Abstract
A total of 215 Streptomyces strains were screened for their capacity to regio- and stereoselectively hydroxylate β- and/or α-ionone to the respective 3-hydroxy derivatives. With β-ionone as the substrate, 15 strains showed little conversion to 4-hydroxy- and none showed conversion to the 3-hydroxy product as desired. Among these 15 Streptomyces strains, S. fradiae Tü 27, S. arenae Tü 495, S. griseus ATCC 13273, S. violaceoniger Tü 38, and S. antibioticus Tü 4 and Tü 46 converted α-ionone to 3-hydroxy-α-ionone with significantly higher hydroxylation activity compared to that of β-ionone. Hydroxylation of racemic α-ionone [(6R)-(−)/(6S)-(+)] resulted in the exclusive formation of only the two enantiomers (3R,6R)- and (3S,6S)-hydroxy-α-ionone. Thus, the enzymatic hydroxylation of α-ionone by the Streptomyces strains tested proceeds with both high regio- and stereoselectivity.
Ionones and their derivatives are important intermediates in the metabolism of terpenoids, e.g., in carotenoid biosynthesis, and have been isolated from many sources (1a, 11). Compounds with a trimethylcyclohexane building block constitute essential aroma elements in many plant oils and thus have attracted the attention of the flavor and fragrance industry (3). Further, ionone derivatives, e.g., 3-hydroxy-β-ionone, could prove valuable intermediates for the chemoenzymatic synthesis of carotenoids, e.g., for astaxanthin and zeaxanthin (5).
Microbial transformation of α- and/or β-ionone to a number of hydroxy and oxo derivatives has been reported for several fungal strains (2, 4, 8, 9, 18), mainly of the genus Aspergillus, but not for bacterial strains. 3-Hydroxy-α-ionone was observed, among other metabolites, when Cunninghamella blakesleeana ATCC 8688 (2) or Aspergillus niger JTS 191 (18) was used.
Many species of the order Actinomycetes are known to catalyze a broad spectrum of xenobiotic transformations. Several cytochrome P-450-dependent monooxygenases from Streptomyces strains, which catalyze the hydroxylation of a wide range of substrates, have been investigated on the molecular level (12) and thus provide an interesting potential as biocatalysts for specific hydroxylation reactions by recombinant techniques.
As a first step in this direction, we now report the screening of 215 Streptomyces strains for their capacity to hydroxylate β- and/or α-ionone to the respective 3-hydroxy derivatives in a regio- and stereoselective manner. The structure and stereochemistry of the main biotransformation product were characterized unequivocally by nuclear magnetic resonance (NMR) spectroscopy.
MATERIALS AND METHODS
Materials.
Substrates for the biotransformation reactions were purchased from Merck (racemic α-ionone) and Fluka (β-ionone), respectively. BASF AG (Ludwigshafen, Germany) kindly provided 3- and 4-hydroxy-β-ionone as references. Soybean meal, yeast extract, and nutrient broth were obtained from Difco.
Strains.
For the first screening round, 215 Streptomyces strains were selected from the collection of the Institute of Microbiology/Biotechnology, University of Tübingen, Tübingen, Germany. Streptomyces griseus ATCC 13273 was purchased from the American Type Culture Collection. Streptomyces hygroscopicus Lu 1537 was kindly provided by the BASF AG.
Media.
Three complex media and one synthetic medium were investigated for strain cultivation and bioconversion of ionones (all amounts given per liter): medium A (20 g of soybean meal, 20 g of mannitol [pH 7.5]), medium B (8 g of nutrient broth, 10 g of yeast extract, 5 g of glucose [pH 7.0]), medium C [5 g of (NH4)2SO4, 0.5 g of MgSO4 · 7H2O, 0.05 g of MnSO4 · H2O, 3.6 g of K2HPO4, 1.5 g of KH2PO4, 2 g of glucose, 0.2 g of FeSO4 · H2O, 10 mg of ZnSO4 · 7H2O, 3 mg of MnCl2 · 4H2O, 30 mg of H3BO3, 20 mg of CaCl2 · 6H2O, 1 mg of CuCl2 · 2H2O, 2 mg of NiCl2 · 6H2O, 3 mg of Na2MoO4 · 2H2O, 500 mg of Titriplex III (pH 7.0)], and medium D (15 g of glucose, 15 g of soybean meal, 5 g of corn steep liquor, 5 g of NaCl, 2 g of CaCl2 [pH 7.0]).
