Natural-Product-Inspired Compounds as Countermeasures against

Article pubs.acs.org/jnp

Cite This: J. Nat. Prod. XXXX, XXX, XXX−XXX

Natural-Product-Inspired Compounds as Countermeasures against the Liver Carcinogen Aflatoxin B1 Adam C. Carter,† Jarrod B. King,† Allison O. Mattes,† Shengxin Cai,† Narender Singh,‡ and Robert H. Cichewicz*,† †

Natural Products Discovery Group, Institute for Natural Products Applications and Research Technologies, Department of Chemistry & Biochemistry, Stephenson Life Science Research Center, University of Oklahoma, 101 Stephenson Parkway, Norman, Oklahoma 73019, United States ‡ CFD Research Corporation, 701 McMillian Way, Suite D, Huntsville, Alabama 35806, United States Downloaded by UNIV AUTONOMA DE COAHUILA at 17:19:06:358 on May 28, 2019 from https://pubs.acs.org/doi/10.1021/acs.jnatprod.9b00290.

S Supporting Information *

ABSTRACT: Aflatoxin B1 (AfB1) ranks among the most potent liver carcinogens known, and the accidental or intentional exposure of humans and livestock to this toxin remains a serious global threat. One protective measure that had been proposed is employing small-molecule therapeutics capable of mitigating the toxicity of AfB1; however, to date, these efforts have had little clinical success. To identify molecular scaffolds that reduce the toxicity of AfB1, we developed a cell-based high-throughput high-content imaging assay that enabled our team to test natural products (pure compounds, fractions, and extracts) for protection of monolayers and spheroids composed of HepG2 liver cells against AfB1. The spheroid assay showed notable potential for further development, as it afforded greater sensitivity of HepG2 cells to AfB1, which is believed to better mimic the in vivo response of hepatocytes to the toxin. One of the most bioactive compounds to arise from this investigation was alternariol-9-methyl ether (1, purified from an Alternaria sp. isolate), which inspired the synthesis and testing of several structurally related molecules. Based on these findings, it is proposed that several types of natural and synthetic polyarene molecules that have undergone oxidative functionalization (e.g., compounds containing 3-methoxyphenol moieties) are promising starting points for the development of new agents that protect against AfB1 toxicity. flatoxins are highly potent carcinogenic metabolites produced by several types of fungi including Aspergillus f lavus and Aspergillus parasiticus.1,2 Since their discovery in the 1960s following the poisoning of over 100 000 turkeys and other poultry in England,1,3 a great deal of research has been carried out pertaining to this family of toxins with regard to their impact on human and animal health. Among the aflatoxins reported to date, aflatoxin B1 (AfB1) is widely considered to be one of the most carcinogenic naturally occurring compounds known to humankind.4 AfB1 and related aflatoxins are believed to be responsible for elevated hepatocellular carcinoma rates observed in many parts of the world including Africa, eastern Asia, and parts of South America.5 It is estimated that 4.5 billion people worldwide are exposed to aflatoxins in their diet and that aflatoxins are responsible for up to 28% of all hepatocellular carcinoma cases globally.6 The potency and ease of access to AfB1 and related analogues has sparked concerns that mycotoxins such as AfB1 may be employed or developed for use in biological warfare, as was the case in Iraq during the 1980s.7 In addition to their negative effects on human health, aflatoxin contamination has led to the loss of millions of dollars of agricultural products in the United States alone.8 Thus, AfB1 and its analogues

A

© XXXX American Chemical Society and American Society of Pharmacognosy

continue to play major roles in the global health and safety of humans and other animals. In adult humans, conversion of AfB1 to its chemically reactive form, AfB1-8,9-epoxide, through oxidative metabolism involving several isoforms of cytochrome P450 enzymes (CYP) (Figure 1), leads to many of the detrimental cancerrelated effects associated with AfB1 exposure. Several CYP isoforms have been implicated as responsible for the metabolic

Figure 1. Metabolic activation of aflatoxin B1 (AfB1) to chemoactive aflatoxin B1-8,9-epoxides. Received: March 29, 2019

A

DOI: 10.1021/acs.jnatprod.9b00290 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

compounds that afforded in vitro cellular protection against the toxic effects of AfB1.18,19 Further inspection showed that these reports focused exclusively on using colorimetric assays to measure cell metabolic activity (i.e., Alamar blue, MTT, intracellular ATP) as end points for assessing cell viability. Our preliminary studies indicated that these assay methods were imperfect because of their poor abilities to detect low levels of AfB1 toxicity, as well as their failure to assess overall cell health (e.g., cell morphology). Specifically, our microscopy-based observations of cells treated with AfB1 alone and with putative protective agents revealed two problems: (i) cells treated with even low-level (noncytotoxic) concentrations of AfB1 displayed substantial morphological changes that were not immediately reflected in their metabolic activity, and (ii) in some instances, the application of “protective” compounds to AfB1-treated cells resulted in viable, but severely altered hepatocytes that did not resemble healthy control cells (e.g., reduced cell areas, cells appeared angular as opposed to rounded, cells grew in isolation rather than appearing as islands of cells, and more). These observations alerted us that a different method was required to identify protective molecules that could maintain cell viability, while also sustaining healthy cell morphologies. Accordingly, we determined that a high-content imaging approach using an Operetta system (PerkinElmer) would provide a reasonable solution. Specifically, high-content imaging afforded the opportunity to simultaneously collect quantitative cell viability data, as well as qualitatively assess cell morphologies.20 To determine an appropriate cell density for high-content imaging studies, a series of experiments were performed examining the impact of different cell-seeding strategies on assay performance. It was determined that 2 × 104 HepG2 cells per well in a 96-well microtiter plate offered ample numbers of cells per field to reproducibly quantify cell viability, while it also prevented overcrowding, so that the morphologies of most individual cells were discernible. Under these assay conditions, healthy HepG2 cells grew as islands of polygonal-shaped cells (Figure 3). Using the Operetta image analysis software, we automated the determination of cell shape and area, which provided two important assay-design benefits. First, this approach allowed us to detect minor perturbations in cell shape (an indicator of overall cell health), and second, cell area could now be used as a metric for estimating viability. The correlation between cell area and viability was made known to us during the data exploration phase of our assay development process; we had observed that cell viability (as determined by MTT and SRB assays) of a cell population following AfB1 exposure was strongly correlated with average cell size. Thus, cell area was chosen as a novel surrogate metric for testing purposes because of its relationship to cell viability. Next, the concentration−response effects of AfB1 on HepG2 cell viability was investigated with the goal of determining the EC50 value of the toxin at 48 h post-treatment. An EC50 value of 30 μM was established for AfB1, which was in the range of other reported in vitro assay systems.18,19 The Z-factor for the assay was determined to be 0.82, indicating that the assay design provided a high likelihood of detecting bioactive samples under high-throughput screening conditions. Testing Two Putative AfB1 Protective Compounds: Oltipraz and N-Acetylcysteine. Previous reports for two compounds, oltipraz and N-acetylcysteine (Figure 2), suggested that these compounds offered protection from AfB1-induced cellular toxicity.21,22 However, these results have been called into question by a subsequent investigation12

