Synthesis, resolution, and absolute configuration of novel tricyclic benzodiazepines

Article abstract of DOI:10.1016/j.tetasy.2004.02.028

The synthesis, resolution, and stereochemical characterization of novel piperazino-, thiazino-, and oxazinobenzodiazepines are described. The absolute stereochemistry of the heterotricycles was determined by using X-ray crystallography and the enantiomeri

Full text of DOI:10.1016/j.tetasy.2004.02.028

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 15 (2004) 1259–1267
Synthesis, resolution, and absolute configuration
of novel tricyclic benzodiazepines
Jay M. Matthews,a,* Alexey B. Dyatkin,a Mary Evangelisto,b Diane A. Gauthier,a
Leonard R. Hecker,a William J. Hoekstra,a Fuqiang Liu,b Brenda L. Poulter,a
Kirk L. Sorgib and Bruce E. Maryanoffa
aJohnson & Johnson Pharmaceutical Research & Development LLC, Spring House, PA 19477-0776, USA
bJohnson & Johnson Pharmaceutical Research & Development LLC, Raritan, NJ 08869, USA
Received 18 November 2003; revised 21 January 2004; accepted 27 February 2004
Abstract—The synthesis, resolution, and stereochemical characterization of novel piperazino-, thiazino-, and oxazinobenzodiaze-
pines are described. The absolute stereochemistry of the heterotricycles was determined by using X-ray crystallography and the
enantiomeric purity was determined by using Pirkle-solvent NMR techniques and chiral HPLC.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1
1. Introduction
determined by H NMR with Pirkle’s solvent7 and by
chiral HPLC. The absolute configurations of the pipe-
razino- and thiazinobenzodiazepines were determined
by X-ray crystallography, while the absolute configu-
ration of the oxazinobenzodiazepine was determined by
vibrational circular dichroism (VCD), as reported ear-
lier.8
The benzodiazepine chemotype has been found in
numerous biologically active compounds, such as opiate
antagonists, anti-inflammatory agents, psychomotor
depressants, vasopressin receptor antagonists, and
antibiotics.1–3 This quality of the benzodiazepine ring
system has engendered its classification as a ‘privileged
structure’ for drug discovery.4 Although the synthesis
of various heterocyclo[1,4]benzodiazepines has been
reported,5 there are no publications on the synthesis of
piperazino-, thiazino-, or oxazinobenzodiazepines, such
as 1a–c. Recently, we discovered compounds based on a
racemic benzodiazepine backbone with exciting biolo-
gical activity.6 Our biological evaluation and structure–
activity relationship (SAR) development was expedited
by the synthesis of racemic products; however, we ulti-
mately needed to assess the biological properties of the
individual enantiomers to define the potency, pharma-
cology, and toxicology. While an enantioselective syn-
thesis of each enantiomer would be extremely helpful to
this end, the most feasible route appeared to be a clas-
sical resolution of each benzodiazepine. Herein, we
describe the synthesis, resolution, and stereochemical
characterization of benzodiazepine tricycles 1a–c. The
enantiomeric purity of each resolved tricycle was
N
X
a: X = NMe
b: X = S
N
H
c: X = O
1
2. Results and discussion
The syntheses of the respective benzodiazepines are
outlined in Schemes 1 and 2. Methyl 4-methylpiper-
azine-2-carboxylate dihydrochloride 5a was prepared as
shown in Scheme 1. Bis-protected piperazine 2 was
prepared from piperazine-2-carboxylic acid according to
literature precedent.9 Deprotection of
2 provided
piperazine intermediate 3, which was reductively ami-
nated with aqueous formaldehyde to provide N-methyl
intermediate 4, in 96% overall yield for the two steps.
Catalytic hydrogenolysis of 4 provided racemic methyl
piperazate 5a in high yield.
* Corresponding author. Fax: +1-215-628-3297; e-mail: jmatthew@
prdus.jnj.com
0957-4166/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2004.02.028

1260
Boc
J. M. Matthews et al. / Tetrahedron: Asymmetry 15 (2004) 1259–1267
3
X
H
N
X
CH Cl , Et N
N
2
2
3
TFA
DCE
1
6/7
5
N
CO R
2
N
H
CO R
O
NaB(OAc) H
3
N
Z
CO2Me
2
CH Cl
2
2
N
Z
CO2Me
10
O
37% HCHO
99%
Cl
86%
5
3
2
O N
a: X = NMe, R = Me
b: X = S, R = Et
c: X = O, R = Me
2
NO
8
2
6
Me
N
Me
N
68-78%
MeOH
2 HCl
O
Pd-C (10%)
N
Z
CO2Me
N
H
CO2Me
LiAlH
Fe
N
4
X
HCl, H
2
AcOH
53-89%
4
5a
THF, 65°C
70-99%
96%
N
H
O
Scheme 1. Synthesis of methyl piperazate 5a.
7
Methyl thiomorpholine carboxylate 5b10 and morpho-
lino amino ester 5c11 (both racemic) were prepared
according to literature procedures. As shown in Scheme
2, acylation of amino esters 5a–c with 2-nitrobenzoyl
chloride provided amides 6a–c, which were subsequently
cyclized to lactams 7a–c using iron-mediated condi-
tions.12 Hydride reduction of lactams afforded racemic
benzodiazepines 1a–c. To identify a suitable resolving
agent/solvent for providing single enantiomers of race-
mates 1a–c, we initially tested several enantiomerically
pure acids and solvent conditions as part of a 5 · 10
screening matrix (Table 1) with each heterocycle, 1a–c.
Individual reactions that resulted in crystals were filtered
and cracked with aqueous base to separate the enan-
tiomerically enriched free base from the resolving agent.
After finding an effective acid for the resolution (see
below), each opposite enantiomer was obtained by
crystallization with the opposite, applicable chiral acid.
N
N
X
X
resolution
(see text)
H
N
H
N
H
(S)-(+)-1
1
Scheme 2. Synthesis and resolution of 1a–c.
on H NMR with chiral solvating agents.7 In the NMR
studies we employed Pirkle’s solvent, 2,2,2-trifluoro-1-
(9-anthryl)ethanol (TFAE). Figure 1 (panel A) depicts
the spectra of racemate 1a in the absence of TFAE. The
aromatic triplet between d 6.81–6.86 (identified through
COSY and NOESY experiments) is the signal that we
focused on to determine the initial enantiomeric purity.
