Imidazole-derived 2-[N-carbamoylmethyl-alkylamino]acetic acids, substrate-dependent modulators of insulin-degrading enzyme in amyloid-β hydrolysis

Article abstract of DOI:10.1016/j.ejmech.2014.04.009

Insulin degrading enzyme (IDE) is a highly conserved zinc metalloprotease that is involved in the clearance of various physiologically peptides like amyloid-beta and insulin. This enzyme has been involved in the physiopathology of diabetes and Alzheimer's disease. We describe here a series of small molecules discovered by screening. Co-crystallization of the compounds with IDE revealed a binding both at the permanent exosite and at the discontinuous, conformational catalytic site. Preliminary structure-activity relationships are described. Selective inhibition of amyloid-beta degradation over insulin hydrolysis was possible. Neuroblastoma cells treated with the optimized compound display a dose-dependent increase in amyloid-beta levels.

Full text of DOI:10.1016/j.ejmech.2014.04.009

European Journal of Medicinal Chemistry 79 (2014) 184e193
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Imidazole-derived 2-[N-carbamoylmethyl-alkylamino]acetic acids,
substrate-dependent modulators of insulin-degrading enzyme in
amyloid-b hydrolysis
,
b
,
c
,
d
,
,
b
,
c
,
d
,
,c,g,h
Julie Charton a
e, Marion Gauriot a
e, Qing Guo k, Nathalie Hennuyer b
,
Xavier Marechal a
e, Julie Dumont a
e, Malika Hamdane b
,
,
b
,
c
,
d,
,
b
,
c
,
d
,
,
i,
j
Virginie Pottiez a
e, Valerie Landry a
e, Olivier Sperandio e f, Marion Flipo a
,
,
b
,
c
,
d,
,
b
,
c
,
d,
,
,
b,c,d,e
Luc Buee b j, Bart Staels b
h, Florence Leroux a
e, Wei-Jen Tang k,
,
i,
,
c
,
g
,
,
b,
c
,
d
,
Benoit Deprez a
1, Rebecca Deprez-Poulain a
,
b
,
c
,
d
,
e
, ,
,
b
,
c
,
d,
e
, ,
1
*
*
a INSERM U761 Biostructures and Drug Discovery, Lille, France
b Univ Lille Nord de France, Lille F-59000, France
c Institut Pasteur de Lille, IFR 142, Lille F-59000, France
d PRIM, Lille F-59000, France
e CDithem Platform/IGM, Paris, France
f Inserm UMR-S 973/MTi, University Paris Diderot, Paris, France
g INSERM U1011 Nuclear Receptors, Cardiovascular Diseases and Diabetes, Lille F-59000, France
h European Genomic Institute for Diabetes (EGID), FR 3508, Lille F-59000, France
i INSERM U837 Neurodegenerative Diseases and Neuronal Death, Lille F-59000, France
j CHRU, Lille F-59000, France
k Ben-May Institute for Cancer Research, The University of Chicago, W421 Chicago, IL, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Insulin degrading enzyme (IDE) is a highly conserved zinc metalloprotease that is involved in the
clearance of various physiologically peptides like amyloid-beta and insulin. This enzyme has been
involved in the physiopathology of diabetes and Alzheimer’s disease. We describe here a series of small
molecules discovered by screening. Co-crystallization of the compounds with IDE revealed a binding
both at the permanent exosite and at the discontinuous, conformational catalytic site. Preliminary
structureeactivity relationships are described. Selective inhibition of amyloid-beta degradation over
insulin hydrolysis was possible. Neuroblastoma cells treated with the optimized compound display a
dose-dependent increase in amyloid-beta levels.
Received 7 February 2014
Received in revised form
31 March 2014
Accepted 4 April 2014
Available online 4 April 2014
Keywords:
Enzymes
Medicinal chemistry
Structureeactivity relationships
Amyloid-beta peptides
Inhibitors
Ó 2014 Elsevier Masson SAS. All rights reserved.
X-ray diffraction
1. Introduction
Abbreviations: EtOAc, ethyl acetate; AcOH, acetic acid; Boc, tert-butoxycarbonyl;
CH3CN, acetonitrile; DCE, dichloroethane; DCM, dichloromethane; DIEA, N,N-dii-
sopropylethylamine; DMF, dimethylformamide; DMSO, dimethylsulfoxide; EDCI, N-
ethyl-3-(3-dimethylaminopropyl)carbodiimide; Et3N, triethylamine; EtOH, ethanol;
HOBt, N-hydroxybenzotriazole; MeOH, methanol; PBS, phosphate buffered saline;
PTSA, para-toluenesulfonic acid; rt, room temperature; SAR, structureeactivity
relationship; TFAA, Trifluoroacetic anhydride; THF, tetrahydrofuran; TMB, 3,30,5,50-
tetramethylbenzidine.
Insulin-degrading enzyme (IDE) is an atypical, highly conserved
and ubiquitous metalloprotease of the M16 family [1,2]. It is mainly
intracellular [3] but a small amount is trafficked extracellularly [4].
IDE shows a unique structure [5,6]: it is organized in two 56 kDa
halves which are joined by a 28 amino acid residue loop [7], and
enclose a large, almost spherical, catalytic chamber, named “crypt”.
Using X-ray crystallography Tang et al. have shown that sub-
strates bind to two distant sites in the chamber. To let substrates in,
* Corresponding authors. INSERM U761 Biostructures and Drug Discovery, Lille,
France.
E-mail addresses: benoit.deprez@univ-lille2.fr (B. Deprez), rebecca.deprez@
univ-lille2.fr (R. Deprez-Poulain).
ꢀ
the enzyme opens. The binding of substrates to the exosite, 30 A
1
Home pages: http://www.insulysin.org, http://www.deprezlab.fr.
http://dx.doi.org/10.1016/j.ejmech.2014.04.009
0223-5234/Ó 2014 Elsevier Masson SAS. All rights reserved.

J. Charton et al. / European Journal of Medicinal Chemistry 79 (2014) 184e193
185
away from the catalytic Zinc, promotes a conformational change of
the substrate that is critical for cleavage at the catalytic site [8,9]
Also, recently, via X-ray crystallography, an unexpected displace-
ꢀ
ment (swinging-door) of a subdomain of IDE that creates an 18 A
opening to the chamber, that permits the entry of short peptides
[10].
Along with insulin, substrates of IDE include amyloid-
b (Ab)
[11], Insulin-Growth Factor-II IGF-II [12,13], glucagon [11], so-
matostatin [14] most of which are amyloidogenic [15] (Fig. 1A).
Ubiquitin [16] and CGRP [17] are also shown as IDE substrates. IDE
prefers hydrophobic or basic residues at P1 and P’1 and substrates
that lack a positive charge at the C-terminus [1]. Interestingly, be-
side its peptidolytic role, IDE interacts with and regulates the
proteasome complex [18]. Also, the activity of IDE can be modu-
lated by the binding of ATP to the catalytic chamber [19].
Silencing IDE expression with siRNA reduces insulin-mediated
protein degradation [20]. In several animal models, deletion of
Ide gene or mutations in the gene result in elevated insulin levels
and glucose intolerance, associated with elevated A
[21]. Also, transgenic overexpression of IDE in neurons results in
significantly reduced levels of A in brain and retards plaque for-
b in the brain
b
mation in amyloid precursor protein (APP) transgenic mice [22]. In
addition, Ide gene was linked to type-2 diabetes (T2D) and Alz-
heimer’s disease (AD) in humans [23,24].
Small organic molecules are complementary to genomic or
transcriptomic interventions, because they are systems modulators
and not “erasers” of protein activity. They help to understand the
target’s function and can be transcribed into therapeutically agents
by modulation of the proteolytic profile (i.e. inhibition or activa-
tion), the distribution pattern and the chronically or temporarily
inhibition of IDE. Leissring et al. described the first substrate-based
zinc-binding hydroxamate inhibitors of IDE [25]. However, their
hydroxamate group [26] combined with an arginine residue limit
their use as pharmacological probes. The poorly bioavailable sur-
amin [27] and two other compounds identified in a cell-free assay,
were reported to be activators [28]. Also, by molecular modelling,
Çakir et al. have discovered compounds that activate the hydrolysis
of several substrates of IDE and Kukday et al. described activators
specific of the rat orthologue [29].
