New application of firefly luciferase − it can catalyze the enantioselective thioester formation of 2-arylpropanoic acid

Authors


D.-I. Kato, Graduate School of Engineering, University of Hyogo, 2167 Shosha, Himeji, Hyogo 671-2201, Japan
Fax: +81 79 267 4891
Tel: +81 79 267 4969
E-mail: kato@eng.u-hyogo.ac.jp

Abstract

We introduce a new application of firefly luciferase (EC 1.13.12.7). The firefly luciferases belong to a large superfamily that includes rat liver long-chain acyl-CoA synthetase (LACS1). LACS1 is the enzyme that is involved in the deracemization process of 2-arylpropanoic acid and catalyzes the enantioselective thioester formation of R-acids. Based on the similarity of the reaction mechanisms and the sequences between firefly luciferase and LACS1, we predicted that firefly luciferase also has thioesterification activity toward 2-arylpropanoic acid. From an investigation using three kinds of luciferases from North American firefly and Japanese fireflies, we have confirmed that these luciferases exhibit an enantioselective thioester formation activity and the R-form is transformed to a thioester in preference to the S-form in the presence of ATP, Mg2+, and CoASH. The enantiomeric excesses of unreacted recovered acid and thioester were determined by chiral phase HPLC analysis and the resulting 2-arylpropanoyl-CoAs were identified by high resolution mass spectroscopy. The Km and kcat values of thermostable luciferase from Luciola lateralis (LUC-H) toward ketoprofen were determined as 0.22 mm and 0.11 s−1, respectively. The affinity of ketoprofen was almost the same of d-luciferin. In addition, the calculated E-value toward ketoprofen was approximately 20. These results suggest that LUC-H could catalyze the kinetic resolution of 2-arylpropanoic acid efficiently and would be a new option for the preparation of optically active 2-substituted carboxylic acids.

Abbreviations
ACS

acyl-CoA synthetase

ee

enantiomeric excess

fenoprofen

2-(3-phenoxyphenyl)propanoic acid

flurbiprofen

2-(3-fluoro-4-phenyl-phenyl)propanoic acid

ibuprofen

2-(4-isobutylphenyl)propanoic acid

ketoprofen

2-(3-benzoylphenyl)propanoic acid

LACS1

long-chain acyl-CoA synthetase

LUC-H

thermostable luciferase from Luciola lateralis

naproxen

2-(6-methoxy-2-naphthyl)propanoic acid

tropic acid

2-hydroxymethylpropanoic acid

Firefly luciferase is a well-known enzyme that concerns in the bioluminescence reaction. It catalyzes the oxidation of firefly luciferin with molecular oxygen in the presence of ATP and Mg2+, resulting in luminescence [1–3]. The stereoselectivity of this bioluminescent reaction was investigated in detail by Seliger et al. [4], who found that the d-form is the specific substrate for the light emission reaction whereas the l-form is not used for the light-producing reaction. The bioluminescence reaction of firefly luciferase is composed of two reaction steps. The first step is the activation of the carboxyl group to form luciferyladenylate and the second is the light emission reaction via the oxidation of this intermediate. The substrate activation mechanism in the initial step is commonly observed in adenylate-forming enzymes, such as aminoacyl-tRNA synthetases, acyl-CoA synthetases (ACSs), and nonribosomal peptide synthetases(Fig. 1) [5]. In comparison with the amino acid sequences of firefly luciferases and these enzymes, it has become apparent that luciferase shares the most significant similarities with ACSs. In addition, the crystal structures of Photinus pyralis and Luciola cruciata luciferases, which were determined by Conti et al. [6] and Nakatsu et al. [7], respectively, were confirmed to have the same framework as that of ACSs. These results indicate that firefly luciferase evolved from the same ancestral enzymes as ACSs and acquired a luminescent system in the course of evolution. The supporting evidence of this hypothesis was reported by Airth et al. [8]. They showed the bifunctionality of firefly luciferase (i.e. it could catalyze the thioester formation of dehydroluciferin as well as a bioluminescent ability). More recently, Oba et al. [9] found that firefly luciferase also has the ability to form thioesters of long-chain fatty acids. This report indicated that firefly luciferase also has an acyl-CoA synthetase activity toward an unnatural substrate. In addition, Nakamura et al. [10] showed that luciferase exhibits bimodal action depending on the chirality of luciferin, in that it transforms l-luciferin into luciferyl-CoA whereas it oxidizes the d-form to oxyluciferin.

