Among those newly identified GABA receptor associated proteins, GABARAP is the best studied. It was cloned in a yeast two-hybrid system using the intracellular loop of γ2 as bait (Wang et al. 1999). The interaction is specifically limited to γ1, γ2S, and γ2L, but not other GABAA receptor subunits (Wang et al. 1999; Nymann-Andersen et al. 2002b). GABARAP belongs to a family including GEC1 (first isolated in Guinea-pig Endometrial Cells, estrogen-induced 1.8 kb RNA coded protein) or GABARAPL1; GATE-16 (Golgi-associated ATPase enhancer of 16 kDa) or GABARAPL2 or GEF-2 (Ganglioside expression factor 2); GABARAPL3; GABARAPL4; LC3 (Microtubule-associated protein light chain 3); yeast Apg8 or Aut7p; and C. elegans LGGs (LC3, GABARAP and GATE-16) protein. Figure 1 describes the amino acid conservation among the members of this family. All the proteins in this family share some common features: first, they are evolutionally conserved in eukaryotic cells from yeast to mammals. For example, at the amino acid level, all orthologues (mouse, rat and bovine) of GABARAP present 100% identity with their human counterparts, indicating a very high level of conservation in evolution and suggesting a critical function of these proteins in mammalian cells.
Figure 1. Sequence alignment of Human GABARAP, GEC1, GATE-16, LC3, Yeast Apg8/Aut7, and their homologs from C. elegans LGGs.
Download figure to PowerPoint
Second, the existing studies of their three-dimensional structures showed that the proteins in this family share strong similarity to ubiquitin (Paz et al. 2000; Bavro et al. 2002; Coyle et al. 2002; Knight et al. 2002; Sugawara et al. 2004), exhibiting a compact fold consisting of four or five-stranded β-sheets with two α-helices on either side. The outer strands of the β-sheets are aligned antiparallel to the inner strands and helices α3 and α4 are located on one side of the sheet. The presence of α1 and α2 helices is a unique feature of the GABARAP family, which is lacking in ubiquitin. These helices are attached non-covalently to the ubiquitin-like core by a number of interactions, e.g. hydrophobic interactions and hydrophilic interactions including hydrogen bonds and salt-bridges. The first two α-helices (N-terminal a.a. 1–22) have been shown to form a hydrophobic patch suitable for protein–protein interfaces (Bavro et al. 2002; Coyle et al. 2002). In GABARAP and GEC1, this domain was defined as interacting with microtubules (Wang and Olsen 2000; Mansuy et al. 2004). Another hydrophobic patch found in GABARAP, composing of Ile21, Pro30, Tyr49, Leu50, Val 51, Lue55, Phe60, Leu63, and Phe-104, might be involved in dimerization or oligomerization, interaction with the γ2 subunit, and/or N-ethylamaleimide-sensitive factor (NSF) (Coyle et al. 2002; Nymann-Andersen et al. 2002a; Leil et al. 2004).
Third, the GABARAP family is not only structurally similar to ubiquitin, but also undergoes a similar post-translational cascade modification process as occurs in ubiquitinylation (Hochstrasser 2000). Yeast Apg8 was the first identified member in this family undergoing this process. During the modification, the last amino acid in the C-terminus is first cleaved by the activating cysteine protease Apg4, leaving glycine116 exposed. This glycine is conserved in all the proteins in this family. Then it is activated by Apg7 and transferred to Apg3, which are E1 and E2 enzymes, respectively (Komatsu et al. 2001; Yamazaki-Sato et al. 2003). Finally, it conjugates to another biomolecule, catalyzed by an as yet unidentified E3. Apg8 and the GABARAP family are not attached to proteins but rather to the membrane lipid phosphatidylethanolamine (PE) (Ichimura et al. 2000). The lipid PE can also be removed by Apg4, releasing the C-terminal glycine exposed for a new cycle (Hemelaar et al. 2003). The C-terminal glycine is critical for this modification because mutation of glycine to alanine in Apg8 prevents cleaving off the last amino acid (Ichimura et al. 2000). The homologues of Apg4, Apg7 and Apg3 were recently cloned in human and verified as the respective enzymes involved in the modification of mammalian counterparts of Apg8 (Tanida et al. 2001; Tanida et al. 2002; Hemelaar et al. 2003; Scherz-Shouval et al. 2003; Yamazaki-Sato et al. 2003; Tanida et al. 2004). Similar to Apg8, LC3 was conjugated to PE at the end of this process, which was also demonstrated in a [3H]PE incorporation experiment (Kabeya et al. 2004). Although the final conjugated adduct remains unknown for GATE-16 and GABARAP, several lines of evidence suggest that it is lipid-related (Kabeya et al. 2004). Accumulated evidence has shown that lipopeptides have an increased tendency to insert into membranes compared to the same peptide lacking the lipid moiety. Therefore, this modification must be one facet of the important scheme regulating membrane trafficking and cell surface stability (Casey 1995). However, no E3 enzyme has been identified for this modification. Given the fact that the members in this family also have their own unique functions, the E3 enzymes were hypothesized to be the key factor differentiating their location by conjugating them to different lipids, or different pools and locations.
