Depsides: Lichen Metabolites Active against Hepatitis C Virus

A thorough phytochemical study of Stereocaulon evolutum was conducted, for the isolation of structurally related atranorin derivatives. Indeed, pilot experiments suggested that atranorin (1), the main metabolite of this lichen, would interfere with the lifecycle of hepatitis C virus (HCV). Eight compounds, including one reported for the first time (2), were isolated and characterized. Two analogs (5, 6) were also synthesized, to enlarge the panel of atranorin-related structures. Most of these compounds were active against HCV, with a half-maximal inhibitory concentration of about 10 to 70 µM, with depsides more potent than monoaromatic phenols. The most effective inhibitors (1, 5 and 6) were then added at different steps of the HCV lifecycle. Interestingly, atranorin (1), bearing an aldehyde function at C-3, inhibited only viral entry, whereas the synthetic compounds 5 and 6, bearing a hydroxymethyl and a methyl function, respectively, at C-3 interfered with viral replication.


Introduction
Hepatitis C virus (HCV) is a small, enveloped virus of genus Hepacivirus, from the Flaviviridae family. About 170 million people worldwide are chronically infected with HCV, and there are three to four million new infections each year [1], principally through blood transmission. The high genetic variability of the virus has hindered vaccine development. However, new treatments have emerged in the last few years, with the discovery of direct-acting antivirals (DAAs), which were added to the previous standard care procedure based on the use of pegylated interferon (IFN) and ribavirin (RVB). Clinical trials currently underway suggest that IFN-free regimens combining several DAAs will cure almost all cases of hepatitis C, making it possible to avoid the severe adverse effects of IFN treatment [2]. Nevertheless, the current high cost of such treatments may restrict them to only the wealthiest patients and nations [3]. There is, therefore, still a need to develop alternative and complementary approaches to treatment for a large proportion of patients.
It has recently been shown that natural compounds may be a source of anti-HCV drugs [4,5], but the presence of such active molecules has never been investigated in lichens. Lichens are symbiotic associations between at least one fungal and an algal and/or cyanobacterial partner. Some 1,035 secondary metabolites have been isolated from the 18,500 lichen species from the Arctic to tropical habitats described to date, and many more compounds remain to be characterized [6]. Most of these compounds are polyketides, polyphenols, quinones or terpenoids, presumably of fungal origin [6,7], and their biological activities remain largely underexplored. However, a few have been shown to have antibiotic, antimycobacterial, anti-inflammatory, analgesic, antipyretic, or antiproliferative effects [7]. Antiviral activities have also been reported for some secondary lichen metabolites, such as (+)-usnic acid, sekikaic acid and anthraquinones, targeting Junin and Tacaribe arenaviruses [8], respiratory syncytial virus [9] and herpes simplex virus type 1 [8], respectively. In a pilot study, we discovered that atranorin seemed to inhibit HCV. As this depside is a major compound of the genus Stereocaulon, we conducted a phytochemical investigation on a crude extract of S. evolutum. Seven structurally related atranorin derivatives were isolated, together with seven other secondary metabolites. Two atranorin analogs were also synthesized. The relationship between structure and antiviral activity against HCV was discussed with these molecules.

