Phenothiazine inhibitors of trypanothione reductase as potential antitrypanosomal and antileishmanial drugs

Article abstract of DOI:10.1021/jm960814j

Given the role of trypanothione in the redox defenses of pathogenic trypanosomal and leishmanial parasites, in contrast to glutathione for their mammalian hosts, selective inhibitors of trypanothione reductase are potential drug leads against trypanosomiasis and leishmaniasis. In the present study, the rational drug design approach was used to discover tricyclic neuroleptic molecular frameworks as lead structures for the development of inhibitors, selective for trypanothione reductase over host glutathione reductase. From a homology-modeled structure for trypanothione reductase, replaced in the later stages of the study by the X-ray coordinates for the enzyme from Crithidia fasciculata, a series of inhibitors based on phenothiazine was designed. These were shown to be reversible inhibitors of trypanothione reductase from Trypanosoma cruzi, linearly competitive with trypanothione as substrate and noncompetitive with NADPH, consistent with ping-pong bi bi kinetics. Analogues, synthesized to define structure-activity relationships for the active site, included N-acylpromazines, 2-substituted phenothiazines, and trisubstituted promazines. Analysis of K(i) and I₅₀ data, on the basis of calculated log P and molar refractivity values, provided evidence of a specially favored fit of small 2-substituents (especially 2-chloro and 2-trifluoromethyl), with a remote hydrophobic patch on the enzyme accessible for larger, hydrophobic 2-substituents. There was also evidence of an additional hydrophobic enzymic region available to suitable N-substituents of the promazine nucleus. K(i) data also indicated that the phenothiazine nucleus can adopt more than one inhibitory orientation in its binding site. Selected compounds were tested for in vitro activity against Trypanosoma brucei, T. cruzi, and Leishmania donovani, with selective activities in the micromolar range being determined for a number of them.

Full text of DOI:10.1021/jm960814j

148
J . Med. Chem. 1998, 41, 148-156
Articles
P h en oth ia zin e In h ibitor s of Tr yp a n oth ion e Red u cta se a s P oten tia l
An titr yp a n osom a l a n d An tileish m a n ia l Dr u gs†
Cecil Chan,‡ Hong Yin,‡ J acqui Garforth,‡ J ames H. McKie,‡ Rabih J aouhari,‡ Peter Speers,‡
Kenneth T. Douglas,*,‡ Peter J . Rock,§ Vanessa Yardley,§ Simon L. Croft,§ and Alan H. Fairlamb|
Department of Pharmacy, University of Manchester, Oxford Road, Manchester M13 9PL, U.K., Department of Medical
Parasitology, London School of Hygiene & Tropical Medicine, Keppel Street, London WCIE 7HT, U.K., and Department of
Biochemistry, Medical Science Institute, University of Dundee, Dundee, DD1 4HN, U.K.
Received December 2, 1996X
Given the role of trypanothione in the redox defenses of pathogenic trypanosomal and
leishmanial parasites, in contrast to glutathione for their mammalian hosts, selective inhibitors
of trypanothione reductase are potential drug leads against trypanosomiasis and leishmaniasis.
In the present study, the rational drug design approach was used to discover tricyclic neuroleptic
molecular frameworks as lead structures for the development of inhibitors, selective for
trypanothione reductase over host glutathione reductase. From a homology-modeled structure
for trypanothione reductase, replaced in the later stages of the study by the X-ray coordinates
for the enzyme from Crithidia fasciculata, a series of inhibitors based on phenothiazine was
designed. These were shown to be reversible inhibitors of trypanothione reductase from
Trypanosoma cruzi, linearly competitive with trypanothione as substrate and noncompetitive
with NADPH, consistent with ping-pong bi bi kinetics. Analogues, synthesized to define
structure-activity relationships for the active site, included N-acylpromazines, 2-substituted
phenothiazines, and trisubstituted promazines. Analysis of Ki and I50 data, on the basis of
calculated log P and molar refractivity values, provided evidence of a specially favored fit of
small 2-substituents (especially 2-chloro and 2-trifluoromethyl), with a remote hydrophobic
patch on the enzyme accessible for larger, hydrophobic 2-substituents. There was also evidence
of an additional hydrophobic enzymic region available to suitable N-substituents of the
promazine nucleus. Ki data also indicated that the phenothiazine nucleus can adopt more
than one inhibitory orientation in its binding site. Selected compounds were tested for in vitro
activity against Trypanosoma brucei, T. cruzi, and Leishmania donovani, with selective
activities in the micromolar range being determined for a number of them.
In tr od u ction
an arsenical it is no surprise that melarsoprol is toxic,
the drug itself leading to the deaths of 4-8% of patients
treated with it.1 Nifurtimox and benznidazole are the
two major drugs available for Chagas’ disease prevalent
in South America (caused by T. cruzi), and there is
continuing debate about their effectiveness and safety.
Against Leishmania species the main drugs in use are
the pentavalent antimonials (sodium stibogluconate and
meglumine antimoniate) and, as well as toxicity prob-
lems, resistance is now reported to these. Pentamidine
has been used to treat this family of diseases but lengthy
courses are required and toxicity is a problem.2 Thus,
approaches to develop new drugs are sought in many
laboratories. A metabolic difference between the patho-
gen and the mammalian host recently discovered 3 may
provide a means of developing a selective antiparasitic
drug. Moreover, it may even prove possible to combat
all three diseases, African Trypanosomiasis, Chagas’
disease, and leishmaniasis, with a single agent.
Trypanosomiasis and leishmaniasis are major third-
world diseases, with many millions of new infections
presenting annually. Present chemotherapies are in-
adequate, toxic, or both with current drugs including
the arsenicals, nifurtimox, and pentamidine.1 The
trivalent arsenical drug melarsoprol is still widely used
against the second stage of African trypanosomiasis
(sleeping sickness) in which the parasitic pathogen
(Trypanosoma brucei rhodesiense or T. brucei gambi-
ense) has invaded the central nervous system, a stage
refractory to the drugs suramin and pentamidine.1 As
* To whom correspondence should be addressed. Tel: 0161 275 2386.
Fax: 0161 275 2396. E-mail: ktd@fs1.pa.man.ac.uk.
† Abbreviations: GR, glutathione reductase; GSSG, glutathione
disulfide; GSH, reduced glutathione; I50, concentration of inhibitor
required to reduce velocity by 50% under conditions defined in text or
table legend; Ki, inhibition constant; log P, logarithm to base 10 of the
partition coefficient for octanol-water (calculated as described); T[S]2,
trypanothione disulfide; T[SH]2, dihydrotrypanothione; TR, trypan-
othione reductase.
Glutathione is responsible for many cellular protec-
tion activities including those against free radicals and
oxygen-derived species. In the course of this action
glutathione disulfide (Figure 1) is formed from glu-
‡ University of Manchester.
§ London School of Hygiene & Tropical Medicine.
| University of Dundee.
X Abstract published in Advance ACS Abstracts, December 15, 1997.
