Enantioselective synthesis of diversely substituted quaternary 1,4-benzodiazepin-2-ones and 1,4-benzodiazepine-2,5-diones

Article abstract of DOI:10.1021/ja0640142

Benzodiazepines are privileged scaffolds in medicinal chemistry, but enantiopure examples containing quaternary stereogenic centers are extremely rare. We demonstrate that installation of the di-(p-anisyl)methyl (DAM) group at N1 of 1,4-benzodiazepin-2-on

Full text of DOI:10.1021/ja0640142

Published on Web 11/02/2006
Enantioselective Synthesis of Diversely Substituted
Quaternary 1,4-Benzodiazepin-2-ones and
1,4-Benzodiazepine-2,5-diones
Paul R. Carlier,* Hongwu Zhao, Stephanie L. MacQuarrie-Hunter,
Joseph C. DeGuzman, and Danny C. Hsu
Contribution from the Department of Chemistry, Virginia Tech, Blacksburg, Virginia 24060
Received June 7, 2006; E-mail: pcarlier@vt.edu
Abstract: Benzodiazepines are privileged scaffolds in medicinal chemistry, but enantiopure examples
containing quaternary stereogenic centers are extremely rare. We demonstrate that installation of the di-
(p-anisyl)methyl (DAM) group at N1 of 1,4-benzodiazepin-2-ones and 1,4-benzodiazepine-2,5-diones derived
from enantiopure proteinogenic amino acids allows retentive replacement of the C3-proton via a
deprotonation/trapping protocol. A wide variety of carbon and nitrogen electrophiles function well in this
reaction, providing the corresponding quaternary benzodiazepines with excellent enantioselectivity.
Deprotonation/trapping experiments on a pair of diastereomeric 1,4-benzodiazepine-2,5-diones provide
evidence for a key role of conformational chirality in these enantioselective reactions. Acidic removal of
the DAM group is fast and high-yielding and can be performed selectively in the presence of a N-Boc
indole. Thus the synthesis of quaternary benzodiazepines with diverse N1 functionality can now be
accomplished.
Scheme 1. Enantioselective Deprotonation/Alkylation of
1,4-Benzodiazepin-2-ones and 1,4-Benzodiazepine-2,5-diones
Introduction
Benzodiazepines are privileged scaffolds in medicinal chem-
istry,1 prompting the development of combinatorial methods for
the synthesis of 1,4-benzodiazepin-2-ones,2 1,4-benzodiazepine-
2,5-diones,3 and related heterocycles.4,5 Since R-amino acids
are the key starting materials that supply the configuration at
C3 in these heterocycles, it is not surprising that benzodiazepines
featuring a quaternary stereogenic center at C3 have received
little attention. To provide a concise synthesis of these com-
pounds we developed enantioselective deprotonation/alkylation
routes to quaternary 1,4-benzodiazepin-2-ones6 and proline-
derived 1,4-benzodiazepine-2,5-diones.7 Treatment of (S)-1a
with base at -78 °C, followed by addition of benzyl bromide,
gives (R)-2a in 97% ee with retention of configuration (Scheme
1). A similar sequence applied to (S)-3a at -100 °C gives
quaternary 1,4-benzodiazepine-2,5-dione (S)-4a in 95% ee. The
high enantioselectivities observed in these reactions are at-
tributed to “memory of chirality”:8-10 deprotonation destroys
the sp3-hybridized stereogenic center of the starting materials
but generates an enantiopure, conformationally chiral enolate
(1) Evans, B. E.; et al. J. Med. Chem. 1988, 31, 2235-2246.
(2) Ellman, J. A. Acc. Chem. Res. 1996, 29, 132-143.
(3) Boojamra, C. G.; Burow, K. M.; Thompson, L. A.; Ellman, J. A. J. Org.
Chem. 1997, 62, 1240-1256.
that racemizes slowly on the alkylation reaction time scale. The
requirement for an i-Pr group at N1 in these reactions clearly
limits the appeal of this methodology, since this group cannot
(4) Herpin, T. F.; Van Kirk, K. G.; Salvino, J. M.; Yu, S. T.; Labaudiniere, R.
F. J. Comb. Chem. 2000, 2, 513-521.
(5) Wu, Z.; Ercole, F.; FitzGerald, M.; Perera, S.; Riley, P.; Campbell, R.;
Pham, Y.; Rea, P.; Sandanayake, S.; Mathieu, M. N.; Bray, A. M.; Ede,
N. J. J. Comb. Chem. 2003, 5, 166-171.
(10) The term chirality in the phrase “memory of chirality” originally coined
by Fuji (Kawabata, T.; Yahiro, K.; Fuji, K. J. Am. Chem. Soc. 1991, 113,
9694-9696) properly refers to the sense of chirality, or configuration.
