Methyl-substitution of an iminohydantoin spiropiperidine β-secretase (BACE-1) inhibitor has a profound effect on its potency

Article abstract of DOI:10.1016/j.bmcl.2015.06.082

The IC₅₀ of a beta-secretase (BACE-1) lead compound was improved ～200-fold from 11 μM to 55 nM through the addition of a single methyl group. Computational chemistry, small molecule NMR, and protein crystallography capabilities were used to com

Full text of DOI:10.1016/j.bmcl.2015.06.082

Accepted Manuscript
Methyl-substitution of an iminohydantoin spiropiperidine β-secretase
(BACE-1) inhibitor has a profound effect on its potency
Melissa Egbertson, Georgia B. McGaughey, Steven M. Pitzenberger, Shaun R.
Stauffer, Craig A. Coburn, Shawn J. Stachel, Wenjin Yang, Jim C. Barrow, Lou
Anne Neilson, Melody McWherter, Debbie Perlow, Bruce Fahr, Sanjeev
Munshi, Timothy Allison, Katharine Holloway, Harold G. Selnick, ZhiQiang
Yang, John Swestock, Adam J. Simon, Sethu Sankaranarayanan, Dennis
Colussi, Katherine Tugusheva, Ming-Tain Lai, Beth Pietrak, Shari Haugabook,
Lixia Jin, I.-W. Chen, Marie Holahan, Maria Stranieri-Michener, Jacquelynn J.
Cook, Joseph Vacca, Samuel L. Graham
PII:
S0960-894X(15)00687-3
DOI:
http://dx.doi.org/10.1016/j.bmcl.2015.06.082
BMCL 22872
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
2 May 2015
22 June 2015
23 June 2015
Please cite this article as: Egbertson, M., McGaughey, G.B., Pitzenberger, S.M., Stauffer, S.R., Coburn, C.A.,
Stachel, S.J., Yang, W., Barrow, J.C., Neilson, L.A., McWherter, M., Perlow, D., Fahr, B., Munshi, S., Allison, T.,
Holloway, K., Selnick, H.G., Yang, Z., Swestock, J., Simon, A.J., Sankaranarayanan, S., Colussi, D., Tugusheva,
K., Lai, M-T., Pietrak, B., Haugabook, S., Jin, L., Chen, I.-W., Holahan, M., Stranieri-Michener, M., Cook, J.J.,
Vacca, J., Graham, S.L., Methyl-substitution of an iminohydantoin spiropiperidine β-secretase (BACE-1) inhibitor
has a profound effect on its potency, Bioorganic & Medicinal Chemistry Letters (2015), doi: http://dx.doi.org/
10.1016/j.bmcl.2015.06.082
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Graphical Abstract
To create your abstract, type over the instructions in the template box below.
Fonts or abstract dimensions should not be changed or altered.
Methyl-substitution of an iminohydantoin
spiropiperidine β-secretase (BACE-1)
inhibitor has a profound effect on its potency
Leave this area blank for abstract info.
Melissa Egbertson, Georgia B. McGaughey, Steven M. Pitzenberger, Shaun R. Stauffer, Craig A. Coburn,
Wenjin Yang, Shawn Stachel, Lou Anne Neilson, Melody McWherter, Debra Perlow, Bruce Fahr, Sanjeev
Munshi, Timothy Allison, Katharine Holloway, Harold G. Selnick, James C. Barrow, ZhiQiang Yang, John
Swestock, Adam J. Simon, Sethu Sankaranarayanan, Dennis Colussi, Katherine Tugusheva, Ming-Tain Lai,
Beth Pietrak, Shari Haugabook, Lixia Jin, I.-W. Chen, Marie Holahan, Maria Stranieri-Michener, Jacquelynn J.
Cook, Joseph Vacca, Samuel L. Graham

Bioorganic & Medicinal Chemistry Letters
journal homepage: www.els evier.com
Methyl-substitution of an iminohydantoin spiropiperidine β-secretase (BACE-1)
inhibitor has a profound effect on its potency
Melissa Egbertsona@*, Georgia B. McGaugheyb@, Steven M. Pitzenbergerc, Shaun R. Stauffera@, Craig A.
Coburna@, Shawn J. Stachela, Wenjin Yangd@, Jim C. Barrowa@, Lou Anne Neilsona@, Melody
McWhertera@, Debbie Perlowa@, Bruce Fahrd@, Sanjeev Munshie@, Timothy Allisone@, Katharine
Hollowayb, Harold G. Selnicka@, ZhiQiang Yanga, John Swestockf@, Adam J. Simong@, Sethu
Sankaranarayanang@, Dennis Colussig@, Katherine Tugushevag@, Ming-Tain Laig@, Beth Pietrakg@, Shari
Haugabookg@, Lixia Jinh@, I.-W. Chenh@, Marie Holahani, Maria Stranieri-Micheneri, Jacquelynn J.
