A formal catalytic asymmetric synthesis of (+)-biotin with modified cinchona alkaloids

Article abstract of DOI:10.1055/s-2001-16748

A formal catalytic asymmetric synthesis of (+)-biotin was realized. The key steps involve a catalytic, highly enantioselective and quantitative desymmetrization of a meso cyclic anhydride followed by a one-pot chemoselective reduction to form the optically active lactone intermediate in the Goldberg - Sternbach (+)-biotin synthesis.

Full text of DOI:10.1055/s-2001-16748

SPECIAL TOPIC
1737
A Formal Catalytic Asymmetric Synthesis of (+)-Biotin with Modified
Cinchona Alkaloids
A
symmetric Syn
huBiotin lho Choi, Shi-Kai Tian, Li Deng*
Department of Chemistry, Brandeis University, Waltham, MA 02454-9110, USA
Fax +1(781)7362516; E-mail: deng@brandeis.edu
Received 14 June 2001
and practicality of the Goldberg–Sternbach strategy
(Scheme 1) results from its ability to take advantage of a
readily available symmetric bifunctional starting material,
fumaric acid (1), for the efficient construction of the bicy-
Abstract: A formal catalytic asymmetric synthesis of (+)-biotin
was realized. The key steps involve a catalytic, highly enantioselec-
tive and quantitative desymmetrization of a meso cyclic anhydride
followed by a one-pot chemoselective reduction to form the optical-
ly active lactone intermediate in the Goldberg–Sternbach (+)-biotin clic thiolactone intermediate 5, and subsequently, to capi-
synthesis.
talize on the sterically well-differentiated two faces of the
cis-fused bicyclic ring in 5 to realize the controlled cre-
ation of the third stereogenic center during the introduc-
tion of the side chain to give biotin (6).
Key words: biotin, vitamin, cinchona alkaloids, asymmetric catal-
ysis, enantioselective desymmetrization
The potential of the Goldberg–Sternbach approach has,
however, not yet been fully realized due to the lack of an
efficient method for the desymmetrization of meso inter-
mediates such as acid 2, anhydride 3, or diester 7 to form
optically active lactone 4 or thiolactone 5. Highly efficient
routes for the preparation of biotin (6) from either optical-
ly active 4 or 5 have been established.5–10 Matsuki and
Chen recently reported enantioselective reduction of an-
hydride 3 and thioanhydride 9 to the corresponding lac-
tone 4 (90% ee)11 and thiolactone 5 (98.5% ee),12
respectively. In spite of their brevity and high enantiose-
The unusual structure, biological significance, and com-
mercial importance of biotin (6) have made it a fascinat-
ing target for total synthesis both in academia and industry
for more than fifty years.1 The total synthesis of biotin has
become a test case for evaluation of new strategies that are
conceived upon emergence of new synthetic methods.
Among numerous approaches reported to date, the first
synthesis of biotin (6) described by Goldberg and Stern-
bach, after subsequent modifications, is still one of the
best syntheses of biotin.2–4 The unsurpassed efficiency
O
asymmetric
reduction
BnN
NBn
Na2S.9H2O
O
O
S
9
O
O
O
S
O
asymmetric
reduction
CH3COSK
HOOC
3 steps
Ac2O, reflux
BnN
NBn
BnN
NBn
BnN
NBn
BnN
NBn
DMF, 150 °C
COOH
O
HOOC
COOH
O
O
O
O
O
4
1
5
3
2
asymmetric
alcoholysis
3 steps
O
ester reduction
/lactonization
O
O
asymmetric
hydrolysis
HN
NH
( )
BnN
NBn
BnN
NBn
COOH
ROOC
COOR
ROOC
COOH
S
4
8
7
6: (+)-Biotin
Scheme 1
Synthesis 2001, No. 11, 28 08 2001. Article Identifier:
1437-210X,E;2001,0,11,1737,1741,ftx,en;C03601ss.pdf.
