Pharmacomodulation of the Antimalarial Plasmodione: Synthesis of Biaryl- and N-Arylalkylamine Analogues, Antimalarial Activities and Physicochemical Properties

With the aim of increasing the structural diversity on the early antimalarial drug plasmodione, an efficient and versatile procedure to prepare a series of biaryl- and N-arylalkylamines as plasmodione analogues is described. Using the naturally occurring and commercially available menadione as starting material, a 2-step sequence using a Kochi-Anderson reaction and subsequent Pd-catalyzed Suzuki-Miyaura coupling was developed to prepare three representative biphenyl derivatives in good yields for antimalarial evaluation. In addition, synthetic methodologies to afford 3-benzylmenadione derivatives bearing a terminal -N(Me)2 or -N(Et)2 in different positions (ortho, meta and para) on the aryl ring of the benzylic chain of plasmodione were investigated through reductive amination was used as the optimal route to prepare these protonable N-arylalkylamine privileged scaffolds. The antimalarial activities were evaluated and discussed in light of their physicochemical properties. Among the newly synthesized compounds, the para-position of the substituent remains the most favourable position on the benzyl chain and the carbamate -NHBoc was found active both in vitro (42 nM versus 29 nM for plasmodione) and in vivo in Plasmodium berghei-infected mice. The measured acido-basic features of these new molecules support the cytosol-food vacuole shuttling properties of non-protonable plasmodione derivatives essential for redox-cycling. These findings may be useful in antimalarial drug optimization.


