Products from Enzyme-Catalyzed Oxidations of Norcarenes

Abstract

Recent studies revealed that norcarane (bicyclo[4.1.0]heptane) is oxidized to 2-norcarene (bicyclo[4.1.0]-hept-2-ene) and 3-norcarene (bicyclo[4.1.0]hept-3-ene) by iron-containing enzymes and that secondary oxidation products from the norcarenes complicate mechanistic probe studies employing norcarane as the substrate (Newcomb, M.; Chandrasena, R. E. P.; Lansakara-P., D. S. P.; Kim, H.-Y.; Lippard, S. J.; Beauvais, L. G.; Murray, L. J.; Izzo, V.; Hollenberg, P. F.; Coon, M. J., accompanying article). In the present work, the product profiles from oxidations of 2-norcarene and 3-norcarene by several enzymes were determined. Most of the products were identified by GC and GC-mass spectral comparison to authentic samples produced independently; in some cases, stereochemical assignments were made or confirmed by 2-D NMR analysis of the products. The enzymes studied in this work were four cytochrome P450 enzymes, CYP2B1, CYPΔ2E1, CYPΔ2E1 T303A, and CYPΔ2B4, and three diiron-containing enzymes, soluble methane monooxygenase (sMMO) from Methylococcus capsulatus (Bath), toluene monooxygenase (ToMO) from Pseudomonas stutzeri OX1, and phenol hydroxylase (PH) from Pseudomonas stutzeri OX1. The oxidation products from the norcarenes identified in this work are 2-norcaranone, 3-norcaranone, syn- and anti-2-norcarene oxide, syn- and anti-3-norcarene oxide, syn- and anti-4-hydroxy-2-norcarene, syn- and anti-2-hydroxy-3-norcarene, 2-oxo-3-norcarene, 4-oxo-2-norcarene, and cyclohepta-3,5-dienol. Two additional, unidentified oxidation products were observed in low yields in the oxidations. In matched oxidations, 3-norcarene was a better substrate than 2-norcarene in terms of turnover by factors of 1.5 to 15 for the enzymes studied here. The oxidation products found in enzyme-catalyzed oxidations of the norcarenes are useful for understanding the complex product mixtures obtained in norcarane oxidations.

Mechanistic probes have been used for years to reveal details about reaction mechanisms in chemistry and biology. The concept of a mechanistic probe study is that a short-lived intermediate can be revealed by a characteristic rearrangement of a probe substrate that is observed in the reaction products. One compound used as a probe substrate for studies of enzyme-catalyzed oxidations is norcarane, bicyclo[4.1.0]heptane (1). During the course of probe studies with norcarane in our laboratories,¹ we observed a large number of minor oxidation products that appeared to be derived from norcarane as judged by GC retention times and mass spectra. Eventually, we realized that, in addition to expected alcohol products, the enzymes were oxidizing norcarane to both 2-norcarene (2) and 3-norcarene (3), and that these alkenes were further oxidized to give small amounts of secondary oxidation products.² Details of the norcarane oxidation studies are reported in the accompanying paper.³ In this paper, we report studies of the oxidations of both 2-norcarene and 3-norcarene by several iron-containing enzymes. Most of the products from oxidations of these substrates were synthesized independently, providing firm identifications. The results of this work are used in the accompanying paper³ to aid in identification of the many products from enzyme-catalyzed oxidations of norcarane.

Results and Discussion

Products from norcarene oxidations

Iron-containing enzymes oxidize norcarenes to give a complex mixture of products. Chart 1 shows a collection of oxidation products that we have characterized. Compounds 4-13 and 16 are singly oxidized products of the norcarenes, whereas ketones 14 and 15 are doubly oxidized products. Most of the compounds in Chart 1 were synthesized independently so that the products from the enzyme-catalyzed oxidations could be identified by GC and GC-mass spectral comparisons. In some cases, the stereochemical assignments in the compounds were not certain, and we confirmed the stereochemistry with NMR experiments. In addition to the compounds shown in Chart 1, two other minor products were detected in enzyme-catalyzed oxidation reactions of 2 and 3 but were not fully characterized.

Chart 1.

Ketones 4 and 5 were prepared by oxidation of the corresponding alcohols (Scheme 1) as previously reported by Chan and Rickborn.⁴

Scheme 1.

