Anal. Chem. 2003, 75, 255-260
Fluorophoric Assay for the High-Throughput Determination of Amidase Activity Erik Henke†,‡ and Uwe T. Bornscheuer*,†
Institute of Chemistry & Biochemistry, Department of Technical Chemistry & Biotechnology, Ernst-Moritz-Arndt-University Greifswald, Soldmannstr. 16 D-17487 Greifswald, Germany, Institute of Technical Biochemistry, University of Stuttgart, Allmandring 31, D-70569 Stuttgart, Germany
An assay has been developed for the high-throughput identification of amidase activity. Amines released from the enzyme-catalyzed hydrolysis of corresponding amides were detected by the formation of a fluorescent dye by coupling with 4-nitro-7-chloro-benzo-2-oxa-1,3-diazole (NBD-Cl). Using this format, 22 lipases and esterases were tested for their ability to hydrolyze aromatic substituted N-acylamines in a microtiter plate format. Identified active enzymes were further characterized toward a broad range of compounds to determine the influence of substrate structure on activity. For recombinantly produced esterases, it could be shown that the assay works with high reproducibility and sensitivity, even in the presence of amino acids and proteins present in culture media and cell debris.
Modern biocatalysis research has generated a need for simple, inexpensive, and fast assay systems for high-throughput screening of enzyme activity. Novel methods to improve enzyme properties, such as directed evolution1,2 and increasing efforts to exploit the natural diversity of biocatalysts obtained, for example, by DNA shotgun cloning from environmental sources such as soil samples,3 are mainly based on the generation of large expression libraries (usually 104-107 individual enzymes/mutant). These are then usually screened for activity using microtiter plates on the basis of assays in combination with absorbance or fluorescence measurements, because this provides high reliability and throughput. One interesting target in biocatalysis is the kinetic resolution of chiral racemic amines with hydrolases, because enantiomerically pure amines are key building blocks in pharmaceutical and pesticide synthesis with a considerable market size. Thus, it is not surprising that several attempts were made in the past to employ hydrolytic enzymes in the synthesis of enantiomerically * Corresponding author. Phone: (+49) 3834-86-4367. Fax: (+49) 3834-864346. E-mail:
[email protected]. Web: http://www.chemie.unigreifswald.de/∼biotech. † Ernst-Moritz-Arndt-University. ‡ University of Stuttgart. (1) Bornscheuer, U. T.; Pohl, M. Curr. Opin. Chem. Biol. 2001, 5, 137143. (2) Arnold, F. H.; Volkov, A. A. Curr. Opin. Chem. Biol. 1999, 3, 54-59. (3) Miller, C. A. Inform 2000, 11, 489-495. 10.1021/ac0258610 CCC: $25.00 Published on Web 12/11/2002
© 2003 American Chemical Society
pure amines. This included amidases (EC 3.5) and proteases (EC 3.4)4,5 as well as lipases (EC 3.1.1.4).6-10 Moreover, the BASF AG (Ludwigshafen, Germany) has recently established a large-scale process (>2000 metric tons/ year) for the resolution of a variety of amines based on a highly stereoselective acylation with a methoxy acetic acid ester catalyzed by a lipase from Burkholderia plantarii.11,12 However, the currently used methods for the detection of enzymatic activity toward amides or amines are based on classical sequential approaches in which a handful of biocatalysts are used in small-scale experiments followed by standard analysis using gas chromatography or HPLC. This makes a thorough investigation of a broad range of substrates and reaction conditions very tedious and time-consuming. On the other hand, it has already been shown that