Identification of C-Terminal Domain Residues Involved in Protein Kinase A-Mediated Potentiation of GluR6

Abstract

Glutamate receptors are the major excitatory receptors in the vertebrate central nervous system and have been implicated in a number of physiological and pathological processes. Previous work has shown that glutamate receptor function may be modulated by protein kinase A (PKA)-mediated phosphorylation, although the molecular mechanism of this potentiation has remained unclear. We have investigated the phosphorylation of specific amino acid residues in the C-terminal cytoplasmic domain of the rat GluR6 kainate receptor as a possible mechanism for regulation of receptor function. The C-terminal tail of rat GluR6 can be phosphorylated by PKA on serine residues as demonstrated using [γ-³²P]ATP kinase assays. Whole cell recordings of transiently transfected HEK293 cells showed that phosphorylation by PKA potentiates whole cell currents in wildtype GluR6 and that removal of the cytoplasmic C-terminal domain abolishes this potentiation. This suggested that the C-terminal domain may contain residue(s) involved in the PKA-mediated potentiation. Single mutations of each serine residue in the C-terminal domain (S815A, S825A, S828A, and S837A) and a truncation after position 855, which removes all threonines (T856, T864, and T875) from the domain, do not abolish PKA potentiation. However, the S825A/S837A mutation, but no other double mutation, abolishes potentiation. These results demonstrate that phosphorylation of the C-terminal tail of GluR6 by PKA leads to potentiation of whole cell response, and the combination of S825 and S837 in the C-terminal domain is a vital component of the mechanism of GluR6 potentiation by PKA.

Glutamate receptors (GluRs) are the major excitatory receptors in the vertebrate central nervous system (CNS) and have been implicated in a number of normal CNS functions including synaptic plasticity, and pathological processes such as epilepsy, Parkinson’s disease, schizophrenia, and ischemic cell death (Dingledine et al., 1999). GluRs are ligand-activated membrane receptors and consist of two large families: metabotropic (linked through G proteins to downstream second messengers) receptors termed mGluRs and ionotropic (current passing) receptors referred to as iGluRs (Dingledine et al., 1999). iGluRs have been subdivided based upon their agonist specificities and sequence homology into N-methyl-D-aspartate (NMDA) α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid (AMPA), and kainate receptors (Lerma, 2006). Each receptor is believed to be composed of a tetramer of subunits that oligomerize to form functional channels in the cell membrane (Dingledine et al., 1999). To date, five mammalian kainate receptor subunits (GluR5-7, KA1-2) have been identified based upon their high affinity for this compound (Bettler et al., 1990, Egebjerg et al., 1991, Werner et al., 1991, Bettler et al., 1992, Herb et al., 1992).

