To shed more light on the binding of

To shed more light on the binding of benzophenanthridine to GlyT1, we studied how the replacement of certain residues around the sanguinarine sensitive cysteine 475 previously identified (Jursky et al., 2012) affects GlyT1 inhibition of by these alkaloids. Based on these results we propose a putative model of how alkaloids bind to the intracellular part of GlyT1.
Materials and methods
Results In previous work we found that both sanguinarine and chelerythrine inhibit glycine uptake mediated by transporter GlyT1 (Jursky and Baliova, 2011). Further studies found that the inhibition can be eliminated by mutating GlyT1c cysteine 475 alone (Jursky et al., 2012). To study the interaction of alkaloids with GlyT1 in more detail, we constructed a structural homology model of human GlyT1 using an inward-open LeuT template available in the PDB. This conformation was chosen because of the exclusive accesibility of cysteine 475 from the intracellular (cytoplasmic) side of GlyT1 in this conformation (Jursky et al., 2012). Because of their aromatic character, the benzophenanthridines might use stacking interactions. For this reason, we performed DNA site-directed mutagenesis on the coding triplets at all aromatic amino-acids in the vicinity of cysteine 475 (Fig. 1). In order to keep the transporter functional, the native residues were replaced with either the most similar amino-acids or the amino-acids found in corresponding positions of the closest family members. To verify that the mutations did not greatly perturb the basic GlyT1 trafficking and glycine transport, the protein reductase enzyme and activity of each mutant was tested using both Western blotting and glycine uptake measurement (Fig. 2A). Those mutants that exhibited proper trafficking, giving a typical glycosylated immunoreactive band of the correct size (75kDa) and a significant glycine uptake above background, were then assayed for glycine uptake in the presence and absence of alkaloids. Comparing the effects of the alkaloids on the glycine uptake of both wild type and mutated transporters with these of the untreated control allowed us to infer the potential role of the mutated residues in the binding of benzophenanthridines to GlyT1. In most mutants we observed a level of inhibition very similar to the wild type and an insensitive to alkaloids when cysteine 475 was mutated (Fig. 2B). The exception is the mutation of tyrosine 370 to phenylalanine, which significantly increased the ability of both sanguinarine and chelerythrine to inhibit GlyT1. This indicates that this residue might also be involved with C475 in the interaction of benzophenanthridines with GlyT1 (Fig. 2B). Fig. 2B also shows that substitution of tyrosine 370 to leucine did not replicate the effect of the phenylalanine substitution, suggesting the potential importance of aromaticity. Benzophenanthridines also have aromatic character. To investigate if this property is important for the inhibition of GlyT1 we eliminated the aromaticity of sanguinarine by reducing it to dihydrosanguinarine with sodium borohydride (Brossi and Borer, 1965, Psotova et al., 2006). This treatment completely abolished sanguinarine’s ability to inhibit GlyT1 (Fig. 2C). To investigate whether the increased inhibition introduced by the Y370F mutation results from the creation of a new, artificial alkaloid interaction site, we made a double GlyT1 mutant bearing both C475S and Y370F mutations. Surprisingly, even though this mutant produced correct immunoreactive band in the Western blot (Fig. 2A), its ability to transport glycine was severly impaired. The subsequent, stepwise replacement of cysteine 475 with glycine, alanine and valine showed that the substitution C475V produced an active transporter even when the Y370F mutation was introduced simultaneously. Further analysis of the C475V and C475V/Y370F mutants showed that not only had the C475V mutation recovered the glycine transport ability, but it had also made the transporter insensitive to benzophenanthridines, both C475V and C475V/Y370 mutants exhibited transport activities comparable to the wild-type control. These results indicate that the Y370F mutation does not introduce a new binding pocket and that this mutation, by itself, is not sufficient for binding the alkaloid, though it is most probably involved.