Isolation of Flavonoids from Deguelia duckeana and Their Effect on Cellular Viability, AMPK, eEF2, eIF2 and eIF4E

Preparations of Deguelia duckeana, known in Brazil as timbó, are used by indigenous people to kill fish. Reinvestigation of its extracts resulted in the isolation and identification of 11 known flavonoids identified as 3,5,4’-trimethoxy-4-prenylstilbene (1), 4-methoxyderricidine (2), lonchocarpine (3), 4-hydroxylonchocarpine (4), 4-methoxylonchocarpine (5), 5-hydroxy-4’,7-dimethoxy-6-prenylflavanone (6), 4’-hydroxyisolonchocarpine (7), 4’-methoxyisolonchocarpine (8), 3’,4’,7-trimethoxyflavone (9), 3’,4’-methylenedioxy-7-methoxyflavone (10), and 2,2-dimethyl-chromone-5,4’-hydroxy-5’-methoxyflavone (11). Except for 1, 3, and 4 all of these flavonoids have been described for the first time in D. duckeana and the flavanone 6 for the first time in nature. Compounds 2, 3, 4, 7, 9, and 10 were studied for their potential to induce cell death in neuronal SK-N-SH cells. Only the chalcone 4 and the flavanone 7 significantly induced lactate dehydrogenase (LDH) release, which was accompanied by activation of caspase-3 and impairment of energy homeostasis in the MTT assay and may explain the killing effect on fish. Interestingly, the flavone 10 reduced cell metabolism in the MTT assay without inducing cytotoxicity in the LDH assay. Furthermore, the flavonoids 2, 3, 4, 7, and 10 induced phosphorylation of the AMP-activated protein kinase (AMPK) and the eukaryotic elongation factor 2 (eEF2). The initiation factor eIF4E was dephosphorylated in the presence of these compounds. The initiation factor eIF2alpha was not affected. Further studies are needed to elucidate the importance of the observed effects on protein synthesis and potential therapeutic perspectives.


Introduction
Flavonoids possess a broad variety of different biological activities, among which their antioxidative properties are of special interest. Reactive oxygen radicals are harmful to biomembranes, but can additionally play a role as mediators in different signaling pathways [1,2]. Depending on their structural features, flavonoids such as prenylated flavones or chalcones, can also be involved in cytotoxic processes [3,4].
The genus Deguelia which belongs to the Fabaceae family can be found in tropical South America and shows a predominance of prenylated flavonoids and stilbenes [5]. Prenylated stilbenes as well as The genus Deguelia which belongs to the Fabaceae family can be found in tropical South America and shows a predominance of prenylated flavonoids and stilbenes [5]. Prenylated stilbenes as well as chalcones and isoflavonoids have been isolated from Derris rariflora (syn. Deguelia nitidula) [6]. Similar compounds have also been found in the roots of Deguelia hatschbachii [7]. Stilbenes with a slight effect on seed germination and plant growth have been described from the leaves of Deguelia refuescens var. urucu [8], whereas prenylated isoflavonoids with a slight antibacterial and antifungal activity have been reported from Deguelia longeracemosa [9]. In leaves of Deguelia utilis prenylated chalcones and stilbenes, which have cytoprotective properties in a neuronal cell line, have been found [10].
Preparations of Deguelia duckeana (Fabaceae) are known in Brazil as timbó and used by indigenous people for killing fish. However, the compounds responsible for these cytotoxic effects as well as their mode(s) of action are still unknown. Initial phytochemical investigations revealed three chalcones and a stilbene derivative as constituents, but no studies on their cytotoxic activities have been performed [11]. This prompted us to reinvestigate D. duckeana and to perform the first studies on which way the isolated flavonoids may affect cell life using the neuronal cell line SK-N-SH.

