Sequence‐Based Prediction of Promiscuous Acyltransferase Activity in Hydrolases

Abstract Certain hydrolases preferentially catalyze acyl transfer over hydrolysis in an aqueous environment. However, the molecular and structural reasons for this phenomenon are still unclear. Herein, we provide evidence that acyltransferase activity in esterases highly correlates with the hydrophobicity of the substrate‐binding pocket. A hydrophobicity scoring system developed in this work allows accurate prediction of promiscuous acyltransferase activity solely from the amino acid sequence of the cap domain. This concept was experimentally verified by systematic investigation of several homologous esterases, leading to the discovery of five novel promiscuous acyltransferases. We also developed a simple yet versatile colorimetric assay for rapid characterization of novel acyltransferases. This study demonstrates that promiscuous acyltransferase activity is not as rare as previously thought and provides access to a vast number of novel acyltransferases with diverse substrate specificity and potential applications.


Materials
Benzyl alcohol and 2-phenylethanol were purchased from FLUKA (Buchs, Switzerland) and vinyl acetate from ACROS ORGANICS (Geel, Belgium).All other chemicals and solvents were purchased from Sigma, VWR, or Carl Roth and were used without further purification.Synthetic genes were codon-optimized for expression in Escherichia coli and subcloned into pET-28a(+) by BioCat GmbH (Heidelberg, Germany).Est8 was obtained from BRAIN AG (Zwingenberg, Germany) and was expressed from pET-26 and PestE from pET-21, respectively.All plasmids encoded C-terminal His6-tags.

Expression and Purification of His6-tagged bHSLs
Chemically competent E. coli BL21(DE3) cells were transformed with expression vectors and plated on LB agar containing 50 µg/mL kanamycin.Pre-cultures (5 mL of LB containing 50 µg/mL kanamycin) were inoculated with single colonies and incubated overnight (37°C, 180 rpm).TB medium (200 mL containing 50 µg/mL kanamycin) was inoculated with 0.2% (v/v) of the pre-culture.The cultures were incubated (37°C, 180 rpm) until they reached an OD600 of 0.6.Protein expression was induced by addition of isopropyl-β-D-thiogalactoside to a final concentration of 0.4 mM, followed by incubation at 20°C (180 rpm) for 16 h.Cells were harvested by centrifugation at 4000 g and 4°C for 30 min and washed with 30 mL of 50 mM potassium phosphate (pH 7.4).Washed cell pellets were stored at -20°C for later use.
Cell pellets were resuspended in ~20 mL lysis buffer (50 mM potassium phosphate, 300 mM sodium chloride, pH 8.0) and lysed by sonification on ice (5 cycles of 1 min sonication (40% intensity, 50% pulsed cycle) followed by 1 min incubation on ice) using a SONOPULS HD 2070 (BANDELIN electronic GmbH & Co. KG, Berlin, Germany).Lysates were clarified by centrifugation (10000 g, 4°C, 30 min).The His6-tagged proteins were purified by immobilised metal-affinity chromatography using 3 ml of Roti ® garose-His/Ni Beads (Carl Roth, Karlsruhe, Germany).The Ni-NTA resin was washed with deionized water and equilibrated with lysis buffer.The lysates were applied by gravity and the flow-through discarded.Unspecifically bound proteins were removed by excessive washing of the resin with washing buffer (50 mM potassium phosphate, 300 mM sodium chloride, 20 mM imidazole, pH 8.0).The recombinant proteins were eluted with approximately 6 ml of elution buffer (50 mM potassium phosphate, 300 mM sodium chloride, 300 mM imidazole, pH 8.0).Elution fractions were stored at 4°C after being desalted using PD-10 desalting columns (GE Healthcare, UK) equilibrated with storage buffer (200 mM potassium phosphate, pH 8.0).For crystallization experiments, Est8 was further purified by gel filtration on a Superdex200 16/60 column (GE Healthcare, Freiburg, Germany) with storage buffer.Protein concentrations were determined by measuring absorbance at 280 nm using a NanoDrop TM (Thermo Fisher, Germany) device.Theoretical extinction coefficients and molecular weights were calculated from the protein sequences using ExPASy ProtParam. [1]If necessary, proteins were further concentrated using Vivaspin™ concentrators with 10 kDa cut-off (Sartorius, Germany).

