Brasilterpenes A–E, Bergamotane Sesquiterpenoid Derivatives with Hypoglycemic Activity from the Deep Sea-Derived Fungus Paraconiothyrium brasiliense HDN15-135

Five bergamotane sesquiterpenoid derivatives, brasilterpenes A–E (1–5), bearing an unreported spiral 6/4/5 tricyclic ring system, were isolated from the deep sea-derived ascomycete fungus Paraconiothyrium brasiliense HDN15-135. Their structures, including absolute configurations, were established by extensive spectroscopic methods complemented by single-crystal X-ray diffraction analyses, electronic circular dichroism (ECD), and density-functional theory (DFT) calculations of nuclear magnetic resonance (NMR) data including DP4+ analysis. The hypoglycemic activity of these compounds was assessed using a diabetic zebrafish model. Brasilterpenes A (1) and C (3) significantly reduced free blood glucose in hyperglycemic zebrafish in vivo by improving insulin sensitivity and suppressing gluconeogenesis. Moreover, the hypoglycemic activity of compound 3 was comparable to the positive control, anti-diabetes drug rosiglitazone. These results suggested brasilterpene C (3) had promising anti-diabetes potential.


Introduction
Diabetes is a chronic metabolic condition marked by high blood glucose (or blood sugar) levels, which can cause catastrophic damage to the heart, blood vessels, eyes, kidneys, and nerves over time [1]. Over the last few decades, the prevalence of diabetes has progressively increased, and a global consensus has been made to halt the rise in diabetes and obesity [2]. Hence, new therapeutic techniques and medications are urgently required. Marine-derived sesquiterpenoids have attracted much attention due to structural diversity and biological activities [3][4][5][6]. According to studies, they have hypoglycemic activity and can trigger pancreatic β-cell regeneration [7,8], suggesting that sesquiterpenoids could be the leading compounds for anti-diabetes drug development.
Brasilterpene A (1), was isolated as colorless crystals. Its molecular formula was determined as C 16  , and one carbonyl carbon (δ C 168.6). These groups accounted for three out of the six degrees of unsaturation, requiring three additional rings in 1.
The relative configurations of 1 and 2 were determined by key NOESY correlations as shown in Figure 3. The NOESY correlation between H-9 and H 3 -13 in 1 and 2 indicated the E geometry of 10 double bond. The NOESY cross-peaks of H-14/H-1/H-10 in 1 indicated they were coplanar. Due to the rigidity of bicyclo[3.1.1]heptane ring system and the NOESY correlation of H-14/H-4a, the relative of compound 1 was suggested to be 1S*,5R*,6S*,9S*, and 14S*. However, there is no reliable evidence to determine the relative configuration of C-3. To distinguish the two possible epimers 1S,3R,5R,6S,9S,14S-1 (1a) and 1S,3S,5R,6S,9S,14S-1 (1b), the computations of 1 H and 13 C NMR chemical shifts for 1a and 1b were performed (Tables S7 and S8) [31]. The DP4+ probability analysis indi-Mar. Drugs 2022, 20, 338 4 of 14 cated that 1b appeared consistent with the NMR experimental data with a probability of 100% (Table 2) [32]. The absolute configuration of 1 was further confirmed by ECD calculations [23] ( Figure 4A) and the X-ray diffraction analysis ( Figure 5) with a Flack parameter of -0.05 (8). In compound 2, the NOESY correlations of H-10/H 3 -16, H-5/H-9, and H-3/H-14/H-4a (δ H 2.34) suggested that 2 was probably the C-14 epimer of 1 ( Figure 3). Finally, the absolute configuration of 2 was determined by the good match of the experimental and calculated ECD curves ( Figure 4A). absolute configuration of 2 was determined by the good match of the experimental and calculated ECD curves ( Figure 4A).            