ORIGINAL RESEARCH

Electronic descriptors for the antimalarial activity
of sulfonamides

Nélio H. Nicoleti1 • Augusto Batagin-Neto2
• Francisco C. Lavarda3

Received: 22 September 2015 / Accepted: 9 May 2016 / Published online: 20 May 2016

� Springer Science+Business Media New York 2016

Abstract As an interesting class of materials for designing

new antimalarial drugs, sulfonamides have shown potential

for several pharmacological applications. In this study,

multivariate data analyses were employed to correlate the

antimalarial activity reported for a group of sulfonamide

derivatives with electronic structure descriptors obtained

through quantum mechanical calculations. A simple classi-

fication rule based on a discriminant function was obtained,

which is able to correctly classify 94 % of the compounds as

active or non-active. The obtained function combines only

two electronic descriptors and provides valuable insights

into the design of new derivatives with improved anti-

malarial potency, as well as identifies possible active sites on

the structure of sulfonamides.

Keywords Sulfonamides � Antimalarial �
Electronic structure calculations �
Quantitative structure–activity relationship

Introduction

Malaria is an infectious disease caused by protozoa of the

genus Plasmodium and is transmitted to humans by the

Anopheles mosquito. Since this malady is mainly present in

tropical regions and primarily affects poor people in

developing countries, its occurrence is often associated

with socioeconomic problems (World Health Organization,

2011). In spite of the recent progress, there are not yet

vaccines available for clinical treatment against malaria

and the problem is even greater due to increasing parasite

resistance (Aguiar et al., 2012; Flannery et al., 2013;

Jensen et al., 2012).

In order to contain the epidemic, different technologies,

methods and drugs are currently being used. Such proce-

dures range from using nets treated with insecticides

(Eisele et al., 2011; Metropolis et al., 2014) to employing a

combination of drugs with distinct effects on the parasites

(Desgrouas et al., 2014; Guiguemde et al., 2014; Muta-

bingwa, 2005; Santelli et al., 2012). However, even with

these actions tests indicate a continuous increase in the

resistance of the parasites (Kümpornsin et al., 2014; Mita

et al., 2014; Perakslis, 2014; Winzeler and Manary, 2014),

evidencing the urgent need for new drugs.

In this context, we propose here a new study of a family

of sulfonamides previously investigated by Elslager et al.

(1984), and Agrawal and collaborators (Agrawal et al.,

2001a, b; Singh and Agrawal, 2008). Given the low cost of

production and especially the versatility of the synthesis of

these compounds, they can be an interesting option for

malaria treatment (Kumar Parai et al., 2008).

In general, the sulfonamide functional group has an

impressive effectiveness and assuredness history in medi-

cine. It usually presents interesting pharmacokinetic prop-

erties being well absorbed by several routes of

administration, presenting good penetration into tissues and

fluids, and being easily metabolized (Kahn, 2005; Spoo and

Riviere, 2001). In particular, in the last decade it has been

widely employed in the treatment of malaria (Barea et al.,

2011; Primas et al., 2012; Salahuddin et al., 2013), and

& Augusto Batagin-Neto

abatagin@itapeva.unesp.br

1 Federal Institute of Education, Science and Technology of

São Paulo, Piracicaba, SP 13414-155, Brazil

2 Campus Experimental de Itapeva, UNESP - Univ Estadual

Paulista, Itapeva, SP 18409-010, Brazil

3 Departamento de Fı́sica, Faculdade de Ciências, UNESP -

Univ Estadual Paulista, Bauru, SP 17033-360, Brazil

123

Med Chem Res (2016) 25:1630–1638

DOI 10.1007/s00044-016-1596-9

MEDICINAL
CHEMISTRY
RESEARCH

http://crossmark.crossref.org/dialog/?doi=10.1007/s00044-016-1596-9&domain=pdf
http://crossmark.crossref.org/dialog/?doi=10.1007/s00044-016-1596-9&domain=pdf


currently there are more than 100 products containing this

substance on the market (Smith and Jones, 2008).

In this report, we evaluate the possible relationships

between the electronic structure and antimalarial activity of

some molecules of the group 2,4-diamino-6-quinazoline

sulfonamides. A quantitative structure–activity relationship

(QSAR) study of a subset of these molecules was con-

ducted by Agrawal and collaborators (Agrawal et al.,

2001a, b; Singh and Agrawal, 2008) by using topological

parameters. In our study, we conducted an intensive

investigation of the electronic structure by employing 65

descriptors derived solely from quantum mechanical cal-

culations. We believe that the analysis of the electronic

structure can bring a clearer explanation of the mechanisms

associated with the antimalarial activity of these com-

pounds and also delineate essential features for the design

of new active compounds.