Fig. 2. X-ray crystal structure of hIDE-CF co-crystallized with 1 (PDB code: 4DTT). A)
general view, B) detailed interactions at the exosite C) detailed interactions at the
catalytic site. Zn (red sphere); O (red); N (blue); C (cyan for IDE N-terminal domain,
grey for IDE C-terminal domain, orange for 1); hydrogen contacts in dotted lines. Ar-
rows point residues shown to be critical by mutagenesis. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of
this article.)
hydrolytic profile of IDE as well as their effect on neuroblastoma
cells expressing amyloid-b.
Herein, we describe the discovery of drug-like ligands of IDE
with an atypical binding mode and report their impact on the
2. Results and discussion
2.1. Hit identification
2.1.1. Screening of a 2000-member library
A library of 2000 drug-like compounds (Supporting Information
Fig. S1 and [30]) was screened in a fluorescence assay for the in-
hibition of the hydrolysis of dual labelled substrate (Ab16e23
Km ¼ 100 M) by IDE. Compounds showing dose-dependent in-
hibition were defined as hits. 1 (BDM41367, Fig. 1B), was the most
active compound (IC50 b16e23: 2.9 M). Interestingly, the dosee
:
m
A
m
response curve plateaus at 50% inhibition (Fig. 1C). A study of the
enzymatic mechanism evidenced that 1 is a reversible, partial and
competitive inhibitor of the hydrolysis of
(Supporting Information Fig. S2).
Ab16e23 by IDE
2.1.2. Binding of the hit
Using X-ray crystallography Tang et al. have shown that sub-
strates bind to two distant sites in the chamber. To understand how
1 interacts with IDE, we co-crystallized it catalytically active IDE
enzyme and solved the structure by molecular replacement (PDB
code ¼ 4DTT, Supplemental information and Table S1). Surprisingly,
in the co-crystal, 1 is observed at these two binding sites (Fig. 2A)
[31]. At the exosite (Fig. 2B), the imidazole ring forms a hydrogen
bond with Glu341, a residue that binds the N-terminus of IDE
substrates. The amide and amine functions interact with Gly361
Fig. 1. A) IDE’s substrates, B) structure of 1 (BDM41367) inhibitor of labelled Ab16e23
hydrolysis C) by IDEwt (C IC50
¼ 2.9
mM).
Ab16e23

186
J. Charton et al. / European Journal of Medicinal Chemistry 79 (2014) 184e193
and Gln363 main chains. At the catalytic site, 1 interacts with res-
idues from both the N- and C-terminal domains (Fig. 2C). The
carboxylate group of 1 completes the zinc coordination sphere
formed by His108, His112 and Glu189. 1 also forms hydrogen bonds
with side chains of Tyr831 and Asn139 and the main chain of
Val833 via its imidazole ring. The benzyl group of 1 makes a hy-
drophobic contact with the side chain of Phe115. The negatively
charged Glu111 is located in the vicinity of the ionizable imino
function of 1. In line with this observation, when 1 is co-crystallized
with the mutated, catalytically inactive E111Q enzyme, it is found
only at the exosite (PDB code ¼ 2YPU, Supporting Information
Table S1, Fig. S3). The binding of 1 to the exosite was further sub-
stantiated by an enzymatic assay using the E341A exosite mutant.
In this mutant, the glutamate shown to interact with 1 in the
exosite is replaced with a neutral alanine [16]. With this enzyme, 1
iminodiacetic precursors were prepared by alkylation of iminodi-
acetic acid with alkyl bromides (Scheme 1) or by acylation using a
3-step procedure: i. formation of the diester using MeOH and SOCl2,
ii. acylation of the amine by benzoylchloride, iii. saponification of
esters (Scheme 1). Synthesis of desired compounds proceeded first
by in situ conversion of the corresponding N-alkylated of N-acyl-
ated iminodiacetic precursors into a cyclic anhydride with TFAA in
acetic anhydride. Anhydrides are then reacted with a histidine or
phenylalanine derivative to give compounds 1e3, 6 and 7 after a
deprotection step if needed (Scheme 1). Finally, amide analogues 4,
5 were obtained by reacting methyl esters 1 and 3 respectively with
methylamine, in refluxed methanol overnight.
2.2.2. Structureeactivity relationships
Table 1 gathers results of the preliminary SARs. We found that
homologation of the benzylalkyl chain consistently improves IC50
(3 versus 1, 5 versus 4). Also, the replacement of the methyl ester by
a methylamide group led to more potent compounds (5 versus 3, 4
versus 1). Compound 5, bearing the phenylpropyl chain, is the most
behaves as a full inhibitor (IC50 Ab16e23: 5.8 mM, Fig. 3).
Comparing the X-ray structures obtained with proteins that
differ at the catalytic site on one hand, and enzyme inhibitions
measured with protein that differ at the exosite on the other hand,
we propose the following hypothesis to explain the peculiar
behaviour of 1. IDE has an enclosed catalytic chamber. During
catalysis, IDE needs to open to give access to substrates or modu-
lators to the catalytic chamber [5,6]. In the open state, the non
contiguous catalytic site (formed by both IDE-N and IDE-C residues)
is not fully assembled. However, in this open conformation, the
exosite remains as a binding site for 1. Upon capture of the substrate
and closure of the catalytic chamber, 1 could dissociate and
compete with the substrate either at the exosite or at the re-
assembled catalytic site. The equilibrium between 1 bound to the
exosite and to the catalytic site in a 1:2 ligand:site ratio could
explain why 1 behaves as a partial inhibitor in the wild type IDE and
as a full inhibitor in the exosite-mutated enzyme [32].
potent compounds (BDM43079, IC50
A
b16e23 ¼ 100 nM). In this
series, all active compounds plateau at 50%. This is consistent with
their retained ability to bind to the exosite that we observed in X-
ray crystal structures (Supporting Information Fig. S3). Expectedly,
the replacement of the imidazole ring by a phenyl ring (6) or the
replacement of the benzyl by a benzoyl (7), completely abolish
activity in line with the binding requirements described for the
enzyme and the E111Q mutant.
2.3. Characterization of 5
2.3.1. Substrate-dependent activity
The effect of the best inhibitor 5 on the hydrolytic profile of IDE
was assessed using five unlabelled native substrates (Fig. 4).
Expectedly 5 inhibits the hydrolysis of amyloid-b1e40 by IDE (IC50
2.2. Design and synthesis of analogues
A
b1e40: 3.1 ꢀ 2.0
effect of 5 depends on the nature of the substrate. Indeed, it pro-
M) and
mM) (Fig. 4A). However, in contrast with EDTA, the
2.2.1. Synthesis
motes the hydrolysis of insulin (EC50 insulin: 0.53 ꢀ 0.39
m
We made a limited number of variations at key interaction
points of 1 with IDE to confirm its binding mode. Non commercial
IGF-II (Fig. 4B, C), and has no effect on the hydrolysis of somato-
statin or glucagon (Fig. 4D, E).
The absence of inhibition of insulin hydrolysis could be due to
the much higher affinity of this substrate to the enzyme
(Km ¼ 85 nM) [33]. Also, to bind to the exosite, amyloid-
b uses two
acidic residues while insulin and IGF-II has rather neutral residues.
The changes in the molecular surface of the exosite of 1-bound-IDE
may contribute to the differences in substrate susceptibility to the
ligands. The activating effect observed could also be mediated by
the reduction of IDE aggregation observed in a Small-Angle X-Ray
scattering experiment (Supporting Information Fig. S4).
2.3.2. Selectivity of 5
Before using 5 in cellular systems, we checked its selectivity
against a panel of related metalloproteases including hNEP and
hECE, two other metalloproteases implicated in the hydrolysis of
amyloid-b. 5 proved to be selective with IC50s above 100 mM for
hNEP, hECE, hACE, hACE2, hTACE, hMMP1 and hMMP13.