Fourth, all the proteins are involved in intracellular trafficking. GATE-16 (GABARAPL2) was first purified from a bovine brain cytosol based on a cell-free intra-Golgi transport assay (Legesse-Miller et al. 1998). It directly binds NSF and GOS-28, a Golgi-specific v-SNARE, and is believed to be a component of the intra-Golgi transport machinery (Legesse-Miller et al. 1998; Sagiv et al. 2000; Muller et al. 2002). GATE-16 has been shown to enhance the ATPase activity of NSF, which, in turn, stimulates the recruitment of GATE-16 to the unpaired GOS-28 in an ATP-dependent or ATPase-independent manner. This interaction protects the labile, unpaired GOS-28 from proteolysis. Recently, it has been shown that the function of GATE-16 as a SNARE protector is essential not only for intra-Golgi transport, but also for the homotypic fusion of post-mitotic Golgi fragments (Sagiv et al. 2000; Muller et al. 2002; Elazar et al. 2003). GEC1 (GABARAPL2) was originally identified as a novel early estrogen-regulated gene (Nemos et al. 2003). The first N-teminal 22 a.a. of GEC1 show 77% identity with the tubulin-binding domain of GABARAP and was shown to provide the domain for interaction with tubulin. Morever, it promotes tubulin assembly and displays tubulin bundling activity. Like GABARAP and GATE-16, GEC1 also interacts with NSF (Chen et al. 2006). In vitro binding analysis revealed an association between GEC1 and GABAA receptor γ2 subunit and tubulin at the same time (Mansuy et al. 2004). A recent study demonstrated that GEC1 also interacts directly with the κ-opioid receptors, but not µ, δ-opioid receptors. Expression of GEC1 dramatically increased total and surface κ-opioid receptors (Chen et al. 2006). All these findings strongly suggest that GEC1 may be the second protein of the GABARAP family implicated in receptor membrane trafficking via microtubules (Mansuy et al. 2004).
LC3 was first identified as a protein that copurified with microtubule-associated protein 1A and 1B from rat brain (Mann and Hammarback 1994). The lipidated form of LC has been shown to locate in membranes and was proposed as a marker of autophagosomes (Kabeya et al. 2000). Yeast Apg8 is well accepted as an autophagic factor (Ichimura et al. 2000), but it also takes part in multiple intracellular membrane trafficking processes by interacting with different v-SNARE molecules involved in ER to Golgi transport and in vacuolar fusion (Legesse-Miller et al. 2000). C. Elegans LGG-1 has also been used as the autophagasomes marker (Melendez et al. 2003; Roudier et al. 2005; Rowland et al. 2006). Autophagosomes normally transport bulk cytoplasm to the lysosome for degradation. Autophagy is recently identified as a novel degradative pathway for GABAA receptors (Rowland et al. 2006). In C. elegans, the GABAA receptors lacking presynaptic innervation, but not acetylcholine receptors, traffic to autophagosomes after endocytic removal from the cell surface (Rowland et al. 2006). The formation of autophagasome might serve as the trafficking destination of GABAA receptor-containing endocytic vesicles, thus allowing the coordinated degradation of cytoplasmic and membrane-bound postsynaptic proteins.
