GSG1L-containing AMPA receptor complexes are defined by their spatiotemporal expression, native interactome and allosteric sites

Perozzo, Amanda M.; Schwenk, Jochen; Kamalova, Aichurok; Nakagawa, Terunaga; Fakler, Bernd; Bowie, Derek

doi:10.1038/s41467-023-42517-7

Download PDF

Article
Open access
Published: 26 October 2023

GSG1L-containing AMPA receptor complexes are defined by their spatiotemporal expression, native interactome and allosteric sites

Nature Communications volume 14, Article number: 6799 (2023) Cite this article

1418 Accesses
1 Citations
Metrics details

Subjects

Abstract

Transmembrane AMPA receptor regulatory proteins (TARPs) and germ cell-specific gene 1-like protein (GSG1L) are claudin-type AMPA receptor (AMPAR) auxiliary subunits that profoundly regulate glutamatergic synapse strength and plasticity. While AMPAR-TARP complexes have been extensively studied, less is known about GSG1L-containing AMPARs. Here, we show that GSG1L’s spatiotemporal expression, native interactome and allosteric sites are distinct. GSG1L generally expresses late during brain development in a region-specific manner, constituting about 5% of all AMPAR complexes in adulthood. While GSG1L can co-assemble with TARPs or cornichons (CNIHs), it also assembles as the sole auxiliary subunit. Unexpectedly, GSG1L acts through two discrete evolutionarily-conserved sites on the agonist-binding domain with a weak allosteric interaction at the TARP/KGK site to slow desensitization, and a stronger interaction at a different site that slows recovery from desensitization. Together, these distinctions help explain GSG1L’s evolutionary past and how it fulfills a unique signaling role within glutamatergic synapses.

Gating and modulation of a hetero-octameric AMPA glutamate receptor

Article 02 June 2021

AMPA receptor anchoring at CA1 synapses is determined by N-terminal domain and TARP γ8 interactions

Article Open access 23 August 2021

GluD1 is a signal transduction device disguised as an ionotropic receptor

Article 16 June 2021

Introduction

Claudins were first discovered in chicken liver as being integral membrane proteins necessary for the formation of tight junctions¹. Since then, an extended family of claudin proteins has been identified that fulfills diverse roles in numerous cell types and organisms from nematodes to humans². In the mammalian brain, claudins have evolved to perform specialized roles at glutamatergic synaptic junctions where they regulate neurochemical transmission mediated by the excitatory neurotransmitter, L-glutamate (L-Glu)^3,4. Critical claudin proteins found at these synapses include AMPA receptor (AMPAR) auxiliary proteins such as the Type I and II transmembrane AMPAR regulatory proteins (TARPs)⁵, as well as germ cell-specific gene 1-like protein (GSG1L), a distant homolog of TARPs^6,7,8. TARPs and GSG1L assemble with the pore-forming subunits of AMPARs, which promotes the forward trafficking of receptors to the plasma membrane and modifies both their gating behavior^9,10 and pore properties^11,12,13,14.

TARPs and GSG1L are structurally related, both possessing the archetypical claudin membrane topology of cytoplasmic N- and C-termini, four transmembrane domains and two extracellular loops, with the length of the first loop (ExL1) being the primary distinction between them^15,16. Despite a similar overall topology, TARPs and GSG1L tend to exert opposing positive or negative effects on AMPARs, respectively, in terms of recovery from desensitization^6,7, ion permeation¹⁷, and synaptic transmission^18,19,20. These functional distinctions most likely reflect their evolutionary past, with GSG1L emerging later than TARPs as an AMPAR auxiliary subunit⁸. How their marked functional differences can be explained in structural terms, however, has evaded any clear explanation. Although some aspects of TARP modulation of AMPAR gating have been tied to the KGK motif, an evolutionarily-conserved regulatory site on the lower D2 lobe of the AMPAR ligand-binding domain (LBD)^21,22,23, direct experimental evidence for a bona fide regulatory site for GSG1L is lacking.

Proteomic and structural studies of recombinant and native AMPAR complexes have established that TARPs associate with AMPAR tetramers in either a 2:4 or 4:4 arrangement^{6,22,24,25,26,27} and often partner with cornichons (CNIHs) in hetero-octameric assemblies, especially in the hippocampus^28,29,30,31. In contrast, only two GSG1L subunits have been proposed to bind per tetramer, at least in cryo-EM studies of recombinant protein^15,32. Based on short-term plasticity observed at corticothalamic synapses, there is an emerging view that individual neurons may establish glutamatergic synapses that are specific for GSG1L¹⁹; however, the native composition of AMPAR-GSG1L complexes remains to be fully explored.

Here, we present an interdisciplinary and comprehensive study which addresses the spatiotemporal expression profile, native interactome, and functional properties of AMPAR-GSG1L complexes. Unlike TARPs, which are highly expressed across most brain regions throughout development^5,33,34, we find that GSG1L expression is restricted to specific regions and developmental stages in the rodent brain. High-resolution proteomic analyses on native receptor complexes reveal that GSG1L assembles into a unique set of low abundance AMPARs with distinct subunit composition. Contrary to conventional understanding, we show that the main actions of GSG1L are not mediated via the KGK motif, but rather through a separate, evolutionarily-conserved allosteric site. Furthermore, we demonstrate that AMPAR-TARP and AMPAR-GSG1L assemblies rely on subunit coordination between pore-forming and auxiliary subunits, with a dominant role for GluA2, to fine-tune channel gating.

Results

GSG1L expression is restricted to specific brain regions and developmental timepoints in the rat brain

Recent work has shown that GSG1L is highly expressed in the anterodorsal (AD) and anteroventral (AV) nuclei of the anterior thalamus (AT) throughout development, where it modulates short-term plasticity¹⁹. However, a detailed spatiotemporal expression profile of GSG1L remains to be fully characterized. To systematically examine the expression of GSG1L in the brain from P14 to P240, we used a transgenic rat reporting lacZ expression under the control of the endogenous GSG1L promoter^19,35. Subsequent chromogenic reaction by X-gal produced a dark blue stain, which provides a visual readout of GSG1L promoter activity. The lacZ signal therefore serves as a proxy for GSG1L expression (Fig. 1 and Supplementary Fig. 1).

**Fig. 1: GSG1L has a dynamic, region-specific expression pattern throughout development, with increased expression in the adult rat brain.**

Intact whole brains of homozygous GSG1L knockout (KO) rats at P14, P24, P42, and P120 were compared with corresponding wildtype (WT) littermates taken as negative controls (Fig. 1a). As expected, no staining was observed in WT controls. By contrast, GSG1L reporter rat brains displayed an evident developmental increase in surface cortical expression of lacZ, with lowest levels at P14 and a continual increase until P120, which was the last timepoint tested. Interestingly, lacZ expression was restricted to defined regions of the cortex, forming a clear boundary of lacZ activity which extended caudally during development (Fig. 1a). Notably, we observed strong lacZ expression in the external granular layer (EGL) of the cerebellum at P14 (Fig. 1a). The EGL contains precursors of granule cells that migrate and differentiate by P20 into the internal granular layer of the cerebellum³⁶. Indeed, at later developmental timepoints, we observed high lacZ expression in the granular layer (GL) of the cerebellum (Fig. 1c). Taken together, whole brain staining of the GSG1L KO reporter rat suggests that GSG1L expression levels increase during development, except for limited brain regions, such as the cerebellum, where expression is persistent.

To maximize X-gal penetration and examine GSG1L expression in deeper brain structures, we next generated 300 μm-thick coronal sections from P14, P21, P60, P180, and P240 GSG1L KO brains and stained with X-gal post-sectioning (Fig. 1b and Supplementary Fig. 1). Remarkably, lacZ activity was localized to cortical layer 2/3 (L2/3) neurons, with a clear developmental delay. Staining representative of GSG1L was undetectable in L2/3 at P14 and P21 but was observed at P60 with highest levels at P180-P240, consistent with in situ hybridization data³⁷. In the striatum, we found a similar age-dependent increase in lacZ expression with undetectable levels at P14 and P21. At P60, there was strong staining in the caudate putamen (CP) persisting throughout and up to P240. LacZ staining was restricted to the CP with undetectable levels in the globus pallidus (GPe) (Fig. 1b). Our findings reveal that developmental expression of GSG1L is non-uniform across the cortex; while in L2/3 it is restricted to adults, staining in the endopiriform nucleus (EPN) and infralimbic area (ILA) is present throughout postnatal development (Fig. 1b).

Based on the prominent striping pattern observed in the cerebellum of intact whole brains (Fig. 1a, ref. P42), lacZ expression in the cerebellum was further investigated, here using sagittal sections at P14, P21, and P60 (Fig. 1c and Supplementary Fig. 1). We observed distinct lacZ activity in the granular layer (GL) of the cerebellum, while signals were undetectable in the molecular layer (ML) (Fig. 1c). Strikingly, lacZ expression was restricted to certain cerebellar lobules, with distinct segregation to anterior lobes at all three developmental timepoints. Altogether, these observations demonstrate that GSG1L exhibits brain-region, and possibly cell-type-specific, expression during development with global levels increasing throughout maturation.

Composition of AMPAR-GSG1L complexes in the mature rat brain

Following characterization of the expression profile of GSG1L, we next aimed to identify the precise molecular constituents of native GSG1L-containing AMPAR assemblies. In the mammalian central nervous system (CNS), AMPARs are multiprotein signaling complexes located at extrasynaptic and synaptic sites³⁸. The core complex is composed of four pore-forming GluA proteins and up to four auxiliary subunits (TARPs, CNIHs, and/or GSG1L) that bind to two pairs of binding sites^6,24,29. While it is known that TARP γ8 and CNIH2 co-assemble into the same AMPAR complex, at least in the hippocampus^28,31, the binding partners of GSG1L-containing AMPARs remain largely unexplored. Therefore, we studied the set of interaction partners of GSG1L (interactome; Fig. 2) in the rodent brain using our established multi-epitope affinity purification-mass spectrometry (meAP-MS) approach⁶ (see Methods).