2.4. Cellular activity
2.4.1. Synthesis and characterization of precursor 8
To assess the effect of 5 on cells that natively express IDE, we
used SH-SY5Y neuroblastoma cells that produce amyloid- from
b
hAPP751 with the Swedish mutation (APPswe) or hAPP696 (APPwt)
(Fig. S5) [34]. As one could expect from its zwitterionic behaviour, 5
does not cross easily cell membranes. We thus synthesized the
methyl ester 8 (BDM43124, Fig. 5, Scheme 2). Compound 8 was
Fig. 3. 1 is an inhibitor of the dual labelled Ab16e23 substrate hydrolysis by IDEwt (C
IC50 ¼ 2.9
m
M) and IDE E341A (- IC50 ¼ 5.3
mM).

J. Charton et al. / European Journal of Medicinal Chemistry 79 (2014) 184e193
187
Scheme 1. Synthesis of carboxylic acids 1e7. (Reactants and conditions: a) ReBr, MeOH, DIEA or TEA or KOH, room temp., 2e12 h b) trifluoroacetic anhydride 2% in acetic anhydride,
20e70 ꢂC, 5 h, 100%. c) histidine derivative or phenylalanine methyl ester, anhydrous DIEA, anhydrous DMF, Argon, room temp., overnight. d) TFA 5%/DCM, triisopropylsilane, 1 h,
room temp. e) MeNH2/EtOH, MeOH, reflux. f) i. 80/20 MeOH/SOCl2, overnight, room temp. ii. Benzoylchloride, DIEA, CH2Cl2 overnight, room temp. iii. NaOH, MeOH, H2O.)
synthesized similarly to analogues 1e7. Opening of the anhydride
was performed with methanol, then the resulting monoacid com-
pound was reacted with the histidine derivative with EDCI and
HOBt. Deprotection of the trityle function in acidic conditions gave
8 (Scheme 2).
The carboxylic acid is essential for binding to the catalytic site
but not to the exosite. In agreement with this binding mode
observed in this series, 8 is inactive regarding Ab16e23 hydrolysis
contrasted effects on insulin and amyloid-
enzymatic assays.
b hydrolysis in cell-free
Altogether, this work brings new modulators of IDE that
differentiate from previously described inhibitors [39] by their
binding mode and their selectivity toward Amyloid beta hydrolysis,
leaving intact insulin degradation in cell models. Moreover, in a
cell-based assay, using 8 as a selective pharmacological reagent, we
proved that IDE is involved in the degradation of amyloid-b 1e40
(IC50 > 100
hydrolysis (50% of 5 activity at 100
m
M) and shows some residual activation of insulin
and 1e42. This is in line with previous implication of IDE in Alz-
heimer’s disease by genomic and genetic experiments.
mM). This compound has a
higher logD7.4 while retaining a good solubility and is cell perme-
able (0.2 10ꢁ6 cm sꢁ1). We checked that it hydrolyses readily to give
back the active acid 5 (conversion half lives ¼ 6 min in plasma and
11.3 h in cell culture medium) (Fig. 5) [35].
Along with the results on cell-models, the stability, solubility
and selectivity properties of 5 and its precursor 8 (BDM43124),
support their use as a probes for the exploration of the role of IDE in
the amyloid pathway without affecting the insulin pathway.
More importantly our results strongly consolidate previous
observations that pointed out IDE as an important player in the
amyloid degradation and encourage the search for drugs that
would either activate directly IDE, or increase its expression, to
clear amyloid peptide efficiently.
2.4.2. Cellular activity of 8
SH-SY5Y cells [36] treated with 8 display a dose-dependent
increase in levels of both Ab1e40 and
(Fig. 6) [37]. Differences
1e42
between 24 and 48 h may reflect the slow permeation kinetic of 8.
No effect on APP holoproteins and on CTFs was observed, which is
consistent with the inhibition of amyloid-
firms the implication of IDE in the catabolism of amyloid-
We also use -pancreatic cells and hepatocytes which we show
express IDE. 8 did neither affect the level of insulin secreted by
b
degradation. This con-
4. Experimental section
b
[38].
b
4.1. Biology
b
-
pancreatic cells nor modify the repression of PEPCK by insulin on
hepatocytes. This result suggests that 8 does not inhibit insulin
degradation by IDE in a cellular context and confirm the substrate
selectivity observed in the biochemical assays (Supporting
Information Fig. S6). The activating effect observed in a biochem-
ical setting, which we linked to the reduction of aggregation of the
enzyme, may not be relevant in a much more complex cellular
context.
4.1.1. In vitro IDE activity assay
In vitro IDE activity was measured with a quenched substrate
ATTO 655- Cys-Lys-Leu-Val-Phe-Phe-Ala-Glu-Asp-Trp. Human re-
combinant IDE was cloned and purified as previously described.
Briefly, human IDE (1.87 ng/m
L) was incubated 10 min at 37 ꢂC with
compound in Hepes 50 mM, NaCl 100 mM, pH 7.4 and the enzy-
matic reaction is started by adding the substrate (final concentra-
tion 5 mM). After 30 min, samples (1% DMSO final) are excited at
635 nm and fluorescence emission at 750 nm is measured on a
Victor3 V1420 Perkin Elmer spectrophotometer. All measurements
were carried out as 8-point dose response curves and are reported
as the average of at least three independent measurements. EDTA
was used as a reference inhibitor (100% inhibition at 2 mM).
3. Conclusion
Altogether, we report here the discovery and characterization of
small dual ligands of IDE. Unlike the previously described hydrox-
amate inhibitors, the lead compound 1 binds to the catalytic site
ꢀ
and a distal exosite in the N-terminal domain of IDE 30 A away from
4.1.2. Cell culture of SH-SY5Y stable cell lines
the zinc atom [5].
Human neuroblastoma cell lines SH-SY5Y expressing either
human APP751 with the Swedish mutation (APPSw) or human
APP696 (SY5Y-APPwt) were generated as previously described [34].
The optimized compound 5 modulates the proteolytic profile of
IDE in
a substrate-dependent manner. In particular, it has

188
J. Charton et al. / European Journal of Medicinal Chemistry 79 (2014) 184e193
Table 1
Inhibition of the hydrolysis of Ab16e23 by IDE by 1e7.
Cmpd
R1e
Ar
eR2
IC50 Ab16e23 hydrolysis [m
M]a
1 (BDM41367)
eOCH3
2.9 ꢀ 1.9
1.4 ꢀ 1.3
0.6 ꢀ 0.9
0.6 ꢀ 0.6
0.1 ꢀ 0.5
>100
2
eOCH3
eOCH3
eNHCH3
eNHCH3
eOCH3
3
4
5 (BDM43079)
6
7
eOCH3
>100
a
IC50 values are given as the mean standard error from 3 to 7 independent experiments performed in quadruplicate.
Cells were maintained in Dulbecco’s modified Eagle medium
(DMEM, GIBCO BRL, France) supplemented with 10% foetal calf
serum, 2 mM L-glutamine, 1 mM non-essential amino-acids, 50
and then incubated with a Horseradish peroxidase-conjugated sec-
ondary antibody (Goat anti-rabbit A4914 from SigmaeAldrich and
Horse anti-mouse from Vector Laboratories). Finally, peroxidase ac-
tivity was revealed with the ECL detection kit and visualized with
HyperfilmÔ ECLÔ (Amersham/GE Healthcare France). Quantitative
units/mL penicillin/streptomycin (Invitrogen, France) in a 5% CO2
humidified incubator at 37 ꢂC.
determinations of amyloid
dium at the end of drug treatments were performed using human
amyloid 1e42 and amyloid 1e40 Elisa kits (Invitrogen). Following
initial experiments for determining optimal dilution factors, dosages
b 1e42 and amyloid b 1e40 in cell me-
4.1.3. Cellular assay in the amyloid pathway
Cells were seeded in 6-well plates (1.94105 cells/well). 24 h later,
cells were replenished with 2 mL of medium containing either
vehicle or drug at 0.1 mM or 10 mM (from a stock solution at 10 mM in
DMSO). Following 24 h and 48 h of drug exposure, the cell medium
b
b
were performed with 100
(undiluted sample) and 10
m
m
L of cell medium for amyloid
L of cell medium for amyloid
b
b
1e42
1e40
(dilution factor of 1:10). The data was obtained in duplicate from
three independent experiments and normalized to untreated cells.
