Substituted isoquinolines by Noyori transfer hydrogenation: Enantioselective synthesis of chiral diamines containing an aniline subunit

Article abstract of DOI:10.1021/jo990594s

Transfer hydrogenation using the Noyori catalyst 5-Ts is effective for the enantioselective hydrogenation of imines containing fully substituted nitrogen groups (12 or 13). Analogues such as 11c could not be reduced in practical yield, apparently due to product inhibition of the catalyst. Asymmetric transfer hydrogenation of the aniline imine 8a was possible, but required impractical purity levels for the substrate, and the nitro analogue 7 could not be reduced efficiently. The best results were obtained with the bromophenyl imine 20. In the case of 20b, the product 21b was formed with 98.7% ee, and the material could be upgraded to >99% ee by crystallization of the hydrochloride salt. Reaction of 21b with NH₃ or MeNH₂ in the presence of Cu/CuCl gave the chiral anilines 10b or 23b. The latter substance is comparable to the commercially available 1 as a chiral proton donor for amide enolates and provides access to the hitherto unavailable enantiomeric series.

Full text of DOI:10.1021/jo990594s

6724
J . Org. Chem. 1999, 64, 6724-6729
Su bstitu ted Isoqu in olin es by Noyor i Tr a n sfer Hyd r ogen a tion :
En a n tioselective Syn th esis of Ch ir a l Dia m in es Con ta in in g a n
An ilin e Su bu n it
E. Vedejs*
University of Wisconsin, Madison, Wisconsin 53706
P. Trapencieris and E. Suna
Organic Synthesis Institute, Riga, Latvia LV 1006
Received April 6, 1999
Transfer hydrogenation using the Noyori catalyst 5-Ts is effective for the enantioselective
hydrogenation of imines containing fully substituted nitrogen groups (12 or 13). Analogues such
as 11c could not be reduced in practical yield, apparently due to product inhibition of the catalyst.
Asymmetric transfer hydrogenation of the aniline imine 8a was possible, but required impractical
purity levels for the substrate, and the nitro analogue 7 could not be reduced efficiently. The best
results were obtained with the bromophenyl imine 20. In the case of 20b, the product 21b was
formed with 98.7% ee, and the material could be upgraded to >99% ee by crystallization of the
hydrochloride salt. Reaction of 21b with NH3 or MeNH2 in the presence of Cu/CuCl gave the chiral
anilines 10b or 23b. The latter substance is comparable to the commercially available 1 as a chiral
proton donor for amide enolates and provides access to the hitherto unavailable enantiomeric series.
Sch em e 1
As part of a study involving the use of chiral aniline 1
as an asymmetric proton donor, we have been interested
in developing an efficient enantioselective synthesis of
o-aminoarylisoquinolines 2.1 In particular, we hoped to
prepare the unsubstituted analogue 2 with R ) H. This
substance should provide easy access to a variety of
N-substituted analogues of 1, chiral anilines that could
be used to probe structural and pKa effects in the
asymmetric protonation of enolates. An efficient route to
isoquinoline derivatives related to 2 has recently been
described by Noyori et al. via the asymmetric transfer
hydrogenation of imines such as 3.2 The reported conver-
sion from 3 to 4 was catalyzed by the ruthenium complex
5-Np s (Ar ) 1-naphthyl) and proceeded in good yield and
with promising enantioselectivity (84% ee). The corre-
sponding process for our purposes would have to be
performed with ortho-substituted imines 7 or 8 (X ) NO2
or NH2) or with analogous imines where X might be a
leaving group that can be converted into nitrogen func-
tionality. Ideally, asymmetric hydrogenation would be
performed with substrates having Z ) H. However, Z )
OCH3 should not interfere in the intended asymmetric
protonation applications using derivatives of 10b, and
this substitution pattern may be necessary for high
enantioselectivity in the Noyori reductions. Thus, the
dimethoxy series (b series; Z) CH3O, Scheme 1) as well
as the parent system (a series; Z ) H) was selected for
the initial attempts to prepare analogues of 1.
The most direct route to 10 would be via reduction of
the imines 7 or 8. The nitro imines 7 were easily prepared
by standard Bischler-Napieralski cyclization from 6,3
and reduction with Fe/HCl gave the imino anilines 8a
and 8b (Scheme 1; 68 and 86% yield, respectively).
However, the product anilines were contaminated with
an intensely yellow impurity that could not be removed
by chromatography.4 With impurity levels near 20%,
transfer hydrogenation did not occur within 16 h at 20
°C. Repeated crystallizations reduced the amount of the
contaminant to 2%, but with considerable material loss.
When 8a containing ca. 2% of the impurity was treated
(1) (a) Vedejs, E.; Lee, N.; Sakata, S. T. J . Am. Chem. Soc. 1994,
116, 2175. (b) Vedejs, E.; Kruger, A. W. J . Org. Chem. 1998, 63, 2792.
(2) Uematsu, N.; Fujii, A.; Hashiguchi, S.; Ikariya, T.; Noyori, R. J .
Am. Chem. Soc. 1996, 118, 4916. Hashiguchi, S.; Noyori, R. Acc. Chem.
Res. 1997, 30, 97.
(3) Ott, H.; Hardtmann, G.; Denzer, M.; Frey, A. J .; Gogerty, J . H.;
Leslie, G. H.; Trapold, J . H. J . Med. Chem. 1968, 11, 777.
