Rp 18

Sl-D extract (4.3 g) was adsorbed into 5 g of silica gel (silica gel 60, Merck, Darmstadt, Germany) and added to a glass column packed with silica gel (30 g, 70–230 mesh, Merck, Darmstadt, Germany). The circumvention system consisted of a gradient of hexane-ethyl acetate with polarity increments of 10%, collecting 36 fractions of 100 mL each. The fractions were concentrated at reduced pressure and monitored with normal phase chromatography and revealed with cerium sulfate. Fractions showing similarity in their chemical characteristics were grouped into 4 fractions (SlD-1 to SlD-4).
The SlD-3 fraction (584 mg) was then adsorbed in silica gel (RP-18, Merck, Darmstadt, Germany) for separation using a pre-packed reverse phase column (10 g, RP-18, Merck, Darmstadt, Germany) as a stationary phase. The elution system consisted of 100% acetonitrile (HPLC grade, Merck, Darmstadt, Germany) and polarity increments of 10% water, collecting volumes of 7 mL. Separation was monitored with reverse phase thin layer chromatography (RP-18, Merck, Darmstadt, Germany) and revealed with cerium sulfate. In fraction 21–22, the diterpene 15,16-epoxy-10-β-hydroxy-neo-cleroda-3,7,13(16), 14-tetraene-17,12R:18,19-diolide (1, 10 mg) was isolated. Fractions 26–28 had salviandulin A (2, 25 mg), and in fractions 37 and 38, a flavone called eupatorin (3, 85.2 mg) was identified (Figure 2).

The organic fraction (9.5 g) was adsorbed in silica gel (10 g normal phase), applied to a glass column with silica gel (5.0 × 40 cm, 30 g, 0.04–0.06-mm mesh, Merck, Darmstadt Germany), and eluded with hexane/ethyl acetate with 5% ascendant polarity, collecting 71 subfractions of 50 mL each. Subsequently, the subfractions were concentrated in a rotary evaporator under reduced pressure and grouped according to their similarity by TLC into 17 subfractions (C1F1-C1F17). The C1F9 fraction (834 g) was absorbed with silica gel (RP-18, Merck, Darmstadt, Germany) and fractionated on a glass column (350 × 20 mm) packed with silica gel (10 g, RP-18, Merck, Darmstadt, Germany). The elution system was water; acetonitrile, using four mixtures: 70:30; 65:35, 40:60; and methanol (100%), and obtained 16 sub-fractions of volumes of 10 mL. In fraction 3 and 4, compound (1) was isolated and identified by NMR and MS (see supplementary data Figures S1–S4). Compounds (2–4) and (6) were identified by comparison of HPLC data with standards and luteolin 7-O-rhamnoside (5) was identified by mass spectrometry (supplementary data, Figure S5). The HA extract, fractions and subfractions were analysed by TLC on silica gel 60 under light at 254 and 360 nm (supplementary data, Figures S11–S17). All structures identified in the bioactive sub-fractions are shown in Figure 3.

A MX-500 Bruker spectrometer was used to measure one-dimensional (1D) and two-dimensional (2D) nuclear magnetic resonance (NMR) spectra. The chemical shifts (δ) were calculated (ppm) relative to TMS and J scalar coupling constants reported in Hz. MS analyses were carried out on an Agilent triple quadrupole 6410 QQQ LC/MS mass spectrometer with an ESI ion source (nebulizer gas pressure is 60 psi, gas temperature is 350 °C, and flow rate is 12 L/min), operating in the negative and positive scan modes of ionization through direct infusion method using methanol–water (4:6 v/v) at a flow rate of 0.5 mL/min. Separations and purifications of secondary metabolites were carried out by using column chromatography either on silica gel 70–230 mesh or RP-18 (E. Merck, Darmstadt, Germany). RP-18 (Merck) and pre-coated silica gel 60 F254 TLC plates were used to check the fractions, and the spots were detected by UV light and by spraying with ceric sulphate and sulfuric acid reagent followed by heating on a hot plate (TLC plate heater III CAMAG, Muttenz, Switzerland). Analytical and reagent grade solvents were obtained from Sigma-Aldrich (St. Louis, MO, USA). NMR deuterated methanol (CD₃OD) and dimethylsulfoxide (DMSO-d₆) were purchased from Cambridge Isotope Laboratories (Tewksbury, MA, USA).