Efficient production of raspberry ketone via 'green' biocatalytic oxidation

Article abstract of DOI:10.1016/j.tet.2003.10.019

For the development of a 'green' oxidation method, the transformation of 4-(p-hydroxyphenyl)butan-2-ol (rhododendrol) into 4-(p-hydroxyphenyl) butan-2-one (raspberry ketone) was used as a model reaction. Different lyophilized cells of Rhodococcus spp. have been screened for their ability to perform the desired oxidation. Rhodococcus equi IFO 3730 and R. ruber DSM 44541 were able to use acetone as a hydrogen acceptor in a hydrogen transfer-like process. The oxidation can be performed at substrate concentrations up to 500 g/L.

Full text of DOI:10.1016/j.tet.2003.10.019

Tetrahedron 59 (2003) 9517–9521
Efficient production of raspberry ketone via ‘green’
biocatalytic oxidation
*
Birgit Kosjek, Wolfgang Stampfer, Ruud van Deursen, Kurt Faber and Wolfgang Kroutil
Department of Chemistry, Organic and Bioorganic Chemistry, University of Graz, Heinrichstrasse 28, A-8010 Graz, Austria
Received 2 September 2003; revised 7 October 2003; accepted 8 October 2003
Abstract—For the development of a ‘green’ oxidation method, the transformation of 4-(p-hydroxyphenyl)butan-2-ol (rhododendrol) into
-(p-hydroxyphenyl)butan-2-one (raspberry ketone) was used as a model reaction. Different lyophilized cells of Rhodococcus spp. have been
4
screened for their ability to perform the desired oxidation. Rhodococcus equi IFO 3730 and R. ruber DSM 44541 were able to use acetone as
a hydrogen acceptor in a hydrogen transfer-like process. The oxidation can be performed at substrate concentrations up to 500 g/L.
q 2003 Elsevier Ltd. All rights reserved.
1. Introduction
physical processes (such as distillation and extraction of
natural sources) or by enzymatic/microbial processes
transforming precursors isolated from Nature.
1
3
The availability of environmentally benign oxidation
methods for the oxidation of sec-alcohols to the correspond-
ing ketones still represents a significant problem for
synthetic organic chemistry. Whereas, chemical protocols
are still bounded to the use of toxic hypervalent metals or
expensive reagents (e.g. Swern-oxidation), biocatalytic
methods are hampered by the instability of alcohol
dehydrogenases towards elevated concentrations of organic
materials, such as substrate, product and co-solvents. As a
consequence, large-scale bio-oxidations employing sec-
alcohol dehydrogenases are restricted to the use of
The main natural source of raspberry ketone is raspberry
(Rubus idaeus, up to 3.7 mg/kg of fresh berries) although the
compound is also present in other fruits, such as peach,
grapes, apples, various berries, vegetables (e.g. rhubarb),
1
4,15
and in the bark of trees (e.g. yew, maple, and pine).
However, the amount of raspberry ketone obtainable from
these sources is to low and is too expensive to be of
industrial significance. Due to the wide area of applications
(e.g. in aroma formulations of kiwi, cherry, strawberry) as
well as the high price for the natural flavour compound
1
–6
fermenting cells
and are thus limited by low substrate
Recently, we developed an asymmetric
biocatalytic hydrogen transfer protocol employing
7
–9
21
concentration.
(EUR 600–2000 kg ) in comparison to the synthetic
1
16
analogue (EUR 10 kg2 ), the food industry shows a strong
1
7
lyophilized cells of Rhodococcus ruber DSM 44541 using
2-propanol as hydrogen donor (for ketone reduction) and
acetone as hydrogen acceptor (for alcohol oxidation) as
interest in economic processes for natural products. The
possible natural precursor of raspberry ketone is 4-(p-
hydroxyphenyl)-2-butanol 2 (rhododendrol, betuligenol),
which is most abundant as the corresponding 2-glycosides
(glucoside, mannoside) in the bark of birch (Betula spp.),
rhododendron (Rhododendron spp.), maple (Acer spp.), fir
1
0–12
co-substrates.