Statistical analyses were performed using simple and

multiple linear regressions (SLR and MLR) and other

multivariate methods such as principal component analysis

(PCA) and linear discriminant analysis (LDA) (Hair et al.,

1998). No significant regression equation could be obtained

via SLR and MLR studies. PCA data allow a clear sepa-

ration of the compounds into two subgroups of molecules,

according to the degree of activity. In addition, LDA pro-

vides a significant rule that allows the correct classification

of a high percentage of the compounds as active or inac-

tive. Finally, a simple rule for achieving new active com-

pounds is proposed. Based on the proposed rules, some

successful theoretical tests with substituents of known

electronic influence on the common structure were carried

out.

Materials and methods

A group of 16 sulfonamide derivatives with potential

antimalarial activity was evaluated (Agrawal et al., 2001a,

b; Singh and Agrawal, 2008). Figure 1 presents the basic

structure of 2,4-diamino-6-quinazoline sulfonamides,

which is common to all the studied compounds.

Table 1 presents the chemical structure of the sub-

stituents attached to the basic structure of sulfonamide

derivatives, as well as the experimental antimalarial

activity associated with each. The activity index that we

attempt to correlate with quantum descriptors of the com-

pounds is given by the difference in lifetime, in days, of

treated and untreated mice after infection (DMST) (Agra-

wal et al., 2001a, b; Singh and Agrawal, 2008).
Fig. 1 Basic structure of 2,4-diamino-6-quinazoline sulfonamides.

R1, R2, and X represent distinct substituents (Table 1)

Table 1 Description of the structure of the sulfonamide derivatives and the activity index measured in days

ID Substituents Activity index DMST

–NR1R2 X

1 N(C2H5)2 H 3.30

2 N(C2H5)2 Cl 2.30

3 N(CH2CH2CH3)2 H 0.30

4 N(CH2CH2OH)2 H 0.30

5 N(CH3)CH(CH3)2 H 0.70

6 N(CH3)CH2CH2N(C2H5)2 H 0.10

7 N(CH2)5 H 4.40

8 N(CH2)4 H 5.00

9 N[(CH2)2]2O H 4.70

10 N[(CH2)2]2S H 2.50

11 N[(CH2)2]2NCH3 H 1.00

12 N[(CH2)2]2NC(=O)OC2H5 H 0.20

13 4Cl–C6H4NH H 0.70

14 3Br–C6H4NH H 0.30

15 3Cl–C6H4NCH3 H 0.30

16 C6H5NCH3 H 0.50

Med Chem Res (2016) 25:1630–1638 1631

123



All the calculations, including geometry optimization,

were carried out via the semiempirical molecular orbital

method Austin Model 1 (AM1) (Dewar et al., 1985), which

is implemented in the MOPAC package (Stewart, 1990).

The choice of this approach was based on preliminary

conformational studies employing varied methods

(semiempiricals: MNDO, MNDO/d, PM3, and AM1; and

DFT with B3LYP functional and 6-31G* basis set) carried

out for the model molecule cyclohexanesulfonamide. AM1

was the semiempirical method that better reproduced the

structural features of the model compounds (Malešič et al.,

1997; Ojala et al., 2001), presenting an average deviation

of 4.9 %, just a little bit superior to the DFT approach

(3.7 %).

For QSAR studies, the geometry optimization was

considered complete when achieving gradient norms below

0.01 and no negative force constants were observed. The

calculations were performed in vacuo by employing a

restricted Hartree–Fock (RHF) approach.

A collection of 65 electronic indexes, most of them

related to the energy and electron density of the frontier

molecular orbitals, was employed as molecular descriptors

for each derivative. The descriptors are divided into two

groups: (1) global indexes, related to the whole molecule:

heat of formation (DHF), total electric dipole moment (DT),

total energy (ET), electronic energy (EE), nuclear energy

(EN), the energies of the highest occupied molecular orbital

(EHOMO), the level below (EHOMO-1), the lowest unoccu-

pied molecular orbital (ELUMO), the level above

(ELUMO?1), the difference in energy (ELUMO - EHOMO),

and the chemical hardness (EHOMO ? ELUMO)/2, and (2)

local indexes, related to the ith atom or to the bond of all

i and j chemically bonded atoms of the basic structure

(Fig. 1): atomic charge (CHARi), total electron population

(TEPi) and partial electron population in the HOMO and

LUMO (PEPi
H and PEPi

L, respectively), and bond orders

(BOi-j) (all originating from Mulliken population

analysis).

