Two general routes to 1,4-disubstituted-2,3,4,5-tetrahydro- 1H-3-benzazepines.

Article abstract of DOI:10.1021/ol0067641

[structure] Two general routes to 1,4-disubstituted-2,3,4, 5-tetrahydro-1H-3-benzazepines are described. Both routes utilize an appropriately functionalized phenethylamino alcohol as the penultimate intermediate: the first route makes use of the reductive amination of a benzyl alkyl ketone with alpha-(aminomethyl)benzyl alcohol, while the second route utilizes the addition of a Grignard reagent to the oxazolidine derived from a substitued phenylacetaldehyde and alpha-(methylaminomethyl)benzyl alcohol. In all cases studied, the cis-1,4-disubstituted-2,3,4, 5-tetrahydro-1H-3-benzazepine was obtained as the major product.

Full text of DOI:10.1021/ol0067641

ORGANIC
LETTERS
2000
Vol. 2, No. 25
4099-4102
Two General Routes to
1,4-Disubstituted-2,3,4,5-tetrahydro-
1H-3-benzazepines
Samuel W. Gerritz,* Jennifer S. Smith, Suganthini S. Nanthakumar,†
David E. Uehling, and Jeffery E. Cobb
Glaxo Wellcome, Inc., FiVe Moore DriVe, P.O. Box 13398,
Research Triangle Park, North Carolina 27709
swg25518@glaxowellcome.com
Received October 23, 2000
ABSTRACT
Two general routes to 1,4-disubstituted-2,3,4,5-tetrahydro-1H-3-benzazepines are described. Both routes utilize an appropriately functionalized
phenethylamino alcohol as the penultimate intermediate: the first route makes use of the reductive amination of a benzyl alkyl ketone with
r-(aminomethyl)benzyl alcohol, while the second route utilizes the addition of a Grignard reagent to the oxazolidine derived from a substitued
phenylacetaldehyde and r-(methylaminomethyl)benzyl alcohol. In all cases studied, the cis-1,4-disubstituted-2,3,4,5-tetrahydro-1H-3-benzazepine
was obtained as the major product.
Chemical entities based on the 1-phenyl-2,3,4,5-tetrahydro-
1H-3-benzazepine1 nucleus have been shown to bind to the
dopaminergic family of G-protein-coupled receptors (GPCRs).2
For example, SCH 23390 (1) is a 300 picomolar antagonist
of the dopamine D1 receptor.2c As part of a drug discovery
effort seeking small molecule ligands for other GPCRs, we
became interested in the preparation of 1-phenyl-4-alkyl-3-
benzazepines as exemplified by 2 (R2 ) alkyl). Despite the
large number of known analogues2,3 of 1, a general route to
1,4-disubstituted-3-benzazepines has not been reported.4
Typically, 1-phenyl-3-benzazepines (2, R2 ) H) have been
synthesized via the acid-mediated electrophilic cyclization
of an amino alcohol precursor (3, R2 ) H).2 These amino
alcohol precursors are readily prepared by the reaction of
styrene oxide with the appropriate N-methylphenethylamine.
The logical extension of this approach for the preparation
of 1,4-disubstituted-3-benzazepines (2, R2 ) alkyl) was
unsuccessful because the requisite N-methyl-1-alkyl-2-phen-
ethylamines provided poor yields (<30%) of 3 (R2 ) alkyl)
upon reaction with styrene oxide.5 To circumvent this
† Current address: Vertex Pharmaceuticals, 130 Waverly Street, Cam-
bridge, MA 02139.
(1) For the remainder of this Letter, we will refer to 2,3,4,5-tetrahydro-
1H-3-benzazepines as 3-benzazepines.
(2) Reviews: (a) Weinstock, J.; Hieble, J. P.; Wilson, J. W., III. Drugs
Future 1985, 10, 645. (b) Barnett, A.; Gold, E. H.; Chipkin, R. E.; Iorio,
L. C. AdV. Biosci. 1990, 77, 1. (c) Iorio, L. C.; Barnett, A.; Billard, W.;
Gold, E. H. AdV. Exp. Med. Biol. 1986, 204, 1.
(3) (a) Walter, L. A.; Chang, W. K. U.S. Patent 3393192, 1968; Chem.
Abstr. 1969, 96507. (b) Kasparek, S. AdV. Heterocycl. Chem. 1974, 17, 45.
(c) Berger, J. G.; Chang, W. K.; Clader, J. W.; Hou, D.; Chipkin, R. E.;
McPhail, A. T. J. Med. Chem. 1989, 32, 1913. (c) Chumpradit, S.; Kung,
M. P., Billings, J. J.; Kung, H. F. J. Med. Chem. 1991, 34, 877.
(4) The synthesis of a 1-phenyl-4-methyl-3-benzazepine is described in
ref 3a, but no additional 4-position substituents were reported.
10.1021/ol0067641 CCC: $19.00 © 2000 American Chemical Society
Published on Web 11/22/2000

