H. Novel and improved enzymes through directed evolution, in Wiley Encyclopedia of Chemical Biology. John Wiley and Sons: NJ, 2008. van Beilen, J. B.; Duetz, W. A.; Schmid, A.; Witholt, B. Practical issues in the application of oxygenases. Trends Biotechnol., 2003, 21, 170-177. Munro, A. W.; Leys, D. G.; McLean, K. J.; Marshall, K. R.; Ost, T. W.; Daff, S.; Miles, C. S.; Chapman, S. K.; Lysek, D. A.; Moser, C. C.; Page, C. C.; Dutton, P. L. P450 BM3: the very model of a modern flavocytochrome. Trends Biochem. Sci., 2002, 27, 250-257. McLean, K. J.; Sabri, M.; Marshall, K. R.; Lawson, R. J.; Lewis, D. G.; Clift, D.; Balding, P. R.; Dunford, A. J.; Warman, A. J.; McVey, J. P.; Quinn, A. M.; Sutcliffe, M. J.; Scrutton, N. S.; Munro, A. W. Biodiversity of cytochrome P450 redox systems. Biochem. Soc. Trans., 2005, 33, 796-801. Fasan, R.; Chen, M. M.; Crook, N. C.; Arnold, F. H. Engineered alkanehydroxylating cytochrome P450(BM3) exhibiting nativelike catalytic properties. Angew. Chem. Int. Ed. Engl., 2007, 46, 8414-8418. Labinger, J. A.; Bercaw, J. E. Understanding and exploiting C-H bond activation. Nature, 2002, 417, 507-514. Glieder, A.; Farinas, E. T.; Arnold, F. H. Laboratory evolution of a soluble, self-sufficient, highly active alkane hydroxylase. Nat. Biotechnol., 2002, 20, 1135-1139. Fasan, R.; Meharenna, Y. T.; Snow, C. D.; Poulos, T. L.; Arnold, F. H. Evolutionary history of a specialized P450 propane monooxygenase. J. Mol. Biol., 2008, 383, 1069-1080. Kubo, T.; Peters, M. W.; Meinhold, P.; Arnold, F. H. Enantioselective epoxidation of terminal alkenes to (R)- and (S)-epoxides by engineered cytochromes P450 BM-3. Chemistry, 2006, 12, 1216-1220. Haines, D. C.; Tomchick, D. R.; Machius, M.; Peterson, J. A. Pivotal role of water in the mechanism of P450BM-3. Biochemistry, 2001, 40, 1345613465. Meinhold, P.; Peters, M. W.; Chen, M. M.; Takahashi, K.; Arnold, F. H. Direct conversion of ethane to ethanol by engineered cytochrome P450 BM3. Chembiochem, 2005, 6, 1765-1768. Xu, F.; Bell, S. G.; Lednik, J.; Insley, A.; Rao, Z.; Wong, L. L. The heme monooxygenase cytochrome P450cam can be engineered to oxidize ethane to ethanol. Angew. Chem. Int. Ed. Engl., 2005, 44, 4029-4032. Sohn, H.; Kim, Y. S.; Jin, U. H.; Suh, S. J.; Lee, S. C.; Lee, D. S.; Ko, J. H.; Kim, C. H. Alteration of the substrate specificity of Thermus caldophilus ADP-glucose pyrophosphorylase by random mutagenesis through errorprone polymerase chain reaction. Glycoconjug. J., 2006, 23, 619-625. Ballicora, M. A.; Erben, E. D.; Yazaki, T.; Bertolo, A. L.; Demonte, A. M.; Schmidt, J. R.; Aleanzi, M.; Bejar, C. M.; Figueroa, C. M.; Fusari, C. M.; Iglesias, A. A.; Preiss, J. Identification of regions critically affecting kinetics and allosteric regulation of the Escherichia coli ADP-glucose pyrophosphorylase by modeling and pentapeptide-scanning mutagenesis. J. Bacteriol., 2007, 189, 5325-5333. Jourden, M. J.; Clarke, C. N.; Palmer, A. K.; Barth, E. J.; Prada, R. C.; Hale, R. N.; Fraga, D.; Snider, M. J.; Edmiston, P. L. Changing the substrate specificity of creatine kinase from creatine to glycocyamine: evidence for a highly evolved active site. Biochim. Biophys. Acta., 2007, 1774, 1519-1527. Fellner, T.; Piribauer, P.; Ogris, E. Altering the holoenzyme composition and substrate specificity of protein phosphatase 2A. Methods Enzymol., 2003, 366, 187-203. Jeppsson, M.; Traff, K.; Johansson, B.; Hahn-Hagerdal, B.; GorwaGrauslund, M. F. Effect of enhanced xylose reductase activity on xylose consumption and product distribution in xylose-fermenting recombinant Saccharomyces cerevisiae. FEMS Yeast Res., 2003, 3, 167-175. Zeng, Q. K.; Du, H. L.; Wang, J. F.; Wei, D. Q.; Wang, X. N.; Li, Y. X.; Lin, Y. Reversal of coenzyme specificity and improvement of catalytic efficiency of Pichia stipitis xylose reductase by rational site-directed mutagenesis. Biotechnol. Lett., 2009, DOI: 10.1007/s10529-10009-19980-x. Watanabe, S.; Kodaki, T.; Makino, K. Complete reversal of coenzyme specificity of xylitol dehydrogenase and increase of thermostability by the introduction of structural zinc. J. Biol. Chem., 2005, 280, 10340-10349. Koch, D. J.; Chen, M. M.; van Beilen, J. B.; Arnold, F. H. In vivo evolution of butane oxidation by terminal alkane hydroxylases AlkB and CYP153A6. Appl. Environ. Microbiol., 2009, 75, 337-344. Peters, M. W.; Meinhold, P.; Glieder, A.; Arnold, F. H. Regio- and enantioselective alkane hydroxylation with engineered cytochromes P450 BM-3. J. Am. Chem. Soc., 2003, 125, 13442-13450. Fadnavis, N. W.; Radhika, K. R.; Vedamayee Devi, A. Preparation of enantiomerically pure (R)- and (S)-3-amino-3-phenyl-1-propanol via resolution