Previous studies suggest that iGluR function may be modulated by protein phosphorylation (Raymond et al., 1993, Wang et al., 1993, Wang et al., 1994, Roche et al., 1996, Lee et al., 1998). Basal and induced phosphorylation of AMPA and kainate GluRs by protein kinase A (PKA), protein kinase C (PKC), and Ca⁺⁺/calmodulin-dependent protein kinase II (CaM kinase II) have been demonstrated (McGlade-McCulloh et al., 1993, Moss et al., 1993, Raymond et al., 1993). Phosphorylation of NMDA receptors has similarly been reported (Chen and Huang, 1992, Tingley et al., 1993, Lieberman and Mody, 1994). Mutational studies, immunochemical assays, phosphoamino acid analysis and phosphopeptide mapping suggest that basal and kinase-dependent phosphorylation of GluRs occurs primarily on serine and threonine residues (Blackstone et al., 1994, Omkumar et al., 1996, Roche et al., 1996, Barria et al., 1997, Leonard and Hell, 1997, Mammen et al., 1997, Tingley et al., 1997, Lee et al., 1998). PKA-mediated potentiation of whole cell currents has been demonstrated in human embryonic kidney (HEK293) cells transiently transfected with rat GluR6 cDNA (Raymond et al., 1993, Wang et al., 1993). Raymond et al. (1993) demonstrated that intracellularly applied PKA catalytic subunit (PKA_cat) increased the amplitude of glutamate-induced whole cell current. Although the specific site(s) of phosphorylation was (were) not identified, potentiation was abolished by a S684A point mutation and by inclusion of an inhibitor of PKA (PKI) in the pipette solution. Wang et al. (1993) observed potentiation of kainate-induced currents in GluR6-transfected HEK293 cells when PKA_cat was directly perfused into individual cells. A S684A point mutation decreased PKA_cat-induced current potentiation while a S666A mutation had no effect. The double mutation S684A/S666A abolished PKA_cat-induced potentiation. These experiments were reported at a time when the assumed transmembrane topology of the glutamate receptor placed the N- and C-termini and the S1 region in an extracellular location and S2 intracellularly (Hollmann et al., 1989, Keinänen et al., 1990). This topology assumed that the sites of interest in these studies (i.e., S684 and S666) were intracellular and were therefore exposed to PKA_cat and PKI. Subsequent immunocytochemical, phosphorylation, and N-glycosylation studies (Hollmann et al., 1994, Wo and Oswald, 1994, Bennett and Dingledine, 1995, Wo and Oswald, 1995a, Wo and Oswald, 1995b) indicate an extracellular location for S2, thereby placing positions 684 and 666 in an extracellular location (Figure 1). Reconciliation of the findings of Raymond et al. (1993) and Wang et al. (1993) with the corrected transmembrane topology (Hollmann et al., 1994, Wo and Oswald, 1994, Wo and Oswald, 1995a) of the GluR6 subunit requires one or a combination of the following possibilities: (1) extracellular phosphorylation of sites 684 and 666 by kinases in the extracellular milieu (or by PKA released from lysed cells in the Raymond paper), (2) intracellular phosphorylation of GluR6 en route from the endoplasmic reticulum (ER) to the cell membrane, or (3) the presence of one or more additional intracellular sites for PKA-mediated phosphorylation of GluR6. The goal of the present study was to identify putative site(s) in the cytoplasmic C-terminal domain involved in mediating PKA phosphorylation.

Figure 1.

Schematic of the domains of the GluR6 receptor. ATD refers to the amino terminal domain and M1 through M4 refer to membrane-associated regions of the protein. The structure shown is the S1S2 domain of GluR6 (1S7Y; Mayer, 2005). The vertical lines in the sequence illustrate where the truncations were made for Figure 3 (line before S815; GluR6ΔC-term) and Figure 2, 4, and 5 (line before T856; GluR6Δ855).

While the C-terminal tail of rat GluR6 does not contain a primary PKA consensus site (R-R-X-S/T, where X is any amino acid; Zetterqvist et al., 1976, Kennelly and Krebs, 1991), it does contain a number of sites that could interact with PKA based upon similarity of neighboring residues to the classic recognition motif (Figure 1). Although the PKA consensus sequence is relatively well defined, previous work has shown that PKA can phosphorylate target peptides whose amino acid sequence differs somewhat from this sequence (Kemp et al., 1975, Smith et al., 1999). Our experimental approach was to determine whether the isolated C-terminal tail of GluR6 is phosphorylated by PKA, and if so, whether serine or threonine residues in the C-terminus were phosphorylated. Mutant receptors carrying the specific serine/threonine residues were expressed in HEK293 cells and the effects of altering specific residues on PKA-mediated potentiation of whole cell currents were determined. By this approach, we have identified two serine residues in the C-terminal tail of GluR6; together these two serine residues mediate fully the potentiation of the whole cell current response, once they are phosphorylated by PKA.

Methods

Molecular biology

Mutations were introduced into WT rat GluR6 (Q form; kindly provided by Dr. Steve Heinemann, Salk Institute) using the QuikChange^® XL Site-Directed Mutagenesis Kit (Stratagene). Oligonucleotides were synthesized and verified by the Bioresource Center, Cornell University.