activation of AfB1 including CYP3A4, CYP3A5, and CYP1A2; however, some uncertainty remains as to which CYP isoform(s) is the most critical for causing in vivo toxicity when genetic and comorbidity factors are considered.9−11 Once AfB1 is converted to its reactive electrophilic form, the compound reacts with cellular targets (e.g., DNA, RNA, and proteins), forming covalent bonds. One of the most problematic of these covalent modifications involves the reaction of AfB1-8,9-epoxide with DNA, which ultimately gives rise to mutations that increase the risk for tumor development, especially in the liver.2,11 Because of its severe toxic effects, research has focused on the development of protective measures to counteract the carcinogenic properties of AfB1. These efforts have largely followed two general strategies: (i) reducing exposure to the AfB1 and (ii) blocking its metabolic transformation to the activated epoxide. Among the described chemoprotective agents, two of the most widely reported are chlorophyllin and oltipraz (Figure 2), which appear to have limited efficacy

Figure 2. Structures of previously studied putative chemoprotective agents chlorophyllin, oltipraz, and N-acetylcysteine.

in humans at risk for dietary exposure to AfB1.12−16 While these compounds have not seen marked successful use in humans, they do serve as inspiration for the development of new and more effective chemoprotective agents that are capable of mitigating the effects of AfB1 toxicity. While AfB1’s metabolism, disposition, and clinical toxic manifestations vary among different animals, age groups, and people of different genetic backgrounds,11 our study was focused on addressing the constellation of acute and shortterm toxic effects imposed by AfB1 on human adults under the threat of receiving a bolus dose of the toxin. To address a key feature posed by AfB1 in the context of this threat (i.e., liver toxicity), we developed a bioassay system to enable the identification of new natural products that reduced the cellular toxicity of AfB1 toward human liver-derived cells (HepG2). Historically, the HepG2 cell line has been used as a proxy for hepatocyte function and drug metabolism since it is known to expresses several CYP enzymes, albeit at levels that are markedly lower than those observed in intact human liver tissues.17 In this report, we detail the development and testing of our new assay system and describe the identification of several natural and related synthetic molecules that afforded in vitro protection against AfB1 toxicity in monolayers and spheroids composed of HepG2 cells.

■

RESULTS AND DISCUSSION Assay Development and Optimization. At the outset of this study, our team identified a small number of published peer-reviewed methods that described the identification of B



Article

Figure 3. Two of the top bioactive samples (hits) from screening fungal and bacterial extracts for substances affording protective effects against AfB1 toxicity (30 μM) in monolayers of HepG2 cells. The cells simultaneously challenged with AfB1 and treated with the extract from plate 58, well B1 (source of compounds 1−3) exhibited partially restored cell morphologies compared to the cells treated with AfB1 alone (note how the cells treated only with AfB1 are smaller and more angular in appearance). In comparison, the cells treated with AfB1 and the extract from plate 59, well E6 (source of compounds 4 and 5) exhibited increased cell area, but were morphologically dissimilar to the cells receiving no AfB1 (vehicle-only treated controls).

Figure 4. Oltipraz and N-acetylcysteine provided no discernible protection from AfB1 toxicity in HepG2 cell monolayers. Error bars represent ±1 SD calculated from a two replicate measurements of live cell area.

bacterial extracts (tested at 33.4 μg/mL) and 644 fungal extracts (tested at 16.7 μg/mL) were screened. These concentrations were chosen as reasonable starting points based on previous experience applying our natural product extract library to screening programs conducted in the lab. While none of the bacterial samples were found to be active (even when the initial concentration was doubled from 33.4 μg/mL to 66.8 μg/mL), 21 of the fungal samples (3.3%) showed protective activities (protection was defined as the percent change in live cell area for cells treated with AfB1 and test compounds/extracts versus cells treated with vehicle only; compounds/extracts that afforded >50% changes in relative live cell areas were considered protective). A second round of testing was undertaken to confirm the protective effects of the bioactive samples resulting in 17 extracts that showed protective activities at 16.7 μg/mL (i.e., increased live cell area and partially restored HepG2 cell morphology so that cells

showing that administration of oltipraz causes adverse effects in human patients including gastrointestinal symptoms, phototoxicity, thermal sensitivity, and paresthesias. N-Acetylcysteine administration can cause adverse effects as well, ranging from nausea to death.23 In our assay system, neither compound showed protective effects against AfB1 (Figure 4), with oltipraz actually increasing the toxic effects of AfB1. Interestingly, we observed that both compounds caused increases in HepG2 cell area when administered alone, which may have been a contributing confounding factor in prior tests. Nevertheless, it is also possible that the protective effects of oltipraz and N-acetylcysteine are strictly manifested in vivo where other, system-level biological processes may be at play. Screening of Natural Product Extracts. An exploratory screen was carried out to test natural product extracts derived from bacteria and fungi for the presence of compounds that offered protection against AfB1 toxicity. For these tests, 276 C



Article

Figure 5. Structures of fungal secondary metabolites 1−6 tested in this study.

