A novel one-step diastereo- and enantioselective formation of trans- azetidinones and its application to the total synthesis of cholesterol absorption inhibitors

Article abstract of DOI:10.1021/jo990428k

An efficient and practical asymmetric process was developed for the synthesis of azetidinone-based cholesterol absorption inhibitors. Key to this synthesis was the discovery of a novel one-step diastereo- and enantioselective formation of trans β-lactams starting from commercially available 3(S)-hydroxy-γ-lactone. Various trans β-lactams can be prepared in good yields and with better than 95:5 enantio- and diastereoselctivity. A Lewis acid-catalyzed aldol condensation and a highly enantioselective oxazaborolidine-catalyzed chiral reduction completes the side chain.

Full text of DOI:10.1021/jo990428k

3714
J . Org. Chem. 1999, 64, 3714-3718
A Novel On e-Step Dia ster eo- a n d En a n tioselective F or m a tion of
tr a n s-Azetid in on es a n d Its Ap p lica tion to th e Tota l Syn th esis of
Ch olester ol Absor p tion In h ibitor s
Guangzhong Wu,* YeeShing Wong, Xing Chen, and Zhixian Ding
Chemical Process Research and Development, Schering-Plough Research Institute,
2015 Galloping Hill Road, Kenilworth, New J ersey 07033
Received March 9, 1999
An efficient and practical asymmetric process was developed for the synthesis of azetidinone-based
cholesterol absorption inhibitors. Key to this synthesis was the discovery of a novel one-step
diastereo- and enantioselective formation of trans â-lactams starting from commercially available
3(S)-hydroxy-γ-lactone. Various trans â-lactams can be prepared in good yields and with better
than 95:5 enantio- and diastereoselctivity. A Lewis acid-catalyzed aldol condensation and a highly
enantioselective oxazaborolidine-catalyzed chiral reduction completes the side chain.
In tr od u ction
diastereo- and enantioselective formation of trans â-lac-
tams starting from commercially available 3(S)-hydroxy-
γ-lactone.
Azetidinones are a very important class of compounds
possessing a wide range of biological activities1 and are
also shown to be useful synthons for various pharma-
ceutical products.2 Recently, trans azetidinones 1 and 2
have been found to be potent cholesterol absorption
inhibitors (CAI) and have shown efficacy in clinical trials
in reducing cholesterol levels.3 Various synthetic methods
have been developed to establish the two chiral centers
on the â-lactam ring.4 These include chiral auxiliary
based chemistry,5a chiral pool derived methodologies,5b
and catalytic asymmetric synthesis.5c For the synthesis
of 1, a variant of Evan’s auxiliary-based chemistry was
developed.5a In addition to the recovery of the auxiliary,
this method required several steps to construct the
â-lactam ring. For the synthesis of 2, we designed and
developed a short and practical process which provides
an easy access for this type of â-lactam structure. Key to
this synthesis was the discovery of a novel one-step
Resu lts a n d Discu ssion
Our synthetic strategy was to construct the â-lactam
ring with the desired chiral centers followed by creation
of needed side chain. The benzylic chiral center on the
side chain could be obtained via an oxazaborolidine6 or
CBS-catalyzed reduction of a ketone precursor. For the
synthesis of the â-lactam ring, we envisioned a one-step
reaction between (S)-3-hydroxy-γ-lactone, 4, and an ap-
propriately substituted imine. A condensation reaction
between an open-chain hydroxy ester and an imine are
well documented but they give cis-â-lactams as the major
product.5b Although condensation of lactone 4 with an
aldehyde is known,7 reaction with an imine has not been
reported. However, this type of reaction offers an op-
portunity to form â-lactams in one step. Lactone 4 is
readily derived from a glucose and is available commer-
ciallly.8 The existing hydroxy chiral center on the rigid
lactone ring should induce the desired C-3 chiral center
(Scheme 1).
(1) (a) Katritzky, A. R.; Rees, C. W.; Scriven, E. F. V. Comprehensive
Heterocyclic Chemistry II, Vol. 1B; Padwa, A., Ed. Pergamon, Elsevier
Sci. LTD: Oxford, U.K., 1996; p 536 and references therein. (b) Morin,
R. B.; Gorman, M. Chemistry and Biochemistry of â-Lactam Antibiotics,
Vols. 1-3; Academic: New York, 1982.
(2) (a) Ojima, I. Acc. Chem. Res. 1995, 28, 383. (b) Banik, B. K.;
Manhas, M. S.; Bose, A. K. J . Org. Chem. 1993, 58, 307.