Initial screening for β-ionone bioconversion.
The first screening round for an overall ionone hydroxylation potential of the individual Streptomyces strains was performed with β-ionone as the substrate. Precultures (5 ml of medium A) were inoculated from slant agar stocks and incubated for 2 days at 28 to 30°C. Medium A (100 ml) was inoculated with 3 ml of the precultures and incubated at 28 to 30°C with shaking (150 rpm). β-Ionone (0.1% [wt/vol]) was added to each of the cultures after 2 days, and cultivation continued for 5 days. Cells were then separated by filtration. A 2-ml aliquot of the culture supernatant was extracted with ethyl acetate-hexane (3:2) or diethyl ether, and the extract was analyzed by thin-layer chromatography (TLC). Control cultivations were carried out analogously but without addition of β-ionone. The stability of β-ionone in medium A was tested under incubation conditions identical to those as employed for strain cultivation.
Bioconversion of α- and β-ionone.
The strains Streptomyces griseus ATCC 13273 and S. hygroscopicus Lu 1537, selected during the initial screening round for their ability to convert β-ionone, were employed for a second screening round under the same cultivation conditions. A 1-ml aliquot was withdrawn every other day, and the culture supernatant was extracted as described above and analyzed for conversion products by TLC and gas liquid chromatography (GLC). After 10 to 12 days of cultivation, the cultures were filtered, and the supernatant was subjected to the final product analysis. Bioconversion of α-ionone was performed as described above for β-ionone.
Extraction of conversion products from culture supernatant.
Larger volumes of culture supernatant (e.g., 100 ml) were extracted twice with 50 ml each of diethyl ether, and the extracts were washed with a saturated aqueous NaCl solution, dried with MgSO4 or Na2SO4, and then evaporated to yield the unused substrate and conversion products.
Analytical methods.
Samples for TLC analysis were spotted onto TLC plates (silica gel layer thickness, 0.2 mm; Silica Gel 60 F254; Merck), and the plates were developed with hexane-ethyl acetate (3:2). Product spots were visualized first by fluorescence quenching at 254 nm and then by spraying with a 2.5% (wt/vol) vanillin solution in 95% ethanol-H2SO4 and subsequent heating. Authentic β- and α-ionone and 3- (Rf, 0.37) and 4-hydroxy-β-ionone (Rf, 0.33) were employed as standards.
Gas chromatograms were run on a Carlo Erba MEGA 5300 gas chromatograph, equipped with a flame ionization detector (FID), a Spectra Physics Labnet Version 3.5 integrator system, and a 20-m-long glass capillary column coated with a chiral polysiloxane phase modified by chemically bonded α- and β-cyclodextrin (0.38 and 0.34%, respectively) (16) (temperature program, 100°C [1-min isotherm] with increases from 100 to 220°C [4°C/min]; forepressure, 4 × 104 Pa H2). Samples were applied to the gas chromatograph in CH2Cl2 solution.
NMR spectra were run in CDCl3 solution on a Bruker (Karlsruhe-Forchheim, Germany) ARX 500 spectrometer (nominal frequencies of 500.13 MHz for 1H and 125.77 MHz for 13C).
RESULTS AND DISCUSSION
Hydroxylation of β-ionone.
In the first screening round, 215 different Streptomyces strains (from the strain collection of the Institute of Microbiology/Biotechnology, University of Tübingen) were tested for their potential to hydroxylate β-ionone in position 3, i.e., for their ability to transform β-ionone to 3-hydroxy-β-ionone. Twenty Streptomyces strains were selected at random and cultivated on a small scale (50 ml) in four different media (medium A to medium D [see Materials and Methods for medium composition]) in the presence of β-ionone. Medium A proved best for both strain cultivation and β-ionone biotransformation and was therefore selected for all further cultivation studies. TLC analysis showed bioconversion of β-ionone to more-polar products, e.g., hydroxylated or oxygenated products, for 13 of the 215 Streptomyces strains.
