Towards the structural characterisation of equilibrative nucleoside transporters

Human equilibrative nucleoside transporters (ENTs) represent a major pharmaceutical target for cardiac, cancer and viral therapies. Understanding the molecular basis for transport is crucial for the development of improved therapeutics through structure-based drug design. ENTs have been proposed to utilise an alternating access mechanism of action, similar to that of the major facilitator superfamily. However, ENTs lack functionally-essential features of that superfamily, suggesting they may use a different transport mechanism. Understanding the molecular basis of their transport requires insight into diverse conformational states. In this work I explored distinct but complementary approaches towards the stabilisation of ENTs, with the goal of structural characterisation.

First, I demonstrated that yields of wild-type human ENT isoform 1 (hENT1) obtained using previously optimised methodologies (expression in Spodoptera frugiperda cells and a two-step IMAC/protease purification) were not amenable to further investigations due to instability and poor yields. Second, I explored stabilising point mutations of hENT1. I identified four distinct variants of hENT1 with single- point mutations at the large intracellular loop (ICL6) and transmembrane helix 7 (TM7) that stabilise the apo-state (∆Tm 0.7-1.5 °C). Furthermore, I found that two variants that specifically stabilise the inhibitor-bound state (∆∆Tm 3.0-5.0 °C), supporting the role of ICL6 and TM7 in hENT1 gating. I also found that variant T336A (TM8) binds the inhibitor nitrobenzylthioinosine with a 7-fold lower affinity than wild-type. Therefore, this residue may help determine inhibitor and substrate sensitivity. Finally, I identified a homologue of hENT1 from the ascomycete Byssochlamys spectabilis (BsENT). BsENT is significantly more stable than hENT1 (ΔTm of 17.3 ± 1.9 °C (p = <0.0001)) and can be purified from insect cells with high, homogenous and monodisperse yields, 3-5-fold higher than hENT1. Furthermore, initial crystallisation trials suggest that BsENT is amenable to structural characterisation. Thus, BsENT is a promising candidate for homology modelling of hENT1.