**Fig. 2: GSG1L is integrated in a subpopulation of native AMPARs with distinct composition.**

Affinity isolation of GSG1L-containing protein complexes from solubilized brain membranes of WT rats (P59) was performed with three different anti-GSG1L antibodies. Source material from GSG1L KO rats served as a stringent negative control. We determined the abundances of all affinity purified proteins and calculated normalized abundance ratios between WT and KO to evaluate the specificity of co-purifications (i.e., target-normalized ratios (tnRs)). Data from all experiments visualized by t-distributed stochastic neighbor embedding (t-SNE) uncovered the proteins that consistently and specifically co-purified with GSG1L by their clustered co-localization with the target (Fig. 2a and Supplementary Fig. 2). Accordingly, GSG1L displayed robust and exclusive interaction with known constituents of the AMPAR interactome⁶: the pore-forming GluA1-4 proteins, a select set of auxiliary subunits (TARPs γ2, γ3, γ8, CNIH2) and three distinct constituents of the receptor periphery (cysteine‐knot AMPAR modulating protein 44 (CKAMP44 or Shisa9), proline-rich transmembrane protein 1 (PRRT1), and leucine-rich repeat transmembrane protein 4 (LRRT4)). Noteworthy, none of the AMPAR interactome constituents involved in ER-located biogenesis, including FRRS1L, CPT1c or ABHD6^30,39, were co-purified, suggesting that GSG1L can only bind either to GluA tetramers ready for ER exit or to AMPAR assemblies inserted into the plasma membrane at synaptic/extrasynaptic sites.

For more refined evaluation of the composition of GSG1L-containing AMPARs, the MS-derived abundance values of the main interactors were plotted relative to that of GSG1L (Fig. 2b). GluA2 appeared in an almost 1:1 ratio with GSG1L, while GluA1 and GluA3 were less abundant, pointing towards co-assembly of GSG1L into di- or tri-heteromers as reported in cryo-EM analysis of native AMPARs²⁹. Moreover, we estimated the ‘average stoichiometry’ of GSG1L in AMPARs by determining the abundance ratio of GSG1L and GluA tetramers (sum of all co-purified GluA1-4 proteins divided by four). These results indicated a stoichiometry close to two (1.74 ± 0.31), consistent with structural studies on recombinant GSG1L-bound receptors (e.g.,³²). The lower abundances determined for TARPs and CNIH2, determined under conditions that are known to preserve stoichiometries⁶ (CL-47 buffer), suggest that although GSG1L may form ‘conventional’ hetero-octameric assemblies with these auxiliary subunits^6,30, it also assembles as the sole inner core constituent in TARP/CNIH-free AMPARs (Fig. 2b).

The relative contribution of GSG1L-containing AMPARs in the rat brain was further analyzed by two-step APs. First, the near-complete pool of GSG1L-containing AMPARs was isolated with anti-GSG1L antibodies, and subsequently, a mixture of anti-GluA1-4 antibodies extracted the entirety of the remaining AMPARs. The quantitative evaluations of MS-analysed AP eluates showed that GSG1L participates in about 5% of all AMPARs in the mature brain (Fig. 2c). Of note, this subpopulation has a particular composition as LRRT4, a known postsynaptic cell adhesion molecule involved in synapse formation⁴⁰, showed preferred binding to GSG1L-containing AMPARs (Fig. 2c).

Members of the claudin superfamily differentially modify AMPAR desensitization

Given the somewhat unique distribution and composition of native AMPAR-GSG1L complexes, we next wanted to further probe the functionality of isolated receptors. To date, the structural basis underlying differential regulation of AMPARs by TARPs (Fig. 3a) and GSG1L (Fig. 3b) is poorly understood. Therefore, we sought to resolve how modulation of AMPAR gating by these claudin proteins can be distinguished from a structure-function perspective. To do so, we first expressed Type I (γ2) and Type II (γ7) TARPs, as well as GSG1L, with homomeric GluA2Q_flip AMPARs in HEK293 cells and measured the rate of onset and the extent of desensitization (Fig. 3 and Supplementary Table 1). For comparison, we studied a different class of auxiliary protein, CKAMP44, which also affects AMPAR desensitization⁴¹.

**Fig. 3: TARPs and GSG1L differentially modify AMPAR desensitization.**

Consistent with previous reports^42,43,44,45, the prototypical TARP, γ2, concomitantly slowed entry into and attenuated equilibrium desensitization. Rates into desensitization of GluA2 AMPARs with and without γ2 were best fit as the sum of two exponential functions, with the more prominent fast component corresponding to 8.2 ± 0.4 ms (97.7% of response) for GluA2 alone and 12.0 ± 1.1 ms (70.7% of response) for GluA2/γ2 (Fig. 3c, d and Supplementary Table 1). Equilibrium desensitization (i.e., steady-state/peak response) significantly increased by 15-fold from 1.4 ± 0.2% for GluA2 alone (n = 20) to 22.2 ± 1.9% with γ2 (n = 23) (Fig. 3e).

Similar to TARP γ2, γ7 and GSG1L both slowed entry into AMPAR desensitization, albeit to a more modest degree (Fig. 3c)^{7,21,46,47,48}. The more robust effect of γ7 was to enhance equilibrium desensitization, resulting in a 4-fold increase relative to GluA2 alone, whereas GSG1L had no appreciable effect (Fig. 3e and Supplementary Table 1). Although the time constants of the fast component of desensitization were similar between GluA2 alone compared to A2 + γ7 and A2 + GSG1L, the contribution of the slow component increased by 3-fold (to 7.1%) and 14-fold (to 33.0%) for γ7 and GSG1L, respectively (Fig. 3d and Supplementary Table 1). TARPs γ2 and γ7, as well as GSG1L, exerted a modest slowing of GluA2 deactivation kinetics (Supplementary Table 1). In contrast, CKAMP44 behaved in an opposite manner to the claudin-related proteins, inducing profound receptor desensitization, as noted previously^49,50. Co-assembly of GluA2 with CKAMP44 accelerated desensitization kinetics by 1.7-fold to 4.8 ± 0.2 ms (n = 11) (Fig. 3d). The speeding up of desensitization kinetics was accompanied by a significant 6-fold reduction in the equilibrium response to 0.25 ± 0.05% of the peak (n = 11) (Fig. 3e).

Taken together, these data establish that while all members of the claudin superfamily studied here slow desensitization rates of GluA2 AMPARs, albeit to different degrees, they do not all attenuate equilibrium desensitization (i.e., GSG1L). By contrast, CKAMP44 exerts an opposing modulation on channel gating, enhancing desensitization.

Both TARPs and GSG1L attenuate AMPAR desensitization through the KGK site

Previous work has shown that mutation of the evolutionarily-conserved KGK motif on the lower lobe of the LBD (Fig. 4a–c) disrupts TARP γ2-dependent slowing of AMPAR entry into desensitization²¹. Whether the KGK motif serves as a common regulatory site for other AMPAR auxiliary proteins, particularly claudins that also attenuate desensitization, has yet to be explored. We therefore compared how mutation of the KGK site (i.e., KGK to single D, herein termed 3D)²¹ impacts the gating kinetics of AMPARs in complex with TARPs γ2 and γ7, GSG1L, and CKAMP44 (Fig. 4d–f and Supplementary Table 1).

**Fig. 4: Claudin-related proteins attenuate AMPAR desensitization through the KGK motif.**

As expected²¹, the 3D mutation significantly attenuated the effect of TARP γ2 on the equilibrium response, resulting in a 4-fold reduction to 5.3 ± 0.7% (n = 25) (Fig. 4d, e) and also disrupted the slowing of desensitization kinetics by γ2 (Fig. 4d, f). Noted previously²¹, the 3D mutant itself exhibited faster decay kinetics relative to A2 WT receptors (Fig. 4f and Supplementary Table 1) but indistinguishable steady-state/peak responses (Fig. 4e). For this reason, we also compared the desensitization kinetics of 3D receptors bound by auxiliary subunits with 3D receptors alone (Fig. 4f). In keeping with this, the fast component of desensitization exhibited by 3D/γ2 receptors was 7.9 ± 0.7 ms (n = 25), remaining similar to 3D receptors in the absence of TARP γ2 (5.9 ± 0.2 ms, n = 18) (Fig. 4f), and the slow component was nearly 2-fold faster (Supplementary Table 1). The same general trend was observed for other claudin-related proteins, whereby mutation of the KGK site diminished the effects of both γ7 and GSG1L on entry into desensitization (Fig. 4d, f). For AMPARs co-expressed with γ7, the equilibrium response significantly decreased from 5.2 ± 1.3% to 0.92 ± 0.23% (n = 7) (Fig. 4e), which was accompanied by an increased contribution of the fast component of desensitization from 92.9% (WT) to 97.4% (3D) (n = 7) (Fig. 4f). For AMPARs co-expressed with GSG1L, the contribution of the fast component increased from 67.0% (WT) to 91.0% (3D), leading to an overall acceleration of current decay (Fig. 4f and Supplementary Table 1). More profound desensitization was also manifested by a decrease in the steady-state response from 2.1 ± 0.3% to 0.75 ± 0.24% (n = 11) (Fig. 4e). The 3D mutant also attenuated the modest slowing of deactivation by the claudin proteins studied (Supplementary Table 1). In contrast, mutation of the KGK site did not impact the modulatory capacity of CKAMP44, as it continued to induce profound desensitization of 3D receptors, further speeding entry kinetics (n = 6) (Fig. 4d, f).

Together, these data demonstrate that the evolutionarily-conserved KGK motif is an allosteric site that is the common target of claudin-related proteins (γ2, γ7, GSG1L), but not CKAMP44. The ability of CKAMP44 to regulate AMPAR gating independent of the KGK motif further reveals that this auxiliary protein must operate through a distinct structural pathway.