advantage of the supplies of these compounds that were already available in our pure compound library, we confirmed that 1 and 2 were indeed highly active, affording 67% and 66% maximum relative live cell areas (MRLCAs) following AfB1 exposure, while 3 did not protect the HepG2 cells (MRLCA = 15%) (Table 1). Further testing of molecules 1 and 2 in the

appeared more like untreated-control samples) (refer to Figure 3 for examples of cellular phenotypes generated by active samples). Seven of the bioactive fungal extracts that provided improvements in HepG2 cell morphology (AfB1-treated cells appeared similar to cells not exposed to AfB1) were selected, and the source fungi were subject to small-scale fermentation. Fungi were grown for 4 weeks in 250 mL flasks on monolayers of Cheerios breakfast cereal, at which point the cultures were extracted with EtOAc and subjected to partitioning against H2O. The organic phases were retained, the resulting mixtures of natural products were subjected to LC-MS analyses, and the natural-product-containing residues were tested in the assay to generate concentration−response curves (1.25 to 50 μg/mL). Three of the samples (laboratory codes 58B1, 55C11, and 59E6) showed good concentration−response effects; however, LC-MS revealed that two of the samples (58B1 and 55C11) contained nearly identical natural product profiles. This left two samples, 58B1 and 59E6 (Figure 3), which were selected for further study. Notably, sample 58B1 provided greater retention of HepG2 cell morphology compared to the vehicleonly controls, while sample 59E6, which afforded comparable live cell areas, lacked the ability to restore normal HepG2 cell morphology. Both fungal isolates were subjected to ITS sequencing, revealing their likely identities as an Alternaria sp. (source of extract 58B1) and an Aspergillus sp. (source of extract 59E6). Purification of Compounds 1−3 from Alternaria sp. Isolate 58B1. The small-scale extract produced from the Alternaria sp. (58B1) was subjected to HPLC chromatography over C18, and the resulting fractions were deposited into a deep-well 96-well plate. After removal of the solvent, the organic residues were resuspended in DMSO, and aliquots were transferred by a pin tool to fresh 96-well microtiter plates for bioassay testing. Our team has employed this methodology as a strategy for accelerating the process of purifying bioactive natural products from microorganisms and plants, and details are provided elsewhere.24 Three wells exhibited protection from AfB1 toxicity, and their contents were analyzed by LCMS. The first two active wells contained mixtures of two major-component molecules, while the third active well contained one major component. We were able to quickly identify all three compounds as alternariol-9-methyl ether (1), altenuisol (2), and alternariol (3) (Figure 5) by dereplication against our lab’s internal database of fungal metabolites. Taking

Table 1. Maximum Protection Afforded by Compounds 1− 14 from AfB1 Toxicity in HepG2 Monolayers, Expressed as Relative Live Cell Areas ±1 SD (Compared to Vehicle-Only Treated Cells)a compound 1 2 3 4 5 6 7 8 9 10 11 12 13 14

MRLCAb (%)

maximum concentration tested (μM)

± ± ± ± ± ± ± ± ± ± ± ± ± ±

50 50 0.5c 0.5c,d 5c 50 5c 12.5c 50 50 50 5c,d 0.5c,d 0.5c

67.0 65.7 15.2 64.9 20.5 68.3 37.0 61.6 53.9 60.2 41.0 63.9 72.0 33.5

1.1 4.7 5.8 6.8 2.0 14.5 0.1 6.1 1.5 8.2 3.4 0.7 4.6 0.3

When monolayers of cells were exposed to 30 μM AfB1, it resulted in relative live cell areas that were ∼30% compared to vehicle-only treated cells. Accordingly, MRLCA (maximum relative live cell areas) values of >50% were implemented as the threshold for substances to qualify as potentially protective against AfB1. bMRLCA values were obtained by taking the live cell area recorded for the concentration of a test substance that afforded the greatest total live cell area and dividing that value by the live cell area obtained for vehicle-only treated cells [(live cell area for treated cells/live cell area for cells treated with vehicle only) × 100%]; data are expressed as percentages ±1 SD, n = 2. cThe live cell area did not increase at higher concentrations tested. dToxicity was observed at 50 μM. a

absence of AfB1 showed that compound 2 was mildly cytotoxic to the HepG2 cell line, while 1 exhibited no overt cytotoxic effects. Upon comparing the structures of compounds 1 and 2 versus 3, we observed that the foremost difference between these metabolites was the presence of a methoxy group at the 9-position of 1 and 2. On the basis of these observations, we D



Article

Figure 6. Structures of synthetic and commercial compounds and 7−14 tested in this study.

CYP1A227,28 and are more readily accessible via synthesis than the polycyclic systems comprising compounds 1 and 2, we decided to explore the activities of additional benzophenones containing 3-methoxyphenol moieties. For these reasons, oxybenzone (8) and benzophenone-6 (9) (Figure 6) were tested using the HepG2 monolayers, resulting in MRLCAs of 61.6% (12.5 μM) and 53.9% (50 μM), respectively. This suggested that increasing the number of 3methoxyphenol units in a molecule did not necessarily lead to further increases in protection. Importantly, these observations provided evidence that benzophenones may serve an alternative scaffold for disrupting the cellular toxicity caused by AfB1. Published studies exploring other classes of compounds reported to afford protection against AfB1 (e.g., flavonoids and stilbenoids) revealed that small-molecule-dependent inhibition of CYP1A2 and functionally similar CYPs was highly reliant on the positioning of the pendant hydroxy and alkoxy substituents among these chemical scaffolds.29−31 To determine if an alternative substitution pattern on a benzophenone would increase or diminish protection, we synthesized and tested 4hydroxy-3-methoxybenzophenone (10) (Figure 6). This compound exhibited an MRLCA of 60.2% (50 μM), indicating that o-methoxyphenol units were acceptable for inclusion on this scaffold. However, when ketone 10 was reduced to yield 4hydroxy-3-methoxybenzhydrol (11) (Figure 6), the MRLCA resulting from this compound dropped to 41.0% (50 μM), suggesting that the carbonyl group is an important contributor to protective activity. Given the planarity of many reported CYP inhibitors, as well as the planarity of compounds 1, 2, and 4, we hypothesized that testing other planar aromatic scaffolds might prove valuable. Accordingly, we identified 1-hydroxy-3methoxyxanthone (12) as a candidate xanthone scaffold (Figure 6) to probe because of its structural similarity and increased planarity compared to 8. Compound 12 exhibited a similar MRLCA (63.9% at 5 μM) to that observed for 8; however, this compound also showed signs of toxicity, thus diminishing further interest in this compound series. While reviewing published accounts of small-molecule CYP inhibitors, the compound ritonavir (13)32 (Figure 6) caught

