Evidence for multiple, distinct ADAR-containing complexes in Xenopus laevis

Abstract

ADAR (adenosine deaminase acting on RNA) is an RNA-editing enzyme present in most metazoans that converts adenosines in double-stranded RNA targets into inosines. Although the RNA targets of ADAR-mediated editing have been extensively cataloged, our understanding of the cellular function of such editing remains incomplete. We report that long, doublestranded RNA added to Xenopus laevis egg extract is incorporated into an ADAR-containing complex whose protein components resemble those of stress granules. This complex localizes to microtubules, as assayed by accumulation on meiotic spindles. We observe that the length of a double-stranded RNA influences its incorporation into the microtubule-localized complex. ADAR forms a similar complex with endogenous RNA, but the endogenous complex fails to localize to microtubules. In addition, we characterize the endogenous, ADAR-associated RNAs and discover that they are enriched for transcripts encoding transcriptional regulators, zinc-finger proteins, and components of the secretory pathway. Interestingly, association with ADAR correlates with previously reported translational repression in early embryonic development. This work demonstrates that ADAR is a component of two, distinct ribonucleoprotein complexes that contain different types of RNAs and exhibit diverse cellular localization patterns. Our findings offer new insight into the potential cellular functions of ADAR.

The RNA-binding complex ESCRT-II in Xenopus laevis eggs recognizes purine-rich sequences through its subunit, Vps25

“…To determine whether ESCRT-II recognizes specific sequence motifs, we performed CLIP-Seq using the UV-crosslinked material that immunoprecipitated with Vps25 under denaturing conditions. Using the Illumina MiSeq platform, we obtained 11,167 unique, non-rRNA reads, which we aligned to a custom transcriptome derived from the Xenopus laevis 7.0 genome and our own sequencing data (35). The unique reads aligned to 1640 transcripts (Table S1 and S3), of which 196 transcripts had more than one aligning read.…”

“…The remaining reads were aligned to X. laevis rRNA genes using Bowtie allowing for up to three mismatches, and the aligning reads were removed. The remaining reads were aligned to sequences derived from a custom gene model (35) allowing up to three mismatches and retaining the best strata of alignments using the parameters -m 3 -best -strata. Reads aligning to more than two transcripts or to antisense strands were removed.…”