Chirality is retained in the intermediates whether they racemize or remain
enantiopure; the same is obviously true of the products. Perhaps a better
term would be “memory of configuration”.
(6) Carlier, P. R.; Zhao, H.; DeGuzman, J.; Lam, P. C.-H. J. Am. Chem. Soc.
2003, 125, 11482-11483.
(7) MacQuarrie-Hunter, S.; Carlier, P. R. Org. Lett. 2005, 7, 5305-5308.
(8) Fuji, K.; Kawabata, T. Chem.sEur. J. 1998, 4, 373-376.
(9) Zhao, H.; Hsu, D. C.; Carlier, P. R. Synthesis 2005, 1-16.
9
10.1021/ja0640142 CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 15215-15220
15215

A R T I C L E S
Carlier et al.
Scheme 2. Synthesis of N-DAM-1,4-Benzodiazepin-2-ones
(S)-1d, 5d-8d
be easily removed. However, we have shown that a large N1-
substituent is required for excellent enantioselectivities at -78
°C;6,11 compound (S)-1b, which features a smaller but potentially
replaceable PMB substituent at N1, requires the use of undesir-
able, cryogenic reaction temperatures (-100 to -109 °C) to
achieve adequate enantioselectivity for deprotonation/alkyla-
tion.11 Therefore, to allow the enantioselective synthesis of
quaternary 1,4-benzodiazepin-2-ones and 1,4-benzodiazepine-
2,5-diones that feature diverse N1 substitution, we sought to
employ a bulky (secondary) N1 substituent that could be easily
removed after the enantioselective deprotonation/alkylation step.
In this paper we report that the di-(p-anisyl)methyl (DAM) group
ably fills this role, providing quaternary benzodiazepines in
higher enantioselectivities and yields than previously attained
with the i-Pr and PMB substituents. Following deprotonation/
alkylation, the DAM group can be removed with mild acid,
providing the corresponding N-H quaternary benzodiazepines
that can then be further functionalized at N1.
a 1.2 equiv of NaH (5d-7d) or 1.5 equiv of NaH (1d) was used. 2.0
equiv of NaH was used.
b
Table 1. Enantioselective Deprotonation/Alkylation of Ala-Derived
N-DAM-1,4-Benzodiazepin-2-one (S)-1d at -78 °C
Results and Discussion
We envisioned an ideal N1-substituent for enantioselective
deprotonation/alkylation of 1,4-benzodiazepin-2-ones and 1,4-
benzodiazepine-2,5-diones to have three characteristics. First,
it must have steric bulk comparable to a secondary alkyl group,
so that the derived enolate would not readily racemize at -78
to -100 °C. Second, it must be installed on the N-H precursor
through the use of a highly reactive donor, since we have shown
that reaction at 0 °C within hours is necessary to prevent
racemization during N1 alkylation.6 Last, a nonreductive depro-
tection was required due to the reductive sensitivity of the imine
moiety in 1,4-benzodiazepin-2-ones. After exploration of several
alternatives, only the DAM group filled these requirements.
DAM is a well-known protecting group for â-lactam amides12,13
but has been less commonly used for larger lactams and acyclic
amides.14,15 In our hands, yields of the N-DAM amides proved
critically dependent upon the purity of the DAM-Br.16 The best
procedure involved treatment of the corresponding alcohol
(DAM-OH) with acetyl bromide in dry benzene at room
temperature, removal of the reactants in vacuo, and recrystal-
lization from hexane.
Syntheses of the N-DAM-1,4-benzodiazepin-2-ones (S)-1d,
5d-8d are described in Scheme 2. N-H-1,4-Benzodiazepin-2-
ones 1c, 5c-8c were prepared from the corresponding N-Boc
amino acids and 2-aminobenzophenones by a modification of
Shea’s protocol17 that we have previously described.11 DAM-
Br reacted with the sodium salts of (S)-1c, 5c-8c within 2-4
h at 0 °C to give the corresponding N-DAM derivatives (S)-1d,
5d-8d in good yield and typically >98% ee. Although
deprotonation with 1.2-1.5 equiv of NaH gave good yields
(66-81%), in several cases, the use of 2 equiv of NaH gave
slightly higher yields and had no deleterious impact on
entry
Ea
product
t (h)
% yieldb
% eec
1
2
3
4
5
6
7
8
9
Bn
4-MeC6H4CH2-
N-Boc-indol-3-yl-CH2- (-)-10d
(-)-2d
(-)-9d
4
80 (72) 99 (97) R
6
4
4
2
5
82 (76) >99.5 (94) Rd
96
>99.5
allyl
(+)-11d
(-)-12d
(+)-13d
88 (76) 99 (94)
86e (0) 98 (na) R
65e,f (0) 96 (na)
EtO2CCH2-
Et
-CN
N3
(+)-14d 0.16 95e
(+)-15d 0.5 88
(-)-16d 0.16 94
99
>99.5
>99.5
-N(Boc)-NHBoc
a Electrophiles (equiv) used: BnBr (10), 4-MeC6H4CH2Br (10), N-Boc-
indol-3-ylmethyl bromide (2.5), allyl bromide (10), BrCH2CO2Et (10), EtI
(20), tosyl cyanide (2.0), trisyl azide (2.5), BocNdNBoc (5). b The yield
in parentheses corresponds to that of the N-i-Pr analogue obtained from
(S)-1a.6 c Enantiomeric excess was measured by chiral stationary-phase
HPLC; value in parentheses corresponds to that of the N-i-Pr analogue
obtained from (S)-1a.6 Where indicated, absolute stereochemistry was
determined by correlation; retentive substitution is assumed by analogy in
other cases. d (R) Stereochemistry of (-)-9d was deduced on the basis of
similar sign of rotation and HPLC elution order compared to (-)-2d. e The
benzodiazepine enolate was added to the electrophile. f LDA provided
superior yields of 13d and was used in place of KHMDS.
