Stereoselective enzymatic reduction of keto-salinosporamide to (-)-salinosporamide A (NPI-0052)

Article abstract of DOI:10.1016/j.tetlet.2007.02.017

Salinosporamide A (NPI-0052, 1), a highly potent 20S proteasome inhibitor, has been prepared from its ketone precursor (2) by asymmetric enzymatic reduction. The yields are quantitative with complete stereoselective conversion to the desired product, with no evidence for the undesired diastereomer. This process should lead to new synthetic strategies for the total synthesis of 1.

Full text of DOI:10.1016/j.tetlet.2007.02.017

Tetrahedron Letters 48 (2007) 2537–2540
Stereoselective enzymatic reduction of keto-salinosporamide
to (À)-salinosporamide A (NPI-0052)
Rama Rao Manam, Venkat R. Macherla* and Barbara C. M. Potts
Nereus Pharmaceuticals, Inc., 10480 Wateridge Circle, San Diego, CA 92121, USA
Received 19 January 2007; revised 3 February 2007; accepted 6 February 2007
Available online 9 February 2007
Abstract—Salinosporamide A (NPI-0052, 1), a highly potent 20S proteasome inhibitor, has been prepared from its ketone precursor
(2) by asymmetric enzymatic reduction. The yields are quantitative with complete stereoselective conversion to the desired product,
with no evidence for the undesired diastereomer. This process should lead to new synthetic strategies for the total synthesis of 1.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Salinosporamide A (NPI-0052; 1) is a potent inhibitor of
the 20S proteasome, that is currently in clinical trials for
the treatment of cancer.1–3 Structurally, 1 comprises a
b-lactone-c-lactam bicyclic ring system substituted with
methyl, cyclohex-2-enylcarbinol, and chloroethyl sub-
stituents that give rise to specific and mechanistically
important interactions within the proteasome active
site.4 This dense functionality, which includes 5 contigu-
ous chiral centers, makes 1 an extremely attractive and
challenging synthetic target. The first enantioselective
total synthesis of 1 was reported by Corey and co-workers
in 2004 starting from S-threonine.5 Several alternative
strategies have been published, with all routes converg-
ing upon a common penultimate compound that bears
a hydroxyl group in place of chlorine.6–8 We envisioned
a new strategy to complete the total synthesis of 1
through its ketone precursor, ‘keto-salinosporamide’
(2), and developed a suitable method for this transforma-
tion. Using ketoreductase enzymes, the ketone group of
2 can be stereoselectively reduced to produce 1 in quan-
titative yields. Here we report the details of this reaction.
reagents gave little to no product or resulted in degrada-
tion, except for NaBH4, which still offered no stereo-
selectivity for 1, and in some cases concomitantly
reduced the b-lactone ring to give compound 4. At
reduced temperatures, C-5 epimer 3 was favored, while
increased temperatures generally resulted in over-
reduction to 4; the latter could usually be minimized
by decreasing the number of equivalents of reducing
agent to <1. Given the lack of selectivity for 1, we
explored the potential of ketoreductase enzymes to
execute the required stereoselective reduction.
Enzymes are well known catalysts for stereoselective
reductions. There are many ketoreductases available
for screening in an attempt to identify a suitable enzyme
for a desired transformation.9 Keto-salinosporamide (2)
was screened against a panel of 98 NAD(P)H-dependent
ketoreductase enzymes using a spectrophometric assay
to indicate consumption of NAD(P)H cofactor as a
result of enzymatic reduction of the ketone.9b Through
this screening process, two NADH-dependent keto-
reductases, KRED-EXP-C1A and KRED-EXP-B1Y,
were shown to be catalytically active against our sub-
strate, with specific activities of 0.28 and 0.26 U/mg,
respectively. Small scale (1–2 mg) reactions were per-
formed on these two active enzymes10 using sodium for-
mate and formate dehydrogenase (FDH) as a cofactor
recycler. The products were extracted with EtOAc after
1 h and analyzed by HPLC, which indicated that both
enzymes selectively produced the desired product 1
(KRED-EXP-C1A: 68% conversion; KRED-EXP-
B1Y: 66% conversion) with about 20% starting material
remaining. These reactions were subsequently per-
formed at a 60 mg scale using the same FDH recycler,
Earlier, we reported the semi-synthesis of keto-salinos-
poramide (2) from 1 and the diastereoselective NaBH4
reduction of 2 to produce 3, the C-5 epimer of 1, with
about 90% de.3 The present study was designed to selec-
tively generate 1 from 2. We evaluated many commer-
cially available reagents for this stereoselective
reduction under a variety of reaction conditions (Table
1), none of which gave 1 as a single product. Most
*
Corresponding author. Tel.: +1 858 200 8334; fax: +1 858 587
4088; e-mail: vmacherla@nereuspharm.com
0040-4039/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2007.02.017

