SARs for the Antiparasitic Plant Metabolite Pulchrol. Part 2: B- and C-Ring Substituents

Neglected tropical diseases affect most of the underprivileged populations in tropical countries. Among these are chagas and leishmaniasis, present mainly in South and Central America, Africa and East Asia. Current treatments are long and have severe adverse effects, therefore there is a strong need to develop alternatives. In this study, we base our research on the plant metabolite pulchrol, a natural benzochromene which has been shown to possess antiparasitic activity against Trypanosoma and Leishmania species. In a recent study, we investigated how changes in the benzyl alcohol functionality affected the antiparasitic activity, but the importance of B- and C-ring substituents is not understood. Fifteen derivatives of pulchrol with different substituents in positions 1, 2, 3, and 6 while leaving the A-ring intact, were therefore prepared by total synthesis, assayed, and compared with pulchrol and positive controls. The generated series and parental molecule were tested in vitro for antiparasitic activity against Trypanosoma cruzi, Leishmania braziliensis, and L. amazonensis, and cytotoxicity using RAW cells. Substantial differences in the activity of the compounds synthesized were observed, of which some were more potent towards Trypanosoma cruzi than the positive control benznidazole. A general tendency is that alkyl substituents improve the potency, especially when positioned on C-2.


Introduction
Neglected Tropical Diseases (NTDs) native to tropical regions, affect around 1 billion people in 149 countries [1], most of them part of underprivileged populations [2][3][4]. Some NTDs are associated to parasites from the family Trypanosomatidae, among them Trypanosoma cruzi which causes the chagas disease, while several Leishmania species are responsible for leishmaniasis. Chagas disease (transmitted by a Triatominae bug) affects around 8 million people, mainly in South and Central America [5]. It may remain asymptomatic for decades until the heart tissue is eventually damaged sufficiently to cause death [6,7]. The antiparasitic drugs benznidazole and nifurtimox are used to treat chagas, and are efficient when given immediately after infection. However, they lose efficiency with time, and present serious adverse effects [8,9]. Leishmaniasis (transmitted by Phlebotomine sand flies) is found mainly in Africa, Latin America and East Asia. It exists in three forms: cutaneous, mucocutaneous, and visceral, and some 700,000 to 1 million new cases occur annually [10]. The treatment, mainly pentavalent The benzo[c]chromene pulchrol (1), found in the roots of Bourreria pulchra, [22] is known as "Bakalche" in Yucatan (Mexico) and used to treat cutaneous diseases, injuries, viral infections and fevers [23,24]. Pulchrol (1) was found to possess antiparasitic activity against Leishmania braziliensis, L. amazonensis, L. mexicana and Trypanosoma cruzi [22,25]. A synthetic route was developed in 2014 [26,27], and recently we reported the effects that transformations of its A-ring benzyl alcohol functionality have against T. cruzi epimastigotes, L. amazonensis promastigotes and L. braziliensis promastigotes, as well as their cytotoxicity towards mammalian cells (assayed in RAW cells) [28]. We are now interested to expand the previous study [28] with compounds having various substituents in the B-and C-rings. Our aim is to use the synthetic routes and some of the intermediates used for the synthesis of pulchrol (1), to obtain new derivatives. The only position available for exchange in the B-ring is C-6, and we were especially interested in examining the role of the alkyl substituents. Ring C has theoretically four positions open for substitution, but in practice only three. Here we focused on the presence and position of a methoxy group, as well as various alkyl substituents. As a result, we have prepared 15 new analogues (8a-10g, 10a-10h) and we have tested them for antiparasitic activity towards T. cruzi, L. amazonensis and L. braziliensis together with 1 and the positive controls benznidazole and miltefosine. The cytotoxicity towards mammalian cells of all the compounds was determined with murine macrophage cells (RAW). This and additional studies of the antiparasitic activities of pulchrol analogues will eventually provide us with a model that can be used for the design of more potent antiparasitic structures with lower cytotoxicity.