enantiopurity of the N-DAM product. Note that benzhydryl
bromide proved unreactive toward the sodium salt of 1c under
these conditions, even in the presence of HMPA.
Use of our original protocol (LDA/n-BuLi/HMPA/THF) for
deprotonation/alkylation of (S)-1d gave the desired products,
but only in fair to moderate yields. After examination of a range
of bases, the KHMDS protocol employed for N-PMB derivative
1b11 was found to provide both operational simplicity and
superior yields. Results with alanine- (Ala-) derived 1,4-
benzodiazepin-2-one (S)-1d are shown in Table 1.
Treatment of (S)-1d with KHMDS for 30 min in THF/HMPA
at -78 °C produces an enolate that reacts well with a range of
active carbon electrophiles (entries 1-4). Compared to our
published deprotonation/alkylations of N-i-Pr analogue (S)-1a,
both yields and enantioselectivities are significantly improved
(cf. entries 1, 2, and 4). Enantioselectivities exceed 98%, and
chemical yields range from 80% to 96%. The advantages of
the DAM group are also evident in reaction with ethyl
(11) Carlier, P. R.; Lam, P. C.-H.; DeGuzman, J.; Zhao, H. Tetrahedron:
Asymmetry 2005, 16, 2998-3002.
(12) Kawabata, T.; Kimura, Y.; Ito, Y.; Terashima, S. Tetrahedron Lett. 1986,
27, 6241-6244.
(13) Kobayashi, Y.; Takemoto, Y.; Ito, Y.; Terashima, S. Tetrahedron Lett. 1990,
31, 3031-3034.
(14) AbuSbeith, K.; Bleasdale, C.; Golding, B. T.; Kitson, S. L. Tetrahedron
Lett. 1992, 33, 4807-4810.
(15) Henneuse, C.; Boxus, T.; Tesolin, L.; Pantano, G.; Marchand-Brynaert, J.
Synthesis 1996, 495-501.
(16) Kice, J. L.; Rudzinski, J. J. J. Org. Chem. 1987, 52, 2414-2421.
(17) Hart, B. R.; Rush, D. J.; Shea, K. J. J. Am. Chem. Soc. 2000, 122, 460-
465.
9
15216 J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006

Enantioselective Synthesis of Quaternary Benzodiazepines
A R T I C L E S
Table 2. Enantioselective Deprotonation/Alkylation Reactions of
Ala-, Phe-, Abu-, and Met-Derived N-DAM-1,4-benzodiazepin-2-
ones (S)-5d-8d at -42 °C
bromoacetate and ethyl iodide (entries 5-6), where the corre-
sponding alkylations of the N-i-Pr analogue (S)-1a failed. In
these cases, inverse addition of the enolate to the electrophile
also proved important. Reaction with the sp-hybridized carbon
electrophile tosyl cyanide18 introduces the CN group in 95%
yield and 99% ee (entry 7). Racemic glycine-derived 3-cyano-
1,4-benzodiazepin-2-ones have been evaluated as antianxiety
agents,19 but to the best of our knowledge, quaternary 3-cyano-
1,4-benzodiazepin-2-ones are heretofore unknown. Similarly,
nitrogen functionalization at C3 is quite common in 1,4-
benzodiazepin-2-one cholecystokinin antagonists,1 Ras farne-
sylation inhibitors,20 and antiarrhythmic agents,21 but quaternary
examples have not been described. Knowing that the potassium
enolate of a glycine-derived 1,4-benzodiazepin-2-one undergoes
azidation21 with trisyl azide,22 we explored use of this reagent.
