Antiplasmodial Activity of Nitroaromatic and Quinoidal ... - Science Direct

Archives of Biochemistry and Biophysics Vol. 393, No. 2, September 15, pp. 199 –206, 2001 doi:10.1006/abbi.2001.2487, available online at http://www.idealibrary.com on

Antiplasmodial Activity of Nitroaromatic and Quinoidal Compounds: Redox Potential vs Inhibition of Erythrocyte Glutathione Reductase Philippe Grellier,* Jonas Sˇarlauskas,† Zˇilvinas Anusevicˇius,† Audrone˙ Maroziene˙,† ˇ e˙nas† ,1 Chantal Houee-Levin,‡ Joseph Schrevel,* and Narimantas C *Laboratoire de Biologie Parasitaire et Chimiothe´rapie, Muse´um National d’Histoire Naturelle, IFR 63, 61 rue Buffon, 75231 Paris Cedex 05, France; †Institute of Biochemistry, Mokslininku 12, Vilnius 2600, Lithuania; and ‡LCP, UMR 800 CNRS, Universite Paris-Sud, Bat. 350, 91405 Orsay, France

Received March 13, 2001, and in revised form June 11, 2001; published online August 21, 2001

Prooxidant nitroaromatic and quinoidal compounds possess antimalarial activity, which might be attributed either to their formation of reactive oxygen species or to their inhibition of antioxidant enzyme glutathione reductase (GR, EC 1.6.4.2). We have examined the activity in vitro against Plasmodium falciparum of 24 prooxidant compounds of different structure (nitrobenzenes, nitrofurans, quinones, 1,1ⴕ-dibenzyl-4,4ⴕ-bipyridinium, and methylene blue), which possess a broad range of singleelectron reduction potentials (E 71 ) and erythrocyte glutathione reductase inhibition constants (K i(GR) ). For a series of homologous derivatives of 2-(5ⴕ-nitrofurylvinyl)quinoline-4-carbonic acid, the relationship between compound K i(GR) and concentration causing 50% parasite growth inhibition (IC 50 ) was absent. For all the compounds examined in this study, the dependence of IC 50 on their K i(GR) was insignificant. In contrast, IC 50 decreased with an increase in E 71 and positive electrostatic charge of aromatic part of molecule (Z): log IC 50 (␮M) ⴝ ⴚ(0.9846 ⴞ 0.3525) ⴚ (7.2850 ⴞ 1.2340) E 71 (V) ⴚ (1.1034 ⴞ 0.1832) Z (r 2 ⴝ 0.8015). The redox cycling activity of nitroaromatic and quinoidal compounds in ferredoxin:NADP ⴙ reductase-catalyzed reaction and the rate of oxyhemoglobin oxidation in lysed erythrocytes increased with an increase in their E 71 value. Our findings imply that the antiplasmodial activity of nitroaromatic and quinoidal compounds is mainly influenced by their ability to form reactive oxygen species, and much less significantly by the GR inhibition. © 2001 Academic Press 1 To whom correspondence should be addressed. Fax: 370 2 729196. E-mail: [email protected].

0003-9861/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

Key Words: Plasmodium falciparum; nitrofuran; quinone; methylene blue; glutathione reductase; oxidative stress.

The emergence of resistance of malarial parasite Plasmodium falciparum to available drugs, e.g., chloroquine, resulted in the demand for new antimalarial drugs and an understanding of their mechanisms of action. Malaria-infected human erythrocytes are under a tolerable degree of oxidative stress, which appears to promote growth and differentiation of P. falciparum (1). However, the enhancement of oxidative stress by the inherited erythrocyte deficiency in NADPH-regenerating glucose-6-phosphate dehydrogenase, or by the action of prooxidant compounds, suppresses the parasite growth, since the latter appears to be more sensitive to reactive oxygen species than the host cells (2–5). The antimalarial activity of nitrothiophenes and other nitroaromatic compounds was attributed to the formation of reactive oxygen species during flavoenzymecatalyzed redox cycling reactions and/or oxyhemoglobin oxidation (2, 3). An analogous mechanism has been proposed for the action of methylene blue and 10-arylizoalloxazines (2). Antimalarial 2-hydroxy-3-alkyl-1,4naphthoquinones, which act as specific inhibitors of mitochondrial bc 1 complex of P. falciparum (6), are also able to participate in redox cycling reactions and induce methemoglobin formation (2, 4). The prooxidant events may be important in the action of widely used antimalarial agents like amodiaquine and pyronardine, since they may form oxidation products of quinone imine structures (7); besides, hydroxylated metabolites of primaquine also autoxidize to quinone 199

200

GRELLIER ET AL.