S0022-2623(96)00814-X CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/15/1998

Phenothiazine Inhibitors of Trypanothione Reductase
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 2 149
F igu r e 1. Structures of glutathione disulfide and trypan-
othione in its disulfide form.
tathione reductase (GR). Trypanosomes do not contain
GR but rather an analogous enzyme, trypanothione
reductase (TR), and its substrate is trypanothione
disulfide, T[S]2 (Figure 1).3,4 Parasitic TR differs from
host GR in processing T[S]2 but not glutathione disulfide
(GSSG). Conversely, host GR does not process T[S]2.5,6
This mutual substrate exclusivity indicated that selec-
tive ligand design should be possible, making TR an
important potential target for drug design against
parasitic diseases involving trypanosomes and/or leish-
manias.3,6,7 Effective inhibition of TR would compro-
mise the parasites’ redox defenses, making them more
sensitive to redox-damage drugs, such as nifurtimox.
For example, a trypanothione reductase inhibitor might
be an effective antitrypanosomal and/or antileishmanial
drug in its own right or coadministration with a redox-
active drug such as nifurtimox might allow use of
lowered doses of the latter. Consequently, we undertook
the development of novel drug leads by a priori design
of enzyme inhibitors, selective for TR over human GR.
Our preliminary results provided strong, selective in-
hibitors based on either tricyclic heterocyclic nuclei8 or
peptides,9 neither family being obviously related to the
structure of the substrate for TR. We now report in
more detail on one family of tricyclic TR inhibitor, the
phenothiazines. A further impetus to developing such
a group of TR inhibitors was the literature reports that
several tricyclic neuroleptics, including the phenothi-
azines, possess weak, but established, antitrypanosomal
and antileishmanial activities.10-20
F igu r e 2. Schematic summary of regions of the active-site of
trypanothione reductase used in the design of the phenothi-
azine-based inhibitors (not to scale).
correspondence between the two structures, especially
for the core and active site: two regions of dissimilarity
were found in surface loops, not in the active site. The
active sites of human GR and TR differ in the hydro-
phobic pocket of the latter22 which complements the
spermidine portion of the substrate (Figure 2). In GR
this region has polar, positively charged side chains to
neutralize the carboxylate ion sites of GSSG. In addi-
tion, there is a second hydrophobic pocket (not used by
trypanothione as substrate), discovered for TR when
N,N-bis(benzyloxycarbonyl)-L-cysteinylglycyl-3-(dime-
thylamino)propylamide disulfide was found to be a good
alternative substrate.27 This second hydrophobic pocket
(called the Z-site), tentatively located from molecular
graphics analysis of the region of F396 (Figure 2), was
used to help design a TR-specific peptide inhibitor.9
Peptides make poor leads as drugs, and we needed to
develop alternative lead structures. Initially we docked
bicyclic (naphthalene) and tricyclic (anthracene) aro-
matic anchors in the spermidine hydrophobic pocket
formed by F114, W21, etc., but these were soon rede-
veloped to resemble well-known tricyclic neuroleptics,
the imipramines and the phenothiazine family.8 In view
of their easier synthetic access compared to that of the
imipramines, the phenothiazines were chosen to probe
structure-activity effects on TR inhibition and for
testing against the parasites.
Ap p r oa ch
Our initial approach to inhibitor design was guided
by a computer model of TR, constructed by modification
of GR using homology modeling,8 judged successful
based initially on its power to rationalize substrate
selectivities of GR and TR and to predict selective
inhibitors.8 The crystal structures of TR from Crithidia
fasciculata21-23 and recently T. cruzi,24,25 including the
TR:mepacrine complex,26 have been solved. Comparing
modeled and experimental coordinates showed a close
Resu lts
We have already shown8 that with the disulfide
substrate (T[S]2 or N,N′-bis(benzyloxycarbonyl)-L-cys-
teinylglycyl-3-(dimethylamino)propylamide disulfide27)
as variable substrate, trifluoperazine is a linear, com-
petitive inhibitor with a Ki value of 21.9 ( 1.7 µM. This
pattern of inhibition was confirmed by detailed analysis

150 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 2
Chan et al.
Ta ble 2. Inhibition by N-Substituted Phenothiazines,
2-Chlorophenothiazines, and 2-(Trifluoromethyl)phenothiazines
of T. cruzi Trypanothione Reductase under the Conditions
used in Table 1a
Ta ble 1. Inhibition by 2-Substituted Promazines of
Recombinant T. cruzi Trypanothione Reductase
(Approximately 0.3 µg‚mL-1) in 0.02 M HEPES Buffer
Containing 0.15 M KCl, 1 mM EDTA, 0.1 mM NADPH, and
0.12 mM Trypanothione Disulfide at 25 °Ca
S
S
R1
N
R
R1
N
(CH2)3NMe2
reference
no.
substitution pattern
I50 (µM)
to material
reference to
material
no.
R1
I50 (µM)
R1 ) H, R)
17
18
19
20
21
22
23
H
NI at 1000
*1412 ( 305
*1176 ( 319
*1028 ( 186
992 ( 112
Sigma-Aldrich
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
H
CF3
Cl
*108 ( 14
*110 ( 18
*35.4 ( 7.7
502 ( 20
Sigma-Aldrich
Sigma-Aldrich
Sigma-Aldrich
(CH2)3NH2
b
c
d
b
CH2C(CH3)2N(CH3)2
(CH2)3NHCH3
(CH2)2C(NH2)dNH
CH2CHCH3N(CH3)2
COCH3
b
CCH3dNOCH3
COCH2CH3
CO(CH2)2CH3
CCH3dNOBz
CO(CH2)4CH3
CH2OC6H4-p-C(CH3)3
COOH
351 ( 95
Scheme 1
357 ( 30
b
b
*271 ( 48
Sigma-Aldrich
Sigma-Aldrich
301 ( 86
CH2CHCH3CH2N(CH3)2 *249 ( 70
204 ( 16
Scheme 1
Scheme 1
Scheme 1
Scheme 2
R1 ) CF3, R )
171 ( 50
24
25
*73.7 ( 16.7
Sigma-Aldrich
Sigma-Aldrich
*161 ( 29
*3911 ( 306
*1187 ( 221
*458 ( 96
*348 ( 62
219 ( 13
(CH2)3N N(CH2)2OH
*72.0 ( 13.9
CH2OH
CONH2
COOCH3
CH2NH2
c
d
d
d
(CH2)3N NCH3
R1 ) Cl, R )
CO(CH2)2COOH
COCH2NEt2
COCH2NHPh
CO(CH2)3NEt2
COCH2Cl
CO(CH2)5NEt2
CO(CH2)2NEt2
COCH2Br
CO(CH2)2CONHNHPh
CO(CH2)2CONHCH2Ph
CO(CH2)2CH3
(CH2)2NMe2
(CH2)2C(NH2)dNH
(CH2)3NH2
(CH2)3NHCH3
(CH2)3NdN(CH2)2CH3
(CH2)3N
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
1484 ( 340
829 ( 219
256 ( 88
320 ( 79
862 ( 186
457 ( 57
802 ( 328
251 ( 80
151 ( 53
188 ( 24
95
*150 ( 11
307 ( 65
*139 ( 7
*126 ( 4
*95.5 ( 25.5
*80.8 ( 13.2
*81.3 ( 15.6
Scheme 3
e
e
f
g
h
i
j
CHdNOH
145 ( 12
Scheme 2
a
Compounds marked with an asterisk in the I50 column were
shown to not inhibit human erythrocyte glutathione reductase.
