Chiral-pool synthesis of 1,2,4-trisubstituted 1,4-diazepanes as novel σ1 receptor ligands

Article abstract of DOI:10.1016/j.bmc.2017.07.027

Starting from enantiomerically pure amino acids, 1,4-diazepanes with various substituents in 1, 2, and 4-position were synthesized following the late stage diversification strategy. The key step in the formation of the seven-membered ring was the intramolecular EDC coupling of amino acids 15, 26, and 39. The configuration in 2-position does not influence the σ₁ affinity and selectivity over related receptors. A cyclohexylmethyl or a butyl group are the preferred substituents in 4-position, whereas a methyl moiety in 2-position and a (substituted) benzyl moiety in 1-position result in the highest σ₁ affinity. These results fit nicely to the reported σ₁ pharmacophore models. The compounds did not inhibit the structurally related fungal enzyme sterol Δ^8,7-isomerase, but showed inhibition of diverse enzymes in late cholesterol biosynthesis at high concentrations. In a screening against more than 50 target proteins, (2S)-1-benzyl-4-(4-methoxybenzyl)-2-methyl-1,4-diazepane ((S)-28b, K_i(σ₁) = 0.86 nM) showed a clean receptor profile. The dose dependent potentiation of electrically stimulated contractions of guinea pig vas deferens indicates σ₁ agonistic activity of (S)-28b. Even at a dose of 100 mg/kg (S)-28b did not induce severe toxic or behavioral effects in the Irwin screen. Clear cognition enhancing effects were observed for (S)-28b after inducing amnesia by scopolamine.

Full text of DOI:10.1016/j.bmc.2017.07.027

Accepted Manuscript
Chiral-pool synthesis of 1,2,4-trisubstituted 1,4-diazepanes as novel σ receptor
1
ligands
Lena Fanter, Christoph Müller, Dirk Schepmann, Franz Bracher, Bernhard
Wünsch
PII:
S0968-0896(17)31206-3
DOI:
http://dx.doi.org/10.1016/j.bmc.2017.07.027
BMC 13864
Reference:
To appear in:
Bioorganic & Medicinal Chemistry
Received Date:
Revised Date:
Accepted Date:
8 June 2017
6 July 2017
13 July 2017
Please cite this article as: Fanter, L., Müller, C., Schepmann, D., Bracher, F., Wünsch, B., Chiral-pool synthesis of
1
,2,4-trisubstituted 1,4-diazepanes as novel σ receptor ligands, Bioorganic & Medicinal Chemistry (2017), doi:
1
http://dx.doi.org/10.1016/j.bmc.2017.07.027
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1
Chiral-pool synthesis of 1,2,4-trisubstituted 1,4-diazepanes as
novel σ
1
receptor ligands
a
b
a
b
Lena Fanter, Christoph Müller, Dirk Schepmann, Franz Bracher, Bernhard
a,c
Wünsch*
a
Institut für Pharmazeutische und Medizinische Chemie der Universität Münster,
Corrensstraße 48, D-48149 Münster, Germany
Tel.: +49-251-8333311; Fax: +49-251-8332144; E-mail: wuensch@uni-muenster.de
b
Department für Pharmazie, Ludwig-Maximilians-Universität München,
Butenandtstr. 5-13, D-81377 München, Germany
c
Cells-in-Motion Cluster of Excellence (EXC 1003 – CiM), Westfälische Wilhelms-
Universität Münster, Germany
Abstract
Starting from enantiomerically pure amino acids, 1,4-diazepanes with various
substituents in 1, 2, and 4-position were synthesized following the late stage
diversification strategy. The key step in the formation of the seven-membered ring
was the intramolecular EDC coupling of amino acids 15, 26, and 39. The
configuration in 2-position does not influence the σ affinity and selectivity over
1
related receptors. A cyclohexylmethyl or a butyl group are the preferred substituents
in 4-position, whereas a methyl moiety in 2-position and a (substituted) benzyl
moiety in 1-position result in the highest σ
reported σ pharmacophore models. The compounds did not inhibit the structurally
related fungal enzyme sterol ∆ -isomerase, but showed inhibition of diverse
1
affinity. These results fit nicely to the
1
8/7

2
enzymes in late cholesterol biosynthesis at high concentrations. In a screening
against more than 50 target proteins, (2S)-1-benzyl-4-(4-methoxybenzyl)-2-methyl-
1
,4-diazepane ((S)-28b, K
dose dependent potentiation of electrically stimulated contractions of guinea pig vas
deferens indicates σ agonistic activity of (S)-28b. Even at a dose of 100 mg/kg (S)-
8b did not induce severe toxic or behavioral effects in the Irwin screen. Clear
i
(σ
1
) = 0.86 nM) showed a clean receptor profile. The
1
2
cognition enhancing effects were observed for (S)-28b after inducing amnesia by
scopolamine.
Key words: 1,2,4-trisubstituted 1,4-diazepanes; chiral pool synthesis; late stage
diversification; σ ligands; σ
1
agonists; selectivity; cholesterol biosynthesis; sterol
8/7
∆
-isomerase; Irwin test; cognition enhancement;
1. Introduction
Based on the molecular weight, ligand binding profile and distribution in the
organism, the class of σ receptors was subdivided into two subtypes, which were
1-4
termed σ
1
and σ
2
receptor. In contrast to the σ
2
subtype, the σ receptor has been
1
cloned from different tissues and species, such as mouse and rat brain, guinea pig
liver, human brain and placenta. The human gene encodes for a unique protein with
2
23 amino acids leading to a molecular weight of 25.3 kDa. A similar mammalian
protein could not be detected, but unexpectedly, the σ receptor shares a 30 %
identity and 67 % similarity with the yeast enzyme sterol-∆ -isomerase. Despite the
1
8/7
5-9
strong similarity, the σ receptor does not have any sterol isomerase activity.
1

3
Not only the sequence, but also the structure of the σ receptor is unique: Two
1
transmembrane helices are connected by a 50 amino acid loop. The N-terminus
with only 10 amino acids is rather short, whereas the C terminus is rather big
containing the ligand binding domain. Together with small parts of the second
transmembrane domain the C-terminal domain forms two steroid binding domain
1
0-12
like structures (SBDL1 and SBDL2).
The similarity of the σ
yeast sterol ∆ -isomerase, the presence of two SDBL domains and the moderate
affinity of some neuroactive steroids such as progesterone and
norepiandrosterone indicate that σ receptors play a role in the field of steroids.
Very recently, the X-ray crystal structure of the σ receptor has been solved. In
contrast to all reports so far, only one transmembrane domain was found with the N-
1
receptor and the
8/7
σ
1
1
1
13
and C-termini of the σ receptor being located on opposite sides of the membrane.
1
The σ
reticulum (ER) associated with mitochondria (MAM) and the plasma membrane.
In the ER as well as the cell membrane, the ligand binding domain containing C-
terminus are oriented towards the cytoplasm. This may indicate a reversed polarity
of the receptor in these membranes. It has been reported that the σ receptor aids
the IP receptor in the ER membrane in a chaperone-like manner regulating proper
Ca -signaling. Moreover, it is able to modulate the opening state of some cation
1
receptor is predominantly found in the membrane of the endoplasmic
14-16
1
3
2+
+
+
2+
channels (K , Na , Ca channels) and the activity of some neurotransmitter
systems (e.g. the glutamatergic neurotransmission). However, the exact signal
transduction pathway following σ receptor activation has not been identified
1
1
7,18
yet.

