In order to avoid deficient translation, the lipase gene was ordered from Geneart (Regensburg, Germany) as synthetic gene with optimized codon usage for expression in E. coli. Different vectors were chosen (pET32b+ (Trx-tag), pET39b+ (DsbA-tag), pET40b+ (DsbC-tag) and pET44b+ (NusA-tag)) to achieve tag-mediated soluble expression of the corresponding proteins. Appropriate primer pairs were used to amplify the lipase gene by standard PCR methodology using Pfu DNA polymerase (Promega, Fitchburg, USA) and to prolong the sequence by the restriction sites for NcoI or SacI (forward) and XhoI (reverse). After purification of the PCR products, double digestion of the PCR fragments as well as all vectors was performed in accordance to the manufacturer’s protocol. Finally, lipasegene and vectors were ligated by T4 ligase (Thermo Fisher Scientific, Waltham, USA) at 22˚C overnight. The resulting constructs were transformed into E. coli DH5α and correctness of all constructs was confirmed by sequencing of the entire coding sequence (Eurofins MWG Operon, Ebersberg, Germany).

Functional plasmids were transformed into E. coli BL21 (DE3) and cells were grown overnight at 37˚C in 5 mL LB medium with the appropriate antibiotic. Antibiotic concentration was 100 µg/mL for ampicillin and 50 µg/mL for kanamycin, respectively. The preparatory culture was transferred into the main culture (500 mL TB media in a 2 L Erlenmeyer flask), grown for several hours at 37˚C and protein expression was induced with IPTG (0.1 mM final concentration) at an OD₆₀₀ of 1.5. Low temperature expression was chosen at 17˚C for 24 h and 150 rpm. The resulting proteins were named Trx-lipase, DsbA-lipase, DsbC-lipase and NusA-lipase, respectively.

The cells were harvested by centrifugation (7500 × g for 5 min at 4˚C), washed in 50 mM Tris-HCl buffer (pH 8.0) and finally resuspended in 5 mL of the same buffer. Cell lysis was done by sonification and the soluble protein fraction was separated by centrifugation (15,000 × g for 20 min at 4˚C).

Purification was achieved by IMAC using an ÄKTA FPLC system (GE Healthcare, Chalfont St Giles, UK) equipped with a 5 mL Ni-NTA superflow cartridge (Qiagen, Venlo, Netherlands). For all steps a flow rate of 5 mL/min was applied. Prior to the purification, the column was equilibrated with 5 - 15 mL equilibration buffer (50 mM Tris-HCl, 0.3 M NaCl, 20 mM imidazole, pH 8.0). The protein containing sample was supplemented with 20 mM imidazole, filtered through a sterile filter (0.22 µm) and injected into the chromatography system. After sample loading, the column was washed with 15 mL equilibration buffer. Elution of the target protein was obtained by an elution gradient from 20 - 500 mM imidazole over 10 mL and a final step using 5 mL elution buffer (50 mM Tris-HCl, 0.3 M NaCl, 500 mM imidazole, pH 8.0). Protein containing fractions were pooled and residual imidazole was removed by a desalting step using a HiPrep 26/10 Desalting column (GE Healthcare, Chalfont St Giles, UK). The column was equilibrated with desalting buffer (1 mM Tris-HCl, pH 8.0) and desalting occurred over two column volumes at a flow rate of 5 mL/min. Afterwards, the protein preparation was lyophilized and stored at −80˚C.

Lipase lyophilisates were resolved in the appropriate buffer. Protein concentrations were determined by Bradford assay and a stock solution (100 µg/mL) was further diluted according to the specific assay. In general, enzyme activity assays contained 1780 µL of the corresponding buffer, 20 µL enzyme solution and 200 µL substrate solution.

Substrate specificity was determined using a set of p-nitrophenyl esters with saturated carboxylates ranging from C₄ to C₁₆ (p-nitrophenyl butyrate (pNPB), p-nitrophenylcaprylate (pNPC), p-nitrophenyllaurate (pNPL), p-ni- trophenylpalmitate (pNPP)). Esters were dissolved in 2-propanol to form the substrate solution (1 mM). In these experiments 50 mM Tris-HCl pH 8.0, supplemented with 1% triton X-100 as surfactant, served as standard buffer. The absorption at 405 nm (absorption maximum of p-nitrophenol (pNP)) was investigated over 2 min and the slope was calculated to determine specific activities. The molar extinction coefficient of pNP for the different pH values and buffers was determined separately (data not shown).

