Bicyclic carbamates as inhibitors of papain-like cathepsin proteases

Article abstract of DOI:10.1016/j.bmcl.2006.12.014

A 6-oxa-1-aza-bicyclo[3.2.1]octan-7-one system inhibits the proteolytic activity of several cysteine proteases belonging to the papain family. In vitro mechanistic studies and in silico calculations suggest that the minimal π-overlap between the bridgehea

Full text of DOI:10.1016/j.bmcl.2006.12.014

Bioorganic & Medicinal Chemistry Letters 17 (2007) 1254–1259
Bicyclic carbamates as inhibitors of papain-like cathepsin proteases
Robert Epple,* Hugo D. Urbina, Ross Russo, Hong Liu, Daniel Mason,
Badry Bursulaya, Christine Tumanut, Jun Li and Jennifer L. Harris
Department of Drug Discovery, The Genomics Institute of the Novartis Research Foundation, 10675 John Jay Hopkins Drive,
San Diego, CA 92121, USA
Received 3 November 2006; revised 4 December 2006; accepted 4 December 2006
Available online 9 December 2006
Abstract—A 6-oxa-1-aza-bicyclo[3.2.1]octan-7-one system inhibits the proteolytic activity of several cysteine proteases belonging to
the papain family. In vitro mechanistic studies and in silico calculations suggest that the minimal p-overlap between the bridgehead
nitrogen and the carbonyl leads to a considerable weakening of the urethane system, making it susceptible to nucleophilic attack
from the active site thiol group. The resulting covalent adduct is slowly hydrolyzed, releasing the hydroxypiperidine product of
the inhibitor. Synthesis and testing of a set of analogs led to variable protease subtype selectivities ranging from micromolar to
nanomolar potencies.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Most cathepsins are endopeptidases belonging to the
papain-like family of cysteine proteases.1,2 Cathepsins
are generally concentrated in lysosomal and endosomal
compartments and play a role in a broad array of phys-
iological processes. Among them are the rather non-spe-
cific ‘house-keeping’ functions such as degradation of
both internalized and cellular proteins.3 However, more
specialized functions in the processing of enzymes and
hormones have been recently attributed to cathepsins
and they have been implicated in multiple disease pro-
cesses.4 Indeed, inhibition of specific cathepsins is be-
lieved to have therapeutic implications on pathological
conditions associated with cellular homeostasis, apopto-
sis, tumor invasion and metastasis, bone resorption, and
antigen presentation.5–11
have been routinely employed in the inhibition of prote-
ases by covalent interaction of enzyme with inhibitor.
Aldehydes, semicarbazones, nitriles, and ketones usually
lead to reversible inhibition, whereas halomethylke-
tones, epoxides, and Michael acceptors are among the
functionalities used for irreversible inhibition.
To date relatively few protease inhibitors have success-
fully progressed through clinical trials. Major issues
are the often non-drug-like properties of the peptidic
portion of the inhibitor and/or the lack of selectivity,
often attributed to non-specific interactions of the elec-
trophilic functionality with endogenous nucleophiles
in vivo.16 In order to circumvent these drawbacks com-
monly seen with the classical approach of substrate-
based drug design, novel molecular concepts need to
be explored. We report the investigation of a strained
bicyclic carbamate system discovered by high through-
put screening (HTS).
Small molecule inhibitors of proteases have been the
subject of several excellent and extensive reviews.12–18
Many protease inhibitors, including those for cathep-
sins, consist of a peptide-substrate mimetic that medi-
ates binding to specific pockets in the enzyme and an
electrophilic ‘warhead’ that makes crucial contacts to
the catalytic residues. The nature of the latter determines
the classification of these small molecules into reversible
and irreversible inhibitors. Several electrophilic groups
We performed a HTS of an unbiased library of roughly
1 million compounds based on monitoring the cleavage
rate of
a coumarin-labeled synthetic tetrapeptide
highly reactive to cathepsin S.19 While many of the con-
firmed hits displayed an electrophilic moiety and/or a
highly peptidic structure commonly seen in protease
inhibitors, compound 1 stood out by its unusually low
molecular weight and its non-peptidic nature. The
structure of 1 was tentatively assigned as in Figure 1,
by tracing the hit back to its origin. It is a cyclized
Keywords: Cathepsins; Protease inhibitors; Carbamates.
