Analysis of stereoelectronic properties, mechanism of action and pharmacophore of synthetic indolo[2,1-b]quinazoline-6,12-dione derivatives in relation to antileishmanial activity using quantum chemical, cyclic voltammetry and 3-D-QSAR CATALYST procedures

Bioorganic & MedicinalChemistry 10 (2002) 1979–1989

Analysis of Stereoelectronic Properties, Mechanism of Action and

Pharmacophore of Synthetic Indolo[2,1-b]quinazoline-6,12-dione

Derivatives in Relation to Antileishmanial Activity Using

Quantum Chemical, Cyclic Voltammetry and 3-D-QSAR

CATALYST Procedures

Apurba K. Bhattacharjee,* David J. Skanchy, Barton Jennings,^yThomas H. Hudson,

James J. Brendle and Karl A. Werbovetz^{

DivisionofExperimentalTherapeutics, WalterReedArmyInstituteofResearch, 503RobertGrantAvenue, SilverSpring, MD20910, USA

Received 22 May 2001; accepted 12 December 2001

Abstract—Severalindool[2,1- b]quinazoline-6,12-dione (tryptanthrin) derivatives exhibited remarkable activity at concentrations

below 100 ng/mL when tested against in vitro Leishmania donovani amastigotes. The in vitro toxicity studies indicate that the

compounds are fairly well tolerated in both macrophage and neuronal lines. An analysis based on qualitative and quantitative

structure–activity relationship studies between in vitro antileishmanial activity and molecular electronic structure of 27 analogues of

indolo[2,1-b]quinazoline-6,12-dione is presented here by using a combination of semi-empirical AM1 quantum chemical, cyclic

voltammetry and a pharmacophore generation (CATALYST) methods. A modest to good correlation is observed between activity

and a few calculated molecular properties such as molecular density, octanol–water partition coeﬃcient, molecular orbital energies,

and redox potentials. Electron transfer seems to be a plausible path in the mechanism of action of the compounds. A pharmaco-

phore generated by using the 3-D QSAR of CATALYST produced a fairly accurate predictive model of antileishmanial activity of

the tryptanthrins. The validity of the pharmacophore model extends to structurally diﬀerent class of compounds that could open

new frontiers for study. The carbonylgroup of the ﬁve- and six-membered rings in the indolo[2,1- b]quinazoline-6,12-dione skeleton and

Introduction

cases of leishmaniasis occur each year, which in the

1

Diseases caused by protozoalparasites have an over-

whelming impact on public health throughout the

world, particularly in the tropics and subtropics. The

leishmaniases represent a spectrum of disease resulting

from diﬀerent species belonging to the genus Leishma-

nia, a protozoalparasite transmitted by the bite of the

phlebotomine sandﬂy. Clinical manifestations of the

infection range from cutaneous and mucocutaneous to

visceral leishmaniasis. An estimated 1.5–2 million new

visceralmanifestation is often fatalif untreated.

An

additional 350 million people in 88 countries around the

world are threatened by leishmaniasis, and visceral

leishmaniasis alone has claimed an estimated 100,000

lives in Sudan in recent outbreaks.²Nonavailability of

satisfactory chemotherapeutic agents and failure to

develop an eﬀective vaccine are considered to be two

stumbling blocks in the combat of this disease.³

The current chemotherapy of leishmaniasis relies upon

the use of pentavalent antimony compounds that

require parenteraladministration of high doses and a

lengthy course of treatment resulting in marked increase

of serious side eﬀects and decreasing eﬃcacy. Two

pentavalent antimonial drugs, sodium stibogluconate

(Pentostam) and meglumine antimonate (Glucantime) are

the current treatments of choice for leishmaniases and

have been the choice for the past 50 years. Such heavy

*Corresponding author. Tel.: +1-301-319-9043; fax: +1-301-319-9449;

e-mail: apurba.bhattacharjee@na.amedd.army.mil (A.K. Bhattacharjee);

werbovetz@dendrite.pharmacy.ohio-state.edu (K.A. Werbovetz)

^yPresent address: Millennium Pharmaceuticals, Inc., 38-2 Sidney

Street, Cambridge, MA 02139, USA.

^{Present address: Division of MedicinalChemistry and Pharmacog-

nosy, College of Pharmacy, The Ohio State University, 500 West 12th

Ave., Columbus, OH 43210-1291, USA.

PII: S0968-0896(02)00013-5

1980

A. K. Bhattacharjee et al. / Bioorg. Med. Chem. 10 (2002) 1979–1989

metalpharmacool gy is found to have side eﬀects such as

nausea, diarrhea, convulsions, and even cardiotoxicity.⁴

Antimony-resistant strains have also been reported.⁵Thus,

there is a clear need for development of less toxic drugs

that are eﬀective against all forms of leishmaniases.

compounds showed high eﬃcacy against Leishmania.

The present study deals with 27 such analogues (Chart

1) that were tested for in vitro antileishmanial activity

against L. donovani.

However, the mode of action of these compounds is yet

unknown, and very little molecular level information is

found in literature to supplement a rational selection of

these compounds. Since the ability of drugs to interact

with the recognition sites in receptors results from a

Much eﬀort and attention have therefore been focused

on the discovery and development of new and less toxic

chemotherapeutic agents in recent years.^3,6Indolo[2,1-b]-

quinazoline-6,12-dione (tryptanthrin) is a compound

with a long history⁷and is well documented to possess

antibacterialactivity against a variety of pathogenic

bacteria, particularly the causative agent of tubercu-

losis.^8,9Recently, indolo[2,1-b]quinazoline-6,12-dione

and its substituted derivatives were revisited in a

screening campaign conducted by the Walter Reed

Army Institute of Research, Silver Spring, MD (USA).

The compounds displayed remarkable in vitro anti-

leishmanial activity against Leishmania donovani.