Figure 3). Then, the absolute configuration was determined to be 1S,5S,6S,9S, and 14R by comparing its experimental and calculated ECD spectra ( Figure 4C). As a class of unusual bergamotane-sesquiterpenoid derivates, brasilterpenes A-E (1-5), were proposed to be biosynthesized by the farnesyl diphosphate (FPP) synthesis pathway (Scheme 1). The bergamotane-sesquiterpenoid skeleton was cyclized from FPP via nerolidyl diphosphate (NPP) followed by a bisabolyl cation intermediate [33]. Then compounds 1-5 were generated by further oxidation, nucleophilic attack of OH-9, and methylation. Due to the flexibility of nucleophilic attack direction during the formation of the furan ring, compounds 1-4 appeared in pairs as C-14 epimers. To screen these compounds for hypoglycemic activity, the β-cell-specific ablation transgenic zebrafish, Tg(-1.2ins:htBid TE−ON ; LR) [34], was applied. In this transgenic line, the proapoptotic protein human truncated Bid (htBid) is target expressed in the β cells under the control of the tetracycline and ecdysone-inducible system ( Figure 6A) [35]. After the zebrafish is incubated with doxycycline (Dox) and tebufenozide (Tbf), the β cells of the zebrafish can be induced to express the htBid protein, which results in the ablation of β cells and increased free blood glucose ( Figure 6B, Teton− vs. Teton+). This model was then used to assess all isolated compounds (1-5). After the incubation of β-cell-ablated To screen these compounds for hypoglycemic activity, the β-cell-specific ablation transgenic zebrafish, Tg(-1.2ins:htBid TE−ON ; LR) [34], was applied. In this transgenic line, the proapoptotic protein human truncated Bid (htBid) is target expressed in the β cells under the control of the tetracycline and ecdysone-inducible system ( Figure 6A) [35]. After the zebrafish is incubated with doxycycline (Dox) and tebufenozide (Tbf), the β cells of the zebrafish can be induced to express the htBid protein, which results in the ablation of β cells and increased free blood glucose ( Figure 6B, Teton− vs. Teton+). This model was then used to assess all isolated compounds (1)(2)(3)(4)(5). After the incubation of β-cell-ablated zebrafish larvae with 10 µM of these five compounds for 24 h, the total glucose level from each group was measured. As shown in Figure 6B, the free glucose level of the β-cell-ablated group (Teton+) was 502.8 ± 12.2 pmol/larva, while compounds 1 and 3 significantly reduced the glucose levels to 449.3 ± 15.1 and 420.4 ± 2.3 pmol/larva, respectively. The hypoglycemic activity of compound 3 was comparable to the positive control, anti-diabetes drug rosiglitazone (RGZ, with a glucose level of 395.6 ± 6.6 pmol/larva). Moreover, compounds 1 and 3 did not show any toxic effect to the zebrafish larvae up to 200 µM ( Figure 6C). The hypoglycemic activity of compounds 1 and 3 may be related to the S configuration at C-14, which was the unique structural difference compared to 2 and 4. Compound 3 was more active than 1, which indicated that the presence of the hydroxyl group at C-3 may reduce the blood glucose level. Similarly, compound 4 was more active than 2, which illustrated the above reduction. Furthermore, compound 5 was more active than 4, which suggests that the endocyclic △ 2 double bond may increase hypoglycemic activity.
The hypoglycemic effects may be due to increased β-cell mass or increased glucose uptake. To survey whether compounds 1 and 3 induce β-cell regeneration in Tg(-1.2ins:htBid TE−ON ; LR), the Tg(-1.2ins:htBid TE−ON ; LR) was crossed with Tg(-1.2ins:H2BmCherry) whose β cells were labeled with mCherry fluorescence. Without the induction, the β-cell number of the double transgenic Tg(-1.2ins:htBid TE−ON ; LR); Tg(-1.2ins:H2BmCherry) larvae was 29.0 ± 1.6 ( Figure 7, Teton−). While the β-cell number significantly decreased to 8.8 ± 1.6 after induction (Figure 7, Teton+). However, treated with The hypoglycemic activity of compounds 1 and 3 may be related to the S configuration at C-14, which was the unique structural difference compared to 2 and 4. Compound 3 was more active than 1, which indicated that the presence of the hydroxyl group at C-3 may reduce the blood glucose level. Similarly, compound 4 was more active than 2, which illustrated the above reduction. Furthermore, compound 5 was more active than 4, which suggests that the endocyclic 2 double bond may increase hypoglycemic activity.