Figure 1.

 Two-step mechanism of adenylate-forming enzymes, light emission reaction of firefly luciferase and thioesterification reaction of acyl-CoA synthetase. The initial step of these enzymatic reactions is the activation of the carboxyl group to form acyl-adenylate intermediate. In the case of thioesterification reaction, the acyl group is then transferred to the thiol group of CoASH to form a thioester (acyl-CoA), whereas light emission reaction was achieved by a multistep oxidation with molecular oxygen of this intermediate followed by the production of oxyluciferin.

In 1990, Suzuki et al. [11] reported that P. pyralis luciferase showed a significant sequence similarity with rat liver long-chain acyl-CoA synthetase (LACS1). LACS1 is the enzyme that is involved in the deracemization process of 2-arylpropanoic acid [12,13]. Deracemization is a reaction that inverts the chirality of either enantiomer of a racemate to the other antipode, resulting in an optically active compound starting from a racemic mixture. In the case of rat liver, the deracemization process is realized by three enzymatic reactions, such as thioesterification, epimerization, and hydrolysis of the thioester [14]. In these processes, LACS1 catalyzes the thioesterification reaction and this reaction is the only step that proceeds in an enantioselective manner [15,16]. Thus, the enantiomeric ratio in the reaction mixture gradually shifts to one enantiomer with the progression of the reaction.

The absolute configuration of compounds sometimes has a strong influence on the biological activity. For example, 2-arylpropanoic acid constitutes an important group of physiologically active compounds, because both enantiomers display different activities in vivo. The S enantiomers are generally more important because they are active as nonsteroidal anti-inflammatory drugs [17]. On the other hand, the R enantiomer of flurbiprofen has recently been paid much attention because of its anticancer activity as well as a potent reducer of beta amyloid, the main constituent of amyloid plaques in Alzheimer's disease [18]. Thus, the preparation of optically pure enantiomers is extremely important and many kinds of enzymatic approaches have been tried so far [19–21].

In the present study, we examined whether firefly luciferase has thioesterification activity that can enantiodifferentiate unnatural substrates, such as 2-arylpropanoic acid. Based on the similarity of the reaction mechanisms and the sequences between firefly luciferase and LACS1, we predicted that firefly luciferase also has thioesterification activity toward 2-arylpropanoic acid. From the investigation using three kinds of luciferases from the North American firefly and Japanese fireflies, we have confirmed that these luciferases exhibit an enantioselective thioesterification activity and the R-form is transformed to a thioester in preference to the S-form in the presence of ATP, Mg2+, and CoASH.

Results and Discussion

Construction of a simple assay system

To confirm whether luciferase could truly catalyze the enantioselective thioesterification reaction of 2-arylpropanoic acid, racemic 2-phenylpropanoic acid, which has the simplest structure of 2-arylpropanoic acids, was selected as a substrate and the enantiomeric excess (ee) of the unreacted acid and thioester were measured using a chiral phase HPLC column (ee exhibits how excessive the amount of one enantiomer is compared to the other in an optically active compound; ee value can be determined by the equation: ee (%) = |R − S|/(R + S) × 100, where R and S are the respective moles of enantiomers in a compound). The produced thioester, however, could not be applied to the chiral phase HPLC column because of its polarity. Thus, the thioester and unreacted acid had to be separated before the analysis. To separate these two compounds, a simple protocol was constructed. When the thioester was produced by the reaction, a large hydrophilic group was introduced into the molecule as compared with a hydrophobic benzene ring of the substrate. When the reaction was stopped by hydrochloric acid, the carboxylate would be protonated to form a free acid, which would be easily extracted into an organic layer. On the other hand, the thioester would remain in the aqueous phase because of the hydrophilicity of its structure. Thus, the unreacted acid and the produced thioester can be separated by a simple extraction procedure. In addition, if luciferase has an ability to distinguish the absolute configuration of the starting material, the ee of the compounds in the two layers will increase.