4
19-23
The σ
1
receptor is involved in several neurological and psychiatric disorders.
Inhibition of the σ
1
receptor was associated with antipsychotic, analgesic and anti-
addictive activity. Several clinically used antipsychotics such as haloperidol and
24
quetiapin show high σ
II of clinical trials for the treatment of neuropathic pain.
receptor agonists display antidepressant effects and ameliorate memory and
1
affinity. The pyrazole derivative S1RA is currently in phase
25,26
On the contrast, σ
1
27
cognition deficits in some neurodegenerative disorders (e.g. Parkinson’s and
28,29
Alzheimer’s Disease).
High σ
1
affinity was found for the antidepressants
30
31
fluvoxamine and opipramol as well as the acetylcholinesterase inhibitor
32
donepezil, which is used for the treatment of Alzheimer’s disease. Furthermore,
some human tumor cell lines show high expression levels of σ receptors.
Incubation of these cell lines with σ receptor antagonists resulted in inhibition of
cell growth up to cell death indicating the high potential of σ receptor ligands for
1
1
1
33-35
tumor therapy.
The class of σ
Among them, 1,4-disubstituted piperazines represent a prominent subgroup of
potent σ receptor ligands. For example, 1-butyl-4-(3,4-dichlorophenethyl)piperazine
1, Figure 1) binds with high affinity at σ
on the aryl substituent 1,4-dibenzylpiperazines 2 also exhibit high σ
high selectivity over the σ subtype (e.g. 2a: Aryl = 4-BrC : K(σ ) = 0.78 nM; 2b:
Aryl = 4-IC : K(σ ) = 0.80 nM). The cyclohexyl moiety directly attached at a
piperazine N-atom is able to increase both σ and σ affinity. The tetralinylbutyl
derivative 3 interacts in picomolar range with the σ receptors (K(σ ) = 0.036 nM,
(σ ) = 14.6 nM). Introduction of a methoxy group in 5-position of the tetraline ring
1
receptor ligands contains structurally very diverse compounds.
1
36
(
1
receptors (K
i
(σ
1
) = 0.55 nM). Depending
1
affinity and
2
6
H
4
i
1
37
6
H
4
i
1
1
2
1
i
1
K
i
2

5
and shortening the tetramethylene spacer by one CH
2
moiety resulted in PB28 with
38
reversed σ receptor selectivity (K
i
(σ
1
) = 13.6 nM, K
i
2
(σ ) = 0.34 nM).
Figure 1: Published piperazine-based ligands 1-4 show high σ
Expansion of the piperazine ring of 4 by a CH -moiety led to 1,4-diazepanes 5 with
increased σ affinity and increased selectivity over the σ subtype. Novel 1,4-
1
receptor affinity.
2
1
2
diazepanes 6 should be available starting from enantiomerically pure α-amino
acids.
Recently, the high σ
derivatives 4 was reported. An additional aromatic moiety in the N-4 substituent
seems to be favorable for high σ receptor affinity, as shown for the dibenzyl
derivative 4a (R = CH Ph, K(σ ) = 38 nM). Ring expansion of the 2-
hydroxymethyl) substituted piperazines 4 resulted in 1,4-diazepanes 5
homopiperazines) with even increased σ receptor affinity. The σ affinity of the
1
receptor affinity of 2-(hydroxymethyl) substituted piperazine
1
39
2
i
1
(
(
1
1

6
dibenzyl derivative 5a (R = CH
2
Ph, K
i
(σ
1
) = 7.4 nM) is about 5-fold higher than the
σ
1
affinity of the analogous piperazine derivative 4a. In addition to the increased σ
1
affinity, the σ /σ selectivity was increased remarkably by the ring expansion from 5-
1
2
40
fold for piperazine 4a to 53-fold for 1,4-diazepane 5a. It is speculated that the
different orientation of the three substituents at the seven-membered ring of 5,
which is conformationally more flexible than the six-membered piperazine chair of 4,
is responsible for the increased σ affinity and selectivity of 1,4-diazepanes 5.
1
In this project we aim at the expansion of the relationships between the structure
and the σ receptor affinity and σ /σ selectivity of 1,4-diazepanes
homopiperazines). In particular the substituents in 1-, 2- and 4-position of the
1
1
2
(
seven-membered 1,4-diazepanes 6 will be varied systematically. Moreover, the
influence of the configuration at 2-position of 6 will be investigated. In order to
prepare 1,2,4-trisubstituted 1,4-diazepanes 6 a chiral pool synthesis starting from
enantiomerically pure α-amino acids is planned (Figure 1). In addition to σ
receptor affinity, the effect on the ergosterol and cholesterol biosynthesis of the
,2,4-trisubstituted 1,4-diazepanes will be investigated. In order to learn more about
the activity of σ ligands the most promising candidate(s) will be carefully
investigated in vitro and in vivo in various animal models.
1
and
σ
2
1
1
2. Synthesis
In the first series of ligands (S)-phenylalanine ((S)-7) was used as chiral starting
compound leading to 1,4-diazepanes 17 and 18 with a benzyl moiety in 2-position
(
Schemes 1 and 2). After transformation of (S)-phenylalanine ((S)-7) into its methyl
41,42
43,44
ester (S)-8,
the benzylamine (S)-9
was prepared by condensation of (S)-8

7
with benzaldehyde and subsequent NaBH reduction. For the construction of the
4
seven-membered ring, the secondary amine (S)-9 was acylated with 3-
chloropropionyl chloride. Performing this transformation with pyridine at -20 °C for 3
h provided the propionamide (S)-10 in 82 % yield. Triethylamine as base, higher
temperatures than -20 °C and longer reaction times led to increasing amounts of the
elimination product (S)-12.
Scheme 1: Synthesis of 1,4-diazepane-2,5-dione (S)-16a from (S)-phenylalanine
(
(S)-7) via the azide route. Reagents and reaction conditions: (a) CH
RT, 4 h, >99 %. (b) 1. PhCH=O, toluene, NEt , molecular sieves 5 A, RT, 1 h; 2.
NaBH , CH OH, 0 °C, 10 min, 79 %. (c) ClCH CH COCl, pyridine, CH Cl , -20 °C, 3
h, 82 %. (d) NaI, acetone, microwave irradiation, 97 %. (e) NaN
3
OH, SOCl
2
,
3
4
3
2
2
2
2
3
, DMF, RT, 24 h, 73
.
3
%
. (f) H
2 3
, 3 bar, Pd/C, CH OH, RT, 3 h, then BF
2
OEt , RT, 15 h, 63 %. The same
reactions were performed with (R)-phenylalanine ((R)-7) leading to the enantiomeric
azidopropionyl derivative (R)-13.

