UPLC-Q-TOF-MS/MS Analysis of Seco-Sativene Sesquiterpenoids to Detect New and Bioactive Analogues From Plant Pathogen Bipolaris sorokiniana

Seco-sativene sesquiterpenoids are an important member of phytotoxins and plant growth regulators isolated from a narrow spectrum of fungi. In this report, eight seco-sativene sesquiterpenoids (1–8) were first analyzed using the UPLC-Q-TOF-MS/MS technique in positive mode, from which their mass fragmentation pathways were suggested. McLafferty rearrangement, 1,3-rearrangement, and neutral losses were considered to be the main fragmentation patterns for the [M+1]+ ions of 1–8. According to the structural features (of different substitutes at C-1, C-2, and C-13) in compounds 1–8, five subtypes (A–E) of seco-sativene were suggested, from which subtypes A, B/D, and E possessed the diagnostic daughter ions at m/z 175, 189, and 203, respectively, whereas subtype C had the characteristic daughter ion at m/z 187 in the UPLC-Q-TOF-MS/MS profiles. Based on the fragmentation patterns of 1–8, several known compounds (1–8) and two new analogues (9 and 10) were detected in the extract of plant pathogen fungus Bipolaris sorokiniana based on UPLC-Q-TOF-MS/MS analysis, of which 1, 2, 9, and 10 were then isolated and elucidated by NMR spectra. The UPLC-Q-TOF-MS/MS spectra of these two new compounds (9 and 10) were consistent with the fragmentation mechanisms of 1–8. Compound 1 displayed moderate antioxidant activities with IC50 of 0.90 and 1.97 mM for DPPH and ABTS+ scavenging capacity, respectively. The results demonstrated that seco-sativene sesquiterpenoids with the same subtypes possessed the same diagnostic daughter ions in the UPLC-Q-TOF-MS/MS profiles, which could contribute to structural characterization of seco-sativene sesquiterpenoids. Our results also further supported that UPLC-Q-TOF-MS/MS is a powerful and sensitive tool for dereplication and detection of new analogues from crude extracts of different biological origins.