Different statistical methods were employed to evaluate

possible structure–activity correlations: simple and multi-

ple linear regressions (SLR and MLR), principal compo-

nent analysis (PCA), and linear discriminant analysis

(LDA). In particular, PCA and LDA have shown to be

interesting multivariate methods for pattern recognition

and classification, having been successfully employed in

QSAR studies of a variety of compounds (Autreto and

Lavarda, 2008; Batagin-Neto and Lavarda, 2014; Naranjo-

Montoya et al., 2014).

SLR and MLR studies were performed with the aid of a

proprietary software for statistical analyses. Combinations

of up to three electronic descriptors (independent variables)

were performed in MRL, involving the complete data set.

The quality of the correlations was evaluated, first, by the

Pearson correlation parameter (R) between experimental

and predicted values of the dependent variable. Distinct

functional forms of DMST indexes were considered in the

regressions: DMST, 1/DMST, and log(DMST) (as a

dependent variable). For each one of these cases, 45,825

combinations with up to three descriptors were performed.

Aiming to evaluate the similarities and differences

between the electronic structures of the compounds, the

PCA study was performed at two different levels: (1) full

PCA, involving the whole data set (16 cases and 65 elec-

tronic descriptors) and (2) reduced PCA, involving subsets

of descriptors (up to 5 independent variables). In reduced

PCA, 45,825 distinct combinations of descriptors were

considered. The calculations were automatically performed

through specifically developed software. The compounds

were divided into two subgroups based on their anti-

malarial potency for classification: (1) non-active com-

pounds: with DMST\ 2.0. (compounds 3–6 and 11–16)

and (2) active compounds: with DMST C 2.0. (compounds

1–2 and 7–10). The choice of the threshold value (2 days)

was based on the average value of the antimalarial potency

of the data set that was 1.66 days.

LDA calculations were performed with the aid of the

statistical package SPSS 11.0.1 (SPSS Inc., 2001). In order

to obtain a minimum set of electronic indexes for the

determination of the discriminant function (DF), the step-

wise method with Mahalanobis distance criterion (vali-

dated by F statistics) was employed. Additionally, a cross-

validation method (leave-one-out classification) was

employed in order to critically evaluate the robustness of

the obtained models. For this purpose, each molecule was

tested by using the classification rule derived from a subset

containing all the other molecules.

Results and discussion

Simple and multiple linear regressions

Despite the number of combinations evaluated, only low

correlation parameters could be observed (r\ 0.9) in the

linear regressions studies that are not good enough to

propose a significant prediction rule for the antimalarial

potency of the compounds. However, even with the

absence of a significant linear relationships, it is possible to

obtain some interesting clues from the more representative

results (0.7\ r\ 0.9) that can help in the subsequent

analyses. For example, it was noticed that the most relevant

equations were often associated with local molecular

indexes of ring B, what could indicate a potential reactive

region on the sulfonamide’s main structure.

Nevertheless, the absence of significant linear correla-

tions suggests that exploratory and classificatory methods

1632 Med Chem Res (2016) 25:1630–1638

123



could be more appropriate to propose a predictive rule for

antimalarial activity of the molecules.

Principal component analysis

Figure 2 shows a plot of the first and second principal

components (PC1 and PC2) coming from the full PCA

study.

As shown, in spite of the PC1 and PC2 components

determining around 70 % of the total variance, there is no

evident separation between active and inactive subgroups.

An interesting feature that deserves to be highlighted is the

distinction of compound 2 with respect to the other mole-

cules. As shown in Table 1, this derivative is the only one

that presents a chlorine atom at position X (X = Cl), which

provides a distinct electronic structure to this molecule,

evidenced in the PC2 scores.

Reduced PCA was performed in order to investigate

whether smaller sets of electronic descriptors could

account for the antimalarial activity of the derivatives.

Since it is based on few descriptors, the interpretation of

the results via this approach is more direct than in full

PCA.

Figure 3 shows a plot of PC1 and PC2, relative to the

reduced PCA that better discriminates the two subgroups of

molecules. The variables involved are the electronic energy

(EE), which represents the portion of the total energy due

only to the electrons of the molecule, and the fraction of the

lowest unoccupied molecular orbital (LUMO) that is

located on atom 9, i.e., the partial electronic population,

referred to as LUMO, on atom 9 (PEP9
L).

