Identification of Phytochemicals in Bioactive Extracts of Acacia saligna Growing in Australia

Acacia saligna growing in Australia has not been fully investigated for its bioactive phytochemicals. Sequential polarity-based extraction was employed to provide four different extracts from individual parts of A. saligna. Bioactive extracts were determined using in vitro antioxidant and yeast α-glucosidase inhibitory assays. Methanolic extracts from barks, leaves, and flowers are the most active and have no toxicity against 3T3-L1 adipocytes. Compound isolation of bioactive extracts provided us with ten compounds. Among them are two novel natural products; naringenin-7-O-α-L-arabinopyranoside 2 and (3S*,5S*)-3-hydroxy-5-(2-aminoethyl) dihydrofuran-2(3H)-one 9. D-(+)-pinitol 5a (from barks and flowers), (−)-pinitol 5b (exclusively from leaf), and 2,4-di-t-butylphenol 7 are known natural products and new to A. saligna. (−)-Epicatechin 6, quercitrin 4, and myricitrin 8 showed potent antioxidant activities consistently in DPPH and ABTS assays. (−)-Epicatechin 6 (IC50 = 63.58 μM), D-(+)-pinitol 5a (IC50 = 74.69 μM), and naringenin 1 (IC50 = 89.71 μM) are the strong inhibitors against the α-glucosidase enzyme. The presence of these compounds supports the activities exerted in our methanolic extracts. The presence of 2,4-di-t-butylphenol 7 may support the reported allelopathic and antifungal activities. The outcome of this study indicates the potential of Australian A. saligna as a rich source of bioactive compounds for drug discovery targeting type 2 diabetes.


Introduction
Acacia saligna (Labill.) H. L. Wendl. is a Western Australian species previously known as A. cyanophylla Lindl [1]. A. saligna plants growing in Saudi Arabia, Egypt, Tunisia, and other parts of Africa have been shown to have various bioactive phytochemicals. Flavonoids from the flower were shown to have antifungal, antioxidant, antiacetylcholinesterase, and antibacterial activity [2][3][4]. Volatile phytochemicals of A. saligna, possessing allelopathic activity, indicate their potential to be green herbicides [5,6]. The leaf extracts containing polyphenols have demonstrated antibacterial and antifungal activities, while some isolated compounds have exhibited antioxidant and cytotoxicity against liver cancer cells [7][8][9]. Two recent reports on the ethanolic crude extract of barks showed antifungal and antioxidant activities [10], and α-glucosidase inhibitory activity [11]. A. saligna has also been used in the Middle East, Africa, and South America as ruminants' fodder [12][13][14]. The utility of A. saligna as animal feed indicates its low toxicity and high nutritional benefit. The dose-response DPPH scavenging activities of all extracts from flowers, leaves, and bark expressed in percentage of activity and IC 50 are presented in Supplementary Tables S1 and S2 for vitamin C. Figure 1a displays the dose-response curves of the most active BK-MeOH extract. Table 2 shows the IC 50 value of DPPH scavenging of each extract. The BK-MeOH extract (IC 50 = 94.24 ± 19.89 µg/mL) has the highest antioxidant activity, followed by LF-MeOH (IC 50 = 190.1 ± 59.15 µg/mL) and FL-MeOH (IC 50 = 331.5 ± 17.21 µg/mL). Compared to vitamin C (49.97 ± 10.76 µg/mL), the decreasing antioxidant activities of the methanolic extracts can be expressed as BK-MeOH > LF-MeOH > FL-MeOH. This trend agrees with the previously reported finding that polar organic solvent extract seemed to have better antioxidant activity due to its high polyphenols content [24].
This trend agrees with the previously reported finding that polar organic solvent extract seemed to have better antioxidant activity due to its high polyphenols content [24].
Using a similar DPPH method, the crude ethyl acetate extract from flowers collected in Tunisia was shown to have an IC50 of 67 µg/mL [2], while the water flower extract from Egyptian A. saligna showed poor activity with an IC50 of 461.7 µg/mL [3]. Elansary et al. [4] showed that their crude methanolic extract of leaves collected in Saudi Arabia has a potent antioxidant activity with an IC50 of 17 µg/mL. The crude methanolic extract from barks collected in Egypt was reported [10] to have an IC50 of 10.1 µg/mL. The variation in activities may mainly be attributed to each extract's different chemical compositions affected by the growing conditions [25] and methods of extraction and assay.  The dose-response ABTS •+ radical scavenging activities of all extracts are summarised in Supplementary Tables S3 and S4 for vitamin C-positive control. Figure 1b displays the dose-response curves of ABTS •+ scavenging percentage for BK-MeOH extract. The IC50 values for the three extracts are listed in Table 2. Similar to the DPPH scavenging activity, the trend of antioxidant activities in this ABTS radical assay indicates that all methanolic extracts exert higher activities than their counterparts, compared to vitamin C. The decreasing antioxidant activities of the methanolic extracts can be expressed as BK-MeOH > LF-MeOH > FL-MeOH, compared to vitamin C. Interestingly, BK-MeOH was slightly more active than vitamin C against ABTS •+ radical.  Using a similar DPPH method, the crude ethyl acetate extract from flowers collected in Tunisia was shown to have an IC 50 of 67 µg/mL [2], while the water flower extract from Egyptian A. saligna showed poor activity with an IC 50 of 461.7 µg/mL [3]. Elansary et al. [4] showed that their crude methanolic extract of leaves collected in Saudi Arabia has a potent antioxidant activity with an IC 50 of 17 µg/mL. The crude methanolic extract from barks collected in Egypt was reported [10] to have an IC 50 of 10.1 µg/mL. The variation in activities may mainly be attributed to each extract's different chemical compositions affected by the growing conditions [25] and methods of extraction and assay.