Indolo[2,1-b]quinazoline-6,12-dione and derivatives can

be easily synthesized by various methods that are dis-

cussed in a later section. The parent compound can also

be produced by Candida lipolytica when grown in media

containing an excess of tryptophan, hence the name

tryptanthrin. To further improve the in vitro activity, a

series of additionalazatryptanthrin anaolgues were

designed and tested incorporating a pyridine moiety for

ring A with one or two nitrogen atoms. Many of these

combination of steric and electronic properties,^10À12

a

study of the molecular electronic structure of the

indolo[2,1-b]quinazoline-6,12-dione derivatives should

provide better insights of the mechanism of action of

the compounds and aid the design of more eﬃcient

analogues. Moreover, a three-dimensional pharmaco-

phore modelgeneration woudl be usefulto identify the

structuralrequirements for antielishmanialactivity in

this series and also be useful for 3-D queries to search

databases of proprietary and/or commercially available

compounds. Thus, in continuation of our eﬀorts and

interest in the design of antileishmanial chemothera-

peutic agents,^6,13,14we present here a detailed analysis

of stereoelectronic properties in relation to structure–

activity relationships and mechanism of action of the

above 27 tested tryptanthrins using a combination of

15

semi-empiricalAM1 quantum chemica,l

cyclic vol-

tammetry,¹⁶and 3-D-QSAR CATALYST¹⁷methods.

Chart 1.

Compd

1-

2-

3-

4-

X

1⁰

2⁰

3⁰

4⁰

1

2

3

4

5

6

7

8

9

CH

–N¼

CH

–N¼

CH

–N¼

CH

C¼O

CH

C–ClCH

CH

CCH(OCH₃)₂

C–I

CH

–N¼

C–OCF₃

C–ClCH

CCH₂OCH₃

C-ClCH

CH

CCO₂CH₂CH₃

CCHOCH₃–

CH(CH₃)₂

C–ClCH

C–F

CH

CS(CH₂)₂OH

CH

–N¼

CH

–N¼

CH

10

11

12

13

14

15

16

17

18^a

19

20

21

22

23

24

25

CH

–N¼

CCH₃

C-F

CH

–N¼

CH

C¼O

CH

–N¼

C¼O

CH

–N¼

CH

C¼O

C–Br

C–ClCH

C–F

C–ClCH

C–F

CH

C–N<(CH₂)₄>N–CH₃

CH

C–NCH₃(CH₂)₂OH

CH

C¼O

CH

C-F

CH

C¼O

CH

C¼O

CH

C-Cl

CH

C¼O

CH

C–N–Cyclohexyl

C¼O

C–ClCH

C–F

CH

C–I

C–F

C–CH₃

CH

C¼O

CH

C¼O

C-OCH₃

CH

C-OCH₃

C¼O

C=C(CN)₂

C¼C–

Phenyl

C¼O

C–F

26

27

CH

C–ClBr

CH

C–F

CH

C–N<(CH₂)₄>N–CH₃

CH

C–F

CH

Br

C

¼O

^aIn this compound, the aromatic N of ring B has been replaced with a N–H group.

A. K. Bhattacharjee et al. / Bioorg. Med. Chem. 10 (2002) 1979–1989

1981

Results and Discussion

axenic amastigotes of the compounds are presented in

Table 1.

Antileishmanial activity and toxicity to in vitro

mammalian cell lines

Inspection of Table 1 indicates an increasing trend of

activity with increasing molecular density and decreas-

ing octanol/water partition coeﬃcients (log P) for a

large number of the tryptanthrin analogues but no

quantitative linear relationship could be established

between the two properties and activity. However, the

observed trend indicates that density and lipophilicity of

the molecule may be contributing for potent activity of

the compounds. More dense molecules being more

potent may imply less susceptibility for metabolic

degradation and longer survival time for transport into

the parasite infected cells facilitating a better parasite

killing in the macrophages. But there remains a concern

because this property could impart toxicity to the potent

analogues as well. The observed decreasing trend of

calculated log P with increasing activity may imply a

role of lipophilicity of the compound for potent activity

since the ability of a compound to cross biological

membranes is frequently related to its octanol/water

partition coeﬃcient. The octanol/water partition coeﬃ-

cients (log P) for the analogues were calculated using

the Villars’s scale as implemented in SPARTAN.¹⁹

Twenty-seven tryptanthrin analogues were screened in

vitro against L. donovani axenic amastigotes as descri-

bed previously.^6,12,14These compounds displayed

remarkable antileishmanial activity, with 13 of the

agents studied displaying IC₅₀values below 100 ng/mL.

Compounds 1 and 2 were the most active against

Leishmania in this study, with IC₅₀values of 16 ng/mL

(Table 1). As a means of comparison, the clinically useful

antileishmanial agent amphotericin B possesses an IC₅₀

value in this assay of 416 ng/mL.¹⁴Eighteen of these

compounds were also tested for toxicity against murine

J774 macrophages and rat neuronal NG-108–15 cells

(Table 2). In almost all cases, the tryptanthrins were

considerably more toxic to the parasites than to the mam-

malian cell lines, and were frequently an order of magnitude

more active against L. donovani. The best selectivity was

observed with compound 6, which was 69-fold more toxic

to the parasites than to both mammalian cell lines.

Analysis of stereoelectronic properties

Molecular level information obtained from semi-

empirical (AM1) quantum chemical calculations on the

optimized electronic structure of the compounds were

investigated to explore quantitative as well as qualita-

tive relationships between chemical structure and bio-

logical activity in the light of QSAR analysis developed

by Hansch.¹⁸The fundamentalhypothesis of QSAR is

that biological properties are functions of molecular

structure. It is reasonable to assume that certain con-

formations of a molecule can only bind to the speciﬁc

site of the receptor. The compounds that have both the

necessary functionalgroups and right conformation wil

only interact with the target. Thus, molecules with

similar structures and electronic properties are likely to

have similar biological activity. Accordingly, molecular

properties such as (a) the structuralparameters, (b)

molecular density, (c) log P, (d) molecular orbital ener-

gies, and (e) electrostatic potentials were calculated and

the results examined.

The correlation between activity and orbital energies

particularly the lowest unoccupied molecular orbital

(LUMO) energy is quite remarkable. Inspection of both

Table 1 and Figure 1 indicates that more negative

LUMO energy favors potent activity in the tryptan-

Table 1. Experimentalantielsihmanialactivities of the compounds

with calculated molecular properties and redox potential values

Compd

IC₅₀

(ng/mL)

Density Log P LUMO Redox 1 Redox 2

potential potential

(eV)

(mV)

