Neurochemical Research

, Volume 33, Issue 8, pp 1459–1465

Post-translational Regulation of l-Glutamic Acid Decarboxylase in the Brain

Authors

    • Department of Basic Science, Charles E. Schmidt College of Biomedical ScienceFlorida Atlantic University
    • Department of Basic Science, Charles E. Schmidt College of Biomedical ScienceFlorida Atlantic University
Original Paper

DOI: 10.1007/s11064-008-9600-5

Cite this article as:
Wei, J. & Wu, J. Neurochem Res (2008) 33: 1459. doi:10.1007/s11064-008-9600-5
  • 438 Views

Abstract

Gamma-aminobutyric acid (GABA) is the primary inhibitory neurotransmitter in the central nervous system. GABA is converted from glutamic acid by the action of glutamic acid decarboxylase (GAD). There are two forms of GAD in the brain, GAD65 and GAD67, referring to a molecular weight of 65 and 67 kDa, respectively. Perturbations in GABAergic neurotransmission have been linked to a number of neurological disorders. Since GAD is the rate-limiting enzyme in controlling GABA synthesis, it is important to understand how GAD is regulated in the brain. It is known that GAD function can be regulated at the transcriptional/translational and post-translational levels. This review focuses briefly on the recent advances in revealing the post-translational regulation of GAD function including protein phosphorylation, palmitoylation and activity-dependent cleavage. The results from these studies have improved our understanding of the regulation of GAD function in the brain.

Keywords

Post-translational modificationGAD65GAD67PhosphorylationTruncated GADCalpainPalmitoylation

Abbreviations

CSP

Cysteine string protein

GAD

Glutamic acid decarboxylase

HIP14

Huntingtin interacting protein 14

HSP70

Heat shock protein 70

IDDM

Insulin-dependent diabetes mellitus

PAT

Palmitoyl acyltransferase

PPT

Protein palmitoyl thioesterase

SPS

Stiff person syndrome

SV

Synaptic vesicles

VGAT

Vesicular GABA transporter

Introduction

Gamma-aminobutyric acid (GABA), the major inhibitory neurotransmitter in the mammalian central nervous system, is synthesized by l-glutamate acid decarboxylase (GAD) (EC 4.1.1.15). Unlike the other neurotransmitter synthesizing enzymes, there are two forms of GAD in the brain, GAD65 and GAD67, referring to a different molecular weight of 65 and 67 kDa respectively [1].

It is not clear why two forms of GAD are needed in the brain. But it is well established that GAD65 and GAD67 have different subcellular localizations and functions in the brain. GAD65 is found predominantly to be associated with synaptic vesicle membranes in the nerve terminals and therefore could mediate fast, point-to-point GABA neurotransmission [2]. GAD67 is evenly distributed throughout the cell and GABA synthesized by GAD67 is used for the other functions such as trophic factor for neuronal development or energy source [3, 4]. New evidence from structural studies of GAD65 and GAD67 indicated that GAD67 is constitutively active and is responsible for the basal GABA production while GAD65 is transiently activated in response to the extra demand of GABA in neurotransmission [5]. Studies of transgenic mice revealed that genetic knockout of GAD67 is lethal [6], while GAD65 knockout mice could still survive but are susceptible to seizures, anxiety and epilepsy, indicating an impairment in GABA neurotransmission [7].

We have demonstrated that GAD65, but not GAD67, forms a protein complex consisting of heat shock protein 70 (HSP70), vesicular GABA transporter (VGAT) and cysteine string protein (CSP) [8, 9]. This protein complex provides a structural basis for a functional coupling model between GABA synthesis and transport into synaptic vesicles that we proposed previously [8, 9]. This model illustrates an important mechanism that would obviously be an efficient way to rapidly load GABA vesicles, especially when GABA is in high demand. This model could also provide molecular/theoretical basis for several reports from the other laboratories. Firstly, GAD65KO mice are more susceptible to seizure, anxiety and epilepsy [2, 10], indicating that there is an impairment in GABA transmission due to the lack of GAD65; secondly, physiological studies in GAD65KO mice revealed that both the quantal size and frequency of GABA-mediated miniature inhibitory postsynaptic currents (IPSCs) appear to be normal, but the release of GABA is reduced during sustained stimulation [2], indicating that GAD65 is responsible for rapid refill of GABA content in synaptic vesicles; thirdly, when GAD65 was expressed in excitatory glutamatergic neurons of the substantia nigra in a Parkinson’s disease rat model, it could induce a phenotypic shift from excitatory to inhibitory response with enhanced GABA release. However, when GAD67 was expressed, the phenotype was much less obvious and the excitatory response was still predominant [11], further supporting the role of GAD65 in neurotransmission.

