Production of the Apoptotic Cellular Mediator 4-Methylthio-2-oxobutyric Acid byUsing an Enzymatic Stirred Tank Reactor with in Situ Product Removal

Miguel Garcıa-Garcıa, Irene Martı nez-Martınez, AÄ lvaro Sanchez-Ferrer, andFrancisco Garcıa-Carmona*

Department of Biochemistry and Molecular Biology-A, Faculty of Biology, University of Murcia, Campus Espinardo,E-30071 Murcia, Spain

D-Amino acid oxidase (DAAO) was used to study the oxidative deamination of racemic mixturesof D,L-methionine in its soluble and immobilized forms and thus obtain the correspondingR-ketoacid. The soluble enzyme form was obtained from aTrigonopsisVariabilis CBS 4095 extract,free of L-amino acid oxidase, and was co-immobilized with a 200-fold excess of catalase toavoid the undesirable side reaction of H2O2 with theR-keto acid, which would otherwise renderits corresponding decarboxylated acid, the 3-methylthiopropionic acid (MTPA). With thisbiocatalyst, quantitative conversion (>98%) ofD-methionine into theR-keto acid 4-methylthio-2-oxobutyric acid (MTOB) and into MTPA was achieved using 5 mg‚mL-1 of biocatalyst at pH8.0, 25 °C, and pure oxygen at 3 vvm. A stirred tank reactor with in situ product removal(STR-ISPR) was developed to avoid conversion of MTOB into MTPA. The reaction mediumwas re-circulated through a strong anion exchange column (Amberlite IRA-400). This resultedin the complete removal of MTOB from the reaction medium. After the reaction, the reactionproducts were eluted sequentially with water (L-methionine), 10 mM HCl (MTPA), and 0.5 MHCl (MTOB). After elution, MTOB was crystallized to its sodium salt.

1. Introduction

The production ofR-keto acids represents an ideal applicationof immobilized enzyme technology since chemical synthesisprocedures are tedious and generally result in low yields (1, 2).The difficulty involved in preparing the keto acid analogues ofessential amino acids has limited large scale clinical testing inthe management of a chronic renal failure, such as acute uremia(3-5), in which keto acids may serve as effective nitrogen-deficient nutritional substitutes for many of the essential aminoacids, by decreasing or forestalling the frequency of dialysistreatment (1). In addition, theseR-keto acids are gainingimportance as nutraceuticals (1).

Some keto acids can be prepared from their correspondingamino acids by an enzymatic reaction involvingD- or L-aminoacid oxidases. The latter is only commercially available fromsnake venom, whereas,D-amino acid oxidase (DAAO) is readilyavailable from different microorganisms, especially by fermen-tation of the yeastsRhodotorula gracilisand TrigonopsisVariabilis. (For review, see refs6 and7)

D-Amino acid oxidase (EC 1.4.3.3) is a flavin-adenindinucleotide (FAD)-dependent flavoprotein that catalyzes theoxidation of D-amino acids into imino acids and hydrogenperoxide with the liberation of a proton (Figure 1). The iminoacid formed is rapidly hydrolyzed to yield the correspondingR-keto acid and ammonium ion. However, the nascentR-ketoacid is nonenzymatically converted to the corresponding car-boxylic acid by the presence of the above-mentioned hydrogenperoxide. Thus, this last reactive side product is very disad-vantageous for the process ofR-keto acid production. Addition-ally, hydrogen peroxide has a strong denaturating effect on

proteins and, therefore, influences the operational stability ofthe DAAO. Catalase, therefore, has been used in free (8),immobilized, or co-immobilized (9) forms with DAAO toproduceR-keto acid.

In spite of the general scheme described above (Figure 1)for the DAAO oxidation ofD-amino acids, only three have beenenzymatically transformed into their correspondingR-keto acidsby the DAAO-catalase system. The most studied is thetransformation ofD-phenylalanine into its correspondingR-ketoacid (phenylpyruvic acid) (8-10). Higher yields were obtained

* Address correspondence to this author. Fax:+34968364147. E-mail:gcarmona@um.es.