Ki values, where determined, were of a linear competitive type.
For 1 the Ki value was 59.1 ( 5.9 µM (coefficient of variation 7.9%),
for 2, 30.2 ( 2.6 µM (coefficient of variation 8.5%), and for 3, 10.8
b
( 1.1 µM (coefficient of variation 10.2%). Schnitt,J . ; Boitard,
Scheme 3
Scheme 3
Scheme 3
J .; Comoy, P.; Hallot, A.; Suquet, M. Bull. Soc. Chim. Fr. 1957,
938; Chem. Abstr. 1962, 56, 478. Rhone Poulenc, FR 70, 804
addition to FR 1173133. c Chem. Abstr. 1960, 54, 2372i. Rhone
k
l
d
Poulenc GB 816538, 1959. Chem. Abstr. 1964, 60, 5515b. Cusic,
J . W.; Lowrie, H. S. (G. D. Searle & Co.) U.S. Pat. 3112310, 1962.
m
n
o
d
p
for 1, 2, 3, 22, 23, 24, 25, and 41 (data not shown, but
Ki values are given in Table 2). As TR is a two-substrate
enzyme, with NADPH as the second substrate, we also
determined the patterns of inhibition for NADPH as the
variable substrate, with the results shown in Figure
3a-c (Supporting Information). The pattern of inhibi-
tion of TR by trifluoperazine from the diagnostic set of
plots in Figure 3 is uncompetitive against NADPH as
variable substrate.28 The value of K′i determined using
the uncompetitive model was 31 µM.
Values of I50 are recorded in Table 1 for a range of
different 2-substituents. In Table 2 values of I50 (and
in some cases Ki) for the effects of multiple substitution
of the tricyclic nucleus are collected for both ω-(N,N-
dimethylamino)propyl and acetyl substitutions on the
tricyclic bridging nitrogen. Table 3 collects inhibition
data for the effect of altering the nature of the tricyclic
bridging nitrogen for a series with fixed 2-substituents
in the tricyclic nucleus. Compounds indicated by as-
terisks in Tables 1-3 gave no detectable inhibition of
human glutathione reductase up to concentrations of 1
mM (0.1 mM for a few compounds of limited solubility).
In Figure 4 (Supporting Information) log(1/I50) is
plotted against log P values calculated for the 2-sub-
stituents in Table 1. Figure 5 (Supporting Information)
shows a similar plot of the inhibition data for molecular
refractivity values calculated for these 2-substituents.
There was no significant correlation of log(1/I50) with
log P or the molar refractivities for the substituents in
(CH2)2NEt2
a
NI ) no inhibition. Ki values were of the linear competitive
type for the following compounds: 22, Ki ) 216 ( 33 µM
(coefficient of variation 15.1%); 23, Ki ) 127 ( 11 µM (coefficient
of variation 8.6%); 24, Ki ) 21.2 ( 3.3 µM (coefficient of variation
15.6%); 25, Ki ) 23.0 ( 1.8 µM (coefficient of variation 7.9%); 41,
b
Ki ) 18.7 ( 2.1 (coefficient of variation 11.2%). Godefroi, E. F.;
Wittle, E. L. J . Org. Chem. 1956, 21, 1163. Rhone-Poulenc, Patent
DE 1012915, 1953. c G. D. Searle & Co. U.S. Pat. 2590125, 1947;
d
U.S. Pat. 2687414, 1948. Chem. Abstr. 1964, 60, 5515b. G. D.
Searle & Co., U.S. Pat. 3112310, 1962. e Chem. Abstr. 1957, 51,
501a. Gailliot, P.; Robert, J . Rhone Poulenc U.S. Pat. 2738399,
1956. f Yale, H. L. J . Am. Chem. Soc. 1955, 77, 2270. Chem.
g
Abstr. 1990, 113, 59071. El-Ezbawy, S. R.; Alshaikh, M. A. J .
Chem. Technol. Biotechnol. 1990, 47, 209-18; Chem. Abstr. 1967,
67, 90823. G. D. Searle & Co. U.S. Pat. 332046, 670516, 650118.
Chem. Abstr. 1961, 55, 6484c. Gritsenko, A. N.; Zhuralev, S. V.;
Skorodumov, V. A. Uch. Zap. Inst. Farmakol. Khimioter. Akad.
h
i
Med. Nauk. USSR 1958, No.1, 13-26. Kano,H.; Makisumi, Y.
Shionogi Kenkyusho Nempo 1957, 7, 511-516; Chem. Abstr. 1958,
52, 10094. Shurawlew, S. V.; Grizenko, A. N. Zh. Obshch. Khim.
1956, 26, 3385; Engl. Trans. 3769. j Chem. Abstr. 1968, 68, 49530.
Yarmukhametova, D. Kh.; Kudryavstev, B. V.; Buzlama, V. S. Izv.
Akad. Nauk. SSSR, Ser. Khim. 1967, 7, 1506. k Chem. Abstr. 1996,
64, 11220b. Rhone-Poulenc U.S. Pat. 2645640, 1951; BASF DE
1011887, 1955; SK and F NL 6505736, 1969. Chem. Abstr. 1963,
59, 11518h. Yoshitomi Pharm. Pat. J P 6783, 1958. Sawizkaya,
N. V.; Czizin, Y. S.; Stchvkina, M. N. Zh. Obshch. Khim. 1956,
l
m
n
26, 2900-2905. Engl. Transl. S 3227. Rhone-Poulenc U.S. Pat.
2830987, 1956. o Rhone-Poulenc GB Pat. 780193, 1955. Morisa-
p
wa, K.; Terai, Y.; Kawahara, S.; Ichikawa, K.; Oka, Y. Yakugaku
Kenkyu 1957, 29, 161-3; Chem. Abstr. 1959, 53, 538145.

Phenothiazine Inhibitors of Trypanothione Reductase
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 2 151
Ta ble 3. Inhibition by Polysubstituted Phenothiazines of T. cruzi Trypanothione Reductase under Conditions Used in Table 1
R3
S
R2
R1
N
R
R4
no.
R1
R2
R3
R4
R
I50 (µM)
reference to material
44
45
46
47
48
49
50
51
52
H
Cl
H
H
H
H
H
H
H
H
H
(CH2)3NMe2
(CH2)3NMe2
(CH2)3NMe2
(CH2)3NMe2
(CH2)3NMe2
COCH3
COCH3
COCH3
COCH3
*765 ( 57
472 ( 156
175 ( 60
a
COCH3
Cl
Cl
COCH3
COCH3
CCH3dNOBz
H
H
C, H, N by HRMS, f
C, H, N by HRMS, f
C, H, N by HRMS, f
chlorpromazine sulfoxide
b
c
d
e
199 ( 78
Cl
O
*76.6 ( 63
436 ( 37
COCH2Cl
COCH3
Cl
H
*2811 ( 855
402 ( 40
COCH3
COCH3
COCH3
49.1 ( 8.7
a
b
Chem. Abstr. 1955, 49, 3268i. Rhone Poulenc U.S. Pat. 2645640. Mousseron, M.; Kamenka, J . M.; Legendre, P.; Stenger, A. Chim.