hypothesized that the presence of the methoxy moiety was an important contributor to the protective effects of 1 and 2. Purification of Compounds 4 and 5 from Aspergillus sp. Isolate 59E6. The extract produced from the small-scale preparation of the probable Aspergillus sp. isolate (59E6) was processed, and the fractions were tested for protection against AfB1 toxicity as described for the Alternaria sp. isolate (vide supra). Two fractions were identified as bioactive, and their contents analyzed by LC-MS. These wells contained relatively pure, single metabolites that were identified as rubrofusarin B (4) and funalenone (5) (Figure 5) by LC-MS-guided dereplication and comparisons with samples previously deposited in our pure compound library. Using the available purified material, we determined that 4 and 5 afforded 65% and 21% MRLCA values, respectively, at test concentrations of 50 μM (Table 1). Assessing the Protective Properties of Compounds 6−14. When examined collectively, metabolites 1−5 are relatively small, planar, aromatic molecules that are characteristic of the types of substrates reported as inhibitors of CYP1A2.25 These types of compounds are thought to readily fit in the relatively narrow binding pocket of this enzyme.26 On the basis of this observation, we investigated a series of structurally similar molecules as candidates for protecting HepG2 cells against AfB1-induced toxicity. Based on the lack of activity observed for compound 3, the methoxyphenol moiety found in compounds 1 and 2 was thought to be necessary to afford protective activity. Therefore, we proceeded to test another secondary metabolite from our library, monomethyl sulochrin (6) (Figure 5), which is characterized as a benzophenone. Compound 6 exhibited an MRLCA of 68.3% (50 μM), which was comparable to the results observed for the other bioactive fungal metabolites identified in our testing. However, when testing the synthetic compound 3,5-dimethoxyphenol (7) (Figure 6), which contains two m-methoxyphenol moieties, we observed very little activity, as evidenced by a low MRLCA value (37.0% at 5 μM with no increase at higher concentrations), suggesting that compound 7 may be too small and/or at least two aromatic rings might be required to provide activity. Since benzophenones are known substrates of E



Article

our attention since it represented a significant departure from the arene-containing systems we had investigated. Further fueling our interest in this compound was the fact that 13 is widely reported to inhibit CYP3A4, which has also been implicated in the activation of AfB1 into its toxic epoxide form.2,11 To test the hypothesis that 13 may protect cells from AfB1, its activity was assessed in our assay system. This test revealed that 13 afforded strong protection against AfB1 (72.0% relative live cell area at 5 μM), suggesting that CYP3A4 and its functional homologues may also be viable targets for limiting the toxicity of AfB1. These observations were supportive of earlier studies.9 Combining this observation with other reports citing nitrogen-containing heterocycles as inhibitors of CYPs,33−36 we attempted to combine a heterocycle system with an arene to determine if an amalgamation of these moieties would offer protection from AfB1-induced toxicity. Accordingly, the compound 2-hydroxy4-methoxy-N-(3-pyridinylmethyl)benzamide (14) (Figure 6) was prepared and tested. However, this molecule afforded little to no protection (33.5% MRLCA at 0.5 μM with no increase at higher concentrations tested) to HepG2 cells. Since our studies had already generated several other promising compounds, we did not pursue this or other classes of compounds further. Testing a HepG2 Spheroid Model for Assessing AfB1 Protection. While our assay system proved capable of detecting compounds that afforded protection to HepG2 cells that had been exposed to AfB1, we were aware that certain limitations remained in our approach. Specifically, CYP expression in HepG2 monolayers is known to be reduced in comparison to CYP expression/activity levels under in vivo conditions.17,19,37,38 Furthermore, methods to consistently and uniformly upregulate CYPs in vitro remain challenging,38 and our own experiments with several small-molecule promotors of CYP activity proved unreliable or provided only small increases in putative CYP activity (data not shown). Problems such as these were not only apparent in our studies but were observed in virtually every published account assessing the in vitro toxicity of AfB1 (i.e., the amount of toxin required to cause cell death was consistently higher than what was required to manifest toxicity in vivo).37,39 Considering these problems, we were intrigued by reports that CYP expression is increased in HepG2 cells when they are grown as spheroids.40,41 Spheroids, which are a form of 3D cell culture in which cells are allowed to aggregate free of a noncellular substrate,42 have been used in a number of assay systems because of their abilities to better mimic in vivo cell behaviors (e.g., gene expression, drug susceptibility, drug penetration, drug metabolism, morphology, and more).40−44 As such, we proceeded with further experiments utilizing HepG2 spheroids in place of the HepG2 monolayers. Our preliminary experiments creating and testing HepG2 spheroids revealed straightaway a remarkable 15× increase in the sensitivity of these cells to AfB1 (LC50 of 2 μM in spheroids versus 30 μM in monolayers). The increased sensitivity to AfB1 treatment appeared to be in agreement with the reported increase in CYP; however, CYP expression and activity levels were not directly quantified. Next, we proceeded to test several of the top-performing monolayer-protecting compounds (Table 1) in the spheroid model (Figure 7). Two of the topmost performing compounds, 1 and 12, afforded strong protection (95% and 72% relative live cell areas, respectively, at 10 μM) from AfB1 exposure (Figure 7). Moreover, micro-

Figure 7. Relative live cell areas of HepG2 spheroids treated with AfB1 and test compounds (10 μM). Compounds were considered protective if they enabled spheroids to achieve relative live cell areas that were >50% compared to vehicle-only treated spheroids. Error bars are expressed as ±1 SD, n = 3.

scopic analysis of the cells contained in the treated spheroids indicated that these cells retained morphologies that were indistinguishable from the unchallenged controls (Figure 8). Using our newly developed high-content assay, we screened a library of bacterial and fungal crude extracts, resulting in the identification of three fungus-derived natural products that afforded protection against AfB1 toxicity. In support of this assay, we found the images collected using the Operetta to be invaluable; these data enabled us to concentrate on active samples that preserved the natural cellular morphology of healthy liver cells. Using natural products as a starting point, we were able to identify other structurally similar fungal secondary metabolites that also exhibited strong protective activities. Additionally, we used synthetic methods to further explore bioactive scaffolds and obtain valuable insights concerning the structural requirements affecting activity and identify paths toward further scaffold optimization. While we did not confirm the mechanism(s) by which the compounds afford protection to HepG2 liver cells, the increased sensitivity of HepG2 spheroids to AfB1 and the increased protection afforded by these compounds in these spheroid assays suggested that some of the bioactive molecules may be acting wholly or in part through inhibition of one or more CYPs. This potential mechanism, as well as the potent protective activity observed for 13, a known CYP3A4 inhibitor, suggested that the use of selective or semiselective inhibitors of CYPs (e.g., CYP1A2, CYP3A4, and CYP3A5) may provide a viable route toward the development of chemoprotective measures that defend against AfB1 toxicity. Ideally, such compounds may be used to prevent liver damage and reduce the risk of acute liver toxicity and hepatocellular carcinoma in people and animals at risk for exposure to bolus doses of AfB1 and its toxic analogues.