Figure 1. General scheme for the oxidative deamination of amino acidracemic mixtures byD-amino acid oxidase. For details see the text. Inthe case of methionine (R) H3C-S-CH2-CH2-), its correspondingR-keto acid is 4-methylthio-2-oxobutyric acid (MTOB) and its corre-sponding decarboxylated compound is 3-methylthiopropionic acid(MTPA). End products of the reactions are boxed.

187Biotechnol. Prog. 2008, 24, 187−191

10.1021/bp0702424 CCC: $40.75 © 2008 American Chemical Society and American Institute of Chemical EngineersPublished on Web 12/20/2007

using an enzymatic reactor than with permeabilizedR. graciliscells containing both DAAO and catalase (10, 11).

Another R-keto acid, pyruvic acid, was produced fromD-alanine (10) in a stirred tank reactor containing DAAO-catalase co-immobilized in Affi-Gel 10. However, its industrialapplication was hindered by several factors, such as the cost ofthe catalyst, its low reusability, the cost of the substrate, andthe low price of the product. In the case of this last factor, theopposite is true for theR-keto acid, 4-methylthio-2-oxobutyricacid (MTOB) (see R in Figure 1), whose price is 80 times higherthan the correspondingD-methionine. The profit could be evenhigher, if a low-cost racemicD,L-methionine were used, sincethe stereospecificity of DAAO is absolute. No activity versusL-amino acids has ever been detected (8). In addition, this

product is of increasing interest since, at least in the case ofmethionine-dependent tumors (70% of all tumors), tumor cellsare deficient in MTOB and, therefore, unable to undergoapoptosis (12). Thus, MTOB becomes an indirect inhibitor ofcell growth in culture via the methional metabolic pathway (12).

Despite the interest of MTOB in cancer research, noenzymatic procedure for the quantitative transformation ofD,L-methionine into MTOB has been developed, apart from the firsttrials carried out with immobilizedT. Variabilis whole cells oncalcium alginate with manganese oxide (13). This complexsystem, together with the general drawbacks of using metalsand whole cells, generates additional problems, including severelimitation in oxygen diffusion and the existence of otherundesirable enzymatic activities (9). Trost and Fischer (8)proposed the use of the immobilized DAAO with the externaladdition ofA. nigercatalase, with which they obtained conver-sions of 73% of MTOB and 27% of its corresponding decar-boxylated compound, the 3-methylthiopropionic acid (MTPA)(see R in Figure 1). This system has the problem of increasingcost due to the use of a non-reusable enzyme (catalase) and,specially, the difficulty of a complete removal of the solubleprotein from the final product.

The aim of this work was to design an enzymatic processfor the quantitative oxidative deamination of the racemicD,L-methionine into MTOB by co-immobilization of DAAO andcatalase on an oxirane support, prior to its isolation as a solidsodium salt.

2. Materials and Methods2.1. Chemicals.All chemicals were commercially available,

of reagent grade, and purchased from Sigma-Aldrich-Fluka(Madrid, Spain). Eupergit C was kindly supplied by RohmPharma (Germany) and used as described by the supplier.3-Methylthiopropionic acid was acquired from Chemos GmbH(Germany).

Figure 2. Oxidative deamination ofD,L-methionine by co-immobilizedDAAO-catalase onto Eupergit C using HPLC. The reaction medium(100 mL) at 25°C contained 20 mMD,L-methionine in water, pH 8.0,20 U of DAAO, and was aerated with pure oxygen (3 vvm).

Figure 3. Characterization of the biocatalyst. pH (A), temperature (B), substrate (C), and oxygen flow (D) dependences of DAAO-catalase biocatalyst.The reactions were performed under the standard reaction described in the Materials and Methods section.

188 Biotechnol. Prog., 2008, Vol. 24, No. 1

2.2. Enzymes.DAAO was obtained fromTrigonopsisVari-abilis CBS 4091 grown under conditions to induce DAAO (14).The enzyme was purified by ammonium sulfate fractioniationbetween 30% and 55% as described by Sa´nchez-Ferrer et al.(15), obtaining a specific activity of 5 U‚mg-1. This enzyme(0.2 U), 40 catalase units (fromMicrococcus lysodeikticus,Sigma-Aldrich-Fluka), and 1 mg of dry Eupergit C (Sigma) weremixed with 1 M phosphate buffer pH 8.0 for 48 h followingthe immobilization conditions described by Kra¨mer and Steckam(16). Immobilization yield (enzyme activity units) was 94%.