Herein we report an extended screening
for further strains and demonstrate the applicability of this
system on a preparative-scale synthesis of 4-(p-hydroxy-
phenyl)butan-2-one (raspberry ketone, 1), which represents
the principal flavor component of raspberries. The latter
compound was chosen as model substrate, since the
microbial oxidation of the corresponding alcohol
1
8–20
(Abieta spp.), and yews (Taxus spp.).
is achieved using the b-glycosidase activity of various
Deglycosilation
1
5
micro-organisms or plants yielding alcohol 2, which is
subjected to bio-oxidation to obtain natural raspberry
ketone 1.
(
rhododendrol 2) obtained from natural sources, such as
the bark of various birch spp., represents the key step in the
synthesis of natural raspberry ketone 1. According to
regulations within the European Community, flavor sub-
stances labeled ‘natural’ may only be obtained either by
Besides its application in flavor and fragrance formulations,
raspberry ketone 1 is used in an entirely new section of the
food and cosmetics industry, called ‘cosmaceutical
industry’, a combination of cosmetics and nutraceuticals.
Therein, it is a constituent of skin-lightening cosmetic, and
Keywords: biocatalysis; Rhodococcus ruber; oxidation; hydrogen transfer;
raspberry ketone.
w
weight loss advancing dietary supplements (Vitarosso ),
1–23
2
due to its capability to burn subcutaneous fat. Since
the preference of the consumer for ‘natural’ labelled
*
Corresponding author. Tel.: þ43-316-380-5350; fax: þ43-316-380-9840;
e-mail: wolfgang.kroutil@uni-graz.at
0040–4020/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2003.10.019

9518
B. Kosjek et al. / Tetrahedron 59 (2003) 9517–9521
food/cosmetics additives is constantly increasing, an
economic biocatalytic process for raspberry ketone from
natural sources is needed.
cells proved to be inefficient. Thus, addition of rhododen-
drol (3.3 g/L) to a two days old fermentation broth caused
the cells to agglomerate, and no conversion was observed.
Likewise, attempts towards enzyme induction by substi-
tution of glucose in the fermentation medium by rhododen-
drol was unsuccessful, indicating that rhododendrol cannot
be utilized as a carbon source and/or acts as growth
inhibitor. Due to these limitations, experiments with
fermenting cells were discontinued. Since lyophilized
cells can be stored for at least one year without significant
loss of activity, most of the following experiments were
performed using the lyophilized cell preparation.
2
. Results and discussion
A screening for strains capable of oxidizing rhododendrol
rac-2 was performed using rehydrated lyophilized cells in
phosphate buffer, while the substrate was added directly as
solid without organic co-solvent to give a final concen-
tration of 20 g/L. All Rhodococcus spp. investigated
oxidized the alcohol to the corresponding ketone (Table 1)
at varying rates (conversion of 16 to 32% after 24 h).
For the assessment of the method for preparative appli-
cations, the maximum substrate concentration is of crucial
importance. Thus, measuring the relative apparent initial
activity of lyophilized cells at a constant level of acetone
(10% v/v) with increasing substrate concentration, it was
found that the concentration of rac-rhododendrol could be
increased up to 500 g/L (3.0 mol/L) (Fig. 1). Interestingly,
at 5% v/v acetone, the activity maximum was observed at
Table 1. Screening of Rhodococcus spp. for bio-oxidation activity (without
co-substrate)
a
Entry
Strain
Conversion (%)
1
2
3
4
5
Rhodococcus ruber DSM 44539
Rhodococcus ruber DSM 44540
Rhodococcus ruber DSM 43338
Rhodococcus equi IFO 3730
16
19
27
30
32
1
80 g/L of rac-2, while the reaction did not go to
completion, which can be explained by the fact that the
co-substrate was not in excess. The efficiency of comparable
systems in terms of substrate concentration reported for the
bio-oxidation of rac-rhododendrol using Candida and
Rhodococcus ruber DSM 44541
a
Lyophilized cells in phosphate buffer (pH 8.0, 50 mM, 248C) containing
rac-2 (20 g/L), 24 h.
1
5
24
Acetobacter sp. or Pichia farinosa in aqueous buffer is
much lower, namely 0.2 and 1.5 g/L, respectively.