imines which undergo flavoenzyme-catalyzed redox cycling (8). Irrespective of a large amount of data, the quantitative structure–activity relationships in the antiplasmodial activity of prooxidant compounds are poorly understood. The toxicity of nitroaromatics and quinones against mammalian cells and bacteria often increases upon an increase in their single-electron reduction potential (E 71 ) value (9 –12). The relationship ⌬ log IC 50/ ⌬E 71 ⬵ ⫺10 V ⫺1, where IC 50 is the compound concentration for 50% cell survival, indicates that the main factor of cytotoxicity is redox cycling and oxidative stress (10). Thus, if other factors are of minor importance, one may expect an increase in the antiplasmodial activity of nitroaromatic and quinoidal compounds with their redox potential. However, the antimalarial activity of some redox active compounds, such as 10arylizoalloxazines (13–15) and methylene blue (16, 17), was also attributed to their inhibition of antioxidant flavoenzyme glutathione reductase (GR, 2 EC 1.6.4.2), which catalyzes the reduction of glutathione disulfide (GSSG) at the expense of NADPH. It is assumed that both human erythrocyte and P. falciparum GR play important roles for the intraerythrocyte growth of parasites, protecting it from oxidative stress (5, 13–17). Since nitroaromatic and quinoidal compounds may efficiently inhibit GR from various sources (18 –22), it is necessary to assess the relative importance of redox cycling and GR inhibition in their antiplasmodial activity. The solution of this problem may provide some guidelines in the rational design of antimalarial and, possibly, trypanocidal drugs, in view of the inhibition of antioxidant flavoenzyme trypanothione reductase (EC 1.6.4.8) by quinones and nitroaromatics (5, 21–25). In this work, we have shown that the antiplasmodial activity of model prooxidant compounds of different structure (nitrobenzenes, nitrofurans, quinones, 1,1⬘dibenzyl-4,4⬘-bipyridinium, and methylene blue) was determined mainly by the single-electron reduction potential and electrostatic charge of molecules, and not by glutathione reductase inhibition. MATERIALS AND METHODS Materials. Recombinant human erythrocyte GR, overexpressed in Escherichia coli and isolated as previously described (26), was a generous gift of Dr. Katja Becker (Biochemie-Zentrum, Heidelberg University, Germany). The enzyme concentration was determined spectrophotometrically using ␧ 464 ⫽ 11 mM ⫺1 cm ⫺1. Ferredoxin: 2 Abbreviations used: GR, glutathione reductase; FNR, ferredoxin: NADP ⫹ reductase; Hb-Fe 2⫹O 2, oxyhemoglobin; Hb-Fe 3⫹, methemoglobin; GSSG, glutathione disulfide; Q, quinone; ArNO 2, aromatic nitrocompound; E 71 , single-electron reduction potential at pH 7.0; k cat, catalytic constant; k cat/K m, steady-state bimolecular rate constant; K i(GR), inhibition constant toward glutathione reductase; P, octanol/water partition coefficient; IC 50, compound concentration causing 50% parasite growth inhibition; Z, electrostatic charge.

NADP ⫹ reductase (FNR, EC 1.18.1.2) from Anabaena PCC 7119 prepared as described previously (27) was a generous gift of Dr. Marta Martinez-Julvez and Prof. Carlos Gomez-Moreno (Zaragoza University, Spain). The concentration of FNR was determined using ␧ 459 ⫽ 9.4 mM ⫺1 cm ⫺1. The derivatives of 2-(5⬘-nitrofurylvinyl)-quinoline-4-carbonic acid a–i (Fig. 1) were synthesized by condensation of 5-nitrofuran-2-aldehyde with 2-methylquinoline-4-carbonic acid, subsequent preparation of 2-(5⬘-nitrofurylvinyl)-quinoline-4-carbonic acid chloranhydride, and its amidation by corresponding amines according to adapted procedures (28). All compounds were characterized by melting point, thin-layer chromatography, and 1H NMR. Other compounds were purchased from Sigma or Aldrich, and were used as received. Freshly prepared suspensions of erythrocytes from healthy patients, obtained from Vilnius Blood Transfusion Center, were washed twice by centrifugation, resuspended in 0.01 M K-phosphate (pH 7.0), containing 0.137 M NaCl, 0.0027 M KCl, 10 mM glucose, and 1 mM EDTA, and stored at 4°C. The stock suspension of erythrocytes was used within 7–10 days of preparation. Enzymatic assays. Steady-state rates of GR-catalyzed NADPH oxidation were determined spectrophotometrically using ⌬␧ 340 ⫽ 6.2 mM ⫺1 cm ⫺1 in a Hitachi-557 spectrophotometer. Assays were performed in 0.05 M Hepes, pH 7.0, containing 1.0 mM EDTA, at 25°C. At saturating substrate concentrations (100 ␮M NADPH, 1 mM GSSG), the catalytic constant (k cat) of GR was equal to 190 s ⫺1. For inhibition experiments, the reaction rates were determined either at a fixed NADPH concentration (100 ␮M) and varied GSSG (40 –500 ␮M), or at a fixed GSSG concentration (1 mM) and varied NADPH (10 –50 ␮M), in the presence of 4 –5 concentrations of inhibitor, and in its absence. The inhibition constant (K i(GR)) was obtained from dependence of 1/k cat on the inhibitor concentration. Nitro and quinone reductase activities of GR and FNR were determined according to NADPH oxidation rate, in the presence of 150 ␮M NADPH and various amounts of nitro compounds and quinones. For the studies of oxyhemoglobin (Hb-Fe 2⫹O 2) oxidation, the erythrocytes were lysed in buffer solution containing 40 ␮g/ml digitonin, and the Hb-Fe 2⫹O 2 concentration was adjusted to 7.5–30 ␮M (␧ 577 ⫽ 15.37 mM ⫺1 cm ⫺1 (28)). Kinetics of Hb-Fe 2⫹O 2 conversion to methemoglobin (Hb-Fe 3⫹) was monitored according to absorbance rise at 630 nm and absorbance decrease at 577 nm (⌬␧ 630 ⫽ 3.46 mM ⫺1 cm ⫺1, ⌬␧ 577 ⫽ 10.55 mM ⫺1 cm ⫺1 (29)), after addition of excess oxidant (molar ratio 1:10 – 100), at 37°C. The oxidation of NADH or NADPH (500 ␮M) by quinoidal or nitroaromatic compounds in the presence of lysed erythrocytes (initial hematocrite 1%) was monitored according to absorbance decrease at 340 nm (⌬␧ 340 ⫽ 6.2 mM ⫺1 cm ⫺1) at 37°C, using a 0.2-cm optical path cell. The reaction mixture was stirred by aeration. Pulse-radiolysis studies. Pulse-radiolysis studies of chinifur (compound i, Fig. 1) were carried out using the equipment setup at the Institut Curie-Biologie (Orsay) described elsewhere (30). The samples of chinifur (30 –100 ␮M) in deaerated 0.02 M K-phosphate and 0.1 M K-formate solution (pH 7.0) were irradiated with an average dose per pulse 10 Gy, which ensured initial COO ⫺ concentrations of 6.2 ␮M. The rate constant of chinifur reduction by COO ⫺ and the bimolecular rate constant of dismutation of chinifur radical were calculated as described elsewhere (31), monitoring the transient absorbance rise and decay of the radical. The E 71 of chinifur has been calculated from the one-electron transfer equilibrium constants with 1,1⬘-dibenzyl-4,4⬘-bipyridinium ion or tetramethyl-1,4-benzoquinone (31). In vitro P. falciparum culture and drug assays. The chloroquineresistant strain FcB1 of P. falciparum was maintained in continuous culture on human erythrocytes as described by Trager and Jensen (32). In vitro antiplasmodial activity was determined using a modification of the semiautomated microdilution technique of Desjardins et al. (33). Stock solutions of test compounds were prepared in DMSO. Drug solutions were serially diluted with culture medium and added to asynchronous parasite cultures (1% parasitemia and

