A chemo-enzymatic expeditious route to racemic dihexanoyl (2R*,3R*,4R*)-dehydroxymethylepoxyquinomycin (DHMEQ), the precursor for lipase-catalyzed synthesis of the potent nuclear factor-κB inhibitor, (2 S,3S,4 S )-DHMEQ

Article abstract of DOI:10.1248/cpb.c12-00417

In order to synthesize the potent nuclear factor (NF)- κB inhibitor, (2S ,3 S ,4 S )-dehydroxymethylepoxyquinomycin (DHMEQ), in a large scale, a new route for its corresponding racemic precursor, dihexanoyl (2R*,3R*, 4R*)- DHMEQ, was developed. By employi

Full text of DOI:10.1248/cpb.c12-00417

1220
Note
Chem. Pharm. Bull. 60(9) 1220–1223 (2012)
Vol. 60, No. 9
A Chemo-Enzymatic Expeditious Route to Racemic Dihexanoyl
(2R*,3R*,4R*)-Dehydroxymethylepoxyquinomycin (DHMEQ),
the Precursor for Lipase-Catalyzed Synthesis of the Potent Nuclear
Factor-κB Inhibitor, (2S,3S,4S)-DHMEQ
,a
Ryohei Kobayashi,a Kengo Hanaya,a Mitsuru Shoji,a Kazuo Umezawa,b and Takeshi Sugai*
a Department of Pharmaceutical Science, Keio University; 1–5–30 Shibakoen, Minato-ku, Tokyo 105–8512,
b
Japan: and School of Medicine, Aichi Medical University; Nagakute, Aichi 480–1195, Japan.
Received May 7, 2012; accepted June 13, 2012
In order to synthesize the potent nuclear factor (NF)-κB inhibitor, (2S,3S,4S)-dehydroxymethylepoxy-
quinomycin (DHMEQ), in a large scale, a new route for its corresponding racemic precursor, dihexanoyl
(2R*,3R*,4R*)-DHMEQ, was developed. By employing both hydroquinone and benzoquinone intermediates,
the total yield, reproducibility, and synthetic steps were improved and the synthetic cost was reduced.
Key words dehydroxymethylepoxyquinomycin; process chemistry; hydroquinone; oxidation; benzoquinone;
epoxidation
(2S,3S,4S)-Dehydroxymethylepoxyquinomycin (DHMEQ) originally reported dimethyl ether 2.
(1a) is a potent nuclear factor (NF)-κB inhibitor with anti-in-
By applying 48% hydrobromic acid, demethylation of 5a
flammatory and anticancer activities in animals.1) (2S,3S,4S)- proceeded efficiently to give 5b. The workup was simple,
Isomer possesses approximately 10-fold greater activity than as most of the product precipitated from the chilled reaction
its enantiomer, and thus, an effective separation of enantio- mixture upon completion of the demethylation at both posi-
mers is required to develop (2S,3S,4S)-1a for clinical use. In tions. Reduction of the nitro group by catalytic hydrogenation
a previous study, we have successfully separated the enan- proceeded in a quantitative manner. Acylation of the free
tiomers by lipase-catalyzed kinetic resolution2) (Chart 1). Di- amine of 6a with acetylsalicyloyl chloride and pyridine in
hexanoyl (2R*,3R*,4R*)-DHMEQ (1b) is the substrate of the tetrahydrofuran (THF) selectively provided 6b. The use of a
enzymatic hydrolysis, and the development of an expeditious stronger base such as triethylamine resulted in the formation
route for 1b is expected.
of a byproduct, due to overacylation of one of the phenolic
Two problems exist in previously reported procedures. They hydroxy groups.
required the use of expensive hypervalent iodine oxidants2,3)
Oxone® (potassium peroxymonosulfate) with a catalytic
(from 2 to 3, Chart 2) and had low reproducibility in the ep- amount of tetra-n-butylammonium bromide4) afforded 7a in
oxidation step3) (from 3 to 4). The present synthesis beginning 60% yield (four steps, from 5a, Chart 2) as yellow prisms (mp
with readily available starting material 5a was designed to 123.8–124.0°C). Therefore, the use of expensive hypervalent
solve the above-mentioned problems using 6b as a synthetic iodine oxidant could be avoided. This approach also allowed
intermediate. This allowed oxidation of the hydroquinone application of cerium(IV) ammonium nitrate (CAN).5) Upon
under much milder conditions, than that needed for the scale-up, CAN-mediated oxidation proceeded to give 7a in
Chart 1. Lipase-Catalyzed Kinetic Resolution of ( )-1b the Precursor of (2S,3S,4S)-DHMEQ
The authors declare no conflict of interest.
