Characterization of seaweed hypoglycemic property with integration of virtual screening for identification of bioactive compounds

Journal of Functional Foods xxx (xxxx) xxxx

Contents lists available at ScienceDirect

Journal of Functional Foods journal homepage: www.elsevier.com/locate/jff

Characterization of seaweed hypoglycemic property with integration of virtual screening for identification of bioactive compounds Yao Xian China,1, Xin Chena,1, Wan Xiu Caoa, Yurizam Sharifuddinb, Brian D. Greenc, ⁎ Phaik Eem Limb,d, Chang Hu Xuea, Qing Juan Tanga, a

Human Health Research Laboratory, College of Food Science and Engineering, Ocean University of China, Qingdao 266003, Shandong, China Institute of Biological Sciences, Faculty of Science, University of Malaya, 50603 Kuala Lumpur, Malaysia c Institute for Global Food Security, School of Biological Sciences, Queen’s University Belfast, Belfast BT9 5AG, UK d Institute of Ocean and Earth Sciences, University of Malaya, Kuala Lumpur 50603, Malaysia b

A R T I C LE I N FO

A B S T R A C T

Keywords: Bioactive compound Hypoglycaemia Seaweed Virtual screening Molecular docking Surflex-Dock

Seaweeds have been studied extensively for their nutritional values but their potential nutraceutical application remained underutilized due to uncharacterized bioactive compounds. Here we demonstrated that water extracts of Kappaphycus, Halimeda, Padina and Sargassum were able to improve insulin resistance, reduced hyperglycemia and protect liver and pancreatic tissue from HFD-induced damage in mice, with both Padina and Sargasssum displayed more significant results than the other two seaweeds. A list of potential bioactive compounds was then composed by virtual screening of 276 compounds detected by LC-MS on selected Padina fractions using molecular docking by Surflex-Dock. Further analysis determined punicate as the most potent bioactive compound that inhibits both glucosidase and dipeptidyl-peptidase-4 enzymes. In conclusion, we discovered novel in vivo hypoglycemic activity in Halimeda and several potential α-glucosidase and DPP-4 inhibitors in Padina via virtual screening, demonstrating the efficacy of molecular docking to facilitate discovery of novel bioactive compounds.

1. Introduction

hyperglycemia, which leads to various health complications such as retinopathy, neuropathy and kidney failure (Bornfeldt & Tabas, 2011; Jellinger, 2009). Management of hyperglycemia has thus become a primary diabetes care strategy, targeting various phase and factors of glucose metabolism including α-glucosidase, dipeptidyl peptidase-4 (DPP-4), the incretin hormones gastric inhibitory polypeptide (GIP) and glucagon‐like peptide‐1 (GLP‐1) (Bonadonna, Borghi, Consoli, & Volpe, 2016; Derosa & Maffioli, 2012). Seaweed are edible large marine algae and represent a rich source of bioactive compounds. Recent studies have discovered a large range of potential bioactivities by seaweed and its components including antihyperglycemia and anti-hyperlipidemia (Chin et al., 2015, 2019; Paxman et al., 2008; Sharifuddin, Chin, Lim, & Phang, 2015). However, due to the large number of compounds present in a sample, isolating the bioactive compound is both laborious and costly, typically involving a series fractionation and chromatography separation (Sasidharan, Chen, Saravanan, Sundram, & Yoga Latha, 2011). Hence, many reports of bioactivities are on consumption of whole plant or extracts without isolating and identifying the bioactive compound that were responsible for the bioactivities. This unclarity limits the applicability of the results

Studies have repeatedly demonstrated the important role of diet towards human health, with growing evidence of dietary small bioactive molecules influencing various metabolic pathways and immune systems. Thus, utilizing dietary intervention to combat various chronic disease such as obesity, diabetes and even Parkinson’s disease has garnered much research interest (Schwiertz et al., 2010). Diabetes mellitus is a chronic metabolic disease that involved insufficient production of insulin (Type 1) or decreased insulin sensitivity (Type 2) (American Diabetes Association 2017). Globally, there are more than 422 million of diabetic sufferers with 1.6 million death directly attributed to the disease annually (WHO, 2018). Compared to Type 1, Type 2 diabetes mellitus (T2DM) poses the greater health problem as it accounts for nearly 90% of the global diabetes cases (Zheng, Ley, & Hu, 2018). The development of T2DM is closely linked to lifestyle, with obesity and physical inactivity being the largest risk factors (NIDDK, n.d.). One of the characteristics of T2DM is impaired glucose tolerance which is frequently associated with insulin resistance. The reduced response towards insulin would contribute to

⁎

Corresponding author. E-mail address: [email protected] (Q.J. Tang). 1 Co-first author. https://doi.org/10.1016/j.jff.2019.103656 Received 4 July 2019; Received in revised form 16 October 2019; Accepted 28 October 2019 1756-4646/ © 2019 Elsevier Ltd. All rights reserved.

Please cite this article as: Yao Xian Chin, et al., Journal of Functional Foods, https://doi.org/10.1016/j.jff.2019.103656

Journal of Functional Foods xxx (xxxx) xxxx

Y.X. Chin, et al.

2.3. Animal care and experimental design

to real life scenarios and impedes the development of a natural bioproduct for precision nutrition. The tremendous progression in bioinformatics has seen development of computer-aided drug design technology, which can be either direct (Structure-based Drug Design, SBDD) or indirect (Ligand-based Drug Design, LBDD). Virtual screening is an approach in SBDD whereby the 3D structures of small molecules is retrieved from large databases to find those that fit the binding sites of the receptor using docking program such as Sybyl-Flexx, Glide and others. Molecular docking is an important step in virtual screening, whereby small molecules are docked to the target sites of proteins according to their three-dimensional conformations and interaction with the catalytic sites of the protein ligands. After docking, the binding affinity are scored to find the optimum pairing of molecules via energy calculations, leading to a selection of plausible bioactive compounds considered to be appropriate for the given receptor (Sliwoski, Kothiwale, Meiler, & Lowe, 2014). Advances in algorithm have vastly improved the reliability of the predictions, making virtual screening a useful complementary tool for experimental high throughput screening (Spitzer & Jain, 2012). This study aims to explore using virtual screening technology to identify key bioactive compounds in four genera of seaweed (Kappaphycus, Halimeda, Padina and Sargassum) that would provide a shortlist of compounds for further investigation. The four seaweeds were previously tested positive in vitro for α-glucosidase and DPP-4 inhibitions. Here we further tested the in vivo anti-diabetic effects of the seaweeds in C57BL/6J mice fed with high-fat diet (HFD) and subsequently selected a fraction from Padina fractionation for liquid chromatography–mass spectrometry (LC-MS) analysis. Potential compounds were then selected using virtual screening by molecular docking with Surflex-Dock and compared with reports in literature to determine its feasibility.

