Substrate specificity of native and mutated cytochrome P450 (CYP102A3) from Bacillus subtilis

Journal of Biotechnology 108 (2004) 41–49

Substrate specificity of native and mutated cytochrome P450 (CYP102A3) from Bacillus subtilis Oliver Lentz, Vlada Urlacher, Rolf D. Schmid∗ Institut für Technische Biochemie, Universität Stuttgart, Allmandring 31, D-70569 Stuttgart, Germany Received 29 January 2003; received in revised form 3 November 2003; accepted 18 November 2003

Abstract Within the Bacillus subtilis genome sequencing project, two monooxygenases (CYP102A2 and CYP102A3) were discovered which revealed a similarity of 76% to the well-known cytochrome P450 BM-3 (CYP102A1) of Bacillus megaterium. All enzymes are natural fusion proteins consisting of a heme domain and a reductase domain. We here report the cloning, expression and characterization of B. subtilis enzyme CYP102A3. The substrate specificity of this enzyme is similar to that of B. megaterium CYP102A1, which hydroxylates medium-chain fatty acids in subterminal positions. A double mutant was prepared that hydroxylates a number of other substrates, which do not bear any resemblance to the natural substrate of this enzyme family. © 2003 Elsevier B.V. All rights reserved. Keywords: Cytochrome P450 CYP102; Bacillus megaterium; Bacillus subtilis; Fatty acid hydroxylation; Alkane hydroxylation; Arene hydroxylation

1. Introduction P450 monooxygenases play a key role in primary and secondary metabolic pathways and in drug detoxification. Cytochromes P450 are heme-thiolate proteins which are widely distributed in animals, plants and microorganisms (Nelson et al., 1996). Abbreviations: ABTS, 2 ,2-azino-di-(3-ethyl-benzathiazoline6-sulphonic acid); BM-3, cytochrome P450 CYP102A1 from Bacillus megaterium; HRP, horseradish peroxidase; 8-pNCA, ␻(p-nitrophenyl)octanoic acid; 10-pNCA, ␻-(p-nitrophenyl)decanoic acid; 12-pNCA, ␻-(p-nitrophenyl)dodecanoic acid; KPi, K2 HPO4 / KH2 PO4 buffer ∗ Corresponding author. Tel.: +49-711-685-3192; fax: +49-711-685-3196. E-mail address: [email protected] (R.D. Schmid).

Their unique hydroxylation capacity provides the opportunity to use these biocatalysts for the synthesis of fine chemicals (Stegeman and Lech, 1991; Juchau, 1990; Guengerich et al., 1996), as they even act on non-activated carbon–hydrogen bonds and hydroxylate such compounds as alkanes, fatty acids, terpenes and steroids, often exhibiting high regio- and stereoselectivity. P450 CYP102A1 from Bacillus megaterium, a natural fusion protein between a heme and a reductase domain, hydroxylate substrates 100–1000 times faster than eukaryotic P450 systems (Boddupalli et al., 1990). Thus, it was obvious to investigate other natural fusion bacterial P450s. The Bacillus subtilis genome sequencing project revealed six cytochrome P450 enzymes (bioI, cypA, cypC, cypX, yetO and

0168-1656/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.jbiotec.2003.11.001

