Quinone-Amino Acid Conjugates Targeting Leishmania Amino Acid Transporters

The aim of the present study was to investigate the feasibility of targeting Leishmania transporters via appropriately designed chemical probes. Leishmania donovani, the parasite that causes visceral leishmaniasis, is auxotrophic for arginine and lysine and has specific transporters (LdAAP3 and LdAAP7) to import these nutrients. Probes 1–15 were originated by conjugating cytotoxic quinone fragments (II and III) with amino acids (i.e. arginine and lysine) by means of an amide linkage. The toxicity of the synthesized conjugates against Leishmania extracellular (promastigotes) and intracellular (amastigotes) forms was investigated, as well their inhibition of the relevant amino acid transporters. We observed that some conjugates indeed displayed toxicity against the parasites; in particular, 7 was identified as the most potent derivative (at concentrations of 1 µg/mL and 2.5 µg/mL residual cell viability was reduced to 15% and 48% in promastigotes and amastigotes, respectively). Notably, 6, while retaining the cytotoxic activity of quinone II, displayed no toxicity against mammalian THP1 cells. Transport assays indicated that the novel conjugates inhibited transport activity of lysine, arginine and proline transporters. Furthermore, our analyses suggested that the toxic conjugates might be translocated by the transporters into the cells. The non-toxic probes that inhibited transport competed with the natural substrates for binding to the transporters without being translocated. Thus, it is likely that 6, by exploiting amino acid transporters, can selectively deliver its toxic effects to Leishmania cells. This work provides the first evidence that amino acid transporters of the human pathogen Leishmania might be modulated by small molecules, and warrants their further investigation from drug discovery and chemical biology perspectives.


Introduction
Leishmania is an obligate intracellular protozoan parasite responsible for leishmaniasis, a debilitating disease found in more than 88 countries, including all Mediterranean area [1]. Resulting infection in humans ranges from chronic skin ulcers, via erosive mucosal disease with progressive destruction of the nasopharynx and severe facial disfigurement, to a life-threatening systemic infection with hepatosplenomegaly [1]. Global yearly incidence of leishmaniasis approaches 2 million new cases with an estimated 59,000 deaths predominantly in India, Bangladesh, Nepal, and Sudan [2], representing a significant health problem in tropical and subtropical regions of the world. Despite the epidemiological importance in the developing world, such as the possible impact in Western societies, the drugs used for the treatment of this disease are unsatisfactory due to high cost, toxicity, problems in administration and resistance issues [3]. In view of the foregoing facts, there is still an urgent need for new and more effective drugs and innovative drug design strategies. In this context, biology oriented synthesis (BIOS) of natural product-derived and -inspired scaffolds may show promising results in finding new lead structures for chemical biology and medicinal chemistry research [4]. Indeed, this approach recognizes natural products as evolutionary selected and biologically pre-validated starting points in chemical space to be suitably deployed for library generation [4]. On this basis, the natural occurring lapachol was used to generate the natural product-inspired skeleton shown in Figure 1, and this was subjected to modification to yield a small library of potential hits [5]. In a