*To whom correspondence should be addressed. e-mail: sugai-tk@pha.keio.ac.jp
© 2012 The Pharmaceutical Society of Japan

September 2012
1221
Chart 2. Preparation of Benzoquinone Intermediate 7a
Chart 3. Epoxidation of 7a and the Transformation of the Products to Dihexanoyl DHMEQ 1b
56% yield (two steps, from 6a). However, extractive separation corresponding deacetylated phenol 10b (22:78) in a combined
of the desired product from highly polar and dark-colored by- yield of 60%. To maintain a homogeneous state of dissolved
products was difficult. Thus, Oxone®-mediated oxidation was MMPP throughout the reaction, water was added as co-solvent
more superior than CAN for conducting the procedure.
for the THF solution of the substrate. At workup, the pH of
For epoxidation of precursor 3 by treatment with H2O2 un- the mixture was adjusted to 6.2, which facilitated product iso-
der strongly basic conditions, the first step involved hydrolysis lation. Phthalic acid and unreacted monoperoxyphthalic acid
of phenolic acetate, followed by 1,4-nucleophilic attack by remained in the aqueous phase as salts.
hydroperoxide on 8 (Chart 3). The desired epoxidation was
Moreover, MMPP-mediated epoxidation could be performed
slow due to equilibrium between 8 and an enolate 9; attack by in non-aqueous solvent such as anhydrous N,N-dimethylfor-
HOO− on the latter didn’t occur. Meanwhile, hydrolytic cleav- mamide (DMF), which readily dissolves MMPP. Total yield
age of the amide C–N bond in 3 was a major side reaction. of the epoxidized products in DMF was better than that in
However, when using the new precursor, benzoquinone 7a, THF–water. However, the reaction produced an intractable
which was more electron-deficient than monoacetal 3, other mixture of the desired products and tautomeric forms (11a and
weakly acidic epoxidizing agents could be used under milder 11b), which were determined by NMR. These byproducts had
conditions.
long lifetimes, and didn’t readily isomerize to 10a and 10b
Epoxidation of 7a with inexpensive magnesium mono- under neutral and acidic conditions. Furthermore, they were
peroxyphthalate (MMPP)6) in a mixture of THF and water readily decomposed by hydrolysis of amide bonds under basic
successfully gave a mixture of an acetate 10a and the conditions.

1222
Vol. 60, No. 9
In addition, buffered neutral conditions were used for ep- developed. The ability to use two characteristic intermediates,
oxidation. However, the color of the solution of benzoquinone a highly oxidizable hydroquinone 6b and the benzoquinone
7a as well as epoxides 10a and 10b tended to darken quickly, 7a, and the improved reaction conditions enabled preparation
even in a phosphate buffer at a neutral pH. Moreover, MMPP- of 1a with 19% overall yield in eight steps. This new route
mediated epoxidation of this substrate was very slow in the avoids the use of expensive reagents, heavy metals, and rare
phosphate buffer.
metals. The scheme involves only two column chromatogra-
To simplify the purification of the desired products in the phy purification steps and is applicable to large scale synthesis
later synthetic steps, complete consumption of acetylated ben- with lower cost and time requirements than other methods.
zoquinone 7a was desirable. At the same time, it was neces- Thus, the diester preparation with lipase-assisted enantiomeric
sary to suppress the undesired hydrolysis to 7b (Chart 2) as resolution can greatly improve the preparation of (2S,3S,4S)-
much as possible, because epoxidation of the free phenolic DHMEQ.
form 7b was very slow. Coordination of the acetate carbonyl
oxygen to magnesium ion was considered to be important.
The reaction conditions using aqueous THF satisfied the
above-mentioned requirements.
Experimental
General All mps are uncorrected. IR spectra were mea-
sured as films for oils or as attenuated total reflection (ATR)
1
The final step was selective reduction of the benzoquinone on a Jeol FT-IR SPX60 spectrometer. H-NMR spectra were
carbonyl groups. Reported examples3) strongly suggested the measured on a VARIAN 400-MR spectrometer. Silica gel
importance of a free phenolic hydroxy group, for the regio- 60N (spherical and neutral, 100–210µm, 37560-79) of Kanto
and stereoselective reduction to (2R*,3R*,4R*)-1a. The product Chemical Co. was used for column chromatography. Prepara-
should consist of only 10b, though the phenolic acetate in tive TLC was performed with Merck Silica Gel 60 F254 plates
10a was partially hydrolyzed to form 10b during epoxidation. (0.5mm thickness, No. 5744).