The hypoglycemic effects may be due to increased β-cell mass or increased glucose uptake. To survey whether compounds 1 and 3 induce β-cell regeneration in Tg(-1.2ins:htBid TE−ON ; LR), the Tg(-1.2ins:htBid TE−ON ; LR) was crossed with Tg(-1.2ins:H2BmCherry) whose β cells were labeled with mCherry fluorescence. Without the induction, the β-cell number of the double transgenic Tg(-1.2ins:htBid TE−ON ; LR); Tg(-1.2ins:H2BmCherry) larvae was 29.0 ± 1.6 ( Figure 7, Teton−). While the β-cell number significantly decreased to 8.8 ± 1.6 after induction (Figure 7, Teton+). However, treated with 10 µM, compounds 1-5 did not change the β-cell number in each group (Figure 7), which suggested that the hypoglycemic activity of compounds 1 and 3 was not due to β-cell regeneration. The hypoglycemic mechanism of compounds 1 and 3 was then investigated. The insulin signaling sensitivity, which is frequently indicated by the phosphorylation of Akt, was first determined. The Akt phosphorylation of 10 μM of compounds treated β-cellablated Tg(-1.2ins:htBid TE−ON ; LR) was detected by Western blot. As shown in Figure 8A, the levels of Akt phosphorylation in 1 and 3 were increased, which suggested increased insulin sensitivity. The genes involved in glucose homeostasis were also assessed, including insa (insulin a), gcga (glucagon a), pck1 (phosphoenolpyruvate carboxykinase 1), and g6pc1a.2 (glucose-6-phosphatase catalytic subunit 1a, tandem duplicate 2). Ablation of β cells significantly decreases insa expression ( Figure 8B, Teton+ vs. Teton−1). However, compounds 1 and 3 treatment did not change the insa levels ( Figure 8B). For the gluconeogenic genes, gcga, pck1, and g6pc1a.2, compound 1 treated larvae significantly decreased gcga, and 3 have a similar trend but without significance ( Figure 8C). Compound 3 significantly down-regulated the pck1 expression, and 1 has a similar trend with the p value of 0.076 ( Figure 8D). Compound 3 treatment reduced the g6pc1a.2 level, albeit there is no significance ( Figure 8E). These data suggested that compounds 1 and 3 suppressed gluconeogenesis. The hypoglycemic mechanism of compounds 1 and 3 was then investigated. The insulin signaling sensitivity, which is frequently indicated by the phosphorylation of Akt, was first determined. The Akt phosphorylation of 10 µM of compounds treated β-cell-ablated Tg(-1.2ins:htBid TE−ON ; LR) was detected by Western blot. As shown in Figure 8A, the levels of Akt phosphorylation in 1 and 3 were increased, which suggested increased insulin sensitivity. The genes involved in glucose homeostasis were also assessed, including insa (insulin a), gcga (glucagon a), pck1 (phosphoenolpyruvate carboxykinase 1), and g6pc1a.2 (glucose-6-phosphatase catalytic subunit 1a, tandem duplicate 2). Ablation of β cells significantly decreases insa expression ( Figure 8B, Teton+ vs. Teton−1). However, compounds 1 and 3 treatment did not change the insa levels ( Figure 8B). For the gluconeogenic genes, gcga, pck1, and g6pc1a.2, compound 1 treated larvae significantly decreased gcga, and 3 have a similar trend but without significance ( Figure 8C). Compound 3 significantly downregulated the pck1 expression, and 1 has a similar trend with the p value of 0.076 ( Figure 8D). Compound 3 treatment reduced the g6pc1a.2 level, albeit there is no significance ( Figure 8E). These data suggested that compounds 1 and 3 suppressed gluconeogenesis. Rosiglitazone (RGZ) was used as a control compound. All the values shown are means ± SEM from three independent experiments. * p < 0.05, ** p < 0.01, and *** p < 0.001 by one-way ANOVA.

