Bioactive secondary metabolites from endophytic strains of Neocamarosporium betae collected from desert plants

Endophytic fungi from desert plants belong to a unique microbial community that has been scarcely investigated chemically and could be a new resource for bioactive natural products. In this study, 13 secondary metabolites (1–13) with diverse carbon skeletons, including a novel polyketide (1) with a unique 5,6-dihydro-4H,7H-2,6-methanopyrano[4,3-d][1,3]dioxocin-7-one ring system and three undescribed polyketides (2, 7, and 11), were obtained from the endophytic fungus Neocamarosporium betae isolated from two desert plant species. Different approaches, including HR-ESI-MS, UV spectroscopy, IR spectroscopy, NMR, and CD, were used to determine the planar and absolute configurations of the compounds. The possible biosynthetic pathways were proposed based on the structural characteristics of compounds 1–13. Compounds 1, 3, 4, and 9 exhibited strong cytotoxicity toward HepG2 cells compared with the positive control. Several metabolites (2, 4–5, 7–9, and 11–13) were phytotoxic to foxtail leaves. The results support the hypothesis that endophytic fungi from special environments, such as desert areas, produce novel bioactive secondary metabolites.


Introduction
Plant endophytic fungi are one of the significant sources of novel bioactive compounds (Tan and Zou, 2001;Zhang et al., 2006;Kharwar et al., 2011;Yang et al., 2012;Adnani et al., 2017). and can produce bioactive drug molecules, such as the anticancer compound taxol (Kharwar et al., 2011). Furthermore, Alternaria oxytropis can biosynthesize the neurological toxin swainsonine to protect its host plant against livestock (Braun et al., 2003;Cook et al., 2011;Cook et al., 2013;Grum et al., 2013;Cook et al., 2014). Large-scale chemical investigations during the last two decades have lowered the chance of obtaining novel bioactive compounds from plant endophytic fungi inhabiting common environments. Therefore, fungi from special biotopes have become hot spots for isolating new natural products (Tian et al., 2017;Zhang et al., 2018a). Compared with plants in common environments, desert plants must develop strategies to adapt to extreme conditions, such as strong ultraviolet radiation, high-concentration saline-alkali stress, drought, and large temperature fluctuations. Endophytic fungi inhabiting desert plants can produce secondary metabolites, with different biological or ecological functions, to help the host plants tolerate extreme conditions during their mutualistic symbiotic relationship. Thus, endophytic fungi inhabiting desert plants belong to a unique fungal community that can biosynthesize novel bioactive secondary metabolites. Recently, our group chemically investigated fungi collected from desert plants across Northwest China and isolated a series of small molecules with novel skeletons and various biological activities (Zhang et al., 2018b;Song et al., 2019;Tan et al., 2019;Li et al., 2020;Zhang et al., 2021;Xu et al., 2022). Thirteen secondary metabolites (1-13) (Figure 1), including four undescribed structures (1, 2, 7, and 11), with different biosynthetic origins were obtained from the endophytic fungus Neocamarosporium betae (synonym: Phoma betae), isolated from Suaeda glauca Bunge (Chenopodiaceae) and Nitraria roborowskii Kom. (Zygophyllaceae). This study describes the extraction and purification, structural characterization, biosynthesis, and biological evaluation of these compounds.

Strain and fermentation
Neocamarosporium betae (CGMCC3.19915; GenBank Accession: ON945545) isolated from S. glauca Bunge and N. betae (CGMCC3.20844; GenBank Accession: ON945546) isolated from N. roborowskii Kom. were identified by Dr. Bing-da Sun. Endophytic fungi were cultured on PDA at 25°C for 1 week. Fresh mycelia were incubated in a solid medium (80 ml of distilled water and 60.0 g of rice in 500 ml flasks) to be fermented at room temperature for 4 weeks.

Cytotoxic evaluation using the MTT assay
The cytotoxicity of compounds 1-13 was tested against three cancer cell lines, namely, MCF7, B16, and HepG2, using the MTT assay. cis-Platinum was used as a positive control. Cytotoxic assessment was as described .

Phytotoxicity assay
The phytotoxicity of different concentrations of the pure compounds was tested on the leaves of green foxtail and corn using a leaf puncture test (Li et al., 2020). Fresh leaves were cut into 4-cm rectangles using scissors, washed three times with 75% alcohol, and dried in a sterile glass Petri dish, and sterile wet filter paper was placed on the bottom. Next, 5-mm diameter filter paper discs, prepared using a hole punch, were disinfected for 30 min at 121°C. The compounds were prepared as 1 mg/ml solutions with dimethyl sulfoxide. On each leaf segment, three benign lesions were made equidistant from each other using forceps. A disc was placed over the lesion, and 20 ml of dimethyl sulfoxide or test solution was added to each disc. The experiment was repeated three times and the results were recorded after 72 h.

