The scavenging of 1,1-diphenyl-2-picrylhydrazyl (DPPH) radicals was used in this investigation, as antioxidants act to inhibit this oxidation. In the DPPH study, antioxidants are able to reduce the stable DPPH radicals to the yellow colored diphenyl-picrylhydrazine. Compound 4 presented a minor inhibitory effect of 14.6% at 100 μM in the DPPH assay compared to vitamin C (Table 1). The ferrous ion chelating activities of L. tulipifera compounds are also shown in Table 1. Ferrozine quantitatively produced complexes with Fe²⁺. In the presence of chelating reagents, such as the test samples, the complex is disturbed resulting in a lightening of the typical red color. Compound 4 showed a minor level of Fe²⁺ scavenging effect of 14.7% whereas EDTA presented a strong scavenging ability. In the reducing power assay, the color of the test solutions changed from yellow to different shades ranging between green and blue depending upon the anti-oxidative capacities of these antioxidants. The presence of compounds 1 to 4, similar to the reference antioxidant substances, induced the reduction of the Fe³⁺/ferricyanide complex to the ferrous form. Table 1 summarizes the reducing power of the four pure constituents at 100 μM, which was moderate compared to 3-tert-butyl-4-hydroxyanisole (BHA) at the same dose (OD₇₀₀ = 0.98). Next, we measured the inhibitory effect of four constituents in an in vitro mushroom tyrosinase inhibition assay (Table 1). Compared to the inhibition of 69.2% displayed by kojic acid that is a commonly used human tyrosinase inhibitor in the cosmetic industry compound 1 showed a medium to minor (20.5%) inhibition of mushroom tyrosinase .

Nowadays the main reasons behind the ineffective medical therapies responsible for the high death ratios to cancer patients are from malignant cancer proliferation and metastasis. Therefore, it is vital and valuable for biologists to develop novel medicinal agents for anti-cancer treatment improving the effectiveness of treatment. The melanoma cell line A375.S2 used in the study, is metastatic and widely used in many anti-melanoma studies [20,33]. Accordingly, A375.S2 is suitable as a representative target in the study. The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was used to investigate the stimulating of cell death caused by the tested compounds. The cell proliferation assay examined the anti-cell proliferation of L. tulipifera compounds on human melanoma A375.S2 cells after a 24 h treatment. The samples were treated with 100 μM of four compounds as demonstrated in Figure 2. We observed that cells treated with the tested concentrations of compounds 1 and 4 exhibited more than 50% of cell inhibitory effects for up to 24 h.

We next examined the effects of the four compounds on cell cycle progression in A375.S2 cells after 24 h of treatment by PI staining followed with flow cytometry. The accumulation of G2/M population is considered as a biomarker for DNA damage and growth inhibition [21]. As presented in Figure 3, compound 3 induced a moderate accumulation of G2/M populations compared with the vehicle control, whereas no significant changes were seen in other compounds’ results. There are studies to show that genotoxic agents cause the G2/M-arrest and eventually cellular apoptosis. For example, costunolide induces G2/M-arrest and apoptosis in breast cancer cells, MDA-MB-231 [21]. However, no significant sub-G1 population was found in compound 1–4 treated cells, indicating that the dose at 100 μM of compounds inhibited the proliferation and cellular migration of A375.S2 cells effectively without inducing cell death.

To investigate whether treatments of L. tulipifera compounds modulate the migration of melanoma cells, the wound-healing assay was performed. Figure 4A showed that the migration of A375.S2 melanoma cancer cells was inhibited by compounds 1–4, as demonstrated by the denuded zones at 20-h post-treatment. Figure 4B demonstrated the quantitative analysis on the inhibition of migration ability by the four compound-treatments compared with the vehicle control. The calculated denuded zone (indicating the migratory ability of A375.S2 cells) for compounds were 100 ± 5%, 16.14 ± 6%, 79.8 ± 7%, 27.0 ± 6.6% and 45.5 ± 8%, respectively. Among these compounds, 2 and 3 exerted the most anti-migration effect.

We further investigated whether L. tulipifera compounds modulate the level of intracellular ROS. Flow cytometer-based DCFDA staining, which is oxidized by ROS to the highly fluorescent DCF was used to detect the intracellular ROS, H₂O₂. The calculated effective ROS scavenge ability of compounds 1–4 were 1.0 ± 0.4%, 59.2 ± 81.9%, 34.594 ± 18.0%, 623.2 ± 71.8% and 15.6 ± 7.6%, respectively (Figure 5). Among these four compounds, compound 3 induced a dramatically increased level of ROS in A375.S2, whereas no significant change was observed in the other L. tulipifera compounds-treated cells. According to our above results, compound 3 exerts multiple bio-effects on A375.S2, including anti-migration, induction of G2/M accumulation and up-regulation of intracellular ROS without inhibiting the proliferation of A375.S2 cells, suggesting the chemoprevention potential of compound 3.