8
All attempts to react the chloropropionamide (S)-10 with primary amines to form
1,4-diazepanediones failed due to slow nucleophilic substitution and fast elimination
of HCl giving the acrylamide (S)-12. Therefore, the leaving group was improved by
replacement of the Cl substituent of (S)-10 by an I-atom (Scheme 1). The
iodopropionamide (S)-11 was accomplished via a Finkelstein reaction of (S)-10 with
NaI in acetone under microwave irradiation. In order to avoid basic conditions and
elimination, (S)-11 was reacted with highly nucleophilic but less basic NaN
to afford the azidopropionamide (S)-13 in 73 % yield. Formation of acrylamide (S)-
2 was not observed under these reaction conditions. Reduction of the azide with
in the presence of Pd/C yielded the primary amine (S)-14a, which reacted
3
in DMF
1
H
2
spontaneously to form 1,4-diazepanedione (S)-16a. Treatment with the Lewis acids
.
3
Ti(OiPr)
4
or BF
OEt led to complete cyclization affording the 1,4-diazepanedione
2
(
S)-16a in 63 % yield. Unfortunately, the 1,4-diazepane (S)-17a could not be
isolated after LiAlH reduction of (S)-16a due to purification problems.
4
In literature the synthesis of (S)-16a has already been reported by cyclization of the
45
carboxylic acid analog of methyl ester (S)-14a. Intramolecular Staudinger reaction
46,47
or treatment with coupling agents
yielded the 1,4-diazepane-2,5-dione (S)-16a.
In contrast to the chloropropionamide (S)-10, the iodopropionamide (S)-11 reacted
with primary amines to afford the secondary amines (S)-14b-d in 89-93 % yields
(
(
Scheme 2). Due to the mild reaction conditions, elimination towards acrylamide
S)-12 could not be observed. However, all attempts to induce the intramolecular
aminolysis of the aminoesters (S)-14b-d by heating to reflux or addition of Lewis
.
3
acids (BF
2
OEt , Ti(OiPr)
4
, AlCl ) failed to give the 1,4-diazepanediones (S)-16b-d.
3

9
Scheme 2: Synthesis of diversely substituted 1,4-diazepanes derived from (S)-
1
phenylalanine. Reagents and reaction conditions: (a) R NH
2
, DMF, RT, 24 h, 89-93
.
%
. (b) LiOH, THF/H
h, 56-79 %. (d) LiAlH
chloroethyl) chloroformate, THF, -78 °C, 10 min, then CH
f) R’CH=O, CH Cl or ClCH CH Cl, NaBH(OAc)
CH CN, K CO
reflux, 3 h, 56 % ((S)-17g), 21 % ((S)-17h). (i) (S)-17j (R = n-Bu), NH
CH OH, reflux, 15 min, 87 %. (j) HCH=O, CH Cl , NaBH(OAc) , RT, 3 h, 88 %
18c). (k) R-X, CH CN, K CO , reflux, 24 h, 22-41 % (18b/18d). The corresponding
2
O, RT, 0.5-1 h. (c) EDC HCl, HOAt or HOBt, CH
2
Cl
1
2
, DMF, RT,
1
4
. THF, reflux, 16-18 h, 61-89 %. (e) (S)-17b (R = PMB), (1-
3
OH, reflux, 40 min, 99 %.
1
(
2
2
2
2
3
, RT, 1-3 h, 72-84 %. (g) R -X,
3
2
3
, microwave irradiation, 29-73 %. (h) rac-2-phenyloxirane, CH
3
OH,
1
4
HCO
2
, Pd/C,
3
2
2
3
(
(
3
2
3
R)-configured enantiomers were prepared analogously starting with the
#
iodopropionyl derivative (R)-11. Only the (R)-configured enantiomer (R)-17k was
prepared.

1
0
Therefore, a two-step one-pot cyclization was envisaged comprising hydrolysis of
the esters (S)-14b-d with LiOH and subsequent cyclization of the resulting amino
acids (S)-15b-d with coupling agents. After several optimization reactions, the
.
coupling agent EDC HCl, in the presence of HOBt or HOAt, turned out to furnish
both the highest yields (56-79 %) and enantiomeric purity (see part enantiomeric
purity) of the 1,4-diazepanediones (S)-16b-d. LiAlH reduction of (S)-16b-d
4
provided the 1,2,4-trisubstituted 1,4-diazepanes (S)-17b-d (Scheme 2).
In order to further modify the substituents in 4-position via a late-stage building
48,49
block,
the p-methoxybenzyl (PMB) group in 4-position of (S)-17b was cleaved
off selectively with (1-chloroethyl) chloroformate at -78 °C and subsequent heating
50
in methanol. The secondary amine (S)-17a was isolated in 99 % yield and served
48,49
as central intermediate for late stage diversification,
i.e. introduction of various
substituents in 4-position. The introduction of various substituents was performed by
51
reductive alkylation with aldehydes and NaBH(OAc)
alkylation with alkyl halides in the presence of K CO
-phenyloxirane. Since the secondary amine (S)-17a was able to attack
-phenyloxirane at either the 2- and 3-position, two regioisomeric 2-hydroxyethyl
3
, by microwave assisted
2
3
and by reaction with racemic
2
2
derivatives with the phenyl moiety either in 1- ((S)-17g) or 2-position ((S)-17h) were
formed, which were separated by flash chromatography. Both regioisomers were
isolated as 1:1-mixtures of diastereomers with (S)-configuration in 3-position of the
1,4-diazepane ring and both (R)- and (S)-configuration in the hydroxyethyl side
chain originating form racemic 2-phenyloxirane (Scheme 2).

1
1
In order to learn more about relationships between the structure and σ receptor
affinity homologous phenylalkyl groups were introduced at 4-position (17b-f).
Additionally, alkyl moieties (17i-k) were attached at this position. The dimethylallyl
moiety of 17k was selected, since it is part of the prototypical σ
1
agonist
(
+)-pentazocine.
Modifications at 1-position of the 1,4-diazepane ring were performed exemplarily for
the 4-butyl derivative (S)-17j. Hydrogenolytic removal of the benzyl moiety in
1
-position with ammonium formate and Pd/C resulted in the secondary amine
S)-18a, which was N-methylated with formaldehyde and NaBH(OAc) or alkylated
with alkyl halides to provide the 1,2,4-trisubstituted 1,4-diazepanes (S)-18b-d
Scheme 2).
(
3
(
The corresponding enantiomers (R)-17a-j and (R)-18a-d with (R)-configuration in
-position of the 1,4-diazepane ring were prepared in the same manner starting with
2
(
(
R)-phenylalanine ((R)-7). In contrast to the synthesis of (S)-17c and (S)-17d, the
R)-configured enantiomers (R)-17c and (R)-17d were prepared by reductive
alkylation of secondary amine (R)-17a with benzaldehyde and phenylacetaldehyde,
respectively, and NaBH(OAc)
3
.
Next, the influence of the substituent in 2-position of the 1,4-diazepane ring should
be investigated. For this purpose the benzyl moiety of the phenylalanine derivatives
17 and 18 was replaced by a methyl group leading to the amino acid alanine as a
chiral starting material.