*
Corresponding author. Tel.: +858 812 1720; fax: +858 812 1848;
e-mail: repple@gnf.org
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.12.014

R. Epple et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1254–1259
1255
cyclization of 6 required a two step procedure: first the
piperidine-nitrogen was selectively carbonylated to the
corresponding chloroformate using triphosgene and tri-
ethylamine, then the 3-hydroxy group was deprotonated
using sodium hydride, which led to spontaneous intra-
molecular cyclization to give racemic trans-4-(2,3-di-
methyl-phenoxy)-6-oxa-1-aza-bicyclo[3.2.1]octan-7-one
8. Attempts to cyclize the chloroformate intermediate of
7 to the bicycle 9 were unsuccessful, probably due to the
unfavorable boat conformation of the piperidine ring re-
quired for intramolecular ring-closure.
O
N
O
1
O
OMe
OMe
Figure 1. Compound 1, the result of intramolecular cyclization of
4-aryloxy-3-hydroxypiperidines with phosgene.
derivative of 4-aryloxy-3-hydroxypiperidines such as
ifoxetine, a selective serotonin reuptake inhibitor that
has been studied more than 25 years ago at Ciba–Geigy
as an antidepressant.20–22
In order to invert the stereocenter at position three of
compound 4, we performed a Mitsunobu reaction using
p-nitrobenzoic acid as the nucleophile. Alternatively,
the inversion could be accomplished by an oxidation/re-
duction sequence using PyrÆSO3 and Red-Al, but with
inferior yields. Saponification of intermediate 10 in
methanolic base followed by deprotection gave piperidine
11 (cf. ifoxetine). Intramolecular cyclization analogous to
the above described procedure gave cis-4-(2,3-dimethyl-
phenoxy)-6-oxa-1-aza-bicyclo[3.2.1]octan-7-one 12.
First we resynthesized compound 1 to confirm its struc-
tural identity including its relative stereochemistry
(Scheme 1).23 According to synthetic procedures de-
scribed by Paioni,20 the tetrahydropyridine 2 was con-
verted to the racemic Cbz-protected epoxide 3 in two
steps. Opening of the epoxide 3 (ROH = 2,3-dimethyl-
phenol) under basic aqueous conditions gave a mixture
of regioisomers. As expected for a favored trans-diaxial
ring-opening, the ratio of 4-aryloxy piperidines 4 versus
3-aryloxypiperidines 5 was about 5:1. The regioisomers
were chromatographically separated and hydrogenolyti-
cally deprotected to give the ‘open’ forms 6 and 7. The
Analogs 13–16 (Table 1) were synthesized in a similar
fashion to the synthesis of 12. In the case of analog 13,
thiophosgene was used instead of phosgene in the cycliza-
tion step, for analogs 15 and 16 the corresponding alcohol
(benzylalcohol for 15, isopropanol for 16) was used as
solvent together with NaH in the opening of epoxide 3.
Cbz
N
Cbz
Cbz
N
O
Cbz
H
Br
N
N
c)
a), j)
N
a), b)
20
OR
OH
OH
O
OH
OR
OH
2
3
l), m), n)
5
4
17
k)
g), h), i)
d)
d)
O
N
O
O
H
N
H
N
H
N
H·HBr
12
N
21
OR
OH
OH
OH
OH
OR
OR
Br
18
11
6
7
e), f)
e), f)
e), f)
O
N
O
O
N
O
N
O
O
O
O
N
OR
OR
OR
Br
19
12
8
9
Scheme 1. Reagents and conditions: (a) CbzCl, 0.5 M Na2CO3/dioxane 5:2, rt, 5 h; (b) m-CPBA, DCM, 0 ꢁC–rt, 7 h; (c) ROH, 2 M NaOH/MeCN
1:4, reflux, 48 h; (d) 1 atm H2, Pd/C (cat), EtOH, rt, 3 h; (e) triphosgene, NEt3, DCM, 0 ꢁC, 1 h; (f) NaH, rt, 18 h; (g) PPh3, DEAD, p-nitrobenzoic
acid, THF, 50 ꢁC, 48 h; (h) NaOH, MeOH, rt, 1 h; (i) 1 atm H2, Pd/C (cat), EtOH, rt, 3 h; (j) OsO4, NMO, citric acid, t-BuOH/H2O 1:1, 50 ꢁC, 12 h;
(k) HBr, HOAc, 0 ꢁC–rt, 3 h; (l) 2,3-dimethylphenol, K2CO3, MeCN, 60 ꢁC, 12 h; (m) MeNH2, MeOH, 50 ꢁC, 2 h; (n) CDI, DMAP, benzene, reflux,
4 h. Compounds 3–21 are racemic mixtures.

