First Identification of 12β-Deoxygonyautoxin 5 (12α-Gonyautoxinol 5) in the Cyanobacterium Dolichospermum circinale (TA04) and 12β-Deoxysaxitoxin (12α-Saxitoxinol) in D. circinale (TA04) and the Dinoflagellate Alexandrium pacificum (Group IV) (120518KureAC)

Saxitoxin and its analogues, paralytic shellfish toxins (PSTs), are potent and specific voltage-gated sodium channel blockers. These toxins are produced by some species of freshwater cyanobacteria and marine dinoflagellates. We previously identified several biosynthetic intermediates of PSTs, as well as new analogues, from such organisms and proposed the biosynthetic and metabolic pathways of PSTs. In this study, 12β-deoxygonyautoxin 5 (12α-gonyautoxinol 5 = gonyautoxin 5-12(R)-ol) was identified in the freshwater cyanobacterium, Dolichospermum circinale (TA04), and 12β-deoxysaxitoxin (12α-saxitoxinol = saxitoxin-12(R)-ol) was identified in the same cyanobacterium and in the marine dinoflagellate Alexandrium pacificum (Group IV) (120518KureAC) for the first time from natural sources. The authentic standards of these compounds and 12α-deoxygonyautoxin 5 (12β-gonyautoxinol 5 = gonyautoxin 5-12(S)-ol) were prepared by chemical derivatization from the major PSTs, C1/C2, produced in D. circinale (TA04). These standards were used to identify the deoxy analogues by comparing the retention times and MS/MS spectra using high-resolution LC-MS/MS. Biosynthetic or metabolic pathways for these analogues have also been proposed based on their structures. The identification of these compounds supports the α-oriented stereoselective oxidation at C12 in the biosynthetic pathway towards PSTs.


Introduction
Saxitoxin (STX, 1, Figure 1) is one of the most potent voltage-gated sodium channel (Na v ) blockers [1]. More than 50 natural STX analogues, which are known as paralytic shellfish toxins (PST) [2,3], have been reported [4]. These toxins are produced by some species of freshwater cyanobacteria and marine dinoflagellates [5,6]. Due to their potent physiological effects and unique structures, chemical and biological aspects of STX and its analogues have been studied intensively by chemists and pharmacologists [7,8]. The first biosynthetic study of STX was a feeding experiment conducted by Shimizu et al. [9], in which stable isotope-labeled acetic acid and amino acids were used as essential substrates for PST-producing cyanobacteria and dinoflagellates. Next, Neilan's group [10] Figure 1. The structures of natural analogues of saxitoxin (STX, 1). 12β-deoxyGTX5 (12α-GTXol 5) (2) and 12β-deoxySTX (12α-STXol) (3) are the analogues identified in natural sources for the first time in this study.

Screening for New PSTs in the PST-Producing Cyanobacterium D. circinale (TA04) and the Dinoflagellate A. pacificum (Group IV) (120518KureAC)
Screening for new STX analogues was performed on the semi-purified extracts from the cyanobacterium D. circinale (TA04) and the dinoflagellate A. pacificum (Group IV) (120518KureAC) (see Sections 4.5 and 4.6 for a description of the preparation of the semipurified cell extracts) using high resolution (HR)-LCMS, with both hydrophilic interaction liquid chromatography (HILIC) [27] and reversed phase (RP) separations in the positive mode. The screening using RP separation resulted in the identification of the unknown peaks in each organism. In D. circinale (TA04), an unknown peak detected at 13  Screening for new STX analogues was performed on the semi-purified extracts from the cyanobacterium D. circinale (TA04) and the dinoflagellate A. pacificum (Group IV) (120518KureAC) (see Sections 4.5 and 4.6 for a description of the preparation of the semipurified cell extracts) using high resolution (HR)-LCMS, with both hydrophilic interaction liquid chromatography (HILIC) [27] and reversed phase (RP) separations in the positive mode. The screening using RP separation resulted in the identification of the unknown peaks in each organism. In D. circinale (TA04), an unknown peak detected at 13 cyanobacteria as described above (LWTX1, LWTX4 (4), LWTX6, and 12β-deoxyGTX3 in Figure 1), this compound was supposed to be 12-deoxyGTX5.
Another unknown peak was detected in D. circinale (TA04) at 12.9 min and also in A. pacificum (Group IV) (120518KureAC) at 12
NOESY1D experiments for 12α-deoxyGTX5 (5) were also conducted as described above, and no significant NOE was detected between H5 and H12. Energy-minimized molecular models of 2 and 5, which were calculated using Spartan'18 (Wavefunction, Irvine, USA), with molecular mechanics, Merck Molecular Forcefield (MMFF), estimated that the distances between H5 and H12 in 2 and 5 were 2.56 and 3.04 Å, respectively ( Figure 4). These data all supported stereochemical assignment at C12 in synthetic 2, 3, and 5.

