A membrane‐bound esterase PA2949 from Pseudomonas aeruginosa is expressed and purified from Escherichia coli

Pseudomonas aeruginosa strain 1001 produces an esterase (EstA) that can hydrolyse the racemic methyl ester of β‐acetylthioisobutyrate to produce the (D)‐enantiomer, which serves as a precursor of captopril, a drug used for treatment of hypertension. We show here that PA2949 from P. aeruginosa PA01, a homologue of EstA, can efficiently be expressed in an enzymatically active form in E. coli. The enzyme is membrane‐associated as demonstrated by cell fractionation studies. PA2949 was purified to homogeneity after solubilisation with the nonionic detergent, Triton X‐100, and was shown to possess a conserved esterase catalytic triad consisting of Ser137–His258–Asp286. Our results should allow the development of an expression and purification strategy to produce this biotechnologically relevant esterase in a pure form with a high yield.

Pseudomonas aeruginosa strain 1001 produces an esterase (EstA) that can hydrolyse the racemic methyl ester of b-acetylthioisobutyrate to produce the (D)-enantiomer, which serves as a precursor of captopril, a drug used for treatment of hypertension. We show here that PA2949 from P. aeruginosa PA01, a homologue of EstA, can efficiently be expressed in an enzymatically active form in E. coli. The enzyme is membrane-associated as demonstrated by cell fractionation studies. PA2949 was purified to homogeneity after solubilisation with the nonionic detergent, Triton X-100, and was shown to possess a conserved esterase catalytic triad consisting of Ser137-His258-Asp286. Our results should allow the development of an expression and purification strategy to produce this biotechnologically relevant esterase in a pure form with a high yield.
Carboxyl esterases (EC 3.1.1.1) and lipases (EC 3.1.1.3), usually referred to as lipolytic enzymes, hydrolyse ester bonds of a wide range of lipidic substrates [1]. They exert various cellular functions in animals, plants and microorganisms [2][3][4]. The present classification scheme consists of 15 bacterial lipase families, which contain several hundreds of protein sequences and three dimensional structures [5,6]. Despite relatively low amino acid sequence similarity (typically below 20%) lipolytic enzymes have a common a/b-hydrolase fold, characterised by a central hydrophobic sheet composed of mostly eight b-strands, which are connected by a-helices [7]. These enzymes are characterised by a typical GXSXG-consensus motif containing the catalytic serine, which forms the catalytic triad with histidine and aspartate residues [8]. Glycine residues of the GXSXG-motif account for the localisation of the catalytic serine on top of a sharp turn preceded by an a-helix followed by a b-strand. This structural motif, named nucleophilic elbow, is essential for substrate hydrolysis and represents one of the best conserved structural motifs among a/b-hydrolases. The catalytic triad and the residues forming the oxyanion hole, which stabilises the tetrahedral intermediates formed during the hydrolysis, are structurally strongly conserved in all a/b-hydrolases [9].
Although many structural features of a/b-hydrolases are similar, these enzymes show a wide range of functionalities [10]. In recent years, the number of biochemical and structural data about lipolytic enzymes are rapidly increased, however, the understanding of structure-function relationship is still limited [11]. Nowadays, lipolytic enzymes are broadly used as industrial biocatalysts, since many of them are fairly stable in harsh conditions that are often needed for industrial bioprocess [12,13], for example, high temperatures, presence of organic solvents or high ionic strength buffers. Moreover, their enantio-, stereoand regioselectivity allow controlled organic synthesis reactions using complex unnatural substrates, as, for example, precursors for the biopharmaceuticals paclitaxel [14], naproxen [15] and captopril [16]. The latter is the first drug used for the treatment of hypertension, congestive heart failure and diabetic nephropathy, which acts by inhibiting angiotensin-converting enzyme [16]. Presently, captopril is still prescribed and distributed under the trade name Capoten Ò . Its chemical synthesis requires the optically pure key intermediate, (D)-b-acetylthioisobutyric acid (DAT) as the inhibiting potency of the (D)-enantiomer is 100-fold higher than of the (L)-enantiomer [16,17]. To overcome the laborious and expensive chemical synthesis of optically active DAT, Sakimae et al. [18] screened for microorganisms producing esterases, which can stereoselectively hydrolyse the racemic methyl ester of b-acetylthioisobutyrate [18]. The functional screening revealed several strains capable of DAT synthesis with Pseudomonas strains producing DAT of highest optical purity [18].
The esterase EstA from P. aeruginosa 1001 was cloned, overexpressed and shown to produce DAT [19,20]. However, expression of EstA in E. coli resulted in low enzymatic activity presumably caused by formation of protein aggregates, which could not be completely abolished by fusion with maltose-binding protein [19,20]. Here, we report cloning, expression and purification of the protein encoded by open reading frame pa2949 from P. aeruginosa PA01, which is homologous to EstA from P. aeruginosa 1001. We have developed a system for efficient expression of highly active PA2949 in common laboratory strain E. coli BL21(DE3). Biochemical analysis showed esterase but no lipase activity of PA2949, and enzyme activity could be inhibited by serine-hydrolase inhibitors. Furthermore, we showed that PA2949 is localised in the membrane of E. coli prompting us to develop a detergent-based purification method, which yielded mg amounts of enzymatically active protein.
Our data demonstrates that PA2949 can be functionally expressed, easily purified and adequately stabilised thus making it available for a range of different biotechnological applications.

