Mycothiazole (MYZ) (Figure 1) was isolated and purified from the Tongan marine sponge Cacospongia mycofijiensis as described in the experimental section. Four other compounds—latrunculin A, dendrolasin, laulimalide, and zampanolide—were also isolated from the same sponge species [16].

The concentrations of MYZ that produced 50% inhibition (IC₅₀) of MTT reduction for cell lines or CSFE proliferation for T cells were determined using eleven established cell lines and primary murine splenic T-cells (Table 1). These cells showed diverse responses to MYZ, with most cell lines (HeLa, P815, HeLaS3, 143B, B16, 4T1, MDCK, and RAW264.7) and the murine T-cells being highly sensitive, and three cell lines (HL-60, LN18 and Jurkat) being relatively resistant to the drug. For example, HL-60 cells had a mean IC₅₀ value of 12 μM; whereas, HeLa were highly sensitive with a mean IC₅₀ of 0.36 nM (Table 1, Figure 2A,B). In contrast, P815 and MDCK cells exhibited a biphasic response (Table 1, Figure 2C,D); thus, low nanomolar concentrations of MYZ inhibited MTT responses of these cells by about 50%; whereas, more extensive inhibition was observed at micromolar concentrations, suggesting both high and low-sensitivity mechanisms of action. Preliminary attempts to investigate cell recovery by washing out MYZ for 24 h after a 24 h exposure showed that some cell lines (e.g., 143B) remained sensitive after washout, others (P815) were partially reversed in their ability to proliferate, and some, like HeLa, completely recovered from the effects of the drug (data not shown). ³H-thymidine uptake was used to confirm that the MTT assay was measuring a proliferative response and not just metabolic depression of non-dividing cells (Figure 3).

MYZ has been shown to inhibit complex I of MET, and therefore cell lines lacking MET were used to investigate how this would effect the MYZ potency. Ablation of MET function in HL-60 cells by mitochondrial genome deletion (HL-60ρ⁰ cells) had little effect on the IC₅₀ value for MYZ (26 μM compared to 12 μM in the parental cell line) (Table 1, Figure 2A). On the other hand, the IC₅₀ for MYZ with HeLaρ⁰ cells (Table 1, Figure 2B) was 78 μM compared to 0.36 nM in the parental cell line—a 217,000-fold decrease in sensitivity in the absence of MET function. In all cases, ρ⁰ cells were highly insensitive to the action of MYZ with IC₅₀ values in the 26–78 μM range. In P815ρ⁰ cells the initial 50% reduction in MTT response to nanomolar concentrations of MYZ was lost. An IC₅₀ value of 22 μM was observed, the MTT reduction occurring in the same concentration range as the micromolar part of the biphasic response in P815 cells (17 μM).

To distinguish cytotoxic from cytostatic effects, trypan blue dye exclusion was used with HeLa, P815, and HL-60 cells (Figure 4). MYZ treatment at concentrations as high as 25 μM had little effect on the viability of HeLa and P815 cells, even after 48 h incubation with the drug. With HeLa cells, 50% cytotoxicity was seen after 48 h exposure to 100 μM MYZ (Figure 4A), and with P815 cells, 25–30 h exposure to 50 μM MYZ resulted in 40% loss of viable cells (Figure 4B). Thus, the cytotoxic effects of MYZ were only seen at concentrations four or more orders of magnitude above the IC₅₀ concentration determined in the MTT assay. In contrast, treatment of HL-60 cells with MYZ at a concentration comparable to its IC₅₀ in the MTT assay (25 μM) resulted in complete loss of viability (Figure 4C). Cytotoxic effects of MYZ with HL-60 cells were evident as early as 5 h following exposure.

To investigate the kinetics of metabolic inhibition by MYZ, MTT assays were performed on P815 cells treated with MYZ for 0, 12, 24, and 48 h (Figure 5). There was an unexpected lag period of 12 h before the MTT response began to be affected by MYZ. The threshold concentration required to give an effect on metabolism was about 0.1 nM MYZ. As expected, inhibition at each concentration increased with longer exposure times.

