Isolation and Biological Activity of Iezoside and Iezoside B, SERCA Inhibitors from Floridian Marine Cyanobacteria

Marine cyanobacteria are a rich source of bioactive natural products. Here, we report the isolation and structure elucidation of the previously reported iezoside (1) and its C-31 O-demethyl analogue, iezoside B (2), from a cyanobacterial assemblage collected at Loggerhead Key in the Dry Tortugas, Florida. The two compounds have a unique skeleton comprised of a peptide, a polyketide and a modified sugar unit. The compounds were tested for cytotoxicity and effects on intracellular calcium. Both compounds exhibited cytotoxic activity with an IC50 of 1.5 and 3.0 μΜ, respectively, against A549 lung carcinoma epithelial cells and 1.0 and 2.4 μΜ against HeLa cervical cancer cells, respectively. In the same cell lines, compounds 1 and 2 show an increase in cytosolic calcium with approximate EC50 values of 0.3 and 0.6 μΜ in A549 cells and 0.1 and 0.5 μΜ, respectively, in HeLa cells, near the IC50 for cell viability, suggesting that the increase in cytosolic calcium is functionally related to the cytotoxicity of the compounds and consistent with their activity as SERCA (sarcoplasmic/endoplasmic reticulum Ca2+-ATPase) inhibitors. The structure–activity relationship provides evidence that structural changes in the sugar unit may be tolerated, and the activity is tunable. This finding has implications for future analogue synthesis and target interaction studies.


Introduction
Calcium ions (Ca 2+ ) regulate various cellular processes by activating gene transcription, cell proliferation and migration [1,2]. The cytosolic levels of calcium are maintained at sub-micromolar concentrations, while the extracellular levels are at high millimolar concentrations. Most of the intracellular Ca 2+ is stored in the endoplasmic reticulum (ER), which is enriched in calcium-binding proteins [1]. Calcium is exchanged between the cytosol and the ER in three different ways: channels, which allow transport from the ER to the cytosol; exchangers of Na + /Ca 2+ , which use the Na + gradient to transport Ca 2+ ; and pumps (sarcoplasmic/endoplasmic reticulum Ca 2+ -ATPase, SERCA), which transport Ca 2+ against the concentration gradient into the ER [2,3]. In recent years, it has been reported that cancer cells have altered expression of SERCA pumps, making them a potential target. Inhibition of the SERCA pump leads to high levels of Ca 2+ in the cytosol, which is toxic and, if prolonged, results in growth arrest and apoptosis of the cells [4,5].
Iezoside (1) (Figure 1) was previously isolated from a filamentous marine cyanobacterium collected from Ie Island, Okinawa, Japan, that was identified based on morphology as Leptochromothrix valpauliae. Molecular phylogenetic analysis indicated that while the cyanobacterium was related to Leptochromothrix spp., it was phylogenetically distinct and likely represented a new species ( Figure S1 in reference 13) [13]. Iezoside consists of three moieties: a polyketide unit, a peptide unit and a rhamnose sugar. The planar structure, double bond geometry and relative configuration of the sugar unit were determined by NMR. The absolute configuration of the molecule was determined by amino acid analysis, Mosher's method for the sugar moiety and circular dichroism for the polyketide unit. The uncertain C18-C19 configuration required the total synthesis (16 steps LLS, 4.4% total yield) to conclusively assign the complete 3D structure. Using computational methods and electronic circular dichroism (ECD), the compound was limited to two potential isomers, 18R,19R or 18S,19R. The total synthesis indicated that 1 had the 18R,19R configuration. Iezoside (1) was identified as a SERCA inhibitor that promotes cell cycle delay and induces apoptosis pathways, with an IC50 for cell viability of 6.8 nM against HeLa cells and a Ki of 7.1 nM on SERCA1a [13].
We previously reported the discovery of lagunamide D (3) (Figure 1), a 26-membered macrocycle from a collection of marine cyanobacteria comprised of mainly Dichothrix sp. and Lyngbya sp. Compound 3 was shown to be a mitochondrial cytotoxin, with potent antiproliferative activity and an IC50 of 7.1 nM against A549 lung adenocarcinoma cells [23,24]. The same extract and fraction that yielded 3 was further investigated based on its NMR profile. Here, we report the isolation of iezoside (1) and a new, O-demethyl analogue, iezoside B (2) (Figure 1) [13]. Both compounds were isolated in sub-milligram quantities. Iezoside (1) was four-fold more abundant than iezoside B (2), and logically, we first determined the structure of the major compound, which provided clues about the structural differences of the new minor analogue. Our structure determination of iezoside B (2) relied on the NMR comparison in DMSO-d6, which required us to first conclusively demonstrate that the structures (except the methylation pattern) are identical. Since iezoside (1) from Japan was shown to be a cytotoxic and potent SERCA inhibitor, 2 was tested for its cytotoxicity and its effect on intracellular calcium. We compared the potencies of 1 and 2 side-by-side to interrogate the effect of the methylation pattern in the sugar moiety on its activity.

