Discovery of pyrrolidine-based β-secretase inhibitors: Lead advancement through conformational design for maintenance of ligand binding efficiency

Article abstract of DOI:10.1016/j.bmcl.2011.11.024

We have developed a novel series of pyrrolidine derived BACE-1 inhibitors. The potency of the weak initial lead structure was enhanced using library-based SAR methods. The series was then further advanced by rational design while maintaining a minimal ligand binding efficiency threshold. Ultimately, the co-crystal structure was obtained revealing that these inhibitors interacted with the enzyme in a unique fashion. In all, the potency of the series was enhanced by 4 orders of magnitude from the HTS lead with concomitant increases in physical properties needed for series advancement. The progression of these developments in a systematic fashion is described.

Full text of DOI:10.1016/j.bmcl.2011.11.024

Bioorganic & Medicinal Chemistry Letters 22 (2012) 240–244
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of pyrrolidine-based b-secretase inhibitors: Lead advancement
through conformational design for maintenance of ligand binding efficiency
Shawn J. Stachel a, , Thomas G. Steele a, Alessia Petrocchi a, Sharie J. Haugabook d, Georgia McGaughey c,
⇑
M. Katharine Holloway c, Timothy Allison b, Sanjeev Munshi b, Paul Zuck e, Dennis Colussi d,
Katherine Tugasheva d, Abigail Wolfe d, Samuel L. Graham a, Joseph P. Vacca a
a Department of Medicinal Chemistry, Merck Research Laboratories, Merck & Co., PO Box 4, West Point, PA 19486, USA
b Department of Structural Biology, Merck Research Laboratories, Merck & Co., PO Box 4, West Point, PA 19486, USA
c Chemisty Modeling & Informatics, Merck Research Laboratories, Merck & Co., PO Box 4, West Point, PA 19486, USA
d Department of Neuroscience, Merck Research Laboratories, Merck & Co., PO Box 4, West Point, PA 19486, USA
e Screening and Protein Science, Merck Research Laboratories, Merck & Co., PO Box 4, West Point, PA 19486, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
We have developed a novel series of pyrrolidine derived BACE-1 inhibitors. The potency of the weak
initial lead structure was enhanced using library-based SAR methods. The series was then further
advanced by rational design while maintaining a minimal ligand binding efficiency threshold. Ultimately,
the co-crystal structure was obtained revealing that these inhibitors interacted with the enzyme in a
unique fashion. In all, the potency of the series was enhanced by 4 orders of magnitude from the HTS lead
with concomitant increases in physical properties needed for series advancement. The progression of
these developments in a systematic fashion is described.
Received 9 October 2011
Revised 4 November 2011
Accepted 7 November 2011
Available online 12 November 2011
Keywords:
BACE-1
Inhibitor
Ó 2011 Elsevier Ltd. All rights reserved.
Pyrrolidine
Ligand binding efficiency
Rational design
Conformational design
Alzheimer’s disease (AD) is rapidly becoming one of the largest
unmet medical needs in the developed world. This issue is further
exacerbated by the burgeoning population of the age group most at
risk for the disease. More than 5 million Americans over the age of
65 are estimated to be affected, with costs for treatment to be up-
ward of $20 trillion dollars over the next 40 years.1 These facts cou-
pled with the realization that there is currently no disease
modifying treatments available serve as further impetus in the
search for an effective therapeutic agent. As b-secretase (BACE-1)
is responsible for the initial cleavage event in the catabolism of
the amyloid precursor protein (APP) and the foundation of the
amyloid cascade, it has gained a prominent role in the quest for
a disease modifying treatment for AD.2
The first generation of BACE-1 inhibitors were mainly peptidic
in character. The scissile amide bond was replaced with either a
hydroxyethyl amine (HEA), hydroxyethylene (HE), or reduced
amide transition state isostere. Inhibitors based on this design
principle have historically displayed poor pharmacokinetic proper-
ties as well as P-glycoprotein (P-gp) efflux liabilities. In an effort to
address these shortcomings, second generation BACE-1 inhibitors
have centered on novel templates that rely on non-linear struc-
tures to fill the active site.3 To identify novel non-peptidic scaffolds
that have higher potential to achieve acceptable pharmacokinetic
properties researchers have resorted to intensive screening of their
internal compound collections while also leveraging fragment
based screening and other novel methods for lead identification.