Bacterial expression of GluR6 C-terminal tail-Glutathione S-transferase (GST) fusion protein

cDNA encoding the WT C-terminal tail of rat GluR6 (G809-A877) was generated by PCR (Saiki et al., 1988). PCR fragments were subcloned into EcoR1 and BamH1 sites of the PGEX-2T plasmid vector (Amersham Pharmacia Biotech) predigested with the same restriction enzymes, and the resultant constructs were sequenced. This construct was used as a template for site directed mutagenesis. WT and mutant constructs were then transformed into E. coli BL21 (Cohen et al., 1972), cultured at 37°C in LB broth (Fisher Scientific, Fair Lawn, NJ) supplemented with 100 μg ml⁻¹ ampicillin to an OD₅₆₀ of 0.7 and induced with 250 μM isopropyl-1-thio-β-D-galactopyranoside for four hours at 37°C. Bacteria were harvested and lysed in ice cold PBS by three freeze-thaw cycles followed by sonication. The lysed bacterial extract was centrifuged at 12,000 × g for 20 minutes and the supernatant was incubated with glutathione-Sepharose 4B (Amersham Pharmacia Biotech) for one hour. Sepharose beads with bound GluR6 tail-GST fusion protein were then washed and incubated with thrombin (Amersham Pharmacia Biotech; 1 unit/sample) at 4°C for 16 hours where necessary to cleave GST from GluR6 tail peptide. Samples were subjected to SDS-PAGE to verify presence of GluR6 tail-GST fusion protein and successful thrombin cleavage. Thrombin treated and untreated protein samples were then used as substrates in parallel for [γ-³²P]ATP/kinase assays.

[γ-³²P]ATP/PKA assay

PKA phosphorylation reactions were carried out for 30 min at 37°C using 1 μl PKA_cat (100 U ml⁻¹), 25 μl GluR6 tail-GST fusion protein preparation, kinase buffer (40 mM HEPES, 40 mM MgCl₂, 50 mM β-glycerophosphate, 4 mM Na₃VO₄, 4 mM dithiothreitol, 2 mM phenylmethylsufonyl fluoride, 2 mM benzamidine, adjusted to achieve the same total volume for all tubes), 5 μM nonradioactive ATP and 9 μM [γ-³²P]ATP (20 μCi/tube; New England Nuclear, Boston, MA). Following the kinase reaction, the mixture was diluted 1:1 in 2X SDS loading buffer (100 mM Tris, pH 6.8, 4% SDS, 20% glycerol, 0.25% bromophenol blue) and resolved by SDS-PAGE. The degree of phosphorylation was evaluated by autoradiography.

Cell Culture and Transfection

HEK293 cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM; Life Technologies, Gaithersburg, MD) supplemented with 10% (v/v) fetal bovine serum (FBS; Life Technologies), 100 U ml⁻¹ sodium penicillin G, and 100 μg ml⁻¹ streptomycin sulfate in a humidified 5% CO₂ incubator at 37°C. When 60% confluent, cells were cotransfected with either WT or mutant rat GluR6 cDNA in pcDNA1/Amp expression vector (Invitrogen, Carlsbad, CA) and green fluorescent protein (GFP) cDNA in the pEGFP-N1 plasmid (Clontech Laboratories, Palo Alto, CA) using the calcium phosphate method (Chen and Okayama, 1987). One microgram of GluR6 DNA was used for each 35 mm dish. GFP was used as a marker of gene expression in transiently transfected cells (Plautz et al., 1996, Subramanian and Srienc, 1996). GFP cDNA concentration was maintained at 50% that of GluR6 concentration to increase the positive predictive value of GFP. Transfection efficiency of GFP-cotransfected cells was evaluated after 36 hours by visual inspection of GFP fluorescence. For use in experiments, cells were collected, resuspended in growth medium, and replated at lower density 30 min prior to study to facilitate lifting cells off of the dish. Immediately prior to study, the replated cells were washed twice with Dulbecco’s phosphate buffered saline (PBS; Life Technologies) and were bathed in the extracellular solution (145 mM NaCl, 3 mM KCl, 1 mM CaCl₂, 2 mM MgCl₂, 10 mM HEPES, and 5 mM glucose buffered to pH 7.4) for our study.