Isolation and Structure Elucidation
The cyanobacterial assemblage was lyophilized and extracted using 1:1 EtOAc/MeOH followed by 1:1 EtOH:H2O and further partitioned between EtOAc and H2O, followed by partitioning the polar layer between n-BuOH and H2O. The EtOAc

Isolation and Structure Elucidation
The cyanobacterial assemblage was lyophilized and extracted using 1:1 EtOAc/MeOH followed by 1:1 EtOH:H 2 O and further partitioned between EtOAc and H 2 O, followed by partitioning the polar layer between n-BuOH and H 2 O. The EtOAc soluble portion was further subjected to fractionation using reversed-phase chromatography to yield 0.8 mg of iezoside (1) and 0.2 mg of iezoside B (2).  (1) from the Okinawan cyanobacterium were published in acetone-d 6 after we had completed our analysis [13]. Briefly, a tripeptide unit was determined by COSY and HMBC correlations as Thz-Leu-N-Me-Ala (Table 1, Figure 2). The broadening of the H-10 (δ H 4.83), H-11 (δ H 1.29) and H-12 (δ H 2.78) as well as the lack of HSQC signals were observed in DMSO-d 6 , similarly to the published 1 [13]. The broadening was attributed to the slow conformational exchange of the N-Me-Ala residue. The polyketide unit was established using the COSY and HMBC correlations in Figure 2 and Table 1, similarly as described [13]. H-21 (δ H 6.14) had to be part of a trans double bond (J = 15. soluble portion was further subjected to fractionation using reversed-phase chromatography to yield 0.8 mg of iezoside (1) and 0.2 mg of iezoside B (2). Iezoside (1) was isolated as a colorless oil with a molecular formula of C37H59N3O7S ([M+H] + 690.4104, calcd. for C37H60N3O7S, 690.4152) as determined by HRESIMS. The structure was established in three parts by NMR analysis in DMSO-d6, while data for iezoside (1) from the Okinawan cyanobacterium were published in acetone-d6 after we had completed our analysis [13]. Briefly, a tripeptide unit was determined by COSY and HMBC correlations as Thz-Leu-N-Me-Ala (Table 1, Figure 2). The broadening of the H-10 (δH 4.83), H-11 (δH 1.29) and H-12 (δH 2.78) as well as the lack of HSQC signals were observed in DMSO-d6, similarly to the published 1 [13]. The broadening was attributed to the slow conformational exchange of the N-Me-Ala residue. The polyketide unit was established using the COSY and HMBC correlations in Figure 2 and Table 1, similarly as described [13]. H-21 (δH 6.14) had to be part of a trans double bond (J = 15.7 Hz) with H-20 (δH 5.29). H-16 (δH 6.09) was part of a conjugated diene with s-trans configuration (J = 11.2) with H-15 (δH 6.32), which also showed HMBC correlation to C-26 (δC 14.  Iezoside B (2) was isolated from the same fraction as iezoside (1) and had a molecular formula of C36H57N3O7S as determined by HRESIMS ([M+H] + 676.3990 calcd. 676.3954) with 10 degrees of unsaturation. Compound 2 differs by 14 mass units (CH2) from 1, which is attributed to the lack of the methoxy group at C-31 (Figures 1-3), as detailed below. Iezoside B (2) was isolated from the same fraction as iezoside (1) and had a molecular formula of C 36 H 57 N 3 O 7 S as determined by HRESIMS ([M+H] + 676.3990 calcd. 676.3954) with 10 degrees of unsaturation. Compound 2 differs by 14 mass units (CH 2 ) from 1, which is attributed to the lack of the methoxy group at C-31 (Figures 1-3), as detailed below.
The structure of 2 was established based on 1D and 2D NMR analysis in DMSO-d 6 (Table 1, Figure 3). The thiazole ring was established based on characteristic 1 H and 13 C NMR chemical shifts (Table 1)  The structure of 2 was established based on 1D and 2D NMR analysis in DMSO-d6 (Table 1, Figure 3). The thiazole ring was established based on characteristic 1 H and 13 C NMR chemical shifts (Table 1), coupled with COSY and HMBC data. H-1 (δH 7.70) showed a COSY correlation to H-2 (δH 7.58) and an HMBC correlation to C-3 (δC 173.6), supporting the thiazole structure. The HMBC correlation of H-4 (δH 5.19) to C-3 extended the structure by a leucine unit, based on the COSY correlations of H-4 to 4-NH (δH 8.56) and to H-5 (δH 1.77) and of H-5 to H-6 (δH 1.65). H-6 showed COSY correlations to H-7 (δH 0.91) and H-8 (δH 0.87).
The adjacent N-Me alanine unit was connected based on the HMBC correlation of H-11 (δH 1.28) and 4-NH to C-9 (δC 170.6). Similarly, as observed and reported for 1, broadening of the proton signals occurred, and the unit also did not show correlations in the HSQC spectrum [13].
The polyketide unit (C13-C29) was established as three spin systems with three trisubstituted double bonds and one disubstituted double bond that could be connected due to HMBC correlations. The first unit (C22-C25) was established from the COSY correlations of H-25 (δH 0.93) to H-24 (δH 2.07) and of H-24 to H-23 (δH 5.48), which belonged to a trisubstituted double bond. The HMBC from H-23 to C-29 (δC 12.1) and of H-29 (δH 1.63) to the non-protonated C-22 (δC 131.8) established that C-22 was methylated.