was collected, discarded from cell debris by centrifugation (1000 ꢃ g,
10 min) and was kept at ꢁ80 ꢂC for ELISA dosages of amyloid
b 1e42
and amyloid
b 1e40. For preparation of total cellular proteins,
adherent cells were washed with PBS and scraped with a rubber
policeman in 50 L of RIPA buffer (Thermo Scientific) containing
4.2. Chemistry
m
protease inhibitors (Complete Mini, Roche Molecular Biochemicals,
Meylan, France). The cell homogenate was sonicated and stirred
10 min at 4 ꢂC. Cell lysate was recovered in supernatant after
centrifugation at 12,000 ꢃ g, 10 min at 4 ꢂC. Protein concentration
was determined using the BCA protein assay kit (Pierce) and samples
were kept at ꢁ80 ꢂC. Proteins were mixed to LDS Sample Buffer
containing a reducing agent (Invitrogen NP-009) and boiled at 100 ꢂC
4.2.1. General information
NMR spectra were recorded on a Bruker Avance 300 spec-
trometer. Chemical shifts are in parts per million (ppm) and were
referenced to the residual proton peaks in deuterated solvents. The
assignments were made using one dimensional (1D) 1H and 13C
spectra (classical or Jmod) and two-dimensional (2D) HSQC, HMBC,
ROESY and COSY spectra. Mass spectra were recorded with a LCe
MSeMS tripleequadrupole system (Varian 1200ws) or a LCMS
(Waters Alliance Micromass ZQ 2000). LCMS analysis was per-
for 10 min.12
SDS-polyacrylamide gel (Novex NuPAGEÒ Invitrogen), blotted onto
nitrocellulose membrane (0.44 m Hybond and Hybond-Phosphate
from Amersham/GE Healthcare France). For APP-CTF analysis,
30 g of total proteins were loaded on Tris-Tricine SDS-poly-
mg of total proteins were loaded onto a 4e12% Bis-Tris-
m
formed using a C18 TSK-GEL Super ODS (2 mm particle size column,
dimensions 50 mm ꢃ 4.6 mm). A gradient starting from 100% H2O/
0.1% formic acid and reaching 20% H2O/80% CH3CN/0.08% formic
acid within 10 min (method A) or 5 min (method B) at a flow rate of
1 mL/min was used. Preparative HPLC were performed using a
Varian PRoStar system using an OmniSphere 10 column C18
250 mm ꢃ 41.4 mm Dynamax from Varian, Inc. A gradient starting
from 20% CH3CN/80% H2O/0.1% formic acid and reaching 100%
m
acrylamide gel electrophoresis (Criterion precast gel, Bio-Rad Bio-
research division, Marnes la coquette, France). Proteins were
transferred to nitrocellulose membrane (0.22 mM Hybond, Amer-
sham Biosciences). Membranes were blocked andincubated with the
appropriate antibody according to Supporting Information Table 2

J. Charton et al. / European Journal of Medicinal Chemistry 79 (2014) 184e193
189
Fig. 4. Modification of hIDE peptidolytic profile by 5 (BDM43079) measured on full unlabelled native substrates. A) amyloid-b1e40, B) insulin C) IGF-II, D) somatostatin, E) glucagon.
Values are given as the mean of 3 experiments. [hIDE] ¼ 7.5
m
g/mL. [5] ¼ 100 M; [EDTA] ¼ 50 mM. EDTA inhibits hIDE independently of the substrate’s nature by chelating the zinc,
m
whereas 5 behaves either as an activator or an inhibitor depending of the substrate.
CH3CN/0.1% formic acid at a flow rate of 80 mL/min or 20% MeOH/
80% H2O/0.1% formic acid reaching 100% MeOH/0.1% formic acid
was used. Purity (%) was determined by reversed phase HPLC, using
UV detection (215 nm), and all compounds showed purity greater
than 95%; Melting points were determined on a Büchi B-540
apparatus and are uncorrected. All commercial reagents and sol-
vents were used without further purification. Organic layers ob-
tained after extraction of aqueous solutions were dried over MgSO4
and filtered before evaporation under reduced pressure. Purifica-
tion yields were not optimized. Thick layer chromatography was
bromo-3-phenylpropane (348 mL, 2.28 mmol). The mixture was
stirred at room temperature for 10 h and then was refluxed for 24 h.
The solvent was evaporated and the crude product was purified by
preparative HPLC to give the compound (103 mg, 27%) Purity 99%;
1H NMR (DMSO-d6)
J ¼ 7.5 Hz, 2H), 2.65 (t, J ¼ 7.5 Hz, 2H), 3.41 (s, 4H), 7.13e7.26 (m, 5H).
d
ppm: 1.67 (dt, J ¼ 7.5 Hz, 2H), 2.56 (t,
performed with Silica Gel 60 (Merck, 40e63 mm).
4.2.2. (Carboxymethyl-phenethyl-amino)-acetic acid
To a stirred solution of iminodiacetic acid (500 mg, 2.7 mmol) in
MeOH (30 mL) was added KOH (454 mg, 8.1 mmol) and (2-
bromoethyl)benzene (549 mg, 2.97 mmol). The mixture was
refluxed overnight, cooled at room temperature and filtrated. The
filtrate was evaporated and the crude product was purified by
preparative HPLC to give the compound as a white solid (230 mg,
36%). Purity 100%; 1H NMR (DMSO-d6)
d
ppm: 7.22 (m, 5H), 3.46 (s,
ppm: 172.5,
4H), 2.86 (m, 2H), 2.70 (m, 2H). 13C NMR (DMSO-d6)
d
140.0, 128.6, 128.3, 125.9, 55.8, 54.7, 33.8. LC tR ¼ 1.12 min, MS
(ESIþ): m/z ¼ 238 (M þ H)þ.
Fig. 5. Probe profile for 5 and its methyl ester 8. MW in g molꢁ1; Solubility: in PBS
from a DMSO stock solution; LogD7.4: PBS/octanol partition from a DMSO stock solu-
tion; Plasma stability: mixed gender mouse plasma lithium heparin; Cell medium
stability: William’s E medium supplemented with 10% foetal calf serum.
4.2.3. [Carboxymethyl-(3-phenyl-propyl)-amino]-acetic acid
To a stirred solution of iminodiacetic acid (277 mg, 2.08 mmol)
and TEA (870 mL, 6.24 mmol) in methanol (10 mL) was added 1-

190
J. Charton et al. / European Journal of Medicinal Chemistry 79 (2014) 184e193
Scheme 2. Synthesis of precursor 8. (Reactants and conditions: a) trifluoroacetic anhydride 2% in acetic anhydride, 20e70 ꢂC, 5 h, 100%. b) MeOH, DIEA, reflux, overnight; c) H-L-
His(1-Trt)-NHMe, EDCI, HOBt, DIEA, DMF, room temp., overnight; d) TFA 5%/DCM, triisopropylsilane, 1 h, room temp.)
Fig. 6. A
b
Cellular responses to 8 (BDM43124) methyl ester precursor of 5. Effect of 0.1 or 10
mM of 8 on extracellular levels of Ab1e40 (A, B) and Ab1e42 (C, D) in different APP SY5Y
strains at 24 h (,) or 48 h (-). A) and C) APPswe, B) and D) APPwt. 8 increases A
b
peptides levels in a dose-dependent manner. Results are normalized to A levels in vehicle
b
treated cells. *p < 0.05, **p < 0.01.
13C NMR (DMSO-d6)
d
ppm: 172.6, 141.9, 128.3, 126.1, 125.7, 57.5,
J ¼ 0.9 Hz, 1H), 7.30 (m, 5H), 8.26 (d, J ¼ 0.9 Hz, 1H), 8.34 (d,
53.7, 32.5, 28.8. LC tr ¼ 2.76 min; MS (ESIþ): m/z ¼ 252 (M þ H)þ.
J ¼ 8.0 Hz,1H); 13C NMR (DMSO-d6)
d ppm: 172.3,171.7, 170.4, 138.2,
137.9, 129.0, 128.4, 127.4, 116.0, 57.6, 56.7, 53.6, 52.2, 51.6, 28.4. LC
tR ¼ 2.30 min, MS (ESIþ): m/z ¼ 375 (M þ H)þ. HRMS (m/z): [MHþ]
calcd. for C18H23O5N4, 375.1668; found, 375.1665.
