Identification of new targets of Drosophila pre-mRNA adenosine deaminase

Shuli Xia1, Jinghua Yang2, Yingjun Su2, Jiang Qian3, Enbo Ma4 and Gabriel G. Haddad1,5

1 Department of Pediatrics, Section of Respiratory Medicine, Yale University School of Medicine, New Haven, Connecticut
2 Department of Surgery, Yale University School of Medicine, New Haven, Connecticut
3 Department of Molecular Biophysics and Biochemistry, Yale University School of Medicine, New Haven, Connecticut
4 Department of Molecular and Cell Biology, University of California, Berkeley, California
5 Departments of Pediatrics and Neuroscience, Albert Einstein College of Medicine, Bronx, New York


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Adenosine deaminase acting on RNA (ADAR) in Drosophila and mammals has recently become the target of numerous investigations. It is now clear that this protein has a number of functions in the nervous system. Indeed, the mutation of ADAR in Drosophila (dADAR) results in many pathological and physiological changes, such as sensitivity to hypoxia and neuronal degeneration. To understand the full scope of dADAR function, it is crucial to identify new dADAR targets. A polyclonal antibody against inosine was developed and used to enrich inosine-containing mRNAs. The efficiency of immunoaffinity purification was confirmed for the Q/R editing site of GluR-B pre-mRNA that has been edited by ADAR2 to generate inosines at the editing site. This approach was applied to enrich inosine-containing mRNAs from total mRNAs of wild-type and dADAR mutant flies, respectively. The enriched mRNA portion was then amplified and hybridized with Drosophila cDNA arrays. With this method, over 500 mRNAs were identified as potential dADAR targets by showing a higher amount in the enriched mRNA portion from wild-type flies than from dADAR mutant flies. The occurrence of A-to-G conversion in these mRNAs was further analyzed by comparing over 7,000 Drosophila cDNAs sequences with their genomic sequences. A final list of 62 candidates was generated from the overlap of the two approaches. Twelve genes from the final list were further examined by sequencing the RT-PCR products of these genes from wild-type and dADAR mutant flies. Seven of the 12 genes were proven to have A-to-G changes in the wild-type but not in mutant flies. We conclude that the combination of immunoaffinity enrichment of inosine-containing mRNA, DNA microarrays, and sequence comparison could facilitate the discovery of new dADAR substrates, which in turn allows us to better understand the targets of dADAR and the biological function of A-to-I RNA editing in flies.

Drosophila adenosine deaminase acting on RNA; dADAR; A-to-I RNA editing; dADAR target; microarray


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
RNA EDITING TAKES PLACE in the RNA sequence after transcription is completed. It was first reported as a posttranscriptional change in the mitochondrial RNA of trypanosomes (5, 30). Some well-identified examples of editing include modification of adenosine (A) to inosine (I) or cytidine (C) to uridine (U) by deamination (7, 27). The consequence of RNA editing is to alter genomic information and, consequently, protein function. If such base changes occur in the coding region of mRNAs, then it could result in protein alterations and molecular diversity (8, 19).

The majority of the A-to-I modifications have been so far limited to transcripts found in the nervous system (25, 28). Most of these transcripts encode ligand- or voltage-gated ion channels and G protein-coupled receptors, such as Na+ channel (29), Ca2+ channel (31), glutamate receptor (GluR-B) (7, 12, 17), and the serotonin receptor (6). The best-characterized example of A-to-I editing is in the GluR-B gene. After A-to-I modification, the codon for glutamine, CAG, is changed to CIG. Since the translation machinery interprets I as G, this conversion results in coding arginine (CGG) rather than glutamine. This single amino acid change can have a profound influence on function. For example, Ca2+ channel permeability as well as the function of the receptor is much altered (15, 16, 35). Many known A-to-I conversions have been serendipitously found by noticing the A-to-G difference among cDNA products or between cDNA and genomic sequences during the course of cloning. Given the amount of inosine in mRNA isolated from various tissues, e.g., more than 5% adenosine is converted to inosine during inflammation, the possibility exists that many additional RNAs are targets for editing by these editases (25, 37).

A family of adenosine deaminases that act on RNA (ADAR) has been identified (14, 23, 36). Although there are three ADAR isoforms in mammalian tissues, only one Adar gene has been found in Drosophila, and this is mainly expressed in the brain (18). Specific A-to-I RNA editing has been reported for several Drosophila ion channel genes (18, 24). For example, we (18) and others (24) have previously reported that dADAR mutant flies that lack dADAR activity exhibit behavioral deficits including hypoxia sensitivity, temperature-sensitive paralysis, and neurodegeneration. Although some of these behavioral changes of dADAR mutant flies could be explained by the existing targets, we have found that other phenotypic manifestations could not be accounted for by the editase targets that are already known. For example, we have found that dADAR mutant flies are more resistant to oxidative stress than wild type. Therefore to better understand the full scope of dADAR function, it is crucial to identify new targets of dADAR using these dADAR mutant flies. With the combination of microarray and sequence analysis, we have systematically identified a number of new dADAR substrates.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Fly stocks.
Canton-S (C-S) flies were used as the wild type. dADAR mutant flies (ADAR–/–) were generated via X-ray mutagenesis as described elsewhere (10). All of these flies were raised on a standard food medium at temperatures of 23–25°C.

