Synthesis of a potent photoreactive acidic γ-secretase modulator for target identification in cells

Article abstract of DOI:10.1016/j.bmc.2012.08.034

Supramolecular self-assembly of amyloidogenic peptides is closely associated with numerous pathological conditions. For instance, Alzheimers disease (AD) is characterized by abundant amyloid plaques originating from the proteolytic cleavage of the amyloid precursor protein (APP) by β- and γ-secretases. Compounds named γ-secretase modulators (GSMs) can shift the substrate cleavage specificity of γ-secretase toward the production of non-amyloidogenic, shorter Aβ fragments. Herein, we describe the synthesis of highly potent acidic GSMs, equipped with a photoreactive diazirine moiety for photoaffinity labeling. The probes labeled the N-terminal fragment of presenilin (the catalytic subunit of γ-secretase), supporting a mode of action involving binding to γ-secretase. This fundamental step toward the elucidation of the molecular mechanism governing the GSM-induced shift in γ-secretase proteolytic specificity should pave the way for the development of improved drugs against AD.

Full text of DOI:10.1016/j.bmc.2012.08.034

Bioorganic & Medicinal Chemistry 20 (2012) 6523–6532
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis of a potent photoreactive acidic
for target identification in cells
c-secretase modulator
Andreas Rennhack a, Thorsten Jumpertz b, Julia Ness b, Sandra Baches b, Claus U. Pietrzik c,
a,
Sascha Weggen b, , Bruno Bulic
⇑
⇑
a Research Group Chemical Biology of Neurodegenerative Diseases, Center of Advanced European Studies and Research, Ludwig-Erhard-Allee 2, D-53175 Bonn, Germany
b Department of Neuropathology, Heinrich-Heine-University, D-40225 Duesseldorf, Germany
c Molecular Neurodegeneration Group, Institute of Pathobiochemistry, University Medical Center of the Johannes Gutenberg-University, D-55099 Mainz, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Supramolecular self-assembly of amyloidogenic peptides is closely associated with numerous pathological
Received 30 May 2012
Revised 23 July 2012
Accepted 8 August 2012
Available online 31 August 2012
´
conditions. For instance, Alzheimers disease (AD) is characterized by abundant amyloid plaques originat-
ing from the proteolytic cleavage of the amyloid precursor protein (APP) by b- and -secretases. Com-
pounds named -secretase modulators (GSMs) can shift the substrate cleavage specificity of -secretase
c
c
c
toward the production of non-amyloidogenic, shorter Ab fragments. Herein, we describe the synthesis of
highly potent acidic GSMs, equipped with a photoreactive diazirine moiety for photoaffinity labeling.
Keywords:
Alzheimer´s disease
Proteases
Secretase inhibitors
Secretase modulators
Photoreactive probes
Diazirine
The probes labeled the N-terminal fragment of presenilin (the catalytic subunit of
a mode of action involving binding to -secretase. This fundamental step toward the elucidation of the
molecular mechanism governing the GSM-induced shift in
the way for the development of improved drugs against AD.
c-secretase), supporting
c
c
-secretase proteolytic specificity should pave
Ó 2012 Elsevier Ltd. All rights reserved.
Target identification
Drug development
1. Introduction
lar amyloid plaques in the cerebral cortex and hippocampus.6–8
Amyloid plaques are essentially formed by the self-assembly of
the 42 amino acids-long Ab42 peptides, produced by cleavage of
APP at synaptic terminals by aspartyl proteases (Fig. 1).9,10 After
shedding of the extracellular domain of APP by the b-secretase,
the membrane-bound C99 fragment is subsequently processed
Over the last decade, regulated proteolysis has emerged as an
important cellular activity involved in various biological processes
such as cell differentiation, stress response or transcriptional regu-
lation. Two percent of the human genome encodes for proteolytic
enzymes, and deregulations of the ‘degradome’ have been linked
to various disorders, such as haemophilia, obesity and cancer.1,2
Notably, the familial hereditary form of Alzheimers disease
(FAD), which accounts for ca. 1–2% of cases of Alzheimers disease
(AD), is associated with mutations in the genes encoding for pre-
within its transmembrane domain by the
c-secretase complex.
Interestingly, the substrate is not cleaved at a single position;
heterogeneous fragments are secreted and range in lengths from
37 to 43 amino acids.11,12 Investigations of the proteolytic mecha-
nism providing a predominantly 40 amino acids-long peptide are
´
´
senilins, the catalytic subunits of
c
-secretase and the amyloid pre-
complicated by the multimeric nature of the c-secretase complex
cursor protein (APP).3–5
which includes, besides the presenilins 1 or 2 (PS1/2) N- and C-ter-
minal fragments forming the catalytic core, multiple accessory pro-
teins (APH-1, nicastrin, PEN-2).13,14 While the impact of the Ab42
peptide on cell viability has been extensively investigated in vitro
AD is a progressive neurodegenerative disorder characterized by
the presence of intracellular neurofibrillary tangles and extracellu-
and in vivo, the physiological role of the c-secretase and its multiple
Abbreviations: Ab, amyloid-b peptide; GSM,
-secretase inhibitor; APP, amyloid precursor protein; PS1/2, presenilin 1/2; PS1-
NTF, N-terminal fragment of PS1; AD, Alzheimer’s disease; FAD, familial Alzheimer’s
disease; APH-1, anterior pharynx defective-1; PEN-2, presenilin enhancer-2; NCT,
nicastrin; NSAID, non-steroidal anti-inflammatory drug.
c-secretase modulator; GSI,
other substrates remains unclear.9,10 Altogether, avoiding the
formation of amyloid plaques remains a promising therapeutic
approach for FAD and, plausibly, for sporadic AD.15,16 Unfortu-
nately, previous clinical trials operating on full inhibition of
c
⇑
Corresponding author. Tel.: +49 228 9656401.
E-mail address: bruno.bulic@caesar.de (B. Bulic).
c-secretase with the objective to ‘trim down’ the production of
0968-0896/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmc.2012.08.034

6524
A. Rennhack et al. / Bioorg. Med. Chem. 20 (2012) 6523–6532
Figure 1. Schematic representations of (a) sequential cleavage of APP substrate by b- and c-secretases leading to the formation of Ab40 and Ab42 peptides. (b) c-Secretase
complex with PS1-NTF (purple), APH-1 (blue), nicastrin (light green), PEN-2 (orange) and a substrate molecule (dark green).
the culprit amyloidogenic Ab42 peptides failed due to severe side
effects likely linked to broad inhibition of the constitutive functions
of
c
-secretase. Interestingly, compounds termed
c
-secretase mod-
-secretase
ulators (GSMs) preserve the constitutive activity of
c
and only shift the cleavage specificity toward less amyloidogenic
fragments such as Ab38.17 Clearly, unlike the quantitative effect
of c-secretase inhibitors, the qualitative action of GSMs may allow
the protease to keep on performing its physiological functions.
Since the early identification of GSMs in the class of non-steroi-
dal anti-inflammatory drugs (NSAIDs), intensive efforts in
medicinal chemistry to improve their low blood–brain barrier per-
meability and moderate potency have yielded second-generation
GSMs in two distinct classes (acidic and non-acidic GSMs), named
after a characteristic carboxylic acid.18–21 However, their molecular
mechanism of action remains largely unknown. Strong evidence
gained from in vitro investigations with partially purified protease
Figure 2. Left: Parent acidic phenylpiperidine-type GSM GSM-1. Right: Derived
photoprobe AR243 with photoreactive diazirine moiety and biotin-tag for affinity
purification.
using our biotinylated probe provided evidence for direct binding
support a direct action of the compounds on the c-secretase com-
to the c
-secretase complex at the PS1-NTF epitope.35 This is consis-
plex itself.22–24 Furthermore, cross-competition binding assays be-
tween radiolabeled inhibitors and NSAID GSMs indicated that
transition-state inhibitors were displaced in a non-competitive
manner, suggesting different binding sites and an allosteric mode
of action for GSMs.22,25 Moreover, fluorescence lifetime imaging
has been used to monitor conformational changes at the PS1 epi-
tent with recent data obtained from investigations in the class of
imidazole-based non-acidic GSMs showing binding to Pen2 and
PS1-NTF.36,37 Moreover, comparable photoaffinity labeling studies
with benzophenone- and azide-equipped acidic modulators, but
no diazirine-GSM, appeared in the recent literature.34,38
The 3-phenyl-3-(trifluoromethyl)-3H-diazirine (in the follow-
ing referred to as diazirine) photoreactive group was primary se-
lected for its interesting characteristics. That is low steric bulk,
high lipophilicity and long wavelength of activation (ꢀ350 nm).