Electrophysiological recording

The whole cell current recording configuration of the patch clamp technique was used (Hamill et al., 1981). Pipettes were pulled from borosilicate glass capillary tubes to a tip diameter of 2–3 μm and a resistance of 2–3 MΩ when filled with the pipette solution (140 mM KCl, 10 mM HEPES, 11 mM EGTA, 1 mM CaCl₂, 2 mM MgCl₂, 2 mM tetraethylammonium, and 4 mM ATP, buffered to pH 7.3). PKA_cat (100 U ml⁻¹; Sigma Chemical Company, St. Louis, MO) was added to the pipette solution when potentiation by phosphorylation was tested. Recordings were obtained using an Axopatch 200B amplifier (Molecular Devices Corporation, Sunnyvale, CA) and passed through an internal low pass filter with a cutoff frequency of 5 kHz before being digitized at a sampling frequency of 20 kHz and analyzed using PClamp 6 or 8 (Molecular Devices Corporation). All recordings were obtained at a holding potential of −60 mV at 21°C using the cells between 36 and 96 hours after transfection. After achieving the whole cell patch clamp configuration, the cells was lifted off the dish and placed within 50 μm of the 150 μm diameter outflow of a U-tube application device (Udgaonkar and Hess, 1987). The solution exchange time for this device, determined by using open pipette tip measurements of voltage changes in response to 500 mM CsCl, was approximately 0.9 ms. Given unstirred layers and the shape of the cell, the 10–90% solution exchange time was 1–3 ms (Francis et al., 2001). The first agonist application (1 mM glutamate) was achieved within two minutes of obtaining whole cell configuration. Subsequent agonist applications were delivered at 4 min intervals.

Data analysis

The PKA-mediated potentiation was typically maximal or near maximal at 6 min, so the ratio of the 6 min to 2 min current amplitude was used a measure of the effect of PKA. This accounted for the effects of variable GluR6 expression. The Wilcoxon rank sum test was used to determine statistical significance of differences in normalized current amplitudes among GluR6 constructs. The two tailed Fisher exact test was used to determine the statistical significance of differences in percentages of cells demonstrating potentiation among GluR6 constructs.

Results

[γ-³²P]ATP/PKA assay

Previous work suggests that kinase-mediated modulation of ionotropic glutamate receptor function may involve direct phosphorylation of residues within the C-terminal domain (Kohr and Seeburg, 1996, Roche et al., 1996, Lee et al., 1998, Zheng et al., 1998, Banke et al., 2000). PKA, a kinase that has been implicated in the modulation of glutamate receptor function and more specifically in the potentiation of agonist-induced currents in GluR6 (Raymond et al., 1993, Wang et al., 1993, Banke et al., 2000), phosphorylates peptides primarily on serine and threonine residues (Kennelly and Krebs, 1991, Songyang et al., 1994). The C-terminal domain of rat GluR6 contains four serine residues (S815, S825, S828, and S837) and three threonine residues (T856, T864, and T875; Figure 1). Although the C-terminal tail of rat GluR6 does not contain a primary PKA consensus site (R-R-X-S/T, where X is any amino acid; Zetterqvist et al., 1976, Kennelly and Krebs, 1991, Songyang et al., 1994), it does contain a number of sites that could interact with PKA based upon charge distribution and similarity of neighboring residues to the established recognition motif. To test whether PKA can phosphorylate the C-terminal domain of GluR6, and, if so, to determine whether phosphorylation occurs on serine or threonine residues, γ [³²P]ATP/PKA assays using bacterially expressed WT and mutant C-terminal domain GST-fusion proteins as substrate were performed. As shown in Figure 2A, the WT C-terminal tail of GluR6 was phosphorylated by PKA. The S815A/S825A/S828A/S837A quadruple mutation abolished phosphorylation of bacterially expressed GST-fusion protein (Figures 2A and C). However, truncation of the C-terminal segment immediately C-terminal to K855 (removing T856, T864, and T875; Glur6Δ855), did not abolish phosphorylation. These findings suggest that one or more of the serine residues in the WT construct were likely phosphorylated. Thrombin cleavage of GST from phosphorylated WT and Glur6Δ855 fusion protein complexes resulted in a loss of signal intensity when compared to uncleaved peptides. This suggests that the 8 kD phosphorylated C-terminal portion of the GST fusion peptide migrates to the nonresolvable dye front once cleaved by thrombin and demonstrates that phosphorylation occurs on the C-terminal segment peptide. As a control, GST alone was not phosphorylated by PKA (data not shown). Taken together, these results indicated that the peptide mimicking the C-terminal tail of GluR6 can be phosphorylated by PKA on one or more of its serine residues. We then constructed a series of mutant receptors based on these results, expressed these holoreceptors in HEK293 cells and determined if specific serine residues in the C-terminal region of GluR6 were required for PKA phosphorylation, using whole cell recording.