ABTS •+ Radical Assay
The dose-response ABTS •+ radical scavenging activities of all extracts are summarised in Supplementary Tables S3 and S4 for vitamin C-positive control. Figure 1b displays the dose-response curves of ABTS •+ scavenging percentage for BK-MeOH extract. The IC 50 values for the three extracts are listed in Table 2. Similar to the DPPH scavenging activity, the trend of antioxidant activities in this ABTS radical assay indicates that all methanolic extracts exert higher activities than their counterparts, compared to vitamin C. The decreasing antioxidant activities of the methanolic extracts can be expressed as BK-MeOH > LF-MeOH > FL-MeOH, compared to vitamin C. Interestingly, BK-MeOH was slightly more active than vitamin C against ABTS •+ radical.

Inhibition of Yeast A-Glucosidase Enzyme Assay
The dose-response inhibitory activities of the plant extracts and the positive control acarbose against the yeast α-glucosidase enzyme are shown in Supplementary Tables S5-S8  and Figures S2-S5. The detectable IC 50 values are listed in Table 3. Non-alcoholic extracts, except aqueous barks extract (BK-H 2 O), are less active than the methanolic samples. BK-MeOH extract showed superior inhibitory activity (IC 50 4.373 ± 0.24 µg/mL) compared to BK-H 2 O and the two methanolic counterparts. This value is comparable with the inhibitory activity of crude ethanolic bark and leaf extracts of the South African A. saligna with IC 50 of 2.35 µg/mL and 3.64 µg/mL, respectively [11]. Notably, the inhibitory activity of crude ethanolic leaf extract is more potent than our LF-MeOH.

Toxicity of Bioactive Methanolic Fractions against 3T3-L1 Adipocytes
FL-MeOH, LF-MeOH, and BK-MeOH were tested for their toxicity against 3T3-L1 adipocytes using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) assay. The 3T3-L1 adipocytes were used in this study because they are the ideal cell model suitable for investigating the antidiabetic activities of A. saligna. The cell line can provide an excellent model of white adipose tissue to investigate glucose uptake, lipogenesis, and glycogen synthesis under an insulin-resistant state [26]. Figure 2a shows the results of the MTT assay on 3T3-L1 adipocytes treated with four different concentrations of FL-MeOH (25-200 µg/mL) for 24, 48, and 72 h. The cell viability was estimated between 93 and 116%. FL-MeOH showed no toxic effects at the highest test concentration (200 µg/mL). Similarly, LF-MeOH and BK-MeOH also showed no toxicity against 3T3-L1 adipocytes at 200 µg/mL after incubation for 72 h (Figure 2b,c).

