Cyclic AMP-dependent protein kinase phosphorylates group III metabotropic glutamate receptors and inhibits their function as presynaptic receptors

Abstract

Recent evidence suggests that the functions of presynaptic metabotropic glutamate receptors (mGluRs) are tightly regulated by protein kinases. We previously reported that cAMP-dependent protein kinase (PKA) directly phosphorylates mGluR2 at a single serine residue (Ser843) on the C-terminal tail region of the receptor, and that phosphorylation of this site inhibits coupling of mGluR2 to GTP-binding proteins. This may be the mechanism by which the adenylyl cyclase activator forskolin inhibits presynaptic mGluR2 function at the medial perforant path-dentate gyrus synapse. We now report that PKA also directly phosphorylates several group III mGluRs (mGluR4a, mGluR7a, and mGluR8a), as well as mGluR3 at single conserved serine residues on their C-terminal tails. Furthermore, activation of PKA by forskolin inhibits group III mGluR-mediated responses at glutamatergic synapses in the hippocampus. Interestingly, beta-adrenergic receptor activation was found to mimic the inhibitory effect of forskolin on both group II and III mGluRs. These data suggest that a common PKA-dependent mechanism may be involved in regulating the function of multiple presynaptic group II and group III mGluRs. Such regulation is not limited to the pharmacological activation of adenylyl cyclase but can also be elicited by the stimulation of endogenous G(s)-coupled receptors, such as beta-adrenergic receptors.

Fig. 1

PKA phosphorylates mGluR7 in the hippocampus and mGluR4 in the cerebellum. Autoradiographs showing PKA-induced phosphorylation of mGluR7 immunoprecipitated from total hippocampal proteins (a), and phosphorylation of mGluR4 immunoprecipitated from total cerebellum proteins (b). The molecular weights of the predominant phosphorylated band in (a) and the upper most band in (b) are consistent with that of mGluR7 and mGluR4, respectively, as determined by western blot analysis. The multiple bands in (b) may represent degradation products of mGluR4. In both cases, phosphorylation of the receptors was inhibited by the presence of a selective inhibitor of PKA (1 μM PKI). 10 U of purified PKA was used per 100 μg protein in the presence of [γ-³²P]ATP. Each graph is representative of three independent experiments. Numbers beside arrows pointing to the graphs refer to the molecular weight markers.

Fig. 2

C-terminal tails of group III mGluRs are PKA substrates. The autoradiograph (representative of three independent experiments) shows that PKA phosphorylated all group II and III mGluRs tested (mGluR6 not tested). Protein (GST alone) expressed from the pGEX vector yielded no significant phosphorylation. One microgram of protein was used for each lane. Numbers besides arrows pointing to the graph refer to the molecular weight markers.

Fig. 3

MALDI-MS confirmed m7CTX and m4CTX were phosphorylated by PKA with a ~80 m/z mass shift. (a) m7CTX construct (GPLGSPNSHP ELNVQKRKRS FKAVV-TAATM SSRLSHKPSD RPNGEAKTEL CENVDPNSPA AKKKYVSYNN LVI) was HPLC purified and analyzed by MALDI-TOF MS using α-cyano-4-hydroxycinnamic acid as the matrix (calculated m/z = 7962.0 before phosphorylation, left panel, and m/z = 8042.0 after incorporation of a single phosphate group, right panel). (b) m4CTX construct (GPLGSPNSEQ NVPKRKRSLK AVVTAATMSN KFTQKGNFRP NGEAK-SELCE NLETPALATK QTYVTYTNHA I) was analyzed as described for m7CTX (calculated m/z = 7735.7 before phosphorylation, left panel, and m/z = 7815.7 after incorporation of a single phosphate group, right panel).

Fig. 4

Identification of phosphorylation sites of m7CTX and m4CTX. (a) and (c) Total ion current trace of precursor ion scanning LC-MS experiment for m7CTX and m4CTX tryptic peptides. The mass spectrums of phosphopeptides present in the tryptic digests are shown in the inserts. The only phosphopeptides S(PO₃H)FK from m7CTX corresponds to residues 20–22 (SFK) in m7CTX and Ser862 in mGluR7 cDNA sequence, and that of m4CTX, S(PO₃H)LK, corresponds to residues 18–20 in m4CTX and Ser859 in mGluR4 cDNA sequence. (b) and (d) The UV traces of the same runs. T5-PO₃H stands for the phosphorylated peptide (residues 20–22 (S(PO₃H)FK) in m7CTX or residues 18–20 (S(PO₃H)LK) in m4CTX) which elutes very early as the back shoulder of the solvent peak.

