Synergistic effect of mutagenesis and truncation to improve a polyesterase from Clostridium botulinum for polyester hydrolysis

The activity of the esterase (Cbotu_EstA) from Clostridium botulinum on the polyester poly(ethylene terephthalate) (PET) was improved by concomitant engineering of two different domains. On the one hand, the zinc-binding domain present in Cbotu_EstA was subjected to site-directed mutagenesis. On the other hand, a specific domain consisting of 71 amino acids at the N-terminus of the enzyme was deleted. Interestingly, a combination of substitution of residues present in the zinc-binding domain (e.g. S199A) synergistically increased the activity of the enzyme on PET seven fold when combined to the truncation of 71 amino acids at the N-terminus of the enzyme only. Overall, when compared to the native enzyme, the combination of truncation and substitutions in the zinc-binding domain lead to a 50-fold activity improvement. Moreover, analysis of the kinetic parameters of the Cbotu_EstA variants indicated a clear shift of activity from water soluble (i.e. para-nitrophenyl butyrate) to insoluble polymeric substrates. These results evidently show that the interaction with non-natural polymeric substrates provides targets for enzyme engineering.

present in members of the I.5 lipase family, both enzymes, Cbotu_EstA and PfL1, contain a zinc-binding domain, previously found only in lipases from thermophilic organisms. A major difference between these two enzymes is the presence of an extra domain at the N-terminus of Cbotu_EstA, which is neither present in PfL1 nor in other members of the I.5 lipase family 28 . The building of the model lacking 71 amino acids from the N-terminus of Cbotu_EstA, showed the presence of a hydrophobic patch and higher dynamics of the lid for interfacial activation 28 . Interestingly, this truncated enzyme (del71Cbotu_EstA) was able to hydrolyze PET, in contrast to the wild-type enzyme Cbotu_EstA. On the other hand, mutations in the zinc-binding domain likewise led to higher activity on aromatic polyesters. Therefore, it was exciting to investigate whether the combination of the two modifications would synergistically improve the activity of this enzyme.

Results and Discussion
Engineering of the Zn-binding and N-terminal domains of Cbotu_EstA. Based on the analysis of the 3D structure of Cbotu_EstA ( Fig. 1), which was previously solved at a resolution of 1.2 Å, (PDB-Code 5AH1) 18 , we hypothesized that concomitant modification of the Zn-binding and N-terminal domain could improve the activity on the synthetic polyester PET. Individual modification in these regions has previously been shown to increase the activity of Cbotu_EstA on synthetic polyesters 28,29 . Substitution of amino acids present in the so-called entrance to the zinc ion cavity has an influence on the lid opening during interfacial activation in members of the I.5 lipase family such as the lipase from Geobacillus stearothermophilus (BTL2) 23 . On the other hand, Cbotu_EstA contains 71 amino acids at the N-terminus which covers the lid structure of the enzyme. Removal of this extra domain which is not present in other members of the I.5 lipase family would expose a hydrophobic patch which could improve the adsorption of the enzyme on hydrophobic surfaces and generally improve dynamics. Indeed, truncation of the N-terminal extra domain was shown to improve hydrolysis of polyester 28 . In this paper, these two approaches were combined to evaluate a potential synergistic effect on PET hydrolysis.

