Constrained peptidomimetics reveal detailed geometric requirements of covalent prolyl oligopeptidase inhibitors

Article abstract of DOI:10.1021/jm901013a

Prolyl oligopeptidases cleave peptides on the carboxy side of internal proline residues and their inhibition has potential in the treatment of human brain disorders. Using our docking program FITTED, we have designed a series of constrained covalent inhibitors, built from a series of bicyclic scaffolds, to study the optimal shape required for these small molecules. These structures bear nitrile functional groups that we predicted to covalently bind to the catalytic serine of the enzyme. Synthesis and biological assays using human brain-derived astrocytic cells and endothelial cells and human fibroblasts revealed that these compounds act as selective inhibitors of prolyl oligopeptidase activity compared to prolyl-dipeptidyl-aminopeptidase activity, are able to penetrate the cells and inhibit intracellular activities in intact living cells. This integrated computational and experimental study shed light on the binding mode of inhibitors in the enzyme active site and will guide the design of future drug-like molecules.

Full text of DOI:10.1021/jm901013a

6672 J. Med. Chem. 2009, 52, 6672–6684
DOI: 10.1021/jm901013a
Constrained Peptidomimetics Reveal Detailed Geometric Requirements of Covalent Prolyl
Oligopeptidase Inhibitors
Janice Lawandi,† Sylvestre Toumieux,† Valentine Seyer,† Philip Campbell,† Sabine Thielges,† Lucienne Juillerat-Jeanneret,*,‡
and Nicolas Moitessier*,†
†Department of Chemistry, McGill University, 801 Sherbrooke Street W, Montreꢀal, Queꢀbec, Canada H3A 2K6, and ‡University Institute of
Pathology, CHUV-UNIL, Rue du Bugnon 25, CH-1011 Lausanne, Switzerland
Received July 9, 2009
Prolyl oligopeptidases cleave peptides on the carboxy side of internal proline residues and their
inhibition has potential in the treatment of human brain disorders. Using our docking program
FITTED, we have designed a series of constrained covalent inhibitors, built from a series of bicyclic
scaffolds, to study the optimal shape required for these small molecules. These structures bear nitrile
functional groups that we predicted to covalently bind to the catalytic serine of the enzyme. Synthesis
and biological assays using human brain-derived astrocytic cells and endothelial cells and human
fibroblasts revealed that these compounds act as selective inhibitors of prolyl oligopeptidase activity
compared to prolyl-dipeptidyl-aminopeptidase activity, are able to penetrate the cells and inhibit
intracellular activities in intact living cells. This integrated computational and experimental study shed
light on the binding mode of inhibitors in the enzyme active site and will guide the design of future drug-
like molecules.
Introduction
only to achieve selective inhibition of POP activity over other
prolyl-specific proteases but also to reveal the optimal shape,
size, and electronic requirement of a potent POP inhibitor. To
reach these goals, we have exploited an approach integrating
structures available from X-ray crystallography data, docking
predictions, higher level computations (e.g., density functional
theory, DFT), and chemical tools. We have combined and
applied all of these techniques, also evaluating the inhibition of
enzymatic activity, to reveal insight into binding and inhibition
modes that each of them alone cannot predict. Using this
strategy, we have designed a series of pseudopeptidic and
peptidomimetic inhibitors, many of which are built around
bicyclic scaffolds. These scaffolds closely mimic the known
prolyl oligopeptidase inhibitor Cbz-Pro-prolinal (1, Figure 1).
We report herein docking predictions of a series of inhibi-
tors into the active site of POP, followed by the synthesis and
biological evaluationofthedesignedinhibitorsinhumancells.
Of the entire series of compounds, one bicyclic scaffold
exhibits high nanomolar inhibition of prolyl oligopeptidase
activity and high selectivity over prolyl-dipeptidyl aminopep-
tidase activity in human brain astrocyte-derived cells, human
brain-derived endothelial cells, and human fibroblasts. Com-
bining results from the computational modeling and biologi-
cal evaluation, optimal inhibitor geometry is revealed.
Prolyl oligopeptidase (POP,a EC 3.4.21.26) is a large (about
80 kDa), highly conserved, and widely distributed prolyl-
selective serine endoprotease. Unlike other serine proteases,
prolyl oligopeptidase can accommodate proline residues (and
less frequently alanine residues) in its active site, cleaving short
peptideson the carboxysideofthese residues.1 Overthe last 15
years, abnormal POP activity, particularly in many regions of
the brain,2 has been linked to a number of diseases.3-5 As a
consequence, POP is believed to be a target for the treatment
of neurodegenerative (e.g., Alzheimer’s disease) and psychia-
tric (e.g., bipolar disorder6) disorders. A number of POP
inhibitors, including covalent inhibitors such as Cbz-Gly-
prolinal (1), Cbz-Pro-prolinal (2),7 and JTP-4819 (3),8,9 and
noncovalent inhibitors (e.g., SUAM-1221 (4)10 and S-17092-1
(5)11) have been prepared and evaluated in various experi-
mental models. Some of these POP inhibitors were found to
improve memory and learning12 in animal models of brain
disorders and to prevent amyloid-like depositions.13-15 A few
of these inhibitors also exhibited good pharmacokinetic and
safety profiles and were able to penetrate the blood-brain
barrier.16 However, none has yet reached advanced clinical
trials, and drug candidates are yet to be developed.
Although many POP inhibitors have been reported, a better
understanding of the geometric and electronic requirements for
a potent and selective POP inhibitor is still necessary in order to
design improved POP inhibitors. Our aim in this study was not
Results
Computer-Aided Design of Constrained Inhibitors. We
combined structural knowledge from reported inhibitor
structures, docking experiments, and evaluation of synthetic
feasibility to design a series of inhibitors. To date, no crystal
structure for the human isoform of prolyl oligopeptidase has
been reported. However, a structure of POP from porcine
brain (PDB code: 1H2Y), which demonstrates 97% identity
*To whom correspondence should be addressed. For N.M.: phone,
514-398-8543; fax, 514-398-3797; E-mail, nicolas.moitessier@mcgill.ca.
For L.J-J.: phone, þ41 21 314 7173; fax, þ41 21 314 7115; E-mail:
lucienne.juillerat@chuv.ch.
a Abbreviations: DFT, density functional theory; POP, prolyl oligo-
peptidase; DPP, dipeptidyl peptidase.
r
pubs.acs.org/jmc
Published on Web 10/12/2009
2009 American Chemical Society

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6673
Figure 3. Docked designed covalent inhibitor (8b, green) overlaid
with crystal structure of Cbz-Pro-prolinal (2, yellow).
as shown by Bayly and co-workers on a set of nitrile deriva-
tives.22 As the scoring of covalent bonds would require more
advanced and much more time-consuming techniques to be
accurate, ourstrategy cannotrelyongenerated scores. Instead
we wished to design compounds that maintained the covalent
bond and at least 2 of the 3 major interactions observed with
these pseudopeptides. We have already successfully applied
this strategy to the design of metalloenzyme inhibitors, as the
metal coordination is not well scored either.23
Figure 1. Selected POP inhibitors.
Typically, a docking program used in conjunction with a
scoring function predicting the binding affinity of the docked
ligand gives high scores to good binders. However, none of
the currently available scoring functions can handle covalent
bonds formed upon docking. Thus, we visually inspected the
docked poses, hypothesizing that if the bound pose retained
the key interactions observed in the crystal structure, the
docked ligand should be active. This assumption has been
successfully exploited in other medicinal chemistry programs
such as the development of nanomolar metalloprotease
inhibitors.23 Compound 2 binds to POP through two hydro-
gen bonds with the side chains of an arginine (Arg643) and a
tryptophan (Trp595), while the five-membered ring proline
ring sits on the aromatic ring of the Trp595 side chain. We
predicted, using FITTED, that when bound to POP, the
bicyclic structure 8b should retain the key interactions made
by 2 (Figure 3). We therefore prioritized 8b for immediate
synthesis. Synthetic routes enabling the synthesis of a small
set of analogues of 8b, with different ring sizes and heteroa-
toms, were developed. By varying only a few atoms of the
bicyclic scaffold, small changes in geometry and chemical
properties were calculated. For example, DFT calculations
were carried out on the analogous scaffolds 8b and 9b and
revealed significant differences in the electronic distribu-
tions, making the nitrile of 9b less reactive (DFT results
not shown).
To evaluate the impact of the conformational constraint
of the bicyclic structures, we also designed and prepared a
series of pseudopeptides, closely resembling the bicyclic
scaffolds, in order to compare the bioactivity of the flexible
dipeptide to that of the corresponding rigid bicyclic scaf-
folds. From the docking study, we predicted that the five-
membered proline ring of compounds 11a and 12a,b sits on
the aromatic side chain of Trp595. Our docking studies also
indicated that this interaction should be disrupted when a
six-membered ring (14) is introduced; a loss of potency is
expected. We chose to also synthesize 14 and test it to
validate our docking predictions.
Figure 2. Pseudopeptidic inhibitors inspired from 1 and 7.
with the human form, is available and can be used for
docking experiments.17
As illustrated in Figure 2, this ligand-based design led us to
the selected scaffold-based inhibitors, designed to mimic
Fmoc-Ala-pyrrolidine-2-carbonitrile (7), previously shown
to be a noncompetitive inhibitor selective for POP over DPP-
IV,18 and Cbz-Gly-prolinal (1). Several pieces of evidence18
including a crystal structure of Cbz-Pro-prolinal bound to
POP,17 indicate that both inhibitors 1 and 7 are most likely
reacting with the active site serine. For this work, we
hypothesized that the designed inhibitors will also bind
covalently to POP and covalent docking was carried out.
We used the latest version of our docking program FITTED
(version 2.6)19-21 to dock various covalent inhibitors, mainly
built around rigid, bicyclic scaffolds. However, we first had
to modify the program to perform covalent docking. With
the novel implementations in the program, FITTED can now
automatically identify reactive functional groups in the
ligands (e.g., aldehydes, nitrile) and then dock these ligands,
evaluating both the noncovalent and covalent binding
modes simultaneously. As mentioned in the design section,
we planned to develop constrained structures that would
mimic pseudopeptidic inhibitors. The major interaction of
these compounds is the covalent bond. However, the scoring
of the binding of covalent inhibitors requires considering the
reactivity of the reactive group (i.e., nitrile or aldehyde),
which can vary significantly from one compound to another
Chemistry
Synthesis of the Pseudopeptidic Inhibitors. The series
of pseudopeptides, incorporating either a piperidine-2-