Isolation of Pure Compounds Active Fractions
The most bioactive extracts were selected for compound isolation by column chromatography. Isolated compounds were structure elucidated by NMR and HRMS analysis. We have successfully isolated ten compounds through this effort, as shown in Figure 3. Elansary et al. [4] reported the non-cytotoxic activity of their crude methanolic extract of A. saligna leaves from Saudi Arabia against HEK-293 (non-cancerous human embryonic kidney cells). Their MTT assays revealed no significant toxicity against HEK-293 cells at the highest test concentration of 400 µg/mL. Buttner et al. [11] also reported the non-toxicity of their leaf and bark extracts of A. saligna from Saudi Arabia against Caco-2 cells at the highest test concentration of 300 mg/mL. Although our non-toxic results against murine 3T3-L1 cells cannot be directly compared to these reported results; however, their findings have highlighted the non-toxicity potential of A. saligna.

Isolation of Pure Compounds Active Fractions
The most bioactive extracts were selected for compound isolation by column chromatography. Isolated compounds were structure elucidated by NMR and HRMS analysis. We have successfully isolated ten compounds through this effort, as shown in Figure 3. Among these, compounds 2 and 9 are novel. Their isolation and identification are discussed in the following sections.

Isolation of Pure Compounds Active Fractions
The most bioactive extracts were selected for compound isolation by column chromatography. Isolated compounds were structure elucidated by NMR and HRMS analysis. We have successfully isolated ten compounds through this effort, as shown in Figure 3. Among these, compounds 2 and 9 are novel. Their isolation and identification are discussed in the following sections.

