Advertisement

Evidence for a Site-specific Cytidine Deamination Reaction Involved in C to U RNA Editing of Plant Mitochondria *

  • Wei Yu
    Affiliations
    Institut für Genbiologische Forschung Berlin GmbH, Ihnestrasse 63, D-14195 Berlin, Federal Republic of Germany
    Search for articles by this author
  • Wolfgang Schuster
    Correspondence
    To whom correspondence should be addressed: Institut für Genbiologische Forschung Berlin GmbH, Ihnestrasse 63, D-14195 Berlin, FRG. Tel.: 49-30-83000764; Fax: 49-30-83000736;
    Affiliations
    Institut für Genbiologische Forschung Berlin GmbH, Ihnestrasse 63, D-14195 Berlin, Federal Republic of Germany
    Search for articles by this author
  • Author Footnotes
    * This work was supported by grants from the Deutsche Forschungsgemeinschaft, the Bundesministerium für Forschung und Technologie, and the Human Frontiers Science Program. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore by hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Open AccessPublished:August 04, 1995DOI:https://doi.org/10.1074/jbc.270.31.18227
      Transcripts of higher plant mitochondria are modified post-transcriptionally by RNA editing. To distinguish between the mechanisms by which the cytidine to uridine transition could occur a combined transcription/RNA editing assay and an in vitro RNA editing system were investigated. Mitochondria isolated from etiolated pea seedlings and potato tubers were supplied with [α-32P]CTP to radiolabel the mitochondrial run-on transcripts. High molecular weight run-on transcripts were isolated and hydrolyzed, and nucleotide identities were analyzed by one- and two-dimensional thin layer chromatography. The amount of label comigrating with UMP nucleotides increases with extended incubation times. Analogous products were obtained by incubation of [α-32P]CTP or [5-3H]CTP radiolabeled in vitro transcripts with a mitochondrial lysate from pea mitochondria. 5-3H label of the cytosine base was detected in the UMP spot after incubation of in vitro transcripts with mitochondrial lysate. These results are consistent with a deamination reaction involved in this post-transcriptional C to U modification process. To prove that cytidines are deaminated specifically in vitro transcripts were reisolated after incubation and analyzed by reverse transcription-polymerase chain reaction. Sequence analysis clearly shows that only cytidines at editing sites are edited while residual cytidines are not modified and suggests that site-specific factors are involved in RNA editing of plant mitochondria.