6674 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Lawandi et al.
Scheme 1a
Scheme 2a
a (a) Boc2O, Et3N, DMAP, CH2Cl2, 40%; (b) NH3 (aq) 28%, THF,
60°C, 89%; (c) (CF3CO)2O, Et3N, THF, 0 °C, 74-83%; (d) CF3CO2H,
CH2Cl2, 0°C, 29-53%; (e) RCOOH (Cbz-Gly-OH or Cbz-Ala-OH or
Cbz-D-Ala-OH), Et3N, HOBt, EDCI, CH2Cl2, 30-65%; (f) Boc2O,
Et3N, 1,4-dioxane/H2O, 82%; (g) ClCO2Et, Et3N, THF, -15°C then
NH3 (aq) 28%, rt, 83%; (h) (CF3CO)2O, Et3N, CH2Cl2, 0 °C, 52-68%;
(i) HCl, 1,4-dioxane, 90%.
carbonitrile or pyrrolidine-2-carbonitrile moiety, was synthe-
sized as described in Scheme 1. To synthesize Cbz-Gly-L/D-
pyrrolidine-2-carbonitrile (11a,b) and Cbz-L/D-Ala-L-pyrroli-
dine-2-carbonitrile (12a,b), the pyrrolidine nitrile portions
were prepared from L- and D-Pro methyl esters using pre-
viously reported procedures (see Experimental Section). Then
the carbonitrile derivatives were coupled to Cbz protected Ala
and Gly residues. In parallel, we also prepared the racemic
compound 14 from pipecolinic acid 13. The procedures used
to prepare the pyrrolidine-2-carbonitriles were not as success-
ful on pipecolinic acid 13. We therefore used slightly different
methods to make 14 (listed in steps f-i of Scheme 1).
Synthesis of the Bicyclic Scaffolds. Following the synthesis
of the pseudodipeptides, we synthesized the (5,5) fused
scaffolds as enantiopure compounds or diastereomeric mix-
tures enriched in the diastereomer that we predicted to be
most active from the docking and the initial results from the
biological testing (Scheme 2).24-28 After protection of the
amine and acid functions of allyl Gly rac-15,29 we oxidized
the alkene to the corresponding aldehyde rac-16 in quanti-
tative yields.30 Reaction of this aldehyde (rac-16) with L-Cys-
OMe at room temperature, and then at 50 °C for several days
in pyridine, afforded the first two bicyclic structures 17a and
17b after careful chromatographic separation. Alternatively,
we repeated this same procedure to prepare the desired
bicyclic structures 17c and 17d as a separable 1:1 diastereo-
meric mixture. The same approach was applied to other
amino acids, and we synthesized the scaffolds 18a and 18b
from rac-16 and L-Ser-OMe and 19a and 19b from rac-16 and
L-Thr-OMe.
a (a) TMSCl, MeOH, 0 °C to rt; (b) CbzCl, Et3N, 0 °C; (c) O3, CH2Cl2,
then DMS; (d) L-Cys-OMe, pyridine, molecular sieves, then pyridine,
50 °C, 4 days, 48% 17a,b (over 3 steps); (e) L-Ser-OMe, pyridine,
molecular sieves, then pyridine, 50 °C, 4 days, 49% of 18a,b (over 3
steps); (f) (i) D-Cys-OH, pyridine, molecular sieves, then pyridine, 50 °C,
4 days, (ii) TMSCHN2, CH2Cl2 33% of 17c,d (over 4 steps); (g) L-Thr-
OMe, pyridine, molecular sieves, then pyridine, 50 °C, 4 days, 35% of
19a,b (over 3 steps); (h) L-homo-Cys-OH, pyridine, 50 °C, 4 days;
(i) TMSCHN2, CH2Cl2, 32% of rac-20 (over 4 steps).
Scheme 3a
To probe the geometric requirement for optimal binding
and to probe the computational predictions, the size of the
right-hand five-membered ring was expanded to 20 as a
mimic of 14. rac-20 was prepared following synthetic routes
similar to the one used previously (Scheme 2). In contrast to
the synthesis of the 5,5-bicyclic structures, none of the other
possible diastereomers were isolated in a large enough
amount to enable any further transformations. In addition,
a (a) CbzCl, 4 N NaOH, 0 °C to rt, 21 h; (b) (CH2O)n, cat. pTsOH,
toluene, reflux, 3h; (c) (COCl)2, cat. DMF, CH2Cl2, quant; (d) LiAl-
[OtBu]3H, THF; (e) (i) L-CysOMe HCl, pyridine, 24 h, (ii) 50 °C, 3 days;
3
(f) K2CO3, MeOH, 6-8% over 6 steps.
our docking studies predicted a distorted binding mode and
further efforts to make the other isomers were deemed
unnecessary.

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6675
Figure 4. Selected NOE identifying the stereochemistry of the
isolated compounds.
Scheme 4a
Figure 5. Comparison of 1H NMR spectra, confirming stereo-
chemistry of isolated diastereomers. The peaks refer to the two
hydrogen atoms on carbon no. 5 (see 8b in Figure 4).
relies on the intermediate 22, which upon treatment
with L-Cys-OMe afforded the desired structure in a single
step (Scheme 3).
Functionalization of the Scaffolds to Covalent Inhibitors.
The covalent functionality was added to each of the scaffolds
once formed (Scheme 4). For this purpose, we converted
each of the methyl esters into the corresponding amides, 24a
through 28b, using ammonia,31 which we then transformed
into the corresponding nitriles with good to excellent yields
by dehydration.32
Confirming the Stereochemistry of the Various Diastereo-
mers. Given that 17a-d are reported in the literature,25,33
providing stereochemistry identified by NOE experiments, we
compared NMR data from our purified compounds to
the reported data. For all other bicyclic compounds, both
COSY and NOESY experiments were performed to identify
a (a) NH3, MeOH, rt, quant; (b) (CF3CO)2O, Et3N, THF, rt,
27-81%.
Two additional diastereomeric scaffolds (23a and 23b)
were prepared from L- and D-Glu (21a,b) using a modified
strategy reported by Subasinghe et al.26 This strategy

6676 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Lawandi et al.
Table 1. Inhibition of POP Activity and DPP-IV Endoprotease Activity in Cell Extracts
Z-Gly-Pro-AMC (POP activity)a
Gly-Pro-AMC (DPP IV activity)a
human astrocyte-
derived cells b
human brain-derived
endothelial cellsc
human
fibroblastsd
human astrocyte-
derived cells
human brain-derived
endothelial cells
human
fibroblasts
compd
Z-Pro-NH2
Z-Pro-CN
Boc-Pro-CN
Pro-CN
11a
þ
-
-
-
þ
þ
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
nd
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
þþþ
þþþ
þþþ
þþþ
þþ
þ
þþþ
þþþ
þþþ
þþþ
þ
þþþ
þþþ
þþþ
þþþ
þ
11b
12a
12b
14
17a
17b
-
-
-
-
þ
-
-
-
þ
24a
24b
-
-
8a
8b
þ
þþ
þþþ
þ
þ
þþþ
þ
þþþ
8c
8d
-
þ
þþ
þ
9a,b
-
-
þ
29a,b
rac-30
31a
þ
þ
-
-
-
þ
þþ
þþ
þþ
þþ
31b
þþ
þþ
a -, No inhibition at 100 μM; þ, 10-50% inhibition at 100 μM; þþ, >50% inhibition at 100 μM; þþþ, >90% inhibition at 20 μM; nd, not
determined. b LN18, LN229, LNZ308: human glioblastoma cells. c HCEC, human brain-derived endothelial cells. d PO03, human lung-derived
fibroblasts; PG98/5, human skin-derived fibroblasts.
Table 2. Inhibition of POP Endoprotease Activity in Intact Living Cells
Z-Gly-Pro-AMC (POP activity)a
the isolated diastereomers (Figure 4). In addition, only one
stereogenic center is made throughout the synthesis while the
other two already appear in the starting amino acids. When
racemic allyl glycine was used, two stereogenic centers needed to
be assigned.
To further confirm the stereochemistry of the scaffolds, we
simply compared the 1H NMR spectra of the novel scaffolds
to those of 8a-b (Figure 5).
human astrocyte- human brain-derived
derived cells
human
fibroblasts
compd
endothelial cells
Z-Pro-NH2
Z-Pro-CN
Boc-Pro-CN
Pro-CN
11a
þ
-
-
-
-
-
-
þ
-
-
-
-
þþþ
þþþ
þþþ
þþþ
þ
þþþ
þþþ
þþþ
þþþ
-
-
-
-
-
þþþ
þþþ
þþþ
þþþ
-
Biological Evaluations
11b
12a
Enzyme Inhibition in Cell Extracts. POP activity was
measured using the fluorogenic substrate Z-Gly-Pro-AMC
and prolyl-dipeptidyl-aminopeptidase (DPP IV) activity
with the fluorogenic substrate Gly-Pro-AMC using the hu-
man brain astrocyte-derived LN18, LN229, and LNZ308
glioblastoma cells, the human brain-derived HCEC endothe-
lial cells and the human lung-derived PO03 and skin-derived
PG98/5 fibroblasts. The initial inhibition screening experi-
ments were performed using cell extracts, obtained by ex-
tracting cell layers in PBS-0.1% Triton X-100 to determine
the potential of the developed inhibitors to inhibit total
cellular enzyme activities (Table 1).
12b
14
17a
17b
-
-
-
-
þ
þ
24a
24b
-
-
8a
8b
-
þ
þ
þþþ
þþþ
þþþ
8c
8d
-
-
-
-
-
-
-
-
þ
9a,b
-
þ
29a,b
rac-30
31a
þ
-
-
-
All compounds were found to be very selective for POP
activity over DPP IV activity in cell extracts.
þþ
þ
þþ
31b
a -, No inhibition at 100 μM; þ, 10-50% inhibition at 100 μM; þþ,
>50% inhibition at 100 μM; þþþ, >90% inhibition at 20 μM.
þþ
þþ
þþ
Enzyme Inhibition in Intact Living Cells. To evaluate the
potential of these molecules to inhibit membrane-inserted
enzymes as well as to penetrate cells and inhibit intracellular
enzymatic activities, the inhibition of POP activity was also
evaluated with the same compounds in intact living cells
(Table 2).
ing the potency of 11a, 11b, 12a, 12b, and 8b as POP
inhibitors, the IC50 values were determined in cell extracts
and in intact cells (Figure 6 and Table 3).
As shown in Table 2, the inhibitors not only inhibit POP
activity in cell extracts but also this proteolytic activity in
intact living cells. Out of all the compounds tested, com-
pounds 11a, 11b, 12a, 12b, and 8b were the best inhibitors of
POP activity. To obtain more detailed information concern-
Compounds 12a and 11a are the most potent inhibitors of
POP activity of our series, displaying low nanomolar IC50
values. They are also very selective over DPP IV activity,
both in cell extracts and in intact cells (data not shown).
The inhibition of POP activity, as defined by IC50, is about

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6677
Figure 6. Dose-response inhibition of POP activity by 11a, 11b, 12a, 12b, and 8b in intact cells (PBS) and in cell extracts (PBS-Triton).
10-fold higher in cell extracts than in intact cells, possibly
reflecting the partial efficacy of these inhibitors of crossing
the cell membrane.
may share the endoproteolytic activity attributed to POP.
Highly specific POP inhibitors therefore need to be designed
in order to further explore this enzyme and its function.
As our long-term aim is to develop POP inhibitors target-
ing POP expressed in the brain, and as most of the informa-
tion concerning POP and its inhibitors was obtained in brain
models, we evaluated the inhibitory potency of our com-
pounds in human brain-derived cells of astrocytic origin
(glioblastoma cells) and endothelial cells as a model of cells
forming the blood-brain barrier. As POP and POP activity
are widely expressed in many other cells, we also evaluated
the inhibitory potential of the inhibitors in human fibro-
blasts of the lung or the skin, as models for non-brain-
derived cells. Selectivity for endoproteolytic POP activity
over exoproteolytic DPP IV activity can easily be obtained
due to the different chemical properties of their catalytic site
(DPP IV binds to charged terminal residues while POP does
not). Selectivity for other prolyl-specific endoproteolytic
activities is challenging and in fact rarely reported.40
Constrained POP Inhibitors. In contrast to noncovalent
competitive inhibitors which can adjust their orientation and
translation in the enzyme binding site, the orientation of
covalent inhibitors is tightly controlled by the forming
covalent bond and any seemingly minor change in shape
can have a significant impact on the binding affinity. The
docking study and biological testing of the scaffold-based
structures revealed 8a and 8b as potent POP inhibitors. The
biological evaluation of the analogues provided more in
Discussion
POP and Other Prolyl Peptidases. The family of serine
peptidases able to cleave peptide bonds after a prolyl residue
comprises exopeptidases such as the cell plasma-membrane
anchored DPP IV/CD26 and FAP-R/seprase, and the intra-
cellular DPP8, DPP9, and endopeptidases, mainly repre-
sented by the widely distributed intracellular prolyl
oligopeptidase (POP). POP and/or POP-like activity have
been found in many regions of the brain2,34 and has been
associated with the cleavage of oligopeptides involved in
memory and the processing of amyloid precursor protein to
β-amyloid.15 Studies on animal models suggested that POP
inhibition may lead to improved memory and learning in
rodents,13,35 to reverse scopolamine-induced amnesia in
rats,36 and to enhance cognition in Parkinson’s disease.37
Human POP levels are reduced in different stages of depres-
sion, increased in maniac and schizophrenic patients, and the
enzyme is reactivated to normal levels by antidepressants.38
Conflicting data regarding altered POP activity13,39 have
been reported in patients with Alzheimer’s disease. Thus,
POP activity may represent a target for the treatment of
neurological disorders. However, FAP-R/seprase, and possi-
bly the DPP8/DPP9 proteases, also display postprolyl hydro-
lytic endoproteolytic activities.34 Therefore, several proteases

