Isolation, Bioactivity, and Molecular Docking of a Rare Gastrodin Isocitrate and Diverse Parishin Derivatives from Gastrodia elata Blume

Gastrodia elata Blume (G. elata) is a well-known medicine food homology plant widely used in treating neurological disorders such as Alzheimer’s disease (AD). Here, undiscovered gastrodin derivatives were systematically studied. Seven novel gastrodin derivatives (1–7), including a unique gastrodin isocitrate (1) and six differently substituted parishin derivatives (2–7), were isolated. Structural identification was mainly based on 1D and 2D NMR data, high-resolution ESI-MS data, and HPLC analysis. Notably, the stereochemistry of 1 was further elucidated by ECD calculations. Compounds 1 and 6 showed neuroprotective effects on the H2O2-induced PC12 cell injury model. Molecular docking analysis exhibited that 1 and 6 had good affinities with three popular AD-related targets. These findings not only enriched the chemical diversity but also revealed potential active components in G. elata.


INTRODUCTION
Alzheimer's disease (AD) is a common neurodegenerative disorder characterized by a gradual decline in cognitive function, memory impairment, and behavioral abnormalities. 1,2As the aging population increases worldwide, AD has emerged as a crucial public health issue that poses a significant threat to the health and life of the elderly. 3Unfortunately, there are no genuinely effective interventions available to treat, reverse, or delay the progression of AD.Therefore, the need for innovative bioactive molecules focused on the treatment or targeted prevention of AD has become increasingly urgent.Recently, an increasing number of people have adopted the concept of "food as medicine" in disease prevention, treatment, and health care. 4umerous epidemiological studies have indicated an association between consuming foods that are rich in polyphenols and preventing chronic diseases. 5−8 The promising medicine food homology plant Gastrodia elata Blume (G.elata) is an excellent example. 9,10−14 This has stimulated further exploration of the potentially active gastrodin derivatives in G. elata.
In this study, we performed a systematic and comprehensive investigation of gastrodin derivatives with potential activity in G. elata.First, gastrodin derivatives were isolated and purified by silica gel column chromatography (CC), microporous resin chromatography, and high-performance liquid chromatography (HPLC).Then, their structures were elucidated by NMR and HR-ESI-MS data, derivatization-HPLC analyses, and electronic circular dichroism (ECD) calculations.Next, a reactive oxygen species (ROS)-induced neuronal injury model, using H 2 O 2 to induce rapid contraction and apoptosis in PC12 cells, was applied for neuroprotective activity evaluation in vitro. 15Finally, molecular docking was used to predict the binding energies of the target proteins.Herein, we report the structure elucidation, neuroprotection, and molecular docking of gastrodin derivatives.
2.2.General Experimental Procedures.HR-ESI-MS data were measured using an Agilent 6230 accurate mass time-offlight mass spectrometer with an Agilent Eclipse Plus C18 RRHD column (2.1 × 100 mm, 1.8 μm).All semipreparative HPLC separations were equipped with an Agilent 1100 multiple-wavelength absorbance detector with a mobile phase using the H 2 O/ACN system and detected on a BEH phenyl column (250 × 10 mm, 5 μm, Waters) at 210 and 254 nm.All NMR data were collected by Bruker Ascend 600 NMR using methanol-d 4 as a solvent.A Rudolph Research Analytical Autopol I automatic polarimeter was used for recording optical rotations, a Beckman UV−Vis Spectrometer (DU-800) was used for acquiring UV spectra, and an Agilent Cary 600 series Fourier Transform Infrared (FT-IR) Spectrometer was used for obtaining IR spectra (KBr).

Plant Material.
The rhizome of G. elata was purchased from Guangyuan City of Sichuan Province, China, and authenticated by one of the authors, Dr. Ying Liu.

Derivatization-HPLC Analysis.