42

O. Lentz et al. / Journal of Biotechnology 108 (2004) 41–49

yrhJ), whose biological function is only partially known. CypC (formerly known as ybdT) is a 48 kDa protein which hydroxylates myristic acid in ␣- and ␤-position in the presence of H2 O2 (Matsunaga et al., 1999). BioI is a 45 kDa protein involved in biotin biosynthesis in B. subtilis (Stok, De Voss, 2000). There is no data or publication about cypA and cypX at the moment. The proteins coded by yetO and yrhJ were recently classified as P450 CYP102A2 and P450 CYP102A3 and are involved in the fatty acid metabolism of B. subtilis (Gustafsson et al., 2001; Lee et al., 2001). Like CYP102A1 from B. megaterium, they are water-soluble fusion proteins (Mr 119 kDa), which contain a heme and an FAD-/FMN-dependent reductase domain within the same polypeptide chain. In terms of the electron transfer from NADPH to the heme iron, P450 CYP102A2 and P450 CYP102A3 can thus be regarded as self-sufficient. The similarity of 76% (CYP102A2) and 77% (CYP102A3) to P450 CYP102A1 makes the comparison of substrate preferences and specificity of these enzymes of some interest. Cytochrome P450 BM-3 (CYP102A1) from B. megaterium, an enzyme which is among the most studied prokaryotic P450 monooxygenases, catalyzes the subterminal hydroxylation of fatty acids with a chain length of C12 –C22 . It preferentially hydroxylates in the ␻-1–3 positions with high enantioselectivity in the ␻-1 and ␻-2 positions (98% R, 2% S) (Truan et al., 1999). By introduction of mutations into this enzyme’s binding site, the substrate spectrum was increased significantly. CYP102A1 mutant A74G F87V L188Q was able to convert octane, ionones and naphtalene. The hydroxylation of octane resulted in a mixture of 2-, 3- and 4-octanol with a conversion rate of 93% (Appel et al., 2001). In this paper, we describe the cloning of the wild-type CYP102A3 protein from B. subtilis and the introduction of mutations at positions F88 and S189 (corresponding to F87 and L188 of CYP102A1).

2. Materials and methods 2.1. Materials All chemicals were of analytical grade or higher quality and were purchased from Fluka (Buchs,

Switzerland) or Sigma (Deisenhofen, Germany). Capric and lauric acid were obtained from Henkel (Düsseldorf, Germany). 2.2. Bacterial strains, vector construction and mutagenic PCR The E. coli strain DH5␣ [supE44, lacU169 (80lacZ M15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1] (Clontech, Heidelberg, Germany) was used for cloning, E. coli strain BL21(DE3) (F− , ompT, hsdSB (rB − mB − ) gal dcm (DE3)) (Novagen, Bad Soden, Germany) for protein expression. The P450 yrhJ gene was amplified by PCR using genomic DNA of B. subtilis (DSM 402) as template and primers which introduced a NcoI site at the 5 end and a NotI site at the 3 end. The primers were 5 -gcatccatggccatgaaacaggcaagcgcaatac-3 and 5 -gcatgccgcggccgccattcctgtccaaacgtc-3 . The obtained PCR fragment was inserted downstream of the strong, IPTG-inducible T7 promoter of pET20b(+) resulting in plasmid pET20b-yrhJ. The reaction conditions involved a 50 ␮l reaction volume containing 25 pmol of each primer, 5 ng of isolated genomic DNA, 2.5 U of Taq polymerase, 2.5 nmol of each dNTP. The reaction was started at 95 ◦ C (7 min) and thermocycled for 30 cycles: 95 ◦ C (1.5 min), 55 ◦ C (1.5 min), 72 ◦ C (2.75 min), with a final extension at 72 ◦ C (7 min). PCR products were controlled by subsequent sequencing. The S189Q (single mutant; primers 5 -ggcgatgaatcaacaaaaaagactgggcctgc-3 and 5 -gcaggcccagtcttttttgttgattcatcgcc-3 ) and F88V S189Q (double mutant; primers 5 -gggggagatggcttagttacaagctggacgcacg-3 and 5 -cgtgcgtccagcttgtaactaagccatctccccc-3 ) mutants were created by subsequent site-directed mutagenesis of the CYP102A3 wild-type gene using the Stratagene QuickChange Kit (Stratagene, La Jolla, CA, USA). Site-directed mutagenesis involved the exchange of a single codon to replace a certain amino acid. The reaction conditions involved a 50 ␮l reaction volume containing 5 ␮l DMSO, 1 pmol of each primer, 0.8 ␮g of template plasmid DNA, 2.5 U of Pfu turbo polymerase, 0.75 nmol of each dNTP. The reaction was started at 95 ◦ C (2 min) and thermocycled for 15 cycles: 95 ◦ C (1 min), 58 ◦ C (F88V, 1 min) or 55 ◦ C (S189Q, 1 min), 68 ◦ C (10 min), with a final extension at 68 ◦ C (10 min). All PCR product solutions were treated with 20 U Dpn I at 37 ◦ C for 1 h to digest the initial, non-mutated

O. Lentz et al. / Journal of Biotechnology 108 (2004) 41–49

43

template DNA prior to the transformation into E. coli DH5␣ or BL21(DE3) cells.

mination of the whole protein content. The purified fractions were immediately used for activity tests.

