Potential α-Glucosidase Inhibitors from the Deep-Sea Sediment-Derived Fungus Aspergillus insulicola

Three new phenolic compounds, epicocconigrones C–D (1–2) and flavimycin C (3), together with six known phenolic compounds: epicocconigrone A (4); 2-(10-formyl-11,13-dihydroxy-12-methoxy-14-methyl)-6,7-dihydroxy-5-methyl-4-benzofurancarboxaldehyde (5); epicoccolide B (6); eleganketal A (7); 1,3-dihydro-5-methoxy-7-methylisobenzofuran (8); and 2,3,4-trihydroxy-6-(hydroxymethyl)-5-methylbenzyl-alcohol (9), were isolated from fermentation cultures of a deep-sea sediment-derived fungus, Aspergillus insulicola. Their planar structures were elucidated based on the 1D and 2D NMR spectra and HRESIMS data. The absolute configurations of compounds 1–3 were determined by ECD calculations. Compound 3 represented a rare fully symmetrical isobenzofuran dimer. All compounds were evaluated for their α-glucosidase inhibitory activity, and compounds 1, 4–7, and 9 exhibited more potent α-glucosidase inhibitory effect with IC50 values ranging from 17.04 to 292.47 μM than positive control acarbose with IC50 value of 822.97 μM, indicating that these phenolic compounds could be promising lead compounds of new hypoglycemic drugs.