GSG1L slows recovery from desensitization through a separate allosteric site

TARP γ2 has been shown to speed up recovery from AMPAR desensitization (depending on the composition of the pore-forming subunits)^15,38,42,51. In contrast, GSG1L profoundly slows the recovery process^6,7,15; as yet, the precise structural basis for this distinction remains unknown. Given that mutation of the KGK site affected both γ2- and GSG1L-mediated slowing of entry into desensitization, we investigated whether this site was also involved in recovery from desensitization (Fig. 5 and Supplementary Table 2). CKAMP44 was also of interest as it has been shown to induce a profound slowing of recovery⁴¹, but whether it does so by the same structural mechanism as GSG1L is unclear.

**Fig. 5: GSG1L slows recovery from desensitization through a unique regulatory site.**

In agreement with the literature, TARP γ2 modestly sped up GluA2 AMPAR recovery from desensitization (τ_recovery) from 22.3 ± 2.0 ms (n = 10) to 16.7 ± 0.9 ms (n = 6), while GSG1L and CKAMP44 had the opposite effect (Fig. 5a, b). The recovery time course with CKAMP44 could also be well fit by a mono-exponential function (138.4 ± 19.4 ms, n = 8), whereas GSG1L co-expression resulted in a bi-exponential recovery, with fast and slow time constants of 37.5 ± 5.4 ms (16.0%) and 322.5 ± 34.0 ms (84.0%), respectively (n = 9) (Fig. 5b and Supplementary Table 2).

When the KGK site was mutated, γ2 modulation of recovery from desensitization was lost (n = 17) (Fig. 5c, d and Supplementary Table 2), suggesting that interactions with the KGK motif are critical for γ2 to regulate both the rates into and out of AMPAR desensitization. Interestingly, mutation of the KGK motif did not disrupt the ability of GSG1L to slow recovery. The recovery time course with GSG1L was bi-exponential, resulting in time constants of 50.5 ± 12.0 ms (15.9%) and 300.2 ± 23.7 ms (84.1%) (n = 7) (Fig. 5c, d). CKAMP44 was similarly unaffected, slowing recovery of GluA2 3D AMPARs by 7-fold to 148.9 ± 23.0 ms (n = 5) (Fig. 5c, d). These observations reveal that γ2 accelerates exit from desensitization through the KGK site, whereas GSG1L and CKAMP44 slow recovery through a different structural pathway.

To evaluate other possible regulatory sites, we focused on the base of the AMPAR LBD as it may be accessible to the ExL1 of GSG1L^15,32. Based on this, we identified residue Glu658 (E658) which is also evolutionarily-conserved (Fig. 4a) and present across GluA1-4 subunits. This residue was screened by mutagenesis, amongst others, due to its charged nature and location on the lower D2 lobe of the LBD in proximity to the KGK motif (Fig. 4b). Strikingly, we observed that mutation of E658 to Lys (i.e., E658K) eliminated the ability of GSG1L to slow recovery, resulting in a single recovery time constant of 48.0 ± 4.6 ms (n = 8) (Fig. 5e, f and Supplementary Table 2). Importantly, GSG1L continued to slow desensitization and deactivation kinetics of E658K AMPARs, similar to WT receptors, indicating that the mutation did not disrupt GSG1L assembly (Supplementary Fig. 3 and Supplementary Table 1). TARP γ2 continued to accelerate recovery from desensitization of E658K receptors from 47.0 ± 3.3 ms (n = 6) to 32.0 ± 2.6 ms (n = 7) (Fig. 5e, f), demonstrating that the E658 residue fulfills a role that is distinct from the KGK site. CKAMP44 regulation was also unaffected by the E658K mutation (180.1 ± 18.0 ms, n = 8) (Fig. 5e, f), which signifies that GSG1L and CKAMP44 must slow recovery from desensitization through separate allosteric sites. Together, these findings uncover that GSG1L, unlike TARPs, regulates the onset and recovery from AMPAR desensitization through two distinct allosteric sites on the AMPAR LBD. Furthermore, the regulatory effect of the E658 residue is specific to slow recovery/exit from the desensitized state imparted by GSG1L.

TARPs also modify GluA1/A2 AMPAR heteromer gating via the KGK motif

Since the majority of native AMPARs are heteromers composed of GluA1/A2 subunits⁵², we next used our regulatory sites found in GluA2 homomers (Figs. 4 and 5) to better understand how TARPs (Fig. 6) and GSG1L (Fig. 7) modulate GluA2(R)-containing heteromeric gating. Heteromerization was confirmed in each recording by testing for the loss of cytoplasmic polyamine block (Supplementary Fig. 4 and Methods), as we have performed previously²⁴. To date, most studies have focused on TARPs where TARP γ2 (or γ8) is tethered to either GluA1 or GluA2, or both, revealing that the degree of TARP modulation of AMPARs is greater with 4 TARP subunits than with 2 subunits^24,26,53,54. Given this, we tethered TARP γ2 to both GluA1 and GluA2(R) (see Methods) to maximize the TARP phenotype, which allowed us to interrogate the individual contributions of GluA1 and GluA2 to channel gating by way of sequentially mutating the KGK site in each subunit (Fig. 6a, cartoon and Supplementary Table 3).

**Fig. 6: GluA2 dominates gating in heteromeric AMPAR-TARP complexes, which is mediated through the KGK site.**

**Fig. 7: Heteromeric AMPAR-GSG1L complexes exhibit slow recovery from desensitization that is mediated through the EK site.**

Fully-TARPed heteromers (A1/γ2 + A2/γ2) exhibited a pronounced equilibrium response of 24.0 ± 2.1% and a fast time constant of desensitization of 7.7 ± 0.7 ms (n = 6), which were in good agreement with the responses evoked by A1/A2 heteromers freely expressed with γ2 (n = 7) (Fig. 6b, c and Supplementary Table 3). These observations reaffirm previous work^21,24,54 that tethering AMPAR subunits to TARPs does not affect the normal functional characteristics of the receptor. Surprisingly, mutation of the KGK motif on GluA1 had no effect on TARP modulation of either the steady-state response or desensitization kinetics (n = 10) (Fig. 6b, c and Supplementary Table 3). Given that the 3D mutation in GluA1 homomers disrupted TARP modulation (Supplementary Fig. 5 and Supplementary Table 3), this finding may be explained by a functional dominance of the GluA2 subunit. In keeping with this, subsequent mutation of the KGK motif on GluA2 elicited a significant decrease in the steady-state response by nearly 2-fold to 13.5 ± 1.1% and a concomitant speeding up of desensitization kinetics to 4.9 ± 0.3 ms (n = 9) (Fig. 6b, c). Interestingly, mutation of KGK on both GluA1 and GluA2 led to a further reduction in the steady-state response to 5.6 ± 0.6% (n = 8), revealing a synergy that occurred only when all pore-forming subunits were mutated (Fig. 6b). Desensitization kinetics, however, did not speed up beyond mutating GluA2, suggesting that the functional interaction at GluA2 is sufficient to mediate TARP effects on gating kinetics (Fig. 6c and Supplementary Table 3).

Taken together, two important conclusions emerge from these data. Firstly, our observations underscore a dominant role for the GluA2 subunit in TARP modulation of AMPAR heteromer gating, in line with recent structural and functional studies^24,26,31. As alluded to above, our results cannot be explained by an insensitivity of the GluA1 subunit to TARP modulation via the KGK site, since the 3D mutation in GluA1 homomers abolishes γ2 modulation (Supplementary Fig. 5). Accordingly, our data suggest a functional asymmetry in the architecture of the AMPAR heteromer that determines how TARPs affect the GluA1 and GluA2 subunits during channel gating. Secondly, our data reveal that all four KGK sites need to be mutated to achieve the greatest loss of TARP modulation, demonstrating coordination between subunits within the AMPAR-TARP complex during activation.

Slow recovery from desensitization is coordinated by GSG1L via GluA2 in AMPAR heteromers

We next examined the structural basis underlying GSG1L-mediated slow recovery from desensitization of GluA1/A2 heteromers (Fig. 7a₁-d₁ and Supplementary Table 4). To do so, we employed a similar mutagenesis strategy as described in Fig. 6, this time with the E-to-K mutation (EK; residue 651 in GluA1, 658 in GluA2) (ref. Figs. 4 and 5). Important to note, GSG1L was freely expressed with GluA1 and GluA2 subunits in these experiments, since tethering GSG1L disrupted its canonical modulatory properties (Supplementary Figs. 6–8 and Supplementary Tables 3-5). Thus, the responsiveness of receptors to the partial agonist kainate (KA) was also assessed to confirm GSG1L incorporation into AMPAR complexes, as AMPAR heteromers alone are relatively insensitive to KA (Fig. 7a₂-d₂ and Supplementary Table 5).

As with GluA2 homomers, co-assembly of GSG1L with GluA1/A2 heteromers resulted in a slow, bi-exponential recovery time course with fast and slow time constants of 23.7 ± 3.8 ms (20.1%) and 547.6 ± 62.7 ms (79.9%), respectively (n = 6) (Fig. 7a₁). When EK was mutated on GluA1 only, GSG1L-bound heteromers still exhibited a slow recovery time course but it was best fit by a mono-exponential function, with a τ_recovery of 338.3 ± 36.7 ms (n = 7) (Fig. 7b₁). Although this recovery did not exactly match the bi-exponential time course of WT heteromers co-expressed with GSG1L (blue fit line), it was still relatively slow compared to WT and EK mutant heteromers alone (Supplementary Table 4). This distinction may hint at some functional contribution by GluA1 masked by the dominance of GluA2. Indeed, subsequent mutation of the EK site on GluA2 lead to a greater attenuation of GSG1L’s effect on the recovery time course, with fast and slow time constants of 80.6 ± 10.7 ms (54.7%) and 446.0 ± 20.5 ms (45.3%), respectively (n = 7) (Fig. 7c₁). Of note, the contribution of the slow component to the recovery time course was much smaller relative to WT complexes (Supplementary Table 4). Notably, mutation of EK on both GluA1 and GluA2 subunits nearly eliminated slow recovery mediated by GSG1L, resulting in a single τ_recovery of 129.8 ± 8.0 ms (n = 9) (Fig. 7d₁). Altogether, these data demonstrate a dominant role for GluA2 in dictating the time course of recovery from desensitization of AMPAR-GSG1L heteromers.