limitation, we have developed two novel and complementary
routes to 3 (R2 ) alkyl). The first route is outlined in eq 1,
Scheme 1
in which 3 (R2 ) alkyl) arises from the reductive amination
of an alkyl benzyl ketone. The retrosynthesis described in
eq 2 utilizes a Grignard addition to an appropriately
substituted oxazolidine to afford 3 (R2 ) alkyl). The
successful application of these approaches to the synthesis
of 1,4-disubstituted-3-benzazepines is reported herein.
Initially, we focused on the “reductive amination” route
described in eq 1, and our synthesis commenced with the
preparation6 of benzyl ketones 4a-f. The conversion of
commercially available 3,4-dimethoxyphenylacetic acid to
its corresponding Weinreb amide7 5 proceeded smoothly
under standard conditions. As described in eq 3, addition of
methylmagnesium bromide to 5 provided ketone 4a in a 93%
yield. The generality of this method allowed the synthesis
of ketones 4b-f with comparable yields.
C(4) methyl group and the pseudoaxial C(2) hydrogen of
cis-9a. Conversely, the observation of an NOE between the
pseudoaxial C(4) and C(2) hydrogens of trans-9a is consis-
tent with a 1,4-trans stereochemical relationship. We sub-
sequently confirmed the relative stereochemistry of cis-9a
via X-ray crystallography of a closely related analogue.
To eliminate the N(3) reductive alkylation step, we
explored the use of R-(methylaminomethyl)benzyl alcohol
in place of 6, but we found that only ketone 4a underwent
successful reductive amination with the more hindered amino
alcohol. While the use of 6 in the reductive amination step
adds a step to the overall synthetic scheme, it does facilitate
the introduction of a variety of N(3) substituents at a late
stage in the synthesis. We have successfully utilized alde-
hydes and simple ketones (i.e., acetone) for the reductive
alkylation of 8a-d.
Entries 1-8 of Table 1 summarize the 3-benzazepines
synthesized via reductive amination of 4a-d with 6. In all
cases, a 3:1 ratio of cis:trans 3-benzazepines was obtained,
suggesting that the C(4) substituent plays a minor role in
controlling the stereochemical outcome of the acid-mediated
cyclization.
Although this “reductive amination” route to 1,4-disub-
stituted-3-benzazepines approach is quite general, it has the
significant disadvantage that it requires the installation of
the R2 group of 3 at an early stage in the synthesis. In
addition, sterically or electronically demanding ketones (i.e.,
4e and 4f) did not undergo reductive amination with 6. To
As described in Scheme 1, ketone 4a was reacted with
R-(aminomethyl)benzyl alcohol (6) under standard reductive
amination conditions8 to afford amino alcohol 7a as a 1:1
mixture of diastereomers. Upon treatment with neat meth-
anesulfonic acid at 0 °C, 7a cyclized to provide the desired
4-methyl-1-phenyl-3-benzazepine 8a as an inseparable mix-
ture of diastereomers. Reductive alkylation of N(3) of 8a
with formaldehyde afforded 9a as a 3:1 mixture of readily
separable cis and trans diastereomers in a 59% yield over
three steps. The relative stereochemistry was assigned by
NOESY analysis as summarized in Figure 1. The expected
pseudoequatorial orientation of the phenyl substituent at C(1)
of cis- and trans-9a was confirmed by the observation of an
NOE between the pseudoaxial C(1) hydrogen and the
pseudoequatorial hydrogen at C(2). The relative stereochem-
istry of the C(1) and C(4) substituents of cis-9a was assigned
through the observation of an NOE between the pseudoaxial
(5) This synthetic step is reported in ref 3a without yield information.
(6) All new compounds reported herein provided analytical data con-
sistent with their proposed structures.
(7) Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 22, 3815.
(8) Borch, R. F.; Bernstein, M. D.; Durst, H. D. J. Am. Chem. Soc. 1971,
93, 2897.
Figure 1. NOEs observed between the C(1), C(2), and C(4)
substituents of cis- and trans-9a.
4100
Org. Lett., Vol. 2, No. 25, 2000

