Design and synthesis of BACE-1 inhibitors utilizing a tertiary hydroxyl motif as the transition state mimic

Article abstract of DOI:10.1016/j.bmcl.2009.06.065

Two series of drug-like BACE-1 inhibitors with a shielded tertiary hydroxyl as transition state isostere have been synthesized. The most potent inhibitor exhibited a BACE-1 IC₅₀ value of 0.23 μM.

Full text of DOI:10.1016/j.bmcl.2009.06.065

Bioorganic & Medicinal Chemistry Letters 19 (2009) 4711–4714
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design and synthesis of BACE-1 inhibitors utilizing a tertiary hydroxyl motif
as the transition state mimic
Fredrik Wångsell a, Francesco Russo a, Jonas Sävmarker a, Åsa Rosenquist b,c
,
Bertil Samuelsson c,d, Mats Larhed a,
*
a Organic Pharmaceutical Chemistry, Department of Medicinal Chemistry, Uppsala University, BMC, Box 574, SE-751 23 Uppsala, Sweden
b Department of Chemistry, Linköping University, SE-581 83 Linköping, Sweden
c Medivir AB, Lunastigen 7, SE-141 44 Huddinge, Sweden
d Department of Organic Chemistry, Arrhenius Laboratory, Stockholm University, SE-106 91 Stockholm, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
Two series of drug-like BACE-1 inhibitors with a shielded tertiary hydroxyl as transition state isostere
Received 6 May 2009
Revised 16 June 2009
Accepted 16 June 2009
Available online 21 June 2009
have been synthesized. The most potent inhibitor exhibited a BACE-1 IC50 value of 0.23 lM.
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Alzheimer’s disease
BACE-1
Enzyme inhibitor
Transition state mimetic
Alzheimer’s disease (AD) is the most common form of dementia
among the elderly and it is currently believed that approximately
37 million people worldwide have the disease.1 Two hallmarks
for AD are accumulation and extracellular aggregation of insoluble
amyloid plaques (Ab) deposits, along with intracellular neurofibril-
lary tangles in the brain. The human aspartic protease BACE-1 has
long been regarded as a therapeutic target for the reduction of Ab
formation.2 BACE-1 is the initial protease that processes amyloid
precursor protein (APP) in the pathway leading to Ab.3 The finding
that BACE-1 knockout mice were unable to produce Ab provided an
in vivo validation of BACE-1 as one of the proteases responsible for
Ab production in the brain.4,5 Thus, inhibition of BACE-1 is a prom-
ising strategy in the search for an effective AD treatment.
Inhibition of aspartic proteases has been challenging and in the
case of BACE-1 the first generation of inhibitors were inspired by
knowledge gained in the development of HIV-1 protease and renin
inhibitors.6 The first publically available enzyme/inhibitor co-crys-
tal X-ray structure (PDB-code: 1FKN) further spurred the discovery
process.7 Despite extensive research and even though numerous
potent and selective inhibitors have been found, no BACE-1 inhib-
itors have yet reached the market. Several reported inhibitors suf-
fer from poor bioavailability, insufficient membrane permeability
and blood-brain barrier penetration due to their peptidic
character.1,4
Previous work from our laboratory has demonstrated that HIV-
1 protease inhibitors containing a shielded tertiary hydroxyl group
as transition state mimic have improved Caco-2 cell membrane
permeability while retaining good inhibitory properties.8 Inspired
by these results, we sought to develop BACE-1 inhibitors contain-
ing a tertiary hydroxyl group as the central transition state isoste-
re, together with a benzyl or a phenylethylene group in the
P1-position. A flexible and robust synthetic route was developed,
which gave access to two different tertiary statine-like moieties
(Fig. 1). Both structural classes were used in combination with a
substituted isophthalamide containing an inverted amide bond
and various selected substituents in the P20-P30 position, designed
to bind into the enzyme subsites S20-S30.