We have previously reported a novel series of 2-aminothiazole
BACE-1 inhibitors that was identified using a high concentration
high throughput screen (HTS) of the Merck sample collection.4 In
addition to this aforementioned described hit, 3,4-disubstituted
pyrrolidine 1, as a mixture of trans diastereomers, was also
identified as a low potency inhibitor of BACE-1 (Fig. 1). While the
intrinsic potency of 1 was very weak in the initial enzymatic assay
(IC50 = 240 lM) we viewed the unique pyrrolidine structural scaf-
fold as having potential to achieve a final compound devoid of some
of the physical property limitations present in other inhibitor series.
The synthetic route also allowed for facile access to the pyrrolidine
scaffold using a dipolar cycloaddition reaction of the appropriate
trans-cinnamate ester with an azomethine ylide precursor.5 Using
the HTS hit 1 as a starting point we envisaged that the amide linkage
would serve as an ideal starting point for rapid SAR development
though library-based compound synthesis. However, the SAR
⇑
Corresponding author.
E-mail address: shawn_stachel@merck.com (S.J. Stachel).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.11.024

S. J. Stachel et al. / Bioorg. Med. Chem. Lett. 22 (2012) 240–244
241
O
O
2
O
N
N
O
O
N
IC50 = 47 uM
O
Br
Br
N
H
N
N
H
1
2
H
IC50 = 240 uM
IC50 = 47 uM
O
O
N
O
N
O
N
Br
Br
N
H
N
O
O
H
N
3
S
H
4
5
IC50 = 4 uM
IC50 = 19 uM
IC50 = 57 uM
Figure 1. Elaboration of HTS hit.
proved very limited in this region and the 2-phenyl morpholine
analog (2; IC50 = 47 lM) was found to be the best fit during our
screen. Concomitantly, we explored the effect of halogen substitu-
tion on the western phenyl ring which served two functions. First,
bromine substitution would serve as a synthetic handle with which
we hoped to further optimize the potency. Secondly, we hoped the
electron density of a suitably positioned bromine atom would en-
able a better density fit should a low resolution co-crystal structure
be obtained and would provide a vantage point to guide further ana-
log design. With compound 2 in hand, we subsequently performed a
Suzuki coupling library using the pendant aryl bromide as the func-
tional handle as describedabove. Gratifyingly, compound 3 emerged
from this screen with a 60-fold improvement in potency over the ini-
tial HTS lead 1. Unfortunately, while improvements in intrinsic po-
tency were realized the increased enzyme potency came at the
expense of ligand binding efficiency. While HTS hit 1 possessed a
LBE of 0.19, analog 3 exhibited a slightly lower LBE of 0.17. Since
additional potency would need to be achieved before the series
could be viewed as a serious lead class we viewed this drift in LBE
as a move in the wrong direction and resolved to keep molecular
weight low and focus our efforts with maximal ligand efficiency in
mind going forward.
In the campaign advancing the series to biphenyl analog 3, we
continued to explore the eastern portion of the molecule through
additional amide SAR. As noted above, the SAR was particularly
restricted but we observed that in addition to the 2-substituted mor-
pholine (2), the corresponding carbon analog, that is pyrrolidine 4,
possessed similar potency. Interestingly the 4-phenyl piperidine
analog (5) also emerged from the scan of this region, albeit with an
even more stringent SAR profile (Fig. 2). With these two analogs in
hand we contemplated the effects of conjoining the two substitution
patterns into a single molecular entity. We subsequently synthe-
sized the hybrid cis-2,4-diphenyl analog 6 and were rewarded by a
synergistic increase in potency to 600 nM, representing a 400-fold
increase in potency from the starting HTS hit in addition to
increasing the LBE to 0.25.