Identification of 12β-deoxySTX (3)(12α-STXol) in A. pacificum (Group IV) (120518KureAC) and D. circinale (TA04)
The peaks of synthetic 12β-deoxySTX (3) and 12α-deoxySTX standards were observed at 12.6 min and 14.1 min, respectively, on the EIC at m/z 284.1466 ± 0.02 of RP-LCMS ( Figure 7A). The retention times of the unknown peak detected at 12.8 min in A. pacificum (Group IV) (120518KureAC) ( Figure 7B) and at 12.9 min in D. circinale (TA04) ( Figure 7C) were close to that of synthetic 3 (12.6 min), suggesting that these peaks correspond to 12β-deoxySTX ( Figure 8B) were almost identical with those detected for synthetic 3 ( Figure 8A) (the differences were within 0.002 Da), supporting identification of 3 in this species. The potential 12β-deoxySTX (3) peak detected in semi-purified D. circinale (TA04) extract has not been confirmed using the MS/MS spectrum because of the low intensity of this peak. The peak detected at 13.3 min in D. circinale (TA04) extract ( Figure 7C) was interpreted as the desulfated 12β-deoxyGTX5 (2) ion ([M-SO3+H] + ) because of the close retention time to that of 2 (13.2 min, Figure 5C).  The peaks of synthetic 12β-deoxySTX (3) and 12α-deoxySTX standards were observed at 12.6 min and 14.1 min, respectively, on the EIC at m/z 284.1466 ± 0.02 of RP-LCMS ( Figure 7A). The retention times of the unknown peak detected at 12.8 min in A. pacificum (Group IV) (120518KureAC) ( Figure 7B) and at 12.9 min in D. circinale (TA04) ( Figure 7C) were close to that of synthetic 3 (12.6 min), suggesting that these peaks correspond to 12β-deoxySTX (3). The HR-MS detected at m/z 284.1443 in D. circinale (TA04) and at m/z 284.1463 in A. pacificum (Group IV) (120518KureAC) (calcd. m/z 284.1466) for these peaks (Figures S12 and S13) supported identification. The presence of 12β-deoxySTX (3) in A. pacificum was further confirmed by comparing the LC-MS/MS spectrum with that of authentic 3 (Figure 8). The HR-MS of the major fragment ions detected in the MS/MS spectrum of 3 in A. pacificum ( Figure 8B) were almost identical with those detected for synthetic 3 ( Figure 8A) (the differences were within 0.002 Da), supporting identification of 3 in this species. The potential 12β-deoxySTX (3) peak detected in semi-purified D. circinale (TA04) extract has not been confirmed using the MS/MS spectrum because of the low intensity of this peak. The peak detected at 13.3 min in D. circinale (TA04) extract ( Figure 7C) was interpreted as the desulfated 12β-deoxyGTX5 (2) ion ([M-SO3+H] + ) because of the close retention time to that of 2 (13.2 min, Figure 5C).
Lukowski et al. [18] functionally expressed GxtA, Rieske oxygenase which is involved in β-hydroxylation at C11, and SxtSUL, O-sulfotransferase of 11β-hydroxyl STX, and SxtN, which is functionalized as N-sulfotransferase of the carbamoyl group in STX. SxtSUL was obtained from the cyanobacterium Microseira wollei, and SxtN came from Aphanizomenon sp. NH-5. Since we identified 12β-deoxyGTX5 (2) and 12β-deoxySTX (3) in D. circinale (TA04) in this study, a homologous enzyme with SxtN in D. circinale (TA04) should have a similar functionality to that of M. wollei. A possible biosynthetic route from 4 to 2, 3, and 12β-deoxyGTX3 which was previously found in D. circinale (TA04), is proposed in Figure 9 although the predicted compound 5 (11β-hydroxy-12β-deoxySTX) has not yet been identified. GxtA, SxtSUL, and SxtN were predicted to be involved in these reactions. The presence of 3 (in this study) and 4 in Dolichospermum circinale (TA04) [26] and Alexandrium pacificum (120518KureAC) [27] suggested the functional similarity of the enzyme that catalyzes O-carbamoylation of 4 in these organisms. Further biochemical and analytical studies are needed to compare similar biosynthetic reactions between cyanobacteria and dinoflagellates.
Mar. Drugs 2022, 20, 166 9 of 13 reactions. The presence of 3 (in this study) and 4 in Dolichospermum circinale (TA04) [26] and Alexandrium pacificum (120518KureAC) [27] suggested the functional similarity of the enzyme that catalyzes O-carbamoylation of 4 in these organisms. Further biochemical and analytical studies are needed to compare similar biosynthetic reactions between cyanobacteria and dinoflagellates. Figure 9. Possible biosynthetic routes from 4 to 2, 3, and 12β-deoxyGTX3 [29]. The predicted intermediate 6 has not been identified in a natural source, including PST-producing organisms.

