Peer Review #2 of "Screening of a natural compound library identifies emodin, a natural compound from Rheum palmatum Linn that inhibits DPP4 (v0.1)"

1 College of Life Sciences, Fujian Normal University, Fuzhou, P. R. China 2 Engineering Research Center of Marine Biological Resource Comprehensive Utilization, Third Institute of Oceanography, State Oceanic Administration, Xiamen, P. R. China 3 Engineering Research Center of Marine Biological Resource Comprehensive Utilization, Third Institute of Oceanography, State Oceanic Administration, Subject Areas, Xiamen, P. R. China 4 Engineering Research Center of Marine Biological Resource Comprehensive Utilization, Third Institute of 10 Oceanography, State Oceanic Administration, Xiamen, P. R. China 5 Fujian Provincial Key Laboratory of Innovative Drug Target Research, School of Pharmaceutical Sciences, Xiamen, P. R. China


21
results showed that emodin, a natural compound from Rheum palmatum Linn, inhibited DPP4 activity 22 with an in vitro IC 50 of 5.76 µM without inhibiting either DPP8 or DPP9. A docking model revealed that 23 emodin binds to DPP4 protein through Glu205 and Glu206, although with low affinity. Moreover, 24 emodin treatment (3, 10 and 30 mg/kg, P.O.) in mice decreased plasma DPP4 activity in a dose-25 dependent manner. Our study suggests that emodin inhibits DPP4 activity and may represent a novel 26 therapeutic for the treatment of type 2 diabetes.
27 Main article text 28 Introduction 29 Type 2 diabetes mellitus (T2DM) is a metabolic disease associated with insulin resistance and pancreatic 30 β-cell failure (Defronzo 2009). Therefore, enhancing pancreatic insulin secretion while protecting 31 pancreatic β-cells represents a promising therapeutic approach for the treatment of type 2 diabetes. 32 Glucagon-like peptide 1 (GLP-1) is one of the incretin hormones released from cells in the gastrointestinal 33 tract in response to nutrient absorption. Incretin hormones, especially GLP-1, regulate post-prandial 34 insulin secretion by inhibiting glucagon release and stimulating insulin biosynthesis and secretion 35 (Baggio & Drucker 2007). In T2DM patients, GLP-1 is critical for glucose homeostasis (Mulvihill & 36 Drucker 2014). 37 115 To screen for novel DPP4 inhibitors from natural compounds, we established a natural compound library 116 comprising 155 naturally derived compounds, in which 131 were isolated and purified from Chinese 117 herbal medicines, and 24 were from marine organisms. DPP4 screening was first conducted on these 155 118 natural compounds by following a DPP4 screening assay kit protocol. All compounds (10 µM) were 119 screened for DPP4 inhibitory activity. The results suggested that emodin showed greater than 50% 120 inhibition in the DPP4 activity assay at 10 µM. Two other compounds were ruled out because of auto-121 fluorescence ( Figure 1A). 122 To further validate this finding, a dose response experiment was performed to test the inhibitory 123 activity of emodin on DPP4. Emodin was shown to inhibit DPP4 activity in vitro with an IC 50 of 5.76 µM 124 and Ki of 0.85 ( Figure 1B). The DPP4 antagonist sitagliptin was used as a positive control, which showed 125 an IC 50 of 21.78 nM ( Figure 1C), a value similar to those from previous reports (Kim et al. 2005). 126 Considering that emodin is an anthraquinone, and many naturally occurring anthraquinones have 130 Anthraquinone compounds inhibit DPP4 activity but not DPP8 or DPP9 activity 131 We further investigated a series of anthraquinone compounds in the DPP4 activity assay to identify 132 potent DPP4 inhibitors in this class. Aloe-emodin, rheochrysidin, chrysophanol and rhein were tested by 133 using the same DPP4 assay format (Table 1) (Table 1). These results showed that emodin was the most 136 effective anthraquinone in inhibiting DPP4 activity. In addition, the Ki and binding energy for these 137 compounds have also been listed in Table 1. Because of the high similarity between DPP4 and DPP8/9 138 and the reported toxicity of DPP8 or DPP9 inhibition in animal studies (Lankas et al. 2005), we tested the 139 anthraquinone compounds in DPP8 and DPP9 activity assays. All of the compounds were tested at 100 140 µM, and none showed activity against either DPP8 or DPP9. Rhein showed a very weak activity on 141 DPP8, with an IC50 greater than 100 µM. The biological function of rhein on DPP8 is minimal compared 142 to other DPP8 inhibitors. This result suggests that emodin is a relatively selective inhibitor against DPP4 143 (Figure 2). To confirm this finding, we conducted a molecular docking assay.
144 Emodin binds to Glu205 and Glu206 of DPP4 protein in a docking model 145 The active site of DPP4 consists of Arg125, Glu205, Glu206, Typ547, Trp629, Tyr666, and His740 146 according to the crystal structure template of DPP4 with a small molecular inhibitor (PDB code: 2ONC) 147 (Feng et al. 2007). Our docking model revealed that the negatively charged hydroxyl group of emodin is 148 engaged in tight H-bonding with Glu205 and Glu206 (Fig. 3A), suggesting a mechanism of binding of 149 emodin to the DPP4 active site. The binding modes showed that emodin was bound to the active site of 150 DPP4 with the hydroxyl moiety but did not form hydrogen bonds with other amino moieties such as 151 Tyr547 or Trp629 (Ji et al. 2014;Kim et al. 2005), which may affect the activity of emodin. The 152 compounds with similar structure as emodin that have hydroxyl group at similar site could also form H-153 bond with Glu205 and Glu206, and these compounds (aloe-emodin and rhein) also showed DPP4 154 inhibitory activity. In comparison, compounds without the hydroxyl group at R2 location (rheochrysidin 155 and chrysophanol) lack the ability to form H-bond with DPP4 at active site, thus they showed weakest 156 DPP4 inhibitory activity.Following dialysis assay suggested emodin binding to the DPP4 active site in a

223
To further understand the binding mechanism of emodin and DPP4, we conducted a molecular 224 docking model by simulating the emodin binding mode at the DPP4 active site. The docking assay 225 revealed that emodin interacts with the DPP4 active site and forms H-bonds with Glu205 and Glu206 at 226 the active site of DPP4.
227 Based on the in vitro data and the docking model, we subsequently conducted animal experiment by 228 orally administering emodin to Balb/C mice and ob/ob (-/-) mice. Plasma DPP4 activity was inhibited by 229 emodin administration in a dose-dependent manner, and the blood glucose levels were decreased in both 230 mice strains.  Manuscript to be reviewed