INTRODUCTION
Seco-sativenes are a member of sesquiterpenoids possessing a unique bicyclo[3.2.1]octane ring system and different substitutions including glycosylation, methylation, and acylation; different heterocyclic rings such as lactone, furan, and pyran ring; and diverse oxygenation sites (hydroxylation) on the core skeleton, increasing the chemical diversity. The structural differences of seco-sativenes mainly lie in the diverse substituents at C-1, C-2, and C-13 (Li et al., 2020a). From the structural features, it is implied that seco-sativenes come from a sesquiterpene pathway but not from a direct farnesyl pyrophosphate cyclization product. Rearrangement and oxidative cleavage reactions might play a pivotal role in the biosynthetic pathway, which is supported by the isolation of different precursors and intermediates (Mayo et al., 1962aLi et al., 2020a,b). Fungus B. sorokiniana is known for producing a variety of secondary metabolites, with sesterterpene, cyclic peptides, and sesquiterpenoids as the most representative classes (Nihashi et al., 2002;Ali et al., 2016;Qader et al., 2017;Phan et al., 2019). Qader et al. (2017) isolated three seco-sativene sesquiterpenoids from B. sorokiniana, in which helminthosporal acid and helminthosporol displayed a strong phytotoxic effect on lettuce seed germination and toxicity against brine shrimps, and helminthosporal acid also showed antifungal activity. Phan et al. (2019) isolated and elucidated 12 seco-sativene sesquiterpenoids including a new seco-sativene sesquiterpenoid and three new sativene analogues from B. sorokiniana, in which helminthosporic acid and dihydroprehelminthosporol displayed weak necrotic activity against wheat leaves and helminthosporol showed an inhibitory effect on seed germination. Seco-sativene analogues displayed strong phytotoxic effects on cereals and gramineous plants, (Ludwig et al., 1956;Ludwig, 1957;Mayo et al., 1961Mayo et al., , 1962bMayo et al., , 1963Spencer, 1965;Katsumi et al., 1967;Taniguchi and White, 1967;White and Taniguchi, 1972;Pena-Rodriguez et al., 1988;Pena-Rodriguez and Chilton, 1989;Qader et al., 2017;Phan et al., 2019) whereas others possessed plant-growth-promoting biological activities to rice, lettuce, cucumber, and wheat seedlings (Briggs, 1966;Hashimoto et al., 1967;Nukina et al., 1975;Pena-Rodriguez and Chilton, 1989;Miyazaki et al., 2017Miyazaki et al., , 2018Qader et al., 2017). In addition, some seco-sativene sesquiterpenoids also possessed antifungal, cytotoxic, and toxic effects, and other analogues could inhibit the growth of the malaria-causing protozoan of Plasmodium falciparum and exhibited certain anti-NO production activities (Li et al., 2020b). The novel core skeleton and diverse biological activities attracted us to chemically investigate this unique member of sesquiterpenoids. Recently, a series of new seco-sativene sesquiterpenoids were isolated from the endophytic fungus Cochliobolus sativus (teleomorph: Bipolaris sorokiniana) inhabiting in a desert plant, Artemisia desertorum, and their structures were mainly determined by NMR experiments, X-ray diffraction, and high-resolution mass analysis. Helminthosporic acid (2) could promote plant leaf growth, whereas cochliobolin F, helminthosporic acid (2), drechslerine B (8), and helminthosporal acid displayed strong phytotoxic effects on corn leaves (Li et al., 2020a). However, the traditional isolation method was used as the main technique for the isolation of seco-sativene sesquiterpenoids, (Ramos et al., 2019) which precluded discovery of new/novel analogues of secosativene sesquiterpenoids. Thus, efficient approaches for mining novel structures of seco-sativene sesquiterpenoids are urgent.
Mass spectrometry, especially tandem mass spectrometry, has been one of the most important physicochemical approaches for the characterization of secondary metabolites due to its rapidity and sensitivity (Jin et al., 2018;Liang et al., 2018;Conceição et al., 2020;Scupinari et al., 2020). Molecular weight and formula are often inconclusive for metabolite identification; however, fragmentation patterns represent a specific feature for a certain structural class. Chen et al. (2018) applied neutral loss scan in QqQ-MS and molecular formula calculation in UPLC-Q-TOF-MS to detect amorfrutin analogues, which provided the idea of detection and structural dereplication in the complex crude extract. Yang et al. (2017) used UPLC-Q-TOF-MS/MS coupled with neutral loss scan and diagnostic ions to analyze the secondary metabolites of Schisandra chinensis. Ahad et al. (2020) combined UPLC-Q-TOF-MS with SCX-SPE to achieve the enrichment and structural identification of the same skeleton metabolites. Thus, much evidence demonstrated that fragmentation patterns coupled with UPLC-Q-TOF-MS/MS analysis were an efficient and convenient tool for the detection and dereplication of similar metabolites.
A series of seco-sativene sesquiterpenoids (1-8, Figure 1) were isolated and elucidated in our previous work (Li et al., 2020a). According to the structural features (of different groups at C-1, C-2, and C-13), five subtypes of seco-sativenes were suggested (subtypes A-E) (Figure 1). Interestingly, each subtype of the structure has the same diagnostic daughter ions in the mass spectrometric profile, which could provide a reliable approach to analyze structures of seco-sativenes and target potent new analogues. To date, no investigations about electrospray tandem mass/mass of seco-sativenes sesquiterpenoids were reported. The potential application prospect and unique skeleton of secosativenes prompted us to investigate the mass spectrometric cleavage mechanisms of this unique member of sesquiterpenoids.
In this report, the UPLC-Q-TOF-MS/MS fragmentation rules of seco-sativene sesquiterpenoids (1-8) were presented; some known and new seco-sativene sesquiterpenoids were detected from the extract of the plant pathogen fungus Bipolaris sorokiniana based on UPLC-Q-TOF-MS/MS analysis. Two new (9 and 10) and two known (1 and 2) seco-sativene sesquiterpenoids were then isolated and elucidated by HR-ESI-MS and NMR spectra, and the antioxidant activities of these seco-sativene sesquiterpenoids (1, 2, 9, and 10) were assessed.

