Triazole-linked reduced amide isosteres: An approach for the fragment-based drug discovery of anti-Alzheimer's BACE1 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.05.007

In the course of a β-site APP-cleaving enzyme 1 (BACE1) inhibitor discovery project an in situ synthesis/screening protocol was employed to prepare 120 triazole-linked reduced amide isostere inhibitors. Among these compounds, four showed modest (single digit micromolar) BACE1 inhibition. Our ligand design was based on a potent reduced amide isostere 1, wherein the P ₂ amide moiety was replaced with an anti-1,2,3-triazole unit. Unfortunately, this replacement resulted in a 1000-fold decrease in potency. Docking studies of triazole-linked reduced amide isostere A3Z10 and potent oxadiazole-linked tertiary carbinamine 2a with BACE1 suggests that the docking poses of A3Z10 and 2a in the active sites are quite similar, with one exception. In the docked structures the placement of the protonated amine that engages D228 differs considerably between 2a and A3Z10. This difference could account for the lower BACE1 inhibition potency of A3Z10 and related compounds relative to 2a.

Full text of DOI:10.1016/j.bmcl.2011.05.007

Bioorganic & Medicinal Chemistry Letters 21 (2011) 3992–3996
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Triazole-linked reduced amide isosteres: An approach
for the fragment-based drug discovery of anti-Alzheimer’s BACE1 inhibitors
Christopher J. Monceaux a, Chiho Hirata-Fukae b, , Polo C.-H. Lam c, Maxim M. Totrov c, Yasuji Matsuoka b,
,
⇑
,à
Paul R. Carlier a,
⇑
a Department of Chemistry, Virginia Tech, Blacksburg, VA 24061, USA
b Department of Neurology, Georgetown University Medical Center, Building D, Suite 177, 4000 Reservoir Rd. NW, Washington, DC 20057, USA
c Molsoft LLC, 11199 Sorrento Valley Rd., CA 92121, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
In the course of a b-site APP-cleaving enzyme 1 (BACE1) inhibitor discovery project an in situ synthesis/
screening protocol was employed to prepare 120 triazole-linked reduced amide isostere inhibitors.
Among these compounds, four showed modest (single digit micromolar) BACE1 inhibition. Our ligand
design was based on a potent reduced amide isostere 1, wherein the P2 amide moiety was replaced with
an anti-1,2,3-triazole unit. Unfortunately, this replacement resulted in a 1000-fold decrease in potency.
Docking studies of triazole-linked reduced amide isostere A3Z10 and potent oxadiazole-linked tertiary
carbinamine 2a with BACE1 suggests that the docking poses of A3Z10 and 2a in the active sites are quite
similar, with one exception. In the docked structures the placement of the protonated amine that engages
D228 differs considerably between 2a and A3Z10. This difference could account for the lower BACE1 inhi-
bition potency of A3Z10 and related compounds relative to 2a.
Received 28 February 2011
Revised 28 April 2011
Accepted 2 May 2011
Available online 8 May 2011
Keywords:
BACE1 reduced amide inhibitors
Microtiter plate-based screening
‘Click’ chemistry
Diazotransfer
Ó 2011 Elsevier Ltd. All rights reserved.