6678 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Lawandi et al.
Table 3. IC50 Values [nM] for the Inhibition of POP Endoproteolytic Activity in Intact Living Cells and in Cell Extracts
LN229 (nM)
LNZ308 (nM)
HCEC (nM)
compd
intact cells
cell extract
intact cells
cell extract
intact cells
cell extract
11a
11b
12a
12b
8b
325
3400
28
20
430
3
590
>4000
29
20
430
5
180
2800
20
95
690
7
7500
2000
1400
500
4700
2500
1200
200
6600
1300
2000
700
Figure 7. Effect of cyclization on inhibitory potency.
depth information about the required functional groups and
stereochemical features of the inhibitors. First, comparing
8a/b to 17a/b and 24a/b revealed that the nitrile functional
group is essential for optimal binding, most likely through
the postulated covalent bond with the catalytic serine. Then
the predicted drop in activity observed from 11a to 14 and
from 8a to rac-30 confirmed that mimicking the Pro ring of
Cbz-Gly-prolinal (1) with a larger ring results in loss of
binding affinity. Unlike with the five-membered ring systems
mimicking Pro, this expanded six-membered ring cannot fit
into the binding site without clashing with the enzyme or
disrupting the key hydrogen bonds and hydrophobic inter-
actions. The bound pose is distorted when compared to the
crystal structure, as revealed by our docking studies.
Figure 8. (a) Structure of Cbz-Pro-prolinal cocrystallized with
POP. (b) Docked structures of 12a, 12b, 8a, and 8b (yellow) super-
posed to the crystal structure of Cbz-Pro-prolinal (green).
Furthermore, the replacement of the sulfur atom of 8b by
an oxygen atom (in 9b) resulted in a complete loss of activity.
Although reduction in inhibitory potency has been observed
when moving from thiooxazoline to oxazoline,41 such a large
drop in affinity is unexpected. Again, we hypothesize that
these changes are magnified by the fact that these inhibitors
bind covalently. This seemingly subtle change actually leads
to reduced ring size, C-S bonds being significantly longer
than C-O bonds. As for rac-30, the increase to a six-
membered ring resulted in a different shape of the whole
molecule, hence a different binding affinity. To further probe
the differences between these two structures, DFT calcula-
tions were carried out and revealed a significantly different
electronic distribution and a much less polarized (i.e., less
reactive) nitrile group in 9b when compared to 8b. This
reduced reactivity may prevent the covalent binding.
The stereochemistry of the molecular scaffolds was also
critical for the inhibition of POP endoproteolytic activity (8a
vs 8b, 8c, or 8d). As expected, the natural amino acid
configuration is required at the nitrile position (8a vs 8c, 8b
vs 8d, or 11a vs 11b) in order to mimic the natural Pro-
containing substrate of the enzyme. More interestingly, the
optimal configuration of the left-handed stereogenic center
(carbon 6 in Figure 7) of the scaffold mimics the D-amino
acid configuration (8a vs 8b). To investigate this specific
observation, we prepared the corresponding pseudopeptides
12a and 12b (Figure 7). When moving from Gly to L-Ala (11a
to 12a), the inhibition increased by an order of magnitude.
More interestingly, when the configuration of the Ala resi-
due was inverted (12a to 12b) (Figure 7), inhibition was still
observed but dropped by a factor of 500. Similarly, when the
Pro stereogenic center is inverted, 11a versus 11b, inhibition
is observed, although with a higher IC50 for the D-isomer.
This data suggests that the stereochemistry at these two
centers have an impact on the potency.
As shown in Figure 8, the proposed binding mode of 12a is
very similar to the experimentally observed binding mode of
Cbz-Pro-prolinal. The proline ring stacks on Trp595 and
interacts with Phe476, while the terminal phenyl ring inter-
acts with Phe173 and the two hydrogen bonds between
carbonyls of the inhibitor and the enzyme are retained. In
contrast, 12b does not fit as nicely in the binding site and a
hydrogen bond is lost, resulting in a significant drop in
potency. When constraining 12a and 12b into 8a and 8b, a
loss of activity is observed from 12a to 8a, while a slight

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6679
Figure 9. Effect of cyclization on inhibitory potency.
increase is observed from 12b to 8b. We rationalize that the
stereochemistry of the carbon C1, which is imposed by
the synthetic pathway influences the binding (Figures 8
and 9). Comparing the crystal structure of Cbz-Pro-prolinal,
cocrystallized with POP (PDB code: 1H2Y), and the docked
conformations of 12a, 12b, 8a, and 8b, shows that,
when compared to 8a, 8b adopts a shape that more closely
replicates the interactions of the enzyme with Cbz-Pro-
prolinal (Figure 8). This observation is consistent with the
measured inhibitory potencies of these inhibitors on POP,
where we noted that 8b was more potent than all the other
bicyclic structures. Interestingly, the binding mode and more
specifically of the position of the proline ring of 12b and 8b
are fairly similar while the binding mode of 12a and 8a are
quite different as are the measured potencies. In practice, the
cyclization of 12b corresponds to the substitution of the two
hydrogen atoms highlighted by gray spheres of Cbz-D-Pro-
prolinal in Figure 9 by a covalent bond and to the production
of 8b. In contrast, the proline β-carbon (shown in yellow)
is not positioned to be covalently linked to the γ-carbon of
the Cbz-L-Pro-prolinal. As a result, when 12a is cyclized
into 8a, 8a cannot adopt the bioactive conformation of 12a.
Further docking studies with 31c, an epimer of 31a, re-
vealed that the binding mode should be very similar to Cbz-
Pro-prolinal and 31c is therefore expected to be a tight binder.
Unfortunately, all our synthetic efforts to make 31c proved
unsuccessful. This last observation will now guide our future
synthetic efforts to make constrained bicyclic POP inhibitors.
POP Activity and Selectivity in Intact Cells. The inhibitory
potency of these compounds was maintained in intact living
cells (Table 2). These experiments demonstrate that the
developed inhibitors are penetrating the cells, a key feature
if POP inhibitors have to be efficient as therapeutic agents for
neurodegenerative diseases because their target enzyme is
intracellular. Moreover, when assessed for inhibition toward
DPP IV activity, none of the compounds exhibited inhibitions
at concentration as high as 100 μM. This data clearly demon-
strates the high selectivity of 8a and 8b for POP activity.
Overall, these molecules achieved excellent selectivity for
inhibition of POP endoproteolytic activity over inhibition of
DPP IV exoproteolytic activity in soluble enzyme activities
extracted from relevant cell models and in human intact
living astrocyte-derived cells, human brain-derived endothe-
lial cells, and human primary fibroblasts.
ments leading to more potent inhibitors of prolyl oligopeptidase.
We first used a novel feature of our docking program FITTED to
dock covalent inhibitors to the binding site of prolyl oligopepti-
dase and to select a lead scaffold for synthesis. On the basis of
this docking study, two were selected for synthesis (8a and 8b).
To validate our predictions, several analogues of different sizes,
stereochemistry, and shapes were also prepared. Exhaustive
biological evaluation confirmed our prediction and identified
8b as a potent, highly selective, and cell-permeant POP inhibitor
(IC50 of 200-700 nM). Through this series of seemingly similar
compounds, we demonstrated that scaffolding is very sensitive
to seemingly subtle changes such as the substitution of a sulfur
atom by an isosteric oxygen atom. The docking studies also
identified the configuration of the stereogenic center at the ring
junction as a limiting factor for optimal activity and will lead our
design of the second generation of more potent inhibitors.
Experimental Section
Synthesis. All commercially available reagents were used with-
˚
out further purification unless otherwise stated. 4 A mole-
cular sieves were dried at 100 °C prior to use. Optical rota-
tions were measured on a JASCO DIP 140 in a 1 dm cell at
20 °C. FTIR spectra were recorded using a Perkin-Elmer Spec-
1
trum One FT-IR. H and 13C NMR spectra were recorded on
VarianMercury400MHz, 300MHz, orUnity500 spectrometers.
Chemical shifts are reported in ppm using the residual of chloro-
form as internal standard (7.26 ppm for 1H and 77.160 ppm for
13C, respectively). Thin layer chromatography visualization was
performed by UV or by development using KMnO4, H2SO4/
MeOH or Mo/Ce solutions. Chromatography was performed on
silica gel 60 (230-40 mesh). Low resolution mass spectrometry
was performed by ESI using a Thermoquest Finnigan LCQDuo.
High resolution mass spectrometry was performed by EI peak
matching (70 eV) on a Kratos MS25 RFA double focusing mass
spectrometer or by ESI on a IonSpec 7.0 T FTMS at McGill
University. Prior to biological testing, reversed phase HPLC was
used to verify the purity of compounds on an Agilent 1100 series
instrument equipped with VWD-detector, C18 reverse column
(Agilent, Zorbax Eclipse XDB-C18 150 mm ꢀ 4.6 mm, 5 μm),
UVdetectionat254nm. AllmeasuredpuritiesarelistedinTable4
in the Experimental Section “HPLC Analysis of Purity.” All
tested compounds were g95% pure, except for compounds 14,
17a, and 24b (g90% pure), which were not active according to
our biological tests.
(2S)-N-(tert-Butoxycarbonyl)proline Methyl Ester.42 To a
solution of L-proline methyl ester hydrochloride (1.00 g, 6.06
mmol) in 10 mL of dry CH2Cl2 was added Boc2O (2.90 g,
13.3 mmol), Et3N (0.92 mL, 6.61 mmol), and DMAP (0.81 g,
6.64 mmol) at 0 °C under Ar. After 18 h, the crude mixture was
washed with aqueous HCl 0.5 M, saturated aqueous NaHCO3
solution, H2O, and brine. The organic phase was dried over
Conclusion
We have used a series of constrained peptidomimetics to
better understand some important inhibitor shape require-