Modifications in the B-Ring
The synthetic routes used to prepare derivatives with variations in ring B were partly based on an already published synthetic route to yield pulchrol [26]. The common intermediate 4 was used as the starting material for the synthesis of all ring B derivatives (see Scheme 1). The benzo[c]chromene pulchrol (1), found in the roots of Bourreria pulchra, [22] is known as "Bakalche" in Yucatan (Mexico) and used to treat cutaneous diseases, injuries, viral infections and fevers [23,24]. Pulchrol (1) was found to possess antiparasitic activity against Leishmania braziliensis, L. amazonensis, L. mexicana and Trypanosoma cruzi [22,25]. A synthetic route was developed in 2014 [26,27], and recently we reported the effects that transformations of its A-ring benzyl alcohol functionality have against T. cruzi epimastigotes, L. amazonensis promastigotes and L. braziliensis promastigotes, as well as their cytotoxicity towards mammalian cells (assayed in RAW cells) [28]. We are now interested to expand the previous study [28] with compounds having various substituents in the B-and C-rings. Our aim is to use the synthetic routes and some of the intermediates used for the synthesis of pulchrol (1), to obtain new derivatives. The only position available for exchange in the B-ring is C-6, and we were especially interested in examining the role of the alkyl substituents. Ring C has theoretically four positions open for substitution, but in practice only three. Here we focused on the presence and position of a methoxy group, as well as various alkyl substituents. As a result, we have prepared 15 new analogues (8a-10g, 10a-10h) and we have tested them for antiparasitic activity towards T. cruzi, L. amazonensis and L. braziliensis together with 1 and the positive controls benznidazole and miltefosine. The cytotoxicity towards mammalian cells of all the compounds was determined with murine macrophage cells (RAW). This and additional studies of the antiparasitic activities of pulchrol analogues will eventually provide us with a model that can be used for the design of more potent antiparasitic structures with lower cytotoxicity.

Modifications in the B-Ring
The synthetic routes used to prepare derivatives with variations in ring B were partly based on an already published synthetic route to yield pulchrol [26]. The common intermediate 4 was used as the starting material for the synthesis of all ring B derivatives (see Scheme 1).
Derivative 8a was prepared by reducing the ester group in 4 to the alcohol 5, which was treated with NaSEt in dry DMF at 110 • C to obtain an ortho demethylated phenol, which was not isolated as an intermediate as it spontaneously cyclized and was deprotected to the desired product 8a, the 6-demethylated analogue of 1, albeit in low yields (7%). The monosubstituted analogues 8b-8e were prepared by reducing the ester functionality of 4 to the aldehyde 7, which by alkyl addition was transformed to the corresponding secondary alcohol. Cyclization using PBr 3 in the presence of LiI gave the desired compounds [29]. The products 8b-8e were obtained as racemic pairs, and the enantiomers were separated by HPLC with a normal phase semipreparative chiral column. The pure enantiomers were obtained in low yields (less than 10%). The determination of the absolute configuration of 8b-8e, which could have been done with the secondary alcohols by the Mosher's method, was not attempted as the enantiomers were approximately equipotent (vide infra). The absolute configuration of C-6 does not appear to influence the potency. The 6.6-diethyl and 6.6-dibutyl analogues 8f and 8g were prepared from 6a and 6b, based on the pulchrol synthetic route [26]. The new alkyl groups were introduced by a double addition step to the ester group in 4 using the corresponding organolithium reagents. For the cyclization of 6a and 6b to 8f and 8g, an excess in hydroiodic acid was used, to avoid the formation of cannabidiol type byproducts as a result of elimination [28]. Nevertheless, the byproducts 9a and 9b appeared together with 8f, and 9c and 9d with 8g. The desired products were obtained by HPLC purification, with moderate yields (8f 30% and 8g 56%).