All the experimental procedures were done in accordance to guidelines published by the National Institutes of Health (Guide for the Care and Use of Laboratory Animals, 8th edition). Approval for animal study was granted by the Committee on the Ethics of Animal Experiments of Ocean University of China (Approved protocol ID SCKK2012-0001). A power analysis was done using G*Power (version 3.1.9.2) and it was determined that at least seven animals per group is required to achieve statistical power of 0.95 with Type I error rate set at 0.05. Seventy-two male, SPF-grade, four weeks old C57BL/6J mice (14–18 g, n = 9) were purchased from Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). Each mouse was individually housed with free access to food and water on a 12 h light-dark cycle at 21–24 °C and 40–55% relative humidity. The beddings (wood shavings) were changed 3 times a week and the room were cleaned daily and disinfected with mild detergent every week. Animals were regularly checked for any signs of distress such as restlessness, dulling of fur and diarrhea. After acclimatized for a week on standardized chow diet (AIN93M), the mice were randomly assigned into 8 groups (n = 9). A group of mice (Control) fed with a low-fat diet (LFD; Research Diets Inc. D12450J) and the remaining 7 groups were fed with a HFD (Research Diets Inc., D12451J). After 8 weeks, each group of mice were respectively gavage daily with 200 μL of saline (Control and Model, 0.9% w/ v), respective seaweed water extracts (each 100 mg/kg body weight, b.w.), and positive controls Acarbose and berberine (each 30 mg/kg b.w.) for 4 weeks. Upon the completion of experiment, mice were fasted for 12 h before collecting blood samples via retro-orbital sinus. Mice were then sacrificed via cervical dislocation. Liver, muscle and pancreas were collected and weighted before storing at −80 °C along with samples of muscle tissues.

2. Material and methods

2.4. Oral glucose intolerance test (OGTT)

2.1. Preparation of samples

At week 12 of the animal study, mice were fasted for 12 h with water. Each mouse was then gavage 2 g/kg b.w. of glucose and blood sample were taken via lateral tail veins at time points of 0, 30, 60, 90 and 120 min. Blood were let clotted at RT for 30 min before obtaining serum through centrifugation (2000g, 10 min, 4 °C). Glucose concentrations in serum were determined using commercial kit (Biosino Bio-Tech, Beijing, China) according to manufacturer’s instructions.

The four genera of seaweed (Kappaphycus, Halimeda, Padina and Sargassum) for research were provided by the Institute of Ocean and Earth Sciences (IOES) of University of Malaya. The seaweeds were washed with running tap water to remove impurities such as mud and small mollusks before rinsing with distilled water. Washed seaweeds were then dried air-dried in a drying cabinet at 30 °C for 48 h. Dried seaweeds were kept in airtight bags until further use.

2.5. Biochemical analysis The glucose and insulin levels of serum were measured respectively using colorimetric kit (Biosino Bio-Tech, Beijing, China) and ELISA kit (Biocalvin, Suzhou, China) following manufacturer’s instructions. Insulin resistance was calculated using the Homeostatic Model Assessment of Insulin Resistance (HOMA-IR) with the formula: Insulin activity (mU/L) × Glucose concentration (mg/dL). Glycogen levels for liver and muscle tissues were measured using commercial kits (Biocalvin, Suzhou, China) according to manufacturer’s instructions.

2.2. Preparation of extracts Water extracts were prepared from dried seaweed using method as described previously. Briefly, dried seaweed were powdered using a conventional miller. Seaweed powder was added to boiling distilled water at a 1:50 ratio, and let steep for 30 min on a roller mixer. The mixture was then centrifuged at 3000g at room temperature for 20 min. The supernatant was then removed and condensed using rotatory evaporator. Condensed water extract was then oven-dried at 45 °C and stored at 4 °C until further use. For Kappaphycus, the method was modified to remove the hydrocolloid carrageenan. The seaweed powder was bagged with filter cloth and added to distilled water (1:50 w/v) at 75 °C with constant stirring for 4 h. Carrageenan was precipitated out of the water by adding 95% ethanol (3:1 ratio). Precipitated carrageenan was removed and the remaining solution condensed using rotatory evaporator. The condensate was then centrifuged (5000g, 20 min, 4 °C) to remove lingering traces of carrageenan. The supernatant was then oven-dried (45 °C) to get the water extract of Kappaphycus which was stored at 4 °C.

2.6. Histology Liver and pancreas tissues were fixed in 4% paraformaldehyde for 3 days, processed and embedded in paraffin wax. Thin sections (5 μm) were cut and placed on slides. Hepatic and pancreatic tissues were then stained with hematoxylin and eosin (H&E) and morphology observed under microscope (BX41, Olympus). 2.7. RNA extraction Total RNA was extracted from tissues using standard TRIzol (Invitrogen) method. In brief, 50–100 mg of frozen intestinal, hepatic 2

Journal of Functional Foods xxx (xxxx) xxxx

Y.X. Chin, et al.

at 1 mM and test samples were added into each well as triplicates. Subsequently, 20 µL of DPP-4 enzyme (G-CLONE, Beijing, China) was added to each well and the plate was incubated at 37 °C on a shaking incubator at 250 rpm for an hour. After an hour, 300 µL of 3 mM acetic acid was added to stop the reaction and fluorescence activity was read at excitation and emission wavelengths 351 nm and 430 nm respectively.

and muscle tissues were homogenized in 1 mL of TRIzol reagent at RT using a tissue homogenizer (Bioprep-24 homogenizer, Allsheng). After incubation for 5 min at RT, 200 μL chloroform were added to the homogenates and continue with incubation on ice for another 10 min. The samples were then centrifuged (12,000g, 10mins, 4 °C) and the aqueous layer was transferred to new tubes. Equal volume of isopropanol was then added to the samples and the mixtures were centrifuged (12,000g, 10 min, 4 °C) after incubation on ice for 10 min. The supernatants were discarded and pellet was resuspended in 1 mL of 75% ethanol. The samples were then centrifuged (7500g, 5 min, 4 °C) and air-dried for 5–10 min to obtain pellets of RNA which were resolubilized in 100 μL of DEPC-treated water. The purity, concentration and integrity of extracted RNA were checked using Nanodrop (Thermo Fisher, Waltham, Massachusetts, USA) and agarose gel electrophoresis.