2.3. Enzyme Assays.The activity of DAAO was determinedaccording to a peroxidase-phenol-4-aminoantipyrine colori-metric assay at 505 nm (17). The standard reaction mixture (1mL) at 37°C contained 20 mMD-methionine, 7 mM phenol,0.4 mM 4-aminoantipyrine, and 4.4 U of peroxidase. Thereaction was started by adding a maximum of 0.1 U of DAAOinto oxygen saturated 50 mM potassium phosphate buffer pH8.0. One unit of DAAO corresponds to the formation of 1µmolof H2O2 per minute under the above-mentioned conditions.

The activity of immobilized DAAO-catalase was followedby RP-HPLC (Agilent 1100 Series with a diode array detector).Measurement of methionine, its correspondingR-keto acid(MTOB), and carboxylic acid (MTPA) was performed on a 5µm Kromasil C8 column (150 mm× 4.6 mm i.d., Ana´lisisVınicos S.L., Spain) at a flow-rate of 1 mL‚min-1 and a fixedwavelength of 215 nm in a mobile phase consisting of 15 mM

KH2PO4, 30% methanol, and 10 mM tetrabutylammoniumhydrogen sulfate (TBAHS), pH 6.5. Retention times ofD,L-methionine, MTPA, and MTOB were 2.0, 5.3, and 7.0 min,respectively. Biotransformations were performed in a continu-ously stirred tank reactor (Metrohm 718 STAT Tritino, Madrid,Spain) with 100 mL working volume. The pH-controlledreaction mixture was mechanically stirred (250 rpm), aeratedwith pure oxygen (3 vvm), and kept at 25°C. ImmobilizedDAAO-catalase (20 DAAO U) was added to start the reactionand pH was automatically adjusted to pH 8.0 by adding 3 Mammonia. Samples were filtered and an aliquot of 20µL wasinjected and analyzed by RP-HPLC in order to determinesubstrate and product concentrations based on the correspondingcalibration curves carried out with the standards.

At 25 °C in 100 mL the standard reaction medium for theimmobilized DAAO-catalase contained 20 mMD,L-methioninein water, 0.5 g of wet biocatalyst (20 U), and pure oxygen at 3vvm.

2.4. Isolation of Reaction Products.When all the substratehad been converted, the solution was filtered and the im-mobilized enzyme was washed. The reaction products wereisolated from the reaction mixture by ion exchange chroma-tography on a strong anion exchanger Amberlite IRA 400(Rohm & Haas) (bed volume 20 mL; bv) equilibrated with waterat a flow of 0.8 bv‚min-1. Under this water flush,L-methioninewas eluted, while theR-keto acid and the carboxylic acidremained bound to the resin. These two compounds weresequentially eluted by increasing concentrations of HCl. MTPAwas eluted with 22 bv of 10 mM HCl and MTOB with 22 bvof 0.5 M HCl.

A stirred tank reactor with in situ product recovery (STR-ISPR) (18) was also used. The reaction medium was pumpedcontinuously thorough the same IRA 400 column. When theconversion was finished, the column was disconnected andMTOB was eluted under the above-mentioned conditions.

MTOB sodium salt was obtained by combining its fractionsand drying them. The viscous solution thus obtained wasneutralized to pH 7.5 with 0.2 N NaOH and dried to give thesodium salt of MTOB. The purity was confirmed by NMR ina Bruker 400-MHz spectrometer (1H NMR, D2O) (δ ppm): 3.02(t, 2H, -CH2-), 2.7 (t, 2H,-CH2-), 2.04 (s, 3H, CH3S), anda value of 95% was obtained. Optical rotation ofL-methioninewas measured using a Jasco P-1020 at 28°C. A value of 26.19-((0.02)° was obtained with the standard (Sigma) and with thesample.