10.1021/jo990594s CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/18/1999

Substituted Isoquinolines by Noyori Transfer Hydrogenation
J . Org. Chem., Vol. 64, No. 18, 1999 6725
Sch em e 2
with 1 mol % of the catalyst 5-Np s and HCO2H/NEt3,
reduction did take place and the expected diamine 10a
was obtained in 75% yield and with 71% ee. In another
attempt using a different batch of catalyst, the ee value
mysteriously increased to 83% ee, and similar inconsis-
tencies were encountered in attempts to hydrogenate 3.
No substantial difference between the two batches of
5-Np s could be detected using NMR methods, but the
limited solubility of both 5-Np s and its precursor (C6H6)2-
RuCl25 complicates catalyst assay using spectroscopy. We
made no attempt to define the source of the problem in
catalyst preparation because relatively consistent results
were obtained with a modified Noyori catalyst 5-Ts (Ar
) MeC6H4).6 This catalyst reduced purified 8a (2%
contaminant) in 66% yield (85% ee). However, if the
starting aniline 8a contained higher levels of the yellow
contaminant, then conversion decreased, suggesting that
the impurity inhibits the catalyst. Since the yellow
contaminant was difficult to remove and the enantiose-
lectivity was modest, attention was turned to alternative
approaches as described below. Most of the subsequent
studies were performed using 5-Ts as the catalyst for
asymmetric transfer hydrogenation.
One desirable option is to perform the asymmetric
reduction at the nitro imine stage (7) in the hope that
the nitro group might survive treatment with HCO2H/
Et3N and catalyst 5-Ts. Unfortunately, this variation also
encountered poor conversion to products. At the 1 mol %
catalyst loading level, 7a was not reduced at a useful rate,
while 7b afforded ca. 20% of 9b after 13 h. Increased
conversion resulted using higher catalyst loading, but
practical yields were not obtained. On the other hand,
the enantioselectivity measured for the product 9b was
excellent (97.1% ee). This result did not solve the
preparative problem, but high ee was encouraging be-
cause it indicated that the nature of the ortho substituent
is important and might be used to control the asymmetric
hydrogenation. Other ortho-substituted systems were
therefore explored (Scheme 2). The initial focus was on
the dimethoxy series, corresponding to the b series of
Scheme 1, and the goal was to identify a dominant ortho
substituent that would allow sufficient control over the
asymmetric hydrogenation to afford good enantioselec-
tivity with or without the methoxy substituents.
ee was obtained in the reduction step, but conversion
remained problematic.
Since catalyst inhibition had been a recurring theme
in the reductions, we suspected that the problem might
be due to the sulfonamide N-H group. Blocking this site
would minimize the risk that the product amino sulfona-
mide might act as a bidentate ligand for ruthenium.
Accordingly, 11c was converted into the N-methoxy-
methyl or N-benzyl derivatives 12c and 13c. The reduc-
tions were still rather slow and required 3-4 days using
6-7.5% of the catalyst 5 for 60-75% conversion to
products 14c or 15c. In both cases, excellent enantiose-
lectivity of 98% ee or better was obtained.
After considerable effort, 15c could not be debenzylated
without at least some (4-15%) racemization (palladium-
catalyzed hydrogenolysis conditions). On the other hand,
treatment of 14c with dilute HCl produced the desired
16c together with a byproduct (tentatively, 17) that was
resistant to hydrolysis. This problem could be avoided
by converting the crude reduction product to the N-
trifluoroacetamide 18c, hydrolyzing the latter to 19c, and
cleaving the trifluoroacetyl group using K2CO3/H2O-
MeOH. The sequence succeeded to the extent that a
family of potentially useful, chiral aniline derivatives
16c-e having enhanced N-H acidity became available
from the corresponding imines 12c-e with high enan-
tiomeric purity (>99% ee after crystallization). However,
the number of steps necessitated by the problems with
catalyst inhibition, as well as the complications with the
protecting group chemistry, combined to make this
approach relatively laborious.
Conversion of 8b into the corresponding N-tosyl de-
rivative proceeded smoothly in pyridine with TsCl to give
11c in 89% yield (Scheme 2). However, attempted asym-
metric hydrogenation again was frustrated by poor
conversion. Thus, 11c gave no product within 16 h using
2 mol % of catalyst 5-Ts and only 11% conversion after
72 h with 7.5 mol % of catalyst. As before, excellent g96%
(4) The yellow contaminant could not be separated well enough to
give satisfactory analysis or spectra, but one small sample was obtained
having a 3:2 ratio of oxygen/nitrogen according to elemental analysis,
consistent with structure i, below. Analogous indazole N-oxides are
reported as yellow solids: Reissert, A.; Lemmer, F. Chem. Ber. 1926,
59, 351. Behr, L. C. J . Am. Chem. Soc. 1954, 76, 3672. Behr, L. C.;
Alley, E. G.; Levand, O. J . Org. Chem. 1962, 27, 65.
One additional series of imine derivatives 20 (o-
bromophenyl series) was examined, and this eventually
provided the best route to diamines 10 (Scheme 3).
Asymmetric hydrogenation of 20b proceeded without any
(5) Bennett, M. A.; Smith, A. K. J . Chem. Soc., Dalton Trans. 1 1974,
233.