Scheme 2
Table 1. Synthesis of 1,4-Disubstituted-3-benzazepines
cis:trans
yield,
b
entry
routea
R1
R2
R3
ratio
%
1
2
4
6
7
RA
RA
RA
RA
RA
RA
OO
OO
OO
OO
OO
OO
OO
OO
OO
OO
Me
Me
Me
Me
iPr
Bn
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Et
OMe
OMe
OMe
OMe
OMe
OMe
OMe
OMe
OMe
OMe
OMe
OMe
H
3:1
3:1
3:1
3:1
3:1
3:1
3:1
3:1
3:1
3:1
54
48
50
52
44
49
68
72
66
74
49
54
42c
38c
42c
47c
Bn
iBu
Me
Bn
Me
Et
Bn
iBu
iPr
Ph
8
9
Our initial studies confirmed that addition of a Grignard
reagent to 12 yielded the desired amino alcohol product. As
described in Scheme 3, addition of methylmagnesium
10
11
12
13
14
15
16
17
18
n.d.d
n.d.
cis only
cis only
cis only
cis only
Scheme 3
Me
Et
iPr
Bn
H
H
H
a RA ) reductive amination; OO ) oxazolidine opening. b Yields
calculated over 4 steps (RA route) or 3 steps (OO route). c Does not include
the yield of spiropiperidine dione (approximately 20%). d n.d. ) not
determined.
overcome these limitations, we devised the route shown in
eq 2 in which the R2 substituent of 4 is introduced via the
addition of a Grignard reagent to the oxazolidine derived
from R-(methylaminomethyl)benzyl alcohol and a substituted
phenylacetaldehyde. The reaction of Grignard reagents with
oxazolidines to afford amino alcohol derivatives is well
precedented,9 although this methodology has not been applied
to the synthesis of 1,4-disubstituted-3-benzazepines.10
Our synthesis commenced with the conversion of Weinreb
amide 5 to 3,4-dimethoxyphenylacetaldehyde with DIBALH
in THF at -78 °C (Scheme 2). The unstable aldehyde was
not isolated but instead, following a Rochelle’s salt workup,
was treated immediately with R-(methylaminomethyl)benzyl
alcohol (11) in toluene and concentrated in vacuo to afford
oxazolidine 12 as a 1.2:1 mixture of diastereomers. Oxazo-
lidine 13 was synthesized in an analogous fashion from
Weinreb amide 10. Oxazolidines 12 and 13 were not stable
to silica gel chromatography, but they would tolerate an
aqueous workup and subsequent storage as a solution in
toluene.
bromide to 12 provided the amino alcohol 14a as a 1:1
mixture of diastereomers, which upon treatment with neat
methanesulfonic acid readily cyclized to provide a 3:1
mixture of the aforementioned 4-methyl-1-phenyl-3-benz-
azepine diastereomers cis-9a and trans-9a. 3-Benzazepines
9b-d were synthesized by analogy. These 3-benzazepines
were previously synthesized by the reductive amination of
4a-d (vide supra), a route which required four steps from
Weinreb amide 5. Even though the oxazolidine route saves
only one linear step from 5, the introduction of the C(4)
substituent in the penultimate step greatly streamlines the
synthesis of C(4) analogues. Moreover, C(4) substituents
which were too sterically or electronically demanding (i.e.,
isopropyl or phenyl) to undergo reductive amination with
R-(aminomethyl)benzyl alcohol were easily introduced via
the oxazolidine route, affording 9e and 9f, respectively.
Unfortunately, we were not able to unambiguously assign
the stereochemistry of 9e and 9f, and so the ratio of cis to
trans diastereomers is unknown. Entries 9-14 of Table 1
summarize the 3-benzazepines synthesized via the opening
of oxazolidine 5.
(9) (a) Senkus, M. J. Am. Chem. Soc. 1945, 67, 1515. (b) Goodson, L.
H.; Christopher, H. J. Am. Chem. Soc. 1950, 72, 358. (c) Neelakantan, L.
J. Org. Chem. 1971, 36, 2256. (d) Takahashi, H.; Niwa, H.; Higashiyama,
K. Heterocycles 1988, 27, 2099. (e) Davidsen, S. K.; Chu-Moyer, M. Y. J.
Org. Chem. 1989, 54, 5558.
(10) It should be noted that analogous oxazolidines have been utilized
for the synthesis of 1-phenyl-3-benzazepine (e.g., 3, R2 ) H), although in
this case the oxazolidines were simply reduced to amino alcohol 4 (R2 )
H) via borohydride reduction: (a) Labaw, C. S.; Tremper, A. W. Eur. Pat.
Appl. 177326, 1985; Chem. Abstr. 1985, 60593. (b) Coote, S. J.; Davies,
S. G.; Middlemiss, D.; Naylor, A. Tetrahedron Lett. 1989, 30, 3581.
The cyclization of 14a-d (and 8a-d) to 9a-d consis-
tently provided a 3:1 mixture of cis and trans 3-benzazepines.
Org. Lett., Vol. 2, No. 25, 2000
4101

