Characterization of New Gambierones Produced by Gambierdiscus balechii 1123M1M10

The benthic dinoflagellate genus Gambierdiscus is the primary producer of toxins responsible for ciguatera poisoning (CP), a food intoxication endemic in tropical and subtropical areas of the world. We used high-performance liquid chromatography tandem high-resolution mass spectrometry (HPLC-HRMS) to investigate the toxin profile of Gambierdiscus balechii 1123M1M10, which was obtained from Marakei Island (2°01′N, 173°15′E), Republic of Kiribati, located in the central Pacific Ocean. Four new gambierone analogues including 12,13-dihydro-44-methylgambierone, 38-dehydroxy-12,13-dihydro-44-methylgambierone, 38-dehydroxy-44-methylgambierone, and desulfo-hydroxyl gambierone, and two known compounds, gambierone and 44-methylgambierone, were proposed by analyzing their fragmentation behaviors and pathways. Our findings provide new insights into the toxin profile of Gambierdiscus balechii 1123M1M10, which can be used as a biomarker for species identification, and lay the foundation for further toxin isolation and bioactivity studies of gambierones.


Introduction
Gambierdiscus is a genus of marine autotrophic epi-benthic dinoflagellate that grows on the surface of macroalgae, corals, and sand grains [1][2][3]. This dinoflagellate has gained scientists' attention due to its production of ladder-shaped polyether toxins, including ciguatoxins (CTXs) [4,5], maitotoxins (MTXs) [6][7][8], gambierones [9][10][11], gambierol [9], gambieric acids [10], and gambieroxide [11]. Most of these toxins can bioaccumulate in the food chain, and the consumption of fish or shellfish contaminated with these toxins can cause ciguatera food poisoning (CFP), which is the most common non-microbial foodborne illness occurring in the tropical and subtropical regions of the world [12]. Patients with CFP may suffer from gastrointestinal, cardiological, and long-lasting neurological symptoms [13]. It is estimated that nearly 50,000 people are affected by CFP annually [14]. However, no effective CFP prevention and treatment strategy has been presented thus far, which is attributable to a lack of pure toxins and a poor understanding of their toxicity. Although these toxins are notoriously toxic, they have potential medicinal value because of their desirable biological activities. For instance, gambieric acids show significant antifungal properties [10], while gambierol inhibits voltage-gated potassium channels [15,16]. Maitotoxins (MTXs) enhance calcium ion influx across cell

Investigation of Toxin Profile in Gambierdiscus balechii 1123M1M10
Some studies and our previous report have demonstrated that G. balechii 1123M1M10 has CTX-like toxicity, suggesting that it may be the producer of CTXs or gambierones [8,41,46]. To characterize the toxin profile of G. balechii 1123M1M10, an untargeted metabolomics study was conducted using electrospray ionization time-of-flight mass spectrometry (ESI-TOF-MS) operated with an information-dependent acquisition (IDA) method. The analysis performed in positive mode yielded more chemical information than in the negative mode, and was therefore used for further investigation. It was reported that the general fragmentation pathways of ladder polyether compounds were the dissociation of one covalent C-C bond, along with one etry (ESI-TOF-MS) operated with an information-dependent acquisition (IDA) method. The analysis performed in positive mode yielded more chemical information than in the negative mode, and was therefore used for further investigation. It was reported that the general fragmentation pathways of ladder polyether compounds were the dissociation of one covalent C-C bond, along with one or two covalent C-O bonds, and the typical feature of their fragmentations was the consecutive loss of H2O (18 Da), since the dissociation of polarized C-O bonds occurs more easily than C-C bonds [47]. According to the analyses of MS 2 fragments-including the fragment of the side chain in the ring I m/z 109, specific fragments m/z 233 or m/z 219, sulfate loss (m/z 959 or 945), and successive water lossessix putative gambierone analogues were found from fractions 3 to 8 from a total of 18 fractions obtained from the Sephadex LH-20 chromatography performed in this study. Four of them were tentatively proposed as novel compounds (desulfo-hydroxyl gambierone (2), 38-dehydroxy-12,13-dihydro-44-methylgambierone (3), 12,13-dihydro-44methylgambierone (4) and 38-dehydroxy-44-methylgambierone (6)) and were eluted at 5.35 min, 6.00 min, 6.01 min, and 6.29 min, respectively ( Figure 1A). The other two compounds were proposed as gambierone (1) and 44-methylgambierone (5), based on the comparison of the fragmentation behaviors and retention times with the published data and standards [40,48]. The retention time of putative 44-methylgambierone was 6.28 min, consistent with the standard, while the putative gambierone (retention time: 5.26 min) was eluted earlier than the standard (retention time: 5.92 min) (Figure 1). Under the same analysis conditions, the putative gambierone and gambierone standard had similar mass spectra and the same specific product ions, indicating that they may have the same structures with different configurations. Of the six gambierone analogues, four of them were not detected by the IDA method in the negative ESI mode due to their low concentration in the sample. In the ESI (−)-MS 2    Compound 1 ([M + NH 4 ] + m/z 1042.5006, C 51 H 80 NO 19 S + , ∆ 3.3 ppm) was considered to have the same structure with different configurations as gambierone according to the comparison of their retention times and fragment ions. The full-scan mass spectrum of compound 1 showed proton, ammonium, and sodium adduct precursor ions at m/z 1025, 1042, and 1047, respectively, which corresponded with the standard but with different ion ratios ( Figure 2A). However, compound 1 showed similar product ions with gambierone in the MS 2 spectra ( Figure 2B). The proposed fragmentation pathways of this compound are discussed below ( Figure 2C). The fragmentation of the ammonium loss plus the water loss produced the proton adduct ion at m/z 1025 and the pseudo-molecular ion [M + H − H 2 O] + at m/z 1007, which is consistent with that of the standard under the same condition. The product ion at m/z 962, which was not present in the MS 2 spectrum of the standard, was proposed as the pseudo-molecular ion [M + NH 4 − SO 3 ] + with a mass difference (∆ppm)  [48]. The product ion at m/z 867 was generated by the cleavage of the right-side chain and the broken C-C bond was between C-2 and C-3. In addition to the cleavage of the side chain, the common cleavage of ring I was also observed in compound 1. The product ion at m/z 711 was formed by the cleavage of one covalent C-C bond and one C-O bond in ring I (C 33 -C 34 and C 37 -O), plus the further loss of SO 3 (80 Da). The fragment ion at m/z 219 was produced by the cleavage of the C-O bond (C-37) along with the C-C bond between C-34 and C-35. The product ion at m/z 161 was formed by the cleavage of the C-O bond located at the C-37 and C-C bond (C-36 and C-37), plus the further loss of H 2 O (18 Da). Furthermore, the cleavage of the ring G (C 27 -C 28 and C 26 -O) plus the loss of SO 3 (80 Da) generated the fragment ion at m/z 567. The product ion at m/z 457 was formed by the dissociation of one covalent C-C bond (C-22 and C-23) and one C-O bond (C-27) in the ring F, and the elimination of SO 3 (80 Da). The product ion at m/z 621 was formed by the dissociation of one covalent C-C bond between C-13 and C-14, and one C-O bond located at C-15 in the ring C. The fragmentation pathways were proposed based on MS 2 data, and their mass differences (∆ppm) were less than 10 ppm (Table S1).
are discussed below ( Figure 2C). The fragmentation of the a loss produced the proton adduct ion at m/z 1025 and the p H2O] + at m/z 1007, which is consistent with that of the stand The product ion at m/z 962, which was not present in the M was proposed as the pseudo-molecular ion [M + NH4 − (∆ppm) of 1.9 ppm. The precursor ion at m/z 1025 yielded th was attributed to the elimination of the neutral loss of SO m/z 927, 909, 891, 873, and 855 were generated by the loss 3 H2O (54 Da), 4 H2O (72 Da), and 5 H2O (90 Da), respect consistent with the gambierone standard and reported MS m/z 867 was generated by the cleavage of the right-side ch was between C-2 and C-3. In addition to the cleavage of the age of ring I was also observed in compound 1. The produc the cleavage of one covalent C-C bond and one C-O bond plus the further loss of SO3 (80 Da). The fragment ion at cleavage of the C-O bond (C-37) along with the C-C bond product ion at m/z 161 was formed by the cleavage of the and C-C bond (C-36 and C-37), plus the further loss of H cleavage of the ring G (C27-C28 and C26-O) plus the loss of S ment ion at m/z 567. The product ion at m/z 457 was form covalent C-C bond (C-22 and C-23) and one C-O bond (C-2 nation of SO3 (80 Da). The product ion at m/z 621 was form covalent C-C bond between C-13 and C-14, and one C-O b C. The fragmentation pathways were proposed based on M ences (∆ppm) were less than 10 ppm (Table S1).   