Sugar-based enantiomeric and conformationally constrained pyrrolo[2,1- c ][1,4]-benzodiazepines as potential GABAA ligands

Article abstract of DOI:10.1021/jm101244n

Synthesis of a library of pyrrolo[2,1-c][1,4]-benzodiazepines derived from spiro bicyclic d- or l-proline analogues containing a d- or l-fructose moiety was developed. The l-fructose moiety was obtained by using a new synthetic pathway starting from l-arabinose through a six steps synthesis in 18% overall yield. Molecular modeling calculations and DNMR studies showed that d- and l-fructose-based pyrrolobenzodiazepines exhibit a rigid (P)- and (M)-helical conformation, respectively, in which the C-11a substituent was always pseudoequatorial. Additionally, pyrrolobenzodiazepines functionalized with a chloride, bromide, nitro, or amino group in the benzene ring, with or without N-methylation and with or without protection of sugar alcohol groups, allowed a relationship between the molecular structure and biological activity to be established. The conformation of the diazepam ring was not the sole key player influencing binding affinities, and the sugar moiety can in some cases increase the binding activity, possibly by participating in the binding event. Finally, these compounds have increased the understanding of the differential recognition of (M)-/(P)-helical benzodiazepines on GABA_A receptor.

Full text of DOI:10.1021/jm101244n

ARTICLE
pubs.acs.org/jmc
Sugar-Based Enantiomeric and Conformationally Constrained
Pyrrolo[2,1-c][1,4]-Benzodiazepines as Potential GABAA Ligands
Ana C. Araꢀujo,*,†,‡ Amꢀelia P. Rauter,† Francesco Nicotra,‡ Cristina Airoldi,‡ Barbara Costa,‡ and
Laura Cipolla‡
†Centro de Química e Bioquímica, Departamento de Química e Bioquímica da Faculdade de Ci^encias da Universidade de Lisboa,
Lisboa, Portugal
‡Dipartimento di Biotecnologie e Bioscienze, Universitꢁa degli Studi di Milano-Bicocca, Milano, Italia
S Supporting Information
b
ABSTRACT: Synthesis of a library of pyrrolo[2,1-c][1,4]-benzodiazepines
derived from spiro bicyclic D- or L-proline analogues containing a D- or L-
fructose moiety was developed. The L-fructose moiety was obtained by using
a new synthetic pathway starting from L-arabinose through a six steps
synthesis in 18% overall yield. Molecular modeling calculations and DNMR
studies showed that D- and L-fructose-based pyrrolobenzodiazepines exhibit a
rigid (P)- and (M)-helical conformation, respectively, in which the C-11a
substituent was always pseudoequatorial. Additionally, pyrrolobenzodiaze-
pines functionalized with a chloride, bromide, nitro, or amino group in the
benzene ring, with or without N-methylation and with or without protection
of sugar alcohol groups, allowed a relationship between the molecular
structure and biological activity to be established. The conformation of the
diazepam ring was not the sole key player influencing binding affinities, and
the sugar moiety can in some cases increase the binding activity, possibly by participating in the binding event. Finally, these
compounds have increased the understanding of the differential recognition of (M)-/(P)-helical benzodiazepines on GABAA
receptor.
’ INTRODUCTION
Anxiety disorders are highly prevalent disabling disorders,
which frequently turn into chronic clinical conditions.1,2 Benzo-
diazepines are fast-acting and effective antianxiety agents3 and
the most commonly prescribed anxiolytics. However, their side
effects such as sedation and, following chronic administration,
development of tolerance, consecutive abuse liability, and with-
drawal symptoms, render their use problematic in the long term
treatment of anxiety disorders.2 Hence, development of new
anxiolytic agents which retain the rapid anxiolytic potential of
benzodiazepines, but lack their unfavorable side effects, remains a
challenge.
Most antianxiety drugs work by modulating neurotransmitters
in the brain, and in particular, benzodiazepine derivatives act on
the γ-amino butyric acid receptor type A (GABAA receptor), a
ligand-gated chloride ion channel that mediates the effects of the
Figure 1. Conformational equilibrium in (3R)- and (3S)-1,4-benzodia-
zepine-2,5-dione.
of the sign of the torsion angle around C3-N,48 as illustrated in
major inhibitory neurotransmitter γ-amino butyric acid (GABA)
in the central nervous system.2,4,5 Furthermore, benzodiazepine
anxiolytics, contrary to GABA, do not open directly the receptor
channel gate but are known to induce a conformational change at
the GABAA receptor.6 Additionally, benzodiazepines exhibit an
interesting conformational enantiomerism which influences their
ability to bind to the benzodiazepine binding site.7 In particular,
the seven-membered ring of 1,4-benzodiazepine has a boat
conformation, referred to as P (plus) or M (minus) on the basis
Figure 1 for (3R)- and (3S)-1,4-benzodiazepine-2,5-dione. The
effect of the conformational changes on the binding ability has
been investigated by several groups, which highlighted that (3R)-
and (3S)-1,4-benzodiazepines have different binding affinities to
the GABAA receptor.7-9 From single crystal X-ray analysis7 and
computational studies of 1,4-benzodiazepines ring inversion,10 it
Received: September 26, 2010
Published: February 07, 2011
r
2011 American Chemical Society
1266
dx.doi.org/10.1021/jm101244n J. Med. Chem. 2011, 54, 1266–1275
|

Journal of Medicinal Chemistry
ARTICLE
Scheme 1. Key Steps toward D-Fructose-Based Pyrrolobenzodiazepines20-23 a
a Reagents and conditions: (a) dry DMF, reflux, 2 h; (b) (i) dry DMF, reflux, 2 h, (ii) CH3I, Cs2CO3, dry DMF, rt, 2 h; (c) Pd(OH)2/C (10% w/w),
CH3OH/EtOAc, 24 h; (d) BCl3 (1.0 M in CH2Cl2), rt, 1 h.
was observed that both enantiomers have a preferred conforma-
tion in which R2 is pseudoequatorially oriented (Figure 1). In
addition, the S-enantiomer with (M)-conformation was always
more active than the corresponding R-enantiomer with (P)-
conformation in [3H]diazepam binding assays and in vivo
pharmacological tests.7-9 Therefore, it was suggested that the
primary source of binding selectivity of 1,4-benzodiazepine
enantiomers was the differential recognition of ligand conforma-
tions by the receptor, with a preferred recognition of (M)-1,4-
benzodiazepine conformation.8,9 Hence, benzodiazepine ring
helical chirality and consequent stereospecific GABAA receptor
binding is important for the development of new lead com-
pounds. Consequently, great efforts have been devoted to
the synthesis of conformationally constrained benzodiazepine
derivatives.10,11 An enantioselective synthesis of 1,4-benzodiaze-
pine-2,5-diones derived from proline (pyrrolo[2,1-c][1,4]benzo-
diazepines) has been previously accomplished and afforded
indeed rigid benzodiazepines.12 Additionally, some hybrid ben-
zodiazepine structures containing monosaccharides have also
been reported by other research groups.13-18 However, the
ability of proline and monosaccharide derived benzodiazepines
to bind to GABAA receptor and their behavior as anxiolytic
agents has been scarcely studied so far.7,19 Herein we describe the
synthesis of a library of enantiomeric and conformationally
constrained pyrrolo[2,1-c][1,4]-benzodiazepines spiro-linked
to a five-membered ring carbohydrate and their preliminary
evaluation as GABAA receptor ligands.
pyrrolobenzodiazepines 4-7 with (11aR) configuration in good
yields. Subsequent hydrogenolysis with Pd(OH)2/C resulted in
debenzylation and reduction of the nitro group, affording the
sugar-based pyrrolo[2,1-c][1,4]-benzodiazepines 10-13, accord-
ing to Araꢀujo et al.22 D-Fructose-based pyrrolobenzodiazepines 8
and 9 were obtained in a two-step reaction by condensation of
compound 2 with 5-chloro- or 5-bromoisatoicanhydride, followed
by N-methylation with iodomethane and cesium carbonate in
DMF. To avoid removal of the chloro/bromo substituent, deben-
zylation was accomplished with BCl3 (1.0 M in CH2Cl2) to give
compounds 14-15 in good yields.
