Novel Sesquiterpene and Diterpene Aminoglycosides from the Deep-Sea-Sediment Fungus Trichoderma sp. SCSIOW21

Six new sesquiterpene aminoglycosides, trichaspside F (2) and cyclonerosides A–E (5–9), two new diterpene aminoglycosides, harzianosides A and B (10, 11), and three known sesquiterpenes, trichodermoside (1), cycloneran-3,7,10,11-tetraol (3), and cyclonerodiol (4), have been isolated from the n-butanol extract of Trichoderma sp. SCSIOW21 (Hypocreaceae), a deep-sea-sediment-derived fungus. The structures and relative configurations of the new compounds were determined using spectroscopic techniques and comparisons with those reported in the literature. The absolute configurations of the aglycone part of cyclonerosides A–E (5–9) were tentatively proposed based on optical rotation and biogenic considerations. Cyclonerosides A–E (5–9) represent the first glycosides of cyclonelane-type sesquiterpenes generated from Trichoderma. The NO-production-inhibitory activities were evaluated using macrophage RAW264.7 cells. Among the isolated compounds, trichaspside F (2) and cyclonerosides B–E (6–9) exhibited the strongest NO-production-inhibitory activities with IC50 values of 54.8, 50.7, 57.1, 42.0, and 48.0 µM, respectively, compared to the IC50 value of 30.8 µM for the positive control (quercetin). When tested for anti-fungal activities against several pathogenic fungi, none of the compounds exhibited significant activities at a concentration of 100 µM.


Introduction
Filamentous fungi have larger genomes and more biosynthetic gene clusters (BGCs) compared to bacteria, which leads to greater chemical diversity, making them one of the most important resources for natural product drug discovery [1]. However, with the intense study carried out over the last century, the high rediscovery rate from terrestrial fungi has seriously hindered new drug development from this resource. Research on natural products from marine fungi has received a great amount of research attention in recent years. A total of 346 compounds have been characterized from marine-sediment-derived fungi during 2005-2015 [2,3], in which almost half of the compounds (157) are reported to exhibit anti-microbial and anti-cancer activities. Further, 246 compounds have been reported from marine-sediment-derived fungi during 2016-2020 [4]. Among them, 12 compounds exhibit antiviral activity, 57 compounds have anti-microbial activity, and 62 compounds display cytotoxicity. In addition, 70 compounds have cytoprotective activity, including anti-infection, anti-oxidant, and neuroprotective activities, which demonstrates the diversity of the research being conducted on compound activity. However, anti-microbial and anti-cancer activities are still of significant interest in marine fungi natural products. Approximately 200 more compounds exhibiting anti-microbial and anti-cancer activities were reported in 2019-2022 from marine fungi [1]. Considering the best-selling drug lovastatin as an example, which contains hypolipidemic activity and was first discovered from a Penicillium species fungal strain, more bioassays should be prepared to screen for natural products in marine fungi to discover new active compounds with potential applications.
The number of Trichoderma spp. fungi isolated from nature was limited to dozens just a few decades ago. However, the introduction of molecular evolutionary approaches has led to the rapid expansion of the Trichoderma taxonomy, resulting in the discovery of~50 new species each year. By July 2020, 361 Trichoderma species had been successfully cultivated and their DNA barcoded [5]. Trichoderma species are considered to be treasure troves of natural products. By 2008, 186 compounds had been identified from Trichoderma [6], while 203 compounds were characterized between 2009 and 2020, including terpenoids, cyclopeptides, diketopiperazines, alkaloids, and polyketides [7]. Among filamentous fungi, Trichoderma species are some of the dominant producers of terpenoids. However, glycosylated sesquiterpenes and diterpenes have been rarely reported in the literature [8].
Anti-microbial, anti-microalgae, anti-cancer, and phytotoxic activities have been reported for sesquiterpenes and diterpenes isolated from Trichoderma species [7]. Cyclonerane sesquiterpenes also exhibit nematocidal activity [13]. However, the effects of sugar moieties on the activities of these compounds remain under debate. The sugar moiety in bisabolenetype sesquiterpenes has no effect on their growth-inhibition activities against marine phytoplankton species [11], and glycosylation of trichothecene sesquiterpenes appears to reduce their anti-fungal and anti-microalgae activities [9]. In contrast, the amino sugar moiety in bisabolane sesquiterpenes is indispensable for their activities against several aquatic pathogenic bacteria [14]. Trichodermoside, a bisabolane sesquiterpene glycoside, has been shown to weakly inhibit the growth of human HeLa cells [10].