As can be seen in Table 1 (entry 8), excellent yield and
enantioselectivity were obtained. Furthermore, reaction of the
potassium enolate with of (S)-1d with di-tert-butyl azodicar-
boxylate23 gave the corresponding protected hydrazine in
excellent yield and enantioselectivity (Table 1, entry 9). We
propose that the stereochemistry of the deprotonation/alkylation
reactions depicted in Table 1 is uniformly retentive, based on
correlation of (-)-2d, (-)-9d, and (-)-12d to the corresponding
quaternary amino acids (vide infra).
entry
R1
R2
Ea
product
HMPA (n equiv) % yield
% eeb
1
2
3
4
5
6
7
8
9
Me
Me
Me
Bn
Bn
Bn
Bn
Bn
Et
H
H
H
Bn
Bn
Bn
(-)-17d
(-)-17d
(-)-17d
(+)-2d
(+)-2d
(+)-2d
6
6
0
6
6
0
6
6
6
6
6
6
6
6
78
98 Rc
11d,e nd
19e
79
nd
>99.5 S
Cl Me
Cl Me
Cl Me
17d,e nd
23e
58
68
65
58
86
67
33
80
nd
92
Cl allyl (+)-18d
Cl -CN (+)-19d
96
Cl Bn
(-)-20d
94 Rf
94
10 Et
11 Et
12 MeSCH2CH2- Cl Me
13 MeSCH2CH2- Cl allyl (-)-24d
14 MeSCH2CH2- Cl -CN (-)-25d
Cl allylg (-)-21d
Cl -CN (+)-22d
96
87
72
87
(-)-23d
a Inverse addition of the enolate to the electrophile (equiv): MeI (10),
BnBr (10), allyl bromide (10), tosyl cyanide (2.0). b Enantiomeric excess
was measured by chiral stationary-phase HPLC; nd ) not determined. Where
indicated, absolute stereochemistry was determined by correlation; in all
other cases, retentive substitution is assumed by analogy. c (R) Stereochem-
istry of (-)-17d was deduced on the basis of similar sign of rotation and
HPLC elution order compared to chloro-substituted analogue (-)-2d.
d Reaction with only 1.2 equiv KHMDS. e Reaction time was identical to
that of entries 1 and 4 (1.5 h). f (R) Stereochemistry of (-)-20d was deduced
on the basis of similar sign of rotation and HPLC elution order compared
to Ala derivative (-)-2d. g Reaction was performed with allyl iodide, which
proved superior to allyl bromide in this case.
Extension of this protocol to 1,4-benzodiazepin-2-ones de-
rived from other amino acids was then performed. Unexpectedly,
yields with phenylalanine- (Phe-), aminobutyric acid- (Abu-),
and methionine- (Met-) derived 1,4-benzodiazepin-2-ones were
low, suggesting that steric hindrance in the enolate was slowing
alkylation. After considerable experimentation we found that
preparation of the enolates in DME at -42 °C (acetonitrile/
CO2) and inverse addition to the electrophile gave acceptable
yields without significantly reducing enantioselectivity (Table
2).
need to incorporate HMPA in alkylations of lithium enolates
of 1,4-benzodiazepin-2-ones.6
As a first example of this modified protocol we present Ala
derivative (S)-5d, which lacks the 7-chloro substituent present
in 1d and all our previously published substrates.6,11,24 With
benzyl bromide as the electrophile, (S)-5d gave the desired
product in 78% yield and 98% ee (Table 2, entry 1; cf. Table
1, entry 1 for reaction of 1d). Thus the chloro substituent in 1d
is not required for good enantioselectivity or yield. The need
for excess KHMDS in these reactions can be seen from
comparing entries 1 and 2 and entries 4 and 5: when the amount
of KHMDS is decreased from 2.5 to 1.2 equiv, very low
consumption of starting material is observed and the yields of
17d and 2d decrease significantly.25 A similar outcome is
observed when HMPA is omitted from the reaction (cf. entries
1 and 3 and entries 4 and 6); we have previously reported the
The enolate generated from Phe-derived 1,4-benzodiazepin-
2-one (S)-6d reacts at -42 °C in DME with methyl iodide, allyl
bromide, and tosyl cyanide in moderate to good yields with
excellent enantioselectivity (92->99.5% ee; Table 2, entries
4, 7, and 8). Similarly good results are obtained for benzylation,
allylation, and cyanation of the enolate generated from Abu-
derived benzodiazepine (S)-7d (Table 2, entries 9-11). Finally,
for reasons unknown, reactions of the Met-derivative (S)-8d
proceed with somewhat lower enantioselectivity (72-87% ee;
Table 2, entries 12-14).