Hence, since it is established that R-amino acids are the key
starting materials that provide the configuration at C-11a of
pyrrolo[2,1-c][1,4]benzodiazepines, it is desirable to synthesize
the enantiomeric sugar-based pyrrolobenzodiazepines with
(11aS) configuration, starting from the L-proline analogue 24
(Scheme 2). Following the synthetic route for the D-proline
analogue 2, compound 24 was originated from L-fructose deri-
vative 22. One of the possible approaches to obtain this com-
pound relies on the use of L-fructose as starting material.
However, because this is an expensive and unnatural sugar,
alternative synthesis of L-fructose has been elaborated from other
carbohydrates of the L-series, for example, by chain elongation of
the corresponding arabinonic acid, arabinose or glyceraldehyde,
or by inversion of configuration of sorbose.24-28 Wedescribeherein
a new synthesis of the R/β-methyl 3,4,6-tri-O-benzyl-R,β-L-
arabino-hex-2-ulo-2,5-furanoside (22), starting from L-arabinose
16 (Scheme 2) in 18% overall yield. This efficient approach took
advantage of the stereochemistry of the arabinonolactone 17,
which was easily prepared by oxidation of the anomeric position
of compound 16 with bromine in water.29 Benzylation of
lactone 17 using benzyl 2,2,2,-trichloroacetimidate under acidic
conditions afforded compound 18. Introduction of the formyl
group was then accomplished via insertion of a dithiane and
subsequent hydrolysis. Arabinonolactone 18 was reacted with
’ RESULTS AND DISCUSSION
1. Synthesis. Bicyclic spiro analogue D-proline 2 was obtained
through a 15-step pathway in 40% overall yield starting from
D-fructose 1, in accordance with previously published material
(Scheme 1).20,21 Condensation of the D-proline analogue 2
with isatoic anhydrides type 3 afforded the D-fructose-based
1267
dx.doi.org/10.1021/jm101244n |J. Med. Chem. 2011, 54, 1266–1275

Journal of Medicinal Chemistry
ARTICLE
Scheme 2. Synthetic Steps toward Sugar-Based L-Proline Starting from L-Arabinose and the Following Synthesis toward
L-Fructose-Based Pyrrolobenzodiazepinesa
a Reagents and conditions: (a) Br2, K2CO3, H2O, rt, 2 h, 98% yield; (b) CCl3C(dNH)OBn, TfOH, dioxane, 0 °C then rt, 3 h, 65% yield; (c) 1,3 dithiane,
ꢀ
n-BuLi (1.6 M hexane), THF, -78 °C, 2 h, 64% yield as a mixture of anomers (1:4 R:β ratio); (d) CH3OH, H2SO4, ms 3 Å, THF, reflux
40 °C, 12 h, 67% yield as a mixture of anomers (2:1 R:β ratio); (e) DBDMH, acetone, -20 °C, 15 min; (f) NaBH4, EtOH, 0 °C, 30 min, 65% yield as a
mixture of anomers (over two steps); (g) [i] BTSFA, CH3CN, reflux 50 °C, 6 h, [ii] AllSi(CH3)3, TMSOTf, 0 °C then rt, 2 h, 76% yield for R anomer
(8:2 R:β); (h) detailed synthesis can be found in the Supporting Information; (i) dry DMF, reflux, 2 h; (j) [i] dry DMF, reflux, 2 h, [ii] CH3I, Cs2CO3,
dry DMF, rt, 2 h; (k) Pd(OH)2/C (10% w/w), CH3OH/EtOAc, 24 h; (l) BCl3 (1.0 M in CH2Cl2), rt, 1 h.
Scheme 3. Synthetic Steps toward Pyrrolo[2,1-c][1,4]benzodiazepines Derived from the Free Amino Acid D- and L-Proline
2-lithium-1,3-dithiane followed by methylation of the free hydro-
xyl group in compound 19 using catalytic amounts of concd
sulfuric acid in refluxing THF. The dithiated intermediate 20 was
hydrolyzed by 1,3-dibromo-5,5-dimethylhydantoin (DBDMH),
giving aldehyde 21 according to a procedure by Davis et al.30 The
R/β-mixture of the crude aldehyde 21 was finally reduced to the
primary alcohol 22 by treatment with sodium borohydride in
ethanol. Elongation at the anomeric position of compound 22
was accomplished through in situ silylation of the primary
hydroxyl group, followed by Lewis acid mediated allylation by
allyltrimethylsilane, affording the R-anomer 23 as the major
compound, isolated in 76% yield. Synthesis of the L-proline
analogue 24 was then performed as previously described for its
enantiomer 2. The sugar-based free amino acid 24 was isolated in
a total yield of 3.3% over 18 steps.
25-30, Scheme 2) was accomplished by heating suitably func-
tionalized isatoic anhydrides of type 3 in dry DMF with
L-proline analogue 24. As mentioned previously for the corre-
sponding enantiomers, N-methylation was performed to obtain
benzodiazepine derivatives 29 and 30 and debenzylation was
performed accordingly, affording compounds 31-36. The
pyrrolo[2,1-c][1,4]benzodiazepines 37-40, derived from D- or
L-proline, were also produced following the general procedure
described for compound 3919 (Scheme 3) in order to be used as
reference compounds for the biological assays.
2. Conformational Analysis. Conformational analysis of the
water-soluble sugar-based pyrrolo[2,1-c][1,4]benzodiazepine-
2,5-diones 10, 11, and 31 was performed by molecular modeling
calculations and experimental NMR studies at variable tempera-
ture (dynamic NMR, DNMR). Previous data collected on hybrid
pyrrolobenzodiazepines derived from D-fructose, compounds 10
(R1 = R2 = R3 = H) and 11 (R1 = CH3, R2 = R3 = H), indicated
Synthesis of the novel sugar-based pyrrolo[2,1-c][1,4]benzo-
diazepine-2,5-diones with (11aS)-configuration (compounds
1268
dx.doi.org/10.1021/jm101244n |J. Med. Chem. 2011, 54, 1266–1275

Journal of Medicinal Chemistry
ARTICLE
Table 1. Computational Calculation of the Dihedral Angle
C5a-C9a-N,10-C11 by Molecular Mechanics and
Dynamics Computations
as GABAA receptor ligands has been performed. In parti-
cular, theirability to displace [3H]flunitrazepamfrom the receptor
was tested by a classical competition binding assay, using rat
cortical membranes.33 The ability of the pyrrolobenzodiazepine
Dihedral
scaffolds to bind to GABAA receptor, at
a concen-
conformer
(oxygen-substituent) (kcal/mol)
Total energya Anglea,b C5a-C9a-N,
tration of 100 μM, was determined using the radioligand
[3H]flunitrazepam at a concentration of 1 nM. As control,
[3H]flunitrazepam without pyrrolobenzodiazepines was used to
determine a comparative baseline.
compd
10-C11 (deg)
(11aR)-10
pseudoaxial
25.028
22.319
þ16.91
þ17.32
pseudoequatorial
Initial binding assays were performed on D-fructose-based
pyrrolobenzodiazepines 4-7 and 10-13, which presented a
(P)-helical chirality (Scheme 1). The effect of the presence of
the N-methyl and benzene ring substituents, such as -NO2
and -NH2, were evaluated. Despite having the “incorrect” con-
formation, some of thesynthesizedcompounds showed affinity for
GABAA receptor. Data showed that compounds 7, 10, and 13
possess affinity for GABAA receptor since they significantly
inhibited the binding of [3H]flunitrazepam in the μM range of
concentration (Table 2). Results indicate that a methyl group at
position 10 together with a substituent on the benzene ring is
important for binding affinity. Subsequently, binding assays were
performed with compounds 8-9 and 14-15 in order to evaluate
the effect of a halogen (-Cl and -Br) substituent on the benzene
ring. Data showed that a halogen substituent in combination with
free hydroxyl groups have significant effect on binding activity
(14-15). Of all the active compounds (7, 10, 13-15) only
compound 7 is O-benzylated; hence almost all the sugar-based
pyrrolobenzodiazepines that presented the highest binding affi-
nities were water-soluble.