The remaining unit was determined as a 3-O-methyl-methylpentose based on the COSY correlations of H-30 (δH 4.55) to H-31 (δH 3.75) and of that to H-32 (δH 3.12), which in turn had a COSY correlation to H-33 (δH 3.28), and that to H-34 (δH 3.48), which finally The adjacent N-Me alanine unit was connected based on the HMBC correlation of H-11 (δ H 1.28) and 4-NH to C-9 (δ C 170.6). Similarly, as observed and reported for 1, broadening of the proton signals occurred, and the unit also did not show correlations in the HSQC spectrum [13].
The polyketide unit (C13-C29) was established as three spin systems with three trisubstituted double bonds and one disubstituted double bond that could be connected due to HMBC correlations. , which was connected to the methylated C22-C25 unit as shown by HMBC correlation of H-29 to C-21 (δ C 138.0). E-configuration of the C20-C21 double bond was confirmed by proton-proton coupling constant (J = 15.7 Hz).
The remaining unit was determined as a 3-O-methyl-methylpentose based on the COSY correlations of H-30 (δ H 4.55) to H-31 (δ H 3.75) and of that to H-32 (δ H 3.12), which in turn had a COSY correlation to H-33 (δ H 3.28), and that to H-34 (δ H 3.48), which finally had a COSY correlation to H-35 (δ H 1.13). The HMBC correlation of H-37 (δ H 3.31) to C-32 (δ C 80.7) guided the positioning of the O-Me, while the HMBC correlation of H-30 to C-19 (δ C 78.0) completed the planar structure. The configuration of C-20/C-21 and C-15/C-16 for both 1 and 2 was determined as trans and s-trans, respectively, based on the J coupling constants (15.7 Hz and 11.2 Hz, respectively). The absolute configuration of C-18/C-19 was determined by comparison of its circular dichroism (CD) spectra to the literature, as shown in Figure 4 [13]. Compound 2 shows a similar Cotton effect as 1. Iezoside (1) shows a negative Cotton effect at 258 nm (∆ε −3.6) and a positive at 234 nm (∆ε +5.9), while iezoside B (2) shows a negative Cotton effect at 272 nm (∆ε −1.2) and a positive at 232 nm (∆ε +8.9). According to the published data for 1, the C18/C19 configuration would be either 18R,19R or 18S,19R [13]. The NMR data for 1 was initially collected in DMSO-d 6 and later compared and matched to the literature values for iezoside (1) in acetone-d 6 as the planar structures and relative configuration of the compounds were identical ( Figure S8, Table S1) [13]. By comparison of the 1 H NMR spectra, the isolated compound matched the natural iezoside and not the 18S,19R-unnatural analogue; therefore, the absolute configuration was determined as 18R,19R [13]. The key differences between the two diastereomers were the H-27 and H-29 shifts as well as the upfield shift of the H-28 from δ H 1.12 to δ H 0.96. Coupling constants and NOESY data supported the relative configuration of the sugar moiety as rhamnose (Figures 2 and 3 The configuration of C-20/C-21 and C-15/C-16 for both 1 and 2 was determined as trans and s-trans, respectively, based on the J coupling constants (15.7 Hz and 11.2 Hz, respectively). The absolute configuration of C-18/C-19 was determined by comparison of its circular dichroism (CD) spectra to the literature, as shown in Figure 4 [13]. Compound 2 shows a similar Cotton effect as 1. Iezoside (1) shows a negative Cotton effect at 258 nm (Δε −3.6) and a positive at 234 nm (Δε +5.9), while iezoside B (2) shows a negative Cotton effect at 272 nm (Δε −1.2) and a positive at 232 nm (Δε +8.9). According to the published data for 1, the C18/C19 configuration would be either 18R,19R or 18S,19R [13]. The NMR data for 1 was initially collected in DMSO-d6 and later compared and matched to the literature values for iezoside (1) in acetone-d6 as the planar structures and relative configuration of the compounds were identical ( Figure S8, Table S1) [13]. By comparison of the 1 H NMR spectra, the isolated compound matched the natural iezoside and not the 18S,19Runnatural analogue; therefore, the absolute configuration was determined as 18R,19R [13]. The key differences between the two diastereomers were the H-27 and H-29 shifts as well as the upfield shift of the H-28 from δH 1.12 to δH 0.96. Coupling constants and NOESY data supported the relative configuration of the sugar moiety as rhamnose (Figures 2 and  3 [13]. Finally, the optical rotation ([α] 20 D = +18.2 (c 0.05, CHCl3)) was compared to the reported data for 1 ([α] 25 D = +56 (c 0.38, CHCl3)), [13] suggesting the same absolute configuration. The data for 2 ([α] 20 D = +18.7 (c 0.013, CHCl3)) matched closely and in sign with those of 1, suggesting 2 has the same absolute configuration as 1 [13].