Cooki@, Joseph Vaccaa@, Samuel L. Grahama@.
aMedicinal Chemistry Department, WP14-2 Merck and Co. West Point, PA 19486, USA
bMolecular Modeling, WP53-3 Merck and Co. West Point, PA 19486, USA
c NMR Structure Elucidation, Process and Analytical Chemistry, WP14-1 Merck and Co. West Point, PA 19486, USA
dSunesis Pharmaceuticals, 395 Oyster Point Blvd. Ste. 400 South San Francisco, CA 94080, USA
eStructural Biology, WP 14-2 Merck and Co. West Point, PA 19486, USA
fProcess Chemistry, WP 14-1 Merck and Co. West Point, PA 19486, USA
gPharmacology, WP 26-1 Merck and Co. West Point, PA 19486, USA
hDrug Metabolism, WP 75-B Merck and Co. West Point, PA 19486, USA
iImaging Research, WP 44c Merck and Co. West Point, PA 19486, USA
@for current address, see reference
1
ARTICLE INFO
ABSTRACT
Article history:
Received
Revised
The IC50 of a beta-secretase (BACE-1) lead compound was improved ~200-fold from 11 µM to
55 nM through the addition of a single methyl group. Computational chemistry, small molecule
NMR, and protein crystallography capabilities were used to compare the solution conformation
of the ligand under varying pH conditions to its conformation when bound in the active site.
Accepted
Available online
Chemical modification then explored available binding pockets adjacent to the ligand.
A
strategically placed methyl group not only maintained the required pKa of the piperidine
nitrogen and filled a small hydrophobic pocket, but more importantly, stabilized the
conformation best suited for optimized binding to the receptor.
Keywords:
Amyloid
2009 Elsevier Ltd. All rights reserved
.
beta secretase
spiropiperidine
magic methyl effect
12
Beta-secretase (BACE-1) is
a
transmembrane aspartyl
quite hydrophilic.11
Previous leads had tended toward high
protease that has served as a promising therapeutic target for the
treatment of Alzheimer’s disease (AD) for many years.
molecular weight (>500 Dalton), high polar surface area (>100
(2-5) 2 3 4 5
14
Å3), and strong P-gp transport susceptibility.13 Therefore the
BACE-1 is an attractive target as its inhibition results in a
reduced production of Aβ40-42 thereby decreasing the amount of
amyloid accumulation, which appears to be one of several critical
steps in the process of neuronal degeneration. 3 6 7 Currently, there
are several BACE-1 inhibitors in late-stage clinical trials that will
shed further light on whether a reduction of amyloid burden will
result in an overall improvement in cognitive function.8 Since
BACE-1 is primarily located in the brain, a small molecule
discovery of the BACE-selective spiropiperidine lead 1 (Table 1,
BACE-1 IC50 22 µM, Cathepsin D IC50 = >1.6 µM, MW= 416,
PSA = 47.3) from a high-throughput screen of the Merck sample
collection offered a unique opportunity to explore a compound
with inherently superior baseline physical characteristics.
In-house optimization of 115 led to compounds 2 and 3 (Table
1, BACE-1 IC50 = 11 and 2.8 µM, respectively) that also
demonstrated good selectivity over Cathepsin D (Cathepsin D
IC50 = 220 µM, >505 µM respectively). The first X-ray crystal
structure of a spiropiperidine (compound 3)16 bound to the
inhibitor would ideally need to possess the appropriate
10
physicochemical properties for CNS penetration.9
This is a
daunting task as the BACE-1 active site is large (>1000 Å3) and

BACE-1 enzyme revealed a hitherto undiscovered binding mode
for a BACE-1 inhibitor. In the co-crystal structure, compound 3
was observed to interact with the catalytic aspartic acids via
hydrogen bonding to intermediary water molecules, rather than
by direct contact. This method of binding induced a conformation
change in the flexible “flap” hairpin loop (enzyme residues Val69
to Trp76), significantly expanding the volume of the active site
pocket.
because of the poor apparent permeability (Papp). Other means
were sought to improve potency while improving ligand
efficiency and physical properties, with an inhibitor of low
molecular weight as a key goal. It was proposed that a detailed
characterization of the binding and native conformational
preferences of the iminohydantoin-spiropiperidine structure
would provide new ways to approach the problem.
Figure 1. X-ray co-crystal structure of BACE-1 and spiropiperidine
analog 5. The surface of the “flap” residue Gln73 (gray stick) and water
molecules not associated with the catalytic aspartic acids have been removed
for clarity.
The numerous (> 170) publicly available crystal structures of
inhibitor/BACE-1 complexes demonstrate that the active site is
inherently flexible, accommodating a wide variety of ligands.17
Most published reports have relied on structure based drug
design for optimization toward a static site pose. In contrast, this
paper describes a more iterative approach, in which information
from X-ray crystallography influenced investigations of ligand
conformational preferences in the absence of the enzyme, which
in turn guided new analog designs. Interestingly, and presumably
as a result of the interplay between enzyme pocket environment
and ligand binding conformation, the ligand adopts a chair
conformation with an intramolecular steric interaction, in which
the fluorophenyl ring is positioned in close proximity to the
piperidine ring. MM2 modeling of this conformer showed it is of
higher energy than other conformations available to the system
(twist boats, other chairs).18 Additional analogs were sought to
explore the series further, with both potency and the capacity to
serve as tools for X-ray, modeling and NMR as objectives.