Compound 5 ([M + H] + m/z 1039.4865, C52H79O19S + , Δ 6.3 ppm) could unambiguously be identified as 44-methylgambierone by comparing the retention times and mass spectra with the standard and analyzing the proposed fragmentation pathways. The full-scan mass spectrum of compound 5 showed proton, ammonium, and sodium adduct precursor ions at m/z 1039, 1056, and 1061, respectively, which were consistent with the reference standard ( Figure 3A). Under the same MS 2 condition, compound 5 showed almost the same MS 2 ions as 44-methylgambierone ( Figure 3B). The proposed fragmentation pathways of the putative 44-methylgambierone were discussed below ( Figure 3C). The precursor ion yielded product ions at m/z 1021, 1003, 985, and 967, which were attributed to the elimination of the sequential neutral loss of H2O (18 Da) from the precursor ion at   19 S + , ∆ 6.3 ppm) could unambiguously be identified as 44-methylgambierone by comparing the retention times and mass spectra with the standard and analyzing the proposed fragmentation pathways. The full-scan mass spectrum of compound 5 showed proton, ammonium, and sodium adduct precursor ions at m/z 1039, 1056, and 1061, respectively, which were consistent with the reference standard ( Figure 3A). Under the same MS 2 condition, compound 5 showed almost the same MS 2 ions as 44-methylgambierone ( Figure 3B). The proposed fragmentation pathways of the putative 44-methylgambierone were discussed below ( Figure 3C). The precursor ion yielded product ions at m/z 1021, 1003, 985, and 967, which were attributed to the elimination of the sequential neutral loss of H 2 O (18 Da) from the precursor ion at m/z 1039. The product ion at m/z 959 was formed by the neutral loss of SO 3  , and 851, respectively. Except for the neutral losses, there were many fragments obtained by the cleavage of the right-side chain. For instance, the fragment ion at m/z 901 was achieved with the cleavage of the C-C bond between C-38 and C-39, and then the elimination of H 2 O (18 Da) produced the product ion at m/z 883, while the elimination of SO 3 (80 Da) plus the further sequential loss of H 2 O (18 Da) produced fragment ions at m/z 821, 803, 785, and 767. The small fragment ion at m/z 123 was formed by the cleavage of the C-C bond between C-39 and C-40, and then the elimination of the methyl group produced the fragment ion at m/z 109. The product ion at m/z 95 was generated by the breaking of the C-C bond located at C-40 and C-41. In addition, the cleavage of the ring I was also observed in this study. The product ion at m/z 693 was formed from the fragment ion at m/z 959 by the elimination of C 16 H 26 O 3 (266 Da), which corresponded to the dissociation of one covalent C-C bond between C-33 and C-34, and one C-O bond located at C-32 in the ring I. The sequential loss of H 2 O (18 Da) from the product ion at m/z 693 produced fragment ions at m/z 675, 657, and 639. Fragment ion at m/z 831 was also obtained by the cleavage of the ring I but at different sites (C 35 -C 36 and C 37 -O), and then the elimination of SO 3 (80 Da) plus the further sequential loss of H 2 O (18 Da) produced fragment ions at m/z 751, 733, and 715, respectively. Furthermore, the cleavage of ring I generated five small fragments at m/z 247, 233, 215, 193, and 175, which have been commonly observed in previous studies [40,48]. The fragment ion at m/z 303 was obtained by the cleavage of one covalent C-C bond (C-30 and C-31) and one C-O bond (C-29) in the ring H plus the neutral loss of H 2 O (18 Da). Furthermore, the cleavage of the ring G generated two moieties. Then, the loss of SO 3 (80 Da) and sequential losses of H 2 O (18 Da) from the left moiety at m/z 647 generated two fragment ions at m/z 567 and 549, while the elimination of two and three H 2 O (18 Da) from the right moiety produced the fragment ions at m/z 357 and 339. The fragmentation pathways were proposed based on MS 2 data, and their mass differences (∆ppm) were less than 10 ppm (Table S2). and their mass differences (∆ppm) were less than 10 ppm (Table S2).      