Synthesis of ring-A serjanic acid derivatives and their cytotoxic evaluation through the brine shrimp lethality assay (BSLA)

2.1 General procedures

Reagents were obtained from Sigma-Aldrich (Milwaukee, Wisconsin, USA) and were used without any purification. Analytical and preparative thin layer chromatography (TLC) was performed on silica gel (15–40 μm PF₂₅₄) 0.25 and 0.5 mm thick plates respectively (supplied by Merck), and the spots were visualized by UV-B light or by spraying with AcOH–H₂O–H₂SO₄ (74:16:10) mixture, followed by heating. Column chromatography was performed using silica gel 230–400 mesh. All compounds were fully characterized by IR, HR-EI-MS, ¹H NMR and ¹³C NMR spectroscopy. Infrared spectra were recorded on a Thermo Scientific Serie Nicolet iS10 FTIR spectrophotometer, preparing the samples as potassium bromide disks (1% w/w). EI-HR-MS were recorded on a JEOL JMS-AX505WA double focus mass spectrometer, using EI mode (70 eV) and PFTBA as internal reference, ESI-HR-MS were acquired on a ThermoScientific LTQ-Orbitrap-LC-MS spectrometer. ¹H and ¹³C NMR spectra were recorded on a Bruker Avance 300* spectrometer at 300 and 75 MHz respectively, using CDCl₃ as solvent and TMS as internal reference, unless otherwise stated. Melting points were determined using a Krüss-Optronic apparatus and are not corrected.

2.2 Isolation of serjanic acid (1)

Dried and ground fruits of Phytolacca icosandra L., (2.5 kg) were extracted in a Soxhlet apparatus using MeOH as solvent. The methanol extract was concentrated under reduced pressure affording 300 g of a crude brown residue. The residue was submitted to column chromatography on silica gel using solvent mixtures of increasing polarity (CHCl₃, EtOAc, MeOH) to afford 15 fractions (A–O). Fraction “K” (34.2 g), was dissolved in MeOH–H₂O (7:3 v/v) mixture and concentrated H₂SO₄ was added until the solution reached pH = 2, then the solution was refluxed for 6 h. The precipitate formed after hydrolysis was filtered off and dried to afford a solid residue (23.5 g) which was submitted to column chromatography on silica gel using CHCl₃–(CH₃)₃CO (9:1 v/v) as eluent. This procedure led to four sub-fractions (A–D). Sub-fractions “B” and “C” of the chromatography process were combined on the basis of their chromatographic profiles, and the combined sub-fractions afforded 4.7 g of serjanic acid (1) (0.2% of dry mass) as a white amorphous solid, m. p. 288–290 °C (lit.: 288–291 °C [24]). All spectroscopic data were in agreement with the literature values [25].

2.3 Synthesis of 3-oxo-serjanic acid (2)

To a solution of compound 1 (2.5 g, 4.99 mmol) in acetone (250 mL) at 0 °C was added Jones reagent (K₂Cr₂O₇–H₂SO₄) drop wise until yellow-orange colour persisted. The reaction was monitored by TLC till its completion in 2 h. The excess of Jones reagent was eliminated by adding MeOH to the solution and the solution was filtered to remove insoluble residue. The filtrate was concentrated under vacuum to afford a white amorphous solid. The solid was recrystallized from ethanol to afford compound 2 as colourless needles (2.32 g, 93%), m. p. 276–278 °C. All spectroscopic data was in agreement with the literature values [25].

2.4 General procedure for the synthesis of indolo-triterpenes 8–13

To a portion of compound 2 (0.12 mmol) dissolved in glacial AcOH (2 mL) and kept under agitation, a solution of phenyl hydrazines 3–7 (0.12 mmol) dissolved in glacial AcOH (0.5 mL) was added dropwise. The solution was refluxed for 4 h until the end of reaction by TLC monitoring. After cooling, 40 mL of distilled water was added and the resulting solution was extracted with CH₂Cl₂ (3 × 10 mL). The organic phase was washed with 5% NaOH (2 × 20 mL) and brine (2 × 20 mL), dried with MgSO₄ and concentrated under vacuum to obtain the crude reaction product. It was subjected to preparative TLC plates on silica gel (15–40 mm PF₂₅₄), using hexane-ethyl acetate (2:3 or 1:1 v/v) as solvent mixture to afford the purified product as a solid.

