Biochemical characterization of a family IV esterase with R-form enantioselectivity from a compost metagenomic library

A novel family IV esterase (hormone-sensitive lipase, HSL) gene, est15L, was isolated from a compost metagenomic library. Encoded Est15L comprised 328 amino acids with a molecular weight of 34,770 kDa and was an intracellular esterase without a signal peptide. The multiple sequence alignment (MSA) of Est15L with other family IV esterases showed conserved regions such as HGG, DYR, GXSXG, DPL, and GXIH. Native Est15L was a dimeric form from the results of size exclusion chromatography. It was optimally active at 50 ℃ and pH 9.0, indicating alkaline esterase. However, it showed a low thermostability with half-lives of 30.3 at 30 ℃ and 2.7 min at 40 ℃. It preferred p-nitrophenyl butyrate (C4) with Km and Vmax values of 0.28 mM and 270.8 U/mg, respectively. Est15L was inhibited by organic solvents such as 30% methanol, isopropanol, and acetonitrile with residual activities of 12.5, 0.9, and 0.3%, respectively. It was also inhibited by 1% SDS and 1% PMSF; however, Est15L maintained its activity at 1% Triton X-100 and EDTA. Est15L was inhibited by Cu2+, Zn2+, Mn2+, Co2+, Fe2+, and Na+. In addition, Est15L hydrolyzed glyceryl tributyrate with a residual substrate amount of 43.7% at 60 min but could not hydrolyze the oils (fish and olive) and glyceryl trioleate. Interestingly, Est15L showed significant enantioselectivity toward the R-form with a residual substrate amount of 44.6%, lower than that of the S-form (83.5%). Considering its properties, Est15L can be a potential candidate for chemical reactions, such as the synthesis of pharmaceutical compounds.


Introduction
Esterase (EC 3.1.1.1.) is a lipolytic enzyme that hydrolyzes ester bonds to carboxylic acid and alcohols. Bacterial lipolytic enzymes were first classified into eight families by Arpigny and Jaeger according to conserved amino acid sequence motifs and biochemical properties [1]. The lipolytic enzymes have been studied, and recently, family XIX was reported [2].
The family IV esterase is also called hormone-sensitive lipase (HSL) because it showed epinephrine-sensitive activity in human adipose tissue [3]. Family IV esterase belongs to an alpha/beta hydrolase and has β-sheet structures covered by α-helices [4]. The family IV esterase has two domains: a cap domain and a catalytic domain. The role of the cap domain is unknown, but there is a report that the cap domain of family IV esterase is deeply related to the recognition of substrates [5], and the catalytic domain has a catalytic triad: serine (S) in the GXSXG motif, aspartic acid (D), and histidine (H) [6].
Esterase has numerous applications. In particular, esterase can be used for chemical reactions, such as transesterification or production of biodiesel [7], and for ester prodrugs, which have been focused on the application for drug delivery systems to avoid metabolism and side effects [8]. For example, the 2nd-generation

A positive esterase clone from the compost metagenomic library
The metagenome was obtained from Yonghyun Nonghyub Compost Factory (Sachon, Korea), and its library was constructed using the fosmid vector [19]. From this library, 19 esterase-positive clones were obtained on LB agar plates containing 1% glyceryl tributyrate for 15 h at 37 ℃. They were mixed, digested with a restriction enzyme, cloned with plasmid pUC19, and 18 positive subclones were obtained [19]. By sequencing, nine different lipolytic enzymes were identified, and some of them were reported [19,23,24]. In this study, a positive clone YH-E15 was selected for further study.