After the above mentioned separation procedure, the ee of the materials in the organic layer was measured. For this experiment, three kinds of luciferases, P. pylalis luciferase, L. cruciata luciferase, and thermostable luciferase from Luciola lateralis (LUC-H), were selected. LUC-H is a mutant enzyme of L. lateralis luciferase, with improved thermostability and resistance to a kind of surfactant as compared with the wild-type [22] and it has been used for hygiene monitoring and biomass assays based on the ATP-bioluminescence method [23,24]. In the presence of ATP, Mg2+, and CoASH, the optical purity of the recovered acid increased gradually to the S-form (Table 1, entries 1–3). In particular, the ee value exceeded over 35% when the luciferases from Japanese fireflies, L. cruciata and LUC-H, were used. These results demonstrate that the luciferases of the Japanese fireflies can differentiate the chirality of 2-phenylpropanoic acid, transforming the R-form to a thiol ester with CoASH. Photinus pyralis firefly, however, exhibits little enantioselectivity toward this substrate. This result is of some interest because these three enzymes exhibit significant amino acid sequence similarity with each other. As LUC-H exhibited the highest enantioselectivity toward 2-phenylpropanoic acid, further investigations were mainly performed using this enzyme.

Table 1.   Enantioselectivity of firefly luciferase catalyzed thioesterification reaction of 2-arylpropanoic acid. The substrate was incubated at 30 °C for 60 min with LUC-H in the presence of ATP, CoASH and Mg2+. The ee of the recovered acid was determined by chiral phase HPLC analysis; the ee of the thioester was determined by chiral phase HPLC analysis after the hydrolysis of thioester linkage to form the corresponding acid. ND, not detected.
EntrySubstrateOrigin of luciferaseRecovered acid (% ee)Thioester (% ee)
 12-phenyIpropanoic acidPhotinus pyralis4.6 (S)6.2 (R)
 22-phenyIpropanoic acidLuciola cruciata35.2 (S)32.2 (R)
 32-phenyIpropanoic acidLUC-H42.4 (S)67.3 (R)
 4FlurbiprofenLUC-H92.4 (S)51.9 (R)
 6IburprofenLUC-H96.1 (S)55.1 (R)
 6KetoprofenPhotinus pyralis44.7 (S)32.5 (R)
 7KetoprofenLuciola cruciata98.0 (S)52.0 (R)
 8KetoprofenLUC-H99.6 (S)44.6 (R)
 9NaproxenLUC-H99.6 (S)69.5 (R)
102-phenyIbutanoic acidLUC-H57.6 (S)36.0 (R)
11Tropic acidLUC-H8.3 (R)> 99.9 (S)
122-phenylpentanoic acidLUC-HRacemateND
132-phenyl-3-methylbutanoic acidLUC-HRacemateND
142-(4-chlorophenoxy)propanoic acidLUC-HRacemateND
152-methyl-3-phenylpropanoic acidLUC-HRacemateND

LUC-H catalyzed kinetic resolution of 2-arylpropanoic acid

To expand the applicability of the present reaction, other 2-arylpropanoic acids, such as flurbiprofen, ibuprofen, ketoprofen, and naproxen, were selected and investigated. These acids constitute an important group of physiologically active compounds. Although the enantiomers of these compounds have been already prepared via lipase catalyzed kinetic resolution and recrystallization of diastereomeric mixtures, the yield and the ee of the products are not sufficiently high [25]. In case of LUC-H catalyzed thioester formation, however, its enantioselectivity toward these profen compounds was significantly high and the ee of the recovered acids were over 90% ee (Table 1, entries 4, 5, 8 and 9). The absolute configuration of the recovered acids was also S, similar to the case of 2-phenylpropanoic acid.

Because the ee of the thioesters could not be determined as they existed, the thioester bond was hydrolyzed and the ee of the resulting free acid was measured. The two kinds of hydrolytic conditions, chemical hydrolysis under high pH and enzymatic hydrolysis under neutral pH, were examined, because there was some fear of racemization by the chemical hydrolysis. The pKa of the α-methine proton of the 2-arylpropanoyl-CoA is approximately 10.3 [14] and the pH in the chemically hydrolyzed reaction mixture was > 10. The produced acid was extracted by an organic solvent and its ee was determined by chiral phase HPLC analysis. Contrary to our expectations, the ee values were almost the same regardless of the conditions (data not shown). Based on these results, both hydrolytic methods are applicable to give free acid from the thioester. In addition, the enatiomer ratios of these compounds were partial to the R-form. These results give support to the results of the HPLC analysis of unreacted acids, which indicate that the thioester formation reaction proceeds in the R-enantioselective manner (Fig. 2).