Isolation and Structural Identification of Compounds from FL-MeOH
FL-MeOH has been prioritised for pure compound isolation because of its superior biological activities and non-toxicity. Two successive column purification of FL-MeOH (Supplementary Figure S6) was performed to give five pure fractions.
Fraction FL-MeOH-A1 is a yellow solid. Complete spectral data analysis of FL-MeOH-A1 revealed that it is naringenin 1 (Figure 3). The NMR data provided in Supplementary Table S9 agrees with that reported in the literature [27,28]. The work by Al-Huqail et al. only reported the presence of naringenin 1 in the water extract of A. saligna flowers via their HPLC analysis. Although naringenin 1 has been isolated from other plants [29], it is the first time that we have isolated pure naringenin 1 from the flowers of Australian A. saligna. Our study showed that the specific optical rotation of naringenin 1 was [α] 23 = −16.68 • (c 0.1, EtOH), which is slightly lower than the previous report [α] 22 = −14.7 • (c 0.36, EtOH). This finding confirms the stereochemistry of laevorotatory (−)-or (2S)-flavanone comparable to the reported naringenin 1 [28].
The attachment of the sugar moiety at 7-O of aromatic ring A was confirmed by C1"-C7 connectivity observed in the 2D HMBC NMR experiment ( Figure 4a). Moreover, NOESY NMR ( Figure 4b) showed strong cross-peaks between H1" of the sugar and H8 of aromatic ring A, representing a closer space configuration. The IR spectra gave prominent bands at 3300.70 (OH stretching), 2920.64 (sp 3  FL-MeOH-B1b was isolated as a yellow powder. A complete spectral data interpretation (Supplementary Table S10) of FL-MeOH-B1b was confirmed to be isosalipurposide 3 ( Figure 3), which agrees with the literature [2,31]. Isosalipurposide 3 was previously isolated from flowers of A. saligna by various groups [2,32,33]. FL-MeOH-B2b was obtained as a yellow powder. Its complete spectral data analysis of this fraction revealed the structure to be quercetin-3-O-rhamnoside, also known as quercitrin 4 ( Figure 3). The NMR data of FL-MeOH-B2b (Supplementary Table S11) agrees with that reported in the literature [34]. Quercitrin 4 was previously identified in the leaves [8]; it is noteworthy that this is the first time that quercitrin 4 has been isolated from the flowers of Australian A. saligna.
FL-MeOH-B3b was isolated as a white solid. Complete spectral data analysis (Supplementary Table S12) of this fraction revealed a structure identical to those reported for D-(+)-pinitol 5a (Figure 3). The optical rotation of FL-MeOH-B3b was found to be [α] 23 [35]. This inositol ether was first isolated in the sugar pine (Pinus lambertiana) and could occur in various plants with both enantiomers [36,37]. It was first documented in the related genus A. nilotica [38]. However, it has never been isolated as a single enantiomer of pinitol from A. saligna flower until now. The compound has been reported as an antidiabetic compound [39,40].
Fractions FL-MeOH-A2, -B1a, -B2a, -B3a, and -C (Supplementary Figure S6) were not further purified due to their low quantities and high impurities. LF-MeOH-A1 is a colourless solid. Complete spectral data analysis of the compound revealed that it is (−)-epicatechin 6 ( Figure 3), which has spectroscopic data (Supplementary Table S13) identical to that reported [41]. The isolation of compound 6 from the leaves of A. saligna was previously reported by E-Toumy et al. [7]. However, this study did not provide the compound's specific optical rotation and absolute stereochemistry. In our work, the optical rotation of LF-MeOH A1 was determined to be [α] 23 = −69.6 • C (c 0.1, MeOH), identical to the literature value [41] of (−)-epicatechin 6 isolated from other plants.
LF-MeOH-A3 is a yellow solid identified by complete spectral data analysis (Supplementary Table S14) as 2,4-di-t-butylphenol 7 [42,43] (Figure 3). This compound has been known as a natural toxin and isolated from different groups of organisms, including plants, e.g., sweet potatoes [42] and pine trees [44]. However, for the first time, our group isolated it from the leaves of A. saligna. It has also been identified as an antioxidant [45], anticancer, antiviral [46], antibacterial, and antifungal [43]. Moreover, as 2,4-di-t-butylphenol 7 demonstrated allelopathic activities against weeds and lettuces [47,48], this compound could also be linked to the reported herbicide properties of A. saligna. Indeed, previous works have demonstrated the strong allelopathic [5,49,50] and antifungal activities [4] of A. saligna leaves. However, none of those reports identified 2,4-di-t-butylphenol 7 and tested its potential.
LF-MeOH-B2 is a yellow powder, similar to FL-MeOH-B2b, and was fully confirmed by spectral data to be quercitrin 7 ( Figure 3).
The LF-MeOH-C2b is a yellow powder. Complete spectral data analysis revealed that C2 is myricetin-3-O-rhamnoside (Myricitrin) 8 ( Figure 3). NMR data listed in Supplementary Table S14 are identical to those reported in the literature [51].
The LF-MeOH-C3 was isolated as a white solid. Complete NMR and HRMS analysis initially revealed that this compound is identical to D-(+)-pinitol 5a. However, its optical rotation  Table 5). The cross-peaks correlation of TOCSY demonstrated three patterns for (1) H5 to H3 and H4 as well as (2) H5 to H6 and H7; and (3) H4 to H3a and H5. HMBC revealed the key correlations, as shown in Figure 5a. This confirms that a five-membered lactone ring is the core of the compound. The aminoethyl substituent attached to the lactone ring at C5 via C6. The second OH group could be at C4 or C3. The chemical shift of C-3 indicates that the OH is appropriately attached to C3. NOESY NMR analysis (Figure 5b) showed strong crossed peaks between H3-H4a and H5-H4b, indicating that H5 is trans to H3. Our spectral data analysis, therefore, concludes LF-MeOH-D to be (3S,5S)-3-hydroxy-5-(2-aminoethyl)dihydrofuran-2(3H)-one 9 or the (3R,5R)-enantiomer (Figure 5c). It is for the first time being isolated as a natural product from the leaves of A. saligna. Nothing in the literature indicates that compound 9 is a known natural product. The absolute configuration of 9 at C3 and C5 can be further confirmed by an X-ray crystallographic study or NMR analysis of Mosher diastereomeric esters of compound 9.   Fractions LF-MeOH-A2, -B1, -B3, -C1, C2a, and -C2c (Supplementary Figure S7) were not further purified as they are insufficient in quantity and purity.  Fractions LF-MeOH-A2, -B1, -B3, -C1, C2a, and -C2c (Supplementary Figure S7) were not further purified as they are insufficient in quantity and purity.