2.3. Enzyme preparation

2.5. Enzyme activity assay (pNCA-assay)

The enzymes used for hydroxylation experiments were prepared as follows: 500 ml Luria Bertani (LB low salt) medium supplemented with 100 ␮g/ml ampicillin for plasmid selection was inoculated with 5 ml of an overnight culture of recombinant E. coli BL21(DE3) cells. The overnight culture contained 1% glucose to prevent basal expression. The shaking flask was incubated at 200 rpm and 37 ◦ C. At an optical density (OD578 nm ) of 0.8–1.0, P450 expression was induced by adding IPTG to a concentration of 0.4 mM. The cells were grown at 30 ◦ C and at 150 rpm, higher stirrer speeds resulted in enzyme inactivation. Cells were harvested by centrifugation (8000 g/15 min) and stored at −20 ◦ C until further use.

The pNCA-assay is a spectrophotometric assay based on ␻-(p-nitrophenolate)carboxylic acids. The hydroxylation of these surrogate substrates results in the formation of a semiacetale that dissociates in an oxo-carboxylic acid and the yellow p-nitrophenolate (Schwaneberg et al., 1999). For all activity assay procedures 5 ␮l of a 15 mM DMSO solution of the respective pNCA substrate was used in disposable cuvettes. After the addition of 865 ␮l of Tris–HCl buffer (50 mM, pH 8.0) and 30 ␮l of P450 extract solution (∼2.5 ␮M), the samples were pre-incubated 5 min before the reaction was started by the addition of 100 ␮l of an aqueous 6 mM NADPH solution. Further details are described elsewhere (Schwaneberg et al., 1999, Schwaneberg, 1999).

2.4. Preparation of E. coli BL21(DE3) cell extracts and further purification Up to 3.5 g (wet weight) BL21(DE3) cells suspended in 7.5 ml Tris–HCl buffer (25 mM, pH 7.5, with 1 mM EDTA as protease inhibitor), were thawed on ice, and sonified in an ice bath for 2 × 2 min (Branson Sonifier W250, Dietzenbach, Germany; output level 80 W, duty cycle 30%). The suspension was centrifuged at 32 500×g for 30 min. Afterwards, the crude extract was passed through a 0.22 ␮m Sterivex-GP filter (Millipore, Eschborn, Germany); P450 content was measured by CO-differential spectra (Omura and Sato, 1964). An ÄKTAexplorer system (Amersham Pharmacia Biotech, Uppsala, Sweden) was used for further purification. The cell-free extract was loaded onto a self-packed XK16/20 (60 mm × 16 mm) column containing DEAE650M anion exchange material (TosoHaas, Stuttgart, Germany) pre-equilibrated with a 25 mM Tris buffer, pH 7.5 and eluted with a step-gradient of 130/250/1000 mM NaCl (flow rate 4 ml/min). The protein solution was concentrated using Centricon filter units (Millipore, Eschborn, Germany) with a 50 kDa filter unit and desalted with a PD10 gel filtration column (Amersham Pharmacia Biotech, Uppsala, Sweden). The enzyme solution showed a purity of up to 67% checked by comparing the results of CO-differential spectra and BCA deter-

2.6. NADPH-assay The NADPH-assay was performed as described elsewhere (Appel et al., 2001). The observed rates of NADPH consumption were corrected for the slow background reaction, i.e. the oxidization of the reduced forms of CYP102 FAD/FMN domain to the heme domain by air and electron transfer in the absence of substrate (Daff et al., 1997). 2.7. Determination of uncoupling Reduced oxygen intermediates like H2 O and H2 O2 can also be generated from reducing equivalents of NADPH. Production of hydrogen peroxide during substrate turnover was measured by the HRP-ABTS assay developed by Childs and Bardsley and described by Tsotsou et al. (2002). 2.8. Enzymatic hydroxylation and product isolation For the conversion reactions, 50 mM substrate solution in DMSO (50 ␮l), 50 mM KPi buffer (pH 7.8, 4.59 ml) and P450 extract solution (to an enzyme end concentration of 0.125 ␮M) were combined. After a pre-incubation time of 5 min, the reaction was started by adding 110 ␮l of an aqueous 12 mM NADPH so-