Figure 2.

Phosphorylation of the GST-C terminal peptide of GluR6. (A) The first lane is the wildtype peptide conjugated to GST. In the second lane, the GST has been cleaved from the peptide and the peptide ran at the tracking dye (not shown). The fourth lane shows the peptide for which all four serines were mutated to alanine (GST conjugate) and the fifth lane is the same construct treated with thrombin. The sixth lane is the peptide truncated before T856 (GST conjugate; GluR6Δ855) and the seventh lane is the same construct treated with thrombin to cleave the GST. (B) The first lane is the wildtype conjugated to GST, the second lane is a similar construct with the S825A mutation, and the third lane is the S837A mutation. (C) Quantitation of the autoradiograms. To control for variability between replicates, values were normalized to the intensity for wildtype. Significance was tested with a one-sample t-test. Significant differences (p < 0.05) are indicated with an asterisk.

Whole cell recording

C-terminal tail truncation

To test whether the C-terminal domain of GluR6 was required for PKA-mediated potentiation of GluR6, we compared whole cell current responses in HEK293 cells transiently transfected with WT rat GluR6 cDNA with those obtained in cells transfected with a mutant GluR6 cDNA construct in which the entire cytoplasmic C-terminal domain was deleted. WT GluR6 showed a time-dependent potentiation of whole cell current upon serial application of 1 mM glutamate when PKA was included in the intracellular pipette buffer (Figure 3A). However, truncation of the total C-terminal domain (GluR6ΔC-term) rendered the mutant insensitive to the intracellular presence of PKA (Figure 3B) and led to a rundown in current, suggesting that the C-terminal domain is essential to the mechanism of PKA-mediated potentiation of GluR6.

Figure 3.

The effect of truncation of the GluR6 C-terminus before S815. When PKA is included in the pipette, the peak current for WT GluR6 increases as a function of time. When the protein is truncated before S815 (removal of all serines and threonines from the C-terminal tail, current rundown is observed even in PKA.

Selective serine and threonine mutagenesis

The functional role of specific C-terminal serine and threonine residues in PKA-mediated potentiation of GluR6 was investigated using site-directed mutagenesis. Whole cell currents elicited by application of 1 mM glutamate six min after attaining whole cell configuration were normalized to those obtained upon initial (two min) agonist application. The ratio is multiplied by 100, and the mean of all replicates is reported below as “normalized current.” The results of the [γ-³²P]ATP PKA assays (Figure 2A) suggest that if the effects of PKA on GluR6 were the result of phosphorylation of residues within the C-terminal tail domain, they were likely to be mediated by phosphorylation of at least one serine residue. To pinpoint specific serine residues, single serine-to-alanine mutations were introduced at S815, S825, S828 and S837. Examples of two and six min agonist applications are shown for WT and S815A, S825A, S828A, and S837A constructs (Figure 4) to demonstrate that all of these single serine mutants could be potentiated in the presence of PKA (summarized in Figure 6). Whole cell recordings of HEK293 cells transfected with WT rat GluR6 cDNA demonstrated potentiation in 64% of cells studied (n = 33), with a normalized current of 131%. Exclusion of PKA from the pipette buffer resulted in a lack of potentiation in any cells studied (n = 5), with a normalized current of 82%. That is, in the absence of PKA, current rundown was consistently observed. Whole cell current potentiation was retained in spite of serine-to-alanine mutations at either S815 (60% of cells potentiated, normalized current: 121%, n = 25), S825 (60% of cells potentiated, normalized current: 121%, n = 30), S828 (83% of cells potentiated, normalized current: 158%, n = 12) or S837 (52% of cells potentiated, normalized current: 121%, n = 27), suggesting that these residues are not individually required for PKA-mediated potentiation in GluR6. Although the percent potentiation for S825A and S837A were numerically less than WT, the difference was not statistically significant; however, these values differed significantly from currents in the absence of PKA (Figure 6A).