Diazirines have been successfully used as photolabeling reagents
since their first synthesis by Brunner et al. in 1980 and their appli-
cation has been reviewed extensively.39–43 We hereafter describe
the design and synthesis of GSM-1 derived, diazirine-equipped
photoreactive probes as well as a preliminary structure–activity
relationship around this compound class.
tope of c-secretase, indicating a closure of a hydrolytic site upon
incubation with GSMs.26,27 In contrast to these findings, several re-
search teams reported the binding of NSAID GSMs to the APP sub-
strate rather than to the c
-secretase.28–30 However, the very low
potencies and specificities of the used GSMs prompted further
investigations questioning the validity of the proposed APP binding,
with strong evidence for self-aggregation and promiscuous bind-
ing.31–33
We therefore embarked on the synthesis of diazirine-modified
probes for photoaffinity labeling based on the highly potent sec-
ond-generation acidic modulator GSM-1 (Fig. 2). The design criteria
included: low steric bulk, high potency, stability of the photoreac-
tive moiety and photolysis wavelength compatible with experi-
ments in live cells. After optimization of a diazirine-compatible
synthetic route and elucidation of a preliminary structure–activity
relationship for the GSM-1 scaffold, we could successfully synthe-
size the AR243 probe equipped with a photoreactive diazirine moi-
ety and a biotin tag for affinity purification, along with a nanomolar
modulatory activity and high structural fidelity to the parent com-
pound (Fig. 2). However, the biotin tag, with its relative high polar-
ity and steric bulk, required the incorporation of an adequate spacer
and, in this respect, copper-catalyzed click chemistry presents an
attractive alternative.34 Nonetheless, photolabeling experiments
2. Results
2.1. Chemistry
2.1.1. Diazirine-compatible Cu(I)-catalyzed A3-coupling
The synthesis of the parent modulator GSM-1 is described in the
patent literature and involves the reaction of a-phenyl amines with
benzaldehydes under Dean–Stark conditions prior to the addition of
an organozinc reagent.44 However, the route proved impractical for
diazirine-modified GSM-1 derivatives because the diazirine group
appeared thermally labile. We therefore investigated metal-
catalyzed three-component coupling reactions with copper, iron,

A. Rennhack et al. / Bioorg. Med. Chem. 20 (2012) 6523–6532
6525
indium or gold replacing the organozinc reagent with a terminal
alkyne to widen the substrate scope and to provide milder reaction
conditions (A3-coupling, Scheme 1).45–48 From these catalysts only
copper(I) appeared to be suitable for our diazirine-based approach
and after several optimization rounds, we could obtain the desired
coupling products in satisfactory yield. In the successful setup three
equivalents of copper(I)-bromide with respect to the aldehyde were
used and the reactions were run at 90 °C for 24 h in toluene with a
2:1:1 stoichiometry with respect to amine, aldehyde and alkyne
in the presence of activated 4 Å-molecular sieves (General method
A; see Section 5). Lowering the reaction temperature to 60 °C and
increasing the reaction time to 48 h could further prevent thermal
decomposition of the diazirine and was therefore chosen for the
synthesis of photoreactive derivatives (General method B; see
Section 5). Increasing the stoichiometry of alkynes and amines
provided only minor improvement. The products of the three-
component couplings were subsequently purified and submitted
to further manipulations on side chains to provide the final com-
pounds (as described below and in the Supplementary data).
acid was necessary as we planned to link the biotin tag via an es-
ter bond for reasons of higher lipophilicity compared to the more
common primary amide-link (see also Section 3.1 and Scheme 1).
Required terminal alkynes were synthesized from commercial 5-
hexyn-1-ol by acylation with biotin or acetic anhydride,49 yield-
ing esters 6 and 7 (Scheme 2B), respectively. Amino-alkyne 9
was obtained via the free amine 8 starting from commercial
N-(5-hexynyl)phthalimide by treatment with hydrazine hydrate
and reprotection using Boc-anhydride (Scheme 2B). Non-commer-
cial benzaldehydes were synthesized as outlined in Scheme 3.
Briefly, 3-methoxy-4-(4-methyl-1H-imidazol-1-yl)benzaldehyde
10 was obtained as described in the patent literature from 4-
(5)-methylimidazole and 4-fluoro-3-methoxy-benzaldehyde.50
Following the protocol from Chen et al.,51 we obtained the
diazirine-equipped benzaldehyde 14 starting from commercial
1,4-dibromobenzene (Scheme 2). The so obtained intermediate
4-trifluoroacetylbenzoic acid 11 was esterified with methanol to
12, followed by oxime formation and tosylation to provide 13.
Subsequently, treatment with liquid ammonia in DCM afforded
the corresponding intermediate diaziridine. The sensitive oxida-
tion to obtain the diazirine was finally accomplished under Swern
conditions.52 Reduction of the methylester to the aldehyde using
DIBAL or milder reducing agents such as SDBBA unfortunately
resulted in decomposition of the diazirine.53 However, its
precursor, the trifluoromethyl diaziridine-methyl benzoate, was
found to be stable toward SDBBA-mediated reduction. Therefore,
undertaking the reduction step before the diaziridine-oxidation,
we could obtain the aldehyde 14 with a 40% overall yield.
2.1.2. Synthesis of building blocks
Amine 1 was obtained according to the patent literature in
racemic and enantiomeric pure forms, and served as a starting
material for the silyl-protected analog 5, which was synthesized
via the intermediates 2 and 3 (Scheme 2A).44 Esterification of 3
with 2-trimethylsilylethanol and subsequent hydrogenolytic
N-debenzylation of the intermediate 4 yielded the desired protected
ester 5. Orthogonal protection of the piperidine-bound carboxylic
Scheme 1. Synthesis of phenylpiperidine-type GSMs via A3-coupling and subsequent side chain modifications. R1–R3 as well as numbering of stereocenters refer to Table 1.
Reagents and conditions: (i) CuBr, activated 4 Å-mol sieves, PhMe, 90 °C, 24 h or 60 °C, 48 h; (ii) TBAF, THF, rt, 2 h; (iii) LiOH–H2O, THF/H2O/MeOH, rt, 2 h; (iv) Lindlar’s cat,
quinoline, H2, 1 atm, PhMe, rt, 7 h; (v) 4 M HCl, dioxane, rt, 2 h; (vi) biotin, EDAC, DIPEA, DMF, rt, 2 h (for detailed synthetic procedures and spectroscopic data the reader is
referred to the Supplementary data).

6526
A. Rennhack et al. / Bioorg. Med. Chem. 20 (2012) 6523–6532
establish a preliminary structure–activity relationship for straight
and inverse GSMs.
By reacting the enantiomeric pure forms of amine 1 with 4-chlo-
robenzaldehyde and 3-methylbut-1-yne followed by ester hydroly-
sis, we could obtain the enantiomers BB25 and BB23, which
minimally differed from the parent GSM-1 structure by an internal
alkyne required for this milder synthesis route (Scheme 1 andTable
1). CHO cells stably overexpressing wild-type APP and PS1 were
incubated with the compounds to determine their potency against
Aß42 and selectivities versus Aß40 (all) and Aß38 (only for selected
compounds) (Table 1 and Supplementary Figs. S1(A–F) and S2) by
sandwich ELISA of cell-culture supernatants.3 The ELISA data for
BB25 and BB23 revealed that the alkyne introduction was well tol-
erated, as reflected by activities comparable to the parent modula-
tor GSM-1 (IC50 = 68 4 nM (Ab42) Table 1, entry 1; BB25:
IC50 = 87 8 nM (Ab42); BB23: IC50 = 300 0 nM (Ab42); entries 2
and 3). In addition, BB25 showed the typical profile of an acidic
GSM, selectively lowering Aß42 while raising Aß38 and having no
significant impact on Aß40 levels (Supplementary Figs. S1A, B and
S2). Each enantiomer displayed a distinct potency, as often observed
for drugs targeting chiral domains within biomolecules. As reported
for NSAID-like acidic GSMs, esterification of the characteristic car-
boxylic acid of BB25 to give BB25-Me reversed the activity profile
of the compound, turning the modulation from selective Ab42 low-
ering (IC50) to Ab42 elevating (EC50), in a yet incompletely under-
stood manner (BB25-Me: EC50 = 2.5 0.3 lM (Ab42); entry 4 and
Supplementary Fig. S2). 19,54
Scheme 2. (A) Synthesis of silyl-protected amine 5. Reagents and conditions: (i)
BnBr, KI, CH3CN, rt, 15 min; (ii) LiOHÁH2O, THF/H2O/MeOH, rt, 2 h;44 (iii)
In an effort to come closer to the parent GSM-1 structure,
hydrogenation of the alkyne in BB23 and BB25 was probed under
various conditions, ending unfortunately in decomposition
through dehalogenation. However, we were able to isolate BB22
and BB24 after partial hydrogenation using Lindlar’s catalyst
(Scheme 1), bearing a cis-alkene in its side-chain. Interestingly,
the potency of the more active enantiomer BB24 (IC50 = 30 1 nM
(Ab42); entry 5 and Supplementary Fig. S2) improved over the
parent compound BB25 whereas the potency of the opposite enan-
tiomer BB22 further decreased (IC50 = 616 12 nM (Ab42); entry 6
and Supplementary Fig. S2). However, due to only moderate gain in
potency and restricted compatibility of hydrogenation reactions
with diazirines, we selected BB25 for further probe-development
as detailed below.
