*2.1. Chemistry*

The biomimetic syntheses of thiaplidiaquinones A and B and their corresponding thiazine regioisomers **3** and **4** *via* precursor quinones **5** and **6** (Figure 2) have been reported elsewhere [5].

**Figure 2.** Structures of thiazine regioisomers **3** and **4** and precursor quinones **5** and **6**.

### *2.2. Biology*

#### 2.2.1. Inhibition of ROS Generation

It was originally reported that thiaplidiaquinones A and B displayed a strong accumulation of intracellular ROS in Jurkat cells, 97% of cells for **1** and 93% for **2**, relative to untreated cells [4]. In the present study, the presence of ROS in Jurkat cells was determined using dihydrorhodamine 123 (DHR123), a cell permeable probe that becomes fluorescent when oxidized to rhodamine 123 in the presence of ROS [7]. Once the cells were loaded with DHR123, they were then treated with test compounds at a range of concentrations, and the mean fluorescent intensity determined by flow cytometry. In Figure 3 it is shown that even at a top test concentration of 100 μM, natural products **1** and **2** exhibited only modest levels of accumulation of intracellular ROS, particularly in comparison to precursor quinones **5** and **6** and the dioxothiazine regioisomer of thiaplidiaquinone A **3**. Of the four dioxothiazine-containing compounds (**1**–**4**), **3** was clearly the most potent generator of intracellular ROS in this assay.

**Figure 3.** Comparison of DHR fluorescence by Jurkat cells treated with **1**–**6**.

#### 2.2.2. Apoptosis *vs.* Necrosis in Jurkat Cells

It was reported previously that natural products **1** and **2** exhibited similar levels of cytotoxicity towards Jurkat cells with IC50 ~3 μM using a standard cell viability assay [4]. In addition, by using propidium iodide (PI) as a probe, from the observation of treated cells having a significant loss of nuclear DNA in combination with ROS production it was concluded that both **1** and **2** were inducing apoptosis. However, as this approach looked at the average cell population it does not provide the opportunity to clearly differentiate between apoptosing and necrosing cell subpopulations. In our hands, natural products **1** and **2**, and regioisomers **3** and **4**, exhibited significantly less potent cytotoxicity towards Jurkat cells, with toxicity only apparent at concentrations close to 100 μM. For our assessment of the mechanism of cell death, we used the combination of Annexin V-FITC and PI stains to allow us to evaluate the ability of test compounds **1**–**6** to induce apoptosis or necrosis (or both) in Jurkat cells. After 24 h treatment in the presence of 100 μM of each test compound, cells were analyzed by flow cytometry to identify live (AnV−/PI−), apoptosing (AnV+/PI−), necrotic (AnV−/PI+) and late apoptotic (AnV+/PI+) cells.

As shown in Figure 4, natural products thiaplidiaquinone A (**1**) and B (**2**) exhibited quite different profiles, with **1** inducing both necrosis (PI+/AV−) and apoptosis (PI−/AV+) and **2** causing cell death almost exclusively *via* apoptosis. A similar profile was observed for the dioxothiazine regioisomers **3** and **4**, with an even clearer distinction between cell death by necrosis for **3**, while cell death induced by **4** was dominantly *via* apoptosis. These results identify that placement of the geranyl side chain in ring-D of the molecule specifically dictates the mechanism of cell death (**1**/**3** *vs.* **2**/**4**) with a geranyl chain at the C-3 position resulting in death by apoptosis in comparison to a geranyl chain at the C-4 position which results in necrotic cell death. In direct contrast to the dioxothiazine-containing compounds, the profiles of both precursor quinones **5** and **6** were dominated by late stage apoptosis (PI+/AV+). From the observations made, it is evident that the presence of the dioxothiazine ring modulates the activity of the compounds (*i.e.*, **1**–**4** *vs.* **5**/**6**) whereas the placement of one of the geranyl side chains in the structure specifically dictates the mechanism of cell death (**1**/**3** *vs.* **2**/**4**). It should be noted that the induction of ROS did not clearly correlate with either apoptotic or necrotic cell death indicating that ROS production appears to be a poor indicator of the route of cell death for the current test compounds. This is not unexpected as necrosis is also linked to ROS production [8,9], thereby confounding the use of ROS as an indicator of the route of cell death.

**Figure 4.** Summary plots of flow cytometry data using Annexin V-FITC and PI to determine mode of cell death for **1**–**6** (necrotic PI+/AnV<sup>−</sup>; apoptotic PI−/AnV+; dead/late apoptosis PI+/AnV<sup>−</sup>).