As can be seen through PC1 scores, it is possible to

separate the molecules into two groups associated with the

antimalarial potency of the compounds; only compound 5

is misclassified. As a matter of fact, this compound has

already been considered an outlier in the work of Agrawal

et al. (2001b), which is indeed reinforced by the results

presented in Fig. 3.

Despite the fact that the PCA study does not allow for

the proposition of a classification rule, it suggests that the

antimalarial activity of the sulfonamide derivatives can be

associated with the electronic descriptors EE and PEP9
L.

Since these variables present similar loadings for PC1

construction (*0.707), both are supposed to contribute

with similar relevance to data dispersion and then to the

clustering observed in Fig. 3.

Linear discriminant analysis

In LDA, the same criterion considered in the PCA studies

was employed to define active and non-active subgroups of

compounds. Equation 1 shows the discriminant function

(DF), which provides better distinction between these

subgroups. Similar to PCA, the descriptors involved are EE

and PEP9
L.

DF ¼ 3:07� 10�4 EEð Þ þ 143:814 PEPL
9

� �
� 11:984; ð1Þ

where EE should be given in electron Volts (eV).

The DF presented has a statistical significance of 98.9 %

(Wilk’s Lambda = 0.497, v2 = 9.091 with 2 degrees of

freedom, and p\ 0.011). Following the scores obtained

from Eq. 1, a cutoff parameter of 0.243 can be defined for

the classification of the compounds. This parameter defines

the central value between the centroids of active and non-

active groups and allows classifying the compounds

according to their DF scores. So, compounds with DF

scores higher than 0.243 can be classified as active mole-

cules, while derivatives with DF\ 0.243 can be identified
Fig. 3 Scores of PC1 and PC2 from the most significant reduced

PCA. The involved descriptors are EE and PEP9
L

Fig. 2 PC1 and PC2 scores from the full PCA study

Med Chem Res (2016) 25:1630–1638 1633

123



as non-active. By Eq. 1 and the cutoff parameter, using a

leave-one-out classification it was possible to classify

correctly around 94 % of the molecules, which indicates

the relevance of the proposed rule.

Figure 4 illustrates the dispersion of the DF scores for

(a) active, (b) non-active, and (c) the whole set of com-

pounds; the cutoff parameter suggested for compound

classification is also identified. The centroids of active and

inactive subsets are located at 1.215 and -0.729,

respectively.

As can be seen, similar to PCA, only compound 5 is

misclassified, reinforcing the hypothesis that this derivative

is in fact an outlier (Agrawal et al., 2001b). Given this

possibility, all the above-presented analyses were repeated,

excluding compound 5 from the data set, and the same

results were obtained.

Achieving high activity

Although the obtained DF is an interesting option for

testing the activity tendency of already synthesized mole-

cules, it would be more useful if it provided some clues

regarding plausible substitutions on R1 and R2 ligands

(Fig. 1), in order to obtain derivatives with improved bio-

logical properties.

Both, LDA and PCA, have indicated the relevance of the

descriptors EE and PEP9
L in the distinction of active and

non-active compounds. In particular, the DF (Eq. 1) sug-

gests that active compounds must present high EE and/or

PEP9
L values (since higher DF scores are associated with

active derivatives). Such descriptors are associated with the

molecular energy due to the electrons present in the sys-

tems (and their interactions), and the fraction of the

LUMO’s density located on atom 9 of the basic sulfon-

amide structure, respectively.

In order to identify the relevance of each one of these

descriptors in the classification of the compounds, we have

evaluated them individually. Figure 5 shows the relation-

ship of EE and PEP9
L with the DMST values. The solid

horizontal lines represent a guide to the eyes referring to

the limit between active and inactive compounds

(DMST C 2.0: active derivatives and DMST\ 2.0: non-

active derivatives). As can be seen, the descriptor EE plays

a key role in explaining the antimalarial potency of the

compounds, in such a way that active and non-active

molecules can be discriminated just by considering the

cutoff value EE
cutoff = -27,310 eV (dashed vertical line

presented in Fig. 5a). Note that molecule 5 is again the

only misclassified molecule (it presents an electronic

energy typical of an active molecule; however, it belongs

to the non-active group).

