Novel non-peptidic vinylsulfones targeting the S2 and S3 subsites of parasite cysteine proteases

Article abstract of DOI:10.1016/j.bmcl.2009.08.098

We describe here the identification of non-peptidic vinylsulfones that inhibit parasite cysteine proteases in vitro and inhibit the growth of Trypanosoma brucei brucei parasites in culture. A high resolution (1.75 A) co-crystal structure of 8a bound to cruzain reveals how the non-peptidic P2/P3 moiety in such analogs bind the S2 and S3 subsites of the protease, effectively recapitulating important binding interactions present in more traditional peptide-based protease inhibitors and natural substrates.

Full text of DOI:10.1016/j.bmcl.2009.08.098

Bioorganic & Medicinal Chemistry Letters 19 (2009) 6218–6221
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel non-peptidic vinylsulfones targeting the S2 and S3 subsites
of parasite cysteine proteases
Clifford Bryant a,b, Iain D. Kerr d, Moumita Debnath d, Kenny K. H. Ang a,b, Joseline Ratnam a,b
,
Rafaela S. Ferreira c, Priyadarshini Jaishankar a,b, DongMei Zhao a,b, Michelle R. Arkin a,b,c
,
James H. McKerrow b,e, Linda S. Brinen b,d, Adam R. Renslo a,b,c,
*
a Small Molecule Discovery Center, University of California, San Francisco, CA 94158, United States
b Sandler Center for Basic Research on Parasitic Disease, University of California, San Francisco, CA 94158, United States
c Department of Pharmaceutical Chemistry, University of California, San Francisco, CA 94158, United States
d Department of Cellular and Molecular Pharmacology, University of California, San Francisco, CA 94158, United States
e Department of Pathology, University of California, San Francisco, CA 94158, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
We describe here the identification of non-peptidic vinylsulfones that inhibit parasite cysteine proteases
in vitro and inhibit the growth of Trypanosoma brucei brucei parasites in culture. A high resolution
(1.75 Å) co-crystal structure of 8a bound to cruzain reveals how the non-peptidic P2/P3 moiety in such
analogs bind the S2 and S3 subsites of the protease, effectively recapitulating important binding interac-
tions present in more traditional peptide-based protease inhibitors and natural substrates.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 12 July 2009
Accepted 31 August 2009
Available online 3 September 2009
Keywords:
Protease inhibitors
Cruzain
Rhodesain
Vinylsulfone
There is an urgent need for new safe and effective drugs to treat
the trypanosomal diseases such as Human African Trypanosomia-
sis (HAT) and Chagas’ disease. Existing drugs for these conditions
are sub-optimal with regard to efficacy and/or safety, while the
lone truly modern drug available for HAT, eflornithine, is difficult
to administer in poor rural settings, and is effective only against
Trypanosoma brucei gambiense. Current efforts to address this un-
met medical need range from the repurposing of azole antifungals1
in Chagas’ disease to the investigation of a variety of recently iden-
tified biochemical targets.2,3 Prominent among the latter are pa-
pain-like cysteine proteases such as cruzain in Trypanosoma cruzi
and rhodesain and TbCatB in Trypanosoma brucei.4 There is now
good precedent for targeting cysteine proteases in human disease
seeing as inhibitors of related human proteases have progressed
into late stage clinical trials (e.g., cathepsin K inhibitor odanacati-
nib5 for osteoporosis). Among parasite protease inhibitors being
studied is the cruzain inhibitor K777 (1), currently in preclinical
development as a potential treatment for Chagas’ disease.6
marily hydrophobic and hydrogen bonding interactions), but as pep-
tides also possess well recognized liabilities as potential drugs. These
include susceptibility to enzymatic or chemical hydrolysis of peptide
bonds and the metabolism of amino acid side chains, issues that can
potentially be addressed through chemical modification of the amino
acidbackboneand/orsidechain. Wedescribehereourinitialeffortsto
identify non-peptidic parasite cysteine protease inhibitors that reca-
pitulate the binding interactions of substrate-like peptidic inhibitors
but have a greater likelihood of possessing drug-like properties.
