Adenosine deaminases acting on RNA (ADARs): RNA-editing enzymes

Abstract

Adenosine deaminases acting on RNA (ADARs) were discovered as a result of their ability extensively to deaminate adenosines in any long double-stranded RNA, converting them to inosines. Subsequently, ADARs were found to deaminate adenosines site-specifically within the coding sequences of transcripts encoding ion-channel subunits, increasing the diversity of these proteins in the central nervous system. ADARI is now known to be involved in defending the genome against viruses, and it may affect RNA interference. ADARs are found in animals but are not known in other organisms. It appears that ADARs evolved from a member of another family, adenosine deaminases acting on tRNAs (ADATs), by steps including fusion of two or more double-stranded-RNA binding domains to a common type of zinc-containing adenosine-deaminase domain.

Figure 1

Primary structures of ADARs and ADATs (adenosine deaminases acting on tRNAs), showing arrangements of dsRNA-binding domains and deaminase domains. The deaminase motifs (I-III) consist of four or five highly conserved amino acids surrounding each of the cytosines and histidines that bind zinc. Proteins have been drawn to scale and are aligned at the catalytic glutamate residue (E) in the active site in adenosine deaminase motif I. The deaminase domains of ADAR and ADAT1 proteins are larger than the core deaminase domain (which we define as the region of homology between Tad2/ADAT2 and Tad3/ADAT3) because of the insertion of sequence between the second and third zinc-chelating residues, separating deaminase motifs II and III, and probably also because of fusion with a carboxy-terminal subdomain of unknown origin. Abbreviation: aa, amino acids.

Figure 2

A phylogenetic tree of core deaminase sequences. Protein sequences that are predicted from genomic sequences rather than from full cDNA sequences are indicated by asterisks. The vertebrate genes listed include one from the crab-eating monkey Macacus fascicularis. The fish genomes examined are from the zebrafish Danio rerio and pufferfish, Takifugu rubripes and Tetraodon fluviatilis. Two tunicate genomes have been examined, from Ciona intestinalis and Ciona savigni. One squid (Loligo pealeii) is included (J. Rosenthal, personal communication). The insect genomes are of the two Drosophila species, Drosophila melanogaster and Drosophila pseudoobscura, and the malaria mosquito Anopheles gambiae. Alignments were made using T-COFFEE, the boundaries of the core deaminase domain were specified and a tree based on the alignment was generated using MEGA. Bootstrap values are given on all branches.

Figure 3

An alignment of core deaminase domains from vertebrate ADAR1 and ADAR2 proteins, showing residues characteristic for each. Asterisks indicate cysteines and histidines that chelate zinc, as in Figure 1, and the three active-site motifs are bracketed.