Alterations in the level of GABA or GAD in CNS have been linked to a number of neurological disorders, including Huntington’s disease [12], Parkinson’s disease [11, 13], anxiety [10] and epilepsy [2]. In addition, GAD65 has been implicated as an autoantigen in several human autoimmune diseases, such as insulin-dependent diabetes mellitus (IDDM) [10], Stiff-person syndrome [14] and Batten disease [15]. Interestingly, GAD65 and GAD67 are differentially altered following a pathological stimulus. For example, significant increases in the number of GAD67-immunoreactive neurons in the external and internal segments of the globus pallidus while no significant difference in the number of GAD65-immunoreactive neurons was observed in the monkeys rendered Parkinsonian by systemic MPTP administration [16]. In a separate rat model with seizures induced by kainate, an up-regulation of GAD67 mRNA, but not GAD65 mRNA, was observed in dentate granule cells following seizures [17]. Decreased GAD67 expression was also observed in schizophrenia [18, 19].

Regulation of GAD can be addressed at the transcriptional/translational and post-translational levels. In this review, we will focus on the post-translational regulation of GAD in three aspects: protein phosphorylation, palmitoylation and cleavage of GAD depending upon neuronal activity.

Phosphorylation of GAD

Reversible phosphorylation of proteins plays central roles in the regulation of most critical cellular processes which occurs in both prokaryotic and eukaryotic organisms. It is well known that many enzymes and receptors become activated or inactivated by phosphorylation and dephosphorylation.

We have shown that protein phosphorylation and dephosphorylation play an important role in regulation of GAD activity in the brain. GAD65 and GAD67 activity appear to be regulated in opposite manner by phosphorylation. GAD65 is activated upon phosphorylation, while GAD67 is inhibited by phosphorylation. Soluble GAD, presumably GAD67, was inhibited by protein phosphorylation and its activity was markedly increased by phosphatase treatment [20]. On the other hand, the membrane associated GAD, presumably GAD65, is markedly activated by protein phosphorylation through a membrane associated protein kinase and inhibited under dephosphorylation conditions [21]. Later, these findings were further confirmed using a well-defined GAD65 and GAD67 system, i.e., highly purified recombinant human GAD65 and human GAD67 [22]. The kinases and phosphatases that are involved remain largely unknown. Identification of the protein kinases/phosphatases involved is very important in revealing the signal pathway that regulates GABA neurotransmission. In vitro studies showed that protein kinase A (PKA) is responsible for phosphorylation and inhibition of GAD67 activity while calcineurin is the phosphatase responsible for dephosphorylation and activation of GAD67 [20, 22]. In addition, analysis of the phosphorylated GAD67 by matrix assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry revealed that threonine 91 is phoaphorylated by PKA [22]. GAD65 is activated by phosphorylation mediated by protein kinase Cε (PKCε) and inactivated by protein phosphatase 2A in vitro [22]. The phosphorylation site involved has not been identified. It is interesting to see whether PKCε would also regulate GAD65 activity in vivo. Numerous studies have shown the functional link between PKC isoforms and changes in the neurotransmitter release [23, 24]. The 83.4-kDa PKCε isoform is particularly interesting, because it is expressed at very low levels in several normal tissues but is expressed at much higher levels in the brain, especially the presynaptic nerve terminals [25]. PKCε can be translocated to the membrane fraction in rat cortical synaptosomes upon activation [26]. The evidence for the specific involvement of PKCε in brain functions is emerging [27]. Therefore, it is conceivable that upon neuronal stimulation, PKCε translocates from cytosol to synaptic vesicle membrane or filamentous actin [28] and this translocation brings PKCε close enough to phosphorylate and activate GAD65. Whether and how PKCε actually regulates GAD65 activity in vivo needs to be further investigated.