the activity of many enzymes depends strongly on the substrate structure; for example, the choice of acyl donor in lipase-catalyzed acylation of amines is very crucial for success.13,14 We here describe a novel format for the determination of enzymatic hydrolytic activity toward amides that can be performed in a high-throughput format and, thus, allows the investigation of a broad range of substrates and a screening within large enzyme libraries. EXPERIMENTAL SECTION Reagents and Materials. All chemicals were purchased at the highest purity available from Fluka (Buchs, Switzerland) and (4) Brieva, R.; Rebolledo, F.; Gotor, V. J. Chem. Soc., Chem. Commun. 1990, 20, 1386-1387. (5) Kitaguchi, H.; Fitzpatrick, P. A.; Huber, J. E.; Klibanov, A. M. J. Am. Chem. Soc. 1989, 111, 3094-3095. (6) Gotor, V.; Menendez, E.; Mouloungui, Z.; Gaset, A. J. Chem. Soc., Perkin Trans. 1 1993, 9, 2453-2456. (7) Jaeger, K. E.; Liebeton, K.; Zonta, A.; Schimossek, K.; Reetz, M. T. Appl. Microbiol. Biotechnol. 1996, 46, 99-105. (8) Kanerva, L. T.; Csomos, P.; Sundholm, O.; Bernath, G.; Fulop, F. Tetrahedron: Asymmetry 1996, 7, 1705-1716. (9) Pozo, M.; Gotor, V. Tetrahedron 1993, 49, 4321-4326. (10) Puertas, S.; Brieva, R.; Rebolledo, F.; Gotor, V. Tetrahedron 1993, 49, 40074014. (11) Liese, A.; Seelbach, K.; Wandrey, C. Industrial Biotransformations; WileyVCH: Weinheim, 2001. (12) Balkenhohl, F.; Ditrich, K.; Hauer, B.; Ladner, W. J. Prakt. Chem. 1997, 339, 381-384. (13) Smidt, H.; Fischer, A.; Fischer, P.; Schmid, R. D. Biotechnol. Tech. 1996, 335-338. (14) Wagegg, T.; Enzelberger, M. M.; Bornscheuer, U. T.; Schmid, R. D. J. Biotechnol. 1998, 61.
Analytical Chemistry, Vol. 75, No. 2, January 15, 2003 255
Sigma-Aldrich (Deisenhofen, Germany). Chiral amines and amides were a gift from Bayer AG (Monheim, Germany). Lipases and esterases were used as delivered by the suppliers: Chirazymes L2, L8, L9, L10, E2, E3, and E4 from Roche, (Penzberg, Germany); Lipase R, Lipase AYS, Lipase A Amano 6, Lipase AK Amano 20, Lipase D, Lipase AH, and Acylase from Amano (Nagoya, Japan); Novozyme SP535 from Novozymes (Copenhagen, Denmark); and PLE from Sigma-Aldrich. Recombinant enzymes (PFE I,15,16 BsubpNBE,17 BsteE and BsubE,18 and GCL19) were prepared by expression in Escherichia coli or Pichia pastoris, as described previously. For higher accuracy, all pipetting steps were performed with a Beckman 2000 pipetting robot (Beckman, Fullerton, CA). Fluorescence was determined using a BMG Fluostar microtiterplate reader (BMG, Offenburg, Germany), and absorbance measurement was performed with an Ultraspec 3000 photometer (AmershamPharmacia, Uppsala, Sweden). Amidase Kinetics. To 800 µL of phosphate buffer (50 mM, pH 7.5) in 1-mL cuvettes, 50 µL of amide (20 mM) and 50 µL of 4-nitro-7-chloro-benzo-2-oxa-1,3-diazole (NBD-Cl) (40 mM), both dissolved in DMSO, were added. Reactions were started by addition of 100 µL of enzyme solution in phosphate buffer. Conversions were followed by absorbance measurements at 475 nm every 2 min over a period of 1 h, and the slope dA/dt was calculated automatically. In parallel, the background adsorption was recorded under the same conditions without addition of substrate. Activity was calculated from difference in the slope between the biotransformation reaction and the background reaction using an extinction coefficient of ) 18 500 M-1 cm-1. Specific enzymatic activity was expressed on the basis of the protein content as determined using the bicinchoninic acid assay (BCA Kit, Pierce, Rockford, IL). Microtiterplate Assay. Into a 96-well microtiter plate (black, polystyrene), 180 µL/well of a 5 mM emulsion of the amide (phosphate buffer 50 mM; pH 7.5; 10% (v/v) DMSO; 2% gum arabic) were pipetted. After addition of 20 µL/well of cell lysate or hydrolase solution (in phosphate buffer, 50 mM; pH 7.5), the plates were tape-sealed and incubated in a shaking incubator at 37 °C, 200 rpm for 20-24 h. Afterward, 50 µL of a 20 mM NBDCl solution in 1-hexanol was added and the plates were incubated for an additional hour. Fluorescence (excitation, 485 nm; emission, 538 nm) was then determined with the microtiter plate fluorescence reader. Construction and Screening of Recombinant Enzyme Variant Libraries. This was essentielly performed as described before.15,17 Random mutagenesis was performed by error-prone PCR20 and DNA shuffling.21 Mutated esterase encoding genes were ligated into the expression vector and transformed into E. coli DH5R. Colonies were transferred into 384-well microtiterplates (15) Henke, E.; Bornscheuer, U. T. Biol. Chem. 1999, 380, 1029-1033. (16) Krebsfa¨nger, N.; Zocher, F.; Altenbuchner, J.; Bornscheuer, U. T. Enzyme Microb. Technol. 1998, 22, 641-646. (17) Henke, E.; Pleiss, J.; Bornscheuer, U. T. Angew. Chem., Int. Ed. 2002, 41, 3211-3213. (18) Henke, E.; Bornscheuer, U. T. Appl. Microbiol. Biotechnol. 2002, 60, 320326. (19) Catoni, E.; Schmidt-Dannert, C.; Brocca, S.; Schmid, R. D. Biotechnol. Tech. 1997, 11, 689-695. (20) Vartanian, J. P.; Henry, M.; Wain-Hobson, S. Nucleic Acids Res. 1996, 24, 2627-2631. (21) Stemmer, W. P. C. Nature 1994, 370, 389-390.
256
Analytical Chemistry, Vol. 75, No. 2, January 15, 2003
using a picking robot (Biorobotics, Cambridge, U.K.) and grown in LB-Amp media supplemented with 5% (v/v) DMSO. These plates were used as stocks, enabling direct storage at -80 °C. Expression of esterase variants was performed by replica plating in 96-well microtiterplates containing LB-amp broth supplemented with L-rhamnose (2 mg/mL) to induce expression, followed by incubation at 37 °C for 20 h. Cells were disrupted by three freeze/ thaw cycles. The cell lysate in the wells was directly used for activity determination in the microtiterplate assay described above. Kinetic Resolution of p-Chloro-1-(phenylethyl)-methoxyacetamide (1). A 100-µL portion of the amide solution in DMSO (100 mM) was added to 900 µL of sodium phosphate buffer (50 mM; pH7.5). After addition of 10 mg of crude extract of a hydrolase, the mixture was incubated for 20 h at 40 °C in a thermoshaker. Afterward, substrates and products were extracted with 1 mL of diethyl ether. Conversion and enantiomeric excess of substrate (%eeS) or product (%eeP) were determined by gas chromatography using a chiral column (cyclodextrine-β-3P (Macherey-Nagel, Du¨ren, Germany), carrier: H2, 40 kPa; split 1:100, 120 °C, 25 min; 10 °C/min; 160 °C; 2 °C/min; 200 °C, 8 min). Enantioselectivity was calculated according to Chen et al.22 RESULTS AND DISCUSSION An assay suitable for high-throughput screening of large libraries of (recombinant) enzymes sets some requirements for the detection reaction. The assay should not involve an elaborate purification of the enzymes, especially if the biocatalyst is directly produced by cultivation in the microtiter plate (MTP). For the determination of amidase activity, a highly specific detection method for amines is necessary, because nondesired side reactions with amino acids, peptides, and proteins must be excluded. Furthermore, the detection reaction must be carried out in aqueous solution or at least in a biphasic system with a waterimmiscible organic solvent to avoid tedious and time-consuming extraction steps. Finally, the derivatization must occur at temperatures compatible with the enzymatic