44

O. Lentz et al. / Journal of Biotechnology 108 (2004) 41–49

lution. NADPH consumption was monitored by measuring the absorbance at 340 nm. After reaching the value of the negative control, NADPH was added once again. In the case of the fatty acids, the pH was adjusted to 2.0 after completion of the reaction by the addition of HCl (5 M) with concomitant precipitation of the fatty acids and their hydroxy derivatives. The slurry was extracted twice with dichloromethane (2.5 ml each). The organic phases were combined, dried with sodium sulfate, and evaporated to complete dryness. Prior to derivatization 100 ␮l of dichloromethane were added. pH adjustment is not necessary after octane hydroxylation. The reaction mixture was extracted three times with diethylether (2.5 ml each) containing 1 mM of 1-decanol as internal standard. The organic phases were dried with sodium sulfate and evaporated to a residual volume of 100 ␮l. This solution was used directly for GLC analysis. The hydroxylation of naphtalene was performed in the same manner. The residual product solution was analyzed with a special TLC analysis method using Fast Blue RR salt as indicator (Jork et al., 1993). 2.9. GLC analysis Since the observed NADPH consumption does not differentiate between main and side reactions, it is mandatory to examine the structure and amount of hydroxylated product by GLC analysis. For this purpose, the extracts were analyzed by capillary GLC with co-injection of authentic standards and compared to previously described data (Appel et al., 2001; Lentz et al., 2001). For the derivatization of fatty acids, N-methylN-trimethylsilyl-heptafluorobutyramide (10 ␮l) was added to a 40 ␮l aliquot of the diethylether/dichloromethane extract solution. The mixture was incubated at ambient temperature for 15 min. Gas chromatograms (1 ␮l samples) were run on a Fisons Mega Series GC (Fisons Instruments, Mainz, Germany), equipped with split injector and flame ionization detector (FID), and an Optima 5 column (25 m × 0.25 mm, film thickness 0.25 ␮m; Macherey & Nagel, Düren, Germany). Forepressure was 50 kPa H2 for octanol and 100 kPa for fatty acid analysis, injector temperature 280 ◦ C (octanol)/300 ◦ C (fatty acids) and detector temperature 300 ◦ C. Temperature program for all

fatty acids: 80 ◦ C (2 min isothermal) −80 to 250 ◦ C (10◦ /min); for octane: 40 ◦ C (5 min) −40 to 100 ◦ C (2 ◦ C/min) −100 to 250 ◦ C (10 ◦ C/min).

3. Results and discussion The mutations A74G, F87V and L188Q have been proven useful in CYP102A1 with regard to broadening its substrate spectrum (Appel et al., 2001). The corresponding mutations were also introduced in CYP102A3 that already contains a glycine residue at position 75 (corresponding to A74G of CYP102A1). We used the wild-type enzyme, the F88V, S189Q mutants and the F88V S189Q mutant for the hydroxylation experiments to check which mutation is responsible for the widened substrate spectrum. The aim was to identify monooxygenases with properties which would make them suitable candidates for industrial application. High stability at ambient temperature and high regioselectivity in substrate hydroxylation constituted the desired features. 3.1. Differences in stability and substrate specificity between B. megaterium CYP102A1 and B. subtilis CYP102A3 We explored the properties of the P450 monooxygenase CYP102A3 of B. subtilis and its two specifically designed mutants using the pNCA-assay and compared them to the properties of P450 CYP102A1 of B. megaterium. pH, co-solvent, temperature and buffer salt stability values of CYP102A3 were compared to the CYP102A1 values previously described by Schwaneberg (1999). There were only minor differences in temperature and pH stability between the two enzymes. For CYP102A1, 30 min incubation at 36 ◦ C lead to an increase in activity of up to 60%; no such effect was observed for CYP102A3. The residual activity of CYP102A3 at ambient temperature, measured over a period of 7 days, was higher than that of the B. megaterium enzyme (Fig. 1). CYP102A3 was not as stable as CYP102A1 when kept in saline solutions. P450 CYP102A1 retains 80% of its activity in solutions with 400 mM NaCl whereas CYP102A3 activity fell below this level when the NaCl concentrations were above 150 mM. This instability caused problems in purifying CYP102A3 using