(3) (a) Rosenblum, S. B.; Huynh, T.; Afonso, A.; Davis, H. R.; Yumibe,
N.; Clader, J . W.; Burnett, D. J . Med. Chem. 1998, 41, 973. (b) Clader,
J . W.; Burnett, D. A.; Caplen, M. A.; Domalski, M. S. Dugar, S.;
Vaccaro, W.; Sher, R.; Browne, M. E.; Zhao, H.; Burrier, R. E.;
Salisbury, B.; Davis, H. R., J r. J . Med. Chem. 1996, 39, 3684. (c)
McKittrick, B. A.; Dugar, S.; Clader, J . W.; Davis, H. J r.; Czarniecki,
M. Bioorg. Med. Chem. Lett. 1996, 6, 1947.
(4) For
a review: (a) Georg, G. I. The Organic Chemistry of
Our early work started with generation of a dianion
of lactone 4 with lithium diisopropylamide (LDA) at -35
°C in THF followed by addition of an imine and hexameth-
ylphosporic triamide (HMPA) to give predominately the
two adducts shown in Table 1. A warm up of the reaction
did not drive the cyclization to completion. The low
reactivity of the two intermediates is most likely due to
the formation of stable lithium aggregates.9 We specu-
lated that an additional lithium salt may be able to form
â-Lactams; VCH: New York, 1993. (b) Patai, S.; Rappoport, Z.
Synthesis of Lactones and Lactams; Ogliaruso, M. A., Wolf, J . F., Eds.;
J ohn Wiley and Sons: New York, 1993. (c) Reference 1a.
(5) (a) Thiruvengadam, T. K.; Tann, C. H.; Lee, J .; McAllister, T.;
Sudhakar, A. U.S. Patent 5306817, 1994; Chem. Abstr. 1994, 112,
157423e. Chen, L.; Zaks, A.; Chacklamanil, S.; Dugar, S. J . Org. Chem.
Soc. 1996, 61, 8341. (b) Barbaro, G.; Battaglia, A.; Guerrini, A.;
Bertucci, C. Tetrahedron: Asymmetry 1997, 8, 2527. Georg, G. I.;
Akgun, E. Tetrahedron Lett. 1990, 31, 3267. Tschan, D. M.; Fuentes,
J . E.; Lynch, J . E.; Laswell, W. L.; Volante, R. P.; Shinkai, I.
Tetrahedron Lett. 1988, 29, 2779. Oguni N.; Ohkawa, Y. J . Chem. Soc.,
Chem Commun. 1988, 1376. (c) Wu, G.; Tormos, W. J . Org. Chem.
1997, 62, 6412. Fujieda, H.; Kanai, M.; Kambara, T.; Iida, A.; Tomioka,
K. J . Am. Chem. Soc. 1997, 119, 2060.
(7) Shieh, H.-M.; Prestwich, G. D. J . Org. Chem. 1981, 46, 4321.
(8) (a) Hollingsworth, R. I. U.S. Patent 5,292,939. (b) Synthon Corp.,
3900 Collins Road, Lansing, MI 48910.
(6) Corey, E. J .; Bakshi, R. K.; Shibata, S.; Singh, V. K. J . Am. Chem.
Soc. 1987, 109, 7925.
10.1021/jo990428k CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/21/1999

Formation of trans-Azetidinones
J . Org. Chem., Vol. 64, No. 10, 1999 3715
Ta ble 2. On e-Step F or m a tion of Tr a n s â-La cta m s
Sch em e 1
trans-3:cis-3 HPLC
in solution
isolated
item
Ar1
Ar2
yield, % yield, %a
Ta ble 1. Meta l a n d Tem p er a tu r e Effects on
Dia ster eoselectivity
3a
3b
3c
3
4-BnOPh- 4-MeOPh-
4-BnOPh- Ph-
90:10
86:14
99:1
90:10
92:8
n/a
70
75
70
68
57
23
70
70
64
62
51
Ph-
Ph-
4-BnOPh- 4-FPh-
3d
3e
Ph-
4-FPh-
4-FPh-
4-FPh-
97:3
a
In all cases, the ratios of trans-3:cis-3 in isolated products were
better than 95:5.
temp, convn,
trans-3: yield,
procedure was developed to further enhance the diaster-
eomeric ratio in isolated 3 from 90:10 to better than 95:
5. Under the optimized conditions, the reaction was
scaled up to 300 g smoothly and the isolated yield
improved from 35% to 64%. This reaction represents the
first one-step enantio- and diastereoselective, high yield-
ing, and practical synthesis of trans â-lactams starting
from (S)-3-hydroxy-γ-lactone.
b
no.
base
°C
%
I:II
cis-3a
%
1
2
3
4
5
Et2Zn/LDA (1:1)
LDA
-25
-25
80
99
65
98
99
11:89 40:60
73:23 87:13
86:14 99:1
69:31 90:10
79:21 90:10
7
76
46
68
78
NaHMDA/LiHMDA -25
LDA
LDA
-15
-35
These are the ratios in reaction mixture. b These are total yield
of trans- and cis-3 as determined by HPLC.
a
This one-step azetidinone formation is quite general
for arylimines, and various trans â-lactams can be
prepared as shown in Table 2. Three salient features
deserve mention. First, the reaction works well with both
electron-withdrawing and -donating substituents on the
aromatic rings. Second, the diastereoselectivity of six
reactions ranges from 86:14 to 99:1 in crude products.
