ADAR2 A-->I editing: site selectivity and editing efficiency are separate events

Abstract

ADAR enzymes, adenosine deaminases that act on RNA, form a family of RNA editing enzymes that convert adenosine to inosine within RNA that is completely or largely double-stranded. Site-selective A-->I editing has been detected at specific sites within a few structured pre-mRNAs of metazoans. We have analyzed the editing selectivity of ADAR enzymes and have chosen to study the naturally edited R/G site in the pre-mRNA of the glutamate receptor subunit B (GluR-B). A comparison of editing by ADAR1 and ADAR2 revealed differences in the specificity of editing. Our results show that ADAR2 selectively edits the R/G site, while ADAR1 edits more promiscuously at several other adenosines in the double-stranded stem. To further understand the mechanism of selective ADAR2 editing we have investigated the importance of internal loops in the RNA substrate. We have found that the immediate structure surrounding the editing site is important. A purine opposite to the editing site has a negative effect on both selectivity and efficiency of editing. More distant internal loops in the substrate were found to have minor effects on site selectivity, while efficiency of editing was found to be influenced. Finally, changes in the RNA structure that affected editing did not alter the binding abilities of ADAR2. Overall these findings suggest that binding and catalysis are independent events.

Figure 1

Schematic RNA secondary structure of the GluR-B stem–loop containing the R/G site. The R/G site is marked in bold. The arrows denote the mutations, nucleotide and position downstream of the R/G site, made to investigate the importance of internal loops for selective R/G editing.

Figure 2

Comparison of R/G site-selective editing by recombinant rat ADAR1 and ADAR2. RNA substrates were sequenced after in vitro editing, RT–PCR and cloning. (A) Distribution of edited sites in the clones. I represents an edited adenosine. The number of clones with the same editing pattern is indicated to the right. (B) Black bars represent the percentage of edited clones displaying selective editing at the R/G site only. Gray bars represent the percentage of edited clones with R/G site editing and additional edited sites. Light gray bars represent the percentage of edited clones with no R/G site editing but with editing at other sites.

Figure 2

Comparison of R/G site-selective editing by recombinant rat ADAR1 and ADAR2. RNA substrates were sequenced after in vitro editing, RT–PCR and cloning. (A) Distribution of edited sites in the clones. I represents an edited adenosine. The number of clones with the same editing pattern is indicated to the right. (B) Black bars represent the percentage of edited clones displaying selective editing at the R/G site only. Gray bars represent the percentage of edited clones with R/G site editing and additional edited sites. Light gray bars represent the percentage of edited clones with no R/G site editing but with editing at other sites.

Figure 3

The importance of internal loops for selective R/G editing. ADAR2 editing of wild-type RNA is compared to editing of four mutant RNA. (A) Distribution of edited sites in the clones. I represents an edited adenosine. The number of clones with the same editing pattern is indicated to the right. (B) Black bars represent the percentage of edited clones displaying selective editing at the R/G site only. Gray bars represent the percentage of edited clones with R/G editing and additional edited sites. Light gray bars represent the percentage of edited clones with no R/G site editing but with editing at other sites.

Figure 3

The importance of internal loops for selective R/G editing. ADAR2 editing of wild-type RNA is compared to editing of four mutant RNA. (A) Distribution of edited sites in the clones. I represents an edited adenosine. The number of clones with the same editing pattern is indicated to the right. (B) Black bars represent the percentage of edited clones displaying selective editing at the R/G site only. Gray bars represent the percentage of edited clones with R/G editing and additional edited sites. Light gray bars represent the percentage of edited clones with no R/G site editing but with editing at other sites.

Figure 4

(A) Fluorescence emission spectra of RNA samples. RNA oligonucleotides, 27 nt long, were used to make R/G substrate analogs with a 2-AP replacing the R/G site adenosine. Fluorescence intensity is plotted as a function of scanned emission wavelength (λ_ex 310 nm) for 0.8 µM single-stranded 2-AP, 0.8 µM duplex 2-AP wild-type (2-AP WT) and 0.8 µM duplex G56 mutant (2-AP G56). The spectra are the average of nine scans. The buffer spectrum has been subtracted in all spectra. (B) Gel mobility shift assays showing RNA duplex formation and ADAR2 binding. Lane 1, single-stranded 2-AP; lane 2, duplex of 2-AP WT; lane 6, duplex of 2-AP G56; lanes 2–5, increasing amounts of ADAR2 were added to 20 fmol of wild-type 2-AP RNA duplex; lanes 7–9, ADAR2 added to 2-AP G56 duplex, as described for lanes 2–5. Gel retardation was analyzed on 8% native polyacrylamide gels.

Figure 4

(A) Fluorescence emission spectra of RNA samples. RNA oligonucleotides, 27 nt long, were used to make R/G substrate analogs with a 2-AP replacing the R/G site adenosine. Fluorescence intensity is plotted as a function of scanned emission wavelength (λ_ex 310 nm) for 0.8 µM single-stranded 2-AP, 0.8 µM duplex 2-AP wild-type (2-AP WT) and 0.8 µM duplex G56 mutant (2-AP G56). The spectra are the average of nine scans. The buffer spectrum has been subtracted in all spectra. (B) Gel mobility shift assays showing RNA duplex formation and ADAR2 binding. Lane 1, single-stranded 2-AP; lane 2, duplex of 2-AP WT; lane 6, duplex of 2-AP G56; lanes 2–5, increasing amounts of ADAR2 were added to 20 fmol of wild-type 2-AP RNA duplex; lanes 7–9, ADAR2 added to 2-AP G56 duplex, as described for lanes 2–5. Gel retardation was analyzed on 8% native polyacrylamide gels.