Unlike ubiquitin, proteins in the GABARAP family do not form covalently bound polymers at the end of this ubiquitinylation-like modification. Instead, they might form non-covalent homo-dimers or oligomers. The crystal structure shows that GABARAP exists in two conformations: the monomeric conformation at low protein concentration and a head-to-tail oligomeric conformation in high salt conditions (Coyle et al. 2002). Nymann-Andersen et al. (2002a) reported that GABARAP self-associates and dimerizes in physiological salt concentrations. However, they found no evidence for higher order complexes larger than a dimer. By using deletion constructs of GABARAP and GABARAP-derived synthetic peptides, they limited the self interaction domain to amino acid 41–51, which overlaps with the GABAA receptor γ subunit binding domain (35–52 a.a). It is possible that dimerization regulates the availability of free GABARAP for association with trafficking ‘cargo’ proteins, like GABAA receptors. Further, if both GABARAP molecules interact via this domain with their dimer partner, higher order oligomers would be prevented.
Amino acids 35–52 of GABARAP were identified as the γ2 subunit binding domain (Nymann-Andersen et al. 2002a; Leil et al. 2004). GABARAP clusters have been shown to co-localize with the clusters of GABAA receptors in cultured neurons (Wang et al. 1999). Leil et al. (2004) further confirmed the co-localization, demonstrating that over-expressed GABARAP co-localized with the γ2 subunit in the cytoplasmic perinuclear regions of cultured hippocampal neurons. Although GABARAP did not appear to co-localize extensively with the surface GABAA receptors, some GABARAP was detected to co-localize with synaptophysin, an established marker for synaptic vesicles. GABARAP shows two aspects of effects on the γ2-containing GABAA receptors: (i) regulating the ion channel properties, possibly by promoting receptor clustering; and (ii) manipulating the cell surface expression, possibly by increasing the intracellular trafficking of receptors. In quail fibroblast QT-6 cells, a recombinant expression system, GABARAP clusters GABAA receptors; these clusters demonstrated lower apparent affinity for GABA, faster deactivation, and slower desensitization at a given GABA concentration (Chen et al. 2000). Disruption of microtubules prevents the clustering of both GABARAP and GABAA receptors (Chen et al. 2000; Wang and Olsen 2000), indicating that the clustering of GABARAP might be important for the properties of GABAA receptors. In addition, the N-terminal microtubule binding domain (a.a. 1–22) and γ2 subunit are also required for the clustering effect of GABARAP (Chen et al. 2000). The similar results were demonstrated in recombinant HEK293 cells expressing GABARAP with α1β2γ2S subunits in certain ratios. GABARAP altered desensitization, deactivation, and diazepam potentiation of GABA-mediated currents. However, the authors speculated that GABARAP does not alter receptor kinetics directly but by facilitating surface expression of γ2-containing GABAA receptors, instead of αβ pentamers lacking γ2 (Boileau et al. 2005). In L929 fibroblast cells expressing the α1β2γ2S subunits with GABARAP, the channel conductance was significantly higher than that expressing the same three subunits alone, even reaching super-high conductances observed in native receptors (Everitt et al. 2004), suggesting that GABARAP might participate in the organization of GABAA receptors on the membrane.
In cultured neurons, GABARAP mainly localized in intracellular compartments, such as endoplasmic reticulum and Golgi structures, with only a small fraction in synaptic membrane co-localizing with GABAA receptors (Kneussel et al. 2000; Kittler et al. 2001; Leil et al. 2004). Some GABARAP puncta were found to co-localize with the γ2 subunit near the surface of the soma at points of synaptic contact labeled by GAD65, a marker for GABAergic presynaptic terminals, suggesting that GABARAP might participate in transporting the γ2 subunit to the cell surface, or even to synapses (Leil et al. 2004). Electron microscope evidence reveals no anchoring role for GABARAP at GABAA receptor synaptic membranes but frequent co-localization with GABAA receptors in intracellular vesicles located near synaptic membranes (Kittler et al. 2001). Recent studies confirmed that GABARAP enhances the cell surface expression of γ2-containing GABAA receptors in both heterologous expression systems and cultured hippocampal neurons. Over-expression of GABARAP increased net surface levels of GABAA receptor in Xenopus laevis oocytes as shown by both increased GABA currents and surface-expressed protein (Chen et al. 2005). This GABARAP stimulation of GABA currents required the γ2 subunit and the microtubule-binding domain (a.a. 1–22) as well as the intact polymerization of microtubules, indicating the effect of GABARAP was based on the interaction with microtubules (Chen et al. 2005). However, GABARAP co-expression did not alter the general properties of GABAA receptors such as sensitivity to GABA or benzodiazepines in oocytes. Increased surface GABAA receptors were also detected in cultured hippocampal neurons by immunofluorescence and in COS-7 cells by flow cytometer. Mutagenesis of the key amino acids responsible for the γ2 subunit interaction blocks this enhancement effect and the interaction with NSF (Leil et al. 2004). The mechanism of this enhancement still remains unknown. It might be that GABARAP increases trafficking of receptor toward the cell membrane along microtubules, facilitates the insertion, or stabilizes the surface receptors.