Figure 4.

Selected current traces from GluR6 WT, single serine to alanine mutations, and GluR6Δ855. In all cases, PKA was included in the pipette and measurements were taken 2 min and 6 min after formation of the whole cell patch.

Figure 6.

(A) Average percent change in maximal current for 6 min following the formation of the whole cell patch versus 2 min following formation of the patch. The error bars represent standard errors. (B) Percent of cells that showed potentiation for each of the conditions tested. The asterisk in both (A) and (B) indicates a statistically significant difference from wildtype (p < 0.05). Triple refers to the S815A/S825A/S837A mutation, and GluR6Δ855 refers to the truncation before T856.

To determine whether the C-terminal domain threonine residues are involved in PKA-mediated potentiation of GluR6, a truncation mutation was introduced into rat GluR6 immediately C-terminal to K855 (GluR6Δ855). This mutation, which removed all three threonine residues in the C-terminal tail domain, did not abolish potentiation (Figure 4), which was seen in 78% of cells studied (normalized current: 160%, n = 18), suggesting that the threonine residues are not essential for GluR6 potentiation by PKA.

Taken together, the whole cell recording results of the single serine and GluR6Δ855 mutations ruled out the possibility of either single serine residues or any combination of threonine residues being involved in mediating the phosphorylation-induced whole cell current potentiation. Instead, these results suggested a possibility that multiple serine residues may be required for PKA-mediated potentiation of GluR6. Initially, we engineered the S815A/S825A/S837A GluR6 (triple) mutant construct. In all but one case, the six-minute current amplitude was smaller than the two-minute current amplitude (normalized current: 78%, n = 8; Figure 5), suggesting that some combination of these serine residues (S815, S825, and S837) is required for PKA-mediated potentiation. To determine which combination of the three candidate serine residues was involved in potentiation, we engineered three double serine-to-alanine mutants on a GluR6Δ855 background. Examples of these whole-cell recording experiments are shown in Figure 5. The S815A/S825A mutation failed to abolish potentiation (56% of cells potentiated, normalized current: 136%, n = 9), as did the S815A/S837A mutation (57% of cells potentiated, normalized current: 176%, n = 7), suggesting that neither of these combinations of serine residues is essential to the mechanism of PKA-mediated potentiation of GluR6. In contrast, all of the cells transfected with the S825A/S837A mutant construct failed to show potentiation (normalized current: 55%, n = 6), suggesting that the combination of both S825 and S837 was required for GluR6 potentiation by PKA. Consistent with these findings, both the S825A and S837A mutations significantly decreased the phosphorylation of a GST-C terminal peptide construct (Figures 2B and C).

Figure 5.

Selected current traces from a GluR6 triple serine to alanine mutation and double serine to alanine mutations (on the GluR6Δ855 background). In all cases, PKA was included in the pipette and measurements were taken 2 min and 6 min after formation of the whole cell patch.