1256
R. Epple et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1254–1259
Table 1. Determined IC50 values in micromolar of compounds 1–21 for inhibition of various cathepsins24
X
O
N
OR
Compound
X
O
Strel
R
CatB
0.046
CatS
0.473
CatC
1.727
CatL
1.270
CatK
2.438
CatH
>30
CatX
0.553
CatF
5.908
CatV
3.317
8
trans
12
13
14
O
S
cis
cis
cis
0.001
0.519
0.005
0.016
0.696
0.017
0.018
0.606
0.029
0.022
0.994
0.026
0.038
1.833
0.104
2.653
>100
6.591
0.197
7.149
0.326
0.331
>10
0.131
5.166
0.158
O
0.487
15
16
O
O
O
cis
0.002
0.008
0.174
0.298
0.403
0.191
0.263
0.230
0.511
1.267
>10
>10
0.547
1.102
3.760
1.826
1.289
0.894
cis
19
6
trans
OR = Br
0.158
>100
4.553
>100
>10
5.816
>100
>30
>100
>100
>10
>10
>10
>100
>100
>100
>100
>100
11
21
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
Strel, relative stereoconformation.
An alternative synthetic route via dihydroxylation led to
the protected tetrahydropyridine 17. Selective cyclization
and alkylation of the deprotected diol failed to give the
desired bicyclic carbamate analogs. Treatment of the
epoxide 3 with HBr selectively gave bromide 18 which
could be cyclized to the carbamate 19, but any attempts
to displace the bromide with phenols using inorganic
bases in DMF (DMSO) were unsuccessful. The control
compound 21 was synthesized in three steps starting from
epibromohydrin 20.
values than 12 or 14. An exception is the low single-digit
nanomolar activity of the benzyl-substituted analog 15
against cathepsin B, representing the most selective
inhibitor for cathepsin B in the series (100-fold).
Next we investigated the stability of these low molecular
weight protease inhibitors. Both stereoisomers 8 and 12
were isolated as white, stable solids. The compounds
were also stable in aqueous basic or neutral solutions,
while spontaneous ring-opening was observed under
fairly acidic conditions (pH < 2). To the best of our
knowledge there are only two reports in the literature
on similar ‘anti-Bredt’ bicyclic systems.25,26 The bicyclic
carbamate is predicted to be substantially less stable
than usual carbamates due to substantial ring strain
and minimal p-overlap between the bridgehead nitrogen
and the carbonyl group. The relative thermodynamic
stability of the bicyclic urethane 12 was calculated to
be >20 kcal lower in comparison to the monocyclic ana-
log 11,27 rendering it susceptible for nucleophilic attack.
Besides other factors affecting inhibitory potency such as
decreasing the degrees of rotational freedom in the bicy-
cle 12, this greatly enhanced susceptibility for nucleo-
philic attack may play a crucial role especially in the
case of covalent inhibition.
An in depth profiling24 of analogs 8 and 12–16 revealed
that these cyclic carbamates showed inhibitory activity
across a panel of human cysteinyl cathepsin proteases
(Table 1). Within this protease family, the determined
IC50 values were lowest for cathepsin B, followed by
cathepsins S, C, L, and K. Moderate activities were seen
against cathepsins V, X, F, and H. The compounds were
shown to be inactive against other families of cysteine
proteases such as caspases and against members of the
serine protease family such as thrombin and trypsin
(data not shown).
Generally, there is a remarkable drop in IC50 values
when comparing the trans to the cis configured stereo-
isomers (cf. 8 vs 12). Also, analog 12, which contains
the more hydrophobic dimethylphenyl group, seemed
to be consistently more active in the cathepsin panel
when compared to the phenyl-substituted analog 14.