In silico docking
Alzheimer’s disease (AD) is a progressive neurodegenerative
disease whose first clinical manifestation is cognitive impairment,
followed by dementia, and eventually death. Data from the 2000
US census indicates that AD currently affects more than 4.5 million
Americans.1 According to the Alzheimer’s Association, this afflic-
tion is the 6th leading cause of death in the United States.2 The
amyloid cascade hypothesis postulates that AD results from an
accumulation of the 40 and 42 amino acid residue peptides, Ab
or Ab40/42, which form insoluble aggregates. These aggregates or
plaques are implicated in subsequent neurodegenerative processes
that lead to cognitive impairment.3 Since formation of Ab requires
proteolytic cleavage of the amyloid precursor protein (APP) by the
b-site APP cleaving enzyme 1 (BACE1), BACE1 appears to be a log-
ical target to slow or halt the progression of AD.4
Coburn et al. at Merck reported that reduced amide 1 exhibited
8 nM potency to inhibit BACE1 (Fig. 1).5 We sought to further re-
duce the peptidic nature of 1 by replacing the P2 amide (boxed)
with a 1,2,3-triazole heterocycle. 1,2,3-Triazoles have been shown
to be effective amide surrogates in applications such as acetylcho-
linesterase inhibitors6 and HIV-1 protease inhibitors.7
The large dipole moment, combined with the ability of the sp2
nitrogen atoms to serve as hydrogen bond acceptors, make the
1,2,3-triazole an appealing amide mimic.7 The potent Merck inhib-
itor 2a8 replaced the P2 amide of 1 with an oxadiazole, simulta-
neously exchanging the reduced amide isostere with a tertiary
carbinamine. It thus seemed possible that simple anti-triazole
replacement of the P2 amide in 1 might also be tolerated.
An X-ray crystal structure of 2a bound to BACE1 (PDB ID:
2IRZ) demonstrated that the oxadiazole nitrogens mimic the P2
amide carbonyl in receiving an H-bond from an important resi-
due (Q73) within the active site of the BACE1 enzyme (PDB ID
2IRZ),8 and it seemed likely that the nitrogen atoms of a 1,2,3-
triazole could play the same role. Additionally, the incorporation
of a 1,2,3-triazole unit would allow facile inhibitor assembly via
the copper-catalyzed acetylene azide cycloaddition reaction
(CuAAC); in this regard we wanted to apply the high throughput
screening (copper-catalyzed microtiter plate screening) tech-
nique of Wong, Sharpless, and Fokin.9,10 In this way we hoped
to rapidly interrogate both subtle and major P- and P0-side
changes in the structure of triazole-linked analogs of 1. Note that
after we had concluded our work, a Merck team in 2010 re-
ported a 1,2,3-triazole-linked tertiary carbinamine inhibitor 2b,
which surprisingly offered dramatically reduced inhibition po-
⇑
Corresponding authors.
E-mail addresses: yasuji.y.matsuoka@gsk.com (Y. Matsuoka), pcarlier@vt.edu
(P.R. Carlier).
tency relative to 2a (16.3 vs 0.012 l
M, respectively).11 We offer
further commentary on the relative potencies of 2a and 2b in
the modeling section below.
Senior Research Fellow Center, Ehime University, Shitsukawa, Toon, Ehime, Japan.
Present address: Neural Pathway Discovery, GlaxoSmithKline, Singapore 138667,
à
Singapore (Y.M.).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.05.007

C. J. Monceaux et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3992–3996
3993
Figure 1. Reduced amide 1, heterocycle-linked tertiary carbinamines 2a and 2b, and HEA 3–4 inhibitors of BACE1.
Our approach differed from that of Merck’s triazole 2b in two
ways (Fig. 2). Firstly, the warhead of the inhibitor would incorpo-
rate a reduced amide isostere rather than a tertiary carbinamine.
Secondly, we decided to synthesize acetylene fragments that com-
prise the P3–P2 (i.e., left) portion of 1, and use the azide fragments
and ( )-A10b.12 Since 1,4-diethynyl benzene is bifunctional and
could make triazole dimers, for our screening experiments we gave
it two different designations based on the equivalents used for the
CuAAC reaction (A8 1 equiv; A9 3.6 equiv).
The selected azide components Z1–Z12 are shown in Scheme 1.
