Zanthoxylum schreberi

Plazas, Erika, Casoti, Rosana, Murillo, Monica Avila, Costa, Fernando Batista Da & Cuca, Luis Enrique, 2019, Metabolomic profiling of Zanthoxylum species: Identification of anticholinesterase alkaloids candidates, Phytochemistry 168, pp. 1-13 : 8-10

2.4. Isolation of biomarkers from bark of Zanthoxylum schreberi

According to our OPLS-DA results and comparing the exact masses and fragmentation patterns, we concluded that the variables 419POS and 232POS could correspond to putative berberine and chelerythrine. Since, these alkaloids were previously reported as cholinesterase inhibitors ( Brunhofer et al., 2012; Ahmed et al., 2015), we selected an extract with high occurrence of these compounds for isolation and further evaluation of their activity against both enzymes, which would allow us to validate our statistical model. Thus, from the bark of Z. schreberi by bio-guided three compounds were isolated, which presented positive reaction with Dragendorff and showed AChE inhibition in TLC bioautography. The isolated alkaloids were analyzed by LC-MS in the same conditions of the metabolomic profiling and as well as by 1 H-NMR. The alkaloid A2 (Rt: 18.42 min), with exact mass 336.1229 and molecular formula C 20 H 18 NO 4 + showed the same ions in the MS 2 spectrum as the variable 419POS (m/z 321.0940, 320.0920, 306.0724, 292.0968 and 278.078). In the 1 H-NMR, A2 showed characteristic signals of the protoberberine scaffold and by comparison with data from previous reports this compound was fully identified as berberine ( Jeon et al., 2002). The alkaloid A3 (Rt: 17.89 min), presented a 1 H-NMR profile similar to berberine, but without the signal in ?? 6.1 ppm (s, 2H) that corresponds to methylenedioxy moiety and three signals in ?? 4,20, 4,10 and 4,02 ppm (s, 3H, each one), which indicate the presence of three methoxy substituents. In the HRMS analysis, A3 showed an exact mass of 338.1387 with molecular formula C 20 H 20 NO 4 + and the same fragmentation ions in MS 2 as the variable 123POS ( Table 2 ). By comparing the spectral data, A3 was identified as columbamine. Likewise, the alkaloid A1 (Rt: 18.66 min), with exact mass 348.1299 and molecular formula C 21 H 18 NO 4 +, showed the same profile in the MS 2 spectrum as the variable 232POS, with ions m/z 321.0940, 320.0920, 306.0724, 292.0968 and 278.0780, and based on the 1 H-NMR results, this compound was successfully identified as chelerythrine.

Although berberine and chelerythrine have been previously reported as cholinesterase inhibitors, in the present work it was also determinate the dual anti-cholinesterase activity in order to support our statistical results. The inhibition curves were determined at specified substrate concentration in the presence of different concentrations of A1, A2 and A3 ( Fig. 7 ).

The IC 50 values were calculated after fitting the curves using nonlinear regression function. The values are shown as mean of three independent experiments each performed by triplicates ( Table 3 ). Our results are in accordance with those reported in literature for these compounds. Berberine is one of the isoquinoline alkaloids with the highest inhibition activity against AChE reporting IC 50 values between 0.01 and 2.00 μM. However, berberine has been shown to be less active against BChE, the IC 50 values reported for horse-serum BChE are highest than 3 μM ( Jung et al., 2009; Murray et al., 2013). In contrast, columbamine, another protoberberine alkaloid, is less potent than berberine against AChE with reported IC 50 value of 5 μM ( Tsai and Lee, 2010) which has not previous reports against BChE inhibition. With regards to our results, columbamine shows comparable activity against both enzymes Ee AChE and Eq BChE with IC 50 values of 3.752 ± 0.160 and 2.048 ± 0.088 respectively. These results are an important finding in the search to multi-target compounds considering the role played by BChE in AD physiopathology. Although, AChE is the main responsible for the hydrolysis of the acetylcholine neurotransmitter, recent evidence ( Lane et al., 2006), has shown that BChE is also involved in the cholinergic regulation. In fact, during the progress of AD the activity of AChE decreases, while that of BChE increases gradually to keep the catalytic pathway.

Furthermore, chelerythrine has been reported as less active than berberine against both enzymes with IC 50 values in the range of 1–10 μM, which is in concordance to our results ( Konrath et al., 2013). However, kinetic studies showed that chelerythrine presents a mixed inhibition mechanism against Ee AChE. Moreover, molecular docking models suggest that this alkaloid binds both the catalytic (CAS) and peripheral anionic sites (PAS) of the TcAChE ( Torpedo californica ) and this characteristic has been linked with the inhibition of Aβ fibrillogenesis induced by AChE ( Brunhofer et al., 2012). Therefore, if all the benzophenanthridine alkaloids present this characteristic binding mode, they would present some high multimodal potential as amyloid aggregation inhibitors.

Considering that chelerythrine (232POS), berberine (419POS) and columbamine (123POS) were recognized as potential cholinesterase inhibitors, in the S-plot, the isolation and identification of these alkaloids from the bark of Z. schreberi , as well as, their dual inhibitory activity allowed us to validate our statistical model. Thus, our metabolomic profiling allows a rapid correlation of the chemical composition and the anticholinesterase activity of extracts, therefore decreasing the time spent in the isolation procedure, including the isolation of inactive compounds. Moreover, predictive statistical models were built in order to uncover the anticholinesterase potential of non-tested extracts. Therefore, our results constitute an important contribution to the search of naturally occurring anticholinesterase alkaloids using a targeted screening strategy based on the chemical profile of crude extracts.

Figures

Fig. 7. Dose-response inhibition curves of the isolated alkaloids (A1, A2 and A3) against EeAChE and EqBChE. The IC50 values were calculated after fitting the curves using non-linear regression function of GraphPad Prism. The curves are shown s mean of three independent experiments performed in triplicates.