Ther. 1996, 7, 397-402. c Zupancic, B. G. Kem. Ind. 1976, 25 (4), 223-8; Chem. Abstr. 1977, 86, 55367. Chem. Abstr. 1991, 115, 232258.
d
Merck U.S. Pat. 5036067. e Chem. Abstr. 1986, 104, 15081. Merck Frosst Eur. Pat. EP 306639.
Ta ble 4. Sensitivity of T. brucei Bloodstream Form
Trypomastigotes and L. donovani Amastigotes and T. cruzi
Trypomastigotes in Mouse Peritoneal Macrophages to
Phenothiazines
linear competitive versus T[S]2 and uncompetitive ver-
sus NADPH is consistent29 with a ping-pong bi bi
mechanism for TR, a model which has been suggested
30-32
from initial velocity substrate plots for T[S]2
and
ED50 (µM)a
the NADPH-analogue inhibitor 2′,3′-cyclic NADPH.31
Relative to compounds 1, 2, and 3 (with H, CF3, or C1
in the 2-position, respectively), almost any substitution
weakens binding. We return to this point later. Re-
placing a 2-acetyl group 4 progressively by 2-propionyl
6, 2-butanoyl 7, and 2-(p-tert-butylphenoxy)methyl 10
substituents increases the inhibition strength steadily
with 10 binding approximately 3 times as tightly as 4.
The simplest explanation of this is that there is a
hydrophobic site available to interact with such 2-sub-
stituents. This is not evident merely by inspecting 1,
2, and 3 where the major changes are polarity and, to
a lesser extent, steric effects. The slightly better
inhibition by 8 compared to 5 may also indicate a
hydrophobic region, which can be used selectively by
2-substituents.
compound (code)
T. brucei
L. donovani
T. cruzi
2 (triflupromazine)
3 (chlorpromazine)
4 (77009/CC64)
5 (CC111)
6 (CC62)
7 (CC70)
8 (CC112)
9 (CC87)
10 (RJ 15)
15 (79009/CC201)
16 (75009/CC196)
44 (BA 21752)
45 (CC35)
46 (CC124)
47 (CC126)
51 (CC80)
52 (CC22)
33 (CC127)
34 (CC104)
3.2
3.69
6.42
10-30
10.1
3.26
1.56
>30
>30
0.37
7.35
3.01
10-30
1.29
3.5
>30
>30
0.97
15.2
>30
>30
3.83
>30
<3, >1
4.90
T
>30
>30
10-30
16.5
>10
T
>30
>30
>30
>30
>30
T
>30
T
>30
48.4
>30
12.1
6.78
>30
>30
5.97
34.3
>30
>30
3.30
>30
10
10-30
29.5
>30
>30
>30
>30
>30
10-30
T
>30
>30
54.9
>30
5.9
Figure 4 (Supporting Information) shows that if clear
outliers (CO2H, CH2OH, and Cl) are omitted, there is a
weak trend to increased inhibitory strength for higher
values of P, but I50 decreases only 1.2-fold for a 5-fold
change in log P (based on the correlation equation for
the data in Figure 4). The outliers in Figure 4 can be
further examined using the molar refractivity plot of
Figure 5 (Supporting Information). (Essentially similar
plots are obtained for surface area, volume, polarizabil-
ity, etc. for the 2-substituent, but as these and the molar
refractivity are not independent parameters; the latter
has been chosen simply to illustrate the point.) Two
major regions are evident in Figure 5. For molar
refractivities of less than approximately 15 Å3, the data
are collected into a box labeled region I. Although there
is a very small increase in inhibition strength as the
molar refractivity increases, the value of log(1/I50) is
almost independent of molar refractivity. For molar
refractivities greater than approximately 15 Å3 (box
labeled region II). The major distinction between sub-
stituents in regions I and II is size (easily determined
if the size of the longest axis of the lowest energy
conformation of the 2-substituents involved is plotted
(data not shown) instead of molar refractivity). For
35 (CC100)
36 (CC90)
18 (BA 18971)
24 (fluphenazine)
39 (GP 38889)
a
T indicates toxic to the macrophages.
the 2-position if the whole data set were included. For
the compounds referred to in region II of Figure 5
(Supporting Information), the data were correlated by
the following equation:
log(1/I50) ) 2.89 + 0.02MR
(r ) 0.865, F ) 17.86)
The activities of a number of the compounds in Tables
1-3 against T. brucei, L. donovani, and T. cruzi are
given in Table 4.
Discu ssion
The reversible linear competitive inhibition kinetics
observed with this family of structures is consistent with
the design process which located the lead nucleus in the
hydrophobic pocket of the active site formed by residues
M113, A343, and W21.8 The pattern of inhibition of

152 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 2
Chan et al.
the stronger inhibition by the 2-CH2NH2 (15) than the
2-CH2OH (12) substituent. If the aminomethyl sub-
stituent had bound in its protonated form (reasonable
+
as the pKa of RCH2NH3 is approximately 10 in free
solution), a stronger interaction would have been ex-
pected if E18′ were in the CO2- form. Other groups with
potential to hydrogen bond with E18′ (e.g. 16, 5), as
donors and as acceptors, also failed to give the increase
in binding strength expected of such interactions. Thus
the 2-substituents may be directed to the face of the
hydrophobic pocket away from E18′ (i.e. toward the Met-
113) or not large enough to access E18′. The strength
of binding of 2-chloro, and to a lesser extent 2-CF3,
relative to the other small 2-substituents indicates that
a particularly favorable interaction/molecular fit is
possible for them. This interaction may be provided,
at least in part, by the side chain of M113. We discuss
later the X-ray data for the TR-mepacrine complex.26
The bridging nitrogen can tolerate many substituents
(Table 2). However, within this broad statement are
covered several separate influences on binding for
detailed substitution differences. Relative to the ω-(N,N-
dimethylamino)propyl group (1, 2, and 3), almost any
change is deleterious to inhibition (compare 1 with 17-
23 or 3 with 38-43). A tertiary amine is important as
inhibition weakens by 1 order of magnitude on replacing
an ω-NMe2 group (as in 1 or 3) by NHCH3 (as in 20 or
40) or NH2 (as in 18 and 39). One substitution which
maintains inhibitory strength is the introduction of a
substituted piperazine ring (24 and 25 each bind ap-
proximately 1.5-fold more strongly than 2, based on Ki
values). The piperazine is a specially favored case as
replacement of the terminal dimethylamino group of 3
by a piperidine (42) weakens inhibition. It is possible
that for 24 and 25 the extension of the distal nitrogen
(further from the tricyclic nucleus) by CH2CH2OH and
CH3, respectively, enables a remote hydrophobic region
to be accessed. Going from N-alkyl to N-acyl substitu-
tion of the nuclear nitrogen adversely affects inhibition.
