The endoplasmic reticulum membrane complex promotes proteostasis of GABAA receptors

Summary The endoplasmic reticulum membrane complex (EMC) plays a critical role in the biogenesis of tail-anchored proteins and a subset of multi-pass membrane proteins in the endoplasmic reticulum (ER). However, because of nearly exclusive expression of neurotransmitter-gated ion channels in the central nervous system (CNS), the role of the EMC in their biogenesis is not well understood. In this study, we demonstrated that the EMC positively regulates the surface trafficking and thus function of endogenous γ-aminobutyric acid type A (GABAA) receptors, the primary inhibitory ion channels in the mammalian brain. Moreover, among ten EMC subunits, EMC3 and EMC6 have the most prominent effect, and overexpression of EMC3 or EMC6 is sufficient to restore the function of epilepsy-associated GABAA receptor variants. In addition, EMC3 and EMC6 demonstrate endogenous interactions with major neuroreceptors, which depends on their transmembrane domains, suggesting a general role of the EMC in the biogenesis of neuroreceptors.


INTRODUCTION
Acquiring the correct transmembrane topology is essential for the function of membrane proteins, which consist of about 30% of the eukaryotic proteome. The endoplasmic reticulum membrane complex (EMC) plays a critical role in the insertion of membrane proteins into the lipid bilayer of the endoplasmic reticulum (ER) (Chitwood et al., 2018;Guna et al., 2018;Shurtleff et al., 2018). The EMC is ubiquitously expressed and highly conserved (Volkmar and Christianson, 2020;Wideman, 2015). Ten subunits (EMC1-10) have been identified in the EMC in mammals: EMC1, EMC3-7, and EMC10 are membrane-spanning, whereas EMC2, EMC8, and EMC9 are soluble (EMC8 and EMC9 are structurally redundant and mutually exclusive) ( Figure 1A) (O'Donnell et al., 2020). Recent cryo-electron microscope (cryo-EM) structures showed the overall similar organization of the human EMC (Pleiner et al., 2020) and yeast EMC (Bai et al., 2020): both have a large ER luminal region, 12 transmembrane helices, and a smaller cytosolic region. Research has shown that EMC1-3, 5, and 6 are the core subunits because their depletion leads to co-translational degradation of other subunits, malfunction in the assembly of the full mature EMC, and loss of EMC's overall activity in U-2 OS cells (Volkmar et al., 2019).
A growing number of EMC-dependent client membrane proteins continue to be reported (Volkmar and Christianson, 2020). The EMC was first described in yeast in a genetic screen for protein folding factors (Jonikas et al., 2009). Later, the EMC was identified in the screen of the interactome for the mammalian ER-associated degradation (ERAD) network (Christianson et al., 2011). Recently, the EMC's important role has been recognized in inserting tail-anchored membrane proteins post-translationally and a subset of multi-pass membrane proteins co-translationally. The EMC acts as an insertase for tail-anchored proteins with moderately hydrophobic transmembrane domains . Subsequently, the EMC was reported to facilitate the insertion of the first transmembrane domain of certain G-protein coupled receptors (GPCRs), such as b-adrenergic receptors (Chitwood et al., 2018). Furthermore, quantitative proteomics analysis comprehensively identified EMC-dependent membrane protein clients, showing that the EMC enables the biogenesis of membrane proteins with destabilizing features, such as polar and charged residues-containing transmembrane domains (Shurtleff et al., 2018;Tian et al., 2019). The idea that the EMC could act as a more general factor to facilitate the biogenesis of membrane proteins should be further investigated because the EMC regulates the function of a variety of membrane proteins, including cystic fibrosis transmembrane conductance regulator (CFTR) in yeast, acetylcholine receptors in C. elegans, and rhodopsins and transient receptor potential (TRP) channels in Drosophila (Louie et al., 2012;Richard et al., 2013;Satoh et al., 2015).
We focus on proteostasis maintenance of neurotransmitter-gated ion channels . Because of the nearly exclusive expression of such membrane proteins in the CNS, previous screenings in yeast and mammalian cells did not identify the potential involvement of the EMC in their biogenesis (Shurtleff et al., 2018;Tian et al., 2019). Interestingly, EMC6 was shown to positively regulate the expression of acetylcholine receptors and the response to GABA in C. elegans (Richard et al., 2013). Therefore, here we evaluated the role of the EMC in the biogenesis of GABA A receptors, the primary inhibitory ion channels in the mammalian CNS (Akerman and Cline, 2007). Functional GABA A receptors are assembled as a pentamer in the ER from eight subunit classes: a1-6, b1-3, g1-3, d, ε, q, p, and r1-3 (Sequeira et al., 2019). The most common subtype in the human brain contains two a1 subunits, two b2/b3 subunits, and one g2 subunit ( Figure 1B) (Laverty et al., 2019;Zhu et al., 2018). Individual subunits must fold into their native structures in the ER and assemble with other subunits correctly on the ER membrane to form a heteropentamer (Barnes, 2001;Jacob et al., 2008;Martenson et al., 2017). Only properly assembled receptors exit the ER, traffic through the Golgi for complex glycosylation, and reach the plasma membrane for their proper function (Lorenz-Guertin and Jacob, 2018). The proteostasis network, which contains chaperones (such as BiP and calnexin), ERAD factors (such as Hrd1, Sel1L, and VCP), and trafficking factors (such as LMAN1), orchestrates the folding, assembly, degradation, and trafficking of GABA A receptors, which is essential for their function (Di et al., 2013;Di et al., 2016;Fu et al., 2019;Wang et al., 2022b;Wang et al., 2013). Loss of function of GABA A receptors causes a variety of neurological diseases, including epilepsy and intellectual disability (Hernandez and Macdonald, 2019). Furthermore, numerous clinical variants in a single subunit cause subunit protein misfolding and/or disrupt assembly of the pentameric complex in the ER (Fu et al., 2022). Such mutant subunits are retained in the ER and subjected to excessive ERAD, being dislocated into the cytosol and degraded by the proteasome. This leads to decreased cell surface localization of the receptor complex and imbalanced neural circuits. Because the proteostasis network orchestrates the biogenesis of GABA A receptors, regulating the expression or function of their proteostasis network components, such as GABA A receptor-interacting chaperones and ERAD factors, can fine-tune the functional surface expression of GABA A receptors. Specifically, activating the pro-folding chaperones or inhibiting the ERAD factors genetically or pharmacologically has the potential to restore the surface trafficking and thus function of pathogenic GABA A receptors, providing a therapeutic strategy to ameliorate related neurological diseases (Wang et al., 2022a;Wang et al., 2014). For example, overexpression of BiP, a pro-folding Hsp70 family chaperone in the ER, is sufficient to promote the anterograde trafficking and surface expression of pathogenic GABA A receptors containing the A322D mutation in the a1 subunit (Di et al., 2013). Consistently, increasing the BiP protein levels pharmacologically through the application of a small molecule BiP activator called BIX leads to enhanced functional surface expression of this variant (Fu et al., 2018).
Here we aim to understand the role of each individual subunit of the EMC on the proteostasis maintenance of GABA A receptors. We found that EMC3 and EMC6 have the most prominent effect on increasing the protein levels of endogenous GABA A receptors. Furthermore, the interactions between the EMC and GABA A receptors are dependent on the transmembrane domains. Overexpressing the EMC is sufficient to restore the function of certain disease-associated variants of GABA A receptors.

