Potential antiplatelet aggregation metabolites from the discarded sorghum (Sorghum bicolor L.) root

Abstract Sorghum (Sorghum bicolor L.) is the fifth largest crop in the world and has potential health benefits, but vast quantities of sorghum roots are discarded after harvest. Based on the previous antiplatelet aggregation for this species, two new multi-substituted 3H-indole alkaloids sorghumine A (1) and sorghumine B (2), together with 14 known compounds (3–16), were found from the water extract of sorghum roots. Compounds 1–2 were identified by analyzing their spectroscopic data and physic and chemical properties, and the absolute configuration was further determined by electronic circular dichroism (ECD) analysis and calculations. 1–2, 4, 6–8 and 13–15 showed significant inhibition of platelet aggregation induced by adenosine diphosphate. 2–4, 6–9 and 11 showed significant inhibition of platelet aggregation induced by collagen. 4–6, 8, 10–11 and 16 showed significant inhibition on platelet aggregation induced by thrombin. Furthermore, molecular docking showed that active compounds can bind to P2Y12 and COX-1 receptors in platelet. Graphical Abstract


Introduction
Platelet aggregation is a pivotal factor leading to thrombosis and frequently results in phlebothrombosis, cardiopathy, stroke and a variety of other cardiovascular and cerebrovascular diseases (Minno et al. 2013). Antiplatelet aggregation is a standard approach for the treatment of thrombotic diseases (Qiao et al. 2014). Many antiplatelet aggregation drugs have been used clinically such as aspirin, clopidogrel and ticagrelor, despite their side effects and limited impact (Slevin et al., 2012). Hence, discovering new antiplatelet aggregation drugs has always been a desirable scientific effort especially with more effective, safer and fewer side effects.

Results and discussion
In HMBC spectra ( Figure S1), the correlations from H-3 to C-2, C-4, C-5 and C-9 as well as those from H-8 to C-4, C-6 and C-7 suggested that 1 was 2,3,5,6,7-pentasubstituted 3H-indole core. The 1 H-1 H COSY correlations of H-19/H-23 and H-20/H-22, together with the HMBC correlations from H-19/H-23 to C-3, C-20 and C-21 confirm the presence of a para-substituted phenol unit and linked with the C-3 of 3H-indole core. The 1 HÀ 1 H COSY correlations of H-13/H-17 and H-14/H-16 as well as the HMBC correlations from H-13/H-17 to C-11, C-15, from H-14/H-16 to C-12 and C-15 and from H-10 to C-5, C-7 and C-11 suggested the presence of a 1-(4-hydroxyphenyl) ethan-1one moiety and linked with the C-6 of 3H-indole core. The HMBC correlations from -OCH3 to C-3 0 and the coupling constant (J ¼ 7.8 Hz) of H-1 0 (d H 4.59) suggested that the sugar is 3 0 -O-methyl-b-D-glucopyranose. The joint position of 3 0 -O-methyl-b-D-glucopyranose was further supported by the HMBC correlations from H-1 0 to C-5. In addition, the other -OCH 3 linked with the C-2 of 3H-indole core was supported by the HMBC correlations from -OCH 3 to C-2. In order to further verify the structure of sugar units, acid hydrolysis was subsequently performed and analysis of its HPLC retention times (Ito et al. 2004;Hu et al. 2015), ESI-MS data and rotation value. The results show that the acid hydrolysis product of the HPLC retention times (t R 7.20 min), ESI-MS m/z 191.18 [M-H]and rotation value (½a 20 D þ52.4) consistent with standards (3 0 -O-methylb-D-glucopyranose). Finally, the absolute configuration of 1 (2 R) was determined by ECD calculations with time-dependent density functional theory (TD-DFT) similar to the experimental data ( Figure S1). Sorghumine B (2) was obtained as a brown powder and has the same molecular formula as 1. Upon comparing their 1 D- (Table S1) and 2 D-NMR spectrum, we see that the difference is mainly that the chemical shift of H-3 shifted up-field to 0.23 ppm in 2. The opposite optical value and Cotton effects in the ECD spectrum with 1 further implied that 2 was the configurational isomer of 1. The sugar was identified in the same method as the sugar in compound 1. The sugar was identified as 3 0 -O-methylb-D-glucopyranose by the HMBC correlations from -OCH3 to C-3 0 , the coupling constant (J ¼ 7.8 Hz) of H-1 0 (d H 4.65), HPLC analysis (Ito et al. 2004;Hu et al. 2015) of acid hydrolysis as well as the ESI-MS m/z 191.18 [M-H]and the specific rotation of the sugar part was þ53.6. Finally, the absolute configuration of 2 (2S) was determined by ECD analysis and calculation ( Figure S1).