The ratios are further enhanced after crystallization.
Third, good isolated yields were achieved for all examples
except 3a where solubility of the imine was poor. It is
interesting to compare our results with a previous report
where an open-chain 3-hydroxybutyrate was used as a
starting material.11 The predominant products in both
cases have the same absolute stereochemistry for the
â-lactam ring despite the opposite stereochemistry for the
side-chain alcohols. This clearly demonstrates the effect
of the butyrolactone ring on the dianion configuration and
reactivity. A brief investigation indicated that this reac-
tion does not work well with imines bearing alkyl groups.
With a good trans â-lactam method in hand, we turned
our attention to elaborating the side chain of desired
chirality. We envisioned an oxidation of the diol to an
aldehyde followed by an aldol condensation to complete
the side-chain framework (Scheme 2). Thus, the 95:5 ratio
of 3 was first subjected to an oxidative cleavage with
NaIO4 to form aldehyde 5. The cis isomer was not stable
under the reaction conditions and epimerized at C-3 to
give the opposite enantiomer of the desired product,
resulting in a 100% trans â-lactam (90% ee.). LDA-
promoted direct condensation of 4-fluoroacetophenone
with aldehyde 5 gave a complex mixture, presumably due
to enolate exchange. We solved this problem by carrying
out the reaction under Mukaiyama aldol condensation
conditions. First, enolization of 4-fluoroacetophenone
with LDA followed by trapping with TMSCl gave enol
trimethylsilyl ether in 95% yield. A TiCl4-catalyzed
reaction of enol trimethylsilyl ether with 5 afforded both
diastereomers of hydroxy ketone 6. The condensation
mixed aggregates and increase the reactivity.10 To our
delight, addition of LiCl does accelerate the cyclization
of these adducts to give a mixture of two â-lactams in
35% isolated yield. The imine condensation reaction gives
exclusive control over C-3 but only moderate diastereo-
selectivity at the C-4 (or benzylic) position with an initial
ratio of 73:27 (I:II) in favor of the desired SSS intermedi-
ate. The intermediate ratio was determined by HPLC on
the quenched samples. We also observed that in this
reaction the rate of cyclization of intermediate I to trans-3
is about four times as fast as that of II to cis-3, enhancing
the ratio of trans to cis from 73:27 (I:II) to 87:13 (trans-
3:cis-3). To improve both the selectivity and the yield,
this reaction was studied carefully using different metals,
solvents, additives, concentrations, and temperatures.
Examination of the cation effect on the diastereoselec-
tivity indicated that the weaker the coordination ability
the better the selectivity as shown in Table 1. For
example, a 1:1 ratio of Et2Zn/LDA gave an 11:89 ratio of
I:II while 1:1 ratio of NaHMDA/LiHMDA reversed the
ratio to 86:14. Hence, either the trans or the cis isomer
can be obtained from this condensation reaction depend-
ing upon which metal is used. Chelation with either the
enolate anion or the alkoxy anion may have played a role
in the change of the selectivity.
Next we optimized the reaction conditions using LDA
because it gave the best yield. N,N′-Dimethylpropyl-
eneurea (DMPU) was introduced to replace the toxic
HMPA. In the absence of either HMPA or DMPU, no
addition takes place. It was also found that polar
solvents, such as DMF, increased the reaction rate
significantly. The selectivity could be further improved
by carrying out the reaction at -35 °C. Reaction tem-
peratures below -35 °C resulted in poor solubility of
imines and slow reaction rate. A simple precipitation
(9) (a) Gilchrist, J . H.; Collum, D. B. J . Am. Chem. Soc. 1992, 114,
794.
(10) McGarrity, J . F.; Ogle, C. A.; Brich, Z. Loosli, H. R. J . Am.
Chem. Soc. 1985, 107, 1810.
(11) Georg, G. I.; Kant, J .; Gill, H. S. J . Am. Chem. Soc. 1987, 109,
1129.

3716 J . Org. Chem., Vol. 64, No. 10, 1999
Wu et al.
materials were purchased commercially. Melting points are
not corrected.
Sch em e 2
Gen er a l Meth od P r ep a r a tion of 3. To a dry 5-L three-
necked flask equipped with a mechanical stirrer, thermometer,
and addition funnel were added sequentially 500 mL of THF,
400 mL of DMPU, and 120 mL (0.92 mol) of diisopropylamine.
To the cooled mixture at -40 to -45 °C was added dropwise
368 mL (0.92 mol) of 2.5 M n-BuLi hexane solution. After 20
min, 47 g (0.46 mol) of lactone 4 diluted in 250 mL of THF
was introduced and the reaction was agitated at -40 to -45
°C for 2 h. To the resulting mixture was added dropwise a
solution of 100 g (0.33 mol) of imine in 1 L of DMF over a
period of 30 min. The reaction was maintained at -25 to -30
°C for 14-18 h and warmed to -13 to -17 °C for another 4 h.