Using flow cytometry of PI-stained cells, the effects of MYZ on progression of HeLa cells through the cell cycle were investigated. MYZ after 16 h exposure had no effect on cell cycle progression at concentrations up to 100 nM, three orders of magnitude greater than the MTT IC₅₀, (Table 2). Synchrony of HeLa cells with 2 mM thymidine block, then release also showed no significant changes in cell cycle distribution (data not shown). A similar lack of a cell cycle effect was seen with HL-60 cells at concentrations from 0.1–50 μM (n = 3 preparations) (Figure S2 of Supplementary Data). To ensure a block was not missed in the HeLa cells, concentrations up to 20 μM and exposures up to 48 h were also tested, and again, no treatment resulted in cell cycle arrest (Figure 6).

DCF (2,7-dichlorofluorescein) [17] was used to determine whether MYZ affected the production of reactive oxygen species (ROS) (Figure 7). After 24 h treatment with MYZ, decreased ROS levels were seen in both HeLa and P815 cells (Figure 7A,D) compared with untreated control cells, but by 48 h, the DCF response had returned to control levels (Figure 7B,E). With HeLaρ⁰ at 48 h (Figure 7C) and P815ρ⁰ at 24 h (Figure 7F), some lesser shifts in DCF fluorescence were also observed. Dichloroacetate (DCA) was chosen as a positive control for increasing ROS in the cells; however, it proved to be too toxic to the cells at the two concentrations tested (20 μM and 2 μM).

The mode of action of MYZ remained elusive until recently when Morgan et al. reported it to be a MET complex I inhibitor as well as an HIF-1 inhibitor [5,6]. These authors reported a large range in sensitivity of different cell lines to growth inhibition by MYZ, with the T47D breast cancer cell line being the most sensitive and another breast cancer cell line MDA-MB-231 being the least sensitive. McLaughlin [11] has also reported highly selective activity of acetogenins, potent complex I inhibitors, in different cell lines. When screened against the NCI 60 cell line panel, MYZ showed potent activity against the small and non-small lung cancer cell lines DMS114 (13 nM IC₅₀) and NCI-H23 (250 nM IC₅₀); however, there was a broad range of activity in other cell lines from 0.1–100 μM [4]. In HCT-116 colon cancer cells, MYZ had an IC₅₀ of 3.8 nM [18]. In our study, MYZ also showed a broad range of activity in different cancer cell lines. In another NCI study of MET complex 1 inhibitors [10], leukemic cell lines such as K562 were particularly sensitive, yet in our study, although splenic T cells were very sensitive, the leukemic cell lines HL-60 and Jurkat showed a high level of resistance to MYZ.

The nanomolar response in sensitive cells suggests a direct link between NADH flux and MET, a principal site of NADH oxidation in respiring cells. To investigate this link, a panel of ρ⁰ cells was used. These cells lack mitochondrial DNA, and therefore have defective electron transport function at complexes I, III, and IV as well as having a non-functional cytochrome oxidase (complex V, [15]). Relative to MYZ-sensitive parental cells (0.36–14 nM IC₅₀ range), ρ⁰ cells for these same cell lines had low sensitivity to MYZ in the MTT assay with IC₅₀ values in the range of 26–78 μM. This amounts to a 2500–217,000-fold difference in MTT sensitivity between parental and ρ⁰ cells. With HL-60 cells, on the other hand, the IC₅₀ values were similar between the parental (12 μM) and HL-60ρ⁰ (26 μM) cells. Thus, having functional MET made little difference to the effect of MYZ in this cell line. It was therefore concluded that MET was not necessary for the micromolar MYZ sensitivity seen in those cells.