General procedures
All solvents for chromatography were purchased from Sigma-Aldrich (France). Thin-layer chromatography (TLC) was carried out on silica gel plates (Merck silica gel 60F 254 ) with the B, C, and G standard solvent systems for the identification of substances from lichens [10]. The spots were first visualized under UV light and then after spraying with anisaldehyde-H 2 SO 4 reagent. Circular chromatography and column chromatography (CC) were carried out on silica gel (40-63 μm, Kieselgel 60, Merck, 7667). Medium-pressure liquid chromatography (MPLC) was conducted on a SPOT Liquid Chromatography Flash apparatus (Armen Instrument), using prepacked silica gel or RP-18 columns (Chromabond, Merck). 1 H, 13 C and 2D NMR spectra were recorded on a BrukerAvance III400 or a Bruker 300 spectrometer with deuterated solvents. High-resolution mass spectrometry (HR-MS) measurements were carried out to determine exact mass, on a Bruker Maxis 4G, MicrO-Tof Q 2, a Thermo-fisher Q-Exactive or a Waters Q-Tof 2 mass spectrometer for chemical ionization. HPLC-ESI and ESI-MS n mass spectra were obtained with a LCQ Deca ion trap mass spectrometer (Thermo Finnigan) equipped with an ESI source, in negative ionization mode. The Ion Trap MS system was coupled to an HPLC system, including a Surveyor autosampler (Thermo Finnigan), a 1100 Series binary pump (Agilent Technologies), a UV-6000 DAD UV-visible detector (Thermo Finnigan) and a Zorbax Eclipse XDB-C18 column (150 mm x 2.1 mm, Agilent Technologies) maintained at 30°C by a thermostat.

Ethics statement
Specimens of S. evolutum Graewe were collected from siliceous rocks in Saint Just (Ille et Vilaine, France), by F. Le Dévéhat, in November 2011. No specific permits were required for the described field studies in Saint Just (Ille et Vilaine). The research sites are not privately owned or protected in any way and field studies did not involve endangered or protected species. A voucher specimen (JB/10/121) has been deposited in the Herbarium of the Department of Pharmacognosy and Mycology of the University of Rennes 1 (France).

Detection of compound 2 by LC-ESI-MS/MS in acetone and ethyl acetate extracts
Dried ground S. evolutum powder (1 g) was macerated at room temperature in acetone or ethyl acetate (15 mL) for 24 h. This extraction process was repeated three times. Simultaneously, pure atranorin (1 g) or a dried ethyl acetate extract (100 mg) was macerated in acetone (15 mL or 1.5 mL) for one week. Compound 2 was detected in extracts by HPLC-ESI, as previously described [11].

Hemisynthesis of atranorin derivatives
Methyl-8-hydroxy-4-O-demethylbarbatate (5): a solution of atranorin (93.5 mg, 0.25 mmol) in methanol (10 mL, 0.025 M) was stirred at 0°C for a few minutes. Sodium borohydride (18.9 mg, 0.5 mmol) was then added and the mixture was stirred for a further 1 h at room temperature. The reaction was quenched by adding a few drops of saturated ammonium chloride. The solvent was evaporated off and the residue was purified on a small column of silica gel and eluted with a solvent system of petroleum ether-CH 2 Cl 2 (1:9), followed by CH 2 Cl 2 (100%). Compound 5 (79.3 mg, 84%) was obtained as a white crystalline solid [10,12].
Methyl-4-O-demethylbarbatate (6): a solution of methyl-2,4-dihydroxy-3,6-dimethylbenzoate (compound 7) (147 mg, 0.75 mmol) in anhydrous CH 2 Cl 2 (3.75 mL, 0.2 M) was added dropwise, at-70°C, to a 1.0 M solution of boron tribromide in anhydrous CH 2 Cl 2 (3 mL, 3 mmol). The reaction mixture was stirred for 12 h at room temperature. It was then hydrolyzed with 10 mL of 2 N HCl solution and subjected to extraction with ethyl acetate. The organic phase was washed with 2 N HCl and then with a saturated NaCl solution, dried over MgSO 4 , filtered and evaporated. The residue was passed through a silica gel column, with 20% ethyl acetate-petroleum ether as the eluent. β-orcinol carboxylic acid (103.1 mg, 74%) was obtained as a white crystalline solid [10]. Trifluoroacetic anhydride (0.5 mL, 8 mmol) was added to a solution of β-orcinol carboxylic acid (98.3 mg, 0.54 mmol) in anhydrous diethyl ether (5 mL, 0.1 M) under nitrogen, and the mixture was stirred for ten minutes at room temperature. We then added compound 7 (88.2 mg, 0.45 mmol) and the mixture was stirred for 24 h at room temperature. The solvent was removed under low pressure and the residue was purified by chromatography on deactivated silica gel with 30% ethyl acetate/n-hexane, with a gradual increase in polarity to 50% ethyl acetate/n-hexane. Methyl-4-O-demethylbarbatate (6) was obtained as a white solid (24 mg, 15%) [13].