Binding assays were also performed on L-fructose-based
pyrrolobenzodiazepines 25-36. Data illustrated in Table 3 re-
inforced the previous results obtained for the corresponding
enantiomers. Hence, N-methylation and a substituent on posi-
tion 7 in the benzene ring are fundamental factors for GABAA
receptor binding affinity, where the most important factor is the
substituent on the benzene ring. The biological data demonstrate
that -NO2 and -NH2 substituents are powerful enough to
produce effect even in the absence of the N-methyl and/or
O-benzyl groups. L-Fructose-based pyrrolobenzodiazepines with
a -Cl substituent (29 and 35) presented higher biological
activity compared to those with a -Br substituent (30 and 36).
Pyrrolobenzodiazepines without the sugar moiety were also
characterized in receptor binding studies with [3H]flunitrazepam
(37-40, Table 4). Compound 40 presenting (M)-helical con-
formation shows higher binding activity than pyrrolobenzodia-
zepine 38 with (P)-helical conformation indicating that the type
of conformation is important. However, for the sugar-based
pyrrolobenzodiazepines, compound 10 with a (P)-helical con-
formation shows higher binding activity compared to the corre-
sponding (M)-helical compound 31. Contrary to unconstrained
benzodiazepines, the effect of the helical conformation seems
more complex.
(11aR)-11
pseudoaxial
30.438
28.235
þ22.73
þ24.57
pseudoequatorial
(11aS)-31
pseudoaxial
25.680
22.838
-16.58
-17.85
pseudoequatorial
a Calculated by ChemBio3D Ultra by applying an MM2 force field and
molecular dynamics computation (step interval = 2 fs; frame interval =
1 fs; heating/cooling rate = 1.000 kcal/atoms/ps; target temperature =
300 K). Total energy and dihedral angle C5a-C9a-N,10-C11 values
calculated with the two methods are consistent. b The dihedral
angle average between the two conformers resulted þ17° and þ24°
for compound 10 and 11, respectively, corresponding to the helical
chirality (P) and -17° for compound 31, corresponding to the helical
chirality (M).
that they adopt a rigid conformation around the benzodiazepine
ring, where H-11a has a pseudoaxial position.22 A positive
dihedral angle C5a-C9a-N,10-C11 confirmed the helical
chirality (P) in both compounds (Table 1). Additional NMR
experiments run at variable temperature, to a maximum value of
363 K, showed that no conformational equilibrium of the
benzodiazepine ring is present (for more details, see the Support-
ing Information). In all reported cases, the coalescence tempera-
ture for the pseudoequatorial/-axial equilibrium is always around
333 K.31,32
The conformational analysis study of the novel hybrid L-
fructose-based pyrrolobenzodiazepine 31 (R1 = R2 = R3 = H)
showed a unique preferential conformation of the benzodiaze-
pine ring. As expected, this pyrrolobenzodiazepine adopts the
(M)-helical conformation, which is the opposite to that assumed
by its enantiomer, pyrrolobenzodiazepine 10, because this con-
formation places the pyrrolidine substituent in a pseudoequator-
ial orientation on the benzodiazepine ring. The dihedral angle
C5a-C9a-N,10-C11 was also determined and found to be
-17°, corresponding to the helical chirality (M) (Table 1 and
Figure 2). NMR experiments run at variable temperature
(Figure 3) showed no conformational equilibrium of the benzo-
diazepine ring, i.e. no protons presented coalescence between
298 and 363 K, in contrast to quaternary 1,4-benzodiazepin-2-
ones, which exists as mixtures of the (M)- and (P)-
conformers.11,12
The sugar-based pyrrolobenzodiazepines 10, 11, and 31
appeared always as single conformers regarding the diazepine
ring, and the sterically demanding substituent present on this ring,
the sugar-linked pyrrolidine, always assumes a pseudoequatorial
orientation. The conformational homogeneity of the diazepine
ring in the synthesized proline-derived benzodiazepines can
be attributed to an amide resonance-imposed requirement for
the proline moiety to occupy a pseudoequatorial orientation on
the benzodiazepine ring, as reported in other cases.11,12
The steric hindrance effect of the sugar moiety was evaluated
by comparison of the pyrrolobenzodiazepines with the sugar-
based pyrrolobenzodiazepines with the corresponding substitu-
ents. The binding correlated well with two exceptions. First,
pyrrolobenzodiazepine 28 showed 17% reduction of the radi-
oligand specific binding, while the corresponding pyrrolobenzo-
diazepine derivative without sugar moiety (compound 40)
showed an increased reduction (37%), allowing for the possibi-
lity of steric hindrance of the sugar moiety. Second, comparison
of pyrrolobenzodiazepine 37 with the corresponding sugar-based
3. GABAA Receptor Binding Assay. A preliminary biological
evaluation of the synthesized sugar-based pyrrolobenzodiazepines
1269
dx.doi.org/10.1021/jm101244n |J. Med. Chem. 2011, 54, 1266–1275

Journal of Medicinal Chemistry
ARTICLE
Figure 2. Oxygen-substituent pseudoaxial (a) and pseudoequatorial (b) conformers of the pyrrolo ring of compound 31.
Figure 3. DNMR spectra (298K-363K) of molecule 31 dissolved in D2O, 32 scans. Spectrum 1, 298 K; spectrum 2, 303 K; spectrum 3, 313 K;
spectrum 4, 323 K; spectrum 5, 333 K; spectrum 6, 343 K; spectrum 7, 353 K; spectrum 8, 363 K.
pyrrolobenzodiazepine (compound 10) shows that the pyrrolo-
benzodiazepine is not active, whereas the corresponding sugar
analogue shows activity. Also, comparison between pyrroloben-
zodiazepine 38 and the corresponding sugar analogue (com-
pound 7) shows slightly increased binding for the sugar analogue,
indicating no steric hindrance of the sugar moiety. Consequently,
the sugar moiety in the S-compounds can have a potential steric
effect, whereas in the R-compounds the sugar moiety does not
seem to inflict significant steric hindrance on the binding to the
receptor and can even increase the binding affinity, indicating a
possible positive participation.
’ CONCLUSIONS
A library of sugar-based pyrrolobenzodiazepines was synthe-
sized with constrained enantiomeric (M)- and (P)-conforma-
tion. The locked enantiomeric conformations were achieved by
controlling the absolute stereochemistry of carbon C-11a by
connecting the sugar derivatized pyrrolo moiety (proline de-
rivative) to the benzodiazepine structure. The library consisted of
sugar moieties with free hydroxyl groups or O-benzyl groups
(polar and apolar inhibitors), with or without N-methylation and
a variety of substituents on position 7 of the benzodiazepine.