reaction. We selected 4-nitro-7-chloro-benzo-2-oxa-1,3-diazole (NBD-Cl, Figure 1) as derivatization reagent. Benzooxadiazole derivatives were previously used as fluorescence tagging reagents for amines,23,24 for example, to enhance enantiomeric separation in chromatography,25,26 but also for alcohols, phenols, thiols,27,28 and carboxylic acids.29 In contrast to the derivatization of weaker nucleophiles, the reaction of NBD-Cl with amines can be performed in aqueous solution at room temperature. The yellow dye formed can be detected directly spectrophotometrically (λmax ) 475 nm, ) 18 500 M-1 cm-1 in phosphate buffer pH 7.5, 10% (22) Chen, C.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am. Chem. Soc. 1982, 104, 7294-7299. (23) Ghosh, P. B.; Whitehouse, M. W. Biochem. J. 1968, 108. (24) Imai, K.; Uzu, S.; Kanda, S. Anal.Chim. Acta 1994, 290, 3-20. (25) Guo, X.; Fukushima, T.; Li, F.; Imai, K. Analyst 2002, 127, 480-484. (26) al-Kindy, S.; Santa, T.; Fukushima, T.; Homma, H.; Imai, K. Biomed. Chromatogr. 1998, 12, 276-280. (27) Imai, K.; Fukushima, T.; Yokosu, H. Biomed. Chromatogr. 1994, 8, 107113. (28) Birkett, D. J.; Price, N. C.; Radda, G. K.; Salmon, A. G. FEBS Lett. 1970, 6, 346-348. (29) Toyo’oka, T.; Ishibashi, M.; Takeda, Y.; Nakashima, K.; Akiyama, S.; Uzu, S.; Imai, K. J. Chromatogr. 1991, 588, 61-71.
Figure 1. Derivatization of amines with 4-nitro-7-chloro-benzo-2-oxa-1,3-diazole (NBD-Cl).
(v/v) DMSO for (R,S)-2-phenylethylamine). Thus, the direct determination of enzyme kinetics is possible. Higher sensitivity can be reached using the fact that the dye is highly fluorescent in organic solvents (excitation 480 nm; emission 538 nm). The hydrolytic enzymatic reaction can be performed in the media of choice by just adding the hydrolase to the substrate in a buffered solution or as a suspension. After an appropriate reaction time, a solution of NBD-Cl in a waterimmiscible solvent is added, followed by further incubation to allow for reaction with the amine and extraction into the organic phase. In addition, the reaction is not affected by organic cosolvents, such as DMSO, or emulsifying agents, such as gum arabic. To verify if the assay system is suitable for quantification of the released amine under the test conditions, a calibration curve was recorded (data not shown). A linear range between 40 and 200 µM (R,S)-1-phenylethylamine could be observed, where the amine concentration complies with the fluorescence signal. In this range, 1 µM derivated (R,S)-1-phenylethylamine corresponded to 480 relative fluorescence units (rfu, amplification set to 11 on a Fluostar, BMG, Offenburg, Germany) with a standard deviation 150) and pig liver esterase (E ) 28) showed excellent respectively good enantioselectivity; the other esterases were almost nonselective. 258
Analytical Chemistry, Vol. 75, No. 2, January 15, 2003
CONCLUSIONS This novel assay enables the fast and accurate screening of large enzyme libraries and commercial preparations on hydrolytic activity toward amides. The great advantage of the assay is that
Table 1. Hydrolytic Activity of Esterases toward r-Substituted Aryl- and Alkyl Ethyl Amides 1-19 Using the NBD-Cl Assaya substrate
Chirazyme E3
Chirazyme E4
PLE
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
9585
26706
42693
40259
9902 62953 37576
49538
1838
rec. BsubE
rec. BsteE
rec. BsubpNBE 41405
8732
48859
38364
5670 6866
12247
10537
7377 15314 7690
19284
5622
8126
rec. PFE I
36859
a All reactions were performed in 96 well MTPs as described in the Experimental Section. For clearity only relative fluorescence units (rfu) values for active enzyme/substrate combinations are given.