Sequence analysis of the insert DNA in the positive clone
DNA sequences of the esterase-positive clone were determined using the Sanger dideoxy method by Solgent (Daejeon, Korea). From this sequence, the ORF similar to esterase was confirmed, and its amino acid sequence was analyzed by BLASTp of NCBI (http:// www. ncbi. nlm. nih. gov). Prediction of signal peptide was performed using SignalP 5.0 in CBS (http:// www. cbs. dtu. dk/ servi ces/ Signa lP/). Molecular weight and pI were predicted using the ExPASy ProtParam tool (http:// web. expasy. org/ protp aram). Clustal W method of DNA/MAN (Lynnon Biosoft, version 4.11, Quebec, Canada) was used to analyze multiple sequence alignment, and the neighborjoining method in MEGA version X [26] was used to construct the phylogenetic tree. Similarities between the identified enzyme Est15L and other enzymes were calculated using DNA/MAN.

Preparation of crude enzymes
The clone YH-E15 was cultured in 200 mL of LB broth containing 50 μg/mL of ampicillin and incubated for 15 h at 37 ℃ and 200 rpm. The cell was collected from the cultured medium by centrifuging at 4 ℃ and 6,000 × g for 15 min. The collected pellet was washed two times with 20 mL of 20 mM Tris-HCl (pH 8.0) buffer by centrifugation at 4 ℃ and 6,000 × g for 5 min, resuspended with 5 mL of the same buffer, sonicated (amplitude of 38%, pulse on for 1 s and pulse off for 1 s) three times using a microtip sonicator (VCX500, Sonics & Materials, Newtown, CT, USA), and then centrifuged at 4 ℃ for 15 min at 6,000 × g. The supernatant was collected as a crude extract.

Purification of Est15L
Before the purification step, the crude enzyme was centrifuged for 15 min at 6,000 × g and 4 °C. The supernatant was loaded to a HiTrap Q anion exchange column in a BioLogic LP system (Bio-Rad, Hercules, CA, USA) with 20 mM Tris-HCl (pH 8.0), and the buffer flowed with a linear gradient with a high buffer containing 1 M NaCl at 1 mL/min for 1 h 30 min. The active fractions were pooled, dialyzed with 50 mM sodium phosphate (pH 7.0) containing 1.5 M (NH 4 ) 2 SO 4 for the t-butyl HIC column as a second column, loaded to the column, and eluted with a high buffer containing 1.5 M (NH 4 ) 2 SO 4 with a linear gradient at 1 mL/min for 1 h. To confirm the native molecular mass of the enzyme, Sephacryl S-200 size exclusion chromatography was performed using 50 mM sodium phosphate (pH 7.0) containing 0.15 M NaCl at a flow rate of 0.5 mL/min for 4 h. β-Amylase, bovine serum albumin (BSA), and trypsinogen (200, 66.4, and 24.0 kDa, respectively) were used as standard markers. During purification, active fractions were loaded on 11.5% acrylamide gel, and then SDS-PAGE was performed [27]. The concentration of protein was determined by Bradford assay using BSA as a standard [28].

Enzyme assays
The standard esterase assay was performed using 1 mM p-NP butyrate in 50 mM Tris-HCl (pH 8.0). The amount of p-nitrophenol as the product was observed continuously by kinetic mode in a spectrophotometer (OPTI-ZEN, K-Lab, Daejeon, Korea) for 2 min at 25 ºC at 400 nm. The molecular extinction coefficient of p-nitrophenol used was 16,400 /M/cm at pH 8.0. The production of 1 µmol p-nitrophenol per minute was defined as one unit of an enzyme.
The acetyl-or butyryl-cholinesterase activity was measured by the Ellman method using ATCI or BTCI, respectively, as the substrate, as previously described [29]. Briefly, the enzyme was added to 100 mM sodium phosphate containing 0.5 mM DTNB and 0.5 mM ATCI or BTCI, respectively, and the absorbance of the reaction mixture was observed continuously at 412 nm for 15 min at 25 ºC using kinetic mode in spectrophotometer (OPTIZEN).