Modifications in the C-Ring
Analogues modified in the C-ring (see Figure 2 for the structures of compounds 10a to 10h) were prepared based on the procedure used to synthesize pulchrol (1) [26], but with different methoxylated phenyls forming the biaryl (corresponding to 4) through a Suzuki coupling reaction. The major difference in the biaryl formation was that the reaction time in the microwave reactor had to be increased from 30 min to 60 min. The yields were generally good, varying from 75% to 92%. During the cyclization most of the alkyl substituted analogues were obtained in better yields (72% to 85%) than the derivatives 10a and 10b substituted with methoxy groups. method, was not attempted as the enantiomers were approximately equipotent (vide infra). The absolute configuration of C-6 does not appear to influence the potency. The 6.6-diethyl and 6.6-dibutyl analogues 8f and 8g were prepared from 6a and 6b, based on the pulchrol synthetic route [26]. The new alkyl groups were introduced by a double addition step to the ester group in 4 using the corresponding organolithium reagents. For the cyclization of 6a and 6b to 8f and 8g, an excess in hydroiodic acid was used, to avoid the formation of cannabidiol type byproducts as a result of elimination [28]. Nevertheless, the byproducts 9a and 9b appeared together with 8f, and 9c and 9d with 8g. The desired products were obtained by HPLC purification, with moderate yields (8f 30% and 8g 56%).

Modifications in the C-Ring
Analogues modified in the C-ring (see Figure 2 for the structures of compounds 10a to 10h) were prepared based on the procedure used to synthesize pulchrol (1) [26], but with different methoxylated phenyls forming the biaryl (corresponding to 4) through a Suzuki coupling reaction. The major difference in the biaryl formation was that the reaction time in the microwave reactor had to be increased from 30 min to 60 min. The yields were generally good, varying from 75% to 92%. During the cyclization most of the alkyl substituted analogues were obtained in better yields (72% to 85%) than the derivatives 10a and 10b substituted with methoxy groups. The biological activities of all the synthesized derivatives are given in Table 1, while the 1 H-and 13 C-NMR chemical shifts of the assayed compounds are given in Tables 2 and 3. (1) The natural product pulchrol has previously been investigated and shown to be toxic towards Trypanosoma and Leishmania parasites in vitro. The highest activity was reported towards T. cruzi epimastigotes (IC50 18.5 μM) which is comparable to the potency shown by the drug benznidazole (19.2 μM) that currently is used to treat the chagas disease. Pulchrol also showed moderate leishmanicidal activity against L. braziliensis and L. amazonesis promastigotes, with IC50 values of 59.2 μM and 77.7 μM, respectively. The effects that modifications of the benzylic alcohol functionality of 1 have on the antiparasitic activity were studied previously [28], and especially esters of the alcohol increased the potency significantly. In this study, we evaluate how modifications on ring B and C affect the antiparasitic activity against T. cruzi, L. braziliensis and L. amazonensis. As a comparison, the cytotoxicity to mammalian murine macrophage cell lines (RAW) was determined, and the quotas IC50 RAW cells/ IC50 parasite is given as the selectivity index (SI) in Table 1.

Antiparasitic Activity of Pulchrol
The biological activities of all the synthesized derivatives are given in Table 1, while the 1 H-and 13 C-NMR chemical shifts of the assayed compounds are given in Tables 2 and 3.   Table 3. 13 C-NMR chemical shifts (in ppm) for the assayed compounds 1, 3a-h, 4a-l, 5a-e and 6 determined at 100 MHz in CDCl 3 . The assignments were made with 2D NMR spectroscopy, COSY, HMQC and HMBC experiments.

Antiparasitic Activity of Pulchrol (1)
The natural product pulchrol has previously been investigated and shown to be toxic towards Trypanosoma and Leishmania parasites in vitro. The highest activity was reported towards T. cruzi epimastigotes (IC 50 18.5 µM) which is comparable to the potency shown by the drug benznidazole (19.2 µM) that currently is used to treat the chagas disease. Pulchrol also showed moderate leishmanicidal activity against L. braziliensis and L. amazonesis promastigotes, with IC 50 values of 59.2 µM and 77.7 µM, respectively. The effects that modifications of the benzylic alcohol functionality of 1 have on the antiparasitic activity were studied previously [28], and especially esters of the alcohol increased the potency significantly. In this study, we evaluate how modifications on ring B and C affect the antiparasitic activity against T. cruzi, L. braziliensis and L. amazonensis. As a comparison, the cytotoxicity to mammalian murine macrophage cell lines (RAW) was determined, and the quotas IC 50 RAW cells/ IC 50 parasite is given as the selectivity index (SI) in Table 1.