Table 2. Enantioselectivity of Five Hydrolases in the Kinetic Resolution of 1 as Determined by GC Analysis
enzyme
activity mU/mg protein
rec. BsubpNBE PLE Chirazyme E3 Chirazyme E4 Chirazyme L2, lyo. (CAL-B)
58 64 142 273 860
a
enantiomeric excess %eeS %eeP 19 29 11 14 >95
42 91 60 12 >99
conversion [%]
E
enantiopreference
31 24 16 53 50
3 28 4 1 .150
S R S R R
Reaction time 20 h at 40 °C.
the hydrolysis itself can be performed under conditions that are optimal for enzyme activity. Furthermore, no protein purification steps or separations from culture medium or cells is required. Thus, a direct assaying of enzyme-producing microorganisms in a microtiter plate format is possible. Furthermore, the assay has no limitation regarding the substrates, facilitating the direct screening using the compound of interest. Because the only substances needed for the detection reaction are NBD-Cl and a water-immiscible solvent (e.g., 1-hexanol), the test system is also very economical. The sensitivity of the assay under the described conditions is sufficient for most applications. For more difficult substrates or when an even lower detection limit is required, the more reactive fluoro derivative (NBD-F) can be applied; however, the much higher price (∼800 times more expensive than NBDCl) makes this substance less applicable for use in a highthroughput environment, despite the reported up to 500 times reactivity as compared to NBD-Cl.31 Only little background was observed, which led us to assume that autohydrolysis of NBD-Cl to 7-hydroxy-4-nitro-benzooxadiazole is neglectible, but also the known shift to shorter wavelengths in the absorption and fluorescence spectra of the hydrolysis product compared to the amino derivatives might have made this assay more reliable.28 The observed low disturbance by cellular (31) Miyano, H.; Toyo’Oka, T.; Imai, K. Anal.Chim. Acta 1985, 170, 81-87.
amines and amino acids and proteins might be due to the better solubility of the hydrophobic substrate amines in organic solvents as the derivatization reaction occurs in the organic layer. Thus, the use of the less reactive NBD-Cl instead of NBD-F and the resulting higher selectivity might also be an advantage. One limitation is the disadvantage that the direct determination of enzyme kinetics by fluorescence measurement is possible only for enzymes that tolerate high concentrations of water-immiscible solvents in a two-phase system. At least, the format allows the rapid identification of active biocatalysts, and enzyme kinetics can be determined spectrophotometrically. In principle, the determination of the (apparent) enantioselectivity by using optically pure (R)- and (S)-amides in two parallel assay reactions should also be possible. Using this newly established assay, further insights into the reactivity of carboxylester hydrolases toward amides could be achieved. Surprisingly, a high rate of the tested carboxylester hydrolases showed activity toward amides, which is in contrast to the literature in this field. This indicates that amidase activity is a more common feature of carboxylester hydrolases than assumed until now. Thus, a detailed activity study of already identified carboxylester hydrolases seems to be a much more promising way to detect enzymes with interesting features than the generation of new activity via directed evolution, as we had Analytical Chemistry, Vol. 75, No. 2, January 15, 2003
259
tried here. With several thousand sequences of esterase and putative esterases in today’s databases (most of them are not yet characterized), there is a high chance for researchers to find new applicable biocatalysts. Especially interesting is the fact that esterases seem to be much more suitable for the hydrolytic cleavage of amides than the interfacially activated lipases. That activity depends much more on the substrates’ structures than in the hydrolysis of carboxylesters is in accordance with previous results.14 We found that the caproyl derivative of the phenyl alkylamines showed the best activity with most hydrolases. The observation that activity depends more on the polarity of the substrate and on the chain length of the leaving acid moiety rather than on sterical demands provides researchers with additional
260
Analytical Chemistry, Vol. 75, No. 2, January 15, 2003
guidelines for the further improvement of applications by substrate engineering. ACKNOWLEDGMENT The authors thank Prof. Dr. R. D. Schmid (Institute of Technical Biochemistry, University of Stuttgart) for helpful discussion. This work was financed by the German Research Foundation (DFG, Bonn, Germany, Grant Bo 1475/2-1).
Received for review June 17, 2002. Accepted November 12, 2002. AC0258610