The sugar units of new compounds 1−7 were determined according to the reported method. 16Each compound (0.5 mg) was hydrolyzed in 0.5 M HCl at 95 °C for 2 h followed by neutralization with 0.5 M NaOH.After complete drying in vacuo, the residue was reacted with L-cysteine methyl ester in pyridine (0.1 mL; 5 mg/ mL) for 1 h at 60 °C.Next, o-toryl isothiocyanate dissolved in pyridine (0.1 mL; 5 mg/mL) was added and reacted for 1 h at the same temperature.Derivatization of D-(+)-glucose and L-(−)-glucose was carried out in the same way, while the reaction mixture without glucose was used as a blank control.Direct HPLC analysis of all reaction mixtures was conducted using a VisionHT C18 HL analytical column (250 × 4.6 mm, 5 μm; Grace) at 250 nm, and the mobile phase was 0.1% FA-containing 25% ACN.The flow rate and running time were 0.8 mL/min and 14 min, respectively.The absolute configurations of the sugar units in the new compounds were determined by a comparison of the t R values with those of the authentic standards.
2.6.ECD Calculations.The conformational analysis of compound 1 was conducted as follows.First, the 10 lowest energy state conformations were provided by the MMFF94S force field calculation in Sybyl-X 2.0 (5.0 kcal/mol).Second, the possible conformers were calculated and ranked using polarizable conductor calculation model by ORCA5.0.1 software at the B3LYP-D3(BJ)/6-31G* level. 17Subsequently, the parameters for the first 60 excited states of the main conformations were calculated at the PBE0/def2-TZVP level using timedependent density functional theory (TD-DFT). 18Finally, the ECD curves of possible isomers were simulated, and the absolute configuration of 1 was determined by a comparison of the experimental ECD curve to the simulated curves.

Cell Culture.
To investigate the neuroprotective effects of the isolated compounds, a H 2 O 2 -induced injury model was constructed in PC12 cells (Shanghai Institute of Biochemistry and Cell Biology, China).Cells were cultured in Dulbecco's modified Eagle's medium (Gibco, USA) supplemented with 10% fetal bovine serum (Gibco, USA), 5% horse serum (Gibco, USA), and 1% penicillin−streptomycin, then inoculated into 96well cell culture plates at a density of 1.0 × 10 4 per well, and incubated at 37 °C with 5% CO 2 for 24 h.The culture media of the control group and model group were replaced with 100 μL of a new culture media.Afterward, 100 μL of the test compounds (5 μM, 10 μM, and 20 μM) was added to the experimental group and incubated at the same temperature and CO 2 conditions for 4 h.After that, H 2 O 2 (10 μL, final concentration of 600 μM) was added to the model group, experimental group, and positive control (Trolox), and the culture was continued for 24 h.Finally, Cell Counting Kit-8 solution (Dojindo, Japan) was added to each well, and incubation was continued at 37 °C for 1 h.Immediately afterward, the optical density (OD) values were recorded with an enzyme labeling instrument at 450 nm.
The structures of all of the compounds were manually constructed, and the predicted binding mode was defined as the conformation with the lowest free energy.AutoDock vina docking 1.1.2software was used to dock the 3D minimumenergy ligands and proteins.
3.1.1.Gastrodin Isocitrate.Compound 1 was isolated as a light yellow, amorphous powder.The molecular formula was assigned as C 33 H 42 O 19 from HR-ESI-MS (Figure S1) at m/z 741.2274 (calcd 741.2248).Its IR spectrum showed the presence of hydroxyl groups (3356 cm −1 ), carboxylic groups (1736 cm −1 ), and aromatic rings (1612 and 1512 cm −1 ).As shown in Table 1, the ICA was deduced from the 1 H and 13 C chemical shifts, coupling constants (J), HSQC, and HMBC associations.The connection sequences were determined by 1 H− 1 H COSY correlations among H-1 (δ 4.44), H-2 (δ 3.44), and H-3 (δ 2.60/2.80), the HMBC correlations from H-1 to δ C 173.6 and δ C 172.  