The key findings of this study are 1) Intracellular accumulation of the δ subunit was distinctly slower than that of the γ2 subunit, 2) Internalization of the δ subunit was modulated by network activity but independent of ligand-binding, similar to that of the γ2 subunit, 3) BDNF reduced intracellular accumulation of the δ subunit, and 4) BDNF increased surface expression of the δ subunit through activation of TrkB receptors, PLCγ and PKC.
Internalization of the δ subunit has not been studied in the past. Here we report that the rate of intracellular accumulation of the δ subunit was distinct from that of the β2/3 and γ2 subunits. The δ subunit accumulated in the intracellular compartment slowly and reached a peak between 3–6 hr. In contrast, peak intracellular accumulation of the γ2 subunit occurred within 30–60 min. Recent findings have demonstrated differential surface expression of GABAR subunits under pathological conditions. Prolonged seizures of status epilepticus reduced surface expression of the β3 and γ2 subunits, whereas surface expression of the δ subunit either increased or remained unaltered (Goodkin et al., 2008
; Terunuma et al., 2008
). In contrast, alcohol intoxication reduced surface expression of the δ subunit and increased surface expression of the γ2 subunit (Liang et al., 2007
). These studies suggested that trafficking of the δ and γ2 subunit-containing receptors is differentially modulated in pathological conditions. The data presented here show that the trafficking of these subunits is distinct even under the physiological condition.
Internalization of the δ subunit was modulated by neuronal activity; treatment of slice cultures with TTX reduced internalization and increased surface expression, suggesting a negative correlation between activity and surface expression of the δ subunit. In contrast, in the animals in status epilepticus, surface expression of the δ subunit increased or remained unaltered, suggesting a correlation between activity and surface expression of the δ subunit (Goodkin et al., 2008
; Terunuma et al., 2008
). However, in these studies surface expression was determined only after an hour of increased neuronal activity, whereas we have used 6 hr TTX treatment. Bicuculline did not alter internalization of the δ subunit, suggesting that binding to the GABA had minimal effect on internalization of these receptors. Our previous studies suggest a similar ligand independent internalization of the γ2 subunit as well (Goodkin et al., 2008
). Hence the difference in the kinetics of internalization of the δ and γ2 subunit is not likely to be due to difference in GABA affinity for receptors composed of these subunits.
The δ and γ2 subunits are mutually exclusive in vivo
(Araujo et al., 1998
) and may have different intracellular accumulation rates. The β2 and δ subunits are known to co-assemble and form functional receptors (You and Dunn, 2007
; Herd et al., 2008
). If these subunits are part of the same receptors, their internalization ought to be similar. However, we found that the rate of intracellular accumulation of the δ and β2/3 subunits differed. The reasons for this difference are presently unclear, though, several possibilities may account for the difference. One possibility is that the majority of the δ subunit may assemble with the β1 subunit. Immunochemical studies in cultured hippocampal neurons support this possibility. In these neurons immunoreactivity of the β1 subunit and the δ subunit appeared diffuse and was mostly concentrated in the cell soma, whereas immunoreactivity of the β2/3 subunits appeared distinct and spread in the cell soma as well as processes (Mangan et al., 2005
Another possibility is that there may be two pools of the β2/3 subunits: a fast-trafficked pool assembled with the γ2 subunits, and a slow-trafficked pool assembled with the δ subunit. Homomeric β3 receptors have been observed in recombinant systems (Wooltorton et al., 1997
; Taylor et al., 1999
) though whether such receptors are also present in vivo
is not known. In the hippocampus, αβ3 receptors are also expressed (Mortensen and Smart, 2006
). Hence internalized β2/3 subunits can originate from three pools of membrane expressed receptors: αxβ2/3δ, αxβ2/3γ2 and αxβ3. A slow-trafficked fraction of the β2/3 subunits assembled with the δ subunit may not be detected against the large population of the fast-trafficked fraction of the β2/3 subunits. It is unclear whether the β2 and β3 subunits have distinct rate of intracellular accumulation in neurons, though studies using recombinant receptors suggest that both the subunits are trafficked quickly (Connolly et al., 1999
; Kittler et al., 2000
; Cinar and Barnes, Jr., 2001
). Determining the percentage of the δ subunit assembled with the β1, β2 and β3 subunits under basal conditions may help to explain distinct rates of intracellular accumulation observed in our studies.
