ABSTRACT: ADARs, short for adenosine deaminases acting on RNA, were first found by their characteristic activity of deaminating up to 50% of all adenosines in double-stranded RNA in vitro. C. elegans has two ADARs, only one is found in the fly, and mammals express three ADARs. Importantly, the spectrum of physiological roles of these enzymes in the different organisms remains largely unknown. ADARs are ~750-1150 amino acids in size, feature 2-3 N-terminal double-stranded RNA binding motifs, an enzymatic domain related to that of prokaryotic cytidine deaminases, and a ~190 residue extended C-terminal domain of unknown function. The mammalian ADARs are detectable in many if not all tissues, with relatively prominent expression in the central nervous system, predicting tissue- or cell type-specific tasks for the mammalian ADARs. Studies over the last decade collectively indicated that one function of ADARs is the editing of nuclear transcripts to introduce select amino acid substitutions in proteins. Arguably the best-studied example for this type of RNA editing (‘A-to-I editing’) in the mammal is performed by ADAR2 and occurs in primary transcripts for the GluA2 subunit (also known as GluR-B or GluR2 subunit) of AMPA receptors mediating fast excitatory neurotransmission. A particular glutamine codon (CAG) in the exonic sequence for the inner lining of the glutamate-activated AMPA receptor is converted by deamination of the central adenosine into an unusual arginine codon (CIG; the resultant inosine is seen as guanosine by the protein translation machinery). This substitution of the gene-encoded glutamine codon by the edited arginine codon is found in nearly 100% of all GluA2 cDNA derived from brain mRNA. RNA editing in this instance ensures that GluA2-containing AMPA receptors become impermeable to Ca2+. In addition to switching a key functional determinant in GluA2 to nearly 100%, ADAR2 has been implicated in generating numerous amino acid substitutions at lower levels in various mammalian neurotransmitter receptors and voltage-gated ion channels, primarily but not exclusively of nervous tissue. Most of these substitutions elicit distinct functional consequences, which led to the notion that site-selective RNA editing by ADAR2 may ensure the functional fine-tuning of the affected transmembrane proteins. Such fine-tuning should reflect the expansion of the protein sequence directed by the respective gene into sequence subpopulations, each differing in particular functional aspects, such as transmitter efficacy, desensitization, or voltage threshold. If this notion is correct, ADAR2 knockout should elicit phenotypes, which might range from subtle to severe, depending on the physiological role(s) of the ADAR2 targets in different tissues. ADAR2 knockout mice, however, die soon after birth from severe epileptic seizures, a consequence of the changed ion conductance properties of AMPA receptors containing unedited GluA2. Fortunately, mice lacking ADAR2 are robust and have normal life span if they carry GluA2 alleles genetically altered to contain the particular arginine codon for a Ca2+ impermeable ion channel pore of AMPA receptors, thus rescinding the need for ADAR2-mediated GluA2 transcript editing. These mice are thus eminently suitable to study the phenotypic consequences of failure to edit all ADAR2 targets excepting GluA2. We therefore tested cohorts of ADAR2-/-/GluA2(R)/(R) mice in 16 well-established paradigms for central and peripheral phenotypic abnormalities. We report that we found few differences to control mice (GluA2(R)/(R)) and hence, the lack of ADAR2 appears not to affect most of the physiological functions tested.