O. Lentz et al. / Journal of Biotechnology 108 (2004) 41–49

45

Fig. 1. Residual activity of CYP102A1 of B. megaterium and CYP102A3 of B. subtilis at room temperature (25 ◦ C). The residual activity was determined with the pNCA-assay directly after preparation of the cell-free extract (t = 0) and repeated every 24 h with the enzyme solution stored at room temperature. Storage and reaction conditions were the same for both enzymes.

anion exchangers. However, the comparison of activity and specificity between the purified enzyme and the cell-free extract revealed no differences. We therefore concluded that additional purification could be left out in the tests mirroring substrate activity. The greatest difference between CYP102A1 and CYP102A3 relates to their activity when different amounts of co-solvent are present in the reaction mixture. Co-solvents are often used to solubilize hydrophobic substrates in aqueous reaction mixtures. It could be shown that the B. subtilis enzyme was more active when up to 25% (v/v) DMSO was present in the reaction mixture. The residual activity of CYP102A1, however, already dropped to 11% at 10% (v/v) DMSO (Fig. 2). The reason for this is not known and there is no further data like X-ray diffraction available at the moment. Substrate specificity towards the pNCA-substrates indicated that CYP102A3 was better suited to hydroxylate shorter-chain substrates than CYP102A1 (Table 1). Wild-type CYP102A1 did not show any activity towards 8-pNCA; CYP102A3 revealed some activity but the mutant CYP102A3 F88V S189Q re-

vealed an activity that was twice as high as that of the CYP102A3 wild-type. 3.2. Hydroxylation of fatty acids The wild-type enzyme CYP102A3 catalyzes the conversion of medium-chain fatty acids as shown in Tables 2 and 3. According to Lentz et al. (2001), the wild-type CYP102A3 and the mutants were expected to be able to catalyze the conversion of capric acid (C10:0) since they show activity towards the similar-sized 8-pNCA. Indeed, GLC analysis identified 7-, 8-, 9- and a small amount of 10-hydroxycapric acid (in the case of the F88V S189Q mutant) as products (Table 3) with a conversion rate of only 28%. Lauric (C12:0) and palmitic acid (C16:0) were also hydroxylated by CYP102A3 but with far higher conversion rates compared to CYP102A1 of B. megaterium (e.g. 93% versus 40% for lauric acid). With the exception of palmitic acid, the regioselectivity of the hydroxylation process was almost equal to that found in the wild-type CYP102A3 and the S189Q single mutant. This indicates that the mutation at position

46

O. Lentz et al. / Journal of Biotechnology 108 (2004) 41–49

Table 1 Kinetic parameters of P450 CYP102A1 (wild-type and A74G F87V L188Q mutant) of B. megaterium and CYP102A3 (wild-type and F88V S189Q mutant) of B. subtilis for pNCA substrates P450 enzyme KM (␮M) CYP102A1 CYP102A3 CYP102A1 CYP102A3 kcat (s−1 ) CYP102A1 CYP102A3 CYP102A1 CYP102A3

12-pNCA

10-pNCA

8-pNCA

WT WT mutant mutant

7.2 10.4 15.1 6.5

42.0 24.5 22.7 6.4

n.d. 61.8 197.6 31.9

WT WT mutant mutant

1.7 3.1 2.2 1.3

9.7 1.5 5.9 1.3

n.a. 0.4 4.3 3.5

2.28 × 105 6.4 × 104 2.60 × 105 2.03 × 105

n.a. 6 × 103 2.2 × 104 1.1 × 104

kcat /KM (s−1 M−1 ) CYP102A1 WT CYP102A3 WT CYP102A1 mutant CYP102A3 mutant

2.44 2.98 1.50 2.00

× × × ×

105 105 105 105

n.d.: no enzyme activity detected; n.a.: not applicable. Table 2 Reaction rates of the B. subtilis CYP102A3 enzyme Substrate