Cell culture
Huh-7.5.1 cells [14] were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 4.5 g/l D-glucose, 4 mM L-glutamine (Life Technologies), supplemented with 100 units/mL penicillin (Life Technologies), 100 μg/mL streptomycin (Life Technologies), non-essential amino acids (Sigma-Aldrich), 1 mM Hepes (Life Technologies) and 10% heatinactivated fetal calf serum (FCS, Hyclone). Lichen compounds and control HCV inhibitors (erlotinib from Abcam and telaprevir from Euromedex) were dissolved in dimethyl sulfoxide (DMSO). In cell viability and HCV in vitro propagation experiments, the final concentration of DMSO was systematically adjusted to 0.1%. We checked that this amount of DMSO had no effect on the biological cycle of HCVcc in vitro (data not shown).

Virus production and titration
HCVcc was generated from the FL-J6/JFH-5'C19Rluc2AUbi construct, a monocistronic, fulllength HCV genome that expresses Renilla luciferase [15]. It was produced and titrated as described elsewhere [16,17], except for the readout for luciferase activity measurement. Cell viability was evaluated with the Cell Proliferation Reagent WST-1 (Roche), according to the manufacturer's instructions. Luciferase assays were performed according to the manufacturer's instructions (Promega) and measurements were performed on a Centro XS3 LB960 luminometer (Berthold Technologies).

Data analysis
Cell viability and viral replication data from at least three independent experiments were analyzed. The IC 50 values of the compounds were determined from dose-response curves generated with seven to 12 concentrations of each compound. Curves were fitted to the data points by nonlinear regression analysis and IC 50 values were interpolated from the resulting curves with GraphPad Prism 5.00 software.