Characterization of the enzyme
The standard enzyme assay was characterized using p-NP butyrate with slight modification. For the optimum temperature experiment, the buffer was preheated to 20, 30, 40, 50, 60, and 70 ℃ prior to assay. For optimum pH, 50 mM Universal buffer (boric acid/ citric acid/ trisodium orthophosphate) for pH 6.0 to 12.0 was used. The molecular extinction coefficients at each pH were used as previously described [18]. For thermostability, the enzyme was heated for 0, 5, 10, 20, 30, and 60 min at 30 and 40 ºC, added to the assay mixture, and its residual activity was measured.
Ions such as NaCl, KCl, MgCl 2 , CaCl 2 , BaCl 2 , MnCl 2 , FeCl 2 , CoCl 2 , CuCl 2 , and ZnCl 2 were added to the assay solution at 2 or 5 mM to confirm the effects of ions. Methanol, isopropanol, and acetonitrile were added to the assay solution to have a final concentration of 5 or 30% to confirm the effect of the organic solvent. Effects of detergents (such as SDS and Triton X-100) were observed at a concentration of 1%. The effects of inhibitors, such as phenylmethylsulfonyl fluoride (PMSF) and ethylenediaminetetraacetic acid (EDTA), were observed at a concentration of 1 mM.
Lipid hydrolysis activity was measured with a pH shift assay using oils (fish and olive oil) and glyceryl triesters (glyceryl tributyrate and glyceryl trioleate) as substrates [30]. The enzyme was reacted with the substrate in 20 mM Tris-HCl (pH 8.0) containing 0.1% phenol red, and its absorbance at 560 nm was observed continuously using kinetic mode in the spectrophotometer (OPTIZEN) at 25 ºC for 60 min at 5 min intervals. The enantioselectivity was measured using 1% (R)-methyl-3-hydroxy-2-methyl-propionate or (S)-methyl-3-hydroxy-methyl propionate as a substrate for the pH shift assay [18].

Sequence analysis and multiple alignments of Est15L
Due to DNA sequencing for the positive clone YH-E15, it was revealed that insert DNA comprised 2,587 bp, and an open reading frame (ORF) was predicted to be an esterase. The ORF was 987 bp in length and named est15L. The encoded Est15L esterase comprised 328 amino acids with molecular weights of 34,770 Da with no signal peptide, and its predicted theoretical pI value was 4.57. Est15L has been deposited under the accession number of OK336712 in GenBank. In BLASTp, Est15L showed the highest homology (85.03%) to alpha/ beta hydrolase of Sphingorhabdus sp. (GenBank accession number, MBF6602187) obtained from metagenome-assembled genomes isolated from diarrhea affected cattle B. Conversely, alpha/beta hydrolase of Sphingorhabdus sp. Showed a similar identity (100%) with Est8L (QZA73595), which was obtained from the compost metagenomic library [24]. Est15L showed relatively low identity to other reported enzymes, and enzymatic properties were characterized in further study.
In the phylogenetic tree, it was confirmed that Est15L is a novel member of family IV esterase (i.e., HSL) ( Fig. 1).

Purification of Est15L
Est15L was bound to HiTrap Q and eluted with the linear gradient (Fig. 3A). Specific activity was increased 5.50 times (24.86 U/mg), compared to the crude extract (4.52 U/mg), with a yield of 24.8%. In size exclusion chromatography using Sephacryl S-200, Est15L was eluted at 61.5 mL, with an increased specific activity of 160.3 U/ mg ( Fig. 3B; Table 1).
In SDS-PAGE, the predicted band of Est15L, corresponding to about 34.9 kDa, was detected in the fractions from Sephacryl S-200 in an activity-dependent manner (Fig. 4). However, it showed smear bands around the major band, along with some minor bands at 100, 75, 48, and 27 kDa. Est15L was partially purified using the two chromatographies. Est15L did not bind to other resins (such as CHT-II, High S, and t-butyl HIC) and showed low yields of less than 8.7%.
From the elution volume, the molecular mass of Est15L was calculated to be 67.2 ± 8.8 kDa, and it can be predicted that native Est15L existed as a dimeric form ( Fig. 3B; Table 2).