Because the salicylamide C–N bond is labile under basic con-
2-Nitrohydroquinone, 5b To a solution of 5a (15.0g,
ditions, we planned an enzymatic hydrolysis of acetate in 10a 82.0mmol) in 48% hydrobromic acid (100mL) was heated
under neutral conditions, which was prompted by the fact that to reflux for 3h and then cooled in an ice-water bath. The
the phenolic ester in 1b was readily removed by the action precipitates were collected by filtration to afford 5b (10.6g)
of a lipase.2) Burkholderia cepacia (Amano PS-IM) lipase- and the mother liquor was extracted 4 times with AcOEt
catalyzed hydrolysis of 10a in a mixture of THF and water (ethyl acetate). The combined organic layer was washed with
provided 10b in 83% yield from the mixture of 10a and 10b. saturated aqueous NaHCO3 and brine, dried over anhydrous
Longer reaction time lowered the pH of the reaction mixture, Na2SO4, and concentrated in vacuo to afford 5b (2.10g). Total
1
which caused non-enzymatic hydrolysis of the salicylamide recovery of 5b was 12.7g (quantitative) as red solid. H-NMR
C–N bond. Moreover, the resulting strongly acidic salicylic (400MHz, acetone-d6) δ: 7.06 (d, J=9.0Hz, 1H, Ar-H), 7.24
acid itself accelerated further hydrolytic decomposition of (dd, J=2.9, 9.0Hz, 1H, Ar-H), 7.48 (d, J=2.9Hz, 1H, Ar-H),
10b. However, conventional buffered conditions could not be 8.71 (s, 1H, OH), 10.01 (s, 1H, OH). This was employed for
applied for this hydrolysis because of the unstable properties the next step without further purification.
of epoxides with phosphate ions.
2-Aminohydroquinone, 6a To a solution of crude 5b
At the workup of the hydrolysis, the accompanying acetic (10.6g) in MeOH (methanol, 100mL) was added 10% pal-
acid should carefully be removed by partitioning between ladium on activated charcoal (1.06g), and the mixture was
organic phase and brine. Otherwise, we faced an unexpected stirred for 2h at room temperature under H2 atmosphere.
serious problem in the next step. After performing the se- After removal of insoluble materials, the filtrate was concen-
quential reduction and acylation of the crude mixture with trated in vacuo to afford 6a (8.52g, quantitative) as gray solid.
hexanoic anhydride to access dihexanoate 1b, we were really 1H-NMR (400MHz, CD3OD) δ: 6.12 (dd, J=2.9, 8.4Hz, 1H,
confused with the profile of the product. It was contaminated Ar-H), 6.32 (d, J=2.9Hz, 1H, Ar-H), 6.55 (d, J=8.4Hz, 1H,
with acetates 1d and 1e. After the completion of the lipase- Ar-H). This was employed for the next step without further
catalyzed hydrolysis, acetate ion persisted in the reaction purification.
mixture even after the further step of acylation. If the acetic
2-N-[(2-Acetoxybenzoyl)amino]cyclohexa-2,5-diene-1,4-
acid remained in the reaction mixture, it gave the mixture dione, 7a To a solution of crude 6a (3.50g) in THF (140mL)
containing undesired 1d and 1e. It was due to the presence of and pyridine (6mL) was added acetylsalicyloyl chloride
acetic hexanoic mixed anhydride which was caused by the nu- (5.84g, 29.4mmol) with stirring. After stirring for 5min at
cleophilic action of acetate ion to hexanoic anhydride.
room temperature, to the mixture was added aqueous HCl
All the problems above were resolved by an alternative (1mol/L, 40mL) and the resulting mixture was further stirred
approach using isopropyl alcohol as a solvent, in which the for 5min. Then, the mixture was extracted 3 times with
lipase-catalyzed reaction was converted from hydrolysis to AcOEt. The combined organic layer was washed with brine,
transesterification.7) The resulting isopropyl acetate was vola- dried over Na2SO4, and concentrated in vacuo. The residue
tile and easily removed through the usual workup procedure was diluted with a mixture of acetonitrile (46mL) and H2O
involving filtration by Celite® and concentration of the filtrate. (23mL). To the solution was added Oxone® (12.1g, 19.6mmol)
This procedure efficiently provided 10b, which was reduced and tetra-n-butylammonium bromide (903mg, 2.8mmol),
with NaBH(OAc)3 and subsequently acylated with hexanoic and the mixture was stirred for 30min at room temperature.
anhydride to produce 1b.2) The yield of 1b from 7a was 31% After removal of insoluble material, the filtrate was extracted
in four steps.