2538
R. R. Manam et al. / Tetrahedron Letters 48 (2007) 2537–2540
Table 1. Reduction of 2a with NaBH4
NaBH4 (# equiv)
Reaction conditions
Product ratio
Solventb
Temperature (ꢁC)
Time (min)
1
3
4
1
1
1
2
1
0.5
0.25
1
0.5
0.25
0.25
0.5
Monoglyme + 1% water
À78
À10
0
rt
rt
rt
rt
rt
rt
rt
rt
rt
0
14
14
14
8
8
8
5
95
50
33.3
0
10
50
50
0
10
40
0
50
0
Monoglyme + 1% water
Monoglyme + 1% water
Monoglyme + 1% water
Monoglyme + 1% water
Monoglyme + 1% water
Monoglyme + 1% water
IPA + 1% water
IPA + 1% water
IPA + 1% water
IPA + 5% water
IPA
30
33.3
50
45
50
50
50
60
50
10
40
30
20
33.3
50
45
0
8
0
10
12
8
8
8
50
30
10
10
10
0
0.5
IPA
8
70
0.25
IPA
rt
8
No reaction
1
0.5
0.25
1 + LiCl
1 + LiCl
1 + LiCl
1 + CeCl3
1 + CeCl3
1 + CeCl3
THF + 1% water
THF + 1% water
THF + 1% water
Monoglyme + 1% water
Monoglyme + 1% water
Monoglyme + 1% water
Monoglyme + 1% water
Monoglyme + 1% water
Monoglyme + 1% water
rt
rt
rt
10
12
8
10
10
10
10
10
10
50
50
30
5
27.2
10
5
0
50
70
95
36.4
30
95
50
60
50
0
0
À78
0
0
36.4
60
0
25
20
10
À78
0
25
20
10
a Degradation or little to no product was observed using the following reagents. (1) NaBH4 on 10% basic Al2O3, (2) K-Selectride, (3) KS-Selectride,
(4) BTHF-(R)-CBS, (5) BTHF-(S)-CBS, (6) NaBH(OAc)3, (7) (CH3)4NBH(OAc)3, and (8) iPrMgCl.
b Methyl and ethyl ester derivatives were formed when MeOH and EtOH were used, respectively.
which gave product 1 after 1 h with about 80% conver-
sion. The products were extracted with EtOAc and char-
acterized by LC–MS and NMR to confirm the structure
of 1, albeit with only 50–60% recovery under these
exploratory conditions.
but subsequent evaluation of the aqueous extract
revealed that a portion of the product had decomposed
to 5, which is an expected hydrolysis product that forms
in aqueous solution.3,11 Furthermore, the isolated yields
on a 50–100 mg scale were only 40%, further suggesting
decomposition. In a similar way, when KRED-EXP-
B1Y ketoreductase was used on 10 mg scale, the conver-
sion from 2 to 1 was about 90% complete after 1 h based
on organic extract analysis, but the analysis of the mix-
ture of aqueous and organic extracts (1:1) indicated the
presence of degradation product 5 in about 20%. When
the concentrations of the BIY enzyme and GDH were
doubled, the conversion was 90–95% after 40 min, with
minimal decomposition (2–5%) and isolated yields of
85–90%.12 Decomposition was also minimized when
biphasic solutions (50% aqueous t-BuOAc, n-BuOAc,
TBME) were used, but the percent conversion was gener-
ally very low even with longer reaction times (20–24 h),
except in 50% aqueous TBME (80% conversion; Table
2). We note that both KRED-EXP-C1A and KRED-
EXP-B1Y were used as lyophilized powders from crude
cell lysates, suggesting the potential for improved specific
To optimize the reaction conditions, these transforma-
tions (Scheme 1) were performed on 10–100 mg scale
using KRED-EXP-C1A or KRED-EXP-B1Y ketore-
ductases and glucose dehydrogenase (GDH) recycler in
place of FDH; all other reaction conditions were similar
to those used previously. The reactions were monitored
by an analytical HPLC method in which the two possible
products, 1 and 3, were completely resolved. The results
are summarized in Table 2. When KRED-EXP-C1A
ketoreductase was used, the conversion from 2 to 1 was
70% complete after 1 h on a 10 mg scale. Based on HPLC
analysis of the organic extract, the conversion was 90%
complete when the reaction time was increased to 3 h,
H
H
KRED (C1A/B1Y)
O
OH
O
pH 6.9
H
N
H
N
32-39 oC
O
O
H
H
H
OH
OH
O
O
OH
O
OH
O
H
N
H
N
H
N
N
AD
NA
DH
O
CH3
CH3
COOH
O
O
OH
CH3
O
GDH
Glucose
O
Gluconic acid
Cl
Cl
CH3
2
1
Scheme 1. Enzymatic reduction of 2 to 1 with NADH dependent
ketoreductases.
Cl
5
3
4
Cl