6680 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Table 4. HPLC Analysis of Purity
Lawandi et al.
boxylic acid (4.26 g, 18.6 mmol) and Et3N (2.60 mL, 18.7 mmol)
in dry THF (25 mL) under Ar at -15 °C was added dropwise a
solution of ethyl chloroformate. The resulting reaction mixture
was stirred for 1 h. Then 28% aqueous NH3 (6 mL) was then
added dropwise and the solution was warmed to rt and stirred
for another 4 h. The solution was concentrated and the residue
extracted with EtOAc, washed with 10% aqueous citric acid, a
saturated aqueous NaHCO3 solution, then brine, dried over
Na2SO4, filtered, and concentrated in vacuo to give rac-N-(tert-
butoxycarbonyl)-piperidine-2-carboxamide (3.53 g, 83% yield,
oil). 1H NMR (400 MHz, CDCl3) δ 6.12 (m, 2H), 4.70 (m, 1H),
4.11 (m, 1H), 2.70 (m, 1H), 2.25 (m, 1H), 1.46 (s, 9H), 1.70-1.38
(m, 5H).
compd analysis condition retention time (min.)
purity (%)
11a
11b
12a
12b
14
B
B
B
B
B
A
A
D
D
C
C
A
A
A
B
B
B
B
4.3
4.5
5.9
5.7
8.7
3.8
3.3
5.2
4.7
7.3
6.7
3.5
3.8
3.0
3.7
7.6
5.8
6.7
98.7
98.8
98.0
98.2
92.5
17a
17b
24a
24b
8a
91.9 (9:1 of 17a:17b)
96.7 (1:9 of 17a:17b)
98.0
93.6
97.9
99.7
rac-2-Piperidine-2-carbonitrile (HCl Salt).46 A solution of
rac-N-(tert-butoxycarbonyl)-piperidine-2-carboxamide (0.100 g,
0.48 mmol) in 1,4-dioxane (5 mL) was treated with 4 N HCl
for 45 min at rt. The reaction mixture was then filtered, washed
with Et2O, and concentrated in vacuo to afford the HCL salt of
rac-2-piperidine-2-carbonitrile (0.064 g, 92% yield, white
powder). H NMR (400 MHz, CD3OD) δ 3.84 (d, 1H), 3.38
(d, 1H), 3.03 (t, 1H), 2.24 (d, 1H), 1.94-1.86 (m, 2H), 1.72-1.69
(m, 3H).
8b
8c
98.9 (18:1 of 8c:8d)
96.0 (1:4 of 8c:8d)
95.8 (1:10 of 9a:9b)
98.1 (1:21 of 29a:29b)
95.7
8d
9a,b
29a,b
rac-30
31a
31b
1
97.6
97.3
Na2SO4, filtered, concentrated in vacuo, and purified by flash
chromatography (petroleum ether/ethyl acetate; 85/15) to give
(2S)-N-(tert-butoxycarbonyl)proline methyl ester (0.552 g, 40%
yield). 1H NMR (CDCl3, 300 MHz) δ 4.21 (dd, lH), 3 70 (s, 3H),
3.58-3.30 (m, 2H), 2.30-1.75 (m, 4H), 1.40 (s, 9H).
General Procedure for Dipeptide Coupling. Anhydrous Et3N
(2 equiv) was added dropwise to a stirred solution of amino acid
(1 equiv) (TFA or HCl salt), N-benzyloxycarbonyl-glycine or
N-benzyloxycarbonyl-alanine or N-benzyloxycarbonyl-D-ala-
nine (1 equiv) and HOBt (0.6 equiv) in CH2Cl2 under Ar.
The resulting solution was cooled to 0 °C. A solution of
1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride
(1.05 equiv) in CH2Cl2 was then added dropwise. The reaction
mixture was warmed up to rt, stirred until completion (tlc
monitoring, 16 to 20 h), and then washed with 0.2 N aqueous
HCl, saturated aqueous NaHCO3, dried over Na2SO4, filtered,
and concentrated in vacuo. Flash chromatography afforded the
dipeptides.
(2S)-N-(tert-Butoxycarbonyl)pyrrolidine-2-carboxamide.43 (2S)-
N-(tert-Butoxycarbonyl)proline methyl ester (0.552 g, 2.4 mmol)
was heated in a mixture of THF and 28% aqueous NH3 (2/9, v/v)
at 60 °C for 18 h under Ar. The reaction mixture was concentrated
to give Boc-L-Pro-NH2 (0.460 g, 89%), used in the next step
1
without any further purification. H NMR (400 MHz, CDCl3)
δ 7.50-7.01 (m, 2H), 4.87 (m, 1H), 4.35 (m, 1H), 3.44 (m, 1H),
2.00-1.75 (m, 4H), 1.43 (s, 9H).
(2S)-N-(tert-Butoxycarbonyl)-pyrrolidine-2-carbonitrile.43
TFA anhydride (0.39 mL, 2.80 mmol) was added to a solution
of (2S)-N-(tert-butoxycarbonyl)pyrrolidine-2-carboxamide (0.501
g, 2.33 mmol) and Et3N (0.779 mL, 5.59 mmol) in dry THF
(6 mL) under Ar at 0 °C. The reaction was stirred at 0 °C for
another 6 h. After addition of CH2Cl2, the organic phase was
washed with a saturated aqueous NaHCO3 solution, dried over
Na2SO4, filtered, and concentrated in vacuo to afford (2S)-
N-(tert-butoxycarbonyl)-pyrrolidine-2-carbonitrile (0.393 g, 86%,
brown oil) used in the next step without any further purification.
1H NMR (400 MHz, CDCl3) δ 4.43 (m, 1H), 3.54-3.30 (m, 2H),
2.30-1.98 (m, 4H), 1.49 (s, 9H).
N-Benzyloxycarbonylamino-glycine-2S-pyrrolidine-2-carbo-
nitrile (11a). Following the general procedure, (2S)-pyrrolidine-
2-carbonitrile (TFA salt) (0.080 g, 0.38 mmol) reacted with Cbz-
Gly-OH afforded after flash chromatography (CH2Cl2/acetone;
gradient of 98/2 to 95/5) N-benzyloxycarbonylamino-glycine-
2S-pyrrolidine-2-carbonitrile (11a) (63 mg, 58%). 1H NMR
(CDCl3, 300 MHz) δ 7.41-7.27 (m, 5H), 5.75 (s, 1H), 5.12
(s, 2H), 4.74 (m, 1H), 4.05 (dd, 1H), 3.94 (dd, 1H), 3.59 (m, 1H),
3.41 (m, 1H), 2.35-2.07 (m, 4H). 13C NMR (CDCl3, 75 MHz)
δ 167.6, 156.5, 136.4, 128.6, 128.3, 128.1, 118.1, 67.1, 46.7, 45.5,
43.6, 30.0, 25.2. HRMS (ESIþ) calcd for [C15H17N3O3 þ Na]þ:
310.11621. Found: 310.11591.
N-Benzyloxycarbonylamino-L-alanine-2S-pyrrolidine-2-carbo-
nitrile (12a). Following the general procedure, (2S)-pyrrolidine-
2-carbonitrile (TFA salt) (0.075 g, 0.36 mmol) reacted with Cbz-
L-Ala-OH afforded, after flash chromatography (petroleum
ether/ethyl acetate; gradient of 5/2 to 5/3), N-benzyloxycarbo-
nylamino-L-alanine-2S-pyrrolidine-2-carbonitrile (12a) (37 mg,
34%). 1H NMR (CDCl3, 300 MHz) δ 7.39-7.29 (m, 5H), 5.59
(d, 1H), 5.08 (dd, 2H), 4.78 (m, 1H), 4.48 (m, 1H), 3.66 (m, 2H),
2.34-2.11 (m, 3H), 1.65 (m, 1H,), 1.38 (d, 3H). 13C NMR
(CDCl3, 75 MHz) δ 171.8, 155.8, 136.4, 128.7, 128.3, 128.1,
118.2, 67.1, 48.4, 46.6, 46.5, 29.9, 25.4, 18.4. HRMS (ESIþ)
calcd for [C16H19N3O3 þ Na]þ: 324.13186. Found: 324.13168.
N-Benzyloxycarbonylamino-D-alanine-2S-pyrrolidine-2-carbo-
nitrile (12b). Following the general procedure, (2S)-pyrrolidine-
2-carbonitrile (TFA salt) (0.080 g, 0.38 mmol) reacted with Cbz-
D-Ala-OH afforded after flash chromatography (hexanes/ethyl
acetate; gradient of 5/2 to 6/4) N-benzyloxycarbonylamino-D-
alanine-2S-pyrrolidine-2-carbonitrile (12b) (56 mg, 49%). Mix-
ture of rotamers. 1H NMR (CDCl3, 400 MHz) δ 7.41-7.28
(m, 5H), 5.70 (d, 1H), 5.40 (d, 1H), 5.11 (m, 2H), 4.67 (m, 1H),
4.53 (m, 1H), 3.87-3.78 (m, 1H), 3.65-3.42 (m, 1H), 2.42-2.06
(m, 4H), 1.45(d, 3H), 1.34 (d, 3H). 13C NMR(CDCl3, 75 MHz) δ
171.5; 155.6, 136.3, 128.7, 128.3, 128.1, 118.1, 67.2, 67.1, 48.5,
(2S)-Pyrrolidine-2-carbonitrile (TFA Salt).44 (2S)-N-(tert-
Butoxycarbonyl)-pyrrolidine-2-carbonitrile (0.202 g, 1.03 mmol)
was dissolved in dry CH2Cl2 (5 mL) under Ar and TFA (2 mL)
was added at 0 °C. The reaction was stirred at 0 °C for 2 h. The
solvent was evaporated and the product was crystallized in
Et2O, yielding the white TFA salt of (2S)-pyrrolidine-2-carbo-
nitrile (0.110 g, 51% yield, white powder). 1H NMR (400 MHz,
CD3OD) δ 4.92 (s, 1H), 4.68 (t, 1H), 3.47-3.35 (m, 2H),
2.52-2.43 (m, 1H), 2.32-2.07 (m, 3H).
rac-N-(tert-Butoxycarbonyl)-piperidine-2-carboxylic Acid.45
DL-Pipecolinic acid (3.00 g, 23.2 mmol) and Et3N (3.24 mL,
23.2 mmol) were dissolved in 1,4-dioxane (32 mL) and H2O
(21 mL). Boc2O (5.58 g, 25.5 mmol) was added, and the solution
was stirred for 24 h at rt. The solvent was evaporated, and the
residue was extracted in EtOAc. The solution was washed with
aqueous 5% HCl and brine. The organic layer was dried over
Na2SO4, filtered, and evaporated to obtain rac-N-(tert-
butoxycarbonyl)-piperidine-2-carboxylic acid (4.39 g, 82%
yield, white powder). 1H NMR (CDCl3) δ 10.93 (s, 1H),
4.92-4.74 (bd, 1H), 3.98-3.91 (m, 1H), 2.95 (m, 1H), 2.20
(s, 1H), 1.66-1.23 (m, 14H).
rac-N-(tert-Butoxycarbonyl)-piperidine-2-carboxamide.43 To
a
solution of rac-N-(tert-butoxycarbonyl)-piperidine-2-car-