phenotypic based screening, the most promising derivative resulted naphthoquinone I (Fig. 1), with an IC 50 value (1.26 mM) against Leishmania axenic amastigotes only four time higher than the reference drug miltefosine (0.31 mM) [5]. However, it showed a selectivity index (SI) with respect to mammalian cells of only 4.7 [5]. This indicated I as a potential hit candidate for the development of compounds endowed with a better activity and, at the same time, better selectivity. Toward this goal, due to the lack of target information driving a target-based approach, we were forced toward a ligand-based design strategy. For this reason while fishing target(s) of I through chemical proteomics techniques [6], we envisaged a 'targeting' approach [7,8,9] as a suitable strategy to generate novel molecules endowed with an improved profile. Indeed, drug targeting, i.e. uptake of a drug via parasite-specific pathways, has been advocated as a chemotherapeutic strategy to selectively inhibit drug targets that have equally sensitive counterparts in the host [8].
It is well known that Trypanosomatid parasites (including Trypanosoma and Leishmania spp.) depend on their hosts for several nutrients, such as glucose, purines, amino acids, vitamins and lipids, necessary for their replication and survival [10]. In order to acquire these molecules, the parasites express plasma membrane transporters that mediate their intake [10]. These transporters can potentially be exploited to carry toxic compounds inside parasite cells, leading to selective toxicity delivered via selective transport systems [7,8,9]. Fairlamb [11] and co-workers first recognized the potential of the P2 purine transporter to carry drugs into trypanosomes [12]. From these findings, membrane transporters have been utilized by several medicinal chemists to design anti-parasitic agents with improved profiles [13,14,15,16,17,18,19,20,21,22,23,24].
Furthermore, beside the selective accumulation of drugs into parasite cells, the design and development of new transporterdirected compounds could find application in the therapeutic area [25,26]. In addition, from a chemical biology perspective, such compounds can be used as chemical probes to study the role of amino acid transporters as drug targets in Leishmania.
Transporter-directed compounds are typically chemical conjugates obtained by linking two moieties: (i) a so called haptophore, which is recognized by the transporter and (ii) a fragment that may have anti-parasitic activity (toxophore) [27].
Arginine (Arg) and lysine (Lys) are essential amino acids for Leishmania, whereas in mammals arginine is conditionally essential, depending on the health status and developmental stage of the organism [28]. Shaked-Mishan et al. and Inbar et al. have identified the arginine (LdAAP3) and lysine (LdAAP7) transporters from Leishmania donovani. These proteins translocate arginine and lysine at high-affinity and specificity. LdAAP3 expression and activity are highly dependent on arginine availability [29,30,31].
We hypothesized LdAAP3 and LdAAP7 can be targeted with small molecules by (i) selective delivery of cytotoxic molecules into the parasite, or (ii) by alteration of nutrients intake. In addition, this approach can be useful to elucidate the complex biology of parasite transporters.
Here, we designed new conjugates in which the leishmanicidal naphthoquinone moiety of I was coupled to Arg and Lys pendant groups (Figs. 1 and 2). We found, for the first time, that some conjugates inhibited transport and a few were cidal to L. donovani promastigotes and amastigotes via competitive inhibition with their natural substrates.