Fig. 5 Relationship between DMST parameter and the descriptors: a PEP9
L and b EE

Fig. 4 Discriminant function scores of active, inactive, and the whole

set of compounds. Dotted line indicates the cutoff parameter, 0.243

1634 Med Chem Res (2016) 25:1630–1638

123



In spite of subgroup separation observed in Fig. 5a, it is

important to stress that both the descriptors, PEP9
L and EE,

strongly contribute to construct the DF (Eq. 1), presenting

similar standardized coefficients (0.821 for EE and 0.706

for PEP9
L). In general, it is noticed that the information

provided by PEP9
L descriptor (Fig. 5b) improves the sta-

tistical significance of the subgroups distinction provided

by EE (Fig. 5a). It suggests that the conjunction of these

two parameters is relevant to distinguish between active

and non-active compounds, and the best classification rule

must be based on the DF and its cutoff value. Nevertheless,

in order to get more information regarding the physical

relevance of these parameters and its influence on the

biological potency of the compounds, it could be interest-

ing to evaluate them individually.

For the set of molecules considered in this study, as

shown in Fig. 6, the larger is the number of electrons in the

system, lower is the electronic energy of the compound. In

general terms, it occurs because ground-state (and stable)

molecules generally present occupied orbitals with nega-

tive energies (ei
occ). Since EE predominantly depends on a

summation of ei
occ values (subtracted by the twice counted

coulomb and exchange contributions), each electron added

to the system leads to a reduction in the EE descriptor.

Indeed, as can be seen, higher values of EE are typically

associated with compounds presenting a reduced number of

electrons in the valence shell. Given the fact that the set of

molecules studied in this work presents a common basic

structure (Fig. 1), a reduced number of electrons can be

achieved by choosing appropriate substituents on sites R1,

R2, and X.

This trend is in fact evidenced by comparing the

molecules 8 and 6, whose DMST are, respectively, the

largest (5.00) and the lowest (0.10) in the whole set. The

very active compound 8 contains 106 valence electrons and

N(CH2)4 as the substituent group, while the potentially

inactive compound 6 contains 132 electrons in the valence

shell and N(CH3)CH2CH2N(C2H5)2 as the substituent

group. Compound 8 does not just present the higher EE

value, but it also presents the smallest ligand. This fact can

suggest that the substituent may play an important role in

the antimalarial potency of the derivatives, in such a way

that small ligands can be linked to a more efficient anti-

malarial effect (probably due to sterical effects). In this

context, the electronic descriptor EE is highlighted as rel-

evant in the correlation studies, simply because it is the

electronic index that best reflects the difference of the

ligand’s volume among the compounds.

Figure 7 shows the relationship between DMST and the

volume of the substituents attached to each compound

(Volsub) (obtained via VEGA ZZ package Pedretti et al.,

2002; Pedretti et al., 2003; Pedretti et al., 2004) that

reinforces this hypothesis.

As can be seen, compounds with smaller ligands gen-

erally present higher antimalarial potencies. Volsub is

linked to the number of atoms present on the substituent

and the van der Waals radius of these atoms (being this last

associated with the electronic structure of the atom). In this

sense, substituents with low Volsub tend to present a

reduced number of atoms and/or atoms with lower van der

Waals radius, and both of these features lead to molecules

with a reduced number of electrons in the valence shell. As

evidenced in Fig. 6, it leads to higher EE values and then to

active derivatives, suggesting that antimalarial potencies of

the compounds can be mainly dominated by sterical

effects.

Despite the relevance of the descriptors Volsub and EE in

describing the antimalarial potency of the compounds, as

discussed before it is also important to address the physical

meaning associated with the presence of PEP9
L in Eq. 1

Fig. 6 Dependence of the electronic energy on the number of

electrons in sulfonamide derivatives

Fig. 7 Relationship between the substituent’s volume and the

antimalarial potency of sulfonamides

Med Chem Res (2016) 25:1630–1638 1635

123



(improving the quality of the DF obtained from LDA). In

his theory about frontier orbitals, K. Fukui suggested that

the electron density of the frontier orbitals carries infor-

mation regarding local reactivity of the molecules. In this

context, since electrophilic species tend to interact more

efficiently with electrons located mainly in the HOMO of

the molecules (most weakly bounded ones), Fukui pro-

posed that sites with higher PEPH values (higher contri-

bution to the formation of the HOMO) could be considered

the most reactive sites for reactions toward electrophiles

(Fukui, 1982). Similarly, considering that nucleophilic

agents tend to accommodate electrons preferentially in the

LUMO of the molecule, the sites with higher PEPL values

can be considered as the most reactive toward these spe-

cies. In Eq. 1, it is observed that active compounds

(DF[ 0.243) tend to present high values of PEP9
L, which

suggests that improved antimalarial activity can be

achieved if site 9 is a good electron acceptor. Such a result

indicates that the mechanism associated with the anti-

malarial activity of sulfonamides can be linked to chemical

reactions on atom 9 (probably toward nucleophiles).