Fungal Material
The fungus was isolated from deep-sea sediment from the Indian Ocean (depth 3824 m, 15.66 • S, 88.00 • E, collected in May 2014). The strain was deposited at the Key Laboratory of Marine Drugs, the Ministry of Education of China, School of Medicine and Pharmacy, Ocean University of China, Qingdao, People's Republic of China. The fungal strain was identified as P. brasiliense according to its morphological characteristics and by comparison of the ITS sequence amplification. The ITS sequence was deposited at GenBank with accession number ON025790.

Fermentation, Extraction, and Isolation
Erlenmeyer flasks (1000 mL) containing 300 mL of 2 # media were directly inoculated with spores of fungus under static condition at room temperature for 30 days. The 2 # media was composed of mannitol (2.0%), yeast extract (0.3%), monosodium glutamate (1.0%), glucose (1.0%), maltose (2.0%), corn steep liquor (0.1%), KH 2 PO 4 (0.05%), MgSO 4 ·7H 2 O (0.03%) in natural seawater (collected from Jiao Zhou Bay, Qingdao, China). The fermentation broth (40 L) was filtered through muslin cloth to separate the supernatant from the mycelia. The mycelia were macerated and extracted with MeOH, then were concentrated to afford an aqueous solution. The fermentation broth combined with the aqueous solution was extracted three times with ethyl acetate and concentrated to give the crude extract (38 g). The

X-ray Crystallographic Analysis of Brasilterpene A (1)
A crystal of compound 1 suitable for X-ray diffraction was obtained by slow evaporation of a solution in MeOH. Single-crystal X-ray diffraction data were measured with an Agilent Gemini Ultra diffractometer with Cu Kα radiation (λ = 1.541 84 Å). The structure was solved by direct methods (SHELXS-97) and refined using full-matrix least-squares difference

Computational Section
Monte Carlo conformational searches were carried out with Spartan'14 software using the Merck Molecular Force Field (MMFF) [36]. Two diastereomers of compound 1 were subjected to conformational analysis and geometrical optimization using B3LYP/6-31+G(d) in DMSO (PCM) in Gaussian 09 [37]. NMR chemical shift calculation was performed with the GIAO method using B3LYP/6-311G+(d,p) in DMSO (PCM). All chemical shifts were Boltzmann averaged, and the shielding constant was used for DP4+ probability calculation based on the method [32]. As for the calculation of ECD, the conformers of compounds 1-4 were initially optimized at the B3LYP/6-31+G(d) in MeOH (PCM), and compound 5 was optimized at the B3LYP/TZVP in MeOH (PCM). The theoretical calculation of the ECD spectra was conducted in MeOH using density functional theory (DFT) with B3LYP and 6-31+G(d) basis set in Gaussian 09 for all conformers of compounds 1-5. ECD spectra were generated using the program SpecDis [38] by applying a Gaussian band shape with a 0.30 eV width for 1, 3, and 4, a 0.40 eV width for 2, and a 0.65 eV width for 5. The calculated spectra were shifted by −6 nm for 1, −3 nm for 2, −14 nm for 3, and −30 nm for 5 to facilitate comparison to the experimental data.
After 48 h of drug-induced β-cell ablation, larvae were rinsed with 0.3X Danieau solution to remove the inducer drugs. Then the zebrafish larvae were cultured in 12-well plates incubated with each compound at a final concentration of 10 µM, respectively. For the Teton− (without β-cell ablation) and Teton+ (β-cell ablation but did not add compounds) group, the DMSO was added as a control.
Free glucose was determined by a glucose assay kit (A22189, Amplex Red Glucose/Glucose Oxidase Assay Kit) similar to Li et al. [39]. After treatment of the compounds for 24 h, a pool of 10 larvae was homogenized in 100 µL of sample buffer. The homogenate was spun at 10,000 rpm for 10 min. Then the supernatant was determined according to the manufacturer's instructions. Each sample was measured for three pools.

Evaluation of Toxicological Effect in Zebrafish Larvae
Five dpf zebrafish larvae were used to evaluate the potential toxicity exerted by compounds 1 and 3 in vivo. Twenty zebrafish larvae were selected and equally placed in 12-well plates and incubated with compounds 1 and 3 from five dpf to six dpf. Each compound with final concentration of 1, 5, 10, 50, and 200 µM was added to the wells, and the DMSO was added separately as the control well. After 24 h of incubation, the mortality and morphological abnormality were observed using a Leica stereomicroscope.

Statistical Analysis
Statistical analysis was performed using GraphPad PRISM 8 software. Results are presented as mean values ± SEM. No outlying values were excluded from the datasets used for statistical analysis. The statistics were performed using one-way ANOVA followed by Bonferroni post hoc test or t-test. p < 0.05 was considered significant.

Conclusions
In summary, five unusual bergamotane sesquiterpenoid derivatives were isolated from the marine-derived fungus P. brasiliense HDN15-135. Compounds 1-5 contained unreported skeleton with a 6/4/5 tricyclic ring system, which was reported for the first time. Compounds 1 and 3 significantly reduced the glucose levels in the hyperglycemic zebrafish model, and the hypoglycemic mechanism of compounds 1 and 3 was proposed to improve insulin sensitivity and suppress gluconeogenesis rather than induce pancreatic β cell regeneration. The study of structure-activity relationship showed that the S configuration at C-14, the presence of endocyclic double bond ( 2 ), and the absence of hydroxyl at C-3 may contribute to the improvement of hypoglycemic activity. Therefore, this study may provide a valuable structural template for the development of hypoglycemic drugs.