4.2.4. (S)-2-[2-(Benzyl-carboxymethyl-amino)-acetylamino]-3-
(1H-imidazol-4-yl)-propionic acid methyl ester (1) (BDM41367)
To a stirred solution of
L-histidine methyl ester dihydrochloride
(324 mg, 1.34 mmol) in DMF (13.4 mL) were added DIEA (704
m
L)
4.2.5. (S)-2-[2-(Carboxymethyl-phenethyl-amino)-acetylamino]-3-
(1H-imidazol-4-yl)-propionic acid methyl ester (2)
(Carboxymethyl-phenethyl-amino)-acetic acid (210 mg,
0.88 mmol) was stirred in a 2% TFAA in acetic anhydride solution at
40 ꢂC during 4 h and solvent were evaporated. To a stirred solution
and N-benzyliminodiacetic anhydride acid (250 mg, 1.22 mmol) in
THF (2.24 mL). The mixture was stirred overnight at room tem-
perature and the solvent was evaporated. The crude product was
purified by preparative HPLC to yield compound 1 (BDM41367) as a
white solid (20%). Purity 100%; 1H NMR (DMSO-d6)
d
ppm: 2.99 (d,
of
anhydrous DMF (10 mL) were added dry DIEA (510
anhydride (0.88 mmol) in anhydrous DMF (2 mL). The mixture was
L-histidine methyl ester dihydrochloride (245 mg, 1.01 mmol) in
J ¼ 6.3 Hz, 2H), 3.23 (s, 2H), 3.27 (s, 2H), 3.59 (s, 3H), 3.71 (d,
mL) and the
J ¼ 13.4 Hz, 1H), 3.77 (d, J ¼ 13.4 Hz, 1H), 4.65 (m, 1H), 6.93 (d,

J. Charton et al. / European Journal of Medicinal Chemistry 79 (2014) 184e193
191
stirred overnight at room temperature under argon and the solvent
was evaporated. The crude product was purified by preparative
HPLC to yield compound 2 as a white solid (200 mg, 52%). Purity
The mixture was stirred at room temperature for 1 h. The solvent
was evaporated and the crude product was triturated in diethyl
ether to give 4 as colourless oil (70%). Purity: 98%; 1H NMR (DMSO-
97%; 1H NMR (DMSO-d6)
d
ppm:8.22 (d, J ¼ 7.8 Hz, 1H), 8.14 (s, 1H,
d6)
d
ppm: 8.95 (s, 1H), 8.39 (d, J ¼ 8.1 Hz, 1H), 8.00 (q, J ¼ 4.5 Hz,
HCOOH), 7.56 (d, J ¼ 1.1 Hz, 1H), 7.28e7.22 (m, 2H), 7.18e7.14 (m,
3H), 6.82 (d, J ¼ 1.0 Hz, 1H), 4.52 (m, 1H), 3.58 (s, 3H), 3.38 (s, 2H),
3.25 (s, 2H), 2.92 (m, 2H), 2.76 (m, 2H), 2.66 (m, 2H); LC
tR ¼ 2.63 min, MS (ESIþ): m/z ¼ 389 (M þ H)þ. HRMS (m/z): [MHþ]
calcd. for C19H25O5N4, 389.1825; found, 389.1820.
1H), 7.33 (m, 5H), 7.28 (s, 1H), 4.57 (m, 1H), 3.91 (s, 2H), 3.51 (s, 2H),
3.49 (s, 2H), 3.14 (dd, J ¼ 5.4 Hz and J ¼ 15.3 Hz, 1H), 2.93 (dd,
J ¼ 8.1 Hz and J ¼ 15.3 Hz, 1H), 2.60 (d, J ¼ 4.5 Hz, 3H); 13C NMR
(DMSO-d6)
d ppm: 170.8, 169.8, 168.4, 135.4, 133.9, 129.8, 129.3,
128.5, 128.1, 116.8, 57.9, 55.9, 53.7, 51.3, 27.3, 25.7; LC: tr ¼ 2.55 min;
MS (ESIþ): m/z ¼ 374 (M þ H)þ. HRMS (m/z): [MHþ] calcd. for
4.2.6. (S)-2-{2-[Carboxymethyl-(3-phenyl-propyl)-amino]-
acetylamino}-3-(1H-imidazol-4-yl)-propionic acid methyl ester
.2TFA (3)
C18H24O4N5, 374.1828; found, 374.1823.
4.2.8. ((((S)-2-(1H-Imidazol-4-yl)-1-methylcarbamoyl-
ethylcarbamoyl)-methyl)-(3-phenyl-propyl)-amino)-acetic acid.
2TFA (5) (BDM43079)
The diacid (0.4 mmol) was stirred in a 2% TFAA in acetic anhy-
dride solution at 40 ꢂC during 4 h and solvent were evaporated. To a
stirred solution of L-(Trt)Histidine methyl ester hydrochloride
[Carboxymethyl-(3-phenyl-propyl)-amino]-acetic
acid
(0.4 mmol) was stirred in a 2% TFAA in acetic anhydride solution at
40 ꢂC during 4 h and solvent were evaporated. To a stirred solution
of L-(Trt)Histidine methyl ester hydrochloride (174 mg, 0.39 mmol)
and DIEA (200 mL, 1.17 mmol) in anhydrous DMF (3 mL) was added
the anhydride (99 mg, 0.39 mmol). The mixture was stirred over-
night at room temperature under argon and the solvent was
evaporated. The crude product was purified by preparative HPLC to
give the compound as a white solid (142 mg, 57%). Purity 98%; 1H
(174 mg, 0.39 mmol) and DIEA (200 mL, 1.17 mmol) in anhydrous
DMF (3 mL) was added the anhydride (0.39 mmol). The mixture
was stirred overnight at room temperature under argon and the
solvent was evaporated. The crude product was purified by pre-
parative HPLC to give the compound as a white solid (63%, formiate
NMR (DMSO-d6)
d
ppm: 1.62 (dt, J ¼ 7.4 Hz, 2H), 2.5 (m, 2H), 2.59 (t,
J ¼ 7.4 Hz, 2H), 2.88 (m, 2H), 3.19 (s, 2H), 3.28 (s, 2H), 3.51 (s, 3H),
4.58 (ddd, J ¼ 6 Hz, 8.1 Hz,1H), 6.62 (s,1H), 6.99e7.38 (m, 21H), 8.49
salt). Purity 92%; 1H NMR (DMSO-d6)
d
ppm: 1.62 (dt, J ¼ 7.4 Hz,
2H), 2.5 (m, 2H), 2.59 (t, J ¼ 7.4 Hz, 2H), 2.88 (m, 2H), 3.19 (s, 2H),
3.28 (s, 2H), 3.51 (s, 3H), 4.58 (ddd, J ¼ 6 Hz, 8.1 Hz, 1H), 6.62 (s, 1H),
6.99e7.38 (m, 21H), 8.49 (d, J ¼ 8.2 Hz, 1H). 13C NMR (DMSO-d6)
(d, J ¼ 8.2 Hz, 1H). 13C NMR (DMSO-d6)
d ppm: 29.3, 29.8, 32.7, 51.8,
51.8, 53.8, 55.1, 57.8, 119.0, 125.6, 127.5, 127.8, 128.0, 128.2, 129.2,
136.3,137.6, 142.1, 170.6, 171.7, 172.7. LC tr ¼ 5.21 min; MS (ESIþ): m/
z ¼ 645 (M þ H)þ. The trityl group was removed in TFA 5%/DCM.
Triisopropylsilane was added until the yellow colour disappears.
