Apart from dynein-based transport of glycine receptors (Maas et al., 2006), the detailed retrograde trafficking route of inhibitory neurotransmitter receptors remains elusive. Here, we identified muskelin as a direct GABAAR α1 subunit binding protein that participates in receptor endocytosis and degradation. Muskelin is a widely expressed intracellular multidomain protein (Adams et al., 1998 and Adams et al., 2000), with high expression levels in hippocampus and cerebellum (Tagnaouti et al., 2007). Our data show that muskelin accompanies receptor transport through different motor protein complexes along both actin filament and MT networks. To identify GABAAR binding proteins that might

participate in the regulation Navitoclax mw of receptor targeting and/or turnover,

we BVD523 applied the LexA yeast two-hybrid system by using the intracellular GABAAR α1 TMIII-TMIV loop sequence (aa 334–420, Figure 1A) as bait. From 2.4 million clones of an adult rat brain library, we identified five putative GABAAR α1 binding partners including a single clone that coded for residues 90–200 of the multidomain protein muskelin (accession number NM_031359) containing a discoidin domain, a lissencephaly-1 (LIS1) homology (LISH) and C-terminal to LisH (CTLH) tandem domain, as well as repeated kelch motifs (Adams et al., 1998) (Figure 1B). The muskelin binding site of GABAAR α1 was mapped through TMIII-TMIV deletion mutants, which identified residues 399–420 as being sufficient for muskelin interaction (Figure 1A). Notably, TMIII-TMIV sequences of GABAAR α2, α3, α5, β2, or γ2 subunits

did not directly bind to muskelin in this assay (Figures 1C and 1D), while USP14 (a positive control) displayed binding (Figure S5 available online). To biochemically substantiate this interaction, we performed GST pull-down and coimmunoprecipitation (co-IP) experiments. Despite GABAAR α1, GABAAR α2 TMIII-TMIV loop-GST fusion proteins also, but not GST alone or fusions to α3, α5, β2 or γ2, displayed specific binding to myc-muskelin derived from HEK293 cells (Figure 1E). GABAAR α2 might associate with muskelin-GABAAR α1 complexes, as it binds to gephyrin (Tretter et al., 2008), which can also interact with muskelin (Figures S1A and S1B); however, GABAAR α2 does heptaminol not seem to be a direct muskelin binding partner (Figures 1C and 1D). Notably, as a control for muskelin-GABAAR α1 binding, deletion of the minimal muskelin-binding motif of GABAAR α1 (aa 399–420) abolished this interaction (Figure 1E). Accordingly, precipitation of endogenous GABAAR α1 led to specific coprecipitation of endogenous muskelin and vice versa with brain lysates (Figures 1F and 1G). Coimmunostaining of hippocampal neurons cultured for 12–14 days in vitro (DIV 12–14) indicated colocalization of muskelin and GABAAR α1 puncta in somata and neurites.