■

EXPERIMENTAL SECTION

General Experimental Procedures. NMR data were collected on Varian 400 and 500 MHz NMR spectrometers. Accurate mass (HRESIMS) data were collected on a Waters SYNAPT G2-Si mass spectrometer. LC-MS data were obtained on a Shimadzu LC-MS 2020 system (ESI quadrupole) coupled to a photodiode array detector, with a Phenomenex Kinetex column (2.6 μm C18 column, 100 Å, 75 × 3.0 mm). The preparative HPLC system utilized SCL10A VP pumps and a system controller with a Luna 5 μm C18 column (110 Å, 250 × 21.2 mm, 10 mL/min), and the analytical and semipreparative HPLC system utilized Waters 1525 binary pumps with Waters 2998 photodiode array detectors and Luna 5 μm C18 F



Article

Figure 8. Protection afforded by compounds 1 and 12 from AfB1-induced toxicity in HepG2 spheroids. Spheroids were treated with calcein AM and Hoechst prior to imaging. In live spheroids, calcein AM was converted to the green-fluorescent calcein product after hydrolysis by intracellular esterases, whereas, in dead cells, calcein was not produced and only the Hoechst-stained (blue) DNA was observed. columns (110 Å, 250 × 4.6 mm, 1 mL/min and 110 Å, 250 × 10 mm, 4 mL/min). All solvents were of ACS grade or better. Chemicals. Thionyl chloride and 2-hydroxy-4-methoxybenzoic acid were purchased from Alfa Aesar. Benzophenone-6 (9), 3picolylamine, 2-(trimethylsilyl)phenyl trifluoromethanesulfonate, and 2,4,6-trihydroxybenzoic acid were purchased from TCI America. Cesium fluoride, vanillic acid, and oxybenzone (8) were purchased from Beantown Chemicals. Oltipraz, dimethyl sulfate, vanillin, and phenylmagnesium bromide were purchased from Sigma-Aldrich. Sodium azide was purchased from Fluka, and potassium carbonate was purchased from Fisher Scientific. Ritonavir (14) was purchased from Selleck Chemicals. N-Acetylcysteine was purchased from VWR. 3,5-Dimethoxyphenol (7) was purchased from Chem Impex International. Aflatoxin B1 Protection Assay. The human hepatocellular carcinoma cell line, HepG2 (HB-8065), was purchased from ATCC. The cells were maintained in Eagle’s minimum essential medium (EMEM) (Gibco) with 5% FetalClone III (Hyclone) and penicillin/streptomycin (50 U/mL, 50 μg/mL) in a humidified incubator in a 5% CO2 atmosphere. Monolayer-based cell assays were performed in 96-well plates by seeding 20 000 cells per well in 100 μL of medium and allowing the cells to attach overnight. Test compounds were added the next day to the wells from freshly prepared DMSO stock solutions. Immediately after compound addition, 100 μL aliquots of a stock solution made from complete EMEM and AfB1 in DMSO were added to the wells, bringing the total well volumes to 200 μL. The cells were incubated for 48 h, at which time viability was determined using a calcein AM and Hoechst 33342 live cell area assay on an Operetta high-content analysis system. To perform this assay, a stock solution of 40 μM calcein AM and 160 μM Hoechst 33342 was prepared in DMSO. The stock solution was

diluted at a ratio of 1:5 in phosphate-buffered saline (PBS), and 5 μL of the diluted solution was added to each well of the microtiter plate. Plates were incubated for 30 min, and the contents of each well were imaged using Operetta. Harmony software was used to calculate the live cell areas by finding all Hoechst-labeled nuclei and assigning live or dead status to each cell based on a threshold of green fluorescence (set to 350 fluorescent units). This assessment was made possible because live cells contain esterases that can cleave the acetoxymethyl (AM) ester group of calcein AM, resulting in the dye’s characteristic bright green glow inside the cells. Vehicle-only treated controls (DMSO) were used to calculate 100% culture viability. The protection afforded by each compound was expressed as the live cell area of the treated cells relative to the live cell area of the vehicleonly controls and expressed as a percentage. A compound was considered protective if the relative live cell area was >50% since ∼30% live cell area was typically observed for the AfB1-treated samples. The spheroid cell assay was carried out in a similar manner with the major difference being the assays were conducted using ultralow attachment plates (Corning 7007). Aliquots of 1000 cells were seeded into each well containing 100 μL of EMEM, and the cells were grown for 4 days until a tight spheroid of cells had formed. Spheroids were treated with test compounds and 100 μL of complete EMEM with AfB1 added. Spheroids were grown for three additional days, followed by viability and live spheroid size determination using the Operetta high-content analysis system. Screening Natural Product Extracts. A collection of 644 fungal extracts dissolved in DMSO was screened at 16.7 μg/mL. Test samples were transferred to test plates using a 1 μL pin tool. The protective activities of the extracts were assessed as described above for the AfB1 protection assay (vide supra). Additionally, a collection of G