Knocking down EMC3 and EMC6 reduces the protein levels and function of endogenous GABA A receptors
We evaluated the effect of each of the ten individual EMC subunits on endogenous GABA A receptor protein levels. As a starting point, we used a mouse GT1-7 hypothalamic neuronal cell line because it is Figure 1. Effect of depleting individual EMC subunits on endogenous GABA A receptor a1 subunit protein levels (A) Schematics of ten EMC subunits. (B) Cartoon Representation of heteropentameric GABA A receptors. The most common subtype in the mammalian CNS consists of a1, b2/b3, and g2 subunits. (C to L) Endogenous total GABA A receptor a1 subunits protein level change on knocking down individual EMC subunits. Mouse hypothalamus GT1-7 neuronal cells were transfected with siRNAs against EMC1 to EMC10, respectively. Two distinct siRNAs targeting each of the ten EMC subunits, designated as EMCn.1 and EMCn.2 (n = 1 to 10), were used to minimize the potential off-target effects. Forty-eight hours after transfection, proteins were extracted and analyzed by western blotting. b-actin was used as the loading control. Normalized band intensity was shown below the images (n = 3). Each data point is presented as mean G SEM *, p< 0.05; **, p< 0.01. NT: Non-targeting scrambled siRNA. iScience Article one of the very few neuronal cells that express endogenous, functional GABA A receptors (with a1 and b3 subunits) (Hales et al., 1992). In addition, GT1-7 is a mature mouse hypothalamic gonadotropin releasing hormone cell line that responds to membrane depolarization (Hales et al., 1994), which is a defining characteristic of neurotransmitter-gated ion channels. To investigate whether there are key subunits of the EMC that play critical roles in the biogenesis of GABA A receptors, we conducted siRNA screening by using two distinct siRNAs targeting each of the ten EMC subunits to minimize the potential off-target effects. GT1-7 cells have high transfection efficiency, which also made them suitable for the siRNA screening experiments (Fu et al., 2019). To further increase knockdown efficiency, we performed two siRNA transfections at 24 and 48 h before harvesting the proteins for SDS-poly acrylamide gel electrophoresis (PAGE) and western blot analysis ( Figures 1C-1L).

ll
Of the five core subunits of the EMC (EMC1-3, 5, and 6) that are critical for its assembly and activity, EMC1, EMC3 and EMC6 positively regulated a1 protein levels (Figures 1C,1E and 1H), whereas the knockdown of EMC2 or EMC5 did not influence the a1 protein levels significantly ( Figures 1D and 1G), suggesting that the formation of the mature EMC might not be a prerequisite for the regulation of GABA A receptor protein levels. The most substantial reduction of the total a1 protein levels was observed with the depletion of EMC3 (p<0.01, Figure 1E) and EMC6 (p<0.01, Figure 1H). This is noted in the figures as decreased EMC3 band intensity for EMC3.1 and EMC3.2 siRNAs-treated samples, compared to the non-targeting (NT) control, to 38 and 37% respectively ( Figure 1E). Correspondingly, when a1 was detected, two decreased band intensities were observed as well, to 41 and 45% respectively ( Figure 1E). We also noted significant effects on EMC6 knockdown (p< 0.01, Figure 1H) as well. This is noted as decreased EMC6 band intensity for EMC6.1 and EMC6.2 siRNAs-treated samples, compared to NT, to 70 and 58% respectively. Correspondingly, when a1 was detected, two decreased band intensities were observed as well, to 75 and 35% respectively ( Figure 1H). Therefore, our results suggested that individual EMC subunit, such as EMC3 and EMC6, is sufficient to positively regulate GABA A receptor protein levels (also see below). EMC3 is the only EMC subunit that exhibits homology to Oxa1 family proteins, which are known membrane protein insertases (Volkmar and Christianson, 2020;Wideman, 2015), and EMC6 plays a critical role in regulating the biogenesis of acetylcholine receptors in C. elegans (Richard et al., 2013).
Among the five peripheral subunits of the EMC (EMC4, 7-10) that are not essential for the complex's stability or assembly, EMC4 and EMC10 positively regulated a1 protein levels ( Figures 1F and 1L), whereas the knockdown of EMC7-9 did not influence the a1 protein levels significantly ( Figures 1I, 1J and 1K), again suggesting a subunit-specific contribution of the EMC. It is worth noting that all cytosolic, soluble EMC subunits (EMC2, 8-9) did not influence a1 protein levels significantly, suggesting that the transmembrane domains in other membrane-spanning EMC subunits could play a critical role in regulating GABA A receptor biogenesis in the ER membrane (also see below).
In addition, we performed experiments to detect endogenous b3 protein levels from EMC3 and EMC6 knockdown samples in GT1-7 cells to further determine their effects on other major subunits of GABA A receptors ( Figure 2A). Knocking down EMC3 by using two distinct siRNAs decreased the EMC3 protein levels Figure 2. Effect of EMC3 and EMC6 on the protein levels and whole-cell patch-clamping currents of endogenous GABA A receptors (A) Mouse GT1-7 neurons were incubated with siRNA against EMC3 or EMC6 for 48 h. Proteins were extracted and analyzed by western blotting; normalized band intensity was shown below the images (n = 3), with b-actin as the loading control.
(B) Mouse GT1-7 neurons were incubated with siRNA against EMC3 or EMC6 for 48 h. Whole-cell patch-clamping was performed on the cells with the IonFlux Mercury 16 ensemble plates at a holding potential of À60 mV. GABA (1 mM) was applied for 4 s, as indicated by the horizontal bar above the currents. The peak currents (Imax) were acquired and analyzed by Fluxion Data Analyzer (n = 6 -10). NT: Non-targeting scrambled siRNA; pA, picoampere. iScience Article to 50 and 25% compared to the non-targeting control ( Figure 2A, left panel) and reduced the endogenous b3 protein levels to 55 and 30%, respectively ( Figure 2A, left panel). Moreover, knocking down EMC6 by using two distinct siRNAs decreased the EMC6 protein levels to 55 and 30% compared to the non-targeting control ( To understand their effects on the function of endogenous GABA A receptors, we carried out whole-cell patch-clamping electrophysiological recording using GT1-7 neurons after siRNA treatment of EMC3 and EMC6. The peak GABA-induced current in the non-targeting siRNA-treated GT1-7 neurons is 837 pA; consistently, knocking down EMC3 or EMC6 decreased the peak current (Imax) to 375 and 573 pA, corresponding to 55 and 32% reduction respectively ( Figure 2B). These results unambiguously demonstrated EMC3 and EMC6 as key EMC subunits to positively regulate protein expression and function of endogenous GABA A receptors.

