*3.2. Bioactivity of* D. orbita *Brominated Indoles*

The predicted Molinspiration bioactivity scores for *D. orbita* brominated indoles identify the ultimate precursor tyrindoxyl sulfate as the most likely pharmacophore. This compound shows potential as a GPCR ligand and enzyme inhibitor (Table 2a). Unlike the intermediate precursor compounds, this polar brominated indoxyl sulfate salt has not been directly tested for cytotoxicity in antibacterial and anticancer screening assays. This is because bioassay guided fractionation of *D. orbita* extracts has revealed most of the activity in the more lipophilic fractions of chloroform extracts and generally no activity is found in the polar methanol water fractions [12,19,26,71,76,77], where tyrindoxyl sulfate is mostly concentrated. Nevertheless, tyrindoxyl sulfate has been present in some of the anticancer extracts showing bioactivity against MCF-7 breast cancer cells *in vitro* [26] and against DNA damaged colon cells *in vivo* [78] and could contribute to the observed activity. Tyrindoxyl sulfate is likely to be metabolized and transported differently to the other less polar compounds *in vivo* (Table 1a). This, along with the predicted enzyme binding activity, suggests that tyrindoxyl sulfate might be worthy of further bioactivity studies.

At the other extreme of polarity (Table 1a), the Tyrian purple pigments have predicted protein kinase receptor interaction (Table 1b). This predicted activity is supported for 6,6 dibromoindirubin, which was shown to be a selective GSK-3 inhibitor, but with limited activity against CDK1/Cyclin B or CDK5/p25 [72,79]. The 6,6 dibromoindigo isomer was not tested in this study and although predicted to have some protein kinase activity (Table 2a), the extreme insolubility of this compound presents problems for bioactivity assessment.

Despite the compatibility with drug-likeness, few of the intermediate brominated indoles from *D. orbita* produced high enough bioactivity scores on Molinspiration to indicate interesting pharmacophores for receptor binding (Table 2a). Nevertheless, purified extracts containing 6-bromoisatin and tyrindoleninone do show broad spectrum antibacterial and anticancer activities [4, 12,19,23,71,77,80]. Of particular interest is the >100 fold selective cytotoxicity towards human lymphoma and female reproductive cancer cell lines (KGN, JAr, OVCAR-3), compared to freshly isolated untransformed peripheral blood monocytes and female granulosa cells [19,22,23,71,80]. Furthermore, these brominated indole derivatives appear to induce apoptosis rather than necrosis in the reproductive cancer cell lines, as indicated by caspase 3/7 activity and DNA fragmentation from TUNNEL staining [19,71]. Preliminary work on these brominated indoles using flow cytometry with propidium iodine and annexin staining indicates they also induce apoptosis in lymphoma cells but not in CaCO2 colon cancer cells [23,77,81]. However, more recent studies on purified 6-bromoisatin and tyrindoleninine indicate they do induce apoptosis in the H2T9 colon cancer cell line [81]. Furthermore, a rodent model for colon cancer using a concentrated extract containing these two brominated indoles shows that apoptosis is induced *in vivo* and unpublished studies indicate that 6-bromoisatin is the main active factor [81]. The mode of action for these brominated indole derivatives is currently unknown and as they are unlikely to bind with the receptors or enzymes listed in Table 2, further studies are required.

The dimeric compound tyriverdin was not predicted to have any bioactivity based on known pharmacophores for receptor or enzyme binding (Table 2a). Nevertheless, this compound has been identified as a potent bacteriostatic agent against a range of human and marine pathogens, using bioassay guided fractionation of *D. orbita* extracts with the flourescein diacetate hydrolysis antibacterial assay [12]. However, further testing of this compound with alternative methods, such as the standard plate dilution assay [12] or the MTS tetrazolium salt cell proliferation assay [26,27] has failed to confirm the antibacterial activity. Additional procedural controls have indicated that tyriverdin can partially quench the green fluorescence of flourescein in the absence of bacterial cells [27]. However, this quenching did not account for all the apparent reduction in fluorescein absorbance, suggesting that tyriverdin may also interfere with esterase activity or some other mechanism of converting flourescein diacetate to flourescien. However, in addition to violating the molecular weight rule for drug-likeness (Table 1a), tyriverdin also has solubility and instability problems. It is only slightly soluble in some solvents, such as chloroform and dichloromethane, but tends to precipitate out of most solvents (e.g., ether extracts [8] and toluene/hexane [27]), then decomposes to Tyrian purple. This low solubility along with its instability in sunlight and high molecular weight make it an unlikely drug-candidate.