From the X-ray structure of 3 bound to the BACE-1 enzyme it
was apparent that the S3 sub-site was not optimally occupied and
may provide an area to obtain additional enzyme-binding
interactions. This was accomplished through the addition of a
meta-biphenyl tolyl group (compound 4, Table 1), where potency
was enhanced 30-fold compared to 3. Unfortunately addition of
the lipophilic aromatic group increased the molecular weight to
524, increased clogP to 5.15, and negatively affected solubility.
While an improvement in ligand binding efficiency (LBE) was
observed between 3 and 4 it is readily apparent that the potency
gains were driven solely through lipophilic interactions as
depicted by the drop in lipophilic ligand efficiency (LLE) value
from 3 to 4. This high lipophilicity was reflected in the large
shift between the enzymatic and cell-based assay where only
moderate cellular activity was observed (sAPPb IC50 =5.2 µM)
despite the enzymatic activity of 110 nM.
X-ray structures of 5 and 6 co-crystallized with BACE-1 were
obtained. As had been observed previously for compound 315, the
X-ray co-crystal of compound 5 16 revealed the piperidine ring of
the spiropiperidine in a sterically-congested chair conformation
(Figure 1).
Solution state NMR studies19 were done to determine
conformational preference(s) of the spiropiperidines in the
absence of enzyme binding. Energetically efficient binding is
expected when the solution conformation matches the bound
conformation. The free base of 7 was first studied in CD2Cl2
solution. 20 Neutral 7 was found by NOE experiments to have a
preferred conformation in which the exo-cyclic imino hydantoin
hydrogen is proximal to the benzylic CH2 and aromatic ring.
These NOEs strongly suggest a hydrogen bond between the
labile imino NH and the basic amine of the piperidine ring.
Vicinal couplings provide additional clues as to the dominant
conformer(s). Due to symmetry, there are four unique vicinal
couplings between the piperidine ring protons. All four were
found to be in the 6-7 Hz range. There is not a single
conformation that satisfies all four coupling restraints.
Considering that the major conformer(s) must also satisfy the NH
NOE restraints, it is clear that the dominant conformation is
actually an equilibrium of the two twist-boat forms of A (Figure
2). One cannot exclude the possibility that the conformer mix
also includes small populations of structures B and/or the free
base form of conformer C.
Table 1. Spiropiperidine BACE inhibitors 1, 2, 3 and 4.
Ideally, additional potency would be achieved with a decrease
in molecular weight and lipophilicity (and be reflected in an
improvement in LBE and LLE). A survey of small polar groups
capable of occupying the S3 sub-site was undertaken (Table 1).
Isopropoxy compound 5 possessed improved potency (enzyme
and cell assays) and lower P-gp susceptibility as compared to 3,
and lower molecular weight, clogP, and improved permeability
(as measured by Papp) and solubility compared to 4. Although
the intrinsic potency of 5 was less than 4, the cell potency was
maintained. The lipophilic ligand efficiency measurement
(LLE), in particular, reflected the improvements from 4 to 5.
Brain/plasma PK experiments in mice showed a ratio of ~25%
for compounds 4 and 5, which was also encouraging.
Compound 7 was also studied in solution under acidic
conditions in order to understand the influence of protonation by
the acidic catalytic site of BACE-1. In the presence of acid, 7
exclusively adopts the more sterically demanding chair form C,
in which the fluorophenyl group is situated above the piperidine
ring. Protonation of the amine prevents hydrogen bonding
between the piperidine amine and the hydantoin amino NH. π-
stacking of the fluorophenyl and benzyl aromatic rings may also
play a role in stabilizing this conformation since a weak NOE
To decrease the overall lipophilicity of the structure a basic
amine was added to 5 to give racemic 6. Compound 6 was more
potent than 3 or 5 in both in vitro assays, and had significantly
improved plasma free fraction, clogP and cell potency (first sub-
micro molar compound) compared to 4. Unfortunately P-gp
susceptibility was increased, and is probably underestimated

was observed between the protons at the para positions of the
benzyl aromatic ring and the fluorophenyl ring.
The target compounds were prepared using Ugi reaction
21
conditions (similar to those already described15 starting with
the racemic 1-benzyl-2-methyl-4-ketopiperidine to give major
and minor diastereomeric reaction products in a ratio of 10:1 and
overall yield of 10%. The major isomer was found to be the cis
analog 8 (where the methyl and aza-fluorophenyl of the
iminohydantoin are in a cis orientation) and the minor trans 9
(Table 2).
Figure 2. NMR studies of analog 7, characterization of the
The solution conformations of 8 and 9 were studied by NMR.
The major cis diastereomer 8 was found to exist under neutral
conditions as a mixture of conformers. The dominant conformer
is a boat where the methyl group is pseudo-equatorial. Under
acidic conditions, 8 transitioned to a complex mixture of
protonated amine isomers (and presumably conformers thereof).
NMR analysis of the minor trans-product 9 indicated it exists
as a mixture of twist-boat conformers (A2) under neutral
conditions. The bioactive chair form (C2) was cleanly produced
under acidic conditions (Figure 4). The alternative, more
sterically demanding chair conformation (B2) was not observed.
The methyl group achieves an energetically favorable equatorial
position in conformation C2 without incurring other interactions
that would be detrimental to binding. Minor isomer 9 was
therefore predicted to be the more potent diastereomer.
free base in organic solution, followed by treatment with acid.