(6) Noyori, R.; Ikariya, T. Personal communication.

6726 J . Org. Chem., Vol. 64, No. 18, 1999
Vedejs et al.
Sch em e 3
Problems were also encountered with conversion of the
imine substrate and yield of the products. Some of the
early hydrogenations gave <25% conversion, but this
difficulty was overcome by the simple expedient of
venting the reaction vessel. Evidently, it is important to
provide an exit for the gases (H2 + CO2) produced by
decomposition of the HCO2H/Et3N reducing reagent as
the hydrogenation proceeds. This version of catalyst
inhibition is a minor problem compared to the examples
discussed earlier where reduction products or impurities
appear to act as catalyst inhibitors.
One complication was encountered to varying degrees
in all of the hydrogenations, increasingly so for the slower
reductions. This involved conversion of the product into
the N-formyl derivative by the Et3N/HCO2H reagent. In
the case of 20b, use of 1 mol % of catalyst 5-Ts resulted
in 8% of 22b at 92% conversion (84% of 21b), and the
amount of 22b increased with time. In a similar experi-
ment using 0.67 mol % catalyst, the reaction was slower
(24 h vs 13 h) and 22b was formed in 19% yield (97%
conversion). Because of this problem, we were unable to
obtain yields much above 70% in a number of experi-
ments, including a preparative scale (18 g) hydrogenation
of 20b. However, the initially formed bromophenyl iso-
quinoline 21b was conveniently isolated and purified as
the hydrochloride salt, 72% from 20b after crystallization
(1.7 g scale; 62% yield on 18 g scale). When 21b was
released from the hydrochloride with aqueous base,
>99% ee was measured for the upgraded material. Thus,
the asymmetric hydrogenation and subsequent crystal-
lization provides material with excellent enantiomeric
purity and in reasonable yield.
With practical access to >99% enantiomerically pure
21b established, the problem of replacing bromide by an
amino group was investigated. This proved to be rela-
tively easy by an adaptation of the method published in
1968 by Ott et al.3 Thus, 21b was treated with liquid NH3
in the presence of copper powder and CuCl in a Parr
reactor, 70 °C, for 5 days. The product 10b was obtained
in 82% yield after crystallization, and with 99% ee. In a
similar process, 21b reacted with CH3NH2/Cu/CuCl to
afford 23b, 80% isolated after crystallization, 99% ee.
of the complications encountered with the o-amino de-
rivatives. Catalyst loading of 0.67 mol % gave 21b with
reasonable conversions of 67% and with excellent 98.7%
ee. When the same conditions were tested with 20a , the
reaction was marginally slower and less selective (94%
ee). However, practical results were realized in each case.
Thus, the o-bromophenyl group provides sufficient control
in the asymmetric hydrogenation step so that useful ee
levels can be achieved with or without the methoxy
substituents.
The asymmetric hydrogenation-amination sequence
provided sufficient 10b to allow reinvestigation of the
sulfonamide derivatives. After secondary amine nitrogen
was protected as the Cbz derivative 25, N-sulfonylation
could be carried out without complications, and depro-
tection (H2/Pd-C) gave sulfonamides 16c-e. The abso-
lute configuration of 16c prepared starting from 20b was
the same as for the material derived from 12c. Similarly,
10b prepared by either route had the same configuration.
The absolute configuration was confirmed by X-ray
crystallography in the case of the crystalline N-formyl
byproduct 22b (anomalous dispersion method). Thus, all
of the hydrogenations correspond to Noyori’s model for
asymmetric induction.2
The asymmetric reductions could be performed as
reported by Noyori et al,2 but several variables deserve
comment, starting with the procedure for catalyst prepa-
5
ration. The method is relatively simple, (C6H6)2RuCl2
+
H2NCHPhCHPhNHSO2Ar; Et3N/i-PrOH, reflux, but the
catalyst is reported to decompose upon attempted puri-
fication. These properties caused some concern when one
batch of 5-Np s was found to give inferior enantioselec-
tivities in some of our early attempts. However, the same
batch of 5-Np s catalyzed the transfer hydrogenation of
20a or 20b with excellent enantioselectivities (20a : 59%
yield, 94% ee; 20b: 71% yield, 98.3% ee). Both catalysts
5-Ts and 5-Np s are easy to store and to handle. Consis-
tent results were obtained using a given batch of catalyst
over a period of several weeks, so the hydrogenation of
the best substrates (o-bromophenyl imines 20) does not
appear to be capricious.
Finally, the diamine 23b was tested in the asymmetric
protonation of the amide enolate 25. The reaction proved
somewhat more sensitive to temperature changes com-
pared to the analogous process using 1. However, addi-
tion of precooled 23b to the enolate gave reproducible
results, and this small modification of the usual pro-
cedure1a gave 26 with 88.8% ee. The corresponding
experiment using the dechloro analogue of 1 results in

Substituted Isoquinolines by Noyori Transfer Hydrogenation
J . Org. Chem., Vol. 64, No. 18, 1999 6727
Ta ble 1. Asym m etr ic Tr a n sfer Hyd r ogen a tion of Im in esa
the sensitivity to product inhibition. Substrates chosen
to minimize this problem (for example, 14, 15, 20) could
be reduced in acceptable yield, although some material
was lost to N-formylation in all cases. However, this is a
small complication in view of the convenience of using
Et3N/HCO2H as the reducing agent. The Noyori proce-
dure is relatively easy, and no special apparatus is
required. In the case of 20b, scale-up to 18 g was carried
out with a 10% loss in yield, but without encountering
other complications. While the method does not tolerate
a broad range of ortho substituents, it works quite well
with the versatile o-Br substrate.10 Combined with the
copper-catalyzed aminations,11 this approach is currently
the method of choice for the synthesis of structures
related to 1.
