Converting human carbonic anhydrase II into a cocaine bensyl ester hydrolase through rational redesign

Enzymes capable of benzoate ester hydrolysis have several potential medical and industrial applications. A variant of human carbonic anhydrase II (HCAII) was constructed, by rational design, that is capable of hydrolysing para-nitrophenyl benzoate (pNPBenzo) with an efficiency comparable to some naturally occuring esterases. The design was based on a previously developed strategy, [G. Höst, L.G. Mårtensson, B.H. Jonsson, Redesign of human carbonic anhydrase II for increased esterase activity and specificity towards esters with long acyl chains, Biochim. Biophys. Acta 1764 (2006) 1601–1606.] in which docking of a transition state analogue (TSA) to the active site of HCAII was used to predict mutations that would allow the reaction. A triple mutant, V121A/V143A/T200A, was thus constructed and shown to hydrolyze pNPBenzo with kcat/KM = 625 (± 38) Ms. It is highly active with other ester substrates as well, and hydrolyzes para-nitrophenyl acetate with kcat/KM = 101700 (± 4800) Ms, which is the highest esterase efficiency so far for any CA variant. A parent mutant (V121A/V143A) has measurable KM values for para-nitrophenyl butyrate (pNPB) and valerate (pNPV), but for V121A/V143A/T200A no KM could be determined, showing that the additional T200A mutation has caused a decreased substrate binding. However, kcat/KM is higher with both substrates for the triple mutant, indicating that binding energy has been diverted from substrate binding to transition state stabilization. Introduction The bulky nature of benzoate esters makes them demanding as substrates for hydrolases. Esterases have been evaluated for their potential for benzoate ester hydrolysis, with possible applications such as mild and selective removal of protecting groups,[2] and synthesis of benzoylated compounds.[3] Another intriguing possibility is to use esterases for treatment of cocaine overdoses. By using an enzyme that hydrolyzes a benzoate ester bond in the cocaine molecule, the level of biologically active drug substance can be lowered rapidly.[4] Notably, several commonly used lipases lack any appreciable activity with model benzoate ester substrates, such as paranitrophenyl benzoate,[5] and many groups perform work on the identification and development of lipase variants with activity against bulky substrates.[5-7] In this report, we expand the range of enzymes able to hydrolyze benzoate esters by adding a human carbonic anhydrase II (HCAII) variant. The physiological role of the zinc enzyme HCAII (CA; carbonate hydro-lyase, EC 4.2.1.1) is to catalyze the reversible hydration of carbon dioxide: CO2 + H2O ↔ HCO3 + H. The details of the catalytic mechanism are described elsewhere.[8] Briefly, a OH bound to the active site Zn acts as a nucleophile, attacking the carbon dioxide substrate which results in a metal bound HCO3. The product is replaced by a H2O molecule, and the enzyme is regenerated when a H is transferred to the bulk solution via a histidine residue. The HCAII active site is catalytically promiscuous, for example being able to catalyze the hydrolysis of certain ester substrates.[9-12] We have previously showed that computational docking of transition state analogues (TSAs) can be used as a tool for designing HCAII mutants, allowing prediction of differences in esterase activity and specificity for variants of HCAII.[1] In this study, we extend these results to rationally design a HCAII variant for hydrolysis of a model benzoate ester. We also characterize the substrate specificity of the mutant for ester substrates of various sizes. Experimental section Mutagenesis The mutant HCAII variants V121A/V143A and V121A/V143A/T200A were constructed using the QuikChange Site-Directed Mutagenesis Kit (Stratagene, USA). Primers were purchased from DNA technology A/S (Denmark). Mutant plasmids were purified and the sequences of the entire coding region were determined to verify correct sequences. The sequencing was performed by GATC (Germany). For reasons discussed elsewhere, [1]we used the plasmid pACA,[13] containing the HCAII gene with the mutation C206S,[14] as template. HCAII C206S is referred to as pseudo wild type HCAII (HCAIIpwt). The stability, the CO2 hydration activity and the esterase activity (hydrolysis of pNPA) of HCAIIpwt are virtually identical to that of wild type HCAII.