General Experimental Procedures
Optical rotations were measured on a 241 polarimeter (PerkinElmer, Waltham, United States). UV-2102 (Unico, Shanghai, China) was used to record UV data. IR spectra were recorded on an FTIR-8400S spectrophotometer (Shimadzu, Kyoto, Japan). NMR data were acquired on a Bruker 500
Time-of-flight MS detection was performed with the Xevo G2-XS QTof system (Waters) combined with an ESI source in positive ion scan mode. The desolvation temperature was set at 450 • C with desolvation gas flow at 900 L/h, and the source temperature was 80 • C. The lock mass in all analyses was leucineenkephalin [(M+H) + = 556.2771], used at a concentration of 200 µl/ml and infused at a flow rate of 10 L/min. Raw data were acquired using the centroid mode, and the mass range was set from m/z 100 to 1,000. The capillary voltage was set at 2.5 kV with 30 V of sample cone voltage. The collision energy was set as 6 eV for low-energy scan and a ramp from 30 to 50 eV for high-energy scan. The instrument was controlled by MassLynx 4.1 software.

Strain and Fermentation
The strain of Bipolaris sorokiniana (strain number: ACCC36805) was isolated from the seed of wheat and provided by the Chinese Academy of Agricultural Sciences. The fungus was grown on PDA (potato dextrose agar) plates at 25 • C for 10 days. Then the fresh mycelium was inoculated into the autoclaving sterilized solid medium with the formula of rice (60.0 g) and distilled water (80 ml) in Fernbach flasks (500 ml) for further fermentation at 25 • C for 30 days.

Extraction and Isolation
The fermented rice substrate was extracted with EtOAc three times, and the solvent was evaporated to dryness under vacuum to afford 200 g of crude extract. The original extract was fractionated on a silica gel column using petroleum ether-acetone (1:0-0:1) progressively to give five fractions (Fr. 1 to Fr. 5). Fr. 14-Acetyl-Drechslerine B (10)

DPPH Scavenging Capacity
Take 15 µl of compounds 1, 2, 9, and 10 with a concentration of 10 mM/L and a serial dilution of seven times, and then mix SCHEME 1 | Possible fragmentation pathway of 1.
Frontiers in Microbiology | www.frontiersin.org with DPPH solution. After 30 min, the remaining amount of the DPPH radical was measured spectrophotometrically at 517 nm. In this test, for comparison, V C was considered as the positive control, and ethanol was considered as the negative control.
, where A 0 is the absorbance of the water, A C is the absorbance of ethanol solution, and A S is the absorbance after adding the sample solution. The IC 50 value was processed by GraphPad Prism 8.