ANTIPLASMODIAL ACTIVITY OF NITROAROMATIC AND QUINOIDAL COMPOUNDS

FIG. 1. Structural formulae of derivatives of 2-(5⬘-nitrofurylvinyl)quinoline-4-carbonic acid used in this work.

1% final hematocrit) in 96-well plates for 24 h, at 37°C, prior to the addition of 0.5 ␮Ci of [ 3H]hypoxanthine (1 to 5 Ci/mmol; Amersham, Les Ulis, France) per well, for 24 h. The growth inhibition for each drug concentration was determined by comparison of the radioactivity incorporated into the treated culture with that in the control culture (without drug) maintained on the same plate. The concentration causing 50% inhibition (IC 50) was obtained from the drug concentration–response curves and the results were expressed as the mean ⫾ standard deviation determined from three independent experiments. The presence of DMSO (ⱕ0.1%) did not inhibit the parasite growth. Partition coefficient calculation and statistical analysis. The octanol/water partition coefficients for nitro compounds and quinones were calculated using ACD log P (version 1) software, a generous gift of Advanced Chemistry Development Inc. (Toronto, Canada). The multiparameter regression analysis was performed using Statistica (version 4.3) software (StatSoft Inc.).

201

acter of inhibition is consistent with the binding of electron-deficient aromatic compounds at the interface of two subunits of GR, at the vicinity of histidines75,75⬘ and phenylalanines-78,78⬘ (34). One should note that the IC 50 of compounds a–i against P. falciparum were in a similar range of concentration (4.3–17.1 ␮M), and did not correlate with their K i(GR), which ranged from 2.5 to 300 ␮M (Table I). We also measured the previously unavailable value of E 71 of compound i (chinifur) (Fig. 1), assuming that E 71 for compounds a–i should be sufficiently close, in view of similar electron-accepting properties of substituents in the 4-position of the quinoline ring (Fig. 1). Briefly, chinifur has been reduced by radiolytically generated anion radicals of formate with a bimolecular rate constant of (8.5 ⫾ 1.1) ⫻ 10 8 M ⫺1 s ⫺1. The transiently formed radical (␭ max ⫽ 480 nm, ␧ 480 ⫽ 11.1 ⫾ 0.24 mM ⫺1 cm ⫺1) decayed in a bimolecular way, with dismutation rate constant of (2.47 ⫾ 0.14) ⫻ 10 8 M ⫺1 s ⫺1. The one-electron transfer equilibrium constants of radicals of chinifur with 1,1⬘-dibenzyl-4,4⬘-bipyridinium (E 71 ⫽ ⫺0.354 V) or tetramethyl-1,4-benzoquinone (E 71 ⫽ ⫺0.244 V) calculated according to the rates of approach to equilibrium (25), were equal to 0.0062 ⫾ 0.0024 and to 0.556 ⫾ 0.109, respectively. The values of E 71 of chinifur, calculated according to the Nernst equation, were equal to ⫺0.221 ⫾ 0.01 and ⫺0.228 ⫾ 0.007 V, using 1,1⬘-dibenzyl-4,4⬘-bipyridinium and tetra-