twice with AcOEt. The combined organic layer was washed
In summary, a scalable synthesis of 1b, the precursor for B. with brine, dried over Na2SO4, and concentrated in vacuo.
cepacia lipase-catalyzed synthesis of (2S,3S,4S)-DHMEQ was The residue was purified by recrystallization from MeOH,

September 2012
1223
after removal of highly polar substance by short silica gel temperature and stirred for further 45min, concentrated in
column chromatography (75g) to afford 7a (4.75g, 60% over vacuo to a half of the original volume, and was then cooled
1
four steps from 5a) as yellow prisms (mp 123.8–124.0°C). H- with an ice-water bath. The precipitates were collected by fil-
NMR (400MHz, CDCl3) δ: 2.48 (s, 3H, CH3), 6.76 (dd, J=2.3, tration to afford crude mixture of 1a (2.12g) as grayish yellow
10.1Hz, 1H, H5), 6.81 (d, J=10.1Hz, 1H, H6), 7.21 (dd, J=0.8, solid. The mother liquor was extracted twice with AcOEt. The
8.2Hz, 1H, Ar-H), 7.36 (ddd, J=0.8, 7.5, 7.8Hz, 1H, Ar-H), combined organic layer was washed with brine, dried over
7.56 (ddd, J=1.6, 7.5, 8.2Hz, 1H, Ar-H), 7.75 (d, J=2.3Hz, 1H, Na2SO4, and concentrated to give a brown solid (1.42g) which
H3), 8.05 (dd, J=1.6, 7.8Hz, 1H, Ar-H). 13C-NMR (100MHz, also contained 1a.
CDCl3) δ: 21.2, 115.6, 123.6, 125.2, 126.5, 131.4, 133.2,
The above-mentioned precipitated crude product (2.12g)
133.7, 138.3, 138.7, 148.3, 163.5, 168.4, 182.9, 187.9. IR cm−1: was suspended in THF (40mL) and to that were added hexa-
3346, 2980, 1772, 1660, 1633, 1591, 1479, 1448, 1326, 1155. noic anhydride (5.0mL, 21.3mmol) and 4-(dimethylamino)-
High resolution (HR)-MS-electron ionization (EI) Calcd for pyridine (30mg). After stirring for 30min at room tempera-
C15H11NO5 285.0637, Found 285.0650.
ture, the mixture was quenched with ice-water (15mL), then
2,3-Epoxy-5-N-[(2-hydroxybenzoyl)amino]cyclohex-5- stirred for 5min at room temperature and extracted twice
ene-1,4-dione, 10b To a solution of 7a (3.00g, 10.5mmol) with AcOEt. The organic layer was washed with aqueous
in THF (25mL) and H2O (25mL) was added MMPP (7.81g, HCl (0.5mol/L), saturated aqueous NaHCO3 and brine, dried
12.6mmol, 1.2eq) with stirring. After stirring for 20min at over Na2SO4, and concentrated in vacuo. The residue was
room temperature, the mixture was adjusted pH 6.2 by add- purified by silica gel column chromatography (120g) with
ing saturated aqueous NaHCO3 and extracted twice with hexane–AcOEt (4:1) to afford 1b (1.25g) as a yellow oil.
AcOEt. The combined organic layer was washed with brine, Independently, similar hexanoylation of the above-mentioned
dried over Na2SO4, and concentrated in vacuo to give a brown solid from the mother liquor (1.42g) provided a mix-
mixture of 7b, 10a and 10b (total 2.79g) as grayish yellow ture containing 1b. The crude product was purified by silica
solid. As the byproduct 7b and product 10b showed same gel column chromatography (35g) with hexane–diethyl ether
Rf on silica gel, the combined yield (10a and 10b) was esti- (1:1) also to afford pure 1b (225mg). The combined yield of
1
1
mated to be 60%, based on H NMR (400MHz, acetone-d6) 1b (1.47g) was 31% over 4 steps. H-NMR (400MHz, CDCl3)
with an internal standard [(1R*,2R*,3R*,6R*,7S*)-2-iodo-4,8- δ: 0.85 (t, J=6.8Hz, 3H, CH3), 0.91 (t, J=6.8Hz, 3H, CH3),
dioxatricyclo[4.2.1.03.7]nonan-5-one], standard signal at δ: 5.11 1.30 (m, 4H, C2H4CH3), 1.34 (m, 4H, C2H4CH3), 1.70 (m, 4H,
(d, J=4.9Hz, 1H)]. 10a: 3.93 (dd, J=2.2, 3.9Hz, 1H, H2), 4.14 (CH2C3H5)2), 2.53 (t, J=7.8Hz, 2H, COCH2), 2.56 (t, J=7.8Hz,
(d, J=3.9Hz, 1H, H3), 7.53 (d, J=2.2Hz, 1H, H6); 10b: 3.90 2H, COCH2), 3.50 (dd, J=2.0, 3.9Hz, 1H, H2), 3.91 (dd,
(dd, J=2.4, 3.9Hz, 1H, H2), 4.09 (d, J=3.9Hz, 1H, H3), 7.60 J=2.9, 3.9Hz, 1H, H3), 5.85 (dd, J=1.5, 2.9Hz, 1H, H4), 7.02
(d, J=2.4Hz, 1H, H6).
