KMI-358 and KMI-370, highly potent and small-sized BACE1 inhibitors containing phenylnorstatine

Article abstract of DOI:10.1016/j.bmcl.2003.12.088

Recently, we reported a novel substrate-based octapeptide BACE1 inhibitor KMI-008 containing hydroxymethylcarbonyl (HMC) isostere as a transition-state mimic. Using KMI-008 as a lead compound, a small-sized and highly potent BACE1 inhibitor KMI-370 (IC₅₀=3.4 nM) was designed and synthesized.

Full text of DOI:10.1016/j.bmcl.2003.12.088

Bioorganic & Medicinal Chemistry Letters 14 (2004) 1527–1531
KMI-358 and KMI-370, highly potent and small-sized BACE1
inhibitors containing phenylnorstatine
Tooru Kimura,a Daisuke Shuto,a Soko Kasai,a Ping Liu,a Koushi Hidaka,a
Takashi Hamada,a Yoshio Hayashi,a Chinatsu Hattori,b Masashi Asai,b
Shinobu Kitazume,c Takaomi C. Saido,d Shoichi Ishiurab and Yoshiaki Kisoa,*
aDepartment of Medicinal Chemistry, Center for Frontier Research in Medicinal Science, Kyoto Pharmaceutical University,
Yamashina-ku, Kyoto 607-8412, Japan
bDepartment of Life Sciences, Graduate School of Arts and Sciences, University of Tokyo, Meguro-ku, Tokyo 153-8902, Japan
cGlyco-chain Functions Laboratory, Supra-biomolecular System Group, Frontier Research System, RIKEN, Wako-shi,
Saitama 351-0198, Japan
dLaboratory for Proteolytic Neuroscience, RIKEN Brain Science Institute, Wako-shi, Saitama 351–0198, Japan
Received 1 December 2003; accepted 26 December 2003
Abstract—Recently, we reported
a novel substrate-based octapeptide BACE1 inhibitor KMI-008 containing hydroxy-
methylcarbonyl (HMC) isostere as a transition-state mimic. Using KMI-008 as a lead compound, a small-sized and highly potent
BACE1 inhibitor KMI-370 (IC50=3.4 nM) was designed and synthesized.
# 2004 Elsevier Ltd. All rights reserved.
1. Introduction
On the basis of knowledge of our HIV-1 protease inhibitor
studies,11,12 we recently reported an octapeptide BACE1
inhibitor 1 (KMI-008) containing an unnatural amino
acid, phenylnorstatine [(2R,3S)-3-amino-2-hydroxy-4-
phenylbutyric acid: Pns], with a hydroxymethylcarbonyl
(HMC) isostere as a transition-state mimic (Fig. 1).13
KMI-008 showed potent inhibition against recombinant
enzyme (IC50=413 nM) and decreased the secretion of
sAPPb (a soluble form of APP generated by the action
of b-secretase) from COS-7 cells co-transfected with
both the APP and BACE1. However, the characteristics
of KMI-008 were not suitable as a drug due to its large
molecular weight, the presence of many natural peptide
bonds, and no protection against exopeptidase. In this
paper, we describe the design and synthesis of truncated
BACE1 inhibitors which are more potent than KMI-008.
Cerebral deposition of amyloid fibrils as senile plaques
is a pathological hallmark of Alzheimer’s disease (AD).1
The predominant constituents of amyloid fibrils are 40-
and 42-residue amyloid b peptides (Ab). Ab is generated
from proteolytic processing of a transmembrane protein,
b-amyloid precursor protein (APP).2,3 Two proteases,
called b- and g-secretases, cleave APP to generate the
Ab peptide. b-Secretase which forms the N-terminus of
Ab has been identified as a novel membrane-bound
aspartic protease, BACE1 (memapsin 2 or Asp-2).4ꢀ7
BACE1 plays a critical role in the progression of AD
since the cleavage of APP by b-secretase is the first step
in Ab formation. Therefore, the development of BACE1
inhibitor is valuable for the elucidation of AD pathol-
ogy. Several transition-state analogue BACE1 inhibitors
modelled on the b-secretase cleavage site have been
reported with relatively low IC50 values.8ꢀ10 However,
the molecular sizes of these inhibitors are too large to be
viable drug candidates.