The Flavonoids from D. duckeana Differently Influence Cell Viability
Because timbó is considered as neurotoxic to fish, the biologic effects of isolated compounds were tested in the neuronal cell line SK-N-SH. Chalcones are known for their cytotoxic activity [4], and the compounds 2, 3 and 4 were first evaluated for their cytotoxic potential. To study whether the 2",2"-dimethylpyrano-or the 3',4'-methylenedioxy structural elements play a role in the cytotoxic activity compounds 7, 9 and 10 were included in the study. The cells were incubated with the respective flavonoid (50 µM) for 24 h and cell death was determined by measuring the release of intracellular lactate dehydrogenase into the supernatant (Figure 2A,B). Only the chalcone 4 and the flavanone 7, which both possess a 4'-hydroxy group and a 2",2"-dimethylpyrano moiety, induced significant membrane damage, whereas the chalcones 2, 3 and the flavones 9 and 10 mediated no cytolytic effect. Compounds 4 and 7 were studied for their LDH release at different concentrations. No significant cell death was observed up to a 10 µM concentration, but only at 50 µM ( Figure 2B). To test whether cell death induced by compounds 4 and 7 is due to apoptosis, activation of caspase-3 was determined by immunoblotting. Cleavage of pro-caspase-3 to its active fragments was observed ( Figure 2C), indicating that cytotoxicity, as determined in Figures 2A,B, may be due to secondary necrosis and a consequence of apoptosis.
Results were verified by the MTT cellular viability assay ( Figure 3A). At concentrations ≥30 μM for compound 4 and 50 μM for compound 7, treated cells displayed a significantly reduced conversion of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) to its formazan salt, indicating a decline in cellular viability. These concentrations are sufficient to induce apoptosis ( Figure 2C) and secondary necrosis (Figures 2A,B). In contrast, compounds 2, 3, and 9 did not reduce conversion of the tetrazolium salt, confirming that these compounds have no effect on cellular viability ( Figure 3A). Interestingly, the flavone 10 which differed from compound 9 by the 3',4'methylenedioxy group diminished the reducing capacity of SK-N-SH cells, as observed by a decline To test whether cell death induced by compounds 4 and 7 is due to apoptosis, activation of caspase-3 was determined by immunoblotting. Cleavage of pro-caspase-3 to its active fragments was observed ( Figure 2C), indicating that cytotoxicity, as determined in Figure 2A,B, may be due to secondary necrosis and a consequence of apoptosis.
Results were verified by the MTT cellular viability assay ( Figure 3A). At concentrations ě30 µM for compound 4 and 50 µM for compound 7, treated cells displayed a significantly reduced conversion of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) to its formazan salt, indicating a decline in cellular viability. These concentrations are sufficient to induce apoptosis ( Figure 2C) and secondary necrosis (Figure 2A,B). In contrast, compounds 2, 3, and 9 did not reduce conversion of the tetrazolium salt, confirming that these compounds have no effect on cellular viability ( Figure 3A). Interestingly, the flavone 10 which differed from compound 9 by the 3',4'-methylenedioxy group diminished the reducing capacity of SK-N-SH cells, as observed by a decline in absorbance ( Figure 3A) without simultaneously killing the cells (Figure 2A). Reduced formazan conversion in the absence of cell death has been associated with impaired metabolism [19]. Despite the limiting data for drawing structure activity relationships the occurrence of the 3',4'-methylenedioxy group may be important for a potential decrease of the metabolic activity in SK-N-SH cells.
Molecules 2016, 21, 192 5 of 11 in absorbance ( Figure 3A) without simultaneously killing the cells (Figure 2A). Reduced formazan conversion in the absence of cell death has been associated with impaired metabolism [19]. Despite the limiting data for drawing structure activity relationships the occurrence of the 3',4'methylenedioxy group may be important for a potential decrease of the metabolic activity in SK-N-SH cells.

Flavonoids from D. duckeana Induce the Phosphorylation of AMPK and eEF2, But Not eIF2alpha
Cells with a low metabolism reduce very little MTT [20]. Therefore, we tested whether the flavone 10 attenuates the cellular metabolism by determining activation of AMPK, an energy-sensing enzyme closely involved in the regulation of energy homeostasis [21]. As shown in Figure 3B, treatment of cells with compound 10 resulted in an increase in AMPK-phosphorylation on its activating loop at threonine-172. Phosphorylation of AMPK at threonine-172 is a consequence of a conformational change due to increased AMP concentrations [22] and a prerequisite of AMPK to