Current Organic Chemistry, 2010, Vol. 14, No. 17 1881

[33]

[34]

[35] [36]

[37]

[38] [39] [40]

[41] [42]

[43] [44]

[45]

[46] [47] [48] [49]

[50]

[51]

[52]

[53] [54]

[55]

[56]

[57]

[58]

[59]

[60]

with immobilized penicillin G acylase. Tetrahedron Asymmetry, 2006, 17, 240-244. Sigmund, A. E.; DiCosimo, R. Enzymatic resolution of (RS)-2-(1aminoethyl)-3-chloro-5-(substituted)pyridines. Tetrahedron Asymmetry, 2004, 15, 2797-2799. Bálint, J.; Schindler, J.; Egri, G.; Hanusz, M.; Marthi, K.; Juvancz, Z.; Fogassy, E. Resolution of methyl-1-phenylethylamines by acidic derivatives of 1-phenylethylamine. Tetrahedron Asymmetry, 2004, 15, 3401-3405. Shin, J.-S.; Kim, B.-G.; Shin, D.-H. Kinetic resolution of chiral amines using packed-bed reactor. Enzyme Microb. Technol., 2001, 29, 232-239. Alexeeva, M.; Enright, A.; Dawson, M. J.; Mahmoudian, M.; Turner, N. J. Deracemization of alpha-methylbenzylamine using an enzyme obtained by in vitro evolution. Angew. Chem. Int. Ed. Engl., 2002, 41, 3177-3180. Reetz, M. T.; Bocola, M.; Carballeira, J. D.; Zha, D. X.; Vogel, A. Expanding the range of substrate acceptance of enzymes: Combinatorial active-site saturation test. Angew. Chem. Int. Ed. Engl., 2005, 44, 4192-4196. Zhao, H.; van der Donk, W. A. Regeneration of cofactors for use in biocatalysis. Curr. Opin. Biotechnol., 2003, 14, 583-589. van der Donk, W. A.; Zhao, H. Recent developments in pyridine nucleotide regeneration. Curr. Opin. Biotechnol., 2003, 14, 421-426. Bommarius, A. S.; Drauz, K. An enzymatic route to L-ornithine from arginine--activation, selectivity and stabilization of L-arginase. Bioorg. Med. Chem., 1994, 2, 617-626. Mayer, K. M.; Arnold, F. H. A colorimetric assay to quantify dehydrogenase activity in crude cell lysates. J. Biomol. Screen., 2002, 7, 135-140. Woodyer, R.; van der Donk, W. A.; Zhao, H. M. Relaxing the nicotinamide cofactor specificity of phosphite dehydrogenase by rational design. Biochemistry, 2003, 42, 11604-11614. Nair, N. U.; Zhao, H. Evolution in reverse: engineering a D-xylose-specific xylose reductase. Chembiochem, 2008, 9, 1213-1215. Granstrom, T. B.; Izumori, K.; Leisola, M. A rare sugar xylitol. Part II: biotechnological production and future applications of xylitol. Appl. Microbiol. Biotechnol., 2007, 74, 273-276. Gupta, N.; Farinas, E. T. Narrowing laccase substrate specificity using active site saturation mutagenesis. Comb. Chem. High Throughput Screen, 2009, 12, 269-274. Bornscheuer, U. T. Methods to increase enantioselectivity of lipases and esterases. Curr. Opin. Biotechnol., 2002, 13, 543-547. Hult, K.; Berglund, P. Engineered enzymes for improved organic synthesis. Curr. Opin. Biotechnol., 2003, 14, 395-400. Turner, N. J. Controlling chirality. Curr. Opin. Biotechnol., 2003, 14, 401406. Ottosson, J.; Rotticci-Mulder, J. C.; Rotticci, D.; Hult, K. Rational design of enantioselective