1
2
Scheme 3: Synthesis of diversely substituted 1,4-diazepanes derived from
(
(
S)-alanine. Reagents and reaction conditions: (a) CH
b) PhCH=O, NaBH(OAc) , ClCH CH Cl, NEt , RT, 5 h, 80 %. (c) ClCH
Cl , -20 °C, 1 h, then RT, 2 h, 61 %. (d) NaI, acetone, microwave
3
OH, SOCl
2
, RT, 2 h, 99 %.
3
2
2
3
2
CH COCl,
2
pyridine, CH
2
2
irradiation, 94 %. (e) PMB-NH
2
, DMF, RT, 24 h, 87 %. (f) LiOH, THF/H
, THF, reflux, 16 h, 78 %. (i)
OH, reflux, 1 h, 94 %.
, reflux, 24 h, 49 %. (k) n-BuBr, CH CN,
, microwave irradiation, 64 %. (l) C 11CH=O, NaBH(OAc) , CH Cl , RT, 2 h,
2
O, RT, 1 h.
.
(
(
(
g) EDC HCl, HOBt, CH
2
Cl
2
, RT, 1 h, 47 %. (h) LiAlH
4
1-chloroethyl) chloroformate, THF, -78 °C, 10 min, then CH
j) Ph(CH Cl, Bu NI, CH CN, K CO
CO
3
2
)
4
4
3
2
3
3
K
2
3
6
H
3
2
2

1
3
1
6
6 %. (m) (S)-28e (R = CH
2
C
5
H
11), NH
4
HCO
2
, Pd/C, CH OH, reflux, 10 min, 93 %.
3
(
(
(
n) RCH=O, NaBH(OAc) , ClCH
3
2
CH
2
Cl, RT, 1 h, 65-67 %. The corresponding
R)-configured enantiomers (R)-28 were prepared analogously starting with
R)-alanine ((R)-19).
42,52,53
The methyl ester (S)-20,
which was obtained by esterification of
enantiomerically pure (S)-alanine ((S)-19), was N-benzylated with benzaldehyde
54,55
and NaBH(OAc)
acylated with 3-chloropropionyl chloride to afford the chloropropionamide (S)-22 in
1 % yield (Scheme 3). In order to avoid HCl elimination and formation of
acrylamide (S)-24, a Finkelstein reaction was performed to afford iodo compound
S)-23, which allowed the nucleophilic substitution with p-methoxybenzylamine to
3
. The resulting secondary amine (S)-21
was subsequently
6
(
give the secondary amine (S)-25b in 87 % yield. Since the intramolecular
aminolysis of the aminoester (S)-25b failed to produce the diazepanedione (S)-27b
directly, the ester (S)-25b was hydrolyzed with LiOH and the resulting amino acid
(
S)-26b was cyclized using the coupling agent EDC. LiAlH reduction of the
4
diazepane-2,5-dione (S)-27b provided the 1,4-diazepane (S)-28b in 78 % yield.
After selective removal of the PMB-protective group in 4-position upon treatment of
(
S)-28b with (1-chloroethyl) chloroformate in THF at -78 °C, various substituents
were introduced at 4-position of the 1,4-diazepane ring. The cyclohexylmethyl
substituted 1,4-diazepane (S)-28e was used exemplarily to modify the substituent in
1-position as well. For this purpose, the 1-benzyl moiety of (S)-28e was cleaved off
hydrogenolytically to obtain the secondary amine (S)-29a, which was further
transformed by reductive alkylation into tertiary amines (S)-29b and (S)-29c
(
Scheme 3). The corresponding (R)-configured 1,4-diazepanes (R)-28a-e were
synthesized in the same manner starting with (R)-alanine ((R)-19).

1
4
Scheme 4: Synthesis of diversely substituted 1,4-diazepanes derived from (S)-
valine. Reagents and reaction conditions: (a) 7 steps analogously to the described
steps for the synthesis of (S)-16b and (S)-27b. Details of the transformations are
given in the Supporting Information, Scheme SI1 and corresponding legend. (b)
LiAlH
min, then CH
RT, 1 h, 71 %.
4
, THF, reflux, 15 h, 61 %. (c) (1-chloroethyl) chloroformate, THF, -78 °C, 10
3
OH, reflux, 40 min, 96 %. (d) C 11CH=O, NaBH(OAc) , ClCH CH Cl,
6
H
3
2
2
In addition to the aromatic benzyl moiety (compounds 17 and 18) and the small
aliphatic methyl group (compounds 28 and 29) in 2-position of the 1,4-diazepane
ring, a branched alkyl substituent (compounds 32) was considered for this position.
41
The proteinogenic amino acid (S)-valine ((S)-30) served as starting material for
the synthesis of these 1,4-diazepanes (S)-32 (Scheme 4). The isopropyl substituted
1
,4-diazepanedione (S)-31b was prepared in the same manner as the benzylated
and methylated analogs (S)-16b-d and (S)-27b. After reduction of the
,4-diazepanedione (S)-31b with LiAlH , the PMB-protective group of (S)-32b was
1
4
removed by (1-chloroethyl) chloroformate to yield the secondary amine (S)-32a,
which was reductively alkylated with cyclohexanecarbaldehyde and NaBH(OAc)
provide the 1,4-diazepane (S)-32c.
3
to
Since the eudismic ratio of the 2-methylated 1,4-diazepanes 28 is close to 1
indicating similar σ
1
affinity of both enantiomers, we focused in the valine series on
the (S)-configured enantiomers (S)-32.

1
5
Enantiomeric purity
Cyclization of the phenylalanine derived amino acids (S)-15b-d represents the key
step in the synthesis of 2-benzyl substituted 1,4-diazepanes (S)-17 and (S)-18.
Therefore, the enantiomeric purity of the key intermediate 1,4-diazepanedione
(
S)-16c was investigated. HPLC analysis of (S)-16c using the chiral column
®
Chiralpak AD (Figure SI1 in Supporting Information) revealed a ratio of
enantiomers (S)-16c and (R)-16c of 93:7. Since it was assumed that the cyclization
of the amino acid (S)-15c is the origin of the small amount of racemization, this
transformation was analyzed carefully. Cyclization of the amino acid (S)-15c with
EDC but without 1-hydroxybenzotriazole (HOBt) led to the enantiomers (S)-16c,
and (R)-16c in the ratio 78:22. Addition of HOBt inhibited racemization and thus the
ratio of enantiomers was increased to 91:9, whereas the corresponding aza
derivative 7-aza-1-hydroxybenzotriazole (HOAt) raised the ratio of enantiomers to
93:7. Cyclization of the amino acid (S)-15c with the phosphonium salt BOP resulted
in a reduced ratio of 79:21. These experiments confirm that the cyclization of the
amino acid (S)-15c represents the reaction step, which is crucial for high
enantiomeric purity. Since the reagent HOAt is only available as a solution in DMF,
56
which can increase racemization during amide formation, HOBt was used as
standard additive in the lactamization reactions.
The same analysis performed with (S)-16b, and (S)-16d revealed similar ratios of
enantiomers of 92:8 (Figure SI2 in Supporting Information) and 91:9 (Figure SI3 in
Supporting Information), respectively. Exemplarily, the enantiomeric purity of the
final product (S)-17i was determined to be 92:8 (Figure SI4 in Supporting
Information). As this ratio is identical with the ratio of enantiomers of the central