Figure 4B), suggesting the existence of two additional hydrogens in the structure of compound 4. The proposed fragmentation pathways of the above-mentioned compound 5 showed that the cleavage of the right-side chain, I, H, and G rings yielded product ions at m/z 95, 123, 109, 233, 215, 193, 175, 303, and 357. These fragment ions, which were diagnostically related to the characterization of right-side moieties, were also observed in compound 4 ( Figure 4C), suggesting that the two compounds shared the same structure at the right side of ring G and the additional hydrogens were likely located at the left side of ring G. Moreover, the fragmentation of putative 44-methylgambierone showed that the cleavage of the C-C bond (C-38 and C-39) at the right-side chain plus the further elimination of H2O (18 Da) and SO3 (80 Da) produced fragment ions at m/z 901, 883, 821, 803, and 785. The same fragment patterns were observed in compound 4, yielding product ions at m/z 903, 885, 823, 805, and 787, which were 2.0 Da higher than those of the fragment ions of putative 44-methylgambierone. The same phenomenon was observed in other left-side moieties produced by the cleavage of rings I, H, and G ( Figure  4C), which supported our hypothesis that two additional hydrogens were located at the left side of ring G. Since both compounds were very similar in ion ratios and in-source two additional hydrogens were proposed to be located at C-12 and C-13 which ensured the intactness of the whole ladder-shaped backbone. Compound 4 was thus proposed to be a new gambierone analogue named 12,13-dihydro-44-methylgambierone. The proposed attributions of the ion formulas, along with mass differences, are presented in Table S3.

Identification and Characterization of New
Mar. Drugs 2022, 20, x FOR PEER REVIEW 8 of 18 fragmentation, two additional hydrogens were proposed to be located at C-12 and C-13 which ensured the intactness of the whole ladder-shaped backbone. Compound 4 was thus proposed to be a new gambierone analogue named 12,13-dihydro-44-methylgambierone. The proposed attributions of the ion formulas, along with mass differences, are presented in Table S3.    Figure 5). The precursor ion [M + H] + at m/z 1023, the neutral loss of SO3, and sequential losses of H2O were observed using the IDA mode, indicating that compound 3 was a gambierone analogue. Because of the low concentration of this compound, the MS 2 spectrum obtained via the IDA mode was not suitable for further fragmentation analysis. To solve this problem, the high-resolution multiple reaction monitoring (MRM HR ) acquisition method was applied to analyze compound 3. The MRM HR workflow is similar to the MRM of the QTRAP system, which uses quadrupole filters to   3 , and sequential losses of H 2 O were observed using the IDA mode, indicating that compound 3 was a gambierone analogue. Because of the low concentration of this compound, the MS 2 spectrum obtained via the IDA mode was not suitable for further fragmentation analysis. To solve this problem, the high-resolution multiple reaction monitoring (MRM HR ) acquisition method was applied to analyze compound 3. The MRM HR workflow is similar to the MRM of the QTRAP system, which uses quadrupole filters to reduce noise and increase selectivity, but the third quadrupole is replaced by TOF, and all fragments generated by a selected precursor can be analyzed in this system. The mass differences of precursor ions and product ions produced by the first SO 3 loss from the precursor ions between compounds 3 and 4 were 18 Da, indicating the presence of dehydroxylation in compound 3. The observation of fragment ions at m/z 123, 109, and 95 in the MS 2 spectra of compounds 3 and 4 indicated that both compounds had the same structure from C-40 to C-46. The product ions at m/z 815, 735, 695, 677, and 569 were formed by the cleavage of rings I and G, suggesting the same structure as compound 4 from C-1 to C-37. The product ion of compound 4 at m/z 903, generated by the cleavage of the C-C bond located at C-38 and C-39, was not observed in the MS 2 spectrum of compound 3, while the product ion at m/z 885 formed by the loss of H 2 O from m/z 903 in compound 4 was observed in compound 3, indicating that dehydroxylation occurred on C-38. The compound was thus named 38-dehydroxy-12,13-dihydro-44-methylgambierone. The proposed attributions of the ion formulas, along with mass differences, are presented in Table S4.
