Advances in chemical and biological characterization of triterpenoid saponins from Anagallis arvensis L. using UHPLC-MS/MS and cell-based assays

Abstract A fraction enriched in triterpenoid saponins (F4) from Anagallis arvensis L. was chemically characterized by UHPLC-ESI-QTOF-MS/MS and NMR analyses. The results proposed the presence of nine monodesmosidic saponins derived from oleanolic acid, including two reported for the first time for this species, 3: 3-O-{β-D-glucopyranosyl (1→4)-[β-D-xylopyranosyl (1→2)]-β-D-glucopyranosyl (1→4)-[β-D-glucopyranosyl (1→2)]-α-L-arabinopyranosyl}-13β, 28-epoxy- 22 acetyl- 28 methoxy-16α, 24-oleananediol and 4: 3-O-{β-D-glucopyranosyl (1→4)-β-D-glucopyranosyl (1→4)-[β-D-xylopyranosyl (1→2)]-β-D-glucopyranosyl (1→4)-[β-D-glucopyranosyl (1→2)]-α-L-arabinopyranosyl}-13β, 28-epoxy- 16α, 30-oleananediol. Furthermore, haemolytic activity was determined by dot-blot autography, and cytotoxicity on human lymphocyte cultures was analysed according to metabolic activity (MTT assay) and membrane integrity (Trypan blue exclusion test). F4 showed mild cytotoxicity (%V > 50% at 100 μg/mL) on human lymphocytes under conditions of activation or not by LPS; moreover, showed haemolytic activity between 50 and 500 μg. This work contributed to phytochemical knowledge of the triterpenoid saponins from A. arvensis, and its cytotoxic effects on normal human cells. Graphical Abstract


Introduction
Anagallis arvensis L. (Primulaceae), popularly known as 'scarlet pimpernel', is a plant widely distributed throughout the world (Al-Snafi 2015), with great current interest due to its medicinal properties such as wound healing, expectorant, antidiuretic, antimicrobial and anticancer, among others (Saleem et al. 2020(Saleem et al. , 2021Yasmeen et al. 2020;Saqib and Janbaz 2021). Several studies have shown that triterpenoid saponins are the major phytochemical constituents in A. arvensis (Mahato et al. 1991;De Napoli et al. 1992;Shoji et al. 1994aShoji et al. , 1994b, and showed interesting biological activities related to its medicinal properties (Sparg et al. 2004;Sober on et al. 2017). However, due to the diversity of phytochemical research approach carried out, the full chemical identification of the secondary metabolites of this species, wich have phytochemical significance, has not been achieved yet.
Considering the high polarity and structural similarity of these compounds, it is usually difficult to isolate single saponin, and its identification by NMR is also difficult, especially when the sugar chain contains more than three residues. The high resolving power achieved in liquid chromatography in tandem with mass spectrometry employing quadrupole time-of-flight, has been demonstrated as a valuable tool for the on-line analysis of known and unknown constituents from plant phytocomplexes; at the same time, UHPLC-ESI-QTOF-MS/MS analyses of triterpenoid saponins of A. arvensis have not been reported.
In this context, we applied a less selective initial extraction (ethanolic extract) compared to previous works, and an optimized partial purification step guided by the differential solubility of this type of compounds, to characterize the fraction obtained by phytochemical conventional methods: the Lieberman Bouchard test, the aphrogen power, the haemolytic activity and the TLC profile revealed with p-AS, and ultra-high performance instrumental methods such as UHPLC-ESI-QTOF-MS/MS, complementing with NMR analysis as much was possible. Furthermore, cytotoxicity on human lymphocytes cultures was determined by MTT assay and Trypan blue exclusion test under two different conditions.