0
0
to incorporate the reduced amide P1–P2 fragment (Figure 2). Note
Azides Z1–Z8 incorporate the P1 amide functionality of Merck re-
0
that triazole 2b is formally derived from a P3–P2 azide and a P1–P1
acetylene.
duced amide 1, and feature the favored benzyl (cf. 1–4) and isobu-
tyl13,14 groups as P1 substituents. Azides Z9–Z12 replace the P1
0
We were cognizant that our synthetic strategy would increase
the distance between the isophthalimide and the basic amine by
one bond relative to 1 and 2a, but believed this change would
not necessarily be deleterious, given the planned variation at
amide with a benzylamine group, a strategy that was successful
for inhibitor 315 and other similar HEA isosteres.14 Two key trans-
formations were used in the syntheses of azides Z1–Z12. The first
was an oxidation/reductive amination sequence of 5 and 6 to give
the secondary amines 7a–d and 8a–d, and 9 to give secondary
amines 10a–d (Scheme 1). Following deprotection of the Cbz or
Boc groups of 7–9, respectively, the substrate contained two amine
groups. As hoped, we found that only the primary amines reacted
in the second key transformation (diazotransfer), affording azides
Z1–Z12 in moderate to good yields. This finding is consistent with
the mechanism proposed by Nyffeler et al.16 but to our knowledge
this is the first example of the diazotransfer reaction performed in
the presence of a secondary amine.
0
0
P3, P1, P1 and P2 . In retrospect we can note that triazole-linked
tertiary carbinamine 2b, which maintains the spacing of oxadiaz-
ole 2a, nevertheless lost significant inhibition potency relative to
2a.
The acetylenes employed are depicted in Figure 3. Acetylenes
A1–A4 were designed to mimic the isophthalimide P3–P2 portion
of 1, 2a, 3, and 4; their syntheses are described in Supplementary
data. Acetylenes A5–A10 were gifts and were incorporated in the
screen to search for new ligands for the P3–P2 region. We have also
recently reported regioselective syntheses of A10a (Ci-symmetric)
With the acetylene and azide building blocks in hand we began
screening an array of triazole-linked reduced amide isosteres using
an in situ synthesis/assay protocol. We based our protocol on the
Cu-catalyed microtiter plate-based synthesis/screening procedure
of Brik et al. making minor modifications.9,17 All 120 possible com-
binations of A1–A10 with Z1–Z12 were prepared and assayed in
this way, using the commercially available HEA isostere inhibitor
4
18 (Inhibitor IV, Calbiochem product catalog no. 565788) as a po-
sitive control. In addition to 4, three of the 120 binary combina-
tions showed significant inhibition at 10 M: A3Z9, A3Z10, and
l
A10Z10. These triazoles were then synthesized by traditional or-
ganic techniques, purified, and assayed. Since A10 was a mixture
of regioisomers, these were separated (the 1,4-diol was designated
A10a and the 1,3-diol A10b) and were used in the CuAAC reaction
with Z10 to afford A10aZ10 and A10bZ10, both as mixtures of dia-
stereomers. Unfortunately, these compounds only showed modest,
single digit micromolar inhibition of BACE1 (Table 1), similar to
that of Merck triazole-linked tertiary carbinamine 2b. In contrast,
HEA inhibitor 4 exhibited an IC50 of 0.020 lM, similar to the liter-
Figure 2. Synthetic strategy employed for 1,2,3-triazole-linked reduced amide
inhibitors.
ature report.18

3994
C. J. Monceaux et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3992–3996
Figure 3. Acetylenes used in the in situ synthesis/assay.
0
0
Scheme 1. Synthesis of the azide components. Reagents and conditions: (a) (COCl)2, DMSO, DIPEA, CH2Cl2 ꢀ78 °C to rt; (b) Ala-NHR2 or 3-R2 -C6H4CH2NH2, MgSO4,
NaBH(OAc)3, (CH2Cl)2; (c) Pd/C, H2, CH3OH; (d) TfN3, CuSO4, H2O/CH3OH/CH2Cl2, 1:3:1; (e) 4 N HCl in 1,4-dioxane.
