Discovery, Semisynthesis, Antiparasitic and Cytotoxic Evaluation of 14-Membered Resorcylic Acid Lactones and Their Derivatives

Ten antifouling 14-membered resorcylic acid lactones 1–10 were isolated previously with low or trace natural abundance from the zoanthid-derived Cochliobolus lunatus fungus. Further optimization of fermentation conditions led to the isolation of two major natural compounds 7 and 8 with multi-gram quantities. By one or two steps, we semisynthesized the six trace natural compounds 1–6 and a series of derivatives 11–27 of compounds 7 and 8 with high yields (65–95%). Compounds 11–13 showed strong antiplasmodial activity against Plasmodium falciparum with IC50 values of 1.84, 8.36, and 6.95 μM, respectively. Very importantly, 11 and 12 were non-toxic with very safety and high therapeutic indices (CC50/IC50 > 180), and thus representing potential promising leads for antiplasmodial drug discovery. Furthermore, 11 was the only compound showed obvious antileishmanial activity against Leishmania donovani with an IC50 value of 9.22 μM. Compounds 11 and 12 showed the values of IC50 at 11.9 and 17.2 μM against neglected Chagas’ disease causing Trypanosoma cruzi, respectively.

P. falciparum lines 3D7 (IC 50 : 20 nM and 1.1 μM) and Dd2 (IC 50 : 8.8 μM and 1.7 μM, respectively), were isolated from a Paecilomyces fungus 11,12 . Hypothemycin and aigialomycin D that exhibited antimalarial activity against P. falciparum lines K1 with IC 50 values of 2.2 and 6.6 μg/mL respectively and cytotoxicity, were isolated from Aigialus parvus 13 . In our previous studies, cochliomycins A, C-F (1)(2)(3)(4)(5), and several analogues (6-10) have been isolated with low or trace yields from the fungus C. lunatus (M351) 17 and C. lunatus (TA26-46) 18 . In view of their potentially biological applications, cochliomycins have attracted the attention of academic groups to synthesize these compounds. The most challenging part of their syntheses was the assembly of the macrolide ring and a significant range of techniques with complicated procedures have been tried and improved [19][20][21][22][23][24][25] . For example, cochliomycin A was firstly synthesized with 22 steps (6.5% overall yields) employing stereo-selective Keck allylation, Juliae-Kocienski olefination and a late-stage RCM reaction as the key steps 19 . Meanwhile, the scarce availability of the natural products including time-consuming isolation protocols has also prompted our study of alternative sources of these materials for additional studies.
Up to now, total syntheses of cochliomycins A-C have been accomplished extensively by Nanda's group 19,24 , Du's group 20,21 , Srihari's group 23 , and Banwell's group 22,25 . Cochliomycin A (1) was firstly synthesized in overall yield of 6.5% in 2012 19 , and then was stereo-selectively synthesized by Wang et al. with 16 steps and only mg scale (23.0 mg) 20 . Cochliomycin B has been synthesized in 4.8% overall yields 21 and the latest progress in the modular total syntheses of 1 (36.8 mg) and cochliomycin B (58.0 mg) were reported with 18 steps 22 . Cochliomycins A-C were in close resemblance to antiplasmodial paecilomycins E and F structurally, and cochliomycin C (2) is a chlorinated derivative of paecilomycin F on aromatic ring at C5 carbon. Total syntheses of paecilomycins E and F were also accomplished and reported during the period 2012-2016 [26][27][28][29][30] . Very recently, cochliomycin C (2) and paecilomycin F were synthesized successively by Srihari's group 23 with 16 steps, Nanda's group 24 with 12 steps in overall yield of 14.1%, and Banwell's group 25 with 11 steps. There is no report on total syntheses of compounds 3-6. Alternatively, semisynthetic strategy could provide an efficient and economical route to obtain natural products for further investigation of their pharmacological activities.
In this study, further optimization of the fermentation conditions of C. lunatus resulted in multi-gram quantities of compounds 7 and 8, and the one or two-step semisyntheses of natural compounds 1-6 and a series of derivatives 11-27 with high yields were reported (Fig. 1). The antiplasmodial, antileishmanial, antitrypanosomal, and cytotoxic activities of these compounds were evaluated in vitro. The preliminary structure-activity relationships were discussed.