i
HO(CH2)2SiMe3, EDAC, DMAP, DCM, rt, 3 h; (iv) Pd/C, H2 1 atm, PrOH, rt, 24 h. (B)
Synthesis of terminal alkynes 6, 7 and 9: (v) biotin, TBTU, DIPEA, DMF, rt, 5 h; (vi)
Ac2O, Py/DCM, rt, 3 h;49 (vii) (H2N)2-H2O, THF, reflux, 3 h; (viii) Boc2O, NaHCO3,
THF/H2O, rt, 2 h (for detailed synthetic procedures and spectroscopic data the
reader is referred to the Supplementary data).
We next investigated the structure–activity relationship
around the lower chloro-aryl moiety as a potential introduction
site for the photoreactive diazirine. By substituting 4-chloro-
benzaldeyde
with
3-methoxy-4-(4-methyl-1H-imidazol-1-
yl)benzaldehyde 10 we obtained AR501, which shares the charac-
teristic imidazole commonly observed in the class of non-acidic
GSMs (Scheme 1 and Table 1).19,50 However, the disappointing
activity for AR501 (IC50 = 1.0 0.1 lM (Ab42); entry 7 and Sup-
plementary Fig. S2) indicated sensitivity to polarity effects at this
site. Interestingly, the corresponding inverse modulator AR499
(EC50 = 1.6 0.2
the methylester analogue of AR501, presented a comparable
activity to BB25-Me (2.5 0.3 M) in raising Aß42, while the
potencies of the respective straight modulators BB25 and
AR501 differed significantly (IC50 = 87 8 nM vs 1.0 0.1 M,
respectively), underlining the apparent disconnect in SAR be-
tween straight and inverse modulators. To further determine
the sensitivity of the site toward polarity effects, we substituted
the chlorine for a trifluoromethyl group using commercial 4-tri-
fluoromethyl benzaldehyde (Scheme 1 and Table 1). The resulting
enantiomers AR262 (IC50 = 50 4 nM (Ab42); entry 9 and Supple-
mentary Fig. S2) and AR261 (IC50 = 350 35 nM (Ab42); entry 10
and Supplementary Fig. S2) displayed similar potency as their
chlorine analogs BB25 and BB23 (87 8 nM and 300 0 nM,
respectively), emphasizing the tolerance of lipophilicity in this
lM (Ab42); entry 8 and Supplementary Fig. S2),
l
Scheme 3. Synthesis of 10 and diazirine-tagged benzaldehyde 14. Reagents and
conditions: (i) K2CO3, DMF, 50 °C, over night;50 (ii) 1. n-BuLi, CF3COOMe, THF,
À78 °C, 1 h/2. n-BuLi, CO2 (s), THF, À78 °C to rt, 2 h;51 (iii) H2SO4, MeOH, 60 °C, over
night; (iv) H2NOHÁHCl, MeOH/Py, reflux, over night; (v) TsCl, NEt3, DCM, 0 °C to rt,
1 h; (vi) NH3 (l), DCM, À78 °C to rt, over night; (vii) SDBBA, THF, 0 °C, 40 min;53 (viii)
l
(COCl)2, DMSO, NEt3, DCM, À78 °C to rt, 1 h.52
.
2.1.3. GSM synthesis and structure–activity relationship
With the optimized protocol for the copper(I)-catalyzed A3-
coupling and building blocks in hand, we were able to synthesize
a subset of 15 structurally divergent GSMs including the desired
photoprobes (Scheme 1 and Table 1). In addition we were able to

A. Rennhack et al. / Bioorg. Med. Chem. 20 (2012) 6523–6532
6527
Table 1
Structures and activities of acidic phenylpiperidine GSMs
b
b
b
Entry
1
Compound
Stereo-chemistrya
R1
R2
R3
IC50/EC50 SEM (Aß42)c
I: S
II: R
III: S
GSM-1
H
68 4 nM
Cl
Cl
Cl
Cl
Cl
I: S
II: R
III: S
2
3
4
5
BB25
H
H
87 8 nM
I: R
II: S
III: R
BB23
300 0 nM
Inverse 2.5 0.3
30 1 nM
I: S
II: R
III: S
BB25-Me
BB24
Me
H
lM
I: S
II: R
III: S
I: R
6
7
BB22
II: S
III: R
H
616 12 nM
Cl
O
N
AR501
rac
rac
H
1.0 0.1 lM
N
N
O
N
8
AR499
Me
Inverse 1.6 0.2 lM
I: S
9
AR262
AR261
II: R
III: S
H
H
50 4 nM
F3C
F3C
I: R
II: S
III: R
10
350 35 nM
11
12
13
AR80
rac
rac
rac
H
H
H
190 14 nM
F3
C
N
N
Biotin
Biotin
O
O
AR179
AR243
1.0 0.0 lM
F3C
500 83 nM
F3
C
N
N
N
N
N
Biotin
NH
14
15
16
AR366
AR367
AR354
rac
rac
rac
Me
H
Inverse 1.1 0.1
l
M
F3C
N
Biotin
NH
2.2 0.1
2.0 0.1
l
M
M
F3C
N
OH
H
l
F3C
N
For dose response curves and selectivity data of selected compounds the reader is referred to the Supplementary data.
a
Stereocenters.
Rn refer to Scheme 1.
b
c
IC50s (straight GSMs) and EC50s (inverse GSMs) were determined as published previously.3
position. Therefore, the 4-position of the lower aromatic ring was
chosen as an anchoring point for the diazirine moiety. Coupling of
4-(3-trifluoromethyl)-3H-diazirin-3-yl) benzaldehyde 14, racemic
amine 1 and 3-methyl-but-1-yne resulted in AR80 as expected
(Scheme 1 and Table 1). Given the modest potency gain of the
enantiomeric pure forms of the modulators, the synthesis was
carried out with racemic 1, providing the racemic photoreactive
compound
AR80
with
preserved
activity
(AR80
IC50 = 190 14 nM (Ab42); entry 11, Supplementary Figs. S1C
and S2). Having successfully placed the photoreactive diazirine
moiety, we next focused on the introduction of a biotin tag for
affinity purification. Since the three-component coupling reaction
proved tolerant to a large diversity of reactants, we attempted to
introduce the biotin tag in the aliphatic side-chain of AR262. To

6528
A. Rennhack et al. / Bioorg. Med. Chem. 20 (2012) 6523–6532
minimize perturbations induced by the bulky biotin tag, we
3. Discussion
selected
a
six-carbon-atoms long spacer for GSM AR179
(IC50 = 1.0
0 lM (Ab42); entry 12, Supplementary Figs. S1D and
3.1. Synthesis and SAR
S2). This analog was obtained by coupling silyl-protected amine
5, 4-trifluoromethyl-benzaldehyde, and biotin-functionalized ter-
minal alkyne 6, followed by selective TBAF-mediated deprotection
of the piperidine-bound carboxylic acid (Scheme 1 and Table 1).