Liriodenine derivatives have been reported to be cytotoxic to many cancer lines including breast cancer, lung cancer and hepatoma cells. Additionally, the anti-proliferative effects of liriodenine and its derivatives may mediate the apoptosis and ROS production within these cells [40,41,42,43]. Although L. tulipifera compounds 1, 3 and 4 exert inhibitory effects on cellular proliferations and migrations of A375.S2 cells, only treatment with compound 3 causes a high level of intracellular ROS, whereas no significant changes of ROS were observed in compound 1 and 4-treated cells. These observations indicate that the compound 1 and 4 inhibition of cellular proliferation and migration is ROS-independent. Previous studies indicated that the anti-cancer effects of liriodenine derivatives were achieved by inducing apoptosis [43]. In our study, at the dose of 100 μM, L. tulipifera compounds 1, 3 and 4 displayed anti-proliferation (Figure 3) and anti-migration (Figure 4) potential but did not induce apoptosis. Interestingly, only compound 3 induced a definitely high level of intracellular ROS, however, there was still no detectable apoptotic cell deaths in compound 3-treated cells (Figure 3). Accordingly, these above observations suggest that there might be novel anti-proliferation and anti-migration pathways triggered by L. tulipifera.

UV spectra were obtained on a Jasco UV-240 spectrophotometer in MeCN. IR spectra were measured on a Hitachi 260-30 spectrophotometer (Hitachi, Tokyo, Japan). ¹H-NMR (400/500 MHz) and ¹³C-NMR (100 MHz), HSQC, HMBC, COSY and NOESY spectra were obtained on a Varian (Unity Plus) NMR spectrometer (Varian, CA, USA). For each sample, 128 scans were recorded with the following settings: 0.187 Hz/point; spectra width, 14,400 Hz; pulse width, 4.0 μs; relaxation delay, 2 s. Low-resolution ESI-MS spectra were obtained on an API 3000 (Applied Biosystems, CA, USA) and high-resolution ESI-MS spectra on a Bruker Daltonics APEX II 30e spectrometer (Bruker, Bremen, Germany). Silica gel 60 (Merck, 70~230 mesh, 230~400 mesh) was used for column chromatography. Precoated silica gel plates (Merck, Kieselgel 60 F-254), 0.20 mm and 0.50 mm, were used for analytical TLC and preparative TLC, respectively, and visualized with 10% H₂SO₄.

The specimen of L. tulipifera was collected from Chiayi County, Taiwan in December, 2007. A voucher specimen was characterized by Dr. Jin-Cherng Huang of Department of Forest Product Compounds, Science and Furniture Engineering, National Chiayi University, Chiayi, Taiwan and deposited in the School of Medical and Health Sciences, Fooyin University, Kaohsiung County, Taiwan.

The air-dried stems of L. tulipifera (9.0 kg) were extracted with MeOH at room temperature and the MeOH extract (187.5 g) was obtained upon concentration under reduced pressure. The MeOH extract was chromatographed over silica gel using CH₂Cl₂/MeOH as eluent to produce seven fractions. Among these seven fractions, part of fraction 4 was subjected to Si gel chromatography eluting with n-hexane/acetone to obtain (-)-glaucine (3) (19 mg) and (-)-norglaucine (4) (24 mg). Part of fraction 5 (7.43 g) was subjected to Si gel chromatography eluting with n-hexane/acetone to obtain liriodenine (1) (36 mg) and (-)-anonaine (2) (14 mg). These compounds were characterized by the comparison of their physical and spectral data (UV, IR, NMR and MS) with values given in the literature.

Liriodenine (1). Yellow needles (CH₂Cl₂); UV λ_max: 224, 248, 265, 308 nm; IR ν_max: 1625, 1022, 933 cm⁻¹; ¹H NMR (500 MHz, CDCl₃): δ 6.38 (2H, s, -OCH₂O-), 7.21 (1H, s, H-3), 7.59 (1H, td, J = 8.0, 1.5 Hz, H-9), 7.77 (1H, td, J = 8.0, 1.5 Hz, H-10), 7.79 (1H, d, J = 5.0 Hz, H-4), 8.60 (1H, dd, J = 8.0, 1.5 Hz, H-8), 8.67 (1H, dd, J = 8.0, 1.5 Hz, H-11), 8.90 (1H, d, J = 5.0 Hz, H-5); ESI-MS m/z: 275 [M]⁺.