ABTS + Scavenging Capacity
Take 15 µl of compounds 1, 2, 9, and 10 with a concentration of 10 mM/L and a serial dilution of seven times, and then mix with ABTS + solution. After 6 min, the remaining amount of the ABTS + radical was measured spectrophotometrically at 405 nm. In this test, for comparison V C served as the positive control, and ethanol served as the negative control.
The clearance rate E is E = [1 -(A S -A 0 )/(A C -A 0 )] × 100%, where A 0 is the absorbance of the water, A C is the absorbance of ethanol solution, and A S is the absorbance after adding the sample solution. The IC 50 value was processed by GraphPad Prism 8.   The high-resolution mass and fragment ions together with the elemental constituents of compound 2 were listed in Table 3.  Figures 5, 6). The highresolution mass and fragment ions together with the elemental constituents of compounds 5 and 6 were listed in Table 4 and Supplementary Table 3 with the elemental constituents of compound 7 are listed in Table 5.
The  Table 6.
With the UPLC-Q-TOF-MS/MS fragmentation mechanisms of 1-8 in hand, it implied that each subtype seco-sativene sesquiterpenoids had a diagnostic daughter ion in the MS profile (subtype A → m/z 175; subtypes B/D → m/z 189; subtype C → m/z 203; subtype E → m/z 187). Though both subtypes B and D had the same diagnostic daughter ion m/z 189, the last cleavage in the sub-type B was the neutral loss of one molecule of CO (−28), whereas the neutral loss of one molecule of H 2 O (−18) was the last cleavage in subtype D, which differentiated these two subtypes B and D. Thus, it could give the possible subtype of secosativene sesquiterpenoids based on the diagnostic daughter ion from the corresponding ESI-Q-TOF-MS/MS data.
Then, the crude extract of the ethyl acetate fraction of the plant pathogen Bipolaris sorokiniana was then analyzed by UPLC-Q-TOF-MS/MS (Supplementary Figure 11). There  (Supplementary Figures 9, 10). The diagnostic ions of 9 and 10 were m/z 189/187 and λ max = 234/215 nm, respectively, implying that structures of 9 and 10 possessed the same subtypes as 7 and 8. The UPLC-Q-TOF-MS/MS fragmentation pathways of these two compounds were nearly the same as those of 7 and 8, except for an additional acetyl group. To confirm this hypothesis, 9 and 10 were isolated from the extract and elucidated by IR, NMR, and HR-ESI-MS spectra. The IR absorption bands at 3,420 cm −1 showed the hydroxyl group in 9, and 2,931 and 1,747 cm −1 revealed the presence of alkyl and ester moieties, respectively, which were also present in that of 10. The 1 H-NMR spectrum revealed the similarity of 9/10 with 7/8, except that the chemical shift values of -CH 2 -12 of 9 and -CH 2 -14 of 10 were down-fielded in 9/10 (Osterhage et al., 2002) and an additional methyl signal was observed in the 1 H-NMR spectrum of 9/10. This implied that the additional acetyl group was connected at C-12 in 9 SCHEME 5 | Possible fragmentation pathways of 8 and 10. and C-14 in 10. The key HMBC correlations from -CH 2 -12/CH 2 -14 and 1 -Me to C-2 (δ C 171.2 in 9/10) supported the conclusion (Figure 2 and Supplementary Figures 12-29). Thus, the structures of 9 and 10 were determined, and their ESI-Q-TOF-MS/MS fragmentation pathways were consistent with those of 7/8 (Schemes 4, 5 and Supplementary Tables 4, 5).

Antioxidant Activity
The antioxidant activities of compounds 1 and 2 were evaluated by the DPPH and ABTS + free radical scavenging test, and the results were presented as IC 50 values. The results demonstrated that compound 1 displayed moderate antioxidant activities with IC 50 of 0.90 and 1.97 mM for DPPH and ABTS + scavenging capacity, respectively. However, compound 2 did not show the obvious antioxidant activity. They were measured by comparing the scavenging ability of DPPH free radical and ABTS + free radical with V C , a well-known potent antioxidant and free radical scavenger with IC 50 of 0.14 and 0.42 mM for DPPH and ABTS + scavenging capacity, respectively ( Table 7).