It is important to note that the kinases that can be tested in in vitro systems are limited, therefore, it is conceivable that other kinases could phosphorylate GAD and regulate its function. From the deduced amino acid sequence, it appears that GAD65 and GAD67 has more putative phosphorylation sites for several protein kinases than identified. Some of these sites and the putative kinases are shown in Fig. 1.
https://static-content.springer.com/image/art%3A10.1007%2Fs11064-008-9600-5/MediaObjects/11064_2008_9600_Fig1_HTML.gif
Fig. 1

Human GAD65 and GAD67 sequence alignment shows identified and putative phosphorylation sites (underlined). Phosphorylation sites are predicted with a NetPhos 2.0 server (http://www.cbs.dtu.dk/services/NetPhos/). Putative kinase consensus motifs highlighted in: grey, PKA [R-X-S/T] or [R-R/K-X-S/T]; yellow, CaMKII [R-X-X-S/T-X]; pink, GSK-3 [S-X-X-X-S(P)]; green, MAP kinase [X/P-X-S/T-P]; turquoise, PKC [R/K-X-S/T] or [S/T-X-R/K], or non-classical PKC [K/R position −6, −4, −2, and F/V/L at +2, +3 to S/T]; red, PKG; dark yellow, casein kianse II (CKII) [X-S/T-X-X-E]. Identified phosphorylation site for PKA in GAD67 is T91 (indicated as boxed). Box 1 indicates the active binding site for cofactor PLP. Other putative phosphorylation sites with unidentified kinase concensus motif are indicated as red. Palmitoylation sites cysteine 30 and 45 in GAD65 are indicated in red box

GAD phosphorylation not only affects its activity but also regulates its subcellular localization. It has been reported that phosphorylation of serine residues 3, 6, 10, and 13 in GAD65 is involved in the anchoring of GAD65 to membrane and has no effect in GAD65 activity [29]. By searching the consensus sequence motif, we found that they are the potential phosphorylation site for mitogen-activated protein kinase (MAPK) and Glycogen synthase kinase 3 (GSK-3) (Fig. 1). GAD65 is isolated both in a cytosolic and synaptic vesicle (SV) membrane anchored form. We have found that under neuronal stimulation, more GAD65 become membrane associated in response to the increased demand of vesicular GABA (unpublished results). Under neuronal stimulation, phosphorylation of GAD65 at the N-terminal sites could facilitate trafficking of GAD65 to SV membrane, leading to increased GABA uptake into SV and release.

Palmitoylation of GAD

Protein palmitoylation refers to the posttranslational addition of a 16-carbon fatty acid, palmitic acid, to the side chain of cysteine, forming a thioester linkage. The enzyme that mediates the palmitoylation of cellular proteins is called palmitoyl acyltransferase (PAT) [30]. The reduced sulphydryl group on GAD (Cysteines 30 and 45) is deprotonated to form a thiolate. The thioester bond between cysteine group (C30 and C45) in GAD65 and palmitate is formed as a consequence of a nucleophilic attack of the thiolate on the α-carbon of palmitoyl-CoA (Pal-CoA). In contrast to PAT, protein palmitoyl thioesterase (PPT) is a small glycoprotein that removes palmitate groups from cysteine residues in lipid-modified proteins. This reversible modification increases protein hydrophobicity and facilitates protein interactions with lipid bilayers, and it can influence protein sorting and function [31]. Moreover, specific physiological stimuli can dynamically alter protein palmitoylation levels, providing an important mechanism for regulating cell development and signaling. Recent work has shown that palmitoylation is important for the regulation of neuronal development and synaptic functions [32]. It has been implicated in processes that include protein sorting, axonal development, presynaptic signaling, G-protein signaling, ion channel clustering and postsynaptic plasticity [33].