Bioinformatic analysis
Amino acid sequences were analysed and aligned using the BLAST search and alignment tool of the Universal Protein Knowledge Base (www.uniprot.org) [21]. Signal peptide cleavage sites were predicted by two different methods, namely the Hidden Markov Model (Signal P-HMM) [22] and neural network (Signal P-NN) [22]. Signal peptides were distinguished from nonsignal peptides by a threshold D-score of SignalP-NN higher than 0.5 and by a threshold C-score of SignalP-HMM higher than 0.95. The transmembrane helix was predicted using the Toppred [23] online tool with a score higher than 0.8.

Cloning, site-directed mutagenesis, expression and purification of PA2949
Restriction endonucleases, Pfu DNA polymerase and bacteriophage T4 DNA ligase (Thermo Scientific, Darmstadt, Germany) reactions were carried out as recommended by the manufacturers. DNA fragments were analysed on 1% (w/v) agarose gels. Plasmid DNA was purified using the InnuPREP DOUBLE pure kit (Analytik Jena, Jena, Germany) or, for genomic DNA from P. aeruginosa PA01 using the DNeasy tissue kit (Qiagen, Hilden, Germany). Used strains and plasmids are listed in Table 1 [24][25][26][27].
The pa2949 gene was amplified by standard PCR using Pfu DNA polymerase, the genomic DNA of P. aeruginosa PA01 as a template and oligonucleotide pair 5 0 -AAACA TATGAAACGATTCCTC-3 0 /5 0 -TCAGAGCTCCACCACC ACCACCACCACGCGACCGGCCAC-3 0 encoding NdeI and SacI restriction sites (underlined) and a C-terminal His 6 -tag (bold). Primers were synthesised by MWG Biotech. The pa2949 gene was cloned into NdeI and SacI restriction sites of pET22b yielding expression plasmid pET-pa2949 (Table 1) allowing for bacteriophage T7-RNA polymerase-dependent expression from the T7 promoter. The mutation of Ser137Ala in PA2949 was performed by the Quik-change PCR method using Pfu DNA polymerase (Invitrogen, Darmstadt, Germany), pET-pa2949 plasmid and the complementary mutagenic oligonucleotide pair 5 0 -TGGCCGGCAACG T C C C C ATGGGCGGG-3 0 /5 0 -CCCG CCCATG G G G C A GTTGCCGGCCA-3 0 (mutated codons are underlined and nucleotides of the wild-type gene are indicated in the subscript). Correctness of plasmids pET-pa2949 and pET-pa2949_S137A was confirmed by DNA sequencing (MWG Biotech, Ebersberg, Germany).
For the expression of PA2949 and PA2949 S137A, E. coli BL21(DE3) cells transformed respectively with pET-pa2949 and pET-pa2949_S137A were grown overnight at 37°C in a Luria-Bertani (LB) medium supplemented with ampicillin (100 lgÁmL À1 ). These cultures were used to inoculate an expression culture in LB medium supplemented with ampicillin (100 lgÁmL À1 ) to an initial OD 580 nm = 0.05. The cultures were grown at 37°C until they reached logarithmic phase (OD 580 nm = 0.5-0.8) and gene expression was induced by addition of isopropyl-b-Dthiogalactosid (IPTG) to a final concentration of 0.4 mM. After 2 h, cells were harvested by centrifugation at 4000 g and 4°C for 20 min and stored at À20°C.