Wild-type

F88V

S189Q

F88V S189Q

Palmitic acid Lauric acid Capric acid Caprylic acid ␤-Ionone Naphtalene Octane

521 586 50 – – – –

599 556 26 – – – –

414 538 149 – – – 52

787 767 2120 41 286 143 722

Values are given in nmol substrate/nmol P450/min, the reaction rates were calculated using an extinction coefficient of NADPH; ε = 6.22 mM−1 . Table 3 Product ratio and conversion rate for the (␻-n) hydroxylation of the respective fatty acids catalyzed by wild-type and mutant B. subtilis CYP102A3 Substrate

Mutant

Conversion rate

Position ␻-4

␻-3

␻-2

␻-1

␻

Capric acid

WT S189Q F88V S189Q

28 82 80

– – –

66 62 20

6 7 20

28 31 56

– – 4

Lauric acid

WT S189Q F88V S189Q

93 80 82

– – 10

57 57 42

35 35 37

8 8 11

– – –

Palmitic acid

WT S189Q F88V S189Q

92 89 66

– – –

32 34 63

33 48 34

35 18 3

– – –

Product ratios and conversion rates are given as percentage of total substrate.

O. Lentz et al. / Journal of Biotechnology 108 (2004) 41–49

47

Fig. 2. Co-solvent stability of CYP102A1 of B. megaterium and CYP102A3 of B. subtilis in a reaction solution containing different amounts of DMSO.

189 has only minor effects on fatty acid conversion. With the second mutation at position 88, which is much closer to the heme, additional hydroxylations in ␻- (capric acid) and ␻-4 position (lauric acid) appeared. The mutation enlarged the reaction pocket and thus allowed for a larger number of substrate conformations. The analysis of the palmitic acid reaction mixture revealed a number of by-products such as keto- or keto-hydroxypalmitic acids. This was also observed by Schneider et al. (1998) for CYP102A1. If one compares the substrate specificities of the wild-type enzyme and the double mutant, it is obvious that the double mutant hydroxylated shorter substrates with higher conversion rates. The F88V S189Q mutant converts 80% of capric acid and only 66% of palmitic acid whereas the wild-type enzyme shows a 28 and 92% conversion rate, respectively. The wild-type and mutant CYP102A3 enzymes have higher fatty acid conversion rates than those of CYP102A1, which range between 38 and 57% (Lentz et al., 2001). The conversion of octanoic acid is severely impaired. The reason for this observation could be the long hydrophobic channel which reaches from the fatty acid recognition site (R47) at the entrance to the central heme. Octanoic acid may

be too polar to pass this region. Ost and co-workers (2000) were able to overcome this problem by introducing another fatty acid recognition/binding site inside the channel at residue 75 of the CYP102A1 enzyme. In many cases, the conversion rates are well below 100%. GLC analysis indicates that the reason is incomplete transformation of the educt, although observation of the NADPH consumption during the catalyzed reaction showed complete oxidation of this cofactor. As we did not detect any hydrogen peroxide or any by-products, water was obviously formed during the catalytic cycle due to abortive oxygen reduction (Guengerich, 2000). 3.3. Hydroxylation of other substrates In contrast to the wild-type enzyme, the double mutant F88V S189Q is also able to convert medium-chain alkanes and aromatic compounds (e.g. naphtalene; Table 2). The TLC analysis method performed after naphthalene hydroxylation showed 1-naphtol as the sole product. This was also found with CYP102A1, which also forms 1-naphtol (Appel et al., 2001).

48

O. Lentz et al. / Journal of Biotechnology 108 (2004) 41–49

Table 4 Product ratio and conversion rate for (␻-n) hydroxylation of octane catalyzed by wild-type and mutant B. subtilis CYP102A3 Conversion rate Compound Position WT S189Q F88V S189Q

– 43 49

Peak 1

Peak 2

Peak 3

4-Octanol ␻-3 – 46 32

3-Octanol ␻-2 – 41 68

2-Octanol ␻-1 – 13 –

Product ratios and conversion rates are given as percentage of total substrate.