1

2

3

4

5

6

7

8

16

24

28

32

36

41

57

60

85

88

0.948

1.289

1.079

1.014

0.928

1.027

1.028

0.999

0.909

1.02

1.047

1.009

1.162

0.997

0.99

1.14

na

À1.69

À1.78

À1.90

À1.64

À1.70

À1.75

À1.74

À1.81

À1.71

À1.92

À1.74

À1.77

À1.82

À1.50

À1.62

À1.51

À1.73

À5.90

À1.54

À1.47

À1.59

À1.44

À1.60

À2.21

À1.34

À1.66

À1.82

na

À551

À464

À622

À572

À546

À576

À558

À575

na

À586

À556

À481

À692

À656

À720

À593

À662

na

À1163

À1102

À1239

À1184

À1197

À1175

À1190

À1184

na

À1242

À1199

À1090

À1324

À1335

À1334

À1219

À1335

na

À1348

À1334

À1405

À1272

na

À1105

À1303

À1159

2.19

3.19

1.91

2.04

2.28

1.18

2.35

1.4

9

The fully optimized geometry the molecules indicates

that the rings A, B, C, and D in the indolo[2,1-b]quin-

azoline-6,12-dione skeleton are planar and super-

imposable (data not shown). From a structural point of

view, a carbonylgroup at the ﬁve-membered C ring is

an important requirement for potent activity because in

the absence of it the activity decreases dramatically

(Chart 1, Table 1). Bulkier groups at the 3⁰-position of

the D ring and a nitrogen atom at either the 2- or 4-

position of the A ring in the indolo[2,1-b]quinazoline-

6,12-dione skeleton seem to support potent activity.

Steric factors associated with the substitutions at the 3⁰-

position of ring D appear to outweigh the electronic

eﬀects for potent antileishmanial activity though the

role of electronic eﬀects at this position due to sub-

stituents such as F, Cl, Br, I, and OCF₃cannot be

totally discounted. Other calculated molecular proper-

ties and the in vitro IC₅₀values against L. donovani

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

3.36

1.35

na

98

150

160

210

220

450

550

590

620

1200

1300

1500

2100

17000

2.35

2.55

2.10

2.54

4.90

3.28

3.38

2.98

2.80

na

0.97

1.031

0.985

1.011

0.998

0.989

0.948

1.241

0.99

À696

À679

À750

À666

na

3.92

4.81

1.36

na

0.937

0.972

1.35

na

À744

À566

Redox 1=oxygen atom of ﬁve-membered ring. Redox 2=oxygen

atom of six-membered ring from cyclic voltammetry experiments.

1982

A. K. Bhattacharjee et al. / Bioorg. Med. Chem. 10 (2002) 1979–1989

thrins. Although the plot of LUMO energy versus IC₅₀

values (Fig. 1) does not indicate an excellent linear cor-

relation (r²=0.5258) the trend is quite apparent. Thus,

electron aﬃnity of the molecule could be an important

factor associated with potent activity. LUMO sites

plotted onto the molecular surface show one prominent

site by the carbonylgroup of the ﬁve-membered ring in

all the 27 compounds and the shape of the LUMO is

found to be large and extended by the carbonyl carbon

atom of the ﬁve-membered ring in the more potent

analogues. The carbonyl oxygen atom at the ﬁve-mem-

bered ring is criticalfor potent activity as absence of this

carbonyl oxygen atom dramatically reduces activity of

the tryptanthrins (Table 1). Therefore, electron transfer

from a receptor site to this carbonylgroup seems to be a

plausible path in the mechanism action of the com-

pounds, as shown in Chart 2. In order to test the

hypothesis, we have performed cyclic voltammetry

experiments and observed a good correlation between

activity and redox potentials as well as LUMO energy

and redox potentials, as described in the following sec-

tion (Table 1).

activity proﬁle of the compounds indicating that elec-

tron aﬃnity is likely to be crucial in the mechanism of

action for these compounds. Tryptanthrin analogues

containing carbonylgroups at the six- and ﬁve-mem-

bered rings displayed two reversible waves with cathodic

and anodic peaks separated by approximately 60 mV

indicating two one-electron transfers (Figs 4 and 5).

Plots of peak current versus (scan rate)^1/2were linear

further conﬁrming the reversible nature of the electron

transfers. Experimental and molecular modeling data

suggest that the more facile (less negative) redox wave

(À0.75 V in unsubstituted tryptanthrin) is due to elec-

tron transfer at the carbonyloxygen on the ﬁve-mem-

bered ring while the more negative wave is due to

electron transfer at the carbonyl oxygen on the six-

membered ring (À1.4 V in unsubstituted tryptanthrin).

This is supported by the observed electrochemical

behavior of the tryptanthrin analogues in which the

carbonylon the ﬁve-membered ring has been repalced

by a benzylic substituent. This compound displays a

single irreversible redox wave below À1.1 V (Fig. 5).

Thus, the signiﬁcance of the carbonyloxygen at the ﬁve-

membered ring is clearly indicated, its absence dramati-

cally reduces potent activity of the compounds. Sub-

stitution at the 3⁰-position with electron withdrawing

substituents has a larger eﬀect on the redox potential of

the less negative wave while substitution at the 4-posi-

tion has a larger eﬀect on the more negative redox

potentialwhich elnds further support to the assertion

that the ﬁrst wave (less negative) is due to electron

transfer at the carbonyloxygen on the ﬁve-membered

ring. A ﬂuorine substituent at the 3⁰-position causes a

shift of +94 mV in the ﬁrst redox potentialand

+70 mV shift in the second (more negative) redox

potentialrealtive to unsubstituted tryptanthrin. A sec-

ond ﬂuorine substituent added at the 3-position causes

an additional+63 mV shift in the ﬁrst potentialand an

additional+116 mV shift in the second potential(Fig.

4). The observation that the more facile reduction

occurs on the carbonyloxygen of the ﬁve-membered

ring while the more negative redox potential is asso-

ciated with electron transfer at the carbonyl on the six-

membered ring is consistent with the calculated LUMO

energies and supports the idea that a facile electron

transfer to the oxygen atom of the ﬁve-membered ring

could be a plausible path in the mechanism of action of

the compounds. As cyclic voltammetry experiments

allow determination of one-electron reduction poten-

tials and LUMO energy reﬂects electron aﬃnities

expressed in terms of one-electron reduction potentials,

these compounds are idealfor assessing the abilty of

simple theoretical methods to reproduce trends in

reduction potentials and, therefore, to anticipate activ-

ity.²¹Thus, as expected, LUMO energy values calcu-

lated for both carbonyl oxygen atoms correlate fairly

well with the measured redox potentials with more

Cyclic voltammetry

Since electrochemical methods can provide reliable

quantitative assessment of electron transfer ability of a

molecule,²⁰cyclic voltammetry experiments on each of

the analogues were performed to measure the redox

potentials associated with electron transfer to the oxy-

gen atoms in compounds. Inspection of Table 1 and the

plots (Figs 2 and 3) indicate that although a good linear

correlation between activity and redox potentials (Fig.