(-)-Anonaine (2). Yellow powder (MeOH); UV λ_max: 230, 272, 310 nm; IR ν_max: 1040, 950 cm⁻¹; ¹H NMR (500 MHz, CDCl₃): δ 2.66 (1H, d, J = 12.0 Hz, H-7α), 2.81 (1H, t, J = 14.0 Hz, H-7β), 2.93~3.06 (3H, m, H-4α, 4β, 5α), 3.40 (1H, m, H-5β), 3.98 (1H, dd, J = 14.0, 4.0 Hz, H-6α), 5.94 and 6.09 (each 1H, d, J = 1.0 Hz, -OCH₂O-), 6.57 (1H, s, H-3), 7.21~7.33 (3H, m, H-8~10), 8.08 (1H, d, J = 8.0 Hz, H-11); ESI-MS m/z: 265 [M]⁺.

(-)-Glaucine (3). Brown powder (MeOH); UV λ_max: 210, 282, 305 nm; IR ν_max: 2800, 1600, 1580, 1310, 1105, 950 cm⁻¹; ¹H NMR (500 MHz, CDCl₃): δ 2.63 (3H, s, N-CH₃), 3.66 (3H, s, C₁-OCH₃), 3.89 (3H, s, C₂-OCH₃), 3.90 (3H, s, C₉-OCH₃), 3.93 (3H, s, C₁₀-OCH₃), 6.60 (1H, s, H-3), 6.78 (1H, s, H-8), 8.09 (1H, s, H-11); ESI-MS m/z: 355 [M]⁺.

(-)-Norglaucine (4). Brown powder (MeOH); UV λ_max: 282, 303 nm; IR ν_max: 2800, 1600, 1585, 1320, 1105, 975 cm⁻¹; ¹H NMR (500 MHz, CDCl₃): δ 3.66 (3H, s, C₁-OCH₃), 3.90 (3H, s, C₂-OCH₃), 3.90 (3H, s, C₉-OCH₃), 3.93 (3H, s, C₁₀-OCH₃), 6.61 (1H, s, H-3), 6.78 (1H, s, H-8), 8.09 (1H, s, H-11); ESI-MS m/z: 341 [M]⁺.

DPPH· is a stable free radial with a violet color (absorbance at 517 nm) that changes its color to light yellow when the free radicals are scavenged [44]. Various concentrations of the four compounds were added to 0.1 mL of stable DPPH (60 μM) solution. When DPPH reacts with hydrogen-donating anti-oxidant, it is reduced, resulting in a decrease in absorbance at 517 nm. The analyzed time interval was 10 min per point, up to 30 min by using UV–vis spectrophotometer (BioTek Co.). Vitamin C was acted as a positive control. The DPPH· radical scavenging activity (%) was determined as:

The ferrous ion chelating potential of the four L. tulipifera compounds was investigated according to a previously described method [44]. Briefly, various test concentrations of samples dissolved in DMSO were added to a solution of 2 mM FeCl₂·4H₂O (0.01 mL). The reaction was initiated by the addition of 5 mM ferrozine (0.02 mL), and the mixture was vigorously shaken and left standing at room temperature for 10 min. The absorbance of the mixture was then read at 562 nm against a blank. EDTA was used as a positive control.

The reducing powers of our natural pure compounds were determined according to the method of [44]. Briefly, various concentrations of test samples were mixed with 67 mM phosphate buffer (pH 6.8, 0.085 mL) and 20% potassium ferricyanide [K₃Fe(CN)₆, 2.5 μL) The mixture was incubated at 50 °C for 20 min, and trichloroacetic acid (10%, 0.16 mL) was then added to the mixture that was then centrifuged for 10 min at 3,000 g. The upper layer of the solution (75 μL) was mixed with 2% FeCl₃ (25 μL), and the absorbance was measured with a 96-well plate spectrophotometer at 700 nm. BHA was used as a positive control. A higher absorbance demonstrates a higher reductive capability.