P2
P1'
O
H
N
SO2Ph
Ph
N
N
H
N
O
P1
P3
Classical cysteine protease inhibitors comprise a di- or tripeptide
linked to an electrophilic warhead that engages the active-site cys-
teine thiol function reversibly or irreversibly. Such inhibitors readily
recapitulate the binding interactions of endogenous substrate (pri-
1
K-777 ( )
Recently, Ellman and co-workers described a non-peptidic class
of cruzain inhibitors discovered via the ‘substrate activity screen-
ing’ approach advanced by that lab.7 In the work described here,
we similarly aimed to survey non-peptidic fragments as potential
P2/P3 moieties, but proceeded by screening the inhibitors directly.
* Corresponding author.
E-mail address: adam.renslo@ucsf.edu (A.R. Renslo).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.08.098

C. Bryant et al. / Bioorg. Med. Chem. Lett. 19 (2009) 6218–6221
6219
While empirical in nature, this effort was guided by the availability
of co-crystal structures of substrate-like inhibitors bound to
cruzain.8,9 Inspection of these structures suggested to us that an
aliphatic or heterocyclic ring might serve as a non-peptidic linkage
to a second pendant hydrophobe that would bind in the enzyme’s
S2 pocket, playing a role analogous to the phenylalanine side chain
in inhibitors like 1. To explore this possibility, carboxylic acids
meeting the above design criteria were acquired from commercial
sources and coupled to the known vinylsulfone amine 2,10 as
illustrated below for a representative series of (hetero)aryl thia-
zoles (3a–c, Scheme 1).
Around 30 non-peptidic vinylsulfones were prepared and IC50
values determined for cruzain, rhodesain, and TbCatB (5 min incu-
bation with inhibitor). True rate constants of rhodesain inactiva-
tion were subsequently determined for the most potent analogs
(Table 1).11,12 A broad range of potencies were observed for these
analogs and some general SAR trends could be discerned. Analogs
bearing a single heteroaromatic ring system were generally devoid
shown). Similarly ineffective were analogs with a heterocyclic ring
linked to a hydrophobe at the ortho position (5 and 6, Fig. 1). In
contrast, analogs possessing more prolate geometries (e.g., 3a–c)
were found to be effective inhibitors of cruzain and quite potent
inhibitors of rhodesain. For example, analog 3a inhibited rhodesain
with IC50 and Ki values superior to those of 1 (Table 1).
We next synthesized a series of similar analogs (4a–d) in which
an isoxazole ring replaces the thiazole ring of 3a–d (Scheme 1). For
these analogs, the requisite isoxazole carboxylic acid starting
materials were prepared in three steps, employing the [3+2] dipo-
lar cycloaddition of nitrile oxides (generated in situ from the
hydroximinoyl chlorides13) with methyl propiolate as a key step
(Scheme 1). The isoxazole analogs 4a–d exhibited a similar SAR
profile as seen for 3a–c, inhibiting rhodesain more potently than
cruzain and conferring no significant inhibition of TbCatB.
The analogs 7a/b and 8a/b represent a second and structurally
distinct series of protease inhibitors that emerged from our survey
of non-peptidic P2/P3 fragments (Fig. 2). These analogs were pre-
pared as diastereomeric mixtures from the coupling of ( ) trans-
2-(4-chlorobenzoyl)cyclohexane-1-carboxylic acid to vinylsulfone
amines. We successfully separated the two diastereomers of 7
and found that one isomer was at least 100-fold more potent than
the other, an encouraging result that implied specific binding inter-
actions, presumably involving the S2 and/or S3 subsites of the pro-
tease active-site.
