Download Computer-Assisted Molecular Modeling of Benzodiazepine and

Computer-Assisted
Molecular
Modeling of Benzodiazepine
and
Thyromimetic
Inhibitors of the HepG2
lodothyronine
Membrane Transporter
Laura Kragie, Maureen
Mary McCourt
L. Forrester,
Vivian Cody, and
Biological Sciences, Faculty of Natural Sciences and Mathematics
(L.K.)
State University of New York at Buffalo
Amherst, New York 14260
Molecular Biophysics (M.L.F., V.C.) and Electron Diffraction
Departments (M.M.)
Medical Foundation of Buffalo
Buffalo. New York 14203
T3 cellular uptake is inhibited
in the presence
of
benzodiarepines
(BZs). The structure-activity
relationship of BZ inhibition correlates strongly with halogen substitution
of the nonfused phenyl ring and
indicates
that this ring is required
for activity. A
structure-activity
series of thyromimetic
(TH) inhibitors of the HepGP iodothyronine
transporter
further
point out the critical importance
of the amino group
of the alanine side chain, its L-stereo configuration,
and the size of the substituents
of the inner and
outer phenyl rings. A third series of compounds,
reported to interact at related sites, were inactive as
HepG2 iodothyronine
transport inhibitors, and therefore the potent inhibitors were restricted to the BZ
and TH compounds.
Using both of these BZ and TH
structure-activity
series along with computer-assisted molecular
modeling
techniques,
we determined which chemical structural components
were
important at the transporter
interaction
site. By superimposing
structures from active chemicals,
excluding residues from poor inhibitors, and incorporating molecular electropotential
data, we developed
a five-point model of BZ conformational
similarity to
the endogenous
transporter
ligand, L-TJ: the alkyl
substitution
at the Nl of the BZ ring seems to simulate the alanine side chain of TJ, and the electronegative halogen and oxygen atoms of substituents
at M/R7/R2’/R4’
of BZ form a pyrimidal pharmacophore that seems to correspond
with the 3-1/5-l/
3’-l/4’-OH
substituents
of T3, respectively.
These
points, suggesting
a tilted cross-bow formation, may
be sites for ligand interaction with the iodothyronine
transporter.
(Molecular
Endocrinology
8: 382-391,
1994)
0ea9-eae9/94/o3a2~391$03.00/0
Molecular Endocrimy
Copyright 0 1994 by TIM Endocrine
INTRODUCTION
In this paper we compare the molecular conformations
of thyroid hormones to benzodiazepines
(BZs), a potential class of membrane iodothyronine
transporter
antagonists. Thyroid hormones have profound effects
on growth, differentiation,
maturation of tissues, and
the turnover of substrates, vitamins, and hormones. In
mammalian species, the active forms are Tq and its
more potent deiodinated metabolite, T3 (1). Facilitated
carrier-mediated
transport
allows the hormones
to
cross the cell membrane and then to interact with
cytosolic, enzymatic, mitochondrial,
microsomal, and
nuclear binding sites (2-4). The iodothyronine
membrane transport process has been studied in many
tissues, and these data show that membrane transport
of TJ is energy dependent, stereo-specific,
and critical
for cellular nuclear binding and metabolism of T3 (for
review see Refs. 5-7).
In a pharmacological
survey of compounds that may
interact with the membrane iodothyronine transporter,
it was shown that compounds of the BZ class, like the
thyromimetics (THs), are reversible inhibitors of the TJ
membrane transporter (8, 9). Because the structureactivity relationship (SAR) describing BZ inhibition of TS
uptake in the HepG2 cell line suggests a site that is
different from the central and peripheral BZ receptor
(BZR) sites, we hypothesized
that the BZs and THs
may directly interact at the TJ binding site on the
membrane transporter.
Using both of these SAM, in
conjunction with molecular modeling strategies, we developed a model for the structural comparison of iodothyronine transport inhibitors and now propose a pharmacophore for ligand interaction with the TJ transport
site.
society
382
lodothyronine
RESULTS
Transporter
Pharmacophore
AND DISCUSSION
in Table 1. This SAR of BZ inhibition reveals that the
nonfused phenyl ring is necessary for activity in our
series. The strongest correlation occurs at the 2’-position of that ring. Although there are no compounds in
our series differing only at R2’, comparisons between
similar compounds show that Cl substitutions have the
greatest activity, followed by F and H. In addition, a
halogen residue at the R4’ and an hydroxyl substitution
at R3 increases inhibitory activity. An alkyl group at Rl
or on the imidazole/triazole
group of the 1,2-annelated
BZ series enhances potency. For the R7 group, a Cl is
preferable over an NO*. However dihalogenated
compounds that are Cl substituted at R7 and R2’ are slightly
less potent than the compound that is Cl substituted at
R2’ and R4’ (Ro22 8349); monohalogenated
compounds Cl substituted at R7 vs. R4’ (Ro5 5115 vs.
diazepam) have equal potency. The carbonyl substituent at R2 of the 1,4-BZ series also enhances potency.
In the resulting SAR, the importance of the halogensubstituted nonfused phenyl ring is highlighted;
this
SAR seems unique when compared to those series
reported for other BZ sites and/or effects (1 O-l 2).
In addition to the BZ SAR series, we studied a limited
series of TH inhibitory compounds. TH compounds can
act as substrates for the carrier protein and are transported into the ceils. These analogs have different
SARs
Figure 1 diagrams the structure and nomenclature
of
the prototypic classical 1,4-BZ and L-T~ compounds.
