Review

TRENDS in Biotechnology

Vol.25 No.5

Enzyme promiscuity: mechanism and applications

Karl Hult and Per Berglund
School of Biotechnology, Department of Biochemistry, Royal Institute of Technology (KTH), AlbaNova University Center, SE-106 91 Stockholm, Sweden

Introductory courses in biochemistry teach that enzymes are specic for their substrates and the reactions they catalyze. Enzymes diverging from this statement are sometimes called promiscuous. It has been suggested that relaxed substrate and reaction specicities can have an important role in enzyme evolution; however, enzyme promiscuity also has an applied aspect. Enzyme condition promiscuity has, for a long time, been used to run reactions under conditions of low water activity that favor ester synthesis instead of hydrolysis. Together with enzyme substrate promiscuity, it is exploited in numerous synthetic applications, from the laboratory to industrial scale. Furthermore, enzyme catalytic promiscuity, where enzymes catalyze accidental or induced new reactions, has begun to be recognized as a valuable research and synthesis tool. Exploiting enzyme catalytic promiscuity might lead to improvements in existing catalysts and provide novel synthesis pathways that are currently not available. Introduction A modern term in enzymology is promiscuity, but what does it mean? As with many modern words, it is used without a well-dened meaning or with different meanings, depending on the author. In this overview, we will dene a promiscuous enzyme as one that does things it is not expected to do, and we will focus on the enzyme promiscuity that has been exploited to expand the use of enzymes in the synthesis of compounds, whereas promiscuity connected to enzyme evolution will only be mentioned briey. Promiscuity comes in many different forms, and we would like to introduce the classication of promiscuity into condition promiscuity, substrate promiscuity and catalytic promiscuity (Box 1). The classication is sometimes not straightforward, and we will see that different types of promiscuity can be combined. One example illustrating this problem is amide formation catalyzed by a lipase: amines are, undoubtedly, promiscuous substrates for a lipase, but is this also an example of catalytic promiscuity and thereby a new reaction for a lipase? Some scientists think so, whereas others regard the two transition states as being too similar to be classied as different reactions. Even if the term promiscuity is new in enzymology, there are numerous old examples in the literature that live up to the denition. An example of promiscuity was
Corresponding author: Hult, K. (kalle@biotech.kth.se). Available online 26 March 2007.
www.sciencedirect.com

demonstrated in the reversal of hydrolytic enzyme activity shown for maltase 1898 [1] and for pancreatic lipase in 1900 [2]. Condition promiscuity was explored by Rona et al. in the early 1930s in a series of experiments using pig liver lipase and pancreatic lipase in organic solvents for resolution of chiral alcohols and esters [3]. A good example of both substrate and catalytic promiscuity is the use of pyruvate decarboxylase to form carboncarbon bonds; this was rst studied in 1921 and is, today, an industrial activity [4]. Enzyme promiscuity has become popular in terms of enzyme evolution, enzyme engineering and biocatalysis [57]. Some scientists regard the promiscuity of enzymes as the starting point for enzyme divergent evolution; thus, low promiscuous activity towards an unnatural substrate might, by a single point mutation, turn the enzyme into a much more procient catalyst, affording a survival benet to the organism. Following the rst benecial mutation, the enzyme then has time to evolve into an even better catalyst [8]. Furthermore, in the laboratory, a promiscuous function can be the starting point for the creation of a new enzyme activity [9]. The evolutionary aspect of promiscuity has recently been discussed in more detail by Tawk et al. they propose that a new promiscuous activity can evolve without negative trade-offs in the native activity, leading to a generalist enzyme that can later on become a specialist [10]. Understanding of substrate and transition state recognition in enzyme catalysis offers great opportunities in the development of new catalytic functions in the scaffold of old enzymes. In nature, enzymes appear to develop to a point at which they are good enough for their task, meaning that the substrate and catalytic specicities have reached a point that satises the needs of the cell. There are, however, important exceptions. Rubisco (ribulose-1,5-bisphosphate carboxylase/oxygenase) evolved under low oxygen pressure and is unable to discriminate between carbon dioxide and oxygen. The consequence is an oxygenase reaction that creates byproducts that the cell has to cope with. In our opinion, this is a case of natural promiscuity that the cell has to live with in the absence of a better alternative. Also, as mentioned above and pointed out by OBrien and Herschlag, promiscuity can be a part of the evolution of new enzyme activity [11]. Gerlt et al. has shown that some enzymes with similar reaction intermediates are able to catalyze each others reactions, sometimes with astonishingly high efciency [12,13]. Furthermore, these enzymes probably originated from the same gene.