Preparation of antibody against inosine.
All the chemicals were purchased from Sigma, St. Louis, MO, unless otherwise indicated. An inosine-specific antibody was raised in rabbits using inosine-BSA as the antigen and purified using inosine-(CH2)6-Sepharose. Antibody against inosine was eluted with glycine (100 mM, pH 2.5) and tetraethylamide (100 mM, pH 11.5), precipitated individually with 45% saturated (NH4)2SO4, and dialyzed twice against 10x volume of PBS. Protein concentration was determined using Bio-Rad protein assay (Bio-Rad, Hercules, CA). Inosine-BSA was made as follows: inosine (0.2 g) was oxidized in 10 ml of 0.1 M NaIO4 for 40 min and followed by adding 0.2 ml of ethylene glycol (16 M) for 5 min. BSA (0.5 g) was dissolved in 10 ml of 100 mM NaHCO3 (pH 8.5) and mixed with oxidized inosine. After shaking overnight, inosine-BSA was dialyzed against running water for 8 h and 10 vol of PBS overnight. Inosine-(CH2)6-Sepharose was made as follows: Sepharose 4B CNBr (3.3 g) (Pharmacia, Piscataway, NJ) was placed in 600 ml of water (adjusted to pH 2.5 with HCl) for 20 min. After removal of the water by filtering, 50 ml of 0.1 M of hexamethylenediamine in 0.1 M of NaHCO3 (adjust pH to 8.3 with HCl) was added, followed by shaking for 2 h, and then filtered to remove the buffer. Meanwhile, 1 g of inosine was oxidized as described above and mixed with Sepharose intermediate and shaken for 2 h. The resin was filtered, and 50 ml of 1.5% NaBH4 was added with shaking overnight. The resin was then washed with large volume of 10 mM of NaHCO3.

Detection of inosine with inosine antibody.
Different amounts of poly(ACU), poly(AGU), poly(G), and poly(I) or a series dilution of mRNA sample were spotted on a nitrocellulose membrane. The membrane was baked at 80°C for 2 h, and then blocked with 5% nonfat dry milk in PBS plus 5 mM EDTA for 1 h. The following steps were the same as in Western blots, except that 5 mM EDTA was included in all procedures. In brief, the membrane was blocked in 5% nonfat dry milk (Carnation, Nestle Food, Glendale, CA) in PBS plus 5 mM EDTA for 1 h. The membranes were then incubated overnight at 4°C in 5% nonfat dry milk/PBS containing the anti-inosine antibody. Membranes were subsequently rinsed five times for 3, 3, 15, 5, and 5 min in 5% nonfat dry milk/PBS/EDTA, and then incubated for 1 h at room temperature in 5% nonfat dry milk/PBS/EDTA containing a secondary antibody conjugated with horseradish peroxidase at a dilution of 1:2,000 (Zymed, San Francisco, CA). Membranes were again rinsed, utilizing the same protocol as above, and signal detection was achieved with the ECL chemiluminescence system (Amersham, Little Chalfont, UK). The density of each dot was quantified with densitometry, as shown in Figs. 1 and 3.



View larger version (36K):
[in this window]
[in a new window]
 
Fig. 1. Immunoaffinity purification of inosine-containing mRNA. An anti-inosine antibody eluted with glycine (pH 2.5) and TEA (pH 11.5) was used to detect poly(ACU), poly(AGU), poly(G), and poly(I) on nitrocellulose membrane at 1.5 x 10–3, 9 x 10–3, 0.056, 0.33, and 2 mg (lanes 1, 2, 3, 4, and 5, respectively). The membrane was blocked with nonfat milk in PBS as for Western blots, except for 5 mM of EDTA, which was included in all procedures. The specificity of the antibody for inosine over guanosine is ~5.0 ± 1 x 103 based on the pH 2.5 results and slightly greater than 1.0 ± 0.2 x 103 based on the pH 11.5 data.

 
Separation of the edited Q/R site GluR-B pre-mRNA by affinity chromatography.
Ten milligrams of inosine-specific antibody purified by inosine-(CH2)6-Sepharose was immobilized and cross-linked with 2 ml of protein A Sepharose Fast Flow (Pharmacia) by incubation at room temperature for 1 h. The beads were then washed with 10 ml of 0.2 M sodium borate (pH 9.0) twice followed by incubation with 20 mM dimethylpimelimidate for 30 min. The beads were further incubated with 0.2 M ethanolamine (pH 8.0) for additional 2 h and resuspended in PBS. All the above procedures were performed at room temperature. One milliliter of the resin was packed in an empty column (Bio-Rad), equilibrated with the wash buffer (1 mM ATP, 1 mM UTP, 1 mM CTP, and 0.02 mM GTP, 5 mM EDTA in PBS). The Q/R site of GluR-B RNA was transcribed and the editing site was labeled in vitro as described (35). Adenosines at the editing sites of GluR-B RNA were converted into inosine by incubation with baculovirus-expressed ADAR2 protein as described elsewhere (36). The ratio of edited and unedited GluR-B RNA was adjusted to ~30%, determined by RNase P1 digestion and thin-layer chromatography (TLC) assay (35). Approximately 0.1 µg of the mixed GluR-B RNA was loaded on the column and washed with 8 ml of the wash buffer, and flow-through fractions were collected at 0.4 ml/tube. The bound RNA was eluted with the wash buffer plus 1 mM of GTP and collected at 0.4 ml/tube. The flow-through fractions and the eluted RNAs were precipitated with 1x vol of isopropanol by addition of 0.5 µg of unlabeled carrier tRNA in each tube. The pellets were resuspended in 50 µl of RNase P1 solution (0.05 U/µl in water), and digested for 1 h at 37°C. The digestion (3.5 µl) was spotted on 10-cm wide cellulose-polyethyleneimine TLC plate and developed in saturated ammonia sulfate, 20 mM sodium acetate, and 2% isopropanol. The labeled adenosine- or inosine-5'-monophosphate was visualized by exposure of X-ray film, as shown in Fig. 2.