* Corresponding author. Tel.: +81-75-595-4635; fax: +81-75-591-
9900; e-mail: kiso@mb.kyoto-phu.ac.jp
Figure 1. Structure of octapeptide BACE1 inhibitor, 1 (KMI-008).
0960-894X/$ - see front matter # 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.12.088

1528
T. Kimura et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1527–1531
with Boc-Dap(Fmoc)-OH and ethyl oxalyl chloride.
After assembly of the peptide-chain, saponification and
following 4M HCl/dioxane treatment gave the desired
product.
3. Results and discussions
In order to determine the minimum structure having
BACE1 inhibitory activity, peptide ladders were
designed by the removal of amino acids from KMI-008
(1) (Table 1). As shown in Table 1, the compounds
lacking P4-P2 positions (2–4) showed no inhibitory
activities at 2 mM. However, the removal of C-terminal
residues had mild influence on the activity (5–8). Espe-
cially, compound 7 had slightly weaker activity (61%)
than KMI-008 (1). These results indicate that the inter-
actions of N-terminal residues were very important
while those of the C-terminus were not significant. They
also suggested that the inhibitors consisting of P4–P10
positions were the minimal requirement for the activity.
.
.
Scheme 1. Reagents and conditions: (a) EDC HCl, HOBt H2O/DMF;
(b) 4 M HCl/dioxane; (c) 20% Et2NH/DMF; (d) Et3N/DMF; (e) 4 M
KOH/DMF.
2. Synthesis
Compounds containing a C-terminal carboxylic acid
were synthesized by the 9-fluorenylmethoxycarbonyl
(Fmoc)-based solid-phase peptide synthesis. For start-
ing resins, p-benzyloxybenzyl alcohol resins or 2-chlor-
otrityl chloride resins coupled with C-terminal Fmoc-
amino acid derivatives were employed. Each synthetic
cycle consisted of the following steps: (1) deprotection
with 20% piperidine in DMF, (2) a coupling with
Fmoc-amino acid derivative using diisopropylcarbodi-
imide (DIPCDI) and 1-hydroxybenzotriazole (HOBt).
In the case of the synthesis of 26 and 27, the P4 ligand
was incorporated by sequential coupling of Na-Boc-Nb-
Fmoc-l-2,3-diaminopropionic acid [Boc-Dap(Fmoc)-
OH] and oxalic acid mono-t-butyl ester. After the pep-
tide-chain elongation, the peptide was cleaved from the
resin by the treatment with TFA in the presence of m-
cresol and thioanisole as scavengers. The product was
finally purified by preparative RP-HPLC and identified
by MALDI-TOF MS. The inhibitory activity of the
synthetic compounds against rh-BACE1 was examined
as described previously.13
To confirm this finding, we attempted the Gly-scan of
each amino acid consisting of octapeptide inhibitor 1
(Table 1). The compounds having a Gly-substitution at
P4-P2 (9–11) showed no inhibition, but compounds 12–
15 substituted at P10–P40 were active (54–73%). It is
clear that the side chains of P4–P2 residues have sig-
nificant interactions with S4–S2 subsites of BACE1.
This is consistent with the previous observation that the
side chain of P30–P40residues are not involved in any
specific interaction in S30–S40 subsites.14
Next, we attempted to optimize the structures of each
position of the inhibitors consisting of P4–P10. The
studies of structure–activity relationship (SAR) focused
on P2 or P3 have already been performed on the devel-
opment of KMI-008 (1), and leucine at P2 and valine at
P3 were the best among the tested compounds.13
Hence, SAR at P10 and P4, corresponding to both
the N- and C-terminals, was carried out to increase the
potency and stability of the inhibitors under physiological
conditions.