Thus, amide 27 is almost 1 order of magnitude poorer
than its alkylamino counterpart, 43. The terminal
amino compound 32 inhibits 6-fold more strongly than
the terminal carboxylic species 26. The terminal methyl
derivative 36 is better still, and its alkyl tail may reach
the Z-site (see Figure 2) avoiding the Glu pincers, with
which the (CH2)3NR2 groups were originally designed8
to interact in their protonated form. For either binding
F igu r e 6. View of the active site of trypanothione reductase
showing 2-(N-(benzyloxy)methylimino)promazine (8) with its
benzyloximo function located in the general region of the Z-site
(around F396′, V58, and V53) and showing a retained interac-
tion of the ω-ammonio function of the tricyclic nitrogen atom’s
side chain with the Glu pincers.
substituents in region I (H, Cl, CF3, CHdNOH, CH2-
NH2, CO2Me, CONH2, COCH3, CH2OH, and CO2H),
inhibitory strength varies by almost 2 orders of mag-
nitude despite the small differences in substituent size.
The very small change in inhibitory power (region II)
for a considerable variation in substituent (and for a
wide range of chemical functionalities) means that
larger substituents can be easily accommodated. The
most straightforward explanation of this is that these
substituents point toward bulk solution. The small but
perceptible increase in inhibition with molar refractivity
(size, Figure 5, Supporting Information) and log P
(Figure 4, Supporting Information) for these substitu-
ents indicates that some relatively nonspecific interac-
tions are possible with a surface hydrophobic patch of
TR for these larger groups. For smaller 2-substituents
of region II (Figure 5), there is an accessible hydrophobic
region, defined in general terms by M113 and Y110. For
larger substituents of region II (e.g. PhCH2ONdC),
modeling indicates that if the tricyclic aromatic nucleus
is sited in the general region of F114/W218 then the
2-substituent can access the Z-site (see Figure 2), but
this necessitates a compromise with poorer contact
available to the tricyclic nucleus (see Figure 6).
-
mode a terminal CO2 group would be disfavored.
Despite inhibition by 30 and 33 (potential alkylating
agents toward TR), there was no evidence of irreversible
inhibition of TR by 30 even on extended incubation at
concentrations up to millimolar. Thus, there can be no
nucleophilic side chains in the immediate vicinity of the
reactive R-ketoalkyl site in 30 and 33, consistent with
the design-hypothesis model (Figure 2). The only
potentially nucleophilic residues being serines, but with
no obvious activation mechanism such as the charge-
relay system of the serine proteases.
For the smaller substituents in region I some very
specific interactions clearly take place. The stronger
inhibition by the free 2-oxime (compared to 5) indicates
that a bonding feature additional to hydrophobic is
available to appropriate 2-substituents. One possibility
considered early was the use of E18′ to bind part of the
ligand (see Figure 2). To try to test this we introduced
a range of substituents. Data in Table 1 are consistent
with E18′ involvement of the inhibitors. For example,
There are indications that the phenothiazine nucleus
can take up more than one orientation in the active-
site hydrophobic pocket which involves W21 and M113.
Thus, changing the substituent on the bridging tricyclic
nitrogen atom from (CH2)3NMe2 to (CH2)3N(C2H4)2Nn-
Pr weakens binding slightly (2-fold) for a 2-chloro
-
if the 2-carboxylic acid (11) binds as its CO2 form, its
poor inhibition (I50 ) 3911 µM and 8-fold weaker than
the amide, 13) would be explicable if E18′ were also
ionized to CO2- at pH 7.25. This is in accord also with

Phenothiazine Inhibitors of Trypanothione Reductase
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 2 153
nucleus (compare Ki values of 3 (10.8 µM) and 41 (18.7
µM)) but strengthens it slightly (by 30%) for 2-trifluo-
romethyl (compare 2 (30.2 µM) with 25 (23.0 µM)). This
type of substitution effect on binding strength (1/Ki) is
readily accommodated if the precise alignments of the
two families (Cl and CF3) of inhibitors in the active site
do not correspond exactly.
further (57-fold relative to 50), giving an I50 of 49 µM
for the triacetyl compound 52.
In summary, the binding of phenothiazine tricyclic
inhibitors of TR provides evidence of multiple modes of
inhibitor acceptance. In addition, the effective area of
the hydrophobic pocket available to inhibitors is not
limited to the hydrophobic wall formed by M113, F114,
and W21 used to recognize the spermidinyl sector of the
TR substrate23 and made use of in the original inhibitor
design paper.8 This region of the active site discrimi-
nates T[S]2 and GSSG23 and is contiguous with another
hydrophobic region (e.g. F296′, P398′, L399′, V58, V53),
not used to recognize the natural substrate, but which
may bind the present tricyclics, some peptide inhibitors9
and a number of alternative substrates.27,33
In an attempt to see if the strong binding properties
of the 2-chloro group could be combined with access to
a hydrophobic region near Tyr-110, derivative (47)
(Table 3) was prepared. With only a 2-Cl group in the
phenothiazine nucleus (3), the I50 value was 35 µM, and
with only a 2-O-benzyloximo function (8) it was 204 µM.
Combining these functions (47) resulted in an I50 value
of 199 µM. Clearly, the favored 2-chloro type of interac-
tion is not possible in the presence of a bulky 8-O-
benzyloximo group. If the 2-Cl group “directs” binding
(e.g. by interaction with M113), then modeling indicates
that the 2-benzyloximo group is unable to optimally
interact with the hydrophobic patch formed by residues
including F396′, L399′, Y110. The large TR active site
allows inhibitors as small as the currently studied
tricyclics considerable scope. An alternative location for
one of the tricyclics (8) is shown in Figure 6, which also
shows how far from the original docking site used for
lead generation8 such a molecule can be located, at least
by molecular modeling, and still provide a theoretical
rationale for binding. In the X-ray structure to 2.9Å
resolution for TR from T. cruzi with mepacrine, an
acridine-based inhibitor,26 of the four TR active sites per
crystal asymmetric unit only one showed electron
density ascribable to the mepacrine inhibitor. The
inhibitor lies in the active site, its tricyclic nucleus in
the region around M113 and W21, and the chloro group
stabilized by interaction with the indolyl NH. The
acridine ring nitrogen atom (N10) may be protonated.26
The alkylamino side chain of mepacrine is extended
toward E18′ and not in the same sense as modeled in
our original design process.8 The phenothiazine ring
system (unprotonated at neutral pH) may bind with a
different orientation to the acridine nucleus in the large
hydrophobic cavity of the active site considering the
kinetic evidence above of more than one family of
binding location even for inhibitors all sharing the
phenothiazine nucleus. The opportunity for the tricy-
clics to adopt more than one bound position within the
general confines of the large TR active site may con-
tribute to a low level of X-ray occupancy for mepacrine.