Figure 2.

 Luciferase catalyzed enantioselective thioesterification of 2-arylpropanoic acid. LUC-H exhibits an enantioselective thioesterification activity and the R-form is transformed to a thioester in preference to the S-form in the presence of ATP, Mg2+, and CoASH.

Substrate specificity of LUC-H catalyzed thioester formation

To investigate the substrate tolerance of LUC-H catalyzed thioester formation, the other substrates were examined (Table 1, entries 10–15). The enantioselective thioester formation reactions also occurred when the α-methyl substituent of 2-phenylpropanoic acid was replaced with an ethyl or hydroxymethyl group. LUC-H could differentiate the chirality of 2-phenylbutanoic acid and the ee value of unreacted acid was almost the same as the case of 2-phenylpropanoic acid. In the case of 2-hydroxymethylpropanoic acid (tropic acid), which has a hydroxyl group on the methyl group of 2-phenylpropanoic acid, the thioester was recovered in enantiomerically pure form, whereas the reaction rate dramatically decreased. In both cases, enantioselectivity was almost the same. Unfortunately, however, bulkier substitutions could not be used by the enzyme; 2-phenylpentanoic acid and 2-phenyl-3-methylbutanoic acid were only recovered as racemates and the no product could be detected in the aqueous phase (Table 1, entries 12 and 13). In addition, the reaction did not proceed by the introduction of a spacer atom between the chiral center and the benzene ring (Table 1, entries 14 and 15). At present, we have no idea why LUC-H is not able to use these compounds. Besides how to distinguish the absolute configuration of substrates, these results were interesting from the view point of the molecular recognition of this enzyme. The 3D structural analysis may shed light on the answers of these questions.

In the case of recombinant rat LACS1, the substrate specificity was not investigated in detail and only three kinds of substrate, ibuprofen, fenoprofen, and flurbiprofen, were examined [16]. With these substrates, flurbiprofen could not be used by LACS1. In the case of LUC-H, however, a series of 2-arylpropanoic acids, including flurbiprofen, were converted to the corresponding thioesters with good enantioselectivity and other derivatives were also converted. These results suggested that LUC-H has a greater potential for enantioselective thioester formation.

Purification and identification of LUC-H produced 2-arylpropanoyl-CoA

The solid-phase extraction method was used to purify the thioesters. Purified compounds were subjected to TLC analysis and the Rf values of each compound were in good agreement with the chemically synthesized authentic samples. These were also identified by the ESI-MS method based on orthogonal TOF-MS. The mass spectra of 2-arylpropanoyl-CoAs were dominated by ions that agreed with the calculated mass of singly charged ions of the corresponding thioester. The results are summarized in Table 2. The detected masses were within the limit of the accuracy of the instrument [26]. These results suggested that the purified compounds were definitely 2-arylpropanoyl-CoAs and that LUC-H is able to catalyze the enantioselevtive thioesterification reaction of 2-arylpropanoic acid.

Table 2.   Identification of 2-arylpropanoyl-CoA produced by LUC-H with TOF-MS analysis. For the detection of 2-arylpropanoyl-CoAs by TOF-MS analysis, see the Experimental procedures section.
EntryThioesterPredicted [M-H]Observed [M-H]Error (p.p.m.)
1Flurbiprofenyl-CoA992.1868992.1832−3.63
2Ibuprofenyl-CoA954.2275954.22881.40
3Ketoprofenyl-CoA1002.19101002.19190.92
4Naproxenyl-CoA978.1011978.19352.43

Cofactor requirement experiments were also performed using ketoprofen as the substrate (Table 3). Based on the ee of the unreacted acids, a dramatical decrease was observed in the absence of ATP and/or CoASH. In addition, no peak of thioester product was detected in the aqueous layer. These results suggested that these cofactors were essential for the reaction and 2-arylpropanoyl-CoA was produced by way of an acyl-adenylate intermediate.