19,20 The absolute configurations of the glucoses were further determined by using HPLC analysis after thiocarbamoyl-thiazolidine derivatization.Only one peak was observed at 10.954 min in compound 1, which was close to that of authentic D-(+)-glucose (10.839 min) (Figure S13); thus, the absolute configuration of glucose should be the D-form.Moreover, each anomeric proton (H-1″) was related to C-4' of each HAr in the HMBC spectrum to form the gastrodin residue (GAr) (Figure 2).H-7′ of each GAr was found to be severally linked to the ICA unit via the HMBC correlations from δ H 4.95 (GAr′) to δ C 173.6 and from δ H 4.95 (GAr″) to δ C 172.2.Additionally, the carbon at δ C 52.3 (C-8') was assigned to a methoxy group in the ICA unit according to chemical shifts and HMBC correlations.To further determine the absolute configuration of 1, the theoretical ECD data were calculated by using TD-DFT (Tables S1−S11).As shown in Figure 3  (m/z 1101.3504,[M − H] − , Figure S14), was found to be a HAr-containing derivative of parishin A (compound 8) based on the difference in the molecular weight of C 7 H 6 O and the variations in the NMR spectra.As demonstrated in Figure 4 and Tables 2 and 3, similar parishin A unit-related signals were detected for a CA unit connected to three GArs via the HMBC correlations from δ H 4.97 (H-7′, GAr′) to δ C 170.9, from δ H 4.94 (H-7′, GAr″) to δ C 174.3, and from δ H 4.97 (H-7′, GAr‴) to δ C 170.9, respectively.In contrast, the presence of extra HAr was indicated by the presence of AA'BB'-type aromatic signals at δ H 6.72 (J = 8.4 Hz) and 7.12 (J = 8.4 Hz) and methylenoxy signals at δ H 4.42 (Table 2).The downfield shift of C-6″ (Δδ C + 8.0) in one GAr (Table 3) correlated with H-6″ (δ H 3.59/3.83)in HSQC was observed (Figure S19), and the HMBC correlation of C-6″ with H-7‴ of HAr suggested that this GAr was substituted by HAr at C-6″ (Figure S20).Symmetrical structural information similar to that of parishin A was embodied in two doublets of methylene groups at δ H 2.77 (2H, d, J = 15.3Hz) and 2.94 (2H, d, J = 15.3Hz) and two carbons at δ C 170.9 and 174.3 in the CA unit, which indicated that the etherified GAr was GAr″.The absolute configuration of the sugar was determined to be β-D-glucose following the same method as that used for 1 (Figure S13).Therefore, 2 was determined to be 1,3-di- Compound 3 was obtained as a yellow amorphous powder and possessed the same molecular formula as 2 according to HR-ESI-MS (Figure S25).Its NMR spectra showed similar HAretherified GAr signals to those of 2, as evidenced by one downfield shift of C-6″ (Δδ C + 8.0) in one GAr (Table 3) and the association of 1 H− 1 H COSY and HMBC (Figure S29− S31).However, in contrast to the CA unit signals of 2, four doublets of methylene groups at δ H 2.78 (1H, d, J = 15.3Hz), 2.95 (1H, d, J = 15.3Hz), 2.79 (1H, d, J = 15.3Hz), and 2.96 (1H, d, J = 15.3Hz) were observed, indicating structural asymmetry, which suggested that the etherified GAr was GAr′.Thus, 3 was identified as 2,3- Compound 4, a light-yellow amorphous powder, had the molecular formula C 54 H 62 O 27 (m/z 1141.3435,[M − H] − , Figure S36).As displayed in Table 3, in addition to the signals of the parishin A unit, 4 exhibited another trans-p-hydroxycinnamic acid residue (trans-HCiAr) according to the AA'BB'-type aromatic signals at δ H 6.82 (J = 8.5 Hz, H-3‴, 5‴) and 7.49 (J = 8.5 Hz, H-2‴, 6‴), a pair of trans double bond signals at δ H 7.69 (J = 16.1 Hz, H-7‴) and 6.43 (J = 16.1 Hz, H-8‴), and an ester carbonyl at δ C 169.1 (C-9‴).Thus, 4 was found to be a trans-HCiAr-containing derivative of parishin A. In addition, two doublets of two sets of methylene groups at δ H 2.78 (2H, d, J = 15.4Hz) and 2.96 (2H, d, J = 15.4Hz) in the CA unit were observed, which indicated symmetry, i.e., the trans-HCiAr was connected to GAr″.The HMBC correlation between H-3″ (δ H    S47). The significant difference in their NMR data lay in the HCiAr, which has a coupling constant of 12.8 Hz between H-7‴ (δ H 6.90) and H-8‴ (δ H 5.88), suggesting that it was a cis-HCiAr, whereas the signals of the parishin A unit were similar (Tables 2 and 3).The clear HMBC correlation between H-3″ (δ H 5.14) of GAr″ and C-9‴ (δ 168.1) of cis-HCiAr was sufficient to help determine the linkage sequence (Figure S49).Thus, 5 was elucidated as 1,3

-di-[4-O-(β-D-glucopyranosyl)benzyl]-2-{4-O-[3-O-(4-hydroxycis-cinnamoyl)-β-D-glucopyranosyl]benzyl}citrate.