2.11. Liquid Chromatography–Mass Spectrometry (LC-MS) analysis Bioactive fractions from flash chromatography was subjected to LCMS analysis to identify the compounds. Briefly, an Agilent 1290 Liquid Chromatography system (Agilent Technologies, Santa Clara, CA, USA) coupled to a 6520 Q-TOF tandem mass spectrometer was used to separate compounds from the samples. The mass detector was a Q-TOF accurate mass spectrometer equipped with an electrospray ionization (ESI) interface, positive mode and controlled by Mass Hunter software. Two µl of the sample comprising mixture of compounds were loaded on a 2.1 mm (i.d) Agilent Zorbax Eclipse Plus C-18 (length 100 mm) analytical column (particle size 1.8 μm) used with a flow rate of 0.25 mL/ min in a solution A (0.1% formic acid in water) and solution B (100% Acetonitrile with 0.1% formic acid). As for negative mode, the solvents used were solution A (0.1% ammonium formate in water) and solution B (100% Acetonitrile). The gradient was run as follows: 5–90% B for 35.0 min, 90–90% B for 6 min, 90–5% for 0.1 min and 5–5% for 6.9 min. The total gradient time for the LCMS run is 48 min. The ionization conditions were adjusted at 300 °C and 4000 V for capillary temperature and voltage, respectively. The nebulizer pressure was 45 psi and the nitrogen flow rate was 10 L/min. All mass spectrometry data were recorded in both positive and negative ion mode. The acquisition rate was at 1.03 spectra/s across the ranges 100–1000 m/z for both positive and negative. All the masses in the sample were than compared to the available plant metabolites internet database (www.plantcyc. org). The analysis only focused on the mass range between m/z 100–1000.

2.8. RT-qPCR For synthesis of complementary DNA (cDNA), 2 μg of RNA were added to 2 μL of random primers (Thermo Scientific, Waltham, Massachusetts, USA) and top up to 12.5 μL with DEPC-treated water. The RNA was heated to 70 °C for 5 min and immediately placed in −20 °C for 1 min. A12.5 μL mixture (5 μL 5X reaction buffer, 2 μL dNTP, 0.5 μL RNase Inhibitor, 1 μL M-MLV, 4 μL DEPC-water) was then added to the RNA solution and vortexed thoroughly. Reverse transcription was done on PCR machine (30 ℃, 10 min; 37℃, 60 min; 90℃, 5 min; maintained at 4℃). The cDNA purity and concentration were check using Nanodrop (Thermo Fisher, Waltham, Massachusetts, USA). The RT-qPCR was run on X960 real-time PCR system (Heal Force, Shanghai, China). The qPCR reaction consists of 10 μL of BrightGreen 2X qPCR MasterMix-S (ABM, Canada), 0.6 μL of 10 μM of primers (Table 1), 2.5 μL of cDNA template and 6.3 μL of nuclease-free water. The thermal profile included an initial denaturation step at 95 °C for 10 min followed by 45 cycles of denaturation at 95 °C for 15 s, annealing at 60 °C for 20 s and extension at 72 °C for 30 s. The mRNA expressions for prohormone convertase (PC3), phosphoenolpyruvate carboxykinase (PEPCK), glucose-6-phophotase (G6Pase) and proglucagon (gcg) were calculated using 2−ΔΔCt method and was presented as fold expression over the average control (N) gene expression taken as one.

2.12. Virtual screening The crystallized protein structures of human α-glucosidase and DPP4 enzymes were retrieved from Protein Data Bank (PDB) to use as base templates for construction of three-dimensional (3D) models of the enzymes. The most energetically and stereochemical favorable models for the two enzymes were build and validated with native ligands (ACR and cyc, for α-glucosidase and DPP-4 respectively) and commercial inhibitors Acarbose and Sitagliptin using ChemBio3D and Surflex-Dock. The molecular structures of compounds identified from LC-MS were obtained from PDB and docked to α-glucosidase and DPP-4 model respectively using Surflex-Dock function in Sybyl-x2.0 program package. The docked compounds were ranked accordingly (total scores and C score, with cutoff points at 6 and 3 respectively) for screening of potential compounds. Compounds that were deemed to be contaminants (i.e. non-plant based) were filtered out from subsequent testing. Purified compounds for top-hits were purchased commercially for further testing.

2.9. Flash chromatography Flash chromatography was conducted on Biotage Isolera system using a toluene-ethyl acetate-methanol gradient. Column used was SNAP Ultra 25 g (Biotage, Cardiff, UK) and the flow rate were set at 25 mL/min, for a total of 60 min. Fractions were then subjected to DPP4 assay for measurements of inhibitory activity. 2.10. DPP-4 assay Measurement for DPP-4 inhibitory activity was done in accordance to method previously described. Briefly, the assay was carried out in Grenier 96 well flat bottom (black) plate. A known DPP-4 inhibitor, the plant alkaloid Berberine (G-CLONE, Beijing, China) was used as a positive control at 1 mg/mL. The substrate, Gly-Pro-AMC (AAT Bioquest Inc., California, USA) was dissolved in phosphate-buffered saline (PBS) Table 1 Primer sequences for RT-qPCR. Gene name

Forward primer

Reverse primer

PEPCK G6Pase gcg PC3

GTAACCCGTTGAACCCCATT TCAGAAGCTGTTCTTGGTCTGAAC GCACATTCACCAGCGACTACA CTTCTTTTCTCTCAGCCCTTCCTAC

CCATCCAATCGGTAGTAGCG GTTCATGGGGATCCCAGAGA TGACGTTTGGCAATGTTGTTC CATTCATTGACAAACTGCCTCTTC

3

Journal of Functional Foods xxx (xxxx) xxxx

Y.X. Chin, et al.