time yield
ee
(%)
entry imine
Z
X
cat. (mol %)
(h)
(%)
1
2
3
4
5
6
7
3
7a
7b
8a
11c
11c
12c
5-Ts (0.5)
5-Ts (1)
5-Ts (1)
8
992
<1
20
842
H
NO2
MeO NO2
NH2
13
13
16
16
72
84
97.1
85
H
5-Ts (2)
66
<1
11
MeO NHTs
MeO NHTs
5-Ts (2)b
5-Ts (7.5)c
96
MeO N(MOM)- 5-Ts (7.5)c
58d >99
Ts
8
9
12d
12e
MeO N(MOM-
5-Ts (7.5)c
84
84
53d
53d
93c
97
Mse
MeO N(MOM)- 5-Ts (7.5)c
(1-Nps)f
10
11
12
13
14
13c
20a
20a
20b
20b
MeO NBnTs
H
H
MeO Br
MeO Br
5-Ts (7.5)
5-Ts (1)
5-Np s (1)
5-Ts (0.67)
72
13
13
13
76g g98h
Br
Br
41
59
67
71
94
94
98.7
98.3
5-Np s (0.67) 13
Exp er im en ta l Section
a
Unless indicated otherwise, all reactions were done in CH2Cl2
at rt, 0.4 M substrate, 6:1 ratio of Et3N/HCO2H azeotrope/
Gen er a l Meth od s. Column chromatography was per-
formed using Acros silica gel (0.06-0.2 mm).
b
substrate; conversion was established by HPLC analysis. 30:1
ratio of Et3N/HCO2H azeotrope/substrate. c 55:1 ratio of Et3N/
Gen er a l P r oced u r e for Asym m etr ic Tr a n sfer Hyd r o-
gen a tion (0.5 m m ol Sca le; Ta ble 1). The 3,4-dihydroiso-
quinoline (0.5 mmol) and Ru catalyst 5 (see Table 1 for loading)
were placed in a 2 mL vial fitted with a septum, outlet needle
(important!), and stirring bar, and CH2Cl2 (1.0 mL, distilled
from CaH2 prior use) was added followed by the formic acid-
triethylamine 5:2 azeotrope12 (0.25 mL). The reaction mixture
was stirred at room temperature (ca. 20 °C) for the apropriate
time (see Table 1), and saturated Na2CO3 solution (0.25 mL)
was carefully added (vigorous gas evolution!). After being
stirred for 10 min, the mixture was partitioned between 8 mL
of CH2Cl2 and 5 mL of Na2CO3 solution. The basic water layer
was washed with 2 mL of CH2Cl2, the organic layers were
combined, washed with water (3 × 2 mL) and brine, dried (Na2-
SO4), filtered, and evaporated (aspirator). The dark red-brown
residue was filtered through a silica gel pad (column: 20 ×
20 mm) using EtOAc (100 mL), solvent was removed (aspira-
tor), and the residue was dried in vacuo (ca. 1 mmHg) over
P2O5 for 12 h. After weighing, all the material was carefully
dissolved in 5 mL of CH2Cl2, transferred into a 25 mL
volumetric flask (with CH2Cl2 rinsing), and filled up to the
mark with the appropriate HPLC eluent. An aliquot of 0.5 mL
was transferred to another 10 or 25 mL volumetric flask and
filled up to the mark with HPLC eluent. The resulting solution
(usually 0.09-0.6 mg/mL, depending on HPLC separation
conditions) was used for HPLC analysis to establish conversion
and enantiomer excess.
d
HCO2H azeotrope/substrate. Isolated yield of purified material.
g
e 1-Naphthylsulfonyl. f 2-Naphthylsulfonyl. Overall yield of N-
h
trifluoroacetamide 18 (R ) Bn) from 13c. Assayed as the
N-trifluoroacetamide 18 (R ) Bn).
90% ee.7 As expected from the configuration of 23b, the
major product is enantiomeric with 26 obtained using 1.
Since only the (R)-enantiomer 1 is commercially avail-
able, the Noyori reduction method described here pro-
vides access to the complementary series of asymmetric
proton donors. The original literature method used in the
synthesis of (R)-1 involves a nonselective reduction step
followed by resolution,3 so a Noyori reduction approach
using the enantiomer of catalyst 5 may be an effective
alternative. Further studies are planned to test deriva-
tives of 10a and 10b in asymmetric anion protonation
and will be described in subsequent publications.