Characterization of variants
Plasmids of Cbotu_EstA variants without extra N-terminal domain and substitutions of residues present in the zinc-binding were transformed into E. coli BL21-Gold(DE3) and expressed at 20 °C with 0.05 mM IPTG to prevent the formation of inclusion of protein aggregation and precipitation 30,31 . SDS-PAGE analysis indicated a purity higher than 90% for all variants of Cbotu_EstA (Fig. 2).
Kinetic parameters of the different variants were calculated (Figs 3 and S1). K M values of the Cbotu_EstA variants were in the range between 0.60 and 2.48 mM ( Table 1). The del71Cbotu_EstA variants showed similar values, while the variants containing the N-terminal domain showed the highest affinity towards p-NPB.
Analysis of further kinetic parameters (V max , k cat and k cat /K M ) of del71Cbotu_EstA_S127A, del71Cbotu_EstA_ W129A, del71Cbotu_EstA_S199A and del71Cbotu_EstA_W274H did not indicate increase in reaction rate, turnover number and kinetic efficiency towards the soluble substrate p-NPB, as contrarily shown for the other del71Cbotu_EstA variants.
The thermal stability of all del71Cbotu_EstA variants decreased when compared to the wild-type enzyme Cbotu_EstA and to the N-terminal domain zinc-binding domain variants. Those variants with substitution of Trp residue with either Ala or His residues (del71Cbotu_EstA_W129A and del71Cbotu_EstA_W274H) showed a higher thermal stability than other variants probably due to a lower hydrophobicity of the side-chain of the substituted amino acids improving solubility and the solvation by water molecules. After 48 h of incubation at 50 °C, only the two variants del71Cbotu_EstA_W129A and del71Cbotu_EstA_W274H showed residual activity of 7.2 and 16%, respectively, while the other variants completely lost their activity, including del71Cbotu_EstA (Fig. 4).
Shift of Cbotu_EstA specificity from soluble to bulky substrates. Elevated temperatures (50 °C) allow polymer chains present at the surface to be more motile, especially in amorphous regions, and hence    improve enzymatic hydrolysis. Activity of the Cbotu_EstA variants on the bulky substrate PET at this temperature was quantified based on the release of terephthalic acid (Ta) and mono-(2-hydroxyethyl) terephthalate (MHET) (Fig. 5a). Cbotu_EstA showed no significant activity on PET most likely due to the low dynamics of the lid opening and the low surface hydrophobicity of the enzyme surface. In contrast, the truncated version of the enzyme, del71Cbotu_EstA, released some Ta and MHET from PET. Substitutions of the zinc-binding domain residues on the other hand did not improve PET hydrolysis when compared to the wild-type enzyme despite the fact that such mutations had previously been described to increase activity on other polyesters 29 (Fig. 5b). However, combination of both modification in variants del71Cbotu_EstA_W129A and del71Cbotu_EstA_ S199A synergistically increased activity on PET by approximately four and seven times, respectively, when compared to del71Cbotu_EstA. Obviously, truncation of the extra domain at the N-terminus is necessary to allow the bulky PET to get in contact with the active site. Hence, only upon truncation, the mutations in the zinc-binding domain had an effect on PET hydrolysis. Both improved dynamics and exposure of a hydrophobic patch upon truncation of the extra domain at the N-terminus of the enzyme could improve interaction with the hydrophobic bulky substrate PET. Compared to the native enzyme, the combination of truncation and mutation in the zinc-binding domain increased the activity by a factor of 50. Interestingly, these two variants, del71Cbotu_EstA_ W129A and del71Cbotu_EstA_S199A did not show significantly improved catalytic efficiency on the small substrate p-NPB while the activity on bulky PET was dramatically improved.
Furthermore, del71Cbotu_EstA_W274H showed a higher release of Ta as compared to both del71Cbotu_ EstA and del71Cbotu_EstA_W129A, but a lack of MHET released (Fig. 5b). The successful tuning of substrate specificity from small water-soluble to large water insoluble substrates is clearly evident for del71Cbotu_EstA_ W274H. This variant showed the lowest catalytic efficiency on p-NPB but a reasonable activity on PET when compared to the native enzyme.

Conclusions
A novel strategy to tune the activity of the esterase Cbotu_EstA from small water soluble esters to bulky water insolubly polyesters was proposed. The combination of the substitution of Trp-129, Ser-199 and Trp-274 in the zinc-binding domain and the truncation of the N-terminal extra domain led to a synergistic improvement of  the enzyme on PET when compared to variants carrying either of the individual modification. The discrepancy between the kinetics of the three variants, namely del71Cbotu_EstA_W129A, del71Cbotu_EstA_S199A and del71Cbotu_EstA_W274H, on the soluble substrate p-NPB and their activity on the bulky substrate PET clearly indicated a shift of enzyme specificity from the water-soluble substrate p-NPB to the water-insoluble polymeric substrate PET. Alhough thermal stability has been affected negatively especially in the case of del71Cbotu_EstA_ S199A variant, and in general for the truncated variants, the combination of the substitution of Trp-129 and Ser-199 and Trp-274 improving dynamics together with the truncation of the N-terminal domain enhancing adsorption led to up to seven fold improved activity of the enzyme on PET, which shows highly active variants, compared to the wild-type enzyme Cbotu_EstA. Detailed analysis of the 3D structure of enzymes considering not only the active site architecture seems to be the key to improve activity on non-natural polymeric substrates such as the man-made polyester PET. Future studies could focus on the application of the concept presented here for a wider range of different polyesters as well as engineered enzymes for improved thermal stability. This would increase the potential of polyesterases both for surface functionalization and in recycling processes.