Table 3.   Cofactor requirements analysis for ketoprofenyl-CoA formation by LUC-H. The substrate was incubated at 30 °C for 60 min with LUC-H. The ee of the recovered acid was determined by chiral phase HPLC analysis; the ee of the thioester was determined by chiral phase HPLC analysis after the hydrolysis of thioester linkage to form the corresponding acid. Relative activity was calculated on the basis of the ee of recovered acid. ND, not detected.
EntryCo-factorRecovered acid (% ee)Thioester (% ee)Relative activity (%)
ATPCoASH
1++98.480.9100
2+0.98ND1.0
3+0.55ND0.6
40.62ND0.6

Kinetic analysis of LUC-H catalyzed thioester formation

The kinetic study of the formation of ketoprofenyl-CoA catalyzed by LUC-H was examined to obtain the detailed reaction information. The efficiency of thioester formation was calculated from the peak areas of the remaining acid and flurbiprofen as an internal standard using reversed phase HPLC analysis. According to the Lineweaver–Burk plot, the Km and Vmax values for racemic ketoprofen were 0.22 mm and 110.3 nmol· min−1·mg−1 protein, respectively (Fig. 3). The kcat value was calculated to be 0.11 s−1. Because the Km value was similar to that of d-luciferin for the bioluminescence reaction (Km = 0.15 mm) (the data are reported in the product data sheet of LUC-H, which was provided by Kikkoman corporation), ketoprofen and d-luciferin could be bound in the active site at the same magnitude of affinity. The Michaelis–Menten parameters of recombinant rat LACS1 for the thioester formation were reported by Sevoz et al. [16] and the values of Km and Vmax for (R)-ibuprofen and (R)-fenoprofen were 1.7 mm, 353 nmol·min−1·mg−1 protein and 0.10 mm, 267 nmol·min−1·mg−1 protein, respectively. From these data, it seems reasonable to assume that the specificity of LUC-H and rat LACS1 toward 2-arylpropanoic acid would be almost the same and hence LUC-H could catalyze the efficient thioester formation as well as recombinant rat LACS1.

Figure 3.

 Kinetic study of ketoprofenyl-CoA synthetic activity on LUC-H. According to the Lineweaver–Burk plot, the Km, Vmax and kcat values for racemic ketoprofen were 0.22 mm, 110.3 nmol· min−1·mg−1 protein, and 0.11 s−1, respectively.

E-value calculation of LUC-H catalyzed thioester formation

The stereoselectivity of recombinant rat LACS1 was very high and the complete separation of the absolute configuration of the substrates was carried out [16]. In the case of LUC-H, however, the enantioselectivity was not perfect and the S-form was also converted to a thioester (data not shown). To confirm the enantioselective ratio of LUC-H catalyzed thioesterification of ketoprofen, the E-value was calculated. The E-value was used to evaluate the enantioselectivity of enzymatic kinetic resolution and was determined by using the ee and yield of the unreacted acid according to the method described by Chen et al. [27]. The calculated E-values in three conversion points are summarized in Table 4. There are little changes for each conversion point. This result exhibits that the produced thioester does not affect the enantioselectivity of the enzyme and the R-acid would be converted to a thioester approximately 20 times faster than the S-acid.

Table 4.   Determination of enantioselectivity (E-value) for ketoprofen by LUC-H catalyzed thioester formation. The substrate was incubated at 30 °C for 60 min with LUC-H. Conversion was calculated from the yield of recovered acid. Yield of the recovered acid was determined by the reversed phase HPLC analysis in the presence of flurbiprofen as an internal standard. ee of the recovered acid was determined by chiral phase HPLC analysis. E-value was calculated based on the conversion (c) and ee of the recovered acid (ee(s)) according to the equation: E = ln[(1 − c)(1 − ee(s))]/ln[(1 − c)(1 + ee(s))][27].
EntryConversion (%)Recovered acid (% yield)ee (%)E -value
124.775.328.820
239.061.053.519
361.838.299.624

pH profile of LUC-H catalyzed thioester formation

The activity-pH profile of the LUC-H catalyzed thioesterification reaction of ketoprofen over a wide range is shown in Fig. 4. The enzyme was highly active in the alkaline pH region with the maximum activity at pH 9–10. In the pH range over 10, the apparent enzymatic activities were reduced markedly. This is because of the spontaneous hydrolysis of the thioester linkage. In comparison with the bioluminescence reaction, the enzyme kept its activity for thioester formation in a wider range of pH and the optimal shifted to the basic region.