Con clu sion s
The Noyori reduction procedure affords excellent enan-
tioselectivities with several of the ortho-substituted imi-
nes (Table 1).2,8 High enantiomer excess was observed
with ortho substituents of varying steric bulk and
electronic properties, including o-Br, o-NO2, and o-N(R)SO2-
Ar (Table 1, entries 3 and 6-15), so the exact source of
the improvement over the parent system (entry 1) is not
clear. However, a simple o-methyl substituent also
improves enantioselectivity,9 so the substituent effect
could be steric in nature. If the ortho substituent is
needed to favor a specific rotamer in one of the competing
diastereomeric transition states for hydrogenation, then
good ee might be expected with the diverse ortho sub-
stituents, but we cannot explain the subtle variations
among the nitrogen substituents.
HP LC An a lysis a n d Ch a r a cter iza tion of Tr a n sfer
Hyd r ogen a tion Red u ction P r od u cts. (S)-1-(2-Nitr op h en -
yl)-6,7-dim eth oxy-1,2,3,4-tetr ah ydr oisoqu in olin e (9b): op-
tical rotation (96.5% ee, HPLC/csp) [R]D ) -105.1 (c ) 1.17,
CHCl3); analytical TLC on silica gel, EtOAc, Rf ) 0.48 (tailing).
Pure material was obtained by crystallization from EtOAc/
hexane: mp 144-145 °C; pale yellow crystals; IR (Nujol, cm-1
)
3313 NH, 1532 NO2, 1367 NO2; 200 MHz NMR (CDCl3, ppm)
δ 7.74-7.69 (1H, m) 7.38-7.22 (2H, m) 7.03-6.98 (1H, m) 6.53
(1H, s) 6.08 (1H, s) 5.43 (1H, s) 3.76 (3H, s) 3.54 (3H, s) 3.05-
2.59 (4H, m) 2.49-2.00 (1H, br s); 13C NMR (50 MHz, CDCl3,
ppm) δ 150.1, 147.8, 147.1, 138.8, 132.1, 131.8, 128.2, 127.9,
127.7, 123.9, 111.3, 110.7, 55.8, 55.7, 55.3, 40.6, 29.0. Anal.
Calcd for C17H18N2O4: C, 64.95; H, 5.78; N, 8.91. Found: C,
64.96; H, 5.84; N, 8.84. HPLC assay for conversion: Symmetry
The principal limitation of the asymmetric hydrogena-
tion procedure at the current level of development is in
(7) Vedejs, E.; Kruger, A. W.; Suna, E. Manuscript in preparation.
(8) (a) For a recent review on asymmetric reduction of cyclic imines,
see: Yurovskaya, M. A.; Karchava, A. V. Tetrahedron: Asymmetry
1998, 3331. (b) Chan, Y. N. C.; Osborn, J . A. J . Am. Chem. Soc. 1990,
112, 9400. Willoughby, C. A.; Buchwald, S. L. J . Am. Chem. Soc. 1994,
116, 8952. Tani, K.; Onouchi, J .; Yamagata, T.; Kataoka, Y. Chem. Lett.
1995, 955. Verdaguer, X.; Lange, U. E. W.; Reding, M. T.; Buchwald,
S. L. J . Am. Chem. Soc. 1996, 118, 6784. Morimoto, T.; Suzuki, N.;
Achiwa, K. Tetrahedron: Asymmetry, 1998, 9, 183. Nishibayashi, Y.;
Takei, I.; Uemura, S.; Hidai, M. Organometallics, 1998, 17, 3420.
(9) We thank Prof. Noyori for informing us of this result and for
providing a sample of catalyst 5-Np s.
(10) The chloro analogue of 20a was reduced with promising ee
(catalyst 5-Ts: 66% yield, 91% ee; catalyst 5-Np s: 56% yield, 86%
ee), but the amine displacement was not attempted.
(11) The copper-catalyzed aminations give lower yields with larger
n-alkylamines; for example, N,N-dimethylethylenediamine afforded
30% of the displacement product under the usual conditions.
(12) Prepared by slow addition of 96% HCOOH (Aldrich, 29.5 mL,
0.75 M) to a cooled neat NEt3 (41.9 mL, 0.3M) and distillation of the
resulting mixture in vacuo (aspirator, 15 mmHg, collected fractions
bp: 91-92 °C, bath temperature 155-160 °C); see: Narita, K.; Sekiya,
M. Chem. Pharm. Bull. 1977, 25, 135.

6728 J . Org. Chem., Vol. 64, No. 18, 1999
Vedejs et al.
C18, 15 cm × 3.9 mm i.d., mobile phase 50% acetonitrile/50%
0.2 M acetate buffer (pH ) 5.0), flow rate 0.8 mL/min, detector
UV 254 nm. Retention time: starting material 7b, 3.6 min;
product 9b, 2.0 min. HPLC/csp: Daicel CHIRALCEL OD, 25
cm × 4.6 mm i.d., mobile phase 10% i-PrOH/90% Hex, flow
rate 0.8 mL/min, detector UV 254 nm. Retention time: 13.9
min (R isomer), minor and 19.8 min (S isomer), major.