To the reaction mixture was added 14 g of LiCl dissolved in
400 mL of DMF. After another 2 h at -15 °C, 200 mL of HOAc
was added to the reaction mixture.
The reaction mixture was poured slowly into a 10-L extrac-
tor containing 2 L of 3 N HCl, 1 L of ice, and 2.5 L of EtOAc.
The mixture was stirred for 15 min and separated into layers.
The aqueous layer was extracted with 2 × 1.0 L of EtOAc.
The combined organic layers were washed with 4 × 2 L of brine
and concentrated. Addition of 250 mL of toluene crystallized
the trans-3. The solid was filtered and dried at 50 °C to give
85.5 g (64% yield) of trans-3. Mp: 119-120 °C. [R]24D ) -69.78
Sch em e 3
1
(c 0.12, THF). H NMR: δ 7.45-7.28 (m, 5H), 7.28-7.18 (m,
4H), 6.95-6.85 (m, 4H), 5.04 (d, J ) 2.0 Hz, 1H), 5.02 (s, 2H),
4.26-4.16 (m, 1H), 3.75-3.70 (m, 1H), 3.70-3.60 (m, 1H), 3.52
(d, J ) 5.0 Hz, 1H), 3.15 (dd, J ) 5.2, 2.0 Hz, 1H), 2.85 (t, J )
5.3 Hz, 1H). 13C NMR: δ 165.6, 158.9, 160.2 & 157.8 (J ) 244
Hz), 136.5, 133.4 and 133.3 (J ) 2.0 Hz), 129.2, 128.6, 128.0,
127.4, 127.3, 118.6 & 118.5 (J ) 7.8 Hz), 115.9 and 115.8 (J )
22.6 Hz), 115.4, 70.0, 69.3, 64.9, 62.7, 56.6. HRMS: 408.1619
(M+ + H), calcd for C24H23FNO4 408.1611. Anal. Calcd for
mixture must be quenched at low temperature to avoid
any retro-aldol reaction. Without isolation, the reaction
mixture was subjected directly to the dehydration condi-
tions using PTSA and molecular sieves to give enone 7
in 75% isolated yield over two steps.
At this point, two different pathways were examined.
The first began with reduction of the double bond in
enone 7 with Wilkinson’s catalyst to give saturated
ketone 8 in 71% yield. A highly enantioselective reduction
of ketone 8 with borane and CBS reagent gave alcohol 9
with a 98:2 selectivity of S:R at the benzylic position.
Crystallization of this product removed 5% of the SSR
diastereomer arising from cis-3. A catalytic hydrogenative
removal of the benzyl group and crystallization was
sufficient to purge the undesired RRS isomer and provide
the final product 2 in 79% yield and excellent chiral
purity (>99% ee).
Alternatively, a palladium-catalyzed double reduction
in EtOH of both the double bond the benzyl group in
enone 7 produced free phenol 10 in 90% yield. A direct
chiral reduction of 10 with CBS reagent gave only 84%
ee. Therefore, a three-step one-pot procedure was devel-
oped to convert the free phenol 10 to the final product
(Scheme 3). Thus, the phenol was protected in situ as
its trimethylsilyl ether using bis-trimethylsilyl urea
(BSU) followed by a highly selective CBS reduction of
the ketone group to give the required alcohol in 97% ee.
The trimethylsilyl group was removed during the acidic
workup, and the product crystallized to afford 2 in 79%
overall yield from 10 and 99.7% ee.
C
24H23FNO4: C, 70.75; H, 5.44; N, 3.44. Found: C, 70.57; H,
5.56; N, 3.41. IR (Nujol): 3330, 2920, 2870, 1740, 1520 cm-1
.
cis-3. 1H NMR δ 7.44-7.36 (m, 4H), 7.35-7.25 (m, 5H), 7.02-
6.94 (m, 4H), 5.26 (d, J ) 2.0 Hz, 1H), 5.05 (s, 2H), 3.95-3.85
(m, 1H), 3.72 (bs, 2H), 2.51 (t 1H). Also confirmed by NOE.
Anal. Calcd for C24H23FNO4: C, 70.75; H, 5.44; N, 3.44.
Found: C, 70.54; H, 5.47; N, 3.56.