Compounds 16, 21 and 28 were synthesized by tradi-
tional solution method. For example, the synthetic route
to 28 is shown in Scheme 1. The 1-(3-dimethylamino-
.
propyl)-3-ethylcarbodiimide hydrochloride (EDC HCl)–
HOBt method was employed for each coupling and the
Na-Boc group was removed with 4M HCl/dioxane. The
P4 ligand was incorporated by two steps of couplings
Table 1. Ladders of BACE 1 inhibitor and Gly-scan
0
0
0
0
P4
Compd
(KMINo)
P4
P3
P2
P1
P1
P2
P3
BACE1
inhibition (%)a
1 (ꢀ008)
2 (ꢀ070)
3 (ꢀ196)
4 (ꢀ251)
5 (ꢀ069)
6 (ꢀ098)
7 (ꢀ122)
8 (ꢀ204)
9 (ꢀ193)
10 (ꢀ194)
11 (ꢀ195)
12 (ꢀ249)
13 (ꢀ250)
14 (ꢀ252)
15 (ꢀ253)
H–
Glu
H
–Val
–Val
H
–Leu
–Leu
–Leu
H
–Pns
–Pns
–Pns
–Pns
–Pns
–Pns
–Pns
–Pns
–Pns
–Pns
–Pns
–Pns
–Pns
–Pns
–Pns
–Asp
–Asp
–Asp
–Asp
–Asp
–Asp
–Asp
–OH
–Asp
–Asp
–Asp
–Gly
–Asp
–Asp
–Asp
–Ala
–Ala
–Ala
–Ala
–Ala
–Ala
–OH
–Glu
–Glu
–Glu
–Glu
–Glu
–OH
–Phe
–Phe
–Phe
–Phe
–OH
–OH
–OH
–OH
–OH
>90
<20
<20
<20
60
46
61
34
<20
<20
<20
54
73
56
H–
H–
H–
H–
H–
H–
H–
H–
H–
H–
H–
Glu
Glu
Glu
Glu
Gly
Glu
Glu
Glu
Glu
Glu
Glu
–Val
–Val
–Val
–Val
–Val
–Gly
–Val
–Val
–Val
–Val
–Val
–Leu
–Leu
–Leu
–Leu
–Leu
–Leu
–Gly
–Leu
–Leu
–Leu
–Leu
–Ala
–Ala
–Ala
–Ala
–Gly
–Ala
–Ala
–Glu
–Glu
–Glu
–Glu
–Glu
–Gly
–Glu
–Phe
–Phe
–Phe
–Phe
–Phe
–Phe
–Gly
–OH
–OH
–OH
–OH
–OH
–OH
–OH
65
a At 2 mM.

T. Kimura et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1527–1531
1529
To study the SAR at the P10 position, we selected com-
pound 7 as a lead compound and prepared compounds
containing different moieties at P10 (Table 2). As shown
in Table 2, compound 16 exhibited weak activity while
compounds 17–20, containing benzoic acid analogues,
were moderately active (52–82% inhibition at 2 mM).
These data revealed that the carboxyl group at the P10
position gave a suitable interaction with the S10 pocket
of BACE1. Interestingly, the phenylamide analogues 19
and 20 were more potent than the benzylamide analo-
gues 17 and 18. Thus, the rigid conformation of the
phenyl ring at the P10 position may be preferred for
binding to the enzyme. Furthermore, compound 21
bearing two carboxyl groups at the P10 position exhib-
ited more potent inhibitory activity (IC50=55 nM) than
KMI-008 (1). It was supposed that the remarkable
increase was caused by the interaction of the P10 moiety
with both active site and flap region of BACE1.
P4 position also has an important effect on the activity
(5 versus 24 and 25). The aspartic acid analogue 24
showed no activity against BACE1, but homoglutamic
acid analogue 25 had slightly decreased inhibitory
activity. Therefore, we designed analogue 26, which has
two carbonyl groups at the side chain corresponding to
the glutamic acid and homoglutamic acid. Surprisingly,
Table 3. BACE1 inhibitory activity (P4 position)
Compd
(KMINo)
R
BACE1 inhibition
(%) at 2 mM
The SAR study focused on P4 was carried out using
heptapeptide inhibitor 5 as a lead compound, which had
moderate potency (Table 3). The inhibitory activity of 5,
22 and 23 indicated that the amino and carboxyl groups
at the P4 position significantly affected the BACE1
binding. In addition, the length of the side chain at the
5 (–069)
60
Table 2. BACE1 inhibitory activity (P10 position)
22 (–103)
23 (–124)
24 (–262)
25 (–322)
<20
<20
<20
43
Compd
(KMINo)
R
BACE1 inhibition
(%) at 2 mM
7 (–122)
61
16 (–205)
17 (–158)
30
52
26 (–300)
>90 (IC50=84 nM)
18 (–293)
19 (–287)
20 (–260)
21 (–360)
64
79
82
>90 (IC50=55 nM)
Figure 2. Structures and BACE1 inhibitory activities of KMI-358 (27)
and KMI-370 (28).