The thiocarbamate 13 had substantially less inhibitory
effects when compared to the ‘oxo’-carbamates. The al-
kyl-substituted analogs 15 (R = benzyl) and 16 (R = iso-
propyl) generally displayed about 10-fold higher IC50
To understand the mechanism of inhibition by these
bicyclic carbamates, compound 12 was examined in a
dialysis study (Fig. 2).28 Briefly, cathepsin B was first
incubated at 37 ꢁC with 0.2 lM 12 (200 · IC50) for com-
plete suppression of catalytic activity (2.9 rfu/sec for
inhibitor treated vs 121 rfu/sec for vehicle control). Then
the mixtures were dialyzed extensively at 4 ꢁC to remove

R. Epple et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1254–1259
1257
50000
25000
0
121
71
51
2.9
0
100
200
300
400
0
100
200
300
400
Sec
Sec
Figure 2. Reversibility of binding of compound 12 to cathepsin B. The rate of conversion of a fluorogenic substrate (Ac-His-Pro-Val-Lys-ACMC) is
monitored using DMSO vehicle (solid diamond) and compound 12 treated (open diamond) cathepsin B before dialysis (left) and after dialysis (right).
The number above each progress curve reflects the individual reaction rate (relative fluorescence units (rfu) per second). The dotted arrows on the
right panel indicate the rate increase during the course of the protease activity assay for compound 12 treated cathepsin B.
unbound inhibitor before the cathepsin activities were
re-measured at 37 ꢁC (51 rfu/sec for inhibitor treated
vs 71 rfu/sec for vehicle control). The apparent recovery
of cathepsin B activity in compound 12 treated sample
after dialysis suggests that compound 12 is a partially
reversible inhibitor. However, a careful examination of
the progress curves in Figure 2 revealed that the sub-
strate conversion was accelerated over the course of
the assay (indicated by the dotted arrowed lines). We
speculated that this rate increase is likely caused by a
temperature-sensitive release of covalently bound com-
pound 12 from the active site, since the substrate conver-
sion of the control was not affected by the incubation
temperature.
Pro76, Ala173, Gly198, and Ala200. In addition the car-
bonyl-O of carbamate 12 forms hydrogen bonds with
the side chains of His199 and Gln23. The calculation
of the free energy of binding33 revealed that the (S,R)-
cis carbamate 12 is by ꢀ6 kcal/mol more stable than
the (S,S)-trans stereoisomer 8. This agrees well with
the observed lower IC50 value for the racemic cis
stereoisomer.
In summary, a series of bicyclic carbamates was identi-
fied as novel, non-peptidic protease inhibitor scaffold
starting from HTS. The scaffold is selective for the
cathepsin family, with subtype selectivities following
the order B > S > C,L,K > V,X,F,H. A preference for
To test this hypothesis, the potential adduct formation
between cathepsin B and inhibitor (compound 12) was
analyzed by mass spectrometry following tryptic/chy-
motryptic digestion.29 Two active site fragments of the
test enzyme (cathepsin B) were identified as 19EIRD
QGSCGSC*W30 and 22DQGSCGSC*W30 from the
untreated cathepsin B sample.30 For inhibitor treated
E
O
N
S
O
E-SH
OH
E-SH
OH
O
HN
N
OR
OR
OR
cathepsin B, both fragments gained
a mass of
247 amu, corresponding to a single covalent conjuga-
tion of 12 to the catalytic cysteine residues (C*). The
stability of this adduct was indeed sensitive to tempera-
ture increase.31 Compared to the control that remained
on ice, a 2-h incubation at 37 ꢁC promoted the dissoci-
ation of this enzyme-inhibitor adduct, resulting in a
4-fold increase in abundance of a metabolite. The multi-
ple reaction monitoring (MRM) characteristics of the
metabolite were identical to those of the suggested dis-
sociation product 11.
Gln23
Cys29
A potential binding mode of compound 12 with cathep-
sin B is shown in Figure 3. It was obtained using a
MCMM/LLMOD conformational search strategy.32
The hybridization of the carbonyl-carbon of 12 was
changed from sp2 to sp3 and placed in covalent contact
with the thiol group of the catalytic cysteine (Cys29). A
formal negative charge was assigned to the carbonyl-ox-
ygen. In this model the dimethylphenyl substituent occu-
pies the hydrophobic P2-pocket, which is defined by
aminoacid residues Gly27, Trp30, Gly73, Tyr75,
His199
Figure 3. Proposed binding mode of compound 12 to cathepsin B
(E-SH). The sulfur of Cys29 (yellow) is in covalent contact with the
carbonyl-carbon of 12. The dotted lines represent H-bonds between
Gln23, His199 and the carbonyl-oxygen of 12.