2, r²=0.5279; Fig. 3, r²=0.5686) is lacking, the trend

for less negative redox potentials of the carbonyl oxygen

atoms in the ﬁve- as well as in the six-membered rings

favoring potent activity is clearly indicated. The corre-

lation proﬁle is quite similar to the LUMO versus

Figure 1. Lowest unoccupied molecular orbital (LUMO) energy

against antileishmanial activity of the tyrptanthrins (r²=0.5258).

Chart 2.

A. K. Bhattacharjee et al. / Bioorg. Med. Chem. 10 (2002) 1979–1989

1983

negative LUMO energies corresponding to less negative

redox potentials. The plots of LUMO energy versus the

redox potentials of the oxygen atom at the ﬁve- and the

six-membered rings show linear regression of r²=0.92

and r²=0.85, respectively (Figs 6 and 7).

of a reliable 3-D QSAR model for antileishmanial

property of the compounds is the correlation of a char-

acteristic and reproducible biological activity to struc-

turalinformation of the respective compound. The

conformationalmodelof the compounds in the training

set of 27 tryptanthrins was used for hypothesis (phar-

macophore) generation that aims to identify the best

three-dimensionalarrangement of chemicalfunctions

explaining the activity variations among the training

set.²³The training set for the present study was pre-

pared with the in vitro activity data and the automatic

generation of hypotheses was performed within CATA-

LYST (HypoGen) using aromatic hydrophobic sites,

hydrogen bond acceptor, hydrogen bond acceptor

(lipid) and ring aromatic sites as the functional features

to describe the antileishmanial activity of the com-

pounds. Calculated stereoelectronic properties of the

compounds provided an idea about the pharmacophoric

modelof the training set and aided in the seelction of

the functionalfeatures in the generation of the hypoth-

esis. It was found that only two of the four features,

hydrogen bond acceptor and hydrogen bond acceptor

(lipid) are embodied in the statistically most signiﬁcant

hypothesis out of the 10 generated hypotheses

(cost=114.0, ﬁxed cost=106.3, RMS=0.71). The RMS

values of the other automatically generated hypotheses

range between 0.7 and 0.8. The dumping score for the

3-D-QSAR using CATALYST

The AM1 optimized geometry of 27 indolo[2,1-b]quin-

azoline-6,12-dione analogue was imported from the

SPARTAN directory into the CATALYST stockroom

to create a training set of antileishmanial tryptanthrins

and conformations were generated using the generalized

CHARMm force ﬁeld as implemented in the program.

The CATALYST force ﬁeld has been documented to do

better than the semi-empiricalquantum chemicalmeth-

ods in predicting the dihedral angles and calculation of

conformations of molecules.²²Molecular ﬂexibility was

taken into account by considering each compound as a

collection of conformers representing a diﬀerent area of

conformational space accessible to the molecule within

a given energy range. The ‘‘best quality searching pro-

cedure’’¹⁷was adopted to select representative con-

formers over a 0–25 kcal/mol range above the computed

global minimum. Since this search procedure in CAT-

ALYST can generate up to 250 conformers for each

molecule within 25 kcal/mol conformational energy

space above the lowest conformation it can identify the

best three-dimensionalarrangement of chemicalfunc-

tions explaining the activity variations among the train-

ing set. The AM1 calculated bond lengths, valence

angles and dihedral angles are found to be in close

agreement with molecular force ﬁeld of the CATA-

LYST (data not shown). A prerequisite for development

Table 2. Toxicity proﬁle of the tryptanthrins

Compd

IC₅₀

Mouse

macrophage

Neuronal

1

2

3

4

5

16

24

28

32

850

—

1000

200

—

200

—

1500

150

—

6

7

8

9

36

41

57

60

85

88

98

2500

—

2500

—

Figure 2. Redox potentialof the carbonyloxygen in the ﬁve-mem-

bered ring versus antileishmanial activity of the tryptanthrins

(r²=0.5279).

1300

200

2500

200

1550

250

850

—

1500

912

—

850

—

400

1500

—

925

700

2500

500

120

350

120

—

1000

2500

560

—

120

170

—

625

2500

—

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

150

160

210

220

450

550

590

620

1200

1300

1500

2100

17000

—

Figure 3. Redox potentialof the carbonyloxygen in the six-membered

ring versus antileishmanial activity of the tryptanthrins (r²=0.5686).

All values are given in ng/mL.

1984

A. K. Bhattacharjee et al. / Bioorg. Med. Chem. 10 (2002) 1979–1989

null hypothesis was 123.4. The total costs of the 10

generated hypotheses varied over a range between 114.0

and 121.0, eight of them in a narrow range of 114.0 to

114.23. This cost factor, known as Conﬁg, is a measure

of the magnitude of the hypothesis space for a given

training set of compounds. If the value exceeds 17, there

are more degrees of freedom in the training set that

CATALYST algorithm can properly handle and the

hypothesis result is likely to be poor.^24,25The Conﬁg

values of the 10 generated hypotheses for the present

training set were 14.78. One of the most potent analo-

gues of the training set such as, 2 maps very well with

the statistically most signiﬁcant hypothesis (Fig. 8)

whereas, the least potent analogue of the series such as,

27 does not map adequately with the hypothesis (Fig.

9). The criticalhydrogen bond acceptor proﬁel by the

carbonyloxygen at the ﬁve-membered ring in the

hypothesis appears to be consistent with the above ste-

reoelectronic analyses where it is clearly shown that

absence of the ﬁve-membered carbonyloxygen atom

will dramatically reduce potent activity of the tryptan-

thrins. Table 3 summarizes the estimated as well as the

measured IC₅₀values for the training set along with the

error values. The error column in Table 3 represents the

ratio of the estimated activity to measured activity, or

its negative inverse if the ratio is less than one. Plot of

the experimental activity values versus the calculated

activity of the training set gave a fairly good correlation

(=0.89) between the experimentalactivity (true activity)

and estimated activity (Fig. 10). Although the plot

shows two concentrated areas of clustered IC₅₀values

indicating some redundancy in the structuraland activ-

ity information to CATALYST we attempted to test the

validity of the hypothesis over the entire training set of

27 compounds. However, removalof some of the above

redundancy of structure and activity information from

the training set did not appear to produce any better

correlation though the plot looked more spread-out

(data not shown).