Introduction
According to the International Diabetes Federation, 537 million people worldwide were diagnosed with diabetes mellitus in 2021, and about 90 percent of them were type 2 diabetes mellitus (T2DM) [1,2]. T2DM is a chronic metabolic disease that is characterized by postprandial hyperglycemia in the case of insulin resistance and relative lack of insulin [3]. The inhibition of α-glucosidase can reduce the cleavage of glucose from disaccharides or oligosaccharides to inhibit postprandial hyperglycemia [4]. Therefore, α-glucosidase is a common therapeutic target for the treatment of T2DM [5]. Currently available α-glucosidase inhibitors, such as acarbose, voglibose and miglitol, have been used to treat T2DM patients. Nevertheless, the use of these drugs has been associated with serious side effects, such as abdominal distension and diarrhea [6,7]. For this reason, the search for natural, efficient and non-toxic α-glucosidase inhibitors provides an attractive strategy for the development of new hypoglycemic drugs.
The 13 C NMR (  , as well as the HMBC correlations from 7-OH (δH 11.33) to C-6, C-7 (δC 153.6), and C-8 (δC 104.5) established the substitution of the aromatic ring D. Furthermore, the HMBC correlations from H-2 (δH 6.83) to C-10 (δC 68.6) and C-16, from H-10 (δH 6.38) to C-2 (δC 89.8), C-11, C-12 and C-16 suggested the presence of two oxygen bridges between C-16/C-2 and C-2/C-10 in ring B, which could be confirmed by the low field chemical shift signal of CH-2 (δC 89.8, δH 6.83). Ring C was established by the HMBC correlations from H-2 to C-4 and C-8, and from H-10 to C-8 and C-9 (δC 196.9). The comprehensive NMR analysis indicated that 1 shared the same oxygen-bridged skeleton with epicocconigrone A (4) [24], with the exception that the appearance of 6-OCH3 in 1 replaced 6-OH in 4, which was supported by the HMBC correlation from 6-OCH3 (δH 3.70) to C-6 (δC 134.7). Thus, the planar structure of 1 was elucidated as shown ( Figure 1), named epicocconigrone C. In the nuclear Overhauser effect spectroscopy (NOESY) spectrum of 1, the correlation between H-2 and H-10 was indicative of their cis relationship ( Figure 2). The absolute configuration of 1 was confirmed by the ECD calculation. Its experimental ECD curve for the absolute configurations of 2S and 10R was consistent with the calculated ECD curve of (2S, 10R) ( Figure 3).   Epicocconigrone D (2) was obtained as a yellow solid. The molecular formula of 2 was determined as C20H20O9 with 11 unsaturated degrees by HRESIMS data at m/z 427.1004 (calcd. 427.1000 for C20H20O9Na, [M + Na] + ), which was supported by the 13 C NMR and DEPT spectral data. The IR spectrum of 2 featured typical absorption bands for hydroxyl (3446 cm −1 ) and conjugated ketone (1626 cm −1 ). The 1 H NMR spectrum (Table 1) of 2 indicated two methyl groups (δH 2.09 and δH 2.18), two methoxy groups (δH 3.57 and δH 3.69), one methylene (δH 4.32, d, J = 12.1 Hz; 4.81 d, J = 12.1 Hz), two oxymethines (δH 5.65 and δH 6.76), and two hydroxyl protons (δH 8.92 and δH 11.46). The 13 C NMR (Table 2) and DEPT spectra revealed 20 carbon atom signals, including one ketone carbonyl (δC 196.8), two oxygenated tertiary carbons (δC 89.9 and δC 70.3), two methoxy carbons (δC 60.3 and δC 60.0), one methylene (δC 55.8), two methyls (δC 11.0 and δC 10.3), and twelve olefinic quaternary carbons. Detailed analysis of 2D NMR spectra of 2 revealed that it had a similar structure to 1. The major differences in 2 were a hydroxymethylene group and a methoxy group substituted at C-12 and C-15, instead of the aldehyde group and the hydroxyl group, respectively, when compared to 1 (Figure 2), which were further confirmed by the HMBC correlations from H2-18 (δH 4.32, 4.81) to C-11 (δC 108.3), C-12 (δC 132.1), and C-13 (δC 118.1), and from 15-OCH3 (δH 3.69) to C-15 (δC 134.6). Thus, the planar structure of 2 was elucidated as shown (Figure 1), named epicocconigrone D. The ROESY correlation between H-2 and H-10 indicated their cis orientation ( Figure 2). The absolute configuration of 2 was understood to be 2S, 10R by comparing the experimental and simulated ECD curves ( Figure 3).    Table 2) and DEPT spectra revealed 20 carbon atom signals, including one ketone carbonyl (δ C 196.8), two oxygenated tertiary carbons (δ C 89.9 and δ C 70.3), two methoxy carbons (δ C 60.3 and δ C 60.0), one methylene (δ C 55.8), two methyls (δ C 11.0 and δ C 10.3), and twelve olefinic quaternary carbons. Detailed analysis of 2D NMR spectra of 2 revealed that it had a similar structure to 1. The major differences in 2 were a hydroxymethylene group and a methoxy group substituted at C-12 and C-15, instead of the aldehyde group and the hydroxyl group, respectively, when compared to 1 (Figure 2), which were further confirmed by the HMBC correlations from H 2 -18 (δ H 4.32, 4.81) to C-11 (δ C 108.3), C-12 (δ C 132.1), and C-13 (δ C 118.1), and from 15-OCH 3 (δ H 3.69) to C-15 (δ C 134.6). Thus, the planar structure of 2 was elucidated as shown (Figure 1), named epicocconigrone D. The ROESY correlation between H-2 and H-10 indicated their cis orientation (Figure 2). The absolute configuration of 2 was understood to be 2S, 10R by comparing the experimental and simulated ECD curves ( Figure 3). Flavimycin C (3) (Table 2) and DEPT spectra displayed 10 well-resolved carbon atom signals, dividing into six quaternary carbons that were assigned to one benzene ring, one methylene (δ C 65.8), one methine (δ C 66.1), one methoxy carbon (δ C 60.2), and one methyl (δ C 9.5). The NMR data of 3 were very similar to those of 8 except for the absence of the methylene signal, and instead, the presence of the methine signal of C-3 (δ H 4.30/δ C 66.1) in 3. Combined with molecular formula, 3 was deduced to be a symmetrical dimeric derivative. The above data suggested 3 was a symmetrical dimer of 8, connecting at C-3/C-10 between the two units ( Figure 2), which was further confirmed by the HMBC correlation from H-3 to C-10. Thus, the planar structure of 3 was confirmed as shown (Figure 1), and named flavimycin C. The 1 H and 13 C NMR spectra (Tables 1 and 2) of this aromatic polyketide dimer only exhibited a set of signals of aromatic polyketide monomer. There were three possible absolute configurations of two chiral carbons C-3 and C-10 in 3. The obvious negative optical activity ([α] 20 D = −70.0) and the Cotton effect indicated that compound 3 was not a mesomer, which implied the possibility of 3R, 10S was excluded. Consequently, the absolute configurations of C-3 and C-10 were the same Mar. Drugs 2023, 21, 157 6 of 11 (3S, 10S or 3R, 10R). The absolute configuration of 3 was understood to be 3R, 10R by comparing the experimental and simulated ECD curves (Figure 3).
The new compounds 1-3 are all aromatic polyketide dimers, particularly compounds 1 and 2 simultaneously featuring consistent 6/6/6/6 heterotetracyclic ring cores and compounds 1-3 co-occurrence in the same marine-derived fungus suggest that they should originate from the same biogenetic pathway. A plausible biosynthetic pathway toward the formation of compounds 1-3 can be proposed by detailed analysis of their structures (Scheme 1).
The new compounds 1-3 are all aromatic polyketide dimers, particularly compounds 1 and 2 simultaneously featuring consistent 6/6/6/6 heterotetracyclic ring cores and compounds 1-3 co-occurrence in the same marine-derived fungus suggest that they should originate from the same biogenetic pathway. A plausible biosynthetic pathway toward the formation of compounds 1-3 can be proposed by detailed analysis of their structures (Scheme 1).