General Information
The reagents were purchased from Merck KGaA (

Harvest and Preparation of D. circinale (TA04) Cell Extract
The toxic strain of the freshwater cyanobacterium D. circinale used in this study is a nonaxenic strain TA04. The field sample of D. circinale was collected at the Tullaroop reservoir, Victoria, Australia, and the TA04 strain was one of the single-trichome isolates prepared by Negri et al. in 1993 [32]. D. circinale (TA04) was provided by Dr. Susan Blackburn, CSIRO, Australia, and cultured in CB' medium (Bicine instead of Tris (hydroxymethyl) aminomethane was added to C medium, and adjust pH to 9.0) (175 mL) in 250 mL plastic tissue culture flasks under the following culture conditions: 16 h light/8 h dark photo-cycle with light provided by LED light (30 µmole photons m −2 s −1 ) at 17 • C for 28 days. The cells were harvested by centrifugation at 4820× g for 15 min at 4 • C, suspended with 2.0 mL of 0.5 M AcOH, and sonicated three times for 30 s on ice. Then, the resulting solution was centrifuged at 4160× g for 15 min at 4 • C. The supernatant was adjusted to pH 7-8 using 2 M NH 3 aq and loaded on an activated charcoal column (5 mL, FUJIFILM Wako Pure Chemical Industries, Ltd., Osaka, Japan). After the column was washed with water (15 mL), PSTs were eluted with AcOH-EtOH-H 2 O (5:50:45, v/v/v, 30 mL). The solvent was removed using a rotary evaporator under vaccum, and the resulting residue was resuspended with 1 mL of 0.05 M AcOH. This solution was filtered through a Cosmospin filter H (0.45 µm, Nacalai Tesque, Inc., Kyoto, Japan). A part of the filtrate was diluted depending on the concentrations of PSTs, and the diluted solution was applied to LCMS.

Harvest and Preparation of Alexandrium pacificum (Group IV) (120518KureAC) Cell Extract
A. pacificum (Group IV) (120518KureAC) was originally isolated at Kure, Hiroshima, Japan, by Dr. Kazuhiko Koike of Hiroshima University in 2012. It was maintained and grown in modified T 1 medium prepared in artificial seawater as 200-mL cultures in 250 mL plastic tissue culture flasks under the following culture conditions: 12 h light/12 h dark photo-cycle with light provided by cool white bulbs (100-150 µmole photons m −2 s −1 ) at 15 • C. Aliquots of cultured cells (18 mL: 2.2 × 10 4 cells mL −1 ) was centrifuged at 810× g for 3 min to pellet the cells. The pellets were re-suspended with 300 µL of 0.5 M acetic acid, and the cells were disrupted by sonication (3 times at 100 Hz for 30 s each, with an interval of 30 s) on ice. The suspension was centrifuged (20,000× g for 5 min at 4 • C), and the supernatant was subjected to ultra-filtration (10,000 Da cut-off, UF-MC). An aliquot (100 µL) of the extract after filtration was transferred to a new tube and mixed with three volumes of THF. The sample was loaded onto Chromabond R HILIC (500 mg, MACHEREY-NAGEL, Düren, Germany) that had been pre-conditioned with 1 mL of MilliQ water and 5 mL of THF. Following loading of the sample, the column was sequentially washed with 3 mL of THF, 3 mL of CH 3 CN, and 3 mL of CH 3 CN-H 2 O-HCOOH (95:5:0.1, v/v/v) [33]. The column was eluted with 3 mL of 0.2 M HCOOH and concentrated under nitrogen stream. The volume was adjusted to 100 µL with MilliQ water and passed through a Cosmospin filter H (0.45 µm).

HR-RP-LCMS and HR-RP-LC-MS/MS Conditions for PSTs Analysis
HR-RP-LCMS was performed on an Inertsustain AQ-C18 column (4.6 i.d. × 250 mm, 5 µm, GL Sciences, Japan) with the mobile phase, HCOOH-H 2 O 0.1:100 (v/v). The flow rate was 0.2 mL/min. The injected volume was 1 µL. The oven temperature was 25 • C. HR-LCMS were recorded on a micrOTOF-Q II mass spectrometer (Bruker Daltonics) equipped with an ESI ion source. The liquid chromatography system used for analysis was a Shimadzu Nexera UHPLC System (Shimadzu). The mass spectrometer conditions were as follows: positive ionization mode; dry gas: nitrogen 7 L/min; dry heater temperature: 180 • C; nebulizer: 1.6 Bar; and capillary: 4500 V. Extracted ion chromatograms (EIC) were presented based on ± 0.02. HR-LC-MS/MS was performed in AutoMS/MS mode setting, with [M + H] + as the precursor ions. The precursor ions were m/z 364.10 for 12β-deoxyGTX5 (2) and m/z 284.11 for 12β-deoxySTX (3) setting the width 3 Da. The sweeping collision energy was 40-60 eV for 2 and 34-52 eV for 3.