View larger version (56K):
[in this window]
[in a new window]
 
Fig. 2. A: first immunoaffinity purification of inosine-containing GluR-B pre-mRNA. GluR-B RNA labeled at the editing site with 32P-labeled was loaded on anti-inosine column. Inosine or adenosine in the editing site of bound and unbound GluR-B RNA was analyzed by thin-layer chromatography (TLC) assay after digestion with RNase P1. Bound RNA was eluted with PBS plus 1 mM of GTP. B: second immunoaffinity purification. Bound GluR-B RNA was again loaded on the same column. Inosine-containing GluR-B RNA was analyzed on TLC as described above. pA and pI: adenosine- and inosine-5'-monophosphate, respectively.

 
Messenger RNA isolation.
mRNA was isolated from fly heads using FastTrack mRNA isolation kit (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. Briefly, fly heads were homogenized in lysis buffer with RNase degrader. Messenger RNA was bound to oligo-dT cellulose and eluted with elution buffer. The quality of mRNA was checked in formaldehyde RNA gel, and the amount of mRNA was measured with a spectrophotometer (GeneQuant, Pharmacia) at 260 nm absorbance.

Isolation of inosine-containing mRNA.
Inosine-containing mRNA was isolated with the anti-inosine affinity column. For immunoaffinity separation, 100 µg mRNA from wild-type or dADAR mutant flies was heated to 70°C for 3 min, quickly put on ice for 5 min, then loaded onto a 1-ml anti-inosine protein A-Sepharose column. The mRNA was allowed to pass the column three times by gravity force. The column was then washed with 20 ml binding buffer (PBS plus 1 mM ATP, 0.1 mM GTP, 1 mM CTP, 1 mM UTP, 5 mM EDTA). The bound mRNA was eluted with 2 ml binding buffer containing 1 mM ITP. The eluted fraction (2 ml) was collected. RNA was recovered by precipitating with 0.5 µg carrier tRNA and 1x vol isopropanol at –70°C for at least 2 h and centrifuged at 16,000 g for 15 min. The pellet RNA was air dried and reconstituted in DEPC water in a proper volume.

In vitro amplification of mRNA.
In vitro amplifying of mRNA was performed according to the method developed by Baugh et al. (4). The first-strand cDNA was synthesized in 10 µl reverse transcription (RT) reaction containing: 100 U SuperScript II (Invitrogen), 100 ng (dT)-T7 primer [5'-GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGG(T)24], 4 µg T4pg32 (USB), and dNTP (0.5 mM) in 1x first-strand buffer (Invitrogen). The second-strand cDNA was carried out in 75-µl reaction containing 20 U DNA polymerase I, 1 U Escherichia coli RNase H, and 5 U E. coli DNA ligase in 1x second-strand buffer (Invitrogen) and incubating at 15°C for 2 h. The double-stranded cDNA was polished by adding 20 U T4 DNA polymerase and incubating for 15 min at 15°C. The reaction was extracted with 1x vol phenol:chloroform (1:1) and Phase Lock Gel (Eppendorf, Westbury, NY). The cDNA was further purified with BioGel P-6 MicroSpin column (Bio-Rad) per the manufacturer’s instructions. The resulting cDNA was precipitated with 5 µg linear polyacrylamide (GenElute-LPA from Sigma), 1/25th vol of 5 M NaCl, and 2.5 vol of 95% ethanol at –20°C for at least 2 h. The in vitro transcription was performed in 40 µl reaction with 160 U T7 RNA polymerase (Promega, San Luis Obispo, CA), 7.5 mM each NTP, and 1x AmpliScribe buffer (Epicentre, Madison, WI). The reaction was carried out at 37°C for 9 h. At the end, amplified mRNA was quantified by absorbance at 260 nm.

Labeling of mRNA and hybridization with Drosophila cDNA array.
RNA was labeled with Cy3 (C-S) or Cy5 (ADAR–/–) according to the protocol described elsewhere (4). In brief, amplified mRNA (3.5 µg) was used as template to synthesize first-strand cDNA with 400 U SuperScript II, 0.1 mM Cy3-dUTP or Cy5-dUTP (Amersham), and 2.5 mM dATP, dCTP, and dGTP, and 1.0 mM dTTP. The reaction was carried out at 42°C for 2 h and terminated with 0.1 M NaOH, 2 mM EDTA. The labeled cDNA was cleaned up with Qiagen (Valencia, CA) PCR purification kit per manufacturer’s instructions. The final hybridization reaction (33 in total) contains: 24.9 µl sample, 4.95 µl of 20x SSC, 2 µl of 10 mg/ml poly-A, and 1.21 µl of 5% SDS. The mixture was boiled at 100°C for 2 min, followed by spinning at 14,000 rpm for 10 min to bring down condensation before applying to an array. The array was covered by a coverslip and put in GeneMachines (San Carlos, CA) Hyb chamber. The chamber was placed at 42°C for 30 min, then incubated at 64°C for 16–18 h. The next day, the slide was washed in a series of SDS/SSC before scanning. The data collected from the scanning were analyzed using GenePix Pro 3.0 (Axon Instruments, Union City, CA). Data were submitted to GEO with the submission numbers GPL358, GSM7798–GSM7799, and GSE529.

Reverse transcriptase PCR (RT-PCR) and DNA sequencing.
Poly(A)+-RNA purified from C-S or ADAR–/– was first digested with RNase-free DNase (Invitrogen) before performing reverse transcription. First-strand cDNA was synthesized using SuperScript II as described above. An aliquot of cDNA was used as template for PCR. Gene-specific primer pair was used to amplify cDNA fragment of chosen gene. The PCR products amplified from C-S or ADAR–/– cDNA were cloned into TOPO TA cloning vector (Invitrogen). For comparison, at least five clones generated from two sets of PCR reactions were sequenced (Keck Technology, Yale School of medicine, New Haven, CT) and analyzed using the GCG program.