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6681
48.4, 47.4, 46.9, 46.5, 46.4, 32.4, 30.1, 25.2, 23.2, 18.6, 18.2.
HRMS (ESIþ) calcd for [C16H19N3O3 þ Na]þ: 324.13186.
Found: 324.13189.
16 (1.10 g, 4.15 mmol) afforded the corresponding bicyclic
scaffolds 17a (434 mg, 30%) and 17b (265 mg, 18%). 17a: H
1
NMR (300 MHz, CDCl3) δ 7.36 (s, 5H), 5.23 (bd, 1H),
5.19-5.17 (m, 1H), 5.13 (s, 2H), 5.08-5.04 (m, 1H), 4.50-4.42
(q, 1H), 3.78 (s, 3H), 3.53-3.36 (m, 2H), 2.77-2.70 (m, 1H),
2.49-2.39 (m, 1H). 13C NMR (75 MHz, CDCl3) δ 175.0, 170.2,
156.1, 136.1, 128.6, 128.3, 128.3, 67.3, 64.0, 58.6, 53.1, 52.2, 37.1,
30.0. HRMS (EIþ) calcd for [C16H18N2O5S]þ: 350.09364.
N-Benzyloxycarbonylamino-glycine-2R-pyrrolidine-2-carboni-
trile (11b). Following the general procedure, (2R)-pyrrolidine-2-
carbonitrile (TFA salt) (0.034 g, 0.16 mmol) reacted with Cbz-
Gly-OH afforded after flash chromatography (CH2Cl2/acetone
gradient of 98/2 to 95/5) N-benzyloxycarbonylamino-glycine-
2R-pyrrolidine-2-carbonitrile (11b) (31 mg, 66%). 1H NMR
(CDCl3, 400 MHz) same as 11a. 13C NMR (CDCl3, 75 MHz)
same as 11a. HRMS (ESIþ) calcd for [C15H17N3O3 þ Na]þ:
310.11621. Found: 310.11621.
N-Benzyloxycarbonylamino-glycine-rac-piperidine-2-carboni-
trile (14). Following the general procedure, rac-2-piperidine-2-
carbonitrile (HCl salt) (0.100 g, 0.691 mmol) reacted with Cbz-
Gly-OH afforded after flash chromatography (CH2Cl2/acetone
(gradient of 98/2 to 95/5) N-benzyloxycarbonylamino-glycine-
rac-piperidine-2-carbonitrile (14) (31 mg, 30%). 1H NMR
(CDCl3, 400 MHz) δ 7.42-7.29 (m, 5H), 5.75 (bs, 1H), 5.68
(bs, 1H), 5.13 (s, 2H), 4.13 (dd, 1H), 3.98 (dd, 1H), 3.70 (m, 1H),
3.31 (m, 1H), 2.08-1.40 (m, 6H). 13C NMR (CDCl3, 125 MHz)
δ 167.4, 156.4, 136.4, 128.7, 128.2, 117.1, 67.2, 42.9, 42.4, 41.8,
28.4, 25.1, 20.5. HRMS (ESIþ) calcd for [C16H19N3O3 þ Na]þ:
324.13186. Found: 324.13187.
1
Found: 350.09310. 17b: H NMR (300 MHz, CDCl3) δ 7.35
(s, 5H), 5.35 (bd, 1H), 5.16-5.14 (m, 2H), 5.12 (s, 2H), 4.67
(q, 1H), 3.77 (s, 3H), 3.37-3.35 (m, 2H), 3.26-3.15 (m, 1H),
2.10-2.00 (m, 1H). 13C NMR (75 MHz, CDCl3) δ 171.3, 169.6,
156.05, 136.1, 128.7, 128.4, 128.2, 67.2, 61.7, 57.8, 54.7, 53.1,
38.5, 35.1. HRMS (EIþ) calcd for [C16H18N2O5S]þ: 350.09364.
Found: 350.09268.
(2S,5R,7S)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-thiabi-
cyclo[3.3.0]octane-2-carboxylic Acid Methyl Ester (17c) and
(2S,5R,7R)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-thiabicy-
clo[3.3.0]octane-2-carboxylic Acid Methyl Ester (17d). Follow-
ing the general procedure and treatment with TMSCH2N2
(5 mL, excess), D-Cys (1.00 g, 6.35 mmol) and rac-16 (1.10 g,
4.15 mmol) afforded the corresponding bicyclic scaffolds 17c
(194 mg, 13%) and 17d (291 mg, 20%). The characterization
data are the same as for 17a and 17b.
General Procedure for the Formation of the Bicyclic Scaffolds.
Synthesis of rac-16. To a suspension of allylglycine (rac-15, 550
mg, 4.78 mmol) in 11 mL of dry MeOH at 0 °C was added
TMSCl (1.9 mL, 15.0 mmol). The resulting mixture was stirred
at 0 °C for 2 h. After stirring for another 12 h at rt, the mixture
was cooled to 0 °C and Et3N (2.70 mL, 19.4 mmol) was slowly
added, followed by CBzCl (0.80 mL, 5.68 mmol). The reaction
was stirred for 4 h at 0 °C, concentrated in vacuo, dissolved in
2 M HCl, and extracted with EtOAc. The solution was then
dried over Na2SO4, filtered, concentrated in vacuo, and used
without any further purification in the next reaction.
(2S,5S,7S)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-oxabi-
cyclo[3.3.0]octane-2-carboxylic Acid Methyl Ester and (18a) and
(2R,5S,7R)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-oxabicy-
clo[3.3.0]octane-2-carboxylic Acid Methyl Ester (18b). Follow-
ing the general procedure, L-Ser-OMe (0.250 g, 1.61 mmol) and
rac-16 (0.250 g, 0.943 mmol) afforded the corresponding bicyclic
scaffolds 18a and 18b as an inseparable mixture (154 mg, 49%).
1H NMR (400 MHz, CDCl3) δ 7.35 (s, 5H), 5.30 (bs, 1H),
5.17-5.19 (d, 1H), 5.13 (s, 2H), 5.09-5.07 (d, 1H), 4.49-4.43
(q, 1H) 3.78 (m, 1H), 2.97-2.90 (m, 1H), 2.69-2.64 (m, 1H),
2.52-2.29 (m, 3H), 1.95-1.86 (m, 1H). 13C NMR (125 MHz,
CDCl3) δ 177.4, 174.1, 170.4, 170.0, 156.1, 136.2, 128.7, 128.4,
90.8, 89.3, 89.2, 70.0, 67.4, 57.2, 55.2, 54.1, 53.0, 52.9, 52.5, 36.1,
31.1. HRMS (ESIþ) calcd for [C16H18N2O6 þ Na]þ: 357.10571.
Found: 357.10533.
Ozone was bubbled through a solution of the oil in CH2Cl2
(50 mL) at -78 °C for 30 min. Excess of ozone was removed by
bubbling argon through the solution. Dimethyl sulfide (2 mL,
excess) was then added. The resulting mixture was stirred at rt
for 16 h, concentrated in vacuo, redissolved in EtOAc, washed
with saturated aqueous NaHCO3, dried over Na2SO4, filtered,
and concentrated to give a colorless oil which was pure enough
to be used in the next step without further purification.
General Procedure for Scaffold Synthesis. A solution of
L-cysteine methyl ester (or D-cysteine, L-serine methyl ester,
L-threonine methyl ester rac-homocysteine) hydrochloride (1.2
to 1.7 equiv) and aldehyde (rac-16) (1.0 equiv) in pyridine was
stirred at rt for 15 h under Ar and then at 50 °C for 3 days. The
excess pyridine was evaporated. When L-cysteine methyl ester,
L-serine methyl ester, and L-threonine methyl ester were used,
the residue was redissolved in CH2Cl2, washed with 5% aqueous
HCl, saturated aqueous NaHCO3, water, and brine, dried over
Na2SO4, filtered, and concentrated in vacuo. When D-cysteine
and rac-homocysteine were reacted, the residue was acidified to
about pH 2 with 1N aqueous HCl and partitioned between
CH2Cl2 and water. The organic layer was washed with brine,
dried over Na2SO4, filtered, and concentrated in vacuo. The
residue was redissolved in THF and treated with a 2.0 M
(trimethylsilyl)diazomethane (TMSCH2N2) solution for 16 h.
The solution was then stirred with saturated aqueous NaHCO3
and extracted with CH2Cl2. The organic phase was dried over
Na2SO4, filtered, and concentrated in vacuo. All crude residues
were purified by flash chromatography (toluene/diethyl ether,
gradient of 9:1 to 8:2) to yield the corresponding bicyclic
scaffolds.
(2S,3R,5S,7S)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-oxa-
6-methylbicyclo[3.3.0]octane-2-carboxylic Acid Methyl Ester and
(19a) (2S,3R,5S,7R)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-
oxa-6-methylbicyclo[3.3.0]octane-2-carboxylic Acid Methyl Ester
(19b). Following the general procedure, L-Thr-OMe (0.680 g,
4.01 mmol) and rac-16 (0.825 g, 3.11 mmol) afforded the
corresponding bicyclic scaffolds 19a and 19b as an inseparable
1
mixture (380 mg, 35%). H NMR (400 MHz, CDCl3) δ 7.33
(s, 8H), 5.60 (bs, 0.5 H), 5.49 (bs, 1H), 5.20-5.19 (bd, 2H), 5.11
(s, 4H), 4.69-4.67 (q, 1H), 4.43-4.42 (q, 0.5 H), 4.26-4.17
(m, 1.5H), 4.11-4.09 (m, 1.5H), 3.76 (d, 6H), 3.08-3.04 (m,
1H), 2.64-2.58 (m, 0.5H), 2.28-2.22 (m, 0.5 H), 1.96-1.89 (m,
2H), 1.49-1.48 (d, 3H), 1.43-1.42 (d, 2H). 13C NMR (100
MHz, CDCl3) δ 177.40, 173.76, 170.19, 169.74, 156.07, 136.14,
128.63, 128.32, 128.25, 128.20, 90.60, 88.79, 79.82, 79.39, 67.28,
67.22, 63.97, 61.59, 54.03, 52.87, 52.45, 35.89, 31.09, 19.85, 19.39.
HRMS (ESIþ) calcd for [C17H20N2O6 þ Na]þ: 371.12136.
Found: 371.12086.
(2R,6R,8R)-1-Aza-8-benzyloxycarbonylamino-9-oxo-5-thiabic-
yclo[4.3.0]nonane-2-carboxylic Acid Methyl Ester (rac-20).
Following the general procedure and treatment with TMSCH2N2
(5 mL, excess), rac-homoCys (0.60 g, 3.50 mmol) and rac-16
(0.568 g, 2.14 mmol) afforded the corresponding bicyclic scaf-
folds rac-20 (250 mg, 32%). 1H NMR (400 MHz, CDCl3)
δ 7.36-7.35 (m, 5H), 5.29 (bd, 1H), 5.19, (d, 1H), 5.12 (s, 2H),
5.07 (dd, 1H), 4.46 (q, 1H), 3.78 (s, 3H), 2.96-2.90 (m, 1H),
2.68 (dt, 1H), 2.49 (m, 1H), 2.39-2.30 (m, 2H), 1.95-1.86
(m,1H). 13C NMR (75 MHz, CDCl3) δ 171.5, 170.3, 156.3,
136.2, 128.7, 128.4, 128.3, 67.3, 54.9, 52.9, 51.9, 51.4, 33.2, 26.8,
25.4. HRMS (ESIþ) calcd for [C17H20N2O5S þ H]þ: 365.11657.
Found: 365.11685.
(2R,5S,7S)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-thiabi-
cyclo[3.3.0]octane-2-carboxylic Acid Methyl Ester (17a) and
(2R,5S,7R)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-thiabicy-
clo[3.3.0]octane-2-carboxylic Acid MethylEster (17b). Following
the general procedure, L-Cys-OMe (1.00 g, 5.84 mmol) and rac-