In a next step, the most active strains Rhodococcus equi IFO
3
730 and Rhodococcus ruber DSM 44541 (entries 4 and 5)
were tested for the ability of accept acetone as hydrogen
acceptor (10% v/v). The tolerance of acetone as co-substrate
at elevated concentrations is crucial to render an efficient
oxidation processes, since it acts as hydrogen acceptor and
increases the solubility of the lipophilic substrate. We were
pleased to see that the conversion increased for both
strains to 52% within 24 h under these conditions. The
In order to prove the viability of the method, rhododendrol
rac-2 was bio-oxidized to raspberry ketone applying an
oxidative kinetic resolution on a preparative scale. Thus,
rac-2 (2.66 g, 160 mmol) was converted employing resting
cells (0.44 g dry cell weight) in phosphate buffer (50 mM,
pH 8.0) at 10% v/v initial acetone concentration, while the
acetone concentration was stepwise increased to finally
reach 20% v/v after 15 h. The reaction was stopped after
44 h to give 1.17 g of isolated ketone 1 after flash
chromatography, which corresponds to a 89% yield with
respect to the theoretical maximum of 50% of a kinetic
resolution. This corresponds to a space-time yield of
10 mmol/L h and a productivity of 2.66 g product per one
gram of cells.
(
(
S)-enantiomer was oxidized preferentially, while the
R)-enantiomer remained behind (Scheme 1). The enantio-
meric excess (ee) of the remaining (R)-alcohol obtained by
R. ruber DSM 44541 was 98% at 52% conversion, which
corresponds to an enantioselectivity (expressed as the
enantiomeric ratio) of E¼91.
In order to investigate the influence of the cell-status on the
catalytic activity, lyophilized cells (obtained by harvesting
the cells in the stationary phase, washing, shock freezing
followed by lyophilization) were compared with freshly
harvested (resting) cells. Resting cells exhibited a 50%
higher activity. For both cell preparations no significant
by-products were found. In contrast, the use of fermenting
In Nature, both enantiomers of rhododendrol 2 occur in
various sources. For instance, in the bark of the silver birch
Betula nana the (R)-enantiomer dominates (ee 94%) while
for Betula saposhnikovii the (S)-enantiomer is in excess (ee
93%). (S)-Rhododendrol is the aglycon part of rhododen-
droketoside, which is an ingredient of the stem bark of the
2
5
Scheme 1. Oxidative kinetic resolution of rhododendrol (rac-2).

B. Kosjek et al. / Tetrahedron 59 (2003) 9517–9521
9519
Figure 1. Apparent initial activities for the oxidation of rac-rhododendrol rac-2 at different substrate concentrations.
Japanese maple tree Acer nikoense which has been used in
traditional folk medicine for the treatment of hepatic
4.72 mol) was dissolved in methanol (2.5 L). While the
reaction was kept at 208C by ice cooling sodium
borohydride (96%, 74 g, 1.888 mol) was added in portions
within 3 h. After completion (16 h) ice (50 mL) was added
and the mixture left stirring for 1 h. The mixture was
carefully acidified using HCl (6 M). Part of the solvent
(0.5 L) was evaporated under reduced pressure before the
mixture was filtered and the precipitant was washed with
diethyl ether. The solvent of the remaining mixture was
evaporated under reduced pressure on a rotavapor at 908C.
The remaining pure yellowish oil was diluted with ethyl
acetate (1:1) and purified by flash chromatography through
silica gel (500 g) and eluted using ethyl acetate. The solvent
was evaporated under reduced pressure and left standing for
one hour to crystallise (white crystals) (650 g, 92%, mp 71–
2
6
disorders and eye disease in Japan.
R)-rhododendrol is of interest for the pharmaceutical
industry, because it possesses anti-inflammatory and
Furthermore,
(
2
7,28
hepatoprotective effects.