0167-7799/$ see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibtech.2007.03.002

232

Review

TRENDS in Biotechnology Vol.25 No.5

Box 1. Types of enzyme promiscuity

 Enzyme condition promiscuity Shown by enzymes with catalytic activity in various reaction conditions different from their natural ones, such as anhydrous media, extreme temperature or pH.  Enzyme substrate promiscuity Shown by enzymes with relaxed or broad substrate specificity.  Enzyme catalytic promiscuity Shown by enzymes catalyzing distinctly different chemical transformations with different transition states. Enzyme catalytic promiscuity can be either: (i) accidental a side reaction catalyzed by the wild-type enzyme; (ii) induced a new reaction established by one or several mutations rerouting the reaction catalyzed by the wild-type enzyme.

Enzymes taken out of the cell can be exposed to reaction conditions and substrates not seen in vivo that will challenge their specicity and might force them to handle substrates and catalyze reactions they were not designed for. This is why there is a low barrier to exploit enzymes as catalysts for unnatural reactions in the laboratory and industry. This could be called accidental in vitro promiscuity [14]. We will now describe the various types of promiscuity, focusing on examples of interest in applied synthesis (Table 1). Condition promiscuity Both condition promiscuity and substrate promiscuity have been explored and used in enzyme-catalyzed synthesis. There are thousands of examples of these, such as reactions performed in organic solvents, in absence of solvents, at high temperature, or at extreme pH. The
Table 1. Examples of promiscuity
Example

exploitation of lipases using modied substrates in organic solvents is particularly pronounced; this combination of substrate promiscuity and condition promiscuity has lead to several industrial applications [15]. Solidgas bioreactors represent another interesting example of condition promiscuity. There is no liquid phase; instead, the enzyme is immobilized in a dry form on a solid support, and the substrates and products are present in the gas phase only. The system enables individual ne-tuning of the thermodynamic activities of substrates and water, and is of great theoretical interest. In addition, the high productivity of the system enables the lipase-catalyzed production of esters at a commercial scale [16]. Early important examples of engineering the condition promiscuity of enzymes come from the group of Frances Arnold at the beginning of the 1990s. They explored the stability and activity of the serine protease subtilisin E in the polar organic solvent DMF where water hydrolysis is suppressed and improved the performance of the enzyme for peptide synthesis with random mutagenesis. The initial enzyme was further improved with three rounds of directed evolution, leading to an enzyme with ten amino acid substitutions [17]. This enzyme showed a 256-fold higher activity than the wild type in 60% DMF. The important interactions contributing to protein stability in non-aqueous solvents have been summarized as design rules by Arnold [18]. Substrate promiscuity As mentioned above, substrate promiscuity has been frequently exploited in the past. Here, we will mention examples of enzymes with strict substrate specicity, which can be destroyed by simple point mutations, and enzymes with an intrinsic high substrate promiscuity, which can be

Type of promiscuity Condition Substrate x x x x x x x x x x

Refs Catalytic Accidental Induced [1] [2] [3] [4] [15] [17] [19] [20] [21] [24] [26] [2729] [31] [32] [3334] [35] [36] [37] [38] [39] [40] [45] [46] [4751]

Maltase: reversal of reaction (1898) Pancreatic lipase: reversal of reaction (1900) Pig liver lipase: reactions in solvent (1931) Pyruvate decarboxylate: used for synthesis of CC bonds (1921) Lipase: transacylation in gas phase Subtilisin E: solventtolerant mutants Alanine dehydrogenase: cofactor usage Phosphite dehydrogenase: cofactor usage Lipase: resolution of chiral acids Lipase: resolution of chiral alcohols Enolase super family: promiscuous activities as part of evolution Acylases catalyzed Markovnicov additions: CN bond formation Lipase catalyzed aldol addition: CC bond formation Serine hydrolases catalyzed Michael addition: CO, CS and CN bond formation Protease catalyzed conjugate additions: CN bond formation Lipase catalyzed conjugate additions: CN bond formation Lipase catalyzed Michael additions: CN and CS bond formation Lipase catalyzed Michael additions: CC bond formation Carbonic anhydrase as a peroxidase Aminopeptidase as an oxidase Aminopeptidase hydrolyzing phosphodiesters Pyridoxalphosphatedependent racemase catalyzed aminotransferase Pyridoxalphosphatedependent racemase catalyzed retroaldol reaction Glycosidase catalyzed synthase reactions
www.sciencedirect.com