Figure 1 (panel B) shows racemate 1a in the presence of
0.45% TFAE in CDCl3. As shown in Figure 1 (panel B),
an approximate enantiomeric shift of 0.2 ppm and a pair
of multiplets was created on addition of 0.45% TFAE
1
Our initial approach for measuring the ees of tricycles
1a–c, and their corresponding enantiomers, was based
Table 1. Resolution of racemates 1a–c
Substrate
Salta
Solventb
Recoveryc (%)
Eed (%)
Configuration
1a
1a
1b
1b
1c
1c
Dibenzoyl-
Dibenzoyl-
(R)-BNPH
(S)-BNPH
D
-tartaric acid
-tartaric acid
Ethanol
Ethanol
96
78
70
72
52
64
P98e
P98e
P98e
94.8
(S)-(þ)-1af
(R)-(ꢀ)-1ag
(S)-(þ)-1bf
(R)-(ꢀ)-1bg
(S)-(þ)-1ch
(R)-(ꢀ)-1cg
L
Methanol
Methanol
Di-p-toluoyl-
D
-tartaric acid
-tartaric acid
Methanol–ether
Methanol–ether
98.3
95.4
Di-p-toluoyl-
L
a The acid–substrate ratio was 1:1, except for 1a, where a 1:2 ratio was utilized. BNHP¼binaphthyl-2,20-diyl hydrogen phosphate. Chiral acids
employed in the screening 5 · 10 crystallization array were (R)-2-hydroxy-5,5-dimethyl-4-phenyl-1,2,3-dioxaphosphorane-2-oxide, (R)-(ꢀ)-4-chlo-
romandelic acid, di-p-tolyl-
(S)-(þ)-2-phenylbutyric acid, (R)-BNHP, and (R)-(ꢀ)-a-methoxy-a-trifluoromethylphenylacetic acid.
D
-tartaric acid, (R)-(ꢀ)-mandelic acid, (1S)-(þ)-10-camphorsulfonic acid,
D-tartaric acid, dibenzoyl-D-tartaric acid,
b Solvents employed in the screening 5 · 10 crystallization array were MeOH, EtOH, i-PrOH, EtOAc, and THF (or a combination).
c The recovery is based on a possible 50% of the desired enantiomer after one recrystallization. Evaporation of the mother liquor followed by multiple
recrystallizations provided higher yields.
d The enantiomeric excess (ee) after one crystallization was determined by chiral HPLC and 1H NMR by utilizing 0.45% (S)-(þ)-2,2,2-trifluoro-1-(9-
anthryl)ethanol (Pirkle’s solvent) as an additive in CDCl3. The enantiomeric purities were determined by using a Hewlett Packard 1100 system
(Chiralcel AS column, 0.46 cm · 25 cm; mobile phase 85:15 hexanes:i-PrOH containing 0.1% diethylamine; detection at 254 nM).
e
1
The opposite enantiomer was not detected using chiral HPLC or H NMR by utilizing 0.45% (S)-(þ)-2,2,2-trifluoro-1-(9-anthryl)ethanol (Pirkle’s
solvent) as an additive in CDCl3.
f The absolute configuration was determined by X-ray crystallography.
g The absolute configuration was assigned by comparison of the sign of its specific rotation to that of the opposite enantiomer.
h The absolute configuration was assigned based on VCD analysis.8

J. M. Matthews et al. / Tetrahedron: Asymmetry 15 (2004) 1259–1267
1261
Figure 1. Downfield regions of three 1H NMR spectra. Panel A: Racemate 1a with no TFAE. Panel B: Racemate 1a with 0.45% TFAE. Panel C:
(S)-(þ)-1a with 0.45% TFAE.
to racemate 1a. This separation was deemed to have
enough analytical utility for an adequate ee measurement
tallization, and the enantiomeric excess could be
improved by further crystallization. The opposite
enantiomers, (R)-(ꢀ)-1a, (R)-(ꢀ)-1b, and (R)-(ꢀ)-1c,
were obtained in a similar way by crystallization of salts
from the opposite acid, with ee’s of P98%, 94.8%, and
95.4%, respectively (Table 1).
(ꢁ3%). Crystallization of 1a with dibenzoyl-
D-tartaric
1
acid and basification provided (S)-(þ)-1a, the H NMR
of which with 0.45% TFAE is shown in Figure 1 (panel
C). This material is highly enriched with an ee of greater
than 97% (ꢁ3%), according to the criterion mentioned
above.
To establish the absolute configuration unambiguously,
crystals of (S)-(þ)-1a and (S)-(þ)-1b were grown for an
X-ray crystallographic structure determination. The
molecular structure of (S)-(þ)-1a along with one mole-
Chiral HPLC confirmed this 1H NMR result in that
compound (S)-(þ)-1a was determined to have P98% ee
1
(Table 1). The same type of H NMR and chiral HPLC
cule of dibenzoyl-D-tartaric acid is shown in Figure 2
measurements were performed on racemates 1b and 1c,
as well, and high enantiomeric excesses were obtained
for (S)-(þ)-1b and (S)-(þ)-1c, respectively (Table 1).
(panel A). By reference to the acid subunit, the absolute
configuration of the tricycle is determined to be S.
Figure 2 (panel B) shows the molecular structure of (S)-
(þ)-1b paired with (R)-BNHP. Interestingly, two
equivalents of the chiral acid are coordinated with each
basic nitrogen, which apparently causes (S)-(þ)-1b to
adopt a more puckered conformation than that present
in (S)-(þ)-1a. By reference to the acid subunit, the
absolute configuration of (S)-(þ)-1b is determined to be
S. Although 1c formed a nice crystalline salt with di-p-
As indicated in Table 1, the resolution of racemate 1a
was accomplished efficiently by using dibenzoyl-D-tar-
taric acid in ethanol to provide compound (S)-(þ)-1a in
48% yield with P98% ee. Racemate 1b was resolved to
afford (S)-(þ)-1b through crystallization with (R)-(ꢀ)-
binaphthyl-2,20-diyl hydrogen phosphate (BNHP) from
methanol in 35% yield and an ee of P98%. Compound
(S)-(þ)-1c was obtained in 35% yield and 98.3% ee by
toluoyl-D-tartaric acid for resolution, resulting in (S)-
(þ)-1c with high ee (98.3%), the crystals were initially
not of sufficient quality for an X-ray diffraction study.
Thus, the absolute configuration of (S)-(þ)-1c was first
assigned by comparing the sign of its specific rotation to
crystallization of the di-p-toluoyl-D-tartrate salt of 1c
from methanol-ether. It should be noted that high
enantiopurities were obtained after just a single crys-
A
B
Figure 2. Panel A: ORTEP plot of the molecular structure of (S)-(þ)-1a with dibenzoyl-
D-tartaric acid (with the atom-numbering scheme). Panel B:
Ball-and-stick model of the molecular structure of (S)-(þ)-1b with (R)-binaphthyl-2,20-diyl hydrogen phosphate (with the atom-numbering scheme).
A single molecule of CH3OH of crystallization is shown (C15, O15).

1262
J. M. Matthews et al. / Tetrahedron: Asymmetry 15 (2004) 1259–1267
that for (S)-(þ)-1a and (S)-(þ)-1b. Also, we employed
vibrational circular dichroism (VCD), which has
become a powerful technique for determining the absolute
configuration for structurally rigid compounds in solu-
tion phase, to corroborate the assignment for (S)-(þ)-1c,
as discussed elsewhere.8 Eventually, we were able to
tion of acid-addition salts of (S)-(þ)-1a and (S)-(þ)-1b
established the absolute configurational assignments.