1
6
intermediate (S)-16b ((S)-16b:(R)-16b = 92:8, Figure SI2 in Supporting
Information), it can be concluded that during the transformation of (S)-16b no further
racemization had occurred.
In addition to the enantiomeric purity of phenylalanine derived 1,4-diazepanes 17
and 18, the enantiomeric purity of the alanine and valine derived 1,4-diazepanes 28,
2
9 and 32 was analyzed. Chiral HPLC of the key intermediates 1,4-
diazepanediones (S)-27b and (S)-31b resulted in ratios of enantiomers of 98:2 and
3 : 17, respectively (Figure SI5 in Supporting Information). This result correlates
nicely with the reactivity of the corresponding amino acids (S)-26b, (S)-39 and (S)-
5b. Whereas the alanine derived amino acid (S)-26b with a small methyl group
8
1
adjacent to the carboxy group reacted faster than the corresponding benzyl
derivative (S)-15b, the isopropyl derivative (S)-39 showed considerably lower
reactivity. A decreased reactivity leads to a prolonged existence of activated
intermediates (activated carboxy group), which increased the amount of
racemization.
3
. Biological activity
receptor affinity
The σ receptor affinity of the amino acid derived trisubstituted 1,4-diazepanes 17,
8, 28, 29, and 32 was determined in radioligand receptor binding assays. Herein,
σ
1
1
1
3
the test compounds compete with the radioligand [ H]-(+)-pentazocine for a limited
number of receptors of a guinea pig brain membrane preparation. The remaining
5
7-59
amount of bound radioligand is measured with a scintillation analyzer.
In Table
1
the σ and σ receptor affinities of all test compounds are summarized.
1
2

1
7
Table 1. Affinities of 1,2,4-trisubstituted 1, 4-diazepane towards σ
1
and σ receptors.
2
#
i
K ± SEM (nM)
σ
1
/ σ
2
1
2
compd.
R
H
H
R
σ
1
σ
2
selectivity
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
S)-17a
R)-17a
S)-17b
R)-17b
S)-17c
R)-17c
S)-17d
R)-17d
S)-17e
R)-17e
S)-17f
R)-17f
S)-17g
R)-17g
S)-17h
R)-17h
S)-17i
R)-17i
S)-17j
R)-17j
R)-17k
S)-18b
R)-18b
CH Ph
>1,000
>1,000
>10,000
>1,000
>1,000
426 ± 72
>1,000
514
-
2
CH Ph
2
-
CH Ph(4-OCH )
2
3
CH Ph
2
29 ± 5.8
42 ± 2.9
71 ± 14
80 ± 21
40 ± 18
52 ± 16
73 ± 14
25 ± 6.9
38 ± 10
28 ± 8.1
234
>34
10
CH Ph(4-OCH )
2
2
2
3
CH Ph
2
CH Ph
CH Ph
2
>14
6.4
11
CH Ph
CH Ph
2
(
(
(
(
(
(
CH ) Ph
2 2
CH Ph
2
426 ± 84
>1,000
193 ± 24
300
CH ) Ph
2
2
CH Ph
2
>19
2.6
12
CH ) Ph
2
3
CH Ph
2
CH ) Ph
2
3
CH Ph
2
CH ) Ph
2
4
CH Ph
2
372 ± 85
423 ± 38
903
9.8
15
CH ) Ph
2
4
CH Ph
2
CH CH(Ph)OH
2
CH Ph
2
3.9
>19
>4
CH CH(Ph)OH
2
CH Ph
2
51 ± 3.5
228
>1,000
>1,000
>1,000
>1,000
>1,000
238 ± 124
169 ± 46
325
CH(Ph)CH OH
2
CH Ph
2
CH(Ph)CH OH
2
CH Ph
2
27 ± 6.4
448 ± 39
280
>35
>2
CH3
CH Ph
2
CH3
CH Ph
2
>3
n-C H
4
9
CH Ph
2
3.7 ± 0.75
3.9 ± 0.58
5.0 ± 0.29
6.2 ± 1.1
13 ± 5.2
64
n-C H
4
9
CH Ph
2
43
CH CH=C(CH
)
3 2
CH Ph
65
2
2
n-C H
(CH ) Ph
2 4
>1,000
942
>150
72
4
9
n-C H
4
9
(CH ) Ph
2 4

1
8
n-C H
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
S)-18c
R)-18c
S)-18d
R)-18d
S)-28b
R)-28b
S)-28c
R)-28c
S)-28d
R)-28d
S)-28e
R)-28e
S)-29b
S)-29c
S)-32b
S)-32c
+)-pentazocine
CH3
120 ± 11
113 ± 24
18 ± 4.0
21 ± 2.9
>1,000
1,000
278
>8
4
9
n-C H
4
9
CH3
8.8
15
n-C H
n-C H
4
9
4
9
n-C H
n-C H
>1,000
>46
53
4
9
4
2
2
2
2
2
2
2
2
9
CH Ph(4-OCH )
2
3
CH Ph
0.86 ± 0.12 46 ± 4.8
0.51 ± 0.06 431
CH Ph(4-OCH )
2
3
CH Ph
845
11
(
(
CH ) Ph
2
4
CH Ph
0.77 ± 0.17 8.6 ± 3.4
0.83 ± 0.06 71 ± 33
0.61 ± 0.12 276
CH ) Ph
2
4
CH Ph
86
n-C H
4
9
CH Ph
452
1005
34
n-C H
4
9
CH Ph
0.37 ± 0.12 372
CH C H
2
6
11
11
11
11
CH Ph
0.22 ± 0.06 7.4 ± 5.0
0.38 ± 0.06 7.0 ± 2.3
CH C H
2
6
CH Ph
18
CH C H
CH Ph(4-OCH ) 1.1 ± 0.17
2
2
3
19 ± 6.4
51 ± 18
>1,000
>1,000
n.d.
17
2
6
CH C H
2
6
CH Ph(4-Cl)
4.9 ± 1.7
20 ± 2.2
6.0 ± 1.0
5.7 ± 2.2
6.3 ± 1.6
89 ± 29
10
CH Ph
CH Ph(4-OCH )
2
3
2
>50
>160
-
CH C H
11
CH Ph
2
2
6
Haloperidol
78 ± 2.3
58 ± 18
13
di-o-tolylguanidine
0.6
#
Generally, the K values were determined in triplicates (n = 3). For compounds
i
showing very low affinity in the first experiment repetitions were not performed (n =
).
n.d. = not determined.
1
Relationship between stereochemistry and σ
low influence of the configuration in 2-position on σ
enantiomers (S)-18c (K = 120 nM) / (R)-18c (K = 113 nM), (S)-17j (K
R)-17j (K = 3.9 nM) and (S)-28e (K = 0.22 nM) / (R)-28e (K = 0.38 nM) represent
1
affinity: The data in Table 1 indicate a
receptor affinity. The pairs of
= 3.7 nM) /
1
i
i
i
(
i
i
i