The enzyme surface depicted in the X-ray structure of
compound 5 (Figure 1) was also examined. A large hydrophobic
space (deep pocket “A” at the center rear) is created in the S1
pocket by the presence of the inhibitor. Ideally, addition of a
small substituent, such as a methyl group, to the spiropiperidine
inhibitor could create a favorable lipophilic interaction at that
site, and at the same time, also help bias the low energy
conformation of the molecule to be that of the bioactive
conformation. Optimally, such a substituent would also have the
added advantage of maintaining the pKa of the piperidine
nitrogen, and thus the key through-water binding interactions
with the catalytic aspartic acids.
Figure 4. Minor trans Ugi product 9 solution NMR conformation study
.
To estimate the possibility of success for this hypothesis, the
attachment of the methyl group to the inhibitor at two sites, one
at the 2-piperidine carbon, the other as a ortho benzyl
substituent, were modeled using the X-ray of compound 6 as a
starting point16. A priori, both positions appeared to possess
substituent vectors that could allow access to the pocket,
however, modeling clearly predicted that the equatorial 2-methyl
piperidine would give the best access to the pocket and thereby
desired potency (Figure 3).
The predicted activities of the two Ugi products was
confirmed in our BACE-1 enzyme assay (Table 2). The trans
diastereomer 9 predicted to have the desired configuration of the
methyl group was ~ 100-fold more potent in versus BACE-1
(BACE-1 pH 6.5 IC50 = 69 nM) compared to the undesired cis
diastereomer 8 (BACE-1 pH6.5 IC50 ~ 7.2 µM). Resolution of 9
into its enantiomers identified 10 as the active enantiomer
(BACE-1 IC50 = 55 nM). Compound 10 is ~200-fold more potent
versus BACE-1 than the des-methyl parent 2. The significant
improvement in potency as a result of the “magic methyl effect”
is reflected in the calculated ligand efficiencies LBE, BEI and
LLE (Table 2).22
Figure 3. Modeling of potential methylated analogs based on
the crystallographically determined bound conformation of
Table 2: Comparison of methylated and un-methylated
spiropiperidine analogs.
compound 6.
Additional advantages could be predicted from the 2-methyl
piperidine substitution, as the energetically preferred
conformational options available to the compound would be
narrowed. A configuration with an equatorial methyl could be
expected to stabilize the chair conformer desired for binding to
the enzyme and destabilize the other possible chair conformers
due to higher-energy methyl-proton gauche and axial-axial
interactions (vida infra).
X-ray crystallography of the active enantiomer 10 bound to
BACE-1 (Figure 5) 16 further confirmed the enzyme structure, the
position of the methyl group extending into the pocket, and the
chair form, as predicted by computational chemistry.
The limitations of gas-phase modeling to capture the
intricacies of enzyme/solvent/ligand interactions are well
recognized, but can serve as a first approximation for study.

When such an analysis is done for the spiropiperidines an
interesting correlation was observed. Computational
comparison of the des-methyl spiropiperidine analogs to
methylated analogs suggests a lowering of energetic costs
Unfortunately, the Ugi sequence was low yielding and
optimization of the synthetic route was needed. Since potassium
isocyanate (KOCN) is poorly soluble in methanol, simply
switching to potassium isothiocyanate, (KSCN), which is soluble
in MeOH, and using higher reaction temperature (70 oC) allowed
greater conversion of the starting materials and reduced side
products (Scheme 1). In addition the ratio of desired to undesired
Ugi diastereomers was improved from 1:10 to 1:2. The yield of
the desired product 18 was thus improved from 2% to 30%.26
required to achieve the bioactive chair conformation upon
24
addition of the methyl group.23
The des-methyl
spiropiperidine template accesses the bioactive conformation
within 4 kcal/mole of the ground state, while the 7-methyl
bioactive conformation is within 1.9 kcal/mol of the ground
state, an improvement of ~ 2.1 kcal. This energy stabilization
is approximately equal to the ∆∆G calculated for methyl
versus des-methyl analogs based on their observed potency,
suggesting that conformational stabilization plays a prominent
role in improving potency. Improved potency is also
attributed to the lipophilic interaction of the methyl group
with the BACE-1 pocket. ITC experiments conducted on
compound 2 revealed that binding of the piperidine to BACE-
1 (Kd 4.2 µM) was associated with ∆H = -5.536 kcal/mol and
∆S of 6.3 kcal/mol25, indicating exothermic binding with a
favored entropic requirement, presumably due to the de-
solvation of waters occupying the active site and surrounding
the ligand. Addition of the methyl group could be expected to
be beneficial in that regard. A more flexible and higher
yielding synthetic route was needed to access analogs to
further probe this interaction (vida infra).