As shown in the screening,
by oxidative kinetic resolution of rac-rhododendrol rac-2,
the (R)-enantiomer (R)-2 is accessible. In order to provide
access to both enantiomers of 2, we synthesized
(
S)-rhododendrol using lyophilized cells of R. ruber DSM
4
2
4541 via biocatalytic reduction of raspberry ketone 1 using
1
-propanol (50% v/v) as hydrogen donor. Enantiomeri-
1
cally pure rhododendrol (S)-2 (ee .99%) was obtained in
7% (0.88 g) isolated yield and 92% conversion on a one
gram scale within 22 h.
8
728C). Spectroscopic data were consistent with those
reported in the literature.
2
9
To simulate a possible oxidative production of ‘natural’
raspberry ketone, starting with enantiomeric pure
1
(
S)-alcohol (S)-2 (which is present in nature), the bio-
H NMR (D -acetone): d 6.93 (2H, d, J¼8.4 Hz, Ar), 6.73
6
catalytic oxidation of the obtained (S)-alcohol (S)-2 using
acetone as hydrogen acceptor (20% v/v) was performed
giving ketone 1, after 17 h, at 95% conversion and 83%
yield.
(2H, d, J¼8.4 Hz, Ar), 3.70 (1H, m, CH–OH), 2.57 (2H, m,
Ph–CH ), 1.64 (2H, m, Ph–CH –CH ), 1.12 (3H, t,
2
2
2
J¼6 Hz, CH3).
All micro-organisms were obtained from culture collections
0
and were cultivated as described previously.
3
We have reported the biocatalytic oxidative kinetic
resolution of racemic rhododendrol 2 furnishing raspberry
ketone 1 and (R)-2 in 98% ee. In a complementary fashion,
bioreduction of 1 gave access to enantiopure (S)-rhododen-
drol. Due to the exceptionally high substrate concentrations,
performed in gram-scale reactions, this methodology is
applicable to the large-scale synthesis of ‘natural’
ingredients for flavor- and fragrance-formulations.
3.2. Screening procedure
Lyophilized cells (60 mg) were rehydrated in phosphate
buffer (1 mL, 50 mM, pH 8.0) at room temperature on a
rotary shaker (130 rpm) in an Eppendorf vial (30 min). rac-
Rhododendrol rac-2 (20 mg, 0.12 mmol) was added and the
reaction mixture was shaken at 248C for 24 h (130 rpm).
The reaction was stopped by extraction with ethyl acetate
(0.5 mL), the organic layer was separated from the cell
debris using centrifugation (13,000 rpm, 5 min) and dried
over Na SO . Conversion was determined by GC analysis.
3
. Experimental
3
(
.1. Raspberry ketone 1 is commercially available
Aldrich)
2
4
For the screening experiments with 10% v/v acetone:
.9 mL buffer was used and the substrate was added
dissolved in acetone (0.1 mL).
0
3
.1.1. Racemic alcohol rac-2. Raspberry ketone 1 (700 g,

9520
B. Kosjek et al. / Tetrahedron 59 (2003) 9517–9521
3
.3. Variation of the substrate concentration
The absolute configuration of rhododendrol was confirmed
by comparison of the optical rotation with literature data:
2
0
20
Rehydrated lyophilized cells (30 mg) in phosphate buffer
0.30 mL, 50 mM, pH 8.0) were mixed with a solution
containing the appropriate amount of rac-rhododendrol rac-
(40 to 300 mg) in buffer (0.24–0.27 mL) and acetone
30–60 mL) to give a total volume of 0.60 mL, the vials
(S)-2: [a] ¼211.8 (c 1.0, EtOH 96%, 77% ee), lit. 213.6
D
(
(S) (c 1.0, EtOH).
2
(
Acknowledgements
were shaken (130 rpm) at 248C for 3.6 h. The reaction was
stopped and analysed as described above.
This study was performed within the Spezialforschungs-
bereich ‘Biokatalyse’ and the Research Centre Applied
Biocatalysis. Financial support by Ciba SC (Basel), Fonds
zur F o¨ rderung der wissenschaftlichen Forschung (Vienna,
project no. F115), TIG, SFG, Province of Styria and the City
of Graz is gratefully acknowledged.