6682 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Lawandi et al.
(S)-Benzyl 5-Oxo-4-(3-oxopropyl)oxazolidine-3-carboxylate
(22a) and (R)-Benzyl 5-Oxo-4-(3-oxopropyl)oxazolidine-3-car-
boxylate (22b). To a solution of L- or D-glutamic acid (5.0 g, 34.0
mmol) in 4 N NaOH (17 mL) at 0 °C was added benzyl
chloroformate (7.8 mL, 54.7 mmol) dropwise over a period of
10 min and then stirred at rt for 16 h. The solution was extracted
with diethyl ether, acidified to pH 2 with 4N HCl, and extracted
with ethyl acetate. The organic phase was then dried over
Na2SO4, filtered, and concentrated in vacuo (5.6 g, 59%, oil).
1H NMR (300 MHz, acetone-d6) δ 7.29-7.41 (m, 5H),
6.64-6.66 (d, 1H), 5.09 (s, 2H), 4.29-4.36 (m, 1H), 2.47-2.53
(m, 2H), 2.17-2.28 (m, 1H), 1.98-2.05 (m, 1H). To a solution of
the crude oil (5.6 g, 20.1 mmol) in toluene (125 mL), under Ar,
was added paraformaldehyde (4.02 g, excess) and p-toluenesul-
fonic acid (0.40 g, 2.12 mmol equiv). The mixture was refluxed
for 3 h with removal of water using a Dean-Stark and then
filtered through a pad of silica and concentrated in vacuo. The
crude product was purified by flash chromatography (Hex/
EtOAc/AcOH, gradient of 8/2/0.01 to 1/1/0.01) to yield the
desired intermediate oxazolidinone (3.02 g, 51%, oil). 1H NMR
(400 MHz, CDCl3) δ 7.39-7.35 (m, 5H), 5.55 (br, 1H),
5.24-5.19 (m, 3H), 4.42-4.39 (t, 1H), 2.57-2.15 (m, 4H). To
a stirred solution of the intermediate oxazolidinone (2.08 g, 7.1
mmol) in CH2Cl2 (50 mL) under Ar were added distilled oxalyl
chloride (0.91 mL, 10.6 mmol) and dimethylformamide (0.055
mL, 0.70 mmol). The solution was stirred for 1 h at rt and then
the solvents were evaporated to give the crude product (quant.
yield), which was immediately dissolved in dry THF (50 mL)
and cooled to -78 °C. To the stirred solution was added lithium
tri-tert-butoxyaluminum hydride (1.89 g, 7.45 mmol) slowly.
The mixture was stirred for 4 h, slowly warming to 0 °C,
quenched with water, and then filtered through celite. The
filtrate was extracted with CH2Cl2, washed with water and
brine, dried over Na2SO4, and concentrated in vacuo to afford
the aldehyde 22a or 22b, which were used without further
purification (1.27 g, 65%, orange oil). 1H NMR (400 MHz,
CDCl3) δ 9.70 (br, 1H), 7.41-7.33 (m, 5H), 5.53 (br, 1H),
5.24-5.15 (m, 3H), 4.39-4.36 (t, 1H), 2.70-2.15 (m, 4H).
(2R,5S,8S)-1-Aza-8-benzyloxycarbonylamino-9-oxo-4-thiabi-
cyclo[3.4.0]nonane-2-carboxylic Acid Methyl Ester 23a. Follow-
ing the general procedure, L-Cys-OMe (0.45 g, 2.6 mmol) and
22a (0.60 g, 2.16 mmol) provided the cyclized intermediate as an
oil. To a solution of this crude mixture in MeOH (25 mL) was
added K2CO3 (0.40 g, 2.9 mmol). After stirring for another 4 h,
the reaction was quenched with water and extracted with
CH2Cl2. The organic phase was washed with brine, dried over
Na2SO4, filtered, and concentrated in vacuo. The crude product
was purified by flash chromatography (Hex/EtOAc, gradient
6/4 to 1/1) to afford the corresponding bicyclic scaffolds 23a
(116 mg, 32%). 1H NMR (300 MHz, CDCl3) δ 7.36-7.31 (m,
5H), 5.51 (br, 1H), 5.13-5.08 (m, 3H), 4.97-4.90 (m, 1H),
4.33-4.16 (m, 1H), 3.76 (s, 3H), 3.41-3.33 (m, 1H),
3.28-3.21 (m, 1H), 2.65-2.53 (m, 1H), 2.43-1.74 (m, 4H).
HRMS (EIþ) calcd for [C17H20N2O5S]þ: 364.10929. Found:
364.10896.
(2R,5S,7S)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-thiabi-
cyclo[3.3.0]octane-2-carboxamide(24a,d). H NMR (400 MHz,
1
(CD3)2CO) δ 7.38-7.36 (m, 5H), 7.11 (bd, 2H), 6.76 (bs, 1H),
5.24-5.23 (dd, 1H), 5.10 (s, 2H), 4.81 (t, 1H), 4.37-4.29 (q, 1H),
3.48-3.47 (m, 2H), 2.58-2.52 (m, 2H). 13C NMR (300 MHz,
(CD3)2CO) δ 174.9, 171.9, 156.8, 137.8, 129.2, 128.8, 128.7, 67.0,
64.9, 60.0, 53.4, 37.0, 31.5. LRMS (ESIþ) calcd for
[C15H17N3O4S þ Na]þ: 358.08. Found: 358.17.
(2R,5S,7R)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-thiabi-
1
cyclo[3.3.0]octane-2-carboxamide (24b,c). H NMR (400 MHz,
(CD3)2CO) δ 7.44-7.26 (m, 5H), 7.22 (bs, 1H), 6.75-6.72 (d,
1H), 6.64 (bs, 1H), 5.16 (t, 1H), 5.10 (s, 2H), 4.91-4.89 (m, 1H),
4.84-4.77 (m (1H), 3.52-3.48 (dd, 1H), 3.91-3.24 (m, 1H),
3.07-3.01 (m, 1H), 2.14-2.05 (m, 1H). 13C NMR (75 MHz,
CDCl3) δ 172.4, 171.4, 156.9, 138.0, 129.2, 128.7, 66.8, 62.0, 60.3,
55.2, 38.2, 34.8. LRMS (ESIþ) calcd for [C15H17N3O4S þ Na]þ:
358.08. Found: 358.15.
(2R,5S,8S)-1-Aza-8-benzyloxycarbonylamino-9-oxo-4-thiabi-
cyclo[3.4.0]nonane-2-carboxamide (28a). H NMR (300 MHz,
1
CDCl3) δ 7.41-7.28 (m, 5H), 7.10 (bs, 1H), 5.79 (d, 1H),
5.33-5.28 (m, 2H), 5.08-5.02 (m, 2H), 4.88 (dd, 1H), 3.69
(dd, 1H), 3.53 (dd, 1H), 3.26 (dd, 1H), 2.34 (ddd, 1H), 2.23
(dt, 2H), 1.79 (ddd, 1H). 13C NMR (75 MHz, CDCl3) δ 171.4,
167.7, 156.7, 135.9, 128.7, 128.5, 128.1, 67.5, 62.1, 61.9, 52.1,
30.8, 28.6, 27.4. HRMS (EIþ) calcd for [C16H19N3O4S]þ:
349.10963. Found: 349.10893.
(2R,5S,8R)-1-Aza-8-benzyloxycarbonylamino-9-oxo-4-thiabi-
cyclo[3.4.0]nonane-2-carboxamide (28b). 1H NMR (400 MHz,
CDCl3) δ 7.40-7.30 (m, 4H), 6.45 (s, 1H), 5.63 (s, 1H), 5.35 (s,
1H), 5.21-5.15 (m, 1H), 5.13 (s, 2H), 4.86 (t, 1H), 4.34-4.20 (m,
1H), 3.67 (dd, 1H), 3.12 (dd, 1H), 2.48-2.38 (m, 1H), 2.38-2.25
(m, 1H), 2.13-2.02 (m, 1H), 1.87-1.74 (m, 1H). 13C NMR (75
MHz, CDCl3) δ 171.0, 169.8, 156.3, 136.3, 128.7, 128.4, 128.3,
67.2, 61.8, 61.1, 51.6, 30.2, 29.8, 25.3, 24.9. HRMS (EIþ) calcd
for [C16H19N3O4S]þ: 349.10963. Found: 349.10863.
General Procedure for Formation of the Nitriles. To a stirred
solution of the amide (either 24a,d, 24b,c, 25a,b, 26a,b, rac-27,
28a, or 28b) in dry THF, under Ar, at 0 °C, was added
TFA anhydride (1.2 equiv) and Et3N (2.0 equiv). The reac-
tion was stirred for 1 to 3 h at rt. The solution was extracted
with CH2Cl2, washed with a saturated NaHCO3, dried over
Na2SO4, and concentrated in vacuo. Further purification of the
residue by flash column chromatography afforded the nitrile
derivative.
(2R,5S,7S)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-thiabi-
cyclo[3.3.0]octane-2-carbonitrile (8a) and (2S,5R,7R)-1-Aza-7-
benzyloxycarbonylamino-8-oxo-4-thiabicyclo[3.3.0]octane-2-car-
bonitrile (8d). Following the general procedure for formation of
the nitrile, 24a (238 mg, 0.71 mmol) or 24d (90 mg, 0.12 mmol)
afforded 8a (114 mg, 51%) or 8d (69 mg, 81%). 1H NMR
(400 MHz, (CD3)2CO) δ 7.42-7.29 (m, 5H), 7.07 (d, 1H),
5.42-5.41 (m, 1H), 5.40 (d, 1H), 5.10 (s, 2H), 4.39 (q, 1H),
3.74-3.68 (m, 1H), 3.52-3.48 (m, 1H), 2.71-2.55 (m, 2H).
13C NMR (75 MHz, (CD3)2CO) δ 175.6, 156.8, 137.9,
129.2, 128.8, 118.1, 67.1, 64.5, 52.2, 48.3, 39.0. HRMS
(ESIþ) calcd for [C15H15N3O3S þ Na]þ: 340.07263. Found:
340.07229.
(2R,5S,8R)-1-Aza-8-benzyloxycarbonylamino-9-oxo-4-thiabi-
cyclo[3.4.0]nonane-2-carboxylic Acid Methyl Ester 23b. As for
23a, L-Cys-OMe (0.95 g, 5.5 mmol), 22b (1.27 g, 4.60 mmol), and
then K2CO3 (0.72 g, 5.2 mmol) afforded after flash chromato-
graphy (Hex/EtOAc, gradient 6/4 to 1/1) the corresponding
bicyclic scaffolds 23b (372 mg, 25%). HRMS (EIþ) calcd for
[C17H20N2O5S]þ: 364.10929. Found: 364.10847.
General Procedure for Formation of the Amides. A stirred
solution of the ester (either 24a-d, 25a,b, 26a,b, rac-27, 28a,b) in
a solution of ammonia in methanol (ca. 7N solution) was
reacted for 1 h at rt. The corresponding amides were obtained
in quantitative yields. Storage leads to partial isomerization.
24a,d, 24b,c, 28a, and 28b were fully characterized. However, the
other amides (25a,b, 26a,b, rac-27) were reacted in the next step
without further purification.
(2R,5S,7R)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-thiabi-
cyclo[3.3.0]octane-2-carbonitrile (8b) and (2S,5R,7S)-1-Aza-7-
benzyloxycarbonylamino-8-oxo-4-thiabicyclo[3.3.0]octane-2-car-
bonitrile (8c). Following the general procedure for formation of
the nitrile, 24b (15 mg, 0.04 mmol) or 24c (55 mg, 0.16 mmol)
1
afforded 8b (11 mg, 79%) or 8c (40 mg, 77%). H NMR (400
MHz, (CD3)2CO) δ 7.38-7.25 (m, 5H), 6.80 (d, 1H), 5.55-5.51
(m, 1H), 5.75 (t, 1H), 5.08 (s, 2H), 4.79-4.72 (m, 1H), 3.49-3.37
(m, 2H), 3.14-3.07 (m, 1H), 2.26-2.18 (m, 1H). 13C NMR (75
MHz, (CD3)2CO) δ 172.4, 156.7, 138.0, 129.2, 128.7, 128.7,
117.5, 114.0, 66.9, 61.6, 55.0, 54.8, 48.1, 37.8, 36.6. HRMS
(ESIþ) calcd for [C15H15N3O3S þ Na]þ: 340.07263. Found:
340.07247.