To assess why the triazole-linked reduced amide inhibitors
exhibited such low inhibition potency, A3Z10 and potent oxadiaz-
ole-linked tertiary carbinamine 2a were docked within the active
site of the BACE1 enzyme. Ligands were removed from five pub-
lished BACE1-inhibitor co-crystal structures (PDB ID: 2PH6,19
2IRZ,8 2B8L,20 2IS0,8 and 2QZL5) and the resulting PDB objects were
converted into ICM (internal coordinate mechanics) objects;21,22
these represent five alternative conformations of BACE1. The best
docking poses of 2a and A3Z10 within the active sites of these five
structures were then found through in silico docking using the ICM
docking module.22,23 The most favorable docking scores were
found for the 2PH6-derived structure, and Figure 4 shows both
2a and A3Z10 overlaid in the active site of this structure of BACE1.
From Figure 4 it is apparent that the left-hand portion of both li-
gands adopt a very similar docking pose. However, one can clearly
see that the ammonium (N1) of A3Z10 is no longer in an optimal
position in which to form a salt bridge to the D228 carboxylate,
which is assumed to be a necessary feature of potent reduced
amide BACE1 inhibitors.5 With regard to the significant loss of po-
tency of triazole-linked tertiary carbinamine 2b relative to 2a, our
docking studies show very similar docking poses of these com-
pounds, but with a significantly reduced docking score for 2b.
Rajapakse et al. have attributed the disparate affinities of 2a and
2b to differences in conformational constraints provided by the
oxadiazole and triazole rings.11 Thus it appears that significant
re-design of our triazole-linked reduced amide isosteres would
be necessary to achieve potent inhibition. It is possible that the tri-
azole, however oriented, is not a good P2 amide surrogate for this
subclass of BACE1 inhibitors.
Since the P3–P2 region of A10b- and A10a-derived triazole
inhibitors differs significantly from that of A3Z10, 1, and 2a, we
also performed docking studies of the diastereomers of A10bZ10
with BACE1. Although neither diastereomer docked correctly in
the 2PH6-derived structure, acceptable docking poses for both
the (1S,10S,30S,40S,60S)- and (1S,10R,30R,40R,60R)-diastereomers were
obtained with the 2IRZ-derived structure. It thus seems possible
that the micromolar inhibition afforded by A10bZ10 could be
attributed to active site occupancy.
In summary, 120 reduced amide isosteres were screened using
a high throughput in situ synthesis/screening protocol. Of these, 4

C. J. Monceaux et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3992–3996
3995
Table 1
IC50 values for the most potent triazole-linked reduced amide inhibitors and control HEA inhibitor 4 (synthesized yield in parenthesis)
Compound
IC50
(lM)
95% CI (lM)
A3Z9
A3Z10
A10aZ10
A10bZ10
4 (Fig. 1)
7.0
2.0
3.4
2.4
5.7–8.6
1.9–2.1
3.2–3.7
2.2–2.7
0.020
0.019–0.021
Figure 4. In silico docking of 2a (off-white) and A3Z10 (green).

3996
C. J. Monceaux et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3992–3996
9. Brik, A.; Muldoon, J.; Lin, Y.-C.; Elder, J. H.; Goodsell, D. S.; Olson, A. J.; Fokin, V.
V.; Sharpless, K. B.; Wong, C.-H. Chem. Biol. Chem. 2003, 4, 1246.
10. Brik, A.; Wu, C. Y.; Wong, C. H. Org. Biomol. Chem. 2006, 4, 1446.
11. Rajapakse, H. A.; Nantermet, P. G.; Selnick, H. G.; Barrow, J. C.; McGaughey, G.
B.; Munshi, S.; Lindsley, S. R.; Young, M. B.; Ngo, P. L.; Katherine Holloway, M.;
Lai, M.-T.; Espeseth, A. S.; Shi, X.-P.; Colussi, D.; Pietrak, B.; Crouthamel, M.-C.;
Tugusheva, K.; Huang, Q.; Xu, M.; Simon, A. J.; Kuo, L.; Hazuda, D. J.; Graham, S.;
Vacca, J. P. Bioorg. Med. Chem. Lett. 2010, 20, 1885.