The mixture was stirred at room temperature for 1 h. The solvent
was evaporated and the crude product was triturated in diethyl
ether to give 3 as a white solid (69 mg, 55%). Purity 99%; 1H NMR
d
ppm: 29.3, 29.8, 32.7, 51.8, 51.8, 53.8, 55.1, 57.8, 119.0, 125.6, 127.5,
127.8, 128.0, 128.2, 129.2, 136.3, 137.6, 142.1, 170.6, 171.7, 172.7. LC
tr ¼ 5.21 min; MS (ESIþ): m/z ¼ 645 (M þ H)þ. To a stirred solution
of methyl ester (0.25 mmol) in methanol (0.5 mL) were added
311 mL of a methylamine solution in ethanol (33%). The mixture was
refluxed overnight. The solvent was evaporated and the crude
product was precipitated in an aqueous solution of formic acid to
give [{[(S)-1-Methylcarbamoyl-2-(1-trityl-1H-imidazol-4-yl)-eth-
ylcarbamoyl]-methyl}-(3-phenyl-propyl)-amino]-acetic acid (49%)
as the formiate salt. Purity: 97%; LC: tr ¼ 5.27 min; MS (ESIþ): m/
z ¼ 644 (M þ H)þ. The trityl group was removed in TFA 5%/DCM.
Triisopropylsilane was added until the yellow colour disappears.
The mixture was stirred at room temperature for 1 h. The solvent
was evaporated and the crude product was triturated in diethyl
ether to give 5 (BDM43079). 2TFA as colourless oil (58%). Purity:
(DMSO-d6) d ppm: 1.79 (m, 2H), 2.57 (m, 2H), 2.90 (m, 2H), 3.06 (dd,
J ¼ 8.6 and 15.4 Hz, 1H), 3.19 (dd, J ¼ 5.4 and 15.4 Hz, 1H), (m, 2H),
3.62 (s, 3H), 3.70 (s, 2H), 3.78 (s, 2H), 4.70 (m, 1H), 7.17e7.30 (m,
5H), 7.41 (s, 1H), 8.86 (d, J ¼ 7.2 Hz, H), 8.98 (s, 1H). 13C NMR (DMSO-
d6)
d ppm: 26.1, 26.8, 32.2, 51.2, 52.4, 54.7, 54.8, 55.7, 117.2, 126.0,
128.3, 128.4, 128.9, 133.7, 141.1, 167.5, 169.7, 170.6. LC tr ¼ 3.06 min;
MS (ESIþ): m/z ¼ 403 (M þ H)þ. HRMS (m/z): [MHþ] calcd. for
C
20H27O5N4, 403.1981; found, 403.1978.
4.2.7. (Benzyl-{[(S)-2-(1H-imidazol-4-yl)-1-methylcarbamoyl-
ethylcarbamoyl]-methyl}-amino)-acetic acid. 2TFA (4)
98%; 1H NMR (DMSO-d6)
d
ppm: 9.05 (d, J ¼ 8.7 Hz, 1H), 9.01 (s, 1H),
8.22 (q, J ¼ 4.8 Hz, 1H), 7.38 (s, 1H), 7.32e7.19 (m, 5H), 4.58 (m, 1H),
4.13 (m, 2Hþ1H), 4.08 (d, J ¼ 6.6 Hz, 1H), 3.17e3.13 (m, 2H þ 1H),
2.96 (dd, J ¼ 8.7 Hz and J ¼ 14.8 Hz, 1H), 2.60e2.57 (m, 3H þ 2H),
To a stirred solution of L-His(Trt)-OMe hydrochloride (1.5 mmol)
in DMF (15 mL) were added DIEA (1 eq) and N-benzyliminodiacetic
anhydride acid (1.4 mmol) in THF (3 mL). The mixture was stirred
overnight at room temperature and the solvent was evaporated.
The crude product was purified by precipitation in H2O as a white
1.92 (m, 2H); 13C NMR (DMSO-d6)
d ppm: 169.5, 167.7, 165.2, 140.5,
134.1, 129.2, 128.4, 128.3,126.1, 116.8, 55.3, 54.9, 54.1, 52.1, 31.8, 27.0,
25.7, 25.3; LC: tr ¼ 3.25 min; MS (ESIþ): m/z ¼ 402 (M þ H)þ. HRMS
(m/z): [MHþ] calcd. for C20H28O4N5, 402.2141; found, 402.2137.
powder (80%) Purity 97%; 1H NMR (DMSO-d6)
d ppm: 2.91e2.94 (m,
2H), 3.24 (s, 2H), 3.26 (s, 2H), 3.52 (s, 3H), 3.77 (s, 2H), 4.58e4.65
(m, 1H), 6.66 (s, 1H), 7.01e7.04 (m, 6H), 7.15e7.18 (m, 5H), 7.33 (s,
1H), 7.35e7.38 (m, 9H), 8.47(d, J ¼ 8.1 Hz, 1H). 13C NMR (DMSO-d6)
4.2.9. (S)-2-(2-(Benzyl-carboxymethyl-amino)-acetylamino)-3-
phenyl-propionic acid methyl ester. (6)
d
ppm: 29.5, 51.7, 51.9, 53.1, 56.9, 57.4, 74.7, 119.2, 127.2, 127.6, 127.8,
To a stirred solution of L-Phe-OMe hydrochloride (1.5 mmol) in
DMF (15 mL) were added DIEA (1 eq) and N-benzyliminodiacetic
anhydride acid (1.4 mmol) in THF (3 mL). The mixture was stirred
overnight at room temperature and the solvent was evaporated.
The crude product was purified by precipitation in H2O to give 6 as
128.1, 128.2, 128.9, 129.3, 136.1, 137.9, 138.1, 142.1, 170.1, 171.6, 172.1.
LC tr ¼ 5.08 min; MS (ESIþ): m/z ¼ 617 (M þ H)þ. To a stirred so-
lution of methyl ester (0.5 mmol) in methanol (1 mL) were added
622 mL of a methylamine solution in ethanol (33%). The mixture was
refluxed overnight. The solvent was evaporated and the crude
product was precipitated. The product was purified by preparative
HPLC to give the formiate salt of (Benzyl-{[(S)-1-methylcarbamoyl-
2-(1-trityl-1H-imidazol-4-yl)-ethyl-carbamoyl]-methyl}-amino)-
acetic acid (100%). Purity: 97%; LC: tr ¼ 4.5 min; MS (ESIþ): m/
z ¼ 616 (M þ H)þ. The trityl group was removed in TFA 5%/DCM.
Triisopropylsilane was added until the yellow colour disappears.
a colourless oil. Yield 40%. 1H NMR (DMSO-d6)
d ppm: 8.13 (d,
J ¼ 8.10 Hz,1H), 7.31e7.16 (m,10H), 4.58 (m,1H), 3.69 (d, J ¼ 13.2 Hz,
1H), 3.64 (d, J ¼ 13.2 Hz, 1H), 3.61 (s, 3H), 3.22 (s, 2H), 3.16 (s, 2H),
3.08 (dd, J ¼ 5.4 and 13.8 Hz, 1H), 2.96 (dd, J ¼ 8.7 and 13.8 Hz, 1H);
13C NMR (MeOD)
d ppm: 173.8, 173.0, 172.5, 137.7, 130.6, 130.3,
129.6, 128.9, 128.1, 59.5, 57.7, 55.6, 54.6, 52.8, 38.3
tR,LCMS ¼ 4.39 min, Purity 100%; MS (ESIþ): m/z ¼ 385 [M þ H]þ.

192
J. Charton et al. / European Journal of Medicinal Chemistry 79 (2014) 184e193
HRMS (m/z): [MHþ] calcd. for C21H25O5N2, 385.1763; found,
evaporated and the crude product was washed with petroleum
385.1756.
ether and diethyl ether to give 8.2TFA as yellow oil (165 mg, 99%).