4. Experimental
obtain X-ray diffractable crystals of 1cÆdi-p-toluoyl-D-
tartaric acid and perform the crystallography, the results
of which agreed with the prior stereochemical assign-
ment. Figure 3 shows the molecular structure of (S)-(þ)-
1c.
4.1. General methods
All commercially available chemicals were used as pur-
chased. Melting points were obtained with using a
Mel-Temp capillary melting point apparatus and are
1
uncorrected. H and 13C NMR spectra were recorded
with a Bruker AC-400 spectrometer (unless otherwise
noted) with TMS as an internal standard. Electrospray
ionization mass spectra (ESI) were obtained using a
Fisons spectrometer (Hewlett–Packard HPLC driven
electrospray MS instrument). The purities of each
compound were determined using a Hewlett–Packard
LC 1100 system (YMC J’Sphere H80 S4 column,
4.0 · 50 mm, 4 mm C18; mobile phase of 90% H2O (0.1%
TFA) to 10% H2O (0.1% TFA) with a flow rate of 1 mL/
min; detection at 220 and 254 nM). The enantiopurities
were determined using a Hewlett–Packard LC 1100
system (Chiralcel AS column, 0.46 · 25 cm; mobile phase
of 85/15 hexanes/i-PrOH containing 0.1% diethylamine,
flow rate of 1 mL/min; detection at 254 nM). Optical
rotations were obtained using a Perkin–Elmer 241
polarimeter. Elemental analyses were conducted by
Robertson Microlit Laboratories. X-ray data were
obtained by Crystalytics Company. For (S)-(þ)-1a and
(S)-(þ)-1b, X-ray data were collected on a computer-
controlled Bruker P4 Single Crystal Diffractometer by
using 1.0ꢁ-wide x-scans and graphite-monochromated
MoKa radiation. For (S)-(þ)-1c, X-ray data were col-
lected on a Bruker SMART CCD Single Crystal Dif-
fraction System by using 0.30ꢁ-wide x-scans and
graphite-monochromated MoKa radiation. All of the
structures were solved and refined with the Bruker
SHELXTL-PC (version 5.0) software package.
Figure 3. Ball-and-stick model of the molecular structure of (S)-(þ)-1c
as a salt with di-p-toluoyl-D-tartaric acid, showing only the ammonium
cation (with its atom-numbering scheme).
In the three X-ray structures, the conformations adopted
by ammonium cations 1a and 1c are similar, but that
adopted by cation 1b differs (Figs. 1–3). In cations 1a
and 1c, the saturated six-membered ring is fused to the
tetrahydrobenzazepine ring in a trans configuration with
the proton on the bridgehead nitrogen in an axial ori-
entation. However, in cation 1b, the saturated six-
membered ring is fused to the tetrahydrobenzazepine
ring in a cis configuration with the proton on the
bridgehead nitrogen axial. In all three compounds, the
six-membered ring consistently has a chair conforma-
tion, as expected, and the tetrahydrobenzazepine ring
has a chair-like conformation, with the unprotonated
aniline nitrogen somewhat pyramidalized. The N-methyl
group in doubly protonated 1a is in an equatorial ori-
entation. The aniline nitrogen in 1b is involved in a
hydrogen-bond donor interaction with a carboxylate
oxygen of the anion. The differences observed between
1a/1c and 1b (phosphate-type derivatives are probably
related to the counteranion in the salt form (diaroyl-
tartrate vs binaphthylphosphate) and the attendant
effects on crystal-packing forces.
4.2. Synthesis of 4-methylpiperazine-2-carboxylic acid
methyl ester 5a
4.2.1. Synthesis of piperazine-1,2,4-tricarboxylic acid 1-
benzyl ester 4-tert-butyl ester 2-methyl ester 3. To 29
(21.8 g, 57.6 mmol), dissolved in dichloromethane
(150 mL) was added triflouroacetic acid (150 mL) and
the reaction was stirred at ambient temperature for 3 h
(monitored by HPLC). The solvent was evaporated in
vacuo and the residual oil was partitioned between ethyl
acetate and 10% NaHCO3. The aqueous phase was
extracted with ethyl acetate (3·) and the organic extracts
were combined and washed with 1 N aqueous HCl (2·).
The combined HCl extracts were washed with ethyl
acetate (1·), carefully neutralized to pH 8 with 1 N
aqueous NaOH, and extracted with dichloromethane
(3·). The organic extracts were combined, washed with
H2O and dried over Na2SO4. Filtration and evaporation
of the solvent in vacuo afforded 3 (13.8 g, 86%), as a
viscous oil. 1H NMR (CDCl3) d 2.70–2.89 (m, 1H),
2.91–3.27 (m, 3H), 3.48–3.56 (t, 1H), 3.71–3.77 (d, 3H),
3. Conclusions
The syntheses of novel heterocyclic benzodiazepines
have been executed. After exploring a panel of crystal-
lization conditions (5 · 10 array), the classical resolution
of benzodiazepines 1a–c was accomplished in high yield
and high enantiomeric excess, as demonstrated by Pir-
1
kle-solvent H NMR and chiral HPLC. X-ray diffrac-

J. M. Matthews et al. / Tetrahedron: Asymmetry 15 (2004) 1259–1267
1263
3.87–3.97 (t, 1H), 4.64–4.76 (m, 1H), 5.09–5.29 (m, 2H),
7.32–7.36 (m, 3H); MS, m=z: 279.0 (Mþ1).
0.7H, H2), 2.60–2.63 (d, J ¼ 11:0 Hz, 0.7H, H40), 2.80–
2.83 (d, J ¼ 11:0 Hz, 0.3H, H40), 3.09–3.12 (m, 0.3H,
H5), 3.17–3.20 (m, 0.7H, H5), 3.29–3.31 (d, J ¼ 11:5 Hz,
0.7H, H20), 3.34–3.40 (m, 0.7H, H50), 3.69 (s, 0.3H, H6),
3.78 (s, 0.7H, H6), 4.19 (s, 0.3H, H1), 4.40–4.42 (d,
J ¼ 13:0 Hz, 0.3H, H50), 5.20 (s, 0.7H, H1), 7.34–7.35
(d, J ¼ 6:5 Hz, 0.3H, H11), 7.47–7.49 (d, J ¼ 7:5 Hz,
0.7H, H11), 7.69–7.74 (m, 1H, H9), 7.79–7.80 (t,
J ¼ 7:0 Hz, 0.3H, H10), 7.86–7.89 (t, J ¼ 7:5 Hz, 0.7H,
H10), 8.17–8.19 (m, 1H, H8); HRMS, m=z: 308.1
(Mþ1); acc. mass: (C14H17N3O5) calcd 308.1233, found
308.1238.