Table 1 lists the substitutions
at sites within the BZ
molecules of our BZ SAR series. The BZ potencies for
inhibition, illustrated by -Log IC& values, are also listed
R3
I
HO
T3
i
Fig. 1. Structure
Table 1. Structural
and Nomenclature
Components
BZ
Triazolam
Lormetazepam
Prazepam
(31)
Lorazepam
(32)
Ro22 8349
Midazolam
Delorazepam
Ro5 4864
Temazepam
(33)
Flurazepam
Ro5 5115
Oxazepam
(34)
Diazepam
(35)
PK 11195
Clonazepam
(36)
Alprazolam
Nordiazepam
Medazepam
(37)
Estazolam
(38)
Chlordiazepoxide
(39)
Flunitrazepam
(40)
Nitrazepam
(41)
Bromazepam
(42)
Ro5 3663
Flumazenil
of BZ and L-T~
of BZR
Ligands
Rl
fused
methylated
CH3
CHPcyclopropyl
H
CH3
fused methylated
H
CH3
CHz
CH&H2N(CH&H&
CH3
H
CH3
H
H
fused methylated
H
fused
fused
triazolo
ring
imidazo
R2
R3
ring
0
0
0
0
ring
0
0
0
0
0
0
0
b
H
OH
H
OH
H
H
H
H
OH
H
H
OH
H
H
H
H
H
H
H
H
H
H
H
H
H
0
ring
0
H
CH3
triazolo ring, without CH3
H
NHCHJ
0
CH3
H
0
H
0
H
0
imidazo ring + COO(C2H5)
triazolo
Iv
Cl
Cl
Cl
Cl
H
Cl
Cl
Cl
Cl
Cl
H
Cl
Cl
H
Non
Cl
Cl
Cl
Cl
Cl
Non
NOz
Br
H
F
R2'
Cl
Cl
H
Cl
Cl
F
Cl
H
H
F
H
H
H
Cl
Cl
H
H
H
H
H
F
H
Nat2’
no C-ring, R5=CH3
no C-ring, R4=CH3,
R4'
H
H
H
H
Cl
H
H
Cl
H
H
Cl
H
H
H
H
H
H
H
H
H
H
H
H
R5=0
-Log
7.38
7.35
6.69
6.54
6.38
6.25
6.12
5.91
5.83
5.82
5.60
5.53
5.50
5.05
4.95
4.94
4.86
4.44
4.37
I&o
+. SEM
+ 0.17
* 0.07
+ 0.09
+ 0.05
+ 0.12
+ 0.06
+ 0.03
k 0.08
f 0.14
f 0.10
f 0.14
f 0.08
f 0.10
f 0.14
+ 0.33
f 0.16
+ 0.16
f 0.14
+ 0.05
<4.3
<4.3
t4.3
<4.3
inactive up to
inactive up to
(N)
(6)
(5)
(5)
(4)
(5)
(4)
(3)
(8)
(3)
(4)
(3)
(3)
(5)
(5)
(4)
(3)
(3)
(3)
(3)
(3)
(5)
(4)
(3)
10 /IM
10 PM
Structurally
analogous
residues numbered
according
to the nomenclature
of classic 1 ,+BZs,
pictured in Fig. 1. I& values listed
as negative log of molar concentrations.
Portions of this table were previously
published
(7). Crystal structure
references
given in
parentheses.
’ N-oxide at position 4.
’ CON(CH3)CH(CH3XC2HS)
and PK 11195 is an isoquinoline
carboxamide
instead of a benzodiazepine.
MOL END0.1994
Vol8 No. 3
384
potencies with regard to their abilities as intracellular
THs, and therefore they may be considered as possible
partial agonists/antagonists
of L-T~ intracellular receptors. Structural substitutions, rank order, and -Log lCs0
values of compounds from this SAR are listed in table
2. The proposed endogenous
ligand/substrate
for the
transporter, TJ, is the most potent of our series. Some
of the Go values must be interpreted cautiously. Because of negative feedback from intracellular hormone
concentrations,
these nonzero experimental conditions
generate IGo values that do not necessarily reflect the
affinity of the protein for the analog (13). In addition, the
potency reported for D-T~ (Sigma Chemicals, St. Louis,
MO) could possibly be inflated due to its probable
contamination with L-T,+ The bromine and chlorine substituted SKF compounds (SKF 94901, SKF 94424, and
SKF 94918) may also be subject to less intracellular
dehalogenation
when compared to the iodinated compounds (SKF 93236, SKF 95050, and SKF 94690). The
potency of 3,5dimethyl,3’-isopropyl-L-T3
(DIMIT) also
could be inflated due to its probable lack of metabolism
by native deiodinases.
The decrease in potency of
reverse TB (rT3) in the 15 PM (*T3) assay relative to the
40 PM (*T3) assay (Table 2) may indicate an inflated
potency due to the probable metabolism of rTB into
3’,3-T2 by the 5’deiodinase.
Indeed, rTB and 3’,3T2
have similar I& values. On the other hand, 3,5-T*
might be less metabolized by the 5deiodinase
relative
to TI, and hence its potency too could potentially be
inflated.
Nevertheless, a rank order can be useful in defining
Table
SARs. The 3’-arylmethyl-substituted
SKF compounds
define the upper size limits of the 3’-substitution
of the
iodothyronines.
The necessity for electronegative
substitutions at the 3’,3,5 sites is not absolute as indicated
by the moderate potency of DIMIT. Overall, the SAR
rank order derived from this brief TH series suggests
the importance of the halogen substituents at the 3’,5,3
sites, the size of the 3’ substituent, the L-configuration
of the amino acid side chain carbon, a sterically restricted zone at the 5’ site, the presence of the side
chain amino group, and the alkyl length of the alanine
side chain.
A number of non-BZ and non-TH compounds that
are aromatic or known to interact with either BZRs or
thyroid hormone systems (9) were also screened for
inhibitory activity. For example, thyroid hormone disorders in rats are associated with changes in cardiac
dihydropyridine
binding site density (14, 15) and large
amounts of phenytoin and bromosulfophthalein
are reported to reduce TJ uptake in cells (for review see Ref.
9). Furthermore, higher concentrations
of BZs can interact with neurotransmitter
receptor systems (16, 17).
Table 3 lists the compounds and their range of assayed
concentrations.
These drugs were all inactive and therefore demonstrate
that, of the chemicals tested, high
potency in inhibiting [‘251]T3 cellular uptake so far is
limited to the THs and the BZs.
Molecular
Modeling
The BZ structure is a nonplanar, seven-member,
heterocyclic ring fused to a planar phenyl ring and a
2. Structural Components of THs
TH
-Log
3'
4'
5'
6'
5
3
I’&
15 PM *Tz
L-Ts(43)
L-3,5-Tz(44)
L-DIMIT
L-rT3(45)
L-3,3'-T2
Triac"
L-T,
Tetrac (46)
D-T;
SKF L-94690
SKF L-94918
SKFL-93236
SKFL-94424
SKF L-94901
SKF L'-95050
3,5-T* proprionic acid
(47)
D-Tg
L-3Sdiiodotyrosine
L-thyronine
D,L-thyronine
H
Isopropyl
H
I
I
I
I
CH*-pyridazinone
CHa-pyridazinone
CH&-OH benzyl
CHP-pyridone
CH2-pyridazinone
CH,-pyridone
H
I
OHHHI
OHHHI
OH H
OH I
OH H
OHHHI
OH I
OH I
OHIHI
OHHHI
OH H
OHHHI
OH H
OH H
OHHI
OHHHI
OH
No outer
H
OH
H
OH
’ Significant contamination with L-T,. Crystal structure
H
H
H
CHB
H
H
H
H
I
I
H
Cl
H
H
Br
Br
I
I
I
CHB
I
I
I
I
I
I
I
Cl
I
Br
Br
H
I
L-alanine
L-alanine
L-alanine
L-alanine
L-alanine
Acetic acid
L-alanine
Acetic acid
o-alanine
L-alanine
L-alanine
L-alanine
L-alanine
L-alanine
L-alanine
Proprionic
HI
I
D-alanine
ring
I
I
H H H
H
L-alanine
H H H
H
D,L-alanine
references given in parentheses.