Cytotoxicity and Effect on Intracellular Calcium
Since compounds 1 and 2 were isolated from the same fraction as the cytotoxic lagunamide D (3), they were originally tested for their cytotoxicity against A549 cells as described for 3 [23]. Iezoside (1) and B (2) were shown to possess micromolar activity against A549 cells. Iezoside (1) was shown to be two-fold more active than 2 with IC50 values of 1.5 and 3.0 μΜ, respectively ( Figure 5A). Both compounds were shown to be non-cytotoxic against primary human small airway epithelial cells (HSAEC), showing

Cytotoxicity and Effect on Intracellular Calcium
Since compounds 1 and 2 were isolated from the same fraction as the cytotoxic lagunamide D (3), they were originally tested for their cytotoxicity against A549 cells as described for 3 [23]. Iezoside (1) and B (2) were shown to possess micromolar activity against A549 cells. Iezoside (1) was shown to be two-fold more active than 2 with IC 50 values of 1.5 and 3.0 µM, respectively ( Figure 5A). Both compounds were shown to be non-cytotoxic against primary human small airway epithelial cells (HSAEC), showing over 70% viability at the highest tested concentration (10 µM) compared to the~20% shown in the A549 line ( Figure 5D), suggesting a potential preference for lung cancer cells over normal small airway epithelial cells. Thapsigargin showed a similar reduced cytotoxicity in HSAEC as shown in Figure S20.
The antiproliferative activity of 1 was originally observed in HeLa cells, concomitant with a change in morphology, an increase in cells in G1 phase and a decrease in cells in S phase, indicating a cell cycle delay [13]. Furthermore, 1 was tested in 39 human cancer cell lines by the Japanese Foundation for Cancer Research 39 (JFCR39), and the fingerprint showed similarities to that of thapsigargin and A23187. Iezoside (1) shows a similar phenotype in HeLa cells as thapsigargin and cyclopiazonic acid, both of which are SERCA inhibitors [14,16,17]. To further investigate that SERCA was the presumptive functional target of 1, the compound was evaluated for its inhibitory effect of SERCA1a via a coupled enzyme assay and downstream effect on the expression of cell cycle regulation and apoptosis proteins [13].
Mar. Drugs 2023, 21, x FOR PEER REVIEW 7 of 12 over 70% viability at the highest tested concentration (10 μM) compared to the ~20% shown in the A549 line ( Figure 5D), suggesting a potential preference for lung cancer cells over normal small airway epithelial cells. Thapsigargin showed a similar reduced cytotoxicity in HSAEC as shown in Figure S20.  With the prior knowledge of 1 being a SERCA inhibitor, both compounds were tested for their effect on cytosolic calcium oscillations in A549 cells, alongside thapsigargin. An increase in cytosolic calcium was observed near the IC 50 for 1 and 2 (0.32 and 1 µM) with an approximate EC 50 of 0.3 µM and 0.6 µM, respectively ( Figure 5B,C, following the same trend as the antiproliferative activity for both compounds (Table 2). The compounds were also tested for their cytotoxicity in HeLa cells, similarly as described in the literature [13]. The compounds showed an IC 50 of 1.0 and 2.4 µM, respectively, at 48 h ( Figure 5E), with compound 1 being 2.3-fold more potent than 