3.2. Seaweed water extracts influenced gluconeogenesis and glycogenesis but not glucose absorption

2.13. Statistical analysis Data were presented as mean ± SEM values unless stated otherwise. Statistical analysis was performed using GraphPad Prism 7 and PASW Statistics 18 software. Significant differences were determined by one-way ANOVA with post hoc Tukey analysis at p values < 0.05.

Among the four seaweed extracts, both brown seaweed (Padina and Sargassum) extracts displayed significant inhibition on DPP-4 activity (Fig. 2A), although only Padina extract showed a modest decline in αglucosidase activity (Fig. 2B). We then measured the mRNA expression levels of selected genes that are involved in gluconeogenesis (Fig. 2C-F). No significant changes were observed in gcg expression while only Kappahycus water extract significantly increased expression of PC3 but not of G6Pase. Halimeda, Padina and Sargassum showed minute increment in PC3 expression and decreased G6Pase expression, although only Padina achieved significance. However, all seaweed water extracts significantly reduced the mRNA expression of PEPCK. Intake of seaweed water extracts did not seem to influence hepatic glycogenesis except for Sargassum which restored the glycogen level to normal. However, all seaweed extracts appeared to have partially increased muscle glycogen level although only Padina achieved significance (Supplementary data, Fig. S2).

3. Results 3.1. Seaweed water extracts reduced body weight and improved glucose tolerance in mice The seaweed water extracts were found to have a positive effect on the obese mice. Body weights of obese mice gavage with the extracts were reduced substantially (Supplementary data, Fig. S1) with the following order from most to least: Halimeda, Kappphycus, Padina and Sargasssum. Both Halimeda and Kappaphycus achieved final body weights that were less than both controls (Acarbose and berberine) while Padina and Sargassum final body weights were comparable to them. Of the four seaweed extracts tested, only two (Kappaphycus and Padina) extracts showed significant reduction in OGTT results, while the other two did not differ from the M group (Fig. 1A and B). We then proceed to investigate if the intake of seaweed extracts would improve insulin resistance. Although only Padina significantly reduced fasting glucose level (Fig. 1C), all four seaweed extracts successfully restored the insulin levels to levels comparable to N group (Fig. 1D) accompanied by significant drop in HOMA-IR values (Fig. 1E).

3.3. Protective effects of seaweed extracts from HFD-induced tissue damage Intake of a HFD induced signs of steatosis in hepatic tissues with appearances of empty vacuoles, distorted pancreatic islets with blurred edges and perforation of inflammatory cells in pancreatic tissues. Generally, all four seaweed extracts exhibited protective effects from HFD in liver and pancreas (Fig. 3A and B). 3.4. Initial screening for active fractions Padina was selected for screening since its extract achieved the best

Fig. 1. The effect of seaweed water extracts (100 mg/b.w.) on oral glucose tolerance test (OGTT) and insulin resistance. (A) OGTT results, (B) The area under curve of OGTT measurements, (C) Fasting Blood Glucose (FBG) levels, (D) Fasting Insulin (FINS) levels, (E) Homeostatic Model Assessment of Insulin Resistance (HOMA-IR). Values are given as mean ± SEM, n = 9. Statistical significances were calculated using one-way ANOVA with Tukey post-test (*p < 0.05, **p < 0.01, ### and ***p < 0.0005); # denotes versus N, * denotes versus M. 4

Journal of Functional Foods xxx (xxxx) xxxx

Y.X. Chin, et al.

Fig. 2. Enzyme activity of (A) Dipeptidyl peptidase-4 (DPP-4), (B) α-glucosidase; and gene expressions of (A) gcg, (B) PC3, (C) G6Pase and (D) PEPCK in mice. Values are given as mean ± SEM, n = 9. Statistical significances were calculated using one-way ANOVA with Tukey post-test (*p < 0.05, ## and **p < 0.01, ### p < 0.0005); # denotes versus N, * denotes versus M.

5

Journal of Functional Foods xxx (xxxx) xxxx

Y.X. Chin, et al.

Fig. 3. Histology at 100x magnification of (A) Liver, with model group showed signs of steatosis (circled); (B) Pancreas with model group showing infiltration of inflammatory cells (arrows). Bar = 100 μm.

positive controls, Acarbose and Sitagliptin (Supplementary data, Fig. S5) for minimal docking energy. The active site for α-glucosidase was determined to be a pocket surrounded by hydrophobic regions with key amino acids Aspartate (Asp)327, Asp542, Asp203, Histidine (His)600 and Arginine (Arg)526 providing the hydrogen bonding site (Supplementary data, Fig. S3). The active site for DPP-4 enzyme was determined on chain A of the dimer and consists of amino acids Glutamate (Glu)205 and Arg125 for hydrogen bonding while amino acids Glu206, Asn710, Phe357, Tyr662 and Tyr666 formed the hydrophobic region (Supplementary data, Fig. S4). The compounds present in P3 and P18 fractions were detected via LC-MS and a list of compounds were composed after identification using Plant Cyc database housed by Plant Metabolic Network (PMN). A total of 276 compounds were identified and docked to active sites of αglucosidase (PBD ID 2QMJ) and DPP-4 enzymes (PBD ID 2P8S-A) using Surflex-Dock. A list of eight potential active compounds (Tables 2 and 3) were composed for the respective enzyme according to the docking scores achieved by each compound.

overall results for anti-diabetic bioactivities. Flash chromatography produced a total of 99 fractions and selected fractions were tested for anti-DPP-4 activity. The inhibitory activity was detected in multiple fractions with continuous peaks forming at least five zones that were suggestive of active compound presence (Fig. 4A) with highest inhibition occurring at fraction 3 and 15. Further analysis (Fig. 4B) after adjusting for equal concentration confirmed the inhibitory effects of fractions number 3 (P3), 4 (P4) and 18 (P18). Fraction 15 was not used as it was colored, suggesting presence of pigments. The fractions P3, P4 and P18 were then subjected to LC-MS analysis to identify the compounds in them.

3.5. Identification and virtual screening of potential bioactive compounds The preferred crystallized structures for α-glucosidase and DPP-4 enzymes modelling were determined as PBD ID 2QMJ (Supplementary data, Fig. S3) and PBD ID 2P8S-A (Supplementary data, Fig. S4) in Protein Data Bank (PBD) after validation with native ligands and the 6

Journal of Functional Foods xxx (xxxx) xxxx

Y.X. Chin, et al.