O
N
6
Br
IC50 = 0.6 uM
N
H
Figure 2. Synergistic binding potency via SAR combination.
the potency enhancing cis-2,4-bis substituted piperidinyl amide
delineated in the non-fused system, an IC50 of 0.15 M was realized
l
in compound 9 while achieving a further enhanced LBE of 0.28.
More importantly, with compound 9 we were finally successful
in obtaining a co-crystal in the BACE-1 active site.6 Analysis of the
crystallographic data revealed several interesting elements of the
interaction between the inhibitor and the enzyme. First, while
not surprising, it was apparent that the pyrrolidine nitrogen was
positioned between the catalytic dyad interacting directly through
a bidentate interaction with D32 and D228 (Fig. 4). This differs
from previously published reduced amide isostere BACE-1 inhibi-
tors,7 in which the secondary amine hydrogen bonds only to the
‘outside’ oxygen (OD2) of D228.
The tetrahydronapthalene portion of the molecule was found to
occupy the S1 region of the enzyme, while the diphenyl piperidine
resides in an area of the enzyme that is not accessed by the stan-
dard Schecter-Berger peptide subsites.8 Here, the 4-phenyl group
(Fig. 5) was neatly packed into a nascent binding pocket under
the flap and adjacent to the S1 pocket. This is created by an upward
motion of the flap accompanied by an outward swing of the side-
chain of Y71, which normally H-bonds to W76 in the flap-closed
form of BACE. This readily explains the lack of substitution toler-
ated in this compacted region. Also apparent is that the 2-phenyl
piperidine substituent is oriented toward the S2’ pocket, allowing
O
O
N
N
Cognizant of the lessons learned early in the series development,
we did not want to pursue further potency enhancements at the
expense of increasing molecular weight and additional biaryl bonds
that could impart undesirable physical properties. As such, we
approached future target design with a keen eye toward maintain-
ing the ligand binding efficiency (LBE) represented by compound
6. One way to effect increasing enzymatic potency is to reduce the
entropy of binding thereby restricting the inhibitor into its bioactive
conformation. Using this approach we were able to effect a 10-fold
increase in binding potency from compound 7 to the 6-member
N
N
H
H
+
8
7
-
IC50 = 4 uM
IC50 = 19 uM
O
N
N
H
9
IC50 = 0.15 uM
spiro-fused analog 8 which resulted in an IC50 = 4
lM whilst main-
taining a favorable LBE of 0.26 (Fig. 3). Ultimately, by employing
Figure 3. Spirocyclization leads to enhanced potency.

242
S. J. Stachel et al. / Bioorg. Med. Chem. Lett. 22 (2012) 240–244
O
O
N
N
Br
Br
N
H
N
10
11
H
IC50 = 0.07 uM
IC50 = 0.024 uM
O
N
O
N
N
N
Cl
H3C
N
N
H
H
12
13
IC50 = 0.029 uM
IC50 = 0.036 uM
Figure 6. Optimization of spirocyclic ring system.
To relieve the strain from this type of configuration the ring must
adopt a twist-boat conformation as its low energy state. As such,
we conducted an investigation of various 2-piperidine substituents
to exploit the effects of A1-3-strain on the enforcement of the twist
boat conformation in the non-spiro system. We found the 2-cyclo-
hexyl substituent to be the optimal group as far as enforcement of
the twist boat conformation in addition to more efficiently filling
the hydrophobic pocket created in the region.10 We incorporated
the newly optimized 2-cyclohexyl substituent into the spiro-system
Figure 4. Interaction of pyrrolidine with catalytic aspartates.
and synthesized compound 11 which showed an IC50 of 0.024 lM.