3.4. Preparative biocatalytic kinetic oxidative resolution
of rac-rhododendrol rac-2
rac-Rhododendrol rac-2 (2.66 g) was mixed with a cell
suspension harvested during the stationary phase (corre-
sponding to 440 mg dry cell weight) in phosphate buffer
(12.0 mL, 50 mM, pH 8.0) and acetone (1.4 mL) in a 50 mL
screw-capped test tube. The mixture was shaken at 308C at
References
1
1
30 rpm (rotary shaker). After 4 and 15 h, acetone (0.80 and
.0 mL, respectively) was added. The reaction was stopped
1. Miyamoto, K.; Ohta, H. Biotechnol. Lett. 1992, 14, 363–366.
2. Fogagnolo, M.; Giovannini, P. P.; Guerrini, A.; Medici, A.;
Pedrini, P.; Colombi, N. Tetrahedron: Asymmetry 1998, 9,
2317–2327.
after 44 h by extraction with ethyl acetate (3£50 mL). The
combined organic layers were dried (Na SO ) and purified
2
4
by flash chromatography to give 1.17 g of the ketone 1
44.5%).
(
3. Villa, R.; Romano, A.; Gandolfi, R.; Sinisterra Gago, J. V.;
Molinari, F. Tetrahedron Lett. 2002, 43, 6059–6061.
4. P e´ rez, H. I.; Luna, H.; Manjarrez, N.; Sol ´ı s, A. Tetrahedron:
Asymmetry 2001, 12, 1709–1712.
3.5. Preparative biocatalytic reduction of raspberry
ketone 1 to (S)-2
5
. P e´ rez, H. I.; Luna, H.; Manjarrez, N.; Sol ´ı s, A. Biotechnol.
Lett. 2001, 23, 1467–1472.
Lyophilized cells of Rhodococcus ruber DSM 44541 (1.2 g)
were rehydrated in phosphate buffer (15 mL, 50 mM, pH
6. Fantin, G.; Fogagnolo, M.; Medici, A.; Pedrini, P.; Fontana, S.
Tetrahedron: Asymmetry 2000, 11, 2367–2373.
7
.5, 140 rpm, rt) for 30 min in a 100 mL round-bottom
flask. Raspberry ketone 1 (1.0 g, 6.1 mmol) dissolved in
-propanol (15 mL) was added. The entire mixture was
7. Kato, Y.; Asano, Y. J. Mol. Catal. B: Enzymol. 2001, 13,
27–36.
2
shaken at 140 rpm and room temperature for 22 h. The
mixture was extracted with ethyl acetate (2£70 mL), the
8. Nakamura, K.; Inoue, Y.; Matsuda, T.; Misawa, I. J. Chem.
Soc., Perkin Trans. 1 1999, 2397–2402.
combined organic phases were dried (MgSO ) and con-
4
9. Liese, A.; Karutz, M.; Kamhuis, J.; Wandrey, C. Biotechnol.
Bioengng 1996, 51, 544–550.
centrated in vacuo. Purification by flash chromatography
yielded 0.88 g (87%) of (S)-2 (ee .99%).
10 Stampfer, W.; Kosjek, B.; Moitzi, C.; Kroutil, W.; Faber, K.
Angew. Chem., Int. Ed. Engl. 2002, 41, 1014–1017.
11. Stampfer, W.; Kosjek, B.; Faber, K.; Kroutil, W. J. Org.
Chem. 2003, 68, 402–406.
3
.6. Preparative biocatalytic oxidation of (S)-2
Lyophilised cells of Rhodococcus ruber DSM 44541 (1.0 g)
were rehydrated in phosphate buffer (15 mL, 50 mM, pH
12. Stampfer, W., Kosjek, B., Kroutil, W., K. Faber (Ciba
Speciality Chemicals Holding Inc., Switzerland), patent EP
filed on March 18, 2002, No. 02 405 204. 5.
7
(
.5, 140 rpm, rt) for 30 min. Enantiopure (S)-rhododendrol
S)-2 (0.88 g, 5.3 mmol) dissolved in acetone (3.1 mL) was
added and the mixture was shaken at room temperature
140 rpm). After 17 h a sample (500 mL) was analyzed by
GC indicating 95% conversion. Ethyl acetate was added
13. The Council of the European Communities: Council Directive
88/388/EEC of 22 June 1988 on the approximation of the laws
of the member States relating to flavourings for use in
foodstuffs and to source materials for their production: http://
europa.eu.int/comm/food/fs/sfp/addit_flavor/flav09_en.pdf.