DISCUSSION
This report analyzed the fragmentation patterns of eight representative seco-sativene sesquiterpenoids (1-8) using UPLC-Q-TOF-MS/MS, and McLafferty rearrangement, 1,3rearrangement, and neural loss (−18, −28) were the main fragmentation patterns. The results indicated that dehydration (−18) occurred easily in seco-sativenes with strong abundance of dehydration peak observed in 1-8, and similar reports were found in other studies (McLafferty and Gohike, 1959;Alén, 1987;Nawamaki and Kuroyanagi, 1996;Sun et al., 2012;Giri et al., 2017;Qian et al., 2018). This may be due to an electron impact inducing fragmentations of alkene monocarboxylic acids to form an active OH ion, which combines an available methylhydrogen atom to one lost molecule of H 2 O (−18) through the McLafferty rearrangement in a six-membered system (Baldas et al., 1969;Alexander et al., 1972 O] + , the molecular ion peak of 1-7 and 9 can be inferred. Compounds 8 and 10 possess a lactone ring at C-1 and C-2, which is different from 1-7 and 9. The special group in 8 and 10 leads to their molecular ion peaks [M+H] + easily being observed in MS profiles. This might be a key signal for differentiating subtype E from other subtypes. Although the McLafferty rearrangement together with 1,3-rearrangement in alkene monocarboxylic molecules to produce a base peak of dehydration has been reported, this was the first report in seco-sativene sesquiterpenoids mass analysis, which provided a base for the seco-sativenes structural elucidation. Diagnostic daughter ions of five subtypes of seco-sativene sesquiterpenoids were provided base on UPLC-Q-TOF-MS/MS analysis in this report. Subtypes A, B/D, and E possessed diagnostic daughter ions at m/z 175, 189, and 203, respectively, whereas subtype C showed a characteristic daughter ion at m/z 187 in the UPLC-Q-TOF-MS/MS profiles. The main difference between subtypes B and D was that the last cleavage was the neutral loss of one molecule of CO (−28) in subtype B, not the neutral loss of one molecule of H 2 O (−18) in subtype D. Diagnostic ions provided signals for the different subtypes of seco-sativenes, and the rearrangement and neutral loss (H 2 O, CO, and HOAc) in the fragmentation patterns provided the possible groups on the structures of seco-sativene sesquiterpenoids. Thus, the structures of the seco-sativenes could be inferred by fragmentation patterns combined with the diagnostic ions and molecular formula based on the UPLC-Q-TOF-MS/MS profile. This report provides a reliable method for the structural analysis of seco-sativene sesquiterpenoids.
Compounds 1, 2, 9, and 10 and other seco-sativenes are a class of phytotoxins. Compound 1 was previously isolated from C. sativus (teleomorph: B. sorokiniana) without phytotoxicity on corn leaves, but helminthosporal acid with an aldehyde group at C-1 (a carboxyl group at C-1 in 1) possessed strong phytotoxic activity (Li et al., 2020a). Therefore, the aldehyde group in helminthosporal acid might be a potential active group for phytotoxicity. Compound 2 exhibited bidirectional regulation activities. On the one hand, it was the gibberellin-like plant growth regulator, which could promote the growth of plant roots and leaves at low concentrations (Qader et al., 2017;Li et al., 2020a). On the other hand, it showed phytotoxicity on corn leaves (Li et al., 2020a). Compounds 9 and 10 did not show obvious antioxidant activity in this report, and no more activities of 9 and 10 were tested due to limited amounts. Compared with 7 and 8, 9 and 10 possess an extra acetyl group at 12-OH and 5-OH, respectively. Osterhage et al. reported that 7 showed inhibitory activity against tyrosine kinase p56, Microbotryum violaceum, Eurotium repens, and Escherichia coli (Osterhage et al., 2002) and that 8 showed antifungal activity against M. violaceum (Osterhage et al., 2002) and phytotoxicity on corn leaves (Li et al., 2020a). Several reports suggested that 5-CH 2 OH might be the potent active group of seco-sativenes sesquiterpenoids for their phytotoxicity (Nakajima et al., 1994;Li et al., 2020b). Therefore, 9 might possess phytotoxicity against barley seeds and corn leaves. But their structure-activity relationships (SARs) need to be further studied. At present, the reports on the activities of secosativene sesquiterpenoids mainly focused on phytotoxicity and growth-promoting effects with few reports about other activities (Li et al., 2020b). To further explore the medicinal value of secosativenes, the antioxidant activity of these compounds (1, 2, 9, and 10) were studied in this report. Only then did compound 1 show moderate activity on DPPH scavenging capacity, which indicated that 13-COOH might be a possible active group, and further biological exploration should be needed in the future. UPLC-Q-TOF-MS/MS spectrometry has evolved to be a mature and common technique, which is now widely used to analyze secondary metabolites from diverse biological resources. Most of researchers used this technology to identify (new/novel) metabolites or dereplicate in different crude extracts (Jin et al., 2018;Wang et al., 2020). Recently, a molecular networking technique based on (U)HPLC-MS/MS combined with different databases was used in the dereplication and targeting of new natural products from diverse biological resources (Watrous et al., 2012;Yang et al., 2013;Aksenov et al., 2017;Hou et al., 2019;Ramos et al., 2019;Rivera-Chaìvez et al., 2019;Shi et al., 2019;Wu et al., 2019;Chao et al., 2020;Zang et al., 2020;Lei et al., 2021;Lin et al., 2021). The molecular networking technique used the known or new compound as the "seed" to realize the visualization of analogues. In the network, MS data were collected from LC-MS and uploaded to the GNPS database for data processing to produce total molecular network profiles. Every node in the same network represented a compound possessing the same core skeleton. Known or new analogues can be quickly inferred according to molecular weight, molecular formula, and fragmentation patterns based on node analysis through searching different databases or house libraries. The targeted isolation of the seed analogues could be realized by searching the location of the "seed" (Klitgaard et al., 2015;Allard et al., 2016;Trautman and Crawford, 2016;Naman et al., 2017;Olivon et al., 2017a,b;Nothias et al., 2018). When a molecular network is combined with fragmentation patterns, the range of metabolites would be narrowed, and the precision of targeted-isolation-compounds would be improved (He et al., 2021). Thus, molecular networking based on the (U)HPLC-MS/MS technique would provide a more convenient approach for dereplication and targeting-isolation of new seco-sativene sesquiterpenoids in the future.