These thirteen plus two additional Streptomyces strains, S. griseus (ATCC 13273), whose cytochrome P-450 monooxygenase system is well characterized (17), and S. hygroscopicus (Lu 1537), which is known to perform many biotransformations (1) were employed in further biotransformation studies with β-ionone as the substrate. β-Ionone conversion was monitored by TLC analysis (not shown); after 10 to 12 days of incubation, conversion of β-ionone came more or less to an end with most strains. Cultivations were stopped, and the product mixture was analyzed by GLC (Table 1).
TABLE 1.
Biotransformation of β- and α-ionone by various Streptomyces strains
Streptomyces straina | Conversion rate (%)b
|
||||
---|---|---|---|---|---|
β-Ionone | 4-Hydroxy-β-ionone | α-Ionone | 3-Hydroxy-α-ionone | Unidentified compoundsc | |
S. arenae Tü 495 | 66 | 33 | 35 | 54 | 1/12 |
S. antibioticus Tü 46 | 74 | 22 | 90 | 7 | 4/3 |
Tü 124 | 86 | 12 | 94 | 0 | 2/6 |
S. griseus ATCC 13273 | 88 | 10 | 49 | 50 | 2/1 |
S. griseus Tü 18 | 88 | 10 | 98 | 0 | 2/2 |
S. griseus Tü 781A | 90 | 8 | 98 | 0 | 2/2 |
S. griseus Tü 17 | 90 | 8 | 94 | <1 | 2/5 |
S. violaceoniger Tü 38 | 91 | 7 | 55 | 44 | 2/<1 |
S. exfoliatus Tü 1424 | 92 | 7 | 97 | 0 | 1/3 |
S. griseoviridis Tü 1963 | 91 | 6 | 81 | 0 | 2/9 |
S. antibioticus Tü 4 | 94 | 5 | 68 | 28 | 1/4 |
S. fradiae Tü 27 | 94 | 4 | 21 | 75 | 2/3 |
S. griseus Tü 16 | 94 | 5 | 98 | 0 | <1/2 |
S. tendae Tü 21 | 95 | 4 | 79 | 0 | <1/21 |
S. hygroscopicus Lu 1537 | 72 | <1 | 99 | 0 | 27/1 |
Streptomyces strain Tü and Lu designations refer to the strain collections of the Institute of Microbiology/Biotechnology (University of Tübingen) and the BASF AG, respectively.
Conversion rates were estimated by GLC analysis of the reaction mixture after 10 to 12 days of incubation. For calculating the product ratio in experiments based on β-ionone, FID values were added for the substrate β-ionone (retention time [Rt], 8.31), the 4-hydroxy product (Rt, 16.36), and one other, not yet structurally characterized compound (Rt, 13.71). In the case of the racemic substrate α-ionone (Rt, 6.24/6.52), FID values of the 3-hydroxy product (Rt, 14.92/15.36) and one other, not yet structurally characterized compound (Rt, 8.74) were added. Individual percentages are given with respect to these sum parameters. It corresponds to a 90% yield (wt/wt) of the initial substrate weight.
The two values are the value from the β-ionone experiment before the slash and the value from the α-ionone experiment after the slash.
Most Streptomyces strains showed low conversion of β-ionone (4 to 10% within 10 to 12 days of incubation), except for S. antibioticus Tü 46 and S. arenae (19 and 33%, respectively). None of the strains, however, converted the substrate to 3-hydroxy-β-ionone as desired. While fungal strains, such as A. niger JTS 191 (8) and IFO 8541 (4), reportedly yielded a complex mixture of β-ionone derivatives when employed under conditions optimized for β-ionone biotransformation, only one major hydroxylation product was formed with all but one of the Streptomyces strains. This could be unequivocally characterized as 4-hydroxy-β-ionone by both a complete NMR analysis (data not given) and by comparison with chemical shift data reported in the literature (13).
Bioconversion of α-ionone.