Sequence analysis.
Genomic and cDNA sequence of Drosophila were downloaded from Berkeley Drosophila Genomic Database (BDGP; Release 2). Comparisons between genomic sequences and the cDNA sequences of specific genes were done by BLASTn.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Specificity of anti-inosine antibody.
The antibody raised against inosine was developed to enrich inosine-containing mRNA from total mRNA. The specificity of the antibody against inosine was tested by polynucleotides immobilized on nitrocellulose membrane. As shown in Fig. 1, antibodies eluted with glycine and tetraethylamide could detect inosine polynucleotides, but not polyguanosine or mixed polynucleotides, poly(AGU) and poly(ACU). The specificity for polyinosine over polyguanosine was estimated to be 1.0 ± 0.2 x 103 at pH 11.5 and greater than 5.0 ± 1 x 103 based at pH 2.5 results.

Immunoaffinity purification of inosine-containing mRNA.
The specificity of the antibody against inosine-containing mRNA over other mRNAs was evaluated using a fragment of GluR-B pre-mRNA, in which adenosine at the editing site (Q/R site) was specifically labeled with 32P. The adenosine at the Q/R site was partially converted into inosine by recombinant editing enzyme ADAR2, so that the ratio of inosine-containing mRNA over mRNA reached ~30:70 (34, 35). This mixture of RNA was loaded onto a protein-A Sepharose column with immobilized anti-inosine antibody. The bound RNA could be efficiently eluted with 0.05 mM inosine monophosphate. We found that it could also be efficiently eluted with 1 mM of GTP. In the latter case, the affinity column can be reused without loss of the binding capacity to inosine-containing mRNA. This may be explained by weak reversible interaction between guanosine and anti-inosine antibody. We therefore included 1 mM of ATP, CTP, and UTP, and 5 µM of GTP in binding or washing buffer to reduce nonspecific interactions. Under typical conditions, ~20% of the input GluR-B RNA was bound to the immobilized antibody.

The ratio between adenosine and inosine in the flow-through and bound RNA was analyzed by TLC after digestion with RNase P1. It is important to highlight that I/A in each fraction directly reflected the ratio between inosine-containing mRNA and mRNA or between edited and unedited GluR-B pre-mRNA. As shown in Fig. 2A, a significant increase of inosine in the bound mRNA was obtained. Efficient separation of inosine-containing mRNA and mRNA was observed in the early fractions, in which no inosine could be detected. In the later flow-through fractions, increasing amount of inosine-containing mRNA could be seen, suggesting that their binding affinity was low and reversible. In contrast, inosine-containing mRNA with high affinity was bound to the column and was eluted with GTP. The ratio of inosine-containing mRNA to mRNA in the elution was greater than 95%. Some inosine-containing mRNAs have low binding affinity, possibly because of a secondary structure that prevents inosine from interacting with the antibody. This is supported by the fact that part of the unbound inosine-containing mRNA could bind again after heat denaturation. In addition, inosine-containing mRNA with high affinity efficiently binds to the column again, and the purity of inosine-containing mRNA increases to up to 98% (Fig. 2B). Thus, although the binding affinity of inosine-containing mRNA to anti-inosine antibody can be affected by secondary structures, it can be improved at least partially by heat denaturation and slowing down of the annealing process.

Isolation of inosine-containing mRNA in Drosophila.
We have applied this immunoaffinity approach to isolate inosine-containing mRNAs from the wild-type (C-S) and the mutant (ADAR–/–) flies. To do this, poly(A)+ RNA was isolated from brains of the wild-type or the mutant flies. On two identical columns, equal amounts of the C-S and ADAR–/– poly(A)+ RNAs (100 µg) were loaded in parallel and washed with the same volume of buffer and eluted with the same volume of buffer containing 1 mM ITP. The eluted inosine-containing mRNA from the C-S or ADAR–/– flies was precipitated. A series of dilutions were spotted on nitrocellulose membrane and detected with anti-inosine antibody and quantified with densitometry (Fig. 3, A and B). Inosine content in the dADAR mutant decreased by about threefold compared with the C-S flies, indicating that inosine-containing mRNAs were enriched in C-S flies. Since our dADAR mutant flies are null-mutant, these data suggest that the excess inosine-containing mRNAs in C-S flies are products of dADAR editing enzyme (24).



View larger version (34K):
[in this window]
[in a new window]
 
Fig. 3. Inosine-rich mRNA in Drosophila. A: far-Western using polyclonal anti-inosine antibody to detect the inosine-rich mRNA in the elutes from the affinity column. About 100 µg of mRNA from wild-type flies (C-S) or dADAR mutant flies (ADAR–/–) were subjected to affinity column with anti-inosine antibody and then eluted by inosine 5'-monophosphate. The elutes were diluted and dotted on nitrocellulose membrane and detected with anti-inosine antibody. B: the density of the dots from C-S was about threefold that of from ADAR–/– under the same dilution, indicating that there was more inosine-rich mRNA in the elutes from the affinity column in C-S than in ADAR–/–.

 
Identification of dADAR-related inosine-containing mRNA in wild-type flies.
To identify potential mRNA targets of dADAR, inosine-containing mRNAs from C-S and ADAR–/– flies were directly compared by hybridization with fly cDNA microarray. mRNA from C-S and ADAR–/– was exposed to an affinity column, and inosine-containing mRNA was eluted. The eluted mRNA portion was amplified by T7 polymerase system (4) and labeled with different fluorescent dyes. We have labeled eluted mRNA from C-S with Cy5 and that from ADAR–/– with Cy3. The labeled probes from C-S and ADAR–/– were then combined and subjected to hybridization with Drosophila cDNA arrays that have over 13,000 genes. The experiments were repeated three times independently. In one experiment, the same amount of probes was hybridized with two different cDNA microarrays to eliminate experimental errors. The correlation coefficient of the hybridization between each microarray was around 0.7. The average of ratio change was obtained and analyzed. More than one-third of the genes (4,000) had fluorescent intensities above background signals. Half of them (>2,000) had a ratio of F535/F635 (ADAR–/–/C-S) between 0.8 and 1.2, indicating that these genes are not differentially expressed. Around 1,000 genes had a ratio of less than 0.75, which corresponds to a false discovery rate of 20 after the significance analysis of microarrays (SAM; Ref. 33). Those genes are present in a higher amount in the C-S eluted mRNA portion than in mutant. Therefore, those genes were considered as potential dADAR targets. With this biochemical approach, a list of about 500 genes were consistently identified as potential dADAR targets (Supplemental Table S1, a list of the top 30, is available at the Physiological Genomics web site).1 Interestingly, a small number of genes were present in a higher amount in the mutant eluted mRNA portion than that in C-S. Since there is only one known adenosine deaminase in flies, we believe that this could be caused by either undefined cotranscriptional events and/or uneven amplification during in vitro transcription or still possibly nonspecificity of the affinity column.