Results and Discussion
The fungal strain was statically cultivated in rice broth containing 3% sea salt and then extracted with n-BuOH. The extract was subjected to silica gel column chromatography, followed by HPLC on an ODS C18 column to obtain 11 compounds ( Figure 1).
Compound 2 was isolated as a colorless gum and its molecular formula was determined to be C23H41NO8 using HRESIMS. The ESI-MS fragment-ion peak observed at m/z 204 was indicative of the presence of an acetamido-sugar moiety [9,11,14]. The 1 H, 13 C, DEPT, and HSQC NMR spectra show four sp 3 methyl, six sp 3 methylene, nine sp 3 methine, and two sp 2 methine signals, as well as peaks corresponding to one sp 3 quaternary carbon and one ketone carbonyl carbon (Table 1) [22.5,, the NMR data were almost identical to those reported for trichobisabolin X [19], a bisabolane-type sesquiterpene, which suggests the presence of an acetamido-substituted sugar residue [9,11,14]. Unfortunately, most of the proton signals in the sugar region overlap in DMSO-d6 (Table 1), and consequently the coupling constants could not be determined. The 1 H and 13 C NMR data were re-acquired in MeOD-d4. The chemical shifts and proton coupling constants in the sugar region [  (Table 1) [11,14]. NOE correlations observed for H-2'/H-4' and H-3'/H-5' further confirmed the relative configuration of the glucopyranoside. The small anomeric-proton coupling constant (3.0 Hz) suggests that the sugar linkage has an α-configuration. The aglycone of 2 was further analyzed using 1 H-1 H correlation spectroscopy

Results and Discussion
The fungal strain was statically cultivated in rice broth containing 3% sea salt and then extracted with n-BuOH. The extract was subjected to silica gel column chromatography, followed by HPLC on an ODS C18 column to obtain 11 compounds ( Figure 1).
Compound 2 was isolated as a colorless gum and its molecular formula was determined to be C 23 H 41 NO 8 using HRESIMS. The ESI-MS fragment-ion peak observed at m/z 204 was indicative of the presence of an acetamido-sugar moiety [9,11,14]. The 1 H, 13 C, DEPT, and HSQC NMR spectra show four sp 3 methyl, six sp 3 methylene, nine sp 3 methine, and two sp 2 methine signals, as well as peaks corresponding to one sp 3 quaternary carbon and one ketone carbonyl carbon (Table 1) , the NMR data were almost identical to those reported for trichobisabolin X [19], a bisabolane-type sesquiterpene, which suggests the presence of an acetamido-substituted sugar residue [9,11,14]. Unfortunately, most of the proton signals in the sugar region overlap in DMSO-d 6 ( Table 1), and consequently the coupling constants could not be determined. The 1 H and 13 C NMR data were re-acquired in MeOD-  (Table 1) [11,14]. NOE correlations observed for H-2'/H-4' and H-3'/H-5' further confirmed the relative configuration of the glucopyranoside. The small anomericproton coupling constant (3.0 Hz) suggests that the sugar linkage has an α-configuration. The aglycone of 2 was further analyzed using 1 H-1 H correlation spectroscopy (COSY) and heteronuclear multiple bond correlation (HMBC) (Figure 2 connect C-1 to C-10 and the C13 methyl group (Figure 2). The two singlet methyl groups are not correlated in the COSY spectrum. The two methyl groups (δ H 0.98 s and 1.03 s) show HMBC correlations with C-10 at δ C 77.4 and another quaternary carbon at δ C 71.6. One additional hydroxy group at δ H 4.03 also exhibits a HMBC correlation with this quaternary carbon. Therefore, the presence of an isopropanol moiety is proposed to be linked to C-10. The aglycone moiety was determined to be the same as that observed in trichobisabolin X upon further comparison of the NMR spectra (Table 1 and Figure 2) [14]. The sugar residue was determined to be attached to C-14 by the mutual HMBC correlations observed between H-1' and C-14, and H 2 -14 and C-1' (Figure 2). H-3/H-5a and H-5a/H-6 ROESY correlations and the identical NMR data obtained for the aglycone suggest the same relative configuration to that observed in trichobisabolin X (Table 1 and Figure 3). Since three bisabolane acetamido glycosides, trichaspsides C-E, have been previously reported [11,14], we assigned compound 