To better understand the binding mode of analogs such as 7 and
8, we crystallized cruzain inhibited by 7 or 8 (as diastereomeric
mixtures). Cruzain inhibited by 8 provided superior crystals and
so we collected diffraction data for these at the Stanford Synchro-
tron Radiation Lightsource (SSRL), BL7-1 (Table 2). Traditionally,
recombinant cruzain has been expressed in Escherichia coli and
purified from refolded inclusion bodies. However, in our experi-
ence the final yield of protein was typically in the low milligram
of enzyme activity (<5% inhibition of cruzain at 1 lM, data not
N
H2N
SO2Ph
Ph
R
OH
S
2
O
a
R=
SO2Ph
Ph
N
N
3b
H
R
N
S
3a
N
O
O
N
N
3c
O
H
N
R
SO2Ph
Ph
R =
F
N
N
4c
N
4d
N
N
N
H
N
H
N
O
4a
4b
SO2Ph
Ph
cruzain inhibition
12.5% at 1 uM
SO2Ph
Ph
O
O
Scheme 1. Synthesis of non-peptidic vinylsulfones 3a–c from commercially
available acids. Reagents and conditions: (a) 2-HCl, HATU, DIEA, DMF, rt. The
preparation of isoxazoles 4a–d was accomplished from commercially available
isoxazole acids (for 4a) or from pyridyl aldehydes as follows for the case of 4d.
Reagents and conditions: (a) 2-pyridinecarboxaldehyde, NH2OH, NaOAc, 95% EtOH,
reflux, 71%; (b) NCS, DMF, 50 °C, then methyl propiolate, Et3N, CH2Cl2, rt, 69%; (c)
1.0 M LiOH, MeOH–THF; (d) 2-HCl, HATU, DIEA, DMF, rt; 81% over two steps.
5
6
2% at 1 uM
Figure 1. Representative examples of ineffective vinylsulfone analogs from the
initial survey of non-peptidic P2/P3 moieties.
Table 1
Enzymatic and biological activities of vinylsulfone analogs
Compd
Protease inhibition IC50
(
l
M)
Rhodesain inhibition kinetics12
Parasite growth inhibition GI50
(l
M)
Cytotoxicity IC50
Jurkat cells
(lM)
Cruzain
Rhodesain
TbCatB
kinact/Ki (Mꢀ1 sꢀ1
)
Kiapp
(l
M)
kinact (sꢀ1
)
T. b. b parasites
K-777 (1)
3a
3b
3c
4a
4b
4c
4d
7a
0.004
10
2
1
2
5
1
6
0.007
<0.006
0.03
0.01
0.01
0.05
0.01
0.05
0.1
4
555,000
84,000
20,100a
25,200
3610
0.78
0.32
—
4.77
13.0
16.2
—
7.4
0.80
4.9
0.029
0.0018
—
0.0079
0.0031
0.0091
—
0.0038
0.00051
0.0007
0.014
7
>100
20
7
>100
10
3% at 10
>100
98
>100
>100
>100
>100
>100
>100b
nd
lM
>100
>100
>100
40
>100
>100
>100
16
8590
348,000a
7720
>100
10
0.05
>100
0.1
9750
2160
5740
12b
nd
7b
8a/b
56
0.1
>100
45
38
8
>100
TbCatB: cathepsin B-like protease in T. brucei.
nd = not determined.
a
Value shown is kass
Tested as the mixture 7a/b.
.
b

6220
C. Bryant et al. / Bioorg. Med. Chem. Lett. 19 (2009) 6218–6221
H
N
H
N
H
N
H
N
1
1
SO2Ph
Ph
SO2Ph
Ph
SO2Ph
SO2Ph
Ph
2
2
O
O
O
O
O
O
O
O
9
Cl
Cl
Cl
Cl
8a/b
7a (1R,2R)
cruzain IC50 ~50 nM
7b (1S,2S)
cruzain IC50 > 10,000 nM
cruzain inhibition
3% at 1 uM
(1R,2R/1S,2S)
Figure 2. Structures of selected non-peptidic vinylsulfones.