Discussion

The present study provides a comprehensive characterization of the AMPAR-GSG1L complex, advancing our understanding of GSG1L in several important ways. First, we demonstrate that GSG1L expression in the rodent brain is region-specific and developmentally regulated, with striking patterns observed in the cerebellum, caudate putamen, and cerebral cortex. Second, we also shed light on the native GSG1L interactome and find that GSG1L binds to all GluA subunits, with an average stoichiometry of two GSG1L proteins per GluA tetramer. Within native AMPAR complexes, GSG1L can either constitute the inner core alone or together with TARP/CNIH subunits. Finally, we provide structural insights into how GSG1L and TARP γ2 differentially modify AMPAR desensitization in homomeric and heteromeric assemblies. We show that slow recovery by GSG1L is not mediated through the KGK site, but instead regulated by a separate, evolutionarily-conserved allosteric site unique to GSG1L. We reveal that coordination between pore-forming subunits and asymmetry underlie AMPAR-claudin gating, with a privileged role for GluA2 in dictating both TARP and GSG1L modulation. As summarized in Fig. 8 and discussed below, together, these features allow AMPAR-GSG1L signaling complexes to fulfill specialized roles at select glutamatergic synapses.

**Fig. 8: The hallmarks of GSG1L-containing AMPAR complexes.**

The first theme that emerges from our work is the unique expression pattern and interactome of native AMPAR-GSG1L complexes. The precise role of GSG1L in the brain represents a relatively new area of study that demands a spatiotemporal mapping of its expression and better understanding of its assembly. For example, initial studies focused on the hippocampus^18,20 where GSG1L promoter activity was recently reported to be low early in development and localized to dentate granule cells (GCs) and CA3 later in development¹⁹ (Supplementary Fig. 1), consistent with in situ hybridization data³⁷. Moreover, evidence from the anterior thalamus (AT) suggests that some neurons establish synapses specific for the AMPAR-GSG1L complex¹⁹. In keeping with this emerging view, our lacZ reporter staining for GSG1L promoter activity shows a high degree of specificity, i.e., expression is not ubiquitous (Fig. 1), and our proteomic data uncover that GSG1L assembles into AMPAR complexes with distinct composition (Fig. 2). Although the signals of the lacZ reporter in the AT are consistent with the functional expression of GSG1L¹⁹ and many signals presented in the current work agree with the Allen Brain Atlas of the mouse brain³⁷, caution is still needed because GSG1L promoter activity may not be faithfully reproduced as a result of insertion of the lacZ reporter in the genome. We suggest that endogenous GSG1L expression is most likely found in regions where our lacZ results and the Allen Brain Atlas agree. The excess of GSG1L in affinity purifications in relation to other auxiliary subunits (Fig. 2b) is suggestive for the surprising, preferred occurrence of GSG1L in TARP/CNIH-free AMPARs. As a note of caution, these complexes might be overrepresented in our experiments, as the binding of GSG1L antibodies could partially destabilize the interaction of other auxiliaries. The apparent solitary role of GSG1L positions it as a unique regulator of AMPAR-mediated neurotransmission. Indeed, functional segregation of AMPAR-auxiliary complexes has been demonstrated at synapses of AD/AV neurons in the AT. TARP γ2 and GSG1L are both expressed in these neurons but exhibit input-specific regulation and likely do not co-assemble¹⁹.

Our study reports that although GSG1L is incorporated into only a subset (about 5%) of all AMPARs in the mature rat brain (Fig. 2c), its expression globally increases throughout development (Fig. 1). One caveat of our proteomic work is that it does not fully capture the dynamics of the GSG1L interactome during development; as such, proteins that assemble at early timepoints or in a developmentally-dependent, cell-type-specific manner may have been overlooked. Nonetheless, in the cortex and caudate putamen (CP), GSG1L exhibits a clear, late developmental onset between P21 and P60, whereas it is expressed earlier in migrating GCs in the cerebellum (Fig. 1). Of note, the CP receives excitatory inputs from the cortex suggesting that GSG1L may be an important constituent of a neuronal circuit widely thought to be involved in motivated behavior⁵⁵. Why GSG1L expression is activated at later developmental stages in these regions is unclear, though CKAMP44, which similarly slows recovery from desensitization, is reportedly expressed throughout the cortex early in development⁴¹. In both the cortex and cerebellum, GSG1L expression is restricted, namely to layer 2/3 and GCs, respectively. Curiously, expression in the cerebellar GL is not uniform across lobules, with GSG1L localizing primarily to the anterior portion which has been linked to sensorimotor function⁵⁶. Although obvious motor deficits were not observed in GSG1L KO mice¹⁹, it does not necessarily mean that its loss does not impact cerebellar-mediated behaviors, such as eyeblink conditioning⁵⁷, as shown for GluA4 KO mice⁵⁸. An intriguing possibility, however, is that GSG1L may be involved in encoding synaptic information in cerebellar GCs, explaining how short-term depression of AMPAR responses can paradoxically affect neuronal gain control^59,60,61.

Interesting to note, the cerebellar GC is abundant in claudin-type AMPAR auxiliary proteins (i.e., TARPs γ2, γ7, and GSG1L) as well as CKAMP39^49,62, with GluA4 as the major AMPAR subunit^63,64. Although there is a low widespread abundance of GluA4 based on our proteomic analyses (GSG1L APs; Fig. 2b), it may, however, correspond to most AMPAR complexes expressed by cerebellar GCs. Our findings also indicate that a prominent constituent of GSG1L-containing AMPARs is the transsynaptic adhesion protein, LRRT4, perhaps implying that this protein plays a role in directing the formation and organization of GSG1L-rich synapses. Altogether, this work presents a valuable foundation for future investigation into the impact of GSG1L on AMPAR response fidelity, excitability, and circuit behavior.

Another important theme stemming from our study is that the lower D2 lobe of the AMPA receptor LBD acts as an allosteric hub for GSG1L and TARPs. Our data uncover that claudin homologs, TARPs and GSG1L, target AMPARs at two functionally distinct sites that regulate entry into and exit out of receptor desensitization. Previous studies have already shown that TARP γ2 regulates the onset of desensitization of GluA2 homomers through the KGK site²¹. The present study extends this finding to both Type I and Type II TARPs, as well as GSG1L, revealing that all these claudins slow the onset of desensitization through the same structural mechanism, albeit to varying degrees (Figs. 3 and 4). The importance of the KGK motif in claudin regulation of AMPAR gating is also extended to GluA1 homomers (Supplementary Fig. 5) and GluA1/A2 heteromers (Fig. 6).

In contrast, γ2 and GSG1L have differential effects on recovery from desensitization (Fig. 5). While γ2 modestly speeds up the recovery process via the KGK site, our data show that the profound slowing of recovery by GSG1L is mediated by a single residue that is conserved amongst all AMPAR subunits. Since TARPs and GSG1L share a common ancestry⁸, it is tempting to speculate that GSG1L’s much weaker effect on the KGK site represents a vestige of its evolutionary past rather than serving any functional purpose. Structural and functional studies have argued that the first extracellular loop of GSG1L, which is comprised of variable sized flexible loops between four β-strands, interacts with AMPARs to slow recovery¹⁵. In particular, the β1-β2 loop contains 49 amino acids that differ between GSG1L and TARPs and, although not fully resolved in cryo-EM structures, these residues are thought to directly interact with AMPARs. The length of this loop is proposed to underlie the functional interactions that distinguish GSG1L and TARPs, with our work now revealing that the different loops may contact distinct regions on the AMPAR LBD (Fig. 8). Furthermore, drastic structural differences in the symmetrical organization of desensitized-state LBD dimers between AMPAR-TARP and AMPAR-GSG1L complexes have been observed; the loss of two-fold symmetry reported for GSG1L may impact its interactions with the LBD³². Taken together, these findings reveal unexpectedly that two important CNS claudin family members, TARP γ2 and GSG1L, have evolved to regulate AMPARs through structurally-distinct and allosterically-separate mechanisms.

In addition to the interface between the extracellular loops and the LBD, functional studies have also demonstrated the importance of AMPAR-TARP interactions at the level of the transmembrane alpha helices and C-termini^65,66, TMD-LBD linkers⁶⁷, and even the NTD⁶⁸. In contrast to TARPs and GSG1L, the structural architecture of the AMPAR-CNIH complex indicates that CNIHs are embedded in the plasma membrane, such that interactions occur exclusively at the TMD⁶⁹. To date, the structure of an AMPAR associated with CKAMPs has yet to be resolved. Although CKAMP44 slows recovery from desensitization, our data demonstrate that the mechanism of allosteric modulation is different from GSG1L.

The final theme relates to the asymmetrical contribution of pore-forming subunits in AMPAR heteromers that is unveiled by claudin-related proteins. Early structural studies revealed that the placement of core subunits within the AMPAR tetramer is critical for channel activation⁷⁰. AMPARs are composed of subunit pairs termed AC (“pore-proximal”) and BD (“pore-distal”), where conformational changes in BD subunits dominate channel gating^26,70. Consequently, the assembly of pore subunits dictates the functional interactions with auxiliary proteins depending on their location within the receptor complex. For example, while TARP γ2 has been shown to engage with the KGK sites of the BD pair via the β4-TM2 loop, the longer β1-β2 loop of TARP γ8 and GSG1L has been shown to simultaneously access the LBDs of adjacent subunit pairs^15,23,71. Until recently, the occupancy of AC and BD pairs in the heteromer structure was not firmly established, but it is now thought that in GluA1/A2 heteromers, A1 and A2 sit at AC and BD positions, respectively^26,29 (Fig. 8). Based on their location, GluA1 and GluA2 exhibit differences in conformational rearrangements at the level of the LBD during open-to-desensitized transitions and contribute asymmetrically to channel gating⁷¹. This state-dependent asymmetry produces gating-state-specific interactions with flexible TARP extracellular loops (shown for TARP γ8), and presumably, GSG1L.