Phytochemical composition
Phytochemical investigations were carried out on whole thalli of S. evolutum, with the aim of isolating metabolites structurally similar to atranorin. Briefly, extractions with solvents of increasing polarity (n-hexane, acetone, tetrahydrofuran) yielded three fractions, which were purified by extensive and repeated column chromatography on silica gel and RP-18, to give 15 compounds (Fig. 1). The structure of the new compound 2 was assigned by detailed spectroscopy and the structures of the known compounds were determined by comparisons with published data.
Compound 2 (a white amorphous powder) was found to have a molecular formula of C 22 (Table 1) and HMQC spectra (S1 Fig.) revealed that compound 2 was very similar to atranorin (compound 1), except that it lacked the aldehyde signal and had additional signals for a ketone group at δ C 198.6 (C-10), a methyl group at δ H 2.28 (3H, s, H-11) and δ C 27.5, and two conjugated aromatic protons at 7.18 (1H, d, J = 16.5 Hz, H-9) and 7.85 (1H, d, J = 16.5 Hz, H-8), indicating the presence of a trans double bond (δ C 128.7 and 133.1). Extensive analysis of the HMBC spectrum (Fig. 2) indicated that the protons of the methyl group at δ H 2.28 were correlated with C-10 (δ C 198.6) and C-9 (δ C 128.7). Moreover, H-9 (δ H 7.18) and H-8 (δ H 7.85) were correlated with the C-10 carbonyl group. These data suggest the presence of a methylvinyl ketone chain. Finally, HMBC correlations from proton H-9 (δ H 7.18) to C-3 (δ C 107.6), together with correlations from H-8 (δ H 7.85) to C-2 (δ C 163.7) and C-4 (δ C 162.6) (Table 1) indicated that compound 2 was an atranorin derivative with a methylvinyl ketone chain substitution at position C-3. Accordingly, compound 2 was formulated as shown (Fig. 2) and was designated ( The structural similarity between compound 2 and atranorin suggested that this compound might be produced during the maceration of S. evolutum in acetone. We tested this hypothesis, by performing LC-MS analysis on freshly prepared crude acetone and ethyl acetate extracts of S. evolutum. Compound 2 was found in the crude acetone extract, but was not detected in the ethyl acetate extract (Fig. 3). Moreover, atranorin and the crude ethyl acetate extract were macerated for one week in acetone without the detection of even trace amounts of 2, indicating that the formation of compound 2 in acetone required the lichen to be present. Stictic acid and its methylvinyl ketone derivative, isidiophorin, were isolated from the acetone extract. These two compounds have been consistently reported to be isolated simultaneously in the acetone fraction [22][23][24]. This suggests that, like compound 2, isidiophorin was probably derived from stictic acid. Two newly reported depsides from the genus Stereocaulon, methyl-3'-methyllecanorate (3) and cladonioidesin (4) [25], were also identified. Interestingly, these depsides and compound 2 differed from atranorin (1) in terms of their C-3 substitutions. Indeed, the substituent on C-3 consisted of an aldehyde group in 1, a methylvinyl ketone moiety in 2, a proton in 3 and a carboxylic group in 4. We therefore synthesized two analogs, one bearing a methyl alcohol group and the other a methyl group (5 and 6 respectively) on C-3, to evaluate the influence of the degree of oxidation of the C-3 substituent on HCV inhibition. The synthesis of these molecules was based on the protocol of Pulgarin [26]. Reduction of the CHO group of atranorin generated compound 5, and esterification of β-orcinol carboxylic acid with compound 7 resulted in compound 6. Finally, four monoaromatic phenols (methyl-β-orcinol carboxylate (7), methyl orsellinate (8) [10], atranol (9) and methyl haematommate (10) [10]), were also isolated.
Thus, the phytochemical investigation of Stereocaulon evolutum extracts and the synthesis of two derivatives yielded a panel of 10 atranorin derivatives for evaluation on hepatitis C virus.

Antiviral activity of atranorin derivatives