3a . The above general method was followed. The ratio of
trans-3a to cis-3a isomer was 90:10, and the isolated yield
of pure trans-3a was 23%. Mp: 135-136 °C. [R]23 ) -63.8 (c
D
1
0.54, MeOH). H NMR: δ 7.45-7.35 (m, 5H), 7.30 (d, J ) 8.7
Hz, 1H), 7.23 (d, J ) 9.1 Hz, 1H), 6.97 (d, J ) 8.7 Hz, 1H),
6.79 (d, J ) 9.1 Hz, 1H), 5.06 (d, J ) 2.4 Hz, 1H), 5.05 (s, 2H),
4.25 (dt, J ) 3.8, 6.0 Hz, 1H), 3.79 (dd, J ) 3.8, 11.3 Hz, 1H),
3.75 (s, 2H), 3.70 (dd, J ) 6.0, 11.3 Hz, 1H), 3.19 (dd, J ) 2.4,
6.0 Hz, 1H). 13C NMR: δ 164.5, 158.5, 155.6, 136.2, 130.4,
129.2, 128.2, 127.6, 127.1, 126.9, 118.1, 114.9, 113.8, 69.6, 69.2,
64.6, 62.2, 56.3, 54.9. Anal. Calcd for C25H25NO5: C, 71.58; H,
6.01; N, 3.34. Found: C, 72.04; H, 6.48; N, 3.50. IR (Nujol):
3320, 2920. 2870, 1730, 1540 cm-1
.
3b. The above general method was followed. The ratio of
trans to cis was 86:14 in the crude product, and 70% pure
trans-3b was isolated. Mp: 158-160 °C. [R]25D ) -40.1 (c 0.33,
1
MeOH). H NMR: δ 7.50-7.25 (m, 11H), 7.20-6.95 (m, 3H),
5.11 (d, J ) 2.3 Hz, 1H), 5.06 (s, 2H), 4.27 (dt, J ) 3.7, 6.1 Hz,
1H), 3.81 (dd, J ) 3.7, 11.2 Hz, 1H), 3.73 (dd, J ) 6.2, 11.2
Hz, 1H), 2.22 (dd, J ) 2.3 6.1 Hz, 1H). 13C NMR: δ 165.2,
158.4, 136.8, 136.2, 129.2, 128.6, 128.2, 127.6, 127.1, 126.9,
123.6, 116.7, 114.9, 69.6, 69.1, 64.6, 62.3, 56.2. Anal. Calcd
for C24H23NO4: C, 74.02; H, 5.95; N, 3.60. Found: C, 73.68;
In summary, we have discovered and developed an
efficient, high-yielding, one-step process for the diastereo-
and enantioselective formation of trans â-lactams and
applied it successfully to the total synthesis of a novel
cholesterol absorption inhibitor.
H, 6.29; N, 3.67. IR (Nujol): 3380, 2920, 1740, 1525 cm-1
.
3c. The above general method was followed. The ratio of
trans to cis was observed as 99:1 in the crude product, and a
70% isolated yield of trans-3c was obtained. Mp: 204-206 °C.
[R]23D ) -89.6 (c 0.17, MeOH). 1H NMR: δ 7.40-7.20 (m, 10H),
5.14 (d, J ) 2.5 Hz, 1H), 4.27 (dt, J ) 3.8. 6.2 Hz, 1H), 3.81
(dd, J ) 3.8, 11.2 Hz, 1H), 3.74 (dd, J ) 6.2, 11.2 Hz, 1H),
3.22(dd, J ) 2.5, 6.2 Hz, 1H). 13C NMR: δ 165.9, 138.2, 137.0,
128.8, 128.5, 127.6, 126.0, 123.0, 116.1, 67.95, 63.61, 62.11,
54.6. Anal. Calcd for C17H17NO3: C, 72.07; H, 6.05; N, 4.94.
Exp er im en ta l Section
All reactions were carried out under nitrogen. 1H and 13C
NMR spectra (300 and 400 MHz) were recorded in CDCl3 and
are referenced to TMS unless otherwise noted. All starting

Formation of trans-Azetidinones
J . Org. Chem., Vol. 64, No. 10, 1999 3717
Found: C, 71.95; H, 6.34; N, 5.04. IR (Nujol): 3370, 2920, 2870,
1H NMR: δ 8.01 (dd, J ) 8.6, 5.5 Hz, 1H), 7.50-7.25 (m, 7H),
7.25 (m, 6H), 7.18 (m, 2H), 7.22 (d, J ) 8.5 Hz, 1H), 6.98 (t, J
) 8.5 Hz, 1H), 5.08 (s, 2H), 4.88 (d, J ) 2.4 Hz, 1H), 4.04-
3.98 (m, 1H).
P r ep a r a tion of 8. To crude 7 (ca. 80 mmol) dissolved in
120 mL of CH2Cl2 was added 2.2 g (2.4 mmol) of the catalyst,
(Ph3P)3RhCl. The mixture was subjected hydrogenation at 60
psi for 18 h. Concentration of the reaction gave a residue of
the product, which was separated by column with hexane and
EtOAc (90:10) to give 27.5 g (71%) 8 as a solid. 1H NMR: δ
7.98 (dd, J ) 8.5, 5.5 Hz, 1H), 7.41 (m, 5H), 7.25 (m, 4H), 7.12
(t, J ) 8.5 Hz, 2H), 6.55 (m, 4H), 5.04 (s, 2H), 4.68 (d, J ) 2.1,
1H), 3.65 (m, 1H), 3.28 (m, 1H), 3.16 (m, 1H), 2.40 (m, 1H),
2.28 (m, 1H).