EMC3 and EMC6 promote anterograde trafficking of GABA A receptors
Because GABA A receptors need to traffic to the plasma membrane for their function, we further tested the hypothesis about the necessity of the EMC with respect to surface trafficking of endogenous GABA A receptors using immunofluorescence staining. We used primary rat cortical neurons because all of the major subunits (a1, b2, b3, and g2 subunits) of functional GABA A receptors are abundantly expressed in the plasma membrane in the cortex (Sequeira et al., 2019). EMC3 and EMC6 depletion was achieved through the lentivirus transduction of four siRNA pools (Ding and Kilpatrick, 2013). At day-in-vitro (DIV) 6 of the neurons, lentivirus transduction was carried out at a MOI of 10, that is, the ratio of lentivirus count to neuron count in each well. Immunofluorescence staining was performed at DIV 12 (Glynn and McAllister, 2006). The application of an anti-GABA A receptor subunit antibody that recognizes the extracellular epitope without a prior membrane permeabilization step using detergent enabled us to label the cell surface proteins of a1, b2/3 and g2 subunits. Application of the Alexa Fluor Dye-conjugated secondary antibody enabled the imaging of the surface subunits using a confocal laser-scanning microscope. Clearly, depletion of EMC3 and EMC6 reduced the surface expression of a1 subunits by 37 and 41%, b2/3 subunits by 61 and 50%, and g2 subunits by 44 and 58% in cortical neurons ( Figure 2C), indicating that EMC3 and EMC6 play a critical role in positively regulating the surface trafficking of all major subunits of endogenous GABA A receptors.
In addition, we carried out surface biotinylation experiments to evaluate the effect of the EMC on exogenously expressed human a1b2g2 GABA A receptors in HEK293T cells. HEK293T cells have been used . Surface proteins were enriched through biotinneutravidin affinity purification, and western blot analysis was applied to detect surface a1 and b2 subunits. Na + /K + ATPase served as loading control of cell surface proteins. To test the total protein expression of GABA A receptors, cells were lysed and total proteins were collected and subjected to SDS-PAGE and western blot analysis (B). b-actin was used as the total protein loading control. Normalized band intensity was shown on the right to the blots (n = 3).
(C) The ratio of the surface/total subunits of GABA A receptors was quantified, as a measure of their surface trafficking efficiency. Data was taken from (A) and (B) for the calculation.
(D) HEK293T cells stably expressing a1b2g2 GABA A receptors were transfected with non-targeting siRNA or siRNAs against EMC3 and EMC6. 48 h after transfection, cycloheximide (CHX) (100 mg/mL), a potent protein synthesis inhibitor, was added to cell culture media for the indicated time. Cells were then harvested, and total proteins were subjected to SDS-PAGE and western blot analysis. Quantification of the a1 band intensity was plotted against the incubation time with CHX (n = 3).
(E and F) EMC3 and EMC6 promote GABA A receptors' trafficking from the ER to Golgi as demonstrated through endoglycosidase H (Endo H) digestion. EMC3 or EMC6 siRNA transfection was applied in HEK293T cells stably expressing a1b2g2 GABA A receptors; 48 h after transfection, proteins were extracted, and subjected to Endo H digestion and western blot analysis. Endo H resistant bands (top two bands in lanes two and 4) represent proteins that have correctly folded in the ER, trafficked to Golgi and fully modified with the N-linked complex glycans, thus becoming resistant to Endo H. On the other hand, acting upon proteins remaining in ER, Endo H may remove the high mannose structure after the asparaginyl-N-acetyl-D-glucosamine on the a1 subunits, generating Endo H sensitive bands (bottom band in lanes two and 4). The Peptide-N-Glycosidase F (PNGase F) enzyme-treated samples served as a control for unglycosylated a1 subunits (lane 5). Quantification of the ratio of Endo H resistant/total a1 band intensity, as a measure of the trafficking efficiency of a1 subunits, was shown on the right (n = 3).
(G) HEK293T cells stably expressing a1b2g2 GABA A receptors were transfected with non-targeting siRNA or siRNAs against EMC3, EMC6, or both EMC3 and EMC6. 48 h after transfection, cells were harvested, and total proteins were subjected to SDS-PAGE and western blot analysis. Quantification of the normalized individual EMC subunit band intensity was shown on the right panels (n = 3). Each data point is presented as mean G SEM *, p< 0.05; **, p< 0.01. NT: Non-targeting scrambled siRNA.