R. R. Manam et al. / Tetrahedron Letters 48 (2007) 2537–2540
2539
Table 2. Conversion of 2 to 1a
2 (mg)
KRED
# equiv (w/w)
GDH
# equiv (w/w)
% Solvent in water
Time
(h)
% Conversion to 1b
10
10
10
C1A
C1A
C1A
C1A
C1A
C1A
C1A
B1Y
B1Y
1
1
1
1
1
1
1
1
1
0.5
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
ꢀ20% DMSO
ꢀ20% DMSO
ꢀ20% DMSO
ꢀ20% DMSO
ꢀ20% DMSO
ꢀ20% DMSO
ꢀ20% DMSO
ꢀ20% DMSO
50% t-BuOAc
1
1
2
3
1
3
4
1
1
20
1
24
1
70
70
85
90
70
80c
90c
90
40
50
0
10
100
100
50
10
10
10
10
B1Y
B1Y
1
1
0.1
0.1
50% n-BuOAc
20
5
50% TBME
24
0.67
0.67
0.67
0.67
80
95
70
95d
90e
10
10
20
50
B1Y
C1A
B1Y
B1Y
2
2
2
2
0.2
0.2
0.2
0.2
ꢀ20% DMSO
ꢀ20% DMSO
ꢀ20% DMSO
ꢀ20% DMSO
a At pH 6.9 using GDH, NAD, glucose.
b Based on HPLC analysis of organic extract.
c Recovered yield 40% after purification by flash column chromatography. Some decomposition product (5) was detected in aqueous layer.
d Recovered yield 90% after purification by flash column chromatography.
e Recovered yield 85% after purification by crystallization.
H.; Neuteboom, S. T.; Richardson, P.; Palladino, M. A.;
Anderson, K. C. Cancer Cell 2005, 8, 407–419.
3. Macherla, V. R.; Mitchell, S. S.; Manam, R. R.; Reed, K.
A.; Chao, T.-H.; Nicholson, B.; Deyanat-Yazdi, G.; Mai,
B.; Jensen, P. R.; Fenical, W.; Neuteboom, S. T. C.; Lam,
K. S.; Palladino, M. A.; Potts, B. C. M. J. Med. Chem.
2005, 48, 3684–3687.
4. Groll, M.; Huber, R.; Potts, B. C. M. J. Am. Chem. Soc.
2006, 128, 5136–5141.
5. Reddy, L. R.; Saravanan, P.; Corey, E. J. J. Am. Chem.
Soc. 2004, 126, 6230–6231.
6. Endo, A.; Danishefsky, S. J. J. Am. Chem. Soc. 2005, 127,
8298–8299.
7. Reddy, L. R.; Fournier, J. H.; Reddy, B. V. S.; Corey, E.
J. Org. Lett. 2005, 7, 2699–2701.
8. Mulholland, N. P.; Pattenden, G.; Walters, I. A. S. Org.
Biomol. Chem. 2006, 4, 2845–2846.
9. (a) Kalaitzakis, D.; Rozzell, D. J.; Kambourakis, S.;
Smonou, I. Eur. J. Org. Chem. 2006, 2309–2313; (b)
Kaluzna, I. A.; Rozzell, J. D.; Kambourakis, S. Tetra-
hedron: Asymmetry 2005, 16, 3682–3689.
10. Protocol for small scale enzymatic reduction: Keto-
salinosporamide (2, 10 mM in DMSO) was mixed with
NAD+ (2 mM), ketoreductase (10 mg/mL, KRED-EXP-
C1A or KRED-EXP-B1Y), sodium formate (20 mM),
FDH-101 (2 mg/mL) for cofactor recycling, and sodium
phosphate buffer (1 mL, 150 mM, pH 6.9). The reaction
mixtures were incubated at 33 ꢁC for an hour, extracted
with ethyl acetate, evaporated to dryness in a speed
vacuum, redissolved in acetonitrile, and analyzed by
HPLC for product formation.
activities for these enzymes in their more highly purified