Antiparasitic Activity Against Trypanosoma cruzi Epimastigotes
Compared to pulchrol (1), the 6,6-didemethyl analogue 8a is considerably less active (66.6 µM), indicating the importance of alkyl substituents on C-6 for the activity against T. cruzi. However, the toxicity to mammalian cells (SI) is also less. The 6-methyl enantiomers 8b and 8c were less potent than 1, as are the 6-ethyl enantiomers 8d and 8e, suggesting the importance of a dialkylated C-6. In 8f the two methyls of 1 have been exchanged for ethyls and compared to 1 as well as benznidazole the antiparasitic activity (10.4 µM) is higher. In addition, 8f showed the highest selectivity (SI 4.2) among all molecules assayed towards T. cruzi in this study. Unlike 8f, analogue 8g with two n-butyl substituents was slightly less potent (22.8 µM) than 1, and less selective than 8f. A possibility is that the compounds for their effect on T. cruzi interact with a lipophilic pocket in a target protein around C-6, although its volume is limited.
Changing the position of the methoxy substituent in the C-ring to positions C-1 (compound 10a) and C-3 (compound 10b) was not beneficial, and the SI-value was lower. A methoxy group in the C-ring is only efficient in position 2, as in pulchrol (1), and the replacement of the methoxy groups in positions 2 and 3 with methyls (compounds 10c and 10d) resulted in equipotent compounds that were more active than 10a and 10b but less active than 1. The analogue with no substituent in the C-ring, 10e, was slightly less potent than 10c and 10d. It is possible that a methoxy group in position 2 enables a hydrogen bond at the target, while a methyl group in position 3 is better than a methoxy or no substituent at all. To further explore the effects of alkyl substituents in the positions 2 and 3, the isopropyl analogues 10f and 10g were prepared and assayed. Both were more potent than 1, and comparable with the 6,6-diethyl analogue 8f. Finally, 10h, with a n-pentyl group in position 2, was found to possess the highest activity towards T. cruzi of all compounds assayed in this investigation (6.4 µM), being approximately three times as potent as pulchrol (1) and the positive control benznidazole. This contradicts the suggestion that the methoxy substituent at C-2 enables a hydrogen bond, and instead propose that the lipophilicity of the pulchrol analogues is correlated with the antiparasitic activity towards T. cruzi.

Antiparasitic Activity Against Leishmania braziliensis Promastigotes
Similar to the results obtained for T. cruzi, the 6,6-didemethyl analogue 8a was considerably less potent than 1, and this is also true for the monomethyl enantiomers 8b and 8c. However, the monoethyl enantiomers 8d and 8e as well as the 6,6-diethyl analogue 8f were more potent towards L. braziliensis and actually slightly more so compared to 1. For the 6,6-dibutyl analogue 8g with the IC 50 -value 29.3 µM this trend is even stronger. Towards L. braziliensis the positioning of the methoxy group in the C-ring at C-3 (10a) or C-1 (10b) instead of C-2 (1), as well as replacing the methoxy substituent at C-2 and C-3 for a methyl (analogues 10c and 10d) results in almost equipotent compounds that are slightly more potent than 1. The analogue without substituents in the C-ring (10e) is less impressive, while the compounds with bigger alkyl substituents at C-2 and C-3 are the most potent towards L. braziliensis. The C-2 isopropyl analogue 10f, as well as 10g (C-3 isopropyl) and 10h (C-2 n-pentyl) were all considerably more potent than 1 towards L. braziliensis, with IC 50 -values between 15 and 20 µM, close to that of the positive control miltefosine. However, their selectivity for the parasite over the mammalian cells was less impressing.