We evaluated the anti-HCV activity of the 10 atranorin-related compounds, using a virus grown in cell culture (HCVcc) with a modification of its genome resulting in expression of the Renilla luciferase reporter gene [15]. Two control inhibitors, erlotinib and telaprevir, were systematically included in each experiment. These two compounds are known to interfere with the entry and replication steps of viral cycle, respectively. The maximum dose of each tested compound was 100 μM, because most of these molecules became insoluble in aqueous solution at higher concentrations. The first experimental procedure used involved the preincubation of human hepatic Huh-7.5.1 cells for 1 h with various doses of the molecules. The cells were then cultured in the presence of both the molecules tested and the virus, for an additional 48 h. At the end of the incubation period, we assessed cell viability and viral replication. This protocol made it possible to test the effects of molecules on both virus entry and replication. Initial observations indicated that most compounds were not cytotoxic ( Table 2 and S2 Fig.). For compounds 1, 4, 8, 9 and 10, high doses led to a minimal loss of signal, due to either weak toxicity or an antiproliferative effect similar to that observed for erlotinib, an epidermal growth factor receptor inhibitor. Conversely, compounds 3 and 5 were more toxic, with estimated halfmaximal cytotoxicity concentrations (CC 50 ) of 82 μM and 106 μM, respectively (Table 2). Remarkably, compound 6 tended to give a slightly, but reproducibly higher viability signal when used at high doses, with a peak at about 50 μM. This probably reflects a modest stimulation of cell proliferation, an effect due solely to the lichen metabolite, because a similar profile was observed in the absence of viral infection (data not shown). At concentrations greater than 50 μM, cytotoxic effects emerged, as shown by the decrease of the viability signal.
With the exception of cladonioidesin (4), all the atranorin derivatives inhibited HCV in a dose-dependent manner. Thus, the half-maximal inhibitory concentrations (IC 50 ) of lichen derivatives ranged between about 10 and 70 μM, whereas those of the controls were consistent with published values [27][28][29] (Table 2). The IC 50 values of these lichen metabolites were similar to those published for other natural molecules [4]. The high standard error of the mean (SEM) observed for the IC 50 of compound 2 probably resulted from the observed instability of the molecule in solution, even if dissolved immediately before use. Monoaromatic phenols and depsides had different IC 50 values ( Table 2). With the exception of compound 2, which was instable, depsides appeared to be more active than monoaromatic phenols, and this finding was confirmed statistically by an analysis of variance (Table 2, S1 Table). The lipophilicity of these lichen metabolites may govern their efficacy in the targeted cells, by influencing their passive diffusion across the cell membrane, and thus their intracellular bioavailability. We used ALOGPS 2.1 freeware to predict the theoretical partition coefficient (logP) of the studied   atranorin derivatives, to obtain an index of lipophilicity. Depsides, which had higher levels of antiviral activity, were also more lipophilic than monoaromatic phenols (S3 Fig.). By contrast, lipophilicity was not a discriminant parameter within the depside family tested. However, our results highlight the importance of the nature of the chemical group at position C-3 in depsides for their anti-HCV activity, as these molecules have identical chemical structures except for this position. The most active depsides (1, R = CHO, IC 50 = 22.3 μM; 5, R = CH 2 OH, IC 50 = 11.8 μM and 6, R = CH 3 , IC 50 = 13.3 μM) were selected for additional investigations. We excluded molecule 3 from further analysis because its selectivity index (SI, ratio of CC 50 to IC 50 ) was below that of compound 5 (Table 2).