SDS/PAGE, zymography and immunodetection
Proteins were analysed by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS/PAGE) under denaturation conditions on 12% (w/v) gels as described by Laemmli [28]. Esterase activity in SDS/PAGE gels was detected by zymography using the fluorescent substrate 4-methylumbelliferyl butyrate (MUB) [29]. The proteins transferred from SDS/PAGE gel to the polyvinylidene difluoride membranes by western blotting [30] were detected using anti-His(Cterm)-HRP antibodies (Invitrogen) according the manufacturers' instructions. The protein concentration was determined by the method of Bradford with bovine serum albumin as a standard [31].

Purification
The total membrane fraction containing PA2949 with a C-terminal His 6 -tag was solubilised with 1% (w/v) Triton X-100 by gentle agitation overnight at 4°C. Solubilised membranes were subjected to ultracentrifugation (30 min, 180 000 g, 4°C) and PA2949 was purified from the supernatant by immobilised metal affinity chromatography (IMAC) using Ni-NTA agarose (Qiagen) [33]. All buffers used were supplemented with (1% w/v) Triton X-100 to keep PA2949 in soluble form. Purified PA2949 samples eluted from the Ni-NTA column were transferred into Tris-HCl buffer (100 mM, pH 8) containing 1% (v/v) Triton X-100 by gel filtration using PD-10 column (GE Healthcare, Solingen, Germany) according to the manufacturers' protocol. The purified protein was then stored at room temperature.

Enzyme activity assays
Enzymatic activities were determined in 96-well microplates by adding 5 lL of enzyme sample to 200 lL of substrate with p-nitrophenyl hexanoate (pNPC 6 ) for esterase and pnitrophenyl palmitate (pNPP) for lipase activity [1].

Temperature optimum
Esterase activities were measured over a range of temperatures from 10°C to 70°C as described previously [34]. Assays were performed in a 96-well microplate by adding 2 lL of enzyme sample to 200 lL of pNPC 6 substrate.

pH and organic solvents stability
Esterase activities of PA2949 incubated for 1 h with buffers of a pH range from 3 to 10.5 [34] or for 3 h with various organic solvents (dimethyl sulfoxide, N, N-dimethyl formamide, methanol, acetonitrile, ethanol, acetone, propan-2ol, diethyl ether, hexane, toluene) [34] were measured in a 96-well microplate by adding 2 lL of enzyme sample to 200 lL of pNPC 6 substrate.

Inhibition
The inhibition of PA2949 was tested according Asler et al. [35] using THL, PMSF, paraoxon (all dissolved in propan-2-ol) and the EDTA (dissolved in H 2 O). Inhibition of PA2949 was performed by incubating enzyme aliquots with the inhibitors for 3 h at the room temperature and subsequent determination of enzyme activity using pNPC6 as the substrate.

Results and Discussion
Open reading frame PA2949 of P. aeruginosa PA01 encodes a putative lipase By searching the Pseudomonas genome database (www.pseudomonas.com) [36] we have identified about hundred genes of P. aeruginosa PA01 encoding putative lipolytic enzymes, among them ORF pa2949. This gene of 948 bp length encodes a protein of M r 34.8 kDa with a predicted Abhydrolase_6 Pfam domain (PF12697) spanning residues 65-299 (Fig. 1A). A BLAST search revealed homology of PA2949 with known esterases and lipases, namely EstA from P. aeruginosa 1001 [19,20], and two lipases from psycrophilic bacteria, Moraxella sp. lipase [37] and Psychrobacter immobilis lipase [38] (Fig. 1B). The genes encoding pa2949 from P. aeruginosa PA01 and estA from P. aeruginosa 1001 share 99% identity (data not shown), and the protein sequences are identical (Fig. 1). The amino acid sequences of Moraxella sp. and P. immobilis lipases are~50% similar to the one of PA2949. The sequence alignment revealed the strictly conserved amino acid Ser137 embedded in the conserved GXSXG-lipase motif, as well as Asp258 and His286 predicted to form the catalytic triad of PA2949.
Expression of PA2949 in E. coli BL21(DE3) In order to obtain enzymatically active and soluble PA2949, we constructed a heterologous expression system using E. coli BL21(DE3) carrying plasmid pET-pa2949. Bacteria were grown at 37°C and expression of pa2949 was induced by addition of 0.4 mM IPTG. SDS/PAGE and western blot analyses revealed expression of a protein with an estimated molecular weight of 35 kDa. In parallel, we measured significantly increased esterase activity in the cell lysate of the expression strain compared with the strain carrying the empty vector (Fig. 2). Additionally, esterase activity of the 35 kDa protein was detected by zymographic analysis. Interestingly, PA2949 did not show activity with palmitic acid p-nitrophenyl ester considered as a typical lipase substrate. In conclusion, we could demonstrate that PA2949 of P. aeruginosa can be functionally expressed in E. coli BL21(DE3) and displays esterase but not lipase activity.