4.2.2. Synthesis of 4-methylpiperazine-1,2-dicarboxylic
acid 1-benzyl ester 2-methyl ester 4. To a solution of 3
(15.5 g, 55.6 mmol), dissolved in 1,2-dichloroethane
(200 mL) was added 37% aqueous formaldehyde (5 mL,
178 mmol) followed by sodium triacetoxyborohydride
(16.5 g, 77.8 mmol), in one portion, and the reaction
mixture was stirred at room temperature until total
consumption of 3 (24–72 h, monitored by HPLC). The
reaction was quenched with 1 N aqueous NaOH and
diluted with dichloromethane. The organic layer was
washed with H2O and dried over Na2SO4. The solvent
was filtered and evaporated in vacuo to afford 4 (16.1 g,
4.3.2. 4-(2-Nitrobenzoyl)thiomorpholine-3-carboxylic acid
methyl ester 6b. According to the general procedure,
reaction of 5b (4.2 g, 20 mmol) and 2-nitrobenzoyl
chloride (3.2 mL; 25.2 mmol) afforded 6b as a viscous oil
1
99%) as an oil. H NMR (CDCl3) d 1.63–2.01 (m, 1H),
2.16–2.21 (m, 1H), 2.26 (s, 3H), 2.68–2.78 (m, 1H), 3.20–
3.38 m, 2H), 3.70–3.76 (d, 3H), 3.88–3.99 (m, 1H), 4.68–
4.80 (d, 2H), 5.09–5.20 (m, 2H), 7.27–7.36 (m, 5H);
HRMS, m=z: 293.1 (Mþ1); acc. mass: (C15H20N2O4)
calcd 293.1501, found 293.1502.
1
(4.4 g, 68%), H NMR (500 MHz, DMSO-d6, 2 amide
rotamers, T ¼ 75 ꢁC) d 1.22–1.25 (m, 0.9H, H7), 1.29–
1.33 (t, J ¼ 7:0 Hz, 2.1H, H7), 2.43–2.46 (d,
J ¼ 13:0 Hz, 0.7H, H4), 2.62–2.71 (m, 0.7H, H60), 2.74–
2.79 (m, 0.3H, H6), 2.89–2.96 (m, 0.7H, H2), 3.03–3.07
(m, 0.7H, H2), 3.33–3.14 (m, 0.7H, H2), 3.17–3.22 (m,
0.3H, H5), 3.46–3.51 (t, J ¼ 12:5 Hz, 0.7H, H50), 3.58–
3.61 (m, 0.7H, H5), 4.19–4.29 (m, 2H, H6), 4.57 (s, 0.3H,
H1), 4.80–4.83 (d, J ¼ 13:5 Hz, 0.3H, H50), 5.61 (s,
0.7H, H1), 7.40 (m, 0.3H, H11), 7.50–7.52 (d,
J ¼ 6:5 Hz, 0.7H, H11), 7.71–7.74 (m, 1H, H9), 7.81–
7.83 (m, 0.3H, H10), 7.86–7.89 (t, J ¼ 7:5 Hz, 0.7H,
H10), 8.19-8.20 (d, J ¼ 8:5 Hz, 1H, H8); HRMS, m=z:
326.2 (Mþ1); acc. mass: C13H14N2O5S calcd 325.0858,
found 325.0856.
4.2.3. Synthesis of 4-methylpiperazine-2-carboxylic acid
methyl ester 5a. Compound 4 (11.4 g, 38.9 mmol) was
suspended in ethanol (70 mL) and conc. HCl (4.2 mL,
136 mmol) was added followed by 10% Pd–C (1.4 g).
The suspension was reacted under a hydrogen atmo-
sphere for 24 h, filtered through a pad of Celite, and
washed several times with ethanol. The combined
organic washes were evaporated in vacuo to afford 5a
1
(8.5 g, 96%) as a viscous oil. H NMR (CD3OD) d 3.07
(s, 3H), 3.43–3.52 (m, 1H), 3.56–3.70 (m, 2H), 3.78–3.88
(m, 2H), 3.95 (s, 3H), 4.10–4.20 (m, 1H), 4.80–4.85
(m, 1H); MS, m=z: 159.2 (Mþ1).
4.3.3. 4-(2-Nitrobenzoyl)morpholine-3-carboxylic acid
methyl ester 6c. According to the general procedure,
reaction of 5c (5.5 g, 38 mmol) and 2-nitrobenzoyl
chloride (5.5 mL; 42 mmol) afforded 6c as a viscous oil
4.3. Synthesis of acylated derivatives 6a–c. General
procedure
1
(8.7 g, 78%); H NMR (500 MHz, DMSO-d6, 2 amide
rotamers, T ¼ 75 ꢁC) d 3.15–3.19 (m, 1H, H5), 3.38–3.45
(m, 1.4H, H50/H4), 3.48–3.50 (m, 0.3H, H4), 3.60–3.62
(d, J ¼ 11:0 Hz, 0.3H, H2), 3.72–3.73 (m, 2H, H2/H6/
H4), 3.80 (s, 1.7H, H6), 3.92–3.95 (d, J ¼ 10:5 Hz, 0.3H,
H40), 4.12–4.14 (d, J ¼ 11:5 Hz, 0.3H, H20), 4.20 (s,
0.3H, H1), 4.32–4.34 (m, 1H, H50/H20), 7.38–7.39 (m,
0.3H, H11), 7.50–7.52 (d, J ¼ 7:5 Hz, 0.7H, H11), 7.72–
7.75 (m, 1H, H9), 7.82 (m, 0.3H, H10), 7.87–7.90 (t,
J ¼ 7:5 Hz, 0.7H, H10), 8.19–8.20 (d, J ¼ 8:5 Hz, 1H,
H8); HRMS, m=z: 295.1 (Mþ1); acc. mass:
(C13H14N2O6) calcd 295.0930, found 295.0931.
To a cyclic amino ester 5a, 5b,10 or 5c,11 suspended in
anhydrous dichloromethane (4 mL/mmol), was added
diisopropylethylamine (3.3 equiv) and the reaction mix-
ture was cooled to 0 ꢁC. A solution of 2-nitrobenzoyl
chloride (1.2 equiv) in dichloromethane (0.66 mL/mmol)
was added dropwise. The reaction was allowed to warm
to room temperature and stirred for 18 h. The reaction
was diluted with dichloromethane, washed with H2O,
and dried over Na2SO4. The reaction was filtered and
the solvent was concentrated in vacuo to afford a crude
product, which was purified by flash chromatography
(hexane/EtOAc as eluent) to afford pure 6a, 6b, or 6c.
4.4. Synthesis of benzodiazepinones 7a–c. General pro-
cedure
4.3.1. 4-Methyl-1-(2-nitrobenzoyl)piperazine-2-carbox-
ylic acid methyl ester 6a. According to the general pro-
cedure, reaction between 5a (11.8 g, 51 mmol) and
2-nitrobenzoyl chloride (8.0 mL; 60 mmol) afforded 6a
as a viscous oil (11.9 g, 76%). 1H NMR (500 MHz,
DMSO-d6, 2 amide rotamers, T ¼ 75 ꢁC) d 1.90–1.95
(m, 0.7H, H4), 1.96–2.02 (m, 0.3H, H4), 2.16 (m, 0.3H,
H2), 2.20 (s, 3H, H3), 2.23–2.28 (dd, J ¼ 7:5, 11.5 Hz,
To acylated amino esters 6a, 6b, or 6c, dissolved in
glacial acetic acid (4.5 mL/mmol), was added iron filings
(10 mol equiv) and the reaction mixture was heated at
110 ꢁC until total consumption of starting material
(4–6 h, monitored by TLC). For 7b or 7c, the reaction
was cooled and poured into water. The precipitate was
filtered, washed with water, and dried in vacuo. For 7a,

1264
J. M. Matthews et al. / Tetrahedron: Asymmetry 15 (2004) 1259–1267
the reaction was carefully neutralized with 10% aqueous
NaHCO3 and the aqueous phase was saturated with
NaCl and extracted several times with ethyl acetate. The
extracts were combined, dried over Na2SO4, filtered, and
concentrated in vacuo to provide crude 7a–c. The solids
were triturated with ether, filtered, and dried in vacuo to
afford pure 7a–c.