x x x x x x x x

x x x x x x x

Review

TRENDS in Biotechnology

Vol.25 No.5

233

increased even more. These examples will show that even if enzymes intrinsically have high or low substrate specicity, they often can be engineered into a promiscuous substrate behavior that suits a specic application. Coenzyme promiscuity For the living cell it is important to keep the NADPH NADP+ pool in a reduced state, whereas the pool of NADH NAD+ must be oxidized. Consequently, dehydrogenases must be able to distinguish between the two cofactor pairs. This is an example of where the cell does not permit promiscuity, even if a few point mutations can induce it. Ashida et al. describe an interesting case in which one point mutation can destroy the NAD+ selectivity or even reverse it to favor NADP+ [19]. Alanine dehydrogenase from Shewanella sp. Ac10 shows a high preference for NAD+ compared with NADP+, with a ratio of specicity constants of 2500. The mutation Asp198Ala creates an enzyme with a decreased KM and an increased kcat for NADP+, leading to specicity for NADP+ of 14 compared with NAD+ altogether, a change of 35 000-times without a drop in total activity. A similar mutation, Asp198Gly, creates an enzyme almost without cofactor specicity. Other dehydrogenases are equally sensitive to simple point mutations. For example, the NAD+ specicity of phosphite dehydrogenase can be abolished by one point mutation and turned into NADP+ specicity by two [20]. Lipase substrate promiscuity Lipases are promiscuous with regard to substrates: they accept a wide range of esters and are extensively used in the resolution of racemic acids and alcohols, although they normally show low enantioselectivity towards chiral acids. The research group around Reetz showed that it was possible to evolve lipases with preference for either the R- or S-enantiomer. They started with the lipase from Pseudomonas aeruginosa, which has no enantioselectivity towards 2-methyldecanoic acid p-nitrophenol ester, and subjected its gene to random mutagenesis to obtain, rst, a lipase with an enantiomeric ratio of 51 favoring the Senantiomer and, later on, a new mutant showing an Rpreference with an enantiomeric ratio of 38 [21]. These experiments were performed without knowledge of the enzyme structure but were later followed up by a paper describing the structure and the effects of the mutations [22]. It has been determined that all lipases have the same preference for secondary alcohols, where the preferred enantiomers have the same geometrical conguration called the Kazlauskas rule [23]. Lipase B from Candida antarctica has a stereospecicity pocket in which the chiral alcohol binds during catalysis. Through a single point mutation, Magnusson et al. were able to change the enantioselectivity 8 106-fold and reverse it from R- to Sselectivity, mainly by increasing the activity towards the S-enantiomer [24]. Interestingly, the change was only because of an increase in the entropic term favoring the S-enantiomer. Although only a small change in the protein is needed to obtain a lipase with inverted enantioselectivity, no enzyme is known from natural sources.
www.sciencedirect.com

Catalytic promiscuity Enzyme catalytic promiscuity refers to the ability of an enzyme active site to catalyze more than one different chemical transformation. Kazlauskas proposes that transformations are different if the types of bonds cleaved and/or made are different, and if the transition states of the two reactions are different [25]. A detailed mechanistic insight can be used to predict promiscuous reactions in an existing enzyme active site. One challenge in current enzyme redesign is to turn a low promiscuous activity of an enzyme into the main activity or to use part of the existing mechanism and divert the reaction path to new products. We will discuss a few such examples in this section. Catalytic promiscuity and enzyme evolution The enzymes mandelate racemase, muconate lactonizing enzyme and enolase share a common partial reaction, namely the Mg2+-assisted a-proton abstraction and enolization of carboxylic acids (Figure 1). The enol intermediate subsequently reacts differently in each enzyme to produce their specic products [14]. The enzymes show a similar bidomain structure in which the active site has a conserved location. With these common features it is easy to accept the hypothesis that these enzymes have evolved by divergent evolution from a common progenitor. In some cases, the evolution has not only changed the substrate specicity but also the reaction specicity. Gerlt et al. found one member of the enolase super-family that had a promiscuous activity as a N-acylamino acid racemase, with a kcat/KM equal to 590 M1s1 for the best substrate, N-acetylmethionine, compared with 2 105 M1s1 for the natural reaction as o-succinylbenzoate synthase. The discovery of enzymes with dual activity indicates that at least some keep their original activity at the same time as new activities start to develop, resulting in an enzyme showing natural catalytic promiscuity [26]. Markovnikov additions catalyzed by a hydrolase accidental catalytic promiscuity The Markovnikov addition is an important reaction in synthetic chemistry. Wu et al. have demonstrated the promiscuous addition of nitrogen-containing heterocycles to vinyl esters, catalyzed by the enzymes D-aminoacylase [27], penicillin G acylase [28] and acylase from Aspergillus oryzae [29] in an organic solvent. Several different N-heterocyclic compounds were used, including purines and pyrimidines, and several reactions reached close to a quantitative yield. No product was observed when using vinyl ethers. This suggests that the carbonyl in the substrate is needed, and the authors propose that its oxygen is bound in an oxyanion hole or, alternatively, coordinated

Figure 1. The common Mg2+ assisted a-proton abstraction and enolization of carboxylic acids in the enolase super family, which can produce racemase, lactonizing enzyme and enolase activities.