12. Monceaux, C. J.; Carlier, P. R. Org. Lett. 2010, 12, 620.
13. Ghosh, A. K.; Shin, D. W.; Downs, D.; Koelsch, G.; Lin, X. L.; Ermolieff, J.; Tang, J.
J. Am. Chem. Soc. 2000, 122, 3522.
14. Ghosh, A. K.; Kumaragurubarana, N.; Hong, L.; Kulkarni, S.; Xu, X. M.; Miller, H.
B.; Reddy, D. S.; Weerasena, V.; Turner, R.; Chang, W.; Koelsch, G.; Tang, J.
Bioorg. Med. Chem. Lett. 2008, 18, 1031.
15. Maillard, M. C.; Hom, R. K.; Benson, T. E.; Moon, J. B.; Mamo, S.; Bienkowski, M.;
Tomasselli, A. G.; Woods, D. D.; Prince, D. B.; Paddock, D. J.; Emmons, T. L.;
Tucker, J. A.; Dappen, M. S.; Brogley, L.; Thorsett, E. D.; Jewett, N.; Sinha, S.;
John, V. J. Med. Chem. 2007, 50, 776.
compounds (A3Z9, A3Z10, A10aZ10, A10bZ10) evidenced signifi-
cant inhibition at 10 M in a cell-free FRET assay, and were further
l
examined, revealing single-digit micromolar IC50 values. Addition-
ally, we have shown that the copper-mediated diazotransfer to pri-
mary amines can be performed in the presence of unprotected
secondary amines. We believe that the synthetic strategy demon-
strated herein will be useful for preparation of functionalized
azides for application in ‘click’ chemistry-based drug discovery of
reduced amide inhibitors for other proteases.
Acknowledgments
This work was supported by the Virginia Alzheimer’s and
Related Diseases Research Award Fund (09-1) and the Virginia
Tech-Georgetown Consortium for Drug Discovery. We thank
Professor K. Barry Sharpless for donations of acetylenes A5–A10,
and Dr. Dawn Wong for assistance with the preparation of Figure 4.
16. Nyffeler, P. T.; Liang, C. H.; Koeller, K. M.; Wong, C. H. J. Am. Chem. Soc. 2002,
124, 10773.
17. For each pair of acetylene and azide, a 20 mM solution of an azide in t-butanol
(100
(100
l
l
L) was combined with a 24 mM solution of an acetylene in acetonitrile
L) and 200 L of water containing copper sulfate (0.5 mM) and sodium
l
ascorbate (5 mM) in a scintillation vial. The vials were purged with argon,
capped and sealed with parafilm, and shaken to achieve homogeneity. After
five days at room temperature, LC/MS analysis of the A2Z1 reaction indicated
greater than 95% consumption of the azide and formation of the expected
triazole. We applied this protocol to prepare all 120 possible combinations of
A1–A10 with Z1–Z12. After five days, these reactions were diluted 31.25-fold
Supplementary data
Supplementary data (synthetic procedures, analytical data, and
details of the BACE1 inhibition assay procedure) associated with
this article can be found, in the online version, at doi:10.1016/
j.bmcl.2011.05.007.
with water to achieve a nominal inhibitor concentration of 160
quantitative conversion of the azide to the triazole. These nominally 160
solutions were then used directly in the BACE1 assay (Perkin Elmer
TruPointTM Beta-Secretase Assay Kit), which affords a nominal 10 M final
concentration of inhibitor in the 96 well plate. Note that the final assay
concentration of Cu and ascorbate in the assay were and 50 M,
lM, assuming
l
M
l
References and notes
5
lM
l
respectively. Control experiments established that these concentrations of Cu
and ascorbate did not affect fluorescence in this FRET assay.
1. Hebert, L. E.; Scherr, P. A.; Bienias, J. L.; Bennett, D. A.; Evans, D. A. Arch. Neurol.
2003, 60, 1119.
18. Stachel, S. J.; Coburn, C. A.; Steele, T. G.; Jones, K. G.; Loutzenhiser, E. F.; Gregro,
A. R.; Rajapakse, H. A.; Lai, M.-T.; Crouthamel, M.-C.; Xu, M.; Tugusheva, K.;
Lineberger, J. E.; Pietrak, B. L.; Espeseth, A. S.; Shi, X.-P.; Chen-Dodson, E.;
Hollloway, M. K.; Munshi, S.; Simon, A. J.; Kuo, L.; Vacca, J. P. J. Med. Chem. 2004,
47, 6447.