1.0 M solution of lithium aluminum hydride in tetra-
hydrofuran (3 equiv), dropwise. On completion of the
addition, the reactions were allowed to warm to ambient
temperature and refluxed until total consumption of
starting material (4–9 h, monitored by TLC). The reac-
tions were cooled to 0 ꢁC, quenched carefully with H2O,
and extracted with ethyl acetate (3·). The organic
extracts were combined, washed with brine, dried over
Na2SO4, filtered, and concentrated in vacuo to afford pure
1a–c.
4.4.1. 2-Methyl-1,3,4,11a-tetrahydro-2H,10H-2,4a,10-
triazabenzo[a,d]cycloheptene-5,11-dione 7a. According
to the general procedure, reaction of 6a (4.7 g, 15 mmol)
with iron filings (8.6 g; 153 mmol) afforded 7a as a white
solid (2.61 g, 69%), mp >230 ꢁC; 1H NMR (DMSO-d6) d
1.92–1.97 (m, 1H), 2.03–2.07 (m, 1H), 2.22 (s, 3H), 2.78–
2.80 (m, 1H), 2.88–2.93 (m, 1H), 3.09–3.11 (d,
J ¼ 12 Hz, 1H), 4.12–4.13 (d, J ¼ 4:9 Hz, 1H), 4.26–4.29
(d, J ¼ 13:2 Hz, 1H), 7.06–7.08 (d, J ¼ 8:1 Hz, 1H),
7.18–7.21 (t, J ¼ 7:6 Hz, 1H), 7.46–7.49 (t, J ¼ 7:3 Hz,
1H), 7.68–7.70 (d, J ¼ 7:7 Hz, 1H), 10.34 (s, 1H); 13C
NMR (DMSO-d6) d 45.84, 50.37, 51.58, 52.73, 120.42,
123.83, 126.58, 130.31, 131.89, 137.04, 167.56, 170.54.
MS, m=z: 246.3 (Mþ1). Anal. Calcd for C13H15N3O2
(245.28): C, 63.66; H, 6.16; N, 17.13. Found: C, 63.19;
H, 6.20; N, 17.01.
4.5.1. 2-Methyl-1,2,3,4,5,10,11,11a-octahydro-2,4a,10-
triazadibenzo[a,d]cycloheptene 1a. According to the
general procedure, reaction of 7a (5.5 g, 23 mmol) and
lithium aluminum hydride–tetrahydrofuran (69 mL;
69 mmol) afforded 1a as a yellow solid (4.8 g, 98%), mp
106–107 ꢁC; 1H NMR (CDCl3) d 1.91 (s, 1H), 2.27–2.30
(m, 4H), 2.52–2.64 (m, 3H), 2.72–2.80 (m, 2H), 2.91–
2.96 (m, 1H), 3.11–3.16 (dd, J ¼ 1:95, 13.1 Hz, 1H),
3.58–3.66 (m, 2H), 3.80–3.81 (s, 1H), 6.71–6.73 (m, 1H),
6.82–6.85 (m, 1H), 7.06–7.09 (m, 1H), 7.12–7.14 (m,
1H); 13C NMR (CDCl3) d 46.01, 52.07, 55.17, 58.93,
62.57, 64.33, 118.15, 120.96, 127.89, 129.32, 130.89,
150.21. MS, m=z: 218.2 (Mþ1). Anal. Calcd for
C13H19N3 (217.31): C, 71.85; H, 8.81; N, 19.34. Found:
C, 71.43; H, 8.74; N, 19.16.
4.4.2. 1,3,4,11a-Tetrahydro-10H,2-thia-4a,10-diaza-di-
benzo[a,d]cycloheptene-5,11-dione 7b. According to the
general procedure, reaction of 6b (4.4 g, 13 mmol) with
iron filings (7 g; 125 mmol) afforded 7b as a white solid
(3.0 g, 89%), mp >230 ꢁC; 1H NMR (CDCl3) d 2.50–
2.2.68 (m, 1H), 2.77–2.82 (m, 1H), 2.87–2.90 (m, 1H),
3.13–3.19 (m, 1H), 3.32–3.39 (m, 1H), 4.17–4.21 (m,
1H), 4.55–4.59 (m, 1H), 7.11–7.13 (d, J ¼ 8:1 Hz, 1H),
7.23–7.26 (t, J ¼ 7:1 Hz, 1H), 7.51–7.55 (t, J ¼ 7:1 Hz,
1H), 7.79–7.81 (d, J ¼ 7:8 Hz, 1H), 10.56 (s, 1H); 13C
NMR (DMSO-d6) d 21.17, 24.40, 35.95, 54.67, 120.86,
124.02, 125.77, 130.96, 132.30, 136.55, 166.60, 169.74.
MS, m=z: 249.9 (Mþ1). Anal. Calcd for C12H12N2O2S
(248.30): C, 58.05; H, 4.87; N, 11.28. Found: C, 57.70;
H, 4.95; N, 11.15.
4.5.2. 3,4,5,10,11,11a-Hexahydro-1H,2-thia-4a,10-diaza-
dibenzo[a,d]cycloheptene 1b. According to the general
procedure, reaction of 7b (3.0 g, 12 mmol) and 1.0 M
LiAlH4–THF (36 mL; 36 mmol) afforded 1b as a yellow
1
solid (1.86 g, 70%), mp 68–70 ꢁC; H NMR (CDCl3) d
2.60–2.75 (m, 5H), 2.96–3.00 (m, 1H), 3.00–3.07 (m,
1H), 3.12–3.15 (m, 1H), 3.33–3.37 (m, 1H), 3.70–3.73 (d,
J ¼ 14:3 Hz, 1H), 3.73 (s, 1H), 3.93–3.96 (d,
J ¼ 14:3 Hz, 1H), 6.71–6.73 (d, J ¼ 7:7 Hz, 1H), 6.83–
6.84 (t, J ¼ 6:4 Hz, 1H), 7.06–7.10 (m, 2H); 13C NMR
(CDCl3) d 28.06, 31.75, 49.92, 52.52, 62.15, 63.08,
118.06, 120.65, 127.94, 128.28, 130.57, 149.52. MS, m=z:
221.9 (Mþ1). Anal. Calcd for C12H16N2S (220.33):
C, 65.41; H, 7.32; N, 12.71. Found: C, 65.16; H, 7.30; N,
12.77.