RESULTS

First, we examined the antiplasmodial activity of a series of homologous nitroaromatic compounds, which were either strong or weak inhibitors of erythrocyte GR. For this purpose, we selected the derivatives of 2-(5⬘-nitrofurylvinyl)-quinoline-4-carbonic acid (Fig. 1), which possess a broad spectrum of bactericidal and antiparasitic activities (28), and inhibit yeast GR (19) and Trypanosoma congolense trypanothione reductase (21). It was found that compounds a–i were uncompetitive inhibitors toward both GR substrates, NADPH and GSSG; i.e., they decreased enzyme k cat and did not affect the steady-state bimolecular rate constants of substrate reaction with GR (k cat/K m) (Fig. 2). The char-

FIG. 2. Inhibition of erythrocyte glutathione reductase by 2-(5⬙nitrofurylvinyl)-quinoline-4-carbonyl-1⬘,3⬘-diaminopropane (compound f, Fig. 1). Concentration of compound f: 0 (1), 3.1 ␮M (2), 6.25 ␮M (3), 12.5 ␮M (4), and 25 ␮M (5); concentration of NADPH, 100 ␮M.

202

GRELLIER ET AL. TABLE I

The Constants of Glutathione Reductase Inhibition (K i(GR)) of Derivatives of 2-(5⬘-Nitrofurylvinyl)-quinoline-4-carbonic Acid a–i (Fig. 1), Their Octanol/Water Partition Coefficients (P), and Their Concentrations Causing 50% P. falciparum Growth Inhibition (IC 50) Compound

K i(GR) (␮M) a

log P b

IC 50 (␮M)

a b c d e f g h i

3 2.5 60 25 ⱖ300 2.5 42.5 25 45

n.d. 2.46 ⫾ 0.30 2.58 ⫾ 0.30 3.11 ⫾ 0.35 2.17 ⫾ 0.30 n.d. n.d. n.d. n.d.

17.1 ⫾ 1.5 4.5 ⫾ 0.3 7.7 ⫾ 0.5 7.4 ⫾ 0.3 7.4 ⫾ 0.3 9.2 ⫾ 0.3 11.1 ⫾ 0.4 6.4 ⫾ 0.4 4.3 ⫾ 0.3

Determined at a fixed NADPH concentration (100 ␮M) and varied concentrations of GSSG. b Not calculated for electrostatically charged compounds. a

methyl-1,4-benzoquinone, respectively. In subsequent analysis, the value of E 71 ⫽ ⫺0.225 V was used for compounds a–i. Next, we examined the activity against P. falciparum of a series of model compounds with a broad range of E 71 values (Table II). These compounds also acted as uncompetitive GR inhibitors. Except methylene blue and 9,10-phenanthrene quinone, the K i(GR) values of model compounds were high, ranging from 71 to 2000 ␮M (Table II).

The analysis of antiplasmodial activity of all the compounds used in this study (Tables I and II), demonstrates that the dependence of log IC 50 on their log K i(GR) is not significant (r 2 ⫽ 0.2296, data not shown). However, there exists a rough linear relationship (r 2 ⫽ 0.4589) between their log IC 50 and their E 71 (Fig. 3A). It appears that the compounds possessing a positive charge in the aromatic part of the molecule (1,1⬘-dibenzyl-4,4⬘-bipyridinium, methylene blue) possess enhanced activity, whereas the negatively charged compounds, e.g., 3,5-dinitrobenzoic acid, are less active (Fig. 3A). It seems that the positively charged alkylamine substitutes of compounds f–i (Table I) do not influence their activity (Fig. 3A). This prompted us to introduce an electrostatic charge of the aromatic part of the molecule (Z) as the second correlation parameter (⫹2 for 1,1⬘-dibenzyl-4,4⬘-bipyridinium, ⫹1 for methylene blue, ⫺1 for 4-nitrobenzoic acid, 3,5-dinitrobenzoic acid, and compound a, and zero for other compounds). This resulted in a markedly improved correlation (Fig. 3B, r 2 ⫽ 0.8015): log IC50 共 ␮ M兲 ⫽ ⫺共0.9846 ⫾ 0.3525兲 ⫺ 共7.2850 ⫾ 1.2350兲 E 71 共V兲 ⫺ 共1.1034 ⫾ 0.1832兲 Z.

[1]

For uncharged compounds b– e (Table I), and 1, 3, 6 –14 (Table II), the log IC 50 vs E 71 dependence is characterized by r 2 ⫽ 0.5800. The introduction of the

TABLE II

Single-Electron Reduction Potentials of Nitroaromatic and Quinoidal Compounds (E 71 ) (35), Their Constants of Glutathione Reductase Inhibition (K i(GR)), Octanol/Water Partition Coefficients (P), and Their Concentrations Causing 50% P. falciparum Growth Inhibition (IC 50) No.

Compound

E 71 (V)

K i(GR) (␮M) a

log P b

IC 50 (␮M)

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

Nitrobenzene 4-Nitrobenzoic acid 4-Nitroacetophenone 1,1⬘-Dibenzyl-4,4⬘-bipyridinium 3,5-Dinitrobenzoic acid 3,5-Dinitrobenzamide m-Dinitrobenzene 4-Nitrobenzaldehyde o-Dinitrobenzene p-Dinitrobenzene Nifuroxime Nitrofurantoin Tetramethyl-1,4-benzoquinone 9,10-Phenanthrene quinone Methylene blue

⫺0.485 ⫺0.425 ⫺0.355 ⫺0.354 ⫺0.350 ⫺0.350 ⫺0.345 ⫺0.325 ⫺0.287 ⫺0.257 ⫺0.255 ⫺0.255 ⫺0.26 ⫺0.12 ⫺0.10 c

⬎2000 800 400 ⬎1000 350 ⬎1000 320 290 ⬎1000 71 200 200 250 12 2.3

1.95 ⫾ 0.40 n.d. 1.42 ⫾ 0.30 n.d. n.d. 0.55 ⫾ 0.20 1.62 ⫾ 0.15 1.56 ⫾ 0.30 1.84 ⫾ 0.20 1.37 ⫾ 0.20 0.73 ⫾ 0.20 ⫺0.54 ⫾ 0.20 2.63 ⫾ 0.30 3.13 ⫾ 0.30 n.d.