Article

276 mammalian microbiome-derived bacterial extracts45 (solubilized in DMSO) was screened at 33.4 and 66.8 μg/mL using a 2 μL pin tool. All samples were screened in duplicate. The Operetta system was used to collect two fields-of-view per well, and the images were saved for analysis. Active samples were defined as extracts that afforded protection, resulting in cells that most closely approximated the vehicle-only treated controls (i.e., extracts and compounds that provided >75% relative live cell area). Purification of Compounds 1−3. The fungal isolate was identified as an Alternaria sp. (internal designation 58B1) based on ITS sequence analysis (GenBank accession number MK038723) and comparison to the BLAST database. The isolate had been obtained via an oral swab taken from a roadkill opossum.45 The fungal isolate was grown for 4 weeks on Cheerios breakfast cereal supplemented with a 0.3% sucrose solution with 0.005% chloramphenicol in a 250 mL flask, at which point it was extracted with EtOAc. The EtOAc extract (55 mg) was prepared at a concentration of 100 mg/mL, and 5 mg (50 μL injection volume) was subjected to semipreparative C18 HPLC (250 mm × 10 mm, 5 μm) with a MeOH/H2O gradient from 10% to 100% MeOH in 30 min with a 10 min MeOH wash. Fractions were collected in a deep-well 96-well plate, and the solvent was removed under vacuum. Each well of the 96-well plate was tested for activity, resulting in three wells that were bioactive in the AfB1 protection assay. Analysis of the PDA chromatogram revealed that each active well corresponded to a well-delineated peak(s). LC-MS analysis of the contents of these wells provided strong matches (i.e., retention times, UV-absorption profiles, and MS data) to compounds already cataloged in our laboratory’s pure compound library (compounds 1−3). To obtain additional material for compound 1, the fungal isolate was grown for 4 weeks on Cheerios breakfast cereal supplemented with a 0.3% sucrose solution with 0.005% chloramphenicol in three large mycobags (Unicorn Bags, Plano, TX, USA). The resulting fungal biomass was homogenized and extracted with EtOAc. The crude extract (28.3 g) was subjected to silica gel vacuum liquid chromatography with elution steps of 1:1 hexanes/DCM, DCM, 10:1 DCM/MeOH, and MeOH, yielding four fractions. The MeOH fraction (7.3 g) was subjected to HP20SS vacuum liquid chromatography and eluted with a step gradient of MeOH in H2O (30%, 50%, 70%, 90%, 100%) and 1:1 DCM/MeOH, yielding a total of six fractions. Upon storage in MeOH at −20 °C, a precipitate was observed in the 100% MeOH fraction from HP20SS VLC, which was removed by centrifugation. The precipitate was subjected to semipreparative HPLC using isocratic elution conditions (60:40 acetonitrile (ACN)/H2O) to yield compound 1 (126 mg), which was identified based on comparison of its HRESIMS, as well as 1H and 13 C NMR data, to values reported in the literature.46 Purification of Compounds 4 and 5. The fungal isolate was identified as an Aspergillus sp. (internal designation 59E6) based on ITS sequence analysis (GenBank accession number MK038724) and comparison to the BLAST database. The fungus was obtained from a soil sample obtained from Malibu, California, as part of the University of Oklahoma, Citizen Science Soil Collection Program.47 The fungal isolate was grown and extracted in the same manner as above to yield 108 mg of fungal extract. This material was subjected to semipreparative HPLC as described above to give fractions that were collected in a 96-well plate. Each well of the 96-well plate was tested for activity, producing two wells that were bioactive in the AfB1 protection assay. Analysis of the PDA chromatograms for the contents of both wells revealed that each active well corresponded to a welldelineated peak. LC-MS analysis of the contents of these wells provided strong matches (i.e., retention times, UV-absorption profiles, and MS data) to compounds already cataloged in our laboratory’s pure compound library (compounds 4 and 5). Synthesis of 4-Hydroxy-3-methoxybenzophenone (10). To a 250 mL round-bottom flask, 1.0 g of vanillin (4.3 mM), 800 mg of sodium azide (12.3 mM), and 15 mL of ACN was added. The mixture was stirred for 5 min, followed by the slow addition of 2 mL of trifluoromethanesulfonic acid (22.6 mM) over a 5 min period. The mixture was stirred for 30 min; then 15 mL of ACN was added, and the crude reaction mixture was placed under vacuum to remove the

solvent. The resulting residue was redissolved in 80 mL of EtOAc and partitioned 3× with 35 mL of H2O. The organic phase was removed under vacuum to yield 1.0 g of crude vanillonitrile. To a dry 20 mL glass vial was added 2 mL of 3 M phenylmagnesium bromide in diethyl ether (6 mM) (diluted with 1 mL of dry THF). The phenylmagnesium bromide was cooled to 0 °C, and 100 mg of crude vanillonitrile in 0.6 mL of dry THF was added dropwise over a 5 min period. The reaction mixture was removed from the ice bath and stirred at room temperature for 25 h. The reaction was heated to reflux for 21 h. Following reflux, the vial was removed from the heat source, the reaction quenched with 3 mL of saturated sodium bicarbonate, and the reaction mixture partitioned against EtOAc (3× 10 mL). The solvent was removed under vacuum and subjected to silica gel vacuum liquid chromatography using a stepwise gradient (10:1 hexanes/EtOAc, 1:1 DCM/hexanes, DCM, 10:1 DCM/MeOH, and MeOH; 300 mL each). The DCM fraction was retained, and the solvent removed under vacuum. The resulting residue was subjected to C18 semipreparative HPLC under isocratic conditions with 45% ACN in H2O to yield 10 (9.2 mg, 6.7% overall yield). The identity of 10 was verified by comparison of its HRESIMS, as well as 1H and 13C NMR spectra to published values.48 Synthesis of 4-Hydroxy-3-methoxybenzhydrol (11). Compound 11 was prepared as described in a previous report with some slight modifications.49 Specifically, to a 20 mL glass vial were added 290 mg of vanillin (1.25 mM) and 1.5 mL of dry tetrahydrofuran (THF), and the mixture was cooled to 0 °C. To the cooled vanillin solution, 1.9 mL of 3 M phenylmagnesium bromide in diethyl ether (5.7 mM) was added, and the mixture was further diluted in 1 mL of dry THF (added dropwise over a 5 min period with stirring). The mixture was stirred at 0 °C for an additional 10 min. The vial was removed from the ice bath and held at room temperature for an additional 21 h. The reaction was quenched with the slow addition of 2 mL of MeOH, and 11 (14.9 mg, 3.4% yield) was purified by C18 semipreparative HPLC. The identity of 11 was verified through comparisons of its HRESIMS, as well as 1H and 13C NMR spectra, to published values.49 Synthesis of 1-Hydroxy-3-methoxyxanthone (12). The preparation of compound 12 was initiated by adding 160 mg of 2,4,6-trihydroxybenzoic acid (0.94 mM), 240 mg of potassium carbonate (1.7 mM), and 5 mL of acetone to a 20 mL glass vial. To this mixture was added 0.16 mL of dimethyl sulfate (1.7 mM), and the mixture was stirred at room temperature for 19 h. To the flask was added 10 mL of H2O, and the reaction mixture partitioned against EtOAc (3 × 10 mL). The organic phases were combined, and the solvent was removed under vacuum. The crude product mixture was purified by C18 semipreparative HPLC under isocratic conditions (40% ACN in H2O) to yield methyl 2,6-dihydroxy-4-methoxybenzoate (88 mg, 47% yield), whose identity was verified by comparison of its HRESIMS data and 1H NMR spectrum to published values.50 To a 20 mL glass vial were added 9 mg of methyl 2,6-dihydroxy-4methoxybenzoate (0.045 mM), 110 mg of cesium fluoride (0.73 mM), 25 μL of 2-(trimethylsilyl)phenyl trifluoromethanesulfonate (0.10 mM), and 1 mL of THF. The mixture was refluxed with stirring for 24 h. After cooling, MeOH and dichloromethane (DCM) were added, and the solvent was removed under vacuum. The residue was redissolved in diethyl ether and washed with H2O (3×). The organic phase was removed under vacuum, and the product was purified from the reaction mixture by C18 semipreparative HPLC (gradient 10− 100% ACN in H2O with 0.1% formic acid) to yield 12 (1.1 mg, 10% yield). The identity of 12 was confirmed by comparison of its HRESIMS data and 1H NMR spectrum to published values.51 Synthesis of 2-Hydroxy-4-methoxy-N-(3-pyridinylmethyl)benzamide (14). To a 2 mL glass vial were added 14 mg of 2hydroxy-4-methoxybenzoic acid (0.08 mM), 20 μL of 3-picolylamine (0.20 mM), 80 μL of triethylamine (0.57 mM), 20 μL of thionyl chloride (0.27 mM), and 400 μL of DCM. The mixture was stirred at room temperature for 23 h; then the solvent was removed under vacuum. The resulting material was redissolved in a mixture composed of equal volumes of DCM and 1 M aqueous hydrochloric acid (3×). The organic solvent was removed under vacuum, and the H