19. Lindsley, S. R.; Moore, K. P.; Rajapakse, H. A.; Selnick, H. G.; Young, M. B.; Zhu,
H.; Munshi, S.; Kuo, L.; McGaughey, G. B.; Colussi, D.; Crouthamel, M.-C.; Lai,
M.-T.; Pietrak, B.; Price, E. A.; Sankaranarayanan, S.; Simon, A. J.; Seabrook, G.
R.; Hazuda, D. J.; Pudvah, N. T.; Hochman, J. H.; Graham, S. L.; Vacca, J. P.;
Nantermet, P. G. Bioorg. Med. Chem. Lett. 2007, 17, 4057.
20. Stachel, S. J.; Coburn, C. A.; Steele, T. G.; Crouthamel, M.-C.; Pietrak, B. L.; Lai,
M.-T.; Holloway, M. K.; Munshi, S. K.; Graham, S. L.; Vacca, J. P. Bioorg. Med.
Chem. Lett. 2006, 16, 641.
21. Abagyan, R. A.; Totrov, M. M.; Kuznetsov, D. A. J. Comput. Chem. 1994, 15, 488.
22. Abagyan, R. A. ICM software manual, Molsoft LLC, San Diego, CA, 2010. http://
www.molsoft.com/man/index.html
2. Mebane-Sims, I. Alzheimer’s
Association 2009, 5, 234.
& dementia: the journal of the Alzheimer’s
3. Hardy, J. A.; Higgins, G. A. Science 1992, 256, 184.
4. Vassar, R.; Bennett, B. D.; Babu-Khan, S.; Kahn, S.; Mendiaz, E. A.; Denis, P.;
Teplow, D. B.; Ross, S.; Amarante, P.; Loeloff, R.; Luo, Y.; Fisher, S.; Fuller, L.;
Edenson, S.; Lile, J.; Jarosinski, M. A.; Biere, A. L.; Curran, E.; Burgess, T.; Louis, J.
C.; Collins, F.; Treanor, J.; Rogers, G.; Citron, M. Science 1999, 286, 735.
5. Coburn, C. A.; Stachel, S. J.; Jones, K. G.; Steele, T. G.; Rush, D. M.; DiMuzio, J.;
Pietrak, B. L.; Lai, M. T.; Huang, Q.; Lineberger, J.; Jin, L. X.; Munshi, S.; Holloway,
M. K.; Espeseth, A.; Simon, A.; Hazuda, D.; Graham, S. L.; Vacca, J. P. Bioorg. Med.
Chem. Lett. 2006, 16, 3635.
6. Bourne, Y.; Kolb, H. C.; Radic, Z.; Sharpless, K. B.; Taylor, P.; Marchot, P. Proc.
Natl. Acad. Sci. U. S. A. 2004, 101, 1449.
7. Brik, A.; Alexandratos, J.; Lin, Y.-C.; Elder, J. H.; Olson, A. J.; Wlodawer, A.;
Goodsell, D. S.; Wong, C.-H. Chem.Biol. Chem. 2005, 6, 1167.
8. Rajapakse, H. A.; Nantermet, P. G.; Selnick, H. G.; Munshi, S.; McGaughey, G. B.;
Lindsley, S. R.; Young, M. B.; Lai, M.-T.; Espeseth, A. S.; Shi, X.-P.; Colussi, D.;
Pietrak, B.; Crouthamel, M.-C.; Tugusheva, K.; Huang, Q.; Xu, M.; Simon, A. J.;
Kuo, L.; Hazuda, D. J.; Graham, S.; Vacca, J. P. J. Med. Chem. 2006, 49, 7270.
23. Totrov, M. M.; Abagyan, R. A. Proteins 1997, 1, 215.