The introduction of the lipophilic diazirine moiety in AR243
(IC50 = 500 83 nM (Ab42); entry 13) appeared to partially com-
pensate the negative impact on potency of the biotin tag in
AR179. We could therefore successfully obtain the complete
photoreactive probe AR243 (Scheme 1 and Table 1) with a puri-
fication tag, nanomolar activity, typical selectivity for an acidic
GSM (Supplementary Figs. S1E and S2) and high scaffold similar-
ity compared with the parent compound GSM-1. In an effort to
further optimize the probe for experimentations in live cells, we
aimed for the stabilization against cellular hydrolases of the
side-chain supporting the biotin tag in AR243. Therefore, the es-
ter bond was replaced by an amide bond in AR367 using the N-
Boc-protected terminal alkyne 9 in conjunction with racemic 1
and 14 in the A3-coupling. HCl-treatment of the resulting cou-
pling product released the free amine that was subsequently cou-
pled to biotin, yielding the methylester AR366 as a potent inverse
Here we report the design and synthesis of GSM-1 derived
diazirine-tagged
c-secretase modulators to elucidate their bind-
ing partner in a cell culture model. We developed an optimized
protocol for the copper-(I)-catalyzed A3-coupling, which was
sufficiently mild to provide compatibility with heat sensitive
diazirines. During the development of our frontrunner probe
AR243 we were able to obtain a small library of straight and in-
verse GSMs that facilitated the establishment of a preliminary
structure–activity relationship. For straight GSMs we found that
increased lipophilicity of the lower aromatic ring and the aliphatic
side chain correlated well with increasing potency as expected for
substances targeting membrane-embedded proteins. This is nicely
illustrated by the disruptive effect of the polar biotin moiety on
potency, partially reversed by introduction of the lipophilic diazi-
rine moiety (see AR179 (IC50 = 1.0 0.0 lM) and AR243
(500 83 nM)). Additionally the presence of hydrogen-bond
donors (Amide-NH: AR367; OH: AR354) in the linker region also
led to a significant loss of potency (2.2 0.1 lM and 2.0 0.1 lM,
respectively). Esterfication of the characteristic carboxylic acid
moiety mediated the switch from straight to inverse modulation.
Among the small subset of three inverse modulators, all showed
comparable potencies on selectively raising Aß42 (Table 1; Entries
5, 7 and 13). Compared to the potencies of their respective straight
GSMs (Table 1; Entries 1, 6 and 14) one might suggest a clear
disconnect between the SAR of the inverse and straight GSMs.
While the principles governing the switch between straight and
inverse modulation are incompletely understood, shared binding
sites are proposed based on photolabeling experiments, competi-
tion assays and the very high scaffold similarity.
photoprobe (EC50 = 1.1 0.1 lM (Ab42); entry 14, Supplementary
Figs. S1F and S2). Saponification then gave the amide analog to
AR243, named AR367 (Scheme 1 and Table 1). Surprisingly, the
amide modification induced
a significant loss of activity
(AR367: IC50 = 2.2 0.1 M (Ab42); entry 15 and Supplementary
l
Fig. S2), pointing to a high sensitivity to the polarity of the linker,
as also previously reported for other photoreactive probes, also
highlighting the advantage of tag-free strategies.34,55 Noteworthy,
the inverse activity of AR366 underlines once more the discon-
nect between straight and inverse modulatory potencies as also
noted previously for AR499. To eventually verify that the hydro-
lysis product of AR243 does not contribute to the activity ob-
served in cells, we synthesized the hydrolysis product AR354
using the alkyne-acetate 7 followed by hydrolysis to obtain the
3.2. Mode of action
The straight and inverse photoprobes AR243 and AR366,
respectively, allowed us to successfully carry out photolabeling
experiments followed by biotin-affinity purification. Selective
binding of both probes to the PS1-NTF was found, suggesting that
both effects, straight and inverse modulation, are mediated via the
free alcohol (Scheme
1
and Table 1). The low potency for
AR354 (IC50 = 2.0 0.1
lM (Ab42); entry 16 and Supplementary
Fig. S2) confirmed the resilience of AR243 towards ester hydroly-
sis and its suitability for labeling experiments in membrane prep-
arations and in live cells.
same c
-secretase subunit.35 This finding is in accordance with sev-
eral independent studies and biochemical investigations.22–24
Indeed, two major compound classes have been identified to
date, acidic and non-acidic imidazole based GSMs, raising ques-
tions about their respective targets.19 We and others have reported
cross-competition binding experiments between acidic and non-
acidic GSMs, thereby providing insights into the binding sites for
these molecules.35,38 Competition between the photoreactive
probe AR243 and the acidic NSAID-like JNJ-40418677 (Fig. 3) indi-
cated a strong overlap in binding sites, as expected for homologous
GSMs.35 However, competitions experiments with non-acidic
GSMs have provided more nuanced results, with a significant dis-
connect between competitiveness and chemical analogy. We ob-
served a strong labeling competition between our acidic probe
AR243 and the imidazole-based non-acidic GSM compound 1
(Fig. 3), whereas structurally analogous and equally potent
E-2012 competed only partially. A similar disconnect has been
independently reported in vice versa competition experiments
with acidic GSMs against the non-acidic probe GSM photoprobe
RO-57-BpB.37 Therefore, it might be tempting to postulate distinct
binding sites for acidic and non-acidic GSMs. However, conclusions
based solely on these experiments should be avoided because of
possible allosteric effects and artifacts linked to compound self-
aggregation at the investigated concentration. Obviously, further
2.2. Photolabeling
With the straight-GSM photoprobe AR243 and the inverse
probe AR366 in hand, we performed the photolabeling experi-
ments followed by purification on streptavidin beads, as de-
scribed in a separate publication.35 Briefly, we demonstrated
that both the straight probe AR243 and the inverse probe
AR366 bound in membrane preparations and in living cells to
the N-terminal fragment of presenilin 1 (PS1), the catalytic sub-
unit of the
studies.34,38 For this, we treated N2a-ANPP cells that stably over-
express APP and all four subunits of the -secretase complex with
0.2 M of the photo-probe AR243 or DMSO vehicle and irradiated
c-secretase complex, in line with several independent
c
l
the live cells with 350 nm UV light. Subsequently, cellular mem-
branes were prepared and proteins labeled with the photo-probe
were affinity purified using streptavidin beads. The purified mate-
rial was resolved on SDS gels and probed by Western blotting
with antibodies against the four subunits of the c-secretase com-
plex or the substrate APP. Both probes selectively labeled the N-
terminal Fragment of PS1, indicating that the effect of straight
and inverse GSMs is mediated via direct binding to this
tase subunit.35
c-secre-

A. Rennhack et al. / Bioorg. Med. Chem. 20 (2012) 6523–6532
6529
Finally, the specificity of the initial
e
Àcleavage at the Ab48 or
Ab49 positions might dictate the downstream outcome of the step-
wise cleavages leading to Ab42 or Ab40 respectively, and a link be-
tween Ab48 and Ab42 has been reported in several PS1 mutants.
Hence, it has been also proposed that the Ab42 lowering effect of
GSMs might be related to a shift in the initial
However, recent experimental data show that the initial e-cleavage
e
Àcleavage site.68
site might not stay permanently upstream of Ab42 and therefore
might not be truly targeted by GSMs.69,70 Altogether, photolabeling
experiments with GSMs have provided extremely valuable infor-
mation indicating direct binding to
c-secretase and an allosteric
mode of action, which inspires further investigations to decipher
the fascinating molecular mechanism of c-secretase.
4. Conclusions
Thephotoreactiveprobe AR243is a novel
c
Àsecretasemodulator
based on GSM-1 with preserved potency. We also reported a preli-
minary structure–activity relationship pointing at preferential lipo-
philic substituents and relative tolerance to steric bulk. We could
demonstrate that the biotin-tagged probe AR243 bound in mem-
brane preparation and in living cells to the N-terminal fragment of
Figure 3. Compounds used for competition experiments. Compound 1, E-2012:
Non-acidic GSMs. JNJ-40418677: Acidic GSM. Note the structural similarity
between Compound 1 and E-2012.
PS1 (the catalytic subunit of the c-secretase complex). Moreover,
studies using mass spectrometry-based proteomic analyses will be
required to accurately map the respective binding sites.
we could obtain an inverse modulator AR366 with micromolar
potency, which labeled the same N-terminal fragment of PS1, pro-
viding an interesting tool for further investigations. Cross-competi-
In contrast, cross-competition experiments with GSMs and
c-secretase inhibitors (GSIs) strongly suggested distinct binding
tion experiments with other
c
Àsecretase modulators point toward
sites. Early cross-competition binding assays between a radiola-
beled transition-state inhibitor and the GSM sulindac sulfide
reported displacement in a non-competitive manner.25 The finding
was verified by a recent report showing unperturbed labeling of
PS1-NTF by photoreactive acidic GSMs in competition with GSIs
directed against the enzymatic active site, substrate-binding site
or an allosteric site.38 Moreover, this simultaneous binding of mod-
ulators and inhibitors allowed Li et al. to elegantly investigate the
GSM-induced conformational changes in PS1 using photoreactive
inhibitors directed towards several catalytic site subsites,
shared binding sites within PS1-NTF, which is in accordance to
independent investigations by other groups.34,35,37,38 However,
while selective modulation of Ab42 production is a promising
´
strategy for drug development for Alzheimers disease, further inves-
tigations are required to better understand the elusive link between
GSM-induced PS1 conformational changes and the shifted proteo-
lytic cleavage specificity of c-secretase.