Table 2
X-ray diffraction data and structure refinement statistics
Data collection
Space group
Cell dimensions
a (Å)
P21
44.69
b (Å)
72.49
c (Å)
60.93
a
(°)
b (°)
(°)
90.0
89.8
90.0
c
Resolution (Å)
Rmerge (%)
I/rI
1.75 (1.84–1.75)
5.7 (13.7)
25 (12)
97.3 (90.0)
7.4 (7.2)
11.5
a
Completeness (%)
Redundancy
Wilson B-factor (Å2)
Figure 3. The crystal structure of the cruzainꢁ8a complex, solved to a resolution of
1.75 Å. The inhibitor is colored gray and the unbiased mFo-DFc electron density is
shown in blue.
Refinement
Rfree/Rfactor (%)
17.6/14.3
4.46
Average B-factor (Å2)
Rms deviations
Bond lengths (Å)
Bond angles (°)
Ramachandran plotb
Favored (%)
Allowed (%)
Outliers (%)
PDB ID
conserved interactions in the S1, S10, and S2 subsites. These include
the formation of two hydrogen bonds to the inhibitor backbone
and another two with the sulfone moiety in the S10 subsite of the
enzyme. Conversely, the presence of a non-peptidic group at P2
in 8a results in the loss of a hydrogen bonding interaction to the
inhibitor backbone that is present in the cruzain complex with 1.
With respect to S3, the non-peptidic P3 moieties of 1 (N-methyl
piperazine) and 8a (4-chlorophenyl) do not make any direct or
water-mediated polar contacts with the enzyme (within a 3.5 Å
cut-off). However, a number of non-polar residues contribute to
the binding of 1 and 8a at the S3–S10 subsites, including Gly65
and Gly 66 at S3, Leu67, Ala138, and Leu160 at S2, Gly23 at S1,
and Met145 at S10. As observed in other structures of cruzain
bound by inhibitors with non-polar P2 substituents, the ‘specific-
ity’ residue at the bottom of the S2 subsite (Glu208) points away
from the S2 pocket. Only in the presence of basic and polar P2
groups (e.g., Arg) is Glu208 seen to point into the S2 subsite to
interact directly with the bound small molecule.8
0.019
1.59
97.2
100
0
3HD3
P P
P P
a
Rmerge =
I (h)j ꢀ (I (h))/
I (h)j, where I (h) is the measured diffraction
intensity and the summation includes all observations.
As defined by Molprobity.16
b
range from a multi-liter culture. More recently, we transitioned to
an expression system in the yeast Pichia pastoris that produces 10–
20 mg of soluble cruzain from four liters of culture. The gene
encoding cruzain was engineered to accommodate mutations at
two predicted glycosylation sites, Ser49Ala and Ser172Gly (mature
domain numbering), preventing the need for deglycosylation of the
yeast-expressed protein.
Non-peptidic inhibitors possessing reasonable enzymatic activ-
ity were subsequently tested for their ability to inhibit the growth
of cultured T. brucei brucei parasites and for general cytotoxicity to
mammalian cells (Jurkat). Gratifyingly, many of the non-peptidic
vinylsulfones were nearly as effective as 1 against cultured para-
sites, while conferring no significant toxicity to Jurkat cells. With
respect to the anti-parasite effects of 3 and 4, the presence of a
nitrogen atom in the pendant aryl ring seems to be important.