Antiparasitic Activity Against Leishmania amazonensis Promastigotes
As can be seen in Table 1, the antiparasitic activity towards L. amazonensis is not improved compared to 1 by replacing the 6,6-dimethyl substituents in 1 for hydrogens (8a), one methyl and one hydrogen (8b and 8c), or one ethyl and one hydrogen (8d and 8f). More potent analogues are the 6,6-diethyl and 6,6-dibutyl analogues with IC 50 -values of 36.9 and 25.4 µM, respectively. This is similar to what was observed with T. cruzi and L. braziliensis (vide supra). For the C-ring analogues, there is a strong variation in the potency depending on the position of the methoxy group, and while the C-3 methoxy analogue 10a is slightly more potent than 1, the C-1 methoxy analogue (10b) is considerably less potent. Methyl groups in positions 2 and 3 (10c and 10d) do not really change things, neither does the nonsubstituted 10e. Again, the most potent analogues are those with larger alkyl groups in positions 2 and 3 (10f, 10g and 10h).

General
1 H-NMR spectra (400 MHz) and 13 C-NMR spectra (100 MHz) were recorded in CDCl 3 with a Bruker Avance II instrument (Bruker Biospin AG, Fällanden, Switzerland). The individual 1D signals were assigned using 2D NMR experiments (COSY, HSQC, HMBC). The chemical shifts are given in ppm with the solvent signal as reference (7.27 ppm for 1 H and 77.0 for 13 C). Infrared spectra were recorded with a Bruker Alpha-P FT/IR instrument (Bruker Biospin AG) with a Diamond ATR sensor as films, and the intensities are given as vw (very weak), w (weak), m (medium), s (strong) and vs (very strong). High resolution mass spectra (HRMS) were recorded with a Waters XEVO-G2 QTOF instrument (Waters Corp, Milford, MA, USA) equipped with electrospray ionization (ESI). A weak solution (10 mg/mL) was leaked into the ionizing unit, and the mass spectrum was recorded. Synthetic reactions were monitored by TLC using alumina plates coated with silica gel and visualized using either UV light and/or spraying/heating with vanillin/H 2 SO 4 . Flash chromatography was performed with silica gel (35-70 µm, 60 Å). Chiral separations were performed by semipreparative HPLC (mod. 1260 Infinity system, column CHIRALPAK ® IB, 4 mL/min, 96:4 hexane/isopropyl, UV detector 254 nm, Agilent (Santa Clara, CA, USA). THF was distilled from sodium, acetonitrile was distilled from CaH 2 and other reaction solvents were dried with Al 2 O 3 . Commercially available compounds were obtained from Aldrich (St. Louis, MO, USA). More detailed data are available in the supplementary materials. the synthesis of 4), BH 3 -THF (1 M, 68 mL, 68.2 mmol) was slowly added to a stirred solution of 3-iodo-4-(methoxycarbonyl)benzoic acid (5.2 g, 17.1 mmol) in dry THF (250 mL) at 0 • C. After 30 h, saturated aqueous NaHCO 3 /H 2 O was added, and the aqueous phase was extracted with ethyl acetate (3 × 250 mL) before drying (Na 2 SO 4 ) and removal of solvent under reduced pressure. Purification by column chromatography (SiO 2 , 4:6 heptane/ethyl acetate) gave (3.65 g, 73%) of the pure product as yellow crystals, identical to that previously reported [26].

Evaluations Against Trypanosoma cruzi
Cultures of Trypanosoma cruzi (epimastigotes, donated by the Parasitology Department of INLASA, Tc-INLASA, city, country), were maintained in medium LIT (pH 7.2), supplemented with 10% FBS and incubated at 26 • C. Medium changes were made every 72 h to maintain a viable parasitic population. Trypanocidal activity was determined according to Muelas-Serrano with some modifications [33]. Samples were dissolved in DMSO (maximum final concentration 1%) at 10 mg/mL. Epimastigotes in logarithmic phase of growth, at a concentration of 3 × 106 parasites/mL, were distributed (100 µL/well) in 96-well flat bottom microtiter plates. Samples at different concentrations (3.1-100 µg/mL) were added (100 µL). Benznidazole (3.1-100 µg/mL) was used as the control drug. Assays were performed in triplicates. The microwell plates were incubated for 72 h at 26 • C. After incubation, a solution of XTT (1 mg/mL) in PBS (pH 7.0 at 37 • C) with PMS (0.06 mg/mL) was added (50 µL/well) and incubated for 4 h at 26 • C. The optical density of each well was measured and the IC 50 values were calculated. A negative control experiments with only 1% DMSO was carried out, showing that the solvent by itself has no antiparasitic activity.