CBS Red u ction to 9. The chiral catalyst was made by
following the standard procedure: trimethylboroxine (28 mg,
0.22 mmol) was added into a solution of diphenylprolinol (75
mg, 0.3 mmol) in toluene (5 mL) and the resultant solution
was heated until refluxing. Toluene was distilled, and another
5 mL of toluene was added and distilled out. The residue was
used directly in the following reaction.
1690, 1460 cm-1
.
3d . The above general method was followed. The ratio of
trans to cis was 92:8 in the crude product, and 62% pure trans-
3d was isolated. Mp: 203-205 °C. [R]23.3 ) -82.1 (c 0.38,
D
1
EtOH). H NMR: δ 7.36-7.32 (m, 5H), 7.23 (dd, J ) 4.7, 9.0
Hz, 1H), 6.92 (t, J ) 9.0 Hz, 1H), 5.10 (d, J ) 2.4 Hz, 1H),
4.25 (dt, J ) 3.8, 6.2 Hz, 1H), 3.79 (dd, J ) 3.8, 11.2 Hz, 1H),
3.70 (dd, J ) 6.2, 11.2 Hz, 1H), 3.21 (dd, J ) 2.4, 5.8 Hz, 1H).
Anal. Calcd for C17H16FNO3: C, 67.76; H, 5.35; N, 4.65.
Found: C, 67.74; H, 5.14; N, 4.56. IR (Nujol): 3380, 2920, 2870,
-1
1720, 1500, 1470 cm
.
3e. The above general method was followed. The ratio of
trans to cis isomer was 97:3 in the crude produc, and 51% pure
trans-3e was isolated. Mp: 135-137 °C. [R]24D ) -61.2 (c 0.43
MeOH). 1H NMR: δ 7.33 (dd, J ) 5.2, 8.6 Hz, 1H), 7.30 (dd, J
) 4.7, 9.0 Hz, 1H), 7.03 (t, J ) 8.6 Hz, 1H), 6.92 (t, J ) 9.0
Hz, 1H), 5.10 (d, J ) 2.4 Hz, 1H), 4.23 (dt, J ) 3.6, 5.8 Hz,
1H), 3.76 (dd, J ) 3.6, 11.3 Hz, 1H), 3.66 (dd, J ) 5.8, 11.3
Hz, 1H), 3.17 (dd, J ) 2.4, 5.8 Hz, 1H). Anal. Calcd for
C
17H15F2NO3: C, 63.95; H, 4.73; N,4.39. Found: C, 63.83; H,
5.03; N, 4.40. IR (Nujol): 3380, 2920, 1725, 1600, 1470 cm-1
.
To a 250 mL oven-dried flask with a magnetic stirrer were
added 6.2 g (12.5 mmol) of 8 and 60 mL of CH2Cl2. To the
resulting solution at -20 °C were added sequentially 0.1 equiv
of the chiral catalyst and 6.3 mL (12.5 mmol) of 2.0 N BH3‚
Me2S over 2 h. The reaction mixture was allowed to warm to
0 °C for 1 h and quenched with methanol. The quenched
solution was concentrated and extracted with CH2Cl2. The
organic layer was concentrated, and the residue was recrystal-
lized from EtOAc and hexanes to give 4.1 g (70%) of product
9. The ee was determined by chiral HPLC. Mp: 132-134 °C.
1H NMR: δ 7.45-7.20 (m, 11H), 7.05-6.90 (m, 6H), 5.05 (s,
2H), 4.72-4.69 (m, 1H), 4.57 (d, J ) 2.2 Hz, 1H), 3.07 (dd, J
) 7.6, 2.2 Hz, 1H), 2.36 (bs, 1H), 2.05-1.85 (m, 4H). 13C
NMR: δ 167.6, 163.4, 160.9, 160.1, 159.1, 157.8, 140.1, 140.0,
136.6, 133.9, 133.8, 129.6, 128.6, 128.1, 127.44, 127.40, 127.3,
127.1, 118.4, 118.3, 115.9, 115.6, 115.5, 115.4, 115.2, 73.0, 70.1,
61.1, 60.3, 36.6, 25.0. Anal. Calcd for C31H27F2NO3: C, 74.53;
H, 5.45; N, 2.80. Found: 74.10; H, 5.43; N, 2.98. IR: 3490,
P r ep a r a tion of 5. To a 2-L three-necked flask equipped
with a mechanical stirrer, thermometer, and addition funnel
were added sequentially 100 g (0.25 mmol) of trans-3 and 800
mL of CH3CN. To the cooled mixture at 10 °C was added
dropwise over a period of 20 min 63 g (0.30 mmol) of NaIO4
dissolved in 800 mL of water while maintaining the temper-
ature below 20 °C. The reaction mixture was allowed to warm
to rt and stirred for 2 h as followed by NMR. The reaction
was poured into a 6-L extractor containing 1.5 L of ice-brine
and 1.5 L of toluene. The layers were stirred and separated,
and the aqueous layer was extracted with 500 mL of toluene.