T. brucei bloodstream form trypomastigotes showed
the highest in vitro sensitivity to the phenothiazines
with several compounds having ED50 values below 5 µM,
8, 15, and 33 being most notable. These rapidly
dividing extracellular culture forms have a doubling
time of around 6 h. In comparison pentamidine, the
positive control drug used, has an ED50 value of ∼0.01
µM. Pentamidine (a polyamine) is not a structural
analogue of the tricyclic family under study but is the
drug currently most effective against T. brucei and used
as a test of relative efficacy of new compounds. The
intracellular amastigote stages of both L. donovani and
T. cruzi had lower sensitivity to the phenothiazines than
T. brucei, with T. cruzi being the least sensitive. Both
of these parasites, grown in murine peritoneal mac-
rophages, have a slower rate of division than extracel-
lular forms. Interpretation of these assays is more
complex as at higher concentrations the compounds
killed both parasites and host cells (T). Chlorprom-
azines 3, 47, and 33 showed interesting selective activity
against L. donovani with ED50 values of <10 µM. In
the same model the ED50 value of the standard drug
sodium stibogluconate is around 5 µg SbV/mL. Only 39
showed any promising activity against T. cruzi with an
ED50 value close to that of the standard drug nifurtimox
(around 3 µM). There was no correlation between ED50
values for in vitro antiparasitic activities from Table 4
and TR inhibitory strength, even using a log-log plot.
Thus if the antiparasitic action of these compunds arises
from TR inhibition, molecular features other than TR
binding are dominant, e.g parasitic cellular penetration,
metabolism. It has been concluded from previous data
for antitrypanosomal/antileishmanial activities of tri-
cyclic compounds that no one structural feature controls
activity.11,19 While the data in this study are consistent
with such a view, the range of activities (Table 4) is
large for any given parasite, and strong antiparasitic
action can certainly be achieved (15 improves on 3 by 1
order of magnitude against T. brucei). Thus structural
optimization of these leads against parasites, while still
to be acheived, is a reasonable goal.
Supplementing a 2-acetyl group already in place (4)
by introducing a second acetyl moiety at the 8-position
(45) makes little difference to the inhibitory strength if
the substituent on the heterocyclic nitrogen atom is
ω-(N,N-dimethylamino)propyl. However, derivatives
with an acetyl group on the heterocyclic nitrogen atom
behave sufficiently differently to indicate membership
of a different class of inhibitors with respect to the
details of their active-site positioning. In the promazine
family (N-(CH2)3NMe2) with an unsubstituted 8-posi-
tion, introduction of 2-Cl (3 cf. 2) improved inhibition
by 3-fold and a 2-COCH3 group (4) weakened inhibition
by 4.6-fold. However, with an N-COCH3 group and an
8-COCH3 both present, substitution of the 2-hydrogen
(of 50) by 2-Cl (51) improved inhibition 7-fold and a
2-COCH3 group (52) increased inhibitory strength still
Exp er im en ta l Section
Promazine, trifluoperazine, and chlorpromazine were pur-
chased from Sigma Chemical Co., Poole, Dorset. Derivatives
4, 8, 17-20, and 37-44 were gifts from Ciba-Geigy, Basel
(additional quantities of 4 were synthesized as described
below), and 48 from Dr. D. Attwood. Other derivatives were
synthesized as described below or as in the original literature
given at the appropriate entry in the tables of kinetic data.

154 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 2
Chan et al.
Sch em e 1
Sch em e 2
Purity was established by TLC in three solvent systems, and
compounds were characterized by high-resolution mass spec-
trometry and NMR spectroscopy.
Sch em e 3
En zym e Isola tion , Assa y, a n d In h ibition Stu d ies. Try-
panothione reductase from T. cruzi was isolated by means of
overexpression of the gene in Escherichia coli J M109 cells
bearing the expression vection pBSTNAV34 as previously
described.8 The enzyme was homogeneous by SDS-PAGE and
had a specific activity identical with that of wild-type TR.5
Enzyme activity was assayed at 25 °C in 0.02 M HEPES buffer,
pH 7.25, containing 0.15 M KCl, 1 mM EDTA, 0.12 mM T[S]2,
and 0.1 mM NADPH5 at an enzyme concentration of ap-
proximately 0.3 µg‚mL-1. Human erythrocyte GR was isolated
from human erythrocytes as described35 and assayed following
literature conditions.36 Assays were conducted spectrophoto-
metrically at 340 nm using a Peltier-thermostated cuvette
holder in a Cary 1E UV-visible spectrophotometer and the
Cary Enzyme Kinetics Software or using a Perkin-Elmer
Lambda 3B instrument thermostated to (0.02 °C by means
of water circulated from a Haake GH Model D8 pump-
thermostating system.
Inhibition type was assessed by analyzing the patterns of
three diagnostic classes of plot: 1/vo versus 1/[So] for various
[I]; 1/vo versus [I] for various [So]; and [So]/vo versus [I] at
various [So]. Values of Ki for competitive inhibition were then
2
determined by direct, weighted (1/vo for weighting) least-
promazine (5, 8) were prepared under standard conditions
using the appropriate alkoxyamine hydrochlorides in pyridine.
2-(Aminomethyl)promazine was synthesized by Scheme 2. N,2-
Diacetylphenothiazine was converted to 2-carboxyphenothi-
azine38 and alkylated with 1 equiv of sodium hydride and
3-(dimethylamino)propyl chloride to give the corresponding
ester. Reduction of the ester function using diisobutylalumi-
num hydride gave 2-(hydroxymethyl)promazine (12). Oxida-
tion of the primary alcohol using Swern’s conditions yielded
the aldehyde, which was converted to the oxime (18). Lithium
aluminum hydride reduction of this oxime afforded the desired
2-aminomethyl compound (17). 2,8-Diacetylphenothiazine39
and 2-acetyl-8-chlorophenothiazine (similarly prepared) were
N-alkylated to the corresponding promazines (45, 46) and the
oxime prepared from the 2-acetyl compound.
N-Acyl-2-chlorophenothiazines were prepared (Scheme 3) by
acylation of the appropriate 2-chlorophenothiazine. N-(3-
Carbomethoxypropionyl)-2-chlorophenothiazine was de-esteri-
fied using sodium hydroxide and the resulting carboxylic acid
(26) converted to the amide (13) and hydrazide (34) by the
ethyl chloroformate method.
Molecu la r Mod elin g. The design of inhibitors was guided
by visualization of particular inhibitor target structures with
the active site of TR using QUANTA 4.1 and CHARMm 4.2
(Molecular Simulations Inc., Burlington, MA, 1995). Calcula-
tions of log P (not determined experimentally) and other
molecular parameters were carried out using the ChemPlus
squares nonlinear regression analysis of the raw data using
the equation for linear competitive inhibition (v ) Vmax[So]/
([So] + Km(1 + [I]/Ki))) using the Grafit program (Erithacus
Software). Values of Vmax and Km were obtained by least-
squares nonlinear regression analysis using Grafit. Values
of I50, the concentration of inhibitor required to give 50%
inhibition under the assay conditions described above, were
determined by nonlinear least-squares fitting of assay velocity
versus inhibitor concentration using the equation derived from
linear competitive inhibition. Reversibility of enzyme inhibi-
tion was assessed using dialysis at 4 °C overnight against
assay buffer (with two changes).