473 ⫾ 113 360 ⫾ 16 172 ⫾ 8 0.153 ⫾ 0.016 390 ⫾ 17 30.3 ⫾ 3.1 50.5 ⫾ 2.4 79 ⫾ 28 11.7 ⫾ 1.1 0.26 ⫾ 0.03 14.7 ⫾ 0.8 12.9 ⫾ 1.3 4.7 ⫾ 0.4 3.9 ⫾ 0.2 0.006 ⫾ 0.001

a b c

Determined at a fixed NADPH concentration (100 ␮M) and varied concentrations of GSSG. Not calculated for electrostatically charged compounds. Calculated from data of Ref. (36), assuming ⌬E 1 /⌬pH ⫽ ⫺59 mV between pH 3.0 and pH 7.0.


203

FIG. 3. (A) The dependence of IC 50 of nitroaromatic and quinoidal compounds against P. falciparum on their single-electron reduction potential (E 71 ). (B) The dependence of IC 50 of nitroaromatic and quinoidal compounds against P. falciparum on their single-electron reduction potential (E 71 ) and electrostatic charge of the aromatic part of the molecule (Z) according to a multiparameter Eq. [1]. The compounds are indicated as in Tables I and II.

compound octanol/water partition coefficient (log P, Tables I and II) as a second parameter did not improve the correlation (r 2 ⫽ 0.5850, data not shown), thus pointing to a limited influence of compound lipophilicity on their activity. Finally, the multiparameter regression analysis of IC 50 dependence on E 71 , Z, and K i(GR) values demonstrates the uncertain role of GR inhibition in the antiplasmodial activity of the compounds examined: log IC50 共 ␮ M兲 ⫽ ⫺共0.9376 ⫾ 0.3637兲 ⫺ 共6.0695 ⫾ 2.1102兲 E 71 共V兲 ⫺ 共1.1405 ⫾ 0.1925兲 Z ⫹ ⫹ 共0.1427 ⫾ 0.1997兲log K i(GR)共 ␮ M兲, 共r 2 ⫽ 0.8065兲.

[2]

Since it is assumed that the linear log IC 50 vs E 71 relationships point to enzymatic redox cycling as the main factor of toxic action of prooxidant compounds (10), our next goal was to determine whether the redox cycling properties of the two most active antiplasmodial compounds, 1,1⬘-dibenzyl-4,4⬘-bipyridinium and methylene blue (Table II), correlate with their E 71 values. For this purpose, we have used ferredoxin:NADP ⫹ reductase which performs a single-electron reduction of quinoidal and nitroaromatic compounds (8, 26), as a model enzyme. FNR catalyzed redox cycling of 1,1⬘-

dibenzyl-4,4⬘-bipyridinium and methylene blue accompanied by oxidation of excess NADPH, with bimolecular rate constants (k cat/K m) of 9.2 ⫻ 10 5 and 2.4 ⫻ 10 4 M ⫺1 s ⫺1, respectively (k cat ⫽ 50 s ⫺1 in the presence of 150 ␮M NADPH). The FNR-catalyzed redox cycling of other compounds used in the present study, was examined previously (27, 37). It is evident that the reactivities of 1,1⬘-dibenzyl-4,4⬘-bipyridinium and methylene blue match the previously demonstrated reactivity vs E 71 relationship (27, 37) (r 2 ⫽ 0.8253, Fig. 4). In addition, methylene blue oxidized NADPH in the absence of enzyme, with rate constant of 6.0 M ⫺1 s ⫺1. In this reaction, the amount of NADPH oxidized significantly exceeds the concentration of methylene blue, indicating that the redox cycling of methylene blue takes place. The redox cycling of 1,1⬘-dibenzyl-4,4⬘-bipyridinium, methylene blue, and several other oxidants used in this study may be observed in lysed erythrocytes as well. Thus, in the presence of lysed erythrocytes (initial hematocrit, 1%), 50 ␮M 1,1⬘-dibenzyl-4,4⬘bipyridinium, tetramethyl-1,4-benzoquinone, or p-dinitrobenzene oxidized 100 –150 ␮M NADH or NADPH during 1 h, whereas 50 ␮M methylene blue oxidized 400 – 450 ␮M NADH or NADPH. High activity of methylene blue resulted mainly from its rapid nonenzymatic oxidation of NAD(P)H; however, the presence of lysed erythrocytes accelerated the reaction by 25–30%. Next, we examined the kinetics of oxyhemoglobin (Hb-Fe 2⫹O 2) oxidation in lysed erythrocytes. The kinetic analysis of the Hb-Fe 2⫹O 2 oxidation by nitroaro-

204

GRELLIER ET AL.