OPEN ACCESS
iScience 25, 104754, August 19, 2022 7 iScience Article extensively for the functional expression of neurotransmitter-gated ion channels because HEK293T cells have low or no endogenous expression of such ion channels, and thus precise control of the subtypes can be achieved from the exogenously expressed subunits (Thomas and Smart, 2005). In addition, HEK293T cells are capable of modeling the proteostasis network that orchestrates the folding, assembly, degradation, and trafficking of neurotransmitter-gated ion channels and thus are a suitable vehicle to elucidate the biogenesis pathways of such channels (Wang et al., 2022b). After 48 h of transfection of both EMC3 and EMC6 siRNAs, surface proteins were enriched through biotin-neutravidin affinity purification (see STAR Methods), and western blot analysis was applied to detect a1 and b2 subunits ( Figure 3A). Significant reduction of cell surface a1 subunits of GABA A receptors was observed when both EMC3 and EMC6 were knocked down compared to NT, to 78% or 34% when using two sets of siRNAs ( Figure 3A, left panel). Similarly, cell surface b2 subunits decreased to 55% or 42% when using two sets of siRNAs ( Figure 3A, right panel), when both EMC3 and EMC6 were knocked down. Because knocking down both EMC3 and EMC6 using siRNA set #2 also decreased the total protein level of a1 to 61.8% and that of b2 to 69.2% ( Figure 3B), we evaluated whether the reduced surface expression of GABA A receptor subunits was a direct consequence of the lowered total receptor expression level in cells. The ratio of surface/total subunits is commonly used as a measure of their surface trafficking efficiency. Knocking down both EMC3 and EMC6 decreased such a ratio for a1 subunits to 61.5% and b2 subunits to 60.7% ( Figure 3C), indicating that EMC3 and EMC6 positively regulate the surface trafficking efficiency of GABA A receptors, which could result from the post-translational effect of EMC3 and EMC6 on the folding and degradation of GABA A receptors.
We hypothesized that knocking down EMC3 and EMC6 would lead to folding and assembly defects of GABA A receptors in the ER and thus their retention in the ER and excessive degradation, causing compromised anterograde trafficking. We evaluated the effect of EMC3 and EMC6 on the degradation kinetics of GABA A receptors using cycloheximide-chase experiments. Remarkably, compared to the non-targeting control, knocking down EMC3 and EMC6 using siRNAs decreased the relative protein abundance of the a1 subunit at each time point during the 5 h chase period after the addition of cycloheximide, a potent protein synthesis inhibitor, in HEK293T cells stably expressing a1b2g2 GABA A receptors ( Figure 3D), These results indicated that depleting EMC3 and EMC6 accelerated the protein degradation of GABA A receptors. Therefore, EMC3 and EMC6 positively regulate the protein stability of GABA A receptors in cells.
Furthermore, we determined the effect of EMC3 and EMC6 on GABA A receptors' trafficking from ER to Golgi through the endoglycosidase H (Endo H) digestion assay. The human a1 subunit of GABA A receptors has two N-linked glycosylation sites at N38 and N138. Endo H resistant bands represent glycoproteins, which have correctly folded in the ER, trafficked to the Golgi and fully modified with the N-linked complex glycans, thus becoming resistant to Endo H. On the other hand, acting on glycoproteins remaining in ER, Endo H may remove the high mannose structure after the asparaginyl-N-acetyl-D-glucosamine on the a1 subunits of GABA A receptors. We carried out EMC3 or EMC6 siRNA transfection in HEK293T cells expressing a1b2g2 GABA A receptors; 48 h after transfection, proteins were extracted, and subjected to Endo H digestion and western blot analysis ( Figures 3E and 3F). The Peptide-N-Glycosidase F (PNGase F) enzyme-treated samples served as a control for unglycosylated a1 subunits ( Figures 3E and 3F, lane 5). Two endo H resistant a1 bands were observed, corresponding to singly or doubly glycosylated a1 at N38 and N138 (( Figures 3E and 3F, lane 5, lanes two and 4). Ratio of Endo H resistant a1 to total a1 represents its ER-to-Golgi trafficking efficiency. Such a ratio decreased to 0.43 with EMC3 siRNA ( Figure 3E, lane 2), comparing to a ratio of 0.59 of NT ( Figure 3E, lane 4). Similarly, the ratio decreased to 0.37 with EMC6 siRNA ( Figure 3F, lane 2), comparing to a ratio of 0.65 of NT ( Figure 3F, lane 4). Therefore, EMC3 and EMC6 have shown capability of positively regulating the stability of GABA A receptors ( Figure 3D) and promoting their anterograde trafficking from the ER to the Golgi post-translationally ( Figures 3E and 3F) and onward to the plasma membrane ( Figures 3A and 3C).
Because EMC3 and EMC6 are the core subunits of the EMC complex, depleting them was reported to disrupt the formation of the EMC complex in U-2 OS cells (Volkmar et al., 2019). Therefore, we evaluated the effect of EMC3 and EMC6 on the stability of the entire EMC complex, which could contribute to the regulation of GABA A receptor biogenesis in HEK293T cells. Knocking down EMC3 or knocking down both EMC3 and EMC6 using siRNAs reduced the protein levels of EMC1, EMC4, EMC7, and EMC10, which are EMC subunits with transmembrane domains, whereas knocking down EMC6 reduced the protein levels of EMC4 and EMC10 ( Figure 3G). Co-immunoprecipitation demonstrated endogenous interactions between EMC3/EMC6 and a number of major neuroreceptors, including GABA A receptors, N-methyl-D-aspartate receptors (NMDARs), and nicotinic acetylcholine receptors (nAChRs) To facilitate the biogenesis of its client membrane proteins, the EMC is expected to interact with them. Therefore, we continued to understand the interactions between EMC3/6 and GABA A receptors. Rat cortical neurons at DIV 12 were harvested to perform co-immunoprecipitation. Endogenous interactions were confirmed between a1 subunits and EMC3/EMC6 ( Figure 4A). In addition, as expected, pulling down a1 subunits led to the detection of b2/3 and g2 subunits of GABA A receptors, indicating the endogenous interactions within the pentameric receptors. Furthermore, because of the important role of the proteostasis network in orchestrating GABA A receptor biogenesis (Wang et al., 2022b), we observed a1-interacting chaperones, including BiP and calnexin, and ERAD factors, including Grp94 and VCP ( Figure 4A). Previously, we demonstrated the role of BiP and calnexin in assisting the folding and the role of Grp94 and VCP in facilitating the degradation of GABA A receptors (Di et al., 2013(Di et al., , 2016Han et al., 2015aHan et al., , 2015b. Furthermore, we evaluated whether the EMC has a more general role in interacting with other major endogenous neurotransmitter-gated ion channels, including nAChRs and NMDA receptors. Both GABA A receptors and nAChRs belong to the Cys-loop superfamily neuroreceptors, sharing a pentameric scaffold , whereas NMDA receptors are tetramers, consisting of two NR1 subunits and two NR2 subunits (NR2A or NR2B, or both) (Hansen et al., 2021). Dysfunction of these receptors leads to neurological, cognitive and psychiatric brain diseases (Shen et al., 2016a;Steinlein, 2012;Wanamaker et al., 2003). Co-immunoprecipitation experiments demonstrated that pulling down EMC3 or EMC6 resulted in the detection of NR1, NR2A, and NR2B subunits of NMDA receptors as well as a7 subunits of nAChRs in Figure 4. Continued ERAD factors (Grp94 and VCP). Neurons were plated onto 10-cm dishes at a density of one million per dish. At DIV 12, proteins were extracted for Co-IP. IgG was used as a negative control during the immunoprecipitation. n = 3. (B) Co-IP from primary rat cortical neurons demonstrated endogenous interactions between EMC3/EMC6 and a number of ion channels, including N-methyl-D-aspartate receptors (NMDARs, including NR1, NR2A and NR2B subunits) and nicotinic acetylcholine receptors (nAChR a7 subunit). n = 3. (C) Co-IP from mouse cortical homogenates, which were prepared from C57BL/6J mice between 8 and 10 weeks of age, demonstrated endogenous interactions between EMC3/EMC6 and selected ion channels. n = 3.
(D) Schematic of the primary sequence of EMC3 and EMC6. R31 and R180 in EMC3 and N22 and D27 in EMC6 were reported to influence the biogenesis of EMC-dependent client proteins.
(E) Mutation of R31A or R180A in EMC3 significantly reduced the interaction of EMC3 with GABA A a1 subunits. The cDNAs of FLAG-tagged EMC3, either in the wild type (WT) form or carrying appropriate mutations of R31A or R180A, were transiently transfected in HEK293T cells stably expressing a1b2g2 GABA A receptors. 48 h after transfection, proteins were extracted from cell lysates and incubated with anti-FLAG M2 magnetic beads. The immuno-purified eluents were separated through SDS-PAGE gel, and western blot analysis was performed to detect a1 subunits and FLAG. Quantification of the band intensity of a1 over FLAG after immunoprecipitation was shown on the right (n = 3).
(F) Mutation of D27A or N22A in EMC6 significantly reduced the interaction of EMC6 with GABA A R a1 subunits. Transfection of cDNAs was applied similarly as in E, however with co-application of FLAG-tagged EMC5 and EMC6 variants in HEK293T cells stably expressing a1b2g2 GABA A receptors. Co-IP and visualization of protein bands were carried out the same way as in E as well. Quantification of the band intensity of a1 over FLAG-tagged EMC6 after immunoprecipitation was shown on the right (n = 3).
(G) Significant increase of the interaction of SEC61a and a1 subunits of GABA A receptors was observed when both EMC3 and EMC6 were knocked down. We carried out siRNA transfection in HEK293T cells stably expressing a1b2g2 GABA A receptors; 48 h after transfection, proteins were extracted from cell lysates and incubated with anti-a1 antibody. The immuno-purified eluents were separated through SDS-PAGE gel, and western blot analysis was performed to detect SEC61a and a1 subunits. Quantification of the band intensity of SEC61a over a1 after immunoprecipitation was shown on the right (n = 3). Each data point is presented as mean G SEM *, p< 0.05; **, p< 0. iScience Article primary cortical neurons ( Figure 4B). In addition, mouse cortical homogenates from C57BL/6J mice between 8 and 10 weeks of age, which have developed mature nervous systems, were used to detect the endogenous interactions between EMC3/EMC6 and selected neuroreceptors in physiologically relevant conditions. Consistently, co-immunoprecipitation assay showed that EMC3 and EMC6 interacted with a1 and b2/b3 subunits of GABA A receptors, NR1 subunit of NMDA receptors, and a7 subunit of the nAChRs in cortical homogenates ( Figure 4C). These experiments successfully demonstrated the interactions between the EMC and a number of major neuroreceptors in both early and mature stages of development, suggesting that the EMC not only assists folding of GABA A receptors but also potentially plays critical roles for the biogenesis of other multi-pass transmembrane receptors in the CNS.
Mutating key transmembrane residues of EMC3 and EMC6 significantly impairs their capability of interacting with GABA A receptors Next, we determined whether the interactions between EMC3/6 and GABA A receptors are through their transmembrane domains. For EMC3, previous literature has shown the importance of positive residues at Arg31 (R31) in transmembrane helix 1 (TM1) and Arg180 (R180) in transmembrane helix 3 (TM3) ( Figure 4D), and the absence of such positive residues destabilize post-and co-translational insertion of EMC-dependent substrates, such as squalene synthase and opioid kappa receptor 1 (Pleiner et al., 2020). For that reason, we hypothesized that a neutral residue at such positions would likewise destabilize the interaction between EMC3 and GABA A receptors. FLAG-tagged WT EMC3 or EMC3 carrying appropriate mutations of R31A or R180A were transiently transfected in HEK293T cells stably expressing a1b2g2 GABA A receptors. The immuno-purified eluents were separated through SDS-PAGE, and western blot analysis was performed to detect a1 subunits and FLAG. As demonstrated by normalized a1 intensity over FLAG, comparing to WT EMC3, significant decrease of the interaction of mutant EMC3 and a1 subunits was observed, to an extent of 0.48 with R31A, or 0.40 with R180A, respectively ( Figure 4E). Therefore, the result indicated that a neutral residue replacement at R31 in TM1 and R180 in TM3 impairs EMC3's capability to interact with GABA A receptors.
Moreover, EMC6 residues in the hydrophilic vestibule are thought to be necessary for the insertion of its substrates ( Figure 4D) (Pleiner et al., 2020) Asp27 (D27) in TM1 has been shown with such possibility, and is conserved across several species such as Homo sapiens, M musculus, and Saccharomyces cerevisiae (Pleiner et al., 2020). In addition, Asn22 (N22) has been identified in TM1 of EMC6 within hydrogen bonding distance of the main chain of EMC5, which contributes to the stabilization of TM1 of EMC6 in the lipid bilayer. Therefore similarly, FLAG-tagged EMC5 and appropriate EMC6 variants were transfected in HEK293T cells stably expressing a1b2g2 GABA A receptors. Co-expression of EMC5 with EMC6 was necessary because EMC5 is required to stably insert EMC6's TM1 (Pleiner et al., 2020). Comparing to WT EMC6, its variants led to significant decrease of its interaction with a1 subunits, to an extent of 0.72 with D27A, or 0.55 with N22A, respectively ( Figure 4F). The result indicated that D27 and N22 in TM1 play critical roles for EMC6's interaction with GABA A receptors. Taken together, we showed that the interactions between EMC3/6 and GABA A receptors are dependent on the key charged/polar residues in the TM domains of the EMC, consistent with the role of the EMC as an insertase for multi-pass transmembrane proteins.
Furthermore, because the SEC61 translocon is known to play a crucial role in the insertion of secretory and membrane polypeptides into the ER co-translationally (Skach, 2009) and the EMC coordinates with SEC61 for the insertion of transmembrane proteins (Chitwood et al., 2018;O'Keefe et al., 2021), we investigated the potential orchestration of SEC61 and EMC on the GABA A receptor biogenesis. We carried out siRNA transfection of both EMC3 and EMC6 in HEK293T cells stably expressing a1b2g2 GABA A receptors. Coimmunoprecipitation experiments showed that significant increase of the interaction of SEC61a and a1 subunits of GABA A receptors was observed when both EMC3 and EMC6 were knocked down, as demonstrated by normalized SEC61a over a1 intensity comparing to NT, to a remarkable extent of 5.3-fold ( Figure 4G). Therefore, the result suggested that on depletion of EMC3 and EMC6, GABA A receptors would be routed to the SEC61 translocon for the insertion into the lipid bilayer.
To further evaluate the EMC's effects on the function of these variants, whole-cell patch-clamping was performed on cells 48 h after transfection using the Fluxion automated patch clamping instrument (see STAR Methods). Increased GABA-induced currents were recorded, as shown in a1(D219N)b2g2 ( Figure 5E), a1(G251D)b2g2 ( Figure 5F) and a1(P260L)b2g2 ( Figure 5G). With EMC3 overexpression, the peak current (Imax) increased from 239 pA to 355 pA and from 185 pA to 285 pA respectively in a1(G251D)b2g2 (Figure 5F) and a1(P260L)b2g2 ( Figure 5G); no significant change was observed in a1(D219N)b2g2 ( Figure 5E), potentially because of its lesion being in the ER lumen, whereas the EMC acts preferably on the transmembrane domains (Chitwood and Hegde, 2019). Moreover, with co-application of EMC5 and EMC6, Imax increased from 270 pA to 550 pA, 239 pA to 325 pA, and 185 pA to 370 pA respectively in a1(D219N) b2g2 ( Figure 5E), a1(G251D)b2g2 ( Figure 5F) and a1(P260L)b2g2 ( Figure 5G). We observed a modest increase in the peak current compared to the substantial increase in the surface expression of a1 subunit variants because the peak current is also determined by multiple electrophysiological parameters (Sallard et al., 2021). The macroscopic current amplitude at a given time point is the product of the number of open channels and their single channel conductance. The open states of GABA A receptors are also influenced by several kinetic parameters, such as activation, desensitization, and deactivation time scales. In addition, GABA A receptors on the cell surface could have binary forms (with a1 and b2 subunits) in addition to triheteromeric forms (with a1, b2 and g2 subunits); these two forms have different single channel conductance. Since the cell surface a1 subunits could have different opening probability and single channel conductance, their quantity is not strictly proportional to the observed macroscopic peak current amplitude. Such a discrepancy was also reported in GABA A receptors containing r1 variants (Hannan and Smart, 2018). Nonetheless, these results indicated that overexpression of EMC3 or EMC5/6 is sufficient to enhance the function of DAVs of GABA A receptors.