The anti-epoxide 7 was previously reported to be formed by reaction of 2-norcarene with peroxyacetic acid,⁵ and we prepared compound 7 by reaction of 2-norcarene with m-chloroperoxybenzoic acid (mCPBA) (Scheme 2). After chromatographic purification, compound 7 was judged to be > 95% pure by NMR spectroscopy. The structure of 7 was confirmed by 1D NOE difference experiments that showed the enhancements in Figure 1. The chemical shift assignments for H_7-endo and H_7-exo were made on the basis of the coupling constants between these two protons and the H₆ proton, which are also shown in Figure 1. These coupling constants are similar to those for the analogous protons in norcarane.⁶

Scheme 2.

Figure 1.

Selected coupling constants (top) and NOE enhancements (bottom) for epoxide 7. The NOE values are for irradiation at H_C1 and H_C2.

When the crude product from the preparative reaction for 7 was analyzed by GC, we observed a minor product in ca. 6% relative yield with a shorter retention time than 7 on a low polarity GC column. The mass spectrum of the minor product was the same as that of epoxide 7. The same product was detected in enzyme-catalyzed oxidations of 2-norcarene. We conclude that the minor product is the syn-epoxide 6 (Scheme 2).

The diastereomeric epoxides 8 and 9 were reported previously.⁷ These compounds were formed in comparable amounts by the reaction of 3-norcarene with mCPBA, and they were separated by chromatography. Reactions of the isolated epoxides with LiNEt₂ gave the known allylic alcohols 10 and 11, respectively (Scheme 3).⁷

Scheme 3.

Oxidation of 3-norcarene with chromium trioxide⁸ gave the known α,β-unsaturated ketone 14 as the major product and the previously unreported isomer 15 as a minor product (Scheme 4). The reaction of ketone 14 with DIBAL-H gave the allylic alcohols 12 and 13, both new compounds, and the previously known⁹ alcohol 16 (Scheme 4).

Scheme 4.

The stereochemical assignments for allylic alcohols 12 and 13 were made from coupling constant data, the results of 1D NOE difference experiments, and the results of molecular modeling studies. Table 1 summarizes the data. For both isomers, H_endo and H_exo at C7 were assigned from the coupling constants between these protons and the bridgehead proton at C1. Molecular mechanics modeling of the two isomers indicated that the distances between the endo and exo protons on C7 and the proton on C2 were shorter for isomer 12 than the respective distances for isomer 13. A series of NOE experiments was performed where the C7 protons on each isomer were selectively irradiated and the absorbance enhancements for the protons on C2 were measured. The NOE enhancements of the H_C2 signal were larger for irradiation of both of the C7 protons in isomer 12, consistent with expectations from the molecular modeling results.

Table 1.

Data Used for Structural Assignments of Isomers 12 and 13.

Property or Feature	12	13
HC1-HC7endo coupling constant (Hz)	5.0	5.0
HC1-HC7exo coupling constant (Hz)	8.0	8.7
HC7exo-HC7endo coupling constant (Hz)	4.5	4.8
HC2-HC7endo distance (Å)a	2.90	3.83
HC2-HC7exo distance (Å)a	3.66	4.34
HC7endo NOE enhancement at HC2 (%)	2.85	1.52
HC7exo NOE enhancement at HC2 (%)	2.50	0

^aDistance between protons in the minimum energy structure calculated by MM2 modeling.

For enzyme-catalyzed oxidation reactions of the norcarenes, the products were quantitated by flame-ionization GC and identified by GC-mass spectral analysis. GC retention times for compounds 4-16 are shown schematically in Figures 2A and 2B. Figure 2B also shows a portion of a typical GC result for a P450 enzyme-catalyzed oxidation of 3-norcarane. The GC analyses were obtained with a low polarity column, which gave the best separations. Mass spectral fragmentation patterns for these compounds are in the Supporting Material for the accompanying paper.³

Figure 2.

(A) GC retention times for compounds 4-16 on a low polarity column (5% phenyl silicone) at 80 °C. (B) Expansion of the congested region of the trace in A with the GC trace of products from oxidation of 3-norcarene by CYPΔ2E1 (gray line) superimposed. The compound numbers are shown above lines that indicate the retention times for each compound. Allylic alcohols 10 and 13 had the same retention times to within 0.01 min.