CONCLUSION
Eight seco-sativene sesquiterpenoids (1-8) were analyzed using the UPLC-Q-TOF-MS/MS technique in positive mode, from which their possible mass fragmentation patterns were suggested, and neural loss, McLafferty rearrangement, and 1,3rearrangement were the main clearage patterns. These eight seco-sativene sesquiterpenoids (1-8) were summarized to be five subtypes according to their structural features. Each subtype possessed a diagnostic daughter ion, which, in return, could contribute to the elucidation of seco-sativene sesquiterpenoids. Based on the fragmentation mechanism mentioned above, some analogues including two potentially new ones were detected. Two known (1 and 2) and two new analogues (9 and 10) were then isolated from the extract of the plant pathogen Bipolaris sorokiniana. Their structures were elucidated mainly by NMR spectra and supported based on their UPLC-Q-TOF-MS/MS analysis. The results demonstrated that diagnostic mass ions of seco-sativene sesquiterpenoids in the UPLC-Q-TOF-MS/MS profiles provided a convenient and high-performance approach for structural characterization and also support that UPLC-Q-TOF-MS/MS is a powerful and sensitive tool for dereplication and detection of new analogues in crude extracts. This study will pave the way for the structural analysis and targeting isolation of seco-sativene sesquiterpenoids in different fungal crude extracts.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

AUTHOR CONTRIBUTIONS
Y-DW, JY, and GD: experiment and writing -review and editing. Y-DW, JY, Y-YL, X-MT, and QL: data collection. S-YY and S-BN: activity experiment. HD: resources. JY, GD, and L-PG: funding acquisition. GD and Y-DW: writing -original draft preparation. All authors have read and agreed to the published version of the manuscript.