AMPA receptors (AMPARs) are responsible for fast excitatory synaptic signalling in the brain. A majority of AMPARs are heterotetramers composed of four pore-forming subunits (GluA1-4), co-assembled with transmembrane AMPAR regulatory proteins (TARPs). Six TARP family members have been described which differ in their effects on receptor trafficking and function (Jackson & Nicoll, 2011). While many central neurons contain more than one type of TARP, it is unknown whether different TARPs can co-associate within with the same receptor and further modify individual AMPAR function.

Structurally, TARPs possess four transmembrane domains and intracellular C- and N- termini. While the TARP’s backbone mediates binding to AMPARs, its C-terminal tail (CT) binds to postsynaptic density protein (PSD-95) through a PDZ binding motif (Bats, 2007). Functionally, TARP family members fall into two categories, type I (gamma-2, -3, -4 and -8) and type II (gamma-5 and -7), according to their ability to rescue excitatory postsynaptic currents (EPSCs) in stargazer cerebellar granule cells (CGCs). Stargazer mice are a spontaneous mutant that lack gamma-2 (stargazin) expression in homozygous individuals. CGCs of these mice lack AMPAR-mediated synaptic currents, suggesting these neurons rely on gamma-2 for AMPAR surface expression and synaptic clustering. Transfecting any type I TARP in stargazer CGCs can compensate for loss of endogenous gamma-2. While only type I TARPs promote AMPARs surface delivery and synaptic expression, TARPs from both subtypes enhance AMPAR channel function. They do this by increasing the AMPAR deactivation- and desensitization time, increasing net cation influx. Of the TARPs, gamma-4 produces the slowest decaying AMPAR-mediated currents.

We made patch clamp recordings from CGCs cultured from postnatal stargazer mice. Transfection of a chimera containing the CT of gamma-2 fused to the backbone of gamma-4 gave rise to miniature EPSCs that decayed remarkably slowly when compared with wild type gamma-2 and gamma-4 (τ_w 22.7 ± 1.7 ms versus 3.2 + 0.2 ms and 4.8 ± 0.6 ms; all n = 5; p = 0.007 and p < 0.001, two-sided permutation t-test, Ho, 2019). Such combined action of the gamma-4 backbone and gamma-2 CT within the chimera suggests that gamma-2 and gamma-4 may use different mechanisms to control AMPARs kinetics. We therefore asked whether gamma-2 and gamma-4 could co-associate within individual AMPARs to confer distinct kinetic properties. To address this, we expressed a chimera containing the CT of gamma-7 fused to the backbone of gamma-4 (gamma-4_7CT) in cultured CGCs. As this chimera lacks the ability to promote the trafficking of AMPARs to synapse, any increase in mEPSCs decay time is expected to reflect the synaptic insertion of receptors containing both gamma-2 and gamma-4_7CT. Indeed, mEPSCs in these neurons displayed significantly slower decay times than those seen in GFP-transfected controls (τ_w 4.2 ± 0.3 ms versus 2.2 ± 0. 3ms, n = 5 and 7; p = 0.0026). Together, these results suggest that, in CGCs that contain endogenous gamma-2 and transfected gamma-4_7CT, a population of synaptic AMPARs contained both TARPs, and that the receptors can show some features derived from each.