DISCUSSION
In this study, we systemically examined the effects of all 10 individual subunits of the EMC complex on regulating endogenous GABA A receptor protein expression. Significant reduction of the protein levels of GABA A receptor subunits, including a1, b2/b3, and g2 subunits, and GABA-induced currents were observed from knocking down EMC3 and EMC6 in neurons (Figures 1 and 2). Knocking down EMC3 or EMC6 reduced the protein levels of several other EMC subunits with transmembrane domains in HEK293T cells ( Figure 3G), Figure 5. Overexpression of EMC3, and EMC5 and EMC6 restores surface expression and whole-cell currents of disease-associated variants of GABA A receptors (A-C). Overexpression of EMC3 and EMC5/6 increased surface expression of a1 subunits of GABA A R in HEK293T cells stably expressing a1(D219N)b2g2 (A), a1(G251D)b2g2 (B) and a1(P260L)b2g2 (C). We carried out cDNA transfection of EMC3, or co-application of EMC5 and EMC6, in corresponding HEK293T cells; 48 h after transfection, surface proteins were enriched through biotin-neutravidin affinity purification, and western blot analysis was applied to detect a1 subunits. Na + /K + ATPase served as loading control of cell surface proteins. Normalized surface a1 band intensity was shown below the images (n = 3).
(D) HEK293T cells stably expressing a1(G251D)b2g2 GABA A receptors were transfected with empty vector control (CTL) or EMC5 and EMC6 cDNAs. 48 h after transfection, cycloheximide (CHX) (100 mg/mL), a potent protein synthesis inhibitor, was added to cell culture media for the indicated time. Cells were then harvested, and total proteins were subjected to SDS-PAGE and western blot analysis. Quantification of the a1 band intensity was plotted against the incubation time with CHX (n = 3).
(E-G) Increased whole-cell patch-clamping currents of GABA A receptors were recorded in HEK293T cells stably expressing a1(D219N)b2g2 (E), a1(G251D) b2g2 (F) and a1(P260L)b2g2 (G). Transfection of cDNA was applied the same way as in (A-C); 48 h after transfection, patch clamping was performed on the cells with the IonFlux Mercury 16 ensemble plates at a holding potential of À60 mV. GABA (100 mM) was applied for 4 s, as indicated by the horizontal bar above the currents. The peak currents (Imax) were acquired and analyzed by Fluxion Data Analyzer (n = 6 -10). Each data point is presented as mean G SEM *, p< 0.05; **, p< 0.01. CTL: Empty vector control sample.