Using a mutational strategy combined with electrophysiology, we investigated this notion with A1/A2 heteromers bound by TARP γ2 or GSG1L auxiliary subunits (Figs. 6 and 7). Mutation of KGK in the GluA1 subunit did not produce a loss-of-function effect on TARP-dependent gating. When the KGK site was mutated in the GluA2 subunit, there was a partial attenuation of TARP modulation on desensitization kinetics and the equilibrium response, which was further pronounced when all AMPAR subunits were mutated. These findings showcase that the functional interaction of TARPs with GluA2/BD is dominant for gating, but a synergy exists between the core subunits. A similar subunit coordination was observed in the context of GSG1L and recovery from desensitization. When the E651 site was mutated in GluA1/AC, GSG1L continued to slow the recovery time course. Only when the equivalent residue was mutated in GluA2/BD was there a partial loss of GSG1L’s effect, which was further attenuated by disrupting functional interactions at all four pore-forming subunits. These findings identify GluA2 as being primed to dominate channel gating and, as such, provide a structural framework to understand how AMPAR heteromers are modulated by all claudin family members.

Methods

Animals for histology

The transgenic GSG1L knockout (KO) rat was among the mutants generated by transposon-based mutagenesis by Kent Hamra at UT Southwestern³⁵, and was purchased and retained in-house. GSG1L KO rats were maintained by crossing GSG1L heterozygous animals in-house. All animal procedures were approved by the Vanderbilt University Animal Care and Use Committee and were in agreement with the NIH and Vanderbilt University guidelines for the care and use of laboratory animals.

Genotyping by PCR of transgenic rats for histology

Genotyping was done as previously described¹⁹. PCR amplification was conducted with the following primers: GSG1L left: acgttgtagtgaccccaagc; GSG1L right: tgcacgcatactacaatga; SFB2: tcatcaaggaaaccctggac.

Staining of whole brains

P14, P24, P42, and P120 GSG1L KO rat brains and corresponding wildtype (WT) rat brains (n = 2) were anesthetized with sodium pentobarbital (Nembutal). The animals were then perfused with normal Rat Ringer solution for 2 min for complete perfusion, followed by 4% paraformaldehyde in 0.1 M phosphate buffer for 4 min. The brains were then obtained and permeabilized with 0.01% sodium deoxycholate and 0.02% Triton X-100 in PBS buffer for 2 h. Following permeabilization, the tissue samples were incubated with X-gal staining solution containing 5 mM K₃[Fe(CN)₆]_, 5 mM K₄[Fe(CN)₆], 2 mM MgCl₂, 0.02% Triton X-100, and 0.1% X-gal in PBS buffer at 37 °C for 8 h in the dark. Stained brains were imaged (MULTIZOOM AZ100M, Nikon) the following day (1 day post-staining).

Staining of fixed brain sections

P14, P21, P60, P180, and P240 GSG1L KO rat brains (n = 2) were anesthetized with sodium pentobarbital (Nembutal). Each animal was then perfused with normal Rat Ringer solution for 2 min for complete perfusion, followed by 4% paraformaldehyde in 0.1 M phosphate buffer for 4 min. Obtained brains were further fixed for 15 min. Fixed brains were subsequently sectioned to generate 300 μm-thick coronal or sagittal sections using a vibratome (Leica VT 1200). The sections were then permeabilized with 0.01% sodium deoxycholate and 0.02% Triton X-100 in PBS buffer for 2 h. Following permeabilization, 300 μm-thick sections were incubated with X-gal staining solution containing 5 mM K₃[Fe(CN)₆]_, 5 mM K₄[Fe(CN)₆], 2 mM MgCl₂, 0.02% Triton X-100, and 0.1% X-gal in PBS buffer at 37 °C for 8 h in the dark. Stained brains were imaged (MULTIZOOM AZ100M, Nikon) the following day (1 day post-staining) and then left for 1 week at 4 °C to enhance the lacZ staining due to residual X-gal in the tissue. It should be noted that efforts to generate reliable anti-GSG1L antibodies for immunohistochemistry (IHC) have so far failed.

Affinity purifications (APs)

Animal procedures were performed in accordance with national and institutional guidelines in accordance with the German law for the welfare of animals and were approved by local authorities (Regierungspräsidium Freiburg X14/14H). Brains from three adult (P59) GSG1L WT and KO rats were dissected and shock-frozen in liquid nitrogen. The following procedures were performed essentially as described in⁶. The brains were cut in pieces and homogenized in 30 ml ice-cold H-buffer (10 mM Tris/HCl pH 7.5, 300 mM sucrose, 1.5 mM MgCl₂, 1 mM EGTA, 1 mM iodoacetamide and protease inhibitors) with Dounce homogenizer. The nuclear fraction was removed by centrifugation (4 min at 1000xg) and the respective supernatants subjected to ultracentrifugation (20 min at 200,000xg). The pellets were homogenized in 30 ml Lysis buffer (5 mM Tris/HCl pH 7.4 supplemented with protease inhibitors) and incubated on ice for 30 min. Additional ultracentrifugation (20 min at 200,000xg) separates soluble proteins from the membrane/membrane-attached proteins. The latter was resuspended in 20 mM Tris/HCl pH 7.4 and loaded on top of a two-layer density gradient (0.5 M /1.3 M sucrose in 10 mM Tris/HCl pH 7.4). After ultracentrifugation (45 min at 30,000 rpm, Sorvall Surespin 630), the membrane protein concentrations were determined by Bradford assay and adjusted to 10 mg/ml.

For each affinity purification (Fig. 2a, b and Supplementary Fig. 2), 1.5 mg of membrane proteins from WT or GSG1L KO rats were solubilized with 1.5 ml CL-47 (Logopharm) supplemented with protease inhibitors (Aprotinin, Leupeptin, Pepstatin A, PMSF). Homogenates were cleared by ultracentrifugation (10 min at 125,000xg) and solubilisates were incubated with 15 μg antibodies pre-coupled to protein A Dynabeads. The following affinity purified GSG1L antibodies were used for the interactome analysis: Ab#1, polyclonal, raised in rabbit against C-term of rat GSG1L (aa304-322), Ab#2, polyclonal, raised in rabbit against rat GSG1L (aa257-278), Ab#3, polyclonal, raised in rabbit against rat GSG1L (aa287-308). Antibodies were incubated for 2 h with solubilisates and subsequently briefly washed two times with 0.5 ml CL-47 buffer. Proteins were eluted with 10 μl of Lämmli buffer w/o DTT. Eluted proteins were shortly separated on SDS-PAGE and silver stained.

For two-step APs (Fig. 2c), 0.5 mg of WT membranes were solubilized in 0.5 ml CL-91⁶. Solubilisates were first incubated for 2 h with a mixture of GSG1L antibodies (20 μg Ab#2, 25 μg Ab#3 and 5 μg Ab#4 (Proteintech, #17328-1-AP)) coupled to protein A Dynabeads to completely affinity isolate GSG1L. After this incubation period, the solubilisate was transferred to a mixture of coupled AMPAR antibodies (25 μg anti-GluA1, #AB1504 Millipore; 20 μg anti-GluA2, #75-002 NeuroMab; 5 μg anti-GluA2/3, #07-598 Millipore; 5 μg anti-GluA3 #182203 Synaptic Systems; 10 μg anti-GluA4 #AB1508 Millipore) for 2 h of incubation. In both cases, after incubation, the beads with antibodies were separated, briefly washed, and proteins eluted as described above. At each step of this experiment, 10 μl aliquots were taken and analysed by SDS-PAGE and Western blotting to control for the complete affinity isolation of AMPARs.

Mass spectrometry (MS)

The eluted proteins of each affinity purification were separated on SDS-PAGE and proteins subsequently visualized by silver staining. The gel lanes were cut and split in two sections to reduce complexity. Proteins were in-gel digested using sequencing-grade modified trypsin (Promega GmbH, Walldorf, Germany) following the procedure described in⁷². Vacuum-dried peptides were dissolved in 20 μl of 0.5% (v/v) trifluoroacetic acid, loaded onto a trap column (C18 PepMap100, 5 μm particles, Thermo Fisher Scientific GmbH, Dreieich, Germany), separated by reversed-phase chromatography via a 10 cm C18 column (PicoTip™ Emitter, 75 μm, tip: 8 μm, New Objective, self-packed with ReproSil-Pur 120 ODS-3, 3 μm, Dr. Maisch HPLC; flow rate 300 nl/min) using an UltiMate 3000 RSLCnano HPLC system (Thermo Scientific), and eluted by an aqueous organic gradient (eluent “A”: 0.5% acetic acid; eluent “B”: 0.5% acetic acid in 80% acetonitrile). MS-analyses were executed on an Orbitrap Elite mass spectrometer with a Nanospray Flex Ion Source (both Thermo Scientific). Precursor signals (LC-MS) were acquired with a target value of 1,000,000 and a nominal resolution of 240,000 (FWHM) at m/z 400; scan range 370–1700 m/z. LC-MS/MS data were extracted using “msconvert.exe” (part of ProteoWizard; http://proteowizard.sourceforge.io/). Peak lists were searched against a UniProtKB/Swiss-Prot database (containing all rat, mouse, and human entries) with peptide mass tolerance ± 5 ppm; fragment mass tolerance ± 0.8 Da using Mascot 2.6.2 (Matrix Science, UK). One missed trypsin cleavage and common variable modifications including S/T/Y phosphorylation were accepted for peptide identification. Significance threshold was set to p < 0.05.