Mar. Drugs 2022, 20, x FOR PEER REVIEW 9 of 18 reduce noise and increase selectivity, but the third quadrupole is replaced by TOF, and all fragments generated by a selected precursor can be analyzed in this system. The mass differences of precursor ions and product ions produced by the first SO3 loss from the precursor ions between compounds 3 and 4 were 18 Da, indicating the presence of dehydroxylation in compound 3. The observation of fragment ions at m/z 123, 109, and 95 in the MS 2 spectra of compounds 3 and 4 indicated that both compounds had the same structure from C-40 to C-46. The product ions at m/z 815, 735, 695, 677, and 569 were formed by the cleavage of rings I and G, suggesting the same structure as compound 4 from C-1 to C-37. The product ion of compound 4 at m/z 903, generated by the cleavage of the C-C bond located at C-38 and C-39, was not observed in the MS 2 spectrum of compound 3, while the product ion at m/z 885 formed by the loss of H2O from m/z 903 in compound 4 was observed in compound 3, indicating that dehydroxylation occurred on C-38. The compound was thus named 38-dehydroxy-12,13-dihydro-44-methylgambierone. The proposed attributions of the ion formulas, along with mass differences, are presented in Table  S4.    Table S5.

Fragmentation Pathways of 38-Dehydroxy-44-Methylgambierone
Compound 6 ([M + H] + m/z 1021.4759, C52H77O18S + , Δ 6.5 ppm) was proposed to be 38dehydroxy-44-methylgambierone, according to its similar features to those of 44-methylgambierone, including retention time, ion ratios, adduct formation, the observation of SO3 (80 Da) loss, specific fragments (m/z 109.0645, Δ 2.8 ppm; m/z 233.1518, Δ 7.7 ppm), and the fragmentation pattern of sequential H2O losses ( Figure 6). The proposed precursor ion [M + H] + at m/z 1021 and product ion [M − SO3] + at m/z 941 produced by the first SO3 loss from the precursor ion were 18 Da lower than those of 44-methylgambierone, suggesting that a hydroxyl group was removed from this compound in comparison with 44-methylgambierone. The observation of fragment ions at m/z 123, 109, and 81 indicated that compound 6 had the same structure as 44-methylgambierone from C-40 to C-46. The produced ions at m/z 813 and 693 revealed that the hydroxyls on C-1, C-2, C-4, and C-5 remained unchanged in compound 6. Then, we deduced that the position of dehydroxylation occurred on C-38. The product ion of 44-methylgambierone at m/z 901 generated by the cleavage of the C-C bond located at C-38 and C-39 was not observed in the MS 2 spectrum of compound 6, while the product ion at m/z 883 formed by the loss of H2O from m/z 901 was observed in compound 6, supporting the hypothesis that dehydroxylation occurred on C-38. The compound was named 38-dehydroxy-44-methylgambierone. The proposed attributions of the ion formulas, along with mass differences, are displayed in Table S5.

Fragmentation Pathways of Desulfo-Hydroxyl Gambierone
Compound 2 ([M + NH4] + m/z 980.5568, C51H82NO17 + , Δ 1.5 ppm) has a similar structure to that of compound 1, which is the putative gambierone. The full-scan mass spectrum of compound 2 showed putative proton, ammonium, and sodium adduct precursor ions at m/z 963, 980, and 985, respectively, and the ion ratio was similar to compound 1 ( Figure 7A). Moreover, the MS 2 spectrum of compound 2 showed similar product ions to compound 1 and other gambierone analogues ( Figure 7B). The observation of the first neutral loss was inferred to be an ammonium loss (m/z 963.5321, Δ −0.9 ppm), and there were no fragments containing a sulfate group. The ammonium adduct precursor ion of compound 2 was 18 Da higher than the ammonium adduct ion at m/z 962 formed by SO3 loss from the precursor ion of compound 1. In addition, the product ion at m/z 945, the first H2O loss proton adduct [M − H2O + H] + , was consistent with the loss of the SO3 proton adduct of gambierone [M − SO3 + H] + . The above information suggests that this gambierone analogue was desulfo-gambierone with an additional hydroxyl group. The product ion at m/z 821 was formed by the cleavage of the C-C bond between C-38 and C-39, corresponding to the fragment ion of 44-methylgambierone yielded by the breakage of the same site plus the loss of SO3, which indicated the same structure of the left moiety in compound 2. The observation of fragment ions at m/z 219, 161, 109, and 81, possessing water loss, indicated that the additional hydroxyl group was located on one of the last six carbon atoms with no exact location determined. The compound was thus named as desulfo-hydroxyl gambierone. The fragmentation pathways and the proposed attributions of the ion formulas, along with mass differences, are presented in Figure 7C and Table S6, respectively.