Introduction
Malaria is caused by four species of protozoan parasites of the genus Plasmodium in tropical and subtropical areas and is curable if treated promptly and adequately. The most dangerous and harmful species, P. falciparum is responsible for malignant malaria with complications such as cerebral malaria or severe anaemia, and is responsible for significant mortality and morbidity including the death of approximately 0.6 million people worldwide by year, mostly African children younger than 5 years [1]. Artemisinin-based Combination Therapy (ACT) is currently the first-line therapy for malaria worldwide; however, a decrease in artemisinin sensitivity, mainly in Southeast Asia, has recently been reported, thus underlining the need for alternative artemisinin-based combinations with novel agents as drug partners [2]. New antimalarials must exhibit pleiotropic mechanisms-of-action or target new metabolic pathways that counterbalance parasite resistance to current drugs. Additionally, there is a continuous demand for new drugs that are not only affordable, safe, available and effective but also easy to synthesize. Our approach is to target the thiol based redox-network of P. falciparum.
Recently, we synthesized a series of 3-benzylmenadione derivatives [3,4], which were identified as potent antimalarial redox-active agents. Among these, the 3-[4-(trifluoromethyl)benzyl]-menadione (called plasmodione, Figure 1) showed the most promising 'early lead' features for drug optimization [5,6]. Mechanistic studies revealed that plasmodione may enter a cascade of redox reactions for drug bioactivation: in its oxidized state, the menadione derivative was proposed to be reduced by the NADPH-dependent glutathione reductases of the Plasmodium-parasitized erythrocytes; in its reduced state, the hydronaphthoquinone was shown to transfer 1e − to methemoglobin (Fe III ) (metHb), the major nutrient of the parasite [3,4,7]. By redox-cycling NADPH and hemoglobin, it was postulated that the increased oxidative stress leads to the death of the parasite, but only in parasitized erythrocytes.
In a previous study, it was found that the mitochondrial electron transport (mETC) chain of P. falciparum blood stages is dispensable for the antimalarial activities of plasmodione and methylene blue in contrast to the 2-hydroxy-naphthoquinone derivative atovaquone [5]. The latter drug is given as a highly efficacious fixed-dose combination (atovaquone-proguanil marketed as Malarone ® ) both for prophylaxis and treatment of multidrug-resistant falciparum malaria. Atovaquone inhibits complex III of the mETC system, inhibits the cell respiratory mechanism and depolarizes the mitochondrial membrane potential [8]. However, the major function of the mETC in P. falciparum blood stage cultures is to regenerate ubiquinone as the electron acceptor of mitochondrial dihydroorotate dehydrogenase (DHODH) which catalyzes a key step in pyrimidine biosynthesis [9]. Atovaquone has been shown to lose its antimalarial activity in the presence of a yeast gene encoding an alternative cytosolic DHODH (yDHODH) that does not require the mETC as electron acceptor [10]. Transgenic P. falciparum strains with or without yDHODH were established providing an excellent tool for the analysis of potential inhibitors of the mETC. The previous study [5] evaluated redox-active 1,4-naphthoquinones, including plasmodione and methylene blue, and were shown to retain potency against the transgenic P. falciparum strain expressing yDHODH, which had lost sensitivity to atovaquone. Furthermore, chemical modification on the menadione moiety of plasmodione derived analogues were functionalized with the atovaquone chain [11], but these hybrid molecules were found inactive against the atovaquone-sensitive Dd2 strain of P. falciparum because 1,4-naphthoquinones plasmodione and atovaquone do not share the same mode-of-action and redox potential values. In addition, atovaquone and plasmodione were clearly shown to result in different morphological changes during the intraerythrocytic parasitic cycle [5]. In agreement with previous reports [12], the morphologies of atovaquone-treated parasites were very similar to the untreated controls. The growth of plasmodione-treated parasites was retarded with a significant increase in abnormal morphologies, predominantly pyknotic ring stages. In contrast to atovaquone, the observations on morphology are in agreement with plasmodione's pronounced activity against ring stages compared to later stages. Other electrochemical parameters for lawsone versus menadione (2-hydroxy versus 2-methyl-1,4-naphthoquinone, see Figure 1) reported in [11,13] support the distinct redox behaviour and mode-of-action of both 1,4-naphthoquinone antimalarials atovaquone and plasmodione. redox behaviour and mode-of-action of both 1,4-naphthoquinone antimalarials atovaquone and plasmodione. The bioavailability and activity/solubility of the plasmodione series is however poor, hindering further drug development. Hence, investigations on its pharmaco-modulation are necessary to determine whether the 4-trifluoromethyl group can be replaced or whether the substituent on the aryl ring can be moved without sacrificing the antimalarial activity. Based on the literature on the activity of numerous 4-or 8-aminoquinolines or phenothiazines, and as a part of our medicinal chemistry program on 1,4-naphthoquinone scaffolds, the present study investigated the antimalarial activity of several plasmodione derivatives in which the 4′-CF3 group was replaced either by a substituted aryl group or by a N-dialkylamino group. Modifications on the menadione core previously provided various plasmodione analogues with varying antiplasmodial activity [14]. In the present study the substitution at the benzylic chain of plasmodione was explored, in particular by using Pd-catalyzed reactions to prepare biphenyl analogues. Biaryls and their aryl-heteroaryl homologues represent an important class of organic compounds found in natural products that were shown to be useful in various applications, ranging from medicinal chemistry [15] to advanced materials and supramolecular chemistry [16][17][18]. For example, the diphenylmethyl chain of the potent substituted quinolone CK-2-68 was replaced by a pyridylaryl chain in the initial lead compound SL-2-25 without sacrificing its antimalarial activity [19,20].
Herein, biaryl-and N-arylalkylamine analogues of plasmodione were synthesized to investigate their antimalarial potential in comparison to plasmodione. For N-substitution, we incorporated N-dimethyl or N-diethyl moieties as found in the approved drugs chloroquine (CQ) [21,22] and related 4-or 8-aminoquinolines (e.g., ferroquine [23,24] or pamaquine [25,26]) and Methylene Blue (MB) [27][28][29] (Figure 2) to learn the structural requirements for antimalarial activity of optimized plasmodione derivatives. We further evaluated the impact of the free amino and N-dialkyl substituents on the physicochemical properties of derivatives 3, 7, 10, 11 and 17-19 in comparison to those of two key reference drugs MB and CQ since pKa values are important for antimalarials that mainly interact with Fe III species. The bioavailability and activity/solubility of the plasmodione series is however poor, hindering further drug development. Hence, investigations on its pharmaco-modulation are necessary to determine whether the 4-trifluoromethyl group can be replaced or whether the substituent on the aryl ring can be moved without sacrificing the antimalarial activity. Based on the literature on the activity of numerous.
4-or 8-aminoquinolines or phenothiazines, and as a part of our medicinal chemistry program on 1,4-naphthoquinone scaffolds, the present study investigated the antimalarial activity of several plasmodione derivatives in which the 4 -CF 3 group was replaced either by a substituted aryl group or by a N-dialkylamino group. Modifications on the menadione core previously provided various plasmodione analogues with varying antiplasmodial activity [14]. In the present study the substitution at the benzylic chain of plasmodione was explored, in particular by using Pd-catalyzed reactions to prepare biphenyl analogues. Biaryls and their aryl-heteroaryl homologues represent an important class of organic compounds found in natural products that were shown to be useful in various applications, ranging from medicinal chemistry [15] to advanced materials and supramolecular chemistry [16][17][18]. For example, the diphenylmethyl chain of the potent substituted quinolone CK-2-68 was replaced by a pyridylaryl chain in the initial lead compound SL-2-25 without sacrificing its antimalarial activity [19,20].
Herein, biaryl-and N-arylalkylamine analogues of plasmodione were synthesized to investigate their antimalarial potential in comparison to plasmodione. For N-substitution, we incorporated N-dimethyl or N-diethyl moieties as found in the approved drugs chloroquine (CQ) [21,22] and related 4-or 8-aminoquinolines (e.g., ferroquine [23,24] or pamaquine [25,26]) and Methylene Blue (MB) [27][28][29] (Figure 2) to learn the structural requirements for antimalarial activity of optimized plasmodione derivatives. We further evaluated the impact of the free amino and N-dialkyl substituents on the physicochemical properties of derivatives 3, 7, 10, 11 and 17-19 in comparison to those of two key reference drugs MB and CQ since pK a values are important for antimalarials that mainly interact with Fe III species. MB, the first synthetic antimalarial dye discovered by Paul Ehrlich in 1891, has recently been rediscovered for its promising activity in drug combinations for the control of transmission of P. falciparum malaria in endemic regions [30] and against Plasmodium vivax [31]. MB is reduced by various NADPH-dependent disulfide reductases, including glutathione reductases from the cytosols of Plasmodium-infected erythrocytes [32]. Malarial parasites normally digest the host hemoglobin in the metHb form in a specialized acidic compartment, called food vacuole, mainly at the intraerythrocytic trophozoite stage. The blue dye is known to cycle between the oxidized form and its electronreduced leucomethylene blue (LMB) form that is involved in intricate redox processes, in particular conversion of metHb to oxyhemoglobin (oxyHb) [29]. As oxyHb is not digested by the parasitic proteases, MB has been proposed to exert its antimalarial action by depriving the parasite from its major nutrient metHb. With its phenothiazine structure, MB displays redox cycling activity able to generate bound (hemoglobin or heme) Fe II from Fe III species, likely responsible for its potent activity in vivo. Upon reduction and different protonation states, MB has the ability to change its structure and display different lipophilicities [29]. Due to its very low pKa (pKa = 0), MB does not accumulate in acidic vesicles but may shuttle in and out the cell compartments depending on its redox state. The final result of MB-catalyzed redox-cycling is an arrest of parasite development [5].
In contrast, the broadly used CQ is a diprotic weak base (pKa = 8.4 and 10.6), which accumulates in the food vacuole and binds to Fe III -heme leading to a stable 1:1 π-π complex (KD ~ 2-15 µM at pH 7.5) [33]. Intraerythrocytic parasites detoxify the exogenous deleterious Fe III -heme by-product (i.e., predominantly existing as a dimeric π-π heme complex) of hemoglobin digestion through biomineralization to an insoluble and much less active crystal, named hemozoin. In the presence of CQ, the final result is the elevation of toxic free heme and inhibition of hemozoin formation. Free heme and heme CQ complexes are therefore thought to kill parasites by inducing severe oxidative stress [34], which successively leads to peroxidation of the parasite membrane lipids, DNA damage, protein oxidation and ultimately death of the parasite.
In continuation of our structure-activity relationships (SAR) studies on the plasmodione series (Scheme 1), the plasmodione bromine derivatives 25a-c were prepared from menadione in moderate to excellent yields (52%-88%) by the described silver-catalyzed decarboxylation reaction between menadione 1 and the corresponding bromophenylacetic acid derivatives 24a-c. MB, the first synthetic antimalarial dye discovered by Paul Ehrlich in 1891, has recently been rediscovered for its promising activity in drug combinations for the control of transmission of P. falciparum malaria in endemic regions [30] and against Plasmodium vivax [31]. MB is reduced by various NADPH-dependent disulfide reductases, including glutathione reductases from the cytosols of Plasmodium-infected erythrocytes [32]. Malarial parasites normally digest the host hemoglobin in the metHb form in a specialized acidic compartment, called food vacuole, mainly at the intraerythrocytic trophozoite stage. The blue dye is known to cycle between the oxidized form and its electron-reduced leucomethylene blue (LMB) form that is involved in intricate redox processes, in particular conversion of metHb to oxyhemoglobin (oxyHb) [29]. As oxyHb is not digested by the parasitic proteases, MB has been proposed to exert its antimalarial action by depriving the parasite from its major nutrient metHb. With its phenothiazine structure, MB displays redox cycling activity able to generate bound (hemoglobin or heme) Fe II from Fe III species, likely responsible for its potent activity in vivo. Upon reduction and different protonation states, MB has the ability to change its structure and display different lipophilicities [29]. Due to its very low pK a (pK a = 0), MB does not accumulate in acidic vesicles but may shuttle in and out the cell compartments depending on its redox state. The final result of MB-catalyzed redox-cycling is an arrest of parasite development [5].
In contrast, the broadly used CQ is a diprotic weak base (pK a = 8.4 and 10.6), which accumulates in the food vacuole and binds to Fe III -heme leading to a stable 1:1 π-π complex (K D~2 -15 µM at pH 7.5) [33]. Intraerythrocytic parasites detoxify the exogenous deleterious Fe III -heme by-product (i.e., predominantly existing as a dimeric π-π heme complex) of hemoglobin digestion through biomineralization to an insoluble and much less active crystal, named hemozoin. In the presence of CQ, the final result is the elevation of toxic free heme and inhibition of hemozoin formation. Free heme and heme CQ complexes are therefore thought to kill parasites by inducing severe oxidative stress [34], which successively leads to peroxidation of the parasite membrane lipids, DNA damage, protein oxidation and ultimately death of the parasite.
In continuation of our structure-activity relationships (SAR) studies on the plasmodione series (Scheme 1), the plasmodione bromine derivatives 25a-c were prepared from menadione in moderate to excellent yields (52%-88%) by the described silver-catalyzed decarboxylation reaction between menadione 1 and the corresponding bromophenylacetic acid derivatives 24a-c. Transition metal catalyzed cross-coupling reactions have revolutionized organic synthesis. Many highly efficient and mild methods have been developed over the last two decades allowing rapid and versatile access to complex molecular structures [42][43][44][45][46][47]. In recent years, the Suzuki-Miyaura reaction in particular has become one of the most important cross-coupling reactions, allowing not only the formation of C(sp 2 )-C(sp 2 ) and C(sp 2 )-C(sp 3 ) but also C(sp 3 )-C(sp 3 ) bonds [44][45][46].
Subsequently, the biaryl menadione derivatives 26-29 (Scheme 2) were successfully synthesized in excellent yields (86%-97%) via a Suzuki-Miyaura coupling reaction. The reaction was performed in dioxane/water (v/v = 4/1) at 80 °C overnight under an atmosphere of argon. Aryl bromide 25c was coupled with the corresponding para-substituted phenylboronic acid in the presence of K2CO3 as base and 4 mol % PdCl2(dppf) as catalyst. K2CO3 was chosen as a mild base because of the low stability of benzylmenadione derivative 25c under basic conditions due to the acidic proton of the CH3 group. Using these reaction conditions, no side product was observed compared to the treatment with stronger bases leading to a rapid decomposition of the starting menadione derivatives. The expected Suzuki-products 26-28 were isolated in excellent yield. This methodology is straightforward and provides easy access to structurally diverse 1,4-naphthoquinones with a biphenyl motif, allowing to study the influence of a second phenyl ring in plasmodione structure on its antimalarial activity. Regarding the 2nd series of N-arylalkylamines, we first prepared the N-dialkylamine analogues of 3-benzylmenadiones ( Figure 2) using the Kochi-Anderson reaction. In earlier investigations to synthesize compound 3 (Scheme 3), the reported conditions used in the silver-catalyzed decarboxylation reaction of menadione 1 with 4-(dimethylamino)phenylacetic acid 2 had proven to be inappropriate and not compatible, leading mainly to degradation of the starting materials because the N-substituent becomes oxidized in the reaction by Ag(II) leading to the formation of Ag(I) and thereby destroying the competent catalyst Ag(II).
It was deemed of interest to explore a new pathway for this purpose. We attempted to achieve the synthesis of para-N-substituted-3-benzylmenadione derivatives 3 via a N-alkylation procedure using two synthetically N-protected routes through: (i) the N-tert-butyloxycarbonyl (N-Boc) and (ii) the N-acetyl protection strategies (Scheme 4). Transition metal catalyzed cross-coupling reactions have revolutionized organic synthesis. Many highly efficient and mild methods have been developed over the last two decades allowing rapid and versatile access to complex molecular structures [42][43][44][45][46][47]. In recent years, the Suzuki-Miyaura reaction in particular has become one of the most important cross-coupling reactions, allowing not only the formation of C(sp 2 )-C(sp 2 ) and C(sp 2 )-C(sp 3 ) but also C(sp 3 )-C(sp 3 ) bonds [44][45][46].
Subsequently, the biaryl menadione derivatives 26-29 (Scheme 2) were successfully synthesized in excellent yields (86%-97%) via a Suzuki-Miyaura coupling reaction. The reaction was performed in dioxane/water (v/v = 4/1) at 80 • C overnight under an atmosphere of argon. Aryl bromide 25c was coupled with the corresponding para-substituted phenylboronic acid in the presence of K 2 CO 3 as base and 4 mol % PdCl 2 (dppf) as catalyst. K 2 CO 3 was chosen as a mild base because of the low stability of benzylmenadione derivative 25c under basic conditions due to the acidic proton of the CH 3 group. Using these reaction conditions, no side product was observed compared to the treatment with stronger bases leading to a rapid decomposition of the starting menadione derivatives. The expected Suzuki-products 26-28 were isolated in excellent yield. This methodology is straightforward and provides easy access to structurally diverse 1,4-naphthoquinones with a biphenyl motif, allowing to study the influence of a second phenyl ring in plasmodione structure on its antimalarial activity. Transition metal catalyzed cross-coupling reactions have revolutionized organic synthesis. Many highly efficient and mild methods have been developed over the last two decades allowing rapid and versatile access to complex molecular structures [42][43][44][45][46][47]. In recent years, the Suzuki-Miyaura reaction in particular has become one of the most important cross-coupling reactions, allowing not only the formation of C(sp 2 )-C(sp 2 ) and C(sp 2 )-C(sp 3 ) but also C(sp 3 )-C(sp 3 ) bonds [44][45][46].
Subsequently, the biaryl menadione derivatives 26-29 (Scheme 2) were successfully synthesized in excellent yields (86%-97%) via a Suzuki-Miyaura coupling reaction. The reaction was performed in dioxane/water (v/v = 4/1) at 80 °C overnight under an atmosphere of argon. Aryl bromide 25c was coupled with the corresponding para-substituted phenylboronic acid in the presence of K2CO3 as base and 4 mol % PdCl2(dppf) as catalyst. K2CO3 was chosen as a mild base because of the low stability of benzylmenadione derivative 25c under basic conditions due to the acidic proton of the CH3 group. Using these reaction conditions, no side product was observed compared to the treatment with stronger bases leading to a rapid decomposition of the starting menadione derivatives. The expected Suzuki-products 26-28 were isolated in excellent yield. This methodology is straightforward and provides easy access to structurally diverse 1,4-naphthoquinones with a biphenyl motif, allowing to study the influence of a second phenyl ring in plasmodione structure on its antimalarial activity.
Following the N-Boc route (i), the commercial 4-aminophenylacetic acid (4) was treated with Boc2O in the presence of Na2CO3 as base to give the corresponding N-Boc-protected phenylacetic acid 5 in 94% isolated yield (route I, Scheme 4). The 3-benzylmenadione intermediate 6 was synthesized in low yield (22%) by one-step Kochi-Anderson reaction between the phenylacetic acid derivative 5 and menadione (1) in the presence of Ag(I) as catalyst and ammonium peroxodisulfate. After removal of the N-Boc protecting group in the presence of trifluoroacetic acid (TFA), compound 7 was isolated in 84% yield as the precursor for N,N-dimethylation of the free amine group in the next step. Finally, the desired product 3 was obtained in 33% yield using Ph3P/DDQ [48] in methanol as an efficient reagent system for the immediate and selective N-alkylation of aromatic amines. However, to improve the synthesis and overall yields of the desired product 3, we studied the effect of the N-protecting group by replacement of the Boc group by an acetyl group. Indeed, by following the N-acetyl route (II), the N-acetyl-protected phenylacetic acid 8 was obtained quantitatively by an acetylation reaction between the free-amine 4 and acetic acid under microwave irradiation conditions [49]. Next, the Kochi-Anderson reaction between the phenylacetic acid derivative 8 and menadione (1) gave access to the desired compound 9 in good yield (54%). The deprotection reaction of the acetyl group was performed in the presence of hydrochloric acid (HCl) under microwave irradiation conditions, affording the desired compound 7 in 76% yield [50]. As shown in Scheme 4, the expected compound N,N-dimethyl-3-benzylmenadione 3 was synthesized in good yield (59%) by one-step reductive amination reaction between the intermediate 7 and formaldehyde in the presence of acetic acid and sodium cyanoborohydride as reducing reagent (59% with NaBH3CN/AcOH versus 39% with DDQ/PPh3) [51]. Use of acetaldehyde instead of formaldehyde under the same reduction reaction conditions (NaBH3CN/AcOH) gave access to a separable mixture of para-N-ethyl-3benzylmenadione (10, 17%) and para-N,N-diethyl-3-benzylmenadione (11, 36%) (Scheme 4).
We then applied the same N-alkylation procedures described in Scheme 4 (N-Boc and -acetyl protection) for the synthesis of meta-N-substituted-3-benzylmenadione derivatives 18 and 19 (Scheme 5). Following route I and starting from the commercial 3-aminophenylacetic acid (12), the meta-N-Boc-benzylmenadione intermediate 14 was obtained in two steps in low yield (7%) by protection of the free benzylamine group with a Boc group followed by the Kochi-Anderson reaction. As the desired compound 14 was obtained in low yield, we thought that the Boc protecting group should be changed to an acetyl group in order to increase the efficacy of this methodology when following route II.
Indeed, the meta-N-acetyl-benzylmenadione intermediate 16 was first isolated in moderate yields (47%) in two steps by N-acetyl protection followed by a subsequent Kochi-Anderson reaction. Then, an HCl-promoted acetyl removal followed by a reductive amination reaction of free amine 17 with formaldehyde and acetaldehyde respectively, resulted in the formation of the desired meta-N,N-dimethyl-3-benzylmenadione (18, 31%) and meta-N,N-diethyl-3-benzylmenadione (19, 22%) respectively (route II, Scheme 5). We then applied the same N-alkylation procedures described in Scheme 4 (N-Boc and -acetyl protection) for the synthesis of meta-N-substituted-3-benzylmenadione derivatives 18 and 19 (Scheme 5). Following route I and starting from the commercial 3-aminophenylacetic acid (12), the meta-N-Bocbenzylmenadione intermediate 14 was obtained in two steps in low yield (7%) by protection of the free benzylamine group with a Boc group followed by the Kochi-Anderson reaction. As the desired compound 14 was obtained in low yield, we thought that the Boc protecting group should be changed to an acetyl group in order to increase the efficacy of this methodology when following route II.
Indeed, the meta-N-acetyl-benzylmenadione intermediate 16 was first isolated in moderate yields (47%) in two steps by N-acetyl protection followed by a subsequent Kochi-Anderson reaction. Then, an HCl-promoted acetyl removal followed by a reductive amination reaction of free amine 17 with formaldehyde and acetaldehyde respectively, resulted in the formation of the desired meta-N,N-dimethyl-3-benzylmenadione (18, 31%) and meta-N,N-diethyl-3-benzylmenadione (19, 22%) respectively (route II, Scheme 5). Next, we synthesized the ortho-N-substituted-3-benzylmenadione derivatives by applying the N-acetyl protection as depicted in Scheme 6 (route II), the ortho-N-acetyl-3-benzylmenadione intermediate was isolated in moderate yield (39%) by a sequential opening of the oxoindole [52] ring 20 to aminoacid 21 followed by a N-acetyl protection [53] of the generated free amine 21 and a final Kochi-Anderson reaction. To synthesize compound 23, the optimal heating and microwave irradiation conditions for the N-acetyl deprotection reaction of compound 22 proved to be inappropriate, leading mainly to the degradation of the starting material 22 (Scheme 6).
Next, we synthesized the ortho-N-substituted-3-benzylmenadione derivatives by applying the N-acetyl protection as depicted in Scheme 6 (route II), the ortho-N-acetyl-3-benzylmenadione intermediate was isolated in moderate yield (39%) by a sequential opening of the oxoindole [52] ring 20 to aminoacid 21 followed by a N-acetyl protection [53] of the generated free amine 21 and a final Kochi-Anderson reaction. To synthesize compound 23, the optimal heating and microwave irradiation conditions for the N-acetyl deprotection reaction of compound 22 proved to be inappropriate, leading mainly to the degradation of the starting material 22 (Scheme 6).