© Georg Thieme Verlag Stuttgart · New York
ISSN 0039-7881

1738
C. Choi et al.
SPECIAL TOPIC
lectivity, both approaches required the use of a large ex- We recently discovered a general and highly enantioselec-
cess of Noyori’s BINAL-H (Figure 1) (4.0 equiv for the tive catalytic desymmetrization of meso and prochiral cy-
conversion of 3 to 4, and 3.0 equiv for the conversion of 9 clic anhydrides catalyzed by commercially available
to 5) as the reducing agent.
mono and bis-cinchona alkaloid derivatives (Figure 2).16
Particularly noteworthy is the effectiveness of this chiral
Lewis base-catalyzed reaction with bulky as well as fuc-
tionalized cyclic anhydrides, which suggests that it may
be able to desymmetrize anhydride 3. Following our pre-
viously reported conditions, we first examined the desym-
metrization of anhydride 3 in diethyl ether with
(DHQD)2AQN as the catalyst and methanol as the nucleo-
phile. While (DHQD)2AQN was found to be a highly ef-
ficient and general catalyst for the desymmetrization of a
wide variety of meso cyclic anhydrides such as cis-2,3-
dimethylsuccinic anhydride (10), it catalyzed the desym-
metrization of 3 to give hemiester 11a in only 59% ee at
room temperature (entry 1, Table).
The bulky size of and the presence of multiple polar and
Lewis basic functionalities in meso intermediates 3 and 7
make the development of catalytic methods for their de-
symmetrization using either enzymes or chiral transition
metal-complexes extremely difficult. For example, al-
though Pig liver esterase (PLE)-catalyzed hydrolysis is
known to be a general and efficient enzymatic method for
the transformation of meso diesters to optically active he-
miesters, it afforded only modest enantioselectivity (75%
ee) in the hydrolysis of meso diester 7 to form hemiester
8.13 Sih devised an enzymatic resolution route to convert
2 to lactone 4 with high enantioselectivity (93% ee), but it
involved an eight-step sequence.14 Attempts to effect
enantioselective ring-opening of anhydride 3 with a chiral After screening other modified mono and bis-cinchona al-
Lewis acid promoter have also met with limited success. kaloids, we found that the catalyst structure-enantioselec-
While Ti-TADDOLates-promoted (Figure 1) alcoholysis tivity profile with anhydride 3 as the substrate is
was highly enantioselective for a broad range of meso cy- substantially different from that with cis-2,3-dimethylsuc-
clic anhydrides, its application to the desymmetrization of cinic anhydride (10) as shown in the Table. However, we
3 resulted in a low enantioselectivity (26% ee).15 Ideally, were pleased to find that DHQD-PHN, although slightly
desymmetrization of meso intermediates along the Gold- inferior to (DHQD)2AQN with simple cyclic anhydrides,
berg–Sternbach route should be accomplished in high is much more effective than (DHQD)2AQN for the de-
enantioselectivity and yield with a catalytic amount of a symmetrization of anhydride 3 to afford 11a in 89% ee at
readily available chiral reagent. Here we present our ef- room temperature (entry 7, Table). The ee was improved
forts towards accomplishing such a goal.