1
9
examples for medium, high and very high affinity ligands showing very similar
affinities. The highest eudismic ratios were found for the ligands 17g (eudismic
σ
1
ratio 5) and 17h (eudismic ratio 9) bearing 2-hydroxyethyl moieties at 4-position.
However, it should be taken into account that these hydroxyethyl substituted
derivatives 17g and 17h represent 1:1-mixtures of diastereomers. The similar
σ
1
affinity of enantiomers can be explained by the high flexibility of the seven-
membered 1,4-diazepane ring allowing the 2-substituent to adapt to the receptor
binding site independent on its orientation at the 1,4-diazepane ring.
Relationships between 4-substituent and σ
proton in 4-position do not show considerable affinity towards σ
concentration of 1 µM. Whereas a small methyl moiety led to slightly increased
affinity (17i), 1,4-diazepanes 17j and 28d with a larger n-butyl 4-substituent
interact in the low nanomolar to subnanomolar range with σ receptors. The
branched and unsaturated methylbutenyl substituent of (R)-17k was tolerated by
the σ receptor as well. 1,4-Diazepanes (S)-28e (K = 0.22 nM) and (R)-28e (K
.38 nM) bearing the saturated cyclohexylmethyl moiety in 4-position represent the
most potent σ ligands of this class of compounds.
1
affinity: 1,4-diazepanes 17a with a
1
receptors up to a
σ
1
1
1
i
i
=
0
1
In contrast to 1,4-diazepanes with saturated and unsaturated (cyclo)alkyl
substituents, 1,4-diazepanes with phenylalkyl substituents in 4-position show
reduced σ
of the side chain does not influence the σ
by analyzing the homologous series of phenylalkyl substituted 1,4-diazepanes 17c-
7f showing K(σ ) values of 25 – 80 nM. A methoxy moiety at the phenyl ring
1
affinity. The distance between the basic N-atom and the phenyl moiety
1
affinity considerably, which can be seen
1
i
1

2
0
seems to slightly increase the σ
the K(σ ) values of the benzyl derivatives 17c (71 / 80 nM) with the K
the methoxybenzyl derivatives 17b (29 / 42 nM). In conclusion, the σ
increases in the following order of substituents in 4-position: H < CH
CH Ph(4-OCH ) < C < CH
1
affinity, which can be observed by comparison of
(σ ) values of
affinity
< (CH
i
1
i
1
1
3
2
) Ph <
n
2
3
4
H
9
2
C
6
11
H .
Relationships between 2-substituent and σ
amino acids phenylalanine (7), alanine (19) and valine (30) resulted in
,4-diazepanes with a benzyl (series 17/18), methyl (series 28/29) and isopropyl
moiety (series 32) in 2-postion. The alanine derivatives 28 with a 1-benzyl moiety
show very high σ affinity (K(σ ) = 0.22 – 0.86 nM). All alanine derived
,4-diazepanes reveal higher σ affinity than their phenylalanine analogs. Even the
affinity of the phenylalanine derived N-butyl derivatives 17j (K(σ ) = 3.7 / 3.9 nM)
affinity of the corresponding alanine derivatives 28d (K(σ ) =
.61 / 0.37 nM). The valine derived 1,4-diazepanes 32 show intermediate σ affinity
1
affinity: Starting the synthesis with the
1
1
i
1
1
1
σ
1
i
1
is exceeded by the σ
1
i
1
0
1
in a range between the alanine and phenylalanine derivatives. The comparison of
the amino acid derived 1,4-diazepanes including serine-derived ligands 5 reveals
increasing σ
1
affinity in the following order: 2-benzyl < 2-isopropyl ≈ 2-
hydroxymethyl < 2-methyl.
Relationships between 1-substituent and σ
of the 1-substitutent on the σ affinity, the most potent phenylalanine derivatives 17j
with a 4-butyl moiety were selected as a starting point. Replacement of the 1-benzyl
moiety of 17j by a small methyl group (18c) led to 30-fold reduced σ affinity (K(σ
1
affinity: In order to study the influence
1
1
i
1
)

2
1
=
120 /113 nM). Only a slight reduction of the σ
,4-diazepanes 18b and 18d with a phenylbutyl or butyl residue instead of the
benzyl moiety in 1-position. In the alanine series the substitution pattern of the
-benzyl moiety of the most potent ligands 28e with a 4-cyclohexylmethyl moiety
was investigated. Introduction of the electron donating methoxy group (29b) and the
electron withdrawing Cl-atom (29c) led to 5-fold and 25-fold reduction of σ affinity
1
affinity was observed for
1
1
1
compared to 29b with a non-substituted 1-benzyl moiety, respectively. According to
these results, a non-substituted benzyl moiety seems to be the best substituent in
1
-position to achieve high σ affinity.
1
Pharmacophore models
The data obtained for the 1,4-diazepanes fit nicely into the pharmacophore models
60,61
62
63
64
developed by R. Glennon,
T. Langer, S. Pricl, and B. Wünsch. The 1,4-
diazepanes contain the required basic N-atom or H-bond donor group. The
substituents at the N-atoms should interact with two lipophilic pockets at the σ
1
receptor, which explains nicely that a proton or a small methyl moiety in 1- or 4-
position do not result in high σ affinity. Optimal substituents for the 1,4-diazepane
N-atoms are the benzyl moiety (1-postion) and the n-butyl and cyclohexylmethyl
moiety (4-position), which occupy the postulated hydrophobic regions of the σ
receptor binding pocket. A large substituent in 2-position such as a benzyl moiety is
less tolerated by the σ receptor than the smaller methyl moiety.
1
1
1
Receptor selectivity
Selectivity towards the σ
2
subtype: In the σ assay the radioligand
2
3
[
H]di-o-tolylguanidine is used as competitor for the test compounds. Rat liver

2
2
membrane preparations are employed as receptor material and σ
1
receptors are
3
masked with (+)-pentazocine, since [ H]di-o-tolylguanidine represents a non-
57-59
selective σ ligand.
Table 1 summarizes the σ
2
receptor affinity of all test
compounds in comparison to their σ
1
affinity.
In general the σ
considerably lower than their σ
S)-17j (K(σ ) = 3.7 nM) and (R)-17j (K
selectivity over the σ receptor, respectively. However, the σ
ligands of this series is even higher due to their very low σ affinity (e.g. (S)-18b
/σ selectivity >150-fold; (R)-18b σ /σ selectivity = 72-fold). The subnanomolar σ
affinity of the alanine derivatives 28, results in excellent σ /σ selectivity (e.g. (R)-
selectivity = 1005-fold), but in case of (S)-28c the high σ affinity (K(σ ) =
.6 nM) leads to reduced σ /σ selectivity of 11-fold. Due to the low σ affinity of the
/σ selectivity values of >50 and >160 were
/σ selectivity the alanine derivatives 28
represent the most promising ligands.
2
receptor affinity of all 1,4-diazepanes investigated in this work is
affinity. The most potent phenylalanine derivatives
(σ ) = 3.9 nM) show 64- and 43-fold
/σ selectivity of some
1
(
i
1
i
1
2
1
2
2
σ
1
2
1
2
1
1
2
28d σ /σ
1 2
2
i
2
8
1
2
2
valine derivatives 32b and 32c high σ
1
2
obtained, respectively. In terms of σ
1
2
Selectivity towards the PCP binding site of the NMDA receptor: It has been shown
that some ligands interacting with σ receptors are closely related to ligands
interacting with the phencyclidine (PCP) binding site within the channel pore of the
1
65-67
N-methyl-D-aspartate (NMDA) receptor.
Therefore, the PCP affinity of the
1,4-diazepanes prepared in this work was determined in receptor binding studies
3
67,68
using the radioligand [ H]-(+)-MK-801.
At a test compound concentration of
1
µM the 1,4-diazepanes 17, 18, 28, 29, and 32 did not compete with the