1270
dx.doi.org/10.1021/jm101244n |J. Med. Chem. 2011, 54, 1266–1275

Journal of Medicinal Chemistry
ARTICLE
Table 2. Binding Competition Studies of D-Fructose-Based
Pyrrolobenzodiazepines with [3H]Flunitrazepam on GABAA
Receptor, Performed on Rat Cortical Membranes
Table 3. Binding Competition Studies of L-Fructose-Based
Pyrrolobenzodiazepines with [3H]Flunitrazepam on GABAA
Receptor, Performed on Rat Cortical Membranes
%[3H]Flunitrazepam
specific bindinga
significance
vs controlb
%[3H]Flunitrazepam
specific bindinga
significance
vs controlb
compd
R1
R2
R3
compd
R1
R2
R3
control
100.00 ( 2.00
111.81 ( 2.33
96.44 ( 8.85
101.01 ( 2.10
74.89 ( 1.50
92.94 ( 4.13
111.2 ( 12.32
78.22 ( 7.43
93.04 ( 12.77
80.17 ( 9.43
70.63 ( 2.36
73.74 ( 7.31
84.21 ( 2.84
control
25
100.00 ( 2.40
96.32 ( 3.56
105.90 ( 4.62
69.60 ( 1.50
82.82 ( 1.24
82.33 ( 2.19
93.20 ( 4.88
109.70 ( 7.71
112.20 ( 8.17
73.86 ( 4.44
83.55 ( 2.52
77.00 ( 2.27
82.17 ( 1.69
4
H
H
Bn
Bn
Bn
Bn
Bn
Bn
H
ns
H
H
Bn
Bn
Bn
Bn
Bn
Bn
H
ns
5
CH3
H
H
ns
26
CH3
H
H
ns
6
NO2
NO2
Cl
ns
27
NO2
NO2
Cl
P < 0.001
P < 0.001
P < 0.01
ns
7
CH3
CH3
CH3
H
P < 0.05
ns
28
CH3
CH3
CH3
H
8
29
9
Br
ns
30
Br
10
11
12
13
14
15
H
P < 0.05
ns
31
H
ns
CH3
H
H
H
32
CH3
H
H
H
ns
NH2
NH2
Cl
H
ns
33
NH2
NH2
Cl
H
P < 0.01
P < 0.01
P < 0.001
P < 0.01
CH3
CH3
CH3
H
P < 0.01
P < 0.05
P < 0.05
34
CH3
CH3
CH3
H
H
35
H
Br
H
36
Br
H
a Values are means ( SEM determined from at least three independent
experiments. b Statistical analysis is performed with Kruskal-Wallis
ANOVA for non parametric values followed by Dunns test; ns means
not statistically significant and P means probability.
a Values are means ( SEM determined from at least three independent
experiments. b Statistical analysis is performed with Kruskal-Wallis
ANOVA for nonparametric values followed by Dunns test. ns means
not statistically significant and P means probability.
Preliminary GABAA receptor competition binding assays
using [3H]flunitrazepam indicated that the conformation of the
diazepam ring is not the key factor for influencing the binding
affinities. The D-fructose-based pyrrolobenzodiazepines displaying
a (P)-helical conformation presented binding affinities similar
to those of the L-fructose-based pyrrolobenzodiazepines with a
(M)-helical conformation. The D-fructose-based pyrrolobenzodia-
zepines, which showed the highest binding activity was N-methy-
lated and carried free hydroxyl groups on the sugar. For the
L-fructose-based pyrrolobenzodiazepines, N-methylation decreased
binding activity and O-benzyl groups on the sugar slightly increased
the binding. The effect of the benzodiazepine substituent also varied
between the D- and L-fructose-based pyrrolobenzodiazepines where
the binding activity increased in the following order NH2 > Cl >
NO2 > Br for the D-series and NO2 > NH2 > Cl > Br for the L-series.
The results further showed that the sugar moiety may or may not
inflict significant steric hindrance on the binding and in some cases
can increase the binding activity, possibly by participating in the
binding event. However, the biological activity of the library is
actually far from that of classical GABAA modulators, which falls in
the nM range; nevertheless, these results indicate that sugar-based
pyrrolobenzodiazepines can be considered as lead compounds and
have increased the understanding of the differential recognition of
(M)-/(P)-helical benzodiazepines on GABAA receptor.
(Merck) with detection with UV light when possible, or charring with a
solution containing concd H2SO4/EtOH/H2O in a ratio of 5:45:45
followed by heating at 180 °C. Flash column chromatography was
performed on silica gel 230-400 mesh (Merck). NMR spectra were
recorded at 400 MHz (1H) and at 100.57 MHz (13C) on a Varian
MERCURY instrument at 300 K or on a Bruker Avance 400 spectro-
meter at 300 K unless otherwise stated. Chemical shifts values (δ) are
reported in ppm downfield from TMS as an internal standard;
J values are given in Hz. Mass spectra were recorded on a MALDI2
Kompact Kratos instrument, with gentisic acid (DHB) as the matrix.
Elemental analyses (C, H, N) were performed on a Perkin-Elmer series
II 2400 analyzer, and all synthesized compounds showed a purity of
more than 95%.
General Procedure for the Synthesis of Sugar-Based Pyrrolobenzo-
diazepines. To a solution of the proline analogue (2 or 24) in dry DMF,
suitably functionalized isatoic anhydride type 3 was added under argon
atmosphere, and the resulting mixture was stirred under reflux until
complete consume of the starting material. The solvent was then
removed under reduced pressure and the residue dissolved in CH2Cl2,
filtered, and evaporated. The crude halogenated benzodiazepines (R2 =
Cl, Br) were used in the next step (methylation), whereas the crude
products derived from R2 = H, NO2 were purified by silica gel flash
chromatography to give the desired product (compounds 4-7 and
25-28).
L-Arabinono-1,4-lactone (17). To a solution of R,β L-arabinose 16
(5 g, 0.033 mol) in water (50 mL) potassium carbonate (6.84 g, 0.050
mol) was added and the resulting mixture was cooled to 0 °C. Then
bromine (3.4 mL, 0.067 mol) was added dropwise under continues
stirring. The reaction was left stirring in the dark for about 1 h at rt. The
solvent was then removed under reduced pressure. The crude residue
was purified by silica gel flash chromatography (EtOAc/EtOH 8:2) to
give the desired product 17 as a yellow oil (4.79 g, 98% yield). MS
’ EXPERIMENTAL SECTION
1. Chemistry. General. All solvents were dried over molecular
sieves (Fluka) for at least 24 h prior to use. When dry conditions were
required, the reactions were performed under argon atmosphere. Thin-
layer chromatography (TLC) was performed on Silica Gel 60 F254 plates
1271
dx.doi.org/10.1021/jm101244n |J. Med. Chem. 2011, 54, 1266–1275

Journal of Medicinal Chemistry
ARTICLE
Table 4. Binding Competition Studies of Pyrrolobenzo-
diazepines Derived from D- and L-Proline with
[3H]Flunitrazepam on GABAA Receptor, Performed on Rat
Cortical Membranes
NMR signals). The isomers were not completely separable by chroma-
tography. The following signal attributions were assigned from a small
amount of pure β-anomer. MS (MALDI-TOF): C30H10O5S2 calculated
1
538.72, observed 539.7 [M þ H]þ. H NMR (400 MHz, CDCl3): δ
(ppm) 7.38-7.18 (m, 15H, OCH2Ph), 4.70 (s, 1H, H-1), 4.62-4.41
(m, 7H, OCH2Ph, H-4), 4.14 (bm, 2H, H-3, H-5), 3.60 (bm, 1H, H-6),
3.17 (bm, 2H, H-20), 2.53 (bm, 2H, H-40), 2.02 (bm, 2H, H-30). 13C
NMR (100.57 MHz, CDCl3): δ (ppm) 137.86, 137.80, 137.50 (Cq Ph),
128.45-127.68 (OCH2Ph), 110.10 (C-2), 84.11, 83.12, 80.23 (C-3,
C-4, C-5), 73.32, 72.94, 72.30 (OCH2Ph), 71.44 (C-6), 53.41 (C-1),
27.48, 27.34 (C-20, C-40), 25.16 (C-30).
%[3H]Flunitrazepam
Methyl 2,3,5-Tri-O-benzyl-1-C-(1,3-dithiane-2-yl)-L-arabinoside
(20). To a solution of 19 (1.0 g, 1.86 mmol) in dry THF (8 mL), ms 3 Å
(1.0 g), MeOH (20 mL), and H2SO4 (40 μL) were added under argon
atmosphere, and the resulting mixture was stirred under reflux (bath
temperature 40 °C) for 12 h. After this time, a solution of 1 M NaOH
was added to the reaction to neutralize the acid, then the solvent was
partially evaporated, and the residue was diluted with CH2Cl2. The
organic layer was dried on sodium sulfate, filtered, and evaporated.