sive” GR substrates, i.e., undergo GR-catalyzed redox cycling reactions (19, 22). Nitroaromatic compounds (Tables I and II) were almost inactive as oxidizing GR substrates, the k cat not exceeding 0.1 s ⫺1. GR reduced quinoidal compounds more rapidly: 9,10-phenanthrene quinone with k cat ⫽ 4 s ⫺1, and k cat/K m ⫽ 2.9 ⫻ 10 4 M ⫺1 s ⫺1; 1,1⬘-dibenzyl-4,4⬘-bipyridinium with k cat ⫽ 1.5 s ⫺1 and k cat/K m ⫽ 1.5 ⫻ 10 4 M ⫺1 s ⫺1; methylene blue with k cat ⫽ 1.0 s ⫺1 and k cat/K m ⫽ 3.9 ⫻ 10 4 M ⫺1 s ⫺1; tetramethyl-1,4-benzoquinone with k cat ⫽ 0.4 s ⫺1 and k cat/ K m ⫽ 9.1 ⫻ 10 3 M ⫺1 s ⫺1. However, in view of relatively low rates, the GR-catalyzed redox cycling reactions of quinoidal compounds do not seem to be important in their antiplasmodial action (22). DISCUSSION

FIG. 4. The dependence of reactivity of nitroaromatic and quinoidal compounds on their single-electron reduction potential (E 71 ) in steady-state reduction of by ferredoxin:NADP ⫹ reductase (A), and oxidation of oxyhemoglobin in lysed human erythrocytes (B). The compounds are indicated as in Tables I and II. The bimolecular steady-state rate constants of reduction of compounds 1–3, 5–14, and i by ferredoxin:NADP ⫹ reductase are taken from Refs. (27, 37).

matic and quinoidal compounds is complex, since the methemoglobin (Hb-Fe 3⫹) formed slows down the reaction (38, 39), Hb-Fe 2⫹O2 ⫹ Q 共ArNO2 兲 ^ Hb-Fe 3⫹ ⫹ O2 ⫹ Q ⫺ 共ArNO 2⫺ 兲,

[3]

where Q is quinone, and ArNO 2 is aromatic nitrocompound. Thus, the approximate values of the rate constants were determined according to the initial reaction rates: 33.0 M ⫺1 s ⫺1 (methylene blue), 27.6 M ⫺1 s ⫺1 (9,10-phenanthrene quinone), 3.0 M ⫺1 s ⫺1 ( p- and odinitrobenzenes), 1.8 M ⫺1 s ⫺1 (m-dinitrobenzene), 1.6 M ⫺1 s ⫺1 (nifuroxime and nitrofurantoin), 0.9 M ⫺1 s ⫺1 (tetramethyl-1,4-benzoquinone), 0.7 M ⫺1 s ⫺1 (chinifur (compound i, Fig. 1)), 0.45 M ⫺1 s ⫺1 (4-nitrobenzaldehyde and 1,1⬘-dibenzyl-4,4⬘-bipyridinium), 0.30 M ⫺1 s ⫺1 (3,5-dinitrobenzoic and 4-nitrobenzoic acid), 0.25 M ⫺1 s ⫺1 (3,5-dinitrobenzamide and 4-nitroacetophenone), and 0.10 M ⫺1 s ⫺1 (nitrobenzene). It is evident that there also exists the linear relationship (r 2 ⫽ 0.8041) between the log of Hb-Fe 2⫹O 2 oxidation rate constant and E 71 of the examined compounds (Fig. 4). With respect to the interaction of quinones and nitroaromatics with glutathione reductase, one should also note that these compounds may act as “subver-

Our data indicate that the antiplasmodial activity of the examined nitroaromatics and quinoidal compounds is mainly influenced by their E 71 values, and the electrostatic charge of their aromatic part, and much less significantly by the inhibition of human erythrocyte GR (Figs. 3A and 3B, and Eqs. [1] and [2]). Evidently, it points to some importance of compound interaction with negatively charged membranes of erythrocytes or/and preferential accumulation of aromatic cations. The formation of reactive oxygen species by these groups of compounds in erythrocytes, the synergism of their action with oxygen, and reactive forms of iron are well documented (2, 40). In particular, the killing of P. falciparum by paraquat (1,1⬘-dimethyl-4,4⬘-bipyridinium) was prevented by desferrioxamine, which prevents the Fenton reaction (40). Although both human erythrocyte and P. falciparum glutathione reductase protect malarial parasites from oxidative stress (10, 11), the inhibitors of erythrocyte GR exhibit comparable activity against P. falciparum enzymes (10 –12); e.g., methylene blue inhibits both GR with sufficiently close K i values (10). However, the micromolar K i(GR) of methylene blue is far above its IC 50 against P. falciparum (Table II). This is also evident for 1,1⬘-dibenzyl4,4⬘-bipyridinium which acts as very weak inhibitors of GR, but possesses high antiplasmodial activity (Table II). In this aspect, it is of some interest to compare the previously reported activity of 10-(4⬘-chlorophenyl)-3methylflavin (13) which inhibits human erythrocyte and P. falciparum GR with K i of 1 and 5 ␮M, respectively (13, 14), with its redox properties. Although the value of E 71 for this compound is unavailable, its twoelectron reduction potential is equal to ⫺0.13 V (14). Assuming that E 71 of flavins are by 0.05 V– 0.10 V more negative than their two-electron reduction potentials (35), the E 71 of 10-(4⬘-chlorophenyl)-3-methylflavin should be between ⫺0.18 and ⫺0.23 V. Thus, as one may judge by the data of Figs. 3A and 3B, the activity