4.4.3. 1,3,4,11a-Tetrahydro-10H,2-oxa-4a,10-diaza-di-
benzo[a,d]cycloheptene-5,11-dione 7c. According to the
general procedure, reaction of 6c (8.67 g, 29.4 mmol)
with iron filings (16.5 g; 294 mmol) afforded 7c as a white
solid (3.67 g, 53%), mp >230 ꢁC; 1H NMR (DMSO-d6) d
2.95–3.00 (m, 1H), 3.44–3.49 (m, 1H), 3.53–3.56 (m,
1H), 3.92–3.95 (m, 1H), 4.05–4.06 (d, J ¼ 3:8 Hz, 1H),
4.14–4.21 (m, 2H), 7.08–7.09 (d, J ¼ 7:3 Hz, 1H), 7.20–
7.24 (t, J ¼ 6:7 Hz, 1H), 7.48–7.51 (t, J ¼ 6:5 Hz, 1H),
7.70–7.71 (m, J ¼ 6:3 Hz, 1H), 10.42 (s, 1H). 13C NMR
(DMSO-d6) d 50.14, 62.57, 64.71, 120.62, 124.04, 126.39,
130.43, 132.03, 136.89, 167.84, 170.16. MS, m=z: 233.3
(Mþ1). Anal. Calcd for C12H12N2O3 (232.24): C, 62.06;
H, 5.21; N, 12.06. Found: C, 61.92; H, 5.24; N, 12.01.
4.5.3. 3,4,5,10,11,11a-Hexahydro-1H-2-oxa-4a,10-diaza-
dibenzo[a,d]cycloheptene 1c. According to the general
procedure, reaction of 7c (12.6 g, 54 mmol) and 1.0 M
LiAlH4–THF (165 mL; 165 mmol) afforded 1c as a yel-
low solid (11.0 g, 99%), mp 111–113 ꢁC; 1H NMR
(CDCl3): d 2.50–2.55 (m, 2H), 2.74–2.85 (m, 2H), 3.04–
3.07 (d, J ¼ 12:8 Hz, 1H), 3.27–3.31 (t, J ¼ 1:9 Hz, 1H),
3.55–3.67 (m, 1H), 3.69–3.81 (m, 4H), 3.82–3.85 (m,
1H), 6.74–6.77 (d, J ¼ 7:7 Hz, 1H), 6.84–6.87 (t,
J ¼ 7:3 Hz, 1H), 7.07–7.09 (m, 1H), 7.10–7.15 (m, 1H);
13C NMR (CDCl3) d 49.52, 53.73, 62.90, 64.28, 67.09,
69.48, 118.36, 121.11, 128.02, 129.17, 130.85, 150.01.
MS, m=z: 205.2 (Mþ1). Anal. Calcd for C12H16N2O
(204.27): C, 70.56; H, 7.90; N, 13.71. Found: C, 70.26;
H, 7.83; N, 13.68.
4.5. Synthesis of benzodiazepines 1a–c. General procedure
To benzodiazepinone 7a, 7b, or 7c in dry tetrahydro-
furan (7.5 mL/mmol), cooled at ꢀ10 ꢁC, was added a

J. M. Matthews et al. / Tetrahedron: Asymmetry 15 (2004) 1259–1267
1265
4.6. Resolution of benzodiazepines 1a–c. General proce-
dures
with cold ether, and dried under vacuo to afford a white
solid (3.4 g, 58%). The solid was partitioned between 1 N
NaOH and ethyl acetate and the layers were partitioned.
The organic phase was washed with water, brine and
dried over Na2SO4. Filtration and evaporation of the
4.6.1. 2-Methyl-1,2,3,4,5,10,11,11a-octahydro-2,4a,10-
triazadibenzo[a,d]cycloheptene, (S)-(þ)- and (R)-(ꢀ)-1a.
To benzodiazepine 1a (3.88 g; 17.8 mmol) in ethanol
solvent in vacuo afforded (S)-(þ)-1c (1.58 g; 26%) as a
20
white solid, ½aꢂ ¼ þ163:6 (c 0.264, MeOH). The
(100 mL) was added dibenzoyl-
D
-tartaric acid (3.2 g;
D
opposite enantiomer was isolated using 1c (0.94 g) and
8.9 mmol) and the suspension was heated until dissolu-
tion. The mixture was cooled and the solvent was
evaporated in vacuo. The residue was redissolved in a
minimum amount of ethanol and heated to initiate
crystallization. Once a solid crystallized, the heat was
removed and the reaction mixture was cooled to ambi-
ent temperature and allowed to sit for 24 h. The resul-
tant crystals were filtered, washed with ethanol, ether,
and dried under vacuo to provide a white solid (2.69 g,
33%). The solid was partitioned between 3 N NaOH and
ethyl acetate. The organic phase was separated, washed
with H2O and brine, and dried over Na2SO4. Filtration
di-p-toluoyl-L-tartaric acid (0.88 g), with the experi-
mental shown above, to afford (R)-(ꢀ)-1c (0.32 g; 32%)
20
as a white solid, ½aꢂ ¼ ꢀ147:6 (c 0.88, MeOH).
D
4.7. X-ray crystallography for (S)-(þ)-1a, (S)-(þ)-1b, and
(S)-(þ)-1c
4.7.1. X-ray crystallography for (S)-(þ)-1a-dibenzoyl-
D-
tartrate. Single crystals of [C13H20N3][(ꢀ)-(S,S)-
C18H13O8] from ethanol are (at ꢀ50ꢁ2 ꢁC) trigonal,
space group P32-C33 (No. 145) with a ¼ 10:231ð2Þ A,
ꢀ
and evaporation of the solvent in vacuo afforded (S)-
20
D
0.808, CHCl3). The opposite enantiomer is isolated by
3
ꢀ
ꢀ
(þ)-1a (0.93 g, 48%) as a white solid, ½aꢂ ¼ þ132:7 (c
c ¼ 24:166ð7Þ A, V ¼ 2190:4ð8Þ A and Z ¼ 3 cation/
anion formula units [dcalcd ¼ 1:309 g cmꢀ3; la(MoKa)
combining 1a (6.2 g) and dibenzoyl-
(0.51 g), with the experimental shown above, to afford
L
-tartaric acid
< 50.7ꢁ].