Figure 4.

 pH activity profile of LUC-H catalyzed thioester formation (○) and bioluminescent reaction (•). The activity of thioester formation was determined by the comparison with the initial velocity of the reaction.

Thioesterification acitivity of L. cruciata and P. pyralis luciferases

From the above experiments (Table 1, entries 1–3), it was known that L. cruciata and P. pyralis luciferases also have an ability to catalyze the enantioselective thioesterification of 2-arylpropanoic acid, although with selectivity inferior to that of LUC-H. To disclose whether these luciferases have an ability to catalyze the enantioselective thioester formation toward other compounds, these enzymes were analyzed. Ketoprofen was used as a substrate. Fortunately, those two luciferases exhibited the enantioselective thioester formation activity (Table 1, entries 6 and 7). The enantiopreference of these enzymes was the R-form, similarly to the case of LUC-H. The reaction efficiency of L. cruciata luciferase is the same level as LUC-H. In addition, P. pyralis luciferase could also differentiate between the enantiomers of the substrate. The produced thioester was also isolated by the solid-phase extraction method and the mass value of the purified thioester was consistent with the calculated value of ketoprofenyl-CoA. In the case of P. pyralis luciferase, the E-value was calculated as 16. This result suggested that the enantioselectivity of P. pyralis luciferase toward ketoprofen was the same as that of LUC-H, although the selectivity toward 2-phenylpropanoic acid was quite different from each other.

Conclusions

We found a unique method for the preparation of optically active 2-arylpropanoic acid using a bioluminescence enzyme. Based on the similarity of the reaction mechanisms and the sequences between firefly luciferase and rat LACS1, we predicted that firefly luciferase would have an enzymatic activity of enantioselevtive thioester formation. The three kinds of luciferases from the North American firefly and Japanese fireflies exhibit an enantioselective thioesterification activity toward 2-arylpropanoic acid and the R-form is transformed to a thioester in preference to the S-form in the presence of ATP, Mg2+, and CoASH. The production of 2-arylpropanoyl-CoAs was confirmed by the high accuracy exact mass measurements by TOF-MS analysis. Thermostable luciferase from L. lateralis (LUC-H) exhibited thioesterification activity in high efficiency and the ee of unreacted recovered acid and thioester were relatively high. The Km and kcat values toward racemic ketoprofen were determined as 0.22 mm and 0.11 s−1, respectively, and the E-value for this substrate was calculated to be 20. These results suggested that LUC-H can be used for the kinetic resolution of 2-arylpropanoic acid and will be the new choice for the preparation of optically active 2-substituted carboxylic acids.

Experimental procedures

General methods

All chemicals were commercially available and used without further purification unless otherwise noted. Analytical TLC was developed on E. Merck Silica Gel 60 F256 plates (0.25 mm thickness). Firefly luciferases from P. pyralis and recombinant luciferase from L. cruciata were purchased from Sigma (St Louis, MO, USA) and Wako (Osaka, Japan), respectively. Thermostable luciferase from L. lateralis (LUC-H) was thankfully gifted from Kikkoman Corporation (Tokyo, Japan). HPLC analysis was carried out with a Shimazu liquid chromatograph (LC-20AT) or CO-8020 (Toso Co., Ltd, Tokyo, Japan). Mass spectra were recorded on a Waters Flight Mass Spectrometer LCT-Premier (Waters Corp., Milford, MA, USA). Centrifugation was carried out by high speed refrigerated microcentrifuges (TOMY, Tokyo, Japan) with TMP-11 rotor. Protein concentration was determined by the method of Bradford [28] with the protein assay kit (Bio-Rad Laboratories, Inc., Hercules, CA, USA) and BSA was used as the standard. Authentic samples of 2-arylpropanoyl-CoA were chemically synthesized according to the described method [29]. The tesB gene of Escherichia coli JM109 was cloned into the NdeI and XhoI sites of pET23C(+) vector by the standard procedure and the resulting plasmid, pTesB, was transformed into E. coli BL21(DE3). The induction of tesB protein expression was performed by the addition of 0.5 mm IPTG at 25 °C in Luria–Bertani medium containing 50 mg·mL−1 of ampicillin.