2. Our IC 50 values are somewhat higher than reported (6.8 nM), although the precise experimental condition differed, including incubation time (72 h) and, possibly, genetic variations in the cell line [13]. Both compounds were shown to reduce cell viability of primary cervical epithelial cells with an IC 50 of 0.02 and 0.05 µM, respectively ( Figure 5H), while thapsigargin was two-fold more potent against primary cervical epithelial cells compared to HeLa cells (0.06 ± 0.04 µM in HeLa cells and 0.03 ± 0.01 µM in the primary cells). This data suggests substantial cell type-dependent differences and that the preference for cancer cells over normal cells cannot be generalized. The spindle-like morphology of HeLa cells that was previously observed for 1 was also noticeable in our experiments ( Figure 6A), suggesting that 2 has a similar mechanism of action [13]. At the tested concentration, 26.5 ± 1.0% of the cells treated with 1 and 14.3 ± 0.4% of the cells treated with 2 showed the spindle-like morphology after 24 h of exposure ( Figure 6B). In the calcium assay using HeLa cells, a similar fold difference was detected with activity at an approximate EC 50 of 0.1 µM and 0.5 µM for 1 and 2, respectively ( Figure 5F,G, Table 2).
The effective concentration observed for both in the intracellular calcium assay is in the same order of magnitude as the viability assay and shows the same fold difference between the two compounds. These data indicate that the O-methyl in position C-31 is not essential for the activity of the compound class ( Figure 5B,C,E,F) ( Table 2), but that the C-31 O-Me group has incremental contribution to the overall activity at the level of target engagement and calcium flux as well as antiproliferative activity. This structure-activity relationship (SAR) suggests that the activity is potentially tunable through modification in the sugar moiety.

General Experimental Procedures
The optical rotations were recorded on a Rudolph Research Analytical Autopol III automatic polarimeter. NMR data were collected on a Bruker Avance II 600 MHz, high resolution 5-mm cryoprobe spectrometer operating at 600 MHz for 1 H and 150 MHz for 13 C, using residual solvent signals (δ H 2.50; δ C 39.5 ppm DMSO-d 6 , δ H 2.50 ppm (CD 3 ) 2 CO) as internal standards. The edited HSQC and HMBC experiments were optimized for 1 J CH = 140 Hz n J CH = 8 Hz, unless indicated otherwise (3 Hz HMBC). HRMS data were obtained using a Bruker Daltonics, Impact II QTOF with electrospray ionization (ESI). Chiral analysis was performed using an Agilent 6230 ESI-ToF and an Applied Biosystems 3200 QTRAP triple quad/linear trap. The effect on intracellular calcium was measured using a Molecular Devices Flexstation instrument. Circular Dichroism data were collected on a Chirascan™ Circular Dichroism Spectrometer (Applied Photophysics, Surrey, UK), with the Pro-Data Chirascan and Pro-Data viewer software 4.7.0.

Biological Material
The cyanobacterial tufts (DRTO 85) were collected from the shallow reef at Loggerhead Key, FL, on 9 May 2015. The cyanobacterial assemblage was microscopically identified as a mixture of Dichothrix sp., Lyngbya sp. and Rivularia sp. [23]. The Smithsonian Marine Station, Fort Pierce, FL, USA, maintains a voucher.