As is apparent in Figure 2, several of the oxidation products from the norcarenes have similar GC retention times. Products 10 and 13 were not resolved under our GC conditions, and the similar retention times for some products resulted in overlapping peaks. In principle, the composition of merged peaks can be determined from GC-mass spectral analysis, but this task is not necessarily simple nor reliable because mass spectra of most of the oxidation products are similar. For norcarene oxidations discussed below, we report the sum of products 10 and 13.

Although we prepared many possible oxidation products from 2- and 3-norcarene, other oxidation products were detected in some of the enzyme-catalyzed oxidation reactions of these substrates. One of the unidentified products appears to have a molecular ion at m/z = 124 (Figure 3A), which is consistent with incorporation of two oxygen atoms and loss of four protons from substrate to give a highly oxidized species with formula C₇H₈O₂. This product is assigned as unknown compound A. It co-eluted with compound 16 and was detected in product mixtures from oxidations of both 2-norcarene and 3-norcarene. In some reactions, unknown A apparently was formed in yields comparable to those of some of the primary oxidation products from the norcarenes.

Figure 3.

(A) Mass spectral fragmentation pattern for unknown product A. (B) Mass spectral fragmentation pattern for authentic product 9. (C) Mass spectral fragmentation pattern observed for unknown product B, which eluted at the retention time expected for epoxide 9, from the product mixture obtained in the CYPΔ2E1-catalyzed oxidation of 2-norcarene.

Another unknown oxidation product was apparent in the reactions of 2-norcarene. Epoxide 9 should not be formed from oxidation of 2-norcarene, but we observed a product with a GC retention time the same as that of epoxide 9 in the product mixtures from most of the oxidations of 2-norcarene. This product, labeled as unknown compound B, not only has the same retention time as compound 9 but also a similar mass spectral fragmentation pattern. Figure 3B shows the fragmentation pattern for authentic epoxide 9, and Figure 3C shows the fragmentation pattern observed at the same retention time as compound 9 from the oxidation of 2-norcarene catalyzed by CYPΔ2E1. The differences in the mass spectral fragmentation patterns are subtle. In authentic 9, the fragment with m/z = 91 is smaller than that with m/z = 95, and the fragment with m/z = 109 is greater than that with m/z = 110. In unknown B, the relative intensities for these two pairs are reversed. The similarity of these fragmentation patterns illustrates the great difficulty one has in attempting to identify a product by the use of a few fragment peaks obtained from single ion monitoring (SIM) analysis of the GC trace. If SIM analysis with limited ion channels were used for the unknown product B (Figure 3C), then one might mistakenly identify that unknown compound as product 9.

Enzyme-Catalyzed Oxidations of Norcarenes

We studied the oxidations of 2-norcarene and 3-norcarene with a representative spectrum of iron-containing enzymes that included four mammalian hepatic cytochrome P450 enzymes and three diiron enzymes. The P450 enzymes, which were expressed in E. coli and purified, were CYP2B1,¹⁰ CYPΔ2B4,¹¹ CYPΔ2E1,¹² and CYPΔ2E1 T303A.¹² The latter three P450s contain short deletions at the N-terminal ends of the enzymes. CYP2B1 (rat) and CYP2B4 (rabbit) are induced in animals by phenobarbital treatment. CYP2E1 is a constitutive and ethanol-inducible P450 linked to alcohol-related liver disease. In the CYPΔ2E1 T303A mutant, a highly conserved threonine located near the active site of the enzyme was replaced with alanine; this conserved Thr is thought to be involved in the proton delivery system in the activation of the P450 enzymes.¹² The diiron enzymes studied were the soluble methane monooxygenase (sMMO) from Methylococcus capsulatus (Bath), one of the more extensively studied diiron enzymes,^13-15 toluene monooxygenase (ToMO) from Pseudomonas stutzeri OX1, and phenol hydroxylase (PH) from Pseudomonas stutzeri OX1. As discussed in the accompanying paper,³ results for norcarane oxidations by the iron-containing enzymes studied in this work and closely related enzymes are available.