It has been reported palmitoylation of cysteines 30 and 45 is critical for post-Golgi trafficking of GAD65 to presynaptic sites and for its relative dendritic exclusion, leading to the presynaptic clustering of GAD65 [34, 35]. Mutation of Cysteine 30 and 45 to Alanine abolished the presynaptic clustering of GAD65 in primary hippocampal neurons [35]. The enzyme involved was later found to be huntingtin interaction protein 14 (HIP14). Although HIP14 was first discovered as a huntingtin interacting protein, using biochemical assays, Huang et al. showed that HIP14 is a neuronal palmitoyl transferase (PAT) and contains a DHHC domain, a cysteine rich region that harbors a conserved tetra-peptide motif, similar to the PATs in yeast [32]. HIP14 shows remarkable substrate specificity for neuronal proteins. Silencing of HIP14 expression by siRNA in cultured neurons reduced axonal clustering of GAD65 [32]. Interestingly, GAD67 has not been reported to be palmitoylated. Palmitoylation of GAD65 could explain why GAD65 is concentrated in the axonal nerve terminals while GAD67 is not. Palmitoylated GAD65 colocalizes with Rab5 in Glogi membranes and axons, and Rab5 regulates the trafficking of palmitoylated GAD65 from Golgi membranes to axons in an endosomal trafficking pathway [34]. However, it is not clear how Rab5 regulates the trafficking of GAD65.

Both GAD67 and GAD65 have N-terminal cysteines, but only GAD65 is palmitoylated. There is no well-defined consensus sequence for palmitoylation other than a requirement for cysteine [36]. Inspection of the GAD65 and GAD67 sequences does not reveal any obvious motif that preferentially direct palmitoylation towards GAD65 (Fig. 1). It is possible that the difference between the three-dimensional structures of GAD65 and GAD67 is the key factor that specifies palmitoylation. GAD65 could expose the N-terminal cysteines for palmitoylation while the N-terminus of GAD67 is not accessible for palmitoylation. The recent structures for GAD65 and GAD67 were obtained in the absence of N-termini of GAD65 and GAD67 [5]. Therefore, further structural studies with full length GAD65 and GAD67 are needed to fully understand the differences in palmitoylation of GAD65 and GAD67. Another interesting hypothesis is that GAD65 and GAD67 could interact with different intracellular proteins. A yet unidentified protein that specifically interacts with GAD65 may facilitate GAD65 palmitoylation, or on the other hand, a protein that specifically interacts with the N-terminus of GAD67 may block the palmitoylation.

Cleavage of GAD

The presence of smaller molecular weight forms of GAD have been reported either derived from RNA splicing or from proteolytic cleavage of full length GAD. For instance, 25 and 44 kDa GAD were reported to be derived from GAD67 RNA splicing [37, 38]. The 25 kDa is enzymatically inactive and is present usually early in the development [37], although it has been reported in adult retina [39]. The 44 kDa GAD is enzymatically active and has been shown to synthesize GABA [38]. The truncated forms of GAD derived from proteolytic cleavage during the course of purification of recombinant GAD from a bacterial expression system [40, 41] have been reported.

Previously, we reported that recombinant human GAD65 could be cleaved into a more stable truncated form in vitro by Factor Xa or mild trypsin treatment [41]. Our in vitro enzyme activity assay indicated that this truncated form is more active than the full length GAD65. N-terminal sequencing of the truncated form revealed that this cleavage occurs between Arginine 69 and Lysine 70. This observation is consistent with the report that the region around 69–70 is a proteolytic hot spot in native GAD65 [42]. It is of interest that in rat pancreatic β-cells, GAD65 has been shown to be released spontaneously from islet cell membrane as a 57–58 kDa hydrophilic soluble fragment in a time dependent manner and the kinetics of this reaction are consistent with an enzymatic cleavage in islet cells [42]. In addition, we also found that GAD67 is cleaved at two specific sites, one at arginine 70 and another one at arginine 90 to produce two truncated forms of GAD67. In the course of the search of endogenous proteases responsible for cleavage of GAD in vivo, it has come to our attention that GAD65 could be cleaved to a truncated form upon neuronal stimulation. Since neuronal stimulation causes Ca2+ influx, we speculate that the cleavage of GAD65 involves Ca2+ and is highly regulated. In addition, GAD is also found to be cleaved in a neuronal activity-dependent manner. The cleavage of GAD only occurred when synaptosomes were under sustained stimulation [42]. Furthermore, the conversion of full-length GAD65 to truncated GAD65 is not the result of random post-mortem degradation, but that it is an intracellular process that is highly regulated [43]. The cleavage is mediated by calpain, a Ca2+-dependent cysteine protease.