Isolation and Structural Identification of Compounds from BK-MeOH
Three main subfractions, BK-MeOH-A1, -B2, and -C2, were isolated (Supplementary Figure S8) from BK-MeOH (300 mg). The NMR data of BK-MeOH-A1 (2.53% w/w) were identified as (−)-epicatechin 6 ( Figure 3). BK-MeOH-B2 (17.83% w/w) is the major component isolated from BK-MeOH, which was assigned to be D-(+)-pinitol 5a. To our surprise, a disaccharide, sucrose, was identified as the component of BK-MeOH-C2 (8.33% w/w) by NMR spectral analysis (Supplementary Table S16). The discovery of D-(+)-pinitol 5a in the flowers and the barks and only (−)-pinitol 5b in the leaves of A. saligna is intriguing. In summary, the presence of these compounds is new from the barks of Australia A. saligna.
IC 50 values reported in the literature obtained from similar DPPH methods; revealed that myricitrin 8 was active with the IC 50 ranging from 2.8 to 165.75 µM [51,54]. While (−)-epicatechin 6 was active with IC 50 in the range of 10.8 to 103.4 µM [55,56]. Similarly, the works by Li et al. [57] and Hong et al. [58] showed that quercitrin 4 has IC 50 values in the range of 4.45 and 107.5 µM. While naringenin 1 was found to have poor activity by Cai et al. [52] with IC 50 of 2 mM. Isosalipurposide 3 was also shown to be active with an IC 50 of 81.9 µM in the DPPH assay [2]. These findings, including ours, reiterate that IC 50 values are considerably variable.
The presence of the three active antioxidants, namely (−)-epicatechin 6 (0.9% w/w), quercitrin 4 (2.86% w/w), myricitrin 8 (5% w/w), logically supports the activity of LF-MeOH observed in both DPPH and ABTS assays. Quercitrin 4 and myricitrin 8 were also found in the leaf extract of Egyptian A. saligna [8]. The potent antioxidant activity of the leaf extract reported by Elansary et al. [4] was extensively exerted by many other flavonoids and polyphenols in the extract, as indicated in their HPLC analysis.
BK-MeOH extract is the most active in both DPPH and ABTS assays ( Table 2). However, only 2.53% w/w of active (−)-epicatechin 6 is present in this extract, in which D-(+)-pinitol 5a (17.83% w/w) is the main component. In this case, the presence of (−)epicatechin 6 may partly explain the high activity exerted by BK-MeOH in both assays. The inconsistent activities of D-(+)-pinitol 5a found between DPPH and ABTS assays would inadequately support the activity exerted by BK-MeOH. GC-MS analysis of BK-MeOH (Supplementary Table S19) revealed the presence of cinnamic acid, lamitol, D-asparagine, and thymidine-5 -monophosphate. However, these compounds do not have the antioxidant activity to support the observed antioxidant activity in BK-MeOH adequately. In comparison, the antioxidant activity of Egyptian crude ethanolic bark extract was reported to be (IC 50 = 10.1 µg/mL) [10]. The potent activity of their bark extract is attributed to the presence of antioxidant compounds such as naringenin 1, kaempferol, rutin, gallic acid, vanillin acid, caffeic acid, ferulic acid, and chlorogenic acid.