Targeted stages of the HCV life cycle
An experimental procedure was designed to determine the inhibitor class (entry or replication) to which compounds 1, 5 and 6 belonged. HCVcc were added to the cell culture during a limited time window (around 17 h). The cells were then washed to eliminate viruses from the inoculum and viral replication was allowed to take place for an additional 30 h. The selected compounds were present in the culture medium during inoculation (I), replication (R), or both these phases (inoculation and replication, I/R). The reference drugs and compound 1 had no effect on cell viability, regardless of the incubation conditions (Fig. 4A). A small anti-proliferative effect was detectable only with high doses of erlotinib. The cytotoxic effect of compound 5 was also reproduced in these new conditions, with a similar CC 50 (~100 μM). Periods of more than 17 h (R and I/R) of incubation with compound 6 resulted in an enhancement of the cell viability signal over a narrow range of high concentrations, probably reflecting the capacity of this molecule to induce cell proliferation, as reported above. At a concentration of 100 μM, this compound induced some cytotoxicity, as demonstrated by a decrease in signal intensity, except in R conditions, in which the toxic effect was shifted to higher concentrations. In terms of antiviral activity, we expected to obtain two inhibition profiles for the luminescence signal, depending   (Fig. 4B). As shown with erlotinib, inhibitors of viral entry are most efficient if present at least during the initial interaction between cells and viruses (conditions I/R and I). The inhibitory effect on the establishment of replication (R conditions) observed at high doses was probably due principally to the inhibition of new infection cycles initiated by virions produced by cells infected with the inoculum. A different profile was obtained for telaprevir. This antiviral drug specifically targets the HCV protease, accounting for its inhibitory action regardless of the experimental conditions (I or R). Its presence in the medium from inoculation onwards prevented the priming of the infection process at the viral polyprotein maturation step, which normally takes place after the translation of the HCV genome issued from the infecting virus. This maturation step is also repeated during the replication phase, for amplification of the viral genome and virion assembly. It is therefore not surprising that inhibition was strongest when telaprevir was present throughout the viral lifecycle (I/R), as there was probably an additive effect. Surprisingly, differences were observed between the lichen metabolites tested. Compounds 5 and 6 had profiles typical of replication inhibitors, whereas compound 1 had effects similar to those of erlotinib (Fig. 4B). However, no inhibitory effect was observed when compound 1 was present in the culture medium only during inoculation. Had compound 1 been soluble enough to allow the evaluation of higher concentrations, its presence during the inoculation stage would probably have led to inhibition, as for erlotinib. These data again highlight the importance of the nature of the chemical moiety at C-3 in depsides for their anti-HCV activity. The chemical group in this position seems to influence not only the potency of the molecule, but also its mechanism of action. Further studies are required to identify the processes involved because, for example, we cannot rule out a role in the inhibition of viral entry for compounds 5 and 6, even though they appeared to act as replication inhibitors in this experiment. It could be hypothesized that the replacement of a methyl or hydroxymethyl by an aldehyde group in position C-3 would lead to a loss of replication inhibition, but maintenance of the ability to inhibit viral entry. We show here, for the first time that compounds from lichens, such as atranorin, may display antiviral activity against HCV. Atranorin has already been described as an antioxidant [30], antinociceptive [31], antimalarial [32] and anti-inflammatory [33,34] compound. Our findings reveal that this molecule and some of its derivatives are potential new lead compounds for the development of anti-HCV drugs. Further studies of the antiviral properties of these molecules are required, but it is already known that atranorin is a common secondary metabolite of lichens, available in great amounts and easy to purify from lichens with a worldwide distribution. The next step will be determining the mechanisms of action of these molecules, which may act directly on viral particles (virucidal effect) or indirectly through effects on viral or cellular factors involved in HCV propagation. Evaluations of the spectrum of activity of these molecules against the various viral genotypes will also be required. It will also be important to determine the capacity of these molecules to limit the emergence of mutant resistant viruses to assess their therapeutic potential. Such studies may also provide information about both the mechanism of action of these molecules and the viral factor targeted. Multitherapy approaches are required to treat hepatitis C [2], so studies exploring the combination of these molecules with the various existing drugs will be required, to determine whether particular combinations have additive, synergistic or antagonistic effects. Despite the tremendous progress in treatment achieved in recent years with the advent of DAAs, studies are continuing and new antiviral drugs are still being discovered [35][36][37][38][39]. For instance, new candidate drugs with different mechanisms of action are currently being investigated [40]. These drugs will make it possible to diversify the treatments on offer, to cut costs, to improve resistance control and to resolve the remaining difficult-to-treat cases, particularly those in non-responders with advanced fibrosis and re-infected transplant patients. This research falls within the scope of efforts to find affordable natural anti-HCV compounds for the treatment of hepatitis C in developing countries, where this disease remains a major public health problem. Results for cell viability are expressed as percentages (mean ± SEM; n = 3) of the mean values obtained for the control culture, in which cells were incubated with vehicle (0.1% DMSO) alone. Erlotinib and telaprevir were used as positive antiviral drug controls inhibiting the entry and replication steps, respectively. Results for viral replication are presented in Table 2 in the main manuscript as IC 50 mean values ± SEM. (DOC) S3 Fig. Relationship between antiviral activity and theoretical partition-coefficients of lichen compounds. One day after seeding at 50,000 cells/cm 2 , Huh-7.5.1 cells were incubated 48 h with HCVcc at a MOI~0.03 in the presence of various concentrations of the tested compounds which were already added to cell culture 1 h before. Viral replication was assessed at the end of the incubation periods to determine the respective IC 50 (mean ± SEM, n = 3, see Table 2 in the main manuscript) of each compounds. The antiviral activities (IC 50 ) of lichen metabolites (depsides in gray circles and monoaromatic phenols in open circles) were represented according to their theoretical partition-coefficients (logP) predicted with the free software ALOGPS 2.1. For clarity, three lichen metabolites were excluded from the analysis: compound 2 for its inaccurate IC 50 value due to its instability, and the inactive compounds 4 and 8.

Supporting Information
(DOC) S1 Protocol. Detailed protocol for extraction and isolation of lichen metabolites. (DOC) S1 Table. Statistical comparison of anti-HCV activity of lichen metabolites (DOCX)