Purification by flash chromatography (petroleum ether/EtOAc 8:2)
afforded anomers 20 (0.688 g, 67% yield, 1.5:1 R:β ratio) as a yellow oil.
The following signal attributions were obtained from the separated R,β
mixture. A 2D NOESY experiment for the R-anomer 20 showed a cross
peak between OCH3 and H-3, and between OCH3 and H-5 allowing
the assignment of the corresponding stereochemistry. R-Anomer: MS
(MALDI-TOF) C31H36O5S2 calculated 552.74, observed 553.7 [M þ
specific
bindinga
significance
vs controlb
compd
R1
R2
control
100.00 ( 1.49
97.00 ( 0.91
80.25 ( 3.68
95.50 ( 2.47
63.40 ( 1.57
(11aR)-37
(11aR)-38
(11aS)-39
(11aS)-40
H
H
ns
CH3
H
NO2
H
P < 0.05
ns
CH3
NO2
P < 0.001
a Values are means ( SEM determined from at least three independent
experiments. b Statistical analysis is performed with Kruskal-Wallis
ANOVA for nonparametric values followed by Dunns test. ns means
not statistically significant and P means probability.
(MALDI-TOF): C5H8O5 calculated 148.11, observed 149.1 [M þ H]þ,
171.1 [M þ Na]þ. Elemental analysis calcd (%): C, 40.55; H, 5.44.
Found: C, 40.61; H, 5.40. 1H NMR (400 MHz, CD3OD): δ (ppm) 4.33
(d, 1H, J = 8.5 Hz, H-2), 4.14-4.09 (m, 2H, H-3, H-4), 3.87 (dd, 1H, J =
13.0, 1.9 Hz, H-5a), 3.65 (dd, 1H, J = 13.0, 4.0 Hz, H-5b). 13C NMR
(100.57 MHz, CD3OD): δ (ppm) 176.41 (CO), 82.82, 75.66, 74.18 (C-
2, C-3, C-4), 60.89 (C-5).
2,3,5-Tri-O-benzyl-L-arabinono-1,4-lactone (18). Benzyl trichloroa-
cetimidate (28 mL, 0.153 mol) and trifluoromethanesulfonic acid
(0.301 mL, 0.003 mmol) were added at 0 °C, under argon atmosphere,
to a solution of 17 (5 g, 0.034 mol) in dry dioxane (100 mL). The
solution was stirred for 3 h, and then quenched with a satd solution of
NaHCO3 and extracted with CH2Cl2. The organic layer was dried on
sodium sulfate, filtered, and evaporated. Flash column chromatography
of the residue (petroleum ether/EtOAc 9:1) afforded 18 as a yellow oil
(9.25 g, 65% yield). MS (MALDI-TOF): C26H26O5 calculated 418.48,
observed 419.5 [M þ H]þ, 457.5 [M þ K]þ. Elemental analysis calcd
(%): C, 74.62; H, 6.26. Found: C, 74.74; H, 6.22. 1H NMR (400 MHz,
CDCl3): δ (ppm) 7.33-7.06 (m, 15H, OCH2Ph), 5.00, 4.97, 4.71, 4.68
(AB, 2H, J = 11.6 Hz, PhCH2O), 4.55 (d, 1H, J = 11.6 Hz, PhCH2O),
4.49-4.41 (m, 2H, PhCH2O) 4.28-4.24 (m, 3H, H-2, H-3, H-4), 3.63
(dd, 1H, J = 11.3, 1.8 Hz, H-5a), 3.50 (dd, 1H, J = 11.4, 3.3 Hz, H-5b).
13C NMR (100.57 MHz, CDCl3): δ (ppm) 172.57 (CO), 137.41,
137.02, 136.72 (Cq Ph), 128.53-127.72 (OCH2Ph), 79.20, 79.09,
78.81 (C-2, C-3, C-4), 72.48, 72.68, 72.47 (OCH2Ph), 67.89 (C-5).
1
H]þ. H NMR (400 MHz, CDCl3): δ (ppm) 7.32-7.26 (m, 15H,
OCH2Ph), 4.79 (s, 1H, H-1), 4.59-4.45 (m, 6H, OCH2Ph), 4.15 (dd,
1H, J = 10.3, 4.5 Hz, H-5), 3.97 (d, 1H, J = 1.5 Hz, H-3), 3.86 (d, 1H, J =
4.2 Hz, H-4), 3.68 (dd, 1H, J = 10.4, 5.4 Hz, H-6a), 3.54 (bm, 1H, H-6b),
3.52 (s, 3H, OCH3), 2.91-2.75 (m, 2H, H-20, H-40), 2.06 (bm, 1H,
H-30a), 1.90 (bm, 2H, H-30b). 13C NMR (100.57 MHz, CDCl3): δ
(ppm) 138.17, 137.88, 137.66 (Cq Ph), 128.87-127.57 (OCH2Ph),
108.59 (C-2), 87.18, 84.33, 82.90 (C-3, C-4, C-5), 73.33, 72.61, 71.67
(OCH2Ph), 70.11 (C-6), 51.46 (C-1), 50.61 (OCH3), 30.78, 29.68 (C-
20, C-40), 26.41 (C-30). β-Anomer: MS (MALDI-TOF) C31H36O5S2
1
calculated 552.74, observed 553.7 [M þ H]þ. H NMR (400 MHz,
CDCl3): δ (ppm) 7.34-7.26 (m, 15H, OCH2Ph), 5.30 (s, 1H, H-1),
4.58-4.47 (m, 6H, OCH2Ph), 4.17 (dd, 1H, J = 13.5, 6.7 Hz, H-5), 4.12
(d, 1H, J = 7.1 Hz, H-3), 4.07 (d, 1H, J = 6.6 Hz, H-4), 3.59 (dd, J = 10.3,
4.8 Hz, H-6a), 3.54 (dd, J = 10.3, 5.4 Hz, H-6b), 3.41 (s, 3H, OCH3),
2.93-2.76 (m, 2H, H-20, H-40), 2.05 (bm, 1H, H-30a), 1.89
(bm, 2H, H-30b). 13C NMR (100.57 MHz, CDCl3): δ (ppm) 138.24,
137.97, 137.75 (Cq Ph), 128.87-127.60 (OCH2Ph), 105.25
(C-2), 85.45, 83.94, 79.25 (C-3, C-4, C-5), 73.68, 72.23, 72.60
(OCH2Ph), 70.39 (C-6), 52.15 (C-1), 49.77 (OCH3), 30.36, 29.69
(C-20, C-40), 25.80 (C-30).
Methyl 3,4,6-Tri-O-benzyl-L-arabino-hexo-2-ulo-2,5-furanoside
(21). To a solution of 20 (0.500 g, 0.904 mmol) in acetone (12 mL) at
25 °C was added a solution of 1.3-dibromo-5.5-dimethylhydantion
(DBDMH) (0.517 g, 1.81 mmol) in acetone (8 mL) at -20 °C. The
solution quickly turned red but soon faded to yellow-orange, and was
stirred for 15 min. At this time, the solution was shaken with a mixture of
satd aq solution of sodium sulfite and CH2Cl2. The organic phase was
washed with sodium bicarbonate, water, and brine, dried (NaSO4), and
concentrated to give compound 21 as yellow oil, which was used in the
next step without further purification.