to 93% when the reaction was carried out at –40 °C. More-
over, the reaction proceeded cleanly allowing both the
isolation of hemiester 11a and the recovery of the catalyst
in quantitative yield without chromatographic purifica-
tion. Further reducing the reaction temperature to –60 °C
resulted in a slight decrease of enantioselectivity (entry
11, Table). The absolute configuration of hemiester 11a
was determined to be 4R and 5S, which is consistent with
our previously proposed stereochemical projection for
modified cinchona alkaloid-catalyzed desymmetrization
of cyclic anhydride.16 The employment of DHQ-PHN led
to the generation of 11b, the antipode of 11a, but surpris-
Ar
Ar
H
O
O
H
O
O
O
OiPr
OiPr
Al
Ti
OEt
O
H
Ar
Ar
(R)-BINAL-H
Ti-TADDOLate
Figure 1 Structures of (R)-BINAL-H and Ti-TADDOLate
DHQD
Ph
DHQD
OMe
N
DHQD
DHQD
O
N
N
N
Me
Ph
H
N
DHQD-MEQ
DHQD(dihydroquinidyl)
(DHQD)2PYR
DHQD-PHN
O
DHQD
DHQD
O
DHQD
OMe
DHQD
N
N
O
N
Cl
O
H
DHQD
N
DHQD-CLB
(DHQD)2PHAL
DHQ(dihydroquinyl)
(DHQD)2AQN
Figure 2 Catalysts used for asymmetric ring-opening of anhydrides 3 and 10
Synthesis 2001, No. 11, 1737–1741 ISSN 0039-7881 © Thieme Stuttgart · New York

SPECIAL TOPIC
Asymmetric Synthesis of (+)-Biotin
1739
Table Asymmetric Ring-Opening of Anhydrides 3 and 10 with
Natural and Modified Cinchona Alkaloids
ingly, in significantly lower ee than that obtained with
DHQD-PHN (entry 12 vs.10, Table).
O
Having succeeded in establishing the first highly efficient
catalytic desymmetrization of anhydride 3, we next fo-
cused our attention on the development of an efficient
conversion of hemiester 11a to lactone 4. A chemoselec-
tive reduction of the carboxylic acid group in hemiester
11a would provide 13 as the precursor for the formation
of lactone 4 to complete a formal catalytic asymmetric
O
BnN
NBn
BnN
NBn
R
S
O
O
O
O
O
O
OH OMe
Catalyst
3
11a
MeOH
Et2O
Me
Me
Me
Me
synthesis of (+)-biotin. Treatment of hemiester 11a with
R
S
17,18
BH3SMe2
followed by ring cyclization of 13 in the
O
O
O
O
presence of aqueous HCl resulted in the formation of lac-
tone 4 in 40–50% isolated yield (Scheme 2). We were es-
pecially disappointed to observe that the ee of lactone 4
was found to fluctuate in a range (44–75%) that was sig-
nificantly lower than the ee (93%) of the starting he-
miester 11a.
OH OMe
12
10
Entry Catalyst (mol%)
T (°C)
ee (%)
of 11a
ee (%)
of 12a
1
2
(DHQD)2AQN (5)
25
59
~0
46
53
41
73
89
90
91
93
88
70b
93
62
72
70
61
69
92
-
This disparity between the enantiomeric excesses of he-
miester 11a and lactone 4 could be caused by a partial ra-
cemization of 11a via a scrambling of the ester group
mediated by a Lewis acidic boron species as shown in
Scheme 2 (11a to 11b). When the borane reduction was
stopped at partial conversion, the ee of the recovered he-
miester 11a was found to decrease from 93% to 79%, thus
confirming that optically active hemiester 11a racemized
during the borane reduction. Given that highly enantio-
merically enriched 11a was obtained from an amine-cata-
lyzed reaction, we examined procedures involving
reduction of carboxylic acids to alcohols under mildly ba-
sic conditions. A slight modification of a one-pot proce-
dure reported by Falorni, utilizing cyanuric chloride, N-
methylmorpholine and NaBH4 as stoichiometric
reagents,19 generated lactone 4 from hemiester 11a in
91% ee and 82% isolated yield, thereby completing a for-
mal catalytic asymmetric synthesis of (+)-biotin
(Scheme 3).5–10
(DHQD)2PHAL (5)
(DHQD)2PYR (5)
quinidine (10)
25
3
25
4
25
5
DHQD-MEQ (10)
DHQD-CLB (10)
DHQD-PHN (10)
DHQD-PHN (20)
DHQD-PHN (20)
DHQD-PHN (20)
DHQD-PHN (20)
DHQ-PHN (20)
25
6
25
7
25
8
0
9
–20
–40
–60
–40
-
10
11
12
-
-
-
a For the ee determination of 12, see supporting information given in
Ref.16
b Compound 11b, the antipode of 11a, is formed as the major enan-
In summary, we have developed a two-step sequence to
accomplish a highly enantioselective conversion of meso
tiomer.