enzymes requires considerations of entropy. Protein Sci., 2001, 10, 1769-1774. Reetz, M. T.; Daligault, F.; Brunner, B.; Hinrichs, H.; Deege, A. Directed evolution of cyclohexanone monooxygenases: enantioselective biocatalysts for the oxidation of prochiral thioethers. Angew. Chem. Int. Ed. Engl., 2004, 43, 4078-4081. Reetz, M. T.; Torre, C.; Eipper, A.; Lohmer, R.; Hermes, M.; Brunner, B.; Maichele, A.; Bocola, M.; Arand, M.; Cronin, A.; Genzel, Y.; Archelas, A.; Furstoss, R. Enhancing the enantioselectivity of an epoxide hydrolase by directed evolution. Org. Lett., 2004, 6, 177-180. DeSantis, G.; Wong, K.; Farwell, B.; Chatman, K.; Zhu, Z.; Tomlinson, G.; Huang, H.; Tan, X.; Bibbs, L.; Chen, P.; Kretz, K.; Burk, M. J. Creation of a productive, highly enantioselective nitrilase through gene site saturation mutagenesis (GSSM). J. Am. Chem. Soc., 2003, 125, 11476-11477. Houde, A.; Kademi, A.; Leblanc, D. Lipases and their industrial applications: an overview. Appl. Biochem. Biotechnol., 2004, 118, 155-170. Reetz, M. T.; Wilensek, S.; Zha, D.; Jaeger, K. E. Directed evolution of an enantioselective enzyme through combinatorial multiple-cassette mutagenesis. Angew. Chem. Int. Ed. Engl., 2001, 40, 3589-3591. Reetz, M. T.; Wang, L. W.; Bocola, M. Directed evolution of enantioselective enzymes: iterative cycles of CASTing for probing protein-sequence space. Angew. Chem. Int. Ed. Engl., 2006, 45, 1236-1241. May, O.; Nguyen, P. T.; Arnold, F. H. Inverting enantioselectivity by directed evolution of hydantoinase for improved production of L-methionine. Nat. Biotechnol., 2000, 18, 317-320. Zha, D. X.; Wilensek, S.; Hermes, M.; Jaeger, K. E.; Reetz, M. T. Complete reversal of enantioselectivity of an enzyme-catalyzed reaction by directed evolution. Chem. Commun., 2001, 2664-2665. Boersma, Y. L.; Pijning, T.; Bosma, M. S.; van der Sloot, A. M.; Godinho, L. F.; Droge, M. J.; Winter, R. T.; van Pouderoyen, G.; Dijkstra, B. W.; Quax, W. J. Loop grafting of Bacillus subtilis lipase A: inversion of enantioselectivity. Chem. Biol., 2008, 15, 782-789. Lentz, O.; Feenstra, A.; Habicher, T.; Hauer, B.; Schmid, R. D.; Urlacher, V. B. Altering the regioselectivity of cytochrome P450 CYP102A3 of Bacillus subtilis by using a new versatile assay system. Chembiochem, 2006, 7, 345350. Trefzer, A.; Jungmann, V.; Molnar, I.; Botejue, A.; Buckel, D.; Frey, G.; Hill, D. S.; Jorg, M.; Ligon, J. M.; Mason, D.; Moore, D.; Pachlatko, J. P.; Richardson, T. H.; Spangenberg, P.; Wall, M. A.; Zirkle, R.; Stege, J. T. Biocatalytic conversion of avermectin to 4 ''-oxo-avermectin: Improvement of cytochrome P450 monooxygenase specificity by directed evolution. Appl. Environ. Microbiol., 2007, 73, 4317-4325.

1882 Current Organic Chemistry, 2010, Vol. 14, No. 17 [61]

[62]

[63]

[64]

[65] [66]

[67] [68]

[69]

[70] [71]

[72]

[73]

[74]

[75]

[76]

[77]

Nair et al.