A
total of 3654 reflections having
2h(MoKa) < 50.7ꢁ (the equivalent of 0.8 limiting CuKa
sphere) were collected; 2865 of these reflections were
unique and gave Rint ¼ 0:029. Lattice constants were
determined with the Bruker XSCANS software package
using 32 centered reflections. Direct methods techniques
were used to solve the structure and the resulting
structural parameters were refined with F 2 data to
convergence by using counter-weighted full-matrix least-
squares techniques and a structural model that incor-
porated anisotropic thermal parameters for all nonhy-
drogen atoms and isotropic thermal parameters for all
hydrogen atoms. The final agreement factors are: R1
(unweighted, based on F )¼0.045 for 1938 independent
reflections having 2h(MoKa) < 50.7ꢁ and I > 2rðIÞ; R1
(unweighted, based on F )¼0.086 and wR2 (weighted,
based on F 2)¼0.089 for all 2865 independent reflections
having 2h(MoKa) < 50.7ꢁ. The absolute configuration
assigned to the chiral cation is based on the known
configuration of the anion; it could not be confirmed
crystallographically since the ‘Flack’ absolute structure
parameter refined to a value of 0(2). Amine hydrogen
atoms H1N and H3N and carboxyl hydrogen atom H2O
were located from a difference Fourier map and refined
as independent isotropic atoms. The remaining hydro-
gen atoms were included in the structure factor
calculations as idealized atoms (assuming sp2- or sp3-
respective carbon atoms and C–H bond lengths of 0.94–
20
D
(R)-(ꢀ)-1a (1.22 g, 39%) as a white solid, ½aꢂ ¼ ꢀ139:8
(c 0.766, CHCl3).
4.6.2. 3,4,5,10,11,11a-Hexahydro-1H,2-thia-4a,10-diaza-
dibenzo[a,d]cycloheptene, (S)-(þ)- and (R)-(ꢀ)-1b. To
benzodiazepine 1b (10.3 g; 46.8 mmol) in methanol
(100 mL) was added (R)-binaphthyl-2,20-diyl hydrogen
phosphate (16.3 g; 46.8 mmol) and the suspension was
heated until dissolution. The reaction mixture was
cooled and the solvent was evaporated in vacuo. The
residue was redissolved in a minimum amount of
methanol and the reaction mixture was allowed to sit for
24 h at ambient temperature. The resultant crystals were
filtered, washed with methanol, ether, and dried in
vacuo to afford a white solid (9.3 g, 35%). The solid was
partitioned between saturated aqueous NaHCO3 and
ethyl acetate. The organic phase was separated, washed
with H2O, brine and dried over Na2SO4. Filtration and
evaporation of the solvent in vacuo afforded (S)-(þ)-1b
20
(3.6 g, 35%) as a white solid. ½aꢂ ¼ þ260:8 (c 0.125,
D
CHCl3). The opposite enantiomer is isolated using 1b
(0.9 g) and (S)-binaphthyl-2-20-diyl hydrogen phosphate
(0.56 g), with the experimental described above, to afford
20
(R)-(ꢀ)-1b (0.32 g, 35%) as a white solid, ½aꢂ ¼ ꢀ280:6
D
(c 0.124, CHCl3).
ꢀ
0.99 A) ‘riding’ on their respective carbon atoms. The
methyl group (C13 and its hydrogens) was refined as a
rigid rotor (using idealized sp3-hybridized geometry and
4.6.3. 3,4,5,10,11,11a-Hexahydro-1H-2-oxa-4a,10-diaza-
dibenzo[a,d]cycloheptene, (S)-(þ)- and (R)-(ꢀ)-1c. To
benzodiazepine 1c (6.16 g; 30 mmol) in methanol
ꢀ
a C–H bond length of 0.96 A) with three rotational
(35 mL) was added di-p-toluoyl-
D
-tartaric acid (5.82 g;
parameters in least-squares cycles. The refined values of
these rotational parameters resulted in N–C–H angles
ranging from 107–111ꢁ. The isotropic thermal parame-
ters for H1N, H3N, and H2O refined to final Uiso values of
0.08(2), 0.03(1), and 0.17(4) A , respectively; those for
the remaining hydrogens were fixed at values of 1.2
(nonmethyl) or 1.5 (methyl) times the equivalent iso-
tropic thermal parameters of the carbon atoms to which
they are bonded. CCDC 233778.13
30 mmol) and the suspension was heated until dissolu-
tion. The reaction mixture was cooled and the solvent
was evaporated in vacuo. The resultant residue was
redissolved in a minimum amount of methanol (35 mL)
and ether (80 mL) was added to give a cloudy solution.
Methanol was added dropwise until clarity was restored.
The solution was allowed to sit at ambient temperature
for 72 h. The resultant crystals were filtered, washed
2
ꢀ

1266
J. M. Matthews et al. / Tetrahedron: Asymmetry 15 (2004) 1259–1267
4.7.2. X-ray crystallography for (S)-(þ)-1b-(R)-binaph-
thyl-2,20-diyl hydrogen phosphate. Single crystals of
[(R,R)-C12H17N2S][(S)-(þ)-(C20H12O4P)] from MeOH
are (at 20ꢁ2 ꢁC) monoclinic, space group P31-C22 (No. 4)
Rint ¼ 0:113. Lattice constants were determined with
Bruker SAINT using peak centers for 765 reflections.
Direct methods techniques were used to solve the
structure and the resulting structural parameters were
refined with F 2 data to convergence using counter-
weighted full-matrix least-squares techniques and a
structural model that incorporated anisotropic thermal
parameters for all nonhydrogen atoms and isotropic
thermal parameters for all hydrogen atoms. Final
agreement factors at convergence are: R1 (unweighted,
based on F )¼0.047 for 3108 independent reflections
have 2h(MoK) < 46.55ꢁ and I > 2rðIÞ; R1 (unweighted,
based on F )¼0.125 and wR2 (weighted, based on
F 2)¼0.081 for all 5734 independent reflections having
2h(MoK) < 46.55ꢁ. The structural model incorporated
anisotropic thermal parameters for all nonhydrogen
atoms and isotropic thermal parameters for all hydrogen
atoms. All amine hydrogen atoms (H5N, H11N, H25N, and
H31N) were located from a difference Fourier synthesis
and refined as independent isotropic atoms. The two
methyl groups (C52, C60, and their hydrogens) were
refined as rigid rotors (using idealized sp3-hybridized
ꢀ
ꢀ
with
a ¼ 10:3195ð8Þ A,
b ¼ 12:1227ð9Þ A,
c ¼
3
ꢀ
ꢀ
12:3904ð10Þ A, b ¼ 108:591ð6Þꢁ, V ¼ 1469:2ð2Þ A and
Z ¼ 2 [dcalcd ¼ 1:358 g cmꢀ3; la(MoKa) ¼ 0.210 mmꢀ1]. A
total of 4483 independent reflections having
2h(MoKa) < 55.0ꢁ (equivalent of 1.0 limiting CuKa
sphere) were collected. Lattice constants were deter-
mined with the Bruker XSCANS software package
using 51 centered reflections. Direct methods techniques
were used to solve the structure and the resulting
structural parameters were refined with F 2 data to
convergence using counter-weighted full-matrix least-
squares techniques and a structural model with incor-
porated anisotropic thermal parameters for all nonhy-
drogen atoms and isotropic thermal parameters for all
hydrogen atoms. Final agreement factors are: R1 (un-
weighted, based on F )¼0.033 for 3455 independent
absorption-corrected reflections have 2h(MoKa) < 55.0ꢁ
and I > 2rðIÞ; and wR2 (weighted, based on F 2)¼0.086
for all 3860 independent absorption-corrected reflec-
tions having 2h(MoKa) < 55.0ꢁ. The methanol solvent
molecule is disordered with two preferred orientations in
the lattice. The major (67.6%) orientation is specified by
H1OS, O1S, C1S, H1Sa, H1Sb and H1Sc; the minor (32.4%)
ꢀ
geometry and a C–H bond length of 0.98 A), which were
allowed to rotate about their C–C bonds in least-squares
cycles. The two hydrogens for the water molecule of
crystallization were located from a difference Fourier
synthesis and included in the structural model as iso-
tropic atoms fixed at these difference Fourier positions.