O
O
O
O
Bn
Bn
N
Bn
N
N
.
aq. HCl
BH3 SMe2
OMe
OH
OMe
OH
O
O
O
O
N
N
N
Bn
Bn
Bn
13
4 (40-50% yield, 44-75%ee)
11a (93% ee)
BLn
BLn
O
BLn
-OMe
O
O
O
Bn
-OMe
Bn
Bn
N
Bn
N
N
OMe
N
OH
O
O
O
O
O
O
OH
O
N
N
N
N
OMe
Bn
Bn
Bn
Bn
O
O
BLn
O
11b
Scheme 2
Synthesis 2001, No. 11, 1737–1741 ISSN 0039-7881 © Thieme Stuttgart · New York

1740
C. Choi et al.
SPECIAL TOPIC
25
4:1, tmajor = 26.9 min, tminor = 20.5 min);23 [ ]D –7.3 (c = 1.56,
O
O
CHCl3).
DHQD-PHN
(20 mol %)
BnN
NBn
93% ee
100% yield
BnN
O
NBn
1H NMR (400 MHz, CDCl3): = 3.59 (s, 3 H), 4.02–4.13 (m, 4 H),
4.97 (d, J =14.8 Hz, 1 H), 5.08 (d, J =14.8 Hz, 1 H), 7.18–7.39 (m,
10 H), 10.53 (br s, 1 H).
13C NMR (100 MHz, CDCl3): = 46.9, 47.0, 52.8, 57.0, 57.5, 128.1,
128.7, 128.8, 128.97, 129.01, 135.6, 160.0, 168.7, 171.4.
(5S)
(4R)
MeOH, Et2O
–40 °C
O
O
O
O
OH OMe
3
11a
1) cyanuric chloride
NMM, THF
2) NaBH4, H2O
3) aq. HCl
To recover the catalyst (DHQD-PHN), K2CO3 was added to the
aqueous layer to adjust the pH value of the solution to 9–11. The re-
sulting mixture was extracted with EtOAc (50 mL), and the organic
layer was washed with brine, dried (Na2SO4) and concentrated to af-
ford the catalyst in quantitative yield and in pure form as shown by
1H NMR. The recovered catalyst was used for a new batch of reac-
tion to give 11a in the same ee and yield as described above.
O
O
See: Ref.5-10
BnN
(6aR)
NBn
HN
NH
91% ee
82% yield
(3aS)
COOH
( )
S
(3aS,6aR)-1,3-Dibenzyltetrahydro-1H-furo[3,4-d]imidazole-
2,4-dione (4)11
O
O
4
6
4
To a solution of cyanuric chloride (159 mg, 0.86 mmol) in THF (3
mL) at r.t. was added N-methylmorpholine (77 mg, 0.74 mmol). A
white suspension was formed. A solution of the hemiester (210 mg,
0.57 mmol, 93% ee) in THF (2 mL) was added to this mixture. The
resulting mixture was stirred for 3 h and then filtered. A solution of
NaBH4 (46 mg, 1.2 mmol) in H2O (1 mL) was added dropwise to
the filtrate at 0 °C. The resulting mixture was stirred for 6 min after
which 2 N aq HCl (5 mL) was added to the mixture. The resulting
mixture was stirred at r.t. for 3 h and then extracted with EtOAc (20
mL). The organic layer was washed with brine, dried (MgSO4), and
concentrated. The residue was purified by chromatography (silica
gel, EtOAc–hexane, 1:4) to give the desired chiral lactone as a white
solid: 151 mg (82%). The ee of the lactone was determined to be
91% (HPLC conditions: Daicel Chiralpak OD, = 254 nm, hex-
Scheme 3
anhydride 3 to chiral lactone 4 in 82% overall yield. The
absolute stereochemistry is controlled via a catalytic,
highly enantioselective and quantitative desymmetriza-
tion of anhydride 3 using a catalytic amount of DHQD-
PHN, a commercially available and fully recyclable mod-
ified cinchona alkaloid. With the new development de-
scribed here, the elegant and practical approach outlined
by Goldberg and Sternbach more than fifty years ago is re-
alized in a catalytic asymmetric synthesis of biotin.