In Vitro Evaluation of α-Glucosidase Inhibitory Activity
All compounds were tested for their α-glucosidase inhibitory activities using a reported method [30], with acarbose as the positive control. The results revealed that compounds 1, 4-7, and 9 showed more potent inhibitory activity (IC50 values ranging from 17.04 ± 0.28 to 292.47 ± 5.87 μM) than acarbose (IC50, 822.97 ± 7.10 μM) ( Table 3). The potent α-glucosidase inhibitory activity of epicocconigrone A (4) and epicoccolide B (6) has been already reported [31]. It could be noted herein that the number of hydroxyl groups of polyhydroxy phenolic compounds was important for α-glucosidase inhibitory activity, as reflected by the low IC50 values of compounds 4 and 6, while structures with fewer hydroxyl groups (compounds 1 and 5) exhibited little activity.

In Vitro Evaluation of α-Glucosidase Inhibitory Activity
All compounds were tested for their α-glucosidase inhibitory activities using a reported method [30], with acarbose as the positive control. The results revealed that compounds 1, 4-7, and 9 showed more potent inhibitory activity (IC 50 values ranging from 17.04 ± 0.28 to 292.47 ± 5.87 µM) than acarbose (IC 50 , 822.97 ± 7.10 µM) ( Table 3). The potent α-glucosidase inhibitory activity of epicocconigrone A (4) and epicoccolide B (6) has been already reported [31]. It could be noted herein that the number of hydroxyl groups of polyhydroxy phenolic compounds was important for α-glucosidase inhibitory activity, as reflected by the low IC 50 values of compounds 4 and 6, while structures with fewer hydroxyl groups (compounds 1 and 5) exhibited little activity.