The synthesis of the final inhibitors is depicted in Scheme 1.9
Starting by protection of 2-hydroxy-3-phenyl-propionic acid
(n = 1) or 2-hydroxy-4-phenyl-butyric acid (n = 2) to render ketals
OH
O
O
O
O
O
H
N
N
H
N
H
N
H
N
H
N
H
P1
P1
OH
OH
O
P1
Statine-like
Type I
Type II
Figure 1. A comparison between transition state isosteres, the commonly used
statine motif and the two transition state cores discussed in this report. Both type I
and II uses a tertiary hydroxyl group in the transition state mimic.
* Corresponding author. Tel.: +46 18 471 4667; fax: +46 18 471 4474.
E-mail address: mats.larhed@orgfarm.uu.se (M. Larhed).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.06.065

4712
F. Wångsell et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4711–4714
O
HO
O
OH
n
O
O
O
O
ii, iii
i
iv
HO
O
O
O
O
O
O
n
n
n
not
isolated
n = 1 or 2
1
3, n=1, 36%
4, n=2, 44%
, n=1, 87%
2, n=2, 99%
P2
P3sp
R1
Ms
Ms
N
N
Type I
8, n=1, R11=F, R22=a (69%)
10, n=2, R1=F, R2=c (72%)
11, n=2, R1=H, R2=c (69%)
P2'-P3'
R1
9
O
O
, n=1, R =F, R =c (70%)
v
vi
O
O
H
N
O
H
N
OH
N
N
H
R2
H
n
O
O
12, n=1, R1=F, R2=d (65%)
13, n=1, R1=F, R2=g (50%)
5, n=1, R1=F, 67%
n
P3
6, n=2, R1=F, 17%
P1
1
vii
7
, n=2, R =H, 21%
14, n=1, R1=F, R2=e (76%)
15, n=1, R1=F, R2=h (77%)
P2
P3sp
Ms
Type II
N
O
O
O
n
P2'-P3'
R2
O
O
n
R4O
R4O
H
N
R2
n
vii, ix,x
xi
R2
HO
N
O
O
H
viii
49-87%
O
P3
R4=TBDMS
P1
2
xii
16
17
-(A), n=2, R2=a (32%)
R4=H
R4=TBDMS
20-(R), n=1, R2=b
21-(S), n=1, R22=b
-(B), n=2, R =a (28%)
18, n=2, R22=c (76%)
27-(R), n=1, R2=b (44%)
34-(R), n=1, R2=b (44%)
22
-(A), n=2, R =b (87%)
2
35-(S), n=1, R22=b (54%)
19
, n=2, R =f (95%)
28
-(S), n=1, R =b (58%)
23-(B), n=2, R2=b (62%)
24, n=2, R2=c 2(49%)
29-(A), n=2, R2=b (31%)
36
-(A), n=2, R =b (74%)
30-(B), n=2, R2=b (59%)
37-(B), n=2, R2=b (69%)
38, n=2, R2=c 2(61%)
25
-(A), n=2, R =f (30%)
2
31
, n=2, R =c (48%)
26-(B), n=2, R2=f (35%)
32-(A), n=2, R2=f (76%)
33-(B), n=2, R2=f (73%)
39
-(A), n=2, R =f (68%)
40-(B), n=2, R2=f (84%)
R2 (P2'-P3' substituents)
O
Ms
N
R5
O
OR3
R1
H
N
H
H
N
H
N
N
H
N
H2N
H2
N
H2N
H2N
H2N
OR3
NH2
O
O
O
O
O
O
, R5 = H
, R3 = Me
e, R3 = H
, R3 = Me
g
1
Aniline-H
, R = H
a
c
d
f
Aniline-F, R1 = F
b, R5 = TBDMS
h, R3 = H
Scheme 1. Reagents and conditions: (i) Pyridinium p-toluenesulphonic acid, dimethoxy propane, CHCl3, 70 °C; (ii) LDA, THF (dry), ꢀ78 °C; (iii) methyl acrylate; (iv) TFA/H2O
(6:1), 80 °C; (v) aniline-H or aniline-F, EDC, HOBt, DIPEA, CH2Cl2 (dry), 40 °C; (vi) R2, 2-hydroxypyridine, DIPEA, THF (dry), reflux; (vii) LiOH, dioxane/H2O (1:1), rt; (viii) R2,
EDC, HOBt, DIPEA, CH2Cl2 (dry), 40 °C; (ix) TBDMS-OTf, Et3N, DCM/THF (1:1), rt; (x) 1 M K2CO3 (aq), THF, MeOH (1:1:3), rt; (xi) aniline-H, EDC, HOBt, DIPEA, CH2Cl2 (dry); (xii)
TBAF, THF, rt.