5. Experimental section
5.1. Chemistry
demonstrating that GSMs allosterically modulate
c-secretase
activity.34 The proposed hydrophilic cavity in PS-1 around TMDs
6 and 7 forming the active site,56–60 indeed lowers the probability
to target this region with GSM-1 analogs, due to the high lipophil-
icity of a scaffold with a cLogP of 7.92.
All reactions were performed under an inert atmosphere of ar-
gon or nitrogen, unless otherwise stated. 1H NMR and 13C NMR
data were obtained on a Bruker Avance DPX 400 or DPX 500. Spec-
tra were recorded at 298 K and the chemical shifts are expressed in
ppm relative to solvent signals (1H NMR: CDCl3 7.26; MeOD 3.31;
DMSO-d6 2.50; CO(CD3)2 2.05; 13C NMR: CDCl3 77.16; MeOD 49.00;
DMSO-d6 39.52; CO(CD3)2 29.84). Compound purity was accessed
by analytical HPLC (Shimadzu LC-10 HPLC-system; 4 Â 150 mm
Despite the valuable information gained from photoaffinity
labeling, the mode of action of GSMs remains obscure. Whereas
the full inhibition of
c-secretase is well understood, the mode of
action of modulators appears difficult to rationalize, and a high-
resolution structure of the tetrameric
c-secretase complex is
eagerly awaited.58 The current model postulates a sequential
cleavage of the C99 substrate each 3–4 amino acids starting at a
M&N Nucleodur Isis 5 lm column) using a gradient system
(t = 0–2 min: 80% H2O + 0.1% TFA; 20% acetonitrile; ramp to 98%
acetonitrile (t = 15 min; 6%/min); hold for 2 min; Injection at
t = 1.5 min) and detection at 220 nm. Preparative HPLC was per-
formed on a Shimadzu LC-8A system using different gradients
(H2O + 0.1% TFA and acetonitrile) at 5 ml/min on a 10 Â 250 mm
Supelco Ascentis C18 column. HRMS-spectra were recorded on a
MAT 95 XL or MAT 90 mass spectrometer from Thermo Finnigan.
Specific rotations were measured on a Autopol I from Rudolph Re-
search Analytical (Hackettstown, NY, USA) in EtOH at stated con-
centrations. All reagents were obtained from Sigma–Aldrich
(Munich, Germany), Alfa-Aesar (Karlsruhe, Germany), Acros
Organics (Geel, Belgium), ABCR (Karlsruhe, Germany) or Fluoro-
chem (Hadfield, UK) and were used as received.
e
–cleavage site near the cytoplasmic border, compatible with an
alpha-helical conformation for the substrate.61–65 Considering that
the signature of GSMs comprises selective lowering of the patho-
genic Ab42 and increase of the shorter Ab38 peptide, it is assumed
that Ab42 and Ab38 might be interdependent, and that GSMs
might promote a further cleavage step on Ab42 to yield Ab38 after
a
tetrapeptide release. However, a loose correlation in the
Ab40:Ab38 ratio in cells expressing different FAD-associated PS1
mutants upon GSM treatment points to independent production
lines for Ab42 and Ab38.11,66 For instance, PS1-mutant cell lines,
´
typically resistant to GSMs Ab42 lowering effect, nevertheless
appear sensitive to the Ab38 elevating effect. Indeed, at least two
separate product lines have been recently confirmed using high-
resolution tandem mass spectrometry, a minor and pathogenic
Ab48-Ab45-Ab42 and a major Ab49-Ab46-Ab43-Ab40.12,67 How-
ever, treatment with sulindac sulfide, a first generation GSM with
low potency, did not allow establishing clearly the source of Ab38.
5.1.1. General methods A and B for A3-coupling
Method A (non-diazirine modulators): Inside a glovebox under
N2-atmosphere, a pressure tight vial was charged with aldehyde

6530
A. Rennhack et al. / Bioorg. Med. Chem. 20 (2012) 6523–6532
(1 equiv), amine (2 equiv) and the respective terminal alkyne
(2 equiv). CuBr (3 equiv) was added, followed by activated 4 Å
molecular sieves (ꢀ1 g per 10 ml solvent). Toluene (7 ml/mmol
aldehyde) was added and the reaction mixture heated to 90 °C un-
der vigorous stirring for 24 h. The crude reaction mixture was fil-
tered twice through Celite and a silica pad eluting with EtOAc
(for biotin containing compounds 5% MeOH in DCM was used).
After evaporation of the organic solvents, the crude extracts were
purified by silica-gel chromatography or preparative HPLC as de-
scribed below and in the Supplementary data. Method B (diazirine
modulators): An equivalent setup as described above was used
reducing the reaction temperature to 60 °C and prolonging the
reaction time to 48 h.
cooled to À78 °C and ammonia (ꢀ2 ml) was condensed into the
solution. The vial was sealed and removed from the cooling after
stirring for 30 min. While warming to rt stirring was continued
over night. The excess ammonia was carefully evaporated and
the formed precipitate filtered off. Evaporation of the filtrate gave
the crude diaziridine that was dissolved in dry THF (60 ml) and
cooled to 0 °C. A freshly prepared 0 °C solution of SDBBA (0.5 M
in THF; 84 ml; 42 mmol) was added over 20 min and the mixture
was stirred for another 20 min and then quenched by the slow
addition of water (5 ml).53 The solvent was evaporated and the oily
residue partitioned between EtOAc and water. Layers were sepa-
rated and the aqueous layer was washed several times with fresh
portions of EtOAc. Combined organic layers are washed with brine,
dried over MgSO4 and evaporated to yield a yellowish oil that was
directly used in the next step. To a À78 °C cold solution of oxalyl-
chloride (3.58 ml, 41.7 mmol) in dry DCM (20 ml) a solution of dry
DMSO (3.11 ml, 43.8 mmol) in dry DCM (20 ml) was slowly added
over 20 min. The resulting solution was stirred for another 10 min
and the crude from above was added over 20 min as a solution in
dry DCM (40 ml). Stirring was continued for 15 min and the reac-
tion was quenched by the slow addition of a solution of triethyl-
amine (5.78 ml, 41.7 mmol) in dry DCM (10 ml). Upon complete
addition the cooling was removed, allowing the reaction mixture
to warm to rt and stirred for an additional hour.52 Water was
added and the layers formed are separated. The aqueous layer
was washed with a fresh portion of DCM, organic layers were com-
bined, washed with brine, dried, filtered and evaporated to obtain a
yellow oil. Purification was performed on silica using pentane/Et2O
(9/1) and afforded 14 as a pale yellow liquid (1.824 g, 40%). Analyt-
ical data were in accordance to literature.72
5.1.2. Synthesis of diazirine-tagged benzaldehyde 14
5.1.2.1. Methyl 4-(2,2,2-trifluoroacetyl)benzoate 12.
4-(2,2,2-
trifluoroacetyl)benzoic acid 11 (5 g, 22.92 mmol) was obtained
according to Chen et al.,51 dissolved in methanol (100 ml) and concd
sulphuric acid (1.22 ml, 22.92 mmol) was added via syringe. The
mixture was heated to 60 °C over night. It was neutralized with 1 M
NaOH aq and concentrated. The obtained residue was partitioned
between sat. NaHCO3 aq and EtOAc. The organic layer was washed
with brine, dried over MgSO4 and evaporated to yield 12 as off-white
solid that was used as obtained (5.20 g, 98%). Analytical data were
in accordance to literature.71
1H NMR (400 MHz, (CD3)2CO d = 3.96 (s, 3H), 8.21 (d, J = 8.5 Hz,
2H), 8.25 (d, J = 8.5 Hz, 2H).
13C NMR (100 MHz, (CD3)2CO d = 53.04, 117.49 (q, J = 290.7 Hz),
130.92 (m), 130.94, 133.94, 137.04, 166.07, 180.82 (q, J = 35.2 Hz).
HRMS (EI; 70 eV) Calcd for C10H7F3O3 (M+) 232.0347. Found
232.0348.
1H NMR (400 MHz, CDCl3) d = 7.35 (d, J = 8.4 Hz, 2H), 7.92 (d,
J = 8.4 Hz, 2H), 10.04 (s, 1H).
