RNA editing

RNA editing refers to targeted sequence alteration of the RNA transcript, resulting in sequence that is different from the underlying DNA template. Originally discovered in trypanosomes as an insertion event of four uridine residues to restore a reading-frame shift in the DNA (Benne et al. 1986), the term has since come to

encompass a diverse set of sequence revisions of the RNA transcript in a wide range of organisms (Gott & Emeson 2000; Knoop 2011). In lower organisms, various types of editing mechanisms have been uncovered since the initial discovery of insertional editing in trypanosomes. For example, paramyxoviruses control the expression of two different isoforms of phosphoprotein by introducing either one or two additional guanosines in the mRNA transcript co-transcriptionally (Cattaneo et al. 1989; S. M. Thomas et al. 1988; Vidal et al. 1990a). The insertion event is always found to occur within a short stretch of guanosines, and thus a stuttering mechanism has been proposed in which the viral polymerase repeatedly copies a cytosine in the template (Vidal et al. 1990b). In

20

2002; Mahendran et al. 1991). Other examples of RNA editing in lower organisms include pyrimidine and purine transitions, transversions of guanosine to cytidine, and conversions of uridine to purines among other sequence revisions in dinoflagellates (Dorrell & Howe 2012; Jackson et al. 2007; S. Lin et al. 2002; Zauner et al. 2004) and U- to-C conversions in placazoa (Burger et al. 2009).

In plants, RNA editing occurs in the form of pyrimidine exchanges in mitochondrial and chloroplast transcripts. Since the initial discovery of C-to-U

substitutions to reconstitute evolutionarily conserved amino acids in wheat and evening primrose mitochondria (Covello & Gray 1989; Gualberto et al. 1989; Hiesel et al. 1989), other instances of C-to-U editing have been observed in the mitochondria of many different land plants (Fang et al. 2012; Hiesel et al. 1994; Malek et al. 1996; Sper-Whitis et al. 1996; Sper-Whitis et al. 1994; W. Wang et al. 2012). Less common than the

commonly observed C-to-U editing, U-to-C exchanges have also been observed in the mitochondria hornworts (Steinhauser et al. 1999; Yoshinaga et al. 1996), lycophytes (Grewe et al. 2009), and some seed plants (Gualberto et al. 1990; W. Schuster et al. 1990). Shortly after the discovery of mitochondrial RNA editing, C-to-U editing of chloroplasts was reported in maize (Hoch et al. 1991) and subsequently in all land plants with the exception of the marchantiid liverworts (Freyer et al. 1997; Inada et al. 2004; Kugita et al. 2003; Tillich et al. 2005; Tsudzuki et al. 2001; Wolf et al. 2004). Studies have demonstrated that chloroplast editing in plants is essential for protein function (Bock et al. 1994) and that specificity is conferred by local sequences (Bock et al. 1996; Chaudhuri et al. 1995; Sutton et al. 1995). Recent efforts to find the mechanism

repeat (PPR) protein family as playing a key and necessary role (S. R. Kim et al. 2009; Kotera et al. 2005; Okuda et al. 2010; Robbins et al. 2009; Yu et al. 2009; Zehrmann et al. 2009; Zhou et al. 2009).

In metazoa, two main types of editing are known to exist: A-to-I changes result from deamination of adenosine to inosine by ADAR, a family of adenosine deaminases that act on RNA (Nishikura 2010; Orlandi et al. 2012; Wulff & Nishikura 2010), and C- to-U differences arise from deamination of cytidine to uridine by ABOBEC1, a member of the AID/APOBEC gene family (Conticello 2012; Keegan et al. 2001; Smith et al. 2012; Wedekind et al. 2003). Inosine is recognized by the translational machinery as guanosine, and thus, A-to-I RNA editing by ADAR is functionally equivalent to A-to-G changes. Discovered initially for its ability to unwind RNA duplexes through adenosine deamination (Bass & Weintraub 1988), ADAR is now recognized for its involvement in post-transcriptional A-to-I editing of various double-stranded RNA substrates, playing a role in proteome diversification (Pullirsch & Jantsch 2010), regulation of gene expression through alternative splicing (Laurencikiene et al. 2006; Rueter et al. 1999; Schoft et al. 2007) and RNA interference (Alon et al. 2012; Borchert et al. 2009; Kawahara et al. 2008; Kawahara et al. 2007; Liang & Landweber 2007; Nishikura 2006; Reid et al. 2008; Yang et al. 2006). Early insights into the function of ADAR came from reports that a glutamine-to-arginine codon change in glutamate receptors controls ion flow in mouse brain (Sommer et al. 1991). Subsequent studies have demonstrated that ADAR likewise directs amino acid changes in neurotransmitters of other organisms (Hoopengardner et al.