Unfortunately, while substantial progress was realized with re-
spect to intrinsic potency toward the enzyme, there was a pro-
nounced potency shift when the compounds were assessed in
our cellular assay. As a matter of reference, while compound 11
exhibited an intrinsic potency of 24 nM, its functional activity in
a cell-based assay was shifted more than 71-fold to 1700 nM. Since
these compounds are very hydrophobic this was not particularly
surprising since initial efforts were grounded in optimization of
the binding which is often most readily accomplished through
hydrophobic interactions.
Having secured a competent and enzymatically potent lead ser-
ies we next sought to introduce a measure of ‘subtle polarity’ in or-
der to decrease the effective shift between the enzymatic and
cellular assay. To accomplish this we introduced a nitrogen atom
into the fused aromatic tetrahydronapthalene ring of 11. Analysis
of the co-crystal structure revealed no significant negative interac-
tion perceived by the introduction of the pyridyl nitrogen. We also
replaced the 2-bromo substituent, with a chlorine atom, as an opti-
mized binding element that would have reduced potential for
untoward electrophilic reactivity. These modifications resulted in
chloropyridine 12 in which the enzymatic potency remained rela-
tively constant at 29 nM, however the corresponding cellular IC50
Figure 5. 4-Phenyl substituent occupies expanded S1 pocket under the flap.
modest changes in structure, with some limitations due to the
vector induced by the ring conformation imposed by the piperidine
(vida infra).
Having obtained the co-crystal structure, we subsequently
‘walked’ a bromide around the constrained system to confirm
whether the binding vectors remained similar to the non-con-
strained system. Through this exercise we were able to effect a fur-
ther 2-fold enhancement of activity to produce compound 10 with
an IC50 = 0.07 lM by the addition of a 4-bromo substituent (Fig. 6).
Surprisingly, the optimal position for bromine placement in the
spiro system was the 4-position in contrast to that of the non-fused
congener that preferred the 2-position. Upon analysis of the
co-crystal structure of 10, the additional potency was easily ratio-
nalized as it fit into a small hydrophobic pocket packing neatly
against the enzyme.
It was also readily apparent from the crystal structure that the
disubstituted piperidine ring occupied a twist-boat conformation.
We had assumed this to be the case prior to obtaining the co-crystal
structure based on the prevailing literature on 2-substituted piper-
idyl amides.9 In these type of systems containing a 2-subtituted
N-acyl piperidine, the prevailing A1,3-strain interaction between
the acyl group and the 2-substituent forces the adjacent 2-substitu-
ent into an axial orientation. With the additional 4-substituent
aligned cis this would result in an unfavorable di-axial arrangement.
O
N
N
Cl
N
H
12
Assay
Value
BACE-1 IC50
sAPPβ_NF IC50
BACE-2 IC50
Renin IC50
29 nM
570 nM
29 nM
a
106000 nM
6212 nM
15 mL/min/kg
44%
Cat D IC50
Clearance (rat)b
%F (rat)c
LBE
0.3
Figure 7. Detailed profile of BACE-1 inhibitor 12. acell-based assay.11 b2 mpk iv
dose. c10 mpk po dose.

S. J. Stachel et al. / Bioorg. Med. Chem. Lett. 22 (2012) 240–244
243
O
O
O
a
c
b
N
H
O
NH2
N
Cl
14
15
16
O
O
EtO2C
OEt
OEt
d
e, f, g
N
N
Cl
Cl
N
N
N
Cl
Bn
18
Boc
17
19
Chiral Separation
O
k, l
h, i, j
O
N
N
N
N
H3C
N
H
Cl
N
20
13
Boc
Scheme 1. Reagents and conditions: (a) methyl propiolate, 170 °C, neat, 33%; (b) POCl3, 62%; (c) triethylphosphonoacetate, NaH, THF, 23% E; (d) N-(methoxymethyl)-N-
(trimethylsilylmethyl)benzylamine, cat. TFA, CH2Cl2, 100% ; (e) 1-chloroethyl chloroformate, CH2Cl2; (f) MeOH; (g) BOC2O, THF, 87% over 3 steps; (h) 2 N NaOH, THF/MeOH,
50 °C; (i) oxalyl chloride, DMF, CH2Cl2; (j) 2-cyclohexyl-4-phenyl-piperidine, triethylamine, CH2Cl2, 70% over 2 steps; (k) NiCl2(dppp), MeMgBr, THF, 43%; (l) TFA, CH2Cl2, 73%.