1 (87%) and 2 (99%), respectively. The quaternary center was intro-
duced by alkylation of compounds 1 and 2 with methyl acrylate in
a Michael addition to give the intermediates 3 (36%) and 4 (44%),
essentially as previously described procedure.10 For the synthesis
of type I final products, intermolecular lactone formation was per-
formed by heating the esters 3 and 4 in aqueous trifluoracetic acid.
The acid was then coupled with aniline-H (R1 = H)11 or aniline-F
(R1 = F)9 using EDC, HOBt and DIPEA in dry CH2Cl2 to give com-
pounds 5–7 (17–67%). Lactones 5–7 was further ring opened with
a selected set of amines (a, c, d, g) in the presence of DIPEA and 2-
hydroxypyridine in refluxing THF, to give the final type I inhibitors
in 50–72% yield (8–13, Scheme 1 and Table 1). Amine a is commer-
cially available while amines c, d, and g were synthesized accord-
ing to standard procedures.9 The methyl ester in products 12 and
13 was hydrolyzed to furnish inhibitors 14 and 15, containing a
free carboxylic acid in 76–77% yield. The inhibitors 8–15 were
isolated and evaluated for BACE-1 inhibition as diastereomeric
mixtures (1:1). To synthesize inhibitors of type II, the acids gener-
ated from the intermolecular lactonization of 3 and 4 were sub-
jected to standard peptide coupling chemistry (EDC, HOBt, and
DIPEA in dry CH2Cl2) with three different amines a, c, and f to
furnish lactones 16–19 in 60–95% yield. When (1S,2R)-1-amino-
indanol (a) was used, the formed diastereomers could be separated
by flash column chromatography to yield the pure diastereomers
16-(A) and 17-(B) in 32% and 28% yield, respectively. The lactones
16–19 were opened with LiOH and the generated tertiary alcohols
were protected with TBDMS-OTf to provide the corresponding
acids 22–26 in 30–87% yield over three steps. To obtain com-
pounds 25-(A) and 26-(B) as pure diastereomers, the correspond-
ing diastereomeric acid was purified by preparative HPLC.
Structures 20-(R) and 21-(S) were prepared according to a litera-
ture procedure.10 Next, acids 20–26 were reacted with aniline-H
using standard peptide coupling conditions to afford the TBDMS-
protected compounds 27–33 in 44–76% yield. Finally, to complete
the synthesis of type II inhibitors the TBDMS-protecting group was
removed using TBAF in THF to furnish the final inhibitors 34–40 in
44–84% yield (Table 1).