25
ane:propan-2-ol, 9:1, tmajor = 37.7 min, tminor = 32.1 min); [ ]D
+56.4 (c = 1.12, CHCl3) {Lit.11 [ ]D25 +52.2 (c = 1.03, CHCl3; 90%
ee)}.
1H and 13C spectra were recorded on a Varian instrument (400 MHz
and 100 MHz, respectively) and internally referenced to tetrameth-
ylsilane signal. Specific rotations were measured on a Jasco Digital
Polarimeter. Liquid chromatography was performed using forced
flow (flash chromatography) of the indicated solvent system on EM
science silica gel 60 (SiO2, 230–400 mesh). HPLC analyses were
perfomed on a Hewlett-Packard 1100 series instrument equipped
with a quaternary pump using OD (Daicel Chiralpak, 250 4.6
mm). UV variable wavelength detector was used to monitor at 254
nm or 280 nm. All reactions were conducted in oven or flame dried
glassware under inert atmosphere of dry N2. Et2O and THF were
distilled from sodium ketyl of benzophenone immediately before
use. Toluene and MeOH were distilled from CaH2 and stored over
molecular sieves type 4 Å. cis-1,3-Dibenzyltetrahydro-2H-fu-
ro[3,4-d]imidazole-2,4,6-trione (3) was prepared according to the
literature procedure.10,20
1H NMR (400 MHz, CDCl3): = 3.92 (d, J = 8.4 Hz, 1 H), 4.08–
4.17 (m, 3 H), 4.31–4.39 (m, 2 H), 4.62 (d, J = 15.6 Hz, 1 H), 5.06
(d, J = 14.6 Hz, 1 H), 7.24–7.39 (m, 10 H).
13C NMR (100 MHz, CDCl3): = 45.4, 47.1, 52.6, 54.6, 70.3, 128.0,
128.3, 128.4, 128.9, 129.0, 129.1, 136.1, 136.2, 158.4, 173.0.
Acknowledgement
We gratefully acknowledge the financial support from NIH (R01-
GM61591), Harcourt General Charitable Foundation, and Research
Corporation (RI-0311).
References
(1) De Clercq, P. J. Chem. Rev. 1997, 97, 1755.
(2) (a) Goldberg, M. W.; Sternbach, L. H. US Patent 2489232,
1949. (b) Chem. Abstr. 1951, 45, 184.
(3) (a) Goldberg, M. W.; Sternbach, L. H. US Patent 2489235,
1949. (b) Chem. Abstr. 1951, 45, 186a.
(4) (a) Goldberg, M. W.; Sternbach, L. H. US Patent 2489238,
1949. (b) Chem. Abstr. 1951, 45, 186h.
(5) Field, G. F. J. Org. Chem. 1978, 43, 1084.
(6) (a) Isaka, I.; Kubo, K.; Takashima, M.; Murakami, M.
Yakugaku Zasshi 1968, 88, 964. (b) Chem. Abstr. 1969, 70,
19984.
(7) (a) Holick, W.; Pauling, H.; Cottong, N. A. Eur. Patent.
Appl. EP 0154225 A2, 1985. (b) Chem. Abstr. 1986, 104,
168273.