OPEN ACCESS
iScience 25, 104754, August 19, 2022 13 iScience Article consistent with the report that depleting core EMC subunits disrupts the EMC complex in U-2 OS cells (Volkmar et al., 2019). It appears that depleting EMC3 downregulated more EMC subunits than depleting EMC6. Therefore, on one hand, individual EMC subunits could stabilize the EMC complex to a different extent, causing their differentiating influence on the protein levels of GABA A receptors. On the other hand, because overexpressing EMC3 or EMC5/6 was sufficient to increase the functional surface expression of several GABA A receptor variants ( Figure 5), it is also possible that EMC3 and EMC6 could execute their functional role on GABA A receptor folding and trafficking independent of the formation of the EMC complex, which would suggest a subunit-specific contribution of the EMC in regulating the biogenesis of a multi-pass transmembrane protein. Intriguingly, consistent results have been reported, showing that the EMC3-EMC6 fusion protein is sufficient to insert the mitochondrial protein Cox2 and nuclear encoded inner membrane proteins (Gü ngö r et al., 2022), suggesting that certain EMC subunits are capable of carrying out the membrane insertase function. EMC3 is homologous to known membrane protein insertases Oxa1 family proteins (Volkmar and Christianson, 2020;Wideman, 2015), and depletion of EMC3 results in ER stress and activates the unfolded protein response (Huang et al., 2021;Tang et al., 2017). Moreover, EMC3 was reported to coordinate the assembly of lipids and proteins required for surfactant synthesis (Tang et al., 2017), maintain differentiation and function of intestinal exocrine secretory lineages (Huang et al., 2021), and play a critical role in mammalian retinal development Zhu et al., 2020). In parallel, in addition to its role in regulating the biogenesis of acetylcholine receptors in C. elegans (Richard et al., 2013), EMC6 was reported to regulate the autophagosome formation, and knocking down EMC6 impaired autophagy (Li et al., 2013). Moreover, the important role of EMC6 as a therapeutic target for cancer and pancreatic inflammatory diseases has been increasingly recognized (Shen et al., 2016b;Tan et al., 2020;Wang et al., 2017;Xiao et al., 2021). Nonetheless, the potential subunit-specific contribution of the EMC to the membrane protein quality control merits further investigation.
Our data support a working model about the role of the EMC in GABA A biogenesis in the ER (Figure 6). The EMC, including EMC3 and EMC6, interacts with the nascent subunits of GABA A receptors through the transmembrane domains. Mechanistic studies revealed that R31 in TM1 and R180 in TM3 of EMC3, and D27 and N22 in TM1 of EMC6 are essential residues for EMC's interaction with GABA A receptors ( Figures 4D-4F), consistent with the structural work of the EMC (Pleiner et al., 2020). Therefore, the EMC facilitates the insertion of the transmembrane domains of GABA A receptor subunits into the lipid bilayer. Subsequently, molecular chaperones, such as BiP and calnexin, promote the productive folding of GABA A receptors (Di et al., 2013;Han et al., 2015b). After the proper assembly of the pentameric receptors on the ER membrane, GABA A receptors engage the trafficking factors, such as LMAN1 (Fu et al., 2019), en route to the Golgi and onward to the plasma membrane. Consistently, we demonstrated that EMC3 and EMC6 promote the anterograde trafficking of GABA A receptors ( Figure 3). Here, we focused on GABA A receptor subtypes in the synaptic sites containing the a1 subunit. Given the substantial diversity of GABA A receptor subunits (a1-6, b1-3, g1-3, d, ε, q, p, and r1-3), numerous subtypes of GABA A receptors containing the combination of subunits can form either in the synaptic sites to trigger fast, transient phasic inhibition or in the extrasynaptic sites to generate persistent tonic inhibition (Farrant and Nusser, 2005). The a1-a3 iScience Article subunits are primarily synaptic, whereas a4, a6, and d subunits are predominantly extrasynaptic (Lorenz-Guertin and Jacob, 2018). Tonic receptors have high affinity to GABA to respond to its ''spill over'' with low conductance. Since synaptic receptors and extrasynaptic receptors have different subunit composition and functional characteristics, it would be of great interest to determine whether the EMC complex plays similar roles in regulating the biogenesis of phasic and tonic GABA A receptors in future studies. Furthermore, since EMC3 and EMC6 interact with major endogenous neurotransmitter-gated ion channels in primary neurons, including Cys-loop receptors and glutamate receptors ( Figures 4A-4C), the EMC could have a general role in the CNS. It would be of great interest to identify the endogenous interactomes of the EMC subunits, such as EMC3 and EMC6, as a way to determine their client membrane proteins in the CNS in the future.
Loss of function of inhibitory GABA A receptors is one of the primary causes of genetic epilepsy. Recent advance in genetics has identified over 150 epilepsy-associated variants in the major subunits (a1, b2, b3, and g2) of GABA A receptors (Absalom et al., 2022;Fu et al., 2022;Hernandez and Macdonald, 2019;Hirose, 2014). Despite the development of numerous anti-seizure drugs, about one-third epilepsy patients are resistant to current drug treatment (Rincon et al., 2021), and many of them have genetic causes (Smolarz et al., 2021). Therefore, there is an urgent need to develop new therapeutic strategy to treat epilepsy, especially drug-resistant epilepsy. Because one major disease-causing mechanism for loss of function of GABA A receptors is their reduced trafficking to the plasma membrane, one promising approach is to adapt the proteostasis network to restore their surface trafficking and thus function (Di et al., 2013(Di et al., , 2021Fu et al., 2018;Han et al., 2015aHan et al., , 2015b. Recently, we showed that pharmacological activation of the ATF6 arm of the unfolded protein response using small molecules AA147 and AA263 restored the functional surface expression of misfolding-prone GABA A variants containing a1(D219N), g2(R82Q), or g2(R177G) by increasing the protein levels of pro-folding chaperones, such as BiP, and the interactions between these chaperones and GABA A variants; importantly, ATF6 activators are more selective for folding defective variants compared to wild type receptors (Wang et al., 2022a). Furthermore, such a selectivity was also reported when other small molecule proteostasis regulators, such as verapamil, dinoprost, and dihydroergocristine, were applied to rescue misfolding-prone GABA A variants (Di et al., 2021;Han et al., 2015b), indicating that adapting the proteostasis network has the promise to achieve selectivity for disease-associated variants. Here, we demonstrated that overexpression of EMC3 and EMC5/6 enhances the functional surface expression of a number of folding-deficient GABA A receptor variants ( Figure 5). Because EMC3 and EMC5/6 interact with both inhibitory GABA A receptors and excitatory NMDARs and nAChRs ( Figures 4A-4C), the EMC complex has the potential to regulate the biogenesis of a number of neuroreceptors that regulate excitation-inhibition balance. Therefore, in order to evaluate the suitability of the EMC as the therapeutic target to rescue pathogenic GABA A receptors, future research is needed to determine whether activating the EMC is capable of achieving selectivity to correct the function of misfolding-prone GABA A receptor variants over wild type receptors.