A major criterion for selection of the DAM group was
removability. The N-DAM group has been removed by oxida-
tion,12,14 reduction,26 and acidic hydrolysis;13 we chose the latter
mode and realized excellent yields and selectivity (Table 3).
Particularly noteworthy is the deprotection of (-)-10d in
1.25% TFA in CH2Cl2 (1.5 h, room temperature), which
provides the desired (-)-10c in quantitiative yield with no
concurrent deprotection of the N-Boc group on the indole.
Hydrolysis of N-H quaternary 1,4-benzodiazepin-2-ones (+)-
2c, (+)-9c, and (-)-12c to the corresponding quaternary amino
acids proceeds under milder conditions (6 M HCl, 125 °C, 2
days) than previously reported for hydrolysis of the N-i-Pr
analogues. Quaternary amino acids (R)-(-)-R-Me-Phe-OH 26,
(18) Kahne, D.; Collum, D. B. Tetrahedron Lett. 1981, 22, 5011-5014.
(19) Ogata, M.; Matsumoto, H.; Hirose, K. J. Med. Chem. 1977, 20, 776-781.
(20) James, G. L.; Goldstein, J. L.; Brown, M. S.; Rawson, T. E.; Somers, T.
C.; McDowell, R. S.; Crowley, C. W.; Lucas, B. K.; Levinson, A. D.;
Masters, J. C. Science 1993, 260, 1937-1942.
(21) Selnick, H. G.; et al. J. Med. Chem. 1997, 40, 3865-3868.
(22) Evans, D. A.; Britton, T. C.; Ellman, J. A.; Dorow, R. A. J. Am. Chem.
Soc. 1990, 112, 4011-4030.
(23) Evans, D. A.; Britton, T. C.; Dorow, R. L.; Dellaria, J. F. J. Am. Chem.
Soc. 1986, 108, 6395-6397.
(24) The chloro substituent in these 1,4-benzodiazepin-2-ones derives from the
2-amino-5-chlorobenzophenone starting material, a commercially available
intermediate used in the production of diazepam.
(25) One possible explanation for the need for excess base is that alkylation
occurs more quickly on a KHMDS/potassium enolate mixed aggregate than
it does on homomeric potassium enolates.
(26) Kobayashi, Y.; Ito, Y.; Terashima, S. Bull. Chem. Soc. Jpn. 1989, 62, 3041-
3042.
9
J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006 15217

A R T I C L E S
Carlier et al.
Table 3. Facile Acidic Removal of the N-DAM Group
Table 4. Conformer Ratios for N-DAM-Substituted Quaternary
1,4-Benzodiazepin-2-ones
% TFA
−
CH2
%
%
eea
entry substrate
R1
E
Cl2 (v/v)
t (h)
yield
1
2
3
(-)-2d Me Bn
25
6.25
1.25
1
1.5
1.5
96 >99.5 R
98 99 R
99 >99.5
mol % R1
equatorial
(-)-9d Me 4-MeC6H4CH2-
(-)-10d Me N-Boc-indol-3-yl
-CH2-
entry
compd
R1
E
R2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
2d
17d
9d
Me
Me
Me
Me
Me
Me
Me
Me
Bn
Et
Bn
Bn
Cl
H
Cl
36
40
39
20
46
4
5
6
7
(+)-11d Me allyl
6.25
6.25
25
1.0
1.5
0.75 94
1.5 99
97
98
98
(-)-12d Me EtO2CCH2-
(-)-20d Et Bn
99 R
98 R
94
4-Me-C6H4CH2-
N-Boc-indol-3-yl-CH2- Cl
10d
11d
12d
13d
16d
18d
20d
21d
23d
24d
15d
14d
19d
22d
25d
(-)-21d Et allyl
6.25
allyl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
EtO2CCH2-
33
a Absolute stereochemistry was determined as listed in Tables 1 and 2.
Et
50
-N(Boc)NHBoc
nda
65
allyl
Bn
allyl
30
40
65
60
(R)-(-)-R-Me-4-Me-Phe-OH 27, and (R)-(-)-R-Me-aspartic
acid 28 were obtained in 72%, 76%, and 85% yield, respectively,
confirming the absolute stereochemistry depicted in Table 1.
The conformational behavior of the quaternary N-DAM- and
Et
CH3SCH2CH2- Me
CH3SCH2CH2- allyl
Me
Me
Bn
Et
N3
70
-CN
-CN
-CN
>99
>99
>99
>99
1
N-H-1,4-benzodiazepin-2-ones, as revealed by H NMR spec-
troscopy, is worthy of comment. Except for the 3-cyano
derivatives, all the quaternary N-DAM-1,4-benzodiazepin-2-ones
appear in 1H NMR spectra as mixtures of R1-pseudoequatorial
and R1-pseudoaxial conformers in slow exchange; we have
previously noted this behavior for quaternary N-Me, N-i-Pr, and
N-PMB analogues.6,11,27 Assignment of the individual conform-
ers is straightforward, based on shielding of the pseudoaxial
group by the benzo ring. Conformer ratios for the N-DAM-
1,4-benzodiazepin-2-ones described in this work are listed in
Table 4.