Article

(13) Mycotoxin control in low- and middle-income countries (IARC Working Group Reports, No. 9); International Agency for Research on Cancer: Lyon (FR), 2015. (14) Deng, J.; Zhao, L.; Zhang, N.; Karrow, N.; Krumm, C.; Qi, D.; Sun, L. Mutat. Res., Rev. Mutat. Res. 2018, 778, 79−89. (15) Kensler, T.; He, X.; Otieno, M.; Egner, P.; Jacobson, L.; Chen, B.; Wang, J.; Zhu, Y.; Zhang, B.; Wang, J.; Wu, Y.; Zhang, Q.; Qian, G.; Kuang, S.; Fang, X.; Li, Y.; Yu, L.; Prochaska, H.; Davidson, N.; Gordon, G.; Gorman, M.; Zarba, A.; Enger, C.; Muñoz, A.; Helzlsouer, K. Biomarkers Prev 1998, 7, 127−134. (16) Egner, P.; Wang, J.; Zhu, Y.; Zhang, B.; Wu, Y.; Zhang, Q.; Qian, G.; Kuang, S.; Gange, S.; Jacobson, L.; Helzlsouer, K.; Bailey, G.; Groopman, J.; Kensler, T. Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 14601−14606. (17) Gerets, H.; Tilmant, K.; Gerin, B.; Chanteux, H.; Depelchin, B.; Dhalluin, S.; Atienzar, F. Cell Biol. Toxicol. 2012, 28, 69−87. (18) Chen, X.; Murdoch, R.; Shafer, D.; Ajuwon, K.; Applegate, T. J. Appl. Toxicol. 2016, 36, 1437−1445. (19) Yokoyama, Y.; Sasaki, Y.; Terasaki, N.; Kawataki, T.; Takekawa, K.; Iwase, Y.; Shimizu, T.; Sanoh, S.; Ohta, S. Biol. Pharm. Bull. 2018, 41, 722−732. (20) Ang, M.; Pethe, K. Cytometry, Part A 2016, 89, 755−760. (21) Bammler, T.; Slone, D.; Eaton, D. Toxicol. Sci. 2000, 54, 30− 41. (22) Valdivia, A.; Martínez, A.; Damián, F.; Quezada, T.; Ortíz, R.; Martínez, C.; Llamas, J.; Rodríguez, M.; Yamamoto, L.; Jaramillo, F.; Loarca-Piña, M.; Reyes, J. Poult. Sci. 2001, 80, 727−734. (23) Sandilands, E.; Bateman, D. Clin. Toxicol. 2009, 47, 81−88. (24) Cai, S.; Risinger, A.; Nair, S.; Peng, J.; Anderson, T.; Du, L.; Powell, D.; Mooberry, S.; Cichewicz, R. J. Nat. Prod. 2016, 79, 490− 498. (25) Sansen, S.; Yano, J.; Reynald, R.; Schoch, G.; Griffin, K.; Stout, C.; Johnson, E. J. Biol. Chem. 2007, 282, 14348−14355. (26) Dutkiewicz, Z.; Mikstacka, R. Bioinorg. Chem. Appl. 2018, 2018, 1−21. (27) Watanabe, Y.; Kojima, H.; Takeuchi, S.; Uramaru, N.; Sanoh, S.; Sugihara, K.; Kitamura, S.; Ohta, S. Toxicol. Appl. Pharmacol. 2015, 282, 119−128. (28) Takemoto, K.; Yamazaki, H.; Nakajima, M.; Yokoi, T. Mutat. Res., Genet. Toxicol. Environ. Mutagen. 2002, 519, 199−204. (29) Shimada, T.; Tanaka, K.; Takenaka, S.; Murayama, N.; Martin, M.; Foroozesh, M.; Yamazaki, H.; Guengerich, F.; Komori, M. Chem. Res. Toxicol. 2010, 23, 1921−1935. (30) Mikstacka, R.; Wierzchowski, M.; Dutkiewicz, Z.; GielaraKorzańska, A.; Korzański, A.; Teubert, A.; Sobiak, S.; Baer-Dubowska, W. MedChemComm 2014, 5, 496−501. (31) Takemura, H.; Itoh, T.; Yamamoto, K.; Sakakibara, H.; Shimoi, K. Bioorg. Med. Chem. 2010, 18, 6310−6315. (32) Zeldin, R.; Petruschke, R. J. Antimicrob. Chemother. 2003, 53, 4−9. (33) Sevrioukova, I.; Poulos, T. Dalton Trans 2013, 42, 3116−3126. (34) Riley, R.; Parker, A.; Trigg, S.; Manners, C. Pharm. Res. 2001, 18, 652−655. (35) Kaur, P.; Chamberlin, A. R.; Poulos, T.; Sevrioukova, I. J. Med. Chem. 2016, 59, 4210−4220. (36) Halpert, J. Annu. Rev. Pharmacol. Toxicol. 1995, 35, 29−53. (37) Nakamura, K.; Aizawa, K.; Aung, K.; Yamauchi, J.; Tanoue, A. Sci. Rep. 2017, 7, 41093. (38) Herzog, N.; Katzenberger, N.; Martin, F.; Schmidtke, K. Journal of Cellular Biotechnology 2015, 1, 15−26. (39) Lauschke, V.; Hendriks, D.; Bell, C.; Anderson, T.; IngelmanSundberg, M. Chem. Res. Toxicol. 2016, 29, 1936−1955. (40) Terashima, J.; Goto, S.; Hattori, H.; Hoshi, S.; Ushirokawa, M.; Kudo, K.; Habano, W.; Ozawa, S. Drug Metab. Pharmacokinet. 2015, 30, 434−440. (41) Takahashi, Y.; Hori, Y.; Yamamoto, T.; Ohara, Y.; Tanaka, H. Biosci. Rep. 2015, 35, No. e00208. (42) Fennema, E.; Rivron, N.; Rouwkema, J.; van Blitterswijk, C.; de Boer, J. Trends Biotechnol. 2013, 31, 108−115.