1258
R. Epple et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1254–1259
17. Vicik, R.; Busemann, M.; Baumann, K.; Schirmeister, T.
Curr. Top. Med. Chem. 2006, 6, 331.
hydrophobic R-groups in cis configuration relative to
the carbamate oxygen was observed, with determined
IC50 values for cathepsin B in the single-digit nanomolar
range. The carbamate functionality in the scaffold is
substantially destabilized due to the ‘anti-Bredt’ nature
of the bicycle, and offers a weak point for nucleophilic
attack by the active site cysteine thiol. The bicycle sub-
sequently undergoes ring-opening and covalently binds
to the catalytic cysteine, leading to inhibition of the
enzyme. This hypothesis is in accordance with the mass
spectrometric analysis of digested enzyme. After incuba-
tion with compound 12, the fragments which include the
catalytic cysteine are covalently linked to the inhibitor.
The non-linear rate recovery seen in dialysis studies
and characterization of metabolites via MRM suggest
that the inhibitor is slowly hydrolyzed off the enzyme
as its ring-opened synthetic precursor 11 at ambient
temperature. A molecular model of cathepsin B and
covalently bound inhibitor 12 further strengthens the
accordance of the proposed binding mode and the
experimental observations.
18. Leung-Toung, R.; Zhao, Y.; Li, W.; Tam, T. F.; Karimi-
an, K.; Spino, M. Curr. Med. Chem. 2006, 13, 547.
19. Harris, J. L.; Backes, B. J.; Leonetti, F.; Mahrus, S.;
Ellman, J. A.; Craik, C. S. Proc. Natl. Acad. Sci. U.S.A.
2000, 97, 7754.
20. Paioni, R. U.S. Patent 4,160,837, 1979.
ˆ
21. Waldmeier, P. C.; Maıtre, L.; Baumann, P. A.; Hauser,
K.; Bischoff, S.; Bittiger, H.; Paioni, R. Eur. J. Pharmacol.
1986, 130, 1.
22. Paioni, R.; Waldmeier, P. C.; Delini-Stula, A. Drugs
Future 1987, 12, 126.
23. Detailed synthetic procedures to compounds 3–21 can be
found in Supplementary material.
24. Enzyme inhibition assays were performed as previously
described (Tully D. C. et al. Bioorg. Med. Chem. Lett.
2006, 16, 1975). Briefly, recombinant human cathepsin
enzymes were used in all enzyme inhibition assays. The
standard assay format contained 50 lM fluorogenic pep-
tide substrate in 100 mM NaOAc, 1 mM EDTA, 0.01%
Brij-35, and 5 mM DTT, pH 5.5, at 37 ꢁC. The enzyme
was preincubated with inhibitor for 20 min before sub-
strate was added to initiate the reaction. The substrate
hydrolysis was monitored by the increase in fluorescence
at an excitation wavelength of 380 nm and an emission
wavelength of 450 nm on a Gemini EM fluorometer. The
reaction progress curve was fitted to the Morrison
equation using PlateKi (BioKin) and the apparent inhibi-
tion constants were reported as IC50 values.
Acknowledgment
We gratefully acknowledge Drs. Phil Alper and Eric
Peters for their helpful discussions and support.
25. Hall, H. K., Jr.; El-Sheikeil, A. J. Org. Chem. 1980, 45,
5325.
Supplementary data
26. Herdeis, C.; Schiffer, T. Tetrahedron 1999, 55, 1043.
27. The strain energy was estimated from energy difference
between open and closed forms of the bicyclic urethane
(11 and 12) and its monocyclic analog 21. The conforma-
tions were obtained via LMOD conformational search
method (Macromodel, Schrodinger, Inc.), employing a
MMFF94 force field and treating solvation implicitly via
the Generalized Born method. The energy calculations
were done using ab initio density functional method
(Jaguar, Schrodinger, Inc.) in which the 6-31G** basis set
and hybrid DFT functional B3LYP were employed. The
aqueous solvation was treated with the Poisson–Boltz-
mann continuum-dielectric method.