Luciferase catalyzed kinetic resolution of 2-arylpropanoic acid

To 100 mm potassium phosphate buffer (pH 7.0), racemic 2-arylpropanoic acid (0.25 mm), ATP (10 mm), CoASH (2 mm), and MgCl2 (10 mm) were added. The final volume of the reaction mixture was adjusted to 500 µL. The mixture was preincubated at 30 °C for 10 min. The reaction was started by the addition of the appropriate amount of firefly luciferase (10–1000 µg protein) and incubated at 30 °C for 60 min. After adding 100 µL of 2 m HCl and 1 mL of diethyl ether to the reaction, the mixture was shaken and centrifuged at 17 610 g for 10 min. The separated organic layer was concentrated in vacuo and the ee was analyzed by HPLC with a Daicel Chiral column (Daicel Chemical Industries Ltd, Tokyo, Japan) using a mixture of hexane/2-propanol/trifluoroacetic acid in an appropriate proportion as the eluent.

The ee of the thioester in the aqueous layer was determined after hydrolysis of the thioester linkage. Two kinds of hydrolytic conditions were examined: chemical hydrolysis, using 120 µL of 2 m NaOH incubation at room temperature for 30 min, and enzymatic hydrolysis, using residual aqueous layer neutralized to pH 7 by addition of 2 m NaOH and 100 µL of cell free extract of tesB overexpressed E. coli and incubation at 30 °C for 30 min. After the hydrolysis, the reaction was quenched by adding 120 µL of 2 m HCl and 1000 µL of diethyl ether. The mixture was shaken and centrifuged at 17 610 g for 10 min. The separated organic layer was concentrated in vacuo and the ee was analyzed by HPLC with a Daicel Chiral column.

The conditions and results of HPLC analysis were: for 2-phenylpropanoic acid, Daicel Chiralcel OD-H column (hexane/2-propanol/trifluoroacetic acid = 100/1/0.1%, 1.0 mL·min−1, 210 nm), tR 26.5 min (R-form) and 32.5 min (S-form); for flurbiprofen, Daicel Chiralcel OD-H column (hexane/2-propanol/trifluoroacetic acid = 100/1/0.1%, 1.0 mL·min−1, 210 nm), tR 31.9 min (R-form) and 37.1 min (S-form); for ibuprofen, Daicel Chiralcel OD-H column (hexane/2-propanol/trifluoroacetic acid = 100/1/0.1%, 1.0 mL·min−1, 210 nm), tR 15.7 min (R-form) and 18.3 min (S-form); for ketoprofen, Daicel Chiralcel OJ-H column (hexane/2-propanol/trifluoroacetic acid = 95/5/0.1%, 1.0 mL·min−1, 254 nm), tR 39.7 min (R-form) and 50.2 min (S-form); for naproxen: Daicel Chiralcel OD H column (hexane/2-propanol/trifluoroacetic acid = 95/5/0.1%, 0.5 mL·min−1, 254 nm), tR 23.3 min (R-form) and 27.4 min (S-form); for 2-phenylbutanoic acid, Daicel Chiralcel OD-H column (hexane/2-propanol/trifluoroacetic acid = 100/1/0.1%, 1.0 mL·min−1, 210 nm), tR 20.0 min (R-form) and 28.4 min (S-form); for tropic acid, Daicel Chiralcel OD-H column (hexane/2-propanol/trifluoroacetic acid = 95/5/0.1%, 1.0 mL·min−1, 254 nm), tR 20.9 min (S-form) and 25.9 min (R-form); for 2-phenylpentanoic acid: Daicel Chiralcel OD-H column (hexane/2-propanol/trifluoroacetic acid = 100/1/0.1%, 1.0 mL·min−1, 210 nm), tR 17.9 min (R-form) and 23.6 min (S-form); for 2-phenyl-3-methylbutanoic acid: Daicel Chiralcel OD-H column (hexane/2-propanol/trifluoroacetic acid = 100/1/0.1%, 1.0 mL·min−1, 210 nm), tR 16.3 min (R-form) and 25.0 min (S-form); for 2-(4-chlorophanoxy)propanoic acid: Daicel Chiralcel OJ-H column (hexane/2-propanol/trifluoroacetic acid = 95/5/0.1%, 1.0 mL·min−1, 254 nm), tR 15.4 min (R-form) and 19.0 min (S-form); for 2-methyl-3-phenylpropanoic acid, Daicel Chiralcel OJ-H column (hexane/2-propanol/trifluoroacetic acid = 95/5/0.1%, 0.5 mL·min−1, 254 nm), tR 20.0 min (R-form) and 22.4 min (S-form).