Chemistry
All reagents and solvents were purchased by Sigma Aldrich and used without further purification, unless otherwise stated. Melting points were measured on a Buchi SMP-20 apparatus and are uncorrected. Direct infusion ESI-MS spectra were recorded on Waters ZQ 4000 apparatus. NMR spectra were recorded on Varian VRX 200 and 400 MHz instruments. Chemical shifts (d) are reported in parts per million (ppm) relative to tetramethylsilane (TMS), and spin multiplicities are given as s (singlet), br s (broad singlet), d (doublet), t (triplet), q (quartet), or m (multiplet). The elemental compositions of the compounds agreed to within 60.4 % of the calculated value. When the elemental analysis is not included, crude compounds were used in the next step without further purification. Chromatographic separations were performed on silica gel (Kieselgel 40, 0.040-0.063 mm; Merck) by flash chromatography. Reactions were followed by thin-layer chromatography (TLC) on glass-backed pre-coated silica gel plates (0.25 mm, 60 F254; Merck), then visualized in an iodine chamber or with a UV lamp. The term ''dried'' refers to the use of anhydrous Na 2 SO 4 . Compounds were named following IUPAC rules as applied by ChemBioDraw Ultra 11.0.

Synthesis of activated amino acids 27-30
N a ,N e -Di-Boc-L-lysine hydroxysuccinimide ester (26) was commercially available. 4-(Boc-amino)butyric acid (33) and 6-(Boc-amino)caproic acid (35) were commercially available and were just subjected to the carboxylic group activation. L-Ornithine hydrochloride (31) (1 g, 5.93 mmol) was dissolved in NaOH (5 mL) and water (5 mL), then a solution of NaOH 4 N (2.73 mL) and a solution of (Boc) 2 O (2.98 g, 13.64 mmol) in 5 mL of THF were added dropwise and stirred for 6 h at room temperature. THF was evaporated and the aqueous solution was made acidic with KHSO 4 and extracted with ethyl acetate (2620 mL). The organic extracts were collected, dried over Na 2 SO 4 and evaporated under vacuum. The crude was crystallized by ethyl acetate/ cyclohexane to give 34 (1.67 g, 85%). N a -Boc-L-arginine (32) (1 g, 3.65 mmol) was dissolved in NaOH 2N (17.5 mL) and dioxane (15 mL) at 0uC. Di-tert-butyl dicarbonate (1.64 g, 7.51 mmol) was added and the resulting suspension was stirred at room temperature for 3 h. Then, further di-tert-butyl dicarbonate (0.82 g, 3.72 mmol) in dioxane (7.5 mL) was added, stirring for additional 39 h. The pH was adjusted to 7 with KHSO 4 and the reaction mixture extracted with ethyl acetate (3620 mL). The organic extracts were collected, dried over Na 2 SO 4 and evaporated under vacuum. The crude was purified by flash chromatography eluting with CHCl 3 /MeOH (9:1) to afford 36 (1.32 g, 76%).
A modified procedure was followed for the synthesis of 1-5. 16   solutions were stirred at 80uC for 24 h, and, after cooling to room temperature, the solvent was evaporated under vacuum. The removal of N-Boc protecting group was carried out as described for 6-15, affording 1 (0.052 g, 55%), 2 (0.033 g, 48%), 3 (0.045 g, 49%), 4 (0.032 g, 43%), 5 (0.056 g, 51%) as trifluoroacetate salts. For elemental analyses of compounds 1-15 see File S1.  Figure 5. Relative toxicity of conjugates 1-15 and parent compounds II and III in axenic amastigotes and promastigotes. Axenic cultures of L. donovani promastigotes and amastigotes were exposed at compound concentration of 1 mg/mL and 2.5 mg/mL, respectively. Thymidine incorporation was measured to monitor cell proliferation. The results show the mean of 2 or 3 independent repeats. doi:10.1371/journal.pone.0107994.g005 Table 1. Effect of selected quinone-amino acid conjugates on amino acid transport in L. donovani promastigotes.

Biological assays of I, II and III
In vitro activity against axenic amastigotes of L. donovani (strain MHOM/ET/67/L82), and cytotoxicity assessment against L6 cells were determined as previously reported [32].

Thymidine incorporation assays
Assays were carried out as described in Shaked-Mishan et al. [35]. Briefly, to 1610 7 cells 1 Ci/mL of [ 3 H]-thymidine was added in 24-well ELISA plates. After a 3 h incubation for promastigotes and 24h for amastigotes, the cells were transferred to a 2 mL tube and washed in 2 mL cold PBS. Then they were suspended in 2 mL lysis buffer (140 mM NaCl, 1.5 mM MgCl 2 , 10 mM Tris-HCl pH 8.6 and 0.5% NP-40). The suspension was vortexed for 10 seconds and then centrifuged for three minutes (6000 g at 4uC). The nuclei-containing pellet was suspended in 2 mL ice-cold 10% trichloroacetic acid (TCA), kept at 4uC for at least 20 minutes and subsequently filtered through GF/C filters (Whatman International Ltd.). The filters were washed with 5 mL 10% TCA and then with 5 mL 95% ethanol. The filters were then subjected to scintillation counting.

Transport assays
Uptake of 25 mM [ 3 H]-arginine (600 mCi/mmol), 10 mM [ 3 H]lysine (1.0 Ci/mmol), or 1 mM [ 3 H]-proline (8.0 mCi/mmol) into mid log phase parasites was determined using the rapid filtration technique [36]. To determine initial rates of transport, transport measurements were performed on 1610 8 promastigotes or amastigotes, which were exposed to radiolabel for up to 5 minutes. The amount of radiolabel associated with the cells was linear with time over the 5 minutes time course of the transport assay. Quinone-amino acid conjugates dissolved in DMSO were added to the cells and pre-incubated for 10 minutes prior to the beginning of transport measurement. Pre-incubation was carried out at 30uC for promastigotes and 37uC for amastigotes.