SDS-PAGE Analysis
The purity of protein samples was analysed by SDS-PAGE.Samples of the purified proteins were denatured by heating (95°C, 10 min) in Laemmli-buffer [2] followed by centrifugation (20800 g, 5 min).The proteins were separated on 12.5% acrylamide gels at a constant voltage of 120 V.The gels were stained using Coomassie Brilliant Blue G-250.

Crystallization of Est8 and Data Collection
Est8 in phosphate buffer (200 mM, pH 8.0) was concentrated to 30 mg/mL and screened for crystallization conditions by mixing 0.3 µL protein with 0.3 µL well solution from JBScreen Classic Kits 1-10.The successful condition (25% PEG 4000, 0.1 M sodium citrate, pH 5.6, and 0.2 M ammonium sulfate) was refined in larger hanging drops (2 µL protein + 2 µL reservoir solution).A crystal from 29% PEG 4000, 0.1 M ammonium sulfate, and 0.1 M sodium citrate, pH 5.6, was cryo-cooled in liquid nitrogen and measured at BESSY beamline 14.2 at 100 K. Data were processed using XDSAPP. [3]

In Silico Prescreening
A sequence library containing more than 20,000 sequences was generated by a BLAST search, using the sequence of 3FAK as query.The sequence library was post-processed using R Studio.Sequences longer or shorter than 300 ± 10 residues were removed from the library.Redundant entries were removed, and the remaining 6,500 sequences were ranked by cap domain hydrophobicity.Hydrophobicity scores were calculated by summing the values from the hydrophobicity scale (reported by Abraham and Leo [8] ) for the 45 N-terminal residues.

In silico Analysis of the Substrate Binding Pockets
The substrate-binding sites in the crystal structures of 4XVC, 3K6K, 3FAK, Est8, 3ZWQ and 1EVQ were identified and analysed using the SiteMap tool [9] of the Schrödinger Maestro software suite (Schrödinger, New York City, USA).In order to guarantee comparability, all structures were processed with the same parameters.A minimum of 15 site points per reported site were set with a total number of 5 site-point groupings using the more restrictive definition of hydrophobicity and the standard grid.Maps were always cropped 4 Å from the nearest site point.In all cases, the actual binding pocket was represented as the top hit.The hydrophobic surface areas of the top-ranked binding site were calculated from that with the isovalue set to -0.4 for all structures.

Colorimetric Acyltransferase Assay
Activity assays were carried out at 25°C in transparent 96-well polystyrene plates.Changes in absorbance at 405 nm were measured using a Tecan Plate Reader (Tecan, Männedorf, SWITZERLAND).Reactions were started by addition of 100 µL of the enzyme solution (in 200 mM potassium phosphate, pH 7.0) to 100 µl of a 2x master mix containing all other components in 200 mM potassium phosphate (pH 7.0).The final reactions contained pNPA (1 mM from a 2 M stock in DMSO, 0.05% DMSO in the final reactions) and 2-phenylethanol or benzyl alcohol (0, 0.25, 0.5, 1.25, 2.5, 5, 10, 15, 20, 25, 37.5, or 50 mM).Reactions were measured in triplicate and corrected by subtraction of chemical background hydrolysis of pNPA in the buffer (background was measured for each concentration of alcohol used).The initial slope was determined as shown in Fig. S4 and the amount of pNPA formed was calculated from an external calibration curve of p-nitrophenol (in 200 mM potassium phosphate, pH 7.0) in the range of 0.0 to 1.0 mM (Fig. S6).The acyl transfer to hydrolysis rate ratios were then calculated by dividing the specific activities obtained for the reactions with varying amounts of alcohol (acyl transfer + hydrolysis) by that obtained for the reaction without alcohol (hydrolysis).The pNPA acyltransferase assay has the advantage of being compatible with virtually any alcohol as acyl acceptor substrate, compared to the alcohol dehydrogenase assay recently published by Mestrom et al. [15] We could not use this assay because the alcohol dehydrogenase precipitates in the presence of 2-phenylethanol.Figure S2: Specific activities for Est8 and homologs in the hydrolysis of vinyl acetate determined using the coupled spectrophotometric assay as described by Mestrom et al. [15] Reactions were performed in 200 mM phosphate buffer and vinyl acetate concentration was 200 mM.    Residues which are more hydrophobic relative to glycine (0) have a positive value, while residues more hydrophilic than glycine were assigned negative values.