Calpain is a well studied intracellular, calcium-dependent cysteine protease. It is widely expressed in the CNS and strictly modulated by an endogenously expressed specific inhibitory protein, calpastatin [44, 45]. There are two ubiquitously expressed calpain isoforms, μ-calpain (calpain-1) and m-calpain (calpain-2). Each calpain is composed of a unique large catalytic subunit (80 kDa) and a common small regulatory subunit (30 kDa). The two isoforms have nearly identical substrate specificities and, so far, have been shown to differ mainly in their calcium requirement for activation in vitro. m-Calpain requires millimolar levels of calcium whereas μ-calpain is active at micromolar concentrations of calcium. Calpain plays a crucial role in mediating adaptive changes in response to a range of environmental signals. Calpain plays important roles in many different physiological processes, such as cell mobility, signal transduction, cell cycle, regulation of gene expression, apoptosis and long-term potentiation [44, 45], as well as in pathological processes [46, 47]. For example, in the CNS calpain has been reported to be involved in several neurodegenerative diseases such as Alzheimer’s disease [48, 49] and Parkinson’s disease [50].

The physiological significance of the cleavage remains elusive. The N-terminus of GAD65 has been reported to be involved in membrane anchoring of GAD65 to synaptic vesicles [29]. It is thus conceivable that the cleavage of full-length GAD65 to truncated GAD65 may decrease or abolish the binding of GAD65 to synaptic vesicle membranes, resulting in an impairment of GABA transmission as predicted from the functional coupling between GAD65 and VGAT [9].

In addition to the essential role played by GAD65 in neurotransmission, it is also a typical autoantigen in several human autoimmune diseases, such as IDDM [51] and SPS [52, 53]. Although both SPS and IDDM are associated with autoantibodies against GAD65, higher autoantibody levels and recognition of a linear N-terminal epitope in the autoantigen GAD65 distinguish SPS from type 1 diabetes [54]. However, little is known about how GAD65 is made available to the immune system during the autoimmune response. It is believed that the N-terminus of GAD65 is masked by membrane interactions under normal conditions, but how the N-terminal epitope becomes exposed in patients with SPS is unclear. We propose that cleavage of GAD65 at the N-terminus by calpain could facilitate the generation of antibodies specific to linear epitopes in these regions. Under continued stimulation or other neurotoxic conditions such as ischemia [48], influx of calcium causes activation of calpain which leads to the cleavage of GAD65, unmasking the N-terminus of GAD65. This N-terminus could then be further secreted out by unknown mechanisms for antibody generation in patients with SPS. This idea is consistent with the report that cytoplasmic processing by calpain is required for presentation of endogenous GAD65 by major histocompatibility complex class II proteins [55]. It would be interesting to see whether calpain is abnormally activated in the brains of individuals with SPS, leading to the conversion of full-length GAD65 to the truncated form.

Future Directions

As discussed above, significant progress has been made in understanding the post-translational regulation of GAD65 and GAD67 in the brain. However, the underlying mechanism is largely unknown. In the aspect of GAD phosphorylation, most of the studies were performed in in vitro systems. Although this in vitro study is invaluable in characterizing the potential for regulation of GAD, it requires further characterization in vivo. Future research will involve at least two aspects of investigation. The first aspect is to extend the in vitro results more precisely into the in vivo area, Identification of the kinases/phosphatases and their related phosphorylation sites are necessary to further understand the physiological significance of GAD phosphorylation. Recent advances in mass spectrometry and other proteomic methods promise to facilitate the identification of the phosphorylation sites in GAD. Generation of phosphorylation site-specific antibodies will greatly facilitate the study of GAD phosphorylation in situ. The second aspect is to establish a functional role for the post-translational modification in response to the physiological or pathological interventions. Pharmacological regulators of the post-translational modification are likely to provide therapeutic targets for GAD-related neurological disorders.

Acknowledgment

This work was supported by the National Institutes of Health (NS37851 to J-Y Wu).

Copyright information

© Springer Science+Business Media, LLC 2008