The combined organic layer was washed with 2 × 500 mL
brine and concentrated to about 500 mL for the next reaction.
The aldehyde is not very stable and gets hydrated quickly. An
analytical sample can be obtained by concentrating the extract
to dryness. [R]20 ) +45° (c 1.9, CHCl3). Mp: 76-78 °C.
D
HRMS: 376.1348 (M + H+), calcd for C23H19FNO3 376.1349.
1H NMR: δ 9.82 (d, J ) 1.3 Hz, 1H), 7.38-7.25 (m, 5H), 7.25-
7.15 (m, 4H), 6.95-6.82 (m, 4H), 5.32 (d, J ) 2.4 Hz, 1H), 4.98
2920, 2850, 1720 cm-1
.
Deben zyla tion to 2. To a flask were added 1 g of Pd-C
(5% by w/w), 11.4 g (181 mmol) of ammonium formate, 18.1 g
(36.3 mmol) of 9, and 250 mL of MeOH carefully at rt under
N2. HOAc was added to adjust the pH to 3-5, and the
resultant mixture was heated at 45-55 °C until the reaction
was finished as determined by TLC (about 2 to 3 h). During
the reaction, the pH was controlled in the range of 3-5 by
adding HOAc. The reaction was filtered and the solvent
evaporated. The residue was dissolved in t-BuOMe and washed
with water. After drying over Na2SO4 and evaporation the
solvent, the product was purified by recrystallization in
t-BuOMe/heptane and MeOH/water to give 11.75 g (79%)
product with 99.4% ee. Mp: 155-157 °C. [R]24D ) -32.6 (c 0.34,
MeOH). 1H NMR: (DMSO-d6) δ 9.54 (s, 1H), 7.32 (dd, J )
8.3, 5.7 Hz, 2H), 7.21 (m, H), 7.35 (m, 4H), 6.77 (d, J ) 8.3 Hz,
2H), 5.3 (d, J ) 4.6 Hz, 1H), 4.82 (d, J ) 2.1 Hz, 1H), 4.50 (m,
1H), 3.10 (m, 1H), 1.70-1.90 (m, 4H). 13C NMR: (DMSO-d6)
δ 167.4, 162.3 & 159.3 (J ) 304.8 Hz), 159.9 and 156.9 (J )
303.3 Hz), 157.5, 142.3 and 142.3 (J ) 2.9 Hz), 134.1 and 134.0
(J ) 2.3 Hz), 128.0, 127.7, 127.7 and 127.6 (J ) 7.1 Hz), 118.4
and 118.3 (J ) 7.9 Hz), 116.0 and 115.8 (J ) 22 Hz), 115.8,
114.9 and 114.7 (J ) 21 Hz), 71.2, 59.7, 59.5, 36.5, 24.6. Anal.
Calcd for C24H21F2NO3: C, 70.40; H, 5.14; N, 3.42; F 9.28.
Found: C, 70.14; H, 5.21; N, 3.53; F 9.34. IR (Nujol): 3260,
(s, 2H), 4.15 (dd, J ) 2.4, 1.3 Hz, 1H). Anal. Calcd for C23H18
FNO3‚1/2H2O: C, 71.86; H, 4.98; H, 3.64. Found: C, 72.04; H,
5.06; N, 3.74. IR: 2920, 1745, 1730, 1620, 1520 cm-1
-
.
P r ep a r a tion of 7. To a 1-L three-necked flask equipped
with a mechanical stirrer, thermometer, and addition funnel
were added at rt a solution of 100 g (0.27 mmol) of aldehyde
6 in 500 mL of toluene and 32 mL (0.27 mmol) of BF3 etherate.
To the cooled mixture at -30 °C was added dropwise 56 g (0.27
mmol) of enol ether. After stirring at -30 °C for 5 min, the
aldol mixture was added dropwise into an ice-cold quench
solution containing 1 L of saturated NaHCO3, 2 L of t-BuOMe,
and 150 mL of hydrogen peroxide (30%). The resulting mixture
was allowed to warm to 15 to 20 °C, and the layers were
separated. The aqueous layer was extracted with 1 L of
toluene. The combined organic layer was washed with 2 × 500
mL of brine and concentrated to about 1 L for dehydration.
An analytical sample was purified on a silica gel column
eluting with EtOAc/hexanes (1:1) to give a diastereomeric
1
mixture of 6 as a solid. H NMR: δ 8.0-7.90 (m, 2H), 7.40-
7.15 (m, 9H), 7.15 (t, J ) 8.5 Hz, 2H), 7.02-6.90 (m 4H), 5.05
(d, J ) 2.4 Hz, 1H), 5.00 (s, 2H), 4.65-4.55 (m, 1H), 3.65-
3.57 (m, 1H), 3.42 (dd, J ) 17.8, 2.4 Hz, 1H), 3.20-3.05 (m,
2H). Anal. Calcd for C31H25F2NO3: C, 72.50; H, 4.91; N, 2.73.