A-to-G changes in Drosophila cDNA.
In the targets of dADAR, A-to-G conversions should also be detectable by comparing the genomic sequence with the cDNAs (32). The integration of microarrays and sequence comparisons could provide more information about dADAR targets. Taking advantage of the Drosophila genome sequence (BDGP Release 2), we compared them with about 7,000 predicted Drosophila cDNAs (BDGP Release 2). With this approach, we could identify the genes that have A-to-G changes in their cDNA sequences. The position of A-to-G changes in each gene was identified by using the following criteria. First, each position must have an A in the genomic sequence but must have a G in the corresponding position of the cDNA sequence. Second, A-to-G changes would not be considered if they are in the 3'-terminal region of cDNA fragment. Third, to avoid random sequence errors, all types of nucleic acid changes in the cDNA sequences were calculated. The cDNA sequences that have nearly equal levels of other nucleotide changes are likely due to sequencing error. Thus we calculated the relative change rate of A-to-G vs. other changes as background and chose the genes with the highest rates as potential dADAR targets. Supplemental Table S2 shows the list of these genes (top 30) sorted by the number of A-to-G changes. About 1,000 genes have been selected as the ratio of A-to-G changes over other changes is less than 1/12. Among them, 800 genes have A-to-G change in their coding region.

Sequencing analysis of the editing sites in identified inosine-containing mRNA.
To overcome the limitation of each method and improve the overall accuracy in the prediction of new dADAR targets, we selected those genes that were both identified with microarrays and sequence alignment as potential dADAR targets for further analysis. A final list of 62 genes (Table 1) was identified as potential dADAR targets with both the affinity column/microarray method and the sequence comparison between genomic sequences and cDNAs. Among the 62 genes, CG18314, which encodes an amine receptor, serves as a positive control, since it was previously identified as a potential dADAR target by sequencing comparison and gene-specific RT-PCR (32).


View this table:
[in this window]
[in a new window]
 
Table 1. Final list of the putative dADAR targets from the integration of biochemical and sequence analysis data

 
If the A-to-G change in the identified inosine-containing mRNAs is the consequence of dADAR action, then adenosine in that position should be present and could be confirmed in the dADAR mutant flies. Since we used anti-inosine antibody to concentrate inosine-containing mRNA in dADAR mutant flies and combined this approach with sequence analysis to get the potential targets in Table 1, we wished to determine whether there were actually A/G differences in these two populations. We therefore directly compared RT-PCR products from wild-type and dADAR mutant flies. Primers for RT-PCR were synthesized to test a group of 12 randomly chosen genes from the 62 genes with different functions (italics in Table 1). Total RNA was isolated from the C-S or ADAR–/– flies, amplified by RT-PCR, and cloned into a vector. With sequencing, we found that 7 of 12 chosen genes were confirmed to have the predicted A in ADAR–/– fly and G in C-S fly cDNA (Table 2). To rule out the potential for background differences between lines, we thoroughly checked the BDGP single nucleotide polymorphism (SNP) database and did not find any in the editing spots of the seven genes of interest in the database. We further chose the target candidate CG18314 and performed sequence analysis of the target loci from cDNA and genomic DNA from the same fly background to confirm RNA editing. The sequences from all of dADAR mutant clones (10) were identical to the genomic one. The sequences from wild-type flies, however, contained at least two edited sites in the sequenced region. The sequence is presented here, with possible editing sites in lower cases: GCAAGCAGAGCTaTCGGGTTGGACCCAGATGACTaGAGTTAGATTTAGA. The editing rates for the 1st and 2nd site are 7/9 and 5/9, respectively.


View this table:
[in this window]
[in a new window]
 
Table 2. Editing of seven potential dADAR targets at certain hot spots

 
Expression of the identified inosine-containing mRNA in the wild-type and the dADAR mutant flies.
Since known substrates of dADAR such as paralytic (para) and cacophony (cac) were not in the final list of potential dADAR target candidates and in the list of immunoaffinity-purified inosine-containing mRNAs (Supplemental Table S1), we reasoned that the identified inosine-containing mRNA by this method included only those relatively abundant genes and that the expression level of the para and cac mRNA may be relatively low. To answer this question, we performed semiquantitative PCR and compared the expression levels of para and cac mRNA and the identified inosine-containing mRNA CG18314 known as amine receptor gene. As shown in Fig. 4, the amine receptor cDNA, but not the para or cac cDNA, could be detected under the same conditions, indicating that in adult flies the expression of para and cac mRNAs were much less than that of the amine receptor.



View larger version (47K):
[in this window]
[in a new window]
 
Fig. 4. Semiquantitative PCR showing the expression level of para, cac, CG18314 (encoding amine receptor). Actin was used as control. In the image, lanes 1, 5, and 9 were for actin; lanes 2, 6, and 10 were for CG18314; lanes 3, 7, and 11 were for para; and lanes 4, 8, and 12 were for cac. Lanes 14 were amplified for 15 cycles; lanes 58 were amplified for 20 cycles; and lanes 912 were for 25 cycles. The expression level of gene CG18314 was much higher in adult flies than that of para or cac.