Conjugates design rationale
Quinones II and III (Fig. 1) were selected as the toxophores to be conjugated with the proper haptophoric amino acids. Thus, II and III were tested against Leishmania axenic amastigotes (strain MHOM-ET-67/L82). Despite they showed leishmanicidal activities (6.02 mM and 8.74 mM, respectively) lower than I (1.26 mM), while retaining the same toxicity against mammalian L6 cells (3.73 mM, 7.03 mM, 5.92 mM respectively), II and III were preferred for the further conjugation approach. This is because the replacement of the ether linkage of I with the amino isosteric groups of II and III might have a positive effect on the chemical and metabolic stability of phenoxy compound.

Conjugates synthesis
II [37] and III [36,38] were prepared according to reported procedures. The novel synthetic route developed for the preparation of quinone-amino acid conjugates 1-15 is depicted in Figure 3. After proper optimization conditions, all the reactions steps were carried out in a parallel fashion using a carousel workstation, with significant improvement of the overall efficiency. The first step was the synthesis of quinones 16-18 (Fig. 3). The Boc amino-protecting group was selected along all the procedure, because of its compatibility with the stability of the amino-quinone bond. Michael-type addition of amines onto p-quinones was adopted as the C-N coupling reaction. There are several reports on the use of this reaction to access 1,4-naphthoquinones possessing a substituted amino group in the 2-position [39,40,41]. However, especially for the transformation of less reactive anilines, catalysts [42], high temperature [37], and long reaction time [36,38] are often required. We developed a simple and straightforward experimental protocol which led to 19-21 in good yields and in parallel fashion. In details, they were synthesized by reacting 1,4-naphthoquinone (22) with three monoBoc-diamines, namely tert-butyl 4-aminophenylcarbamate (23), tert-butyl 4-aminobenzylcarbamate (24) and tert-butyl 4-(aminomethyl)benzylcarbamate (25) in a 1:1 ratio in MeOH for 4 h. Removal of N-Boc protecting groups by TFA afforded compounds 16-18 quantitatively. Then, 1-15 were obtained in a one-pot sequence that involves the coupling of 16-18 to the five protected amino acids (26)(27)(28)(29)(30), activated as NHS esters, and the subsequent removal of the N-Boc protecting groups by acidic treatment. NHS esters are widely used as coupling agents in peptide synthesis and were found suitable for our amide formation [43]. However, the lower reactivity of the aniline group of 16 with respect to the benzyl amine of 17 and 18, required more drastic coupling conditions [44]. Reaction between 16 and amino acids 26-30 was carried out in dioxane at 80uC for 24 h using DMAP as catalyst, while for 17-18 the reaction was easily carried out in dry CH 2 Cl 2 for 5 h at room temperature.
NHS esters were prepared accordingly to standard peptide synthesis procedures; the available amino acids were first protected on the amino groups and then activated on the carboxylic group (Fig. 4). The amino groups of L-ornithine (31) and N a -Boc-Larginine (32) were protected by using slightly different conditions: for 31 the reaction was carried out in THF using (Boc) 2 O and NaOH, while for 32 in dioxane [45] at 0uC. 4-(Boc-amino)butyric acid (33) and 6-(Boc-amino)caproic acid (35) were already commercially available. Then, the carboxylic group of 33-36 was activated by using NHS in presence of DCC and DMAP to afford derivatives 27-30. N a ,N e -Di-Boc-L-lysine hydroxysuccinimide ester (26) was obtained from commercial suppliers.