Purity: 99%; 1H NMR (DMSO-d6)
d ppm: 8.97 (s, 1H), 8.53 (d,
4.2.10. (S)-2-(2-(Benzoyl-carboxymethyl-amino)-acetylamino)-3-
(1H-imidazol-4-yl)-propionic acid methyl ester (7)
Iminodiacetic acid was stirred overnight in a 80/20 MeOH/SOCl2
mixture. The diester obtained after evaporation under reduced
pressure (300 mg, 1.52 mmol) was dissolved in CH2Cl2 (10 mL) and
J ¼ 7.9 Hz, CONH), 8.04 (q, J ¼ 4.7 Hz, CONH), 7.33e7.14 (m, 6H), 4.58
(m, 1H), 3.82 (s, 2H), 3.67 (s, 3H), 3.62 (s, 1H), 3.13 (dd, J ¼ 5.12 Hz
and J ¼ 15.36 Hz, 1H), 2.92 (dd, J ¼ 8.4 Hz and J ¼ 15.3 Hz, 1H), 2.87
(t, J ¼ 7.4 Hz, 2H), 2.59 (d, J ¼ 4.2 Hz, 3H), 2.53 (m, 2H), 1.77 (qt,
J ¼ 7.4 Hz, 2H); 13C NMR (DMSO-d6)
d ppm: 169.8,169.2,167.7,141.2,
DIEA (792
mL, 4.56 mmol) and benzoylchloride (176
mL, 1.52 mmol)
134.4, 129.4, 128.4, 128.3, 126.0, 116.9, 56.3, 54.9, 54.5, 52.1, 51.5,
32.3, 27.3, 27.1, 25.8; LC: tr ¼ 4.07 min; MS (ESIþ): m/z ¼ 644
(M þ H)þ. HRMS (m/z): [Mþ] calcd. for C21H30O4N5, 416.2298;
found, 416.2284.
were added. The mixture was stirred overnight at room tempera-
ture and washed with HCl 0.5 N and NaHCO3 10% solutions. The
organic layer was dried over MgSO4 and evaporated to give the
desired diester as a colourless oil (99%). The former compound
(400 mg, 1.5 mmol) was saponified with NaOH (6 mmol) in MeOH
(5 mL) and H2O (1 mL). After evaporation of MeOH under reduced
pressure, HCl 1 N was added to the residue and the aqueous layer
was extracted 3 times with AcOEt to give (benzoyl-carboxymethyl-
amino)-acetic acid directly used in the next step. White powder.
Yield 98%. tR,LCMS ¼ 2.53 min (5 min gradient), Purity 98%; MS
(ESIþ): m/z ¼ 238 [M þ H]þ. The diacid (0.74 mmol) was dissolved
in trifluoroacetic anhydride 2% in acetic anhydride (3 mL). The re-
action mixture was stirred for 4 h at room temperature then
Acknowledgements
The authors acknowledge assistance from C. Piveteau, S. Das-
sonneville, H. Drobecq and financial support from INSERM, Uni-
versity of Lille 2, Pasteur Institute of Lille, Region Nord Pas de Calais,
EDFR, Etat (0823007, 0823008, 07-CPER 009-01, 2007-0172-02-
CPER/3), ANR (11-JS07-015-01), FRM (DCM20111223046), NIH
GM81539 (to WJT) and PRIM (Pôle de Recherche Interdisciplinaire
pour le Médicament). We thank the LARMNS NMR laboratory,
School of Pharmacy Lille. M.G. is a recipient of a fellowship from
Ministère de la Recherche et de l’Enseignement Superieur. Data
management was performed using Pipeline PilotÔ from Accelrys
Inc.
evaporated. To a stirred solution of
drochloride (188 mg, 0.77 mmol) in DMF (8.0 mL) were added DIEA
(515 L) and the anhydride (0.74 mmol) in DMF (1.0 mL). The
L-histidine methyl ester dihy-
m
mixture was stirred overnight at room temperature and the solvent
was evaporated. The crude product was purified by preparative
HPLC to give 7 as a white solid. Yield 26%. 1H NMR (DMSO-d6)
Appendix A. Supplementary data
d
ppm: 2.85e3.02 (m, 2H), 3.59(s, 1.2H), 3.64 (s, 1.8H), 3.88 (s, 2H),
4.02 (s, 1.2H), 4.08 (s, 0.8H), 4.55 (m, 1H), 6.86 (s, 0.6H), 6.91 (s,
Supplementary data associated with this article can be found in
the online version, at http://dx.doi.org/10.1016/j.ejmech.2014.04.
009. These data include mechanistic studies, enzyme substrates,
supplementary protocols, protein crystallography, NMR spectra,
MOL files and InChiKeys of the most important compounds
described in this article.
0.4H), 7.31e7.44 (m, 5H), 7.69 (s, 0.6H), 7.74 (s, 0.4H), 8.78 (d,
J ¼ 7.5 Hz, 0.6H), 8.89 (J ¼ 7.2 Hz, 0.4H). 13C NMR (MeOD)
d ppm
175.9, 174.9, 172.1, 172.0, 136.7, 135.7, 131.3, 129.6, 127.8, 118.7, 56.6,
55.9, 53.0, 28.3. tR,LCMS ¼ 2.48 min; Purity 98%; MS (ESIþ): m/
z ¼ 389 [M þ H]þ. HRMS (m/z): [MHþ] calcd. for C18H21O6N4,
389.1461; found, 389.1471.
References
4.2.11. ((((S)-2-(1H-Imidazol-4-yl)-1-methylcarbamoyl-ethyl-
carbamoyl)-methyl)-(3-phenyl-propyl)-amino)-acetic acid methyl
ester. 2TFA (8) (BDM43124)
[1] A. Fernandez-Gamba, M.C. Leal, L. Morelli, E.M. Castano, Current Pharma-
ceutical Design 15 (2009) 3644e3655.
[2] L.B. Hersh, Cellular and Molecular Life Sciences 63 (2006) 2432e2434.
[3] M.A. Leissring, W. Farris, X. Wu, D.C. Christodoulou, M.C. Haigis, L. Guarente,
D.J. Selkoe, Biochemical Journal 383 (2004) 439e446.
[4] K. Glebov, S. Schütze, J. Walter, Journal of Biological Chemistry 286 (2011)
22711e22715.
[5] Y. Shen, A. Joachimiak, M.R. Rosner, W.J. Tang, Nature 443 (2006) 870e874.
[6] E. Malito, R.E. Hulse, W.J. Tang, Cellular and Molecular Life Sciences 65 (2008)
2574e2585.
[7] H. Im, M. Manolopoulou, E. Malito, Y. Shen, J. Zhao, M. Neant-Fery, C.Y. Sun,
S.C. Meredith, S.S. Sisodia, M.A. Leissring, W.J. Tang, Journal of Biological
Chemistry 282 (2007) 25453e25463.
[8] N. Noinaj, S.K. Bhasin, E.S. Song, K.E. Scoggin, M.A. Juliano, L. Juliano,
L.B. Hersh, D.W. Rodgers, PLoS ONE 6 (2011) e20864.
[9] M. Manolopoulou, Q. Guo, E. Malito, A.B. Schilling, W.J. Tang, Journal of Bio-
logical Chemistry 284 (2009) 14177e14188.
[10] L.A. McCord, W.G. Liang, E. Dowdell, V. Kalas, R.J. Hoey, A. Koide, S. Koide, W.-
J. Tang, PNAS 110 (2013) 13827e13832.
To a stirred solution of H-L-His(Trt)-OMe (2.5 g, 5.5 mmol) in
methanol (10 mL) were added 4 mL of a solution of methylamine in
ethanol (33%). The mixture was refluxed overnight. The solvent was
evaporated and the crude product was precipitated in water and
filtrated. The crude product was purified by preparative HPLC to
give H-L-His(1-Trt)-NHMe as a white powder (1.43 g, 63%, formiate
salt). Purity: 74%; 1H NMR (DMSO-d6)
d ppm: 8.30 (s, 1H), 7.95 (q,
J ¼ 4.8 Hz, 1H), 7.06e7.43 (m, 15H), 6.63 (s, 1H), 3.53 (dd, J ¼ 5.7 Hz
and J ¼ 7.2 Hz,1H), 2.77 (dd, J ¼ 5.7 Hz and J ¼ 14.4 Hz,1H), 2.63 (dd,
J ¼ 5.7 Hz and J ¼ 14.4 Hz, 1H), 2.54 (d, J ¼ 4.8 Hz, 3H); LC:
tr ¼ 4.79 min; MS (ESIþ): m/z ¼ 411 (M þ H)þ. To a stirred solution
of H-L-His(1-Trt)-NHMe (456 mg, 1.1 mmol) in DMF (5 mL) was
added the anhydride of [Carboxymethyl-(3-phenyl-propyl)-
amino]-acetic acid (1 mmol), HOBt (149 mg, 1.1 mmol), EDCI
[11] I.V. Kurochkin, S. Goto, FEBS Letters 345 (1994) 33e37.