Quantification of proteins

Proteins were quantitatively evaluated according to a procedure described in ref. ⁷³. Briefly, peptide signal intensities (peak volumes, PVs) were extracted from FT full scans and mass calibrated using MaxQuant v1.6.3 (http://www.maxquant.org). Peptide PV elution times were then aligned and assigned to peptides based on matching m/z and elution times (tolerances 2-3 ppm / ± 1 min) as described previously⁷⁴. All protein-specific peptide signal intensities in all runs were further processed to eliminate the influence of PV outliers, false assignments, and gaps by exploring the consistency of PV relations within proteins (i.e., protein-specific PV ratios between and within runs). Abundance_norm spec values (as a measure of molecular abundance) were calculated as described in⁷⁴.

Specificity of co-purifications (Fig. 2a and Supplementary Fig. 2) was determined according to target-normalized abundance ratios (tnRs) of proteins affinity purified from WT versus control (GSG1L KO) (calculated as described in ref. ⁷³). Six distinct anti-GSG1L APs (3 from WT and 3 from GSG1L KO) provided 3 ratios. This information was inspected using the BELKI software suite (https://github.com/phys2/belki). tnR values were visualized by t-distributed stochastic neighbor embedding (t-SNE; Fig. 2a). The high consistency of all datasets was also shown as 2D plots of tnR values (Supplementary Fig. 2). A minimum tnR of 0.25 was used as the indicator for specific interaction. Abundances of all proteins specifically and consistently affinity purified in anti-GSG1L APs were plotted as relative values to the abundance of GSG1L, which was set to 1 (Fig. 2b); data are mean of three experiments. Two-step APs (Fig. 2c) were used to calculate the proportion of GSG1L-containing AMPARs relative to the entirety of AMPARs in the rat brain. First, the protein abundance values in target-depleting GSG1L APs and the subsequent target-depleting GluA1-4 APs were determined for all GSG1L interactors. Relative amounts of proteins in GSG1L and GluA1-4 APs (protein abundance in AP/sum of protein abundances in (GSG1L + GluA1-4 APs)) were calculated.

Plasmids for recombinant electrophysiology

All AMPAR pore-forming subunits correspond to the rat sequence in the pRK5 vector. For GluA2 homomers, the Q/R unedited flip (GluA2Q_flip) isoform was used as indicated. For GluA1 homomers, the flip isoform was also used (GluA1_flip). For the study of heteromers, the Q/R edited flip (i.e., GluA2R_flip) isoform of GluA2 was used. Residue numbering includes the signal peptide. Mutant receptors were generated using site-directed mutagenesis and all new constructs were screened by restriction digest prior to confirmation by sequencing. Auxiliary subunits and AMPARs were co-expressed at a 2:1 or 2:1:1 cDNA ratio for homomers and heteromers, respectively, except for constructs in which mouse TARP γ2 or mouse GSG1L were tethered to the AMPAR subunit in the pRK5 vector (for γ2, refs. ^21,24); fusion notation /γ2 or /GSG1L vs. co-expression notation +γ2 or +GSG1L. For GSG1L fusion constructs, the full-length coding sequence of GSG1L and a short 7 aa linker sequence (same as the γ2 fusion construct, i.e. ELGTRGS⁷⁵) were synthesized by Bio Basic Inc. (Ontario, CA) and sub-cloned into the GluA1 and GluA2 vectors. Species and vectors for co-expressed auxiliary subunits are as follows: human TARP γ7 in pCMV6-XL4 (OriGene), mouse GSG1L in pReceiver-M02 (GeneCopoeia), and mouse CKAMP44 in pRK5 (from J. von Engelhardt). cDNA was co-transfected with a plasmid encoding enhanced green fluorescent protein (eGFP) to identify transfected cells.

Cell culture and transfection

HEK293T/17 cells (ATCC, CRL-11268) were maintained at 37 °C under 5% CO₂ in Minimum Essential Medium with GlutaMAX (i.e., MEM GlutaMAX) supplemented with 10% fetal bovine serum (FBS). Cells were plated at low density (1.6 × 10⁴ cells/ml) on poly-D-lysine-coated 35-mm dishes, and transiently transfected 24-48 h post-plating using the calcium phosphate precipitation method. After 6-12 h, cells were washed twice with divalent PBS and maintained in fresh medium containing 30 μM DNQX or NBQX to minimize auxiliary protein-induced cytotoxicity.

Patch-clamp electrophysiology recordings

All recordings were performed 24-48 h post-transfection on excised outside-out patches. External solution contained (in mM): 150 NaCl, 5 HEPES, 0.1 CaCl₂ and 0.1 MgCl₂, and 2% phenol red at pH 7.3–7.4. Internal solution contained (in mM): 115 NaCl, 10 NaF, 5 HEPES, 5 Na₄BAPTA, 0.5 CaCl₂, 1 MgCl₂, and 10 Na₂ATP at pH 7.3-7.4. For GluA1/A2(R) heteromer recordings, 30 μM spermine was included in the internal solution in place of Na₂ATP; data were only included when the I-V plot was linear, typical of GluA2(R)-containing AMPARs^24,76 (see Supplementary Fig. 4). The osmotic pressure of all solutions was adjusted to 295-300 mOsm with sucrose.

Recording pipettes were composed of borosilicate glass (3-6 MΩ, King Precision Glass, Inc.) coated with dental wax. Agonist solution (10 mM L-Glu or 1 mM KA) was rapidly applied using a piezo-stack driven perfusion system (Physik Instrumente) and solution exchange (<400 μs) was determined by measuring the liquid junction current at the end of each experiment. All recordings were performed using an Axopatch 200B amplifier (Molecular Devices, LLC). The holding potential during recordings was −60 mV or −100 mV. Current records were filtered at 5 kHz and sampled at 25 kHz. Series resistance (3-12 MΩ) was compensated for by 95%. All experiments were performed at room temperature. Data were acquired using pClamp9 software (Molecular Devices, LLC).

Fitting analysis of electrophysiological data

Electrophysiological recordings were analyzed using Clampfit 10.5 (Molecular Devices, LLC). Current decay rates were fit using 1st or 2nd order exponential functions of the form y = A_i*exp(-x/τ_i). For decay rates requiring 2nd order exponential fits, time constants are presented using both the individual components and as weighted means. To measure recovery from desensitization, a two-pulse protocol was used in which agonist was applied at variable interpulse intervals, and the peak amplitude of the second (test) pulse was expressed as a fraction of the peak amplitude of the first (initial) pulse. The recovery time course was fit with mono- or bi-exponential functions, as indicated, to measure τ_recovery (described in⁴⁵; see also Results and Supplementary Tables 2 and 4). In all figures, current traces are normalized unless otherwise indicated. Data were illustrated using Origin 2020 (OriginLab) and Adobe Illustrator.

Statistical analysis of electrophysiological data

Statistical details can be found in the figure legends and Source Data file. Data are presented as mean ± SEM, where n values refer to the number of individual patches. Data points correspond to individual patches for each transfection condition. Transfections were performed at least two independent times. Data were assessed for normality using a Shapiro-Wilks test and appropriate parametric or nonparametric tests were conducted accordingly. For simple pairwise comparisons, a two-tailed unpaired Student’s t-test or Mann-Whitney U test was performed as indicated. For comparisons between multiple groups (parametric), a one-way between-subject ANOVA was performed. Subsequent pairwise comparisons were made using Tukey’s Honestly Significant Difference (HSD) test. For comparisons between multiple groups (nonparametric), a Kruskal-Wallis ANOVA was performed, followed by pairwise Mann–Whitney U tests with application of a Bonferroni-Holm correction. Significance level was set at 0.05 and is denoted as *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001. Exact p-values are provided in the Source Data file for both significant and nonsignificant results, except when p < 0.001. Statistical analysis was conducted using SPSS Statistics (IBM) and custom statistical software kindly provided by Joe Rochford (McGill University).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

Data supporting the findings of this study are available upon request. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository⁷⁷ with the dataset identifier PXD044621 [10.6019/PXD044621]. Referred protein structures have the following PDB accession codes: 5WEO, 5VHY, 1FTJ, 5KBU, 7RYZ. Source data for Figs. 2, 3, 4, 5, 6, and 7 and Supplementary Figs. 2, 3, 5, 6, 7, and 8 are provided with the paper as indicated. Source data are provided with this paper.