In order to cross-validate the observed correlation we

have prepared a test training set of 14 other tryptan-

thrins (Chart 3) that were tested for antileishmanial

activity in the identicalassay by screening in vitro

against L. donovani like the original training set. More-

over, this test set of the tryptanthrins has never seen

before the CATALYST (HypoGen) hypothesis model

that we have generated with the originaltraining set. A

regression analysis was performed with the test set

compounds and the best hypothesis generated from the

Figure 4. Superimposed cyclic voltammograms of tryptanthrin (22,

solid), 3⁰-ﬂouro tryptanthrin (15, dashed), 3⁰,4-ﬂouro tryptanthrin (17,

dotted) showing the eﬀect of electron withdrawing substituents on the

redox potentials. All scans are shown at a scan rate of 50 mV/s.

Figure 6. LUMO energy against redox potentialof the carbonyloxy-

gen in the ﬁve-membered ring with linear regression line (r²=0.9238).

Figure 5. Superimposed cyclic voltammograms of compound 25

(solid) and 3⁰,4-ﬂouro tryptanthrin (17, dotted) showing the eﬀect of

replacing the ﬁve-membered ring carbonyl moiety with a benzylic

substituent. Both scans shown are at a scan rate of 50 mV/s.

Figure 7. LUMO energy against redox potentialof the carbonyloxy-

gen in the six-membered ring with linear regression line (r²=0.8584).

A. K. Bhattacharjee et al. / Bioorg. Med. Chem. 10 (2002) 1979–1989

1985

original training set and a correlation close to 0.8 is

observed between true activity and the estimated activ-

ity values. The estimated as well as the measured IC₅₀

values for the test set tryptanthrins along with the error

values are also presented in Table 3. The actual activity

values are reproduced reasonably well within the limits

of uncertainty (Table 3), thus indicating a good pre-

dictive power of the originalhypothesis or pharmaco-

phore. The more potent analogue in the test set of

tryptanthrins such as 1⁰maps well with the original

hypothesis whereas, the less potent analogues such as 6⁰

and 14⁰do not map adequately with the hypothesis.

catDB utility program as implemented in the software.

The catDB format allows a molecule to be represented

by a limited set of conformations thereby permitting

conformational ﬂexibility to be included during the

database search. However, the search resulted in the

identiﬁcation of more than 300 compounds, and there-

fore required further reﬁnement of our hypothesis in

order to have a shorter list of potential lead compounds.

The reﬁnement was done by incorporating a few addi-

tionalfeatures that were observed from the resutls of

quantum chemical calculations such as a critical vector

direction or electric ﬁeld direction for the hydrogen

bond acceptor and the presence of hydrophobic aro-

matic rings from the electrostatic potential proﬁles. A

database query was thus formulated by superimposing

the in vitro hypothesis and a conformer of 1, one of the

two most potent analogues. Three positional constraints

for the sites of the functionalfeatures were incorporated

and the molecule was transformed into a 3-D shape

with all the functional features that are believed to be

associated with potent activity. The ﬁnalquery now

contains a 3-D shape with three position constraints

representing two hydrogen bond acceptors (lipid), a

hydrogen bond acceptor with a direction vector, and a

ring aromatic (Fig. 11). Using this query as the template

for search, the in-house CIS database led to the dis-

covery of 43 compounds. Interestingly, 16 compounds

from the list were found to be positively active

(IC₅₀<2000 ng/mL) as antileishmanials in our parasite

drug susceptibility assay. We have identiﬁed two com-

Using the generated hypothesis, we have performed a

database search for potentialnew antielishmanialcan-

didates from our in-house ChemicalInformation Sys-

tem²⁶database of over 240,000 compounds. The

ChemicalInformation System database was buitl as a

multi-conformer database in CATALYST using the

Table 3. Estimated and experimentally determined activity values of

the training set and the test set of the compounds

Training

set

(compd)

IC₅₀

(expt)

IC₅₀Error^aTest set IC₅₀IC₅₀Error^a

(est)

(compd) (expt) (est)

1

2

3

4

5

6

7

8

16

24

28

32

36

41

57

60

85

88

98

150

160

210

220

450

550

590

620

1200

30

41

40

53

46

41

1.9

2.6

1.7

1.9

1.4

1.1

1⁰

2⁰

11

26

40

53

20

1.8

20 À1.3

3⁰

20 À2.0

Figure 8. Mapping of one of the two most potent members of the

training set, compound 2, onto the statistically most signiﬁcant

hypothesis.

4⁰

110

2.1

4.2

5.9

4.8

1.3

1.1

5⁰

64

270

870

840

850

900

6⁰

147

173

636

822

1012

41 À1.0

7⁰

42 1.0

34 À1.7

39 À1.5

8⁰

9

9⁰

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

10⁰

11⁰

12⁰

13⁰

14⁰

830 À1.2

390

4.6

1218 1110 À1.1

41 À2.1

1434

2146

3817

860 À1.7

840 À2.5

860 À4.5

240

2.4

2.1

2.8

2.1

300

450

440

180 À1.9

450 2.0

390 À1.1

400 À1.4

430 À1.4

430 À1.5

360 À3.3

1300 2500

1500

2100

1.9

370 À4.0

330 À6.4

170,000 5600 À2.9

IC₅₀all values are given in ng/mL.

^aValues in the error column represent the ratio of the estimated

activity to measured activity, or its negative inverse if the ratio is less

than one.

Figure 9. Mapping of the least potent member of the training set,

compound 27, onto the statistically most signiﬁcant hypothesis.

1986

A. K. Bhattacharjee et al. / Bioorg. Med. Chem. 10 (2002) 1979–1989

undertaken metabolic stability and Caco-2 permeability

studies and exploring the possibility of developing

pharmacophore models on the ADME properties of the

compounds from the experimentaldata.

Conclusion

The study organizes the molecular characteristics that

are both statistically and mechanistically signiﬁcant for

potent antileishmanial activity from a set of diverse

indolo[2,1-b]quinazoline-6,12-dione derivatives. Stereo-

electronic factors of the substituents at the 3⁰-position of

the D ring in indolo[2,1-b]quinazoline-6,12-dione skele-

ton appear to have signiﬁcant eﬀect on potent activity.