28. Dialysis-based reversibility study: 2 nM recombinant
cathepsin B was incubated with 0.2 lM compound 12
(IC50 = 1 nM) or DMSO (control) for 60 min at 37 ꢁC
with slow agitation on a shaker. Inhibition of the
enzymatic activity was monitored according to the
enzyme inhibition assay described in reference 24, by
adding 50 lL of the preincubated enzyme-inhibitor mix-
ture into 50 lL of 200 lM fluorogenic peptide substrate
Ac-His-Pro-Val-Lys-ACMA. Alternatively, the enzyme-
inhibitor mixture was injected into a 3-mL SlideDialyzer
(Pierce, 7500-kDa cutoff) and dialyzed at 4 ꢁC against 1 L
of dialysis buffer (100 mM NaOAc, pH 5.5, 100 mM
NaCl, 1 mM EDTA, 0.001% Brij-35, and 2 mM DTT).
The dialysis buffer was changed three times in 8 h
intervals. After dialysis, the cathepsin B activity was
assayed as described above.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmcl.2006.
12.014.
References and notes
1. McGrath, M. E. Annu. Rev. Biophys. Biomol. Struct. 1999,
28, 181.
2. Turk, V.; Turk, B.; Turk, D. EMBO J. 2001, 20, 4629.
3. Bond, J. S.; Butler, P. E. Ann. Rev. Biochem. 1987, 56,
333.
4. Turk, V.; Turk, B.; Turk, D. Biochim. Biophys. Acta 2000,
1477, 98.
5. Bromme, D.; Kaleta, J. Curr. Pharm. Des. 2002, 8, 1639.
6. Turk, D.; Guncar, G. Acta Crystallogr. 2003, 59, 203.
7. Turk, V.; Turk, B.; Guncar, G.; Turk, D.; Kos, J. Adv.
Enzyme Regul. 2002, 42, 285.
8. Veber, D. F.; Thompson, S. K. Curr. Opin. Drug Disc.
Dev. 2000, 3, 362.
9. Motyckova, G.; Fisher, D. E. Curr. Mol. Med. 2002, 2,
407.
10. Troen, B. R. Drug News Perspect. 2004, 17, 19.
11. Honey, K.; Rudensky, A. Y. Nat. Rev. Immun. 2003, 3,
472.
12. Otto, H.-H.; Schirmeister, T. Chem. Rev. 1997, 97, 133.
13. Leung, D.; Abbenante, G.; Fairlie, D. P. J. Med. Chem.
2000, 43, 305.
29. Enzymatic digestion of cathepsin B for mass spectrometric
analysis: 1 nmol of cathepsin B was denatured in 3 M
guanidine HCl. The cysteine residues were reduced with
TCEP followed by treatment with iodoacetamide. Excess
iodoacetamide was quenched with DTT. The solution was
then diluted to 0.5 M guanidine with 50 mM Tris–HCl,
pH 8, and 10 mM CaCl2. Trypsin was added to the sample
and proteolysis was carried out at 37 ꢁC for 4 h. Then
14. Hernandez, A. A.; Roush, W. R. Curr. Opin. Chem. Biol.
2002, 6, 459.
15. Powers, J. C.; Asgian, J. L.; Ekici, O. D.; James, K. E.
Chem. Rev. 2002, 102, 4639.
16. Abbenante, G.; Fairlie, D. P. Bioorg. Med. Chem. 2005, 1,
71.

R. Epple et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1254–1259
1259
chymotrypsin was added and the reaction was continued
at rt for another 18 h.
30. Procedures and results for the mass spectrometric analysis
of the cathepsin B digests can be found in Supplementary
material.
31. Procedures and results for the temperature-dependent
metabolite monitoring can be found in Supplementary
material.
random changes in torsion angles from the MCMM
method. During the LLMOD structural perturbation,
˚
protein residues within 4 A from the bound ligand were
allowed to move freely. The inhibitor was subjected to
perturbations via TORS command available in Macro-
Model. The employed force field was OPLS-AA 2005 with
distance-dependent dielectric constant 2r and no explicit
solvation.
32. To model the interaction of the cyclic carbamates with
cathepsin B, we used a mixed Monte Carlo Multiple
Minimum (MCMM)/LLMOD (Large Scale LowMode)
conformational search strategy available in MacroModel
(version 8.0; Schrodinger, Inc.). The structural perturba-
tion via the LLMOD method is alternated with the
33. The free energy of binding of compounds was obtained by
subtracting the ligand and receptor energies from that of
the complex. For the free energy calculations an OPLS-
AA force field (Macromodel, Schrodinger Inc.) was
employed and the solvation was accounted for via
Generalized Born method.