In contrast with the EE parameter, it is not easy to

propose substitutions that lead to the desired changes on

the PEP9
L index of the compounds. However, it is reason-

able to consider that this parameter is directly affected by

substitutions on the X position. So, in order to evaluate

what changes could be induced by this substituent, some

complementary calculations were performed by consider-

ing varied groups at position X. For this purpose, the R1

and R2 groups were kept the same, being chosen as the

substituents present on the most active derivative, com-

pound 8 (NR1R2 = N(CH2)4). The X substituents were

chosen based on their polar and resonance effects on ring B

(Hammett substituent constants Carey and Sundberg,

2007): (1) electron releasing (by resonance—ERR or polar

effects—ERP) or (2) electron withdrawing (by reso-

nance—EWR or polar effects—EWP). Table 2 illustrates

the substituents employed and the values obtained for the

parameters PEP9
L, EE and DF (Eq. 1), as well as the status

predicted for the derivatives based on the LDA classifica-

tion rule. The indexes associated with compound 8 are also

presented for comparison.

A decrease in the EE value is observed for all the sub-

stituents (more negative values), in relation to the hydro-

genated compound. On the other hand, distinct effects are

observed on PEP9
L. In general, higher values of this parameter

are observed for substituents that present electron withdraw-

ing as thedominant effect on ringBof sulfonamides (thatwere

predicted as active derivatives), while lower values are

obtained for substituents with ERR or ERP dominant effects

(non-active derivatives). This result indicates that active

compounds could be obtained by an appropriate choice of X

substituent; in particular, electron-withdrawing groups are

good candidates to achieve high activity.

In summary, our approach seems to be more compre-

hensive than the topological indices presented by Agrawal

and collaborators (Agrawal et al., 2001a, b; Singh and

Agrawal, 2008). In their model, the compounds 5, 9, and 11

were considered as outliers, while in our case, only one

outlier is identified (compound 5). The misclassification of

molecule 5 by the other studies can indicate that the

electronic structure is related to the topology of the

molecule, which is an expected finding. However, it seems

that the electronic indexes bring more information than

purely topological parameters.

Conclusions

The antimalarial properties of a group of sulfonamides

were correlated with the electronic properties of these

materials. Simple and multiple linear regressions (SLR and

MLR), principal component analysis (PCA), and a linear

discrimination analysis (LDA) were employed.

Table 2 PEP9
L, EE, and DF parameters obtained for distinct substituents at position X of sulfonamides derivatives

ID Substituents Dominant effect associated with X Parameters Predicted status

–NR1R2 X PEP9
L EE DF

8 N(CH2)4 H – 0.144 -23,747.650 1.487 Active

8A N(CH2)4 CH3 ERRa and ERPb 0.139 -26,185.139 -0.076 Non-active

8B N(CH2)4 OH ERR 0.118 -26,474.527 -3.137 Non-active

8C N(CH2)4 NH2 ERR 0.082 -26,333.040 -8.270 Non-active

8D N(CH2)4 NO2 EWPc 0.174 -31,038.788 3.533 Active

8E N(CH2)4 F ERR and EWPd 0.149 -26,675.258 1.273 Active

8F N(CH2)4 CN EWP 0.168 -27,047.573 3.878 Active

a Electron releasing by resonance; b electron releasing by polar effects; c electron withdrawing by resonance; d electron withdrawing by polar

effects

1636 Med Chem Res (2016) 25:1630–1638

123



From LDA, we found a discriminant function with a

statistical significance of 98.9 %. The function was

obtained from the linear combination of only two elec-

tronic descriptors (among 65 indexes evaluated) and per-

mits correct classifying of 94 % of the studied compounds,

based on their antimalarial properties.

The obtained discriminant function indicates that active

compounds must present high values of electronic energy

and partial electronic population of the LUMO on atom 9

(this result is reinforced by PCA). The electronic energy

plays a major role in determining the degree of activity of

the compounds, which can be associated with the volume

of the substituents. The simplest recommendation for new

active compounds is employing substituents with the

lowest possible volume. The presence of the descriptor

PEP9
L in the DF suggests that this region is a potential

active site for the antimalarial action of the compounds.

Complementary calculations indicate that the attachment of

electron-withdrawing groups at the X position can improve

PEP9
L, leading also to new active compounds.