Hence phenyl substituted analogs 3a and 4a were ineffective
against cultured parasites while pyridyl (3b, 4b, 4d) and pyridazyl
(3c) analogs exhibited antiparasitic effects at low micromolar con-
centration. Interestingly, in our earlier study of non-basic analogs
of 1, a P3 2-pyridyl analog was found to be more effective against
culture T. cruzi parasites than non-pyridyl analogs with superior
enzyme activity.14 Since none of these analogs are predicted to
be significantly protonated at cytosolic or lysosomal pH, proton-
ation state cannot explain the superior parasite activities of the
heteroatom substituted analogs. The effect instead might reflect
intrinsic membrane permeability and/or active transport into par-
The covalently inhibited complex crystallized with two mole-
cules of the mature domain in the asymmetric unit that superim-
pose 214 matching
a-carbons with root mean square distances
(rmsd) of 0.34 Å. The high resolution of the diffraction data
(1.75 Å) allowed us to assign the absolute configuration of the ac-
tive diastereomer 8a (and by extension 7a) as 1R,2R at the P2
cyclohexane ring (both diastereomers possess the S configuration
at P1). Inspection of the cruzainꢁ8a structure reveals that the cyclo-
hexane ring in 8a is located in the S2 subsite of the cruzain active
site while the chlorophenyl ring extends over to form hydrophobic
contact with the S3 subsite (Fig. 3). The importance of cyclohexane
ring stereochemistry in 7/8 is further supported by the finding that
the benzophenone congener 9 (Fig. 2) does not significantly inhibit
cruzain (3% inhibition at 1 lM). Apparently a flat aromatic P2 sub-
stituent as found in 9 is unable to form favorable contacts with the
S2 pocket and/or cannot not properly direct the pendant chloro-
phenyl moiety towards the S3 subsite.
Superimposition of our coordinates for cruzainꢁ8a with our ear-
lier cruzainꢁ(1) crystal structure (PDB ID 2OZ2) reveals a number of

C. Bryant et al. / Bioorg. Med. Chem. Lett. 19 (2009) 6218–6221
6221
asite. Adenosine transporters from Trypanosoma brucei are impli-
cated in the pharmacology of a number of antitrypanosomals, for
example.15
the Department of Energy, Office of Biological and Environmental
Research, and by the National Institutes of Health, National Center
for Research Resources, Biomedical Technology Program, and the
National Institute of General Medical Sciences.
Vinylsulfone analogs 7 and 8 were also examined in the cell-
based assays and found to exhibit anti-parasite effects comparable
to 1 against T. brucei brucei parasites, without significant cytotoxic-
ity to Jurkat cells (Table 1). These analogs were tested as diastereo-
meric mixtures, and so one expects that the active diastereomers
7a and 8a should be as much as twice as potent against parasites.
While siRNA studies have implicated TbCatB as an important target
in T. brucei parasites, the analogs described herein exert a signifi-
cant anti-parasitic effect even in the absence of notable in vitro
activity against this enzyme. As irreversible inhibitors however,
one cannot rule out inhibition of TbCatB by 3, 4, 7, or 8 in the con-
text of parasite culture where the time scale of exposure is much
longer than in a biochemical assay.
In summary, two new classes of non-peptidic vinylsulfone-
based cysteine protease inhibitors have been discovered using an
empirical but structure-guided approach. Many of the non-pep-
tidic analogs exhibit enzyme and parasite activities comparable
to more traditional peptidic inhibitors and none of the non-pep-
tidic species were significantly cytotoxic to Jurkat (mammalian)
cells. A co-crystal structure of the non-peptidic vinylsulfone 8a
(SMDC-256047) bound to cruzain has been solved to high resolu-
tion, demonstrating effective binding to S10–S3 subsites of cruzain.
To our knowledge, this is the first co-crystal structure of a non-
peptidic inhibitor bound to cruzain and this structure should
greatly facilitate the design of more drug-like parasite cysteine
protease inhibitors.
References and notes
1. Benaim, G.; Sanders, J. M.; Garcia-Marchan, Y.; Colina, C.; Lira, R.; Caldera, A. R.;
Payares, G.; Sanoja, C.; Burgos, J. M.; Leon-Rossell, A.; Concepcion, J. L.;
Schijman, A. G.; Levin, M.; Oldfield, E.; Urbina, J. A. J. Med. Chem. 2006, 49(3),
892.
2. Frearson, J. A.; Wyatt, P. G.; Gilbert, I. H.; Fairlamb, A. H. Trends Parasitol. 2007,
23, 589.