Recently, the GABARAP knockout mouse has been shown to be phenotypically normal and exhibits no loss of GABAR synaptic punctate localization (O'Sullivan et al. 2005). This suggests the possibility that GABARAP is not necessary for synaptic localization. Although it traffics γ2-containing GABAA receptors to the surface, increasing surface expression, and under some conditions promotes clustering and synaptic localization, it may not be necessary, but merely accelerates or promotes this process, possibly by mass-action shifting to this state. Alternatively, it could be that the GABARAP trafficking of GABAA receptors to surface, and clustering, are needed for synaptic localization, but, the homologues, or at least one of them, probably GEC1, either also has this function normally, or assumes it as compensation in the GABARAP knockout mouse. However, the homologues GEC1 and GATE-16 did not show increased levels (O'Sullivan et al. 2005).
N-ethylmaleimide-sensitive factor (NSF)
GABARAP interacts and co-localizes with NSF (Kittler et al. 2001; Leil et al. 2004), which plays a central role in general membrane fusion events underlying intracellular trafficking. NSF uses energy from ATP hydrolysis to dissociate SNARE complexes after membrane fusion, allowing the individual SNARE proteins to be recycled for subsequent rounds of fusion. One of the main roles of NSF in intracellular trafficking is the fusion of synaptic vesicles with presynaptic membrane during neurotransmssion (Rothman 1994). In addition, NSF is also enriched in the postsynaptic density, regulating the neurotransmitter receptors in synapses (Walsh and Kuruc 1992; Hong et al. 1994; Puschel et al. 1994). The function of NSF in regulating neurotransmitter receptors is well characterized by the study of NSF and AMPA receptor interaction. NSF has been shown to regulate the surface insertion/stablization of GluR2-containing AMPA receptors and thus AMPA receptor-mediated synaptic transmission (Lin and Sheng 1998; Song et al. 1998; Braithwaite et al. 2000). It was suggested that the interaction between NSF and GluR2 does not promote receptor synthesis or passage through the endoplasmic reticulum or Golgi network, but rather facilitates the synaptic localization of AMPA receptors. GATE-16, a homologue of GABARAP, has been shown to enhance the ATPase activity of NSF and increase intra-Golgi trafficking. It is highly possible that GABARAP shares this ATPase enhancing character and functions as a chaperone helping to target GABAA receptors via intracellular vesicles to the cell surface and synapses.
GABARAP binds with gephyrin in both the yeast two-hybrid system and GST pull-down experiments (Kneussel et al. 2000). Gephyrin is the first isolated GABAA receptor associated protein which is tightly co-localized with both glycine and GABAA receptors at postsynaptic sites (Kneussel and Betz 2000; Luscher and Keller 2004). It is well established that inhibition of gephyrin expression leads to a dramatic decrease in GABAA receptor clustering. Kneussel et al. (2000) reported that clusters of the major subunits of GABAA receptors, α2 and γ2, were completely absent in hippocampal cultures from gephyrin knockout mice. Anti-sense against gephyrin also resulted in a corresponding reduction in density of clusters of α2 and γ2 subunits of GABAA receptors (Essrich et al. 1998). Inhibiting gephyrin expression by siRNA did not modify the total number of GABAA receptors expressed on the neuronal cell surface but significantly decreased the number of receptor clusters. In addition, the clusters that formed in the absence of gephyrin were significantly more mobile compared with those in control neurons, suggesting that gephyrin plays a specific role in reducing the diffusion of GABAA receptors, facilitating their accumulation at inhibitory synapses (Jacob et al. 2005). The interaction between gephyrin and dynein might be the motor for gephyrin recruitment, therefore, responsible for GABAA receptor clustering (Maas et al. 2006).