2
3
radioligand indicating IC50 values higher than 1 µM and thus high σ /PCP selectivity
1
for these ligands. Only the hydroxyethyl derivatives (R)-17g and (R)-17h show
3
replacement of [ H]-(+)-MK-801 in the range of 50 % at a concentration of 1 µM,
which might result in IC50 values in the range of 1 µM. Nevertheless both ligands
are still very selective for the σ receptor over the PCP binding site.
1
Pharmacological characterization of (S)-28b
The (S)-alanine derived 1-benzyl and 4-(p-methoxybenzyl) substituted 1,4-
diazepane (S)-28b with a subnanomolar σ
1
affinity (K = 0.86 nM) and 53-fold
i
selectivity over the σ subtype was selected for further pharmacological
2
characterization. In a screening against 53 selected drug targets (targets see
Supporting Information), the 1,4-diazepane (S)-28b (conc. 1 µM) did not show any
cross reactivity. Only a very weak interaction with histamine H
observed. At a concentration of 1 µM 50 % of the radioligand were displaced from
the histamine H receptor indicating an IC50 value of approx. 1 µM. Altogether, (S)-
8b showed a clean receptor profile.
2
receptor was
2
2
In order to get an idea about the metabolic stability, (S)-28b was incubated with
mouse, rat and human microsomal preparations. After an incubation period of 5
min, less than 50 % of the parent compound remained unchanged indicating fast
metabolic degradation of (S)-28b. To avoid potential interactions the inhibition of the
most important CYP enzymes by (S)-28b was determined. At a concentration of 1
µM the following inhibition of CYP enzymes was recorded: 0 % (CYP 2C9), 38 %
(
CYP 2C19), 60 % (CYP 2D6) and 7 % (CYP 3A4). At a concentration of 60 µM
only 1.6 % of (S)-28b were bound to human serum albumin.

2
4
It has been reported that σ
1
receptor agonists are able to potentiate electrically
stimulated contractions of the guinea pig vas deferens. This effect is dose
69
dependent and can be suppressed by σ
of 30 µM, (S)-28b potentiated the electrically induced contractions indicating
agonistic activity. The potent σ agonist (S)-(+)-3-(3-hydroxyphenyl)-1-
propylpiperidine ((+)-3-PPP, K = 6 nM) was tested as a reference compound. The
agonistic effect of (S)-28b (concentration 30 µM) was 29 % of the effect of
+)-3-PPP (concentration 100 µM). This assay proves clearly the σ agonistic
activity of the 1,4-diazepane (S)-28b.
1
receptor antagonists. At a concentration
1
i
σ
1
(
1
70
The Irwin test was used to analyze the acute toxicity and the effects on behavior
and physiological function of mice of (S)-28b. After i.p. application of 100 mg/kg
body weight of (S)-28b to mice, increased abdominal tone, diarrhea and writhing
were observed. Reduction of the dose to 30 mg/kg body weight resulted only in
diarrhea for the first 60 min. A dose of 1 mg/kg was well tolerated by the mice and
did not lead to any effects. These observations confirm the clean pharmacological
profile of (S)-28b.
71
In the two-phase analgesic formalin assay, (S)-28b did not show significant
effects. Independent of the application of (S)-28b or vehicle, the same licking
behavior was observed during the early phase (0-5 min). In the late phase (10-35
min) a slight but not significant reduction of licking time was determined after
application of the highest dose of 10 mg/kg.
Finally, (S)-28b was tested in a cognition assay, in which object recognition
7
2
following scopolamine (0.3 mg/kg) amnesia was analyzed. The 1,4-diazepane (S)-

2
5
28b showed clear cognition enhancing effects at a dose of 30 mg/kg without any
side effects.
Inhibition of ergosterol and cholesterol biosynthesis
8/7
Due to the structural similarity of the σ
1
receptor with the yeast enzyme sterol-∆ -
5-7
isomerase, the inhibition of this enzyme by selected compounds (Table 2) was
73,74
investigated exemplarily in a cellular assay.
these compounds did not at all inhibit the growth or change the sterol pattern of the
model yeast Yarrowia lipolytica. This finding indicates the high selectivity for the σ
Even at a concentration of 4 µg/mL
1
8/7
receptor over the fungal sterol-∆ -isomerase.
Table 2. Qualitative screening results for cholesterol biosynthesis inhibition:
Accumulated sterols I - IV under incubation with test compound and identified target
enzymes A – D. First line: test compound concentration 1 µM; second line test
compound concentration 50 µM.
accumulated sterol(s)
inhibited
compd.
enzyme(s)
I
II
III
IV
-
-
+
-
-
-
-
-
-
-
-
+
-
-
-
-
-
+
+
+
++
-
+
-
+
-
+
-
++
-
-
+
+
-
-
-
-
-
-
-
-
-
C
(
(
(
(
(
(
(
(
(
S)-17a
R)-17a
S)-17b
R)-17b
S)-17c
R)-17c
S)-17d
R)-17d
S)-17e
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
A, B, C, D
A, C
A, C
A, C, D
A, D
A, C
A
A, C
A
A, B, C
A
-
+
-
A, C
A
A, C
+
toxic at 50 µM
+
toxic at 50 µM
-
+
+
A, C, D