Tyrosinase inhibitory activity was determined spectrophotometrically according to the method described previously [45], with minor modifications. Assays were conducted in a 96-well microplate, an ELSA plate reader (Molecular Devices) being used to determine the absorbance at 490 nm. Kojic acid was used as a positive control. The test substance was dissolved in aqueous DMSO, and incubated with L-tyrosine (2.5 mg/mL) in 50 mM phosphate buffer (pH 6.8). Then, 25 U/mL of mushroom tyrosinase in the same buffer was added, and the mixture was incubated at 37 °C for 30 min. Tyrosinase inhibitory activity was determined at 490 nm by the following equation:

where A is the optical density (OD₄₉₀) without test substance; B is the OD₄₉₀ without test substance, but with tyrosinase; C is the OD₄₉₀ with test substance; and D is the OD₄₉₀ with test substance, but without tyrosinase. The results are listed in Table 1.

Human melanoma cell lines A375.S2 were obtained from the American Type Cell Culture Collection (ATCC, Manassas, VA, USA). It was maintained in monolayer culture at 37 °C and 5% CO₂ in DMEM supplemented with 10% FBS, 10 μg/mL of penicillin, 10 μg/mL of streptomycin, and 0.25 μg/mL of amphotericin B.

The MTT assay was used to determine cell viability and proliferation. The cell lines were seeded in 96-well culture plates (1 × 10⁴ cells/well). After seeding cells for 24 h, various compounds with concentration 100 μM were added. Within 24 h of compound treatments, images of human melanoma A375.S2 cells were taken at suitable time intervals. MTT solution (5 mg/mL and dissolved in phosphate buffered saline; PBS) was diluted 1:10 in culture medium and added to a culture dish followed by an incubation at 37 °C. After 2 h of MTT treatment, the media was removed and each precipitate in a specific dish was dissolved in 100 μL of DMSO to dissolve the purple formazan crystals. After the dishes were gently shaken for 20 min in the dark to ensure maximal dissolution of formazan crystals, the optical density (OD) values of the supernatant were measured at 595 nm. All experiments were repeated at least three times. In consideration of the possible anti-proliferative effects of DMSO, a maximal amount (0.5%) of DMSO was added to culture and used as positive controls. DMSO at this amount was found not to affect the growth of the human melanoma A375.S2 cells.

The cell cycle distribution was determined by propidium iodide (PI) staining as described previously [46]. Briefly, 1 × 10⁶ cells were treated with DMSO as vehicle control or 100 μM of L. tulipifera compounds for 24 h. After treatments, cells were harvested, washed twice with PBS and fixed in 70% ethanol. After centrifugation at 3000 rpm for 5 min at 4 °C, the cell pellets were stained with 10 μg/mL PI (Sigma, St. Louis, MO, USA) and 10 μg/mL RNase A in PBS for 30 min at 37 °C in the dark. The cells were analyzed using a FACScan flow cytometer (Becton-Dickinson, Mansfield, MA, USA) and the results were analyzed using the Cell-Quest software (Becton-Dickinson).

The wound healing assay was described previously [46]. In brief, a total of 3 × 10⁵ A375.S2 cells were seeded onto 12-well plates, treated with PBS (as vehicle control) or indicated concentrations of L. tulipifera compounds, and then grown to complete confluence. A 200-μL plastic pipette tip was used to scratch the culture monolayer and create a clean 1-mm-wide wound area in the A375.S2 confluent culture. After further incubation at 37 °C for 16 h, the wound gaps were photographed. The wound areas were then analyzed and calculated using the software “TScratch” (http://www.cse-lab.ethz.ch). The migration motility of cells was determined as % of vehicle control cells.

It has been reported that the intracellular level of ROS is associated with the inhibition of proliferation and cellular migration by a variety of stresses, including anti-cancer agents. Therefore, to determine whether the anti-proliferative effect of L. tulipifera compounds involves the production of oxidative stress, the changes in endogenous ROS levels were detected using a fluorescent dye DCFDA (Sigma-Aldrich). A total of 1 × 10⁵ A375.S2 cells were seeded onto a 6-well culture plate, treated with or without L. tulipifera compounds for 12 h, respectively. Afterwards, cells were harvested and stained with 100 nM DCFDA for 30 min at 37 °C in PBS, then washed twice with PBS. The fluorescence of DCFDA was measured by a flow cytometer. The excitation wavelength of DCFDA is 485 nm, and the emission wavelength is 530 nm.

All data are the means ± SD from at least triplicate experiments. The significance of the differences was analyzed by a one-way analysis of variance (ANOVA), with p < 0.05 or 0.01 as considered significant.