CH3SCH2CH2- -CN
a nd, not determined due to overlapping resonances.
Scheme 3. Barriers for Conformational Inversion in
Glycine-Derived 1,4-Benzodiazepin-2-ones 29c and 29e27 a
In most cases (entries 1-14), a near 50:50 mixture of
conformers is seen, indicating similar steric demands of the R1
and E substituents. In this context it may appear surprising that
the corresponding N-H quaternary 1,4-benzodiazepin-2-ones
(e.g., 2c, 9c-12c, and 20c, Table 3) present only one set of
resonances in their 1H NMR spectra. However, we attribute the
spectral simplicity of these compounds to averaging caused by
rapid interconversion between the diastereomeric conformers.
This proposal is supported by the experimentally determined
inversion barriers of glycine-derived benzodiazepines 29c and
29e (Scheme 3).
a (M) and (P) helical descriptors of configuration are assigned on the
basis of the sign of the 2-1-7-6 dihedral angle.28
1H NMR spectroscopy (Table 4, entries 15-18). Is this
preference electronic or steric in nature? A strong preference
for the CN-axial conformer in 2-cyanopiperidine has been
attributed to the anomeric effect.30 The anomeric effect may
play some role in the conformational preference of 14d, 19d,
22d, and 25d, but we believe the strong preference for the CN-
axial conformer in these compounds principally derives from
the small size of the cyano group relative to that of R1. For
reference, the equatorial preferences for CN and methyl groups
in a cyclohexane ring are 0.20 and 1.74 kcal/mol, respectively.31
To determine whether the DAM group would be effective
for enantioselective deprotonation/alkylation of 1,4-benzodiaz-
epine-2,5-diones, several substrates derived from proline, thio-
proline, cis-4-hydroxy-D-proline, and trans-4-hydroxy-L-proline
were prepared (Scheme 4). N-H-1,4-Benzodiazepine-2,5-diones
3c, 30c, 31c, and 32c were prepared according to literature
methods32-34 by heating isatoic anhydride with the appropriate
At room temperature, N-H-benzodiazepine 29c presents a
single, averaged resonance for the pseudoequatorial and pseudo-
1
axial CH2 protons in its H NMR spectrum, indicating rapid
interconversion of the enantiomeric (M) and (P) conformers;
decoalescence of these protons is observed at 253 K.27,29 In
contrast, N-benzhydrylbenzodiazepine 29e displays discrete
signals for the pseudoequatorial and pseudoaxial CH2 protons,
which do not coalesce until 458 K.27 Returning to the N-DAM-
3-cyano-substituted benzodiazepines 14d, 19d, 22d, and 25d,
these appear as single, R1-equatorial/CN-axial conformers by
(27) Lam, P. C.-H.; Carlier, P. R. J. Org. Chem. 2005, 70, 1530-1538.
(28) This dihedral angle has the same sign as the C2-C3-N4-C5 dihedral,
which has historically (Konowal, A.; Snatzke, G.; Alebic-Kolbah, T.;
Kajfez, F.; Rendic, S.; Sunjic, V. Biochem. Pharm. 1979, 28, 3109-3113)
been used to designate the sense of conformational chirality of the
benzodiazepine ring.
(30) Booth, H.; Dixon, J. M.; Khedhair, K. A. Tetrahedron 1992, 48, 6161-
6174.
(31) Eliel, E. L.; Wilen, S. H. In Stereochemistry of Organic Molecules; John
Wiley & Sons: New York, 1994; pp 696-697.
(32) Wright, W. B.; Brabander, H. J.; Greenblatt, E. N.; Day, I. P.; Hardy, R.
A., Jr. J. Med. Chem. 1978, 21, 1087-1089.
(29) Linscheid, P.; Lehn, J.-M. Bull. Chim. Soc. Fr. 1967, 992-997.
9
15218 J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006

Enantioselective Synthesis of Quaternary Benzodiazepines
A R T I C L E S
Scheme 4. Synthesis of N-DAM-1,4-Benzodiazepin-2-ones
Derived from (S)-Proline, (S)-Thioproline, cis-4-Hydroxy-D-proline,
and trans-4-Hydroxy-L-prolinea
Table 6. Deprotonation/Benzylation Reactions of Diastereomeric
Substrates 33d and 34d Evidence Conformational Memory at
-100 °C but Not at -78 °C
entry
starting material
T (
°
C)
% yield (41d
+
42d)
ratio 41d:42da
1
2
3
4
33d
34d
33d
34d
-100
-100
-78
95
49
87
82
96:4
8:92
69:31
66:34
-78
a Diastereomer ratios measured by chiral stationary-phase HPLC and
1
confirmed by H NMR.
of the enolate is faster than intermolecular alkylation at this
higher temperature. This -100 °C protocol was then applied
to other alkylations of (S)-3d and thioproline-derived (S)-30d.