The thio-Ugi product 18 could then be converted to 12 by first
transforming the thiocarbonyl to the iminothiochloride with
thionyl chloride and then to the iminoether with methanol, which
was then hydrolyzed to racemic 12 with acetic acid and methanol
(one pot, 3 steps ~45% yield). Resolution on reverse phase chiral
column gave the single enantiomer 12 in 3% overall yield, a 10-
fold improvement over the original route.26
Scheme 1. Thio-Ugi route to compound 12
(a) 17 (racemic)26, 3-fluoroanaline HCl, cyclohexyl
isocyanide, potassium thiocyanate, methanol, 0-70οC. (b) thionyl
chloride (c) methanol (d) HOAc/methanol (e) chiral resolution (f)
HCl/ether
Figure 5. X ray of active enantiomer 10 bound to BACE-1. The surface of
the “flap” residue Gln73 (gray stick) and water molecules not associated with
the catalytic aspartic acids have been removed for clarity.
Further improvement in route flexibility and access to analogs
was found by preparation of the resolved N-CBZ protected
piperidine ketone 20 and formation of the iminohydantoin 24
through a two-step process consisting of a Strecker reaction
followed by cyclization (Scheme 2).
The Strecker was
performed with either TiCl4/TMSCN/DCM or with Zn(CN)2/
HOAc to give a 1:2 or 3:2 ratio of desired (22) to undesired (21)
diastereoisomers. The reaction yield could be further improved
by converting the undesired diastereomer 21 to the desired
isomer 22 by treatment with TMSCN/MeOH. Ring closure could
then be carried out in a two step - one pot procedure consisting of
treatment with Cl3CONCO followed by MeOH/H2O/TEA to give
a mixture of cis 23 and trans 24. The desired enantiomerically
pure imino hydantoin isomer 24 could be isolated from the
starting ketone 20 in 38% overall yield. Treatment of 24 with
neat cyclohexylamine proceeded in 65% yield and the CBZ could
then be removed by hydrogenation and the piperidine alkylated
with different alkyl substituents as chosen. The overall yield of
12 was 3%, similar to the thio-Ugi route, but with a great
improvement in the ability to rapidly explore the synthesis of
analogs.26
In pursuit of a high potency tool compound, two potency-
enhancing groups (equatorial methyl and S3-isopropoxy) were
combined in a single analog. Compound 12 (Table 2) was
initially prepared in low yield (<0.3%) starting from racemic 1-
benzyl-2-methyl-4-ketopiperidine, followed by resolution with
chiral chromatography.15 It was gratifying to see a further 9-fold
and 5-fold improvement in in vitro and cell BACE-1 potency,
respectively, (BACE-1 IC50 = 6 nM, sAPPb IC50 = 63 nM),
compared to 10. Compound 12 was not selective versus BACE-2
(IC50 = 2 nM) but showed good selectivity versus Cathepsin D
(IC50 = 16.3 µM). Polar surface area and cLogP were maintained
at a tolerable values and protein binding still provided a
significant free fraction. Permeability was also maintained. The
increase in MW offset potency increases so that no improvement
in LBE or BEI ligand efficiency was observed compared to 10,
but a good improvement in LLE ligand efficiency was seen.
Using the Strecker route, racemic analogs of compound 12
replacing the methyl group with larger substituents such as ethyl,
propyl, cyclopropyl and allyl were prepared to probe the
lipophilic interaction of the methyl group with the enzyme (Table
2). While ethyl substitution was somewhat tolerated (only ~ 5-
fold loss in potency), propyl, cyclopropyl and allyl substitution
was accompanied by ~50-100-fold losses. Substituents larger
than methyl, which should be more likely to stabilize the
bioactive chair form, were in practice not as good for potency.
The methyl substituent appears to be of optimal steric bulk to
occupy the lipophilic pocket without displacing the rest of the
molecule from its favored enzyme-bound position. Compound

12 was therefore chosen as a potent lead with good properties
with which to explore in vivo efficacy.
of the spiropiperidines and their dependence on pH. The NOE
derived distances demonstrated that pH and configurational
substitution play a significant role in determining conformational
preferences for the spiropiperidine structure. Stabilization of the
conformation best suited for optimized binding to the receptor
via a conformational constraint, coupled with a highly-tailored
occupation of the BACE-1 binding pocket, improved inhibitor
potency by ~200 fold. This was efficiently accomplished by the
addition of one appropriately placed methyl group and
subsequently corroborated through X-ray crystallography.
Scheme 2. Strecker route to compound 12.
With sufficient quantities now synthetically available from the
improved synthetic routes, 12 was characterized for its suitability
to explore Aβ reduction in vivo. The compound was a moderate
P-gp substrate in human, as was revealed when the borderline P-
gp ratio measured at the routine screening compound
concentration (5 µM) was repeated at a more stringent, lower
compound concentration (1 µM) (human P-gp B:A/A:B ratio =
2.7 @ 5 µM, 4.4 @ 1 µM). The compound was a strong P-gp
substrate in mouse (mouse B:A/A:B ratio = 25 @1 µM). A brain
to plasma ratio of 0.25 and reduction of brain Aβ40 and Aβ42 by
30% was observed 3 hours after 12 was dosed IP (5%
methylcellulose vehicle) in mice @ 30 mpk.27 Although
encouraging, the absence of mouse protein binding data
prevented detailed interpretation of these results. As compounds
with high P-gp efflux ratios are not expected to be highly brain
ACKNOWLEDGEMENTS
The authors would like to express their thanks to Chris Cox
and Jenny Wai for their help in preparing the manuscript, and
Peter Orth for preparation and registration of the crystal
structures in the PDB.