The remaining hydrogen atoms were included in the
structure factor calculations as idealized atoms (assum-
ing sp2- or sp3-hybridization of the carbon atoms and
0
orientation is specified by O1S , C1S, H1Sd, H1Se and H1Sf
.
Since the hydroxyl hydrogen for the minor orientation
could not be located from difference Fourier syntheses,
it was not included in the structural model. Hydrogen
atoms H1N and H2N, H2, and H1OS were located from a
difference Fourier map and refined as independent iso-
tropic atoms. The remaining hydrogen atoms of the
cation and anion were included in the structure factor
calculations as idealized atoms (assuming sp2- or sp3-
hybridization of the carbon atoms and C–H bond
ꢀ
C–H bond lengths of 0.95–1.00 A) ‘riding’ on their
respective carbon atoms. The isotropic thermal para-
meters for H5N, H11N, H25N, and H31N refined to final
2
ꢀ
Uiso values of 0.07(2), 0.04(2), 0.13(2), and 0.02(1) A ,
respectively. The isotropic thermal parameters of the
remaining hydrogen atoms were fixed at values 1.2
(nonmethyl and water) or 1.5 (methyl) times the equiv-
alent isotropic thermal parameters of the carbon or
oxygen atom to which they are covalently bonded.
CCDC 233780.13
ꢀ
lengths of 0.93–0.97 A) ‘riding’ on their respective car-
bon atoms. Each alternate orientation for the methyl
group of the disordered methanol solvent molecule (C1S
and its hydrogens) was refined as a rigid rotor (using
idealized sp3-hybridized geometry and a C–H bond
ꢀ
length of 0.96 A), which was free to rotate about the O–
C bond. The isotropic thermal parameters for H1N, H2N
H2, and H1OS refined to final Uiso values of 0.06(1),
,
Acknowledgements
2
ꢀ
0.06(1), 0.05(1) and 0.12(3) A , respectively. The iso-
tropic thermal parameters of the remaining hydrogen
atoms were fixed at values 1.2 (nonmethyl) or 1.5 (me-
thyl) times the equivalent isotropic thermal parameters
of the carbon atoms to which they are covalently bonded.
CCDC 233779.13
We are grateful to Mr. Patrick Sasso for chemical sup-
port. We thank Crystalytics Company for performing
the X-ray crystallography.
References and notes
4.7.3. X-ray crystallography for (S)-(þ)-1c as a salt with
1. Carabateas, P. M. US Patent 3732212, 1973.
di-p-toluoyl-D-tartaric acid. Single crystals of [(S,S)-
C12H17N2O]2 [(S,S)-(C20H16O8)]ÆH2O are (at ꢀ80ꢁ2 ꢁC)
2. (a) Albright, J. D.; Reich, M. F.; Delos Santos, E. G.;
Dusza, J. P.; Sum, F. W.; Venkatesan, A. M.; Coupet, J.;
Chan, P. S.; Ru, X.; Mazandarani, H.; Bailey, T. J. Med.
Chem. 1998, 41, 2442; (b) Ogawa, H.; Yamashita, H.;
Kondo, K.; Yamamura, Y.; Miyamota, H.; Kan, K.;
Kitano, K.; Tanaka, M.; Nakaya, K.; Nakamura, S.;
Mori, T.; Tominago, M.; Yakuuchi, Y. J. Med. Chem.
1996, 39, 3547; (c) Albright, J. D.; Delos Santos, E. G.;
Dusza, J. P.; Chan, P. S.; Coupet, J.; Ru, X.; Mazanda-
orthorhombic, space group P212121-D (No. 19) with
ꢀ
ꢀ
ꢀ
a ¼ 11:057ð2Þ A, b ¼ 13:208ð3Þ A, c ¼ 27:336ð6Þ A, V ¼
3992ð1Þ A and Z ¼ 4 formula units [dcalcd ¼ 1:352 g cm3;
3
ꢀ
la(MoK) < 0.098 mmꢀ1]. A total of 18067 integrated
reflection intensities having 2h(MoK) < 46.55ꢁ were
produced using the Bruker program SAINT software
package; 5734 of these were independent and gave

J. M. Matthews et al. / Tetrahedron: Asymmetry 15 (2004) 1259–1267
1267
rani, H. Bioorg. Med. Chem. Lett. 2000, 10, 695; (d)
Shimada, Y.; Taniguchi, N.; Matsuhisa, A.; Sakamato, K.;
Yatsu, T.; Tanka, A. Chem. Pharm. Bull. 2000, 48, 1644.
3. Thurston, D. E.; Bose, D. S. Chem. Rev. 1994, 94, 433.
4. (a) Horton, D. A.; Bourne, G. T.; Smythe, M. L. Chem.
Rev. 2003, 103, 893; (b) Patchett, A. A.; Nargund, R. P.
Annu. Rep. Med. Chem. 2000, 35, 289; (c) Muller, G. Drug
Discov. Today 2003, 8, 681.
5. (a) Carabateas, P. M. US Patent 3763183, 1973; (b) Kato,
H.; Koshinaka, E.; Arata, Y.; Hanaoka, M. Yakugaku
Zasshi 1974, 94, 1566.
6. Dyatkin, A. B.; Hoekstra, W. J.; Maryanoff, B. E.;
Matthews, J. M. PCT Intl. Patent Appl., WO 0043398.
7. Rothchild, R. Enantiomer 2000, 5, 457.
8. Dyatkin, A. B.; Freedman, T. B.; Cao, X.; Dukor, R. K.;
Maryanoff, B. E.; Maryanoff, C. A.; Matthews, J. M.;
Shah, R. D.; Nafie, L. A. Chirality 2002, 14, 215.
9. Bigge, C. F.; Hays, S. J.; Novak, P. M.; Drummond, J. T.;
Johnson, G.; Bobovski, T. P. Tetrahedron Lett. 1989, 30, 5193.
10. Larsson, U.; Carlson, R. Acta Chem. Scand. 1994, 48, 517.
11. Kogami, Y.; Okawa, K. Bull. Chem. Soc. Jpn. 1987, 60, 2963.
12. Grimm, J.; Liu, F.; Stefanick, S.; Sorgi, K. L.; Maryanoff,
C. A. Org. Proc. Res. Dev. 2003, 7, 1067.
13. Crystallographic data has been deposited with the
Cambridge Crystallographic Data Centre. Copies of the
data can be obtained free of charge on application to
CCDC, 12 Union Road, Cambridge CB2 11EZ UK [fax:
+44 (0)-1223-336033 or e-mail: deposit@ccdc.cam.ac.uk].