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Isolating inosine-containing mRNA using inosine affinity column provides a useful method to characterize the products of A-to-I RNA editing. The same approach may also be used to analyze RNAs with modified nucleotides, for instance, tRNA with rare bases and splicing intermediates with the branched RNA. Using this approach, in combination with sequence analysis of A-to-G changes in databases, inosine-containing mRNAs from the wild-type and dADAR mutant flies have been isolated and compared.

RNA editing is a novel mechanism for generating distinct protein isoforms from a single pre-mRNA (8, 19). Adenosine (A) to inosine (I) pre-mRNA editing is a process that enzymatically modifies A-to-I at single nucleotide positions in specific messengers (13). The enzymes responsible for A-to-I conversion, the adenosine deaminases that act on RNA (ADARs), have been identified in mammals, Drosophila, and Caenorhabditis elegans (3). Lack of A-to-I editing in mice or flies shows severe behavioral defects, suggesting an important role of ADARs in the nervous system (11, 18, 24, 34). However, exactly how the enzymes mediate their function, and what are the RNA targets, is unclear. Up to now, only a few transcripts have been known to be targeted by ADARs. Most of these transcripts encode ligand- or voltage-gated ion channels and G protein-coupled receptors, such as sodium channel, calcium channel, GluR-B, serotonin receptor, etc. (6, 9, 12, 1517, 29, 31, 35). Interestingly, the few reports of specific A-to-I pre-mRNA editing are mostly from nervous system, despite the fact that ADAR activities are detected in nonnervous tissues (25). Moreover, the amount of inosine in mRNA isolated from various rat tissues shows that far more inosine is present than can be accounted for by the few mammalian ADAR substrates (21, 25). Thus it is conceivable that additional RNA targets of A-to-I editing exist (25).

The reason for pursuing Drosophila ADAR (dADAR) targets stems from the facts that 1) wild-type flies have all A-to-I editing found in transcripts encoding ion channels and neurotransmitter receptors such as sodium (para), calcium (dmca1A) and chloride (DrosGluCl-{alpha}) channels and D{alpha}6 (nAChR) receptor and 2) the deletion of dADAR activity resulted in extreme behavioral deficits including hypoxia sensitivity, temperature-sensitive paralysis, and neurodegeneration (18, 24). We have previously found that dADAR mutant flies are more resistant to oxidative stress than wild type. This cannot be explained by the currently known targets, and searching for additional target genes becomes critical.

To better understand ADAR function, we wished to search for new targets of ADARs. Other groups have previously used either biochemical approaches or bioinformatic methods to discover new ADAR targets (20, 22, 32). In our present work, we have identified several new targets of Drosophila RNA editase by taking advantage of the dADAR mutant flies and the combination of microarray and sequence analysis. We have developed a polyclonal antibody against inosine to trap the inosine-containing mRNA in wild-type flies and used the inosine-containing mRNA to hybridize the Drosophila cDNA microarray. Using this approach, we were able to identify 500 potential dADAR targets (Supplemental Table S1). There are two limitations for this biochemical method. First, the use of an anti-inosine antibody and affinity column to trap inosine-containing mRNA in our experiments could give nonspecific binding of mRNA. This could result from the competition between inosine and guanosine. Since there are few inosine but many guanosine in mRNA molecules, even a high-affinity anti-inosine antibody could not guarantee high specificity of binding. From our data, the specificity of the antibody to polyinosine over polyguanosine was about 5,000-fold (Fig. 1). The overall affinity of the antibody to a particular inosine-containing RNA over regular RNA is estimated within a few-fold, which is largely dependent on the number of inosine and the length of RNA (Fig. 3), as indicated by 20-fold excess of GTP elution. Considering that there are no other available techniques to isolate the edited RNA, what we have reported in this manuscript is the first attempt to use this affinity approach for this purpose. We believe that these data are valuable to the investigators in this field. Second, during the amplification process, the ratio of the amount of inosine-containing mRNA could be changed, and therefore the resulting ratio could be different from that of the original one. It is also possible that due to their relative low expression level, some of the known dADAR targets, such as para and cac, were not detected with this method (Fig. 4).

To be able to detect new targets of dADAR, we also compared the genomic and cDNA sequence of a gene using BLASTn, which revealed A-to-G change in about 1,000 genes (Supplemental Table S2). The limitations of the BLASTn are 1) only 7,000 of over 13,000 genes have the corresponding cDNA sequences, and 2) the DNA sequence itself is also associated with sequence errors. About 1% error rate was reported for EST sequencing (26).

With the combination of biochemical and sequence analysis, the limitation of either method can be overcome, and the integration of these two methods provided us with a way to identify new targets with more accuracy. The overlapping of target candidates from the two approaches generated a final list with 62 genes (Table 1). Among them, CG18314, which encodes an amine receptor, was previously identified as a potential dADAR target by sequencing comparison and gene-specific RT-PCR (32). The paper released by BDGP (32) revealed 10 RNA editing sites in this gene by comparison of the genomic sequence (Release 3) with two cDNAs and three ESTs (from Drosophila Gene Collection). In our case, we found five potential RNA editing sites in this gene by sequencing comparison and validated two of them by gene-specific RT-PCR. To further verify the targets candidates, we randomly chose 12 genes from the 62 genes and sequenced their RT-PCR products for wild-type and dADAR mutant flies. DNA sequencing revealed that 7 of 12 genes were proved to have A in mutant and G in wild type (Table 2). The validation of our approach was confirmed by ruling out the possibility that this A-to-G change is due to SNP. We thoroughly checked the BDGP SNP database and did not find any in the editing spots of the seven genes of interest in the database. Given the facts that 1) SNP rate in Drosophila is known to be about 1 in 200–400 bases (13), and 2) A-to-G changes contribute only 1/12 of all possible single nucleotide changes, the rate of A-to-G change due to SNP is about 1/2,000–1/4,000. Since this is far less (an order of magnitude) than the frequency of the A-to-G change in our sequencing of the 7 genes (1/300 bases), the A-to-G change in the seven genes is very likely due to editing by ADAR and not due to single nucleotide polymorphism. This was confirmed by our sequencing of the target candidate CG18314 loci from cDNA and genomic DNA which supported our prediction of RNA editing.