1530
T. Kimura et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1527–1531
Figure 3. Modelled structure of KMI-370 bound to BACE1 and possible hydrogen bonding pattern. The model was constructed using the crystal
structure of BACE1 bound to OM99-2.14
compound 26, which contains Nb-oxalyl-l-2,3-diami-
nopropionic acid, resulted in a further 2–3-fold increase
in potency (IC50=84 nM). This result may be due to the
increase of affinity for the hydrophilic S4 subsite of
enzyme by the two carbonyl groups.
acids, it was expected that KMI-370 would be more
stable to the decomposition by exopeptidase in vivo
than KMI-008 with natural amino acids.
Finally, we examined the effects of KMI-358/-370 on b-
secretase activity in HEK293 cells transfected with
BACE1 (BACE1-HEK293), which are stably expressed
both full length-BACE1 and native APP.15 The secre-
tion of sAPPb in the culture medium was detected using
anti-sAPPb antibody (Fig. 4). KMI-358 and KMI-370
dose-dependently inhibited the secretion of sAPPb from
BACE1-HEK293 cells. The IC50 values of KMI-358
and KMI-370 were 0.65 and 0.20 mM, respectively.
These results demonstrate the potential of our inhibitors
containing HMC isostere as drugs to attenuate Ab-for-
mation in vivo.
From the results of the above SAR studies, we designed
pentapeptide BACE1 inhibitors 27 (KMI-358) and 28
(KMI-370) consisting of P4–P10 moieties (Fig. 2). They
had Nb-oxalyl-l-2,3-diaminopropionic acid at P4 and 3-
aminobenzoic acid or 5-aminoisophthalic acid at the P10
position. Of the tested compounds, KMI-370 (28) was
the most potent inhibitor against BACE1 with an IC50
value of 3.4 nM. In our assay, the IC50 value of KMI-
370 is approximately 20- and 10-fold smaller than those
10
(IC50=65 nM) and OM99-29
0
of P10–P4 statV
(IC50=28 nM), respectively. The modelled structure of
KMI-370 bound to BACE1 (Fig. 3) was prepared from
1FKN (pdb data).14 KMI-370 could fill the active site-
pockets of BACE1 tightly. The carboxylic acids at the
P4 and P10 positions could provide several hydrogen
bonds which contribute to the binding affinity. Through
our SAR study, we found that the pentapeptide inhi-
bitor KMI-370, which had a relatively low molecular
weight (Mw=728), was a very potent inhibitor. Since
both terminals of KNI-370 consist of unnatural amino
4. Conclusion
In conclusion, we have established that a pentapeptide
containing phenylnorstatine is the minimal molecular
size for BACE1 inhibition. Moreover, we have found a
highly potent inhibitor, KMI-370, with an IC50 value of
3.4 nM, which can reduce the secretion of sAPPb from
BACE1-HEK293 cells dose-dependently.
Acknowledgements
This research was supported in part by the Frontier
Research Program of the Ministry of Education,
Science and Culture of Japan, and grants from the
Ministry of Education, Science and Culture of Japan.
The authors thank Dr. Y. Hashimoto (RIKEN) and
Ms. S. Shibakawa for their help in the inhibition assay.
References and notes
1. Selkoe, D. J. Neuron 1991, 6, 487.
2. Selkoe, D. J. Nature 1999, 399, A23.
3. Sinha, S.; Lieberburg, I. Proc. Natl. Acad. Sci. U.S.A.
1999, 96, 11049.
4. Vassar, R.; Bennett, B. D.; Babu-Khan, S.; Kahn, S.;
Mendiaz, E. A.; Denis, P.; Teplow, D. B.; Ross, S.;
Figure 4. Effects of KMI-358/-370 on b-secretase activity in BACE1-
transfected HEK293 cells (n=3).