In all cases, excellent yields and enantioselectivities were
obtained (Table 5, entries 3-11). It is important to note that
addition of HMPA is not necessary to achieve high yields of
the alkylated products from 3d and 30d (Table 5, cf. entries 1
and 3, 4 and 5, 6 and 7, and 8 and 9); in contrast, HMPA is
required for acceptable yields in deprotonation/alkylation of 1,4-
benzodiazepin-2-ones (cf. Tables 1 and 2). All the quaternary
1,4-benzodiazepine-2,5-diones derived from proline and thio-
proline appear as single conformers by 1H NMR spectroscopy.
We attribute this conformational homogeneity to an amide
resonance-imposed requirement for the proline moiety to occupy
a pseudoequatorial orientation on the benzodiazepine ring.7
Removal of the DAM group from (+)-35d, (+)-37d, and (+)-
38d by treatment with 25% TFA in CH2Cl2 is complete within
3 h at room temperature, affording N-H analogues (+)-35c, (-)-
37c, and (+)-38c in 99%, 99%, and 98% yields, respectively.
Retentive transformation of (S)-(+)-3d to (R)-(+)-35d was
proven by hydrolysis of (+)-35c to (R)-(-)-R-benzylproline.35
The diastereomeric substrates 33d and 34d derived from cis-
4-hydroxy-D-proline and trans-4-hydroxy-L-proline feature iden-
tical (R) configuration at the oxygen-bearing stereocenter C2
(pyrrolo[2,1-c][1,4]benzodiazepine numbering) but opposite
configuration at the enolizable carbon C11a. As such, these
substrates can provide another test for the presence or absence
of conformational memory effects in the derived enolates; a
similar strategy has been used to illustrate the importance of
conformational chirality in alkylations of amino acid ester
enolates.36,37 The demonstrated preference for retentive substitu-
tion of 1,4-benzodiazepine-2,5-diones leads to the prediction
that deprotonation/alkylation of 33d would give 41d, whereas
34d would give 42d. However, in the absence of conformational
memory effects both substrates would produce the same enolate
(or mixture of enolates), and thus give the same ratio of products
41d and 42d. In the event, deprotonation/benzylation of 33d at
-100 °C gave a 96:4 ratio of 41d:42d, as expected, favoring
the retentive product (Table 6, entry 1).
a Reagents and conditions: (i) THF, NaH, 0 °C; DAM-Br. (ii) TBDMS-
Cl, imidazole, DMF. Atom numbering is according to standard IUPAC
pyrrolo[2,1-c][1,4]benzodiazepine convention.
Table 5. Enantioselective Alkylation of N-DAM-1,4-Benzodia-
zepine-2,5-diones (S)-3d and (S)-30d Derived from Proline and
Thioproline
entry
A
E
additive
product
% yield
% eea
1
2
3
4
5
6
7
8
9
CH2 Bn
CH2 Bn
CH2 Bn
HMPAb (+)-35d
HMPAb (()-35d
98
90
83
93
83
94
87
65
90
98
89
98 R
0 (-78 °C)c
99 R
na
(+)-35d
CH2 4-Me-C6H4CH2- HMPAb (+)-36d
>99.5 Rd
99 Rd
99 Rd
94 Rd
93
92
99
95
CH2 4-Me-C6H4CH2- na
CH2 2-Ph-C6H4CH2- HMPAb (+)-37d
CH2 2-Ph-C6H4CH2- na
CH2 allyl
CH2 allyl
(+)-36d
(+)-37d
HMPAb (+)-38d
na
na
na
(+)-38d
(+)-39d
(+)-40d
10
11
S
S
Bn
allyl
a Enantiomeric excess measured by HPLC. Where indicated, absolute
stereochemistry was determined by correlation; in all other cases, retentive
substitution is assumed. b Six equivalents of HMPA was added. c Reaction
was carried out at -78 °C. d (R) Stereochemistry was assigned on the basis
of polarimetry and HPLC elution order relative to the simple benzyl
analogue.
proline derivative. Installation of the DAM group proceeded in
excellent yield; in the case of 31c and 32c, protection of the
hydroxyl group was carried out prior to N1-functionalization.