Results and discussion
The molecular formula of 1 was determined as C 13 H 16 O 4 using HR-ESI-MS (m/z 237.1122 [M+H] + ; calcd. 237.1127), with six degrees of unsaturation. Three methyl, two sp 3 methylene (one oxygenated), two sp 3 methine, one quaternary carbon, four olefinic carbon, and one carbonyl group were suggested to be present in the structure of 1 (Table 1) based on the analysis of 1 H, 13 C, and HMBC NMR spectra. These data accounted for all resonances of 1 H-and 13 C-NMR spectra, with three rings, in structure 1. The 1 H-1 H COSY correlations gave an isolated proton spin system corresponding to -CH 2 -4/CH 2 -5/C-6/C-13/13-Me. Complete connectivity was established through HMBC spectrum analysis (Figure 2). The 9-Me/C-9/C-10/C-11/C-12 fragment was constructed based on the HMBC correlations of 9-Me to C-9 and C-10 and of H-10 to C-9, C-11, and C-12; C-12 was connected to C-6, C-7, and C-11 by correlations from H-6 to C-7 (d C = 163.6), C-11, and C-12. The weak, long-range HMBC correlation from 9-Me to C-7 confirmed a 2H-pyran-2,4(3H)dione ring (A ring) in the structure of 1 (Ding et al., 2012). C-2 was linked to C-13 and 2-Me due to the HMBC correlations of 2-Me with C-2 and C-13, and an ether bond was formed between C-2 and C-4 according to the HMBC correlation of H-4 with C-2. Considering the chemical shift values of C-2 (d C = 103.3)/C-11 (d C = 166.3) and degrees of unsaturation, an ether bond might be present between C-2 and C-11. This hypothesis was supported by the weak long-range HMBC correlation from 2-Me to C-11. Thus, the planar structure of 1 was established as possessing a unique 5,6-dihydro-4H,7H-2,6methanopyrano[4,3-d][1,3]dioxocin-7-one skeleton, which was the first report in nature. The relative configuration of 1 was established based on its ROESY correlations (Figure 2), whereas the stereochemistry of 1 was resolved by CD spectroscopy. The ECD spectrum of 1 resembled the calculated ECD curve of (2S, 6S, 13R)-1, but was opposite to that of (2R, 6R, 13S)-1 (Figure 3), which established the absolute configuration of 1 (Figure 1).
The molecular formula of compound 2 was determined to be C 14 H 20 O 6 according to the HR-ESI-MS data (m/z 285.1329 [M +H] + ; calcd. 285.1338). The structure of 2 was mainly elucidated using 2D-NMR spectral correlations. The 1 H-1 H COSY correlations of 2 showed an isolated proton spin system, corresponding to -C-8/C-9-, and the remaining connection was established through the HMBC spectrum. The quaternary carbon (C-7) was connected with C-6, C-8, C-12, and C-13, based on correlations from 13-Me to C-6, C-7, C-8, and C-12. The connection of C-4 with C-3, C-5, and C-11 was achieved through the HMBC correlations of -CH 2 -5 with C-3, C-4, and C-11. HMBC correlations from 1-Me/10-Me to C-2 and C-3 led to the connection of -C-1/C-2/C-3/C-10-. C-5 was connected to the keto-group C-6, whereas 14-OMe was anchored with C-12, according to the correlations from -CH 2 -5 to C-6/C-7 and 14-OMe to C-12, respectively. The chemical shift values and degrees of unsaturation of C-2/C-11, together with the weak long-range HMBC correlation from 1-Me to C-11, indicated the presence of a g-hydroxyl-g-lactone unit in 2, possessing the same unique carbon skeleton found in spicifernin (Nakajima et al., 1990). Spicifernin existed as an equilibrium mixture of keto acid and a g-hydroxyl-glactone in a solvent (Nakajima et al., 1990). A set of double carbon signals was observed in the 13 C-NMR spectrum of 2, suggesting that a pair of stereoisomers might be present in 2 due to the hemiacetal hydroxyl at C-2 (Table 1).
The molecular formula of compound 7 was determined to be C 21 H 34 O 4 according to the HR-ESI-MS spectra (m/z 351.2531, [M +H] + ; calcd. 351.2535), with five degrees of unsaturation. Comparison with the NMR spectra of 8-10 indicated that compound 7 possessed the same carbon skeleton as compounds 8-10, except that an additional terminal double bond (d H-17 = 6.99, J = 16.8 Hz, 10.2 Hz; d H-18a = 6.37, J = 16.8 Hz, 1.2 Hz; d H-18b = 5.71, J = 10.2 Hz, 1.2 Hz) was observed in the NMR spectra (Table 2) (Ichihara et al., 1983a;Ichihara et al., 1983b). The 13 C-NMR chemical shift value of C-16 in compound 7 was d C = 203.3, which was smaller than that for compounds 8 and 10 (Ichihara et al., 1983a;Ichihara et al., 1983b). These findings suggest that the additional terminal olefins were connected with C-16 to form an a, b-unsaturated keto-group, causing the 13 C-NMR chemical shift value to be high-fielded. This hypothesis was further confirmed by the HMBC correlations from CH 2 -18/H-17 to C-16. The 1 H-1 H COSY correlations together with coupling constant analysis determined the connectivity of H-17/CH 2 -18 ( Figure 2). The absolute configuration of compound 7 was the same as that of compound 10, considering their similar CD spectra ( Figure S23, Supplementary data). The molecular formula of compound 11 was determined to be C 10 H 12 O 3 according to the HR-ESI-MS (m/z 181.0853, [M+H] + ; calcd. 181.0865) and NMR data (Table 2), with five degrees of unsaturation. The 1 H-and 13 C-NMR spectra of compound 11 displayed similar resonances as those of compound 13, except that 9-Me in 11 was a singlet peak compared with the doublet (d 9-Me = 2.25) in 13. An additional keto-group signal (d C = 206.8) was observed in the 13 C-NMR spectrum of 13. These resonance differences in compounds 11 and 13 implied that the hydroxyl at C-8 in 13 was transformed into the corresponding keto group in 11. This hypothesis was further supported by the HMBC correlations of 9-Me with C-7 and C-8 (Figure 2), which established the structure of 11.
Neocamarosporium betae is often found as a plant pathogen or endophyte on different plants, from which many bioactive secondary metabolites with different carbon skeletons have been isolated (Xu et al., 2021). The structure of compound 1 possesses a dioxocin-7-one skeleton, which is the first report in Nature. Based on the structural features of 1, the biosynthetic pathway of 1 might originate from two polyketide hybrids. Spiciferones (3-5), spiciferinone (6), and spicifernin (containing the core skeleton of 2) were co-isolated from the plant pathogen Cochliobolus spicifer Nelson, which causes leaf-spot disease in wheat (Nakajima et al., 1991;Nakajima et al., 1992a). Based on isotope-labeling experiments, Nakajima et al. proposed a biogenetic pathway for spicifernin (Nakajima et al., 1992b). Given that compounds 2-6 were isolated from N. betae, the biosynthetic pathways for compounds 2-6 might be the same as that of spicifernin ( Figure  S35). Betaenones (containing the core skeleton of compounds 7-10) act as phytotoxins and were isolated from the pathogenic fungus P. betae Fr., which causes leaf-spot disease in sugar beet, inducing chlorosis in host plant leaves (Ichihara et al., 1983a;Ichihara et al., 1983b;Oikawa et al., 1984b). The structural features of betaenones (decalin scaffold) indicate that the core skeleton was formed by a Diels-Alder reaction ([4 + 2] cycloaddition). The biosynthesis of betaenone B was examined through isotope-labeling experiments, which revealed that the carbon skeleton originated from eight acetate units through a polyketide pathway, and the methyl groups were derived from methionine (Oikawa et al., 1984a).
Recently, the gene cluster of betaenones was identified. The highly reducing polyketide synthase gene with a trans-acting enoyl reductase domain was heterologously expressed to form a decalin scaffold. In addition, a series of post-modification oxidative enzymes were investigated, which allowed for the reconstitution of the betaenone biosynthetic machinery (Ugai et al., 2015).