Fig. 5

Identification of PKA phosphorylation sites with mutation analysis. In vitro phosphorylation of mGluR C-terminus fusion protein with site-directed mutagenesis showing a single serine residue as the phosphorylation site for each mGluR tested, including Ser845 for mGluR3 (a), Ser859 for mGluR4 (b), Ser862 for mGluR7 (c), and S355 for mGluR8 (d). Each autoradiograph is representative of 3 independent experiments, showing PKA-induced phosphorylation of GST fusion proteins.

Fig. 6

Effect of PKA activation on mGluR7 function at the SC-CA1 synapses in the rat hippocampus. (a) 1 mM L-AP4 suppressed field excitatory postsynaptic potentials (fEPSP) dramatically; this mGluR7-mediated effect was inhibited by coapplication of 50 μM forskolin. Calibration: 0.2 mV, 10 ms. (b) Bar graph showing the mean effect of L-AP4 (L) with or without forskolin on fEPSP. The forskolin (F) effect was blocked by a selective inhibitor of PKA, Rp-cAMP (R). n = 4 for each column; error bars show the standard errors of the means; Student’s t-test; *different from L-AP4 alone, p < 0.05.

Fig. 7

Effect of PKA activation on mGluR8 function at the LPP-DG synapses. (a) and (b) fEPSP recordings at the lateral perforant (LPP-DG) synapse from mGluR8-deficient (KO) and wild-type (WT) mice. L-AP4 (20 μM) induced suppression of fEPSP at these synapses in wild-type mice, but had no effect in mGluR8-deficient mice (a). Dose–response curves of L-AP4-induced suppression of fEPSP in both wild-type and mGluR8-deficient mice are shown in (b) (n = 3–4 for each point). (c) and (d) 50 μM forskolin (F) inhibited 20 μM L-AP4 (L)-induced suppression of fEPSP at the LPP-DG synapses in the rat hippocampus, as shown by sample traces of fEPSP recording (c) and a bar graph (d) showing the mean effect across 3 slices. The forskolin effect was blocked by coapplication with 100 μM Rp-cAMPs (R). Student’s t-test; *different from L-AP4 alone, p < 0.05.

Fig. 8

Effect of isoproterenol on presynaptic mGluRs in the rat hippocampus. (a) and (b) 1 μM ISO (I) inhibited 1 μM DCG-IV (D)-induced depression of fEPSP at MPP-DG synapses, as shown by sample traces of recording (a) and a bar graph expressing the mean effect across four slices (b). In the presence of 100 μM Rp-cAMPs (R), Iso failed to reduce DCG-IV-induced depression of fEPSP (a) and (b). *Different from DCG-IV alone; Student’s t-test; p < 0.05. (c) 1 μM ISO (I) inhibited 1 mM L-AP4 (L)-induced suppression of fEPSPs at SC-CA1 synapses. Student’s t-test; *different from L-AP4 alone, p < 0.05. (d) At LPP-DG synapses, 1 or 10 μM ISO did not affect the ability of 20 μM L-AP4 (L) to suppress fEPSP at these synapses. n = 3 for each column.

Fig. 9

Sequence alignments of a C-terminal region in group II and group III mGluRs. The region shown is directly downstream of the C-terminus of the seventh transmembrane domain. Sequences further downstream in the mGluRs listed here contain no PKA consensus sites, and were not shown. (a) Sequence alignment of the C-terminal regions of tested group II and III mGluRs showing single serine residues as PKA phosphorylation sites (in black box). (b) Sequence alignment of the C-terminal regions of non-tested group III mGluRs showing the PKA consensus sites (in black box). Note that sequences shown for mGluR7b and 8b are the same as for mGluR7a and 8a, respectively, because they use splice sites downstream of this region.