2 as trichaspside F. Compounds 5-11 were isolated as colorless gums. Their ESI-MS spectra exhibit representative base peaks at m/z 204, which imply the presence of an acetamido-sugar moiety in each structure [9,11,14]. The sugar region exhibits 1 H and 13 C NMR data that are almost identical to those reported for trichaspside F (2), which reveals the presence of a 2-acetamido-2-deoxy-α-glucopyranosyl moiety in each of these molecules (Tables 2-4). The small coupling constant observed for each aromatic proton (3.0 Hz) suggests all of the sugar linkages have an α-configuration (Tables 2 and 4). methyl groups are not correlated in the COSY spectrum. The two methyl groups (δH 0.98 s and 1.03 s) show HMBC correlations with C-10 at δC 77.4 and another quaternary carbon at δC 71.6. One additional hydroxy group at δH 4.03 also exhibits a HMBC correlation with this quaternary carbon. Therefore, the presence of an isopropanol moiety is proposed to be linked to C-10. The aglycone moiety was determined to be the same as that observed in trichobisabolin X upon further comparison of the NMR spectra (Table 1 and Figure 2) [14]. The sugar residue was determined to be attached to C-14 by the mutual HMBC correlations observed between H-1' and C-14, and H2-14 and C-1' (Figure 2). H-3/H-5a and H-5a/H-6 ROESY correlations and the identical NMR data obtained for the aglycone suggest the same relative configuration to that observed in trichobisabolin X (Table 1 and Figure 3). Since three bisabolane acetamido glycosides, trichaspsides C-E, have been previously reported [11,14], we assigned compound 2 as trichaspside F.        The molecular formula of compound 5 was determined to be C 23 H 41 NO 7 using HRESIMS. The 1 H-1 H COSY analysis shows that the methine at H-10 is connected to the methylene at H 2 -9, and H 2 -9 is connected with another methylene at H 2 -8. The 1-Me methyl group is connected to the methine at H-2 and extends from H-2 to H-6, H 2 -4, and H 2 -5 ( Figure 2). From the HMBC spectrum, two methyl groups (δ H 1.56 s, 1.63 s) are linked to the quaternary carbon at C-11 (δ C 130.1). 14-Me (δ H 1.00 s) exhibits HMBC correlations with C-6 and C-7. 7-OH (δ H 3.89) shows HMBC correlations with C-7 and C-9. These correlations connect the partial structure between C-6 and C-9 ( Figure 2). Both methyl groups at 1-Me and 13-Me show HMBC correlations with the quaternary carbon at C-3 (δ C 86.3) (Figure 2). Detailed 1 H-1 H COSY and HMBC analyses were then conducted to establish the structure of the aglycone (Figures 2, S10 and S18). With the exception of the sugar region, compound 5 exhibits 1 H and 13 C NMR data that are almost identical to those reported for cyclonerodiol (4), a cyclonerane-type sesquiterpene (Tables 2 and 3) [19]. HMBC correlations between H-1' and C-3 establish that the sugar moiety is linked to C-3 (Figure 2), and the ROESY correlations observed for Me-1/H-6 and Me-13/H-2 reveal that the relative configurations of the five-membered ring are identical to those of cyclonerodiol (4) (Figure 3) [20]. Unfortunately, the configuration at C-7 could not be determined. Thus, Mar. Drugs 2023, 21, 7 8 of 15 compound 5 is identified to be 3-O-cyclonerodiol-2-acetamido-2-deoxy-α-glucopyranoside ( Figure 1). As it is the first glycosylated cyclonerane-type sesquiterpene discovered to date, we have assigned compound 5 as cycloneroside A. Cyclonerodiol (4) was obtained in this study with a negative optical rotation of −18, which is similar to the reported −20 of [21]. Considering that cycloneroside A (5) and cyclonerodiol (4) possess the same sesquiterpene carbon skeleton and presumably the same biogenetic pathway (Figure 4), the absolute configurations of the aglycone part of cycloneroside A (5) were thus assigned as that of cyclonerodiol (4) based on the biogenic consideration as 2S, 3R, 6R, 7R [21][22][23]. The absolute configurations of the sugar moiety were not able to be assigned due to the lack of materials. 18 20.  The structures of the three known sesquiterpenes, trichodermoside (1) [10], cycloneran-3,7,10,11-tetraol (3) [25], and cyclonerodiol (4) [19], were determined by comparison with the data reported in the literature.