Fragmentation Pathways of Desulfo-Hydroxyl Gambierone
Compound 2 ([M + NH 4 ] + m/z 980.5568, C 51 H 82 NO 17 + , ∆ 1.5 ppm) has a similar structure to that of compound 1, which is the putative gambierone. The full-scan mass spectrum of compound 2 showed putative proton, ammonium, and sodium adduct precursor ions at m/z 963, 980, and 985, respectively, and the ion ratio was similar to compound 1 ( Figure 7A). Moreover, the MS 2 spectrum of compound 2 showed similar product ions to compound 1 and other gambierone analogues ( Figure 7B). The observation of the first neutral loss was inferred to be an ammonium loss (m/z 963.5321, ∆ −0.9 ppm), and there were no fragments containing a sulfate group. The ammonium adduct precursor ion of compound 2 was 18 Da higher than the ammonium adduct ion at m/z 962 formed by SO 3 loss from the precursor ion of compound 1. In addition, the product ion at m/z 945, the first H 2 O loss proton adduct [M − H 2 O + H] + , was consistent with the loss of the SO 3 proton adduct of gambierone [M − SO 3 + H] + . The above information suggests that this gambierone analogue was desulfo-gambierone with an additional hydroxyl group. The product ion at m/z 821 was formed by the cleavage of the C-C bond between C-38 and C-39, corresponding to the fragment ion of 44-methylgambierone yielded by the breakage of the same site plus the loss of SO 3 , which indicated the same structure of the left moiety in compound 2. The observation of fragment ions at m/z 219, 161, 109, and 81, possessing water loss, indicated that the additional hydroxyl group was located on one of the last six carbon atoms with no exact location determined. The compound was thus named as desulfo-hydroxyl gambierone. The fragmentation pathways and the proposed attributions of the ion formulas, along with mass differences, are presented in Figure 7C and

Toxin Detection in Gambierdiscus balechii 1123M1M10
This study identified six gambierone analogues by comparing their MS 2 spectra with standards and the reported MS data. Their fragmentation pathways were proposed with high accuracy between the monoisotopic mass of the measured ions and the theoretical ones. NMR analyses need to be conducted for further structural elucidation. Herein, gambierone and 44-methylgambierone were quantifiable in the algal crude extracts. The limits of detection (LOD) and quantification (LOQ) of analytical methods were determined by using the signal-to-noise (S/N) ratios of 3:1 and 10:1. The LODs of gambierone and 44methylgambierone analyses were 0.09 ng/mL and 0.2 ng/mL, equating to 0.005 pg/cell and 0.01 pg/cell, and the LOQs were 0.3 ng/mL and 0.8 ng/mL, equating to 0.02 pg/cell and 0.05 pg/cell. The production level of gambierone (0.006 pg/cell) was below its LOQ and 44-methylgambierone was 4.0 pg/cell. 44-Methylgambierone was the predominant gambierone analogue in G. balechii 1123M1M10, accounting for around 40% of the total 6 gambierones with the comparison of their peak areas. 12,13-Dihydro-44-methylgambierone was the second most predominant gambierone analogue, accounting for around 30%, while the proportions of desulfo-hydroxyl gambierone, 38-dehydroxy-44-methylgambierone, 38-dehydroxy-44-methylgambierone, and gambierone were around 9%, 7%, 7%, and 2%, respectively. To date, six types of polyether toxins, including CTXs, MTXs, gambierones, gambierol, gambietic acid and gambieroxide, were found in different Gambierdiscus species, and it appears that not all Gambierdiscus species can produce CTXs, the most important CFP causative toxins (Table 1). Although G. balechii showed CTX-like activity based on the neuroblastoma cell (N2A) assay, no CTXs were detected via HPLC-HRMS analyses [46]. Gambierone, 44-methylgambierone, and other gambierone analogues were likely responsible for the CTX-like activity since these toxins showed weak activities on the sodium channel [8,41]. The bioactivities of gambierone analogues warrant further investigation.