The carbonylgroups of the ﬁve- and six-membered

rings in the tryptanthrin moiety and electron transfer

ability from a receptor is likely to be a crucial step in the

mechanism of action in the compounds. The presence of

a ﬁve-membered carbonyl moiety in the molecule

appears to be a structuralrequirement for potent activ-

ity. Cyclic voltammetry experiments are clearly in

agreement with the quantitative as well as qualitative

structure activity analyses of the compounds. The che-

mically signiﬁcant molecular characteristics disposed in

a three dimensionalspace generated a hypothesis that is

found to be quite satisfactory in correlating true activity

and estimated activity of the tryptanthrins. The critical

hydrogen bond acceptor proﬁle by the carbonyl oxygen

at the ﬁve-membered ring in the generated pharmaco-

phore is consistent with the stereoelectronic and redox

potentialanaylses of structure–activity of the com-

pounds. The validity of the hypothesis extends to struc-

turally diﬀerent class of compounds opening new

frontiers for study. The compounds are found to be less

toxic to mammalian cell lines than to Leishmania in

vitro. Thus, tryptanthrins show remarkable promise for

further study as potentialantielishmanialcandidates.

Figure 10. Plot displaying the experimental IC₅₀values (ng/mL) ver-

sus IC₅₀values estimated by using the statistically most signiﬁcant

hypothesis derived from the training set of 27 tryptanthrins (correla-

tion coeﬃcient=0.8943).

pounds, 28 (7-oxo-7H-benzo/e/pyrimidine-4-carboxylic

acid) (IC₅₀=754 ng/mL) and 29 (1,4,5-trihydroxy

anthraquinone) (IC₅₀=1806 ng/mL) from the list that

have shown potentials for further study (Chart 4). We

are currently employing these two compounds as new

antileishmanial leads starting from structural modiﬁca-

tions to synthesis, design of new analogues and testing.

Although the tryptanthrins possess outstanding activity

against axenic L. donovani amastigotes, these agents have

yet to display activity in Leishmania infected cells,²⁷

a

modelwhich is a better predictor of in vivo antileishmanial

activity than the axenic amastigote assay. We are continu-

ing to explore new strategies to translate the potent intrin-

sic antileishmanial activity of the tryptanthrins into new

candidates for treating leishmaniasis. We have currently

Chart 3.

Compd

1-

2-

3-

4-

X

1⁰

2⁰

3⁰

4⁰

1⁰

2⁰

3⁰

4⁰

5⁰

6⁰

7⁰

CH

–N¼

CH

N¼

CH

–N¼

CH

–N¼

CH

C¼O

CH

C-ClC–I

CH

1,3-Dioxane-2-ylCH

C–(CF₃)₂–OH

CCH₂OCH₃

C–ClCH

CH

C–N<(CH₂)₄>N–CH₃

C—

N<(CH₂)₄CH₃>NH

C–N<(CH₂)₄>N–CH₃

CH

C–N<(CH₂)₄>N–

COOC(CH₃)₃

CH

C–F

CH

8⁰

CH

–NH₂

C–F

CH

C¼O

CH

C–F

C–I

CH

9⁰

10⁰

11⁰

12⁰

13⁰

14⁰

CH

CCH₃

CH

C–F

CH

–OCH₃

CH

C¼O

CH

C–I

CH

C–F

CH

C–N<(CH₂)₄>N–CH₃

CH

CCH₃

A. K. Bhattacharjee et al. / Bioorg. Med. Chem. 10 (2002) 1979–1989

1987

Experimental

Parasite drug susceptibility assay

Chemistry

Chemicals. Tissue culture media, supplements and

chemicals were obtained from the Sigma Chemical

Company. Stock solutions of compounds were prepared

in DMSO and stored at À30 ^ꢀC.

Indolo[2,1-b]quinazoline-6,12-dione and its derivatives

were obtained from PathoGenesis Corporation, Seattle,

WA, USA (now owned by Chiron, Emeryville, CA,

USA). In general, the compounds can be synthesized by

base-catalyzed condensation of substituted isatins and

substituted isatoic anhydrides through a convenient

one-step ﬂexible synthesis as previously reported.²⁸

Cell lines. L. donovani (WHO designation: MHOM/SD/

62/1S-CL2_D) axenic amastigotes were maintained in

modiﬁed RPMI medium as previously described.¹⁴

Indolo[2,1-b]quinazoline-6,12-dione (22). Generalmethod

of preparation: isatin (7.35 g) and isatoic anhydride

(8.15 g) were added to a well-stirred solution of

N-methylpiperidine (0.2 mL) and diisopropyl carbodii-

mide (5 mL) in dry pyridine (35 mL) at 60–65 ^ꢀC. A clear

solution was quickly obtained followed shortly by brisk

evolution of CO₂. When the evolution subsided the

temperature was slowly (during 10 min) increased to

75 ^ꢀC. During this operation tryptanthrin starts crystal-

lizing. To complete the reaction a ﬁnal heating period

(30 min) at 95–100 ^ꢀC is required. Finally the mixture

was cooled and the crystals collected, washed with

MeOH and dried, yield 11.53 g (93%), mp 267–268 ^ꢀC

(lit. mp 265–266 ^ꢀC). Without the drying agent the yield

drops to 68%. Recently two other methods were repor-

ted on the synthesis of tryptanthrins. The ﬁrst method

involves a cyclization of 3-(o-chlorophenyl)-2-methyl-

4(3H)-quinazoline and its meta isomer in liquid ammo-

nia to yield indolo[2,1-b]quinazoline-12(6H)-one which

was oxidized to indolo[2,1-b]quinazoline-6,12-dione.²⁹

The other reported method involves a short-step syn-

thesis as quinazoline alkaloids containing indole skele-

ton via intramolecular aza-Wittig reaction.³⁰Indolo[2,1-

b]quinazoline-6,12-dione is also a naturally occurring

compound was isolated from the Taiwanese medicinal

plant, Strobilanthes cusia and from Polygonum tinc-

torum and Isatis tinctorum.³¹It is produced by C. lipo-

lytica when grown in media containing an excess of

tryptophan.³²

Leishmania drug susceptibility assays. The growth sus-

ceptibility of L. donovani amastigote-like forms to the

compounds was measured in a three day assay using the

CellTiter 96 Aqueous Non-Radioactive Cell Prolifera-

tion Assay Kit (Promega) as previously described.¹⁴

A

SpectraMax Plus microplate reader (Molecular Devices,

Sunnyvale, CA, USA) was used to measure the absor-

bance in each well at 490 nm. IC values (the concentra-

tion of the compound that inhibited parasite growth by

50%) were determined with the software program Soft-

Max Pro (Molecular Devices) using the dose–response

equation y=((aÀd)/(1+(x/c)^b))+d, where x=the drug

concentration, y=absorbance at 490 nm, a=upper

asymptote, b=slope, c=IC and d=lower asymptote.