Found: C, 72.32; H, 5.24; N, 2.80.
2920, 1860, 1715, 1510 cm-1
.
To the above 1 L toluene solution of aldol product were
added 200 g of molecular sieves and 25 g (0.13 mmol) of
p-toluenesulfonic acid monohydrate. This mixture was heated
to 40-50 °C and monitored by proton NMR for about 2-4 h.
The reaction was cooled to 0 °C and filtered through a pad of
MgSO4 and 100 g silica gel. The filtrate was concentrated for
the next step. Alternatively, the concentrated solution was
added to 400 mL of heptane to precipitate enone 7 (99 g, 75%
overall yields). The compound is not stable on silica gel.
HRMS: 496.1706 (M + H+), calcd for C31H24F2NO3 496.1724.
P r ep a r a tion of 10. To a 1-L Parr pressure bottle were
added 0.8 g of palladium on carbon (10%), 1.6 g (19.0 mmol)
of NaHCO3, 16 g (32.3 mmol) of compound 6 in 80 mL of
EtOAc, and 80 mL of methanol. The bottle was shaken under
30 psi of hydrogen pressure for 2 to 3 h and followed by TLC
and HPLC. The reaction mixture was filtered through a pad
of Celite and washed with 200 mL of toluene. The filtrate was
washed with 200 mL of brine and 2 mL of 3 N HCl. After
separation of the layers, the organic layer was concentrated
to give 11.8 g (90% yield) of 10. Mp: 60-62 °C. 1H NMR: δ

3718 J . Org. Chem., Vol. 64, No. 10, 1999
Wu et al.
7.95 (dd, J ) 8.6, 5.5 Hz, 2H), 7.13-7.22 (m, 4H), 7.09 (t, J )
8.6 Hz, 2H), 6.91 (t, J ) 8.6 Hz, 2H), 6.80 (d, J ) 8.6 Hz, 2H),
4.65 (d, J ) 2.1 Hz, 1H), 3.26 (m, 1H), 2.33 (d, 1H), 2.25 (m,
1H)., 13C NMR: δ 197.8, 167.7, 167.7 and 164.3 (J ) 256 Hz),
160.7 and 157.5 (J ) 243 Hz), 156.3, 133.8, 133.0, 130.9 and
130.8 (J ) 9.4 Hz), 129.0, 127.5, 118.6 and 118.5 (J ) 7.7 Hz),
116.2, 116.1 and 115.8 (J ) 23 Hz), 116.0 and 115.7 (J ) 22
Hz), 61.3, 59.7, 35.6, 23.3. Anal. Calcd for C24H19NF2O3‚1/
2H2O: C, 69.23; H, 4.80; N, 3.36; F, 9.13; O, 13.45. Found: C,
were added water and t-BuOMe. The layers were separated.
Concentration of t-BuOMe layer lead to a recovery of >50%
catalyst as the HCl salt after filtration. Crystallization of crude
product from 2-propanol/H2O afforded 1.9 g (79%) of 2 with
>99.4% ee as determined by HPLC on a Chiralcel ODH
column.
Ack n ow led gm en t. We thank Dr. Y. Lu and Mr. M.
Poirier for their help in some of the steps. We thank
Drs. G. Love, M. Mitchell, D. Schumacher, and D.
Walker and Mr. B. Shutts for their suggestions and
proof-reading of this manuscript.
69.75; H, 4.47; N, 2.95; F, 9.11; O, 13.44. IR (Nujol): 3370,
-1
2920, 2860, 1735, 1720, 1680, 1520 cm
.
On e-P ot Rea ction to 2. To a 50 mL oven-dried flask with
a magnetic stirrer were added 2.4 g (5.9 mmol) of compound
10, 10 mL of CH2Cl2, and 0.62 g (3.0 mmol) of bistrimethylsilyl
urea (BSU). After 0.5 h, the reaction was filtered directly into
another 50 mL oven-dried flask containing 0.05 equiv of the
chiral catalyst at -20 °C. To this was added 2.3 mL (4.7 mmol)
of 2 N BH3‚Me2S. The reaction mixture was stirred at -15 to
-20 °C for 3-5 h as monitored by TLC and HPLC. The
reaction was quenched with 10 mL of methanol/HCl, and the
resulting solution was concentrated to a residue. To the residue
Su p p or tin g In for m a tion Ava ila ble: 1H NMR, 13C NMR,
and IR spectra for 2, 3, 3a , 3b, 3c, 3d , 3e, 5, 7, 8, 9, and 10,
also chiral HPLC chromatogram for 2. This material is
available free of charge via the Internet at http://pubs.acs.org.
J O990428K