Toxin Detection in Gambierdiscus balechii 1123M1M10
This study identified six gambierone analogues by comparing their MS 2 spectra with standards and the reported MS data. Their fragmentation pathways were proposed with high accuracy between the monoisotopic mass of the measured ions and the theoretical ones. NMR analyses need to be conducted for further structural elucidation. Herein, gambierone and 44-methylgambierone were quantifiable in the algal crude extracts. The limits of detection (LOD) and quantification (LOQ) of analytical methods were determined by using the signal-to-noise (S/N) ratios of 3:1 and 10:1. The LODs of gambierone and 44-methylgambierone analyses were 0.09 ng/mL and 0.2 ng/mL, equating to 0.005 pg/cell and 0.01 pg/cell, and the LOQs were 0.3 ng/mL and 0.8 ng/mL, equating to 0.02 pg/cell and 0.05 pg/cell. The production level of gambierone (0.006 pg/cell) was below its LOQ and 44-methylgambierone was 4.0 pg/cell. 44-Methylgambierone was the predominant gambierone analogue in G. balechii 1123M1M10, accounting for around 40% of the total 6 gambierones with the comparison of their peak areas. 12,13-Dihydro-44-methylgambierone was the second most predominant gambierone analogue, accounting for around 30%, while the proportions of desulfo-hydroxyl gambierone, 38-dehydroxy-44-methylgambierone, 38-dehydroxy-44-methylgambierone, and gambierone were around 9%, 7%, 7%, and 2%, respectively. To date, six types of polyether toxins, including CTXs, MTXs, gambierones, gambierol, gambietic acid and gambieroxide, were found in different Gambierdiscus species, and it appears that not all Gambierdiscus species can produce CTXs, the most important CFP causative toxins (Table 1). Although G. balechii showed CTX-like activity based on the neuroblastoma cell (N2A) assay, no CTXs were detected via HPLC-HRMS analyses [46]. Gambierone, 44-methylgambierone, and other gambierone analogues were likely responsible for the CTX-like activity since these toxins showed weak activities on the sodium channel [8,41]. The bioactivities of gambierone analogues warrant further investigation.

Species Toxins
Gambierdiscus toxicus [4,10,11,15,29,30,34,37,38] the colored bands. Fractions one and two comprised 40 mL and 10 mL and were collected at a flow rate of 0.3 mL/min. The fractions 3 to 5, 12, 16, and 17 comprised 10 mL and were collected at a flow rate of 0.8 mL/min. The fractions 6 to 11 and 13-15 comprised 5 mL and were collected at a flow rate of 0.8 mL/min. The last fraction, 18, comprised 100 mL and was collected at a flow rate of 0.8 mL/min. All fractions were dried via evaporation under a gentle stream of high-purity nitrogen and redissolved in 50-200 µL 50%-100% methanol. Gambierone and 44-methylgambierone standards were purchased from Laboratorio CIFGA S.A. (Lugo, Spain).

HPLC-QTOF-HRMS Analysis
The non-target analysis of toxins was performed using HPLC-HRMS consisting of an Agilent 1290 UPLC system (Agilent, Palo Alto, CA, USA) and a Sciex X500R QTOF system (AB Sciex, Foster City, CA, USA) operating with the IDA method. A 10 µL aliquot was injected into a Phenomenex Kinetex C18 column (100 × 2.1 mm i.d., 1.7 µm). Gradient elution at a flow rate of 0.2 mL/min was performed using: (A) Mill-Q water containing 0.02% formic acid (Merck, Darmstadt, Germany) and 2 mM ammonium acetate (Sigma Aldrich, MO, USA), and (B) 95% acetonitrile containing 0.02% formic acid and 2 mM ammonium acetate for the IDA method with a positive ESI mode; (A) Mill-Q water containing 2 mM ammonium acetate, and (B) 95% acetonitrile containing 2 mM ammonium acetate for the IDA method with a negative ESI mode. The gradient elution procedure was performed as follows: the initial gradient condition of the metabolite separation started at 30% B and increased to 60% B within 5 min. Then, it increased to 90% B within 30min and to 100% B within 40 min, and was kept for 5 min before returning to 30% B for 1 min. The column was equilibrated at initial gradient conditions for 4 min before the next injection.