Methyl 3,4,6-Tri-O-benzyl-L-arabino-hex-2-ulo-2,5-furanoside
(22). To a 0 °C cooled solution of 21 (≈0.904 mmol) in dry EtOH
(25 mL) was added NaBH4 (0.410 g, 10.85 mmol) under argon
atmosphere. The reaction mixture was left stirring for 30 min and then
quenched by addition of satd aq NH4Cl solution at 0 °C. The resulting
mixture was extracted with EtOAc, and the organic layer was dried on
sodium sulfate, filtered, and evaporated. Purification by flash chroma-
2,3,5-Tri-O-benzyl-1-C-(1,3-dithiane-2-yl)-L-arabinose (19). In
a
100 mL, two-necked, round-bottomed flask equipped with a magnetic
stirring bar, a rubber septum, and an argon filled balloon was placed 1,3-
dithiane (0.432 g, 3.6 mmol) in dry THF (24 mL). The solution was
cooled to -20 °C and n-BuLi (3 mL, 4.8 mmol, 1.6 M in hexane) was
added slowly. After 1.5 h, the resulting mixture was cooled to -78 °C
and added via cannula to a -78 °C solution of 18 (1.0 g, 2.4 mmol) in
dry THF (28 mL). The reaction was stirred for 20 min and quenched
at -78 °C by addition of satd solution of NH4Cl (12 mL). To the
reaction mixture was added EtOAc (80 mL) and then extracted. The
organic layer was dried on sodium sulfate, filtered, and evaporated. Flash
column chromatography of the residue (petroleum ether/EtOAc 8:2)
afforded 19 (0.827 g, 64% yield) as a yellow oil mixture of R and β
anomers (1:4 R:β ratio, as determined from the ratio integrals of the 1H
1272
dx.doi.org/10.1021/jm101244n |J. Med. Chem. 2011, 54, 1266–1275

Journal of Medicinal Chemistry
ARTICLE
tography (petroleum ether/EtOAc 7:3) afforded anomers 22 (0.273 g,
65% yield over two steps, 3:1 R:β ratio) as a yellow oil. The following
signal attributions were assigned from the separated R,β mixture.
R-Anomer: MS (MALDI-TOF) C28H32O6 calculated 464.55, observed
465.5 [M þ H]þ, 503.5 [M þ K]þ. 1H NMR (400 MHz, CDCl3): δ
(ppm) 7.33-7.22 (m, 15H, OCH2Ph), 4.67-4.46 (m, 7H, OCH2Ph,
H-4), 4.16 (t, 2H, H-3), 4.09 (bm, 1H, H-5), 3.78 (dd, 1H, J = 11.9, 6.7
Hz, H-6a), 3.69 (dd, 1H, J = 11.9, 5.2 Hz, H-6b), 3.65 (dd, 1H, J = 10.7,
3.4 Hz, H-1a), 3.54 (dd, 1H, J = 10.7, 3.9 Hz, H-1b), 3.32 (s, 1H, OCH3),
1.78 (bs, 1H, OH). 13C NMR (100.57, CDCl3): δ (ppm) 137.86,
137.63, 137.42 (Cq Ph), 128.49-127.78 (OCH2Ph), 107.69 (C-2),
86.33, 82.37, 80.21 (C-3, C-4, C-5), 73.82, 73.40, 72.42 (OCH2Ph),
69.13 (C-6), 61.49 (C-1), 48.92 (OCH3). β-Anomer: MS (MALDI-
(MestreLab Research Srl). For spectra recorded in D2O, the HDO
chemical shift was calculated according to the equation δ = 5.060 -
0.0122T þ (2.11 ꢀ 10-5)T.34
2. Biological Assays. General. [3H]Flunitrazepam (MW unla-
beled at this specific activity 313.28, C16H12FN3O3), flunitrazepam-
[methyl-3H], has a specific activity of 80.0 Ci/mmol and was purchased
from International PBI SPA (Milan, Italy), American Radiolabeled
Chemicals, Inc. Flumazenil (MW 303.39, C15H14FN3O3), 8-fluoro-
5,6-dihydro-5-methyl-6-oxo-4H-imidazo[1,5-a][1,4]benzodiazepine-
3-carboxylic acid ethyl ester, was purchased from TOCRIS Bio-
science, Tocris Cooksoon Ltd. (Bristol, U.K.) Centrifugation assays
were performed on a centrifuge Biofuge Primo R-Heraeus and on an
ultracentrifuge Beckman XL-90. Filtration assays were performed using
Whatman GF/B glass fiber filters (L = 1-2 cm). Liquid scintillation
counting was performed on a Spectrophotometer Beckman LS65000
with 40% of efficiency for tritium and was used the scintillation fluid
UltimaGold F from PerkinElmer (Boston, USA). The investigations
using experimental animals were conducted in accordance with the
internationally accepted principles for laboratory animal use and
care as found in for example the European Community guidelines
(EEC Directive of 1986; 86/609/EEC). Results are expressed in terms
of percentage of control [3H]Flunitrazepam specific binding and
analyzed by GraphPad Prism using the Kruskal-Wallis ANOVA
for nonparametric data followed by Dunn's test for specific
comparisons.
1
TOF) C28H32O6 calculated 464.55, observed 465.5 [M þ H]þ. H
NMR (400 MHz, CDCl3): δ (ppm) 7.37-7.21 (m, 15H, OCH2Ph),
4.62-4.46 (m, 7H, OCH2Ph, H-4), 4.27 (d, 1H, J = 7.0 Hz, H-3), 4.20
(bm, 1H, H-5), 4.10 (dd, 1H, J = 13.0, 6.9 Hz, H-6a), 4.08 (dd, 1H,
J = 12.0, 6.1 Hz, H-6b), 3.72 (d, 1H, J = 11.9 Hz, H-1a), 3.54 (d, 1H, J =
11.8 Hz, H-1b), 3.30 (s, 1H, OCH3) 2.05 (bs, 1H, OH). 13C NMR
(100.57 MHz, CDCl3): δ (ppm) 137.86, 137.80, 137.74 (Cq Ph),
128.38-127.76 (OCH2Ph), 104.35 (C-2), 84.22, 83.67, 79.70 (C-3,
C-4, C-5), 73.41, 72.97, 72.54 (OCH2Ph), 71.80 (C-6), 62.53 (C-1),
49.57 (OCH3).
3-C-(3,4,6-Tri-O-benzyl-R-L-fructofuranos-2-yl)propene (23). To a
stirred solution of 22 (1 g, 2.15 mmol) in dry CH3CN (8 mL) was added
bis(trimethylsilyl) trifluoroacetamide (BTSFA) (0.568 mL, 2.15 mmol)
at 100 °C under reflux and argon atmosphere. After 3 h, analysis by thin-
layer chromatography indicated that starting material 22 had been
consumed. The reaction mixture was clear and uniform and was allowed
to cool to rt, and AllSi(CH3)3 (0.514 mL, 3.22 mmol) and TMSOTf
(188 μL, 1.07 mmol) were sequentially added at 0 °C and reaction was
left stirring at rt for 1 h. Water (12 mL) was added slowly to hydrolyze
TMS ethers, and then the mixture was neutralized with a 1 M aq solution
NaOH and concentrated in vacuo. The residue was extracted with
CH2Cl2, the organic layer was dried on sodium sulfate, filtered, and
evaporated. Purification by flash chromatography (petroleum ether/
EtOAc 8:2) afforded anomer 23 (0.775 g, 76% yield) as a yellow oil. The
Membrane Preparation. Membranes were prepared from rat cortex
as described by Ahboucha et al.33 The rat frontal cortex was homo-
genized, through a glass/Teflon potter at 1000 rpm, in ice-cold 0.32 M
sucrose, pH 7.4 (20 mL/gr of tissue), and the homogenate was
centrifuged at 2000g for 10 min at 4 °C. The supernatant was then
ultracentrifuged at 140000g for 30 min at 4 °C, and the pellet was
resuspended in Tris-HCl 50 mM pH 7.4 (2 mL/g wet weight) and again
centrifuged at 140000g for 30 min at 4 °C. After the final centrifugation,
the pellet was suspended in 1 mL of ice-cold Tris-HCl buffer (50 mM
pH 7.4) and was manually homogenized through a glass/Teflon potter
and stored frozen at -80 °C. On the day of the assay, the tissue was
thawed, and the pellet, resuspended in Tris-HCl in order to obtain a final
protein concentration of 2 mg/mL, which was used for displacement
binding studies.