Flavonoids from D. duckeana Induce the Phosphorylation of AMPK and eEF2, But Not eIF2alpha
Cells with a low metabolism reduce very little MTT [20]. Therefore, we tested whether the flavone 10 attenuates the cellular metabolism by determining activation of AMPK, an energy-sensing enzyme closely involved in the regulation of energy homeostasis [21]. As shown in Figure 3B, treatment of cells with compound 10 resulted in an increase in AMPK-phosphorylation on its activating loop at threonine-172. Phosphorylation of AMPK at threonine-172 is a consequence of a conformational change due to increased AMP concentrations [22] and a prerequisite of AMPK to adjust intracellular energy levels by inhibiting energy consuming pathways such as global protein synthesis [23,24]. However, compounds 2, 3, 4 and 7 also induced phosphorylation of AMPK ( Figure 3B) indicating that reduced formazan formation ( Figure 3A) and AMPK activation are here independent processes.
Activation of AMPK indicates inhibition of global protein synthesis [23]. To study a possible impact of the tested compounds on translation, we determined their potential to induce phosphorylation of eEF2 at threonine 56, which blocks translational elongation by an AMPK/eEF2K-dependent mechanism and thus protein synthesis [23]. As shown in Figure 4A, the flavonoids 2, 3, 4, 7, and 10 markedly induced phosphorylation of eEF2, which couples AMPK activation with reduced translational elongation. The stilbene 1 and the flavone 11 did not mediate this posttranslational modification of eEF2 at threonine 56. However, up to now no structure activity relationships can be drawn.
The cytotoxic compounds 4 and 7 also induced posttranslational repression of eEF2 ( Figure 4C), but at 50 µM eEF2 phosphorylation at threonine 56 declined in the presence of both compounds (data not shown). This is most probably due to the strong cytotoxic activity of these flavonoids (Figure 2A-C). Altogether, our results indicate, that the cytolytic potential of some flavonoids isolated from D. duckeana is not directly related to their potential to repress translational elongation.
Molecules 2016, 21,192 6 of 11 adjust intracellular energy levels by inhibiting energy consuming pathways such as global protein synthesis [23,24]. However, compounds 2, 3, 4 and 7 also induced phosphorylation of AMPK ( Figure  3B) indicating that reduced formazan formation ( Figure 3A) and AMPK activation are here independent processes. Activation of AMPK indicates inhibition of global protein synthesis [23]. To study a possible impact of the tested compounds on translation, we determined their potential to induce phosphorylation of eEF2 at threonine 56, which blocks translational elongation by an AMPK/eEF2Kdependent mechanism and thus protein synthesis [23]. As shown in Figure 4A, the flavonoids 2, 3, 4, 7, and 10 markedly induced phosphorylation of eEF2, which couples AMPK activation with reduced translational elongation. The stilbene 1 and the flavone 11 did not mediate this posttranslational modification of eEF2 at threonine 56. However, up to now no structure activity relationships can be drawn.
The cytotoxic compounds 4 and 7 also induced posttranslational repression of eEF2 ( Figure 4C), but at 50 μM eEF2 phosphorylation at threonine 56 declined in the presence of both compounds (data not shown). This is most probably due to the strong cytotoxic activity of these flavonoids (Figures  2A-C). Altogether, our results indicate, that the cytolytic potential of some flavonoids isolated from D. duckeana is not directly related to their potential to repress translational elongation. To test the influence of flavonoids from D. duckeana on the initiation phase in the translation process phosphorylation of eIF2alpha was determined in the presence of compounds 2, 3, 4, 7, 9 and 10. eIF2alpha is required in the initiation of translation, as it mediates binding of GTP and the initiator Met-tRNA to the ribosome to form the 43S preinitiation complex. Phosphorylation at serine 51 inactivates eIF2alpha and as a consequence translation comes to halt because initiation is abrogated. As shown in Figure 5A, none of the tested flavonoids were able to mediate posttranslational phosphorylation of eIF2alpha at serine 51. Furthermore, we tested the potential of the flavonoids 2, 3, 4, 7, and 10 on the phosphorylation of eIF4E. Activation of the eukaryotic initiation factor eIF4E is a rate limiting step in cap-dependent translation and is regulated among others by AMPK under energy starvation [25]. Our results, shown in Figure 5B, demonstrate that the flavonoids 2, 3, 4, 7 and To test the influence of flavonoids from D. duckeana on the initiation phase in the translation process, phosphorylation of eIF2alpha was determined in the presence of compounds 2, 3, 4, 7, 9 and 10. eIF2alpha is required in the initiation of translation, as it mediates binding of GTP and the initiator Met-tRNA to the ribosome to form the 43S preinitiation complex. Phosphorylation at serine 51 inactivates eIF2alpha and as a consequence translation comes to halt because initiation is abrogated. As shown in Figure 5A, none of the tested flavonoids were able to mediate posttranslational phosphorylation of eIF2alpha at serine 51. Furthermore, we tested the potential of the flavonoids 2, 3, 4, 7, and 10 on the phosphorylation of eIF4E. Activation of the eukaryotic initiation factor eIF4E is a rate limiting step in cap-dependent translation and is regulated among others by AMPK under energy starvation [25]. Our results, shown in Figure 5B, demonstrate that the flavonoids 2, 3, 4, 7 and 10 repress phosphorylation of eIF4E at serine 209, which is associated with its reduced activity. Phosphorylation of eIF4E at serine 209 is mediated by the ras/MAP-kinase pathway [25], indicating that flavonoids from D. duckeana probably also affect mitogenic or stress stimuli via ERK1/2 or p38 MAP-kinase.
Molecules 2016, 21,192 7 of 11 10 repress phosphorylation of eIF4E at serine 209, which is associated with its reduced activity. Phosphorylation of eIF4E at serine 209 is mediated by the ras/MAP-kinase pathway [25], indicating that flavonoids from D. duckeana probably also affect mitogenic or stress stimuli via ERK1/2 or p38 MAP-kinase.