(S )-1-(2-Am in op h e n yl)-1,2,3,4-t e t r a h yd r oisoq u in o-
lin e (10a ): optical rotation (>99% ee, HPLC/csp) [R]D ) -42.5
(c ) 1.04, CHCl3); analytical TLC on silica gel, 1:1 EtOAc/
hexane, Rf ) 0.2. Pure material was obtained by crystallization
from EtOAc/hexane: mp 113-114 °C; colorless needles; mo-
lecular ion calcd for C15H16N2 224.13139, found m/e ) 224.1312,
error ) 1 ppm; IR (KBr, cm-1) 3388 NH, 3323 NH; 300 MHz
NMR (CDCl3, ppm) δ 7.18-6.98 (5H, m) 6.81 (1H, d, J ) 7.8
Hz) 6.69 (1H, ddd, J ) 7.5, 7.5, 1.5 Hz) 6.62 (1H, dd, J ) 8.1,
1.2 Hz) 5.09 (1H, s) 4.49 (2H, br s) 3.31-3.22 (1H, m) 3.13-
3.00 (2H, m) 2.84-2.75 (1H, m) 2.00 (1H, br s); 13C NMR (75
MHz, CDCl3, ppm) δ 146.2, 137.5, 135.2, 131.2, 129.1, 128.6,
127.3, 127.0, 126.6, 126.0, 117.5, 116.9, 62.0, 43.1, 29.8. Anal.
Calcd for C15H16N2: C, 80.31; H, 7.20; N, 12.49. Found: C,
80.09; H, 7.37; N, 12.43. HPLC/csp assay: Daicel CHIRALCEL
OD, 25 cm × 4.6 mm i.d., mobile phase 5% EtOH/95% Hex/
0.1% Et2NH, flow rate 0.8 mL/min, detector UV 254 nm.
Retention time: starting material 8a , 12.3 min; product 10a ,-
16.0 min (R isomer), minor and 16.9 min (S isomer), major.
(S )-1-(2-Br om op h e n yl)-1,2,3,4-t e t r a h yd r oisoq u in o-
lin e (21a ): optical rotation (97% ee, HPLC/csp) [R]D ) -26.1
(c ) 1.15, CHCl3); analytical TLC on silica gel, 1:3 EtOAc/
hexane, Rf ) 0.24; IR (Nujol, cm-1) 3325 NH; 200 MHz NMR
(CDCl3, ppm) δ 7.60 (1H, dd, J ) 7.4, 1.6 Hz) 7.25-6.97 (6H,
m) 6.76 (1H, d, J ) 7.4 Hz) 5.63 (1H, s) 3.24-2.81 (4H, m)
2.53 (1H, br s). 13C NMR (50 MHz, CDCl3, ppm) δ 143.5, 136.9,
135.7, 132.8, 131.2, 129.1, 128.8, 128.0, 127.3, 126.4, 125.7,
124.6, 60.0, 41.1, 29.6. Hyd r och lor id e sa lt of 21a : optical
rotation (97% ee, HPLC/csp) [R]D ) -12.7 (c ) 1.19, H2O). Pure
material was obtained by crystallization from acetonitrile: mp
190 °C dec. Anal. Calcd for C15H15BrClN: C, 55.49; H, 4.67;
N, 4.32. Found: C, 55.54; H, 4.63; N, 4.27. HPLC assay for
conversion: Daicel CHIRALCEL OJ , 25 cm × 4.6 mm i.d.,
mobile phase 10% i-PrOH/90% Hex, flow rate 1.0 mL/min,
detector UV 254 nm. Retention time: starting imine 20a , 24.2
min; product 21a , 6.4 min (S isomer), major and 7.6 min (R
isomer), minor.
P r ep a r a tive-Sca le Syn th esis of (S)-1-(2-Br om op h en yl)-
6,7-d im eth oxy-1,2,3,4-tetr a h yd r oisoqu in olin e (21b) by
Tr a n sfer Hyd r ogen a tion . Imine 20b (1.73 g, 5 mM) and Ru
catalyst 5-Ts (21.8 mg, 0.037 mM) were placed in a 15 mL
Erlenmeyer flask fitted with a septum, outlet needle (impor-
tant!), and magnetic stirring bar, and 10 mL of CH2Cl2
(distilled from CaH2 prior use) was added, followed by 2.5 mL
of HCOOH-NEt3 5:2 azeotrope.12 After the mixture was
stirred at room temperature for 24 h, 10 mL of saturated Na2-
CO3 solution was added (caution: vigorous gas evolution!), and
stirring was continued for an additional 5 min. The resulting
mixture was diluted with 50 mL of CH2Cl2 and 30 mL of water,
and the layers were separated. The organic phase was washed
with saturated Na2CO3 solution (3 × 10 mL), 10 mL of water,
and 10 mL of brine and dried (Na2SO4). After filtration and
evaporation (aspirator), the residue was purified by flash
chromatography (column: 200 × 35 mm) with 1:1 EtOAc/Hex,
100 mL of dead volume; collected fractions 14-19 (25 mL),
then eluent changed to 3:1 EtOAc/Hex and an additional 450
mL collected to give 1.77 g of yellow-brown oil. The oil was
dissolved in Et2O (100 mL) and placed in a 200 mL round-
bottom flask fitted with an efficient stirring bar, and HCl gas
was passed over the solution until white flakes precipitated.