Fungal Material and Fermentation
The fungal strain A. insulicola was isolated from deep-sea sediments, which were collected from the South China Sea at the depth of 2500 m. After grinding, the sample (1.0 g) was diluted to 10 −2 g/mL with sterile H 2 O, 100 µL of which was spread on potato dextrose agar medium (200.0 g potato, 20.0 g glucose, and 20.0 g agar per liter of seawater) plates containing chloramphenicol as a bacterial inhibitor. It was identified by its morphological characteristics and ITS gene sequences (GenBank accessing No. ON413861), the used primers of which were ITS1 (TCCGTAGGTGAACCTGCGG) and ITS4 (TCCTC-CGCTTATTGATATGC). A reference culture of A. insulicola was deposited at the Hainan Provincial Key Laboratory for Functional Components Research and Utilization of Marine Bio-resources, Haikou, China.

Culture Conditions
The fungal strain A. insulicola was cultured in potato dextrose broth medium (consisting of 200.0 g/L potato, 20.0 g/L glucose, and 1000.0 mL deionized water), and incubated on a rotary shaker (150 rpm) for 72 h at 28 • C. Thereafter, 3 mL of seed broth was transferred to fifty 1000 mL Erlenmeyer flasks containing solid rice medium (each flask contained 80 g rice and 120 mL seawater), used for fermentation. The flasks were incubated under static conditions at room temperature for 30 days.

ECD Calculation
The conformers of compounds were generated using the Confab [32] program ebbed in the Openbabel 3.1.1 software, and further optimized with xtb at GFN2 level [33]. The conformers with population over 1% were subjected to geometry optimization using the Gaussian 16 package [34] at B3LYP/6-31G(d) level and proceeded to calculation of excitation energies, oscillator strength, and rotatory strength at B3LYP/TZVP level in the polarizable continuum model (PCM, methanol). The ECD spectra were Boltzmann-weighted and generated using SpecDis 1.71 software [35].

α-Glucosidase Inhibitory Activity
All the assays were carried out under 0.1 M sodium phosphate buffer (PH = 6.8). The samples were dissolved with DMSO and diluted into a series of gradient concentrations (final concentrations of 6.25, 12.5, 25, 50, 100, 200, 400, and 800 µM). The 10 µL sample was mixed with 100 µL α-glucosidase solution (0.2 U/mL, Sigma) and shaken well, then added to a 96-well plate and placed at 37 • C for 15 min. Subsequently, 40 µL of 2.5 mM 4-nitrophenyl-α-D-glucopyranoside was added and further incubated at 37 • C for 15 min. Finally, the OD value of each well was detected at 405 nm wavelength of microplate reader. Acarbose was used as a positive control. The control was prepared by adding DMSO instead of the sample in the same way as the test. The blank was prepared by adding sodium phosphate buffer instead of 4-nitrophenyl-α-D-glucopyranoside using the same method. The percentage inhibition was calculated using the following equation: % inhibition= [(OD control − OD sample )/(OD control − OD blank )] × 100

Conclusions
In summary, two new tetracyclic cores of integrastatins, named epicocconigrones C-D (1-2), one new dimeric isobenzofuran, named flavimycin C (3), and six known compounds (4)(5)(6)(7)(8)(9) were isolated from fermentation cultures of the deep-sea sediment-derived fungus A. insulicola. The biological evaluation revealed compounds 1, 4-7, 9 exhibited significant α-glucosidase inhibitory with IC 50 values ranging from 17.04 ± 0.28 to 292.47 ± 5.87 µM, among which compound 6 was the most potent α-glucosidase inhibitor, with an IC 50 value 48-fold stronger than positive control acarbose. Comparing the structure of compounds 1, 4, 5 and 6 revealed the α-glucosidase inhibitory activity was greatly enhanced after the hydroxyl group replaced the methoxy group, which further confirmed that polyhydroxy phenolic compounds were efficient α-glucosidase inhibitors, and provided a reference value for the synthesis of novel α-glucosidase inhibitors. In conclusion, the study has enriched the structural diversity of phenolic compounds and provided a promising lead toward the development of novel α-glucosidase inhibitors.

Data Availability Statement:
The authors declare that all relevant data supporting the results of this study are available within the article and its Supplementary Materials file, or from the corresponding authors upon request.