1
8:2 R:β ratio was determined from integration of the crude H NMR
spectra. MS (MALDI-TOF): C30H34O5 calculated 474.59, observed
475.6 [M þ H]þ. Elemental analysis calcd (%): C, 75.92; H, 7.22.
Radioligand Binding Assay. [3H]Flunitrazepam (1 nM) served as
radioligand, and nonspecific binding was determined in the presence of
flumazenil (100 μM) and represented about 1-2% of the total binding.
The water-soluble benzodiazepine derivatives synthesized were diluted
in TrisHCl buffer 50 mM pH 7.4 to obtain the final concentration of
100 μM, whereas stock solutions of the benzodiazepine derivatives with
the benzyl group in the sugar were prepared in ethanol and then diluted
in TrisHCl buffer to obtain the final concentration of 100 μM. The
binding reaction consisted of 0.3 mg of membranes incubated with the
radioligand in the presence or absence of the test compounds for 90 min
at 4 °C. During this incubation period, the appropriate number of filters
were soaked in ice cold buffer. The binding reaction was then stopped by
rapid filtration under vacuum through a glass-fiber filter. The filters were
mounted in the filtration manifold, vacuum was applied, and an aliquot
of the first sample was applied to the first filter. Once the sample
compartment drains, the filter was washed immediately by the rapid
addition of 2 mL portions of ice cold Tris-HCl buffer (two times). After
being washed twice, the filters were dried under a heat lamp for 10 min
and then transferred to scintillation vials and dissolved in 8 mL of
scintillation fluid. Filter bound radioactivity was determined by scintilla-
tion counter, 20 min for each filter. The radioactivity values obtained
from the spectrophotometer in dpm were converted into percentages,
considering the value of specific binding 100%, given by the samples only
with the [3H]flunitrazepam. For all compounds, three independent
experiments were performed in triplicate.
1
Found: C, 76.10; H, 7.35. H NMR (400 MHz, CDCl3): δ (ppm)
7.37-7.24 (m, 15H, OCH2Ph), 5.78 (bm, 1H, H-20), 5.09 (d, 1H, J =
10.2 Hz, H-30a), 4.97 (d, 1H, J = 17.3 Hz, H-30b), 4.75-4.45 (m, 7H,
OCH2Ph, H-4), 4.13 (d, 1H, J = 7.7 Hz, H-3), 3.92 (bm, 1H, H-5), 3.70
(bm, 2H, H-1a, H-6a), 3.50 (m, 2H, H-1b, H-6b), 2.30 (dd, 1H, J = 14.0,
6.5 Hz, H-10a), 2.17 (dd, 1H, J = 14.2, 7.6 Hz, H-10b), 1.83 (bs, 1H, OH).
13C NMR (100.57 MHz, CDCl3): δ (ppm) 138.32, 138.20, 137.73 (Cq
Ph), 133.07 (C-20), 128.71-127.07 (OCH2Ph), 119.17 (C-30), 84.29
(C-2), 86.65, 83.50, 79.34 (C-3, C-4, C-5), 73.65, 73.19, 72.15, 69.61,
64.41 (OCH2Ph, C-6, C-1), 40.29 (C-10).
Molecular Modeling. Conformational analysis of compounds 10, 11,
and 31 was performed applying an MM2 force field in vacuum. Each
model was built with ChemDraw, energy was minimized (minimum rms
gradient = 0.010), and the dihedral angle C5a-C9a-N10-C11 value
calculated using Chem 3D. A molecular dynamics computation was
performed (step interval = 2 fs; frame interval = 1 fs; heating/cooling
rate = 1.000 kcal/atom/ps; target temperature = 300 K) monitoring the
same dihedral angle. The average value was calculated; in all cases, it
corresponded to that found by MM calculations.
DNMR Experiments. 0.5 mg of compound 10, 11, and 31 was
dissolved in 0.6 mL of D2O (293-363 K). An additional DNMR
experiment was performed in CD3CD2OD (173-293 K) for compound
10. 1H and selective 1D-NOESY (data not shown) spectra were
recorded at different temperatures and then processed with MestreNova
1273
dx.doi.org/10.1021/jm101244n |J. Med. Chem. 2011, 54, 1266–1275

Journal of Medicinal Chemistry
ARTICLE
’ ASSOCIATED CONTENT
Memory of Chirality Transformations. J. Org. Chem. 2005, 70, 1530–
1538.
(11) Cartier, P. R.; Zhao, H.; MacQuarrie-Hunter, S. L.; DeGuzman,
J. C. Enantioselective Synthesis of Diversely Substituted Quaternary 1,4-
Benzodiazepin-2-ones and 1,4-Benzodiazepine-2,5-diones. J. Am. Chem.
Soc. 2006, 128, 15215–15220.
(12) MacQuarrie-Hunter, S.; Carlier, P. R. Highly Enantioselective
Synthesis of Rigid, Quaternary 1,4-Benzodiazepine-2,5-diones Derived
from Proline. Org. Lett. 2005, 7, 5305–5308.
S
Supporting Information. Detailed experimental proce-
b
dures and analytical and spectral data for all intermediate and
final compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
(13) Bouhlal, D.; Godꢀe, P.; Goethals, G.; Massoui, M.; Villa, P.;
Martin, P. Synthesis and amphiphilic properties of glycosyl-1,4-benzo-
diazepin-2,5-diones. Carbohydr. Res. 2000, 329, 207–214.
(14) Bouhlal, D.; Martin, P.; Massoui, M.; Nowogrocki, G.; Pilard,
S.; Villa, P.; Goethals, G. Synthesis and epimerization of 10-N-(6-deoxy-
1,2:3,4-di-O-isopropylidene-D-galactopyranos-6-yl)-(11aS)-pyrrolo-
[2,1-c][1,4]benzodiazepin-5,11-dione. Tetrahedron: Asymmetry 2001,
12, 1573–1577.
(15) Abrous, L.; Jokiel, P. A.; Friedrich, S. R.; Hynes, J.; Smith, A. B.;
Hirschmann, R. Novel chimeric scaffolds to extend the exploration of
receptor space: hybrid beta-D-glucose-benzoheterodiazepine structures
for broad screening. Effect of amide alkylation on the course of
cyclization reactions. J. Org. Chem. 2004, 69, 280–302.
Corresponding Author
*Phone: þ46 (0)8 5537 8374. Fax: þ46 (0)8 5537 8468. E-mail:
acaraujo@kth.se. Address: AlbaNova University Center, Division of
Glycoscience, School of Biotechnology, Royal Institute of Technol-
ogy (KTH), 106 91 Stockholm, Sweden.
’ ACKNOWLEDGMENT
This work was supported by Fundac-~ao para a Ci^encia e
Tecnologia (FCT) SFRH/BD/17815/2004 and Ministerio
dell’Universitꢀa e della Ricerca under contract MIUR-PRIN
2005 no. 2005037725. We gratefully acknowledge Dr. Pietro
Fumagalli for his technical support for GABAA receptor binding
assays.
(16) Abrous, L.; Hynes, J.; Friedrich, S. R.; Smith, A. B.;
Hirschmann, R. Design and Synthesis of Novel Scaffolds for Drug
Discovery: Hybrids of β-D-Glucose with 1,2,3,4-Tetrahydrobenzo[e]-
[1,4]diazepin-5-one, the Corresponding 1-Oxazepine, and 2- and 4-Pyr-
idyldiazepines. Org. Lett. 2001, 3, 1089–1092.
’ ABBREVIATIONS USED
(17) Kumar, R.; Lown, J. W. Design, synthesis and in vitro cytotoxi-
city studies of novel pyrrolo[2,1][1,4]benzodiazepine-glycosylated pyr-
role and imidazole polyamide conjugates. Org. Biomol. Chem. 2003, 1,
3327–3342.
(18) Cipolla, L.; Araꢀujo, A. C.; Airoldi, C.; Bini, D. Pyrrolo[2,1-
c][1,4]benzodiazepine as a Scaffold for the Design and Synthesis of
Antitumour Drugs. Anti-Cancer Agents Med. Chem. 2009, 9, 1–31.