Purification and identification of 2-arylpropanoyl-CoA

2-Arylpropanoyl-CoA thioester was purified by solid-phase extraction cartridges (Chromabond C 18 ec, Macherey-Nagel, Düren, Germany)[30]. After the preparation of the aqueous layer as described above, reaction mixture was neutralized by the addition of 2 m NaOH and ammonium acetate was added to a final concentration of 520 mm. After cartridge preconditioning with consecutive washing with methanol, water, and 520 mm ammonium acetate solution (3 mL each), the mixture was loaded onto the cartridge. The column was rinsed with 5 mL of 520 mm ammonium acetate solution to remove free CoASH. The thioester was recovered by elution with 5 mL of distilled water. The eluates were lyophilized and resolved in a small amount of water. Fractions containing the thioester, which were identified with delayed nitroprusside reaction [31], were checked by TLC and TOF-MS analysis. TLC was performed on silica gel plates using the solvent system of 1-butanol/water/acetic acid (60/35/25) and Rf values of 0.57, 0.56, 0.55, and 0.56 were observed for flurbiprofenyl-CoA, ibuprofenyl-CoA, keoprofenyl-CoA, and naproxenoyl-CoA, respectively. Purified thioesters were subjected to ESI-TOF-MS with an LCT-Premier (Waters Corp.). The data were acquired in the negative ESI mode and leucine enkephalin was selected as a reference compound for high accuracy exact mass measurements.

Kinetic analysis

To 100 mm potassium phosphate buffer (pH 7.0), racemic ketoprofen (0.04–1.0 mm), ATP (10 mm), CoASH (2 mm), and MgCl2 (10 mm) were added. The final volume of the reaction mixture was adjusted to 500 µL. The mixture was preincubated at 30 °C for 10 min. The reaction was started by the addition of LUC-H (28.5 µg protein) and incubated at 30 °C for 10 min. After adding 100 µL of 2 m HCl and 1000 µL of diethyl ether to the reaction, the mixture was shaken and centrifuged at 17 610 g for 10 min. The separated organic layer was concentrated in vacuo. The yield of the unreacted carboxylic acid was measured by reversed-phase HPLC using a Mightysil RP-18GP Aqua 250–4.6 (5 µm) (Kanto Corp., Tokyo, Japan) with an isocratic solvent system of water/acetonitrile/trifluoroacetic acid (2/3/0.05%) at a flow rate of 0.5 mL·min−1 at 35 °C. The fractions from the column were monitored at 254 nm and flurbiprofen was used as an internal standard. Retention times of ketoprofen and flurbiprofen were 9.5 min and 13.2 min, respectively. Each assay was performed eight times. The kinetic parameters for racemic ketoprofen were determined by the method of Lineweaver–Burk plots.

E-value calculations

The reaction condition was the same as the kinetic analysis, except that 140 µg protein of LUC-H was used for this study. Each assay was performed three times. The E-value was calculated from the ee and yield of the unreacted acid according to a previously described method [27].

pH profile experiments

The thioester formation activity was determined in the following buffers (100 mm): citric acid–sodium citrate buffer (pH 3.0–4.0), succinate–KOH buffer (pH 4.0–5.5), Me–KOH buffer (pH 5.5–7.0), potassium phosphate buffer (pH 7.0–8.0), Tris/HCl buffer (pH 8.0–10.0), sodium carbonate–sodium hydrogen carbonate buffer (pH 10.0–11.5). The reaction condition was the same as the kinetic analysis, except that other buffer system and 280 µg protein was used for this experiment. Each assay was performed three times and the initial velocity of the thioester formation was determined at each pH. Relative activity was calculated from the comparison of these velocities.

Acknowledgements

This work was partially received the financial support from a Kawanishi Memorial Shinmaywa Education Foundation. DK greatly acknowledges the helpful support of Mr Mikio Bakke of Kikkoman Corporation.

Ancillary