Terpenes are biosynthesized from C-5 building blocks composed of isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP). Prenyl transferase assembles IPP and DMAPP into farnesyl diphosphate (FPP), which is the universal precursor for the sesquiterpenes. FPP is cyclized by various terpene cyclases to form different types of sesquiterpenes. The post-tailoring steps further diversify these structures [29,30]. The plausible biosynthesis routes to cyclonerane-type sesquiterpenes in this study are summarized in Figure 4. Cyclonerodiol (4) has been demonstrated to be the direct product of FPP cyclization by terpene cyclase [21]. The reactions observed in the post tailoring steps may include complex oxidation, dehydration, and glycosylation ( Figure 4). Notably, the Nacetylaminosugar usually needs to undergo additional transamination and acetylation reactions prior to the glycosylation reaction [31,32]. The molecular formula of compound 6 was determined to be C 23 H 41 NO 8 using HRES-IMS. Apart from the sugar region and the additional signals observed for an oxygenated methylene group [δ H 3.54 m, 3.15 m; δ C 68.9], the sp 3 methine signal [δ H 2.11 m, H-3; δ C 43.2], and lack of signals corresponding to the methyl group at C-13 and oxygenated quaternary carbon at C-3 (Tables 2 and 3), compound 6 exhibits NMR data that highly resemble those reported for isoepicyclonerodiol oxide [19]. The additional oxygenated methylene and methine groups are located at C-13 and C-3, respectively, based on the 1 H-1 H COSY correlations observed between H-13 and H-3, and H-3, H-2, and H-4, and the HMBC correlations observed between H-13 and C-3 and C-4, and between 1-Me and C-3 ( Figure 2). The 2D NMR spectra were then carefully analyzed, which confirmed the presence of an aglycone, as shown in Figure 2. The sugar moiety is linked at C-13 according to the HMBC correlations observed between H-1' and C-13, and H-13 and C-1' (Figures 2 and S19-S26). Further, the ROESY correlations observed between Me-1/H-6 and Me-13/H-2 indicate that ring A has the same relative configuration as compound 5. The ROESY correlations observed for H-6/H-10 and H-2/Me-14 (Figure 3), and the almost identical NMR data obtained for ring B suggest the same relative configuration as isoepicyclonerodiol oxide [19]. Thus, compound 6 was identified to be a 13-O-isoepicyclonerodiol oxide-2-acetamido-2-deoxy-α-glucopyranoside, which we assigned as cycloneroside B. The absolute configurations of the aglycone part of cycloneroside B (6) were assigned based on the biogenic consideration as 2S, 3S, 6R, 7R, 10R.