Determination of the stability as well as temperature and pH optimum was performed with the activity assay described above. Noteworthy, no surfactant was supplemented in the buffers and the buffer strength was 50 mM each. Hydrolysis of pNPC served as standard for all measurements. The stability was tested by incubation of the enzyme in buffer at room temperature over several days. The temperature optimum was determined in potassium phosphate buffer pH 8.0. The pH optimum was investigated over a pH range of pH 4 - 11. Therefore, the following buffers were used: citrate buffer (pH 4 - 6), potassium phosphate buffer (pH 6 - 8), Tris-HCl buffer (pH 8 - 9) and CAPS buffer (pH 10 - 11). If buffers with pH < 7 were used, the absorption maximum was measured at 320 nm instead of 405 nm.

Enantioselectivity was determined by the enzymatic hydrolysis of α-methylbenzyl butyrate (MBB) to 1- phenylethanol. The assay was performed in Tris-HCl pH 8.0, the starting concentration of MBB was 400 mM. Reaction vessels were shaken at 25˚C and samples of 200 µL were taken periodically. 1 mL ethyl acetate was added to the samples and intensive mixing transferred the products to the water-free phase. Samples were dried with ammonium sulfate to remove remaining water and analyzed by gas chromatography. R- and S-enantiomer of 1-phenylethanol were separated in a GC 2010 gas chromatograph (Shimadzu, Kyoto, Japan) equipped with an AOC-20i auto injector. For separation of all chiral compounds a FS-Cyclodex β-I/P column (CS—Chromatogra- phie Service GmbH, Langerwehe, Germany) was used. The carrier gas (N₂) was applied with a pressure of 50 kPa and 1 µL of the extracted samples were injected on the column with split flow. The optimized method was 60˚C for 2 min, 60˚C - 90˚C for 5 min (slope: 6˚C/min), 90˚C - 150˚C for 20 min (slope: 3˚C/min) and 150˚C for 2 min. Retention times were 5.4 min for vinyl butyrate, 22.1 min for (R)-1-phenylethanol and 22.6 min for (S)-1-phenylethanol.

Nucleotide and amino acid sequences of the different proteins were found at NCBI and homology arrangements were performed by the BLAST tool of NCBI. The visualized model of the P. stutzeri lipase was obtained from Swiss-Model [16] - [18] and processed and displayed with Yasara v.14.2.12.

The identified gene was synthesized and cloned into different expression vectors. Noteworthy, codon usage of the wild type gene was adapted for expression in E. coli to avoid incorrect translation. Unfortunately, expression without solubilizing tags resulted in insoluble protein located in inclusion bodies (data not shown). Therefore, the expression of fusion proteins with enhanced solubility was chosen [21] , which was formerly described to be transferable to lipases. For example, Narayanan and co-workers reported the soluble expression of a lipase from Burholderia when fused to DsbA or DsbC [22] . Also the soluble expression of lipase B from Candida antarctica was improved by N-terminal addition of a thioredoxin tag [23] . We selected four different tags, namely thioredoxin (Trx, 12 kDa), two thiol: disulfide interchange proteins (DsbA, 23 kDa and DsbC, 25.5 kDa) and the transcription termination/antitermination protein NusA (NusA, 55 kDa). Expression plasmids were constructed resulting in the expression of the lipase/esterase with the respective N-terminal tag. The proteins were named Trx-lipase, DsbA-lipase, DsbC-lipase and NusA-lipase. Overexpression in E. coli BL21 (DE3) resulted in soluble expression of all four fusion proteins (Figure 2(a)). According to their molecular mass, all the constructs were clearly assignable. Since smaller tags would have lower influence on the overall structure of the fusion protein, its enzyme activity was investigated. All the fusion proteins were active with the highest activity for Trx-lipase (Trx-Lip, Figure 2(b)). Because of the smallest tag size and the highest activity in the cell raw extracts, Trx-Lip was selected for further investigation. Purification by IMAC was followed by desalting and lyophilisation and produced the target enzyme in adequate purity.