The substrates 2-norcarene (2) and 3-norcarene (3) were prepared by literature methods involving cyclopropanation of 1,3-cyclohexadiene and 1,4-cyclohexadiene, respectively.^5,16 Purity of the substrate is an issue when quantification of small amounts of products are necessary, and we used preparative GC as the final step in the purification of 2 and 3. The products used in the enzyme studies contained no detectable impurities by NMR spectroscopy and were 99.7% homogeneous (2) and 99.3% homogeneous (3) by analytical GC.

Enzyme-catalyzed oxidations of the norcarenes involved three or four individual experiments with each enzyme and substrate. The norcarenes are good substrates for most of enzymes studied here. For the sMMO enzyme, turnover was limited, probably because the active site is optimized for oxidation of methane, and 10 nmol of sMMO was used in each experiment. The oxidation products were extracted into methylene chloride, and the product mixtures were analyzed by GC for yield quantifications and by GC-mass spectroscopy for identification. Table 2 contains the results. Figure 4 shows representative GC traces for oxidations by a P450 enzyme and two diiron enzymes.

Table 2.

Product Yields from Oxidations of 2-Norcarene and 3-Norcarene.^a

Enzymeb	4	5	6	7	8	9+B	10+13	11	12	14	15	16+A	total
2-norcarene
CYPΔ2E1	NDc	5.62	0.72	0.72	NDc	1.95	26.90	1.35	3.04	4.21	4.35	2.71	51.57
CYPΔ2E1 T303A	NDc	3.89	1.47	1.44	NDc	0.17	8.84	1.03	3.72	6.01	6.19	0.26	33.02
CYP2B1	0.08	7.11	2.94	3.65	NDc	0.32	8.35	0.62	0.54	0.18	0.14	0.91	24.84
CYPΔ2B4	0.15	14.46	3.53	5.02	NDc	0.47	25.38	1.37	1.11	0.27	0.24	2.06	54.06
sMMO	0.08	9.56	<0.08	<0.05	NDc	NDc	2.03	1.08	0.09	0.13	0.15	0.62	13.74
ToMO	0.03	2.20	<0.01	0.09	NDc	0.19	0.12	0.30	0.04	0.17	0.09	NDc	3.23
PH	0.01	0.74	<0.01	1.54	NDc	0.03	0.05	0.03	0.06	0.07	0.06	0.03	2.62
3-norcarene
CYPΔ2E1	NDc	NDc	NDc	NDc	0.58	1.26	17.02	1.26	46.08	4.36	3.48	7.26	81.30
CYPΔ2E1 T303A	NDc	NDc	NDc	NDc	12.29	4.86	8.59	1.79	30.72	6.78	7.20	2.52	74.75
CYP2B1	NDc	0.21	NDc	NDc	13.90	5.85	23.16	1.72	19.30	0.52	0.41	4.13	69.20
CYPΔ2B4	NDc	0.87	NDc	NDc	34.77	11.00	75.11	7.98	82.63	1.07	0.69	16.86	230.98
sMMO	0.15	0.96	NDc	NDc	<0.15	1.66	50.4	0.55	44.48	1.84	0.17	17.74	117.95
ToMO	0.04	0.55	NDc	0.05	1.03	41.61	1.19	0.55	0.50	0.38	0.33	0.61	46.84
PH	0.03	NDc	NDc	0.06	0.39	18.86	0.20	0.08	0.09	0.13	0.12	0.05	20.01

^aProduct yields in nmols; averages of 3 or 4 experiments.

^bCYP = cytochrome P450, sMMO = soluble methane monooxygenase from M. capsulatus (Bath), ToMO= toluene monooxygenase from from P. stutzeri OX1, PH = phenol hydroxylase from from P. stutzeri OX1; oxidation reactions employed 0.5 nmol of P450 enzyme, 1.0 nmol of ToMO or PH, or 10 nmol of sMMO.

^cND = not detected.

Figure 4.

Portions of GC traces of products from oxidation of 2-norcarene (A) and 3-norcarene (B) with toluene monooxygenase from P. stutzeri OX1 (ToMO), soluble methane monooxygenase from M. capsulatus (Bath) (sMMO), and cytochrome P450 2B1 (CYP2B1). The identities of some of the products are indicated with the compound numbers.