The molecular formula of compound 7 was determined to be C 23 H 39 NO 8 using HRESIMS. The IR absorption bands of amide I at 1639 cm −1 (C=O stretching vibration) and amide II at 1555 cm −1 (N-H bending vibration) indicated the presence of a secondary amide group [24]. With the exception of the sugar region, and the lack of the sp 3 (Tables 2 and 3), the 1 H and 13 C NMR data obtained for 7 resemble those of 6 ( Figure S27-S34). The Me-1 proton signal of compound 7 (δ H 1.66) is down-field-shifted when compared to that of 6 [δ H 0.68, d (7.0 Hz)] and appears as a singlet. The methylene signals at position 13 are also low-field shifted (δ H 3.98 and 4.10) relative to those observed for 6 (δ H 3.54, m; δ H 3.15, m) with a geminal coupling observed (J = 12 Hz), which implies the presence of two olefinic carbons at C-2 and C-3 (Table 2). These positions were further validated using the HMBC correlations observed between 1-Me and H-13 and these two olefinic carbons (Figure 2). The sugar moiety is linked at C-13 according to the HMBC correlations observed between H-1' and C-13, and H-13 and C-1' (Figure 2). The ROESY correlation observed between Me-1 and H-13 suggests the presence of a trans double bond. The almost identical NMR data obtained for ring B and the H-6/H-10 ROESY correlation suggest the same relative configuration as that observed for compound 6 ( Table 2 and Figures 3 and S35-S43). Hence, the structure of compound 7 is elucidated ( Figure 1) and assigned as cycloneroside C. The absolute configurations of the aglycone part of cycloneroside C (7) re assigned based on the biogenic consideration as 6R, 7R, 10R.
The molecular formula of compound 8 was determined to be C 23 H 39 NO 8 using HRESIMS. The IR absorption bands at 1653 and 1555 cm −1 indicated the presence of a secondary amide group. With the exception of one additional sp 2 methine [δ H 5.54 (br s); δ C 126.8, CH] and an sp 3 methine [δ H 2.46 m; δ C 41.0, CH], the lack of one sp 2 quaternary carbon at C-2 and one sp 3 methylene group at C-4, which implies that the double bond is positioned between C-3 and C-4, compound 8 exhibits 1 H and 13 C NMR data that resemble those of 7 (Tables 2 and 3 (Figure 3) confirm the position of the new double bond. The HMBC correlations between H-13 and C-3 and C-4, and between 1-Me and C-2 and C-3 also confirm this assignment (Figure 2). The sugar moiety is linked at C-13 according to the HMBC correlations observed between H-1' and C-13, and H-13 and C-1' (Figure 2). The ROESY correlation observed for Me-1/H-6 suggests that these protons are directed toward the same face of the molecule. The other ROESY correlations and NMR data obtained for ring B are almost identical to those observed for compound 6, suggesting they have the same relative configuration (Table 2 and Figure 3). We assigned compound 8 as cycloneroside D. The absolute configurations of the aglycone part of cycloneroside D (8) were assigned based on the biogenic consideration as 2S, 6R, 10R.
The molecular formula of 9 was determined to be C 23 H 39 NO 8 using HRESIMS. The 1 H and 13 C NMR data obtained for the aglycone are very similar to those of 3,7,11trihydroxycycloneran-10-one [25]. Detailed 1 H COSY and HMBC analysis confirms the skeleton (Figures 2 and S44-S51); however, MS reveals an m/z difference of 18, which suggests dehydration between the two hydroxyl groups at C-7 and C-11 (Tables 2 and 3 [25]. The sugar moiety is linked to C-3 according to the HMBC correlation observed between H-1' and C-3 ( Figure 2). The almost identical NMR data obtained for ring A and the ROESY correlations observed for Me-1/H-6 and Me-13/H-2 suggest that compound 9 has the same relative configuration to that of ring A in compound 5 (Table 2 and Figure 3). The ROESY correlation observed at H-2/Me-14 indicates that the two groups are directed toward the same side of the molecule (Figure 3). Thus, the structure of 9 is elucidated ( Figure 1) and assigned as cycloneroside E. The absolute configurations of the aglycone part of cycloneroside E (9) was assigned based on the biogenic consideration as 2S, 3R, 6R, 10R.
Terpenes are biosynthesized from C-5 building blocks composed of isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP). Prenyl transferase assembles IPP and DMAPP into farnesyl diphosphate (FPP), which is the universal precursor for the sesquiterpenes. FPP is cyclized by various terpene cyclases to form different types of sesquiterpenes. The post-tailoring steps further diversify these structures [29,30]. The plausible biosynthesis routes to cyclonerane-type sesquiterpenes in this study are summarized in Figure 4. Cyclonerodiol (4) has been demonstrated to be the direct product of FPP cyclization by terpene cyclase [22]. The reactions observed in the post tailoring steps may include complex oxidation, dehydration, and glycosylation ( Figure 4). Notably, the N-acetylaminosugar usually needs to undergo additional transamination and acetylation reactions prior to the glycosylation reaction [31,32].