2.5 Synthesis of 3-hydroxyimino-olean-12-en-28,30-dioic acid-30-methyl ester (14)

To a solution of 65 mg (0.13 mmol) of ketone 2 dissolved in pyridine (2 mL), hydroxylamine hydrochloride (11.3 mg, 0.16 mmol) was added and the obtained mixture was refluxed for 3 h, cooled and poured into cold water (50 mL) slightly acidified with HCl. The resulted precipitate was filtered, washed with water, dried and recrystallized from absolute ethanol. Colourless needles; yield: 92%; m. p. 262–266 °C. – IR (KBr): ν = 3431, 3055, 2945, 1732, 1709, 1646, 1463, 1211, 769 cm^–1. – ¹H NMR (300 MHz, CDCl₃, 25 °C, TMS): δ = 5.36 (m, 1H, H-12), 3.69 (s, 3H, OMe), 3.10 (m, 1H, H-2a), 2.70 (m, 1H, H-18), 2.22 (m, 1H, H-2b), 1.14, 1.12, 1.11, 1.04, 1.01, 0.73 (6 × s, 18H, CMe). – ¹³C NMR (75 MHz, CDCl₃): δ = 182.8 (C-28), 177.0 (C-30), 167.4 (C-3), 143.1 (C-13), 123.1 (C-12), 55.8 (C-5), 51.9 (OMe), 47.2 (C-9), 45.8 (C-17), 43.7 (C-20), 42.4 (C-18), 42.1 (C-19), 41.6 (C-14), 40.3 (C-8), 39.3 (C-4), 38.3 (C-1), 37.0 (C-10), 33.5 (C-22), 32.4 (C-7), 30.4 (C-21), 28.4 (C-29), 27.7 (C-15), 27.3 (C-27), 25.8 (C-23), 23.5 (C-11), 23.4 (C-24), 23.2 (C-16), 19.1 (C-2), 17.6 (C-6), 16.9 (C-26), 14.9 (C-25). – HRMS ((+)-ESI): m/z = 514.3466 (calcd. 514.3454 for C₃₁H₄₈NO₅, [M+H]⁺).

2.6 Synthesis of pyrazine[2,3-b]-olean-12-en-28,30-dioic acid-30-methyl ester (15)

To a solution of 2 (100 mg, 0.2 mmol) in morpholine (6 mL), was added sulphur (72.4 mg, 2.2 mmol) and ethylenediamine (68 μL, 1.01 mmol) at room temperature. The reaction mixture was heated under reflux for 6 h. After cooling, the mixture was poured into water (50 mL) and extracted with ethyl acetate (2 × 15 mL). The organic layer was washed with water and brine, dried over anhydrous MgSO₄ and concentrated in vacuo. The residue was purified over silica gel preparative TLC plates (hexane-ethyl acetate 7:3 v/v) to afford a white solid. Yield: 64%; m.p: 233–235 °C. – IR (KBr): ν = 3418, 2953, 1731, 1693, 1511, 1460, 1208, 755 cm^–1. – ¹H NMR (300 MHz, CDCl₃, 25 °C, TMS): δ = 8.41 (d, J = 2.4 Hz, 1H, H-2′), 8.26 (d, J = 2.4 Hz, 1H, H-3′), 5.43 (t, J = 3.1 Hz, 1H, H-12), 3.69 (s, 3H, OMe), 2.98 (d, J = 16.3 Hz, 1H, H-1a), 2.72 (dd, J = 3.5, 13.3 Hz, 1H, H-18), 2.51 (d, J = 16.3 Hz, 1H, H-1b), 1.91 (m, 1H, H-9), 1.29, 1.26, 1.18, 1.15, 0.88, 0.83 (6 × s, 18H, CMe). – ¹³C NMR (75 MHz, CDCl₃): δ = 182.7 (C-28), 176.9 (C-30), 159.7 (C-3), 150.5 (C-2), 142.9 (C-13), 142.4 (C-3′), 141.3 (C-2′), 123.2 (C-12), 53.1 (C-5), 51.8 (OMe), 48.1 (C-1), 45.9 (C-17), 45.8 (C-9), 43.7 (C-20), 42.4 (C-18), 42.1 (C-4), 41.7 (C-14), 39.4 (C-19), 39.2 (C-8), 36.6 (C-10), 33.5 (C-22), 32.0 (C-7), 31.6 (C-29), 30.4 (C-21), 28.4 (C-27), 27.7 (C-15), 25.7 (C-23), 24.2 (C-24), 23.3 (C-11), 23.2 (C-16), 20.1 (C-6), 16.7 (C-26), 15.5 (C-25). – HRMS ((+)-ESI): m/z = 535.3439 (calcd. 535.3458 for C₃₃H₄₇N₂O₄, [M+H]⁺).