(8) (a) Sumitomo Chemical Co. Ltd. Jpn. Kokai, Tokyo Koho JP
60100583, 1985. (b) Chem. Abstr. 1985, 103, 141752.
(9) Shimizu, T.; Seki, M. Tetrahedron Lett. 2000, 41, 5099.
(4R,5S)-1,3-Dibenzyl-5-(methoxycarbonyl)-2-oxoimida-
zolidine-4-carboxylic Acid (11a)21,22
Anhyd MeOH (480 mg, 0.60 mL, 15 mmol) was added dropwise to
a mixture of anhydride 3 (504 mg, 1.5 mmol) and DHQD-PHN (151
mg, 0.30 mmol, 20 mol%) in Et2O (75 mL) at –40 °C. The resulting
mixture was stirred for 28 h at –40 °C. The reaction mixture became
clear and the anhydride was completely consumed as shown by
TLC analysis. Aq HCl (1.0 N, 30 mL) was added to the mixture.
The organic phase was collected and the aqueous layer was extract-
ed with EtOAc (100 mL). The combined organic phases were
washed with brine, dried (Na2SO4) and concentrated to give he-
miester 11a (550 mg, ~100%) as a solid. The ee of 11a was deter-
mined to be 93% by HPLC analyses of a diastereomeric mixture of
amide-esters prepared from 11a and (R)-(+)-1-(1-napthyl)ethy-
lamine according to the literature procedure (HPLC conditions: Da-
icel Chiralpak, OD, = 280 nm, 0.6 mL/min, hexane:propan-2-o,
Synthesis 2001, No. 11, 1737–1741 ISSN 0039-7881 © Thieme Stuttgart · New York

SPECIAL TOPIC
Asymmetric Synthesis of (+)-Biotin
1741
(10) Gerecke, M.; Zimmermann, J.-P.; Aschwanden, W. Helv.
Chim. Acta 1970, 53, 991.
(18) Yoon, N. M.; Pak, C. S.; Brown, H. C.; Krishnamurthy, S.;
Stocky, T. P. J. Org. Chem. 1973, 38, 2786.
(11) Matsuki, K.; Inoue, H.; Takeda, M. Tetrahedron Lett. 1993,
34, 1167.
(19) Falorni, M.; Porcheddu, A.; Taddei, M. Tetrahedron Lett.
1999, 40, 4395.
(12) Chen, F.-E.; Huang, Y.-D.; Fu, H.; Cheng, Y.; Zhang, D.-
M.; Li, Y.-Y.; Peng, Z.-Z. Synthesis 2000, 2004.
(13) Iriuchijima, S.; Hasegawa, K.; Tsuchihashi, G. Agric. Biol.
Chem. 1982, 46, 1907.
(14) Wang, Y.-F.; Sih, C. J. Tetrahedron Lett. 1984, 25, 4999.
(15) Jaeschke, G.; Seebach, D. J. Org. Chem. 1998, 63, 1190.
(16) Chen, Y.; Tian, S.-K.; Deng, L. J. Am. Chem. Soc. 2000,
122, 9542.
(20) Lavielle, S.; Bory, S.; Moreau, B.; Luche, M. J.; Marquet, A.
J. Am. Chem. Soc. 1978, 100, 1558.
(21) Matsuki, K.; Inoue, H.; Ishida, A.; Takeda, M.; Nakagawa,
M.; Hino, T. Chem. Pharm. Bull. 1994, 42, 9.
(22) Shimizu, M.; Matsukawa, K.; Fujisawa, T. Bull. Chem. Soc.
Jpn. 1993, 66, 2128.
(23) Hiratake, J.; Inagaki, M.; Yamamoto, Y.; Oda, J. J. Chem.
Soc., Perkin. Trans. 1 1987, 1053.
(17) Theisen, P. D.; Heathcock, C. H. J. Org. Chem. 1993, 58,
142.
Synthesis 2001, No. 11, 1737–1741 ISSN 0039-7881 © Thieme Stuttgart · New York