In a preliminary test with rotenone (Figure S3 of Supplementary Data), MYZ-insensitive HL-60 cells gave IC₅₀ values of 0.41 μM for rotenone in wild type cells and 0.39 μM in HL-60ρ⁰ cells. Thus, the independence of mitochondrial function was similar for both complex I inhibitor drugs. It is possible that resistant cell lines that normally use oxidative phosphorylation to generate ATP, such as HL-60, are able to shift their metabolism to aerobic glycolysis in the presence of a complex I inhibitor like MYZ or rotenone. A shift to aerobic glycolysis, commonly associated with cancer cells, is referred to as the Warburg effect. In a study by Xu et al. [19], a mitochondrial-deficient clone of HL-60 cells (HL-60/C6F) showed enhanced inhibition of cellular ATP levels in the presence of the glycolytic inhibitor 3-bromopyruvate at concentrations of 50–100 μM, indicating that when mitochondrial function is impaired, the cells can alter their metabolism toward aerobic glycolysis. At a higher concentration of 300 μM 3-bromopyruvate, however, both HL-60 and HL-60/C6F cells were completely depleted of ATP and underwent apoptosis.

Our findings with mitochondrial DNA-free cells are novel and suggest that the nanomolar effects of MYZ require MET function, but that MET function is not sufficient for these effects, since the insensitive cell lines were unaffected at nanomolar concentrations of MYZ, yet have normal mitochondrial function. Thus, some other factor or factors not present in insensitive cells appears to be required for the nanomolar effects of MYZ on MET. The nature of these MYZ-priming events and the extent to which they are involved with other complex I inhibitors is at present unknown.

Although MTT dye reduction assays are often referred to as cell proliferation assays, the assay primarily measures cytosolic NAD(P)H oxidoreductases [20,21] and therefore is more correctly a measure of cellular metabolic activity. Using this assay, we showed that MYZ had no effect on MTT responses of sensitive cells until at least 12 h exposure, with an apparent threshold effect at about 0.1 nM MYZ in P815 cells. This contrasts with uncouplers such as FCCP that enhance MET and oxygen consumption within seconds. MET complex inhibitors, like rotenone and myxothiazol, azide and cyanide, also act within minutes to prevent NADH oxidation [22]. Hence, MYZ appears to be an atypical MET complex I inhibitor in its kinetics of action. The reasons for the delay in MYZ action could be due to a slow uptake into the cell, its likely membrane localization, or the need to generate an active metabolite before it can work on its target. In the study by Morgan et al. [6], using both intact and digitonin-permeabilized cells, MYZ, like rotenone, caused an immediate decrease in oxygen consumption by the cells. Just how this metabolic decrease from inhibition of complex I relates to actual cell growth inhibition, as measured in this study, is not known. In human CEM leukemic cells, McLaughlin [11] reported a 2 or 3 day lag period before significant depletion (up to 76% depletion) of ribonucleotide triphosphates occurred following addition of the acetogenin bullatacin.

No cell cycle block was observed in the sensitive HeLa cells after 24 or 48 h exposure, even at MYZ concentrations as high as 20 μM. This effect was not specific to HeLa cells since the non-sensitive HL-60 cells also failed to show the G₂/M arrest seen with rotenone. These results are consistent with a generalized metabolic effect of MYZ in sensitive cells rather than a specific effect on different phases of the cell cycle. The results are opposite to that of the complex I inhibitor rotenone, which arrests cells in G₂/M of the cell cycle [23,24].

In MYZ sensitive cells, the cells remained viable at concentrations of MYZ that extensively reduced metabolism. At higher micromolar concentrations, the MTT response correlated with loss of cell viability in both insensitive cells and in sensitive cells that have a biphasic response to the compound. This introduced the possibility that a second mode of action was involved at these higher concentrations in both types of cells. The second target may only be affected at high concentrations of MYZ because of a low sensitivity binding site, or alternatively, because MYZ is being metabolized to another bioactive compound that reaches its threshold concentration for the second target only at high concentrations. It seems unlikely that increases in ROS levels are involved in the cytotoxicity, since the DCF signal was decreased in HeLa and P815 cells after 24 h treatment with MYZ. It is surprising that MYZ reduced ROS levels in these two cell lines, given that complex I inhibitors like rotenone generally increase O₂^.−, the predominant ROS in cells [10].

