Calculation of anomalous difference Fourier electron density maps revealed the location of zinc ions that Luminespib clinical trial form numerous intermolecular

contacts between protomers in both crystal forms (Figures 7A and S1C). It is likely that these zinc-mediated contacts play a key role in assembly of the two dimer forms observed in the crystal structures, which could explain why GluK3 failed to pack in the canonical arrangement found in many crystal structures for other iGluR LBDs. However, GluK3 LBD glutamate and kainate complexes, which were crystallized in the absence of zinc ions (Venskutonytė et al., 2011, 2012), were also packed in a noncanonical dimer configuration in the P41 space group, highlighting the tendency of KAR LBDs to pack in a variety of nonbiological assemblies, as observed previously for GluK1 and GluK2 (Mayer, 2005; Naur et al., 2005). Relevant to the zinc potentiation of GluK3, zinc ions were bound at a site labeled Zn1, which was created by D759 in all three crystal forms (Figure 7B); the zinc ion at this site was

also coordinated by H762, mutation of which abolished potentiation by zinc, and by H492 and E495 located in helix D of the dimer partner subunit. Additional zinc binding sites, which selleck kinase inhibitor stabilize the alternative dimer assembly, were created by E757 at the base of Levetiracetam helix J together with E757 of its symmetry mate (Zn2), by H444 at the N terminus of helix B (Zn3), by E713 in helix I together with H762 and E766 from a symmetry related molecule (Zn4), by H479 at the C terminus of helix C (Zn5), by the main-chain carbonyl oxygen of Glu495 and the side-chain carboxylate of D499 just after the C terminus of helix D (Zn6), by H479 with its symmetry mate (Zn7), and by E441 with H444 of a symmetry-related molecule (Zn8). For most of these sites, the binding of zinc was characterized by short bond lengths, on the order of 1.9–2.0 Å (Figure 7A). Although the GluK3 LBD dimer arrangements observed here differ from

the canonical arrangement of full-length GluA2 receptors (Sobolevsky et al., 2009), we tested functionally the involvement of H492, which participates in the Zn1 site (Figure 7B), but which is absent in other iGluR subunits. We mutated the histidine into a tyrosine, the equivalent residue in GluK2, and also into an alanine. For both GluK3(H492Y) and GluK3(H492A), zinc still potentiated currents (Figures 7D and 7F). Conversely, GluK2(Y490H) was, like WT GluK2, inhibited by zinc (Figures 7E and 7F). Therefore, H492 does not participate in the functional zinc binding site for GluK3 potentiation. This result is consistent with the fact that substitution of the S1 region of GluK2 to GluK3 does not abolish zinc potentiation (Figures 5C and 5D).