Purity (HPLC) = 95%; tR = 12.54 min (Hydrate smears from
9.50 min to 12.54 min).
13C NMR (100 MHz, CDCl3) d = 28.61 (q, J = 40.6 Hz), 121.95 (q,
J = 275.3 Hz), 127.13, 129.99, 135.40, 136.95.0, 191.20.
5.1.2.2. Methyl 4-(2,2,2-trifluoro-1-((tosyloxy)imino)ethyl)ben-
Purity (HPLC) = 98%; tR = 13.66 min.
zoate 13.
Compound 12 (5.20 g, 22.40 mmol) and hydroxyl-
amine hydrochloride (3.84 g, 56 mmol) were dissolved in a 1:2
mixture (37 ml) of methanol and pyridine and refluxed over night.
Upon complete conversion the solvents were evaporated and the
residue was partitioned between 5% HCl aq and DCM. After resid-
ual pyridine was washed out, the combined aqueous layers were
extracted with DCM and organic layers were combined. It was
washed with brine, dried over MgSO4, filtered and evaporated.
The residue was dissolved in DCM (40 ml) and triethylamine
(7.75 ml, 56 mmol) was added. The mixture was cooled to 0 °C
and tosylchloride (4.83 g, 25.3 mmol) was added in small portions
over 10 min. The cooling bath was removed and the mixture was
allowed to warm to rt during 2 h. It was poured onto 10% HCl aq
and the layers were separated. The aqueous layer was washed with
DCM, organic layers were combined, washed with brine, dried over
MgSO4, filtered and evaporated. The residue was purified on silica
eluting with DCM to obtain 13 (isomeric mixture 2:3) as a white
solid (8.63 g, 96%).
5.1.3. Synthesis of AR243
5.1.3.1. 2-(1-(7-((5-((3aS,4S,6aR)-2-Oxohexahydro-1H-thieno [3,4-d]
imidazol-4-yl)pentanoyl)oxy)-1-(4-(3-(trifluoromethyl)-3H-diazirin-
3-yl)phenyl)hept-2-yn-1-yl)-2-(4-(trifluoromethyl) phenyl)piperidin-
4-yl)acetic acid AR243.
3H-diazirin-3-yl)benzaldehyde 14 (0,52 mmol), 2-trimethylsilyl)ethyl
2-(2-(4-(trifluoromethyl)phenyl)piperidin-4-yl)acetate (390 mg,
1.03 mmol), Hex-5-yn-1-yl 5-((3aS,4S,6aR)-2-oxohexahydro-1H-thieno
[3,4-d]imidazol-4-yl)pentanoate 6 (335 mg, 1.03 mmol) and copper(I)
bromide (222 mg, 1.55 mmol) were reacted in toluene (3.5 ml)
according to general method B. The so obtained crude oil was dissolved
in THF (15 ml) and TBAF (1 M in THF, 1.12 ml, 1.12 mmol) was slowly
added at rt. The reaction was allowed to proceed for 2 h before dilution
with EtOAc and water. The pH was set neutral and the final compound
extracted with EtOAc. The organic layers were combined, washed
with brine, dried over MgSO4, filtered and evaporated. The obtained crude
oil was purified on preparative HPLC to afford AR243 as a white
foam (42 mg, 9.3%).
About 110 mg of 4-(3-(trifluoromethyl)-
5
1H NMR (400 MHz, CDCl3) d = (major isomer) 2.47 (s, 3H), 3.94
(s, 3H), 7.38 (m, 2H), 7.45 (d, J = 8.4 Hz), 7.87 (d, J = 8.1 Hz), 8.13
(d, J = 8.4 Hz).
1H NMR (400 MHz, MeOD) d = 1.39–1.91 (m, 12H), 1.99 (m, 1H),
2.12 (m, 1H), 2.21 (m, 1H), 2.31 (d, J = 6.8 Hz, 2H), 2.36 (t, J = 7.4 Hz,
2H), 2.55 (td, J1 = 6.8 Hz, J2 = 2.1 Hz, 2H), 2.68 (d, J = 12.9 Hz, 1H),
2.89 (dd, J1 = 12.9 Hz, J2 = 4.9 Hz, 1H), 2.98 (m, 1H), 3.10–3.20 (m,
2H), 4.19 (t, J = 6.5 Hz), 4.28 (dd, J1 = 8.1, J2 = 4.5 Hz, 1H), 4.38–
4.50 (m, 2H), 4.90 (s, 1H), 7.32 (d, J = 8.1 Hz, 2H), 7.61 (d,
J = 8.1 Hz, 2H), 7.76–7.89 (m, 4H).
13C NMR (100 MHz, CDCl3) d = (major isomer) 21.88, 52.63,
119.5 (q, J = 277.5 Hz), 127.13, 128.58, 129.37, 129.90, 130.08,
131.07, 133.05, 146.50, 153.49 (q, J = 33.3 Hz), 165.87.
HRMS (ESI-TOF) Calcd for C17H14F3NO5SNa (M+Na) 424.0442.
Found 424.0440.
Purity (HPLC) = 98%; tR = 15.41 min.
13C NMR (100 MHz, MeOD) d = 19.14, 25.98, 25.99, 29.17, 29.47,
29.71, 30.58, 33.31, 34.84, 40.30, 40.71, 41.02, 48.31, 56.97, 59.34,
61.62, 63.39, 65.03, 68.43, 71.58. 94.82, 123.45 (q, J = 275.2 Hz),
125.39 (q, J = 275.5 Hz), 127.64 (q, J = 3.3 Hz), 127.95, 129.92
5.1.2.3. 4-(3-(Trifluoromethyl)-3H-diazirin-3-yl)benzaldehyde
14.
Compound 13 (8.63 g, 21.50 mmol) was dissolved in
DCM (11 ml) and transferred into a pressure tight vial. It was

A. Rennhack et al. / Bioorg. Med. Chem. 20 (2012) 6523–6532
6531
2. Lopez-Otin, C.; Bond, J. S. J. Biol. Chem. 2008, 283, 30433.
(broad), 131.43, 131.59, 132.49 (q, J = 32.5 Hz), 136.60, 143.46,
166.09, 175.40, 175.52 (diazirine-carbon not resolved).
HRMS (ESI-TOF) for C39H43F6N5O5SH+ (M+H) 808.2967. Found
808.2938.
3. Hahn, S.; Bruning, T.; Ness, J.; Czirr, E.; Baches, S.; Gijsen, H.; Korth, C.; Pietrzik,
C. U.; Bulic, B.; Weggen, S. J. Neurochemistry 2011, 116, 385.
4. Kretner, B.; Fukumori, A.; Gutsmiedl, A.; Page, R. M.; Luebbers, T.; Galley, G.;
Baumann, K.; Haass, C.; Steiner, H. J. Biol. Chem. 2011, 286, 15240.
5. Scheuner, D.; Eckman, C.; Jensen, M.; Song, X.; Citron, M.; Suzuki, N.; Bird, T. D.;
Hardy, J.; Hutton, M.; Kukull, W.; Larson, E.; Levy-Lahad, E.; Viitanen, M.;
Peskind, E.; Poorkaj, P.; Schellenberg, G.; Tanzi, R.; Wasco, W.; Lannfelt, L.;
Selkoe, D.; Younkin, S. Nat. Med. 1996, 2, 864.
6. Thal, D. R.; Schultz, C.; Dehghani, F.; Yamaguchi, H.; Braak, H.; Braak, E. Acta
Neuropathol. 2000, 100, 608.
7. Braak, H.; Braak, E.; Bohl, J.; Reintjes, R. Neurosci. Lett. 1996, 210, 87.
8. Bulic, B.; Pickhardt, M.; Mandelkow, E. M.; Mandelkow, E. Neuropharmacology
2010, 59, 276.
Purity (HPLC) = 97%; tR = 13.87 min.
5.1.4. Synthesis of AR366
5.1.4.1. Methyl 2-(1-(7-(5-((3aS,4S,6aR)-2-oxohexahydro-1H-thieno
[3,4-d]imidazol-4-yl)pentanamido)-1-(4-(3-(trifluoromethyl)-3H-dia
zirin-3-yl)phenyl)hept-2-yn-1-yl)-2-(4-(trifluoromethyl) phenyl) pipe
ridin-4-yl)acetate AR366.
