A R T I C L E S
Tellers et al.
overnight age, the vessel was vented to atmospheric pressure, and the
reaction solutions were analyzed by HPLC or SFC.
the two insertion steps. We postulate that rearrangement to the
geometry necessary for insertion is more facile for the â,γ-
substituted olefin 1-Endo, as it is not locked in as rigid a
conformation as R,â-substituted olefin 1-E. The difference in
enantioselectivity can be attributed to the fact that 1-Endo can
coordinate to the metal by both the carboxylate and olefin
groups. The ruthenium-1-Endo intermediate is analogous to
the rhodium-acetamidocinnamate described by Landis and
Halpern.44 This binding mode should offer more stereo-
differentation than the ruthenium-1-E intermediate, where only
the carboxylate group can bind to the metal.57,66,67
Crystallization of 1,1,3,3-Tetramethylguanidinium (2E)-[4-(4-
Chlorobenzyl)-7-fluoro-5-(methylsulfonyl)-1,4-dihydrocyclopenta-
[b]indol-3(2H)-ylidene]acetate (1-E‚TMG). 1,1,3,3-Tetramethylguani-
dine (TMG, 99%, 25.0 mL, 193 mmol) was charged to a slurry of
compound 1-E (70 g, 160 mmol) in isopropylacetate (IPAc, 280 mL)
and MeOH (140 mL) at 23 °C, whereupon the solids dissolved.
Additional IPAc (125 mL) was added with previously generated seed,
resulting in the formation of a slurry. Additional IPAc (125 mL) was
added and the mixture stirred overnight. The slurry was distilled at
constant volume by the simultaneous addition of IPAc. When the
supernatant reached 5 mg/mL, distillation was discontinued and the
slurry cooled in an ice bath. The crystals were filtered and washed
with ice-cold IPAc. Drying yielded 86.0 g of crystalline solid of 96.8
wt % purity. 1H NMR (400 MHz, CD3OD): δ 6.87 (dd, 1H, JHH ) 9.6
Hz, JHF ) 2.7 Hz), 6.76 (dd, 1H, JHH ) 7.9 Hz, JHF ) 2.7 Hz), 6.43 (d,
2H, JHH ) 8.5 Hz), 6.06 (d, 2H, JHH ) 8.5 Hz), 5.36 (s, 2H), 5.21 (t,
JHH ) 2.2 Hz), 2.79 (bm, 2H, JHH ) 3.2 Hz), 2.16 (s, 3H), 2.14 (s,
15H), 2.10 (bm, 2H). 13C{1H} NMR (101 MHz, CD3OD): δ 175.9
(s), 163.3 (s), 157.1 (d, J ) 239 Hz), 148.8 (s), 145.5 (s), 138.9 (s),
136.4 (s), 134.0 (s), 133.5 (d, J ) 5 Hz), 130.0 (s), 129.8 (d, J ) 8.8
Hz), 128.6 (s), 127.9 (d, J ) 7 Hz), 118.1 (s), 115.6 (s), 115.3 (s),
112.5 (s), 112.3 (s), 51.1 (s), 45.1 (s), 40.1 (s), 37.0 (s), 23.4 (s).
Summary and Conclusions
This work describes the implementation of an asymmetric
hydrogenation of a pharmaceutically relevant compound. It is
clear from this example that, despite the increased substrate
complexity, asymmetric hydrogenation can be a practical
industrial tool. Surprisingly, this work represents a rare example
of a mechanistic study on an asymmetric hydrogenation of
industrial relevance.54 It is exactly a result of the increased
substrate complexity that new mechanistic challenges and
opportunities, not available with structurally less diverse
substrates, become available.
(2E)-[5-Bromo-4-(4-chlorobenzyl)-7-fluoro-1,4-dihydrocyclopenta-
[b]indol-3(2H)-ylidene]acetic Acid (1-Br-E). To a solution of [4-(4-
chlorobenzyl)-7-fluoro-5-bromo-1,2,3,4-tetrahydrocyclopenta[b]indol-
3-yl]acetic acid (2-Br, 44.0 g, 0.10 mol) in toluene (450 mL) was added
N-chlorosuccinimide (16.0 g, 0.12 mol) portionwise over 5 min. The
resulting solution was stirred at room temperature for 1 h, after which
acetic acid (11.5 mL, 0.20 mol) was added. The reaction mixture was
stirred for an additional 2 h at 22 °C to crystallize the product. The
slurry was filtered, rinsed with toluene (150 mL), and dried at 40 °C
under vacuum for 24 h to give 40.4 g of the ene acid (93%) as the
pure E isomer. Due to its limited solubility, compound 1-Br-E was
characterized spectroscopically as the TMG salt, 1-Br-E‚TMG. 1H
NMR (400 MHz, CD3OD): δ 7.25 (d, 2H, JHH ) 8.4 Hz), 7.17 (dd,
Experimental Section
Materials. Unless otherwise noted, reagents were purchased from
commercial suppliers and used without further purification. All NMR
spectra were recorded at room temperature unless otherwise noted.