Conclusion
A chemical investigation was conducted on the secondary metabolites produced by the endophytic fungus N. betae that was isolated from two desert plants. A total of 13 polyketides (1-13), including four undescribed molecules (1, 2, 7, and 11), were purified. The new metabolites were examined using HR-ESI-MS, UV spectroscopy, IR spectroscopy, NMR, and CD. The different carbon skeletons of compounds 1-13 enriched the structural diversity of secondary metabolites from N. betae. The possible biosynthetic pathways were suggested according to the structural features of compounds 1-13. In addition, compounds 1, 3, 4, and 9 exhibited strong cytotoxicity toward HepG2 cancer cells, and compounds 2, 4-5, 7-9, and 11-13 displayed phytotoxic activities against foxtail leaves. The results in this report provide further evidence that endophytic fungi inhabiting specific biotopes, such as desert plants, could be new sources of novel secondary metabolites with different biological activities.

Data availability statement
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

Author contributions
SY and GD conceived and designed the study. PL and YT performed the experiments and collected the experimental data. PL, YT, Y-DW, and QL evaluated the activities of all the isolates. JY and D-AS provided funding and analyzed the original draft. X-KX and B-DS provided resources. PL, YT, S-XY, and GD wrote the first draft of the manuscript. PL, YT, and GD revised the manuscript. All authors contributed to the article and approved the submitted version.

Funding
This study was financially supported by the Key Project at Central Government Level: Establishment of the sustainable use for valuable Chinese medicine resources (2060302-2101-18), the National Natural Science Foundation of China (Grant No.