Anti-vitiligo mechanism and the target protein of phenanthridine derivatives
To facilitate the identification of the cellular target (or targets) of HCJA121 and HCJA404, biotinylated derivatives were generated through label-free quantitative proteomic approaches. Axin appeared to act as a binding protein of HCJA121 and HCJA404, which was confirmed using a pull-down assay. Next, to confirm the anti-vitiligo target protein of the phenanthridine derivatives, we performed modelling studies for the complexes axin–HCJA121 and axin–HCJA404. A possible binding site for1 and 2 on DAX, which has important implications on the function of axin, has been investigated using the AutoDock docking program (Morris et al., 2009) with the structure of DAX (PDB, 1WSP) (Schwarz-Romond et al., 2007) as the receptor molecule. Nine conform­ers of HCJA121 and 11 conform­ers of HCJA404 that were outputted from AutoDock onto DAX were analysed; six (HCJA121) and five (HCJA404) of the conformers were localized to the same site formed by the juxtaposition of the residues from two neighbouring protomers. Subsequently, the model with the lowest estimated free energy for binding (−9.35 kcal mol-1 for HCJA121 and −8.42 kcal mol-1 for HCJA404) was selected and modified slightly. The modelled structure of the com­plexes of axin with HCJA121 and HCJA404 revealed that LYS781 and LEU784 in the binding cavity could form contacts with HCJA121 and HCJA404 through electrostatic or hydrophobic interactions (Fig. 3a). To verify the importance of these residues in the binding of compounds, DAX mutants carrying K781A and L784A mutations were generated. As predicted by the model, in the two DAX mutants, HCJA121 and HCJA404 exhibited significantly reduced activation of not only Wnt target genes but also melanogenesis genes (Fig. 3b). Similar to melanogenesis, melanin synthesis was clearly decreased in drug-treated mutated cells (Fig. 3c). However, the effects of 8-MOP on melanogenesis genes and melanogenesis in mutated cells appeared to be unaffected, suggesting that compounds targeting the DAX domain of axin through binding with LYS781 and LEU784 potentiate the axin–LRP6 association to promote Wnt signalling transduction and eventually promote melanogenesis.
These data suggest that HCJA121 and HCJA404 could promote melanogenesis by activating the Wnt signalling pathway. We have previously suggested an autoinhibition hypothesis, according to which the binding of phenanthridine derivatives to the axin C terminus disrupts the interaction with the axin N terminus and thus relieves the autoinhibition of axin. Therefore, active axin binds LRP6 more easily and thus promotes LRP6 phosphorylation when the Wnt signalling pathway is activated 24. Hence, we presumed that these anti-vitiligo phenanthridine derivatives could potentiate the axin–LRP6 association to influence melanin synthesis. To validate this hypothesis, we performed co-immunoprecipitation (co-IP) of axin and LRP6 from the lysates of HCJA121-treated B16 cells (using anti-axin1 and anti-LRP6 antibodies for immunoprecipitation); subsequently, the proteins were separated by western blotting. First, anti-axin1 was employed to capture the axin protein and axin-related complex. Next, LRP6 antibody was used on the protein mixtures captured by the axin1 antibody for capturing LRP6. The spot representing LRP6/anti-LRP6 from the HCJA121-treated group was considerably larger than that from the DMSO-treated group at the same loading quantity of the sample (Fig. 3d), which suggested that axin and LRP6 exhibited strong association in the phenanthridine-derivative-treated cells. The protein complex of axin and LRP6 in drug-treated cells appeared to be more common than it was in the control, which implies that phenanthridine derivatives could potentiate the axin–LRP6 association.