234

Review

TRENDS in Biotechnology Vol.25 No.5

to the zinc ion in the enzyme active site, which, according to them, would polarize the vinyl double bond. The heterocyclic nucleophile is deprotonated by a general base and simultaneously attacks the partial positive charge of the C=C double bond. It is somewhat surprising to note that none or low enantioselectivity was observed in the Markovnikov additions. This is unexpected and hard to explain. Aldol and Michael additions as accidental and induced catalytic promiscuity in a lipase In the reaction mechanism of serine hydrolases, the oxyanion hole is an important feature: it polarizes the carbonyl double bond in the ester substrate and renders the carbonyl carbon susceptible for nucleophilic attack by the active-site serine (Figure 2a). If the serine is removed by mutation, the oxyanion hole is still able to activate carbonyl functions. This is the basis for several of the promiscuous reactions that are shown to take place in the active site of serine hydrolases. Carboncarbon bond formation is important in synthesis, and many biotransformation examples exist [30]. Branneby et al. have shown that lipase B from Candida antarctica catalyzes carboncarbon bond formation through aldol addition, with the substrates hexanal and propanal [31]. The proposed reaction mechanism involves binding of the carbonyl in the oxyanion hole of the lipase while the activesite histidine acts as a base to abstract a a-proton so that a carboncarbon bond can be formed with the second substrate (Figure 2b). The specic activity increased by a factor of four when the active-site serine was replaced with an alanine to avoid the nucleophilic attack present in the natural reaction. Wild-type lipase covalently inhibited with methyl p-nitrophenyl n-hexylphosphonate was completely inactive, conrming that the active site is needed for catalysis of this reaction. The aldolase reaction was carried out in organic solvent, with unnatural substrates and catalyzed by a lipase; it is, therefore, an example that combines all three types of promiscuity: condition, substrate and catalytic. Conjugate addition on a,b-unsaturated carbonyl compounds is long known to be catalyzed by serine hydrolases [32]. Recently, conjugate additions by N-nucleophiles in organic solvents have been studied when catalyzed by a protease [33,34] or by a lipase [35]. Thiols were mainly used as nucleophiles by Carlqvist et al., who proposed that the mechanism involves the active-site histidine as a base that activates the nucleophile for attack at the b-carbon of

the substrate, with the carbonyl oxygen hydrogen of the substrate bonded in the oxyanion hole of the enzyme (Figure 2c) [36]. This was supported by quantum chemical calculations on a minimal model system for C. antarctica lipase B, based on docking studies and molecular dynamics simulations. Furthermore, in that study, the Ser105Ala mutant of C. antarctica lipase B was found to catalyze conjugate additions of thiols and amines on a,b-unsaturated aldehydes, ketones and methyl esters in cyclohexane, toluene or diethyl ether. The reactions followed saturation kinetics with kcat values in the range of 0.001 to 4.000 min1. The enzyme prociency was 107 for the Ser105Ala mutant, a gure approaching that of natural enzymes. The conjugate-addition activity was, in many cases, improved by the Ser105Ala mutation for the tested substrates. Svedendahl et al. has recently shown that Michael additions forming a carboncarbon bond between a 1,3dicarbonyl compound and an a/b-unsaturated aldehyde or ketone are efcient promiscuous reactions for Candida antarctica lipase B [37]. The reactions were run with various substrates in cyclohexane or in absence of solvent, with specic rates recorded in the range of 0.011000 s1. The wild type showed, in all cases, 100-fold lower specic rate than the mutant Ser105Ala. It is interesting to note that the wild type enzyme was able to catalyze all the reactions mentioned above, which could be a good example of accidental promiscuity found in vitro. It is, although, highly improbable that those reactions would occur in nature because it is unlikely that the lipase would encounter such substrates. The increased efciency of the active-site serine-decient mutant is an example of induced catalytic promiscuity. Induced catalytic promiscuity in articial metalloenzymes There are several interesting examples of articial metalloenzymes that can catalyze enantioselective oxidation. Okrasa and Kazlauskas made a thorough study of manganese-substituted carbonic-anhydrase-catalyzed oxidation of styrene analogues. They created the manganese enzyme by dialysis, where the natural Zn2+ was changed to Mn2+. The reaction was run in a buffer containing bicarbonate using hydrogen peroxide as oxidant. The reaction rate was lower for the manganese-substituted carbonic anhydrase than for free manganese or bicarbonate; however, the produced epoxide had enantiomeric