α-Glucosidase Inhibition of Isolated Compounds
The yeast α-glucosidase inhibitory activities of isolated compounds were determined using the same procedure for the extracts. The dose-response inhibitory activities of isolated compounds are presented in Supplementary Figure S9 and Table S20. The IC 50 values of isolated compounds are shown in Table 7 In comparison, compound 9 showed no inhibition against the enzyme across the range of test concentrations. It is noteworthy that D-(+)pinitol is a potent inhibitor and is 2-fold more active than its enantiomer (−)-pinitol 5b. It is important to note that acarbose has been reported to exert more inhibitory activity against mammalian α-glucosidase enzyme than the yeast enzyme. Pacillia et al. [59] reported that naringenin 2 displayed an effective inhibition against the yeast enzyme (IC 50 6.51 µM). However, it was poorly active (IC 50 384 µM) when tested on the rat intestinal glucosidase. They also reported that the positive control acarbose inhibited the rat α-glucosidase more effectively than the yeast enzyme. Therefore, further investigation is required to confirm the inhibitory activity of our extracts and active compounds against the mammalian α-glucosidase enzyme.  The inhibitory activity of (−)-epicatechin 6 against α-glucosidase was reported to have IC 50 values in the range of 0.95 µM to 12.3 mM [56,60,61]. For naringenin 3, variable IC 50 values were also observed in the range of 6.51 to 75 µM [59,62]. Furthermore, the literature indicates that the reported IC 50 values of these compounds and other flavonoids are dispersed and variable [63].
The structure-activity relationships (SAR) investigation carried out by Proença et al. [63] suggested that flavonoids with two phenolic groups at the A or B ring and a hydroxy group at C3 possessed the highest α-glucosidase inhibitory activity. He et al. [64] anḑ Söhretoglu et al. [20] further reiterated that the number of phenolic groups on ring B is vital for the activity. Their docking study indicated that the B ring of the flavonoids located deep inside the active side of the enzyme and the presence of the phenolics significantly im-proved interactions via hydrogen bonding. On the other hand, bulky flavonoid glycosides showed poor inhibition due to their inability to access the binding pocket, which explains the poor activity of 2, quercitrin 4, myricitrin 8 and 9. D-(+)-pinitol is a cyclic polyol known to have highly beneficial effects on inflammation and related diseases, such as T2D [65]. To the best of our knowledge, it is for the first time that both enantiomers of pinitol were shown to be inhibitors against the yeast α-glucosidase enzyme.
It is noteworthy that D-(+)-pinitol 5a (17.83% w/w) is the principal component in BK-MeOH and would be the main contributor to the α-glucosidase inhibitory activity observed in the BK-MeOH (IC 50 = 4.37 ± 0.24 µg/mL) in combination from (−)-epicatechin 6 (2.53% w/w). BK-H 2 O (IC 50 = 23.27 ± 3.88 µg/mL) was also active. However, NMR analysis of this fraction revealed that it contains mainly sucrose. Sucrose is a known α-glucosidase substrate [66]; it might outcompete the intended substrate (p-nitrophenylβ-D-glucopyranose, pNPG) of the assay, resulting in the lower concentration of yellowcoloured p-nitrophenol cleaved by the enzyme. Therefore, the observed inhibitory activity of BK-H 2 O is more likely associated with a fault-positive inhibition.

Sample Collection and Identification
The samples, including leaves, flowers, and stem barks, were collected from 12 Tasman

Sequential Extraction of Plant Parts
The method of extraction was adapted from Subhan [23]. Each extraction was carried out without repetition. Dried flower in powder form (250 g) was first soaked in hexane (1 L) with shaking (80 rpm) at RT for 48 hr, followed by vacuum filtration of the mixture. The resulting filtrate was concentrated under reduced pressure at 35 • C to give FL-hex extract (1.71 g). The solid residue was then air-dried at RT for 12 h and then soaked in dichloromethane (1 L) with the same conditions as above. The resulting filtrate was concentrated under reduced pressure at 35 • C to give FL-DCM extract (1.79 g). The process was repeated for methanol and water to provide the FL-MeOH (26.17 g) and FL-H 2 O (36.31 g). Sequential extraction of dried leaves (250 g) was carried out using the same steps to give LF-hex (3.08 g), LF-DCM (4.98 g), LF-MeOH (25.37 g), and LF-H 2 O (13.32 g), while the sequential extraction of dried barks (250 g) provided BK-hex (0.68 g), BK-DCM (2.12 g), BK-MeOH (18.26 g), and BK-H 2 O (4.34 g), as shown in the Supplementary Figure S1.