The abilities of all of the isolated compounds to inhibit NO production were examined using macrophage RAW 264.7 cells stimulated by LPS [18]. Figure 5 shows that all of the compounds inhibit NO production in a dose-dependent manner. None of them exhibit cytotoxicity at the maximum tested concentration (100 µM) ( Figure 6). A comparison of the data obtained for cyclonerodiol (4) and cycloneroside A (5) suggests that the sugar moiety appears to have little impact on the activity of this type of compound. Trichaspside F (2) and cyclonerosides B-E (6-9) exhibit the strongest abilities to inhibit NO production with IC 50 values of 54.8, 50.7, 57.1, 42.0, and 48.0 µM, respectively. The other compounds in this study show low activity with IC 50 values~100 µM. Our structure-activity relationship analyses reveal that the tetrahydropyran or tetrahydrofuran ring is essential toward improving the activities of cyclonerosides B-E (6-9) when compared with those of cycloneroside A (5), which contains a linear prenyl chain on the right hand side of the structure. The positive control (quercetin) exhibits an NO-production-inhibitory effect with an IC 50 value of 30.8 µM.
various diseases, such as diabetes, cardiovascular diseases, obesity, arthritis, strok cancer [17]. The over-expression of cellular transduction molecules, such as NO, hist and other pro-inflammatory cytokines, leads to chronic inflammation, which even induces inflammation-related diseases. Chemical inhibitors of cell signal transductio can inhibit the excessive release of these harmful cytokines should be useful as po drug candidates [17,33]. There has been a large number of reports on natural pr with NO-production-inhibitory activity isolated from plants, most of which are phe type components [33]. To the best of our knowledge, there have been no reports NO-production-inhibitory activities of sesquiterpene aminoglycosides. These comp do not show any cytotoxicity at the highest test concentration, and further structura ification of these compounds may lead to new, potent, and safe anti-inflammatory The anti-fungal activities were also tested against plant pathogenic fungi inc Helminthosporium maydis, Gibberella sanbinetti, and Penicillium digitatum. None of th pounds exhibit any obvious activity at a concentration of 100 µg/mL.

Fungal Strain and Fermentation
The Trichoderma sp. SCSIOW21 fungal strain was collected from deep-sea sed Inflammatory responses are known to be deeply associated with the pathogenesis of various diseases, such as diabetes, cardiovascular diseases, obesity, arthritis, stroke, and cancer [17]. The over-expression of cellular transduction molecules, such as NO, histamine, and other pro-inflammatory cytokines, leads to chronic inflammation, which eventually induces inflammation-related diseases. Chemical inhibitors of cell signal transduction that can inhibit the excessive release of these harmful cytokines should be useful as potential drug candidates [17,33]. There has been a large number of reports on natural products with NO-production-inhibitory activity isolated from plants, most of which are phenolic-type components [33]. To the best of our knowledge, there have been no reports on the NOproduction-inhibitory activities of sesquiterpene aminoglycosides. These compounds do not show any cytotoxicity at the highest test concentration, and further structural modification of these compounds may lead to new, potent, and safe anti-inflammatory drugs.
The anti-fungal activities were also tested against plant pathogenic fungi including Helminthosporium maydis, Gibberella sanbinetti, and Penicillium digitatum. None of the compounds exhibit any obvious activity at a concentration of 100 µg/mL.

Fungal Strain and Fermentation
The Trichoderma sp. SCSIOW21 fungal strain was collected from deep-sea sediments and deposited at the Laboratory of Microbial Natural Products, Shenzhen University, China. The fungal strain was identified as Trichoderma species according to its morphological characteristics and ITS gene sequence (OP854922). The fungal strain was cultivated statically for 30 d in rice broth containing 3% sea salt. The culture was extracted using an equal volume of n-BuOH.

Spectral Data
Trichaspside F (2)  Data Availability Statement: The authors confirm that the data supporting the findings of this study are available within the article and its Supplementary Materials.