Characterization of Est15L
Est15L was optimally active at 50 ℃ and pH 9.0, indicating Est15L was an alkaline esterase (Fig. 5A, B). In Fig. 1 Phylogenetic tree of Est15L with other esterases/lipase family. The phylogenetic tree was constructed using the neighbor-joining method in MEGA version X with 1,000 bootstrap replications. The accession numbers are accessible using GenBank of NCBI thermostability, Est15L was sensitive to thermal stress with half-lives of 2.7 min at 40 ℃ and 30.3 min at 30 ℃ (Fig. 5C).
Est15L preferred C 4 followed by C 6 , C 8 , and C 2 with relative activities of 69.3, 49.3, and 35.8%, respectively (Fig. 6A). On the other hands, Est15L did not show AChE and BChE activities. In a kinetic study using C 4 , K m and V max values of Est15L were 0.28 ± 0.02 mM and 278.0 ± 10.9 U/mg, respectively (Fig. 6B).
In the presence of 30% methanol, isopropanol, and acetonitrile, Est15L activity was significantly inhibited to 12.5, 0.9, and 0.3%, respectively. In detergent, Est15L activity was maintained to 88.2% at 1% Triton X-100. However, it was extensively inactivated to 0.38% at 1% SDS. In the case of inhibitors, Est15L was stable at 1 mM EDTA with a relative activity of 89.9% but was strongly inhibited to 2.5% by 1 mM PMSF (Fig. 7A).
Est15L efficiently hydrolyzed glyceryl tributyrate with a residual substrate amount of 43.7%. However, no significant hydrolysis activity was observed for olive oil, fish oil, and glyceryl trioleate (Fig. 8A). Interestingly, Est15L showed higher enantioselectivity toward the R-form with a residual substrate amount of 44.6% than toward the S-form with 83.5% after a 60 min reaction (Fig. 8B).