Results and discussion
In the aphrogen power test, the foam produced (8.5 ± 0.5 mm height) persisted for at least 30 min, which indicated the presence of saponins in purified saponins fraction (PS) (Das et al. 2012). The formation of a red ring indicated the presence of type-triterpenoid saponins, when PS reacted with the Liebermann-Burchard reagent.
The chemical structure of compound 3 was deduced by comparison of the fragmentation patterns of known saponins reported previously for this plant [Anagallosaponin IX (2); and Methylanagallosaponin I (8)]. The positive ESI-MS showed ion peak at m/z 1313.6307 [M þ H] þ , which was 30 Da greater than compound 2, suggesting the addition of a C and an O to the aglycone (Supplementary material, Figure  S2). The MS/MS spectrum for compound 3 (Supplementary material, Figure S4A), gave the same five characteristic fragment ions that compounds 2 and 8 at m/z 295 suggesting that they had the same saccharide moieties. 13 C-NMR analysis of the F4 aglycones allowed to observe signals at d 38. 39-39.11 corresponding to C-19 that would confirm that the aglycones are of the oleanane type; a signal was also observed at d 77.24 that would correspond to the epoxy group between positions C-13 and C-28 (Supplementary material, Figure S3); while signals that would indicate that the aglycones have an olean-12-ene skeleton (d 120.8, 144.0 and 210.4) were not observed (Arslan 2014 Figure S6B), and these were consistent with the presence of an acetyloxy group attached to C-22, and a methoxy group attached to C-28 (Supplementary material, Figure S7). Furthermore, at 13 C-NMR spectrum (Supplementary material, Figure S5), a signal at d 55.34 (s) was observed, that could correspond to C-17, different to assigned signals to C-17 for compounds 2 and 8 [d 55.78 (s) and d 56.04 (s) respectively]. The lower shift observed for the C-17 of proposed compound (3) could be due to a slightly greater shielding provided by the methoxy group at C-28 with respect to compound 2, in which in such position the hydroxyl group provided a greater un-shielding by inductive effect; moreover, when comparing with compound 8, an acetyloxy group attached to C-22 instead of the hydroxyl group in such position (at b position) provided a greater unshielding. These data further support the proposed structure for compound 3, that differs in the aglycone respect to compound 2 at C-24 (attached to an OH group) and C-28 (attached to a methoxy group).
Positive ESI-MS of compound 4 showed ion peak at m/z 1387.5960 [M þ H] þ , corresponding to a compound not reported for A. arvensis. The chemical structure of this compound was tentatively deduced by introducing the molecular formula from MS analysis, in PubChem (2004PubChem ( -2022, which resulted in five probable saponins. One of these showed structural coincidences with triterpenoid saponins previously reported for A. arvensis (monodesmosidic, derived from oleanolic acid), and this was supported by experimental values of molecular ion and fragment ions for corresponding peak (Supplementary material, Figure S4B). On the other hand, it was not possible to confirm compound 4 by NMR analysis, due to the superposition of signals generated by the presence of functional groups (-OH) attached to similar positions in aglycones of other saponins in F4. Characteristics interactions for aglycone of the compound 4 are shown in Figure S8 (Supplementary material). The present analysis would provide relevant information for the structural elucidation when the isolation of the proposed compounds can be achieved.
Haemolytic activity of F4 was observed between 50-500 lg tested, as clear halos around the previously loaded points with F4 or positive control (ursolic), contrasting to the red background, and compared to solvent control (see Supplementary material, Table S2 and Figure S9).
Cytotoxic activity on lymphocytes cultures measured by MTT assay (metabolic activity) showed cell viability percentage (%V) greater than 50% up to 100 mg/mL, where cells not activated by LPS were slightly more sensitive at F4 concentrations tested (Supplementary material, Table S2 and Figure S10A). Cytotoxicity assessed by Trypan blue exclusion test (membrane integrity) was more pronounced on lymphocytes not activated by LPS, without reaching viability percent less than 50% up to 100 mg/mL (Supplementary material, Table S2); however, when N cells/mL was comparatively plotted (Supplementary material, Figure S10B), a lower total N cells/mL were detected in tests on LPS activated lymphocytes respect to those not treated with LPS, although the cytotoxic effects were not significantly different (a ¼ 5%). This could mean that F4 presented mild cytotoxic effects on human lymphocytes, regardless of the immunological conditions that we tried to simulate in our in vitro assays.
F4 showed cytotoxic mild effects on human lymphocytes cultures, by affecting %V according to membrane integrity (by staining with Trypan blue) below 50% at 100 mg/ mL, and evidencing also inhibition of metabolic activity (by conversion of MTT to formazan) lower than 50% under similar conditions. Haemolytic activity was demonstrated from 50 to 500 mg F4 on human blood cells.
This work contributed to complete the chemical characterization of the triterpenoid saponins of A. arvensis, accelerating the discovery of new drugs from triterpene saponins, and providing evidence about cytotoxic effects of this type of saponins on normal human cells.