Inhibitor design used a molecular graphics homology model
of T. congolese TR constructed in this laboratory by modifica-
tion of the published coordinates of human erythrocyte GR.37
Toward the end of the project, the published C. fasciculata21-23
X-ray diffraction coordinates were used to replace the modeled
coordinates and rationalize the results obtained, although now
the T. cruzi coordinates are on the Brookhaven database.
Syn th esis. 2-Acylpromazines were prepared (Scheme 1)
by standard acylation of phenothiazine with acyl chlorides
under reflux (2 h) in toluene, followed by Friedel-Crafts
acylation (AlCl3/anhydrous CS2, reflux for 6 h) of the N-acyl
compounds under literature conditions.38,39 N-Deacylation
followed by alkylation of the 2-acylphenothiazines with sodium
hydride and 3-(dimethylamino)propyl chloride afforded the
desired substituted-promazines (4, 6, 7, 9). Oximes of 2-acetyl-

Phenothiazine Inhibitors of Trypanothione Reductase
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 2 155
(7) Schirmer, R. H.; Mu¨ller, J . G.; Krauth-Siegel, R. L. Disulfide-
reductase inhibitors as chemotherapeutic agents: The design
of drugs for trypanosomiasis and malaria. Angew. Chem., Int.
Ed. Engl. 1995, 34, 141-154.
(8) Benson, T. J .; McKie, J . H.; Garforth, J .; Borges, A.; Fairlamb,
A. H.; Douglas, K. T. Rationally designed selective inhibitors of
trypanothione reductase. Biochem. J . 1992, 286, 9-11.
(9) Garforth, J .; McKie, J . H.; J aouhari, R.; Benson, T. J .; Fairlamb,
A. H.; Douglas, K. T. Rational design of peptide-based inhibitors
of trypanothione reductase as potential antitrypanosomal drugs.
Amino Acids 1994, 6, 295-300.
(10) Doyle, P. S.; Weinbach, E. C. The activity of tricyclic antidepres-
sant drugs against Trypanosoma cruzi. Exp. Parasitol. 1989, 68,
230-234.
(11) Evans, A. T.; Croft, S. L. Antileishmanial actions of tricyclic
neuroleptics appear to lack structural specificity. Biochem.
Pharmacol. 1994, 48, 613-616.
(12) Hewlett, E. L.; Pearson, R. D.; Zilberstein, D.; Dwyer, D. M.
Antiprotozoal activity of tricyclic compounds. Science 1985, 230,
1063-1064.
(13) Pearson, R. D.; Manian, A. A.; Harcus, J . L.; Hall, D.; Hewlett,
E. K. Lethal effect of phenothiazine neuroleptics on pathogenic
protozoan Leishmania donovani. Science 1982, 217, 369-371.
(14) Pearson, R. D.; Manian, A. A.; Hall, D.; Harens, J .; Hewlett, E.
L. Antileishmanial activity of chlorpromazine. Antimicrob.
Agents Chemother. 1984, 25 (5), 571-574.
(15) Zilberstein, D.; Dwyer, D. M. Antidepressants cause lethal
disruption of membrane function in the human protozoan
parasite Leishmania. Science 1984, 226, 977-979.
(16) Seebeck, T.; Gehr, P. Trypanocidal action of neuroleptic phe-
nothiazines in Trypanosoma brucei. Mol. Biochem. Parasitol.
1983, 9, 197-208.
(17) Stieger, J .; Seebeck, T. Monoclonal antibodies against a 60 kDa
phenothiazine-binding protein from Trypanosoma brucei can
discriminate between different trypanosome species. Mol. Bio-
chem. Parasitol. 1986, 21, 37-45.
module of Hyperchem Version 4.0 (Hypercube Inc., Waterloo,
Ont., Canada, 1995),40 which uses an additive approach based
on atomic fragment contributions to the overall molecular
property. Partial least squares analyses were performed using
the QSAR facility of SYBYL 6.2 (Tripos Inc., St. Louis, MO,
1995).
Biologica l Test in g: in Vitr o Assa ys. Tr yp a n osom a
br u cei br u cei (str a in S427). Bloodstream form trypomas-
tigotes, cultured in HMI-9 medium,41 were diluted to 105 cells/
mL and aliquots (100 µL) dispensed into wells in a 96-well
microtiter plate. Test compounds diluted in HMI-9 medium
were added in another 100 µL volume to give a final drug
concentration. Each drug concentration was set up in tripli-
cate in a 3-fold dilution series, and cultures were incubated
at 37 °C in 5% CO2/air mixture. After 48 h cells were counted
in a ZM Coulter Counter with a 70 µm aperture or by
Neubauer haemocytometer. ED50 values were calculated by
linear regression analysis from dose-percent inhibition curves.
Leish m a n ia d on ova n i (MHOM/ET/69/L82). Mouse peri-
toneal macrophages were isolated and maintained in Labtek
8 chamber tissue culture slides in RPMI 1640 medium with
10% heat-inactivated foetal calf serum at 37 °C in a 5% CO2
mixture. L. donovani amastigotes were isolated from spleens
of infected hamsters were used to infect cultures. Twenty-
four hours after infection, aliquots of fresh medium containing
test compounds were added in 3-fold dilution series in qua-
druplicate and maintained for 5 days following previously
described methods.42 After 5 days slides were methanol-fixed
and Giemsa-stained, and the proportion of infected macro-
phages was determined in each culture. ED50 values were
calculated by linear regression analysis.
(18) Croft, S. L.; Walker, J . J .; Gutteridge, W. E. Screening of drugs
for rapid activity against Trypanosoma cruzi trypomastigotes
in vitro. Trop. Med. Parasitol. 1988, 39, 145-148.
(19) Hammond, D. J .; Hogg, J .; Gutteridge, W. E. Trypanosoma
cruzi: possible control of parasite transmission by blood transfu-
sion using amphiphilic cationic drugs. Exp. Parasitol. 1985, 60,
32-42.
(20) Hammond, D. J .; Cover, B.; Gutteridge, W. E. A novel series of
chemical structures active in vitro against the trypomastigote
form of Trypanosoma cruzi. Trans. R. Soc. Trop. Med. Hyg. 1984,
78, 91-95.
Tr ypa n osom a cr u zi (MHOM/BR/00/Y). Trypomastigotes,
derived from infected cultures of L6 rat myoblast cells, were
used to infect mouse peritoneal macrophages at a ratio of 5
parasites:1 macrophage and incubated at 37 °C in 5% CO2/air
mixture. Twenty-four hours after infection, cultures were
exposed to fresh medium containing test compounds. Each
compound was tested in a 3-fold dilution series in quadrupli-
cate. After 3 days, cultures were fixed, stained, and analyzed
as described above for L. donovani.
(21) Kuriyan, J .; Kong, X.; Krishna, T. S. R.; Sweet, R. M.; Murgolo,
N. J .; Field, H.; Cerami, A.; Henderson, G. B. X-ray structure of
trypanothione reductase from Crithidia fasciculata at 2.4-
Åresolution. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 8764-8768.
(22) Hunter, W. N.; Bailey, S.; Habash, J .; Harrop, S. J .; Helliwell,
J . R.; Aboagye-Kwarteng, T.; Smith, K.; Fairlamb, A. H. Active
site of trypanothione reductase. A target for rational drug design.