T. Kimura et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1527–1531
1531
Amarante, P.; Loeloff, R.; Luo, Y.; Fisher, S.; Fuller, J.;
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.
J. A.; Nguyen, C.; Koelsch, G.; Ermolieff, J.; Tang, J. J.
Med. Chem. 2001, 44, 2865.
9. Tung, J. S.; Davis, D. L.; Anderson, J. P.; Walker, D. E.;
Mamo, S.; Jewett, N.; Hom, R. K.; Sinha, S.; Thorsett,
E. D.; John, V. J. Med. Chem. 2002, 45, 259.
5. Yan, R.; Bienkowski, M. J.; Shuck, M. E.; Miao, H.;
Tory, M. C.; Pauley, A. M.; Brashier, J. R.; Stratman,
N. C.; Mathews, W. R.; Buhl, A. E.; Carter, D. B.;
Tomasselli, A. G.; Parodi, L. A.; Heinrikson, R. L.;
Gurney, M. E. Nature 1999, 402, 533.
6. Sinha, S.; Anderson, J. P.; Barbour, R.; Basi, G. S.;
Caccavello, R.; Davis, D.; Doan, M.; Dovey, H. F.;
Frigon, N.; Hong, J.; Jacobson-Croak, K.; Jewett, N.;
Keim, P.; Knops, J.; Lieberburg, I.; Power, M.; Tan, H.;
Tatsuno, G.; Tung, J.; Schenk, D.; Seubert, P.;
Suomensaari, S. M.; Wang, S.; Walker, D.; Zhao, J.;
McConlogue, L.; John, V. Nature 1999, 402, 537.
10. Tamamura, H.; Kato, T.; Otaka, A.; Fujii, N. Org.
Biomol. Chem. 2003, 1, 2468.
11. Mimoto, T.; Imai, J.; Tanaka, S.; Hattori, N.; Takahashi,
O.; Kisanuki, S.; Nagano, Y.; Shintani, M.; Hayashi, H.;
Akaji, K.; Kiso, Y. Chem. Pharm. Bull. 1991, 39, 2465.
12. Mimoto, T.; Imai, J.; Tanaka, S.; Hattori, N.; Kisanuki,
S.; Akaji, K.; Kiso, Y. Chem. Pharm. Bull. 1991, 39, 3088.
13. Shuto, D.; Kasai, S.; Kimura, T.; Liu, P.; Hidaka, K.;
Hamada, T.; Shibakawa, S.; Hayashi, Y.; Hattori, C.;
Szabo, B.; Ishiura, S.; Kiso, Y. Bioorg. Med. Chem. Lett.
2003, 13, 4273.
7. Hussain, I.; Powell, D.; Howlett, D. R.; Tew, D. G.;
Meek, T. D.; Chapman, C.; Gloger, I. S.; Murphy, K. E.;
Southan, C. D.; Ryan, D. M.; Smith, T. S.; Simmons,
D. L.; Walsh, F. S.; Dingwall, C.; Christie, G. Mol. Cell.
Neurosci. 1999, 14, 419.
8. (a) Ghosh, A. K.; Shin, D.; Downs, D.; Koelsch, G.; Lin,
X.; Ermolieff, J.; Tang, J. J. Am. Chem. Soc. 2000, 122,
3522. (b) Ghosh, A. K.; Bilcer, G.; Harwood, C.;
Kawahara, R.; Shin, D.; Hussain, K. A.; Hong, L.; Loy,
14. Hong, L.; Koelsch, G.; Lin, X.; Wu, S.; Terzyan, S.;
Ghosh, A. K.; Zhang, X. C.; Tang, J. Science 2000, 290, 150.
15. BACE1-HEK293 cells (80–100% confluent) were replaced
with new serum-free medium (DMEM containing 1 mg/
mL G418) with 100 mM/10 mM/1 mM/100 nM of KMI-
358 or KMI-370, followed by cell incubation for 6 h. The
sAPPb in the culture was concentrated, ran on 7.5% SDS-
PAGE and western blotted. The blots were proved with
anti-sAPPb antibody.