Compound 6, with a molecular formula of C 54 H 62 O 26 , has one oxygen less than 4 (Figure S55).A comparison of their NMR data suggested that they had similar parishin A unit signals, and two doublets were observed at δ H 2.73 (2H, d, J = 15.4Hz) and 2.88 (2H, d, J = 15.4Hz), indicating that 6 had a symmetric structure (Tables 2 and 3 Compound 7 had the same molecular formula as 6 with similar spectral data (Figure S66).By comparison with the NMR data of 6 (Table 2 and Table 3), two doublets for two methylene groups in the CA unit showed that 7 was also a symmetrical molecule.The downfield shift of H-2″ (Δδ H + 1.64) correlated with C-2″ (δ 75.3) suggested that trans-CiAr was linked to the 2″-hydroxyl group of GAr″, which was confirmed by the HMBC correlation of C-9‴ (δ 167.8) with H-2″ of GAr″ (Figure S72).Therefore, the structure of 7 was determined to be 1,3-di-  12), 23 J (13), 24 K (14), 24 and Y (15), 25 were assigned by comparison with NMR data from the literature.It should be noted that parishin K was initially reported and named in 2015, 24 while other publications referred to a different structure without providing concrete evidence to support it. 26.2.Neuroprotective Effects of Gastrodin Derivatives.Next, a H 2 O 2 -induced neuronal injury model in PC12 cells was applied for neuroprotective activity evaluation in vitro (Table S12). 15Here, six representative gastrodin derivatives (1, 6, 7, 8, 14, and 15) were chosen for evaluation and further structure− activity relationship analysis using Trolox as a positive control (Figure 5).The cell viability was significantly reduced to 57.93% within 24 h of exposure to 600 μM H 2 O 2 compared to that of the control group.Remarkably, compound 1 demonstrated a promising neuroprotective effect by significantly reversing H 2 O 2 -induced PC12 cell injury in a concentration-dependent manner, increasing the cell viability to 78.44% at 20 μM (Figure 5A).However, compound 14 (parishin K), an isomer of 1 with a CA backbone, did not confer any protective effect when administered at the same treatment dose.Structural differences between compounds may have an effect on their biological activities.This outcome suggested that the ICA backbone was crucial in gastrodin derivatives, leading to a 20.51% increase in activity.The main difference between ICA and CA is the position of the hydroxyl group linkage on the backbone.In the ICA backbone, the hydroxyl group is located at C-1, while in the CA backbone, it is located at C-2.Thus, differences in the position of the hydroxyl group substituents on the backbone might influence the potential neuroprotective effects of these compounds.Additionally, it was important to mention that ICA, as an active unit, has evidenced its effectiveness in treating AD, 27 countering oxidative stress, 28 and mitigating heavy metalinduced neurotoxicity. 29Moreover, compound 6 also showed a concentration-dependent neuroprotective trend (Figure 5B).Given that the isomers (7 and 15) and the prototype parishin (8) of 6 had little neuroprotective activity, the activities of these compounds seemed to be affected by the esterification position (C-6″ of glucose in GAr).