It is of interest to note that transfected HEK293 cells carrying multiple serine mutations on a GluR6Δ855 background were more likely to show either no response to glutamate application or unstable responses upon prolonged recording. Although the reason for this remains unclear, residues within the C-terminal domain of glutamate receptors have been identified as important trafficking motifs and as protein-protein contact domains with anchoring proteins such as GRIP and PSD-95 (Nishimune et al., 1998, Song et al., 1998, Hayashi et al., 2000, Standley et al., 2000, Scott et al., 2001, Shi et al., 2001, Malinow and Malenka, 2002, Hirbec et al., 2003, Yan et al., 2004, Salinas et al., 2006). It is possible that multiple serine-to-alanine and/or GluR6Δ855 mutations in the C-terminal domain of GluR6 disrupt important interactions with proteins that serve to localize and/or anchor the receptor in the plasma membrane, thereby destabilizing their normal function.

To determine if channel function was altered by the mutations alone, the rise time and desensitization rate were measured for the WT, Triple, and GluR6Δ855 constructs. These constructs were chosen to represent the extremes of serine (Triple) and threonine (GluR6Δ855) mutagenesis. Although rise times for GluR6 are difficult to measure accurately by the cell flow method (Li et al., 2003), no significant difference in apparent rise time or the desensitization rate was seen among these three constructs in the absence of PKA (data not shown). Thus, channel properties seemed to be similar. Consequently, the results obtained with PKA are likely to be a result of receptor phosphorylation.

Discussion

Phosphorylation is an important mechanism of iGluR modulation (Raymond et al., 1993, Wang et al., 1993, Kohr and Seeburg, 1996, Roche et al., 1996, Lee et al., 1998, Zheng et al., 1998, Banke et al., 2000). A variety of phosphokinases have been shown to phosphorylate iGluRs, and PKA-mediated phosphorylation of iGluRs has been demonstrated (Raymond et al., 1993, Wang et al., 1993, Roche et al., 1996). Although controversy regarding the transmembrane topology of iGluRs complicated the interpretation of early studies designed to determine specific sites of phosphorylation by PKA (Raymond et al., 1993, Wang et al., 1993), more recent work suggests that sites of physiological PKA-mediated phosphorylation of iGluRs are intracellular. Our study using [γ-³²P]ATP kinase assays shows that the C-terminal domain of GluR6 can be phosphorylated by PKA, and that serine residues are the sites of phosphorylation. This is consistent with previous work in which residues within the C-terminal domain of iGluRs have been identified as likely sites of phosphorylation. The S845A mutation in the C-terminal domain of GluR1 abolished phosphorylation by PKA (Roche et al., 1996), and forskolin treatment increased phosphorylation of S845 in rat hippocampal slices (Mammen et al., 1997). Site directed mutagenesis and antiphosphopeptide antibodies identified the C-terminal domain residue S831 as a site of phosphorylation by CaM kinase II in GluR1 (Roche et al., 1996, Barria et al., 1997, Mammen et al., 1997). This site is also a PKC target site and, along with the PKA phosphorylation site S845, has been shown to be basally phosphorylated in GluR1 (Mammen et al., 1997). It is interesting to consider that, like S831 in GluR1, which is not a consensus site for CaM kinase II or PKC, the serine residues in the C-terminal domain of GluR6 are not consensus sites for PKA. Both S845 and S831 of GluR1 have been implicated in long-term potentiation (LTP) and long-term depression (LTD). CaM kinase II phosphorylation of S831 has been reported to be required for LTP (Derkach et al., 1999, Lee et al., 2000, Whitlock et al., 2006) and dephosphorylation of S845 has been implicated in LTD (Kameyama et al., 1998, Lee et al., 2000).

PKA has been shown to phosphorylate recombinant GluR6 expressed in HEK293 cells (Raymond et al., 1993). In the same study, PKA potentiated whole cell currents in GluR6, and mutation of S684 and S666 abolished potentiation. These residues were thought to be intracellularly located based upon the accepted transmembrane topology of glutamate receptors at that time (Hollmann et al., 1989, Keinänen et al., 1990). Subsequent studies, however, have shown that S684 and S666 are in fact extracellularly located (Hollmann et al., 1994, Wo and Oswald, 1994, Wo and Oswald, 1995a; Figure 1) and not translocated to the cytoplasm (Basiry et al., 1999) and, thus, should not be accessible to intracellularly-applied PKA. Whether the functional effects of mutation of S864 and S666 are mediated by an allosteric mechanism remains unclear. One clue may come from accessibility studies which suggest that S684 in GluR6 (Basiry et al., 1999) and the corresponding residue in a goldfish kainate binding protein (S271; Wo et al., 1999) are exposed to solvent in the apo and antagonist-bound form but not in the agonist-bound form. This may suggest that this portion of the protein may undergo small changes upon activation and mutation of these results could possibly have inhibited potentiation arising from the phosphorylation of cytoplasmic residues.