Limitations of the study
The current study did not elucidate how the EMC coordinates with other proteostasis network components to orchestrate the biogenesis of GABA A receptors in the ER. In addition, the potential general role of the EMC in regulating the protein quality control of neurotransmitter-gated ion channels in the CNS needs to be further investigated.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:   iScience Article at 1000 rpm and then DMEM was removed while avoiding the pellet. Cell pellets were washed with DPBS and centrifuged again at 1000 rpm for 3 min. The DPBS was removed, and pellets were stored on ice during transport to a À80 C freezer. Cells were lysed with lysis buffer (50 mM Tris, pH 7.5, 150 mM NaCl, and 2 mM n-Dodecyl-B-D-maltoside (DDM) (GoldBio, catalog #: DDM5)) supplemented with complete protease inhibitor cocktail (Roche). Cells were vortexed for 30 s followed by ultrasonication for 30 s for three times. Then they were centrifuged at 15,000 3 g, 4 C for 10 min to obtain the supernatant as total proteins. The protein concentration was measured according to Thermo Fisher MicroBCA kit protocol. Cell lysates were loaded with Laemmli sample buffer (Biorad, catalog #: 1610747) with b-mercaptoethanol (1:10 v/v) and separated through SDS-PAGE gel.

AUTHOR CONTRIBUTIONS
Before proceeding to western blot, gels were made ranging from 8% to 20% resolving gel depending on the size of the protein with 4% stacking gel on top. (v/v) Tween20 (Sigma, catalog #: P7949-500mL). They were further incubated in 5% non-fat milk powder (Nestle Carnation, catalog #: 43875) in TBS-T for 30 min to 2 h. Following the blocking step, the membranes were incubated in 1% milk with the primary antibody added starting at 1:1000 dilution and adjusted accordingly on subsequent runs. The following day, the membrane was washed 3 times with TBS-T for 10 min each and incubated with their secondary antibody (1:10,000) for 1 h. This was followed by 3 more washed with TBS-T. Afterward gels were exposed with Pico PLUS (catalog #: 34578) or Femto (catalog #: 34096) SuperSignal West chemiluminescent substrates from Thermo Scientific for three minutes. After using different exposure times to get optimal images, results were analyzed to quantify band intensity using ImageJ software from the NIH.
Co-immunoprecipitation (Co-IP) iScience 25, 104754, August 19, 2022 iScience Article Sprague Dawley rat E18 brain cortex tissues were obtained from BrainBits, with provided Hibernate EB (HEB) and NbActiv1 media (Springfield, IL). Prior to plating cells, 10-cm dishes or coverslips (Fisher, catalog #: 12-545-80 CIR-1) in a 24-well plate were coated with 4 mL per plate or 500 mL per well of 50 mg/ml poly-Dlysine (PDL, Sigma, catalog #: P6407) at 4 C overnight. The next day PDL was removed and the plates were washed with sterile distilled water two times. Then they were coated with the same volumes of 5 mg/mL Laminin (Sigma, catalog #: L2020) overnight at 4 C. The plates were then put into a 37 C cell incubator while preparing the neurons for 1 h and Laminin was removed immediately before plating neurons.
Neurons were extracted according to instructions from BrainBits. Briefly, connective tissues of cortex were digested with 2 mL of 2 mg/mL of papain (Brain Bits, catalog #: PAP/HE-Ca) at 30 C water bath for 15 min, gently swirling every 5 min. Papain was then removed without disturbing the tissue at the bottom of the tube. The HEB media was then added to the tissue tube. The salinized Pasteur pipette was then used to triturate 30 times very slowly and carefully to not add air bubbles with the tip in the middle of the tube. The undispersed pieces were allowed to settle for one minute. The supernatant with dispersed cells were then transferred to a sterile 15 mL tube. This was spun down at 200 3 g for one minute and the supernatant was discarded. Afterwards 1-2 mL of NbActiv1 was added while carefully avoiding air bubbles.
Neurons were plated at 1 million per 10-cm dish or 25,000 per well of 24-well plate. The neuronal culture media contained the following: 100 mL Neurobasal A (Gibco, catalog #: 21103049), 2 mL B27 (Gibco, catalog #: 17504044), 0.25 mL GlutaMAX (Gibco, catalog #: 35050061) and 1 mL Penicillin-Streptomycin (Hyclone, catalog #: sv30010). Additionally, cytosine b-D-arabinofuranoside hydrochloride (Ara-C hydrochloride) (Sigma, catalog #: C6645) was added from day-in-vitro (DIV) 3 at a final concentration of 2 mM each time as half of the media was changed every three days. At DIV 6, lentivirus transduction was carried out to the neurons as indicated; the multiplicity-of-infection (MOI) was at 10, that is, the ratio of lentivirus count to neuron cells count in each well. At DIV 12, neurons were subjected to protein extraction for co-IP or immunofluorescence staining for confocal microscopy.