Rules of internalization may also change depending on the partnering subunits. Such differential dominance of subunits is known for AMPA receptors (Lee et al., 2004
). In these receptors, internalization of GluR1 and GluR2 subunits is distinct; however, under heteromeric condition GluR1/GluR2 receptors internalize similarly to those consisting of only GluR2 subunits. Homomeric δ subunit-containing GABARs have not been reported under in vivo
conditions however, the δ subunit may dominate other subunits in the pentamer.
Intracellular accumulation depends not only on the rate of internalization, but also the rate of reinsertion and rate of degradation. Fast recycling or degradation of the internalized δ subunit may result in a delay in peak intracellular accumulation. Studies addressing the rate of reinsertion and degradation of the internalized δ subunit may help to explain slow intracellular accumulation.
The β2/3 and γ2 subunits undergo clathrin-dependent endocytosis though non-clathrin mediated endocytosis has also been observed in recombinant receptors (Cinar and Barnes, Jr., 2001
). Clathrin, dynamin, AP2 complex, and amphiphysin regulate internalization of the β2/3 and γ2 subunits [(Kittler et al., 2000
) see reviews (Michels and Moss, 2007
; Jacob et al., 2008
)]. Proteins which mediate the internalization of the δ subunit are not known. So far, only the μ2 subunit of the AP2 complex has been shown to interact with the δ subunit (Kittler et al., 2005
) and the significance of this interaction remains to be examined. A weaker interaction between the δ subunit and the μ2 subunit of the AP2 complex may result in slower removal of the δ subunit from the membrane, resulting in delayed peak intracellular accumulation.
BDNF reduced intracellular accumulation of the δ subunit and increased its surface expression and these effects were independent of activation of TrkB receptors. Although activation of TrkB receptors is known to induce epileptic activity (Rivera et al., 2002
), the effect of BDNF on surface expression of the δ subunit appeared to be independent of this action, because TTX failed to block BDNF effects. In fact, both BDNF and TTX increased surface expression of the δ subunit. Enhanced neuronal excitation is prevalent action of BDNF (Koyama and Ikegaya, 2005
), however it also increases expression of glutamic acid decarboxylase (GAD), a GABA synthesizing enzyme (Yamada et al., 2002
), and the density of inhibitory synapses in the hippocampal slice cultures (Marty et al., 2000
), suggestive of a potentiation of inhibitory neurotransmission. Our results suggest that neuronal activity and BDNF potentially act on the δ subunit via independent mechanisms. BDNF treatment for 6 hr is also known to increase total expression of the α4 subunit in the hippocampal neurons (Roberts et al., 2006
), which is a preferred partner of the δ subunit (Sur et al., 1999
; Peng et al., 2002
). Although BDNF did not alter total expression of the δ subunit, increased expression of the α4 subunit may result in augmentation of surface receptors. Further studies are necessary to understand whether this BDNF effect was through direct modulation of the δ subunit or whether any other GABAR subunit was involved. In the hippocampus, GABARs composed of the δ subunit are major mediators of the tonic inhibition. BDNF-induced increased surface expression of the δ subunit is likely to increase tonic inhibition.
Increased surface expression of the δ subunit in BDNF-treated slice cultures was dependent on activation of PLCγ but not PI3K. This is unlike that of the β2/3 subunits, which involve both PLCγ and PI3K in mediating BDNF effects in the hippocampal neurons (Jovanovic et al., 2004
). PI3K mainly regulates recycling of endosomes (Martys et al., 1996
; Kurashima et al., 1998
) and insertion of GABARs at the membrane (Wang et al., 2003
). The results of the present study suggest that PI3K may not play a significant role in mediating BDNF effects on the δ subunit surface expression. PKC inhibition reduced, and its activation increased, the surface expression of the δ subunit. PKC phosphorylates S409 in β1, 408/409 in β2/3, and S327/343 in γ2 GABARs subunits (Moss et al., 1992b
; McDonald and Moss, 1997
) and a positive correlation between phosphorylation and surface stability has been observed (Kittler et al., 2005
). Similar PKC-mediated phosphorylation of the δ subunit may regulate its stability at the cell surface; however, direct phosphorylation of the δ subunit has not yet been demonstrated. An indirect effect of PKC may include a yet unidentified protein that can regulate membrane stability of the δ subunit. Approximately 30% of the surface-expressed δ subunit had moved to the intracellular space within 4 hr, and BDNF treatment reduced this by half. However, treatment with BDNF resulted in 60–80% increase in the surface expression of the δ subunit. This difference suggests that in addition to internalization, BDNF may also affect the rate of forward trafficking.
In conclusion, our results demonstrate slower intracellular accumulation of the δ subunit compared with that of the β2/3 and γ2 subunits. BDNF effects on surface expression of the δ and γ2 subunits were distinct.