The hygroscopic solid was quickly filtered, washed with dry
Et2O, dried in vacuo (aspirator), and recrystallized from diethyl
ether-acetonitrile to give 1.41 g (72%) of white flakes of the
hydrochloride salt of 21b (hygroscopic!): mp 185 °C dec; [R]D
) -31.8 (c ) 1, H2O); 360 MHz NMR (CDCl3, ppm) δ 10.83
(1H, br s) 9.69 (1H, br s) 7.67 (1H, d, J ) 7.6 Hz) 7.30-7.22
(3H, m) 6.64 (1H, s) 6.11 (1H, s) 5.99 (1H, s) 3.87 (3H, s) 3.63
(3H, s) 3.38-3.21 (3H, m) 3.07-2.99 (1H, m). Anal. Calcd for
C17H19BrClNO2: C, 53.07; H, 4.99; N, 3.64. Found: C, 53.07;
H, 4.96; N, 3.76. A sample of the hydrochloride (100 mg) was
dissolved in 10 mL water, basified with 1 N NaOH, and
extracted with 3 × 5 mL portions of CH2Cl2. The combined
organic extracts were washed with water (5 mL) and brine (5
mL) and dried (Na2SO4). After filtration and solvent removal
(aspirator), 90 mg (99% yield) of colorless amine 21b was
obtained: optical rotation (99% ee, HPLC/csp) [R]D ) -62.4 (c
) 1, CHCl3); analytical TLC on silica gel, EtOAc, Rf ) 0.55;
colorless oil; IR (neat, cm-1) 3315 NH, 1263 (Ar)CO, 1230 (Ar)-
CO; 360 MHz NMR (CDCl3, ppm) δ 7.52 (1H, dd, J ) 7.8, 1.3
Hz) 7.10 (1H, ddd, J ) 7.4, 7.4, 1.3 Hz) 7.03 (1H, ddd, J ) 7.6,
7.6, 1.9 Hz) 6.90 (1H, dd, J ) 7.6, 1.9 Hz) 6.59 (1H, s) 6.21
(1H, s) 5.47 (1H, s) 3.83 (3H, s) 3.63 (3H, s) 3.05 (1H, ddd, J )
12.0, 6.8, 5.2 Hz) 2.99 (1H, td, J ) 12.0, 5.7 Hz) 2.88-2.74
(2H, m) 2.18 (1H, br s); 13C NMR (90 MHz, CDCl3, ppm) δ
148.3, 147.7, 144.1, 133.1, 131.4, 129.0, 128.9, 128.3, 127.4,
124.7, 111.7, 111.1, 58.9, 55.31, 55.26, 39.9, 28.4.
Con ver sion of Ar yl Br om id e 21b t o (S)-1-(2-Am in o-
p h en yl)-6,7-d im et h oxy-1,2,3,4-t et r a h yd r oisoq u in olin e
(10b). According to the method of Ott et al.,3 a Parr pressure
reactor (450 mL) was charged with 21b (7.5 g, 19.5 mM),
copper powder (Fluka, 488 mg, 7.7 mM), and copper(I) chloride
(488 mg, 4.93 mM; washed with 1 N HCl, then carefully with
water; dried in vacuo over P2O5, and stored in the dark), and
75 mL of liquid ammonia was carefully added. The reaction
was heated and stirred at 70 °C (45 bar pressure) for 120 h
and then cooled to room temperature, and ammonia was
evaporated. The crystalline residue was partitioned between
chloroform (100 mL) and water (100 mL). The dark blue water
layer was washed with chloroform (2 × 20 mL), and the
combined organic extracts were washed with water (3 × 30
mL) and brine and dried (Na2SO4). After filtration and solvent
evaporation (aspirator), the residue was filtered through silica
gel (column: 100 × 35 mm) using 800 mL of EtOAc to give
crude product (5.27 g, 95% yield), which was recrystallized
from EtOAc-hexane to yield 4.55 g (82%) of 10b; analytical
TLC on silica gel, EtOAc, Rf ) 0.2; analytical HPLC/csp (Daicel
CHIRACEL OD, 25 cm × 4.6 mm i.d.), mobile phase 20%
EtOH/80% hexane, flow: 0.7 mL/min, retention time 13.0 min
(R isomer), minor and 17.2 min (S isomer), major, ratio 99.5:
0.5 (99% ee). Pure material was obtained by crystallization
from EtOAc/hexane: mp 170-171 °C; colorless crystals; optical
rotation [R]D ) +9.7 (c ) 1, CHCl3); IR (Nujol, cm-1) 3410 NH,
3320 NH, 1235 (Ar)CO; 360 MHz NMR (CDCl3, ppm) δ 7.09
(1H, ddd, J ) 7.7, 7.7, 1.6 Hz) 6.94 (1H, dd, J ) 7.5, 1.6 Hz)
6.68 (1H, ddd, J ) 7.5, 7.5, 1.1 Hz) 6.64 (1H, dd, J ) 8.0, 1.1
Hz) 6.62 (1H, s) 6.33 (1H, s) 5.06 (1H, s) 4.54 (2H, br s) 3.86
(3H, s) 3.64 (3H, s) 3.22 (1H, dt, J ) 12.0, 5.2 Hz) 3. 09-3.02
(1H, m) 2.98-2.90 (1H, m) 2.71 (1H, dt, J ) 16.0, 4.6 Hz)
1.90 (1H, br s); 13C NMR (90 MHz, CDCl3, ppm) δ 147.7,
147.3, 146.2, 130.73, 130.69, 129.3, 128.4, 127.4, 127.3, 117.4,
116.6, 111.6, 110.0, 60.8, 55.8, 42.5, 29.2. Anal. Calcd for
C
17H20N2O2: C, 71.79; H, 7.1; N, 9.85. Found: C, 71.76; H,
7.18; N, 9.87.