Toxicity and transport in Leishmania donovani
Dose-response analysis was carried out in axenic L. donovani promastigotes and amastigotes. To assess the relative toxicity, we first determined the IC 90 values of the synthesized compounds, as well as those of the parent compounds II and III. Exposing L. donovani to 1 mg/mL II or III for 48 h killed 95% of promastigotes, whereas a higher concentration (2.5 mg/mL) was required to obtain a similar effect in amastigotes. Therefore, we decided to test the amino acid conjugates 1-15 at the latter concentrations (Fig. 5). All lysine analogues (1,6,11) were generally less toxic than the parent compounds. Among them, the most effective was 6 that killed 70% of promastigotes and 50% of amastigotes. 1 retained toxicity for promastigotes (50%), but had no effect on amastigotes. 11 was not toxic to promastigotes at all, and was slightly toxic (30%) to amastigotes.
To examine whether the lysine conjugates target lysine transporters we conducted competition experiments to measure transport of lysine in the presence of either 6 or 11. When treating promastigotes with 6 and 11 (in a ratio 1:10 lysine:compound) initial transport rates were reduced to 58% and 25% of control (Table 1, Fig. 6A). In amastigotes treated with 6, initial transport rate was reduced to 47% (Fig. 6C). More importantly, we found that inhibition was dose-dependent (see the 2:1 ratio column in Table 1). This indicated that the compounds are recognized by the lysine transporter and therefore inhibit lysine transport activity.
Arginine conjugates were even less toxic than lysine derivatives. Indeed, in 5, 10 and 15 conjugation abolished II and III toxicity (Fig. 5). However, all three compounds had a strong and dosedependent effect on arginine transport activity (Table 1, Fig. 6B).
Interestingly, GABA conjugate 7 retained the toxicity of the parent compounds against promastigotes, whereas toxicity against amastigotes was reduced to 50%. On the other hand, 2 or 12 completely lacked the toxicity of the parent compounds (Fig. 5). GABA is a close analogue of proline and has been shown to be transported through the Leishmania proline-alanine transporter [46]. To assess whether GABA conjugates 2 and 7 enter promastigotes and amastigotes through LdAAP24, we assayed their inhibitory effect on proline transport. Both 2 and 7 dramatically inhibited proline transport in promastigotes, in spite of the difference in toxicity (Table 1).
In order to validate the hypothesis that transport inhibition was caused by competition between the compounds and the respective amino acids, and not by a general deterioration in cell viability, we also examined the effect of lysine conjugate 6 on the transport of proline. Since lysine is transported by a different transporter than that of proline, there should be no effect of 6 on proline transport. Indeed, we found that proline transport rate was not affected by 6 ( Fig. 7), confirming that the general condition of the cells is not compromised by compound toxicity.
Ornithine conjugate 8 showed reasonable toxicity for promastigotes (70%, Fig. 5), but toxicity for amastigotes was reduced to 30%. The ornithine analogues 3 and 13 were significantly less toxic than the parent compounds II and III.
Since II and III are hydrophobic, it is likely that they enter parasite cells via simple diffusion. The conjugation with hydrophilic amino acids makes these compounds virtually impermeable by diffusion, thus forcing them to enter cells via membrane transporters.
However, both arginine and lysine transporters are highly specific and sensitive to even minor changes in the substrate structures. Nevertheless, transport results indicated that the compounds successfully compete with the natural substrate of the transporters. This could be the result of blocking the recognition site, and not necessarily the result of entry into the cell. LdAAP24 is less substrate-specific, and we speculated that the conjugates would have been translocated more efficiently by this transporter. Indeed, 7 is the most toxic compound among all the conjugates, and a very effective transport inhibitor. In details, 7 was 5-10 fold more efficient than the tested lysine (6 and 11) and arginine (5, 10 and 15) conjugates (Table 1).
To preliminary assess the therapeutic potential of the most toxic compounds against Leishmania, citotoxicity of 6 and 7 in THP1 human macrophage cell line, which serves as ex vivo model for leishmaniasis, was evaluated. The results depicted in Figure 8 clearly indicated that 7 was still toxic to macrophages, though to a lesser extent compared to the parent molecule II. Conversely, 6 did not display any toxicity. These findings nicely suggest that 6, thanks to its selective cytotoxicity towards Leishmania amastigotes with respect to the permissive host cell line, is a potential hit for therapeutic application. Conversely, 7, which is not able to effectively discriminate among parasite and mammalian cells, is not worthy of further therapeutic pursuit.