Toxicity studies using mouse macrophage and neuronal

cells

The toxicity in vitro study was performed against two

mammalian cell lines. A subclone (G8) of the murine

monocyte-like macrophage line J774 was obtained from

Dr. Jose Alunda; Departmento de Sanidad Animal,

Facultad de Veterinaria, Universidad Complutense,

Madrid, Spain. Murine cells were maintained in 75-cm²

tissue culture ﬂasks in Dulbecco’s modiﬁed Eagle med-

ium (GIBCO) supplemented with 10% fetal calf serum,

2 mM l-glutamine 50 mg/mL gentamicin under humidi-

ﬁed 5% CO₂/95% air at 37 ^ꢀC. A subclone (NG-108-15)

of a hybrid rat neuroblastoma/mouse glioma cell line

was the gift of Dr. Marshall Nirenberg; National Heart

Lung and Blood Institute, National Institute of Health,

Bethesda, MD, USA. Neuronal cells were maintained in

75-cm²tissue culture ﬂasks in Dulbecco’s modiﬁed

essentialmedium inculding hypoxanthine-aminopterin-

thymidine (HAT) supplement, 10% fetal calf serum and

50 mg/mL gentamicin under humidiﬁed 5% CO₂/95%

air at 37 ^ꢀC. Toxicity tests were performed in 96-well

tissue culture plates using the protein-binding dye Sul-

forhodamine B (SRB) as described.³³Test compounds

were serially diluted and added to empty wells of the 96-

well plate. Appropriate cells in their respective culture

Figure 11. Display of the shape based query for searching the in-house

CIS database. The essentialeelments are one hydrogen bond acceptor

on the nitrogen atom of ring A, two hydrogen bond acceptors in the

two oxygen atoms, and a hydrophobic function positioned in 3-D

space with respective location constraints for a candidate having high

antileishmanial activity.

Chart 4.

1988

A. K. Bhattacharjee et al. / Bioorg. Med. Chem. 10 (2002) 1979–1989

medium were immediately seeded into the wells.

Appropriate solvent blanks (no compound) were run in

each test. After 72 h under culture conditions, cells were

ﬁxed to the plate by layering 50% TCA (4 ^ꢀC) over the

growth media in each well to produce a ﬁnal TCA con-

centration of 10%. After incubating for 1 h at 4 ^ꢀC, cul-

tures were washed ﬁve times with tap water and air-dried.

Wells were stained for 30 min with 0.4% (w/v) SRB in

1% acetic acid and washed four times with 1% acetic

acid. Cultures were air-dried and bound dye was solu-

bilized with 10mM Tris base (pH 10.5) for 15min on a

gyratory shaker at room temperature. The opticaldensity

(OD) at wavelengths of 490–530 nm was then measured

with the aid of the SpectraMax Plus microplate reader.

search generating up to 250 conformers for each mol-

ecule within 25 kcal/mol conformational energy above

the global minimum energy conformation. The con-

formationalmodelof the training set thus used for

hypothesis (pharmacophore) generation should be bet-

ter equipped to identify the best three-dimensional

arrangement of chemicalfunctions expal ining the activ-

ity variations among the training set. A pharmacophore

search was carried out by setting the default parameters

used in the automatic generation of hypothesis in

CATALYST,¹⁷such as function weight 0.302, mapping

coeﬃcient 0, resolution 297 pm, and activity uncertainty

3. Hypotheses approximating the pharmacophore of the

tryptanthrins are described as a set of functionalfea-

tures such as aromatic hydrophobic, hydrogen bond

acceptor, hydrogen bond acceptor lipid, positively and

negatively ionizable sites distributed within a 3-D space.

These functionalfeatures are represented either by

points or vector geometricalobjects. The direction of

the vector representing a functionalfeature of an atom,

for example the features associated with hydrogen

bonding, is determined from the symmetry, the number

of localized lone pairs, the extended eﬀect of the lone

pairs from their electrostatic potential proﬁles, and the

environment surrounding the atom. There are two pri-

mary methods of generating a hypothesis within CAT-

ALYST. The ﬁrst method is to interactively build

hypotheses by one of several procedures available in the

‘view hypothesis’ workbench based upon the sig-

niﬁcance of the chemicalfunctions and fragments com-

prising the target molecule. The other method

automatically generates a hypothesis from a diverse set

of lead compounds that have activity data from the

same assay. We have used both the methods. The sta-

tisticalrelvance of various generated hypotheses is

assessed on the basis of the cost relative to the null

hypothesis and the correlation coeﬃcients. The hypo-

theses are then used to estimate the activities of the

training set. These activities are derived from the best

conformation generation mode of the conformers dis-

playing the smallest root-mean square (RMS) devia-

tions when projected onto the hypothesis.

Computational procedure

Computationalcaclualtions at the semi-empiricalAM1

level were performed using SPARTAN version 5.0¹⁹

running on a Silicon Graphics Octane workstation.

Conformationalsearch on each of the 27 anaol gues was

performed by multiple rotations of single bonds in the

compounds, thereby generating severalol w energy con-

formers with varying population densities. The most

abundant and the lowest energy conformers were iden-

tiﬁed. The geometry of these conformers was optimized,

and the electronic properties were calculated using the

optimized geometry. Geometry optimization and energy

calculations were performed on the compounds in the

gaseous phase using AM1 semi-empiricalquantum

chemical¹⁵method as implemented in SPARTAN.

Cyclic voltammetry

Cyclic voltammetry experiments were performed on a

CV-50W Voltammetric analyzer with a C2 cell stand

(Bioanalytical systems, West Lafayette IN, USA). A

glassy carbon working, a silver–silver chloride reference

electrode and a platinum auxiliary electrode were used

in a 5 mL glass cell. All samples were dissolved in dry

acetonitrile (Aldrich, St. Louis, MO, USA) with 0.1 M

tetrabutylammonium hexaﬂuorophosphate (Aldrich) as

the supporting electrolyte. Each tryptanthrin analogue

was prepared at a concentration of 1 mM and was

degassed with nitrogen for 5 min prior to analysis.

Samples were run at several scan rates ranging from 20

to 1000 mV/s.¹⁶Linear regression analysis of LUMO

energy against the redox potentialof the carbonyl

oxygen was done using TableCurve 2D (SPSS Inc.,

Chicago, IL, USA).