Figure 2. In serine hydrolases, the carbonyl of the substrate ester is coordinated in the oxyanion hole and attacked by the active site serine (a). Mutation of the active site serine opens up the possibility for other reaction types that depend on polarization of the carbonyl function, such as aldol addition (b) or Michael type additions (c). The oxyanion hole of Candida antarctica lipase B can form three hydrogen bonds to the oxygen of the substrate in the transition state, whereas, in ground state, only two would be developed.
www.sciencedirect.com

Review

TRENDS in Biotechnology

Vol.25 No.5

235

Figure 3. The first part of the reaction mechanism of pyridoxal-phosphate (PLP)-dependent enzymes. The mechanism then diverges, depending on the type of enzyme.

excesses of up to 67%, and no aldehyde product could be seen. Unfortunately, the conversion was low owing to degradation of the anhydrase by reactive intermediates generated during catalysis [38]. Catechol oxidase activity was introduced in an aminopetidase by da Silva and Ming. They prepared a di-Cu2+-substituted aminopeptidase that oxidized catechol with hydrogen peroxide, recording a kcat/KM in the range of 3300 M1s1, which is only ten times lower than that of a natural catechol oxidase [39]. Interestingly, the natural dizinc aminopeptidase produced a 4 1010-fold acceleration of phosphodiester hydrolysis in addition to its natural peptide hydrolytic activity [40]. Thus, this protein scaffold has been shown to perform reactions that pass through totally different transition states: hydrolysis of peptides and phosphodiesters with the natural di-zinc enzyme, and oxidation with the articial di-copper enzyme. All reactions are catalyzed with efciencies of the same magnitude as their natural counterparts. Induced catalytic promiscuity in pyridoxal-phosphatedependent enzymes Pyridoxal-phosphate (PLP) is a versatile cofactor used by more than 140 different enzymes from ve out of the six main enzyme EC classes [41]. These enzymes mainly accept a-amino acids and catalyze reactions such as aracemization, transamination, a- or b-decarboxylation, and retroaldol reactions. Their active site is similar, and the cofactor is covalently bound through an imine linkage

to a lysine residue. The common mechanistic feature is the formation of an external imine formed by the amino group of the substrate and the carbonyl of the cofactor [42,43]. Many of the PLP-dependent enzymes subsequently form a quinonoid intermediate, from which the mechanism diverges depending on the type of PLP-dependent enzyme (Figure 3). Mechanistic details suggested by Toney indicate the consequence of a fully developed quinonoid intermediate for the discrimination between racemization and transamination, and it is clear that small changes are needed in one enzyme to alter the activity into that of another PLP enzyme [44]. Examples of such induced catalytic promiscuity have, for instance, been published by Esaki et al. [45]. Seebeck and Hilvert describe another example of induced catalytic promiscuity of a PLP-dependent enzyme. They converted a racemase into an aldolase by a single point mutation. The mutation, Tyr265Ala, in a racemase from Geobacillus stearothermophilus resulted in a 3 103fold reduction in racemase activity and a 2 105-fold increase in retro-aldol activity with the substrate D-phenylserine, compared with the wild-type enzyme. Replacement of Tyr265 with Ala was expected to promote the nearby His166 to initiate a retro-aldol reaction of a bhydroxy-amino acid, similar to the retro-aldol reaction of the non-related enzyme L-threonine aldolase (Figure 4). Tyr265 is acting as a base in the active site and is necessary for converting L- to D-alanine. Removal of Tyr265 was, therefore, also expected to drastically lower the racemase

Figure 4. The retro-aldol reaction specificity of l-threonine aldolase was engineered into alanine racemase by a single-point active-site mutation [46]. Abbreviation: PLP, pyridoxal-phosphate.
www.sciencedirect.com

236

Review

TRENDS in Biotechnology Vol.25 No.5

activity. Interestingly, the retro-aldol reaction catalyzed by the alanine racemase mutant was found to be highly enantioselective because the conversion of racemic threob-phenylserine stopped at 50% conversion [46]. Induced catalytic promiscuity in glycoside synthesis In 1898, Hill prepared maltose from glucose by equilibrium-controlled synthesis using maltase as a catalyst [1]. Owing to the need for water to dissolve the substrate,

this methodology produces low yields. Kinetically controlled synthesis of glycosides requires activated substrates, and might reach yields of up to 40%. In 1998, Withers and colleagues reported how the yield could be increased considerably by introducing catalytic promiscuity to suppress hydrolysis [47]. This rst paper has led to an abundance of extremely useful glycosynthases [48]. Inverting glycosidases have two carboxylates in the active site: one acts as a nucleophile to form a