Octane was converted by the CYP102A3 mutants with ratios of 43% (S189Q) and 49% (F88V S189Q), respectively. In contrast, mutant F88V was not able to hydroxylate octane. The reaction mixture of the S189Q mutant contained 2-, 3- and 4-octanol as products, whereas the reaction mixture of the F88V S189Q mutant only contained 3- and 4-octanol (Table 4). The absence of 2-octanol was not expected, since a wider reaction pocket should also lead to hydroxylation in this position. 1-Octanol was never produced, neither with the S189Q mutant nor with the F88V S189Q mutant. n-Octane is oxidized at a similar rate as lauric acid, although the overall conversion rate is significantly lower (49% versus 82%, F88V S189Q mutant), which indicates that the uncoupling rate is significantly higher for n-octane as substrate. 3.4. Conclusions and outlook Although B. megaterium CYP102A1 and B. subtilis CYP102A3 are very similar (77%) on the protein level, there are still interesting differences with regard to stability in the presence of the co-solvent DMSO, substrate specificity and conversion rates. In the case of the F88V S189Q mutant of CYP102A3, the oxidation of capric acid yielded a small amount of ␻-hydroxylated product. This has never been observed with CYP102A1 mutants. The substrate specificities of the F88V and S189Q mutants are almost equal to that of the wild-type enzyme, only a very slow octane transformation ability appeared (in the case of S189Q). This leads to the conclusion that both mutations (F88V and S189Q) are necessary to get an enzyme variant with broadened substrate spectrum.

The regioselectivity is an important aspect of the observed hydroxylations. Both enzymes do not hydroxylate at a specific single site (with the exception of the bulky naphthalene). Up to the present, there is only one publication which describes the mutagenesis of CYP102A1 which was able to produce an enzyme variant (F87A) that hydroxylated lauric acid almost exclusively in the ␻-position (Oliver et al., 1997). However, recent investigations could not confirm these results. Only subterminal hydroxylation with mutant F87A was found by us and other groups (Cirino and Arnold, 2002). High regioselectivity of P450 enzymes is attractive for industrial applications, the generation of regioselective mutants thus of great interest. Gene-shuffling based on CYP102A1 and including CYP102A3 might lead to novel enzyme mutants with enhanced temperature stability, co-solvent stability and altered regiospecificity. This aspect is currently under study.

Acknowledgements We gratefully acknowledge the financial support of the BASF AG.

References Appel, D., Lutz-Wahl, S., Fischer, P., Schwaneberg, U., Schmid, R.D., 2001. A P450 BM-3 mutant hydroxylates alkanes, cycloalkanes, arenes and heteroarenes. J. Biotechnol. 88, 167– 171. Boddupalli, S.S., Estabrook, R.W., Peterson, J.A., 1990. Fatty acid monooxygenation by cytochrome P450 BM-3. J. Biol. Chem. 265 (8), 4233–4239. Cirino, P.C., Arnold, F.H., 2002. Regioselectivity and activity of cytochrome P450 BM-3 and mutant F87A in reactions driven by hydrogen peroxide. Adv. Synth. Catal. 344 (9), 932–937. Daff, S.N., Chapman, S.K., Turner, K.L., Holt, R.A., Govindaraj, S., Poulos, T.L., Munro, A.W., 1997. Redox control of the catalytic cycle of flavocytochrome P450BM3. Biochemistry 36, 13816–13823. Guengerich, F.P., Gillam, E.M., Shimada, T., 1996. New applications of bacterial systems to problems in toxicology. Crit. Rev. Toxicol. 26, 551–583. Guengerich, F.P., 2000. Common and uncommon cytochrome P450 reactions related to metabolism and chemical toxicity. Chem. Res. Toxicol. 14 (6), 611–650. Gustafsson, M.C., Palmer, C.N., Wolf, C.R., von Wachenfeldt, C., 2001. Fatty-acid-displaced transcriptional repressor, a