References

Furuse, M., Fujita, K., Hiiragi, T., Fujimoto, K. & Tsukita, S. Claudin-1 and −2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin. J. Cell Biol. 141, 1539–1550 (1998).
Article CAS PubMed PubMed Central Google Scholar
Tsukita, S. & Furuse, M. The structure and function of claudins, cell adhesion molecules at tight junctions. Ann. N. Y. Acad. Sci. 915, 129–135 (2000).
Article ADS CAS PubMed Google Scholar
Jacobi, E. & von Engelhardt, J. Modulation of information processing by AMPA receptor auxiliary subunits. J. Physiol. 599, 471–483 (2021).
Article CAS PubMed Google Scholar
Tikiyani, V. & Babu, K. Claudins in the brain: unconventional functions in neurons. Traffic 20, 807–814 (2019).
Article CAS PubMed Google Scholar
Tomita, S. et al. Functional studies and distribution define a family of transmembrane AMPA receptor regulatory proteins. J. Cell Biol. 161, 805–816 (2003).
Article CAS PubMed PubMed Central Google Scholar
Schwenk, J. et al. High-resolution proteomics unravel architecture and molecular diversity of native AMPA receptor complexes. Neuron 74, 621–633 (2012).
Article CAS PubMed Google Scholar
Shanks, N. F. et al. Differences in AMPA and kainate receptor interactomes facilitate identification of AMPA receptor auxiliary subunit GSG1L. Cell Rep. 1, 590–598 (2012).
Article CAS PubMed PubMed Central Google Scholar
Ramos-Vicente, D. & Bayes, A. AMPA receptor auxiliary subunits emerged during early vertebrate evolution by neo/subfunctionalization of unrelated proteins. Open Biol. 10, 200234 (2020).
Article CAS PubMed PubMed Central Google Scholar
Greger, I. H., Watson, J. F. & Cull-Candy, S. G. Structural and functional architecture of AMPA-type glutamate receptors and their auxiliary proteins. Neuron 94, 713–730 (2017).
Article CAS PubMed Google Scholar
Jackson, A. C. & Nicoll, R. A. The expanding social network of ionotropic glutamate receptors: TARPs and other transmembrane auxiliary subunits. Neuron 70, 178–199 (2011).
Article CAS PubMed PubMed Central Google Scholar
Coombs, I. D. et al. Cornichons modify channel properties of recombinant and glial AMPA receptors. J. Neurosci. 32, 9796–9804 (2012).
Article CAS PubMed PubMed Central Google Scholar
Brown, P., McGuire, H. & Bowie, D. Stargazin and cornichon-3 relieve polyamine block of AMPA receptors by enhancing blocker permeation. J. Gen. Physiol. 150, 67–82 (2018).
Article CAS PubMed PubMed Central Google Scholar
Soto, D., Coombs, I. D., Kelly, L., Farrant, M. & Cull-Candy, S. G. Stargazin attenuates intracellular polyamine block of calcium-permeable AMPA receptors. Nat. Neurosci. 10, 1260–1267 (2007).
Article CAS PubMed PubMed Central Google Scholar
Bowie, D. Polyamine-mediated channel block of ionotropic glutamate receptors and its regulation by auxiliary proteins. J. Biol. Chem. 293, 18789–18802 (2018).
Article CAS PubMed PubMed Central Google Scholar
Twomey, E. C., Yelshanskaya, M. V., Grassucci, R. A., Frank, J. & Sobolevsky, A. I. Structural bases of desensitization in AMPA receptor-auxiliary subunit complexes. Neuron 94, 569–580 e565 (2017).
Article CAS PubMed PubMed Central Google Scholar
Twomey, E. C., Yelshanskaya, M. V. & Sobolevsky, A. I. Structural and functional insights into transmembrane AMPA receptor regulatory protein complexes. J. Gen. Physiol. 151, 1347–1356 (2019).
Article CAS PubMed PubMed Central Google Scholar
McGee, T. P., Bats, C., Farrant, M. & Cull-Candy, S. G. Auxiliary subunit GSG1L acts to suppress calcium-permeable AMPA receptor function. J. Neurosci. 35, 16171–16179 (2015).
Article CAS PubMed PubMed Central Google Scholar
Gu, X. et al. GSG1L suppresses AMPA receptor-mediated synaptic transmission and uniquely modulates AMPA receptor kinetics in hippocampal neurons. Nat. Commun. 7, 10873 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Kamalova, A., Futai, K., Delpire, E. & Nakagawa, T. AMPA receptor auxiliary subunit GSG1L suppresses short-term facilitation in corticothalamic synapses and determines seizure susceptibility. Cell Rep. 32, 107921 (2020).
Article CAS PubMed PubMed Central Google Scholar
Mao, X., Gu, X. & Lu, W. GSG1L regulates the strength of AMPA receptor-mediated synaptic transmission but not AMPA receptor kinetics in hippocampal dentate granule neurons. J. Neurophysiol. 117, 28–35 (2017).
Article CAS PubMed Google Scholar
Dawe, G. B. et al. Distinct structural pathways coordinate the activation of AMPA receptor-auxiliary subunit complexes. Neuron 89, 1264–1276 (2016).
Article CAS PubMed PubMed Central Google Scholar
Twomey, E. C., Yelshanskaya, M. V., Grassucci, R. A., Frank, J. & Sobolevsky, A. I. Elucidation of AMPA receptor-stargazin complexes by cryo-electron microscopy. Science 353, 83–86 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Chen, S. et al. Activation and desensitization mechanism of AMPA receptor-TARP complex by cryo-EM. Cell 170, 1234–1246.e1214 (2017).
Article CAS PubMed PubMed Central Google Scholar
Dawe, G. B. et al. Nanoscale mobility of the apo state and TARP stoichiometry dictate the gating behavior of alternatively-spliced AMPA receptors. Neuron 102, 976–992 (2019).
Zhao, Y., Chen, S., Yoshioka, C., Baconguis, I. & Gouaux, E. Architecture of fully occupied GluA2 AMPA receptor-TARP complex elucidated by cryo-EM. Nature 536, 108–111 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Herguedas, B. et al. Architecture of the heteromeric GluA1/2 AMPA receptor in complex with the auxiliary subunit TARP gamma8. Science https://doi.org/10.1126/science.aav9011 (2019).
Twomey, E. C., Yelshanskaya, M. V., Grassucci, R. A., Frank, J. & Sobolevsky, A. I. Channel opening and gating mechanism in AMPA-subtype glutamate receptors. Nature 549, 60–65 (2017).
Article ADS CAS PubMed PubMed Central Google Scholar
Yu, J. et al. Hippocampal AMPA receptor assemblies and mechanism of allosteric inhibition. Nature 594, 448–453 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Zhao, Y., Chen, S., Swensen, A. C., Qian, W. J. & Gouaux, E. Architecture and subunit arrangement of native AMPA receptors elucidated by cryo-EM. Science 364, 355–362 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Schwenk, J. & Fakler, B. Building of AMPA-type glutamate receptors in the endoplasmic reticulum and its implication for excitatory neurotransmission. J. Physiol. https://doi.org/10.1113/JP279025 (2020).
Zhang, D., Watson, J. F., Matthews, P. M., Cais, O. & Greger, I. H. Gating and modulation of a hetero-octameric AMPA glutamate receptor. Nature 594, 454–458 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Klykov, O., Gangwar, S. P., Yelshanskaya, M. V., Yen, L. & Sobolevsky, A. I. Structure and desensitization of AMPA receptor complexes with type II TARP gamma5 and GSG1L. Mol. Cell 81, 4771–4783 e4777 (2021).
Article CAS PubMed Google Scholar
Menuz, K., Kerchner, G. A., O’Brien, J. L. & Nicoll, R. A. Critical role for TARPs in early development despite broad functional redundancy. Neuropharmacology 56, 22–29 (2009).
Article CAS PubMed Google Scholar
Fukaya, M., Yamazaki, M., Sakimura, K. & Watanabe, M. Spatial diversity in gene expression for VDCCgamma subunit family in developing and adult mouse brains. Neurosci. Res. 53, 376–383 (2005).
Article CAS PubMed Google Scholar
Izsvak, Z. et al. Generating knockout rats by transposon mutagenesis in spermatogonial stem cells. Nat. Methods 7, 443–445 (2010).
Article CAS PubMed PubMed Central Google Scholar
Silbereis, J. et al. Astroglial cells in the external granular layer are precursors of cerebellar granule neurons in neonates. Mol. Cell Neurosci. 44, 362–373 (2010).
Article CAS PubMed PubMed Central Google Scholar
Lein, E. S. et al. Genome-wide atlas of gene expression in the adult mouse brain. Nature 445, 168–176 (2007).
Article ADS CAS PubMed Google Scholar
Hansen, K. B. et al. Structure, function, and pharmacology of glutamate receptor ion channels. Pharmacol. Rev. 73, 298–487 (2021).
Article PubMed PubMed Central Google Scholar
Schwenk, J. et al. An ER assembly line of AMPA-receptors controls excitatory neurotransmission and its plasticity. Neuron 104, 680–692.e689 (2019).
Article CAS PubMed Google Scholar
Roppongi, R. T., Karimi, B. & Siddiqui, T. J. Role of LRRTMs in synapse development and plasticity. Neurosci. Res. 116, 18–28 (2017).
Article CAS PubMed Google Scholar
von Engelhardt, J. et al. CKAMP44: a brain-specific protein attenuating short-term synaptic plasticity in the dentate gyrus. Science 327, 1518–1522 (2010).
Article ADS Google Scholar
Priel, A. et al. Stargazin reduces desensitization and slows deactivation of the AMPA-type glutamate receptors. J. Neurosci. 25, 2682–2686 (2005).
Article CAS PubMed PubMed Central Google Scholar
Tomita, S. et al. Stargazin modulates AMPA receptor gating and trafficking by distinct domains. Nature 435, 1052–1058 (2005).
Article ADS CAS PubMed Google Scholar
Cho, C. H., St-Gelais, F., Zhang, W., Tomita, S. & Howe, J. R. Two families of TARP isoforms that have distinct effects on the kinetic properties of AMPA receptors and synaptic currents. Neuron 55, 890–904 (2007).
Article CAS PubMed Google Scholar
Perozzo, A. M., Brown, P. & Bowie, D. Alternative splicing of the flip/flop cassette and TARP auxiliary subunits engage in a privileged relationship that fine-tunes AMPA receptor gating. J. Neurosci. 43, 2837–2849 (2023).
Article CAS PubMed PubMed Central Google Scholar
Kato, A. S., Gill, M. B., Yu, H., Nisenbaum, E. S. & Bredt, D. S. TARPs differentially decorate AMPA receptors to specify neuropharmacology. Trends Neurosci. 33, 241–248 (2010).
Article CAS PubMed Google Scholar
Kato, A. S., Siuda, E. R., Nisenbaum, E. S. & Bredt, D. S. AMPA receptor subunit-specific regulation by a distinct family of type II TARPs. Neuron 59, 986–996 (2008).
Article CAS PubMed Google Scholar
Kato, A. S. et al. New transmembrane AMPA receptor regulatory protein isoform, gamma-7, differentially regulates AMPA receptors. J. Neurosci. 27, 4969–4977 (2007).
Article CAS PubMed PubMed Central Google Scholar
Farrow, P. et al. Auxiliary subunits of the CKAMP family differentially modulate AMPA receptor properties. Elife 4, e09693 (2015).
Article PubMed PubMed Central Google Scholar
Khodosevich, K. et al. Coexpressed auxiliary subunits exhibit distinct modulatory profiles on AMPA receptor function. Neuron 83, 601–615 (2014).
Article CAS PubMed Google Scholar
Devi, S. P. S., Cheng, Y., Tomita, S., Howe, J. R. & Zhang, W. TARPs modulate receptor-mediated paired-pulse depression and recovery from desensitization. J. Neurosci. 40, 8233–8247 (2020).
Article CAS PubMed PubMed Central Google Scholar
Henley, J. M. & Wilkinson, K. A. Synaptic AMPA receptor composition in development, plasticity and disease. Nat. Rev. Neurosci. 17, 337–350 (2016).
Article CAS PubMed Google Scholar
Miguez-Cabello, F. et al. AMPAR/TARP stoichiometry differentially modulates channel properties. Elife https://doi.org/10.7554/eLife.53946 (2020).
Shi, Y., Lu, W., Milstein, A. D. & Nicoll, R. A. The stoichiometry of AMPA receptors and TARPs varies by neuronal cell type. Neuron 62, 633–640 (2009).
Article CAS PubMed PubMed Central Google Scholar
Shepherd, G. M. Corticostriatal connectivity and its role in disease. Nat. Rev. Neurosci. 14, 278–291 (2013).
Article CAS PubMed PubMed Central Google Scholar
Stoodley, C. J. & Schmahmann, J. D. Evidence for topographic organization in the cerebellum of motor control versus cognitive and affective processing. Cortex 46, 831–844 (2010).
Article PubMed PubMed Central Google Scholar
Keifer, J., Tiwari, N. K., Buse, L. & Zheng, Z. Subunit-specific synaptic delivery of AMPA receptors by auxiliary chaperone proteins TARPgamma8 and GSG1L in classical conditioning. Neurosci. Lett. 645, 53–59 (2017).
Article CAS PubMed PubMed Central Google Scholar
Kita, K. et al. GluA4 facilitates cerebellar expansion coding and enables associative memory formation. Elife https://doi.org/10.7554/eLife.65152 (2021).
Rothman, J. S., Cathala, L., Steuber, V. & Silver, R. A. Synaptic depression enables neuronal gain control. Nature 457, 1015–1018 (2009).
Article ADS CAS PubMed PubMed Central Google Scholar
Wall, M. J. Short-term synaptic plasticity during development of rat mossy fibre to granule cell synapses. Eur. J. Neurosci. 21, 2149–2158 (2005).
Article PubMed Google Scholar
Xu-Friedman, M. A. & Regehr, W. G. Ultrastructural contributions to desensitization at cerebellar mossy fiber to granule cell synapses. J. Neurosci. 23, 2182–2192 (2003).
Article CAS PubMed PubMed Central Google Scholar
Bats, C., Farrant, M. & Cull-Candy, S. G. A role of TARPs in the expression and plasticity of calcium-permeable AMPARs: evidence from cerebellar neurons and glia. Neuropharmacology 74, 76–85 (2013).
Article CAS PubMed PubMed Central Google Scholar
Hollmann, M. & Heinemann, S. Cloned glutamate receptors. Annu. Rev. Neurosci. 17, 31–108 (1994).
Article CAS PubMed Google Scholar
Schwenk, J. et al. Regional diversity and developmental dynamics of the AMPA-receptor proteome in the mammalian brain. Neuron 84, 41–54 (2014).
Article CAS PubMed Google Scholar
Ben-Yaacov, A. et al. Molecular mechanism of AMPA receptor modulation by TARP/Stargazin. Neuron 93, 1126–1137 e1124 (2017).
Article CAS PubMed Google Scholar
Hawken, N. M., Zaika, E. I. & Nakagawa, T. Engineering defined membrane-embedded elements of AMPA receptor induces opposing gating modulation by cornichon 3 and stargazin. J. Physiol. 595, 6517–6539 (2017).
Article CAS PubMed PubMed Central Google Scholar
Riva, I., Eibl, C., Volkmer, R., Carbone, A. L. & Plested, A. J. Control of AMPA receptor activity by the extracellular loops of auxiliary proteins. Elife https://doi.org/10.7554/eLife.28680 (2017).
Cais, O. et al. Mapping the interaction sites between AMPA receptors and TARPs reveals a role for the receptor N-terminal domain in channel gating. Cell Rep. 9, 728–740 (2014).
Article CAS PubMed PubMed Central Google Scholar
Nakagawa, T. Structures of the AMPA receptor in complex with its auxiliary subunit cornichon. Science 366, 1259–1263 (2019).
Article ADS CAS PubMed Google Scholar
Sobolevsky, A. I., Rosconi, M. P. & Gouaux, E. X-ray structure, symmetry and mechanism of an AMPA-subtype glutamate receptor. Nature 462, 745–756 (2009).
Article ADS CAS PubMed PubMed Central Google Scholar
Herguedas, B. et al. Mechanisms underlying TARP modulation of the GluA1/2-gamma8 AMPA receptor. Nat. Commun. 13, 734 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Pandey, A., Andersen, J. S. & Mann, M. Use of mass spectrometry to study signaling pathways. Sci. STKE https://doi.org/10.1126/stke.2000.37.pl1 (2000).
Kocylowski, M. K. et al. A slit-diaphragm-associated protein network for dynamic control of renal filtration. Nat. Commun. 13, 6446 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Bildl, W. et al. Extending the dynamic range of label-free mass spectrometric quantification of affinity purifications. Mol. Cell Proteomics https://doi.org/10.1074/mcp.M111.007955 (2012).
Semenov, A., Moykkynen, T., Coleman, S. K., Korpi, E. R. & Keinanen, K. Autoinactivation of the stargazin-AMPA receptor complex: subunit-dependency and independence from physical dissociation. PLoS One 7, e49282 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Partin, K. M., Bowie, D. & Mayer, M. L. Structural determinants of allosteric regulation in alternatively spliced AMPA receptors. Neuron 14, 833–843 (1995).
Article CAS PubMed Google Scholar
Perez-Riverol, Y. et al. The PRIDE database resources in 2022: a hub for mass spectrometry-based proteomics evidences. Nucleic Acids Res. 50, D543–D552 (2022).
Article CAS PubMed Google Scholar