[14, 15] Although the mutant variants also contain the C206S mutation, they will be referred to as V121A/V143A and V121A/V143A/T200A, respectively. Preparation of enzyme variants Enzyme variants for the kinetic measurements were produced as described previously.[1] The procedure includes IPTG induced expression of mutant enzyme in E. coli BL21(DE3). After cell lysis, the enzymes were purified from centrifuged lysate essentially according to the affinity chromatographic procedure used by Khalifah et al.[16] The purity of the enzyme preparations were checked with SDS-PAGE and the concentrations were determined by measuring the absorbance at 280 nm, using a molecular extinction coefficient of 55400 Mcm.[17] Measurements of nitrophenyl ester hydrolysis Nitrophenyl ester hydrolysis reactions were monitored by measuring the absorbance at the isobestic point of the chromogenic leaving group (i.e. 348 nm for paranitrophenol). Reactions were performed at 25 ̊C in a quartz cuvette with 1 cm lightpath. The substrates were dissolved in acetone, giving a final acetone concentration of 5 %. All measurements were done at pH 8.5 using a 50 mM Tris-SO4 buffer, with the ionic strength kept at 0.1 by adding Na2SO4. Enzyme concentrations were between 0.018 20 μM for HCAIIpwt, V121A/V143A and V121A/V143A/T200A, and substrate concentrations were between 0.03-0.30 mM (the choice of concentrations was restricted by the low solubility of the ester substrates). The apparent second order rate constants, kenz (= kcat/KM when [S]<<KM), were calculated from the initial slopes, that had been corrected for background hydrolysis, using the equation v = kenz·[E]0·[S]0.[18] To investigate if the [S]<<KM assumption is valid for V121A/V143A/T200A, the activities with pNPB, pNPV and pNPC were measured for at least two different substrate concentrations (between 0.1 and 0.2 mM for pNPB and pNPV and between 0.03 and 0.1 mM for pNPC). For all substrates, control experiments were performed in which 25 μM acetazolamide was added to the enzyme-containing samples. The Δε348 values used were 5.15 mMcm (pNPA), 5.26 mMcm (pNPP), 5.26 mMcm (pNPB) and 5.23 mMcm (pNPC).[12] For pNPV and pNPBenzo, Δε348 values of 5.26 mMcm and 5.15 mMcm were used, respectively. Automated docking of transition state analogues to carbonic anhydrase variants Automated docking of transition state analogues (TSA) to the active site of HCAII variants was performed as described previously.[1] The ligand molecules for the docking calculations were designed with the ester bond replaced by a phosphonate group (see Fig. 1), a transition state analogue that is frequently used experimentally to mimic the transition state for the ester hydrolysis reaction.[19] Docking calculations were performed using Autodock 3.0,[20-22] and the TSAs were built using WebLabViewer Pro (Molecular simulations). Mutants were constructed by deleting methyl groups from the wild type HCAII pdb file (pdb accession code 2cba),[23] and all water molecules were deleted except four, which have been found to be important for the docking of ligands to HCAII.[24] The zinc bound hydroxide was not included in the docking, since the oxygen atom of the Zn-OH corresponds to one of the oxygen atoms in the TSA phosphonate group. Zn parameters used were: r = 1.1 Å, ε = 0.25 kcal/mol, q = +2.0 e. For each combination of enzyme and ligand, 100 runs of a Lamarckian genetic algorithm were performed, using a grid consisting of 60*60*60 points (with 0.375 Å between grid points) centred on the macromolecule. Population size was 100 and a maximum number of energy evaluations of 2000000 was used. Default values were used for the other parameters. Initial positions for the ligands were randomized, and all flexible bonds in the ligands were set as active. From each run, the conformation and position of the best docked molecule were retrieved, resulting in 100 docked molecules for each combination of ligand and enzyme variant. Fig. 1. Structures of the ester substrates used for kinetic measurements and a transition state analogue for the hydrolysis of para nitrophenyl benzoate (pNPBenzoTSA). The substrates are paranitrophenyl acetate (pNPA), para-nitrophenyl propionate (pNPP), para-nitrophenyl butyrate (pNPB), para-nitrophenyl valerate (pNPV), para-nitrophenyl caproate (pNPC) and paranitrophenyl benzoate (pNPBenzo).