22

and psychiatric disorders (Bhogal et al. 2011; Grohmann et al. 2010; Sawada et al. 2009; Silberberg et al. 2012; Tan et al. 2009; Zhu et al. 2012). Furthermore, RNA editing by ADAR is necessary for proper development and normal life as ADAR-deficient invertebrates exhibit behavioral defects (Palladino et al. 2000; Tonkin et al. 2002), ADAR1-knockout mice are embryonic lethal (Q. Wang et al. 2000), and ADAR2- knockout mice are viable but die prematurely (Higuchi et al. 2000).

Within the past decade, different approaches have been developed to identify A- to-I editing sites in a global manner. Prior to the development of these genome-wide screens, only a few genes were known to be targets of ADAR (Bass 2002;

Hoopengardner et al. 2003; Morse & Bass 1999). An initial study aimed at global discovery of RNA editing sites searched in large databases of expressed sequence tags and found more than 12,000 sites in 1,600 genes, increasing the number of previously known editing sites in humans by more than two orders of magnitude (Levanon et al. 2004). This work revealed that the majority of A-to-I editing sites in humans are located

within noncoding regions of the gene, typically in Alu elements, a class of short

interspersed elements (SINEs) unique to primates that comprises approximately 10% of the human genome (Batzer & Deininger 2002). With the emergence of next-generation sequencing technology, recent studies have used RNA-Seq to identify editing sites in a global manner across different cell types and organisms (Bahn et al. 2012; Chepelev 2012; Ju et al. 2011; J. B. Li et al. 2009; Park et al. 2012; Peng et al. 2012; Ramaswami et al. 2012). These studies confirm the notion that A-to-I RNA editing in mammalian

systems is widespread, with estimates ranging from approximately 1,000 to over 400,000 in humans (Ramaswami et al. 2012). These studies also show that in humans, the vast

majority of editing sites are within noncoding regions (mainly the 3’UTR and introns) and that codon changes in protein sequences are rare (Kleinman et al. 2012). Expanding the search for editing events beyond processed mRNA transcripts, some next-generation sequencing studies have investigated RNA editing in different species of RNA, such as small RNA, lending insight into crosstalk between the ADAR and RNA intereference pathways (Alon et al. 2012; Warf et al. 2012; D. Wu et al. 2011). Other RNA-Seq studies have focused on the temporal aspect of editing by sequencing nascent RNA, discovering

that the majority of A-to-I editing in Drosophila occurs cotranscriptionally (Rodriguez et

al. 2012).

The second type of RNA editing in metazoa is C-U deamination by APOBEC1 (Keegan et al. 2001). Discovered for its role in producing two distinct forms of

apolipoprotein B: apoB100 and the shorter isoform apoB48, which results from the

conversion of a glutamine codon to a stop codon through C-to-U deamination in apoB

mRNA (S. H. Chen et al. 1987; Hospattankar et al. 1987; Powell et al. 1987; Smith et al. 1997), APOBEC1 is a zinc-dependent cytidine deaminase that achieves editing site specificity through local sequence motifs (Backus & Smith 1994; Hersberger & Innerarity 1998; Hersberger et al. 1999; Shah et al. 1991) and the RNA-binding

“APOBEC1 complementing factor” ACF (Mehta et al. 2000). ApoB is a lipoprotein that

plays a critical role in the transport of cholesterol and triglycerides in the plasma (Chan

1992). In humans, C-to-U editing of apoB mRNA occurs only in the small intestine but

24

apoB100 in modulating LDL levels is important in the development of atherosclerosis, as high LDL levels are a major risk factor for coronary heart disease (Ross 1995). In

contrast, apoB48, which is produced in the small intestine and lacks the carboxyl terminus of apoB100, aids in the synthesis and secretion of chylomicrons (Kane et al. 1980). In contrast to widespread A-to-I RNA editing, only a few additional targets of

APOBEC1 have been identified apart from apoB mRNA, such as the editing of glycine

receptors (Legendre et al. 2009; Meier et al. 2005), NAT1 (novel APOBEC target) mRNA

(Yamanaka et al. 1997), and the neurofibromatosis NF1 mRNA (Skuse et al. 1996).

Recent genome-wide studies using RNA-Seq have uncovered 32 more mRNA target sites of APOBEC1, all of which are in AU-rich segments of the 3’ untranslated regions (3’ UTR) of gene transcripts (Rosenberg et al. 2011). The localization of these editing sites to the 3’ UTR of transcripts may alter binding sites for RNA-binding proteins, abolish or create miRNA seed sequences, or affect additional post-transcriptional processes such as polyadenylation, subcellular localization, or translational efficiency, although further research into the biological impact of these editing events remains to be done.