REFERENCES AND NOTES
penetrant, it is probable that the brain to plasma ratio may have
Experimental conditions for the synthesis of compounds,
NMR data and descriptions of molecular modeling calculations
are available in Supplemental Materials.
29
overestimated the free brain exposure.28
study in ported Rhesus was undertaken.
A more detailed
Compound 12 was studied for its ability to reduce CSF Aβ
30
after IV bolus and 4 hour infusion in ported rhesus (Figure 6).
The CSF/plasma fraction unbound (fu) ratio of 12 at 4 hours after
bolus was 0.79.31 At 8 hours after bolus the compound reduced
CSF Aβ40, Aβ42, and sAPPβ by 54, 62 and 44% compared to
baseline.
Thus, 12 was shown to be a good tool to study the reduction
of CSF Aβ in monkeys. PK studies showed that the piperidine
nitrogen-benzyl methylene linkage was susceptible to cleavage
by metabolic oxidation. Further work in this series focused on
finding analogs with improved metabolic stability and lower P-gp
susceptibility and a future publication outlining this work is
planned. In addition, the series has served as inspiration to other
research groups, who have explored analogously constrained
structures as BACE-1 inhibitors.32
Figure 6. Reduction of Aβ by compound 12 after IV infusion in Rhesus
In summary, a novel mode of binding was discovered by X-
ray crystallography of a spiropiperidine BACE-1 complex.
Modeling revealed the computed high(er) energy of the
spiropiperidine ligand-bound conformation. To better understand
the relative energetic costs required for this binding mode, NMR
experiments were used to study the solution-state conformations

Table 1. Spiropiperidine BACE inhibitors 1, 2, 3 and 4.
No.
1
BACE-1 IC50 (µM)a
22 2.5l
sAPPβ (µM)b
>30
32m
PSAc
47.3
47.9
77.8
49.3
57.2
83.2
47.9
clogPd PB(h)e P-gpf
2.9
Pappg LBEh
0.05
BEIi
4.0
LLEj
-1.2
-1.4
3.3
Brain/plasmak
2
11 0.98
2.1
2.3
5.2
2.9
1.8
92%
3.4
19
25
<10
8
0.06
0.15
0.18
0.18
0.18
4.5
3
2.8 0.12
0.11 0.04
0.45 0.05
0.18 0.08
13n
8.2 1.7
5.2 2.4
5.0 0.2
0.71 0.3
11.3
13.2
12.9
13.3
4
insol
99%
76%
57%
1.8
1.0
4.2
2.5
1.8
0.26
0.23
5
6m
28
1.2
45
3.4
4.9
7
a isolated enzyme IC50 at pH 6.533 34 bcell-based activity in presence of 10% fetal bovine serum.34 cPolar surface area. doctanol/water partition coefficient, ehuman protein binding extrapolated
f
from 10% serum to 100%. B:A/A:B ratio in MDR1 transfected cells at 5 µM compound35 gPermeability measured in porcine control cells, number times 10-6 cms, h,i,jLigand binding
efficiencies calculated from BACE-1 potency, see ref.36 kbrain to plasma ratio (after dosing @ 30 mpk IP in APP-YAC mice37, 0.5% methylcellulose vehicle, 2h time point). l(n=6), mracemic.
n(n=2). Blank = not done.

Table 2: Comparison of methylated and un-methylated spiropiperidine analogs
No.
2
BACE-1 IC50 (µM)a sAPPβ (µM)b PSAc
clogPd PB(h)e P-gp ratiof
Pappg LBEh
BEIi
4.5
LLEj
-1.4
11 0.98
7.2k
32k
47.9
47.9
2.1
92%
3.4l
25
0.06
8
9m
0.069 0.013
0.055 0.004
66.0k
0.80 0.16
0.327 171
3.4
3.2
94
87
3.9
3.4
24
20
10n
11o
12p
13m
14m
15m
16m
47.9
57.2
0.22
0.22
16.2
16.2
4.0
4.5
0.006 0.0005 0.063 0.02
3.7
93
4.4q
24
0.063
0.378
0.660
0.300
0.408
1.5
>10
5.6
f
l
k
m
n
a-e,g-j,See Table 1, human P-gp B:A/A:B ratio at 1 µM compound, P-gp ratio at 5 µM compound, n=2, racemic, early eluting enantiomer of 8, olate eluting
enantiomer
of
8,
psingle
enantiomer,
qP-gp
B:A/A:B
ratio
observed
at
5
µM
was
2.7,
Blank=
not
done.

Scheme 2. Strecker route to compound 12.
a) CBZCl, MeMgBr; b) Zn/HOAc; c) resolution chiral column; d) TiCl4/TMSCN/CH2Cl2 50 0C 30 min.; e) Zn(CN)2/HOAc 50 0C, 1hr;
f) TMSCN/MeOH; g) Cl3CC(O)N=C=O; h) MeOH/H2O/TEA; i) cyclohexylamine, neat; j) H2/Pd/C/MeOH; k) 27(3 steps, see
supplemental data), DMF/K2CO3/70-80 0C (+/- NaI) 16h; l) chiral chromatography to remove 5% undesired enantiomer.