2
6
+
-
+
-
+
+
-
+
++
+
+
A, B, C, D
D
A, B, C, D
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
R)-17e
S)-17f
R)-17f
S)-17i
R)-17i
S)-17j
R)-17j
R)-17k
S)-28c
R)-28c
S)-28d
R)-28d
S)-28e
R)-28e
S)-32b
S)-32c
++
toxic at 50 µM
++
++
+
+
++
+
-
-
+
-
-
+
-
+
-
+
-
-
-
+
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
A, B, C
-
+
+
+
+
+
+
+
+
+
+
+
+
++
+
A, B, C
A, B
A, B, C
A
+
++
A
A, B
A, C
A
A, B
A, B
A, B
+
++
+
++
+
++
-
toxic at 50 µM
+
+
+
+
+
+
+
+
+
+
+
-
+++
+++
+
++
+
++
++
+++
++
+++
+
++
+
++
-
-
+
+
+
-
-
-
+
-
+
-
-
A, B
++
++
+
+
+
++
+
-
-
+
-
A, B, D
A, B, C, D
A, B, C, D
A, B, C, D
A, B, D
A, B
A, B
A, B, C, D
A, B
++
++
+
+
+
+
-
-
A, B, C, D
B
A, B
++
-
-
-
-
B
Accumulated sterols: I cholesta-8-en-3β-ol (zymostenol), II cholesta-8,14-dien-3β-
ol, III cholesta-7-en-3β-ol (lathosterol), IV cholesta-5,7-dien-3β-ol (7-
8/7
24
dehyrocholesterol); target enzymes: A sterol ∆ -isomerase, B sterol ∆ -reductase,
5
7
C sterol ∆ -desaturase, D sterol ∆ -reductase; - sterol not detected, +/++/+++ high
amount of accumulated individual sterols.
Further, the compounds were subjected to a cellular assay for inhibition of enzymes
75
in the post-squalene part of cholesterol biosynthesis. For qualitative screening
detection of target enzymes), two concentrations (1 µM and 50 µM) were initially
tested in a whole cell assay on HL-60 cells. Cholesterol biosynthesis intermediates
(

2
7
were analyzed by GC-MS in a targeted metabolomics fashion after derivatization to
their corresponding trimethylsilyl (TMS) ethers. After treatment with inhibitor, the
accumulated sterols were identified and compared to blank samples (without
treatment) and among each other. A marker sterol for inhibition of the enzyme sterol
8/7
76,77
24
∆
-isomerase (A) is cholesta-8-en-3β-ol (zymostenol, I),
for sterol ∆ -
78
5
reductase (B) cholesta-5,24-dien-3β-ol (desmosterol, II), for sterol ∆ -desaturase
79
7
(
C) cholesta-7-en-3β-ol (lathosterol, III), and for sterol ∆ -reductase (D) cholesta-
80
5
,7-dien-3β-ol (7-dehydrocholesterol, IV). Cholesta-8,14-dien-3β-ol is an indicator
24
for a dual inhibition of the up-stream enzyme sterol ∆ -reductase and the down-
8/7
75,78
stream enzyme sterol ∆ -isomerase
.
Compounds (R)-17b, (S)-17c, (R)-17c, (S)-17d, and (R)-17j were identified as
8/7
selective inhibitors of the human sterol ∆ -isomerase (A) at 50 µM. Only compound
S)-17j was selective and significantly active at the lower concentration of 1 µM.
Hence, this compound was subjected to a quantitative assay for cholesterol
(
75,80
biosynthesis inhibition.
HL-60 cells were incubated with various concentrations
13
of compound (S)-17j in presence of excess 2- C-labeled acetate. Newly
13
synthesized cholesterol could be determined by the incorporation of 2- C-acetate
units. A sigmoidal dose-response curve was obtained by plotting the test compound
concentrations against the peak area of the labelled cholesterol. An IC50 value of
2
7
42 nM was determined (95 % confidence interval: 628-877 nM; goodness of fit R :
0.934), which describes the overall effect on cellular cholesterol biosynthesis. Thus,
8/7
(
S)-17j can be considered as a moderately active inhibitor of ∆ -isomerase in
cholesterol biosynthesis. The best inhibitors of this enzyme are in the low
76
nanomolar range.

2
8
The alanine derivatives 28 with excellent σ
1
activity evolved as multi-enzyme
inhibitors in cholesterol biosynthesis, especially at very high test concentrations.
24
8/7
The enzymes sterol ∆ -reductase (B) and the sterol ∆ -isomerase (A), which are
mainly affected by these compounds, have in common that the enzymatic
conversion goes through carbenium ions (high energy intermediates), which can be
75,76
imitated by protonated aliphatic amines bearing lipophilic substituents.
81
Prominent examples of such multiple inhibitors are triparano and various
76,82
(
di)azasteroids.
4. Conclusion
The good correlation with reported σ
1
pharmacophore models stimulated the
design, synthesis and pharmacological evaluation of a large set of (2S)- and (2R)-
configured 1,2,4-trisubstituted 1,4-diazepanes. It was found that the configuration in
2
-position does not have a great impact on σ
receptors. Alanine derived 1,4-diazepanes with a methyl moiety in 2-position show
the highest σ affinity. Substituted or unsubstituted benzyl moieties in 1-position and
aliphatic (e.g. n-butyl) or cycloaliphatic substituents (e.g. cyclohexyl) in 4-position
seem to be optimal to reach high σ affinity.
1
affinity and selectivity over related
1
1
8/7
Despite the sequence homology between the σ
isomerase, the 1,4-diazepanes did not inhibit the fungal enzyme indicating strong
preference for the σ receptor. However, several 1,4-diazepanes were able to inhibit
1
receptor and the yeast sterol-∆ -
1
8/7
the human sterol ∆ -isomerase at rather high concentration of 50 µM. But only the
butyl substituted 1,4-diazepane (S)-17j (K (σ ) = 3.7 nM) inhibited selectively the
human sterol ∆ -isomerase at a concentration of 1 µM. This led to high nanomolar
i
1
8/7

2
9
inhibition of cholesterol biosynthesis by (S)-17j (IC50 = 742 nM).
The (S)-alanine derived 1,4-diazepane (S)-28b, which binds in the subnanomolar
range at σ
1
receptors (K
i
(σ ) = 0.86 nM), was pharmacologically investigated in
1
more detail. It showed a clean receptor profile in a screening against more than 50
target proteins. In the guinea pig vas deferens model (S)-28b potentiated dose-
dependently electrically stimulated contractions. This behavior indicates σ agonistic
1
activity of (S)-28b. In the Irwin screen (in vivo) (S)-28b did not lead to severe toxic
or behavioral effects even at the very high dose of 100 mg/kg. Obviously, (S)-28b
was well tolerated by mice. After inducing amnesia by application of scopolamine,
(
S)-28b was able to ameliorate cognition deficits of mice.
5
. Experimental, Chemistry
.1. General Methods
5
Unless otherwise noted, moisture sensitive reactions were conducted under dry
nitrogen. THF was dried with sodium/benzophenone and was freshly distilled before
use. Thin layer chromatography (tlc): Silica gel 60 F254 plates (Merck). Flash
chromatography (fc): Silica gel 60, 40–64 µm (Merck); parentheses include:
diameter of the column, eluent, fraction size, R value. Melting point: Melting point
f
apparatus SMP 3 (Stuart Scientific), uncorrected. MS: MAT GCQ (Thermo-
1
Finnigan); IR: IR spectrophotometer 480Plus FT-ATR-IR (Jasco). H NMR (400
13
MHz), C NMR (100 MHz): Unity Mercury Plus 400 spectrometer (Varian); δ in ppm
related to tetramethylsilane; coupling constants are given with 0.5 Hz resolution.
Optical rotation: Polarimeter 341 (Perkin Elmer); 1.0 dm tube; concentration c in
g/100 mL; T = 20 °C; wavelength 589 nm (D-line of Na light); the unit of the specific
T
.
.
-1. -1
rotation ([α]
D
grad mL dm g ) is omitted for clarity.