Evaluations Against RAW Cells
The Raw 264.7 murine macrophage cell line was purchased from the American Type Culture Collection (ATCC-TIB71, ARCC (Manassas, VA, USA). The cells were maintained in DMEM-HG medium supplemented with 10% fetal bovine serum, 100 U/mL of penicillin and 100 µg/mL of streptomycin, and sodium bicarbonate (2.2 g/L) in humidified atmosphere at 37 • C with 5% CO 2 . Samples were dissolved in DMSO and diluted (maximum final concentration of DMSO: 1%) at different concentrations (6.2-200 µg/mL). Medium blank, control drugs and cell growth controls were included to evaluate cell viability. The plates were incubated for 72 h at 37 • C with 5% CO 2 and 3 × 10 4 cells/well. After incubation for the indicated time, the cells were washed, after which 10 µL of resazurin reagent (2.0 mM) was added. They were further incubated at 37 • C for 3 h in a humidified incubator. The IC 50 values were assessed using a fluorometric reader (BioTek (Winooski, VT, USA), 540 nm excitation, 590 nm emission) and the Gen5 software (v. 2017, BioTek). All assays were performed in triplicate.

Conclusions
Fifteen derivatives of pulchrol with modifications on ring B and ring C were prepared and tested towards T. cruzi, L. braziliensis and L. amazonensis, together with 1. The importance of the presence of methyl substituents on ring B was investigated, and the unsubstituted derivative 8a was shown to be less active compared to 1 towards all the three parasites. The effect on bioactivity that just one substituent has on C-6 was different between the parasites. The 6-methyl monosubstituted enantiomers were not more active than 1, suggesting that two methyl substituents instead of one may improve orientation and lipophilic interactions in the binding site. 6-Ethyl monosubstituted derivatives are slightly more potent towards the two Leishmania species, but with T. cruzi they are less potent. The longer the alkyl substituents on C-6 are, the more interesting is the activity. A preference for disubstituted rather than monosubstituted analogues appears to be at hand, but with T. cruzi the 6,6-diethyl analogue is better that the 6,6-dibutyl analogue, and 8f was found to be more potent and selective than the positive control benznidazole. This suggests that additional derivatives with larger and branched alkyl groups at C-6 should be prepared and assayed.
The methoxy group in the C-ring was also shown to play a role for pulchrol's bioactivity, as a derivative without substituents (10e) was considerably less active compared to 1. A methoxy substituent in either C-3 or C-2 appears beneficial compared to C-1, although the differences are not massive. A methyl at C-2 or C-3 instead of a methoxy group has a small impact, although for T. cruzi 1 is still the most potent. Longer and more bulky alkyl substituents in positions C-2 and C-3 (10f, 10g and 10h) are clearly more potent, with all three parasites. The C-2 n-pentyl analogue 10h showed the best activities towards T. cruzi, while 10h together with the C-2 isopropyl analogues 10f and 10g showed the best results with L. braziliensis and L. amazonensis.
Most of the differences in the antiparasitic activity observed in this study can be tentatively suggested to be linked to the lipophilicity of the compounds. However, nothing is known about the molecular targets in these parasites, and to increase our understanding it is necessary to expand our studies in a systematic way. Compared to the QSARs suggested in the previous study of the benzyl alcohol function, we have now a new wish list of compounds to prepare and assay.
The 1D 1 H and 13 C-NMR shifts of the assayed compounds are given in Tables 2 and 3, and as the shifts reflects the electronic conditions in the vicinity of each nucleus they may indicate SARs. However, with the data available in this study, no SARs are obvious from the NMR shifts.