Compound 14 (236 mg, 1.10 mmol), 1
9. Guo, Q.; Wang, Z.; Li, H.; Wiese, M.; Zheng, H. Cell Res. 2012, 22, 78.
10. Zheng, H.; Koo, E. Mol. Neurodegener. 2011, 6, 27.
(663 mg, 2.20 mmol), 9 (434 mg, 2.20 mmol) and copper(I) bromide
(473 mg, 3.3 mmol) were reacted according to general method B.
The crude oil was then treated with 4 M HCl in dioxane (1.38 ml,
5.5 mmol) for 2 h at rt. The solvent was removed in vacuo and the
11. Czirr, E.; Cottrell, B. A.; Leuchtenberger, S.; Kukar, T.; Ladd, T. B.; Esselmann, H.;
Paul, S.; Schubenel, R.; Torpey, J. W.; Pietrzik, C. U.; Golde, T. E.; Wiltfang, J.;
Baumann, K.; Koo, E. H.; Weggen, S. J. Biol. Chem. 2008, 283, 17049.
12. Takami, M.; Nagashima, Y.; Sano, Y.; Ishihara, S.; Morishima-Kawashima, M.;
Funamoto, S.; Ihara, Y. J. Neurosci. 2009, 29, 13042.
residue was dissolved in dry DMF (2.2 ml). DIPEA (575 ll, 3.3 mmol)
13. Wolfe, M. S. Semin. Cell Dev. Biol. 2009, 20, 219.
was added, followed by biotin (403 mg, 1.65 mmol). The mixture was
stirred for 5 min at rt and EDAC (422 mg, 2.2 mmol) was added. The
reaction is left for 2 h and then diluted with EtOAc. Thorough
extraction with water removed residual DMF; it was washed with brine,
dried, filtered and evaporated. The resulting crude material was puri-
fied by prep. HPLC to afford AR366 as off-white foam (226 mg, 25%).
1H NMR (400 MHz, MeOD) d = 1.34–1.47 (m, 3H), 1.52–1.77 (m,
10H), 1.82 (m, 1H), 1.92 (m, 1H), 2.16–2.26 (m, 4H), 2.35–2.47 (m,
4H), 2.66 (d, J = 12.9 Hz, 1H), 2.85 (dt, J1 = 12.9, J2 = 5.1 Hz, 1H), 3.14
(m, 1H), 3.26 (m, 2H), 3.61 (s, 3H), 3.67 (dd, J1 = 11.3 Hz, J2 = 8.3 Hz,
1H), 4.25 (dt, J1 = 7.8 Hz, J2 = 4.8 Hz, 1H), 4.39 (s, 1H), 4.44 (m, 1H),
7.18 (d, J = 8.1 Hz, 2H), 7.62 (d, J = 8 Hz, 2H), 7.64–7.70 (m, 4H).
13C NMR (100 MHz, MeOD) d = 19.06, 26.93, 27.54, 29.30. (q,
J = 40.2 Hz), 29.47, 29.77, 29.92, 32.81, 34.91, 36.83, 39.84, 41.02,
41.56, 43.22, 47.00, 51.99, 56.97, 57.35, 61.58, 63.35, 66.72,
74.64, 90.75, 122.60 (q, J = 273.2 Hz), 125.67 (q, J = 271.1 Hz),
126.76 (q, J = 3.6 Hz), 127.28 129.02, 129.37 (broad), 129.84,
130.80 (q, J = 32.0 Hz), 143.01, 149.65, 166.04, 174.61, 175.96.
HRMS (ESI-TOF) Calcd for C40H46F6N6O4S (M+Na) 843.3103.
Found 843.3069.
14. Bergmans, B. A.; De Strooper, B. Lancet Neurol. 2010, 9, 215.
15. Karran, E.; Mercken, M.; De Strooper, B. Nat. Rev. Drug Discov. 2011, 10, 698.
16. Hardy, J. J. Neurochem. 2009, 110, 1129.
17. Weggen, S.; Eriksen, J. L.; Das, P.; Sagi, S. A.; Wang, R.; Pietrzik, C. U.; Findlay, K.
A.; Smith, T. E.; Murphy, M. P.; Bulter, T.; Kang, D. E.; Marquez-Sterling, N.;
Golde, T. E.; Koo, E. H. Nature 2001, 414, 212.
18. Weggen, S.; Rogers, M.; Eriksen, J. Trends Pharmacol. Sci. 2007, 28, 536.
19. Bulic, B.; Ness, J.; Hahn, S.; Rennhack, A.; Jumpertz, T.; Weggen, S. Curr.
Neuropharmacol. 2011, 9, 598.
20. Oehlrich, D.; Berthelot, D. J. C.; Gijsen, H. J. M. J. Med. Chem. 2011, 54, 669.
21. Wolfe, M. S. J. Neurochem. 2012, 120, 89.
22. Beher, D.; Clarke, E. E.; Wrigley, J. D.; Martin, A. C.; Nadin, A.; Churcher, I.;
Shearman, M. S. J. Biol. Chem. 2004, 279, 43419.
23. Takahashi, Y.; Hayashi, I.; Tominari, Y.; Rikimaru, K.; Morohashi, Y.; Kan, T.;
Natsugari, H.; Fukuyama, T.; Tomita, T.; Iwatsubo, T. J. Biol. Chem. 2003, 278,
18664.
24. Weggen, S.; Eriksen, J. L.; Sagi, S. A.; Pietrzik, C. U.; Ozols, V.; Fauq, A.; Golde, T.
E.; Koo, E. H. J. Biol. Chem. 2003, 278, 31831.
25. Clarke, E. E.; Churcher, I.; Ellis, S.; Wrigley, J. D.; Lewis, H. D.; Harrison, T.;
Shearman, M. S.; Beher, D. J. Biol. Chem. 2006, 281, 31279.
26. Lleo, A.; Berezovska, O.; Herl, L.; Raju, S.; Deng, A.; Bacskai, B. J.; Frosch, M. P.;
Irizarry, M.; Hyman, B. T. Nat. Med. 2004, 1065, 10.
27. Uemura, K.; Lill, C. M.; Li, X.; Peters, J. A.; Ivanov, A.; Fan, Z.; Destrooper, B.;
Bacskai, B. J.; Hyman, B. T.; Berezovska, O. PLoS ONE 2009, 4, e7893.
28. Kukar, T. L.; Ladd, T. B.; Bann, M. A.; Fraering, P. C.; Narlawar, R.; Maharvi, G.
M.; Healy, B.; Chapman, R.; Welzel, A. T.; Price, R. W.; Moore, B.; Rangachari, V.;
Cusack, B.; Eriksen, J.; Jansen-West, K.; Verbeeck, C.; Yager, D.; Eckman, C.; Ye,
W.; Sagi, S.; Cottrell, B. A.; Torpey, J.; Rosenberry, T. L.; Fauq, A.; Wolfe, M. S.;
Schmidt, B.; Walsh, D. M.; Koo, E. H.; Golde, T. E. Nature 2008, 453, 925.
29. Richter, L.; Munter, L. M.; Ness, J.; Hildebrand, P. W.; Dasari, M.;
Unterreitmeier, S.; Bulic, B.; Beyermann, M.; Gust, R.; Reif, B.; Weggen, S.;
Langosch, D.; Multhaup, G. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 14597.
30. Botev, A.; Munter, L. M.; Wenzel, R.; Richter, L.; Althoff, V.; Ismer, J.; Gerling, U.;
Weise, C.; Koksch, B.; Hildebrand, P. W.; Bittl, R.; Multhaup, G. Biochemistry
2011, 50, 828.
31. Beel, A. J.; Barrett, P.; Schnier, P. D.; Hitchcock, S. A.; Bagal, D.; Sanders, C. R.;
Jordan, J. B. Biochemistry 2009, 48, 11837.
32. Barrett, P. J.; Sanders, C. R.; Kaufman, S. A.; Michelsen, K.; Jordan, J. B.
Biochemistry 2011, 50, 10328.
33. Leach, A. R.; Hann, M. M. Curr. Opin. Chem. Biol. 2011, 15, 489.
34. Crump, C. J.; Fish, B. A.; Castro, S. V.; Chau, D. M.; Gertsik, N.; Ahn, K.; Stiff, C.;
Pozdnyakov, N.; Bales, K. R.; Johnson, D. S.; Li, Y. M. ACS Chem. Neurosci. 2011, 2,
705.