Discussion
Transporters are transmembrane proteins that regulate the translocation of non-diffusible small molecules in and out of cells. Most transporters that are actively being pursued in drug discovery, are localized in the central nervous system, and have been targeted by tricyclic antidepressants since the '70s. In the anti-parasitic field, transporters have been much less exploited, even though a few have been showed to translocate Trypanosomatid-toxic compounds [47]. In Leishmania amino acid transporters play a major role in creating and maintaining intracellular pools of their corresponding substrates [38]. Hence, conjugating toxic drugs to Leishmania nutrients, which exploit specific amino acid transporters, might lead to successful approach for targeted toxicity. However, it should be highlighted that this applies only when the structural features essential for activity are preserved in the conjugates. In our case, it is likely that this might occur, as the parent compounds and the conjugates share the same quinone substructure. Indeed, the quinone core, in addition to a possible target-related mechanism, exerts a toxic effect through a general free-radical-generation mechanism [6].
For many years, the design of transporter-directed chemical probes has led to the identification of specific membrane permeases and to the elucidation of the underlying molecular transport mechanism [48]. In absence of crystallographic or NMR structural data, chemical probing has been used as a major mean for functional identification of targeted transport systems.
Hence, properly designed small molecules might be particularly useful for dissecting the role of the recently identified amino acid transporters in Leishmania, both from a drug discovery and a chemical biology viewpoint. To achieve this, we designed a new series of chemical probes able to target Leishmania transporters, utilizing an amino acid motif conjugated to a quinone moiety taken from II and III. The conjugates were evaluated for their toxicity against Leishmania and for their ability to modulate arginine, lysine and proline transport activity. Our analysis indicated that the parent compounds II and III were highly toxic, likely due to their ability to diffuse into cells as hydrophobic compounds. In contrast, all the conjugates were either not toxic or less toxic than II and III, suggesting that they are no longer permeable to cells due to the hydrophilic amino acid pendants. Notably, our studies experimentally verified that the conjugates interfered with amino acid transport. We hypothesize that the toxic conjugates are translocated by the specific transporters. Once inside the parasite cells, they might accumulate and the quinone fragment can exert its toxic activity. It may appear from our data that the compounds act as antileishmanial agents through substrate competition and not necessarily through the compounds' toxicity feature. While arginine and lysine are both essential amino acids for the parasites and preventing their entrance would indeed suffice to inhibit parasite growth, proline is not. Therefore, if toxicity was caused only by competition for transport rather than by the toxicity of the compound itself, GABA conjugates inhibiting proline transport should have exhibited lower toxicity as they inhibit the transport of a non-essential amino acid. In light of this, it is likely that the compounds mechanism of action involves activity of the quinone fragment rather than mere competition by the amino acid fragment.
In our view, the fact that the compounds have lower efficacy in amastigotes (mammalian-stage of the parasite) than in promastigotes (insect stage), as well as their decreased toxicity compared to that of II and III, is compensated for by their improved specificity. In other words, although the compounds are less effective against amastigotes, they are still effective enough to be therapeutically useful. More importantly, they are selectively directed at the parasites' cells and more toxic to them than to the host, as demonstrated for the lysine conjugate 6.
This investigation, although preliminary, suggests that Leishmania amino acid transporters, in both the extracellular (promastigotes) and intracellular (amastigotes) forms, can serve as drug delivery tools. As far as we know, this is the first time that small molecules modulating such Leishmania transporters have been reported; thus, they act as chemical probes useful to study such transporters as drug targets.
In conclusion, the results presented herein may be a starting point for increasing our fundamental understanding of the biological role of L. donovani amino acid transporters and of their potential role as molecular targets in anti-trypanosomatid drug discovery. We are confident that further exploration in both these key issues may ultimately contribute to combating leishmaniases with high efficacy and less host toxicity.