Discussion
In this study, an esterase Est15L, a novel member of family IV (HSL family), was obtained from a compost metagenomic library. Est15L showed the highest similarity (85.03%) to alpha/beta hydrolase of Sphingorhabdus sp. (MBF06602187) and Est8L, and the properties of Est8L were reported in our previous study [24]. Though Est15L and Est8L showed high similarity of amino acid sequence, their properties were different from each other, such as optimum temperature (50 vs. 40 ℃), thermostability (half-lives of 2.7 min at 40 ℃ vs. 3.2 min at 50 ℃), enantioselectivity (R-form vs S-form), and organic solvents effect. It has been reported that some lipolytic enzymes showed different properties in a specific activity, enantioselectivity, ionic effect, and organic solvent effect, despite high similarities of their amino acid sequences [18,24,[31][32][33][34]. In predictions of 3D structure models, Est15L showed the highest identity (33.89%) to Esterase Crystal structure of Chloramphenicol-Metabolizing Enzyme EstDL 136-Chloramphenicol complex (PDB code: 6iey.1.A) homo-dimer form. Est8L showed the highest identity (33.55%) to Esterase Crystal structure of Chloramphenicol-Metabolizing Enzyme EstDL 136 (PDB code: 6aae.1.A) with homo-dimer form. In the predicted model, the N-terminal of Est15L (M1 ~ A18) was not predicted because it did not identify with other family IV esterase; thus, its structure could not be predicted. The Chain A of Est15L comprised 11 α-helix and 8 β-sheets, and Chain B comprised 9 α-helix and 8 β-sheets (Fig. 9A,  C). Moreover, the Chain A of Est8L comprised 12 α-helix and 8 β-sheets, and Chain B of Est8L comprised 11 α-helix and 8 β-sheets (Fig. 9B, D). In the case of β-sheet structures, no difference was found between Est15L and Est8L. However, Est15L has three fewer α-helix structures than Est8L; that is, the motifs F42 ~ T49 of Chain B, A129 ~ E143 of Chain A, and L286 ~ L289 in Est8L were predicted as α-helical structures, whereas Est15L did not show (Fig. 9). Its structural differences might occur in biochemical properties such as optimum temperature, thermostability, and enantioselectivity. By using the model templates as above, docking simulations of Est15L and Est8L with (R)-or (S)-methyl-3-hydroxy-2-methyl-propionate were performed. Interestingly, Est15L showed a higher affinity with the (R)-form (ΔG = -6.31 kcal/mol) than (S)-form (ΔG = -6.23 kcal/mol) at cap domain for the strongest binding, which is considered to have an important role in recognition of the substrate (Additional file 1: Table S1). However, Est8L, showing (S)-form selectivity, also showed a higher affinity with the (R)-form, (ΔG = -6.92 kcal/mol) than with the (S)-form, (ΔG = -6.87 kcal/mol) (Additional file 1: Table S2). In addition, no interaction or binding was predicted between the catalytic triads of the enzymes and their substrates, except that the (S)-form formed a hydrogen bond with Ser176 of Est15L, at a distance of 3.082 Å (ΔG = -6.22 kcal/mol) (Additional file 1: Figure S1). Collectively, though the docking values could not sufficiently support the selectivity, it is suggested that other bindings may contribute to their selectivity at the same time.
Est15L had no signal peptide, likely most of the family IV esterase, suggesting Est15L is an intracellular esterase, but EstA1 had a signal peptide [42].
Est15L showed great enantioselectivity toward the R-form (38.90% higher than the S-form), whereas Est8L showed an S-form preference. The R-form enantioselectivity of Est15L was similar to PestE, a family IV esterase; however, Est15L was different from PestE in specific activity (160.3 vs. 3,910 U/mg), optimum pH (9.0 vs. 7.0), substrate preference (C 4 vs. C 6 ), and optimum temperature (50 vs. 90 ℃) [53,54] (Table 3). Enantioselective esterases can be used for purifying and enriching the specific enantiomer [71]. For example, the lipolytic enzyme from pseudomonas cepacia is a popular catalyst for hydrolysis, transesterification, and esterification of racemic mixtures of secondary alcohols to synthesize important enantiomers [72]. Est15L can be used for purifying the racemic mixtures to enrich the (S)-form enantiomer by selective hydrolysis of the (R)-form enantiomer.
Collectively, Est15L was sensitive to organic solvents, ions, and thermal stress, but it showed good specific activity and enantioselectivity. In this study, Est15L was obtained from a compost metagenomic library, and its     50%, c 15%, d 10%, and e 25%, *** X means it does not have enantioselectivity, S or R means it has S-form or R-form enantioselectivity, respectively. **** (+) Means it is activated by the ion, (−) means it is inhibited by the ion and (0) means that it showed stability by the ion. # Since Est2L is a fusion type esterase, the esterase domain (422 AA) was used for homology analysis enzymatic properties were characterized. Est15L was a novel protein and showed the highest similarity (85.03%) with an alpha/beta hydrolase of Sphingorhabdus sp. and Est8L. Though Est15L and Est8L showed high similarity, their properties were different from each other: optimum temperature (50 vs. 40 ℃), thermostability (half-lives, 2.7 min at 40℃ vs. 3.2 min at 50 ℃), specific activity (160.3 vs. 388.6 U/mg), and enantioselectivity (R-vs. S-form).
Est15L was sensitive to organic solvents such as methanol, isopropanol, and acetonitrile, whereas Est8L was relatively tolerant to them. On the other hands, Est15L had great enantioselectivity with a residual substrate amount of 43.68% for the R-form. We tracked why these differences are occurred by predicting 3D structures and docking simulations, and we found structural differences between Est15L and Est8L, i.e., three less α-helices and longer cap domain of Est15L, which showed higher affinity with R-form in docking simulation. Although Est15L is less stable to ionic and thermal stress than Est8L, the enantioselectivity of Est15L will be more valuable for chemical applications. These results suggest that Est15L is a novel member of family IV esterase and a potential candidate of the chemical reaction or ester prodrugs with enantioselectivity.