3. Results and Discussion

D-Amino acid oxidase is a stereospecific enzyme that canuse a broad spectrum ofD-amino acids used as substrates, withD-methionine being a good substrate for the DAAO fromTrigonopsisVariabilis CBS 4095 (19). However, it has beensuggested thatL-amino acid oxidase (LAAO) is present inT.Variabilis cells (9). In order to test this, an ammonium sulfateenzymatic preparation ofT. Variabilis CBS 4095 was assayedin the presence of both 20 mMD- and L-methionine and wasonly active towardD-isomer.

After co-immobilization of the above-mentioned DAAO withcatalase onto Eupergit C, the biocatalyst was active towardtheD-isomer ofD,L-methionine (Figure 2). The peak at 215 nm,which corresponded to the racemic substrate, decreased(Figure 2), accompanied by a concomitant appearance increaseof two new peaks at about 5.3 and 7.0 min, corresponding to3-methylthiopropionic acid (MTPA) and MTOB, respectively(Figure 2). The conversion ofD-isomer was almost quantitative,

Figure 4. Oxidative deamination ofD,L-methionine by co-immobilizedDAAO-catalase in a stirred tank reactor (A) and in a stirred tank reactorwith in situ product removal (B) by a strong anion exchanger (AmberliteIRA-400 column). Reaction conditions were the same as in Figure 2,except forD,L-methionine (which was 100 mM) and DAAO (whichwas 100 U). (O) D,L-methionine, (3) MTOB, and (0) MTPA.

Biotechnol. Prog., 2008, Vol. 24, No. 1 189

since the initial concentration ofD,L-methionine (20 mM)decreased asymptotically to half of the initial concentration, witha corresponding increase of the MTOB and MTPA concentrationto almost this value (Figure 2).

The activity of the DAAO-catalase biocatalyst was affectedby the pH of reaction medium at 25°C, increasing at morebasic pH values, with a maximum at pH 8.0 (Figure 3A). Afurther increase in pH resulted in a lower amount of MTOBdue to its instability at these more basic pH values. This pH is0.5 unit lower than that described for the oxidation ofD-alanineto its correspondingR-keto acid (pyruvic acid) catalyzed byRhodotorula gracilisDAAO-catalase immobilized onto Affi-Gel 10 (10), and 1.0 unit lower than that shown in the oxidationof D-phenylalanine to phenylpyruvic acid carried out byT.

Variabilis DAAO-catalase immobilized onto glutaraldehyde-agarosa (9). This clearly indicates that the optimum pH foroxidative deamination of amino acids intoR-keto acids dependson the type of amino acid, the source of DAAO, the support,and the post-immobilization technique used.

The activity of the biocatalyst was also dependent on thetemperature in the reaction medium, increasing when thetemperature increased from 15 to 30°C, with a maximum at25 °C (Figure 3B), as previously described in the productionof otherR-keto acids (9, 10). Furthermore, the activity, whichdepended on enzyme concentration, was linear in the 0 to 10mg‚mL-1 range.

The effect of substrate concentration was also evaluated inthe 0 to 100 mM range, showing saturation at concentrationsabove 20 mMD,L-methionine (Figure 3C). The kinetic constants,Vm andKM, were evaluated by nonlinear regression to the data.Values of 202µM‚min-1 and 4.65 mM, respectively, wereobtained. TheKM could not be compared with any other DAAO-catalase systems (9, 10), but theKM was 7-fold higher than theKM calculated for the free enzyme used in this study. Thisincrease could be related to the conformational changes carriedout in the enzyme structure during immobilization. Finally, theinfluence of oxygen flow in the reactor was studied in the 0 to6 vvm range (volume O2/volume medium/ minute), withsaturation reached above 3 vvm (Figure 3D).