Conversely, γ2 knockout mice showed significant reduction of gephyrin clusters (Essrich et al. 1998). Gephyrin clusters were also lost when the γ2 subunit gene was inactivated in mature neurons by conditional knockout in 3-week-old-mice (Schweizer et al. 2003). Thus, the functional association between γ2 subunit-containing GABAA receptors and gephyrin is not limited to development of immature neurons and synaptogenesis, under conditions where GABA can be excitatory.
However, there are also some controversial results regarding the critical role of gephyrin in clustering. GABAergic inhibitory postsynaptic currents and the postsynaptic localization of α1 subunit-containing GABAA receptors in cultured hippocampal neurons, as well as clustering of these receptors in spinal cord sections from gephyrin knockout mice, appear to be unaffected by the loss of gephyrin (Kneussel et al. 2001). In different developmental stages (from 3 DIV to 21 DIV), some β3 and γ2 clusters associated with gephyrin were always detected in extrasynaptic sites and a proportion of synaptic gephyrin clusters were found not associating with GABAA receptor β3 or γ2 immunoreactivity, indicating that gephyrin alone is not sufficient to accumulate GABAA receptors γ2 subunits clusters at synaptic sites (Danglot et al. 2003). Levi et al. (2004) demonstrated that GABA miniature inhibitory postsynaptic currents (mIPSCs) could be recorded from cultured hippocampal neurons from gephyrin knockout mice. The mean amplitude was reduced; the frequency and kinetics remained unchanged. Furthermore, GABAA receptors could form clusters at synapses opposite to GABAergic terminals in neurons lacking gephyrin, albeit at reduced levels compared with control neurons. Cell surface labeling experiment indicated that gephyrin contributes, in part, to aggregation but not insertion or stabilization of GABAA receptor α2 and γ2 in the plasma membrane (Levi et al. 2004). Thus, a major gephyrin-independent component of hippocampal inhibitory synapse formation must exist.
There is no evidence indicating that gephyrin and GABAA receptors directly interact. GABARAP binds with both proteins directly, suggesting that GABARAP might function as the adaptor for the association of gephyrin and the γ2 subunit. Co-expression of both GABARAP and gephyrin results in a recruitment of cytoplasmic gephyrin to GABARAP-rich membrane-associated loci in PC12 cells. Since GABARAP might attach to membrane lipid, e.g., PE, any protein that it binds, such as gephyrin, will become membrane-bound. However, in gephyrin knockout mice, the immunoreactivity of GABARAP showed no significant change in number, size, or localization, suggesting that GABARAP functions prior to or independently of gephyrin (Kneussel et al. 2000).
Glutamate receptor interacting protein (GRIP)
Protein domain prediction demonstrated to us that GABARAP has a postsynaptic density protein (PDZ)-binding domain in the C-terminus (http://scansite.mit.edu/motifscan_seq.phtml). The PDZ domain (Postsynaptic density protein, Discs large, Zonula occludens-1) is a peptide motif of about 90 amino acids initially found in proteins associated with synaptic, septate and tight junctions. PDZ domains have been shown to be present either singly or as repeats in over 100 otherwise unrelated proteins and conceived as scaffolds for assembling multiprotein complexes that facilitate the integration of signaling components and pathways. Some PDZ domains interact with the intracellular C-terminus of ion channels and receptors and thus provide a mechanism for channel clustering. For example, GRIP1, a seven PDZ domain-containing protein, has been shown to interact with AMPA-type glutamate receptors and plays an essential role in their clustering and trafficking (Dong et al. 1997). However, very little evidence has shown the PDZ domain-containing proteins at inhibitory synapses. By yeast two-hybrid screening, GRIP was identified to associate with GABARAP. This association was further verified by GST-pull down and co-immunoprecipitation in COS cells (Kittler et al. 2004a). The PDZ domains 4–7 of GRIP1, corresponding to amino acids 413–794, are responsible for the interaction with GABARAP (Kittler et al. 2004a). GRIP1 has several spliced forms: GRIP1a, GRIP1b (Yamazaki et al. 2001; Yu et al. 2001) and GRIP1c (Charych et al. 2004b). GRIP1b differs from GRIP1a only in an alternate N-terminal region in which the first 18 amino acids of rat GRIP1 are replaced by a new sequence of 19 amino acids, and their downstream sequences appeared to be identical to each other (Yamazaki et al. 2001). The most notable differences between GRIP1a and GRIP1b are some cysteine residues at the N-terminus: GRIP1a contains a cysteine residue at position 8 and 10, while GRIP1b has a single cysteine at position 11. This difference makes the GRIP1b able to be palmitoylated, but GRIP1a is not (Yamazaki et al. 2001). GRIP1c lacks the PDZ domains 1–3 of GRIP1a/b, but contains PDZ domains 4–7. It also contains a specific 35 amino acid N-terminal and a 12 amino acid C-terminal peptide that differ from GRIP1a/b (Charych et al. 2004b). Therefore, GABARAP might bind to all these three isoforms of GRIP1.