2.7 General procedure for the synthesis of the benzylidine derivatives 24–31

To a solution of 2 (100 mg, 0.2 mmol) in ethanol (3 mL), was added KOH (0.6 mmol). The mixture was cooled to 0 °C trough ice bath and added drop wise 0.2 mmol of the corresponding benzaldehyde 16–23 dissolved in 1 mL EtOH. The reaction was carried out at 0 °C for the first 5 min and then at room temperature for 24 h. The mixture was poured into water (10 mL) acidified with HCl (three drops) and extracted with ethyl acetate (2 × 15 mL). The organic layer was washed with water and brine, dried over anhydrous MgSO₄ and concentrated in vacuo. The residue was purified over silica gel preparative TLC plates (hexane-chloroform 1:9 v/v) to afford benzylidines 24–31.

2.8 Brine shrimp lethality assay (BSLA)

The assay was performed as described previously by Meyer [23] with some minor modifications. Brine shrimp eggs (Gulf Breeze^®) were hatched in artificial sea water prepared with commercial salt mixture (Instant Ocean^®), illuminated and oxygenated with an aquarium pump. After 48 h incubation at 26 °C, 10 shrimps were transferred with a Pasteur pipette to three sample vials for each of three doses (100, 25, 5 μg mL⁻¹), for a total of nine vials per sample. The compound samples (10 mg) were dissolved in CHCl₃ (5 mL). Aliquots of testing solutions 250, 62.5 or 12.5 μL (for 100, 25 and 5 ppm doses) were placed on vials of 5 mL and the solvent was evaporated. The residue was redissolved in 5 μL of Tween 80^® and 5 mL of artificial sea water was added. Survivors were counted and the percent deaths at each dose were determined. Control samples were included and assayed simultaneously. IC₅₀ values were calculated from 24 h counts using the probit analysis [26].

3.1 Chemistry

Firstly, SA was isolated from the dried fruits of P. icosandra L. using a procedure detailed in the Experimental Section. The isolated compound was fully characterized by comparison of its spectroscopic data with those previously reported [24, 25] along with its melting point. Subsequently, SA was submitted to a Jone’s oxidation to afford ketone 2 [25]. Compound 2 was then converted into several ring-A indolo-derivatives by its reaction with aromatic hydrazines 3–7 in glacial acetic acid under reflux [27]. Aqueous work-up using NaOH and brine solution followed by simple preparative TLC on silica gel, afforded the purified indolo-triterpenes 8–13 in 58–74% yield. The reaction starting with 3-bromophenylhydrazine yielded a mixture of indoles 9 and 11 in a 2:1 ratio respectively, which could not be separated by sequential preparative silica gel chromatography.

Reaction of ketone 2 with hydroxylamine hydrochloride in pyridine provided the desired C-3 oxime 14 in 92% yield; the compound was precipitated from the solution by adding cold water and filtration in vacuum. Pyrazine 15 was obtained by condensation of ketone 2 in the presence of ethylenediamine and sulphur in morpholine [28]. Aqueous work-up using brine and purification by preparative TLC plates on silica gel afforded 15 in 64% yield.

The synthetic protocol for serjanic acid benzylidine derivatives 24–31 from the intermediate 2 involved a Claisen-Schmidt condensation with the corresponding benzaldehyde 16–23 in the presence of ethanolic KOH at room temperature for 24 h. The work-up was carried out by adding acidified water to precipitate the compounds, extraction with EtOAc and purification by preparative TLC on silica gel. Yields for these compounds were between 40 and 81% (Scheme 1).

Scheme 1:

Reagents and conditions: (a) K₂Cr₂O₇–H₂SO₄, acetone (0 °C); (b) aromatic hydrazine, AcOH, reflux (4 h); (c) NH₂OH·HCl, pyridine, reflux (3 h); (d) sulphur, ethylenediamine, morpholine, reflux (6 h); (e) substituted benzaldehyde, ethanolic KOH, r.t. (24 h).