All synthesized inhibitors were screened against BACE-1 to
determine IC50 values (Table 1).12 The optimized and previously
used isophthalamide moiety13 was exclusively used in the synthe-
sized inhibitors with two modifications, an inverted amide bond
next to the novel tertiary hydroxyl transition state mimic and in
some examples with fluorine in the P3sp para position instead of
hydrogen. This P2-P3 group together with the indanolamine (a)
as P20-P30 substituent, previously used in the development of
HIV-1 protease inhibitors, gave an inhibitor with no BACE-1 activ-
ity (8, IC50 >10 lM). Replacing the indanolamine with the previ-
ously used P2’-P3’ BACE-1 substituent, Val-benzylamide4 (c),
furnished an inhibitor with modest BACE-1 inhibitory data (9,
IC50 = 5.9 lM). The weak inhibitory properties were lost when an

F. Wångsell et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4711–4714
4713
Table 1
BACE-1 inhibition data
P2
N
Ms
P3sp
R1
Type II
P2'-P3'
O
O
n
H
N
HO
R2
N
P2
N
N
H
P3sp
R1
Ms
H
Type I
O
P2'-P3'
P3
O
O
H
N
OH
P1
R2
N
H
N
H
O
n
P3
Compd
n
R1
H
R2
BACE-1 IC50
4.6
(lM)
P1
HO
Compd
n
R1
R2
HO
BACE-1 IC50
>10
(lM)
34-(R)
1
8
1
F
HO
HO
35-(S)
36-(A)
1
2
H
H
2.5
H
N
9
1
2
2
1
F
5.9
O
0.57a
H
N
10
11
12
F
>10
5.6
O
O
HO
H
N
37-(B)
2
2
2
2
H
H
H
H
0.25a,b
H
F
O
O
H
N
>10
H
N
38
0.45a
O
O
O
H
N
H
N
13
14
1
1
F
F
4.0
0.23a,c,d
O
O
39-(A)
O
O
O
OH
H
N
H
N
>10
40-(B)
0.32a,d
O
O
a
CatD Ki >5000 nM.
b
c
Papp (Caco-2) <1 ꢁ 10ꢀ6 cm/s.
H
N
Papp (Caco-2) = 1.2 ꢁ 10ꢀ6 cm/s.
15
1
F
>10
d
OH
CatD was determined on a diastereomeric mixture (1:1).
O
functionalities which were used orthogonally to give inhibitors of
the generic type II (Scheme 1). This modification rendered com-
pounds with an improved activity against BACE-1 (Table 1). Intro-
ducing the indanolamine (a) in the P2-P3 position together with
the benzyl moiety in the P1-position furnished compounds 34-
elongation by one carbon was incorporated in the P1 side-chain to
provide 10 (>10 M). Changing fluorine to hydrogen in the para
P3sp position restored the inhibitory properties, cf. 10 and 11
(>10 M and 5.6 M). Several of the previously reported statine
l
l
l
(R) and 35-(S) (IC50 = 4.6 lM and 2.5 lM). The absolute stereo-
containing BACE-1 inhibitors need an acidic functionality in the
P30 position for potent BACE-1 inhibition.12 Introduction of an
acidic functionality in our inhibitors did not furnish inhibitors with
chemistry of 34-(R) and 35-(S) was determined by X-ray crystal-
lography data (PDB-code: 2UY0) from HIV-1 protease inhibitors
synthesized from the intermediates 20-(R) and 21-(S).10 A one
carbon elongation in the P1-position improved the activity of this
improved activity, 14 (>10 lM) and 15 (>10 lM). The correspond-
ing ester analogues were in fact found to be slightly better; com-
pare 14 and 15 with the methyl esters 12 and 13 (BACE-1 IC50
>10 lM and 4.0 lM). All the inhibitors described above contain
scaffold, compare 36-(A) (IC50 = 0.57
0.25 M) versus 34-(R) and 35-(S). Attaching the common Val-ben-
zylamide (c) in the P20-P30 position provided an equipotent inhib-
lM) and 37-(B) (IC50 =
l
the type I tertiary transition state mimic, representing a structural
change to a tertiary hydroxyl group at the same carbon as the pre-
sumed P1 substituent (Fig. 1). Evidently, this modification was not
well tolerated in the active site of the BACE-1 protease. However,
the central core contains two available protected carboxylic acid
itor, 38 (BACE-1 IC50 = 0.45 lM) while replacing the benzylamine
with an aniline in the P30 position, Val-phenylamide (f), afforded
the pure diastereomeric inhibitors 39-(A) and 40-(B) with BACE-
1 IC50 values of 0.23
lM and 0.32 lM, respectively. The epimers
39-(A) and 40-(B) were equally potent which suggests that in this

4714
F. Wångsell et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4711–4714
series of shielded BACE-1 inhibitors the absolute stereochemistry
at the quaternary center is of minor importance. The most potent
inhibitors from this series were also screened for inhibition of
the anti-target human cathepsin D (Table 1) and were found to
be selective towards BACE-1 (CatD Ki >5000 nM).12 Cell membrane
permeability analysis (Caco-2) of 37-(B) and 39-(A) revealed that
this class of inhibitors have relatively low membrane permeability
(<1 ꢁ 10ꢀ6 and 1.2 ꢁ 10ꢀ6 cm/s, respectively).