IDA methods were performed as follows: mass spectrometry was conducted in IDA mode with a TOF-MS full-scan analysis (0.25 s) and up to 10 dependent MS/MS analyses (0.1 s for each MS/MS analysis) per cycle. The TOF-MS full scan was operated with the mass range of m/z from 100 to 2000, and the dependent MS/MS was operated with the mass range of m/z from 50 to 2000 under the high-resolution mode. Two ions at m/z 132.90490 and 829.53933 were used for full mass calibration and their resolutions were above 21,000 and 29,000. Four ions at m/z 185.12845, 298.21252, 494.33370, and 607.41776 were used for MS 2 calibration and their resolutions were above 21,000, 22,000, 25,000, and 26,000, respectively. Dynamic background subtraction was applied in the IDA criteria for dynamic exclusion. The fragment ions were generated from collision-induced dissociation with nitrogen under standardized collision energy (CE) = 35 V with collision energy spread (CES) = 15 V for the positive ESI mode, and CE = −35 V with CES = 0 V for the negative ESI mode. The other experimental parameters included: nebulizer gas (gas 1), 30 psi; heater gas (gas 2), 40 psi; curtain gas, 25 psi; ion source temperature, 500 • C; ion spray voltage floating, 5500 V for positive ESI mode and −4500 V for negative ESI mode; declustering potential, 80 V for positive ESI mode and −80 V for negative ESI mode; and full MS collision energy, 10 V for positive ESI mode, and −10 V for negative ESI mode. The gas used was nitrogen.
MRM HR methods were performed in the positive ESI mode as follows: mass spectrometry was conducted in the MRM HR mode with a TOF-MS full scan analysis (0.1 s) and TOF MS/MS analyses (0.1 s for each selected precursor ion) per cycle. The TOF-MS full scan was operated with the mass range of m/z from 100 to 2000, and the MS/MS was operated with the mass range of m/z from 50 to 2000 under the high-resolution mode. The fragment ions were generated from collision-induced dissociation with nitrogen under standardized collision energy (CE) = 40 V with collision energy spread (CES) = 15 V. Other experimental parameters included: nebulizer gas (gas 1), 30 psi; heater gas (gas 2), 40 psi; curtain gas, 25 psi; ion source temperature, 500 • C; ion spray voltage floating, 5500 V; declustering potential, 80 V; and full MS collision energy, 10 V. The gas used was nitrogen.

HPLC-QTRAP-HRMS Analysis
The quantification of toxins was performed using HPLC-MS/MS consisting of an Agilent 1290 UPLC system (Agilent, Palo Alto, CA, USA) and a Sciex 5500 QTRAP mass spectrometer (Foster City, CA, USA) operating in a multiple reaction monitoring (MRM) negative ESI mode. A 5 µL aliquot was injected into a Phenomenex Kinetex C18 column (100 × 2.1 mm i.d., 1.7 µm). Gradient elution at a flow rate of 0.2 mL/min was performed using (A) Mill-Q water containing 0.02% formic acid and 2 mM ammonium acetate and (B) 95% acetonitrile containing 0.02% formic acid and 2 mM ammonium acetate. The gradient elution procedure was performed as follows: The initial gradient condition started at 30% B and increased to 100% B at 10 min. Then, it was kept for 0.9 min before returning to 30% B for 0.2 min, with a total run time of 12 min. Further details of the mass parameters of gambierone and 44-methylgambierone are given in Table S7.

Conclusions
Four new and two known gambierones were characterized from G. balechii 1123M1M10 via high-performance liquid chromatography coupled with electrospray ionization tandem mass spectrometry (HPLC/ESI-MS 2 ) analysis, and their fragmentation behaviors and pathways were proposed. 44-Methylgambierone and 12,13-dihydro-44-methylgambierone were the major toxins in this strain. Further studies on the isolation and evaluating bioactivity of these gambierones need to be conducted to investigate their role in CFP.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/md21010003/s1. Figure S1: Original extraction ion chromatograms (XICs) of putative gambierone analogues in fraction three, and XICs of standards of gambierone (5.92 min) at 345 ng/mL and 44-methylgambierone (6.32 min) at 315 ng/mL using the IDA method in the positive ESI mode; Figure S2 Table S1: The proposed attributions of ion formulas along with mass differences (∆ ppm) of compound 1; Table S2: The proposed attributions of ion formulas along with mass differences (∆ ppm) of compound 5; Table S3: The proposed attributions of ion formulas along with mass differences (∆ ppm) of compound 4; Table S4: The proposed attributions of ion formulas along with mass differences (∆ ppm) of compound 3; Table S5: The proposed attributions of ion formulas along with mass differences (∆ ppm) of compound 6; Table S6: The proposed attributions of ion formulas along with mass differences (∆ ppm) of compound 2; Table S7