DPP-4 inhibition 100 80 60 40 20 0

Bl

an k 1 2 3 4 9 15 17 19 20 21 25 30 31 38 41 45 48 50 53 56 59 62 65 68 69 72 74 76 78 79 80 84 83 88 91

DPP-4 Inhibition Percentage (%)

A

Padina fractions 100 80 60 40 20

P1 8 P8 1

P8

P7

P6

P5

P4

P3

la B

P2

0

nk

DPP-4 Inhibition Percentage (%)

B

Padina fractions (1mg/mL) Fig. 4. Screening of active fractions (10 mg/mL) from Padina for DPP-4 inhibitory activity. (A) Inhibitory effects of selected fractions without normalizing concentration. (B) Measurements of DPP-4 inhibition in selected fractions at 1 mg/mL concentration. Values are given as mean ± SEM, n = 9.

linolenate, arachidonate, 2-oleoylglycerol and phytol (22.32 ± 0.71%, 22.10 ± 0.96%, 20.17 ± 0.19%, 10.36 ± 0.12% and 12.74 ± 0.59%, respectively). One compound, 1-octadecanol exhibited negligible DPP-4 inhibition. The top four compounds in Tables 2 and 3 showed consistent molecular interactions with their respective proteins (Fig. 6C–J, Table S1).

Among the eight molecules, subsequent redocking with their target protein (Fig. 5) suggested that the docking positions of phytosphingosine, 1-Octadecanol, sterculate and cannabichromenate were not sufficiently conformed to their respective native ligand of the enzymes and thus removed from the list. Finally, the screened compounds that were picked after validation with original ligand were tested for their inhibitory activity. For αglucosidase inhibition (Fig. 6A), the best bioactivity was achieved by punicate (73.84 ± 1.84%) followed by α-linolenate, arachidonate and phytol (43.83 ± 5.77%, 45.88 ± 1.03% and 38.42 ± 1.10% respectively). One compound, 2-oleoylglycerol displayed negligible α-glucosidase inhibition. Punicate (26.16 ± 0.65%) also showed the most DPP-4 inhibitory activity (Fig. 6B) followed by α-tocotrienol, α-

4. Discussions Rapid advances in biomedical and food sciences has gave credence to the development of dietary intervention as effective strategy against a host of diseases such as metabolic syndrome (Shin, Lim, Sung, Shin, & Kim, 2009), diabetes (Lean et al., 2018) and dementia (Mumme et al.,

Table 2 Docking results for top hits obtained by virtual screening on compounds isolated from Padina (α-glucosidase). Name

Total score

G score

PMF score

D score

Chem score

C score

Phytosphingosine 1-Octadecanol Sterculate Punicate α-Linolenate Arachidonate Phytol 2-Oleoylglycerol

10.2248 8.9657 8.3271 8.2327 8.1595 8.0669 7.4393 7.0768

−209.827 −263.827 −257.348 −202.689 −223.616 −284.302 −258.598 −242.614

−127.702 −95.6923 −101.747 −99.6406 −108.749 −87.0013 −87.8804 −111.001

−198.941 −154.307 −145.915 −143.367 −135.578 −158.03 −158.872 −148.85

−31.3054 −36.9268 −30.6986 −31.9322 −27.5463 −31.4887 −34.0363 −23.5173

4 4 4 4 4 3 3 4

7

Journal of Functional Foods xxx (xxxx) xxxx

Y.X. Chin, et al.

Table 3 Docking results for top hits obtained by virtual screening on compounds isolated from Padina (DPP-4). Name

Total score

G score

PMF score

D score

Chem score

C score

Cannabichromenate 2-Oleoylglycerol Punicate α-Linolenate α-Tocotrienol Arachidonate Phytol 1-Octadecanol

8.6428 7.525 7.374 7.3283 7.2363 6.8396 6.8221 6.5318

−265.471 −257.02 −251.405 −204.389 −243.124 −232.286 −219.856 −174.076

−75.5533 −88.7841 −55.4464 −80.2378 −85.3661 −61.8444 −64.8967 −56.4777

−128.423 −155.312 −130.049 −116.445 −143.135 −122.842 −133.832 −112.935

−33.6563 −31.2229 −25.9228 −24.5107 −33.4279 −22.344 −26.4413 −23.801

4 4 3 3 4 3 3 3

(G6Pase) and phosphoenolpyruvate carboxykinase (PEPCK) are key enzymes in gluconeogenesis. Glucose-6-phophatase catalyzes the final step of gluconeogenesis, generating free glucose (Nordlie & Foster, 2010) while PEPCK is important as a cataplerotic enzyme that removing carbon from citric acid cycle for subsequent use by other biosynthetic and oxidative pathway, including glucose synthesis (Yang, Kalhan, & Hanson, 2009). Only Sargassum achieved significant reduction in G6Pase mRNA expression although both Padina and Halimeda also showed modest decrease. All treatment groups exhibited notable diminished PEPCK mRNA expression compared to HFD model with Sargassum being less effective than the other three seaweeds, which may be due to its lower level of G6Pase. Overall, the animal study revealed that the seaweed water extracts have hypoglycemic effects, with brown seaweeds generally being more potent than Kappaphycus and Halimeda. In order to fully realize the therapeutic potential of the seaweed extracts, we proceed to determine the bioactive compounds for α-glucosidase and DPP-4 inhibition of Padina, which had the best overall results among the seaweeds. Surprisingly, many compounds detected by LC-MS on selected fractions are not known as seaweed compound. Nonetheless, a final list potential compounds for α-glucosidase (5) and DPP-4 (7) inhibition was determined and evaluated after screening by Surflex-Dock. All compounds that showed α-glucosidase inhibitory effect also displayed significant DPP-4 inhibitory effect, albeit at a lower potency. Of all the compounds, punicate exhibited the highest potency for both α-glucosidase and DPP-4 inhibition. Punicic acid is a polyunsaturated fatty acid that is also classified as conjugated linolenic acid and is found in many sources including pomegranate seed oil, gourds and pot marigold. The compound has been reported for many biological effects including anti-obesity, anti-diabetic, anti-inflammatory, antioxidant and hypolipidemia (Aruna, Venkataramanamma, Singh, & Singh, 2016). Punicic acid was also reported to ameliorate insulin resistance in mice (Vroegrijk et al., 2011), which agreed with the findings in our study. However, the inhibition of α-glucosidase and DPP-4 by punicic acid was not described elsewhere. Besides punicate, α-linolenate, arachidonate and phytol also showed significant α-glucosidase and DPP-4 inhibition. Both α-linolenic acid (ω-3), arachidonic acid (ω6) are polyunsaturated fatty acids while phytol is a diterpene alcohol component in chlorophyll (NCBI, n.d.). Alpha-linolenic acid has been linked with increasing insulin sensitivity in the muscles (Bajaj et al., 2005), again showing consistency with our results. Interestingly, the compound 2-oleoylglycerol identified in our study is a monoacylglycerol that acts as a natural ligand to G protein-coupled receptor 119 (GPR119) and is reported to increase GLP-1 secretion in human intestine (Hansen et al., 2011). Our data suggested that 2-oleoylglycerol has modest anti-DPP-4 activity as well. The final compound, α-tocotrienol is considered as vitamin E and is poorly studied, although a study suggested that its presence in plasma could prevent stroke-related neurodegeneration (Khosla et al., 2006). As with the case with punicate, these compounds were not previously associated with inhibition of α-glucosidase and DPP-4. Scores assigned by Surflex-Dock are based on optimized poses of a putative ligand molecule when docked to a binding site. These poses are scored for their binding affinities in the following parameters:

2019). Hence, there is a growing demand for functional food, driving the development of valuable natural products from various sources including plants, fungi and algae. Here we describe the use of SurflexDock to aid the screening of potential bioactive compounds in seaweed water extracts that have anti-diabetic effects in C57BL/6J mice. Research has showed that excess fat cells contributes to insulin resistance, subsequently leading to diabetes mellitus. C57BL/6J mice were chosen as the animal model because of its manifestation of human-like metabolic syndrome including obesity, hyperglycemia and hyperinsulinemia when fed with a HFD (Mouse strain datasheet – 000664 – C57BL/6J, n.d.). The seaweeds (Halimeda, Kappaphycus, Padina and Sargassum) was previously shown as having anti-hyperglycemia or anti-obesity potential (Chin et al., 2015, 2019). Compared to mice fed with a HFD, mice gavaged with water extracts generally showed weight loss after 60 days of treatment. The glucose clearance rate was also returned to a normal level in Kappaphycus group, in agreement with our previous assessment (Chin et al., 2019), with significance improvement achieved by Padina as well. However, all treatment groups showed partial improvement in insulin resistance. Among the four, the brown seaweeds (Padina and Sargassum) performed better than the red (Kappaphycus) and green seaweed (Halimeda), suggesting presence of potent bioactive compound (s) common in brown seaweed. While in vivo improvement of insulin resistance was reported elsewhere for Kappaphycus, Padina and Sargassum (Cyriac & Eswaran, 2016; Park & Han, 2015; Park, Nam, & Han, 2015), this is a novel discovery for Halimeda. Examination of hepatic and pancreatic tissues suggested that treatments generally protect these tissues from HFD-induced damage that may lead to insulin resistance and hyperglycemia (Torres et al., 2011). As with the case of insulin resistance, both brown seaweed extracts performed better the other two seaweed extracts. Furthermore, brown seaweed appeared to increase glycogenesis which is reduced by hyperglycemia (Torres et al., 2011). The hallmark of diabetes, postprandial hyperglycemia, is characterized by a rapid and elevated blood sugar. The glucose level is mainly driven by glycogenolysis and gluconeogenesis, which are regulated by levels of glucagon and insulin. In turn, their secretion and availability is subjected to other factors such as dipeptidyl peptidase-4 (DPP-4) which inactivates the incretin hormone glucagon-like peptide1 (GLP-1) (Mentlein, 2009; Tarantola et al., 2012). Secreted by the gut and derived from proglucagon, GLP-1 stimulates insulin production and inhibits glucagon secretion (De Leon, Crutchlow, Ham, & Stoffers, 2006; Deacon & Ahren, 2011). Our results showed that both water extracts from brown seaweed significantly inhibits DPP-4 activity in the mice. No significance was found for α-glucosidase activity, although Padina, Kappaphycus and Halimeda each showed a lower or similar level to normal control, suggesting glucose absorption were not affected by these seaweeds. The genetic expression of proglucagon (gcg) in intestinal tissues was not generally not affected by the treatments, indicating that its availability for GLP-1 synthesis remain largely the same for all mice fed with a HFD, though Kappaphycus might have elicited an increase in proglucagon conversion to GLP-1 via elevated expression of prohormone convertase (PC3) mRNA. Both glucose-6-phophatase 8

Journal of Functional Foods xxx (xxxx) xxxx

Y.X. Chin, et al.

Fig. 5. The comparison between docking results (yellow) of top hits and the native ligand (grey) of DPP-4 (above) and α-glucosidase (below). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

hydrophobic complementarity, polar complementarity, entropic terms, and solvation terms (Jain, 2003). Our simulation suggested that our bioactive molecules seize the active site of the protein and form bonds

with the key amino acids around the protein catalytic center. The position of the substrate is then further stabilized through hydrophobic interactions. The top four molecules with highest α-glucosidase 9

Journal of Functional Foods xxx (xxxx) xxxx

Y.X. Chin, et al.

Fig. 6. The bioactivities of compounds (1 mg/mL) selected for top hits after docking by Surflex-Dock. (A) α-glucosidase inhibition, (B) DPP-4 inhibition, and the molecular interactions between ligands and (C-F) α-glucosidase, (G-J) DPP-4.

suggesting molecular docking as a useful screening tool. Finally, we reported novel in vivo anti-glucosidase and DPP-4 activities for Halimeda water extract and identified several potential bioactive compounds from Padina with previously unreported inhibition of glucosidase and DPP-4 enzymes.

inhibition bonded with key amino acids of the binding site and other amino acids as well. Similarly, three of the top four molecules with hits for DPP-4 inhibition bonded with key amino acids of the binding site but not with other amino acids. Intriguingly, one of the top hits for DPP4 inhibition, α-tocotrienol did not formed any hydrogen bonding with the key amino acids of DPP-4 enzyme, but rather the molecule was kept in place by multiple hydrophobic interactions with amino acids in the surrounding region. Since these compounds are only from a selected few of fractions from flash chromatography, we planned to utilize the same strategy to search for other, unique bioactive compounds for α-glucosidase and DPP-4 inhibition in other fractions of seaweed. At the same time, it would be interesting to determine whether the identified compounds in this study are native to seaweed.