H
+ SEM (N)
Side chain
6.61
5.28
5.28
4.77
4.68
4.67
6.80(2)
+ 0.20(4)
6.20(2)
6.08 (2)
6.02 (2)
5.42 (2)
+ 0.28(2)
5.19 (2)
+ 0.28(2)
4.93 (2)
k 0.17 (3)
f 0.19 (3)
+ 0.15 (3)
4.60 (2)
4.36 (2)
3.76 (2)
C4.3
5.67 + 0.18 (3)
4.99 f 0.03 (3)
(2)
L-alanine
C4.3
C4.3
40 DM ‘Ts
6.76 f 0.07 (3)
C4.3
(2)
(2)
(2)
lodothyronine
Table
Transporter
3. Screened
Compounds
Drug
BAY K 8644
Nisoldipine
Diltiazem
PN 200-l 10
Theophylline
Adenosine
Prazosin
Phentolamine
Benztropine
Yohimbine
385
Pharmacophore
Which
Did Not Inhibit TJ Uptake
Concentration
range
1 nr+0.1
0.1
0.1
0.1
1opM
0.1-100
lo/.~M
0.5
0.5-10
l-10
1 IlM-1
Drug
PM
/lM
PM
fih4
Quercetin
Rutin
Glyburide
Phenylephrine
Ciproheptidine
Diphenhydramine
Chlorpheniramine
Nomifensine
Protoporphyrin
IX
Bromosulfophthalein
Phenytoin
/.tM
PM
PM
PM
PM
nonfused planar phenyl group substituent. The nonfused phenyl ring is restricted in its freedom of rotation
by the steric interactions with the R7 halogen substituent, resulting in an orthogonal position relative to the
diazepine ring system. The typical 1,4diazepine
conformation is a boat with the C3 atom as the bow point
(Fig. 2). Although this ring has some conformational
flexibility, it does have conformational
stability due to
ring annelation and the presence of the C=N bond.
Given this conformational
flexibility, we sought a minimum energy conformation
and the compounds were
minimized using the MOPAC AM1 molecular orbital
method on SYBYL. Overall there is good agreement
between the minimized structures and the initial crystallographic data. To explore the range of flexibility of
rotation of the nonfused phenyl ring, the C-rings of a
few of the minimized BZ structures were rotated 0’ to
180’ in 10” increments without reoptimization.
The
results of these calculations suggest that these molecules have a relatively flat potential surface, between
torsion angles 90” and 140°, consistent with conformational flexibility within this range.
T3 is an iodine-substituted
tyrosine joined via an ether
linkage to an iodine-substituted
phenol. The ether linkage (with an angle of 120’) fixes the outer phenolic ring
to an orthogonal position relative to the plane of the
inner tyrosyl ring; minimal steric hindrance occurs when
the outer phenolic ring is coplanar and the inner tyrosyl
Fig. 2. Two Perpendicular
Views Illustrating
a Typical
Conformation
The 1,4-diazepine
ring system has a boat conformation.
BZ
Concentration
range
0.5
0.1-100
0.1
1OfiM
lo/.~M
1orM
lOjIM
0.5
0.1-10
0.01-10
0.01-10
/LM
PM
PM
/.LM
/.IM
yM
PM
ring is perpendicular
to the plane of the ether linkage.
The conformation of the outer ring substituents can be
distal (away) or a proximal (near). The alanine side chain
lies perpendicular
to the inner tyrosyl ring; therefore the
alanine side chain and the outer phenyl ring can lie on
the same (cisoid) side of the inner tyrosyl ring or the
opposite (transoid) side (18). Figure 3 illustrates a prototypical L-T~ cisoid conformation
in which the ether
linkage fixes the outer phenolic ring to an orthogonal
position relative to the plane of the inner tyrosyl ring,
and the 3’-iodine is proximal relative to the alanine side
chain. This is the conformation
of TJ used for our
molecular comparisons.
Initially, conformational
comparisons were made between triazolam, the most active BZ of the 1,2-annelated series, with the less active estazolam. Figure 4A
highlights the regions that are unique to the active BZs,
as determined by excluded volume analysis. Here, the
R2’-Cl and the triazole-CH3 are the regions important
for inhibitory potency. In Fig. 48, lormetazepam,
one of
the most active of the 1,4 diazepine analogs, is superimposed over the less active medazepam.
Again, the
analysis highlights the halogen at R2’ and the methyl
group at Rl; it also points out the R2 carbonyl on
Fig. 3. A Typical L-T3-cis Conformation
Where the Ether Linkage Holds the Outer Phenolic Ring to an Orthogonal
Position
Relative to the Plane of the Inner Ring
The 3’-I is shown in the proximal
conformation;
the distal
conformation
would result if the outer, phenyl ring were rotated
180’. The compound
shown is uncharged,
as would be the
predominant
species at pH 7-8.
MOL
388
ENDO.
Vol8
1994
active 1,2-annelated
BZ, paired with a less active BZ, estazolam. Two stereo views of the superimposition
of triazolam
and estazolam,
excluding
their inactive
VdW volumes
and
displaying
their active VdW volumes.
Note those regions which
are unique to the active BZ. 8, Conformational
comparisons
of lormetazepam,
the most active of the 1 ,Cdiazepine
analogs,
with the less active BZ, medazepam.
Two stereo views of the
superimposition
of lormetazepam
and medazepam
where their
inactive
VdW volumes
are excluded
and the active
VdW
volumes
are displayed.
Note those regions
which are unique
to the active BZ. The hydrogens
on the nonfused
phenyl rings
of lormetazepam
and medazepam
were undisplayed
for clarity.
C, Superimposition
of triazolam,
estazolam,
lormetazepam,
and medazepam.
These stereo views show the union of the
two active BZs, triazolam
and lormetazepam,
minus the union
of the two less active of the BZ structures,
estazolam
and
medazepam.
The active residue volumes
are highlighted.
The
hydrogens
on the nonfused
phenyl rings of lormetazepam
and
medazepam
were undisplayed
for clarity.
lormetazepam and the buckling of the A-ring. Figure 4C
combines both the 1,4- and the 1,2-BZ comparisons to
show the baskets of Van der Waals (VdW) volumes
used as our template for the initial fit of TB.