3. Renslo, A. R.; McKerrow, J. H. Nat. Chem. Biol. 2006, 2, 701.
4. Sajid, M.; McKerrow, J. H. Mol. Biochem. Parasitol. 2002, 120, 1.
5. Gauthier, J. Y.; Chauret, N.; Cromlish, W.; Desmarais, S.; Duong le, T.;
Falgueyret, J. P.; Kimmel, D. B.; Lamontagne, S.; Leger, S.; LeRiche, T.; Li, C. S.;
Masse, F.; McKay, D. J.; Nicoll-Griffith, D. A.; Oballa, R. M.; Palmer, J. T.; Percival,
M. D.; Riendeau, D.; Robichaud, J.; Rodan, G. A.; Rodan, S. B.; Seto, C.; Therien,
M.; Truong, V. L.; Venuti, M. C.; Wesolowski, G.; Young, R. N.; Zamboni, R.;
Black, W. C. Bioorg. Med. Chem. Lett. 2008, 18, 923.
6. Engel, J. C.; Doyle, P. S.; Hsieh, I.; McKerrow, J. H. J. Exp. Med. 1998, 188, 725.
7. Brak, K.; Doyle, P. S.; McKerrow, J. H.; Ellman, J. A. J. Am. Chem. Soc. 2008, 130,
6404.
8. Gillmor, S. A.; Craik, C. S.; Fletterick, R. J. Protein Sci. 1997, 6, 1603.
9. Brinen, L. S.; Hansell, E.; Cheng, J.; Roush, W. R.; McKerrow, J. H.; Fletterick, R. J.
Structure 2000, 8, 831.
10. Somoza, J. R.; Zhan, H.; Bowman, K. K.; Yu, L.; Mortara, K. D.; Palmer, J. T.; Clark,
J. M.; McGrath, M. E. Biochemistry 2000, 39, 12543.
11. Tian, W. X.; Tsou, C. L. Biochemistry 1982, 21, 1028.
12. Rhodesain (4 nM) in 100
100 of 10 Z-FR-AMC in the same buffer. Progress curves were
determined for 360 s at room temperature for inhibitor concentrations
ranging from 100 to 0.003 M. Inhibitor dilutions that gave simple
exponential progress curves over wide range of kobs were used to
ll assay buffer was added to inhibitor dilutions in
l
l
lM
l
a
determine kinetic parameters. The value of kobs, the rate constant for loss of
enzymatic activity, was determined from an equation for pseudo first order
dynamics using Prism 4 (GraphPad). When kobs varied linearly with inhibitor
concentration, kass was determined by linear regression analysis. If the
variation was hyperbolic, kinact and Ki were determined from an equation
describing two-step irreversible inhibitor mechanism ½kobs ¼ kinact½Iꢂo=ð½Iꢂoþ
Kꢃi ð1 þ ½Sꢂ =KmÞÞꢂ and non-linear regression analysis.
Acknowledgements
This work was generously supported by funding from the San-
dler Foundation (to A.R.R. and L.S.B.), and the QB3-Malaysia Pro-
gram (to K.H.H.A.). We thank Dr. P. Vedantham for re-
synthesizing a sample of 7a. Part of this research was performed
at the Stanford Synchrotron Radiation Lightsource (SSRL), a na-
tional user facility operated by Stanford University on behalf of
the U.S. Department of Energy, Office of Basic Energy Sciences.
The SSRL Structural Molecular Biology Program is supported by
13. Liu, K.-C.;oShelton, B. R.; Howe, R. K. J. Org. Chem. 1980, 45, 3916.
14. Jaishankar, P.; Hansell, E.; Zhao, D. M.; Doyle, P. S.; McKerrow, J. H.; Renslo, A. R.
Bioorg. Med. Chem. Lett. 2008, 18, 624.
15. De Koning, H. P.; Jarvis, S. M. Mol. Pharmacol. 1999, 56, 1162.
16. Davis, I. W.; Leaver-Fay, A.; Chen, V. B.; Block, J. N.; Kapral, G. J.; Wang, X.;
Murray, L. W.; Arendall, W. B., 3rd; Snoeyink, J.; Richardson, J. S.; Richardson, D.
C. Nucleic Acids Res. 2007, 35, W375.