residue subjected to C18 semipreparative HPLC (gradient 10−100% ACN in H2O with 0.1% formic acid) to yield 14 (1.4 mg, 7% yield). The identity of 14 was verified through comparisons of its HRESIMS data and 1H NMR spectrum to published values.52

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jnatprod.9b00290.

■

NMR spectra for compounds 10−12 and 14 (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: 405-325-6969. Fax: 405325-6111. ORCID

Robert H. Cichewicz: 0000-0003-0744-4117 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This material is based upon work supported by the Army Contracting Command, Aberdeen Proving Ground - Natick Contracting Division (ACC-APG NCD) under Contract No. W911QY-17-C-0008. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the ACCAPG NCD. The Shimadzu LC-ESIMS instrument used for this project was provided in part by a Challenge Grant from the Office of the Vice President for Research, University of Oklahoma, Norman Campus, and an award through the Shimadzu Equipment Grant Program. The Waters LC-ESIMS instrument was obtained in part through an NSF MRI equipment grant (1626372). We are grateful for the contribution of the soil sample from citizen scientist L. Brewster, from which the Aspergillus sp. isolate was obtained. We appreciate the helpful comments and advice provided by J. Phillips during the design and performance of these experiments.

■

REFERENCES

(1) Asao, T.; Büchi, G.; Abdel-Kader, M.; Chang, S.; Wick, E.; Wogan, G. J. Am. Chem. Soc. 1965, 87, 882−886. (2) Eaton, D.; Gallagher, E. Annu. Rev. Pharmacol. Toxicol. 1994, 34, 135−172. (3) Monson, M.; Coulombe, R.; Reed, K. Agriculture 2015, 5, 742− 777. (4) Squire, R. Science 1981, 214, 877−880. (5) Kew, M. J. Gastrointestin. Liver Dis. 2013, 22, 305−310. (6) Liu, Y.; Wu, F. Environ. Health Perspect. 2010, 118, 818−824. (7) Bennett, J.; Klich, M. Clin. Microbiol. Rev. 2003, 16, 497−516. (8) Klich, M. Mol. Plant Pathol. 2007, 8, 713−722. (9) Gallagher, E.; Kunze, K.; Stapleton, P.; Eaton, D. Toxicol. Appl. Pharmacol. 1996, 141, 595−606. (10) Ramsdell, H.; Parkinson, A.; Eddy, A.; Eaton, D. Toxicol. Appl. Pharmacol. 1991, 106, 436−447. (11) World Health Organization & Joint FAO/WHO Expert Committee on Food Additives (83rd, G., Switzerland); Safety evaluation of certain contaminants in food; World Health Organization: Malta, 2018. (12) Sudakin, D. J. Toxicol., Clin. Toxicol. 2003, 41, 195−204. I



Article

(43) Sant, S.; Johnston, P. Drug Discovery Today: Technol. 2017, 23, 27−36. (44) Shah, U.-K.; Mallia, J. d. O.; Singh, N.; Chapman, K. E.; Doak, S. H.; Jenkins, G. J. S. Mutat. Res., Genet. Toxicol. Environ. Mutagen. 2018, 825, 51−58. (45) Motley, J.; Stamps, B.; Mitchell, C.; Thompson, A.; Cross, J.; You, J.; Powell, D.; Stevenson, B.; Cichewicz, R. J. Nat. Prod. 2017, 80, 598−608. (46) Lou, J.; Yu, R.; Wang, X.; Mao, Z.; Fu, L.; Liu, Y.; Zhou, L. Braz. J. Microbiol. 2016, 47, 96−101. (47) Du, L.; Robles, A. J.; King, J. B.; Powell, D. R.; Miller, A. N.; Mooberry, S. L.; Cichewicz, R. H. Angew. Chem., Int. Ed. 2014, 53, 804−809. (48) Huang, J.; Guo, Y.; Jiang, J.; Liu, H.; Ji, Y. Res. Chem. Intermed. 2015, 41, 7115−7124. (49) Cohrt, A.; Jensen, J.; Nielsen, T. Org. Lett. 2010, 12, 5414− 5417. (50) Yuen, T.; Brimble, M. Org. Lett. 2012, 14, 5154−5157. (51) Genoux-Bastide, E.; Lorendeau, D.; Nicolle, E.; Yahiaoui, S.; Magnard, S.; Di Pietro, A.; Baubichon-Cortay, H.; Boumendjel, A. ChemMedChem 2011, 6, 1478−1484. (52) Petrlikova, E.; Waisser, K.; Palat, K.; Kunes, J.; Kaustova, J. Chem. Pap. 2011, 65, 52−59.

J


Natural-Product-Inspired Compounds as Countermeasures against

Recommend Documents