In Vitro Antimalarial Activity
All final isolated compounds were tested for their ability to inhibit the in vitro growth of the multiresistant P. falciparum Dd2 strain by determining the concentration required to inhibit parasite development by 50% (IC50 values). The protocols are described in the experimental section. Two protocols to measure the growth inhibition were used: the tritiated hypoxanthine incorporation method and the SYBRgreen method. This was due to the fact that the radioactive assay was abandoned in the laboratory due to cost issues related to radioactive waste disposal. The growth inhibition measured in the malarial parasite assays and the cytotoxicity against human fibroblasts of all the new compounds are listed in Tables 1 and 2. For comparison, data for plasmodione and its bromo analogue described in our previous publications were also included in the tests. All data are mean values from one to four independent experiments, each consisting of eight drug concentrations in duplicate. a In vitro activity against P. falciparum Dd2 strain is presented as mean IC50 values of benzylmenadione derivatives using the tritiated hypoxanthine incorporation method; b Chloroquine and plasmodione were used as standard drugs. The listed IC50 values of chloroquine, plasmodione and its bromo analogue 25c (named benzylMD 1c and 1a in [3]) are previously reported values.

In Vitro Antimalarial Activity
All final isolated compounds were tested for their ability to inhibit the in vitro growth of the multiresistant P. falciparum Dd2 strain by determining the concentration required to inhibit parasite development by 50% (IC 50 values). The protocols are described in the experimental section. Two protocols to measure the growth inhibition were used: the tritiated hypoxanthine incorporation method and the SYBR green method. This was due to the fact that the radioactive assay was abandoned in the laboratory due to cost issues related to radioactive waste disposal. The growth inhibition measured in the malarial parasite assays and the cytotoxicity against human fibroblasts of all the new compounds are listed in Tables 1 and 2. For comparison, data for plasmodione and its bromo analogue described in our previous publications were also included in the tests. All data are mean values from one to four independent experiments, each consisting of eight drug concentrations in duplicate.