Figure 5. Induced catalytic promiscuity. Glycosidases were transformed into glycosynthases, thioglycoligases or thioglycosynthases by rerouting reaction intermediates achieved by deletion of catalytic amino acid side chains. The different mutations are shown in bold.
www.sciencedirect.com

Review

TRENDS in Biotechnology

Vol.25 No.5

237

glycosylenzyme intermediate, whereas the other one acts as an acid/base catalyst. Removing the nucleophilic carboxylate produces an enzyme that is unable to hydrolyze glycosidic bonds. However, the otherwise intact active site can bind a suitable activated a-glycosyl uoride and an acceptor sugar in such a way that the remaining acid/base carboxylate can catalyze the formation of a glycosidic bond. Because the mutant enzyme is devoid of hydrolytic activity, the product can accumulate without being hydrolyzed. A mutant in which the nucleophilic carboxylate was changed not to an alanine but to serine is especially good as glycosynthase (Figure 5) [49]. An alternative modication of the reaction mechanism is to cripple the ability of the enzyme to hydrolyze the glycosylenzyme intermediate by removing the acid/base carboxylic acid. The glycosyl enzyme can still be formed from substrates with good leaving groups, such as dinitrophenol or uoride; however, water and hydroxyl-containing sugars are too weak nucleophiles to react at an appreciable reaction rate without the acid/base catalytic group. Instead, thiol-containing sugars, which are much more reactive, have to be used. The thioglycosides produced are not hydrolyzed by the enzyme and can, therefore, be prepared in good yields. Withers et al. named this activity thioglycoligase (Figure 5) [50]. In the two examples above, either of the two carboxylates in the active site was removed, but what would happen if both were removed? The double mutant would not be able to form a glycosylenzyme intermediate and, in addition, there would not be an acid/base to facilitate glycoside formation. Again, Withers group conducted the experiments and found an enzyme that could form glycosidic bonds between activated sugars and thiosugars. Presumably, the protein scaffold provided a proximity effect by binding and orienting the substrates in a favorable position for reaction (Figure 5) [51]. Conclusions Enzyme promiscuity, a property shown by many enzymes, is nowadays exploited in applied enzymology. In this review we have discussed some examples of various types of enzyme promiscuity and some applications based on it. We propose that the term, enzyme promiscuity, can be classied into enzyme condition promiscuity, substrate promiscuity and catalytic promiscuity. The applied aspects of condition promiscuity and substrate promiscuity have been explored for a long time, resulting in many industrial applications. Enzyme catalytic promiscuity, however, has only recently been exploited for synthetic applications. Catalytic promiscuity can be further divided into accidental catalytic promiscuity and induced catalytic promiscuity. It is evident that enzyme catalytic promiscuity, and particularly induced enzyme catalytic promiscuity, offers great opportunities in the design and development of new catalytic functions in the scaffold of stable enzymes. This is particularly promising because metagenomics has increased the availability of enzymes from extremophiles, providing new, stable templates for this technology. Exploiting enzyme catalytic promiscuity might lead to new, efcient and stable catalysts with, as yet, unprecedented activity for reactions where no practically useful enzyme alternatives exist today.
www.sciencedirect.com

Acknowledgements
This work was supported by the Swedish Research Council.