To explore the outer ring region of THs, we compared
TB, the most active TH, to the less active compound,
No. 3
T4 (Fig. 5). Conversely, here the excluded volume analysis subtracted the active structure from the inactive
compound, in order to illuminate regions of Tq that are
not in common with TB. The 5’-substituent
highlighted
by this comparison may account for T4’s reduced potency; this suggests that there is a restricted area
beyond the 5’-H of TS.
We went on to compare the electronegative
properties of BZs and THs. In Fig. 6A, L-T~ is compared with
DIMIT, a compound with slightly less potency than LTB. The isopotential map is nearly identical to that of LT3. When we compared
L-T~ to D-T~, we used their
charged side chain conformations
in order to highlight
differences
in the amino acid side chain. Figure 6B
shows the isopotential maps of this pair of stereoisomers and illustrates the significant difference in the
electronegative
profile with the L- vs. o-configurations.
In Fig. 6C we have compared lormetazepam
to L-T~.
Note that, except for the lack of an electronegative
group at BZ’s R4’, the isopotential maps are similar. In
Fig. 6D, we compare triazolam with L-T~ in order to
determine the optimal alignment of the alanine side
chain relative to the triazole group at Rl-R2. This
comparison
shows that the nitrogens in the triazole
group are electronegative
like the R2 carbonyl group in
the 1,4-BZ series and that it is best aligned between
the carboxyl group of the alanine side chain of TJ and
one of its inner ring iodines (3-l). Therefore, when we
ade our fit, we did not alter the rotation of the Cl2 bond from the minimized conformation.
Due to the buckling of the C3 bow of the B-ring,
-ere
is an asymmetry to the BZ molecular structure.
We displayed and compared BZ images of structures
from databases composed of the inverted as well as
the noninverted crystal structures of the BZs and compared their steric volumes and isopotential
maps to
both the transoid and cisoid conformers of L-T~. Figure
7 illustrates our best model for the superimposition
of
the cisoid, proximal conformation of L-T~ and the active
BZs as represented
by lormetazepam
and triazolam.
Based upon the superimposition
of active structures,
combined with exclusion of residues from poor inhibitors that are not in common with the active compounds,
we suggest a model for BZ structural similarity to the
cisoid conformation
of L-T~. Specifically we point out
five overlapping molecular substituents: 1) R2’ halogen
residue occupies the same density as the proximal
iodine on the outer phenolic ring of T3; 2) R4’ halogen
residue occupies the same position as the T3’s outer
4’-OH; 3) R3-OH on the diazepine ring fits the density
Fig. 5. Two Stereo Views of the Conformational
of TB. the Most Active TH with the Less Active
The VdW volume unique to Tq is displayed.
Comparison
Compound,
T.I
lodothyronine
Transporter
387
Pharmacophore
D
Fig. 6. Isopotential
Triazolam
vs.
Maps
Comparing:
A, DIMIT
vs.
L-T~;
B,
D-T~
structural
homology.
B, Same
view
as A but
with
the
halogen and oxygen VdW volumes highlighted. The points for
the RMS
fit were
(TH to lormetazepam):
C4’
to C4’,
Cl’
to
Cl ‘, 5-l to C7, 3-l to R3-OH, alanine chiral carbon to Rl
methyl.
L-T~
(charged
conformers);
C, Lormetazepam
vs.
L-T~;
and D,
L-T~
ig. 7. A, Stereo views of the model for superimposition
of LT3 and the active BZs, lormetazepam and triazolam, revealing
their
vs.
...:‘<.T
..,,
The resulting
RMS fit is 0.838
A.
of the 3-l of T,‘s inner ring; 4) R7 halogen residue
resides in the same position as the 5-l of T,‘s inner ring;
and 5) Rl alkyl or triazole groups align along the amino
acid side chain of TI with the more electronegative
COOH oriented toward the R2 carbonyl (1,4-BZ) or
tertiary amine (triazole/imidazole),
and the Rl alkyl
groups oriented toward T,‘s primary amine (aminoacid
carbon). However, the cisoid distal (phenyl ring rotated
180’) conformation
of L-T~ also can be fitted just as
well to BZ structures with their nonfused ring rotated
180’. This conformation
is slightly less energetically
favorable for the BZs and slightly more favorable for
the TB molecule. Our model must and does accommodate this alternative conformation. The mirror (inverted)
image to the BZ molecular conformation
we have displayed also fits the image of cisoid, proximal L-T~ that
we have displayed in our figures. This fit requires flipping of the L-T~ molecule and instead aligning the
opposite, symmetric inner ring iodine to the R7 or R3
groups of this inverted BZ molecule. In order to keep
the isopotential maps consistent, the amino acid carbon
of L-T~ must be rotated about its axis. This resultant
five-point RMS fit is 1.3 A.
We can also apply other examples from the SAR
rank order to this model. The comparison of D-T~ vs. LT3 examines the effects of the stereoisomerism
of
charge/steric
hindrance. Are the charge effects of the
alanine side chain a necessity or just an enhancement
when compared to the analogous
BZ methyl group?
According to the SAR, L-T* is quite active, but the
deaminated compound, TP proprionic acid, has greatly
MOL
END0.1994
VoI8
No. 3
388
diminished activity. Triac, the deaminated and demethylated acid derivative of T3, shows a moderately diminished activity. Therefore the influence of NH3 on the
chiral carbon is crucial, especially when there are only
two other potential contact points in the pharmacophore. In our model, this amino group aligns with the
methyl groups at Rl, while the carboxyl group aligns
toward the BZ electronegative groups at R2 and R3.
The CH,-cyclopropyl
group on prazepam (see Fig. 8)
is similar in length to the amino acid side chain of L-T~.
and this similarity may explain prazepam’s
potency.
However, the large alkyl groups on flurazepam and
PKll 195, compounds with weaker potencies may indicate the size limitations of the side chain receptor
pocket. The 3’-arylmethyl
substitutions
of the SKF
compounds, combined with the 3’-isopropyl
group of
DIMIT, delimit the size of the TH 3’pocket.
DIMIT’s electrostatic potential map (see Fig. 6A) demonstrates that the electronic properties of two molecules can be similar even though the steric maps may
not. This raises the question of the importance of TH’s
central core of electronegativity
for activity (perhaps
more so than steric factors) at the 35 and 3’ sites.
However, DIMIT might have better potency than other
compounds due to its probable resistance to metabolism. Although the potencies of the SKF compounds
are similar, there may be a slight enhancement
with
halogen substitution at the inner ring 35 positions, the
order following I > Cl > Br. The presence of large
electronegative groups in the outer ring region opposite
the BZ’s R2’ group severely decreases potency, as
demonstrated
by L-T~, rT3, and SKF 905050 (see Fig.