In Vitro Antimalarial Activity
All final isolated compounds were tested for their ability to inhibit the in vitro growth of the multiresistant P. falciparum Dd2 strain by determining the concentration required to inhibit parasite development by 50% (IC50 values). The protocols are described in the experimental section. Two protocols to measure the growth inhibition were used: the tritiated hypoxanthine incorporation method and the SYBRgreen method. This was due to the fact that the radioactive assay was abandoned in the laboratory due to cost issues related to radioactive waste disposal. The growth inhibition measured in the malarial parasite assays and the cytotoxicity against human fibroblasts of all the new compounds are listed in Tables 1 and 2. For comparison, data for plasmodione and its bromo analogue described in our previous publications were also included in the tests. All data are mean values from one to four independent experiments, each consisting of eight drug concentrations in duplicate.

In Vitro Antimalarial Activity
All final isolated compounds were tested for their ability to inhibit the in vitro growth of the multiresistant P. falciparum Dd2 strain by determining the concentration required to inhibit parasite development by 50% (IC50 values). The protocols are described in the experimental section. Two protocols to measure the growth inhibition were used: the tritiated hypoxanthine incorporation method and the SYBRgreen method. This was due to the fact that the radioactive assay was abandoned in the laboratory due to cost issues related to radioactive waste disposal. The growth inhibition measured in the malarial parasite assays and the cytotoxicity against human fibroblasts of all the new compounds are listed in Tables 1 and 2. For comparison, data for plasmodione and its bromo analogue described in our previous publications were also included in the tests. All data are mean values from one to four independent experiments, each consisting of eight drug concentrations in duplicate.

In Vitro Antimalarial Activity
All final isolated compounds were tested for their ability to inhibit the in vitro growth of the multiresistant P. falciparum Dd2 strain by determining the concentration required to inhibit parasite development by 50% (IC50 values). The protocols are described in the experimental section. Two protocols to measure the growth inhibition were used: the tritiated hypoxanthine incorporation method and the SYBRgreen method. This was due to the fact that the radioactive assay was abandoned in the laboratory due to cost issues related to radioactive waste disposal. The growth inhibition measured in the malarial parasite assays and the cytotoxicity against human fibroblasts of all the new compounds are listed in Tables 1 and 2. For comparison, data for plasmodione and its bromo analogue described in our previous publications were also included in the tests. All data are mean values from one to four independent experiments, each consisting of eight drug concentrations in duplicate.

In Vitro Antimalarial Activity
All final isolated compounds were tested for their ability to inhibit the in vitro growth of the multiresistant P. falciparum Dd2 strain by determining the concentration required to inhibit parasite development by 50% (IC50 values). The protocols are described in the experimental section. Two protocols to measure the growth inhibition were used: the tritiated hypoxanthine incorporation method and the SYBRgreen method. This was due to the fact that the radioactive assay was abandoned in the laboratory due to cost issues related to radioactive waste disposal. The growth inhibition measured in the malarial parasite assays and the cytotoxicity against human fibroblasts of all the new compounds are listed in Tables 1 and 2. For comparison, data for plasmodione and its bromo analogue described in our previous publications were also included in the tests. All data are mean values from one to four independent experiments, each consisting of eight drug concentrations in duplicate.

In Vitro Antimalarial Activity
All final isolated compounds were tested for their ability to inhibit the in vitro growth of the multiresistant P. falciparum Dd2 strain by determining the concentration required to inhibit parasite development by 50% (IC50 values). The protocols are described in the experimental section. Two protocols to measure the growth inhibition were used: the tritiated hypoxanthine incorporation method and the SYBRgreen method. This was due to the fact that the radioactive assay was abandoned in the laboratory due to cost issues related to radioactive waste disposal. The growth inhibition measured in the malarial parasite assays and the cytotoxicity against human fibroblasts of all the new compounds are listed in Tables 1 and 2. For comparison, data for plasmodione and its bromo analogue described in our previous publications were also included in the tests. All data are mean values from one to four independent experiments, each consisting of eight drug concentrations in duplicate.

In Vitro Antimalarial Activity
All final isolated compounds were tested for their ability to inhibit the in vitro growth of the multiresistant P. falciparum Dd2 strain by determining the concentration required to inhibit parasite development by 50% (IC50 values). The protocols are described in the experimental section. Two protocols to measure the growth inhibition were used: the tritiated hypoxanthine incorporation method and the SYBRgreen method. This was due to the fact that the radioactive assay was abandoned in the laboratory due to cost issues related to radioactive waste disposal. The growth inhibition measured in the malarial parasite assays and the cytotoxicity against human fibroblasts of all the new compounds are listed in Tables 1 and 2. For comparison, data for plasmodione and its bromo analogue described in our previous publications were also included in the tests. All data are mean values from one to four independent experiments, each consisting of eight drug concentrations in duplicate.

In Vitro Antimalarial Activity
All final isolated compounds were tested for their ability to inhibit the in vitro growth of the multiresistant P. falciparum Dd2 strain by determining the concentration required to inhibit parasite development by 50% (IC50 values). The protocols are described in the experimental section. Two protocols to measure the growth inhibition were used: the tritiated hypoxanthine incorporation method and the SYBRgreen method. This was due to the fact that the radioactive assay was abandoned in the laboratory due to cost issues related to radioactive waste disposal. The growth inhibition measured in the malarial parasite assays and the cytotoxicity against human fibroblasts of all the new compounds are listed in Tables 1 and 2. For comparison, data for plasmodione and its bromo analogue described in our previous publications were also included in the tests. All data are mean values from one to four independent experiments, each consisting of eight drug concentrations in duplicate.             a In vitro activity against P. falciparum Dd2 is presented as mean IC50 values ± standard deviation (SD) determined from (n) independent growth inhibition assays in triplicate using the SYBRgreen ® technique; b Chloroquine was used as reference drug. IC50 values of chloroquine, plasmodione and its bromo analogue 25c (named benzylMD 1c and 1a in [3]) are previously reported values. * Precipitates in DMSO.
99 ± 19 (6) 51.5 b a In vitro activity against P. falciparum Dd2 is presented as mean IC 50 values ± standard deviation (SD) determined from (n) independent growth inhibition assays in triplicate using the SYBR green ® technique; b Chloroquine was used as reference drug. IC 50 values of chloroquine, plasmodione and its bromo analogue 25c (named benzylMD 1c and 1a in [3]) are previously reported values. * Precipitates in DMSO.
To explore the effect of modifications in plasmodione by replacing the substituted aryl by biaryl moieties bearing three distinct functionalities on antiplasmodial activity and cytotoxicity, we tested compounds 6, 25c, 26-28 in the hypoxanthine incorporation test (Table 1). Whatever the nature of the substituent (electron-donating t-Bu, electron-withdrawing NO 2 , protonable substituent NMe 2 ), the potency was lower than that of plasmodione itself. The most active compound was the biaryl functionalized by a p-NMe 2 exhibiting a 10-fold higher IC 50 value (229 nM versus 29 nM for plasmodione).
This last result oriented the work towards the synthesis and testing of plasmodione analogues in which the 4 -CF 3 was replaced by 4 -NMe 2 , 4 -NEt 2 , 3 -NMe 2 , or a 3 -NEt 2 . All intermediates (bearing NH 2 , NHBoc, NHAc, Br groups) were tested in the SYBR green drug assay (Table 2). Interestingly, all benzylmenadione analogues bearing a N-dialkyl or a N-monoalkyl group in 4 or 3 were >10-fold less active than plasmodione itself, suggesting that the lysosomotropic compound concentration in acidic compartments was not a favourable parameter to enhance the antiplasmodial activity of redox-cyclers. In contrast, the non-protonable bromo-or the NH-protected amine analogues were much more potent compared to the N-dialkyl derivatives. The paraand ortho-substitutions were also found the most favourable in the bromo series (82 or 86 nM versus 127 nM for paraor ortho, versus -meta, compared to 58 nM for plasmodione). Unexpectedly, the 4 -NHBoc benzylmenadione was found very active, with an IC 50 value of 109 nM.

Cytotoxicity of Plasmodione Derivatives in Human Cell Assays
All compounds were tested for cytotoxicity on human lung MRC-5 fibroblasts using the Alamar Blue assay (Tables 1 and 2). Most of the benzylmenadione derivatives exhibited a low cytotoxicity as evidenced by the high IC 50 values above 32 µM.

In Vivo Antimalarial Activity
The in vitro most active compound 6 among the newly synthesized derivatives was subsequently tested in the P. berghei mouse model using the chloroquine (CQ)-susceptible ANKA strain (Table 3). For comparative purposes, CQ was also included and dosed at 10 mg/kg intraperitoneally (ip) resulting in a decreased parasitemia by 86.3% and by 76.8% at days 4 and 7. Compound 6 showed significant activity following a 5-day treatment at 50 mg/kg ip resulting in decreased parasitaemia by 75.5% and 81.8% at days 4 and 7. For comparison, the previously described 4-bromo derivative 25c showed 35.8% reduction of parasitemia in P. berghei strain ANKA-infected CD1 mice at day 4 (compound 1a in Table 4 of [3]).