DPPH Scavenging Assay
The DPPH-free radical scavenging study based on a 96-well plate reading approach was performed following Jiang et al. [68] and Chen et al. [69] with slight modifications. Briefly, a 180 µL of DPPH 0.
A 0 is the absorbance of the blank solution (DPPH 0.2 mM + EtOH), A 1 is the absorbance of the sample (sample + DPPH 0.2 mM), and A 2 is the absorbance of the blank sample (sample in EtOH).
The value was then converted into IC 50 (µg/mL) from a graph correlating the sample concentration (mg/mL) and DPPH scavenging activity (%). The results were expressed as mean ± standard error mean (SEM) of three separate experiments (n = 3). The descriptive statistics are analysed in GraphPad Prism 8 (San Diego, CA, USA).

ABTS •+ Radical Decolourisation Assay
The ABTS •+ solution was prepared by generating a reaction between ABTS 7 mM and potassium persulfate 2.45 mM (1:1 of v/v) at room temperature for 16-18 h [70]. The ABTS •+ solution was further diluted to achieve an acceptable measurement at 734 nm [71]. The same experimental procedure used in the DPPH radical scavenging assay was applied to measure the percentage of ABTS •+ radical scavenging. The absorbance was observed using a microplate reader (Tecan Infinite M1000 PRO, Männedorf, Switzerland) at 734 nm.

In Vitro Assay of Yeast α-Glucosidase Inhibition
The enzyme deactivation assay was carried out following the modified microplates method adapted from Ning et al. [72]. A volume of 20 µL of the plant extract in different concentrations or acarbose solution (31.25 to 1000 µg/mL), isolates or acarbose solution (31.25 to 1000 µM), or solvent control was mixed with α-glucosidase (40 µL, 0.075 U/mL in potassium phosphate buffer solution (100 mM, pH 6.8) in 96-well polystyrene plates (Corning, New York City, NY, USA) and then incubated for 15 min at 37 • C. Afterwards, p-NPG solution in the buffer solution (40 µL, 1 mM) was added to the mixture, followed by further incubation for 30 min at 37 • C. The reaction was terminated by adding Na 2 CO 3 solution (100 µL, 200 mM) to the wells. The spectrophotometric observation was then conducted to determine the absorbance of p-nitrophenol released from the reaction under 405 nm wavelength in a microplate reader (Tecan Infinite M1000 PRO, Männedorf, Switzerland). The percentage of inhibition was calculated from the following formula: where A c is the absorbance of the solvent control and enzymatic reaction system and A s is the absorbance of the sample with the enzymatic reaction system. The inhibitory activity was expressed in the value of half minimal inhibitory concentration (IC 50 ).

Effects of Extracts on the Viability of 3T3-L1 Adipocytes
The differentiated 3T3-L1 cells grown in three 96-well microtiter plates (Corning, New York City, NY, USA) were exposed to 100 μL of fresh test solution containing flower, leaf, and bark extracts in a range of concentration of 25-200 μg/mL and incubated for a further 24, 48, and 72 h. After incubation, the solution was replaced with 100 μL of fresh medium containing 10 % MTT solution (5 mg/mL in PBS). The treated cells were then incubated for an additional 4 h. Once finished the last incubation, the MTT solution was replaced by 100 μL of DMSO to solubilise the formazan crystal products. The absorbance was measured at the wavelength of 570 via a multiwell plate reader (Tecan Infinite M1000 PRO, Männedorf, Switzerland). Each concentration was performed twice times whereby each experiment was conducted in triplicate. The percentage of cell viability is expressed in the: Cell viability (%) = absorbance of sample absorbance of control × 100%

Statistical Method
The results were expressed as mean ± standard error mean (SEM) of three independent experiments (n = 3). The results were analysed using a one-way analysis of variance (ANOVA) with Tukey's or Dunnett's post hoc test using GraphPad Prism 8 (Boston, MA, USA). The difference was considered significant at p < 0.05.