PA2949 is localised in the membrane of E. coli
The subcellular localisation of a protein must be considered to develop an efficient purification protocol [39]. As cellular localisation was unknown for PA2949 and its homologues, we first performed a bioinformatic  analysis revealing either a putative N-terminal type I signal peptide spanning amino acids 1-28 or a putative transmembrane helix spanning amino acids 4-24 suggesting a periplasmic, extracellular or membrane localisation. Many membrane proteins containing transmembrane helices display anomalous migration in SDS/PAGE [40] that is often heat inducible [41], and caused by differences in binding of SDS to heat-treated and untreated forms of the protein. Hence, we have tested the effect of temperature on the electrophoretic mobility of PA2949 and have shown that PA2949 migrated faster after incubation at 99°C prior to electrophoresis compared to the sample incubated at room temperature indicating that it is a membrane protein (data not shown). Therefore, we have experimentally separated membranes and soluble fraction of E. coli cells expressing PA2949. For this experiment, we expressed PA2949 in E. coli DH5a under the control of the weak lac promoter rather than the strong T7 promoter to avoid overloading of membranes and prevent mislocalisation [42]. The gene pa2949 was subcloned from plasmid pET-pa2949 into the broad host range vector pBBR1mcs-3 [43] and expression was performed in E. coli DH5a without induction. Cells were disrupted by sonication, membranes were isolated by ultracentrifugation, analysed by SDS/PAGE and western blotting and PA2949 was detected solely in the membrane fraction (Fig. 3). Thus, recombinant PA2949 expressed in E. coli is localised in the cell membrane, however, the localisation of PA2949 in the homologous host P. aeruginosa is still unknown. As E. coli and P. aeruginosa belong to the same class of Gamma-proteobacteria and share evolutionarily conserved signal recognition particles and Sec-translocation systems for targeting membrane proteins [44] we predict membrane localisation of PA2949 in P. aeruginosa as well.