The results from our work have been encouraging. On the one hand, among the seven genes, a few of them encoded G protein-coupled receptors (CG18314) and ion channels (CG11348). On the other hand, using this approach, the accuracy of the prediction of new dADAR targets is more than 50%. This could largely facilitate the course of identifying new targets of dADAR. In summary, the use of multiple techniques and the integration of immunoaffinity enrichment of inosine-containing mRNA, DNA microarrays, and sequence analysis provide us with a way to identify new substrates of dADAR. In addition, this allows us to better understand the role of ADAR and its targets in cell function.


    GRANTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
This work was supported by grants from the National Institutes of Health Grants NS-35918, HL-07778, HD-32573, and GM-60426.


    ACKNOWLEDGMENTS
 
We thank Dr. Kevin White for generously providing us with Drosophila microarray slides. We also thank Dr. Yingyi Xiao for expert assistance with microarrays.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: G. G. Haddad, Dept. of Pediatrics, Albert Einstein College of Medicine, Rose Kennedy Center, Bronx, NY 10461 (E-mail: ghaddad{at}aecom.yu.edu). E-mail for J. Yang: jinghua.yang{at}yale.edu.

10.1152/physiolgenomics.00093.2003.

1 The Supplemental Material for this article (Supplemental Tables S1 and S2) is available online at http://physiolgenomics.physiology.org/cgi/content/full/00093.2003/DC1. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