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2009.06.065.
References and notes
1. Mount, C.; Downton, C. Nat. Med. 2006, 12, 780.
2. Cole, S. L.; Vassar, R. Curr. Alzheimer Res. 2008, 5, 100.
3. Ghosh, A. K.; Gemma, S.; Tang, J. Neurotherapeutics 2008, 5, 399.
4. Silvestri, R. Med. Res. Rev. 2009, 29, 295.
5. Sisodia, S. S.; St George-Hyslop, P. H. Nat. Rev. Neurosci. 2002, 3, 281.
6. Eder, J.; Hommel, U.; Cumin, F.; Martoglio, B.; Gerhartz, B. Curr. Pharm. Des.
2007, 13, 271.
7. Hong, L.; Koelsch, G.; Lin, X. L.; Wu, S. L.; Terzyan, S.; Ghosh, A. K.; Zhang, X. C.;
Tang, J. Science 2000, 290, 150.
8. Ekegren, J. K.; Unge, T.; Safa, M. Z.; Wallberg, H.; Samuelsson, B.; Hallberg, A. J.
Med. Chem. 2005, 48, 8098.
The described synthetic route provided two types of masked
tertiary hydroxyl central cores which were smoothly diversified
with different substituents. In total 15 BACE-1 inhibitors were syn-
thesized. The most potent inhibitor 39-(A) exhibited a BACE-1
value of 0.23 lM and was selective towards CatD, but showed only
low cell permeability. This result reveals that this class of inhibi-
tors in the future needs to be further optimized in order to produce
inhibitors in the nanomolar range with good pharmacokinetic
properties.
9. Experimental details and spectroscopic data can be found in the
Supplementary data associated with this article, available online at
sciencedirect.com.
10. Wu, X. Y.; Öhrngren, P.; Ekegren, J. K.; Unge, J.; Unge, T.; Wallberg, H.;
Samuelsson, B.; Hallberg, A.; Larhed, M. J. Med. Chem. 2008, 51, 1053.
11. Russo, F.; Wångsell, F.; Sävmarker, J.; Jacobsson, M.; Larhed, M., submitted for
publication.
12. Bäck, M.; Nyhlen, J.; Kvarnström, I.; Appelgren, S.; Borkakoti, N.; Jansson, K.;
Lindberg, J.; Nyström, S.; Hallberg, A.; Rosenquist, S.; Samuelsson, B. Bioorg.
Med. Chem. 2008, 16, 9471.
13. 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.;
Holloway, M. K.; Munshi, S.; Simon, A. J.; Kuo, L.; Vacca, J. P. J. Med. Chem. 2004,
47, 6447.
Acknowledgments
Dr. Xiongyu Wu for providing compounds 20-(R) and 21-(S). Dr.
Elizabeth Hamerlink and Dr. Ian Henderson for BACE-1 and CatD
enzyme data, Alexandra Johansson, and Elisabet Lilja for excellent
technical assistance in the production of BACE-1, Dr. Anders
Blomqvist for the original cloning of the human gene for BACE-1.
Aleh Yahorau for conducting the HRMS experiments. Dr. Luke Odell
for linguistic improvements. Finally, we would like to acknowledge
the Swedish Research Council, Alice and Knut Wallenberg’s Foun-
dation and Medivir AB for Caco-2 permeability analyses and finan-
cial support.