Author contributions Y.X.C., X. C., and Q.J.T. designed research with input from Y.S, B.D.G., and C.H.X.; Y.X.C. and X.C. conducted research and data analysis with support from W.X.C., B.D.G., and P.E.L.; Y.X.C. and X.C. wrote the paper and reviewed by Q.J.T. All authors read and approve of the final manuscript.

5. Conclusions Funding In conclusion, Kappaphycus, Halimeda, Padina and Sargassum displayed hypoglycemic effects in mice. All four seaweeds ameliorated insulin resistance and influenced glucose homeostasis by impacting gluconeogenesis besides protecting tissues from HFD-induced damage. The integration of virtual screening in addition to conventional chromatography method has successfully identified several potential bioactive compounds in Padina that are supported by literatures,

This research was funded by the Ocean University of China via National Key R&D Program of China (grant no.: 2018YFC0311201), and University of Malaya via RG109-11SUS and UMQUB2A-2011 research grants. Funding and support from Queen’s University Belfast is also acknowledged.

10

Journal of Functional Foods xxx (xxxx) xxxx

Y.X. Chin, et al.

Declaration of Competing Interest

incretin hormones GLP-1 and GIP. Best Practice & Research: Clinical Endocrinology & Metabolism, 23(4), 443–452. https://doi.org/10.1016/j.beem.2009.03.005. Mouse strain datasheet – 000664 – C57BL/6J. (n.d.). Retrieved April 10, 2019, from https://www.jax.org/strain/000664. Mumme, K. D., von Hurst, P. R., Conlon, C. A., Jones, B., Haskell-Ramsay, C. F., Stonehouse, W., ... Beck, K. L. (2019). Study protocol: Associations between dietary patterns, cognitive function and metabolic syndrome in older adults – A cross-sectional study. BMC Public Health, 19(1), 535. https://doi.org/10.1186/s12889-0196900-4. NCBI (n.d.). PubChem database. Retrieved May 27, 2019, from https://pubchem.ncbi. nlm.nih.gov/. NIDDK (n.d.). Risk factors for type 2 diabetes. Retrieved May 28, 2019, from https:// www.niddk.nih.gov/health-information/diabetes/overview/risk-factors-type-2diabetes. Nordlie, R. C., & Foster, J. D. (2010). A retrospective review of the roles of multifunctional glucose-6-phosphatase in blood glucose homeostasis: Genesis of the tuning/retuning hypothesis. Life Sciences, 87(11–12), 339–349. https://doi.org/10. 1016/j.lfs.2010.06.021. Park, M. H., & Han, J.-S. (2015). Padina arborescens ameliorates hyperglycemia and dyslipidemia in C57BL/KsJ- db/db Mice, a model of type 2 diabetes mellitus. Journal of Medicinal Food, 18(10), 1088–1094. https://doi.org/10.1089/jmf.2014.3375. Park, M. H., Nam, Y. H., & Han, J. S. (2015). Sargassum coreanum extract alleviates hyperglycemia and improves insulin resistance in db/db diabetic mice. Nutrition Research and Practice, 9(5), 472–479. https://doi.org/10.4162/nrp.2015.9.5.472. Paxman, J. R., Richardson, J. C., Dettmar, P. W., Corfe, B. M., Paxmana, J. R., Richardsonc, J. C., ... Corfea, B. M. (2008). Alginate reduces the increased uptake of cholesterol and glucose in overweight male subjects: A pilot study. Nutrition Research, 28(8), 501–505. https://doi.org/10.1016/j.nutres.2008.05.008. Sasidharan, S., Chen, Y., Saravanan, D., Sundram, K. M., & Yoga Latha, L. (2011). Extraction, isolation and characterization of bioactive compounds from plants’ extracts. African Journal of Traditional, Complementary, and Alternative Medicines: AJTCAM, 8(1), 1–10. http://www.ncbi.nlm.nih.gov/pubmed/22238476. Schwiertz, A., Taras, D., Schäfer, K., Beijer, S., Bos, N. A., Donus, C., & Hardt, P. D. (2010). Microbiota and SCFA in lean and overweight healthy subjects. Obesity, 18(1), 190–195. https://doi.org/10.1038/oby.2009.167. Sharifuddin, Y., Chin, Y.-X., Lim, P.-E., & Phang, S.-M. (2015). Potential bioactive compounds from seaweed for diabetes management. Marine Drugs, 13(8), 5447–5491. https://doi.org/10.3390/md13085447. Shin, A., Lim, S.-Y., Sung, J., Shin, H.-R., & Kim, J. (2009). Dietary intake, eating habits, and metabolic syndrome in Korean men. Journal of the American Dietetic Association, 109(4), 633–640. https://doi.org/10.1016/J.JADA.2008.12.015. Sliwoski, G., Kothiwale, S., Meiler, J., & Lowe, E. W., Jr. (2014). Computational methods in drug discovery. Pharmacological Reviews, 66(1), 334–395. https://doi.org/10. 1124/pr.112.007336. Spitzer, R., & Jain, A. N. (2012). Surflex-Dock: Docking benchmarks and real-world application. Journal of Computer-Aided Molecular Design, 26(6), 687–699. https://doi. org/10.1007/s10822-011-9533-y. Tarantola, E., Bertone, V., Milanesi, G., Capelli, E., Ferrigno, A., Neri, D., ... Freitas, I. (2012). Dipeptidylpeptidase–IV, a key enzyme for the degradation of incretins and neuropeptides: Activity and expression in the liver of lean and obese rats. European Journal of Histochemistry, 56(4), e41. https://doi.org/10.4081/ejh.2012.e41. Torres, T. P., Fujimoto, Y., Donahue, E. P., Printz, R. L., Houseknecht, K. L., Treadway, J. L., & Shiota, M. (2011). Defective glycogenesis contributes toward the inability to suppress hepatic glucose production in response to hyperglycemia and hyperinsulinemia in zucker diabetic fatty rats. Diabetes, 60(9), 2225–2233. https://doi.org/ 10.2337/db09-1156. Vroegrijk, I. O. C. M., van Diepen, J. A., van den Berg, S., Westbroek, I., Keizer, H., Gambelli, L., ... Voshol, P. J. (2011). Pomegranate seed oil, a rich source of punicic acid, prevents diet-induced obesity and insulin resistance in mice. Food and Chemical Toxicology, 49(6), 1426–1430. https://doi.org/10.1016/J.FCT.2011.03.037. WHO (2018). Diabetes programme. Retrieved May 28, 2019, from World health organization website: https://www.who.int/health-topics/diabetes. Yang, J., Kalhan, S. C., & Hanson, R. W. (2009). What is the metabolic role of phosphoenolpyruvate carboxykinase? The Journal of Biological Chemistry, 284(40), 27025–27029. https://doi.org/10.1074/jbc.R109.040543. Zheng, Y., Ley, S. H., & Hu, F. B. (2018). Global aetiology and epidemiology of type 2 diabetes mellitus and its complications. Nature Reviews Endocrinology, 14(2), 88–98. https://doi.org/10.1038/nrendo.2017.151.