GRIP is one of the few proteins present at both inhibitory and excitatory synapses although function of GRIP in the inhibitory synapses remains unknown. It has been reported that GRIP1a/b localized to both GABAergic and glutamatergic synapses in cultured hippocampal neurons (Dong et al. 1999; Wyszynski et al. 1999), but not to GABAergic synapses in the intact brain (Wyszynski et al. 1999). GRIP1c is found to be present in excitatory synapses in both cultured neurons and intact brain as demonstrated by immunofluorescence and electron microscope. Contrary to the other forms, it also localizes in GABAergic synapses, suggesting a possible role in GABAergic transmission (Charych et al. 2004b). GRIP1c does not coimmunoprecipitate with any GABAA receptors from brain extract, but with AMPA receptors (Charych et al. 2004b). As mentioned above, there is little evidence that PDZ domain-containing protein is found at inhibitory synapses. In addition, the C-terminus of GABAA receptors extends extracellularly and does not have a PDZ-binding domain. Therefore, GRIP1c might interact with GABAA receptors through GABARAP or other GABAA receptor associated proteins and participate in regulating GABAA receptor trafficking and clustering on the cell membrane. Electron microscopy demonstrated that both GRIP1 and GABARAP were found in the intracellular compartments, e.g., Golgi. The interaction between GRIP1 and GABARAP might exist at the level of intracellular compartments.
Phospholipase C-Related, catalytically Inactive Proteins (PRIP)
The PRIP family has been identified as a group of novel inositol 1,4,5-trisphosphate binding proteins with a domain organization similar to phospholipase C-δ but lacking the enzymatic activity (Kanematsu et al. 2000). PRIP-1 has been shown to regulate Ins (1,4,5) P3-mediated Ca2 + signaling by modulating type 1 inositol polyphosphate 5-phosphatase activity through binding to Ins (1,4,5) P3 (Harada et al. 2005). The family consists of at least two types of protein, PRIP-1 and PRIP-2. They have a number of binding partners, including the catalytic subunit of protein phosphatase 1α and GABARAP (Kanematsu et al. 2002; Uji et al. 2002). Both PRIP-1 and PRIP-2 compete with the γ2 subunit of the GABAA receptor for its binding site on GABARAP (Kanematsu et al. 2002). Northern blot analysis and reverse-transcription polymerase chain reaction showed that PRIP-1 was presented mainly in the brain, whereas PRIP-2 was expressed ubiquitously (Uji et al. 2002). The in situ hybridation revealed that PRIP-1 mRNA is localized in hippocampal pyramidal cells, dentate granule cells, pyramidal and granule cell layers of the cerebral cortex, and in the granule cell and Purkinje cell layers and cerebellar nuclei of the cerebellum. This regional distribution is similar to that of GABAA receptors (Kanematsu et al. 2002). PRIP-1 does not appear to be enriched at inhibitory synaptic sites and the majority of PRIP-1 and GABARAP co-localized as intracellular punctate structures in cultured cortical neurons (Kanematsu et al. 2002). Electrophysiological analysis demonstrated no overt differences in functional cell surface GABAA receptor number in the absence of PRIP-1. In agreement with the above data, binding assays indicated no marked difference in the number of GABA or BZ binding sites between wild type and PRIP-1 knockout mice. Immunoblotting results confirmed that the expression of the major GABAA receptor subunits in hippocampal and cortex neurons does not appear to be affected by the lack of PRIP-1. However, diazepam induced remarkably lower effects in neurons of PRIP-1 knockout mice, as shown by both reduced efficacy and potency. On the contrary, Zn2+ has lost its normal inhibitory effect on GABA-induced effects in hippocampal neurons from PRIP-1 knockout mice. Behaviour analysis revealed that motor coordination was impaired and intraperitoneal injection of diazepam induced markedly reduced sedative and antianxiety effects in PRIP-1 knockout mice. These results indicate that PRIP-1 is essential for the function of GABAA receptors, especially in response to the agents acting on γ2 subunit-containing isoforms (Kanematsu et al. 2002).