First of all, substrate range and specific activities of purified Trx-Lip were determined. Therefore, a set of different p-nitrophenyl esters, ranging from C₄ to C₁₆ was used (Figure 3). Trx-Lip showed maximum activity towards middle chain substrates (C₄ to C₈), whereas esters with unsaturated long chain fatty acids (C₁₂ to C₁₆) were converted very slowly or not at all. The K_M-value for the best substrate (pNPB) was determined to be 0.43 mM calculated by Lineweaver-Burk plot. This indicates a high affinity to short and medium chain fatty acid substrates.

It was suggested that a preliminary classification of esterases and lipases can be done from their substrate spectrum using p-nitrophenyl esters [24] . Esterases hydrolyze small alkyl esters like pNPB, while lipases are able to hydrolyze long chain fatty acid esters like pNPP. Nowadays, this classification is controversially discussed [25] [26] , but still the hydrolytic behaviour can give a first hint. Our findings suggested that Trx-Lip corresponded more to an esterase than to a lipase. Furthermore, interfacial activation of the enzyme could not be observed, which is in accordance to the behaviour of the structurally related lipase from P. mendocina [19] .

Especially for industrial applications, it is necessary to know which conditions can be exposed to a certain enzyme. First, the stability of the recombinant enzyme was tested in buffer. Lyophilized Trx-Lip was resuspended in Tris-HCl buffer pH 8.0 and incubated at room temperature. After 24 h a relative activity of 96% was detected and even after 8 days the lipase still displayed a residual activity of 86% (Figure 4(a)). Additionally, temperature and pH spectra were investigated (Figure 4(b) and Figure 4(c)). The temperature spectrum of Trx-Lip showed a linear slope with increasing temperature and a maximum activity at 50˚C. At higher temperatures the lipase became unstable and lost its activity very fast. The pH range of Trx-Lip was very broad with the highest activity between pH 7.5 - 10 and a maximum at pH 9. The high activity at alkaline pH and the temperature optimum above 40˚C referred this enzyme to typical representatives of lipases and esterases [14] . Additionally, these facts indicate the potential application of the new lipase in industrial processes, for example in the detergent industry.

Enantioselective synthesis of chiral compounds is a major concern in modern pharmaceutical production. Lipases display attractive catalysts because they can either hydrolyse prochiral substrates or convert one enantiomer from a racemic mixture [27] . This increases reactant flexibility and enables different synthetic routes. The possible application of the new enzyme was investigated by the hydrolysis of α-methylbenzyl butyrate to 1-phenylethanol. A commercial lipase from P. stutzeri (Lipase TL^®), which has been shown to process bulky secondary alcohols [28] [29] , was used as positive control. Resulting products were analysed by GC and the amount was calculated from appropriate standard curves (Figure 5).

Not surprisingly, the specific activity of Lipase TL^® (starting activity: 5660 U/mg protein) is higher than that of Trx-Lip (starting activity: 523 U/mg protein). Both enzymes are able to hydrolyse the R-enantiomer with good enantiomeric excess (ee). The ee for Lipase TL^® was calculated to be >99%. In the first 2 h, also Trx-Lip showed ee-values > 93%, but within the next 15 h the ee decreased to 83%. Nevertheless, the exclusive hydrolysis of the R-enantiomer at the beginning indicates not only a stereoselective preference, but also illustrates the ability to perform kinetic resolution of chiral esters. Furthermore, Lipase TL^® is able to perform the (dynamic) kinetic resolution of α-hydroxyketones [29] [30] . In our experiments, Trx-Lip was not able to catalyse kinetic resolution of benzoin, and the sterically less complex α-hydroxypropiophenone could also not be converted (data not shown). Despite the same origin (both enzyme preparations are from Pseudomonas stutzeri), Trx-Lip behaves different from Lipase TL^® and was thereby not a component of the commercial product. This illustrates that Trx-Lip indeed presents a new until now unknown lipase/esterase. It is known that the enantioselectivity of lipases can be influenced by many parameters like the choice of solvent [31] or by immobilization [32] . Optimized conditions could thereby improve the stereoselectivity of the new lipase/esterase. Herein also just one

substrate was tested for enzymatic performance and enantiomeric selectivity. Additional substrate screenings could identify enzyme requirements and will lead to a better understanding of this new enzyme.