The authors declared that there is no conflict of interest. Appendix A. Supplementary material Supplementary data to this article can be found online at https:// doi.org/10.1016/j.jff.2019.103656. References American Diabetes Association, A. D. (2017). Classification and diagnosis of diabetes. Diabetes Care, 40(Suppl 1), S11–S24. https://doi.org/10.2337/dc17-S005. Aruna, P., Venkataramanamma, D., Singh, A. K., & Singh, R. P. (2016). Health benefits of punicic acid: A review. Comprehensive Reviews in Food Science and Food Safety, 15(1), 16–27. https://doi.org/10.1111/1541-4337.12171. Bajaj, M., Suraamornkul, S., Romanelli, A., Cline, G. W., Mandarino, L. J., Shulman, G. I., & DeFronzo, R. A. (2005). Effect of a sustained reduction in plasma free fatty acid concentration on intramuscular long-chain fatty acyl-CoAs and insulin action in type 2 diabetic patients. Diabetes, 54(11), 3148–3153. https://doi.org/10.2337/ DIABETES.54.11.3148. Bonadonna, R. C., Borghi, C., Consoli, A., & Volpe, M. (2016). Novel antidiabetic drugs and cardiovascular risk: Primum non nocere. Nutrition, Metabolism and Cardiovascular Diseases, 26(9), 759–766. https://doi.org/10.1016/j.numecd.2016.05.007. Bornfeldt, K. E., & Tabas, I. (2011). Insulin resistance, hyperglycemia, and atherosclerosis. Cell Metabolism, 14(5), 575–585. https://doi.org/10.1016/j.cmet.2011.07. 015. Chin, Y.-X., Lim, P.-E., Maggs, C. A., Phang, S.-M., Sharifuddin, Y., & Green, B. D. (2015). Anti-diabetic potential of selected Malaysian seaweeds. Journal of Applied Phycology, 27(5), 2137–2148. https://doi.org/10.1007/s10811-014-0462-8. Chin, Y. X., Mi, Y., Cao, W. X., Lim, P. E., Xue, C. H., Tang, Q. J., ... Tang, Q. J. (2019). A pilot study on anti-obesity mechanisms of Kappaphycus alvarezii: The role of native κ-carrageenan and the leftover sans-carrageenan fraction. Nutrients, 11(5), 1133. https://doi.org/10.3390/nu11051133. Cyriac, B., & Eswaran, K. (2016). Anti- hyperglycemic effect of aqueous extract of Kappaphycus alvarezii (Doty) Doty ex. P. Silva in alloxan-induced diabetic rats. Journal of Applied Phycology, 28(4), 2507–2513. https://doi.org/10.1007/s10811015-0762-7. De Leon, D. D., Crutchlow, M. F., Ham, J. Y., & Stoffers, D. A. (2006). Role of glucagonlike peptide-1 in the pathogenesis and treatment of diabetes mellitus. International Journal of Biochemistry & Cell Biology, 38(5–6), 845–859. https://doi.org/10.1016/j. biocel.2005.07.011. Deacon, C. F., & Ahren, B. (2011). Physiology of incretins in health and disease. The Review of Diabetic Studies, 8(3), 293–306. https://doi.org/10.1900/RDS.2011.8.293. Derosa, G., & Maffioli, P. (2012). α-Glucosidase inhibitors and their use in clinical practice. Archives of Medical Science, 8(5), 899–906. https://doi.org/10.5114/aoms. 2012.31621. Hansen, K. B., Rosenkilde, M. M., Knop, F. K., Wellner, N., Diep, T. A., Rehfeld, J. F., ... Hansen, H. S. (2011). 2-Oleoyl glycerol is a GPR119 agonist and signals GLP-1 release in humans. The Journal of Clinical Endocrinology & Metabolism, 96(9), E1409–E1417. https://doi.org/10.1210/jc.2011-0647. Jain, A. N. (2003). Surflex: Fully automatic flexible molecular docking using a molecular similarity-based search engine. Journal of Medicinal Chemistry, 46(4), 499–511. https://doi.org/10.1021/jm020406h. Jellinger, P. S. (2009). Metabolic consequences of hyperglycemia and insulin resistance. Insulin, 4(1), 2–14. https://doi.org/10.1016/S1557-0843(09)80003-1. Khosla, P., Patel, V., Whinter, J. M., Khanna, S., Rakhkovskaya, M., Roy, S., & Sen, C. K. (2006). Postprandial levels of the natural vitamin E tocotrienol in human circulation. Antioxidants & Redox Signaling, 8(5–6), 1059–1068. https://doi.org/10.1089/ars. 2006.8.1059. Lean, M. E., Leslie, W. S., Barnes, A. C., Brosnahan, N., Thom, G., McCombie, L., ... Taylor, R. (2018). Primary care-led weight management for remission of type 2 diabetes (DiRECT): An open-label, cluster-randomised trial. Lancet (London, England), 391(10120), 541–551. https://doi.org/10.1016/S0140-6736(17)33102-1. Mentlein, R. (2009). Mechanisms underlying the rapid degradation and elimination of the

11