To avoid the transformation of MTOB into its correspondingdescarboxylated MTPA, the stirred tank reactor (STR) used inthe previous experiments was transformed into a stirred tankreactor with in situ product removal (STR-ISPR) (17) bycoupling an anion exchange column (Amberlite IRA-400, Rohn& Hass) to the reactor through a membrane pump. Figure 4shows the difference between an STR reaction medium (Figure4A) and the medium found in an STR-ISPR system (Figure4B). No MTOB and MTPA were detected, in the latter due tothe selective binding of both compounds to the Amberlite IRA-400 column. This permits quantitative conversion ofD-methion-ine and an increase in MTOB yield (90% MTOB and 10%MTPA vs 73% MTOB and 27% MTPA in STR). This 73%

Figure 5. Elution profile of products generated during the oxidativedeamination ofD,L-methionine by co-immobilized DAAO-catalase.Reaction conditions were the same as in Figure 3B. At the end of thereaction, the medium was loaded (0.81 bv‚min-1) into the AmberliteIRA-400 column then washed with 37 bv of water, and the productswere sequentially eluted with 22 bv of 10 mM HCl (MTPA) and 22bv of 0.5 M HCl (MTOB). Fractions of 20 mL were collected:L-methionine (fractions 3-15), MTPA (fractions 41-49), and MTOB(fractions 63-77).

Figure 6. NMR spectrum of purified MTOB. The NMR spectrum was measured in D2O with a Bruker 400 MHz spectrometer.

190 Biotechnol. Prog., 2008, Vol. 24, No. 1

MTOB + 27% MTPA yield was also reported by Trost andFischer (8) with D-methionine, with a 10-fold excess of catalaseto DAAO, clearly indicating the benefits of using the in situproduct removal. To take full advantage of the MTOB bindingto the column, the strong anion exchanger was disconnectedfrom the reactor at the end of the reaction, the reaction mediumwas loaded, and a sequential elution of the products was carriedout (Figure 5). All theL-methionine was eluted by simplywashing the column with water, while MTPA was subsequentlyeluted by passing a dilute HCl solution (10 mM). The purity ofL-methionine obtained was 99% as confirmed by a polarimetricstudy when compared with a commercial standard ([R]28

D

+26.19( 0.02). Finally, MTOB was eluted with 0.5 M HCl.The MTOB fractions were combined, dried, and neutralized topH 7.5 with 0.2 M NaOH. This solution was dried again,rendering the solid sodium salt of MTOB, as confirmed byNMR, with a purity of 95% (Figure 6). The overall molar yieldof the process was close to 75%.

Scale-up studies would render higher yields, taking intoaccount that the mass losses due to the liquid transfers arereduced. In addition, the enzymatic conversion is economic sinceit uses only a 200-fold excess of catalase compared with the2260-fold excess described for quantitative oxidation ofD-phenylalanine to phenylpyruvic (9), but still 20-fold more thanthose used by Trost and Fischer (8). The key limitations of thescale-up are the solubility of the substrate and maintaining theoxygen levels during the reaction. This could produce somenegative effects due to the increase in the partial oxygen pressureof the system. Although the concentration of transformedsubstrate is still low, this could be improved by increasing theinitial concentration of the substrate up to 150 mM. To obtainhigher amounts of product, some cosolvent would be requiredin order to improve the solubility of the substrate, althoughfurther substrates would be required.

In conclusion, a simple method for obtaining the sodium saltof MTOB has been developed by using co-immobilized DAAOwith catalase onto Eupergit C in a stirred tank reactor with insitu product removal by coupling Amberlite IRA-400 column.This opens up the possibility of scaling up to obtain the highlyvaluedR-keto acid, MTOB, for the pharmaceutical and foodindustries.

In addition, the experimental methodology developed in thispaper also allows MTPA or a combination of MTPA and MTOBto be produced by changing the level of catalase co-immobilizedwith DAAO from 0 to the level used in this paper. This isimportant since MTPA is not commercially available fromchemical and biochemical suppliers. This would give rise to acommercial process in which three valuable compounds (pureL-methionine, MTPA, and MTOB) can be obtained at the sametime, either in radiolabeled (35S or14C) form or not, dependingon the nature of the initial racemic substrate.

Acknowledgment

This work was partially supported by MCYT and FEDER(BMC2001-0499), MEC and FEDER (BIO2004-00439), andFundacio´n Seneca 00608/PI/04. I.M.M. holds a predoctoral

research grant from the Obra Social Caja de Ahorros delMediterraneo, Spain.