Dr., West Chester, PA 19382; Beth Pietrak: Platform
Technology and Sciences, GlaxoSmithKline, Collegeville, PA;
1 Current contact information for authors: Melissa Egbertson:
Shari Haugabook: NIH Center for Advancing Translational
melissaegbertson@gmail.com; Georgia McGaughey: Vertex
Sciences, 9800 Medical Center Dr, Rockville MD 20850;
Pharmaceuticals, Boston, MA 02210; Shaun Stauffer: Dept. of
Lixia Jin: Department of Pharmacokinetics and Drug
Pharmacology, Vanderbilt University School of Medicine,
Metabolism, Amgen Inc., 1120 Veterans Boulevard, South San
12415-D MRBIV-Langford, Nashville, TN 37232-6600; Craig
Francisco, California 94080, United States. Joe Vacca: Early
Coburn: WuXi AppTec (Shanghai) Co., Ltd. 288 Fute Zhong
Success Partnerships, WuXi AppTec Inc., 1690 Sumneytown
Rd, Shanghai, China; Wenjin Yang, Aldea Pharmaceuticals
Pike, Suite 150 Lansdale, PA 19446
3696 Haven Avenue, Suite C Redwood City, CA 94063; Jim
2 http://www.alz.org/facts/downloads/facts_figures_2015.pdf
Barrow: Dept. of Pharmacology and Molecular Science
3 Vassar, R.; Kuhn, P.H.; Haass, C.; Kennedy, M.E.;
Lieber Institute for Brain Development School of Medicine
Rajendran, L.; Wong, P.C.; Lichtenthaler, S.F. J. Neurochem
Room 336, John G. Rangos, Sr., Building, 855 N. Wolfe
2014 130(1) 4.
Street, Baltimore, MD 21205; Lou Anne Neilson McKown:
4 De-Paula, V.J.; Radanovic, M.; Diniz, B.S.; Forlenza, O.V.
The Horsemanship School 13651 Woodinville-Redmond Rd
Subcell Biochem. 2012 65 329.
NE, Redmond, WA 98052; Melody McWherter: 2 Park Road,
5 Stachel, S. J. Drug Devel. Res. 2009, 70, 100.
Boyertown, PA 19512; Debra Perlow: Clinical Integrated
6 Reitz, C.; Int. J. Alzheimer’s Disease, 2012; 2012:369808.
Programs, PRAH Sciences, 520 Virginia Drive Fort
7 Morris, G.P.; Clark, I.A.; Vissel, B.; Acta Neuropathol.
Washington, PA 19034; Sanjeev Munshi: Business
Commun. 2014, 2, 135.
Development and Licensing, Merck and Co. PO Box 2000,
8 Vassar, R.; Alz. Res.Ther. 2014, 6, 89.
Rahway, NJ 07065; Tim Allison: Labcyte Inc.
9 Kelder, J.; Grootenhuis, P.D.; Bayaa, D.M.; Delbressine,
timothy.j.allison@gmail.com; Bruce Fahr: Celgene 1500
L.P.; Ploemen, J.P. Pharm. Res. 1999, 16, 1514.
Owens St # 600, San Francisco, CA 94158; Harold Selnick,
10 Nienaber, V.; Curr. Top. Med. Chem. 2009, 9, 1688.
selnickh@gmail.com; Adam Simon: Cerora, Inc. 116
11 Estimated volume for BACE-1 sub-sites S2’, S1’, S1, S2,
Research Drive Suite 2207 Bethlehem, PA 18015-4731; Sethu
S3 based on X-ray structure of compound 3. For BACE
Sankaranarayanan: Bristol Meyers Squibb, 1324 Chesapeake
nomenclature see ref 12.
Terrace, Sunnyvale, CA 94089; Dennis Colussi: Dept. of
12 Patel, S.; Vuillard, L.; Cleasby, A.; Murray, C.W.; Yon, J.
Pharmaceutical Sciences, Temple University School of
J. Mol. Bio. 2004, 343, 407.
Pharmacy, 3307 North Broad Street Philadelphia, PA 19140;
13 Stachel S.J.; Coburn C.A.; Steele T.G.; Jones K.G.;
Katherine Tugusheva: Teva Pharmaceuticals, 1027 Squires
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.
P.G.; Stachel S.R.; Yang, W.; Lu, W.; Fahr,
B.;WO2007011833 Jan 25, 2007
27 For procedure see ref. 14 Stanton et.al.
28 Liu, X.; Smith, B.J. J. PET 2008, 325, 349.
14 Stanton M.G.; Stauffer S.R.; Gregro A.R.; Steinbeiser M.;
Nantermet P.; Sankaranarayanan S.; Price E.A.; Wu G.;
Crouthamel M.C.; Ellis J.; Lai M.T.; Espeseth A.S.; Shi X.P.;
Jin L.; Colussi D.; Pietrak B.; Huang Q.; Xu M.; Simon A.J.;
Graham S.L.; Vacca J.P.; Selnick H. J. Med. Chem. 2007; 50;
3431.
29 Rankovic, Z. J. Med. Chem. 2015, 58, 2584.
30 Gilberto, D.B.; Zeoli, A.H.; Szczerba, P.J.; Gehret, J.R.;
Holahan, M.A.; Sitko, G.R.; Johnson, C.A.; Cook, J.J.;
Motzel, S.L. Contemp. Top. Lab. Anim. Sci. 2003, 42, 53.