O. Lentz et al. / Journal of Biotechnology 108 (2004) 41–49 conserved regulator of cytochrome P450 102 transcription in Bacillus species. Arch. Microbiol. 176 (6), 459–464. Jork, H., Funk, W., Fischer, W., Wimmer, H., 1993. Dünnschichtchromatographie: Reagenzien und Nachweismethoden, Band 1b. VCH Verlagsgesellschaft, Weinheim, Germany, pp. 262–265. Juchau, M.R., 1990. Substrate specificities and functions of the P450 cytochromes. Life Sci. 47, 2385–2394. Lee, T.R., Hsu, H.P., Shaw, G.C., 2001. Transcriptional regulation of the Bacillus subtilis bscR-CYP102A3 operon by the BscR repressor and differential induction of cytochrome CYP102A3 expression by oleic acid and palmitate. J. Biochem. 130 (4), 569–574. Lentz, O., Li, Q.-S., Schwaneberg, U., Lutz-Wahl, S., Fischer, P., Schmid, R.D., 2001. Modification of the fatty acid specificity of cytochrome P450 BM-3 from Bacillus megaterium by directed evolution: a validated assay. J. Mol. Catal. B 15, 123– 133. Matsunaga, I., Ueda, A., Fujiwara, N., Sumimoto, T., Ichihara, K., 1999. Characterization of the ybdT gene product of Bacillus subtilis: Novel fatty acid ␤-hydroxylating cytochrome P450. Lipids 34 (8), 841–846. Nelson, D.R., Koymans, L., Kamataki, T., Stegeman, J.J., Feyereisen, R., Waxman, D.J., Waterman, M.R., Gotoh, O., Coon, M.J., Estabrook, R.W., Gunsalus, I.C., Nebert, D.W., 1996. P450 superfamily: update on new sequences, gene mapping, accession numbers and nomenclature. Pharmacogenetics 6, 1–42. Oliver, C.F., Modi, S., Sutcliffe, M.J., Pimrose, W.U., Lian, L.Y., Roberts, G.C., 1997. A single mutation in cytochrome P450 BM3 changes substrate orientation in a catalytic intermediate

49

and the regiospecificity of hydroxylation. Biochemistry 36 (7), 1567–1572. Omura, T., Sato, R.J., 1964. The carbon monoxide binding pigment of liver microsomes. J. Biol. Chem. 239, 2370–2378. Ost, T.W.B., Miles, C.S., Murdoch, J., Cheung, Y.-F., Reid, G.A., Chapman, S.K., Munro, A.W., 2000. Rational re-design of the substrate binding site of flavocytochrome P450 BM3. FEBS Lett. 486, 173–177. Schneider, S., Wubbolts, M.G., Oesterhelt, G., Sanglard, D., Witholt, B., 1998. Controlled regioselectivity of fatty acid oxidation by whole cells producing cytochrome P450 BM-3 monooxygenase under varied dissolved oxygen concentrations. Biotechnol. Bioeng. 64, 3. Schwaneberg, U., Schmidt-Dannert, C., Schmitt, J., Schmid, R.D., 1999. A continuous spectrophotometric assay for P450 BM-3, a fatty acid hydroxylating enzyme and its mutant F87A. Anal. Biochem. 269, 359–366. Schwaneberg, U., 1999. Entwicklung Eines In Vitro Cytochrom P450 CYP102 (BM-3) Enzym-Membran-Reaktors. Dissertation, Universität Stuttgart. Stegeman, J.J., Lech, J.J., 1991. Cytochrome P-450 monooxygenase systems in aquatic species: carcinogen metabolism and biomarkers for carcinogen and pollutant exposure. Environ. Health Perspect. 90, 101–109. Truan, G., Komandla, M.R., Falck, J.R., Peterson, J.A., 1999. P450BM-3: absolute configuration of the primary metabolites of palmitic acid. Arch. Biochem. Biophys. 366, 192–198. Tsotsou, G.E., Cass, A.E.G., Gilardi, G., 2002. High throughput assay for cytochrome P450 BM3 for screening libraries of substrates and combinatorial mutants. Biosens. Bioelectr. 17, 119–131.