References and Notes(1) Fink, D. J.; Falb, R. D.; Bean, M. Production ofR-keto acids by

immobilized enzymes.AIChE Symp. Ser.1978, 74, 18.(2) Cooper, A. J. L.; Ginos, J. Z.; Meister, A. Synthesis and properties

of the R-ketoacids.Chem. ReV. 1983, 83, 321-358.(3) Chantler, C.; El Bishiti, M.; Counahan, R. Nutrirional therapy in

children with chronic renal failure.Am. J. Clin. Nutr.1980, 33,1682-1689.

(4) Schaefer, K.; Erley, C. M.; von Herrath, D.; Stein, G. Calciumsalts of ketoacids as a new strategy for uremic hyperphosphatemia.Kidney Int. Suppl.1989, 27, 136-139.

(5) Battaner, E.; Catala´n, J.; Gala´n, M. A.; De Luis, J. M. Tratamientosustitutivo de la dia´lisis. Ann. Med. Interne1994, 11, 17-21.

(6) Pszczola, D. E. The ABCs of nutraceutical ingredients.FoodTechnol.1998, 52, 30-37.

(7) Pilone, M. S.D-amino acid oxidase: new findings.Cell Mol. LifeSci.2000, 57, 1732-1747.

(8) Trost, E. M.; Fischer, L. Minimization of by-product formationduring D-amino acid oxidase catalyzed racemate resolution ofD/L-amino acids. J. Mol. Catal.2002, 19-20, 189-196.

(9) Fernandez-Lafuente, R.; Rodrı´guez, V.; Guisan, J. M. The coim-mobilitation of D-amino acid oxidase and catalase enables thequantitative transformation ofD-amino acids (D-phenylalanine) intoR-cetoacids (phenypyruvic acid).Enzyme Microb. Technol.1998,23, 28-33.

(10) Buto, S.; Pollegioni, L.; D’Angiuro, L.; Pilone, M. S. Evaluationof D-amino acid oxidase fromRhodotorula gracilisfor the productionof R-keto acids, A reactor system.Biotechnol. Bioeng.1994, 44,1288-1294.

(11) Upadhya, R.; Nagajyothi, B. S. G. Stabilition ofD-amino acidoxidase and catalase in permeabilizedRhodotorula graciliscells andits application for the preparation ofR-ketoacids.Biotechnol. Bioeng.2000, 68, 430-436.

(12) Quasch, G.; Roch, A. M.; Chantepie, J.; Michal, Y.; Fournet, G.;Dumontet, Ch. Methional derived from 4-methylthio-2-oxobutanoateis a cellular mediator of apoptosis in BAF3 lymphoid cells.Biochem.J. 1995, 305, 1017-1025.

(13) Brodelius, P.; Nilsson, K.; Mosbach, K. Production ofR-ketoacids. I. Immobilized cells ofTrigonopsisVariabilis containingD-amino acid oxidase.Appl. Biochem. Biotechnol.1981, 6, 293-308.

(14) Kubicek-Pranz, E. M.; Ro¨hr, M. D-amino acid oxidase from yeastTrigonopsisVariabilis. J. Appl. Biochem.1985, 7, 104-113.

(15) Sanchez-Ferrer, A.; Lo´pez-Mas, J. A.; Garcı´a-Carmona, F. Aprocess for preparing 3-cephalosporanic acid derivatives usingR-ketoacid derivatives. Patent WO 0204353, 2002.

(16) Kramer, L.; Steckam, E. Coimmobilitation ofL-R-glycerophos-phate oxidase with catalase and its application for the synthesis ofdihydroxyacetone phosphate.Tretraedron1997, 53, 14645.

(17) Trinder, P. Determination of blood glucose using an oxidase-peroxidase system with a non-carcinogenic chromogen.J. Clin.Pathol.1969, 22, 158-161.

(18) Lye, G. J.; Woodley, J. M. Application of in situ product-removaltechniques to biocatalic processes.Trends Biotechnol.1999, 17,395-402.

(19) Fischer, L.D-amino-acid oxidase in biotechnology.Recent Res.DeV. Microbiol. 1998, 555, 295317.

Received July 23, 2007. Accepted November 19, 2007.

BP0702424

Biotechnol. Prog., 2008, Vol. 24, No. 1 191