[12] R.A. Roth, M.L. Mesirow, K. Yokono, S. Baba, Endocrine Research 10 (1984)
101e112.
(211 mg, 1.1 mmol) and DIEA (690
mL, 4 mmol). The mixture was
[13] Q. Guo, M. Manolopoulou, Y. Bian, A.B. Schilling, W.-J. Tang, Journal of Mo-
lecular Biology 395 (2010) 430e443.
[14] C. Ciaccio, G.R. Tundo, G. Grasso, G. Spoto, D. Marasco, M. Ruvo, M. Gioia,
E. Rizzarelli, M. Coletta, Journal of Molecular Biology 385 (2009) 1556e1567.
[15] I.V. Kurochkin, FEBS Letters 427 (1998) 153e156.
[16] L.A. Ralat, V. Kalas, Z. Zheng, R.D. Goldman, T.R. Sosnick, W.-J. Tang, Journal of
Molecular Biology 406 (2011) 454e466.
[17] Y.-G. Kim, A.M. Lone, W.M. Nolte, A. Saghatelian, Proceedings of the National
Academy of Sciences of the United States of America 109 (2012) 8523e8527.
[18] R.G. Bennett, J. Fawcett, M.C. Kruer, W.C. Duckworth, F.G. Hamel, Journal of
Endocrinology 177 (2003) 399e405.
stirred overnight. Then 213 mg of EDCI and 200
m
L of DIEA were
added and the resulting slurry was stirred overnight. The solvent
was evaporated. The crude product was purified by preparative
HPLC to give [{[(S)-1-Methylcarbamoyl-2-(1-trityl-1H-imidazol-4-
yl)-ethylcarbamoyl]-methyl}-(3-phenyl-propyl)-amino]-acetic
acid methyl ester (171 mg, 26%). Purity: 92%; LC: tr ¼ 6.14 min; MS:
(ESIþ): m/z ¼ 658 (M þ H)þ. To a stirred solution of triisopro-
pylsilane and DCM (0.5 mL/4 mL) was added the protected com-
pound (171 mg, 0.26 mmol). Then 500
mL of TFA were added. The
[19] N. Noinaj, E.S. Song, S. Bhasin, B.J. Alper, W.K. Schmidt, L.B. Hersh,
D.W. Rodgers, Journal of Biological Chemistry 287 (2011) 48e57.
mixture was stirred at room temperature for 2 h. The solvents were

J. Charton et al. / European Journal of Medicinal Chemistry 79 (2014) 184e193
193
[20] (a) J. Fawcett, W.C. Duckworth, Diabetologia 52 (2009) 1457e1460;
(b) J. Fawcett, P.A. Permana, J.L. Levy, W.C. Duckworth, Archives of
Biochemistry and Biophysics 468 (2007) 128e133.
[21] (a) W. Farris, S. Mansourian, M.A. Leissring, E.A. Eckman, L. Bertram,
C.B. Eckman, R.E. Tanzi, D.J. Selkoe, American Journal of Pathology 164 (2004)
1425e1434;
[27] F. Hoffman La Roche AG. WO 2006066847.
[28] C. Cabrol, M.A. Huzarska, C. Dinolfo, M.C. Rodriguez, L. Reinstatler, J. Ni,
L.A. Yeh, G.D. Cuny, R.L. Stein, D.J. Selkoe, M.A. Leissring, PLoS ONE 4 (2009)
e5274.
[29] (a) B. Cakir, O. Dagliyan, E. Dagyildiz, I. Baris, I.H. Kavakli, S. Kizilel, M. Türkay,
PLoS ONE 7 (2012) e31787;
(b) H. Fakhrai-Rad, A. Nikoshkov, A. Kamel, M. Fernström, J.R. Zierath,
S. Norgren, H. Luthman, J. Galli, Human Molecular Genetics 9 (2000) 2149e
2158.
(b) S.S. Kukday, S.P. Manandhar, M.C. Ludley, M.E. Burriss, B.J. Alper,
W.K. Schmidt, Journal of Biomolecular Screening 17 (2012) 1348e1361.
[30] J. El Bakali, L. Maingot, J. Dumont, H. Host, A. Hocine, N. Cousaert,
S. Dassonneville, F. Leroux, B. Deprez, R. Deprez-Poulain, MedChemComm 3
(2012) 469e474.
[22] M.A. Leissring, W. Farris, A.Y. Chang, D.M. Walsh, X. Wu, X. Sun, M.P. Frosch,
D.J. Selkoe, Neuron 40 (2003) 1087e1093.
[23] R. Sladek, G. Rocheleau, J. Rung, C. Dina, L. Shen, D. Serre, P. Boutin, D. Vincent,
A. Belisle, S. Hadjadj, B. Balkau, B. Heude, G. Charpentier, T.J. Hudson,
A. Montpetit, A.V. Pshezhetsky, M. Prentki, B.I. Posner, D.J. Balding, D. Meyre,
C. Polychronakos, P. Froguel, Nature 445 (2007) 881e885.
[24] (a) L. Bertram, D. Blacker, K. Mullin, D. Keeney, J. Jones, S. Basu, S. Yhu,
M.G. McInnis, R.C. Go, K. Vekrellis, D.J. Selkoe, A.J. Saunders, R.E. Tanzi, Science
290 (2000) 2302e2303;
[31] The electron-density is compatible with a 1:1 ligand-enzyme ratio and a
random occupancy on each site.
[32] Deprez-Poulain et al. EP_20100330093425.
[33] V. Chesneau, M.R. Rosner, Protein Expression and Purification 19 (2000) 91e
98.
[34] (a) Expression of IDE was checked by Q-RT PCR (Supporting Information
Figure S5). (b) V. Vingtdeux, M. Hamdane, S. Begard, A. Loyens, A. Delacourte,
J.C. Beauvillain, L. Buee, P. Marambaud, N. Sergeant, Neurobiology of Disease
25 (2007) 686e696.
(b) J.A. Prince, L. Feuk, H.F. Gu, B. Johansson, M. Gatz, K. Blennow, A.J. Brookes,
Human Mutation 22 (2003) 363e371;
(c) W.Q. Qiu, M.F. Folstein, Neurobiology of Aging 27 (2006) 190e198.
[25] (a) M.A. Leissring, E. Malito, S. Hedouin, L. Reinstatler, T. Sahara, S.O. Abdul-
Hay, S. Choudhry, G.M. Maharvi, A.H. Fauq, M. Huzarska, P.S. May, S. Choi,
T.P. Logan, B.E. Turk, L.C. Cantley, M. Manolopoulou, W.-J. Tang, R.L. Stein,
G.D. Cuny, D.J. Selkoe, PLoS ONE 5 (2010) e10504;
[35] K. Beaumont, R. Webster, I. Gardner, K. Dack, Current Drug Metabolism 4
(2003) 461e485.
[36] {Eketjall, 2013 #2765}.
[37] Non toxic doses as revealed by MTT cell viability assay (data not shown).
[38] M. Behl, Y. Zhang, W. Zheng, Cerebrospinal Fluid Research 6 (2009) 11.
[39] 5 differs from hydroxamic inhibitor Ii1 that binds to uniquely to Zn2þ making
it a pan-substrate inhibitor. A direct comparison shows that the two com-
pounds have similar Ligand Efficiencies (LE), (i.e. 5: LE ¼ 0.29 vs Ii1: LE ¼ 0.21)
C. Abad-Zapatero, J.T. Metz, Drug Discovery Today 10 (2005) 464.
(b) S.O. Abdul-Hay, A.L. Lane, T.R. Caulfield, C. Claussin, J. Bertrand, A. Masson,
S. Choudhry, A.H. Fauq, G.M. Maharvi, M.A. Leissring, Journal of Medicinal
Chemistry 56 (2013) 2246e2255.
[26] M. Flipo, J. Charton, A. Hocine, S. Dassonneville, B. Deprez, R. Deprez-Poulain,
Journal of Medicinal Chemistry 52 (2009) 6790e6802.