Acknowledgments ABN and FCL would like to thank the Brazilian

agency, Fundação de Amparo à Pesquisa do Estado de São Paulo

(FAPESP, proc. 04/13341-1) for financial support.

Compliance with ethical standards

Conflict of interest The authors declare that they have no conflict of

interest.

References

Agrawal VK, Sinha S, Bano S, Khadikar PV (2001a) QSAR studies

on antimalarial 2,4-diamino-6-quinazoline sulfonamides. Acta

Microbiol Immunol Hung 48:17–26

Agrawal VK, Srivastava R, Khadikar PV (2001b) QSAR studies on

some antimalarial sulfonamides. Bioorgan Med Chem

9:3287–3293

Aguiar ACC, Rocha EMMD, Souza NBD, França TCC, Krettli AU

(2012) New approaches in antimalarial drug discovery and

development: a review. Mem Inst Oswaldo Cruz 107:831–845

Autreto PAS, Lavarda FC (2008) Febrifugine derivative antimalarial

activity: quantum mechanical predictors. Rev Inst Med Trop

50:21–24

Barea C, Pabón A, Castillo D, Zimic M, Quiliano M, Galiano S,

Pérez-Silanes S, Monge A, Deharo E, Aldana I (2011) New

salicylamide and sulfonamide derivatives of quinoxaline 1,4-di-

N-oxide with antileishmanial and antimalarial activities. Bioorg

Med Chem Lett 21:4498–4502

Batagin-Neto A, Lavarda FC (2014) The correlation between

electronic structure and antimalarial activity of alkoxylated and

hydroxylated chalcones. Med Chem Res 23:580–586

Carey FA, Sundberg RJ (2007) Structural effects on stability and

reactivity.Advanced organic chemistry. Springer,US, pp 253–388

Desgrouas C, Dormoi J, Chapus C, Ollivier E, Parzy D, Taudon N

(2014) In vitro and in vivo combination of cepharanthine with

anti-malarial drugs. Malar J 13:90–95

Dewar MJS, Zoebisch EG, Healy EF, Stewart JJP (1985) Develop-

ment and use of quantum mechanical molecular models. 76.

AM1: a new general purpose quantum mechanical molecular

model. J Am Chem Soc 107:3902–3909

Eisele TP, Miller JM, Moonga HB, Hamainza B, Hutchinson P,

Keating J (2011) Malaria infection and anemia prevalence in

Zambia’s Luangwa District: an area of near-universal insecti-

cide-treated mosquito net coverage. Am J Trop Med Hyg

84:152–157

Elslager EF, Colbry NL, Davoll J, Hutt MP, Johnson JL, Werbel LM

(1984) Folate antagonists. 22. Antimalarial and antibacterial

effects of 2,4-diamino-6-quinazolinesulfonamides. J Med Chem

27:1740–1743

Flannery EL, Chatterjee AK, Winzeler EA (2013) Antimalarial drug

discovery [mdash] approaches and progress towards new

medicines. Nat Rev Microbiol 11:849–862

Fukui K (1982) Role of frontier orbitals in chemical reactions.

Science 218:747–754

Guiguemde WA, Hunt NH, Guo J, Marciano A, Haynes RK, Clark J,

Guy RK, Golenser J (2014) Treatment of murine cerebral

malaria by artemisone in combination with conventional

antimalarial drugs: antiplasmodial effects and immune

responses. Antimicrob Agents Chemother 58:4745–4754

Hair JF, Tatham RL, Anderson RE, Black W (1998) Multivariate data

analysis. Prentice Hall, New Jersey

Jensen K, Plichta D, Panagiotou G, Kouskoumvekaki I (2012)