Figure S1 :
Figure S1: A) Alignment of the sequences of Est8 and homologous bHSLs.The variable cap domain sequence is highlighted by a yellow rectangle.Residues highlighted in black are identical while grey shading indicates similarity.B) The crystal structure of Est8, showing the position of the cap domain (yellow) relative to the catalytic triad consisting of Ser 146 , Glu 240 and His 270 (purple carbons).Hydrophobic surface area within the substrate binding pocket is highlighted in blue.

Figure S3 :
FigureS3: A) Reaction scheme of the pNP-acyltransferase assay.An acetyl group is transferred from pNPA to the enzyme via nucleophilic attack of the catalytic serine on pPNA, leading to the release of p-nitrophenolate which can be monitored at 405 nm.The acyl-enzyme intermediate formed can be attacked by either a water molecule (hydrolase activity) or an organic nucleophile like an alcohol (acyltransferase activity).A good acyltransferase would preferentially utilize the organic nucleophile over water, resulting in accelerated release of pnitrophenolate in the presence of the alcohol compared to the reaction in its absence.B) GC was used to prove that the pNPA assay detects actual formation of benzyl acetate rather than nonspecific acceleration of pNPA hydrolysis.The increase in rate of 4-nitrophenolate release with increasing concentration of benzyl alcohol correlates well with the amount of benzyl acetate detected by GC.

Figure S4 :
Figure S4: MS measurement after separation via GC of the reaction shown in Fig. S3 to prove that benzyl acetate is formed as product of enzyme-catalysed transesterification in the pNPA-acyltransferase assay.The exact mass of benzyl alcohol is 150.18 g/mol.

Figure S5 :
Figure S5:Slope determination in the pNPA-assay.An example for the reaction of 3ZWQ in the presence of 5 mM benzyl alcohol is shown.The linear slopes were determined so that R 2 in each case is ≥0.99.

Figure S6 :
Figure S6: pNP calibration curve for the quantification of pNP by measuring the absorbance at 405 nm at different concentrations of pNP.

Figure S7 .
Figure S7.Changes in relative activity of bHSLs in the presence of different concentrations of 2-phenylethanol.

Figure
Figure S8.A) Reaction progression of the acetylation of 2-phenylethanol (20 mM) using a ten-fold excess of vinyl acetate (200 mM) as acyl donor in the presence of 1 mg/mL lyophilized crude lysate of WP007.Reactions were carried out in 200 mM potassium phosphate (pH 8.2).WP007 is able to almost fully convert 2-phenylethanol to 2-phenylethyl acetate.B) GC-chromatograms showing the conversion of the substrate to the product over time.

Figure S9 :
Figure S9: Hydrophobicity scale introduced by Abraham & Leo.[8]Residues which are more hydrophobic relative to glycine (0) have a positive value, while residues more hydrophilic than glycine were assigned negative values.

Figure S10 :
Figure S10: Specific activities for hydrolysis and maximum acyltransfer for all investigated bHSLs in the pNPA assay using 2-phenylethanol (A) and benzyl alcohol (B) as acyl acceptor.

Table 1 :
Statistics of the X-ray diffraction data collection and structure refinement.

deviations from ideal geometry
R.m.s.