Purification and biochemical characterisation of PA2949
Purification of membrane proteins requires the usage of detergents for solubilisation as well as to prevent subsequent protein aggregation by stabilising hydrophobic domains, which are naturally embedded in the membrane [45]. Here, we have selected the nonionic detergent Triton X-100 commonly used for purification of membrane proteins from E. coli [46] to extract PA2949 from the membranes. Initially, the total membrane fraction of E. coli BL21(DE3) expressing PA2949 was incubated for 1 h at room temperature with Triton X-100 at concentrations exceeding the critical micellar concentration. Although, Triton X-100 in a concentration range from 0.5 to 2% (w/v) did not reduce the esterase activity of PA2949, the protein was not quantitatively solubilised from the membranes (results not shown). Almost quantitative extraction of PA2949 was achieved after overnight incubation of membranes with detergent without losing esterase activity. Solubilisation of membranes with Triton X-100 and subsequent purification using immobilised metal affinity chromatography yielded~1 mg/L culture / OD 580 nm of pure PA2949 (Fig. 4). Purified PA2949 preparation had 198.8 AE 5.1 UÁmg À1 -specific esterase activity measured with p-nitrophenyl hexanoate at 30°C that corresponds to approximately 60-70% of total esterase activity of membrane fraction of E. coli BL21(DE3)-expressing PA2949. Incubation of PA2949 with various buffers (pH 3.0-10.5) revealed that activity was best retained in Tris-HCl buffer pH 8.0 at room temperature (Table 2), whereas storage at 4°C or freezing (also in the presence of 30% glycerol) lead to the precipitation and inactivation of PA2949.
The similarity of the PA2949 amino acid sequence with psychrophilic esterases (Fig. 1) prompted us to test if PA2949 retained its activity also at low temperatures. Determination of esterase activities at temperatures ranging from 10°C to 70°C revealed 30°C as the optimal temperature (Fig. 5). Interestingly, even at a temperature of 10°C PA2949 retained 47% of its activity measured at 30°C. This data indicate that PA2949 behaves similar to psychrophilic rather than mesophilic enzymes that are usually inactive at low temperatures [47].   Stability in the presence of organic solvents is an essential property of enzymes used as biocatalysts for organic synthesis [13]. Membrane proteins usually contain membrane-embedded domains and soluble domains protruding into the surrounding water milieu, which are differently resistant to hydrophobic (high log P) and hydrophilic (low log P) solvents [48]. Therefore, we have tested the stability of PA2949 after incubation with various solvents (Table 3). Notably, PA2949 was fairly stable in most of the tested solvents, some of them, for example, diethyl ether and methanol even enhanced its activity, but some hydrophilic solvents (acetonitrile, ethanol and propan-2-ol) severely reduced esterase activity.
Binding of substrate-mimicking inhibitors to the enzyme active site can provide valuable data for understanding the catalytic mechanism [49] as well as regio-and enantio preference [50] of lipases. We thus determined the inhibition kinetics of PA2949 using typical lipase inhibitors containing long hydrophobic acyl chains, namely tetrahydrolipstatine (THL) [51], the short acyl chain arylesterase inhibitors paraoxon [52], phenylmethylsulfonyl fluoride (PMSF) [53] and an inhibitor of metal-dependent enzymes, ethylenediaminetetraacetic acid (EDTA) [54]. Resistance towards EDTA indicated that PA2949 is not a metalloenzyme (Fig. 6). All three inhibitors targeting the catalytic serine residue (THL, PMSF and paraoxon) inhibited PA2949 activity, although to a different degree (Fig. 6). Incubation of PA2949 with THL for 3 h resulted in 28% of residual activity, whereas THL, PMSF and paraoxon completely abolished PA2949 activity indicating irreversible inhibition. Phosphonate or sulphonate inhibitors covalently linked to the catalytic serine, mimic the first tetrahedral intermediate (before dissociation of alcohol moiety) and the second tetrahedral intermediate (after dissociation of alcohol moiety) formed during ester hydrolysis respectively [10]. These results provided further evidence that PA2949 contains in its active site a nucleophlic serine. Using site-directed mutagenesis, we constructed variant PA2949 S137A, which was purified (Fig. 4), but did not show any esterase activity. These results are in agreement with the bioinformatic prediction of Ser137 as the catalytic residue. Additionally, our data suggest a rather narrow active site of PA2949 because the bulky inhibitor molecule THL apparently could not efficiently bind to the active site. The preparation of stable PA2949-PMSF and PA2949-paraoxon complexes we have described here will be used for further crystallographic and kinetic studies.
Biotechnological potential of PA2949 PA2949 from P. aeruginosa PA01 is homologous to the esterase EstA from P. aeruginosa 1001, which was previously shown to hydrolyse racemic b-acetylthioisobutyrate methyl ester to form enantiopure DAT [19,20], an important intermediate in the synthesis of pharmaceuticals [16]. However, heterologous expression of enzymatically active EstA turned out to be difficult because of poor solubility and low production yield [20]. EstA was predicted to be a soluble protein with an N-terminal signal sequence instead of a transmembrane helix [19,20]. This misleading assumption resulted in the construction of an expression system where maltose-binding protein (which was shown to enhance the solubility of proteins) was fused to the N-terminus of EstA resulting in blocking of EstA secretion and expression of only poorly active enzyme [20]. Using the pa2949 gene encoding the full-length protein with C-terminal His6-tag we have successfully expressed mg/L culture quantities of active PA2949 located in the membrane of the common laboratory strain E. coli. Subsequent extraction with the nonionic detergent Triton X-100 allowed for the efficient purification of up to 70% of PA2949 from membranes and enabled storage at room temperature without loss of activity. The biochemical characterisation of PA2949 revealed 30°C as the optimal temperature for catalysis (Fig. 5), high stability at pH 7.5-8.5 (Table 2), and notable resistance to various organic solvents (Table 3). Interestingly, a study of DAT production with whole cells of P. putida MR-2068, which expressed a PA2949 homologue revealed pH 7.5 and 45°C as the best conditions [55]. At higher temperatures and in alkaline pH spontaneous hydrolysis of the racemic b-acetylthioisobutyrate methyl ester was observed, thereby reducing the optical purity of DAT [55]. In summary, PA2949 possesses biochemical properties, which match the requirements for the enzymatic synthesis of DAT as well as its subsequent extraction with organic solvent.