The functionally crucial arginine within
The functionally crucial arginine within M2 is not encoded by the gluR2 gene, but rather arises within the pre-mRNA by editing of a codon for the neutral glutamine residue25, 26. RNA editing at the Q/R site is specific to GluR2 and is extremely efficient. In neonatal and adult rat brain, almost 100% of gluR2 mRNA undergoes editing; in embryonic brain, only a small percentage of GluR2 subunits (about 1%) are unedited (Fig. 1). More recently, editing at another site was identified in the extracellular TM3–TM4 loop of Glu-receptor subunits, which determines a switch from the encoded arginine (R) to glycine (G). Editing at the R/G site is specific for GluR2, 3 and 4, and is about 80–90% complete in adult rat MPI-0479605 (Fig. 1). Immediately adjacent to the R/G site, one of two cassettes named flip and flop, each containing 38 amino acids, is introduced in GluR subunits by alternative RNA splicing (Fig. 1). RNA editing at the R/G site and splicing at the flip/flop site are developmentally regulated and are cooperative in controlling desensitization and recovery rates of AMPA receptor responses. The allosteric modulator cyclothiazide strongly attenuates desensitization in flip but not in flop splice variants of recombinant AMPA receptors. Consistent with their extracellular position, neither the R/G site nor the flip/flop cassettes affect Ca2+ permeability of AMPA receptor channels. The Ca2+ permeability of native AMPA receptors varies widely (for review, see 30, 31). Since editing of gluR2 mRNA at the Q/R site is virtually complete under physiological conditions, Ca2+-permeable AMPA receptors arise in neurons only as a consequence of reduced expression of gluR2 mRNA. Studies involving patch-clamp recording and RT-PCR (reverse transcriptase – polymerase chain reaction) demonstrate that AMPA receptor permeability to Ca2+ varies inversely with abundance of gluR2 mRNA in a wide range of cell types. Excitatory principal neurons such as hippocampal17, 18 and neocortical pyramidal cells and dentate-gyrus granule cells exhibit low Ca2+ permeability and more abundant gluR2 mRNA. Hippocampal17, 32 and neocortical GABAergic interneurons and dentate gyrus basket cells display higher Ca2+ permeability and less abundant gluR2 mRNA. Bergmann glia cells of the cerebellum exhibit high Ca2+ permeability and no detectable GluR2. Hippocampal GABAergic interneurons lacking GluR2 are viable and relatively resistant to post-ischemic delayed neurodegeneration. Nitric oxide synthase-positive neurons of cortex, striatum and hippocampus, sharing the common feature of low GluR2 expression, are also relatively resistant in neurodegenerative diseases. In addition, transgenic mice with targeted disruption of the gluR2 gene survive and their principal neurons (that express abundant levels of GluR2 in wild-type animals) are functional. Possible explanations for the survival of neurons that normally lack GluR2 may include compensatory mechanisms for Ca2+ buffering and extrusion (by, for example, expression of Ca2+-binding proteins) or reduced AMPA currents due to expression of receptors with faster and more profound desensitization18, 37. Hence, the GluR2 hypothesis will apply primarily to neurons that normally express Ca2+-impermeable AMPA receptors and that are not required to cope with Ca2+ influx via AMPA receptors. In these cells, acute increases in permeability of AMPA receptors to Ca2+ could account for cell death. High rate of Ca2+ influx through AMPA receptors in cultured rat cerebellar Purkinje cells, in a subpopulation of neocortical neurons39, 40, 41, and in an immortalized rat oligodendroglial cell line leads to enhanced vulnerability to agonist-induced cytotoxicity. Heterozygous transgenic mice engineered for a Q/R editing-deficient gluR2 allele express AMPA receptors with increased Ca2+ permeability, particularly in hippocampal and neocortical principal neurons. The mice develop recurrent seizures and die within the first three weeks of life, with cell loss in the hippocampus. In these animals, unedited GluR2 may contribute to the formation of a greater number of Ca2+-permeable AMPA receptors than in the gluR2 knockout mice. There are other examples where the phenotypes of different knockout mutants of similar design are unexpectedly different, ranging from complete viability to lethality.