5). These data suggest that the distal end of molecule
requires a very precisely oriented and sized fit into the
active transporter site.
The compounds diazepam, Ro22 8349, Ro5 4864,
and Ro5 5115 demonstrate the importance of the halogen sites of BZ in our pharmacophore
criteria. Compounds Cl substituted at R7 and R2’ have slightly less
potency than compounds
Cl substituted at R2’ and
Fig. 8. Stereo Views of the Model for Superimposition
of L-T~
and the Active BZ, Prazepam.
Revealing
the Homology
of the
CH,-Cyclopropyl
Substituent
at Rl to the Alanine Side Chain
on L-T~
Before fitting structures,
the torsion angle of the nonfused
phenyl ring of prazepam
was rotated
from -30”
to -55”
(these angles have equal heats of formation).
Components
fitted were (TH to BZ, respectively):
4’.OH to 4’-H, Cl’ to
Cl ‘, 5-l to C7. C2 (TH inner ring) to C2 (diazepine
ring), alanine
CH, to CH? of Rl group, and alanine chiral carbon to base C
of the Rl cyclopropyl
structure.
The resulting RMS fit is 1.02
A.
R4’; compounds
Cl substrtuted at R7 vs. R4’ have
equal potency. Hence the 2’-halogen
may be slrghtly
more influential than the R7 halogen In the dihalogenated system, and therefore this position was allowed
to be frt less precisely than the other proposed pharmacophore sites.
CONCLUSION
These SAR data and the molecular modelrng of crystallographrcally
derived structures suggest a pharmacophore for the T3 site of the putatrve membrane iodothyronine transporter
of HepG2 cells. To produce a
compound wrth moderate (0.1-l PM) inhibitory activity,
at least three of the following structural components
are needed (BZ/TH. respectively): 1) R2’-halogen/outer
ring iodine; 2) R4’-halogen/outer
ring 4’-OH; 3) R3-OH/
inner ring 3-l; 4) R7-halogen/inner
ring 5-1, 5) Rl alkyl
group or tnazole group/amino
acid side chain with Lconfiguration of the primary amine, and 6) no halogen
at the 5’-outer ring positron of TH. Potentially good BZs
for T3 uptake inhibitors may include modifications of the
BZ triazolam such as an additional 4’-hydroxyl group,
a halogen (Cl or I) or a hydroxyl at R3, and an amino
group on the tnazole’s methyl. To improve on lormetazepam, an alanine side chain could be put at the Rl
positton consistent with the L-configuration,
as well as
adding an OH at the 4’-position of the nonfused phenyl
ring. BZs likely to be ineffective would have a nonfused
phenyl rrng with a 3’- or 2’-halogen in addition to a 5’or 6’-halogen along with substitution of either a halogen
or an OH at R3. These substitutions then would fix the
nonfused ring into an orthogonal
conformation
and
thrust the transord-positioned
halogen into the restricted zone. As well, an alanine side chain In the Dconfiguration at positron Rl or R2 would render the BZ
ineffective.
The binding and TH pharmacophorefor
the T3 nuclear
receptor emphasize the importance of a precise fit of
analogs to the receptor actrve site (19-22). Ion pairing
to the receptor is suggested by the ionrc side chain on
one end and a polar phenolic-OH
on the other. The
nuclear receptor. though, does not distinguish between
the L- or o-configuration
of the side chain. The 3,5
substrtuents define the size of the lipophilic pockets as
well as the stenc constraints that hold the conformation
of the aromatic rings; binding is enhanced by polanzable
groups. A lipophilrc residue of limited size IS required at
the 3’-site and must be capable of being positioned
distal to the nonphenolrc inner ring; here an isopropyl
group binds better than an Iodine. The similarities of
our model to that of the T3 nuclear receptor pharmacophore include the restrictive zone at the 5’-position
(although our SAR seems more restrictive) and increased activity with halogen substttution at 3’,3,5positions. Our differences Include the improvement of
inhibitory activity with halogen substitutron at the 3’,3,5sites relative to substitution with alkyl groups and the
lodothyronine
Transporter
Pharmacophore
absolute requirement
of L-configuration
at the chiral
amino acid carbon. We did not test the importance of
the 4’OH site or ether linkage substitutions in our brief
SAR series, but the outer ring 4’-OH of TS does overlap
the 4’-halogen of BZ in our model of BZ/TH homology.
As well, our 3’-I is fit to the BZs while in the proximal
position, although the distal conformation can just as
easily fit BZs with a rotated C-ring. And as the SAR
and isopotential map comparisons
show, DIMIT and
triac are not the best ligands for the iodothyronine site,
but they are high affinity ligands for the nuclear receptor
(20). Finally, TJ binding in isolated rat hepatic nuclei
show no competition from up to 10 @I triazolam and
lormetazepam
(8). Perhaps the nuclear receptor requires a more precise fit of halogens at the 35 and 3’
positions when compared to the transporter site.
BZs are a class of pharmaceutical
agents that are
primarily used as sedative-hypnotics
and anxiolytics.
The BZRs for these central nervous system actions are
contained in the r-aminobutyric
acid receptor chloride
ion channel complex (23, 24). At least three subtypes
of BZRs have been described that exist in different
brain regions and subserve different physiologic functions (25, 26). There is also a BZR that is not linked to
the y-aminobutyric
acid-BZ-Cl ion channel complex,
termed the “peripheral BZ receptor” (10, 23, 24) although it is present in the central nervous system.
Several models of the central BZR pharmacophore
have been proposed.
Common features of these
models include two proton-accepting
groups separated
by approximately 3.0-3.5 A (11). The condensed aromatic ring is also a common feature for all high affinity
ligands, but it may not be involved directly in the binding
site (11, 12). For inverse agonists of the central BZR,
Allen et al. (27) suggest an aromatic ring system constrained to one plane and containing two hydrogen
binding sites; the antagonists have long substituents
that access regions above or below this aromatic plane.
Of the compounds with classic BZ structure, central
BZRs show stereoselectivity,
are not affected by the
R-l alkyl groups, and compounds with 4’-substitutions
have diminished affinity (28). The peripheral BZR is
distinguished by such varied selective ligands as 4’-Cl
diazepam, the isoquinoline carboxamide
derivative PK
11195, and protoporphyrins,
all of which are devoid of
activity at the central BZR (10). Of the classical BZ
compounds, the peripheral site prefers an alkyl group
of three carbons or less at position 1, a carbonyl at
position 2, a 4’-halogen or methoxy group, and halogens at R2’ and R7; substitution at R3 or R4 reduces
affinity (28).