14. Pabst, A.; Barron, D.; Adda, J.; Schreier, P. Phytochemistry
(
(
(
60 mL) and worked up, as described above, gave 0.72 g
83%) of raspberry ketone 1.
1
990, 29, 3853–3858.
Conversions were determined using a Varian 3900 GC with
a HP Innowax column (30 m£0.25 mm£0.25 mm, 1.0 bar
15. Dumont, B., Hugueny, P., Belin, J.-M., (BFA Laboratoires,
France) 1995 EP 0 707 072 A1, Chem. Abstr. 124:315169y.
16. B o¨ ker, A.; Fischer, M.; Berge, R. G. Biotechnol. Prog. 2001,
17, 568–572.
N ). Temperature program: 1258C/0 min—108C/min–
2
1
758C/0 min–258C/min–2508C/9.5 min. Ketone 1:
T ¼11.6 min, rac-rhododendrol 2: T ¼12.3 min.
17. Krings, U.; Berger, R. G. Appl. Microbiol. Biotechnol. 1998,
4
ret
ret
9, 1–8.
18. Hugueny, P., Dumont, B., Ropert, F., Belin, J.M., Bioflavour,
95 (Dijon), 269-273.
For the determination of the enantiomeric excess of 2,
3
1
samples were acetylated and analysed on a Jasco HPLC
with a Daicel-OD-H column (0.46£25 cm) at 208C using an
array detector. Flow: 0.6 mL/min, eluent: n-heptane/
19. Fushiya, S.; Kabe, Y.; Ikegaya, Y.; Takano, F. Planta Med.
1998, 64, 598–602.
2
-propanol¼9:1, (R)-2: T ¼10.4 min, (S)-2: T ¼11.7
20. Fronza, G.; Fuganti, C.; Pedrocchi-Fantoni, G.; Serra, S.;
Zucchi, G. J. Agric. Food Chem. 1999, 47, 1150–1155.
ret
ret
min, 1: T ¼17.4 min.
ret

B. Kosjek et al. / Tetrahedron 59 (2003) 9517–9521
9521
2
1. Ikemoto, T., Nakatsugawa, H., Yokota, T., (Kanebo Ltd.,
Japan) 1998 JP 10 17462 A2, Chem. Abstr. 128:145160y.
2. Hikima, T., Yokota, T., Ota, C., Hamada, K., (Kanebo Ltd.,
Japan) 2000 JP 2000 95641 A2, Chem. Abstr. 132:255764p.
3. http://www.red-raspberry.org/cosmaceuticals.htm.
4. Falconnier, B., (Pernod Ricard, France) 1999 WO 99/49069,
Chem. Abstr. 131: 242079y.
27. Fushiya, S.; Kabe, Y.; Ikegaya, Y.; Takano, F. Planta Med.
1998, 64, 598–602.
2
28. Fujita, T.; Hatamoto, H.; Iwasaki, T.; Takafuji, S. Phyto-
chemistry 1995, 39, 1085–1089.
2
2
29. Fronza, G.; Fuganti, C.; Mendozza, M.; Rallo, R. S.; Ottolina,
G.; Joulain, D. Tetrahedron 1996, 52, 4041–4052.
3
0. Kroutil, W.; Mischitz, M.; Faber, K. J. Chem. Soc., Perkin
Trans. 1 1997, 3629–3636.
2
2
5. Falconnier, B., Godard, N., Attard, L., Girad, P., (Pernod
Ricard, France) 1998 FR 2 776 301-A1.
3
1. Stampfer, W.; Kosjek, B.; Faber, K.; Kroutil, W. Tetrahedron:
Asymmetry 2003, 14, 275–280.
6. Morikawa, T.; Tao, J.; Ueda, K.; Matsuda, H.; Yoshikawa, M.
Chem. Pharm. Bull. 2003, 51, 62–67.