Our screening results suggest that selective enzymatic hydroxylation of β-ionone at C-3 is difficult if not impossible—a finding which is in line with all earlier reports on the microbial conversion of β-ionone. While stereochemical reasoning implies that the two methyl groups at C-1 should direct any oxidative attack toward C-3 rather than C-2, it is the electronic activation of the allylic hydrogens at C-4 by the C-5⩵C-6 double bond which governs the regiochemistry of β-ionone hydroxylation (Fig. 1). This view is supported by the fact that the main products of β-ionone oxidation found in cultures of Aspergillus were 2- and 4-hydroxy-β-ionone. In the isomeric α-ionone, however, it is C-3 which is in allylic position to the double bond and thus should be most susceptible to oxidative attack. In fact, among the 13 most promising strains from our primary screening capable of β-ionone hydroxylation, 5 hydroxylated α-ionone at C-3 with a much higher hydroxylation activity (Table 1). S. fradiae, for instance, converted 75% of the α-ionone added to the culture medium; i.e., its activity is 19 times higher for the α-isomer than for the β-isomer. Good conversion of both ionone isomers was found for S. arenae.
FIG. 1.
Products of oxidative transformation of β-ionone (A) and α-ionone (B).
GLC and NMR analyses of the bioconversion reaction mixtures produced by these Streptomyces strains clearly established that all Streptomyces strains transform α-ionone to one major hydroxy derivative, not to a product mixture as, e.g., A. niger does (18). A detailed 1H NMR analysis (see below) shows this in fact to be 3-hydroxy-α-ionone. When starting from racemic α-ionone [(6R)-(−)/(6S)-(+)], one would expect to find the four diastereoisomers of the hydroxy product (Fig. 2) in equal amounts if hydroxylation was not stereoselective, as reported, e.g., for A. niger JTS 191 (18). In the chiral-phase gas chromatograms, though, only two major product peaks appear, representing the two enantiomers (3R,6R)- and (3S,6S)-hydroxy-α-ionone (Fig. 2), which merge into one single peak if an achiral phase is used. Enzymatic hydroxylation of α-ionone by the Streptomyces strains thus proceeds with both high regio- and stereoselectivity.
FIG. 2.
Hydroxylation of (6R)- and (6S)-α-ionone with Streptomyces strains.
Structural characterization of (3R,6R)- and (3S,6S)-hydroxy-α-ionone.
One transformation (S. fradiae [Table 1]), where the GLC trace showed sufficiently high turnover of the substrate, racemic α-ionone, was worked up as described above, and the ether extract was dried and evaporated. The oily residue was taken up in CDCl3 (1 ml), the solution was dried again on a molecular sieve, and the extract was filtered carefully and used for the individual 1H and 13C NMR experiments.
The (noise-decoupled) 13C NMR spectrum shows 13 signals for the major product (>80%): one carbonyl resonance, four olefinic carbon resonances, and eight sp3 carbon resonances. The chemical shifts are listed, with the appropriate assignments, in Table 2. In the first report of 13C NMR data on (3R,6R)- and (3S,6S)-hydroxy-α-ionone (14), the resonances of both the quaternary C-1 and the four methyl carbons C-10 to C-13 were assigned incorrectly. One of the sp3 signals appears shifted downfield to 65.40 ppm relative to α-ionone, definitely proving introduction of one hydroxyl function into the substrate. On the straightforward evidence from the 1H and 13C,1H correlation spectroscopy (COSY) NMR spectra, all four methyl groups as well as the tertiary hydrogen at C-6 appear conserved in the transformation product, as do the three olefinic protons. The OH group has to be introduced at either C-2 or C-3. Hydroxylation in allylic position, i.e., at C-3, is established unequivocally from the following NMR arguments. (i) The (geminal) 2J coupling constant between the two diastereotopic protons of the residual methylene group is (−)13.4 Hz; for a CH2 group adjacent to an olefinic π bond, as in position 3, geminal coupling is expected to be 2.5 to 6 Hz more negative. (ii) With a hydroxy group at C-2, the 13C resonance of one of the geminal C-1 methyl groups should appear shifted upfield by 4 to 6 ppm (γ-cis effect). Actually, one methyl resonance moves upfield and one moves downfield, both by ∼2.5 ppm; the axial and the equatorial methyl groups at C-1 thus appear much better differentiated as in the substrate α-ionone.
TABLE 2.