Antiplatelet aggregation activity in vitro
The antiplatelet aggregation activity of compounds 1 2 16 was evaluated against thrombin-, collagen-and adenosine diphosphate (ADP)-induced rat platelet aggregation in vitro.
In the evaluation assay of platelet aggregation induced by collagen ( Figure S3), compound 1 at high and low concentrations also exhibited significant inhibition, compounds 2, 3, 4, 6, 7, 8, 9 and 11 displayed significant inhibition of platelet aggregation in three concentrations. Of these compounds 3 and 10 at 100 lM showed preferable antiplatelet aggregation activity with inhibitions of 62.73 ± 2.53% and 55.64 ± 4.94%, respectively, the inhibitory effect of both was stronger than positive control. On the contrary, compounds 14 and 16 at 25 lM showed significant facilitation of platelet aggregation.
In the evaluation assay of platelet aggregation induced by ADP ( Figure S4), compounds 1, 2,4,6,7,8,13,14 and 15 at three concentrations all showed significant inhibition of platelet aggregation. Compound 1 at 100 lM and 50 lM showed preferable antiplatelet aggregation activity with inhibitions of 30.44 ± 3.81% and 29.72 ± 2.89%, respectively. Similarly, compounds 3, 9, 11 and 16 at high and medium concentrations also showed notable inhibition of platelet aggregation induced by ADP. However, their inhibition of platelet aggregation was lower than that of aspirin.
It can be seen from the above that the antiplatelet aggregation activity of lignans and alkaloids is generally higher than that of other compounds, among which lignans have the highest activity. Lignans have been shown to have significant antiplatelet activity (Qu et al. 2014;Weng et al. 2004), which is consistent with our experimental results. Since compound 1 and compound 2 were only different in configuration, there was no significant difference in their anti-platelet aggregation activity, which could also be known from the results of molecular docking. Only compound 1 had a slightly higher antiplatelet aggregation activity than compound 2 when ADP was used as inducer.

Molecular Docking
Compounds 1 2 16 spontaneously bind to P2Y12 receptor with binding energy from À8.43 to À5.16 kcal/mol. Of these compound 2 binds spontaneously to the receptor relatively easily with a binding energy of À6.83 kcal/mol. In addition, compound 8 has the lowest binding energy of À8.43 kcal/mol. Compound 8

Conclusions
In conclusion, antiplatelet aggregation ingredients were discovered from the discarded sorghum root including two new multi-substituted 3H-indole alkaloids (1-2) and 14 known compounds (3-16). The antiplatelet aggregation activity of all compounds was enhanced and showed different degrees of inhibition on platelet aggregation. Compound 2 showed significant inhibition against collagen and ADP-induced rat platelet aggregation in vitro. Compounds 4, 6 and 8 showed significant antiplatelet aggregation activity against three inducers (thrombin, collagen and ADP). In addition, all compounds bind readily with two key receptor proteins (P2Y12 and COX-1) in the molecular docking test. These discoveries will further promote the use of discarded sorghum roots.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported by the National Natural Science Foundation of China (No. 81860693, U1812403, 82160807), Guizhou Province Innovation Capability Improvement Platform (No. 2020-4013), Characteristic Key Laboratory for the higher education institutions in Guizhou (No. 2020-018).