Because amino alcohol 8 is a 1:1 mixture of diastereomers,
the 3:1 ratio of cis-9 to trans-9 must result from the
equilibration of the C(1) or C(4) center at some point in the
acid-mediated cyclization. The available evidence indicates
that this occurs via formation of the benzylic cation at C(1),
followed by cyclization to form the product 3-benzazepine,
rather than via epimerization of the C(1) or C(4) center after
cyclization. First, resubjecting cis-9 to cyclization conditions
provides no trace of trans-9, thereby ruling out the possibility
of a postcyclization C(1) or C(4) epimerization. A second
piece of supporting evidence is a previous study10b of the
formation of 1-phenyl-3-benzazepines (e.g., 3, where R2 )
H) from the corresponding amino alcohol precursor. In that
case, an enantiomerically pure (>95% ee) amino alcohol
(e.g., 4, R2 ) H) was subjected to the standard cyclization
conditions, and the resulting 3-benzazepine product was
found to possess a very low ee (6%). Taken together, these
results suggest that the C(4) center is unaffected by the
cyclization conditions. Consequently, if the C(4) substituent
could be introduced stereoselectively in amino alcohol 8, then
the resulting cis and trans 3-benzazepines would be expected
to be enantiomerically pure.11
to be the [6,6]-spiropiperidine dienone trans-17a. Neither
the trans diastereomer of 16a nor the cis diastereomer of
17a could be detected by H NMR of the crude cyclization
1
product. Entries 15-18 of Table 1 summarize the 3-benz-
azepines synthesized via the opening of oxazolidine 13.
The synthesis of a spiropiperidine dienone similar to trans-
17a has been reported,12 and it is noteworthy that although
the cyclization precursor is very similar to amino alcohol
15a, the formation of a 3-benzazepine product was not
observed. Moreover, the product spiropiperidine dienone was
reported to be unstable in strong acid or base and rapidly
reverted to starting material upon exposure to 0.1 N aqueous
HCl. We did not observe trans-17a to be unstable.
In summary, we have developed two general and efficient
methods for the synthesis of 1,4-disubstituted-3-benz-
azepines. The reductive amination route (Scheme 1) allows
the rapid analoging of the N(3) position of 3, while the
oxazolidine route (Scheme 3) offers rapid entry into C(4)
analogues of 3. Depending upon the substitution pattern of
the aryl ring of the Weinreb amide precursor (5 or 12), one
of two possible minor products is formed in addition to the
major cis-1,4-disubstituted-3-benzazepine: starting from 5,
the corresponding trans-1,4-disubstituted-3-benzazepine is
provided in ∼10% yield; starting from 12, the corresponding
trans-1,4-disubstituted spiropiperidine dienone is afforded
in ∼20% yield. While side products are generally undesirable
in organic synthesis, the unusual divergence in reaction
pathways reflected by the formation of either trans-1,4-
disubstituted-3-benzazepines or trans-1,4-disubstituted spiropi-
peridine dienones merits further investigation.
We next turned our attention to the synthesis of C(7)-
unsubstituted 3-benzazepines via oxazolidine 13. As outlined
in Scheme 4, addition of methylmagnesium bromide to 13
afforded the corresponding amino alcohol 15a. Treatment
of 15a with neat methanesulfonic acid at 0 °C provided two
products in a 1.8:1 ratio: the desired cis-1,4-disubstituted-
3-benzazepine (cis-16a) and an unanticipated minor product
which, after extensive spectroscopic analysis, was determined
Acknowledgment. The authors thank Judd Berman,
Darryl McDougald, Michael Rabinowitz, Ron Sherrill, and
Jim Veal for helpful discussions and Randy Rutkowske for
2-D NMR experiments. J.S.S. was supported by a Glaxo
Wellcome summer internship (1995).
Scheme 4
Supporting Information Available: Synthetic procedures
for the preparation of 9a from 4a and 16a from 10. This
information is available free of charge via the Internet at
http://pubs.acs.org.
OL0067641
(11) Our asymmetric synthesis of 1,4-disubstituted-3-benzazepines will
be disclosed in a subsequent publication.
(12) Sill, A. D.; Housmyer, C. L.; Gibboney, K. Tetrahedron 1987, 43,
1177.
4102
Org. Lett., Vol. 2, No. 25, 2000