Methanolic extracts of the marine sponge Cacospongia mycofijiensis, collected from ‘Eua, Tonga were fractionated to generate MYZ in greater than 95% purity. All spectroscopic data, including ¹H and ¹³C NMR (Figure S1 of Supplementary Data), optical rotation and high-resolution mass spectrometry were in accordance with previously reported accounts of MYZ [1,3]. The NMR spectra (CDCl₃, 600 MHz) were obtained using a Varian DirectDrive spectrometer equipped with a triple resonance HCN cryogenic probe. High-resolution positive-ion mass spectra were recorded on a Waters Q-TOF Premier™ Tandem Mass spectrometer. Optical rotations were performed using a Rudolph Autopol II polarimeter. Normal-phase column chromatography was carried out with silica gel. Reversed-phase column chromatography was achieved using Supelco Diaion HP20 poly(styrene-divinylbenzene) (PSDVB) chromatographic resin. HPLC was performed using a Rainin Dynamax SD-200 solvent delivery system, with UV detection using a Varian ProStar 335 photodiode array detector. Solvents used for flash normal- and reversed-phase column chromatography were of HPLC or analytical grade quality. All other solvents were purified by glass distillation. Solvent mixtures are reported as % (v/v) unless otherwise stated. Cacospongia mycofijiensis sponge specimens were stored at −20 °C until required for extraction. Frozen sponge (21.0 g) was extracted in MeOH (2 × 100 mL) and loaded onto PSDVB. The loaded extracts were eluted with: (i) 30% Me₂CO/H₂O, (ii) 75% Me₂CO/H₂O and (iii) Me₂CO. A portion of the 75% Me₂CO/H₂O fraction was fractionated further on PSDVB (30–100% Me₂CO/H₂O). The 60–70% Me₂CO/H₂O fractions were further purified using silica gel (CH₂Cl₂–EtOAc). The 10–20% EtOAc/CH₂Cl₂ fractions were finally purified using HPLC (DIOL, 5 μm, 4.6 × 250 mm, 10% IPA/n-hexane) to yield mycothiazole (5.7 mg, t_R = 5.5 min) as a colorless oil with the following properties: [α]_D²⁰ −19.3° (c 0.93, CHCl₃); HRESIMS [M + H]⁺, observed m/z 405.2215, calculated m/z 405.2212 for C₂₂H₃₃N₂O₃S, Δ = +0.7 ppm.

Cell lines used in this study included murine splenic T-cells, adherent cell lines of human (HeLa cervical carcinoma, HeLa S3 semi-adherent clone, 143B osteosarcoma, LN18 glioblastoma), mouse (4T1 breast carcinoma, B16 melanoma), and dog origin (MDCK kidney epithelial), and non-adherent cell lines of human (HL-60 promyelocytic leukemia, Jurkat T-cell lymphoma) and mouse origin (P815 mastocytoma, RAW264.7 monocyte/macrophage). Cell lines were obtained from the following sources: HeLa (Professor Anthony Braithwaite, University of Otago), HeLa S3 (Dr. Alfons Lawen, University of Melbourne), 143B and 143Bρ⁰ (Dr. Mike Murphy, University of Otago), P815 (Dr. John Marbrook, University of Auckland), HL-60 (Dr. Graeme Findlay, University of Auckland), Jurkat (Dr. Thomas Bäckström, Malaghan Institute, Wellington), LN18, 4T1, B16, MDCK, and RAW264.7 (ATCC, Manassas, VA). The parent lines are all phenotypically stable and exhibit properties characteristic of their cell or tissue of origin. HL-60 is regularly tested in our laboratory for respiratory burst activity and differentiation into neutrophils (retinoic acid) or monocytes (PMA or vitamin D₃). Four of the five ρ⁰ cell lines lacking mitochondrial DNA (HL-60ρ⁰, HeLaρ⁰, P815ρ⁰, and B16ρ⁰) used in this study, with the exception of 143Bρ⁰, were generated in our laboratory by prolonged culture of the parental cell lines in the presence of ethidium bromide, followed by culture in its absence [14]. These phenotypically stable cell lines were auxotrophic for uridine and pyruvate. All ρ⁰ cells were tested at regular intervals for loss of MET function by FCCP resistance of WST-1 reduction, and by PCR for absence of mitochondrially-encoded cytochrome b.