of this compound against P. falciparum in a 48-h test (IC 50 ⫽ 2.8 ␮M (13)) agrees with this range of redox potentials. The data of this work do not challenge the concept of the application of GR inhibitors as antimalarial agents. However, in our opinion, it might be confined to the design of redox-inactive inhibitors, since the prooxidant quinoidal or nitroaromatic compounds would exhibit their antimalarial activity mainly through their redox cycling, oxidation of oxyhemoglobin, and, in specific cases, by the inhibition of the electron transport chain (6). The rate of single-electron reduction of these compounds by flavoenzymes dehydrogenases-electrontransferases, e.g., ferredoxin:NADP ⫹ reductase, is not sensitive to their particular structure, and is determined mainly by their E 71 values (Fig. 4). In erythrocytes, the oxidation of excess NAD(P)H by quinones and nitroaromatic compounds observed in a present study may be supported by their single-electron reduction by erythrocyte plasma membrane NAD(P)H:dehydrogenases (41), and by NADH: cytochrome b 5 reductase (2, 42). Mitochondrial NADH:fumarate reductase of P. falciparum seems to be another potential candidate for the redox cycling of quinones and aromatic nitrocompounds (43). In general, the rate of Hb-Fe 2⫹O 2 oxidation by nitroaromatic and quinoidal compounds also increases with their redox potential (Fig. 4). The enhanced antiplasmodial activity of methylene blue (Figs. 3A and 3B) may also be caused by its high rate of nonenzymatic oxidation of NAD(P)H. The data of our work imply that antimalarial agents among quinoidal and nitroaromatic compounds may be selected according to their E 71 values and electrostatic charge. This approach may also be useful in the design and exploitation of trypanocidal quinones and nitroaromatics. Although these groups of compounds inhibit antioxidant flavoenzyme trypanothione reductase or/and act as its “subversive substrates” (21–25, 44), there exists evidence that trypanocidal activity of heterocyclic and 2-hydroxynaphthoquinones increases with an increase in their redox potential (45). However, one should take into account the increased risk in toxicity to mammalian cells and other adverse effects, e.g., methemoglobinemia and hemolitic anemia, associated with an increase in redox potential of redox cycling therapeutic agents (4, 9 –12). ACKNOWLEDGMENTS This work was supported in part by NATO Linkage Grant HTECH.LG 972012. We thank Dr. Katja Becker for her generous gift of human erythrocyte GR; Prof. Carlos Gomez-Moreno and Dr. Marta Martinez-Julvez for their generous gift of FNR; Drs. Rene Bensasson (Museum National d’Histoire Naturelle, Paris) and Elisabeth Davioud-Charvet (Institut Pasteur, Lille) for reading the manuscript and

205

their fruitful discussion; and Advanced Chemistry Development Inc. for their generous gift of ACD log P software.

REFERENCES 1. Ginsburg, H., and Atamna, H. (1994) Parasite 1, 5–13. 2. Vennerstrom, L. J., and Eaton, J. W. (1988) J. Med. Chem. 31, 1269 –1277. 3. Docampo, R. (1990) Chem.-Biol. Interact. 73, 1–27. 4. Lopez-Shirley, K., Zhang, F., Gosser, D., Scott, M., and Meshnick, S. R. (1994) J. Lab. Clin. Med. 123, 126 –130. 5. Schirmer, R. H., Mu¨ller, J. G., and Krauth-Siegel, R. L. (1995) Angew. Chem. Int. Ed. Engl. 34, 141–154. 6. Fry, M., and Pudney, M. (1992) Biochem. Pharmacol. 43, 1545– 1553. 7. Naisbitt, D. J., Williams, D. P., O’Neil, P. M., Maggs, J. L., Willock, D. J., Pirmohamed, M., and Park, B. H. (1998) Chem. Res. Toxicol. 11, 1568 –1595. 8. Vasquez-Vivar, J., and Augusto, A. (1992) J. Biol. Chem. 267, 6848 – 6854. 9. Adams, G. E., Clarke, E. D., Jacobs, R. S., Stratford, I. J., Wallace, R. G., Wardman, P., and Watts, M. E. (1976) Biochem. Biophys. Res. Commun. 72, 824 – 829. 10. Guissani, A., Henry, Y., Lougmani, N., and Hickel, B. (1990) Free Radicals Biol. Med. 8, 173–189. 11. O’Brien, P. J., Wong, W. C., Silva, J., and Khan, S. (1990) Xenobiotica 20, 945–955. 12. Misˇkiniene˙, V., Sergediene˙, E., Nemeikaite˙, A., Segura-Aguilar, ˇ e˙nas, N. (1999) Cancer Lett. 146, 217–222. J., and C 13. Becker, K., Christopherson, R. I., Cowden, W. B., Hunt, N. H., and Schirmer, R. H. (1990) Biochem. Pharmacol. 39, 59 – 65. 14. Scho¨nleben-Janas, A., Kirsch, P., Mittl, P. R. E., Schirmer, R. H., and Krauth-Siegel, L. (1996) J. Med. Chem. 39, 1549 –1554. 15. Becker, K., Kanzok, S., Iozef, R., Tu¨rbachova, I., and Schirmer, R. H. (1999) in Flavins and Flavoproteins 1999. Proceedings of the Thirteenth International Symposium (Ghisla, S., Kroneck, P., Macheroux, P., and Sund, H., Eds.), pp. 853– 856, Rudolf Weber, Berlin. 16. Fa¨rber, P. M., Arscott, L. D., Williams, C. H., Jr., Becker, K., and Schirmer, H. (1998) FEBS Lett. 422, 311–314. 17. Bo¨hme, C. C., Arscott, L. D., Becker, K., Schirmer, R. H., and Williams, C. H., Jr. (2000) J. Biol. Chem. 275, 37317–37323. 18. Grinblat, L., Sreider, C. M., and Stoppani, A. O. M. (1989) Biochem. Pharmacol. 38, 767–772. ˇ e˙nas, N. K., Bironaite˙, D. A., Kulys, J. J., and Sukhova, N. M. 19. C (1991) Biochim. Biophys. Acta 1073, 195–199. ˇ e˙nas, N. K., Kulys, J. J., Medentsev, A. G., and 20. Bironaite˙, D. A., C Akimenko, V. K. (1991) Z. Naturforsch. 46c, 966 –968. ˇ e˙nas, N., Bironaite˙, D., Dicˇkancaite˙, E., Anusevicˇius, Zˇ., Sˇar21. C lauskas, J., and Blanchard, J. S. (1994) Biochem. Biophys. Res. Commun. 204, 224 –229. 22. Blumenstiel, K., Scho¨neck, R., Yardley, V., Croft, S. L., and Krauth-Siegel, R. L. (1999) Biochem. Pharmacol. 58, 1791–1799. 23. Henderson, G. B., Ulrich, P., Fairlamb, A. H., Rosenberg, I., Pereira, M., Sela, M., and Cerami, A. (1988) Proc. Natl. Acad. Sci. USA 85, 5374 –5378. 24. Jockers-Scheru¨bl, M. C., Schirmer, R. H., and Krauth-Siegel, R. L. (1989) Eur. J. Biochem. 180, 267–272. 25. Salmon-Chemin, L., Lemaire, A., de Freitas, S., Deprez, B., Sergheraert, C., and Davioud-Charvet, E. (2000) Bioorg. Med. Chem. Lett. 10, 631– 635.