Acid-Base Properties
The acid-base properties of the relevant plasmodione derivatives were evaluated by absorption spectrophotometric titrations versus pH titrations. The derivatives are characterized by a set of three main absorptions centered at~340 nm,~270 nm and~250 nm, respectively. The electronic transitions at low energies (>300 nm) were attributed to π-π* transitions centered on the 1,4-napthoquinone chromophore (benzoquinoidal structure), while those observed at higher energies (<300 nm) can be ascribed to π-π* benzene or benzyl subunits. Figure 3 depicts the spectral variations measured for compound 3 as a function of pH. The rise of pH does not affect the absorption of higher energies, while those centered below 300 nm are significantly affected by the acidity of the medium. These variations are in agreement with (de)protonation of the ammonium group substituted on the benzylic subunit. The absorption spectrophotometric titrations versus pH for the other plasmodione derivatives are provided as Supplementary Materials (Figures S1-S4).  The statistical processing of the absorption and potentiometric datasets allowed the evaluation of a single protonation constant related to the benzylic ionizable site (i.e., benzylamine and its N-methylated analogues) whatever system considered ( Table 4). As far as the plasmodione derivatives bearing diethyl tertiary amines are concerned (i.e., 10 and 19), the pKa values could not be accurately evaluated due to the precipitation at pH ~5 under our experimental conditions (H2O/DMSO, 91/1 v/v). This is likely due to the higher hydrophobic character brought by the ethyl substituent. The pKa values ranged from 4.3 (compound 17) to 5.07 (compound 11) and are in excellent agreement with the values determined for adequate models such as aniline (pKa = 4.72 [54] in 0.1 M KCl versus compound 7 pKa = 4.74 ± 0.05 or compound 17 pKa = 4.30 ± 0.02) or N-phenyldiethanolamine (pKa = 4.13 [55] in 0.4 M KCl versus compound 3 pKa = 4.87 ± 0.03 or compound 18 pKa = 4.60 ± 0.04). This feature emphasizes the absence of any electronic effect of the menadione core on the acido-basic properties of the benzylamine subunit due to the presence of the methylenic moiety which can be regarded as a poor electronic relay. Taken as an example, Figure 4 depicts the distribution diagrams of the protonated species of compound 3 clearly showing that below pH 5 (anticipated pH of the food vacuole of P. falciparum) the plasmodione analogues bearing an amino substitution of a benzyl position exist predominantly in their ammonium form. Table 4 provides the ratios of the protonated species at pH 5 and 7.4 clearly evidencing accumulation in the food vacuole due to pH gradient.  The statistical processing of the absorption and potentiometric datasets allowed the evaluation of a single protonation constant related to the benzylic ionizable site (i.e., benzylamine and its N-methylated analogues) whatever system considered ( Table 4). As far as the plasmodione derivatives bearing diethyl tertiary amines are concerned (i.e., 10 and 19), the pK a values could not be accurately evaluated due to the precipitation at pH~5 under our experimental conditions (H 2 O/DMSO, 91/1 v/v). This is likely due to the higher hydrophobic character brought by the ethyl substituent. The pK a values ranged from 4.3 (compound 17) to 5.07 (compound 11) and are in excellent agreement with the values determined for adequate models such as aniline (pK a = 4.72 [54] in 0.1 M KCl versus compound 7 pK a = 4.74 ± 0.05 or compound 17 pK a = 4.30 ± 0.02) or N-phenyldiethanolamine (pK a = 4.13 [55] in 0.4 M KCl versus compound 3 pK a = 4.87 ± 0.03 or compound 18 pK a = 4.60 ± 0.04). This feature emphasizes the absence of any electronic effect of the menadione core on the acido-basic properties of the benzylamine subunit due to the presence of the methylenic moiety which can be regarded as a poor electronic relay. Taken as an example, Figure 4 depicts the distribution diagrams of the protonated species of compound 3 clearly showing that below pH 5 (anticipated pH of the food vacuole of P. falciparum) the plasmodione analogues bearing an amino substitution of a benzyl position exist predominantly in their ammonium form. Table 4 provides the ratios of the protonated species at pH 5 and 7.4 clearly evidencing accumulation in the food vacuole due to pH gradient. The statistical processing of the absorption and potentiometric datasets allowed the evaluation of a single protonation constant related to the benzylic ionizable site (i.e., benzylamine and its N-methylated analogues) whatever system considered ( Table 4). As far as the plasmodione derivatives bearing diethyl tertiary amines are concerned (i.e., 10 and 19), the pKa values could not be accurately evaluated due to the precipitation at pH ~5 under our experimental conditions (H2O/DMSO, 91/1 v/v). This is likely due to the higher hydrophobic character brought by the ethyl substituent. The pKa values ranged from 4.3 (compound 17) to 5.07 (compound 11) and are in excellent agreement with the values determined for adequate models such as aniline (pKa = 4.72 [54] in 0.1 M KCl versus compound 7 pKa = 4.74 ± 0.05 or compound 17 pKa = 4.30 ± 0.02) or N-phenyldiethanolamine (pKa = 4.13 [55] in 0.4 M KCl versus compound 3 pKa = 4.87 ± 0.03 or compound 18 pKa = 4.60 ± 0.04). This feature emphasizes the absence of any electronic effect of the menadione core on the acido-basic properties of the benzylamine subunit due to the presence of the methylenic moiety which can be regarded as a poor electronic relay. Taken as an example, Figure 4 depicts the distribution diagrams of the protonated species of compound 3 clearly showing that below pH 5 (anticipated pH of the food vacuole of P. falciparum) the plasmodione analogues bearing an amino substitution of a benzyl position exist predominantly in their ammonium form. Table 4 provides the ratios of the protonated species at pH 5 and 7.4 clearly evidencing accumulation in the food vacuole due to pH gradient.

Discussion
In an attempt to generate plasmodione analogues with improved activity/solubility profiles, we developed synthetic methodologies to prepare a biaryl series through a 2-step sequence using Kochi-Anderson reaction followed by a Pd-catalyzed Suzuki-Miyaura cross-coupling, and a second 3-benzylmenadione series bearing a terminal -N(Me)2 or -N(Et)2 in different positions (ortho, meta and para) on the aryl ring of the benzylic chain of plasmodione through reductive amination, as the optimal route to prepare these protonable N-arylalkylamines.
From the mechanistic point of view, the benzoylmenadiones were proposed to be the key potential metabolites acting in oxidized form as the most efficient subversive substrates of reduced GR described so far and, in reduced form, to transfer electrons to methemoglobin. The reduced species of benzoylmenadiones are assumed to be transported through Fe III complexation into the food vacuole where the electrons are transferred to oxidants (heme and methemoglobin). Regenerated in their oxidized form, the benzoylmenadiones might be transported into the cytosol where they are reduced by the cytosolic GR (either from human erythrocyte or from the parasite) in a continuous redox cycle ( Figure 5). Evidence for cytosol-food vacuole shuttling of plasmodione during redox-cycling is difficult to prove, however, replacing the 4′-CF3 group in plasmodione by protonable terminal -N(Me)2 or -N(Et)2, the aim was to vectorize the final derivatives to the acidic compartments of the malarial parasite, e.g., the food vacuole, where hemoglobin digestion takes place. By partly losing antimalarial potency, these compounds provide indirect proof for the hypothesized cytosol-food vacuole shuttling of plasmodione during redox-cycling. Worthy of mention is the fact that the redox-active MB, which possesses potent antimalarial activity, shares similar physico-biochemical properties to those found for plasmodione and its active analogues, i.e., non-protonable N-dimethylamino groups at the acidic pH

Discussion
In an attempt to generate plasmodione analogues with improved activity/solubility profiles, we developed synthetic methodologies to prepare a biaryl series through a 2-step sequence using Kochi-Anderson reaction followed by a Pd-catalyzed Suzuki-Miyaura cross-coupling, and a second 3-benzylmenadione series bearing a terminal -N(Me)2 or -N(Et)2 in different positions (ortho, meta and para) on the aryl ring of the benzylic chain of plasmodione through reductive amination, as the optimal route to prepare these protonable N-arylalkylamines.
From the mechanistic point of view, the benzoylmenadiones were proposed to be the key potential metabolites acting in oxidized form as the most efficient subversive substrates of reduced GR described so far and, in reduced form, to transfer electrons to methemoglobin. The reduced species of benzoylmenadiones are assumed to be transported through Fe III complexation into the food vacuole where the electrons are transferred to oxidants (heme and methemoglobin). Regenerated in their oxidized form, the benzoylmenadiones might be transported into the cytosol where they are reduced by the cytosolic GR (either from human erythrocyte or from the parasite) in a continuous redox cycle ( Figure 5). Evidence for cytosol-food vacuole shuttling of plasmodione during redox-cycling is difficult to prove, however, replacing the 4′-CF3 group in plasmodione by protonable terminal -N(Me)2 or -N(Et)2, the aim was to vectorize the final derivatives to the acidic compartments of the malarial parasite, e.g., the food vacuole, where hemoglobin digestion takes place. By partly losing antimalarial potency, these compounds provide indirect proof for the hypothesized cytosol-food vacuole shuttling of plasmodione during redox-cycling. Worthy of mention is the fact that the redox-active MB, which possesses potent antimalarial activity, shares similar physico-biochemical properties to those found for plasmodione and its active analogues, i.e., non-protonable N-dimethylamino groups at the acidic pH

