Several studies have demonstrated that ER export of GluAs critically depends on isoform and flip/flop splice variant [9], [41], [42]. We therefore addressed the question whether the increase in GluA surface population by CNIH-2 was similarly affected by using the extracellular epitope tagging approach. As shown in Figure 5A, the ability of GluA1 and GluA2 subunits to reach the cell surface of HeLa cells varied markedly with GluA2i being the subunit that was expressed most and GluA1o the one being expressed least on the cell surface in the absence of CNIH-2. In the presence of CNIH-2, the increase in surface expression of indicated GluA subunits was exactly opposite (Fig. 5B). It was largest for the flop variant of GluA1 (GluA1o; 13.6±1.0; n = 24) and smallest for the flip isoform of GluA2 (GluA2i; 1.4±0.1; n = 12). GluA1i and GluA2o showed intermediate increases in surface expression when co-expressed with CNIH-2 (3.2±0.2 (n = 8) and 2.2±0.1 (n = 12), respectively). These data demonstrate that CNIH-2 can at least partially compensate for splice form-dependent differences in the ER export rates of GluA isoforms.