Conclusions
Altogether, we could extend the knowledge on the flavonoid profile of D. duckeana which fits those known from other species in the genus Deguelia. The isolated flavonoids were shown to have various effects on cells. Compounds 4 and 7, which both possess a 2",2"-dimethylpyrano structure, may contribute to the reported killing of fish because of their proven cytotoxic effects. Moreover, we could demonstrate for the first time that flavonoids such as 2, 3, 4, 7, 9 and 10 affect phosphorylation of eEF2, AMPK and eIF4E and thus influence both translational initiation and elongation. Resulting consequences must be clarified by further experiments to elucidate the applicability of flavonoids in prospective therapeutic settings.

Conclusions
Altogether, we could extend the knowledge on the flavonoid profile of D. duckeana which fits those known from other species in the genus Deguelia. The isolated flavonoids were shown to have various effects on cells. Compounds 4 and 7, which both possess a 2",2"-dimethylpyrano structure, may contribute to the reported killing of fish because of their proven cytotoxic effects. Moreover, we could demonstrate for the first time that flavonoids such as 2, 3, 4, 7, 9 and 10 affect phosphorylation of eEF2, AMPK and eIF4E and thus influence both translational initiation and elongation. Resulting consequences must be clarified by further experiments to elucidate the applicability of flavonoids in prospective therapeutic settings.

Extraction and Isolation
The material from D. duckeana were dried for three days in a hot-air oven 50˝C and separately extracted with CH 2 Cl 2 for roots and with n-hexane for branches in a 2 L flask using an ultrasonic bath (Unique, Indaiatuba, São Paulo, Brazil) for 20 minutes and filtered. The extracts were concentrated in vacuo (40˝C). The CH 2 Cl 2 extract from the roots (8.  Table 1, HMBC and NOESY ( Figure S3).

Cell Culture and Treatment
The human neuronal cell line SK-N-SH was obtained from the American Type Culture Collection (Manassas, VA, USA) and maintained in Eagle's minimal essential medium (EMEM), supplemented with 10% fetal bovine serum (FBS), 100 IU/mL penicillin and 100 µg/mL streptomycin at 37˝C in a humidified atmosphere containing 5% CO 2 (Thermo Fisher Scientific, Waltham, MA, USA). Cells were treated with 0.1-50 µM of the flavonoids after synchronization in FBS-free EMEM containing 100 IU/mL penicillin and 100 µg/mL streptomycin overnight. Control cells were treated with the highest concentration of DMSO as solvent (0.1%).

Cytotoxicity Assay
SK-N-SH cells (1ˆ10 6 cells per well of a 6 well plate) were treated with the flavonoids for 24 h, before cellular supernatants were analyzed for lactate dehydrogenase (LDH) content by the Cytotoxicity Detection Kit (Roche Applied Science, Mannheim, Germany) according to the protocol of the manufacturer. The absorbance at 490 nm was measured by a microplate reader (Model 680, BIO-RAD, Munich, Germany) with the reference wavelength of 690 nm. Total LDH release (100%) was obtained by the treatment of cells with 2% Triton-X100. The relative LDH release is defined by the ratio of LDH released over total LDH in the intact cells.

Metabolic Activity
The MTT viability assay was performed as described by Mosmann [20] Briefly, SK-N-SH cells (50000 cells per well of a 96 well plate) were treated with the respective flavonoid for 24 h and subsequently incubated with 1 mg/mL of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) for 2 h at 37˝C before excess MTT was aspirated. The residual formazan crystals were dissolved in 100 µL of dimethylsulfoxide and quantified at 595 nm, using a microplate reader (Model 680, BIO-RAD, Munich, Germany).
After repeated washing, the specific bands were visualized using horseradish peroxidase-conjugated anti-rabbit IgG and enhanced chemiluminescence reagents (GE Healthcare, München, Germany).

Statistical Analysis
Data are shown as mean˘s.d. Statistical analysis was performed using the GraphPad Prism 5 software (GraphPad, La Jolla, CA, USA) and one-way or two way analysis of variance (ANOVA) followed by the Bonferroni post hoc test. Results were considered significant with p ď 0.05.