On the basis of our previous studies with N-i-Pr analogue
(S)-3a,7 we anticipated the need for cryogenic reaction temper-
atures and an in situ protocol to prevent racemization of the
enolate derived from (S)-3d. In the presence of 10 equiv of
benzyl bromide at -100 °C, (S)-3d reacts with KHMDS to give
the desired benzylated product (+)-35d in 98% yield and 98%
ee (Table 5, entry 1).
Similarly, 34d gave an 8:92 ratio of 41d and 42d, again
favoring the retentive product (Table 6, entry 2). Interestingly,
Interestingly, when the identical reaction is performed at -78
°C, racemic product is obtained, indicating that racemization
(35) Seebach, D.; Boes, M.; Naef, R.; Schweizer, W. B. J. Am. Chem. Soc.
1983, 105, 5390-5398.
(33) Pen˜a, M. R.; Stille, J. K. J. Am. Chem. Soc. 1989, 111, 5417-5424.
(34) Dyatkin, A. B.; Maryanoff, B. E. Org. Prep. Proc. Int. 2002, 34, 652-
655.
(36) Kawabata, T.; Chen, J.; Suzuki, H.; Nagae, Y.; Kinoshita, T.; Chancharunee,
S.; Fuji, K. Org. Lett. 2000, 2, 3883-3885.
(37) Kawabata, T.; Chen, J.; Suzuki, H.; Fuji, K. Synthesis 2005, 1368-1377.
9
J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006 15219

A R T I C L E S
Carlier et al.
Scheme 5. Proposed Formation of Enolate Conformational
Diastereomers from 33d and 34da
ratios obtained from 33d and 34d under these conditions suggest
that alkylation is occurring on a nearly equilibrated mixture of
the (M) and (P) conformers of (2R)-43d. We similarly attribute
the formation of the minor product diastereomers at -100 °C
to a small amount of stereochemical leakage (i.e., (M)-(P)
conformational interconversion).
In conclusion, use of the DAM group at N1 of 1,4-
benzodiazepin-2-ones and proline-derived 1,4-benzodiazepine-
2,5-diones allows construction of quaternary stereogenic centers
in high enantioselectivity via deprotonation/alkylation. The large
steric bulk of the DAM group imparts a significant barrier to
racemization of the conformationally chiral enolate intermedi-
ates, allowing reactions of N-DAM-1,4-benzodiazepin-2-ones
to be carried out at temperatures as high as -42 °C. In contrast,
deprotonation/alkylation reactions of proline-derived N-DAM-
1,4-benzodiazepine-2,5-diones must be carried out at -100 °C,
apparently due to a lower racemization barrier of the derived
enolates.38 Following this retentive substitution process, the
DAM group is easily removed by acidic hydrolysis, thereby
allowing the subsequent installation of diverse N1 functionality.
a (M) and (P) helical descriptors of configuration are assigned on the
basis of the sign of the 11-10-13-12 dihedral angle (IUPAC pyrrolo[2,1-
c][1,4]benzodiazepine numbering).
when the same reactions were carried out at -78 °C, 33d and
34d gave similar product ratios, near 2:1 in favor of 41d (Table
6, entries 3 and 4). To rationalize these divergent results, the
simplest explanation is that the deprotonations of 33d and 34d
give conformationally diastereomeric enolates. On the basis of
the stereoelectronic requirements for deprotonation, and calcula-
tions on the N-i-Pr analogue (S)-3a,7 we propose that deproto-
nation of (11aR)-configured 33d gives an enolate in the (P)
conformation, and deprotonation of (11aS)-configured 34d gives
an enolate in the (M) conformation (Scheme 5).
Acknowledgment. We thank the National Science Foundation
(CHE-021325) and the Virginia Tech Department of Chemistry
for financial support.
Supporting Information Available: Experimental procedures,
characterization data, 1H NMR spectra, and complete refs 1 and
22. This material is available free of charge via the Internet at
http://pubs.acs.org.
JA0640142
At -100 °C, conformational interconversion of (P)-(2R)-43d
and (M)-(2R)-43d is slow relative to intermolecular alkylation,
and thus retentive substitution predominates (Table 6, entries 1
and 2). However, at -78 °C, conformational interconversion is
fast relative to intermolecular alkylation. The similar 41d:42d
(38) Consistent with this analysis, preliminary deprotonation/trapping studies
of N-i-Pr-1,4-benzodiazepin-2-one (S)-1d indicate the potassium enolate
has a racemization barrier of 18 kcal/mol (258 K). In contrast, the same
technique indicates only a 12.5 kcal/mol barrier (173 K) for the potassium
enolate derived from N-i-Pr-1,4-benzodiazepine-2,5-dione (S)-3d (P. C.-
H. Lam, S. L. MacQuarrie-Hunter, and P. R. Carlier, unpublished work).
9
15220 J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006