AMPA receptors commandeer an ancient cargo exporter for use as an auxiliary subunit for signaling.

CNIH-2 increases surface expression of AMPARs. A Quantification of GluA1o surface expression levels by extracellular epitope tagging in HeLa cells expressing GluA1o alone (CTRL) or co-expressing GluA1o and CNIH-2 (CNIH-2). Representative micrographs show an increase in extracellularly HA-tagged GluA1o on the cell surface of HeLa cells when CNIH-2 is co-expressed, visualized by anti-HA immunocytochemistry in non-permeabilized cells. Histogram data are mean surface expression levels ± SEM normalized to CTRL. Asterisk marks a significant difference from CTRL (p<0.01, unpaired Student's t-test; n = 24 for CTRL and CNIH-2, respectively). B Surface biotinylation of HeLa cells expressing GluA1o (CTRL) or co-expressing GluA1o and CNIH-2 (CNIH-2) (n = 6). Note that CNIH-2 co-expression increases both total and surface AMPAR levels. T = total, S = surface, I = internal. Protein load for S is concentrated 10fold. Depletion of the ER-resident lectin calnexin in S serves as a control for specificity of surface membrane biotinylation. C (Top) Representative current traces from somatic outside-out patches evoked by 0.5 mM glutamate (+250 µM TCM to block receptor desensitization) in sham-infected control and CNIH-2 over-expressing CA1 pyramidal neurons of organotypic hippocampal slice cultures DIV 7–10. (Bottom) Quantification of steady-state currents. Data are mean ± SEM. Asterisk marks a significant difference from control (p<0.0001, unpaired Student's t-test; n = 14 and n = 11 for CTRL and CNIH-2, respectively).

Subcellular localization of exogenously expressed CNIH-2. A Representative confocal images of HeLa cells, dissociated hippocampal neurons and glial cells over-expressing CNIH-2. Note the perinuclear accumulation of CNIH-2, which co-localizes with the Golgi markers GalTase-GFP and GM130. B CNIH-2 behaves similar to other Golgi-resident proteins cycling between Golgi and ER, as it co-distributes with GalTase into peripheral Golgi remnants upon nocodazole treatment (10 µM, 4 hrs).

CNIH-2 facilitates ER export of AMPARs. A Representative confocal images of OK cells stably expressing CNIH-2. Co-expression of dominant-negative Sar1 H79G prevents ER export of CNIH-2 leading to its redistribution into the ER. B Quantification of GluA1o surface expression levels by extracellular epitope tagging in the presence of CNIH-2 and either wildtype (WT) Sar1 (white bar) or mutant Sar1 H79G (grey bar). Data are mean increases in surface expression levels by CNIH-2 ± SEM normalized to GluA1o+Sar1 WT or GluA1o+Sar1 H79G without CNIH-2, respectively. Asterisk marks a significant increase in surface expression of GluA1o by co-expression of CNIH-2 (p<0.001, unpaired Student's t-test; n = 12 for both experimental groups). C Quantification of GluA1o surface expression levels in the presence of CNIH-2 and either wildtype dynamin-1 (white bar) or dominant-negative dynamin-1 K44A (grey bar) inhibiting clathrin-dependent endocytosis [38]. Data are mean increases in surface expression levels by CNIH-2 ± SEM normalized as in B (n = 6 for both experimental groups).

CNIH-2 changes glycosylation of GluAs. A Western blot analysis of total and surface populations of GluA2i extracted from HeLa cells by surface biotinylation in the absence (CTRL) or presence of CNIH-2 (CNIH-2) (n = 4). Extensive glycosylation of surface GluA2i during maturation increases its apparent molecular weight. Note the smaller increase upon co-expression of CNIH-2. B Enzymatic deglycosylation analysis of GluA2i surface populations in the presence (+) or absence (−) of CNIH-2. Surface GluA2i remained either untreated (C) or was incubated with either endoglycosidase H (H) or PNGase F (F). Note that upon CNIH-2 co-expression, the GluA2i surface population remains sensitive to endoglycosidase H (n = 2).

Surface trafficking of GluAs by CNIH-2 is splicing-dependent. A Quantification of GluA surface expression levels by extracellular epitope tagging in HeLa cells expressing the indicated GluA subunits. Data are mean ± SEM normalized to GluA1o (GluA1o: n = 24; GluA1i: n = 9; GluA2o: n = 12; GluA2i: n = 12). B Increase in GluA surface expression mediated by CNIH-2 in HeLa cells. Data are mean ± SEM normalized to surface expression of respective GluA subunits without CNIH-2 (GluA1o: n = 24; GluA1i: n = 8; GluA2o: n = 12; GluA2i: n = 12).

CNIH-2 is rendered a surface membrane protein by assembly with AMPARs. A Total (T), surface (S) and internal (I) populations of CNIH-2 in HeLa cells expressing either CNIH-2 alone (CTRL) or together with GluA1o or GluA2i, respectively. S is concentrated 10fold. Note that in the absence of GluAs, CNIH-2 cannot be detected in the surface fraction. However, it is robustly observed in the plasma membrane when co-expressed with GluAs (n = 4). B Total (T), surface (S) and internal (I) populations of CNIH-2 in dissociated hippocampal neurons (DIV 17) transduced with CNIH-2 (+) or GFP (−). S is concentrated 10fold. Both endogenous (−) and over-expressed (+) CNIH-2 can be detected on the cell surface (n = 5). C (Top) Representative current traces recorded in somatic outside-out patches excised from dissociated hippocampal neurons (DIV 16–21) over-expressing either GFP (CTRL, black) or CNIH-2 (CNIH-2, red) upon 1 ms (left panel) and 100 ms applications (right panel) of 10 mM glutamate. (Bottom) Quantification of deactivation and desensitization kinetics. Data are given as mean ± SD. Asterisk denotes a significant difference from control (p<0.01, unpaired Student's t-test; deactivation: n = 10 and 8 for CTRL and CNIH-2, respectively; desensitization: n = 19 and 8 for CTRL and CNIH-2, respectively).