Discussion
In an attempt to generate plasmodione analogues with improved activity/solubility profiles, we developed synthetic methodologies to prepare a biaryl series through a 2-step sequence using Kochi-Anderson reaction followed by a Pd-catalyzed Suzuki-Miyaura cross-coupling, and a second 3-benzylmenadione series bearing a terminal -N(Me)2 or -N(Et)2 in different positions (ortho, meta and para) on the aryl ring of the benzylic chain of plasmodione through reductive amination, as the optimal route to prepare these protonable N-arylalkylamines.
From the mechanistic point of view, the benzoylmenadiones were proposed to be the key potential metabolites acting in oxidized form as the most efficient subversive substrates of reduced GR described so far and, in reduced form, to transfer electrons to methemoglobin. The reduced species of benzoylmenadiones are assumed to be transported through Fe III complexation into the food vacuole where the electrons are transferred to oxidants (heme and methemoglobin). Regenerated in their oxidized form, the benzoylmenadiones might be transported into the cytosol where they are reduced by the cytosolic GR (either from human erythrocyte or from the parasite) in a continuous redox cycle ( Figure 5). Evidence for cytosol-food vacuole shuttling of plasmodione during redox-cycling is difficult to prove, however, replacing the 4′-CF3 group in plasmodione by protonable terminal -N(Me)2 or -N(Et)2, the aim was to vectorize the final derivatives to the acidic compartments of the malarial parasite, e.g., the food vacuole, where hemoglobin digestion takes place. By partly losing antimalarial potency, these compounds provide indirect proof for the hypothesized cytosol-food vacuole shuttling of plasmodione during redox-cycling. Worthy of mention is the fact that the redox-active MB, which possesses potent antimalarial activity, shares similar physico-biochemical properties to those found for plasmodione and its active analogues, i.e., non-protonable N-dimethylamino groups at the acidic pH

Discussion
In an attempt to generate plasmodione analogues with improved activity/solubility profiles, we developed synthetic methodologies to prepare a biaryl series through a 2-step sequence using Kochi-Anderson reaction followed by a Pd-catalyzed Suzuki-Miyaura cross-coupling, and a second 3-benzylmenadione series bearing a terminal -N(Me)2 or -N(Et)2 in different positions (ortho, meta and para) on the aryl ring of the benzylic chain of plasmodione through reductive amination, as the optimal route to prepare these protonable N-arylalkylamines.
From the mechanistic point of view, the benzoylmenadiones were proposed to be the key potential metabolites acting in oxidized form as the most efficient subversive substrates of reduced GR described so far and, in reduced form, to transfer electrons to methemoglobin. The reduced species of benzoylmenadiones are assumed to be transported through Fe III complexation into the food vacuole where the electrons are transferred to oxidants (heme and methemoglobin). Regenerated in their oxidized form, the benzoylmenadiones might be transported into the cytosol where they are reduced by the cytosolic GR (either from human erythrocyte or from the parasite) in a continuous redox cycle ( Figure 5). Evidence for cytosol-food vacuole shuttling of plasmodione during redox-cycling is difficult to prove, however, replacing the 4′-CF3 group in plasmodione by protonable terminal -N(Me)2 or -N(Et)2, the aim was to vectorize the final derivatives to the acidic compartments of the malarial parasite, e.g., the food vacuole, where hemoglobin digestion takes place. By partly losing antimalarial potency, these compounds provide indirect proof for the hypothesized cytosol-food vacuole shuttling of plasmodione during redox-cycling. Worthy of mention is the fact that the redox-active MB, which possesses potent antimalarial activity, shares similar physico-biochemical properties to those found for plasmodione and its active analogues, i.e., non-protonable N-dimethylamino groups at the acidic pH

Discussion
In an attempt to generate plasmodione analogues with improved activity/solubility profiles, we developed synthetic methodologies to prepare a biaryl series through a 2-step sequence using Kochi-Anderson reaction followed by a Pd-catalyzed Suzuki-Miyaura cross-coupling, and a second 3-benzylmenadione series bearing a terminal -N(Me)2 or -N(Et)2 in different positions (ortho, meta and para) on the aryl ring of the benzylic chain of plasmodione through reductive amination, as the optimal route to prepare these protonable N-arylalkylamines.
From the mechanistic point of view, the benzoylmenadiones were proposed to be the key potential metabolites acting in oxidized form as the most efficient subversive substrates of reduced GR described so far and, in reduced form, to transfer electrons to methemoglobin. The reduced species of benzoylmenadiones are assumed to be transported through Fe III complexation into the food vacuole where the electrons are transferred to oxidants (heme and methemoglobin). Regenerated in their oxidized form, the benzoylmenadiones might be transported into the cytosol where they are reduced by the cytosolic GR (either from human erythrocyte or from the parasite) in a continuous redox cycle ( Figure 5). Evidence for cytosol-food vacuole shuttling of plasmodione during redox-cycling is difficult to prove, however, replacing the 4′-CF3 group in plasmodione by protonable terminal -N(Me)2 or -N(Et)2, the aim was to vectorize the final derivatives to the acidic compartments of the malarial parasite, e.g., the food vacuole, where hemoglobin digestion takes place. By partly losing antimalarial potency, these compounds provide indirect proof for the hypothesized cytosol-food vacuole shuttling of plasmodione during redox-cycling. Worthy of mention is the fact that the redox-active MB, which possesses potent antimalarial activity, shares similar physico-biochemical properties to those found for plasmodione and its active analogues, i.e., non-protonable N-dimethylamino groups at the acidic pH

Discussion
In an attempt to generate plasmodione analogues with improved activity/solubility profiles, we developed synthetic methodologies to prepare a biaryl series through a 2-step sequence using Kochi-Anderson reaction followed by a Pd-catalyzed Suzuki-Miyaura cross-coupling, and a second 3-benzylmenadione series bearing a terminal -N(Me)2 or -N(Et)2 in different positions (ortho, meta and para) on the aryl ring of the benzylic chain of plasmodione through reductive amination, as the optimal route to prepare these protonable N-arylalkylamines.
From the mechanistic point of view, the benzoylmenadiones were proposed to be the key potential metabolites acting in oxidized form as the most efficient subversive substrates of reduced GR described so far and, in reduced form, to transfer electrons to methemoglobin. The reduced species of benzoylmenadiones are assumed to be transported through Fe III complexation into the food vacuole where the electrons are transferred to oxidants (heme and methemoglobin). Regenerated in their oxidized form, the benzoylmenadiones might be transported into the cytosol where they are reduced by the cytosolic GR (either from human erythrocyte or from the parasite) in a continuous redox cycle ( Figure 5). Evidence for cytosol-food vacuole shuttling of plasmodione during redox-cycling is difficult to prove, however, replacing the 4′-CF3 group in plasmodione by protonable terminal -N(Me)2 or -N(Et)2, the aim was to vectorize the final derivatives to the acidic compartments of the malarial parasite, e.g., the food vacuole, where hemoglobin digestion takes place. By partly losing antimalarial potency, these compounds provide indirect proof for the hypothesized cytosol-food vacuole shuttling of plasmodione during redox-cycling. Worthy of mention is the fact that the redox-active MB, which possesses potent antimalarial activity, shares similar physico-biochemical properties to those found for plasmodione and its active analogues, i.e., non-protonable N-dimethylamino groups at the acidic pH

Discussion
In an attempt to generate plasmodione analogues with improved activity/solubility profiles, we developed synthetic methodologies to prepare a biaryl series through a 2-step sequence using Kochi-Anderson reaction followed by a Pd-catalyzed Suzuki-Miyaura cross-coupling, and a second 3-benzylmenadione series bearing a terminal -N(Me)2 or -N(Et)2 in different positions (ortho, meta and para) on the aryl ring of the benzylic chain of plasmodione through reductive amination, as the optimal route to prepare these protonable N-arylalkylamines.
From the mechanistic point of view, the benzoylmenadiones were proposed to be the key potential metabolites acting in oxidized form as the most efficient subversive substrates of reduced GR described so far and, in reduced form, to transfer electrons to methemoglobin. The reduced species of benzoylmenadiones are assumed to be transported through Fe III complexation into the food vacuole where the electrons are transferred to oxidants (heme and methemoglobin). Regenerated in their oxidized form, the benzoylmenadiones might be transported into the cytosol where they are reduced by the cytosolic GR (either from human erythrocyte or from the parasite) in a continuous redox cycle ( Figure 5). Evidence for cytosol-food vacuole shuttling of plasmodione during redox-cycling is difficult to prove, however, replacing the 4′-CF3 group in plasmodione by protonable terminal -N(Me)2 or -N(Et)2, the aim was to vectorize the final derivatives to the acidic compartments of the malarial parasite, e.g., the food vacuole, where hemoglobin digestion takes place. By partly losing antimalarial potency, these compounds provide indirect proof for the hypothesized cytosol-food vacuole shuttling of plasmodione during redox-cycling. Worthy of mention is the fact that the redox-active MB, which possesses potent antimalarial activity, shares similar physico-biochemical properties to those found for plasmodione and its active analogues, i.e., non-protonable N-dimethylamino groups at the acidic pH