NMR data for (3R,6R)- and (3S,6S)-hydroxy-α-ionone
Position | δ (ppm)
|
J (1H,1H) (Hz) | |
---|---|---|---|
13C | 1Ha | ||
C-1 | 34.02 | ||
C-2 | 43.98 | 1.328 (ax) | 3J (2-Ha, 3-H) = 6.55 |
1.761 (eq) | 3J (2-He, 3-H) = 5.95 | ||
2J (2-Ha, 2-He) = (−)13.4 | |||
C-3 | 65.40 | 4.197 | 3J (3-H, 4-H) = 2.85 |
5J (3-H, 6-H) = 1.95 | |||
5J (3-H, 13-H) = 1.8 | |||
C-4 | 126.10 | 5.558 | 4J (4-H, 6-H) = 1.75 |
C-5 | 135.26 | ||
C-6 | 54.38 | 2.433 | 3J (6-H, 7-H) = 10.2 |
4J (6-H, 8-H) = 0.7 | |||
4J (6-H, 13-H) = 0.9 | |||
C-7 | 147.51 | 6.476 | 3J (7-H, 8-H) = 15.8 |
C-8 | 133.71 | 6.031 | |
C-9 | 198.35 | ||
C-10 | 27.27 | 2.196 | |
C-11eq | 29.43 | 0.952 | |
C-12ax | 24.64 | 0.815 | |
C-13 | 22.77 | 1.546 | |
3-OH | 1.209 | 3J (OH, 3-H) = 6.4 |
Abbreviations: ax, axial; eq, equatorial.
The single set of 13C resonances for the major product excludes formation of both regio- and diastereoisomers. The 3-OH function thus has been introduced, in the course of the biotransformation not only with very high regioselectivity at C-3 but also with high stereoselectivity trans to the oxobutenyl side chain at C-6. An alternative cis hydroxylation is ruled out by the very small vicinal coupling constant between 3-H and 4-H (2.85 Hz), which definitely excludes equatorial orientation of the residual proton at C-3, with a quasi-axial OH group.
A complete analysis of the 1H,1H long-range coupling pattern of 3-hydroxy-α-ionone (Table 2) provides additional, definitive proof of the quasi-axial orientation of 3-H. This orientation has already been demonstrated with nuclear Overhauser effect difference experiments by Machida and Kilzuchi (6). However, these experiments like those of other researchers have not resolved the small 1H,1H coupling constants by which especially the 3-H and 4-H resonances are split up into highly complex multiplets (6, 7, 10, 14, 15, 18). The 1H NMR data for cis- and trans-3-hydroxy-α-ionone, on the other hand, reported by Yamazaki et al. (18) clearly demonstrate that the 1H chemical shift values of these two diastereoisomeric compounds alone are not sufficiently differentiated for a straightforward stereochemical assignment; for this, a complete coupling analysis is indispensable.
In summary, we established that several Streptomyces strains are able to convert racemic α-ionone in high yield to a racemic mixture of the enantiomeric (3R,6S)- and (3S,6R)-3-hydroxy-α-ionones. The constitution of the hydroxylation products was unequivocally proven by 1H and 13C NMR analysis. Painstaking analysis, in particular of the 1H,1H long-range coupling constants, further showed quasi-axial orientation of the residual hydrogen at C-3, i.e., quasi-equatorial orientation of the newly introduced hydroxyl group.
ACKNOWLEDGMENTS
This work was supported in part by grant ZSP B3.3U from the Federal Ministry of Education, Science and Technology (BMBF), Bonn, Germany, and by the BASF AG, Ludwigshafen, Germany.
REFERENCES
- 1. BASF AG. Unpublished data.