Article
(2S,5S,7S)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-oxabi-
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6683
prepared using Maestro and SMART, a third module of
FITTED. They were subsequently docked using the FITTED
docking engine and defaults parameters. For the covalent dock-
ing to be used, the Ser554 has been selected as a reactive residue.
Biological Evaluations. The human glioblastoma-derived cell
lines LN18, LN229, and LNZ308 were a kind gift of AC Diserens,
Neurosurgery Department, Lausanne, Switzerland, the immorta-
lized human brain-derived HCEC cells have been kindly provided
by D. Stanimirovic, Ottawa, Canada, and the human fibroblasts
PO03 and PG98/5 cells have been prepared in Lausanne from
human lung and skin surgical samples, respectively, according to
protocols accepted by the Hospital Ethics Committee.
CellswereroutinelygrowninDMEMculturemedium contain-
ing 4.5gLglucose, 10%fetal calfserum (FCS), andantibiotics (all
from Gibco, Basel, Switzerland). One to two days before evalua-
tion, cells were seeded in 48-well plates (Costar, Corning, NY) in
complete medium in order to reach confluence on the day of
experiment. On the day of experiment, the culture medium was
removed, and either 200 μL phosphate-buffered saline (PBS, pH
7.2-7.4) was added in half of the wells or 200 μL PBS containing
0.1% Triton X-100 (Fluka, Buchs, Switzerland) was added in the
other half of the wells, for the evaluation of the inhibition of POP
and DPP IV activities in intact cells or cell extracts, respectively.
Experiments were performed in duplicate wells.
cyclo[3.3.0]octane-2-carbonitrile and (9a) (2R,5S,7R)-1-Aza-7-
benzyloxycarbonylamino-8-oxo-4-oxabicyclo[3.3.0]octane-2-car-
bonitrile (9b). Following the general procedure for formation of
the nitrile, 25a,b (17 mg, 0.05 mmol) afforded 9a,b (10 mg, 66%).
1H NMR (400 MHz, CDCl3) δ 7.36 (s, 6H), 5.13-5.30 (m, 3H),
5.12 (s, 3H), 4.77-4.84 (m, 1H), 4.59-4.65 (m, 1H), 4.47-4.52
(m, 1H), 4.38-4.43 (m 0.3H), 4.02-4.24 (m, 1.7H), 3.72-3.78
(m, 0.3H), 3.09-3.16 (m, 1H), 2.67-2.74 (m, 0.3H), 2.36-2.44
(m, 0.3H), 2.01-2.08 (m, 1H). 13C NMR (125 MHz, CDCl3) δ
174.4, 155.9, 135.9, 128.7, 128.5, 128.4, 128.3, 116.8, 116.4, 90.6,
89.1, 70.7, 70.6, 67.5, 53.7, 45.4, 43.4, 35.1. HRMS (ESIþ) calcd
for [C15H15N3O4 þ Na]þ: 324.09548. Found: 324.09527.
(2S,3R,5S,7S)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-oxa-
6-methylbicyclo[3.3.0]octane-2-carbonitrile and (29a) (2S,3R,5S,
7R)-1-Aza-7-benzyloxycarbonylamino-8-oxo-4-oxa-6-methylbic-
yclo[3.3.0]octane-2-carbonitrile (29b). Following the general pro-
cedure for formation of the nitrile, 26a,b (100 mg, 0.30 mmol)
afforded 29a,b (51 mg, 54%). 1H NMR (400 MHz, CDCl3) δ 7.36
(s, 5H), 5.30-5.23 (m, 2H), 5.12 (s, 2H), 4.63 (q, 1H), 4.43 (q, 1H),
4.20 (q, 1H), 3.17-3.09 (m, 1H), 2.71-2.64 (m, 0.1H), 2.41-2.33
(m,0.1H), 2.05-1.97 (m, 1H), 1.53-1.51 (d, 3H), 1.43 (d, 1H). 13
C
NMR (75 MHz, CDCl3) δ 174.4, 155.9, 136.0, 128.7, 128.5, 128.3,
116.5, 116.1, 90.5, 88.9, 80.7, 67.5, 53.5, 51.9, 49.4, 35.3, 18.9, 18.2.
HRMS (ESIþ) calcd for [C16H17N3O4 þ Na]þ: 338.11113.
Found: 338.11096.
The synthetic molecules were dissolved at 10 mg/mL in
methanol and then diluted 1:10 in H2O, and 1 or 5 μL of the
water solution were added to duplicate PBS and PBS-Triton
wells, followed after 5-10 min at room temperature by 1 μL of
Gly-Pro-AMC or Z-Gly-Pro-AMC substrates (1 mg/mL
DMSO, both from Bachem, Bubendorf, Switzerland), final
(2RS,6RS,8RS)-1-Aza-8-benzyloxycarbonylamino-9-oxo-5-thi-
abicyclo[4.3.0]nonane-2-carbonitrile (rac-30). Following the gen-
eral procedure for formation of the nitrile, rac-27 (59 mg, 0.17
mmol) afforded rac-30 (24 mg, 42%). 1H NMR (400 MHz,
CDCl3) δ 7.36 (s, 5H), 5.42 (bs, 1H), 5.20 (bs, 2H), 5.13 (s, 2H),
4.42 (q, 1H), 3.35 (t, 1H), 2.83-2.79 (m, 1H), 2.53-2.48 (m, 1H),
concentration 10 μM. Increase in fluorescence at λex/λem
=
360/460 nm was recorded for 30 min at 37 °C in a thermostatted
multiwell fluorescence reader (Cytofluor, PerSeptive BioSys-
tems, Switzerland). For the determination of IC50, cells in PBS
or in PBS-Triton X-100 were exposed to decreasing concentra-
tions of the inhibitors and then determination of residual
activity was measured and plotted against inhibitor concentra-
tion. IC50 values were determined graphically.
2.41-2.33 (m, 1H), 2.33-2.17 (m, 1H), 2.00-1.94 (m, 1H). 13
C
NMR (100 MHz, CDCl3) δ 170.5, 156.1, 136.0, 128.7, 128.5,
128.3, 115.9, 67.5, 54.4, 51.4, 41.9, 32.8, 28.2, 24.8. HRMS (ESIþ)
calcd for [C16H17N3O3S þ Na]þ: 354.08828. Found: 354.08831.
(2R,5S,8S)-1-Aza-8-benzyloxycarbonylamino-9-oxo-4-thiabi-
cyclo[3.4.0]nonane-2-carboxylic Carbonitrile (31a). Following
the general procedure for formation of the nitrile, 28a (23 mg,
1
0.07 mmol) afforded 31a (9 mg, 39%). H NMR (500 MHz,
Acknowledgment. We thank CIHR, ViroChem Pharma,
NSERC, FQRNT, and the Canadian Foundation for
Innovation for financial support (N.M.). J.L. holds a scholar-
CDCl3) δ 7.42-7.28 (m, 5H), 5.62 (s, 1H), 5.34 (dd, 1H), 5.12 (q,
2H), 4.95 (s, 1H), 4.23 (s, 1H), 3.40 (dd, 1H), 3.31 (dd, 1H), 2.53
(bd, 1H), 2.40 (bd, 1H), 1.97-1.78 (m, 2H). 13C NMR (125
MHz, CDCl3) δ 167.6, 156.6, 136.3, 128.7, 128.7, 128.6, 128.3,
128.2, 116.7, 67.2, 62.3, 52.3, 48.5, 33.2, 29.8, 27.9, 27.7. HRMS
(EIþ) calcd for [C16H17N3O3S]þ: 331.09906. Found: 331.09817.
(2R,5S,8R)-1-Aza-8-benzyloxycarbonylamino-9-oxo-4-thiabi-
cyclo[3.4.0]nonane-2-carboxylic Carbonitrile (31b). Following
the general procedure for formation of the nitrile, 28b (40 mg,
ꢀ ꢀ
ship from Le Fonds Quebecois de la Recherche sur la Nature
et les Technologies. We also thank the Swiss Society for
Multiple Sclerosis for support to L.J-J.’s group.
References
1
(1) Polgar, L. Unusual secondary specificity of prolyl oligopeptidase
and the different reactivities of its two forms toward charged
substrates. Biochemistry 1992, 31, 7729–7735.
(2) Irazusta, J.; Larrinaga, G.; Gonzalez-Maeso, J.; Gil, J.; Meana,
J. J.; Casis, L. Distribution of prolyl endopeptidase activities in rat
and human brain. Neurochem. Int. 2002, 40, 337–345.
0.11 mmol) afforded 31b (10 mg, 27%). H NMR (500 MHz,
CDCl3) δ 7.36 (s, 5H), 5.74 (s, 1H), 5.68 (s, 1H), 5.12 (s, 3H),
4.30-4.18 (m, 1H), 3.30 (dd, 2H), 2.52-2.31 (m, 2H), 2.19-2.02
(m, 1H), 1.92-1.76 (m, 1H). 13C NMR (125 MHz, CDCl3) δ
168.7, 156.1, 136.2, 128.7, 128.4, 128.3, 116.8, 67.2, 61.9, 51.8,
48.5, 33.7, 29.8, 24.9. HRMS (EIþ) calcd for [C17H20N2O5S]þ:
331.09906. Found: 331.09795.
HPLC Analysis of Purity. To verify the purity of all com-
pounds used for biological tests, analytical reverse phase HPLC
was performed on an Agilent 1100 series instrument equipped
with VWD-detector, C18 reverse column (Agilent, Zorbax
Eclipse XDB-C18 150 mm ꢀ 4.6 mm, 5 μm), UV detection at
254 nm. Depending on the compound, two different conditions
were used: condition A (50% acetonitrile, 50% water, 1 mL/
min), condition B (50% methanol, 50% water, 1 mL/min),
condition C (60% water, 40% acetonitrile, 1 mL/min), or
condition D (70% water, 30% acetonitrile, 1 mL/min) (Table 4).
Docking Study. The protein structure (1H2Y) was down-
loaded from the Protein Data Bank and prepared as reported
previously.19 It was next prepared for docking using PROCESS,
a module of our docking program FITTED.21 The ligands were
(3) Szeltner, Z.; Polgar, L. Structure, function and biological relevance
of prolyl oligopeptidase. Curr. Prot. Pept. Sci. 2008, 9, 96–107.
€
€ €
(4) Mannisto, P. T.; Vanalainen, J.; Jalkanen, A.; Garcia-Horsman,
J. A. Prolyl oligopeptidase: A potential target for the treatment of
cognitive disorders. Drug News Perspect. 2007, 20, 293–305.
(5) Brandt, I.; Scharpe, S.; Lambeir, A. M. Suggested functions for
prolyl oligopeptidase: A puzzling paradox. Clin. Chim. Acta 2007,
377, 50–61.
(6) Williams, R. S. B. Prolyl oligopeptidase and bipolar disorder. Clin.
Neurosci. Res. 2004, 4, 233–242.
(7) Cunningham, D. F.; O’Connor, B. Identification and initial char-
acterisation of
a N-benzyloxycarbonyl-prolyl-prolinal (Z-Pro-
prolinal)-insensitive 7-(N-benzyloxycarbonyl-glycyl-prolyl-amido)-
4- methylcoumarin (Z-Gly-Bro-NH-Mec)-hydrolysing peptidase in
bovine serum. Eur. J. Biochem. 1997, 244, 900–903.
(8) Toide, K.; Shinoda, M.; Iwamoto, Y.; Fujiwara, T.; Okamiya, K.;
Uemura, A. A novel propyl endopeptidase inhibitor, JTP-4819,
with potential for treating Alzheimer’s disease. Behav. Brain Res.
1997, 83, 147–151.