Methods
Chemicals, Reagents and Polymers. All chemicals and reagents used in this work were of analytical grade. Buffer components, para-nitrophenyl butyrate (p-NPB), bovine serum albumin (BSA) and HPLC-grade methanol were purchased from Sigma-Aldrich (USA). Films of polyethylene terephthalate (PET), with a thickness of 0.05 mm were purchased from Goodfellow, UK and used as polymeric substrate.
General recombinant DNA techniques and site-directed mutagenesis. All DNA manipulations were performed by standard methods 32 . DpnI restriction enzyme was purchased from New England Biolabs (USA). PCR was performed in a peqSTAR 96 Universal Gradient Thermocycler (Peqlab Biotechnologie GmbH, Austria) using Pfu DNA polymerase (Promega, Germany) with dNTPs purchased from MBI Fermentas (Germany). The plasmids were prepared using Promega mini-and midiprep kits (USA). DNA and protein sequences were analyzed by CLC Main Workbench, version 7.0.3 (Qiagen, Netherlands). Site-directed mutagenesis (SDM) of the mutants of Cbotu_EstA carrying mutations for the residues present on the zinc-binding domain was accomplished using the QuickChange multi site-directed mutagenesis kit (Stratagene) with plasmids previously prepared for each specific substitution 29 and megaprimers for the deletion of the 71 amino acids at the N-terminus of each variant. The following megaprimers were designed so as to anneal on the vector (underlined) and on the desired gene sequence: 5′-TTG TTT AAC TTT AAG AAG GAG ATA TAC ATA GCA TTA TTG GTG GTA ACA ACT ATC CGA TTG-3′ (forward primer) and 5′-CAA TCG GAT AGT TGT TAC CAC CAA TAA TGC TAT GTA TAT CTC CTT CTT AAA GTT AAA CAA-3′ (reverse primer). The following PCR conditions were used: for the first step an initial denaturation was performed for 1 min at 95 °C, followed by 4 cycles for 50 s at 95 °C, with annealing for 30 s at 58 °C, and an extension for 9 min at 72 °C, aside from the final extension of 5 min. For the second step of PCR, 21 cycles instead of 4 were performed mixing the same amount of the two reactions together. In order to remove the template plasmid DNA, DpnI treatment was carried out at 37 °C for 2 h, using 10 U of the restriction enzyme. The products were chemically transformed into E. coli XL-10 competent cells. The plasmids prepared from the positive clones were confirmed by commercial Sanger sequencing (LGC Genomics, Germany) 33 . The correct plasmids were chemically transformed into E. coli BL21-Gold(DE3) for further expression.

Recombinant Expression, Protein Purification and Analysis. Freshly transformed E. coli
BL21-Gold(DE3) cells were incubated and induced for protein expression as previously reported 29 . Purification was performed by IMAC as previously reported 21 . Fractions were pooled and concentrated by Vivaspin-20 column with a molecular weight cut-off (MWCO) of 10,000 Da following the manufacturer's instructions (Sartorius AG, Germany). The buffer was exchanged with 0.1 M Tris-HCl pH 7 using PD-10 desalting columns (GE Healthcare, UK). The protein concentration of the concentrated enzyme solutions was measured using the Bio-Rad Protein Assay Kit (USA) following the Bradford protocol 34 with BSA as protein standard. Sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed following the manufacturer's instructions, using a 4-15% gel. Pre-stained protein Marker IV was purchased from Peqlab (Germany) and used as a molecular mass marker. Proteins were revealed by Coomassie Brilliant Blue method.
Standard Esterase Assay and Thermal Stability Measurements. The measurement of the activity of the wild-type enzyme and different variants was performed using p-NPB as substrate as previously described at 25 °C and pH 7 28 . Kinetic constants were determined in 100 mM potassium phosphate buffer pH 7 at 25 °C using substrate concentration in the range of 0.3 to 6.3 mM following the release of p-NP with absorbance at 405 nm (ε p-NP = 8.31 M −1 cm −1 ) in a TECAN plate reader Infinite M200 PRO (Switzerland). Simple weighted non-linear regression was used to calculate parameters of the Michaelis-Menten equation using SigmaPlot software, version 12.5, with the Michaelis-Menten model (Systat Software Inc., Germany). Differences in thermal stability of different variants were determined at 50 °C, which was determined to be the optimal temperature of the wild-type enzyme Cbotu_EstA 18 in 100 mM potassium phosphate buffer pH 7 for up to 24 h with 0.6 µM enzyme. Aliquots were withdrawn and tested with 100 µM p-NPB to measure the residual activity 35 . One unit of enzyme activity (U) was defined as the amount of enzyme releasing 1 µmol of p-NP per minute under the given experimental conditions. The specific activity was expressed as U mg −1 of protein.
Hydrolysis of polyethylene terephthalate films. Pieces of 0.5 × 1 cm of PET films (thickness 0.05 mm) were cut and washed as previously described 28 . Hydrolysis experiments were performed using 2 µM enzyme in 1 mL potassium phosphate buffer pH 7 with the addition of one piece of PET film. The hydrolysis experiment was conducted for 72 h at 50 °C and 100 rpm. The reaction was stopped as previously described 8  products terephthalic acid (Ta) and mono-(2-hydroxyethyl) terephthalate (MHET) present in the supernatant were determined by high performance liquid chromatography (HPLC) as previously described 29 . A calibration curve of the chemicals was used to calculate the amount of hydrolysis products. Data availability statement. All data generated or analyzed during this study are included in this published article.