The potentiation of AMPA and kainate receptor current amplitude has been suggested to arise from an increased open channel probability (Knapp et al., 1990, Banke et al., 2000). Using nonstationary variance analysis, Banke et al. (2000) identified serine 845 in the C-terminal domain of GluR1 as an important component of the mechanism of potentiation by PKA in that phosphorylation by PKA increases the open probability of recombinant WT GluR1, but not that of the S845A mutant construct (Banke et al., 2000). The open probability of GluR6, measured by nonstationary variance analysis, is similarly increased by PKA, although specific residues involved in the mechanism of this increase have not been identified (Traynelis and Wahl, 1997). The open probability for GluR6 in the absence of PKA has been measured using laser-pulse photolysis to be approximately 0.96 (Li et al., 2003). Although the open probability of GluR6 in the absence of PKA would suggest that the potentiation due to an increase in open probability is unlikely, some potentiation would be possible (0.96 to 1.0) and changes in single channel conductance (or preferential population of higher conductance states; Derkach et al., 1999) cannot be ruled out.

In our studies, truncation of the entire C-terminus of GluR6 (GluR6ΔC-term) abolished the PKA-mediated potentiation of whole cell currents seen in the WT construct (Figure 3). This result supports the existence of sites within the C-terminal domain that are necessary for potentiation of GluR6 by PKA. The GluR6Δ855 mutation, which removes all three of the threonine residues (T856, T864, and T875) in the C-terminal domain, failed to abolish the potentiation of whole cell currents or phosphorylation of the C-terminal peptide, suggesting that these residues are not phosphorylated by PKA. The finding that the S825A/S837A mutation abolishes potentiation of the glutamate-induced GluR6 current by PKA and the S825A and S837A mutations decrease phosphorylation of the C-terminal peptide, suggests that S825 and S837 are involved in PKA-mediated potentiation of GluR6. The possible requirement for two serine residues in potentiation of whole cell current amplitude is unknown. It is possible that this requirement along with the fact that these sites are not strictly PKA consensus sequences may represent a physiological checkpoint in that a stronger stimulus for PKA recruitment may be required to achieve the degree of phosphorylation necessary for a functional effect (i.e., potentiation of current).

Our findings provide insight into the mechanism of potentiation of GluR6 by PKA. As the potentiation of GluR6 by PKA may play a role in a number of developmentally and clinically significant processes, it is hoped that an elucidation of this mechanism will enhance our understanding of the process of learning and memory and improve our ability to diagnose and treat a number of important neurological diseases. From a broader perspective, our findings may provide information regarding the molecular mechanism by which phosphorylation modulates the function of a variety of biologically important proteins.

Abbreviations

AMPA

α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid

ATD

amino terminal domain

CaM kinase II

Ca⁺⁺/calmodulin-dependent protein kinase II

DMEM

Dulbecco’s modified Eagle medium

FBS

fetal bovine serum

GFP

green fluorescent protein

GluR6

kainate receptor subtype 6

GluR6ΔC-term

GluR6 truncated before S815

GluR6Δ855

GluR6 truncated before T856

GST

Glutathione-S-transferase

HEK

human embryonic kidney

iGluR

ionotropic glutamate receptor

LTD

long-term depression

LTP

long-term potentiation

NMDA

N-methyl-D-aspartate

S1S2

extracellular ligand-binding domain of a GluR

PKA

protein kinase A

PKC

protein kinase C

WT

wildtype