6684 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Lawandi et al.
(9) Toide, K.; Shinoda, M.; Miyazaki, A. A novel prolyl endopepti-
dase inhibitor, JTP-4819. Its behavioral and neurochemical prop-
erties for the treatment of Alzheimer’s disease. Rev. Neurosci. 1998,
9, 17–29.
(10) Saito, M.; Hashimoto, M.; Kawaguchi, N.; Shibata, H.; Fukami,
H.; Tanaka, T.; Higuchi, N. Synthesis and inhibitory activity of
acyl-peptidyl-pyrrolidine derivatives toward post-proline cleaving
enzyme; a study of subsite specificity. J. Enzyme Inhib. 1991, 5, 51–
75.
(11) Barelli, H.; Petit, A.; Hirsch, E.; Wilk, S.; De Nanteuil, G.; Morain,
P.; Checler, F. S 17092-1, a highly potent, specific and cell permeant
inhibitor of human proline endopeptidase. Biochem. Biophys. Res.
Commun. 1999, 257, 657–661.
(12) Toide, K.; Shinoda, M.; Fujiwara, T.; Iwamoto, Y. Effect of a
novel prolyl endopeptidase inhibitor, JTP-4819, on spatial memory
and central cholinergic neurons in aged rats. Pharmacol., Biochem.
Behav. 1997, 56, 427–434.
(13) Miyazaki, A.; Toide, K.; Sasaki, Y.; Ichitani, Y.; Iwasaki, T. Effect
of a prolyl endopeptidase inhibitor, JTP-4819, on radial maze
performance in hippocampal-lesioned rats. Pharmacol., Biochem.
Behav. 1998, 59, 361–368.
(14) Kato, A.; Fukunari, A.; Sakai, Y.; Nakajima, T. Prevention of
amyloid-like deposition by a selective prolyl endopeptidase inhi-
bitor, Y-29794, in senescence-accelerated mouse. J. Pharmacol.
Exp. Ther. 1997, 283, 328–335.
(15) Shinoda, M.; Toide, K.; Ohsawa, I.; Kohsaka, S. Specific inhibitor
for prolyl endopeptidase suppresses the generation of amyloid β
protein in NG108-15 cells. Biochem. Biophys. Res. Commun. 1997,
235, 641–645.
(28) Estiarte, M. A.; Rubiralta, M.; Diez, A.; Thormann, M.; Giralt, E.
Oxazolopiperidin-2-ones as Type II’ beta-Turn Mimetics: Synthe-
sis and Conformational Analysis. J. Org. Chem. 2000, 65, 6992–
6999.
(29) Caplan, J. F.; Sutherland, A.; Vederas, J. C. The first stereospecific
synthesis of L-tetrahydrodipicolinic acid: a key intermediate of
diaminopimelate metabolism. J. Chem. Soc., Perkin Trans. 1 2001,
2217–2220.
(30) Kamber, M.; Just, G. gamma-Phosphono-gamma-lactones. The
use of allyl esters as easily removable phosphonate protecting
groups. Can. J. Chem. 1985, 63, 823–827.
(31) Khalil, E. M.; Pradhan, A.; Ojala, W. H.; Gleason, W. B.; Mishra,
R. K.; Johnson, R. L. Synthesis and Dopamine Receptor Modu-
lating Activity of Substituted Bicyclic Thiazolidine Lactam Pepti-
domimetics of l-Prolyl-l-leucyl-glycinamide. J. Med. Chem. 1999,
42, 2977–2987.
(32) Wallen, E. A. A.; Christiaans, J. A. M.; Forsberg, M. M.;
Venalainen, J. I.; Mannisto, P. T.; Gynther, J. Dicarboxylic acid
bis(L-prolyl-pyrrolidine) amides as prolyl oligopeptidase inhibi-
tors. J. Med. Chem. 2002, 45, 4581–4584.
(33) Baldwin, J. E.; Lee, E. Synthesis of bicyclic γ-lactams via oxazo-
lidinones. Tetrahedron 1986, 42, 6551–6554.
(34) Juillerat-Jeanneret, L. Prolyl-specific peptidases and their inhibi-
tors in biological processes. Curr. Chem. Biol. 2008, 2, 97–109.
(35) Morain, P.; Lestage, P.; De Nanteuil, G.; Jochemsen, R.; Robin,
J. L.; Guez, D.; Boyer, P. A. S 17092: a prolyl endopeptidase
inhibitor as a potential therapeutic drug for memory impairment.
Preclinical and clinical studies. CNS Drug Rev. 2002, 8, 31–52.
(36) Toide, K.; Iwamoto, Y.; Fujiwara, T.; Abe, H. JTP-4819: A novel
prolyl endopeptidase inhibitor with potential as a cognitive en-
hancer. J. Pharmacol. Exp. Ther. 1995, 274, 1370–1378.
(37) Schneider, J. S.; Giardiniere, M.; Morain, P. Effects of the prolyl
endopeptidase inhibitor S 17092 on cognitive deficits in chronic low
dose MPTP-treated monkeys. Neuropsychopharmacology 2002, 26,
176–182.
(38) Maes, M. Alterations in plasma prolyl endopeptidase activity in
depression, mania, and schizophrenia: effects of antidepressants,
mood stabilizers, and antipsychotic drugs. Psychiatry Res. 1995,
58, 217–225.
(39) Mantle, D.; Falkous, G.; Ishiura, S.; Blanchard, P. J.; Perry, E. K.
Comparison of proline endopeptidase activity in brain tissue from
normal cases and cases with Alzheimer’s disease, Lewy body
dementia, Parkinson’s disease and Huntington’s disease. Clin.
Chim. Acta 1996, 249, 129–139.
(40) Tran, T.; Quan, C.; Edosada, C. Y.; Mayeda, M.; Wiesmann, C.;
Sutherlin, D.; Wolf, B. B. Synthesis and structure-activity rela-
tionship of N-acyl-Gly-, N-acyl-Sar-, and N-blocked-boroPro
inhibitors of FAP, DPP4, and POP. Bioorg. Med. Chem. Lett. 2007,
17, 1438–1442.
(41) Yoshimoto, T.; Tsuru, D.; Yamamoto, N.; Ikezawa, R.; Furukawa,
S. Structure-Activity Relationship of Inhibitors Specific for Prolyl
Endopeptidase. Agric. Biol. Chem. 1991, 55, 37–43.
(16) Morain, P.; Robin, J. L.; De Nanteuil, G.; Jochemsen, R.; Heidet,
V.; Guez, D. Pharmacodynamic and pharmacokinetic profile of S
17092, a new orally active prolyl endopeptidase inhibitor, in elderly
healthy volunteers. A phase I study. Br. J. Clin. Pharmacol. 2000,
50, 350–359.
€ €
ꢀ
(17) Szeltner, Z.; Rea, D.; Renner, V.; Fulop, V.; Polgar, L. Electro-
static effects and binding determinants in the catalysis of prolyl
oligopeptidase: Site-specific mutagenesis at the oxyanion binding
site. J. Biol. Chem. 2002, 277, 42613–42622.
(18) Li, J.; Wilk, E.; Wilk, S. Inhibition of prolyl oligopeptidase by
Fmoc-aminoacylpyrrolidine-2-nitriles. J. Neurochem. 1996, 66,
2105–2112.
(19) Corbeil, C. R.; Englebienne, P.; Moitessier, N. Docking Ligands
into Flexible and Solvated Macromolecules. 1. Development and
Validation of FITTED 1.0. J. Chem. Inf. Model. 2007, 47, 435–449.
(20) Corbeil, C. R.; Moitessier, N. Docking Ligands into Flexible and
Solvated Macromolecules. 3. Impact of Input Ligand Conforma-
tion, Protein Flexibility and Water Molecules on the Accuracy of
Docking Programs. J. Chem. Inf. Model. 2009, 49, 997–1009.
(21) Corbeil, C. R.; Englebienne, P.; Moitessier, N.; Therrien, E.
FITTED Docking Program; www.fitted.ca 2009.
(22) Oballa, R. M.; Truchon, J. F.; Bayly, C. I.; Chauret, N.; Day, S.;
Crane, S.; Berthelette, C. A generally applicable method for asses-
sing the electrophilicity and reactivity of diverse nitrile-containing
compounds. Bioorg. Med. Chem. Lett. 2007, 17, 998–1002.
(23) Hanessian, S.; Moitessier, N. Sulfonamide-based acyclic and con-
formationally constrained MMP inhibitors: from computer-as-
sisted design to nanomolar compounds. Curr. Top. Med. Chem.
2004, 4, 1269–1287.
(24) Hanessian, S.; McNaughton-Smith, G.; Lombart, H.-G.; Lubell,
W. D. Design and synthesis of conformationally constrained
amino acids as versatile scaffolds and peptide mimetics. Tetrahe-
dron 1997, 53, 12789–12854.
(25) Baldwin, J. E.; Freeman, R. T.; Lowe, C.; Schofield, C. J.; Lee, E. A
[gamma]-lactam analogue of the penems possessing antibacterial
activity. Tetrahedron 1989, 45, 4537–4550.
(42) Jones, K.; Woo, K.-C. A total synthesis of (-)-ruspolinone.
Tetrahedron 1991, 47, 7179–7184.
(43) Tang, F. Y.; Qu, L. Q.; Xu, Y.; Ma, R. J.; Chen, S.-H.; Li, G.
Practical Synthesis of Structurally Important Spirodiamine Tem-
plates. Synth. Commun. 2007, 37, 3793–3799.
(44) Caputo, C. A.; Carneiro, F. d. S.; Jennings, M. C.; Jones, N. D.
Modular syntheses of oxazolinylamine ligands and characteriza-
tion of group 10 metal complexes. Can. J. Chem. 2007, 85, 85–
95.
(45) Heller, B.; Sundermann, B.; Buschmann, H.; Drexler, H. J.; You,
J.; Holzgrabe, U.; Heller, E.; Oehme, G. Photocatalyzed [2 þ 2 þ
2]-Cycloaddition of Nitriles with Acetylene: An Effective Method
for the Synthesis of 2-Pyridines under Mild Conditions. J. Org.
Chem. 2002, 67, 4414–4422.
(46) Johnson, R. L.; Rajakumar, G.; Yu, K. L.; Mishra, R. K. Synthesis
of Pro-Leu-Gly-NH2 analogs modified at the prolyl residue and
evaluation of their effects on the receptor binding activity of the
central dopamine receptor agonist, ADTN. J. Med. Chem. 1986,
29, 2104–2107.
(26) Subasinghe, N. L.; Bontems, R. J.; McIntee, E.; Mishra, R. K.;
Johnson, R. L. Bicyclic thiazolidine lactam peptidomimetics of the
dopamine receptor modulating peptide Pro-Leu-Gly-NH2. J. Med.
Chem. 1993, 36, 2356–2361.
(27) Nagai, U.; Sato, K. Synthesis of a bicyclic dipeptide with the shape
of [beta]-turn central part. Tetrahedron Lett. 1985, 26, 647–650.