Results and Discussion
Progress of the optimization of fermentation. The fungal strain C. lunatus was cultured in a 500 mL flask containing 200 mL liquid medium and incubated at 28 °C for 7 days on a rotary shaker at 120 rpm. The result of analysis showed that 7 and 8 are two predominant products in oligotrophic liquid medium (soluble starch 10 g/L, tryptone 1 g/L, 3% salinity). Then, the medium was selected as the basic medium for further optimization. The level of all the single factors including carbon source, nitrogen source, salinity, pH, inoculum size, and cultural time were studied. Subsequently, orthogonal test was carried out to approach the optimum point. All experiments were carried out in triplicate and the yields of 7 and 8 were calculated as the average values of three independent experiments. The fermented liquid medium was extracted with 200 mL ethyl acetate for each flask three times. The yield of 8 was 155.4 mg/L (soluble starch 10 g/L, NaNO 3 5 g/L, NaOAc 1 g/L, 1% salinity, 10% inoculum size, adjusting initial pH value to 6 with 10% HCl/NaOH, Medium I). The yield of 7 reached 137.8 mg/L (soluble starch 10 g/L, NaNO 3 5 g/L, 0.1% salinity, 15% inoculum size, adjusting initial pH value to 6 with 10% HCl/NaOH, Medium II) (Fig. 2). This supplied with a feasible method to obtain multi-gram quantities of 7 and 8 by fermentation in the laboratory. The advantages of the fermentation method were low cost, time saving, and easy separation by recrystallization. The structures of 7 and 8 were determined by comparing the NMR and MS spectroscopic data with that reported in literature 31,32 . Semisynthesis. Compounds 1-6 and 11-27 were designed and synthesized through one or two-step semisynthetic reaction with high yields ranging from 65-95% from 7 and 8 (Fig. 1). The structures of new derivatives 12, 14-19, 21-22, 24, and 27 were identified by extensive spectroscopic methods including HRESIMS, 1 H NMR, and 13 C NMR. Compound 1 was prepared in yield of 95% through one-step acetonide reaction from 8. Other acetonide derivatives (11, 17, 18, 20, 23, 25, and 26) were also prepared with the same yields. Chlorination of compound 8 with sulfuryl chloride 33 followed by selective reduction with 10% Pd-C under H 2 atmosphere led to the natural compound 2 (68%, yield). Acylation reaction was carried out by reaction of anhydride with the corresponding materials (1-3, and 6-8) in high yields (85-95%). In order to prepare 6, oxidation of the allyl alcohol 8 was examined under various oxidation conditions (PCC, PDC, MnO 2 , DMP, and IBX). However, only DMP and IBX enabled the selective oxidation of the allyl alcohol to give 6 in 65% and 35% yields, respectively (Fig. 3). Interestingly, a fortuitous observation revealed that 6 can slowly proceed in chemical conversion to 3-5 in the process of silica gel column separation. Compounds 3-6 are diastereomers differing from each other by the absolute configurations of the 4′, 5′-diol chiral centers. The proportions of compounds 3-6 were in dynamic equilibrium, and the final ratios of 3-6 were approximately 8: 2: 5: 6 ( Fig. 4). Further study revealed that a subtle chemical conversion of 6 to 3-5 was observed in protic solvents (MeOH, EtOH, and MeOH-d 4 , except for H 2 O) in the stationary state within a week, and similar interconversion of 4 and 5 were also observed under the same conditions, while compound 3 was found to be quite stable (Fig. 5). However, compounds 4-6 were found to be stable in   aprotic solvents (EtOAc, CH 3 CN, acetone, and CH 2 Cl 2 ). Compounds 4-6 were unstable because of the existence of trans-enone moieties and the absolute configurations of the 4′, 5′-diol groups, causing chemical conversions under the condition of protic solvent, such as intramolecular hetero Michael addition 34 and carbon migration of hemiacetal 35 . This conversion played a vital role in obtaining compounds 3-5 for further biological study.