Compounds 1-E,26 2,16 2-Br,18 and Ru(S-BINAP)(OAc)2 were pre-
68
pared according to literature procedures. Resealable pressure vessels
were purchased from Andrews Glass Co., and the Multimax IR
instrument was purchased from Mettler-Toledo.
Preparation of Catalyst 3. A modified version of a literature
procedure was followed.32 In an inert atmosphere glovebox, degassed
methanol (75 mL) was charged to a round-bottom flask containing [(p-
cymene)RuCl2]2 (0.37 g, 0.60 mmol), (S)-BINAP (0.77 g, 1.2 mmol),
and a stir bar. Degassed toluene (25 mL) was added, and the orange
heterogeneous solution was transferred to an ampule with a resealable
Kontes adapter. The ampule was sealed, removed from the glovebox,
and heated for 1-2 h at 50-60 °C with stirring. The clear, orange
solution was brought into a glovebox and stored at room temperature
as a stock solution (approximate molarity ) 0.012).
1H, JHH ) 8.4 Hz, JHH ) 2.0 Hz), 7.15 (dd, 1H, JHH ) 8.4 Hz, JHH
)
2 Hz), 6.96 (d, 2H, JHH ) 8.4 Hz), 5.98 (s, 1H), 5.89 (s, 1H), 3.60 (bs,
2H), 2.95 (s, 12H), 2.79 (t, 2H, JHH ) 5 Hz). 13C{1H} NMR (101 MHz,
CD3OD): δ 174.8 (s), 161.8 (s), 158.0 (s), 155.6 (s), 146.3 (s), 144.7
(s), 137.7 (s), 136.2 (s), 132.4 (s), 130.8 (d, JCF ) 4.8 Hz), 128.3 (s),
127.2 (s), 126.5 (d, JCF ) 10.4 Hz), 116.3 (d, JCF ) 29 Hz), 115.6 (s),
103.8 (d, JCF ) 23 Hz), 102.7 (d, JCF ) 12 Hz), 38.5 (s), 35.6 (s), 22.0
(s). HPLC/MS: m/z for [C20H14BrClFNO2]+ (M + H+) calcd 433.9959,
obsd 433.9957.
Typical Catalyst Screening Protocol. Solid substrate (0.023 mmol)
was weighed into an HPLC vial (2 mL) containing a stir bar. A total
of 14 HPLC vials containing substrate were brought into a glovebox
and charged with degassed methanol (180 µL) and triethylamine (0.023
mmol). To the vials was added the appropriate catalyst solution (0.0023
mmol in 20 µL of MeOH/toluene). The HPLC vials were fitted with
perforated septa caps and transferred to a resealable pressure vessel
(purchased from Andrews Glass Co.) containing sand. The sand was
used as the heat-transfer medium and as a means of preventing the
vials from tipping over. The vessel was sealed, removed from the
glovebox, placed in an oil bath, and attached to a hydrogen/N2/vacuum
manifold. After the solutions were set to stirring and the oil bath was
warmed to 50 °C, the vessel was pressurized to 105 psi with N2 and
then vented to atmospheric pressure. This was repeated three times,
after which the vessel was pressurized with H2 (105 psi). After an
[4-(4-Chlorobenzyl)-7-fluoro-5-(methylsulfonyl)-1,4-dihydro-
cyclopenta[b]indol-3-yl]acetic Acid (1-Endo). To a stirred slurry of
1-E (50.0 g, 0.12 mol) in MeOH (400 mL) was added TMG (13.0 g,
0.12 mol). A stream of nitrogen gas was bubbled into the homogeneous
red solution for 15 min to remove residual oxygen, upon which Ru(S-
BINAP)(p-cymene)Cl2 (120 mL, 0.012 M, 1.2 mol %) was added. The
solution was warmed to 50 °C and stirred for 6 h. After the solution
cooled to room temperature, an isopropyl acetate (25 mL) solution of
acetic acid (7.3 mL, 1.1 equiv, 0.13 mol) was added over 45 min via
addition funnel in order to precipitate 1-E. The solution was stirred an
additional hour and then filtered through a frit. The solvent was removed
under reduced pressure, and the residual oil was redissolved in EtOAc
(250 mL). This solution was washed with 0.1 N HCl (175 mL), water
(150 mL), and brine (100 mL) and then transferred to a round-bottom
flask containing Darco (3 g of Darco KB, 3 g of Darco G-60) and
silica gel (3 g). The Darco and silica gel treatment was employed in
order to remove residual ruthenium. After this slurry was agitated for
30 min, the solution was filtered through a medium frit and the solvent
removed under reduced pressure. To the resulting foam was added
EtOAc (50 mL). Upon this solution being cooled to 0 °C, a white solid
(66) Ashby, M. T.; Khan, M. A.; Halpern, J. Organometallics 1991, 10, 2011-
2015.
(67) A reviewer is acknowledged for suggesting this two-point binding mech-
anism to rationalize the higher enantioselectivity.
(68) Kitamura, M.; Tokunaga, M.; Noyori, R. J. Org. Chem. 1992, 57, 4053-
4054.
9
17072 J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006