Conclusions
Our approach in using sequential polarity-based extraction of A. saligna parts and bioactivity-guided fractionation has fast-tracked the identification of bioactive compounds in A. saligna. Isolation of pure compounds in the active methanolic extracts was greatly simplified, evidenced by the requirement of one or two successive steps in column chromatography. Through this effort, we have successfully isolated ten compounds of different categories. They are (i) isosalipurposide 3, myricitrin 8, and (−)-epicatechin 6 as known compounds isolated from A. saligna; (ii) naringenin 1 and quercitrin 4 as known compounds to exist in A. saligna; however, being isolated in pure form in this work, (iii) D-(+)-pinitol 5a, (−)-pinitol 5b and 2,4-di-t-butylphenol 7 as known natural products found elsewhere, however, are new to this plant, and (iv) naringenin-7-O-α-L-arabinopyranose 2 and (3S*,5S*)-3-hydroxy-5-(2-aminoethyl) dihydrofuran-2(3H)-one 9 as two novel natural products. The antioxidant and α-glucosidase inhibitory activities of the isolated compounds, especially (−)-epicatechin 6, naringenin 1, and D-(+)-pinitol 5b, quercitrin 4, and myricitrin 8, support the activities observed in our methanolic extracts and the reported activities of crude extracts of A. saligna. The presence of 2,4-di-t-butylphenol 7 may also help to explain the reported allelopathic and antifungal activities of A. saligna. The outcome of this study indicates the potential of A. saligna as a rich source of bioactive compounds for drug discovery targeting T2D.

Effects of Extracts on the Viability of 3T3-L1 Adipocytes
The differentiated 3T3-L1 cells grown in three 96-well microtiter plates (Corning, New York City, NY, USA) were exposed to 100 µL of fresh test solution containing flower, leaf, and bark extracts in a range of concentration of 25-200 µg/mL and incubated for a further 24, 48, and 72 h. After incubation, the solution was replaced with 100 µL of fresh medium containing 10 % MTT solution (5 mg/mL in PBS). The treated cells were then incubated for an additional 4 h. Once finished the last incubation, the MTT solution was replaced by 100 µL of DMSO to solubilise the formazan crystal products. The absorbance was measured at the wavelength of 570 via a multiwell plate reader (Tecan Infinite M1000 PRO, Männedorf, Switzerland). Each concentration was performed twice times whereby each experiment was conducted in triplicate. The percentage of cell viability is expressed in the: Cell viability (%) = absorbance of sample absorbance of control × 100%

Statistical Method
The results were expressed as mean ± standard error mean (SEM) of three independent experiments (n = 3). The results were analysed using a one-way analysis of variance (ANOVA) with Tukey's or Dunnett's post hoc test using GraphPad Prism 8 (Boston, MA, USA). The difference was considered significant at p < 0.05.

Conclusions
Our approach in using sequential polarity-based extraction of A. saligna parts and bioactivity-guided fractionation has fast-tracked the identification of bioactive compounds in A. saligna. Isolation of pure compounds in the active methanolic extracts was greatly simplified, evidenced by the requirement of one or two successive steps in column chromatography. Through this effort, we have successfully isolated ten compounds of different categories. They are (i) isosalipurposide 3, myricitrin 8, and (−)-epicatechin 6 as known compounds isolated from A. saligna; (ii) naringenin 1 and quercitrin 4 as known compounds to exist in A. saligna; however, being isolated in pure form in this work, (iii) D-(+)-pinitol 5a, (−)-pinitol 5b and 2,4-di-t-butylphenol 7 as known natural products found elsewhere, however, are new to this plant, and (iv) naringenin-7-O-α-L-arabinopyranose 2 and (3S*,5S*)-3-hydroxy-5-(2-aminoethyl) dihydrofuran-2(3H)-one 9 as two novel natural products. The antioxidant and α-glucosidase inhibitory activities of the isolated compounds, especially (−)-epicatechin 6, naringenin 1, and D-(+)-pinitol 5b, quercitrin 4, and myricitrin 8, support the activities observed in our methanolic extracts and the reported activities of crude extracts of A. saligna. The presence of 2,4-di-t-butylphenol 7 may also help to explain the reported allelopathic and antifungal activities of A. saligna. The outcome of this study indicates the potential of A. saligna as a rich source of bioactive compounds for drug discovery targeting T2D.