Antiplasmodial activity. Results for the in vitro antiplasmodial activity of all of the compounds were shown in Table 1. Compounds 11-13 and 26 exhibited strong antiplasmodial activity with the IC 50 values of 1.84, 8.36, 6.95, and 8.95 μM, respectively. It should be pointed that the introduction of the acetoxy groups in compounds 1 and 8 appreciably change the activity, indicating that adding the acetoxy groups has a positive effect on the antiplasmodial activity. The acetonide functionality in 11 improved the IC 50 value approximately 4-fold over that of 13 with the acetoxy groups at C5′-C6′. This suggests that the acetonide functionality might contribute to the antiplasmodial activity. However, the introduction of the chlorine atom at C5 in 2 and 16-19 was all found to be inactive, indicating that the chlorine atom has a negative effect on the antiplasmodial activity (Fig. 6).
Antileishmanial activity. Antileishmanial activity of all of the compounds was tested in vitro against L.
donovani and were also shown in Table 1. For the enone RALs, compounds 3, 6, 7, 20-22, 26, and 27 displayed obvious antileishmanial activity with IC 50 values ranging from 1.24 μM to 9.11 μM. Additionally, antiplasmodial compound 11 also showed antileishmanial activity with an IC 50 value of 9.22 μM. Antitrypanosomal activity. The bioassays against T. cruzi of all 27 molecules showed little activity against this parasite in initial screenings at 10 μg/mL, except in the cases when the compound was also discovered to be toxic (data not shown). However, compounds 11 and 12 showed IC 50 values of 11.9 and 17.2 μM, respectively, with selectivity indexes of 28 and 88, making 12 a good candidate for in vivo studies.
Toxicity. It should be emphasized that toxicity is a major concern of drug discovery and development. The above active compounds were evaluated for in vitro toxicity against a mammalian kidney cell line (Vero). The selectivity index (SI) was used as the evaluation parameter of drug potential of the test samples by comparing the toxicity on the Vero cell line (CC 50 ) and the selective inhibitory effect on P. falciparum, L. donovani or T. cruzi (IC 50 ) calculated here as CC 50 /IC 50 . Compounds 11-13 and 15 were found to be non-toxic and showed encouraging therapeutic indices, which were far greater than the value requested by the Medicine for Malaria Venture (SI > 10). Specially, antiplasmodial compounds 11 and 12 showed pronounced selectivity indices (SI > 180), and thus represent potentially promising antiplasmodial leads for further development. Compounds 12 and 13, where 12, with a 2-OH group, displayed 3-fold higher therapeutic ratio and equivalent or stronger activities than compound 13, indicating that 2-OH is an important functionality for reducing the toxicity. For the enone RALs, most of them had high antileishmanial efficacy but poor therapeutic indices, indicating that the cis or trans-functionality has a positive contribution to toxicity, and that the antileishmanial activity of these compounds against L. donovani may well be related to their toxicity (Fig. 6).   (Table 2). Compounds 3, 6, 7, 20, 21, 23, 26, and 27 with the enone functionality exhibited selectively cytotoxicity against the above cell lines with IC 50 values less than 10 μM. This further confirmed that the enone functionality was related to the toxicity.

Conclusion
In summary, the natural products 1-6 and a group of derivatives 11-27 were semisynthesized with multi-gram scales of 7 and 8 by one or two steps with high yields. This semisynthetic route provided a convenient source of these RALs for further potentially biological applications. Compounds 11-13 showed strong antiplasmodial activity against P. falciparum, of which 11 and 12 were discovered to be promising non-toxic antiparasitic candidates. The structure-activity relationship analysis indicated that the acetoxy and acetonide groups are required for antiplasmodial activity, while the introduction of chlorine atom at C-5 is not necessary to the activity. Additionally, enone group can lead to an obvious toxicity and a comparatively lower therapeutic ratio, but 2-OH is critical for reducing toxicity. Further studies of selected compounds 11-13 on assessing their in vivo activities and a systematic optimization based upon these identified promising chemical leads are in progress and will be reported in due course.