Extraction and Isolation
The cyanobacterium was lyophilized and extracted using 1:1 EtOAc:MeOH followed by 1:1 EtOH:H 2 O to provide 9.22 g and 11.88 g of extract, respectively. The extract was partitioned between EtOAc and H 2 O, followed by partitioning the polar layer between n-BuOH and H 2 O. The EtOAc fraction was subjected to a silica column using hexanes and increasing amounts of EtOAc, followed by increasing amounts of MeOH in EtOAC to give thirteen fractions. The SiO 2 fraction eluting with 25% MeOH in EtOAc was further purified using a snap C18 column with increasing amounts of MeOH in H 2 O. The fraction eluting at 100% MeOH was further purified using a H 2 O/MeOH gradient (80-90% for 30 min, followed by 100% over 15 min) by HPLC (Phenomenex Synergi 4 µ Hydro-RP 80 Å, 250 × 10 mm, 4 µm; flow rate, 2.0 mL/min; UV detection at 200 nm and 220 nm), from which the fractions at t R 18 min and t R 24 min were further purified. The first fraction was purified using a H 2 O/ACN gradient (60-70% for 30 min, followed by 100% over 10 min) by HPLC (Phenomenex Synergi 4µ Hydro-RP 80 Å column, 250 × 10 mm, 4 µm UV detection at 200 nm, 2 mL/min) and yielded iezoside B (2) (0.2 mg) at t R 25.5 min. The second fraction was further purified using a H 2 O/ACN gradient (60-80% for 30 min, followed by 100% over 10 min) by HPLC (Phenomenex Synergi 4µ Hydro-RP 80 Å column, 250 × 10 mm, 4 µm UV detection at 200 nm, 2 mL/min) which yielded iezoside (1)

Cell Viability Assay
A549 (American Type Culture Collection, ATCC), HSAEC (ATCC), HeLa (ATCC) and primary cervical epithelial (ATCC) cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum at 37 • C humidified air and 5% CO 2 . Cells were seeded (7500/well (A549, HSAEC) and 2000/well (HeLa, primary cervical epithelial cells)) in 96-well plates and allowed to attach overnight before being treated with the compounds or the solvent control (0.5% DMSO). The cells were incubated for 48 h followed by addition of MTT dye, according to the manufacturer's protocol (Promega). IC 50 values were calculated using GraphPad prism software 9.2.0.

HeLa Cell Morphology Study
HeLa (ATCC) cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum at 37 • C humidified air and 5% CO 2 . Cells were seeded 700/well in 384-well plates and allowed to attach overnight before being treated with the compounds or the solvent control (0.5% DMSO). The cells were incubated for 24 h followed by observation under the microscope to quantify the morphological changes.

Intracellular Calcium Assay
A549 and HeLa cells were seeded (10,000/well and 20,000/well, respectively) overnight in black plates. The cells were allowed to attach overnight before the compounds or the solvent control (0.5% DMSO) was added. The FLIPR calcium 6 assay kit was used following the manufacturer's protocol (FLIPR Ca 6, Molecular Devices, San Jose, CA, USA). For the calculation of the approximate EC 50 values, the vehicle control was used as the minimum response, and the maximum detected response before saturating the detector was used as the maxima.

Conclusions
In conclusion, we report the isolation of iezoside (1) and iezoside B (2), the second member of the peptide-polyketide hybrid glycoside, from a collection of a marine cyanobacterial assemblage from Florida. Iezoside (1) was previously isolated from Okinawa, Japan, and now isolated from the Dry Tortugas, Florida, indicating the broad distribution of the compound. Lagunamide D (3), also isolated from the same extract, was proven to be a mitochondrial cytotoxin, which further supports the richness of cyanobacterial extracts [24]. Compound 2 shows similar but slightly weaker activity than 1 in the assays performed, indicating that the methylation on the sugar moiety only minorly impacts the biological effect. The SAR provides evidence that structural changes in the sugar unit may be tolerated and that the activity of this compound class is tunable. This finding has implications for future analogue synthesis and potential target interaction studies.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/md21070378/s1, Figures S1-S15: 1D/2D NMR data of iezoside and iezoside B in DMSO-d 6 and acetone-d 6 ; Figures S16-S21: Bioassay data for thapsigargin, Table S1: Comparison of experimental and literature 1 H NMR data of iezoside in acetone-d 6  Funding: This research was supported by the National Institutes of Health (NIH), NCI grant R01CA172310 and the Debbie and Sylvia DeSantis Chair Professorship (H.L.). Mass spectrometry analysis was supported by NIH grant S10OD021758-01A1.
Data Availability Statement: Data is contained within the article or supplementary materials. Raw data will be made available upon request.