From the total amounts of oxidation products in the final column, 3-norcarene was determined to be a better substrate than 2-norcarene for all of the enzymes studied. For CYPΔ2B4, 5 times as much product was formed from 3-norcarene as from 2-norcarene. For the di-iron enzymes, about an order of magnitude more product was formed in the oxidations of 3-norcarane than in the oxidations of 2-norcarene.

Allylic alcohols 10-13 were the major products formed with all of the P450 enzymes, although ketone 5 was formed in considerable yields in the P450 oxidations of 2-norcarene. For the diiron enzyme sMMO, allylic alcohols were the major product from 3-norcarene, whereas ketone 5 was the major product from 2-norcarene. Interestingly, the diiron enzymes ToMO and PH gave high yields of epoxide 9 in oxidation reactions of 3-norcarene but low yields of epoxides from 2-norcarene. We noted above that unknown product B overlaps with epoxide 9 on GC, but mass spectral analyses confirmed that epoxide 9 was the major component obtained in the oxidations of 3-norcarene.

The oxidation patterns observed for the norcarenes undoubtedly reflect substrate binding preferences in the active site superimposed on the high reactivities expected for the double bonds and the allylic hydrogen atoms. Allylic C-H bonds have bond dissociation energies (BDE's) that are about 12 kcal/mol less than those of analogous C-H bonds in saturated systems,^17,18 and, for the C-H bonds in 3-norcarene, the BDE will be smaller due the adjacent cyclopropyl group.¹⁹

The product distributions for norcarene oxidations found in this work are useful for understanding and identifying the many products found in enzyme-catalyzed oxidations of norcarane, which results in formation of norcarenes by desaturase-type reactions.³ One of the more important observations is that the yields of products 16 and A were relatively large for the P450 and sMMO enzymes. Those two products co-elute with the “radical-derived” rearranged alcohol from oxidations of norcarane in GC analyses and interfere with quantification of the latter, as discussed in the accompanying paper.³

Experimental Section

General

The preparations of 2-norcarane, 3-norcarane, and authentic samples of oxidation products are described in Supporting Information. NMR spectra were recorded in CDCl₃ at 300 or 500 MHz. Proton correlation was resolved with bidimensional COSY experiments using the sequence cosygpqf, and stereochemistry was determined by monodimensional NOE difference experiments. 2-Norcarene and 3-norcarene were purified by preparative GC (10% SE-52 on 60/80 Chromosorb W, 1/4 in × 8 ft column) at 100 °C (isothermal), and the purities of the samples were determined by GC (DB-5, 40 °C) using flame ionization detection. All yields and purities of other authentic samples were determined by GC (DB-5) using flame ionization detection and GC-MS (HP-5MS 5% phenylmethyl siloxane column) with electron impact (EI) ionization. Details of sample preparations are in Supporting Information. Mass spectra are in the Supporting Information for the accompanying article.³

Cytochrome P450-catalyzed oxidations

Cytochrome P450 2B1,¹⁰ P450 Δ2B4,¹¹ P450 Δ2E1,¹² and P450 Δ2E1 T303A¹² enzymes and P450 reductase¹⁰ were expressed in E.coli. and purified as previously described. A stock solution of substrate was prepared from substrates 2 or 3 (2 mg in 100 μL of methanol) and DLPC (2 mg) in 1.0 mL of 50 mM phosphate buffer, pH 7.4; the mixture was sonicated and stored at 0 °C. A stock solution of NADPH was prepared by addition of 15 mg of NADPH to 375 μL of the same buffer, and this solution was also stored at 0 °C. Reaction mixtures were prepared by mixing 0.5 nmol of enzyme, 1.0 nmol of P450 reductase, and 30 μL of DLPC at 0 °C and allowed stand for 10 min. The above buffer was added to the mixture to bring the total volume to 1.89 mL, and the mixture was allowed to stand at 0 °C. After 10 min, 10 μL of substrate stock solution was added, and the reaction mixture was allowed to equilibrate at 26 °C for 5 min. The reaction was initiated by adding 100 μL of the NADPH stock solution. Reactions were maintained at 26 °C for 30 min and then quenched by addition of 1 mL of CH₂Cl₂. The organic phase was separated, and the aqueous phase was extracted with CH₂Cl₂ (2 × 1 mL). The combined organic layer was dried with MgSO₄ and filtered. An internal standard of 1-phenyl-1-propanol was added, and the mixture was concentrated under a stream of nitrogen. The mixture was analyzed by GC (0.32 mm × 30 m capillary column, 5% phenyl silicone bonded phase) and GC-mass spectroscopy (0.25 mm × 30 m, capillary column, 5% phenyl silicone bonded phase).