Our pharmacophore
model shares some characteristics from both the central and peripheral BZR pharmacophores. Our model for the iodothyronine
transporter includes the planar aromatic ring system with
electronegative
substituents.
However, it requires an
additional halogen-substituted
phenyl ring oriented perpendicular to the plane of the aromatic system, and our
sites for the electronegative
substitutions include both
the central BZR-preferred
(7, 2’) and the peripheral
369
BZR-preferred
(4’) positions, as well as at the central/
peripheral BZR-preferred
R2, and the peripheral BZRaversive R3 position (although we have not yet determined its stereoselectivity for this substitution). Like the
peripheral site, we limit alkyl substitution at Rl to one
to three carbons in our model.
The alkyl substitution of BZ may simulate the alanine
side chain of TJ in our model of fit. Electrostatically,
the
substitution
at the BZ Nl may delocalize the alkyl
electrons into the ring system and allow for greater
nucleophilicity at substitutions in the aromatic ring system. As well, the necessity of the carbonyl at R2 allows
for an amide linkage to Nl and prevents its protonation
and the localization of the alkyl electrons (28). The highly
electronegative
halogen and oxygen atoms at R3/R7/
R2’/R4’ of BZ form a pyrimidal pharmacophore
that
seems to correspond
with the 3-l/5-l/3’-l/4’-OH
substitutions of TS. These five points, defining a tilted crossbow formation, may provide proton acceptor sites (such
as protonated lysine residues) for interaction with the
iodothyronine
transporter. The highly sensitive stereospecificity of the TH chiral carbon may suggest hydrogen bonding and/or ion pairing to the active site at the
side chain (BZ-Rl) portion of the molecule.
MATERIALS
AND METHODS
Materials
Tissue culture
media and supplements
were from Sigma
Chemicals and GIBCO (Grand Island, NY). Our source for [‘251]
TB, specific activity of 2200 Ci/mmol.
was New England Nuclear, DuPont (Boston,
MA). BZs were either purchased
from
Sigma (St. Louis, MO) or generously
supplied from Drs. Peter
Sorter (Hoffmann
La Roche, Nutley, NJ) or David Triggle (State
University
of New York, Buffalo, NY). SKF compounds
were a
generous gift from Dr. Tony Underwood
(Smith Kline Beecham,
Welwyn,
England).
The source for the THs (highest
purity
available) L-T~, L-rT3, L-T~, L-T~, D-T~, and DIMIT was Henning/
Berlin (Berlin, Germany)
and for D-T,, 3,5-~-T,,
diiodotyrosine,
D,L- and L-thyronine
was Sigma Chemicals.
3,3,-~-T*
was
originally from Dr. E. C. Jorgensen.
Our HepG2 cell line was a
gift from Dr. B. Knowles. All other chemicals
were from Sigma
and Baker (Phillipsburg,
NJ).
Uptake
Assay
and lCso Values
for Inhibition
Our procedure
to screen for inhibitors
of labeled T3 uptake is
as described (8,9). HepGP cells were grown in multiwell cluster
plates in Dulbecco’s
modified Eagle’s media low glucose with
10% fetal bovine serum. Cell suspensions
were plated into 12to 24-well plates in equal volumes containing
1 04-1 O5 cells per
well and grown to confluence
(l-2 x 1 O6 cells/cm*)
in 4-14
days. Before assay, wells were filled with Hank’s balanced
salt solution plus the drugs or solvent control. To initiate the
assay, [‘251]T3 was added to make a final concentration
of 15
40 PM. At termination
of the assay (30-60
min), wells were
washed with a cold glycine buffer, hydrolyzed
in NaOH. and
then transferred
to vials for counting
in a r-counter.
Each well
is one of triplicates
for each drug or solvent dose. Temperature-dependent
uptake of [‘251]T3 was determined
by subtraction of uptake at 4 C from the uptake at 22 C. The nonspecific
binding/uptake
was defined by uptake at 4 C incubation.
I&
values were derived from dose-response
curves of individual
experiments
using a standard
set of pharmacological
statistical
programs
(29) implemented
on an IBM personal computer.
MOL
390
ENDO.
Vol8
1994
Crystallographic
Databases
A search of the Cambridge
crystallographic
database
(Quest
3D, Database
5.04, October
1992, Medical Foundation,
Buffalo, NY) revealed crystal structures
for 12 of the 25 BZs from
our structure-activity
series. Structures
without available crystal data were created from modifications
of the most homologous published BZ crystal structure
using the SYBYL software
package
(Tripos Associates,
St. Louis, MO). For example,
triazolam
was built from estazolam
and lormetazepam
from
lorazepam.
We used the thyroid analog crystal structures
L-T~
and L-IT3 to build the remaining
analogs of our TH series. All
structures
were energy minimized
by using the MOPAC
AM1
molecular orbital option in SYBYL on a Silicon Graphics
work
station (4D70GTB).
We also explored
the potential
energy
surface for rotation of the fused diazepine
ring (A/B-rings)
and
the nonfused
ring (C-ring).
To do this, we started with the
minimized
structure and rotated the torsion angle between the
fused A/B-rings
and the nonfused
C-ring in 10” increments
form 0” to 180”. At each 10” increment
a single point energy
was calculated,
without reoptimization
of the structure.
Least-Squares
dispensable
technical assistance.
Dr. Kragie also thanks Drs.
Jack Mendelson
and Nancy Mello of the Alcohol and Drug
Abuse Research
Research
Center, McLean
Hospital, Harvard
Medical School, for their current support.
Received
October
28, 1993. Revision received
December
14,1993.
Accepted
December
22,1993.
Address requests for reprints to: Dr. Laura Kragie, Biological
Sciences, Cooke Hall, State University
of New York at Buffalo,
Buffalo, New York 14260-l 300.
This work was supported
by NIH Grants DK-01456
(to L.K.)
and DK-41009
(to V.C.). Portions of this paper were presented
at the 74th Annual Meeting
of The Endocrine
Society,
San
Antonio,
TX, June 1992, and the 23rd National
Medicinal
Chemistry
Symposium,
Division
of the American
Chemical
Society, Buffalo, NY, June 1992.
REFERENCES
Fit and VdW Volumes
Because the fused benzene
ring represents
the only conformationally
stable component
in the BZ ring system, conformational comparisons
between
BZ molecules
were made by
a least squares fit of the fused benzene
ring in each BZ (of
the same stereoisomer)
using the root mean square (RMS) fit
option in SYBYL. The BZs comprised
two structural
groups,
the classical 1,4-BZs
(e.g. diazepam)
and the 1,2-annelated
BZs (e.g. triazolam).