Acknowledgements

The authors wish to acknowledge Dr. William Baker

and Pathogenesis Corporation, Seattle, WA 98119,

USA for providing the compounds used in this study

under a cooperative research agreement. We thank Dr.

Shikha Varma of Accelrys, Burlington, MA 01803,

USA for reading the manuscript and providing valuable

suggestions. We also acknowledge COL John P. Scovill

for his eﬀorts in obtaining the compounds.

3-D-QSAR analyses using CATALYST

The software allows the use of structure and activity

data for a set of lead compounds to create a hypothesis

characterizing the activity of the lead set. The AM1

optimized geometry of each of the 27 tryptanthrins was

imported in CATALYST and was minimized to the

closest local minimum using molecular mechanics based

CHARMm force ﬁeld. Conformational models were

generated that emphasize representative coverage over a

0–25 kcal/mol range above the computed global mini-

mum. CATALYST algorithm allows a conformational

References and Notes

1. World Health Organization. http://www.who. int/emc/dis-

eases/leish/leisdis1.html, 2000.

2. Seaman, J.; Mercer, A.; Sondorp, E. Int. J. Epidemiol 1996,

25, 862.

3. Werbovetz, K. Curr. Med. Chem. 2000, 7, 835.

A. K. Bhattacharjee et al. / Bioorg. Med. Chem. 10 (2002) 1979–1989

1989

4. (a) Cook, G. C. J. Antimicrob. Chemother. 1993, 31, 327.

(b) Olliaro, P. L.; Bryceson, A. D. M. Parasitol. Today 1993,

9, 323.

5. Sacks, D.; Kenney, R.; Kreutzer, R.; Jaﬀfe, C.; Gupta, A.;

Sharma, M.; Sinha, S.; Neva, F.; Saran, R. Lancet 1995, 345,

959.

6. Pitzer, K. K.; Werbovetz, K. A.; Brendle, J. J.; Scovill, J. P.

J. Med. Chem. 1998, 41, 4885.

7. Bird, C. W. Tetrahedron 1963, 19, 901.

8. Mitscher, L. A.; Wong, W. C.; DeMeulenaere, T.; Sulko,

J.; Drake, S. Heterocycles 1981, 15, 1017.

9. Baker, W. R.; Mitscher, L. A. US Patent 5, 441, 955, 1995.

10. Bhattacharjee, A. K.; Karle, J. M. J. Med. Chem. 1996,

39, 4622.

11. Bhattacharjee, A. K.; Majumdar, D.; Guha, S. J. Chem.

Soc., Perkin Trans. 2 1992, 805.

20. Chen, H. Y.; Chen, Y.; Zhu, S. M.; Bian, N. S.; Shan, F.;

Li, Y. Talanta 1999, 48, 143.

21. Hehre, W. J.; Burke, L. D.; Shusterman, A. J.; Pietro, W.

J. Experiments in Computational Chemistry; Wavefunction:

Irvine, CA, 1993; Chapter 3.

22. Grigorov, M.; Weber, J.; Tronchet, M. J.; Jeﬀord, C. W.;

Milhous, W. K.; Maric, D. J. Chem. Inf. Comput. Sci. 1995,

35, 285.

23. Baringhaus, K.-H.; Matter, H.; Stengelin, S.; Kramer, W.

J. Lipid Res. 1999, 40, 2158.

24. Greenidge, P. A.; Weiser, J. Mini Rev. Med. Chem. 2001,

1, 79.

25. Kurogi, Y.; Guner, O. Curr. Med. Chem. 2001, 8, 1035.

¨

26. Chemical Information System; Division of Experimental

Therapeutics, Walter Reed Army Institute of Research: Silver

Spring, MD 20910, 1999.

12. Werbovetz, K. A.; Bhattacharjee, A. K.; Brendle, J. J.;

Scovill, J. P. Bioorg. Med. Chem. 2000, 8, 1741.

13. Bhattacharjee, A. K.; Intl, J. Quant. Chem. 1999, 75, 995.

14. Werbovetz, K. A.; Brendle, J. J.; Sackett, D. Mol. Bio-

chem. Parasitol. 1999, 98, 53.

15. Dewar, M. J. S.; Zoebisch, E. G.; Horsley, E. F.; Stewart,

J. J. P. J. Am. Chem. Soc. 1985, 107, 3902.

16. Smyth, M. R.; Vos, J. G. Analytical Voltammetry;

Elsevier: New York, 1992; Chapter 1.

17. CATALYST Version 4.5; Accelrys: San Diego, CA, 1999.

18. Hansch, C. A. Acc. Chem. Res. 1969, 2, 232.

19. SPARTAN SGI Version 5.1.2; Wavefunction Inc.: Irvine,

CA, 1998.

27. Brendle, J. unpublished data.

28. Mitscher, L. A.; Wong, W. C.; DeMeulenaere, T.; Sulko,

J.; Drake, S. Heterocycles 1981, 15, 1017.

29. Staskun, B.; Wolfe, J. F. S. Afr. J. Chem. 1992, 45, 5.

30. Eguchi, S.; Takeuchi, H.; Matsushita, Y. Heterocycles

1992, 33, 153.

31. Mitscher, L. A.; Drake, S.; Gollapudi, S. R.; Okwute,

S. K. J. Nat. Prod. 1987, 50, 1025.

32. Fiedler, E.; Fiedler, H. P.; Gerhard, A.; Keller-Schierlein,

W. Arch. Microbiol. 1976, 107, 249.

33. Skehan, P.; Storeng, R.; Scudiero, D.; Monks, A.;

McMahon, J.; Vistica, D.; Warren, J. T.; Bokesch, H.;

Kenney, S.; Boyd, M. R. J. Natl. Cancer Inst. 1990, 82, 1107.

Article Doi

Analysis of stereoelectronic properties, mechanism of action and pharmacophore of synthetic indolo[2,1-b]quinazoline-6,12-dione derivatives in relation to antileishmanial activity using quantum chemical, cyclic voltammetry and 3-D-QSAR CATALYST procedures

DOI: 10.1016/S0968-0896(02)00013-5

Source and publish data:

Authors:

Article abstract of DOI:10.1016/S0968-0896(02)00013-5

Full text of DOI:10.1016/S0968-0896(02)00013-5

Products guided by the article

R&D Labs maybe for 13220-57-0

Relevant to this article

Hot Product