3.3.Molecular Docking Simulation.AD is a multifactorial condition disease with an unclear etiology and complex pathogenesis. 30,31Molecular docking simulations will benefit the prediction of the action of drug molecules at the molecular level and provide a theoretical basis for subsequent target identification.−36 Keap1-Nrf2 is a crucial element in regulating oxidative stress and a promising therapeutic target for AD. 37Molecular docking studies provided detailed information on the interactions of 1 and 6 with the Keap1-Nrf2 protein (PDB entry 4L7B), which helped to predict their binding modes, recognize interacting residues, and calculate binding energies (Table S13). 1 and 6 exhibited potential inhibitory effects, with binding energies of −8.9 and −9.2 kcal/mol, respectively (Trolox: −7.5 kcal/mol).The 3D diagrams shown in Figure 6 displayed the interactions of two new active ingredients, 1 and 6, with the Keap1-Nrf2 protein.Additionally, the complexity of AD has prompted a search for more potential targets.Currently, studies on AD are mainly targeting Aβ, 32 and BACE1 has been associated with the production of Aβ. 38 Additionally, APOE is a major genetic risk factor for AD, and its allele APOE4 further increases the risk of AD. 34 Therefore, these two currently popular target proteins, BACE1 (PDB entry 1M4H) and APOE4 (PDB entry 1B68), were also selected for prediction (Figure 6).S13).

Figure 3 .
Figure 3. Experimental and calculated ECD spectra of compound 1.

Table 2 .
1H NMR (600 MHz) Data of Compounds 2−7 and Parishin A(8)in MeOD (δ in ppm, J in Hz) 13ble3.13CNMR (150 MHz) Data of Compounds 2−7 and Parishin A(8)in MeOD (δ in ppm) this study, we performed a systematic investigation of gastrodin derivatives in G. elata.A rare gastrodin isocitrate (1), six new parishin derivatives (2−7), and eight known parishins(8−15)were isolated from G. elata by various chromatographic separation techniques.Their structures were elucidated by a comprehensive analysis of HR-ESI-MS, NMR, and HPLC data, and the stereochemistry of compound 1 was further determined via ECD calculations.Remarkably, compound 1 was the first gastrodin isocitrate with a (1R,2S)-ICA backbone found in nature.The other six new derivatives were parishins with different substitutions, including two etherified parishins (2−3) and four esterified parishins (4−7).Subsequent bioactivity studies indicated that 1 and 6 could reverse H 2 O 2 -induced PC12 cell injury and had potential neuroprotective effects.Molecular docking further predicted their affinities with the target proteins.Our study uncovered potential novel bioactive ingredients in G. elata that deserve further exploration.The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.4c00436.Spectroscopic data of compounds 1−7; HPLC analysis after thiocarbamoyl-thiazolidine derivatization; ECD calculations of compound 1; effects of gastrodin derivatives on the H 2 O 2 -induced PC12 cells injury model; and molecular docking results of representative compounds (PDF) Jian-Lin Wu − State Key Laboratory of Quality Research in Chinese Medicine, Macau Institute for Applied Research in Medicine and Health, Macau University of Science and Technology, Taipa 999078, China; orcid.org/0000-0002-3875-185X;Email: jlwu@must.edu.moNa Li − State Key Laboratory of Quality Research in Chinese Medicine, Macau Institute for Applied Research in Medicine and Health, Macau University of Science and Technology, Taipa 999078, China; orcid.org/0000-0002-0404-6431;Email: nli@must.edu.moState Key Laboratory of Quality Research in Chinese Medicine, Macau Institute for Applied Research in Medicine and Health, Macau University of Science and Technology, Taipa 999078, China Jia-Qian Chen − State Key Laboratory of Quality Research in Chinese Medicine, Macau Institute for Applied Research in Medicine and Health, Macau University of Science and Technology, Taipa 999078, China Shilin Gong − State Key Laboratory of Quality Research in Chinese Medicine, Macau Institute for Applied Research in Medicine and Health, Macau University of Science and Technology, Taipa 999078, China; orcid.org/0000-0001-7707-417XYu-Juan Ban − State Key Laboratory of Quality Research in Chinese Medicine, Macau Institute for Applied Research in Medicine and Health, Macau University of Science and Technology, Taipa 999078, China Li Zhang − State Key Laboratory of Quality Research in Chinese Medicine, Macau Institute for Applied Research in Medicine and Health, Macau University of Science and Technology, Taipa 999078, China Ying Liu − School of Basic Medicinal Sciences and Nursing, Chengdu University, Chengdu 610106, PR China AuthorsJie Zhou −