The most (A) and least (L) active compounds from each group were superimposed
through the RMS
fit option. The SYBYL option Compare
Volumes
subtracted
their VdW volumes. For the BZ group, L was subtracted
from
A to highlight the features
unique to the active compounds.
The volume in common
with the least active compound
was
subtracted
to form the active volume displays (excluded
volume analysis).
A similar methodology
was used for thyroid
hormone
analogs. These THs were RMS fit by superimposing
the ether linkage and both phenyl rings. However,
because
our TH SAR data suggested
that the 5’substituents
of the
inactive compounds
reduced potency,
we subtracted
A from
L to highlight groups unique to the inactive compound.
These
regions were then avoided in the template for T1 fit.
The TS molecule was fit to the VdW surfaces in common to
the most active BZ molecules
(27). This produced
our first
approximation
of T3/BZ fit. We then went on to RMS fit the
following
atoms (TH to BZ): Cl’ to Cl’;
R4’ to R4’; chiral
carbon of the alanine side chain to Rl methyl; 5-l to C7; and
3-l to R3 (see Fig. 1).
lsopotential
No. 3
Maps
We created the isopotential
maps by using the Isopotential
Map option of the SYBYL software
package. This option uses
the MOPAC
geometry
and the charges
that were created
when the structures
were energy minimized.
It computes
and
displays the isopotentials
at discrete points of the molecule.
We calculated
the negative isopotential
maps of the minimum
energy conformers.
The level of the surface of -5.0 kcal was
used for the comparisons
of Fig. 6, A, C, and D. For the
comparisons
between
the charged conformers
of L-T~ and DT3 (Fig. 6B), a level of -17 kcal was chosen for display (30).
Acknowledgments
The authors thank Dr. Darrell Doyle for his sponsorship
of Dr.
Kragie and the financial support
of Boots Pharmaceuticals
(Lincolnshire,
IL). Dr. David Smith provided
assistance
with
the SYBYL software
programs
along with the Medical Foundation graphics department.
Rich Smiehorowski
provided
in-
1. Wartofsky
L. lngbar SH 1991 Diseases of the thyroid.
In:
Wilson JD, Braunwald
E, lsselbacher
KJ, (eds) Harrison’s
Principles
of Internal Medicine,
ed 12. McGraw-Hill,
New
York, pp 1692-l 712
2. DeNayer
P 1987 Thyroid hormone
action at the cellular
level. Horm Res 26:48-57
3. Fliegel L, Newton E, Burns K, Michalak M 1991 Molecular
cloning of cDNA encoding a 55 KDa multifunctional
thyroid
hormone
binding protein of skeletal muscle sarcoplasmic
reticulum.
J Biol Chem 26515496-l
5502
4. Ashizawa
K, Kato H, McPhie P. Cheng S-Y 1990 Regulation of thyroid hormone
binding to its cytosolic
binding
protein
by L-a-alanine.
Biochem
Biophys
Res Comm
167:587-592
5. Kragie L 1993 Neuropsychiatric
implications
of thyroid
hormone
and benzodiazepine
interactions.
Endocr Res
19:1-32
6. Kastellakis
A, Valcan T 1989 Characterization
of thyroid
hormone
transport
in synaptosomes
from rat brain. Mol
Cell Endocrinol
67:231-241
7. Robbins J, Lakshmanan
M 1992 The movement
of thyroid
hormones
in the central nervous system. Acta Med Austriaca 19:21-25
8. Kragie L, Doyle D 1992 Benzodiazepines
inhibit temperature-dependent
L-[‘Z51]triiodothyronine
accumulation
into
human liver, human neuroblast,
and rat pituitary cell lines.
Endocrinology
130:121 l-l 216
9. Kragie L 1992 Requisite structural
characteristics
for benzodiazepine
inhibition
of triiodothyronine
uptake
into a
human liver cell line. Life Sci 51 :PL83-PL88
10. Verma A, Snyder SH 1989 Peripheral type benzodiazepine
receptors.
Annu Rev Pharmacol
Toxicol 29:307-322
11. Villar HO, Davies MF, Loew GH, Maguire PA 1991 Molecular models for recognition
and activation
at the benzodiazepine
receptor:
a review. Life Sci 48:593-602
12. Haefely W, Kyburz
E, Gerecke
M, Mohler H 1985 Recent
advances
in the molecular
pharmacology
of benzodiazepine receptors
and in the structure-activity
relationships
of their agonists
and antagonists.
In: Testa B (ed) Advances in Drug Research.
Academic
Press, New York,
vol 14: 165-323
13. Osty J, Valensi P, Samsom
M, Francon J, Blondeau
J-P
1990 Transport
of thyroid hormones
by human erythrocytes: kinetic characterization
in adults and newborns.
J
Clin Endocrinol
Metab 71 :1589-l
595
14. Kragie L. Kwon Y, Smiehorowski
R 1993 Rat cardiac
calcium channels and their relationships
with fi-adrenergic
and muscarinic
receptors
in hypothyroidism.
Endocr Res
19:57-71
15. Hawthorn
M, Gengo P, Wei X-Y, Rutledge
A, Moran J,
lodothyronine
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
Transporter
Pharmacophore
Gallant S, Triggle D 1988 Effect of thyroid status on padrenoceptors
and calcium channels
in rat cardiac and
vascular
tissue.
Naunyn-Smiedebergs
Arch Pharmacol
3371539-544
Rampe D, Triggle DJ 1986 Benzodiazepine
and calcium
channel function. Trends Pharmacol
Sci 7:461-464
Kenakin TP 1982 The potentiation
of cardiac responses
to adenosine
by benzodiazepines.
J Pharmacol
Exp Ther
222:752-758
Cody V 1980 Thyroid
hormone
interactions:
molecular
conformation,
protein binding, and hormone
action. Endoer Rev 1 :140-l 66
Oppenheimer
JH 1989 Tissue and cellular effects of thyroid hormones
and their mechanism
of action. In: Burrow
G, Oppenheimer
J, Volpe R (eds) Thyroid
Function
and
Disease. WB Saunders,
Philadelphia,
pp 90-l 23
Jorgensen
E 1976 Structure
activity relationships
of thyroxine analogs.
Pharmacol
Ther Part B Gen Syst Pharmacol2:661-682
Cheung EN-y 1985 Thyroid hormone
action: determination of hormone
receptor
interaction
using structural
analogs and molecular
modeling.