(S)-1-(2-Meth yla m in oph en yl)-6,7-dim eth oxy-1,2,3,4-tet-
r a h yd r oisoqu in olin e (23b). A 100 mL pressure reactor was
charged with (S)-bromophenylisoquinoline hydrochloride 21b
(1.92 g, 4.99 mM), copper powder (Fluka, 126 mg, 1.99 mM)
and copper(I) chloride (126 mg, 1.27 mM; washed with 1 N
HCl, then carefully with water; dried in vacuo over P2O5, and
stored in the dark), and 35 mL of liquid methylamine was
added. The reaction was heated at 70 °C for 120 h, methy-
lamine was evaporated, and workup was performed as de-
scribed above for 10b. Filtration through silica gel (column:
50 × 30 mm) using 300 mL of EtOAc gave a product that was
recrystallized from EtOH to yield 1.19 g (80%) of 23b as
colorless crystals; analytical TLC on silica gel, EtOAc, Rf )
0.23-0.59 (tailing); analytical HPLC/csp (Daicel CHIRACEL
OD, 25 cm × 4.6 mm i.d.), flow rate 1.0 mL/min, mobile phase
10% i-PrOH/90% hexane, retention time 13.6 min (S isomer),
major and 17.3 min (R isomer), minor, ratio 99.5:0.5 (99% ee).
Pure material was obtained by crystallization from ethanol:
mp 96-97 °C; colorless crystals; [R]D ) +52.9 (c ) 1, CHCl3);
IR (Nujol, cm-1) 3330 NH, 3320 NH, 1270 (Ar)CO; 300 MHz

Substituted Isoquinolines by Noyori Transfer Hydrogenation
J . Org. Chem., Vol. 64, No. 18, 1999 6729
NMR (CDCl3, ppm) δ 7.21 (1H, ddd, J ) 7.8, 7.8, 1.8 Hz) 6.88-
6.85 (1H, m) 6.65 (1H, s) 6.62-6.60 (2H, m) 6.33 (1H, s) 5.04
(1H, s) 3.86 (3H, s) 3.63 (3H, s) 3.17 (1H, dt, J ) 12.4, 5.6 Hz)
3.01 (1H, ddd, J ) 12.4, 8.0, 5.0 Hz) 2.95-2.82 (1H, m) 2.75
(3H, s) 2.70 (1H, dt, J ) 16.2, 5.0 Hz); 13C NMR (75 MHz,
DEPT, CDCl3, ppm) δ 148.7, 147.6, 147.1, 130.4, 129.3, 128.6,
127.2, 126.7, 115.7, 111.4, 110.4, 110.0, 60.5, 55.8, 55.7, 42.2,
30.4, 29.1. Anal. Calcd for C18H22N2O2: C, 72.44; H, 7.45; N,
9.39. Found: C, 72.50; H, 7.55; N, 9.33.
(R)-1-(2-Meth ylam in oph en yl)-1,2,3,4-tetr ah ydr oisoqu in -
olin e (23a ). The title compound was prepared from 21a as
described for 23b; HPLC/csp assay: Daicel CHIRALCEL OD,
25 cm × 4.6 mm i.d., mobile phase 5% i-PrOH/95% Hex, flow
rate 0.8 mL/min, detector UV 254 nm. Retention time: 23a ,-
10.4 min (R isomer), minor and 11.2 min (S isomer), major;
analytical TLC on silica gel, 1:1 EtOAc/hexane, Rf ) 0.44-
0.19 (tailing). Pure material was obtained by crystallization
from ethanol-water: mp 87-88 °C; colorless plates; optical
rotation (99% ee, HPLC/csp) [R]D ) +5.2 (c ) 1.17, CHCl3); IR
(Nujol, cm-1) 3307 NH; 200 MHz NMR (CDCl3, ppm) δ 7.23-
7.13 (4H, m) 6.93 (1H, dd, J ) 7.8, 1.8 Hz) 6.85-6.79 (1H, m)
6.68-6.61 (1H, m) 6.62 (1H, d, J ) 7.8 Hz) 5.10 (1H, s) 3.31-
2.96 (3H, m) 2.87-2.66 (1H, m) 2.71 (3H, s) 2.1-1.4 (1H, br
s); 13C NMR (50 MHz, CDCl3, ppm) δ 148.6, 137.5, 135.0, 130.7,
128.8, 128.6, 126.8, 126.4, 126.3, 125.7, 115.5, 110.5, 61.8, 42.7,
30.3, 29.6. Anal. Calcd for C16H18N2: C, 80.62; H, 7.63; N,
11.76. Found: C, 79.70; H, 7.69; N, 11.61.
Ack n ow led gm en t. This work was supported by the
National Institutes of Health (GM44724) and by a
NATO travel grant (LG930053). The authors wish to
thank Dr. S. Hitchcock for confirming some of the
hydrogenation and amination results.
Su p p or tin g In for m a tion Ava ila ble: Preparation and
characterization of imine starting materials for Table 1 and
of the sulfonamide-derived reduction products; NMR spectra
for asymmetric hydrogenation derived products described in
the Experimental Section (9b, 10a ,b, 21a ,b, and 23b). This
material is available free of charge via the Internet at
http://pubs.acs.org.
J O990594S