Soluble methane monooxygenase-catalyzed oxidations

The hydroxylase protein (MMOH) was purified from M. capsulatus (Bath) as previously described.^20,21 The regulatory (MMOB) and reductase (MMOR) components were expressed recombinantly in E. coli and purified as reported elsewhere.^22,23 In a typical reaction, a mixture of 10 nmol of MMOH, 20 nmol of MMOB and 5 nmol of MMOR was diluted to a volume of 400 μL with 25 mM potassium phosphate buffer (pH = 7.0). The enzyme was incubated for 1 min at 45 °C, and 10 μL of a solution of substrate (2 mg of substrate in 100 μL of methanol) was added. The oxidation was initiated by the addition of 1.6 μmol NADH. The mixture was gently shaken at 45 °C for 25 min and then extracted with CH₂Cl₂ (3 × 2 mL). The reaction mixtures were worked up and analyzed as described above.

Toluene monooxygenase-catalyzed oxidations

Plasmids containing the genes for the toluene/o-xylene monoxygenase hydroxylase (ToMOH), coupling protein (ToMOB), and Rieske protein were kindly supplied by Prof. Alberto Di Donato. The proteins were expressed in E. coli and purified as described elsewhere.^24,25 The toluene/o-xylene monooxygenase NADH oxidoreductase (ToMOR) was provided by Dr. Valeria Cafaro. The specific activity on phenol was 1.2(1) × 10³ mU/mg of hydroxylase, and the iron content for ToMOH was 4.3(2) as determined by methods reported elsewhere.^24,25 In a typical reaction, a mixture of 1 nmol of ToMOH, 66.57 nmol of ToMOB, 20 nmol of Rieske protein, and 0.1 nmol of ToMOR was diluted to a volume of 2 mL with 20 mM potassium phosphate buffer (pH = 7.4). The mixture was allowed to stand in an ice bath for 10 min. Substrate (10 μL of a solution of 2 mg of 2-norcarene or 3-norcarene in 100 μL of methanol) was added. The mixture was incubated at 26 °C for 10 min while shaking. The oxidation was initiated by the addition of NADH buffer solution; the final concentration of NADH was 1 mM. The mixture was gently shaken at 26 °C for 30 min and then extracted with CH₂Cl₂ (3 × 2 mL). The combined organic phase was dried (MgSO₄) and filtered. The mixtures were analyzed by GC and GC-MS as above.

Phenol hydroxylase-catalyzed oxidations

The hydroxylase component PH/H of phenol hydroxylase complex was purified from E. coli, strain JM109 expressing the entire ph operon cloned into vector pGEM3Z.²⁶ The hydroxylase moiety was purified with a slight modification of a reported procedure.²⁷ The iron content was 4.2(3) per mol protein, and the specific activity on phenol was 515 (SD of 44) mU/mg of hydroxylase, determined as reported elsewhere.²⁴ The coupling (PHM) and reductase (PHP) components were expressed recombinantly in E. coli, strain BL21DE3, and purified as previously described.²⁷ In a typical reaction, a mixture of 1 nmol of PH/H (hydroxylase) and 2 nmol of PHM (coupling protein) was diluted to a volume of 2 mL with 0.1 M Tris-HCl buffer (pH = 7.5). Substrate (10 μl of a solution of 2 mg of 2-norcarene or 3-norcarene in 100 μL of methanol) and 1 mM NADH were added. The mixture was incubated at 25 °C for 5 min while shaking. The oxidation was initiated by the addition of 2 nmol of PHP (oxidoreductase). The mixture was gently shaken at 25 °C for 30 min and then extracted with CH₂Cl₂ (3 × 2 mL). The combined organic phase was dried (MgSO₄) and filtered. The mixtures were analyzed by GC and GC-MS as above.

Supplementary Material

1

Supporting Information Available: Synthetic details and NMR spectra. This material is available free via the Internet at http://pubs.acs.org.