The standard culture medium consisted of RPMI-1640 medium supplemented with 5–10% fetal calf serum, 2 mM glutamate, 25 μg/mL penicillin, 25 μg/mL streptomycin, 50 μg/mL uridine, and 1 mM pyruvate. Murine B16 cells, however, were cultured in complete medium (cIMDM) containing Iscove’s modified Dulbecco’s medium (IMDM), 2 mM Glutamax^®, 10% fetal calf serum, 100 U/mL penicillin, 100 µg/mL streptomycin with 50 μg/mL uridine and 1 mM pyruvate. RAW264.7 cells and CD4/CD8 T cells isolated from the mouse spleen were cultured in a DMEM-based complete T cell medium containing 10% FCS and other supplements as previously described [31]. All cells were grown at 37 °C in an atmosphere of 5% CO₂ in air and high humidity. Adherent cell lines were passaged using TrypLE™ Express (Invitrogen) to detach the cells, washed with medium, and then passaged into fresh medium.

An MTT cell proliferation assay [20,21] was used to measure cell metabolism. Cells reduce the yellow 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenoyltetrazolium bromide (MTT) dye to blue formazan crystals in a manner dependent on intracellular reduced pyridine nucleotides, and therefore the assay is a measure of cell metabolism. Cells were seeded at a density of 1 × 10⁴ cells/well in a 96-well plate and treated with varying concentrations of MYZ for different lengths of time. MTT solution (20 μL of 5 mg/mL in PBS) was added to each well, and after 2 h incubation, 100 μL of solubilizer (10% SDS, 45% N,N-dimethylformamide) was added to lyse the cells and dissolve the blue crystals. The absorbance was measured at 570 nm using a multi-well plate reader. Absorbances were then used to generate concentration-response curves and to calculate IC₅₀ values (the concentration at which MTT reduction is 50% of a non-treated control) using SigmaPlot software.

Cells were seeded in 96-well plates at a density of 1 × 10⁵ cells/mL and treated with MYZ. For the last 8 h of culture, 1 μCi ³H-thymidine was added to each well, and the plates were frozen to stop cell proliferation. After thawing, the cells in each well were transferred to filters using a cell harvester (Tomtec, Hamden, CT) and the radioactivity determined in a liquid scintillation counter (Wallac MicroBeta TriLux).

In primary cultures of splenic T cells, a carboxyfluorescein succinimidyl ester (CFSE) proliferation assay was used to determine the IC₅₀ for growth inhibition, as previously described [31]. Briefly, splenocytes were isolated from female C57BL/6 mice and stained with 1 µM CFSE. Cells were then treated with different concentrations of MYZ in the presence or absence of Con A (3 µg/mL; Sigma Chemical Co.) for 48 h. The cells were then harvested and stained with fluorescently-labelled antibodies (anti-CD4-cychrome c and anti-CD8-PE; BD Biosciences) or isotype controls, according to the manufacturer’s recommendations, and analyzed by flow cytometry using a FACScan (Becton Dickinson). A total of 10,000 events were collected for each assay, and the data were analyzed by CellQuest Pro software (Becton Dickinson).

A trypan blue dye exclusion assay was used to determine cell viability. Cells (1 × 10⁵/mL) were cultured with varying concentrations of MYZ and then stained with trypan blue and counted in a hemocytometer.

Flow cytometric analysis of cell DNA content following staining with propidium iodide (PI) was used to assess progression through the cell cycle. Cells were seeded at 1 × 10⁵ cells/well in 24-well plates and incubated with MYZ for various times. The cells were detached with 50 μL TrypLE^TM Express, washed in PBS, and then incubated with 250 μL of PI solution (0.05 mg/mL PI, 0.1% sodium citrate, 0.1% Triton X-100 in PBS). Samples were scanned and analyzed on a FACScan flow cytometer with BD Cell Quest Pro software.

Mitochondrial ROS was measured by flow cytometry using changes in fluorescence of dichlorofluorescein (DCF). Briefly, cells were prepared as for cell cycle analysis, then 230 μL of 10 μM DCF in PBS was added and left for 30 min. PI (20 μL of 100 μg/mL) was also added to allow gating on only live, PI-negative cells. The samples were then scanned on a FACScan flow cytometer with BD Cell Quest Pro software.