References
1 Hill, A.C. (1898) Reversible zymohydrolysis. J. Chem. Soc. 73, 634658 2 Kastle, J.H. and Loevenhart, A.S. (1900) Concerning lipase, the fatslitting enzyme, and the reversibility of its action. Am. Chem. J. 24, 491525 3 Rona, P. et al. (1931) Studies on the enzymatic ester formation and ester hydrolysis. Biochem. Z. 247, 113145 ber ein Kohlenstoffketten 4 Neuberg, C. and Hirsch, J. (1921) U knu pfendes ferment (Carboligase). Biochem. Z. 115, 282310 5 Hult, K. and Berglund, P. (2003) Engineered enzymes for improved organic synthesis. Curr. Opin. Biotechnol. 14, 395400 6 Bornscheuer, U.T. and Kazlauskas, R.J. (2004) Catalytic promiscuity in biocatalysis: using old enzymes to form new bonds and follow new pathways. Angew. Chem. Int. Ed. 43, 60326040 7 Berglund, P. and Park, S. (2005) Strategies for altering enzyme reaction specicity for applied biocatalysis. Curr. Org. Chem. 9, 325 336 8 Schmidt, D.M.Z. et al. (2003) Evolutionary potential of (b/a)-barrels: functional promiscuity produced by single substitutions in the enolase superfamily. Biochemistry 42, 83878393 9 Aharoni, A. et al. (2005) The evolvability of promiscuous protein functions. Nature Genet. 37, 7376 10 Khersonsky, O. et al. (2006) Enzyme promiscuity: evolutionary and mechanistic aspects. Curr. Opin. Chem. Biol. 10, 498508 11 OBrien, P.J. and Herschlag, D. (1999) Catalytic promiscuity and the evolution of new enzymatic activities. Chem. Biol. 6, R91R105 12 Thoden, J.B. et al. (2004) Evolution of enzymatic activity in the enolase superfamily: structural studies of the promiscuous o-succinylbenzoate synthase from Amycolatopsis. Biochemistry 43, 57165727 13 Yew, W.S. et al. (2005) Evolution of enzymatic activity in the orotidine 50 monophosphate decarboxylase suprafamily: enhancing the promiscuous D-arabino-hex-3-ulose 6-phosphate synthase reaction catalyzed by 3keto-L-gulonate 6-phosphate decarboxylase. Biochemistry 44, 18071815 14 Gerlt, J. et al. (2005) Divergent evolution in the enolase superfamily: the interplay of mechanism and specicity. Arch. Biochem. Biophys. 433, 5970 15 Schmid, A. et al. (2001) Industrial biocatalysis today and tomorrow. Nature 409, 258268 16 Lamare, S. et al. (2004) Solid/gas bioreactors: powerful tools for fundamental research and efcient technology for industrial applications. Green Chem. 6, 445458 17 Chen, K. and Arnold, F.H. (1993) Tuning the activity of an enzyme for unusual environments: Sequential random mutagenesis of subtilisin E for catalysis in dimethylformamide. Proc. Natl. Acad. Sci. U. S. A. 90, 56185622 18 Arnold, F.H. (1990) Engineering enzymes for non-aqueous solvents. Trends Biotechnol. 8, 244249 19 Ashida, H. et al. (2004) Conversion of cofactor specicities of alanine dehydrogenases by site-directed mutagenesis. J. Mol. Cat. B: Enzym. 30, 173176 20 Woodyer, R. et al. (2003) Relaxing the nicotinamide cofactor specicity of phosphite dehydrogenase by rational design. Biochemistry 42, 1160411614 21 Reetz, M.T. (2002) Directed evolution of selective enzymes and hybrid catalysts. Tetrahedron 58, 65956602 22 Bocola, M. et al. (2004) Learning from directed evolution: theoretical investigations into cooperative mutations in lipase enantioselectivity. ChemBioChem 5, 214223 23 Kazlauskas, R.J. (1991) A rule to predict which enantiomer of a secondary alcohol reacts faster in reactions catalyzed by cholesterol esterase, lipase from Pseudomonas cepacia, and lipase from Candida rugosa. J. Org. Chem. 56, 26562665 24 Magnusson, A.O. et al. (2005) An S-selective lipase was created by rational redesign and the enantioselectivity increased with temperature. Angew. Chem. Int. Ed. 44, 45824585 25 Kazlauskas, R.J. (2005) Enhancing catalytic promiscuity for biocatalysis. Curr. Opin. Chem. Biol. 9, 195201 26 Glasner, M. et al. (2006) Evolution of structure and function in the osuccinylbenzoate syhthease/N-acylamino acid racemase family of the enolase superfamily. J. Mol. Biol. 360, 228250