J . Mol. Biol. 1992, 227, 322-333.
(23) Bailey, S.; Smith, K.; Fairlamb, A. H.; Hunter, W. N. Substrate
interactions between trypanothione reductase and N1-gluta-
thionylspermidine disulphide at 0.28-nm resolution. Eur. J .
Biochem. 1993, 213, 67-75.
Ack n ow led gm en t. We are grateful to the Wellcome
Trust (K.T.D., A.H.F., J .A.M.) and to the Medical
Research Council (J .G.) for support and for financial
support from the UNDP/World Bank/WHO Special
Program for Research and Training in Tropical Diseases
(C.C., S.L.C., K.T.D., R.J ., P.J .R., V.Y.). We are also
grateful to Dr. Max Wilhelm (Ciba-Geigy, Basle) for gifts
of several phenothiazines.
(24) Lantwin, C. B.; Schlichting, I.; Kabsch, W.; Pai, E. F.; Krauth-
Siegel, R. L. The structure of Trypanosoma cruzi trypanothione
reductase in the oxidized and NADPH reduced state. Proteins
1994, 18, 161-173.
Su p p or tin g In for m a tion Ava ila ble: Figures 3-5 (4
pages). Ordering information is given on any current mast-
head page.
(25) Zhang, Y.; Bond, C. S.; Bailey, S.; Cunningham, M. L.; Fairlamb,
A. H.; Hunter, W. N. The crystal structure of trypanothione
reductase from the human pathogen Trypanosoma cruzi at 2.3
A resolution. Protein Sci. 1996, 5, 52-61.
(26) J acoby, E. M.; Schlichting, I.; Lantwin, C. B.; Kabsch, W.;
Krauth-Siegel, R. L. Crystal structure of the Trypanosoma cruzi
trypanothione reductase.mepacrine complex. Proteins 1996, 24,
73-80.
(27) El-Waer, A. F.; Douglas, K. T.; Smith, K.; Fairlamb, A. H.
Synthesis of N-benzyloxycarbonyl-L-cysteinylglycine 3-dimethyl-
aminopropylamide disulfide: A cheap and convenient new assay
for trypanothione reductase. Anal. Biochem. 1991, 198, 212-
216.
(28) Cornish-Bowden, A. Principles of Enzyme Kinetics; Butter-
worth: London, 1976; p 59.
(29) Fromm, H. J . Use of competitive inhibitors to study substrate
binding order. In Contemporary Enzyme Kinetics and Mecha-
nism; Purich, D. L., Ed.; Academic Press: New York, 1983; p
243.
(30) Sullivan, F. X.; Sobolov, S. B.; Bradley, M.; Walsh, C. T.
Mutational Analysis of Parasite Trypanothione Reductase: Ac-
quisition of Glutathione Reductase Activity in a Triple Mutant.
Biochemistry 1991, 30, 2761-2767.
Refer en ces
(1) Pepin, J .; Milord, F. The treatment of human African trypano-
somiasis. Adv. Parasitol. 1994, 33, 1-47.
(2) Davidson, R. N.; Croft, S. L. Recent advances in the treatment
of visceral leishmaniasis. Trans. R. Soc. Trop. Med. Hyg. 1993,
87, 130-131.
(3) Fairlamb, A. H.; Blackburn, P.; Ulrich, P.; Chait, B. T.; Cerami,
A. Trypanothione: a novel bis(glutathionyl)spermidine cofactor
for glutathione reductase in trypanosomatids. Science 1985, 227,
1485-1487.
(4) Fairlamb, A. H.; Cerami, A. Metabolism and functions of
trypanothione in the kinetoplastida. Annu. Rev. Microbiol. 1992,
46, 695-729.
(5) Krauth-Siegel, R. L.; Enders, B.; Henderson, G. B.; Fairlamb,
A. H.; Schirmer, R. H. Trypanothione reductase from Trypano-
soma cruzi: purification and characterization of the crystalline
enzyme. Eur. J . Biochem. 1987, 164, 123-128.
(6) Shames, S. L.; Fairlamb, A. H.; Cerami, A.; Walsh, C. T.
Purification and characterization of trypanothione reductase
from Crithidia fasciculata, a new member of the family of
disulfide-containing flavoprotein reductases. Biochemistry 1986,
25, 3519-3526.

156 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 2
Chan et al.
(31) Leichus, B. N.; Bradley, M.; Nadeau, K.; Walsh, C. T.; Blanchard,
J . S. Kinetic isotope effect analysis of the reaction catalyzed by
Trypanosoma congolense trypanothione reductase. Biochemistry
1992, 31, 6414-6420.
(32) Borges, A.; Cunningham, M. L.; Tovar, J .; Fairlamb, A. H. Site-
directed mutagenesis of the redox-active cysteines of Trypano-
soma cruzi trypanothione reductase. Eur. J . Biochem. 1995, 228,
745-752.
(33) El-Waer, A. F.; Smith, K.; McKie, J . H.; Benson, T. J .; Fairlamb,
A. H.; Douglas, K. T. The glutamyl binding site of trypanothione
reductase from Crithidia fasciculata: enzyme kinetic properties
of R-glutamyl-modified substrate analogues. Biochim. Biophys.
Acta 1993, 1203, 93-98.
(34) Meinnel, T.; Mechulam, Y.; Fayat, G. Fast purification of a
functional elongator tRNAmet expressed from a synthetic gene
in vivo. Nucl. Acids Res. 1988, 16(16), 8095-8096.
(35) Krohne-Ehrich, G.; Schirmer, H.; Untucht-Grau, R. Glutathione
reductase from human erythrocytes. Eur. J . Biochem. 1977, 80,
65-71.
(36) Worthington, D. J .; Rosemeyer, M. A. Human glutathione
reductase: Purification of the crystalline enzyme from erythro-
cytes. Eur. J . Biochem. 1974, 48, 167-177.
(37) Karplus, P. A.; Schulz, G. E. Refined structure of glutathione
reductase at 1.54 Å resolution. J . Mol. Biol. 1987, 195, 701-
729.
(38) Baltzly, R.; Harfenist, M.; Webb, F. J . Some phenothiazine
derivatives. The course of the Friedel-Crafts reaction. J . Am.
Chem. Soc. 1946, 68, 2673-2678.
(39) Michels, J . G.; Amstutz, E. D. Studies in the sulfone series. V.
2,8-Diaminophenithiazine-5-oxide. J . Am. Chem. Soc. 1950, 72,
888-892.
(40) Hyperchem 4.0, Hypercube, Inc., 1995 (UnPub).
(41) Hirumi, H.; Hirumi, K. Continuous cultivation of Trypanosoma
brucei bloodstream forms in a medium containing a low concen-
tration of serum protein without feeder cell layers. J . Antimicrob.
Chemother. 1989, 75, 985-989.
(42) Neal, R. A.; Croft, S. L. An in vitro system for determining the
activity of compounds against the intracellular amastigote form
of Leishmania donovani. J . Antimicrob. Chemother. 1984, (14):
463-475.
J M960814J