forms (BioCatalytics, personal communication).
In conclusion, the keto-salinosporamide (2) may serve as
a useful precursor in the total synthesis of 1. We
explored a variety of chemical reagents and reaction
conditions for the stereoselective conversion of 2 to 1
but were unable to achieve results that strongly favored
the desired product. We subsequently screened a library
of ketoreductases and identified two enzymes that can
be utilized to convert 2 to 1 with complete stereoselective
conversion to the desired product, with no evidence for
the undesired diastereomer. The KRED-EXP-B1Y
ketoreductase is superior to KRED-EXP-C1A in the
conversion of 2 to 1. Doubling the concentrations of
B1Y and GDH and decreasing the reaction time resulted
in better yields and minimal decomposition of product
(2–5%). Interestingly, while chemical reagents stereo-
selectively reduced 2 to 3, two ketoreductase enzymes
stereoselectively reduced 2 to the desired product 1.
Acknowledgments
We thank Dr. J. D. Rozzell and his group at BioCata-
lytics, Inc., Pasadena, CA, for enzyme screening and
helpful discussions.
11. Denora, N.; Potts, B. C. M.; Stella, V. J. J. Pharm. Sci., in
press.
References and notes
12. The optimized procedure for enzymatic transformation of
2 to 1: To a solution of 2 (50 mg, 0.16 mmol) in DMSO
(1 mL) in a round bottom flask (25 mL), 5 mL of
potassium phosphate buffer (150 mM, pH 6.9), 100 mg
of ketoreductase EXP-B1Y, 10 mg of GDH-103, 2.5 mL
of glucose (50 mM), and 2.5 mL of NAD (1 mM) were
added. The above reaction mixture was stirred at 37–39 ꢁC
1. Feling, R. H.; Buchanan, G. O.; Mincer, T. J.; Kauffman,
C. A.; Jensen, P. R.; Fenical, W. Angew. Chem., Int. Ed.
2003, 42, 355–357.
2. Chauhan, D.; Catley, L.; Li, G.; Podar, K.; Hideshima, T.;
Velonkar, M.; Mitsiades, C.; Mitsiades, N.; Yasui, H.;
Letai, A.; Ovaa, H.; Berkers, C.; Nicholson, B.; Chao, T.

2540
R. R. Manam et al. / Tetrahedron Letters 48 (2007) 2537–2540
for 40 min and then extracted with EtOAc (2 · 25 mL); the
(42 mg, 85% yield). The structure of 1 was confirmed
by comparison of its mp, specific rotation, and 1H and 13
NMR spectra with those of an authentic sample. Ketore-
ductases KRED-EXP-C1A and KRED-EXP-B1Y, GDH-
103, D-glucose, and NAD grade II free acid were obtained
from BioCatalytics, Pasadena, CA.
combined organic phase was dried over Na2SO4 and
concentrated under reduced pressure to yield a crude
product, which was crystallized in 1:1 acetone: heptane
(6 mL) in a 20 mL scintillation vial (by slow evaporation
under nitrogen gas) to afford 1 as a white crystalline solid
C