Mapping the genome of Plasmodium falciparum on the drug-like

chemical space reveals novel anti-malarial targets and potential

drug leads. Mol BioSyst 8:1678–1685

Kahn CM (2005) The Merck veterinary manual, 9th edn. Merck,

Whitehouse Station

Kumar Parai M, Panda G, Srivastava K, Kumar Puri S (2008) Design,

synthesis and antimalarial activity of benzene and isoquinoline

sulfonamide derivatives. Bioorg Med Chem Lett 18:776–781

KümpornsinK,ModchangC,HeinbergA, EklandEH, Jirawatcharadech

P, Chobson P, Suwanakitti N, Chaotheing S, Wilairat P, Deitsch

KW,KamchonwongpaisanS, FidockDA,KirkmanLA,Yuthavong

Y, Chookajorn T (2014) Origin of robustness in generating drug-

resistant malaria parasites. Mol Biol Evol 31:1649–1660

Malešič M, Krbavčı́č A, Golobič A, Golič L, Stanovnik B (1997) The

synthesis and transformations of some 3-thiocarbamoylthiazo-

lidines. J Heterocycl Chem 34:43–48

Metropolis H, State R, Harcourt P (2014) Effectiveness of insecticide-

treated mosquito nets (Itns) in the control of malaria disease

among slum dwellers in port. Res Hum Soc Sci 4:79–85

Mita T, Ohashi J, Venkatesan M, Marma ASP, Nakamura M, Plowe

CV, Tanabe K (2014) Ordered accumulation of mutations

conferring resistance to sulfadoxine-pyrimethamine in the Plas-

modium falciparum parasite. J Infect Dis 209:130–139

Mutabingwa TK (2005) Artemisinin-based combination therapies

(ACTs): best hope for malaria treatment but inaccessible to the

needy! Acta Trop 95:305–315

Naranjo-Montoya OA, Martins LM, da Silva-Filho LC, Batagin-Neto

A, Lavarda FC (2014) The correlation between electronic

structure and antimalarial activity of tetrahydropyridines.

J Braz Chem Soc 26:255–265

Ojala CR, Ostman JM, Ojala WH, Hanson SE (2001) Molecular and

crystal structures of N-arylglycopyranosylamines formed by

reaction between sulfanilamide and D-ribose, D-arabinose and D-

mannose. Carbohydr Res 331:319–325

Pedretti A, Villa L, Vistoli G (2002) VEGA: a versatile program to

convert, handle and visualize molecular structure on Windows-

based PCs. J Mol Graph Model 21:47–49

Pedretti A, Villa L, Vistoli G (2003) Atom-type description language:

a universal language to recognize atom types implemented in the

VEGA program. Theor Chem Acc 109:229–232

Pedretti A, Villa L, Vistoli G (2004) VEGA—an open platform to

develop chemo-bio-informatics applications, using plug-in

Med Chem Res (2016) 25:1630–1638 1637

123



architecture and script programming. J Comput Aid Mol Des

18:167–173

Perakslis ED (2014) Cybersecurity in health care. N Engl J Med

371:395–397

PrimasN,Verhaeghe P,CohenA,KiefferC,DumètreA,Hutter S,Rault

S, Rathelot P, Azas N, Vanelle P (2012) A new synthetic route to

original sulfonamide derivatives in 2-trichloromethylquinazoline

series: a structure–activity relationship study of antiplasmodial

activity. Molecules 17:8105–8117

Salahuddin A, Inam A, van Zyl RL, Heslop DC, Chen C-T, Avecilla

F, Agarwal SM, Azam A (2013) Synthesis and evaluation of

7-chloro-4-(piperazin-1-yl)quinoline-sulfonamide as hybrid

antiprotozoal agents. Bioorgan Med Chem 21:3080–3089

Santelli AC, Ribeiro I, Daher A, Boulos M, Marchesini PB, dos

Santos RLC, Lucena MBF, Magalhães I, Leon AP, Junger W,

Ladislau JLB (2012) Effect of artesunate-mefloquine fixed-dose

combination in malaria transmission in Amazon basin commu-

nities. Malar J 11:286–297

Singh J, Agrawal V (2008) Use of topological as well as quantum

chemical parameters in modelling antimalarial activity of 2,

4-diamino-6-quinazoline sulphonamides. Oxid Commun

31:17–26

Smith DA, Jones RM (2008) The sulfonamide group as a structural

alert: a distorted story? Curr Opin Drug Discov 11:72–79

SPSS Inc (2001) SPSS for Windows, Version 11.0.1. SPSS Inc,

Chicago

Spoo JW, Riviere JE (2001) Sulfonamides. In: Adams HR (ed)

Veterinary pharmacology and therapeutics. Iowa State Univer-

sity Press, Ames, IO, pp 796–817

Stewart JJP (1990) MOPAC: a semiempirical molecular orbital

program. J Comput Aid Mol Des 4:1–103

Winzeler EA, Manary MJ (2014) Drug resistance genomics of the

antimalarial drug artemisinin. Genome Biol 15:544–555

World Health Organization (2011) World malaria report 2011. World

Health Organization, Genova

1638 Med Chem Res (2016) 25:1630–1638

123


	Electronic descriptors for the antimalarial activity of sulfonamides
	Abstract
	Introduction
	Materials and methods
	Results and discussion
	Simple and multiple linear regressions
	Principal component analysis
	Linear discriminant analysis
	Achieving high activity

	Conclusions
	Acknowledgments
	References