While GluA2-containing calcium-impermeable AMPA receptors (CI-AMPARs) are widely expressed in the central nervous system, the expression of GluA2-lacking calcium-permeable AMPA receptors (CP-AMPARs) is more restricted – they are only found in a subset of neuronal and glial cells or at specific developmental stages. These CP-AMPARs play an important role in the physiology of the brain. Not only do they contribute to basal transmission, but they are also essential to normal development and synaptic plasticity. On the other hand, the over activation or upregulation of CP-AMPARs, has been linked to diverse neurological disorders, including hypoxic/ischemic damage to neurons and glia, chronic pain and drug addiction. As the relative contribution of CP-AMPARs to synaptic transmission is critical in both physiological and pathological processes, it is of interest to understand the molecular mechanisms underlying their regulation. Transmembrane AMPAR regulatory proteins (TARPs) enhance both AMPAR expression and channel function. With this in mind, we asked whether these auxiliary subunits play a role in the differential trafficking of CP- and CI-AMPARs. Much of our knowledge of TARP/AMPAR interactions comes from work on cerebellar granule cells, neurons that normally express only CI-AMPARs. Like granule cells, cerebellar stellate cells express TARPs γ-2 and -7, but contain both CP- and CI-AMPARs. We have examined AMPAR subtype distribution in these two cell types, using patch-clamp recording. In particular, we have considered which TARPs are involved in the synaptic and extrasynaptic expression of different AMPAR subtypes, using the spontaneous mutant mouse stargazer. Experiments on stargazer played a central role in the discovery of TARPs as essential components of the native AMPAR complex. These ataxic mice lack TARP γ-2 and display a characteristic loss of AMPAR-mediated currents at mossy fiber-to-granule cell synapses. This is associated with a near complete absence of surface AMPARs. Closer examination has demonstrated that, γ-2 not only promotes the delivery of AMPARs to the plasma membrane, but also enables the clustering of surface AMPARs at synapse by specifically interacting with PDZ-containing proteins of the postsynaptic scaffold (PSD-95). Six functional TARPs have been identified, according to their ability to rescue AMPAR-mediated EPSCs in stargazer granule cells – TARPs γ-3, -4, -and -8 (which are closely related to γ-2), and the atypical TARPs, γ-5 and -7. The latter interact physically and functionally with AMPARs but fail to restore synaptic transmission in stargazer granule cells. Each of these six TARPs displays distinct, partially overlapping, patterns of expression in the brain. However, it has generally been assumed that only one TARP subtype is present within a given AMPAR complex. TARP γ-2 and -7 are enriched within the cerebellum, where they co-cluster at synaptic contacts [1]. Importantly, no other TARP is expressed in cerebellar granule cells, which would explain why these neurons are particularly susceptible to the loss of γ-2 in stargazer mice. Unlike granule cells, stellate cells in stargazer mice still display AMPAR-mediated EPSCs. However, we find that the contribution of CI-AMPARs to these synaptic currents is dramatically decreased [2]. This implies that CP-AMPARs, unlike CI-AMPARs, can cluster at synapses in the absence of γ-2, or any other typical TARPs. Surprisingly, while functional γ-7-associated CP-AMPARs are present at the surface of stargazer stellate cells, EPSCs are mediated by TARPless CP-AMPARs. Whereas it was believed that γ-7 simply fails to compensate for the lack of γ-2 in stargazer granule cells, we show that γ-7 prevents AMPAR surface expression and synaptic clustering, as knocking-down γ-7 rescues synaptic currents. Our observations indicate that the loss of surface AMPARs in stargazer granule cells lacking γ-2 is caused by the absence of CP-AMPARs and a suppressive effect of γ-7 . Together our results suggest that the relationships between AMPARs and TARPs, and between typical and atypical TARPs, are more complex than originally thought, and that there is still much to learn from stargazer mice about AMPAR trafficking.