31 Plasma fraction unbound (Fu) in Monkey = 12.1%.
32 Brodney M.A.; Barreiro G.; Ogilvie K.; Hajos-Korcsok E.;
Murray J.; Vajdos F.; Ambroise C.; Christoffersen C. Fisher
K.; Lanyon L.; Liu J.; Nolan C.E.; Withka J.M.; Borzilleri
K.A.; Efremov I.; Oborski C.E.; Varghese A.; O'Neill B.T. J
Med Chem. 2012; 55; 9224.
15 Barrow; J.C.; Stauffer; S.R.; Rittle; K.E.; Ngo; P.L.; Yang;
Z-Q.; 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.; Zuck; P.; Levorse; D.A.; Hazuda; D.; Vacca; J.P. J. Med.
Chem. 2008; 51; 6259.
33 Stachel S.J.; Coburn C.A.; Rush D.; Jones K.L.; 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. Bio. Org.
Med. Chem. Lett. 2009; 19; 2977.
16 PDB files for 3 (PDB identifier 3E3W (updated to 3FKT) 5
(4ZPE) 6 (4ZPF) and 10 (4ZPG) BACE-1/inhibitor
complexes are available at the Protein Data Bank
(http://www.rcsb.org/pdb).
34 For a description of in vitro assays see: Pietrak; B.L.;
Crouthamel; M.-C.; Tugusheva; K.; Lineberger; J.E.; Xu; M.;
DiMuzio; J.M.; Steele; T.; Espeseth; A.S.; Stachel; S.J.;
Coburn; C.A.; Graham; S.L.; Vacca; J.P.; Shi; X.-P.; Simon;
A.J.; Hazuda; D.J.; Lai; M.-T. Anal. Biochem. 2005; 342; 144.
35 Hochman, J.H.; Pudvah, N.; Qiu, J.; Yamazaki, M.; Tang,
C.; Lin, J.; Prueksaritanont, T.; Pharm. Res. 2004, 21, 1686.
36 Hopkins, A.L.; Keserű, G.M.; Leeson, P.D.; Rees, D.C.;
Reynolds, C.H.; Nature Rev, Drug Discov. 2014, 13, 105.
LBE = -log BACE-1 IC50 (molar) / number of non-hydrogen
atoms. BEI (binding efficiency index = -log BACE-1 IC50
(molar) / MW in kilodaltons. LLE = -log BACE-1 IC50
(molar) minus clogP.
17 www.rcsb.org search “beta secretase”.
18 MM2 minimization in ChemBio3D of un-protonated
compound 2 protein-bound conformation C (Figure 2) results
in conversion to boat (total energy decrease from 319 to 35.22
kcal/mole). Upon simulated heating, conversion between boat
forms A and chair form B is observed, chair form C is not
observed.
19
NMR methods: Samples were generally studied at a
concentration of ~1mg/mL in the indicated organic solvent.
Data were acquired using a Varian Inova 600 spectrometer
operating with a sample temperature of 25oC. In the case of
compound 7, the sample temperature was also lowered to -
20oC, but no perturbation attributable to a change in
conformers was observed. Standard 1D and 2D pulse
sequences were used. Samples were titrated with acid via the
introduction of small amounts of TFA headspace vapor into
the headspace of the NMR tube followed by mixing.
20 Isopropyl substitution was chosen over cyclohexyl for
simplification of the NMR spectra.
37 Lamb, B.T.; Sisodia, S.S.; Lawler, A.M.; Slunt, H.H.; Kitt,
C.A.; Kearns, W.G.; Pearson, P.L.; Price, D.L.; Gearhart, J.D.
Nat. Genet. 1993, 5 22.
21 Barrow, J.C.; Coburn, C.A.; Egbertson, M.; McGaughey,
G.B.; McWherter, M.; Neilson, L.A.; Selnick, H.G.; Stauffer,
S.R.; Yang, Z.-Q.; Yang, W.; Lu, W.; Fahr, B.; Rittle, K.E.
US 2007/0197571
22 Schönherr H, Cernak T. have recently reviewed additional
examples of the “magic methyl effect” in Angew Chem Int Ed
Engl. 2013, 52, 12256.
23
Calculations done using a distance dependent dielectric of
two in the gas phase with the Monte Carlo algorithm and the
MMFF force field. This energy difference is similar to an
approximate ∆∆G = RTln(Kd1/Kd2) (T=298) calculated
between methyl and des methyl analogs. See supplemental
material for details.
24 MMFF force field: Halgren T.A. et.al. J. Comp. Chem.
1996, 490
25 Ming-Tain Lai, unpublished results.
26 Coburn, C.A.; Egbertson, M.S.; Graham, S.L.; McGaughey,
G.B.; Stauffer, S.R.; Yang, W.; Lu, W.; Fahr, B.; US patent
2007/0021454, July 3 2012. WO 2007/011810 January 25
2007, Coburn, C.A.; Egbertson, M.S.; Graham, S.L.;
McGaughey, G.B.; Stauffer, Rajapakse, H.A.; Nantermet,