206

GRELLIER ET AL.

26. Nordhoff, A., Bu¨cheler, V. S., Verner, D., and Schirmer, R. H. (1993) Biochemistry 32, 4060 – 4066. 27. Anusevicˇius, Zˇ., Martinez-Julvez, M., Genzor, C. G., Nivinskas, ˇ e˙nas, N. (1997) Biochim. Biophys. H., Gomez-Moreno, C., and C Acta 1320, 247–255. 28. Lukevics, E., and Demicheva, L. (1993) Chem. Heterocycl. Comp. (Riga) 3, 291–321 (in Russian). 29. Winterbourn, C. (1985) in CRC Handbook of Methods for Oxygen Radical Research (Greenwald, R. A., Ed.), pp. 137–141, CRC Press, Boca Raton, FL. 30. Favaudon, V., Tourbez, G., Houee-Levin, C., and Lhoste, I. M. (1990) Biochemistry 29, 10978 –10989. 31. Wardman, P., and Clarke, E. D. (1976) J. Chem. Soc. Faraday Trans. I, 69, 1377–1390. 32. Trager, W., and Jensen, J. B. (1976) Science 193, 673– 677. 33. Desjardins, R. E., Canfield, C. J., Haynes, J. D., and Chulay, J. D. (1979) Antimicrob. Agents Chemother. 16, 710 –718. 34. Karplus, P. A., Pai, E. F., and Schulz, G. E. (1989) Eur. J. Biochem. 178, 693–703. 35. Wardman, P. (1989) J. Phys. Chem. Ref. Data 18, 1637–1755.

36. Hay, D. W., Martin, S. A., Ray, S., and Lichtin, N. N. (1981) J. Phys. Chem. 85, 1474 –1479. ˇ e˙nas, N. 37. Misˇkiniene˙, V., Sˇarlauskas, J., Jacquot, J.-P., and C (1998) Biochim. Biophys. Acta 1366, 275–283. 38. Bates, D. A., and Winterbourn, C. C. (1982) Biochem. J. 203, 155–160. ˇ e˙nas, N., and O ¨ llinger, K. (1994) Arch. Biochem. Biophys. 315, 39. C 170 –176. 40. Marva, E., Chevion, M., Golenser, J. (1991) Free Radicals Res. Commun. 12–13, 137–146. 41. Peskin, A. V., and Bartosz, G. (1987) FEBS Lett. 219, 212– 214. 42. Shirabe, K., Fujimoto, Y., Yubisui, T., and Takeshita, M. (1994) J. Biol. Chem. 269, 2952–2957. 43. Fry, M., and Beesley, J. E. (1991) Parasitology 102, 17–26. ˇ e˙nas, N. K., Arscott, D., Williams, C. H., Jr., and Blanchard, 44. C J. S. (1994) Biochemistry 33, 2509 –2515. 45. Goulart, M. O. F., Zani, C. L., Tonholo, J., Freitas, L. R., de Abreu, F. C., Oliveira, A. B., Raslan, D. S., Starling, S., and Chiari, E. (1997) Bioorg. Med. Chem. Lett. 7, 2043–2048.