238

Review

TRENDS in Biotechnology Vol.25 No.5

27 Wu, W-B. et al. (2005) D-Aminoacylase-catalyzed Markovnikov addition of azoles to vinyl esters in organic solvents. Synlett 2433 2436 28 Wu, W-B. et al. (2005) Penicillin G acylase-catalyzed Markovnikov addition of allopurinol to vinyl ester. Chem. Commun. 2348 2350 29 Wu, W-B. et al. (2006) Promiscuous acylases-catalyzed Markovnikov addition of N-heterocycles to vinyl esters in organic media. Adv. Synth. Catal. 348, 487492 30 Breuer, M. and Hauer, B. (2003) Carboncarbon coupling in biotransformation. Curr. Opin. Biotechnol. 14, 570576 31 Branneby, C. et al. (2003) Carboncarbon bonds by hydrolytic enzymes. J. Am. Chem. Soc. 125, 874875 32 Kitazume, T. et al. (1986) Synthesis of optically active triuorinated compounds: asymmetric Michael addition with hydrolytic enzymes. J. Chem. Soc. Chem. Commun. 13311333 33 Cai, Y. et al. (2004) Michael addition of imidazole with acrylates catalyzed by alkaline protease from Bacillus subtilis in organic media. Biotechnol. Lett. 26, 525528 34 Cai, Y. et al. (2004) Alkaline protease from Bacillus subtilis catalyzed Michael addition of pyrimidine derivatives to alpha,beta-ethylenic compounds in organic media. Synthesis 671674 35 Torre, O. et al. (2004) Lipase-catalyzed Michael addition of secondary amines to acrylonitrile. Chem. Commun. 17241725 36 Carlqvist, P. et al. (2005) Exploring the active-site of a rationally redesigned lipase for catalysis of Michael-type additions. ChemBioChem 6, 331336 37 Svedendahl, M. et al. (2005) Fast carboncarbon bond formation by a promiscuous lipase. J. Am. Chem. Soc. 127, 1798817989 38 Okrasa, K. and Kazlauskas, R.J. (2006) Manganese-substituted carbonic anhydrase as a new peroxidase. Chem. Eur. J. 12, 1587 1596

39 da Silva, G.F.Z. and Ming, L-J. (2005) Catechol oxidase activity of diCu2+-substituded aminopeptidase from Streptomyces griseus. J. Am. Chem. Soc. 127, 1638016381 40 Ercan, A. et al. (2006) A moonlighting dizinc aminopeptidase from Streptomyces griseus: mechanisms for peptide hydrolysis and the 4 1010-fold acceleration of the alternative phosphodiester hydrolysis. Biochemistry 45, 1377913793 41 Percudani, R. and Peracchi, A. (2003) A genomic overview of pyridoxalphosphate-dependent enzymes. EMBO Rep. 4, 850854 42 Dunathan, H.C. (1966) Conformation and reaction specicity in pyridoxal phosphate enzymes. Proc. Natl. Acad. Sci. U. S. A. 55, 712716 43 John, R.A. (1995) Pyridoxal phosphate-dependent enzymes. Biochim. Biophys. Acta 1248, 8196 44 Toney, M.D. (2005) Reaction specicity in pyridoxal phosphate enzymes. Arch. Biochem. Biophys. 433, 279287 45 Yow, G-Y. et al. (2003) Conversion of the catalytic specicity of alanine racemase to a D-amino acid aminotransferase activity by a double active-site mutation. J. Mol. Catal. B Enzym. 23, 311319 46 Seebeck, F.P. and Hilvert, D. (2003) Conversion of a PLP-dependent racemase into an aldolase by a single active site mutation. J. Am. Chem. Soc. 125, 1015810159 47 Mackenzie, L.F. et al. (1998) Glycosynthases: mutant glycosidases for oligosaccharide synthesis. J. Am. Chem. Soc. 120, 55835584 48 Perugino, G. et al. (2004) Oligosaccharide synthesis by glycosynthases. Trends Biotechnol. 22, 3137 49 Mayer, C. et al. (2000) The E358S mutant of Agrobacterium sp. bglucosidase is greatly improved glycosynthase. FEBS Lett. 466, 4044 50 Jahn, M. et al. (2003) Thioglycoligases: mutant glycosidases for thioglycoside synthesis. Angew. Chem. Int. Ed. 42, 352354 51 Jahn, M. et al. (2004) Thioglycosynthases: double mutant glycosidases that serve as scaffolds for thioglycoside synthesis. Chem. Commun. 274275

Articles of Interest
Articles of interest in other Trends and Current Opinion journals Mitochondrial transcription and its regulation in mammalian cells Jordi Asin-Cayuela and Claes M. Gustafsson Trends in Biochemical Sciences doi:10.1016/j.tibs.2007.01.003 Historical review: the field of protein methylation Woon Ki Paik, David C. Paik and Sangduk Kim Trends in Biochemical Sciences doi:10.1016/j.tibs.2007.01.006 Life in acid: pH homeostasis in acidophiles Craig Baker-Austin and Mark Dopson Trends in Microbiology doi:10.1016/j.tim.2007.02.005 Synthetic promoters: genetic control through cis engineering Mauritz Venter Trends in Plant Science doi:10.1016/j.tplants.2007.01.002 Fermentation technologies for the production of probiotics with high viability and functionality Christophe Lacroix and Selcuk Yildirim Current Opinion in Biotechnology doi:10.1016/j.copbio.2007.02.002
www.sciencedirect.com