was improved to 570 nM representing a reduced shift between the
two assays of 19-fold (compared to a 71-fold shift for 11). Com-
pound 12 also had the additional attributes of selectivity toward
other related aspartyl proteases, modest in vivo clearance in rats
(15 mL/min/kg) and a respectable oral bioavailability of 44%
(Fig. 7). Finally, replacement of the chloride with a 2-methyl sub-
stituent provided the more basic pyridyl derivative 13, which again
retained a favorable enzymatic potency with an IC50 = 36 nM, but
now the corresponding cellular potency was even further im-
proved to 100 nM—a nominal 3-fold shift in potency between the
enzymatic and cellular assays.
The synthesis of compound 13 is depicted in Scheme 1.
3-Aminocyclohex-2-en-1-ol (14) was condensed under thermal
conditions with methyl propiolate to provide pyridinone 15.12
Transformation of 15 to chloropyridine 16 was readily accom-
plished through the action of phosphorus oxychloride. Horner-Em-
mons olefination of ketone 16 produced the unsaturated ester 17
as a 3:1 mixture of E to Z isomers that were readily separable by
flash chromatography. Formation of the tetrahydronapthalene
was then be realized via a 1,3-dipolar cycloaddition employing
the azomethine ylide generated from N-(methoxymethyl)-N-(tri-
methylsilylmethyl)benzylamine with TFA as the catalyzing agent.
Typically, cycloadditions of tri-substituted olefins are very difficult
without at least two electron withdrawing groups. We were
pleased to find that compound 17 underwent facile cycloaddition
to give pyrrolidine quinoline 18. In contrast, cycloadditions of ear-
lier congeners were much more sluggish and excess equivalents of
the azomethine ylide were necessary to effect full conversion.
1-Chloroethyl chloroformate mediated N-debenzylation of 18 and
conversion to the corresponding N-Boc derivative resulted in com-
pound 19. This transformation resolved any issues with late-stage
hydrogenative removal of the N-benzyl group in the presence of an
aryl chloride, such as in compound 12. Hydrolysis of the ester, acti-
vation of the carboxyl as the acid chloride, and amide formation
produced the corresponding amide 20. Finally, nickel mediated
methylation followed by Boc deprotection produced the desired
compound 13.
structure was dramatically enhanced using library-based SAR
methods. The series was further advanced by maintaining a mini-
mal ligand binding efficiency threshold in compound design. Ulti-
mately, the co-crystal structure was obtained and the co-crystal
structure revealed that these inhibitors interacted with the en-
zyme in a unique fashion and allowed for potent binding in a
non-traditional fashion. To address pharmacokinetic concerns, a
weakly basic nitrogen was introduced resulting in a potent, cell ac-
tive inhibitor with respectable clearance and bioavailability. In all,
the potency of the series was enhanced by 4 orders of magnitude
from the HTS lead with concomitant increases in physical proper-
ties needed for series advancement. Further optimization and
in vivo activity will be the focus of a future disclosure.