Kumar, S.; Scott, E. E.; Liu, H.; Halpert, J. R. A rational approach to reengineer cytochrome P4502B1 regioselectivity based on the crystal structure of cytochrome P4502C5. J. Biol. Chem., 2003, 278, 17178-17184. Pollmann, K.; Wray, V.; Hecht, H. J.; Pieper, D. H. Rational engineering of the regioselectivity of TecA tetrachlorobenzene dioxygenase for the transformation of chlorinated toluenes. Microbiology, 2003, 149, 903-913. Jennewein, S.; Schurmann, M.; Wolberg, M.; Hilker, I.; Luiten, R.; Wubbolts, M.; Mink, D. Directed evolution of an industrial biocatalyst: 2-deoxyD-ribose 5-phosphate aldolase. Biotechnol. J., 2006, 1, 537-548. Bolt, A.; Berry, A.; Nelson, A. Directed evolution of aldolases for exploitation in synthetic organic chemistry. Arch. Biochem. Biophys., 2008, 474, 318-330. Schetter, B.; Mahrwald, R. Modern aldol methods for the total synthesis of polyketides. Angew. Chem. Int. Ed. Engl., 2006, 45, 7506-7525. Demir, A. S.; Ayhan, P.; Sopaci, S. B. Thiamine pyrophosphate dependent enzyme catalyzed reactions: Stereoselective C-C bond formations in water. Clean-Soil Air Water, 2007, 35, 406-412. Pricer, W. E., Jr.; Horecker, B. L. Deoxyribose aldolase from Lactobacillus plantarum. J. Biol. Chem., 1960, 235, 1292-1298. DeSantis, G.; Liu, J. J.; Clark, D. P.; Heine, A.; Wilson, I. A.; Wong, C. H. Structure-based mutagenesis approaches toward expanding the substrate specificity of D-2-deoxyribose-5-phosphate aldolase. Bioorg. Med. Chem., 2003, 11, 43-52. Ran, N. Q.; Frost, J. W. Directed evolution of 2-keto-3-deoxy-6phosphogalactonate aldolase to replace 3-deoxy-D-arabino-heptulosonic acid 7-phosphate synthase. J. Am. Chem. Soc., 2007, 129, 6130-6139. Ran, N.; Draths, K. M.; Frost, J. W. Creation of a shikimate pathway variant. J. Am. Chem. Soc., 2004, 126, 6856-6857. Knop, D. R.; Draths, K. M.; Chandran, S. S.; Barker, J. L.; von Daeniken, R.; Weber, W.; Frost, J. W. Hydroaromatic equilibration during biosynthesis of shikimic acid. J. Am. Chem. Soc., 2001, 123, 10173-10182. Fong, S.; Machajewski, T. D.; Mak, C. C.; Wong, C. H. Directed evolution of D-2-keto-3-deoxy-6-phosphogluconate aldolase to new variants for the efficient synthesis of D- and L-sugars. Chem. Biol., 2000, 7, 873-883. Wada, M.; Hsu, C. C.; Franke, D.; Mitchell, M.; Heine, A.; Wilson, I.; Wong, C. H. Directed evolution of N-acetylneuraminic acid aldolase to catalyze enantiomeric aldol reactions. Bioorg. Med. Chem., 2003, 11, 2091-2098. Williams, G. J.; Woodhall, T.; Farnsworth, L. M.; Nelson, A.; Berry, A. Creation of a pair of stereochemically complementary biocatalysts. J. Am. Chem. Soc., 2006, 128, 16238-16247. Lingen, B.; Kolter-Jung, D.; Dunkelmann, P.; Feldmann, R.; Grotzinger, J.; Pohl, M.; Muller, M. Alteration of the substrate specificity of benzoylformate decarboxylase from Pseudomonas putida by directed evolution. Chembiochem, 2003, 4, 721-726. Osanjo, G.; Dion, M.; Drone, J.; Solleux, C.; Tran, V.; Rabiller, C.; Tellier, C. Directed evolution of the alpha-L-fucosidase from Thermotoga maritima into an alpha-L-transfucosidase. Biochemistry, 2007, 46, 1022-1033. Mackenzie, L. F.; Wang, Q.; Warren, R. A. J.; Withers, S. G. Glycosynthases: mutant glycosidases for oligosaccharide synthesis. J. Am. Chem. Soc., 1998, 120, 5583-5584.