Trends
Pharmacol
Sci
6:31-34
Cody V 1979 Thyroid
hormones-receptor
interactions:
binding models from molecular
conformation
and binding
affinity data. In: Olson E, Christoffersen
R, (eds) Computer
Assisted Drug Design. American Chemical Society, Washington, DC, pp 281-299
Drugan R, Holmes P 1991 Central and peripheral
benzodiazepine
receptors:
involvement
in an organism’s
response to physical and psychological
stress. Neurosci
Biobehav
Rev 15:277-298
Gavish M, Katz Y, Bar-Ami S, Weizman
R 1992 Biochemical, physical and pathological
aspects
of the peripheral
benzodiazepine
receptor.
J Neurochem
58:1589-l
601
Doble A, Martin IA 1992 Multiple benzodiazepine
receptors: no reason for anxiety. Trends Pharmacol
Sci 13:7681
Ruano D, Vizuete M, Cano J, Machado
A, Vitorica J 1992
Heterogeneity
in the allosteric interaction
between
the yaminobutyric
acid (GABA) binding site and three different
benzodiazepine
binding sites of the GABAa/benzodiazepine receptor
complex in the rat nervous
system. J Neurochem 581485-493
Allen M, Tan Y-C, Trudell M, Narayanan
K, Schindler
L,
Martin M, Schultz C, Hagen T, Koehler
K, Codding
P,
Skolnick P, Cook J 1990 Synthetic and computer-assisted
analyses
of the pharmacophore
for the benzodiazepine
receptor inverse agonist site. J Med Chem 33:2343-2357
Wang J, Taniguchi
T, Spector S 1984 Structural
requirements for the binding of benzodiazepines
to their peripheral-type
sites. Mol Pharmacol
25:349-351
Tallarida
R, Murray
R 1986 Manual
of Pharmacologic
Calculations,
ed 2. Springer Publications,
New York
Massa S, Artico M, Mai A, Corelli F, Botta M, Tafi A,
Pantaleoni
P, Giorgi R, Coppolino
M, Gagnotto
A, Skorupska M 1992 Pyrrolobenzodiazepines
and related systems. 2. Synthesis
of properties
of isonoraptazepine
derivatives
J Med Chem 35:4533-4541
Brachtel G, Jansen M 1981 7-Chloro-2,3-dihydro-1
-cycle-
391
propylmethyl-5-phenyl-1,4-benzodiazepam:
prazepam.
Cryst Struct Comm 10:669-672
32 Bandoli G, Clemente
D 1976 Crystal molecular and electronic structure
of an antianxiety
agent: 7-chloro-5-(2chlorophenyl)-l,3-dihydro-3-hydroxy-l,4benzodiazepine-2-one.
J Chem Sci Perkin Trans 2:413-418
33 Galdecki Z, Giowka M 1980 7-Chloro-3-hydroxy-1
-methyl5-phenyl-l,3-dihydro-2H-l,4-benzodiazepin-2-one:
temazepam: Acta Crystallogr
Sect B Struct Sci 36:3044-3048
34 Gilli G. Bertolasi V. Sacerdoti
M. Borea P 1978 7-Chloro1,3-dihydro-3-hydroxy-5-phenyl:2H-l,4-benzodiazepin-2one: oxazepam.
Acta Crystallogr
Sect B Struct
Sci
34:2826-2829
35 Camerman
L, Camerman
N 1972 Diazepam:
7-chloro-1,3dihydro-l-methyl-5-phenyl-2H-1,4-benzodiazepin-2-one.
J Am Chem Sot 94:268-272
36. Chananont
P, Hamor T, Martin I 1979 5-(P-Chlorophenyl)1,3-dihydro-7-nitro-2H-l,4-benzodiazepin-2-one.
Clonazepam. Cryst Struct Comm 8:393-396
37. Gilli G, Bertolasi V, Sacerdoti
M 1978 7-Chloro-2,3-dihydro-1 -methyl-5-phenyl-1
H-l ,4benzodiazepine
(MedazePam). Acta Crystallogr
Sect B Struct Sci 34:3793-3795
38. Kamiya K, Wada Y, Nishikawa
M 1973 8-Chloro-6-phenyl4H-S-triazolo(4,
3a)(1,4)
benzodiazepine.
Chem Pharm
Bull (Tokyo) 21 :1520-l
522
39. Hernstadt
C, Mootz D, Wunderlich
H, Mohrle H 1979 7Chloro-2-methylamino-5-phenyl-3H-l,4benzodiazepine4-oxide
hydrochloride.
J Chem Sot Perkin Trans 2:735738
40. Butcher
H, Hamor T, Martin I 1983 5-(2-Fluorophenyl)1,3-dihydro-1
-methyl-7-nitro-2H-1,4-benzodiazepin-2one: flunitrazepam.
Acta Crystallogr
Sect C Struct Commun 39:1469-l
472
41. Gilli G, Bertolasi
V, Sacerdoti
M, Borea P 1987 7-Nitro1,3-dihydro-5-phenyl-2H-1,4-benzodiazepin-2-one:
nitrazepam. Acta Crystallc-gr
Sect B Struct Sci 33:2664-2667
42. Butcher H, Hamor T, Martin I 1983 7-Bromo-1
.Sdihydro5-(2-pyridyl)-2H-1,4-benzodiazepon-2-one:
bromazepam.
Acta Crystallogr
Sect C Cryst Struct Commun
39:14691472
43. Cody V 1974 Crystal structure
and molecular
conformation of the thyroid hormone
distal 3,5,3’ triiodo-l-thvronine. J Am Chem Sot 96:6720-6725
44. Codv V. Duax WL. Norton
DA 1972 Molecular
conformatibn of the 3,5 diiodo-L-thyronine-n-methylacetamide
complex
(1:l).
Acta Crystallogr
Sect B Struct
Sci
28:2244-2252
45. Okabe N, Fujiwara T, Yamagata
Y, Tomita K-l 1982 The
crystal structure
of a major metabolite
of thyroid hormone
3,3’,5’-triicdo-L-thyronine.
Biochim
Biophys
Acta
717:179-l
81
46. Cody V, Hazel J, Langs DA, Duax WL 1977 Molecular
structure
of thyroxine
analogues:
crystal
structure
of
3,5,3’triiodothyroacetic
and 3,5,3’,5’tetraiodothyroacetic
acid-Ndiethanolamine
(1 :l) complexes.
J Med Chem
20: 1628
47. Cody V, Erman M, De Jarnette
E 1977 Crystal structure
of the active thyroxine
analogue, 3,5 diiodothyropropionic
acid (3-(4-hydroxyphenoxy-3,5-di-iodophenyl)
propanoic
acid. J Chem Res (M) 126:1444-l
465