35. Jumpertz, T.; Rennhack, A.; Ness, J.; Baches, S.; Pietrzik, C. U.; Bulic, B.; Weggen,
S. PLoS ONE 2012, 7, e30484.
Purity (HPLC) = 98%; tR = 13.19 min
Acknowledgments
We thank Karlheinz Baumann and Manfred Brockhaus
(F. Hoffmann-La Roche Ltd, Switzerland) for carboxyl terminus-
specific Ab antibodies, Harrie Gijsen (Janssen Pharmaceutica NV,
Belgium) for compounds 1 and JNJ-40418677, Sangram Sisodia
(University of Chicago, USA) for N2a-ANPP cells, Hiroshi Mori
(Osaka City University, Japan) for antibody PSN2, Edward Koo
(UC San Diego, USA) for various reagents, and Herbert Waldmann
(Max-Planck Institute of Molecular Physiology, Germany) and Gui-
do Reifenberger (University of Duesseldorf, Germany) for encour-
agement and support.
This study was supported by the Competence Network Degener-
ative Dementias of the Federal Ministry of Education (Grant num-
bers 01 GI 1004D, 01 GI 1004C, 01 GI 1004B to C.P., B.B. and S.W.).
36. Kounnas, M. Z.; Danks, A. M.; Cheng, S.; Tyree, C.; Ackerman, E.; Zhang, X.; Ahn,
K.; Nguyen, P.; Comer, D.; Mao, L.; Yu, C.; Pleynet, D.; Digregorio, P. J.;
Velicelebi, G.; Stauderman, K. A.; Comer, W. T.; Mobley, W. C.; Li, Y. M.; Sisodia,
S. S.; Tanzi, R. E.; Wagner, S. L. Neuron 2010, 67, 769.
37. Ebke, A.; Luebbers, T.; Fukumori, A.; Shirotani, K.; Haass, C.; Baumann, K.;
Steiner, H. J. Biol. Chem. 2011, 286, 37181.
38. Ohki, Y.; Higo, T.; Uemura, K.; Shimada, N.; Osawa, S.; Berezovska, O.;
Yokoshima, S.; Fukuyama, T.; Tomita, T.; Iwatsubo, T. EMBO J. 2011, 30, 4815.
39. Dubinsky, L.; Krom, B. P.; Meijler, M. M. Bioorg. Med. Chem. 2012, 20, 554.
40. Blencowe, A.; Hayes, W. Soft Matter 2005, 1, 178.
41. Kotzyba-Hibert, F.; Kapfer, I.; Goeldner, M. Angew. Chem., Int. Ed. 1995, 34,
1296.
42. Fleming, S. A. Tetrahedron 1995, 51, 12479.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2012.08.034.
These data include MOL files and InChiKeys of the most important
compounds described in this article.
43. Brunner, J.; Senn, H.; Richards, F. M. J. Biol. Chem. 1980, 255, 3313.
44. Hannam, J. C.; Kulagowski, J. J.; Madin, A.; Ridgill, M. P.; Seward, E. M. WO/
2006/043064 2006.
References and notes
45. Gommermann, N.; Koradin, C.; Polborn, K.; Knochel, P. Angew. Chem., Int. Ed.
2003, 42, 5763.
46. Li, P.; Zhang, Y.; Wang, L. Chem. Eur. J. 2009, 15, 2045.
1. Puente, X. S.; Sanchez, L. M.; Overall, C. M.; Lopez-Otin, C. Nat. Rev. Genet. 2003,
4, 544.

6532
A. Rennhack et al. / Bioorg. Med. Chem. 20 (2012) 6523–6532
47. Zhang, Y.; Li, P.; Wang, M.; Wang, L. J. Org. Chem. 2009, 74, 4364.
48. Wei, C.; Li, C.-J. J. Am. Chem. Soc. 2003, 125, 9584.
63. Weidemann, A.; Eggert, S.; Reinhard, F. B.; Vogel, M.; Paliga, K.; Baier, G.;
Masters, C. L.; Beyreuther, K.; Evin, G. Biochemistry 2002, 41, 2825.
64. Lichtenthaler, S. F.; Beher, D.; Grimm, H. S.; Wang, R.; Shearman, M. S.; Masters,
C. L.; Beyreuther, K. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 1365.
65. Gu, Y.; Misonou, H.; Sato, T.; Dohmae, N.; Takio, K.; Ihara, Y. J. Biol. Chem. 2001,
276, 35235.
66. Page, R. M.; Baumann, K.; Tomioka, M.; Perez-Revuelta, B. I.; Fukumori, A.;
Jacobsen, H.; Flohr, A.; Luebbers, T.; Ozmen, L.; Steiner, H.; Haass, C. J. Biol.
Chem. 2008, 283, 677.
49. Sharma, S.; Oehlschlager, A. C. J. Org. Chem. 1989, 54, 5064.
50. Fischer, C.; Munoz, B.; Rivkin, A. A. WO/2008/097538 2008.
51. Chen, L. S.; Chen, G. J.; Tamborski, C. J. Organomet. Chem. 1983, 251, 139.
52. Richardson, S. K.; Ife, R. J. J. Chem. Soc., Perkin Trans. 1 1989, 1172.
53. Song, J. I.; An, D. K. Chem. Lett. 2007, 36, 886.
54. Narlawar, R.; Baumann, K.; Czech, C.; Schmidt, B. Bioorg. Med. Chem. Lett. 2007,
17, 5428.
55. Kan, T.; Kita, Y.; Morohashi, Y.; Tominari, Y.; Hosoda, S.; Tomita, T.; Natsugari,
H.; Iwatsubo, T.; Fukuyama, T. Org. Lett. 2007, 9, 2055.
56. Sato, C.; Morohashi, Y.; Tomita, T.; Iwatsubo, T. J. Neurosci. 2006, 26, 12081.
57. Tolia, A.; Chavez-Gutierrez, L.; De Strooper, B. J. Biol. Chem. 2006, 281, 27633.
58. Osenkowski, P.; Li, H.; Ye, W.; Li, D.; Aeschbach, L.; Fraering, P. C.; Wolfe, M. S.;
Selkoe, D. J. J. Mol. Biol. 2009, 385, 642.
59. Takagi, S.; Tominaga, A.; Sato, C.; Tomita, T.; Iwatsubo, T. J. Neurosci. 2010, 30,
15943.
60. Sobhanifar, S.; Schneider, B.; Löhr, F.; Gottstein, D.; Ikeya, T.; Mlynarczyk, K.;
Pulawski, W.; Ghoshdastider, U.; Kolinski, M.; Filipek, S.; Güntert, P.; Bernhard,
F.; Dötsch, V. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 9644.
67. Qi-Takahara, Y.; Morishima-Kawashima, M.; Tanimura, Y.; Dolios, G.; Hirotani,
N.; Horikoshi, Y.; Kametani, F.; Maeda, M.; Saido, T. C.; Wang, R.; Ihara, Y. J.
Neurosci. 2005, 25, 436.
68. Kakuda, N.; Funamoto, S.; Yagishita, S.; Takami, M.; Osawa, S.; Dohmae, N.;
Ihara, Y. J. Biol. Chem. 2006.
69. Mori, K.; Okochi, M.; Tagami, S.; Nakayama, T.; Yanagida, K.; Kodama, T. S.;
Tatsumi, S.-i.; Fujii, K.; Tanimukai, H.; Hashimoto, R.; Morihara, T.; Tanaka, T.;
Kudo, T.; Funamoto, S.; Ihara, Y.; Takeda, M. Psychogeriatrics 2010, 10, 117.
70. Chavez-Gutierrez, L.; Bammens, L.; Benilova, I.; Vandersteen, A.; Benurwar, M.;
Borgers, M.; Lismont, S.; Zhou, L.; Van Cleynenbreugel, S.; Esselmann, H.;
Wiltfang, J.; Serneels, L.; Karran, E.; Gijsen, H.; Schymkowitz, J.; Rousseau, F.;
Broersen, K.; De Strooper, B. EMBO J. 2012, 31, 2261.
71. Uno, H.; Nibu, N.; Misobe, N. Bull. Chem. Soc. Jpn. 1999, 72, 1365.
72. Kwiatkowski, S.; Crocker, P. J.; Chavan, A. J.; Nobuyuki, I.; Haley, B. E.; Watt, D.
S.; Ren-jye, H. Tetrahedron Lett. 1990, 31, 2093.
61. Sastre, M.; Steiner, H.; Fuchs, K.; Capell, A.; Multhaup, G.; Condron, M. M.;
Teplow, D. B.; Haass, C. EMBO Rep. 2001, 2, 835.
62. Yu, C.; Kim, S. H.; Ikeuchi, T.; Xu, H.; Gasparini, L.; Wang, R.; Sisodia, S. S. J. Biol.
Chem. 2001, 276, 43756.