Discussion
In an attempt to generate plasmodione analogues with improved activity/solubility profiles, we developed synthetic methodologies to prepare a biaryl series through a 2-step sequence using Kochi-Anderson reaction followed by a Pd-catalyzed Suzuki-Miyaura cross-coupling, and a second 3-benzylmenadione series bearing a terminal -N(Me)2 or -N(Et)2 in different positions (ortho, meta and para) on the aryl ring of the benzylic chain of plasmodione through reductive amination, as the optimal route to prepare these protonable N-arylalkylamines.
From the mechanistic point of view, the benzoylmenadiones were proposed to be the key potential metabolites acting in oxidized form as the most efficient subversive substrates of reduced GR described so far and, in reduced form, to transfer electrons to methemoglobin. The reduced species of benzoylmenadiones are assumed to be transported through Fe III complexation into the food vacuole where the electrons are transferred to oxidants (heme and methemoglobin). Regenerated in their oxidized form, the benzoylmenadiones might be transported into the cytosol where they are reduced by the cytosolic GR (either from human erythrocyte or from the parasite) in a continuous redox cycle ( Figure 5). Evidence for cytosol-food vacuole shuttling of plasmodione during redox-cycling is difficult to prove, however, replacing the 4′-CF3 group in plasmodione by protonable terminal -N(Me)2 or -N(Et)2, the aim was to vectorize the final derivatives to the acidic compartments of the malarial parasite, e.g., the food vacuole, where hemoglobin digestion takes place. By partly losing antimalarial potency, these compounds provide indirect proof for the hypothesized cytosol-food vacuole shuttling of plasmodione during redox-cycling. Worthy of mention is the fact that the redox-active MB, which possesses potent antimalarial activity, shares similar physico-biochemical properties to those found for plasmodione and its active analogues, i.e., non-protonable N-dimethylamino groups at the acidic pH pK a1 = 1.7 pK a2 = 4.5 pK a3 = 5.9

Discussion
In an attempt to generate plasmodione analogues with improved activity/solubility profiles, we developed synthetic methodologies to prepare a biaryl series through a 2-step sequence using Kochi-Anderson reaction followed by a Pd-catalyzed Suzuki-Miyaura cross-coupling, and a second 3-benzylmenadione series bearing a terminal -N(Me)2 or -N(Et)2 in different positions (ortho, meta and para) on the aryl ring of the benzylic chain of plasmodione through reductive amination, as the optimal route to prepare these protonable N-arylalkylamines.
From the mechanistic point of view, the benzoylmenadiones were proposed to be the key potential metabolites acting in oxidized form as the most efficient subversive substrates of reduced GR described so far and, in reduced form, to transfer electrons to methemoglobin. The reduced species of benzoylmenadiones are assumed to be transported through Fe III complexation into the food vacuole where the electrons are transferred to oxidants (heme and methemoglobin). Regenerated in their oxidized form, the benzoylmenadiones might be transported into the cytosol where they are reduced by the cytosolic GR (either from human erythrocyte or from the parasite) in a continuous redox cycle ( Figure 5). Evidence for cytosol-food vacuole shuttling of plasmodione during redox-cycling is difficult to prove, however, replacing the 4′-CF3 group in plasmodione by protonable terminal -N(Me)2 or -N(Et)2, the aim was to vectorize the final derivatives to the acidic compartments of the malarial parasite, e.g., the food vacuole, where hemoglobin digestion takes place. By partly losing antimalarial potency, these compounds provide indirect proof for the hypothesized cytosol-food vacuole shuttling of plasmodione during redox-cycling. Worthy of mention is the fact that the redox-active MB, which possesses potent antimalarial activity, shares similar physico-biochemical properties to those found for plasmodione and its active analogues, i.e., non-protonable N-dimethylamino groups at the acidic pH pK a1 = 8.4 ± 0.2 pK a2 = 10.6 ± 0. were reported [4,29]; d pK a values were measured in [56]; e Precipitation prevented the accurate determination of the pK a values which were estimated to be~5-5.5. Error = 1σ with σ = standard deviation. * Precipitates in DMSO.

Discussion
In an attempt to generate plasmodione analogues with improved activity/solubility profiles, we developed synthetic methodologies to prepare a biaryl series through a 2-step sequence using Kochi-Anderson reaction followed by a Pd-catalyzed Suzuki-Miyaura cross-coupling, and a second 3-benzylmenadione series bearing a terminal -N(Me) 2 or -N(Et) 2 in different positions (ortho, meta and para) on the aryl ring of the benzylic chain of plasmodione through reductive amination, as the optimal route to prepare these protonable N-arylalkylamines.
From the mechanistic point of view, the benzoylmenadiones were proposed to be the key potential metabolites acting in oxidized form as the most efficient subversive substrates of reduced GR described so far and, in reduced form, to transfer electrons to methemoglobin. The reduced species of benzoylmenadiones are assumed to be transported through Fe III complexation into the food vacuole where the electrons are transferred to oxidants (heme and methemoglobin). Regenerated in their oxidized form, the benzoylmenadiones might be transported into the cytosol where they are reduced by the cytosolic GR (either from human erythrocyte or from the parasite) in a continuous redox cycle ( Figure 5). Evidence for cytosol-food vacuole shuttling of plasmodione during redox-cycling is difficult to prove, however, replacing the 4 -CF 3 group in plasmodione by protonable terminal -N(Me) 2 or -N(Et) 2 , the aim was to vectorize the final derivatives to the acidic compartments of the malarial parasite, e.g., the food vacuole, where hemoglobin digestion takes place. By partly losing antimalarial potency, these compounds provide indirect proof for the hypothesized cytosol-food vacuole shuttling of plasmodione during redox-cycling. Worthy of mention is the fact that the redox-active MB, which possesses potent antimalarial activity, shares similar physico-biochemical properties to those found for plasmodione and its active analogues, i.e., non-protonable N-dimethylamino groups at the acidic pH of the food vacuole, redox potential values in the same range, NADPH-dependent methemoglobin redox-cycling in the presence of the redox-active blue dye [3,29,33].  The reduced species of 3-benz(o)ylmenadione are subsequently assumed to be transported through Fe III complexation into the acidic vesicles (step 3) where reduced species of 3-benz(o)ylmenadione transfer the electrons to oxidants (hematin or metHb, step 4). The final result is an inhibition of hemozoin formation (step 5) and the arrest of parasite development as shown previously [3]. Hence, the antimalarial benzylmenadione derivatives would act as prodrugs of redox-active principles, being cycled in and out of the acidic vesicles in infected red blood cells, thereby oxidizing major intracellular reductants (NADPH) and subsequently reducing oxidants like hematin or metHb. HMS, hexose monophosphate shunt.
The most active derivatives presented in this work are the non-protonable plasmodione analogues, suggesting that lysosomotropic properties are not favourable to optimize the redox-active compounds that are cycled in and out of the food vacuole in parasitized erythrocytes.

Solvents and Reagents
Commercially available starting materials were purchased from Sigma-Aldrich (St. Louis, MO, USA), ABCR GmbH & Co. KG (Karlsruhe, Germany), Alfa Aesar (Haverhill, MA, USA), and Apollo Scientific (Cheshire, UK) and were used without further purification. Solvents were obtained from Sigma-Aldrich and Carlos Erba (Val de Reuil, France). Unless noted, reagent grade chemicals were used for reactions and analytical grade ones for column chromatography and recrystallizations. Blue arrows indicate reduction; red arrows oxidation; dashed arrows uptake processes. The lead 3-benzylmenadione is proposed to be taken up by infected red blood cells (step 1), be reduced in the cytosols of parasitized erythrocytes by human GR (step 2), and then be oxidized at the benzylic chain to 3-benz(o)ylmenadione (3-benz(o)ylMD) in the acidic vesicles or in the food vacuole in heme-catalyzed reactions (not shown). The reduced species of 3-benz(o)ylmenadione are subsequently assumed to be transported through Fe III complexation into the acidic vesicles (step 3) where reduced species of 3-benz(o)ylmenadione transfer the electrons to oxidants (hematin or metHb, step 4). The final result is an inhibition of hemozoin formation (step 5) and the arrest of parasite development as shown previously [3]. Hence, the antimalarial benzylmenadione derivatives would act as prodrugs of redox-active principles, being cycled in and out of the acidic vesicles in infected red blood cells, thereby oxidizing major intracellular reductants (NADPH) and subsequently reducing oxidants like hematin or metHb. HMS, hexose monophosphate shunt.
The most active derivatives presented in this work are the non-protonable plasmodione analogues, suggesting that lysosomotropic properties are not favourable to optimize the redox-active compounds that are cycled in and out of the food vacuole in parasitized erythrocytes.

