GS significantly promote aerobic glucose metabolism in cardiomyocytes
Increased OCR and ATP production are the manifestations of enhanced energy metabolism; thus, we performed targeted metabolomic analysis of glucose metabolites by LC-MS/MS in H9c2 and HUVECs untreated or treated with GS for 48 h. A total of 15 different intermediates of glycolysis and TCA cycle, such as pyruvate, lactate, citrate, and succinate were identified in these two cell lines. In H9c2 cells, GS led to decreases of seven metabolites and increases of eight metabolites, compared with the control group (Fig. 2A) . The analysis of relative intensity for each metabolite showed that GS decreased the levels of fructose 6-phosphate, fructose 1, 6-bisphosphate, 3-phosphoglycerate, dihydroxyacetone phosphate, and phosphoenolpyruvate, and increased the levels of glucose 6-phosphate and pyruvate in H9c2 cells, suggesting that GS increased the conversion of glucose to pyruvate. Moreover, key metabolites of TCA cycle such as citrate, succinate, fumarate, malate, oxaloacetate, and glycolytic metabolite, lactate were clearly increased by GS pretreatment (Fig. 2B) , suggesting that GS mainly accelerated TCA metabolism in the cardiomyocytes. In HUVECs, only six metabolites were upregulated by GS pretreatment, including glucose-6-phosphate, fructose-6-phosphate, pyruvate, succinate, malate, and fumarate. There were no obvious differences in the other nine metabolites between the control and GS groups (Fig. 2C, Supplementary Fig. 5A) . These data suggest that GS mainly increased the conversion from glucose to fructose-6-phosphate and pyruvate, and enhanced the metabolic process from succinate to malate in HUVECs. Collectively, GS had more obvious effects on promoting glucose oxidation in cardiomyocytes than endothelial cells.
Multiple key enzymes regulate glucose breakdown and TCA metabolism reaction, which were examined by Western blot analysis in the control and GS-treated H9c2 cells. GS pretreatment upregulated the levels to the rate-limiting enzymes such as HK-II, PFKP, PKM2, and PDH, and did not change GAPDH expression in H9c2 cells (Fig. 2D, Supplementary Fig. 5B) . Mitochondrial pyruvate carriers, MPC1 and MPC2, which play a potential role in the transport of pyruvate to mitochondria, were increased by GS compared with the control group (Fig. 2D, Supplementary Fig. 5B) . Furthermore, CS, the first and rate-limiting enzyme of the TCA cycle; DLST, one of the core enzymes of α-ketoglutarate dehydrogenase complex; and fumarase, a mitochondrial metabolic enzyme, were significantly higher in GS-pretreated H9c2 cells than the control group (Fig. 2D, Supplementary Fig. 5B) . Two isocitrate dehydrogenase isoforms in cytoplasm (IDH1) and mitochondria (IDH2) were not regulated by GS. In addition, ACO2, which is responsible for catalyzing the isomerization of citrate to isocitrate and SDHA, were decreased by GS pretreatment (Fig. 2D, Supplementary Fig. 5B) . Importantly, the activity of PDH, as a rate-limiting metabolic step for glucose oxidation was measured in untreated- and GS-pretreated H9c2 and HUVECS. We found a significant induction in PDH activity in GS-treated H9c2 cells and no change of PDH activity in GS-treated HUVECs compared with control cells (Fig. 2E) . Taken together, our findings suggest that GS pretreatment mainly accelerates glucose conversion to pyruvate, and increases pyruvate transporting to promote TCA metabolism in the cardiomyocytes. In future experiments, we will focus on the effect and molecular mechanism of GS on cardiomyocytes and neurons dominated by aerobic respiration.
GSincrease mitochondrial biogenesis incardiomyocytes and neurons
To investigate the role of GS in mitochondrial biogenesis, we first examined mitochondrial count in cardiomyocytes and neurons using the Mitotracker probe. We found that pretreatment with GS induced a significant increase in mitochondrial content in a dose-dependent manner in H9c2 cells (Fig. 3A-3B) and primary neurons (Fig. 3A and 3C) . Moreover, high-resolution micrographs of mitochondria were obtained in control and GS-pretreated H9c2 cells using TEM. As shown inFig. 3D-3E , GS pretreatment at 5 μg/mL for 48 h increased the number of mitochondria in the cardiomyocytes and had no effect on the size of myocardial mitochondria, compared with control cells. To determine if the observed changes in mitochondrial respiration and biogenesis translate into altered mitochondrial function, we performed immunoblotting with antibodies against five complexes of the oxidative phosphorylation. Compared with the control group, GS induced increases in complex I-IV levels in H9c2 cells and mainly increased the levels of CII-SDHB and CI-NDUFB8 in PC12 cells (Fig. 3F-3G) . The enhancement of mitochondrial function can increase ROS generation, which damages the cells, so we examined ROS level in untreated- and GS-treated H9c2 cells. Flow cytometry analysis showed that GS had no significant effect on the production of intracellular and mitochondrial ROS in H9c2 cells (Supplementary Fig. 6A) . Additionally, mitochondrial amount and complex protein level were further investigated in HUVECs. We did not observe the effect of GS pretreatment on mitochondrial mass in HUVECs (Supplementary Fig. 6B-6C) . Collectively, GS preferentially promote the mitochondrial biosynthesis of cardiomyocytes and neurons, and not that of endothelial cells.