(19) Wright, W. B., Jr.; Brabander, H. J.; Greenblatt, E. N.; Day, I. P.;
Hardy, R. A., Jr. Derivatives of 1,2,3,11a-tetrahydro-5H-pyrrolo[2,1-
c][1,4]benzodiazepine-5,11(10H)-dione as anxiolytic agents. J. Med.
Chem. 1978, 21, 1087–1089.
(20) Cipolla, L.;Redaelli, C.; Nicotra, F. Synthesis ofa Spiro D-Proline
Analogue Bearing D-Fructose. Lett. Drug Des. Discovery 2005, 2, 291–293.
(21) Nicotra, F.; Panza, L.; Russo, G. Stereoselective access to alpha-
and beta-D-fructofuranosyl C-glycosides. J. Org. Chem. 1987, 52, 5627–
5630.
BTSFA, bis(trimethylsilyl)trifluoroacetamide; DBDMH, 1,3-di-
bromo-5,5-dimethylhydantoin; DMF, N,N-dimethylformamide;
DNMR, dynamic nuclear magnetic resonance; GABA, γ-amino-
butyric acid; MM2, molecular mechanics force field; MS, mass
spectrometry; SEM, standard error of the mean; rms, root-mean-
square; TfOH, trifluoromethanesulfonic acid; THF, tetrahydro-
furan; TMSOTf, trimethylsilyl trifluoromethanesulfonate; TLC,
thin-layer chromatography.
’ REFERENCES
(1) Kessler, R. C.; Berglund, P.; Demler, O.; Jin, R.; Merikangas,
K. R.; Walters, E. E. Lifetime Prevalence and Age-of-Onset Distributions
of DSM-IV Disorders in the National Comorbidity Survey Replication.
Arch. Gen. Psychiatry. 2005, 62, 593–602.
(2) Bateson, A. N. The benzodiazepine site of the GABA(A) receptor:
an old target with new potential? Sleep Med. 2004, 5, S9–S15.
(3) Rudolph, U.; Mohler, H. GABA-based therapeutic approaches:
GABA(A) receptor subtype functions. Curr. Opin. Pharmacol. 2006, 6,
18–23.
(4) Kalueff, A. V. Mapping convulsants’ binding to the GABA-A
receptor chloride ionophore: A proposed model for channel binding
sites. Neurochem. Int. 2007, 50, 61–68.
(22) Araꢀujo, A. C.; Nicotra, F.; Airoldi, C.; Costa, B.; Giagnoni, G.;
Fumagalli, P.; Cipolla, L. Synthesis and Biological Evaluation of Novel
Rigid 1,4-Benzodiazepine-2,5-dione Chimeric Scaffolds. Eur. J. Org.
Chem. 2008, 635–639.
(23) Araꢀujo, A. C.; Nicotra, F.; Costa, B.; Giagnoni, G.; Cipolla, L.
Fructose-fused β-butyrolactones and lactams, synthesis and biological
evaluation as GABA receptor ligands. Carbohydr. Res. 2008, 343, 1840–
1848.
(5) Macdonald, R. L.; Olsen, R. W. GABA(A) receptor channels.
Annu. Rev. Neurosci. 1994, 17, 569–602.
(6) Williams, D. B.; Akabas, M. H. Benzodiazepines induce con-
formational change in the region of the GABAA receptor a1 subunit
M3 membrane-spanning segment. Mol. Pharmacol. 2000, 58, 1129–
1136.
(24) Matsumoto, T.; Enomoto, T.; Kurosaki, T. Facile synthesis of
the next higher ketoses from aldoses. J. Chem. Soc., Chem. Commun.
1992, 610–611.
(25) Bessiꢁeres, B.; Morin, C. Iodomethyl Group as a Hydroxymethyl
Synthetic Equivalent: Application to the Syntheses of D-manno-Hept-2-
ulose and l-Fructose Derivatives. J. Org. Chem. 2003, 68, 4100–4103.
(26) Zhao, M.-X.; Shi, Y. Practical Synthesis of an l-Fructose-
Derived Ketone Catalyst for Asymmetric Epoxidation of Olefins
J. Org. Chem. 2006, 71, 5377–5379.
(27) Sugiyama, M.; Hong, Z.; Whalen, L. J.; Greenberg, W. A.;
Wonga, C.-H. Borate as a Phosphate Ester Mimic in Aldolase-Catalyzed
Reactions: Practical Synthesis of l-Fructose and l-Iminocyclitols. Adv.
Synth. Catal. 2006, 348, 2555–2559.
(28) Gizaw, Y.; BeMiller, J. N. Application of a phase transfer
reaction to the synthesis of L-fructose. Carbohydr. Res. 1995, 266, 81–85.
(29) Box, V. G. S. The Role of Lone Pair Interactions in the
Chemistry of the Monosaccharides. The Mechanisms of the Oxidations
(7) Blount, J. F.; Fryer, R. I.; Gilman, N. W.; Todaro, L. J.
Quinazolines and 1,4-benzodiazepines. 92. Conformational recognition
of the receptor by 1,4-benzodiazepines. Mol. Pharmacol. 1983, 24, 425–428.
(8) Simonyi, M.; Maksay, G.; Kovꢀacs, I.; Tegyey, Z.; Pꢀarkꢀanyi, L.;
€
Kꢀalmꢀan, A.; Otv€os, L. Conformational Recognition by Central Benzo-
diazepine Receptors. Bioorg. Chem. 1990, 18, 1–12.
(9) Maksay, G.; Tegyey, Z.; Simonyi, M. Central benzodiazepine
receptors: in vitro efficacies and potencies of 3-substituted 1,4-benzo-
diazepine stereoisomers. Mol. Pharmacol. 1991, 39, 725–732.
(10) Lam, P. C.-H.; Carlier, P. R. Experimental and Computational
Studies of Ring Inversion of 1,4-Benzodiazepin-2-ones: Implications for
1274
dx.doi.org/10.1021/jm101244n |J. Med. Chem. 2011, 54, 1266–1275

Journal of Medicinal Chemistry
ARTICLE
of Monosaccharides by Bromine, Chromium Trioxide in Acetic Acid,
and Ozone. J. Mol. Struct. 2001, 569, 167–178.
(30) Davis, F. A.; Ramachandar, T.; Chai, J.; Skucas, E. Asymmetric
synthesis of R-amino aldehydes from sulfinimine (N-sulfinyl imine)-
derived R-amino 1,3-dithianes. Formal synthesis of (-)-2,3-trans-3,4-
cis-dihydroxyproline. Tetrahedron Lett. 2006, 47, 2743–2746.
(31) Avdagiꢂc, A.; Lesac, A.; Majer, Z.; Hollosi, M.; Sunjic, V. Lipase-
Catalyzed Acetylation of 3-Substituted 2,3-Dihydro-1H-1,4-benzodia-
zepin-2-ones: Effect of Temperature and Conformation on Enantios-
electivity and Configuration. Helv. Chim. Acta 1998, 81, 1567–1582.
(32) Avdagiꢂc, A.; Lesac, A.; Sunjic, V. First example of the solvent
effect on absolute conformation of chiral 3,3-disubstituted 1,4-benzo-
diazepin-2-ones. Tetrahedron 1999, 55, 1407–1416.
(33) Ahboucha, S.; Araqi, F.; Layrargues, G. P.; Butterworth, R. F.
Differential effects of ammonia on the benzodiazepine modulatory site
on the GABA-A receptor complex of human brain. Neurochem. Int. 2005,
47, 58–63.
(34) Gottlieb, H. E.; Kotlyar, V.; Nudelman, A. NMR Chemical
Shifts of Common Laboratory Solvents as Trace Impurities. J. Org.
Chem. 1997, 62, 7512–7515.
1275
dx.doi.org/10.1021/jm101244n |J. Med. Chem. 2011, 54, 1266–1275