References and notes
1. http://www.alz.org/national/documents/Facts_Figures_2011.pdf
2. Stachel, S. J. Drug Devel. Res. 2009, 70, 100.
3. (a) Barrow, J. C.; Stauffer, S. R.; Rittle, K. E.; Ngo, P. L.; Yang, Z.; Selnick, H. G.;
Graham, S. L.; Munshi, S.; McGaughey, G. B.; Holloway, M. K.; Simon, A. J.; Price,
E. A.; Sankaranarayanan, S.; Colussi, D.; Tugusheva, K.; Lai, M. T.; Espeseth, A. S.;
Xu, M.; Huang, Q.; Wolfe, A.; Pietrak, B. L.; Zuck, P.; Levorse, D. A.; Hazuda, D. J.;
Vacca, J. P. J. Med. Chem. 2008, 52, 6259; (b) Baxter, E. W.; Conway, K. A.; Ludo,
K.; Bischoff, F.; Mercken, M. H.; De Winter, H. L.; Reynolds, C. H.; Tounge, B. A.;
Luo, C.; Scott, M. K.; Huang, Y.; Braeken, M.; Pieters, S. M. A.; Berthelot, J. C.;
Masure, S.; Bruinzeel, W. D.; Jordan, A. D.; Parker, M. H.; Boyd, R. E.; Qu, J.;
Alexander, R. S.; Brenneman, D. E.; Reitz, A. B. J. Med. Chem. 2007, 50, 4261; (c)
Cole, D. C.; Manas, E. S.; Stock, J. R.; Condon, J. S.; Jennings, L. D.; Aulabaugh, A.;
Chopra, R.; Cowling, R.; Ellingboe, J. W.; Fan, K. Y.; Harrison, B. L.; Hu, Y.;
Jacobson, S.; Jin, G.; Lin, L.; Lovering, F. E.; Malamas, M. S.; Stahl, M. L.; Strand,
J.; Sukhdeo, M. N.; Svenson, K.; Turner, M. J.; Wagner, E.; Wu, J.; Zhou, P.; Bard,
J. J. Med. Chem. 2006, 49, 6158.
4. Stachel, S. J.; Coburn, C. A.; Rush, D.; Jones, K. L. G.; Zhu, H.; Rajapakse, H.;
Graham, S. L.; Simon, A.; Holloway, M. K.; Allison, T. J.; Munshi, S. K.; Espeseth,
A. S.; Zuck, P.; Colussi, D.; Wolfe, A.; Pietrak, B. L.; Lai, M. T.; Vacca, J. P. Bioorg.
Med. Chem. Lett. 2009, 19, 2977.
5. (a) Padwa, A.; Dent, W. J. Org. Chem. 1987, 52, 235; (b) Tero, Y.; Kotaki, H.; Imai,
N.; Achiwa, K. Chem. Pharm. Bull. 1985, 33, 2762.
6. A PDB file for the BACE-1/inhibitor complex (PDB identifier 3UFL) has been
deposited with the RCSB Protein Data Bank.
7. 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.; Munshi, S.; Holloway,
M. K.; Espeseth, A.; Simon, A.; Hazuda, D.; Graham, S. L.; Vacca, J. P. Bioorg. Med.
Chem. Lett. 2006, 16, 3635.
In summary, we have developed a novel series of pyrrolidine
derived BACE-1 inhibitors. The potency of the weak initial lead
8. Schechter, I.; Berger, A. Biochem. Biophys. Res. Commun. 1968, 32, 898.
9. Johnson, F. Chem. Rev. 1968, 68, 375.

244
S. J. Stachel et al. / Bioorg. Med. Chem. Lett. 22 (2012) 240–244
10. Unpublished results to be elaborated in a future disclosure
Chen, L.; Lai, M.-T.; Pietrak, B.; DiMuzio, J.; Li, Y.; Xu, M.; Huang, Q.; Garsky, V.;
Sardana, M. K.; Hazuda, D. J. J. Alz. Res. 2005, 7, 139.
12. Pettit, G. R.; Fleming, W. C.; Paull, K. D. J. Org. Chem. 1968, 33, 1089.
11. Shi, X.-P.; Tugusheva, K.; Bruce, J. E.; Lucka, A.; Chen-Dodson, E.; Hu, B.; Wu, G.-
X.; Price, E.; Register, R. B.; Lineberger, J.; Miller, R.; Tang, M.-J.; Espeseth, A.;
Kahana, J.; Wolfe, A.; Crouthamel, M.-C.; Sankaranarayanan, S.; Simon, A.;