Mouse brain homogenization
C57BL/6J mice (Jackson laboratory) at 8-10 weeks were sacrificed, and the cortex was isolated and stored at À80 C. On the day of experiments, the cortex was thawed on ice and homogenized in homogenization buffer (25mM Tris-HCl pH 7.5, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS supplemented with Roche complete protease inhibitor cocktail) using a plastic micro tissue homogenizer, and then sheared by passing 10 times through a 21G needle. Homogenates were centrifuged at 800 g for 10 min at 4 C and supernatants were collected. The pellet was re-homogenized in additional homogenization buffer and centrifuged at 800 3 g for 10 min at 4 C, and the supernatant was collected again. The supernatants were combined and rotated at 4 C for 2 h, and then centrifuged at 18,000 3 g for 30 min at 4 C. The resulting supernatant was collected as mouse brain homogenate, and its protein concentration was determined by a MicroBCA assay. The animal studies followed the guidelines of the Institutional Animal Care and Use Committees (IACUC) at Case Western Reserve University.

Confocal immunofluorescence
To label cell surface GABA A receptors, primary neurons that were cultured on coverslips, were fixed with 2% paraformaldehyde in DPBS, blocked with goat serum for 0.5 h at room temperature, and labeled with 100 mL of appropriate anti-a1 (Synaptic Systems, catalog #: 224203), b2/3 (Millipore, catalog #: 05-474), or g2 (Synaptic Systems, catalog #: 224003) antibodies (1:200) for 1 h without detergent permeabilization. Afterwards, they were incubated at room temperature with 500 mL (1:400) of Alexa 594-conjugated goat anti-rabbit antibody (ThermoFisher, catalog #: A11037), or Alexa 594-conjugated goat anti-mouse antibody (ThermoFisher, catalog #: A11032) for 1 h. Afterwards, cells were permeabilized with saponin (0.2%) for 5 min and incubated with DAPI (1 mg/mL) for 3 min to stain the nucleus. The coverslips were then mounted and sealed. For confocal immunofluorescence microscopy, an Olympus IX-81 Fluoview FV1000 confocal laser scanning system was used. A 603 objective was used to collect images using FV10-ASW software. Quantification of the fluorescence intensity was achieved using the ImageJ software from the NIH.

Biotinylation of cell surface proteins
To perform biotinylation assay, 6-cm dishes were coated in a 300-fold dilution of poly-L-lysine (Fisher, catalog #: ICN15017710) at 2 mL per plate for one hour at 37 C. Plates were rinsed twice with DPBS and allowed to air dry. The cells were plated on the coated dishes and incubated for 48 h post-transfection.

OPEN ACCESS
iScience 25, 104754, August 19, 2022 25 iScience Article The intact cells were washed one time with ice-cold DPBS and incubated with the membrane impermeable biotinylation reagent Sulfo-NHS SS-Biotin (0.5 mg/mL; APExBio, catalog #: A8005) in PBS containing 0.1 mM CaCl 2 and 1 mM MgCl 2 (PBS+ CM) for 30minat 4 C to label surface membrane proteins. In order to quench the reaction, the cells were incubated with 1.5 mL of 50 mM glycine in ice-cold DPBS-CM twice for 5 min at 4 C. They were then washed twice with DPBS. The sulfhydryl groups were then blocked by incubating the cells with 5 nM N-ethylmaleimide (NEM) in DPBS-CM for 15 min at room temperature and then the liquid was removed. Cells were scraped off plates and solubilized overnight at 4 C in lysis buffer (Tris-HCl, 50 mM; NaCl, 150 mM pH 7.5, 2 mM DDM) supplemented with Roche complete protease inhibitor cocktail and 5 mM NEM. The next day the lysates were centrifuged at 16,000 3 g for 10 min at 4 C to pellet cellular debris. The supernatant was saved, and the protein concentration was measured with a MicroBCA assay. Biotinylated surface proteins were affinity-purified from the above supernatant by incubating for 2hat 4 C with 50 mL of immobilized neutravidin-conjugated agarose bead slurry (Fisher, catalog #: PI29200). The samples were then subjected to centrifugation (5,000 3 g, 3 min). The beads were washed three times with buffer (Triton X-100, 1%; Tris-HCl, 50 mM; NaCl, 150 mM, pH 7.5) and three times further without Triton X-100. Surface proteins were eluted from beads by incubating for 30minat room temperature with 80 mL of LSB ⁄ Urea buffer (2x Laemmli sample buffer (LSB) with 100 mM DTT and 6 M Urea, pH 6.8) for SDS-PAGE and Western blotting analysis.

Cycloheximide-chase assay
Cycloheximide-chase assay to evaluate protein degradation in cells was carried out according to published procedures (Di et al., 2016). Briefly, cycloheximide (100 mg/mL), a potent protein synthesis inhibitor, was added to cell culture medium. HEK293T cells were then chased for the indicated time and harvested, and total proteins were extracted and subjected to SDS-PAGE and Western blot analysis.

Endoglycosidase H (endo H) enzyme digestion assay
To remove asparaginyl-N-acetyl-D-glucosamine in the N-linked glycans incorporated on the a1 subunit in the ER, total cell lysates were digested with Endo H enzyme (NEBiolab, catalog #: P0703L) with G5 reaction buffer at 37 C for 1h. The Peptide-N-Glycosidase F (PNGase F) enzyme (NEBiolab, catalog #: P0704L) treated samples served as a control for unglycosylated a1 subunits. Treated samples were then subjected to Western blot analysis.

QUANTIFICATION AND STATISTICAL ANALYSIS
All data are presented as mean G SEM. Statistical significance was evaluated using Student's t-test if two groups were compared and one-way ANOVA followed by post-hoc Tukey test if more than two groups were compared. A p< 0.05 was considered statistically significant. *, p< 0.05; **, p< 0.01.

ll
OPEN ACCESS