Extensive ¹H, ¹³C NMR, IR and HR-MS spectroscopic analysis confirmed the structures of all serjanic acid derivatives. All compounds exhibited in their IR spectra typical signals expected for the functional groups present in the molecules. For instance, all derivatives showed strong absorption bands ν = 3340–3500 cm⁻¹ (O–H stretching), 3050–3060 cm⁻¹ (=C–H stretching), 1693–1732 cm⁻¹ (C=O stretching) and 1208–1221 cm⁻¹ (C–CO–O stretching). Indolo-triterpene, oxime and pyrazine derivatives presented C=N stretching bands between ν = 1457–1463 cm⁻¹, while aromatic C=C stretching bands between 1550 and 1606 cm⁻¹ were observed for indolo-triterpenes, pyrazine and benzylidines.

The ¹H NMR spectra of all compounds revealed signals corresponding to vinylic, methoxy- and methyl hydrogens appearing between δ = 5.36–5.49, 3.69–3.75 and 0.73–1.34 ppm respectively. All compounds except for 14 exhibited two allylic protons as doublets (J ≈ 15 Hz) around δ = 2.17–3.43 ppm, and were assigned to H-1_a and H-1_b of the triterpenoid. Signals of aromatic hydrogens were detected between δ = 6.39–8.64 ppm and the –NH proton in compounds 8–13 was detected between 7.71 and 8.01 ppm. A considerable shift to lower field was observed for methyl groups H-23, H-24 and H-25 of indolo-triterpene and pyrazine derivatives, due to the proximity to the heterocyclic ring. Compound 14 displayed two allylic signals at δ = 3.10 and 2.22 ppm, corresponding to the H-2_a and H-2_b vicinal to the oxime group at C-3 position. Also, a characteristic vinylic proton (H-1″) between δ = 7.23–7.52 ppm was present in derivatives 24–31.

¹³C NMR signals for all triterpenoidal structures were assigned, and the most relevant signals were: carbonylic carbon atoms between δ = 176.3–183.9 ppm of the acid and ester functionalities, vinylic carbons of the C-12/C-13 double bond between 121.8 and 143.5 ppm, methoxy groups between 51.5 and 52.0 ppm and six methyl carbons around 15.0–31.6 ppm. In derivatives 8–13, carbon atoms C-2 and C-3 suffered a downfield chemical shift due to the aromatic ring formation. Signals for C-2 were located around δ = 106.8–108.8 ppm and C-3 signals around 140.9–142.8 ppm, due the proximity with the nitrogen atom. The rest of the carbon atoms belonging to the indolic nucleus resonates between δ = 108.3–137.3 ppm. Elucidation of the oxime 14 was corroborated by detecting the characteristic hydroxyimine carbon at δ = 167.4 ppm and derivative 15 evidenced a downfield chemical displacement to 48.1 ppm of its C-1 position and exhibited the characteristic pyrazine-ring carbons at δ = 159.7 (C-3), 150.5 (C-2), 142.4 (C-3′) and 141.3 (C-2′) ppm. ¹³C NMR signals at δ ≈ 207, 142 and 135 ppm for α,β-unsaturated ketones along with aromatic carbons between 114.0 and 164.3 ppm confirmed the product formation of benzylidines 24–31. Some minor signals, product of the “E/Z” isomerization, where found in the NMR spectra of all benzylidines. Formation of the “E” isomer was assumed, due to kinetic and thermodynamic considerations.

A detailed assignment of the ¹H and ¹³C NMR spectral data for all new compounds is given in the Experimental Section.

3.2 Brine shrimp lethality assay (BSLA)

The toxicity of SA and all derivatives was tested against A. salina. The dose-dependent assays were carried out between 10 and 100 ppm using a negative and positive control; Tween 80^® at this concentration did not affect this bioassay. IC₅₀ values of the compounds are presented in Table 1.

The results showed that almost all derivatives presented lower IC₅₀ values with respect to SA, which can be translate into an increase of the cytotoxic potential of these molecules; in contrast, compound 30 showed an increase of almost 50% in the IC₅₀. Concerning indolo-triterpenes, we observe that all derivatives with the exception of 12 presented higher toxicity than 8. The increase in the toxicity with respect to SA was also observed for the majority of benzilydine derivatives, and it is noteworthy to mention that the presence of an electron-withdrawing group in para-position of the phenyl ring resulted in greater bioactivity.

Derivatives 15, 26 and 31 exhibited the lowest IC₅₀ values. Moreover, the observed toxicity of 15 is slightly lower than that of Podophyllotoxin [21], a well-known bioactive natural compound, making 15 a suitable candidate for further biological and pharmacological studies.