  1. Bass BL. RNA editing and hypermutation by adenosine deamination. Trends Biochem Sci 22: 157–162, 1997.[CrossRef][ISI][Medline]
  2. Bass BL. RNA editing by adenosine deaminases that act on RNA. Annu Rev Biochem 71: 817–846, 2002.[CrossRef][ISI][Medline]
  3. Bass BL, Nishikura K, Keller W, Seeburg PH, Emeson RB, O’Connell MA, Samuel CE, and Herbert A. A standardized nomenclature for adenosine deaminases that act on RNA. RNA 3: 947–949, 1997.[ISI][Medline]
  4. Baugh LR, Hill AA, Brown EL, and Hunter CP. Quantitative analysis of mRNA amplification by in vitro transcription. Nucleic Acids Res 29: E29, 2001.[Medline]
  5. Benne R, Van den Burg J, Brakenhoff JP, Sloof P, Van Boom JH, and Tromp MC. Major transcript of the frameshifted coxII gene from trypanosome mitochondria contains four nucleotides that are not encoded in the DNA. Cell 46: 819–826, 1986.[ISI][Medline]
  6. Burns CM, Chu H, Rueter SM, Hutchinson LK, Canton H, Sanders-Bush E, and Emeson RB. Regulation of serotonin-2C receptor G-protein coupling by RNA editing. Nature 387: 303–308, 1997.[CrossRef][ISI][Medline]
  7. Egebjerg J and Heinemann SF. Ca2+ permeability of unedited and edited versions of the kainate selective glutamate receptor GluR6. Proc Natl Acad Sci USA 90: 755–759, 1993.[Abstract/Free Full Text]
  8. Gott JM and Emeson RB. Functions and mechanisms of RNA editing. Annu Rev Genet 34: 499–531, 2000.[CrossRef][ISI][Medline]
  9. Grauso M, Reenan RA, Culetto E, and Sattelle DB. Novel putative nicotinic acetylcholine receptor subunit genes, D{alpha}5, D{alpha}6 and D{alpha}7, in Drosophila melanogaster identify a new and highly conserved target of adenosine deaminase acting on RNA-mediated A-to-I pre-mRNA editing. Genetics 160: 1519–1533, 2002.[Abstract/Free Full Text]
  10. Haddad GG, Sun Y, Wyman RJ, and Xu T. Genetic basis of tolerance to O2 deprivation in Drosophila melanogaster. Proc Natl Acad Sci USA 94: 10809–10812, 1997.[Abstract/Free Full Text]
  11. Higuchi M, Maas S, Single FN, Hartner J, Rozov A, Burnashev N, Feldmeyer D, Sprengel R, and Seeburg PH. Point mutation in an AMPA receptor gene rescues lethality in mice deficient in the RNA-editing enzyme ADAR2. Nature 406: 78–81, 2000.[CrossRef][ISI][Medline]
  12. Higuchi M, Single FN, Kohler M, Sommer B, Sprengel R, and Seeburg PH. RNA editing of AMPA receptor subunit GluR-B: a base-paired intron-exon structure determines position and efficiency. Cell 75: 1361–1370, 1993.[ISI][Medline]
  13. Hoskins RA, Phan AC, Naeemuddin M, Mapa FA, Ruddy DA, Ryan JJ, Young LM, Wells T, Kopczynski C, and Ellis MC. Single nucleotide polymorphism markers for genetic mapping in Drosophila melanogaster. Genome Res 11: 1100–1113, 2001.[Abstract/Free Full Text]
  14. Kim U, Wang Y, Sanford T, Zeng Y, and Nishikura K. Molecular cloning of cDNA for double-stranded RNA adenosine deaminase, a candidate enzyme for nuclear RNA editing. Proc Natl Acad Sci USA 91: 11457–11461, 1994.[Abstract/Free Full Text]
  15. Lai F, Chen CX, Carter KC, and Nishikura K. Editing of glutamate receptor B subunit ion channel RNAs by four alternatively spliced DRADA2 double-stranded RNA adenosine deaminases. Mol Cell Biol 17: 2413–2424, 1997.[Abstract]
  16. Lee CJ, Kong H, Manzini MC, Albuquerque C, Chao MV, and MacDermott AB. Kainate receptors expressed by a subpopulation of developing nociceptors rapidly switch from high to low Ca2+ permeability. J Neurosci 21: 4572–4581, 2001.[Abstract/Free Full Text]
  17. Lomeli H, Mosbacher J, Melcher T, Hoger T, Geiger JR, Kuner T, Monyer H, Higuchi M, Bach A, and Seeburg PH. Control of kinetic properties of AMPA receptor channels by nuclear RNA editing. Science 266: 1709–1713, 1994.[ISI][Medline]
  18. Ma E, Gu XQ, Wu X, Xu T, and Haddad GG. Mutation in pre-mRNA adenosine deaminase markedly attenuates neuronal tolerance to O2 deprivation in Drosophila melanogaster. J Clin Invest 107: 685–693, 2001.[Abstract/Free Full Text]
  19. Maas S and Rich A. Changing genetic information through RNA editing. Bioessays 22: 790–802, 2000.[CrossRef][ISI][Medline]
  20. Morse DP, Aruscavage PJ, and Bass BL. RNA hairpins in noncoding regions of human brain and Caenorhabditis elegans mRNA are edited by adenosine deaminases that act on RNA. Proc Natl Acad Sci USA 99: 7906–7911, 2002.[Abstract/Free Full Text]
  21. Morse DP and Bass BL. Detection of inosine in messenger RNA by inosine-specific cleavage. Biochemistry 36: 8429–8434, 1997.[CrossRef][ISI][Medline]
  22. Morse DP and Bass BL. Long RNA hairpins that contain inosine are present in Caenorhabditis elegans poly(A)+ RNA. Proc Natl Acad Sci USA 96: 6048–6053, 1999.[Abstract/Free Full Text]
  23. O’Connell MA, Krause S, Higuchi M, Hsuan JJ, Totty NF, Jenny A, and Keller W. Cloning of cDNAs encoding mammalian double-stranded RNA-specific adenosine deaminase. Mol Cell Biol 15: 1389–1397, 1995.[Abstract]
  24. Palladino MJ, Keegan LP, O’Connell MA, and Reenan RA. A-to-I pre-mRNA editing in Drosophila is primarily involved in adult nervous system function and integrity. Cell 102: 437–449, 2000.[ISI][Medline]
  25. Paul MS and Bass BL. Inosine exists in mRNA at tissue-specific levels and is most abundant in brain mRNA. EMBO J 17: 1120–1127, 1998.[Abstract/Free Full Text]
  26. Picoult-Newberg L, Ideker TE, Pohl MG, Taylor SL, Donaldson MA, Nickerson DA, and Boyce-Jacino M. Mining SNPs from EST databases. Genome Res 9: 167–174. 1999.[Abstract/Free Full Text]
  27. Powell LM, Wallis SC, Pease RJ, Edwards YH, Knott TJ, and Scott J. A novel form of tissue-specific RNA processing produces apolipoprotein-B48 in intestine. Cell 50: 831–840, 1987.[ISI][Medline]
  28. Reenan RA. The RNA world meets behavior: A->I pre-mRNA editing in animals. Trends Genet 17: 53–56, 2001.[CrossRef][ISI][Medline]
  29. Reenan RA, Hanrahan CJ, and Barry G. The mle(napts) RNA helicase mutation in Drosophila results in a splicing catastrophe of the para Na+ channel transcript in a region of RNA editing. Neuron 25: 139–149, 2000.[ISI][Medline]
  30. Simpson L, Thiemann OH, Savill NJ, Alfonzo JD, and Maslov DA. Evolution of RNA editing in trypanosome mitochondria. Proc Natl Acad Sci USA 97: 6986–6993, 2000.[Abstract/Free Full Text]
  31. Smith LA, Peixoto AA, and Hall JC. RNA editing in the Drosophila DMCA1A calcium-channel alpha 1 subunit transcript. J Neurogenet 12: 227–240, 1998.[Medline]
  32. Stapleton M, Carlson J, Brokstein P, Yu C, Champe M, George R, Guarin H, Kronmiller B, Pacleb J, Park S, Wan K, Rubin GM, and Celniker S. A Drosophila full-length cDNA resource. Genome Biol 3: RESEARCH0080, 2002.[Medline]
  33. Tusher VG, Tibshirani R, and Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98: 5116–5121, 2001.[Abstract/Free Full Text]
  34. Wang Q, Khillan J, Gadue P, and Nishikura K. Requirement of the RNA editing deaminase ADAR1 gene for embryonic erythropoiesis. Science 290: 1765–1768, 2000.[Abstract/Free Full Text]
  35. Yang JH, Sklar P, Axel R, and Maniatis T. Editing of glutamate receptor subunit B pre-mRNA in vitro by site-specific deamination of adenosine. Nature 374: 77–81, 1995.[CrossRef][ISI][Medline]
  36. Yang JH, Sklar P, Axel R, and Maniatis T. Purification and characterization of a human RNA adenosine deaminase for glutamate receptor B pre-mRNA editing. Proc Natl Acad Sci USA 94: 4354–4359, 1997.[Abstract/Free Full Text]
  37. Yang JH, Luo X, Nie Y, Su Y, Zhao Q, Kabir K, Zhang D, and Rabinovici R. Widespread inosine-containing mRNA in lymphocytes regulated by ADAR1 in response to inflammation. Immunology 109: 15–23, 2003.[ISI][Medline]