Methods
General experimental procedures. NMR I  I  I  I  I  I  nt   2  I  I  I  I  I  I  nt   3  I  I  7.18  I  I  I  nt   4  I  I  I  I  I  I  nt   5  I  I  I  I  I  I  nt   6  I  8.47  I  I  I  I  8.59   7  1.09  5.07  I  I  I  I  I   8  I  I  I  I  I  I  I   9  I  I  I  I  I  I  nt   10  I  I  I  I  I  I  nt   11  I  I  I  I  I  I  I   12  I  I  I  I  I  I  I   13  I  I  I  I  I  I  I   14  I  I  I  I  I  I   Fermentation, extraction, and isolation. The fungal strain was cultivated in 10 L of optimal liquid medium (Medium I and Medium II, respectively) at 28 °C for 7 days on a rotary shaker at 120 rpm. Then the culture was filtered to separate the culture broth from the mycelia. The culture broth was extracted with an equal volume of EtOAc. The combined EtOAc solution was evaporated to dryness under a vacuum to give an EtOAc extract. The EtOAc extract (4.  Antiplasmodial assay. Activity against the causative agent of malaria was performed by culturing human erythrocytes and infecting them with P. falciparum, as described by Trager and Jensen 36 . Briefly, the W2 (Chloroquine resistant) and 3D7 (Chloroquine sensitive) strains of P. falciparum were cultured in RPMI 1640 medium (Sigma-Aldrich, USA) supplemented with 10% human serum (from O+ blood) at a hematocrit of 2% erythrocytes (O+) at 37 °C in a gas mixture of 5% CO 2 , 5% O 2 , and 90% N 2 . Parasites were synchronized by a temperature cycling technique as described by Almanza et al. 37 Malaria bioassays were performed following the procedure of Corbett et al., which used Pico-Green to assess parasite growth inhibition by drugs and used chloroquine as positive control, with an IC 50 of 32.9 nM 38 .
Antileishmanial assay. The anti-leishmania activity was evaluated following the protocol described by Calderon et al. 39 , using the fluorescent DNA intercalator PicoGreen (Invitrogen, USA). The species responsible for visceral leishmaniasis, L. donovani, was used for the assays. Amphotericin B was used as the positive control and had an IC 50 of 76.3 nM 40 .
Antitrypanosomal assay. The antitrypanosomal test used the methods recommended in Romanha et al.
for screening against this parasite with benznidazole as the control drug, with an IC 50 of 3.85 μM 41 . Briefly, the expression of the reporter gene for beta-galactosidase (β-Gal) in the Tulahuen clone C4 of T. cruzi is assessed by colorimetry at 570 nm, which correlates with the parasite growth 42 . Assays were performed on the intracellular amastigote form of the parasite infecting African green monkey kidney (Vero) cells, incubated for 120 h at 37 °C with 5% CO 2 .
SRB and MTT assays. The cytotoxicity against HCT-116, BXPC-3, HeLa, MCF-7, A-549, and HUVEC cell lines was evaluated using the SRB method 43 . The cytotoxicity against K562 and Vero cell lines was evaluated using the MTT method 44 . Adriamycin was used as a positive control. All cells were cultivated in T-75 flasks containing 10% Fetal Bovine Serum (FBS), Dulbeco's Modified Eagle Medium (DMEM) with 2 mM L-glutamine, and 1% penicillin-streptomycin at 37 °C in a humidified atmosphere with 5% CO 2 . For SRB assay, cells in their log phase of growth were seeded into 96 well micro plates (4 × 10 4 cells per well) and followed by treating with test samples at 10 μM or DMSO. After incubation at 37 °C for 72 h, the cells were fixed with the cold 20% (w/v) TCA for 2 h and stained with Sulforhodamine B (SRB) dye for 0.5 h. The protein-bound dye is dissolved in Tris base solution for OD determination at 540 nm on an ELISA Plate Reader. For MTT assay, cells (5 × 10 3 cells per well) were plated in 96 well micro plates. After cell attachment overnight, new medium containing test compounds or DMSO was added. The plate was incubated for 48 h at 37 °C and incubated with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) for additional 2-4 hours. The assay plate was then read at 490 nm on an ELISA Plate Reader. The data were obtained from experiments carried out in triplicate.