- 1a.Enzell C R, Wahlberg I, Aasen J A. Isoprenoids and alkaloids of tobacco. Fortschr Chem Org Naturst. 1977;34:1–79. [Google Scholar]
- 2.Hartman D A, Pontones M E, Kloss V F, Curley R W, Jr, Robertson L W. Models of retinoid metabolism: microbial biotransformation of α-ionone and β-ionone. J Nat Prod (Lloydia) 1988;51:947–953. doi: 10.1021/np50059a022. [DOI] [PubMed] [Google Scholar]
- 3.Krasnobajew V. Terpenoids. In: Kieslich K, editor. Biotechnology. 6a. Biotransformations. Deerfield Beach, Fla: Verlag Chemie; 1984. pp. 97–125. [Google Scholar]
- 4.Larroche C, Creuly C, Gros J-B. Fed-batch biotransformation of β-ionone by Aspergillus niger. Appl Microbiol Biotechnol. 1995;43:222–227. [Google Scholar]
- 5.Loeber D E, Russell S W, Toube T P, Weedon B C L, Diment J. Carotenoids and related compounds. Part XXVIII. Syntheses of zeaxanthin, β-cryptoxanthin, and zeinoxanthin (α-cryptoxanthin) J Chem Soc C. 1971;1971:404–408. [Google Scholar]
- 6.Machida K, Kikuchi M. Norisoprenoids from Viburnum dilatatum. Phytochemistry. 1996;41:1333–1336. [Google Scholar]
- 7.Mayer H, Rüttimann A. Synthese von optisch aktiven, natürlichen Carotinoiden und strukturell verwandten Naturprodukten. IV. Synthese von (3R,3′R,6′R)-Lutein. Helv Chim Acta. 1980;63:1451–1455. [Google Scholar]
- 8.Mikami Y, Fukunaga Y, Arita M, Kisaki T. Microbial transformation of β-ionone and β-methylionone. Appl Environ Microbiol. 1981;41:610–617. doi: 10.1128/aem.41.3.610-617.1981. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Mikami Y, Watanabe E, Fukunaga Y, Kisaki T. Formation of 2S-hydroxy-β-ionone and 4ζ-hydroxy-β-ionone by microbial hydroxylation of β-ionone. Agric Biol Chem. 1978;42:1075–1077. [Google Scholar]
- 10.Mori K. Syntheses of optically active grasshopper ketone and dehydrovomifoliol as a synthetic support for the revised absolute configuration of (+) abscisic acid. Tetrahedron. 1974;30:1065–1072. [Google Scholar]
- 11.Ohloff G. Recent development in the field of naturally occurring aroma components. Fortschr Chem Org Naturst. 1978;35:431–527. [Google Scholar]
- 12.O’Keefe D P, Harder P A. Occurrence and biological function of cytochrome P450 monooxygenase in the actinomycetes. Mol Microbiol. 1991;5:2099–2105. doi: 10.1111/j.1365-2958.1991.tb02139.x. [DOI] [PubMed] [Google Scholar]
- 13.Pabst A, Barron D, Sémon E, Schreier P. A 4-hydroxy-β-ionone disaccharide glycoside from raspberry fruits. Phytochemistry. 1992;31:3105–3107. [Google Scholar]
- 14.Pfander H, Semadeni P A. The synthesis of optically active enriched (+)-(6R)-α-ionone. Aust J Chem. 1995;48:145–151. [Google Scholar]
- 15.Rüttimann A, Mayer H. Synthese von optisch aktiven, natürlichen Carotinoiden und strukturell verwandten Naturprodukten. V. Synthese von (3R,3′R)-, (3S,3′S)- und (3R,3′S, meso)-Zeaxanthin durch asymmetrische Hydroborierung. Ein neuer Zugang zu optisch aktiven Carotinoidbausteinen. Helv Chim Acta. 1980;63:1456–1462. [Google Scholar]
- 16.Scherr O. Ph.D. thesis. Stuttgart, Germany: Universität Stuttgart; 1997. [Google Scholar]
- 17.Trower M K, Lenstra R, Omer C, Buchholz S E, Sarislani F S. Cloning, nucleotide sequence determination and expression of the genes encoding cytochrome P-450soy (soyC) and ferredoxinsoy (soyB) from Streptomyces griseus. Mol Microbiol. 1992;6:2125–2134. doi: 10.1111/j.1365-2958.1992.tb01386.x. [DOI] [PubMed] [Google Scholar]
- 18.Yamazaki Y, Hayashi Y, Arita M, Hieda T, Mikami Y. Microbial conversion of α-ionone, α-methylionone, and α-isomethylionone. Appl Environ Microbiol. 1988;54:2354–2360. doi: 10.1128/aem.54.10.2354-2360.1988. [DOI] [PMC free article] [PubMed] [Google Scholar]