PRIP-1 also binds to the intracellular domains of all GABAA receptor β subunits and weakly with γ2 subunits directly, but not to those of the α1–6, δ or ρ1 subunits. In PRIP-1 knockout mice, forskolin caused less phosphorylation of β3 subunits than in wild-type mice (Terunuma et al. 2004). This reduction is not related to PKA because the PKA activity and the association between PKA and GABAA receptors remain the same in PRIP-1 knockout mice. Instead, the phosphatase activity of PP1α is elevated in PRIP-1 knockout mice, which is in agreement with the result that PRIP-1 is a negative modulator of phophatase (Yoshimura et al. 2001; Terunuma et al. 2004). Further experiments demonstrate that PP1α, PP2A and some unknown phosphatase all participate in dephosphorylation of the β3 subunit but with PP1α playing the predominant role, which explains how PP1α dephosphorylation of the β3 subunit can be negatively regulated in PRIP-1 knockout mice (Terunuma et al. 2004). PRIP-1 itself can be phosphorylated by PKA on T94 and S96. The former site is likely responsible for regulating the dissociation of PP1α from PRIP-1. PRIP-1 is also required for dopamine receptor modulation of GABAA receptor function. Taking this evidence together, PRIP-1 is likely to play a significant role in the phospho-dependent modulation of GABAA receptor activity regulated by endogenous neuronal signaling pathways (Terunuma et al. 2004).
Recent studies indicate that GABARAP plays a more generalized function in cells rather than binding exclusively to neuronal GABAA receptors.
A yeast two-hybrid screen using the cytoplasmic domain of transferrin receptors identified GABARAP as the dominant interactor of transferrin receptors. The interaction between the two proteins is direct and the amino acids 36–117 of GABARAP are necessary, as shown in GST-pull down experiments, indicating that GABAA
receptors and transferrin receptors are docking on either the same site or overlapping sites on GABARAP. However, GABARAP does not affect the endocytosis rate of transferrin receptors, possibly because GABARAP is sequestered in an intracellular compartment and interact with transferrin receptors in a process that is distinct from endocytosis. Immunofluorescence studies indicated that GABARAP is not primarily associated with endocytic vesicles but traffics transferrin receptors along another pathway such as the biosynthetic or degradative pathway (Green et al. 2002
-51-like kinases, ULK1 and ULK2, are homologues of unc
-51 of C. elegans
and Apg1p of yeast (Yan et al. 1998
). Mouse ULK has been shown to be involved in axonal elongation (Tomoda et al. 1999
). By yeast two-hybrid experiment, ULK1 was found to interact with both GATE-16 and GABARAP (Okazaki et al. 2000
). This interaction suggests that these molecules might be important for vesicle transport and axonal elongation in mammalian neurons.
GABARAP was recently found to interact with DEAD (Asp-Glu-Ala-Asp/His) box polypeptide 47 (DDX47) (Lee et al. 2005b
), an RNA helicase responsible in premRNA splicing or ribosome biogenesis (Jankowsky and Jankowsky 2000
). DDX47 induces apoptosis by interaction with GABARAP.
The GABARAP gene was also detected as an insertion in bovine viral diarrhea virus genome from infected cows (Becher et al. 2002
). The insertion of GABARAP served as a processing signal to yield NS3, critical for the cytopathogenicity of the virus. It interacts with the N-terminus of NS3, with the last amino acid cleaved off (Becher et al. 2002
). Besides GABARAP, other genes were also detected to insert into some viral genomes, including ubiquitin, LC3, GATE-16 and the J domain protein Jiv.
By carrying out suppression substractive hybridization to detect differentially expressed genes between a breast cancer cell line and a non-tumorigenic microcell hybrid cell line, GABARAP was identified to display high expression levels in the non-tumorigenic cell line. The down-regulation of GABARAP was also seen in human breast cancer tissue. Over-expression of GABARAP in a breast cancer cell line induced a reduction of growth rate and tumorigenicity, indicating that GABARAP acts as tumor suppressor in breast cancer (Klebig et al. 2005