(dd, J=1.5, 2.0Hz, 1H, H6), 7.10 (d, J=7.8Hz, 1H, Ar-H), 7.35
Then, the above-mentioned crude product (2.79g) was dis- (dd, J=7.8, 7.8Hz, 1H, Ar-H), 7.54 (ddd, J=1.5, 7.8, 7.8Hz,
solved in a mixture of THF (34mL) and isopropyl alcohol 1H, Ar-H), 7.67 (dd, J=1.5, 7.8Hz, 1H, Ar-H), 7.99 (brs, 1H,
(17mL) and to the mixture was added B. cepacia lipase (Ama- NH). Its NMR spectrum was identical with that reported pre-
no PS-IM, 1.50g). The mixture was stirred for 11h at room viously.2)
temperature. After removal of insoluble materials by Celite®,
the filtrate was concentrated in vacuo to afford crude mixture
Acknowledgements This work was supported by Grants-
1
of 10b (2.78g) as grayish yellow solid. H-NMR (400MHz, in-Aid for Scientific Research from the Ministry of Educa-
CDCl3) δ: 3.88 (dd, J=2.0, 3.6Hz, 1H, H2), 3.98 (d, J=3.6Hz, tion, Culture, Sports, Science and Technology of Japan (No.
1H, H3), 6.96 (ddd, J=1.2, 7.6, 8.0Hz, 1H, Ar-H), 7.03 (dd, 20611015 and 23580152), and acknowledged with thanks.
J=1.2, 8.0Hz, 1H, Ar-H), 7.48 (ddd, J=1.2, 7.6, 8.4Hz, 1H, We are thankful to Amano Enzyme Inc. for generous gift of
Ar-H), 7.50 (dd, J=1.2, 8.4Hz, 1H, Ar-H), 7.62 (d, J=2.0Hz, lipase PS-IM. We also thank Dr. Tsuyoshi Saito of the Depart-
1H, H6), 8.83 (brs, 1H, NH), 11.1 (brs, 1H, OH). Its NMR ment of Chemistry, Keio University, for his valuable advice.
spectrum was identical with that reported previously.3) This
was employed for the next step without further purification.
When the epoxidation was performed in anhydrous DMF, the
following signals were also observed. 1H-NMR (400MHz,
acetone-d6) δ: 11a: 3.59 (dd, J=2.1, 4.1Hz, 1H, H2), 4.43 (d,
J=4.0Hz, 1H, H3), 7.23 (d, J=2.1Hz, 1H, H6); 11b: 3.53 (dd,
J=2.0, 3.9Hz, 1H, H2), 4.35 (d, J=3.9Hz, 1H, H3), 7.20 (d,
J=2.0Hz, 1H, H6). Both of them lacked the NH signals.
2,3-Epoxy-4-hexanoyloxy-5-N-[(2-hexanoyloxybenzoyl)-
amino]cyclohex-5-en-1-one, 1b According to the reported
procedure,2) crude mixture of 10b (2.78g) was treated with
NaBH(OAc)3 (4.02g, 19.0mmol) in MeOH (180mL) for
15min at 0°C. The mixture was allowed to warm to room
References
1) Umezawa K., Cancer Sci., 97, 990–995 (2006).
2) Hamada M., Niitsu Y., Hiraoka C., Kozawa I., Higashi T., Shoji M.,
Umezawa K., Sugai T., Tetrahedron, 66, 7083–7087 (2010).
3) Suzuki Y., Sugiyama C., Ohno O., Umezawa K., Tetrahedron, 60,
7061–7066 (2004).
4) Yakura T., Ozono A., Morimoto K., Chem. Pharm. Bull., 59, 132–
134 (2011).
5) Whiteley C. G., Bioorg. Med. Chem., 10, 1221–1227 (2002).
6) Velasco M. A., Synlett, 2005, 1807–1808 (2005).
7) Miyazawa T., Hamada M., Morimoto R., Murashima T., Yamada T.,
Tetrahedron Lett., 48, 8334–8337 (2007).