Received: 12 June, 2009

[78] [79]

[80]

[81]

[82]

[83] [84]

[85]

[86] [87]

[88]

[89] [90] [91] [92]

[93]

[94] [95]

Lin, H.; Tao, H.; Cornish, V. W. Directed evolution of a glycosynthase via chemical complementation. J. Am. Chem. Soc., 2004, 126, 15051-15059. Williams, G. J.; Gantt, R. W.; Thorson, J. S. The impact of enzyme engineering upon natural product glycodiversification. Curr. Opin. Chem. Biol., 2008, 12, 556-564. Park, S. H.; Park, H. Y.; Sohng, J. K.; Lee, H. C.; Liou, K.; Yoon, Y. J.; Kim, B. G. Expanding substrate specificity of GT-B fold glycosyltransferase via domain swapping and high-throughput screening. Biotechnol. Bioeng., 2009, 102, 988-994. Marsden, A. F. A.; Wilkinson, B.; Cortes, J.; Dunster, N. J.; Staunton, J.; Leadlay, P. F. Engineering broader specificity into an antibiotic-producing polyketide synthase. Science, 1998, 279, 199-202. Patel, K.; Piagentini, M.; Rascher, A.; Tian, Z. Q.; Buchanan, G. O.; Regentin, R.; Hu, Z.; Hutchinson, C. R.; McDaniel, R. Engineered biosynthesis of geldanamycin analogs for Hsp90 inhibition. Chem. Biol., 2004, 11, 16251633. Xie, D.; Shao, Z.; Achkar, J.; Zha, W.; Frost, J. W.; Zhao, H. Microbial synthesis of triacetic acid lactone. Biotechnol. Bioeng., 2006, 93, 727-736. Abe, I.; Oguro, S.; Utsumi, Y.; Sano, Y.; Noguchi, H. Engineered biosynthesis of plant polyketides: chain length control in an octaketide-producing plant type III polyketide synthase. J. Am. Chem. Soc., 2005, 127, 12709-12716. Abe, I.; Morita, H.; Oguro, S.; Noma, H.; Wanibuchi, K.; Kawahara, N.; Goda, Y.; Noguchi, H.; Kohno, T. Structure-based engineering of a plant type III polyketide synthase: formation of an unnatural nonaketide naphthopyrone. J. Am. Chem. Soc., 2007, 129, 5976-5980. Arnold, F. H.; Georgiou, G., Directed Evolution Library Creation: Methods and Protocols. 2003, Humana Press Inc. Noren, C. J.; Anthonycahill, S. J.; Griffith, M. C.; Schultz, P. G. A general method for site-specific incorporation of unnatural amino acids into proteins. Science, 1989, 244, 182-188. Reetz, M. T.; Peyralans, J. J.; Maichele, A.; Fu, Y.; Maywald, M. Directed evolution of hybrid enzymes: Evolving enantioselectivity of an achiral Rhcomplex anchored to a protein. Chem. Commun. (Camb), 2006, 4318-4320. Thomas, C. M.; Ward, T. R. Artificial metalloenzymes: proteins as hosts for enantioselective catalysis. Chem. Soc. Rev., 2005, 34, 337-346. Lu, Y.; Garner, D. K.; Zhang, J.-L. Artificial metalloproteins: design and engineering. Wiley Encyclopedia of Chemical Biology, 2008. Keinan, E. Catalytic Antibodies. 1st ed: Wiley-VCH, NY, 2004. Jiang, L.; Althoff, E. A.; Clemente, F. R.; Doyle, L.; Rothlisberger, D.; Zanghellini, A.; Gallaher, J. L.; Betker, J. L.; Tanaka, F.; Barbas, C. F.; Hilvert, D.; Houk, K. N.; Stoddard, B. L.; Baker, D. De novo computational design of retro-aldol enzymes. Science, 2008, 319, 1387-1391. Rothlisberger, D.; Khersonsky, O.; Wollacott, A. M.; Jiang, L.; DeChancie, J.; Betker, J.; Gallaher, J. L.; Althoff, E. A.; Zanghellini, A.; Dym, O.; Albeck, S.; Houk, K. N.; Tawfik, D. S.; Baker, D. Kemp elimination catalysts by computational enzyme design. Nature, 2008, 453, 190-195. Guengerich, F. P.; Macdonald, T. L. Chemical mechanisms of catalysis by cytochromes P450 - a unified view. Accounts Chem. Res., 1984, 17, 9-16. Guengerich, F. P. Common and uncommon cytochrome P450 reactions related to metabolism and chemical toxicity. Chem Res. Toxicol., 2001, 14, 611-650.

Revised: 20 October, 2009

Accepted: 05 November, 2009