Download references

Acknowledgements

We thank Jakob von Engelhardt for kindly providing the CKAMP44 cDNA. We also thank all Bowie lab members, past and present, for thoughtful discussion and suggestions, especially Marika Arsenault, Yuhao Yan and Xin-tong Wang for contributing some patch recordings. We especially thank Mina Moniri and Dr. Phil Biggin (molecular dynamics, Oxford University), Drs. Hana Antonicka and Eric Shoubridge (blue native PAGE analysis, McGill), and Drs. Ryan Alexander and Arjun Bhaskaran (native recordings in the somatosensory cortex, Bowie lab) for performing exploratory experiments that were ultimately not included in this study. We acknowledge the use of the Nikon AZ100M at the Vanderbilt CISR Core Facility. This work was supported by an operating grant from the Canadian Institutes of Health Research (CIHR) (to D.B.), National Institutes of Health (NIH) R01 grants HD061543 and MH123474 (to T.N.), and grants of the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) SFB 1381 (project-ID 403222702), FA 332/15-1 and 16-1 (to B.F.). A.M.P. was supported by Natural Sciences and Engineering Research Council of Canada (NSERC) CGS-M and PGS-D fellowships.

Author information

Authors and Affiliations

Integrated Program in Neuroscience, McGill University, Montreal, QC, H3A 1A1, Canada
Amanda M. Perozzo
Department of Pharmacology and Therapeutics, McGill University, Montreal, QC, H3G 1Y6, Canada
Amanda M. Perozzo & Derek Bowie
Institute of Physiology, Faculty of Medicine, University of Freiburg, Hermann-Herder-Str. 7, 79104, Freiburg, Germany
Jochen Schwenk & Bernd Fakler
Department of Molecular Physiology and Biophysics, Vanderbilt Brain Institute, Vanderbilt University School of Medicine, Nashville, TN, 37232, USA
Aichurok Kamalova & Terunaga Nakagawa
Center for Structural Biology, Vanderbilt University School of Medicine, Nashville, TN, 37232, USA
Terunaga Nakagawa
Signaling Research Centers BIOSS and CIBSS, University of Freiburg, Schänzlestr. 18, 79104, Freiburg, Germany
Bernd Fakler

Authors

Amanda M. Perozzo
View author publications
You can also search for this author in PubMed Google Scholar
Jochen Schwenk
View author publications
You can also search for this author in PubMed Google Scholar
Aichurok Kamalova
View author publications
You can also search for this author in PubMed Google Scholar
Terunaga Nakagawa
View author publications
You can also search for this author in PubMed Google Scholar
Bernd Fakler
View author publications
You can also search for this author in PubMed Google Scholar
Derek Bowie
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

A.M.P and D.B. conceived the project; A.M.P. performed electrophysiology experiments and related data analyses; J.S. performed experiments related to protein biochemistry and proteomic analyses; A.K. performed histology experiments; A.M.P., J.S., and A.K. made figures and evaluated data. T.N., B.F., and D.B. evaluated data and supervised the work. A.M.P. and D.B. wrote the manuscript with the support of all authors.

Corresponding author

Correspondence to Derek Bowie.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Perozzo, A.M., Schwenk, J., Kamalova, A. et al. GSG1L-containing AMPA receptor complexes are defined by their spatiotemporal expression, native interactome and allosteric sites. Nat Commun 14, 6799 (2023). https://doi.org/10.1038/s41467-023-42517-7

Download citation

Received: 15 May 2023
Accepted: 12 October 2023
Published: 26 October 2023
DOI: https://doi.org/10.1038/s41467-023-42517-7

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.