Solvents and Reagents
Commercially available starting materials were purchased from Sigma-Aldrich (St. Louis, MO, USA), ABCR GmbH & Co. KG (Karlsruhe, Germany), Alfa Aesar (Haverhill, MA, USA), and Apollo Scientific (Cheshire, UK) and were used without further purification. Solvents were obtained from Sigma-Aldrich and Carlos Erba (Val de Reuil, France). Unless noted, reagent grade chemicals were used for reactions and analytical grade ones for column chromatography and recrystallizations. When specified, anhydrous solvents were required; dichloromethane (DCM) was distilled over CaH 2 under argon. Tetrahydrofuran (THF) was distilled over sodium/benzophenone under argon or dried by passage through an activated alumina column under argon. 1,4-Dioxane and dimethylformamide (DMF) were purchased anhydrous over molecular sieves from Sigma-Aldrich. (St. Louis, MO, USA) Triethylamine (Et 3 N), diisopropylethyl amine (DIPEA), pyrrolidine, piperidine were distilled over KOH under argon and stored over KOH. All reactions were performed in standard glassware. Microwave reactions were carried out on two different apparatus (Biotage Initiator™, Uppsala, Sweden and CEM, Orsay, France) with comparable results (cross compared); supplier standard microwave vials were used. Crude mixtures were purified either by recrystallization or by flash column chromatography. The latter were performed using silica gel 60 (230-400 mesh, 0.040-0.063 mm) purchased from E. Merck (Kenilworth, NJ, USA). Automatic flash chromatographies were carried out in a Biotage Puriflash apparatus (Uppsala, Sweden) with UV-Vis detection at 254 nm (unless otherwise specified). Monitoring and primary characterization of products were achieved by TLC on aluminium sheets coated with silica gel 60 F254 purchased from E. Merck. Eluted TLC's were revealed under UV (325 nm and 254 nm) and with chemicals. Analytical TLC was carried out on pre-coated Sil G-25 UV 254 plates from Macherey Nagel (Hoerdt, France). Flash chromatography was performed using silica gel G60 (230-400 mesh) from Macherey Nagel.

Instruments
The Nuclear Magnetic Resonance (NMR, 1 H-NMR 300 MHz, 13 C-NMR 75 MHz) spectra were registered on a Bruker Avance 300 apparatus (Wissembourg, France). A Bruker Avance 400 apparatus was used ( 1 H-NMR 400 MHz, 13 C-NMR 100 MHz) for more complex spectra. All examples below are labelled as 300/75 MHz except for a few 13 C spectra (compounds 7, 18, 22). All chemical shifts (δ) are quoted in parts per million (ppm). The chemical shifts are referred to the used partial deuterated-NMR solvent (for CDCl 3 : 1 H-NMR, 7.26 ppm and 13 C-NMR, 77.16 ppm). The coupling constants (J) and the non-equivalence (∆ν) are given in Hertz (Hz). Resonance patterns are reported with the following notations: br (broad), s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), dd (doublet of doublets), AB (AB system), (ABX) (AB system of an ABX) and A 2 B 2 (A 2 B 2 aromatic system). In addition, the following acronyms will be used: C=O carbonyl group; C q : quaternary carbon; CH 2 : secondary carbon; CH 3 : methyl group. Elemental analyses (EA) were obtained at "Service de Microanalyses" at the Institut de Chimie de Strasbourg. Mass spectra (ESI-MS) were obtained on a microTOF LC spectrometer (Bruker Daltonics, Bremen, Germany). High Resolution Mass (HRMS) spectra were measured and fitted with calculated data. Melting points (M.p.) were determined on a Büchi (Rungis, France) melting point apparatus and were not corrected.

Cytotoxicity against Human MRC-5 Cells
MRC-5 SV2 cells are cultured in Earls MEM +5% FCSi. Assays are performed in 96-well microtiter plates, each well containing about 10 4 cells/well. After 3 days incubation, cell viability is assessed fluorimetrically after addition of resazurin and fluorescence is measured (λ ex 550 nm, λ em 590 nm). The results are expressed as % reduction in cell growth/viability compared to untreated control wells and CC50 is determined. Compounds are tested at 5 concentrations (64-16-4-1-0.25 µM). When the CC 50 is lower than 16 µM, the compound is classified as toxic.

Animals
Swiss mice (female-BW~25 g; Janvier France, Le Genest Saint Isle, France) were allocated randomly to five groups of four animals each. Drinking water and food were available ad libitum throughout the experiment. The weight of the individual animals did not differ too much from the group mean.

Infection and Drug Treatment of P. berghei-Infected Mice
P. berghei (ANKA-strain) is maintained in the laboratory by weekly mechanical subpassage in Swiss mice. The infection inoculum was prepared by taking heparinized blood was collected from a clinically ill donor mouse (approximately 20% parasitaemia) and diluted in PBS to obtain an infection inoculum of 0.15 mL with about 4 × 10 8 infected erythrocytes. The infection-inoculum was given intraperitoneally. The untreated infected controls developed severe malaria with most animals showing severe clinical signs on day 4. Chloroquine was used as reference treatment (10 mg/kg ip for 5 days). Treatment with chloroquine resulted in 100% survival until day 7 with low parasitaemia at day 4 post infection. The tested compounds were evaluated after ip dosing at 50 mg/kg for 5 days.

Starting Materials and Solvents
Distilled water was further purified by passing it through a mixed bed of ion-exchanger (R3-83002, M3-83006, Bioblock Scientific, Illkirch, France) and activated carbon (Bioblock Scientific ORC-83005) and was de-oxygenated by CO 2 -and O 2 -free argon (Sigma Oxiclear cartridge) before use. All the stock solutions were prepared in spectroscopic grade dimethylsulfoxide (DMSO, Bioreagent for Molecular Biology, >>99.9%, Sigma) by weighing solid products using a XA105 Dual Range (0.01/0.1 mg-41/120 g) balance Mettler Toledo (Viroflay, France) and the complete dissolution was achieved using an ultrasonic bath. The ionic strength was maintained at 0. (1.79 × 10 −5 M) were carried out by diluting stock solutions freshly prepared in spectroscopic grade DMSO in 40 mL of the corresponding solvent contained in a jacketed cell (Metrohm, Villebon sur Yvette, France) maintained at 25.0 ± 0.2 • C (thermostat E200, Lauda, Roissy Charles de Gaulle, France). The free hydrogen ion concentration was measured with a combined glass electrode (Metrohm 6.0234.500, Long Life) filled with 0.1 M NaCl in water and an automatic titrator system 794 Basic Titrino (Metrohm) connected to a microcomputer (Tiamo light 1.2 program for the acquisition of the potentiometric data). The combined glass electrode was calibrated as a hydrogen concentration probe by titrating known amounts of hydrochloric acid (~1.06 × 10 −1 M from HClO 4 , normapur, 70% min Prolabo) with CO 2 -free sodium hydroxide solution (~9.35 × 10 −2 M from NaOH, BdH Prolabo Chemicals, AnalaR, via VWR distribution) [62]. The HClO 4 and NaOH solutions were freshly prepared just before use and titrated with sodium tetraborate decahydrate (B 4 Na 2 O 7 ·10H 2 O, puriss, p.a, Fluka) and potassium hydrogen phthalate (C 8 H 5 KO 3 , Fluka, puriss, p.a.), respectively, using methyl orange (RAL) and phenolphthalein (Prolabo, purum) as the indicators. The GLEE program [62] was applied for the glass electrode calibration (standard electrode potential E 0 /mV and slope of the electrode/mV pH −1 ) and to check carbonate levels of the NaOH solutions used (<5%). The cell was thermostated at 25.0 ± 0.2 • C by the flow of a Lauda E200 thermostat. A stream of argon, pre-saturated with water vapour, was passed over the surface of the solution. The initial pH was adjusted to~2-3 with HClO 4 (Prolabo, normapur, 70% min) and the absorption spectrophotometric titrations of the plasmodione derivatives versus pH (~3 < pH <~9-10) were carried out by addition of known volumes of NaOH solutions using the automatic titrator of the 794 Basic Titrino device (DET method). After each addition, an absorption spectra was automatically and repeatedly recorded using a CARY 50 spectrophotometer (Varian, Les Ulis, France) fitted with Hellma optical fibers (041.002-UV, Hellma, Plainview, NY, USA) and an immersion probe made of quartz suprazil (Hellma, 661.500-QX) and interfaced (Cetrib) with the potentiometric unit. The distribution curves of the protonated species of the plasmodione derivatives as a function of pH were calculated using the Hyss program [63].

Analysis and Processing of the Spectroscopic Data
The spectrophotometric data were analyzed with the Specfit program [64][65][66] which adjusts the absorptivities and the stability constants of the species formed at equilibrium. Specfit uses factor analysis to reduce the absorbance matrix and to extract the eigenvalues prior to the multiwavelength fit of the reduced data set according to the Marquardt algorithm [67,68].

Conclusions
In summary, new synthetic procedures were successfully developed for the preparation of biaryl analogues of the antimalarial plasmodione in a 2-step sequence based on Kochi-Anderson and Suzuki-Miyaura coupling reactions. 3-Benzylmenadione derivatives bearing a terminal -N(Me) 2 or -N(Et) 2 in the different position of the benzylic chain were prepared in a 4-step synthesis for antimalarial drug evaluation using the multiresistant P. falciparum Dd2 strain. Interestingly, the non-protonable 4 -NHBoc-amino-3-benzylmenadione was selected for its potent in vitro activity, suggesting that cytosol-food vacuole shuttling of redox-cyclers may not need to accumulate in the food vacuole through lysosomotropic effects. This compound also revealed to be active in vivo in P. berghei-infected mice. These observations encourage pursuing further optimization of 3-benzyl-menadiones to improve the pharmacokinetic profile of plasmodione that has been shown to fulfill the criteria required for early antimalarial lead compounds.