      INTRODUCTION

      RNA editing in higher plant mitochondria is a process by which genomically encoded cytidine residues are changed to uridines on the RNA level(
      • Covello P.S.
      • Gray M.W.
      ,
      • Gualberto J.M.
      • Lamattina L.
      • Bonnard G.
      • Weil J.-H.
      • Grienenberger J.-M.
      ,
      • Hiesel R.
      • Wissinger B.
      • Schuster W.
      • Brennicke A.
      ). Although most of the conversions have been found in translated regions, structural RNAs, intron sequences, and also leader and trailer sequences are occasionally modified(
      • Schuster W.
      • Ternes R.
      • Knoop V.
      • Hiesel R.
      • Wissinger B.
      • Brennicke A.
      ). Polypeptides translated from edited transcripts show a higher similarity to their homologues in animals or fungi than those predicted from unedited transcripts(
      • Covello P.S.
      • Gray M.W.
      ,
      • Schuster W.
      ).
      Two biochemically distinct processes have been identified as capable of changing the information at the RNA level. One involves the insertion or deletion of nucleotides(
      • Benne R.
      • van den Burg J.
      • Brakenhoff J.P.J.
      • Sloof P.
      • van Boom J.H.
      • Tromp M.C.
      ), and the other modifies specific nucleotides in the mRNA sequence(
      • Chen S.-H.
      • Habib G.
      • Yang C.Y.
      • Gu Z.W.
      • Lee B.R.
      • Weng S.A.
      • Silberman S.R.
      • Cai S.J.
      • Deslypere J.P.
      • Rosseneu M.
      • Gotto Jr., A.M.
      • Li W.H.
      • Chan L.
      ). RNA editing in kinetoplastid protozoa exemplifies the process of an insertional editing system with the addition and deletion of uridine residues at specific sites. In the current model this editing machinery is guided by small RNAs (guide RNAs) that are complementary to the edited transcripts (
      • Blum B.
      • Bakalara N.
      • Simpson L.
      ) and also serve as a uridine donor for the editing activity(
      • Blum B.
      • Sturm N.
      • Simpson A.M.
      • Simpson L.
      ). Modifications of nucleotides in tRNA or ribosomal RNA sequences, although usually not called RNA editing, can affect the stability of the RNA or modulate the RNA's conformation and identity. In the alternative process RNA editing changes the information content via modification, e.g. in creating a UAA stop-translation codon from a CAA glutamine codon in apolipoprotein B mRNA or by modification of a CAG codon specifying glutamine to an arginine codon (CGG) in neural glutamate-gated calcium channel mRNA(
      • Sommer B.
      • Khler M.
      • Sprengel R.
      • Seeburg P.H.
      ). In the latter example, a double-stranded RNA adenosine deaminase has been suggested as the enzyme responsible for deaminating adenosine to inosine and thus changing a CAG codon to CIG(
      • Higuchi M.
      • Single F.N.
      • Khler M.
      • Sommer B.
      • Sprengel R.
      • Seeburg P.H.
      ). In apolipoprotein B mRNA editing, a cytidine deaminase has been shown to be one component of the enzymatic modification activity(
      • Navaratnam N.
      • Morrison J.R.
      • Bhattacharya S.
      • Patel D.
      • Funabashi T.
      • Giannoni F.
      • Teng B.-B.
      • Davidson N.O.
      • Scott J.
      ).
      Neither biochemical mechanism of the editing reaction in higher plant mitochondria nor the components of the editing machinery are currently known. Three possible mechanisms able to change a cytidine post-transcriptionally could proceed by modification of the base, by base exchange, or by replacement of the complete nucleotide (Fig. 1). The most simple would be creation of a uridine by a site-specific hydrolytic deamination of the cytosine at position 4. Modification of the base pairing ability of the cytidine, e.g. by the attachment of the amino acid lysine at position 2 of the cytidine, would create a lysidine that is read by most enzymes as a uridine. Another possibility is a transglycosylation reaction in which the cytosine is replaced by uracil through breaking and reforming of the glycosyl bond. Alternatively the cytidine nucleotide could be replaced by a uridine nucleotide through cleavage and ligation of the phosphodiester backbone of the RNA chain.
      Figure thumbnail gr1
      Figure 1:Several biochemical mechanisms are known that could change C to U nucleotide identities. Deamination of the cytidine could create a uridine, while modification of the cytosine base can result in a hypermodified base (e.g. lysidine), which is recognized as uridine. Replacement of the cytosine base by uracil in a transglycosylation reaction or complete replacement of the cytidine nucleotide by uridine via deletion and insertion could also change the nucleotide identity.
      To investigate the biochemical nature underlying the RNA editing process in mitochondria of higher plants, we traced the fate of the α-phosphate of the cytidine and a 5-3H label in the cytosine base after incorporation into high molecular weight RNA. A system for coupled transcription/RNA editing was established for this purpose. Results obtained by this system are confirmed by an in vitro RNA editing analysis, in which 32P or 3H radiolabeled in vitro transcripts were incubated with lysates from pea mitochondria. To show that the editing reaction is site-specific we compared edited and unedited templates after incubation with mitochondrial lysate.

      EXPERIMENTAL PROCEDURES

      Isolation of Plant Mitochondria

      Mitochondria were isolated from potato tubers (Solanum tuberosum var. Bintje) or from dark grown shoots of 4-day-old pea seedlings (Pisum sativum var. Progress) by four 5-s pulses on a high speed blender in extraction buffer containing 400 mM mannitol, 1 mM EGTA, 25 mM Tricine
      The abbreviations used are: Tricine
      N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine
      PCR
      polymerase chain reaction
      RT
      reverse transcription.
      (pH 7.2), 10 mM β-mercaptoethanol, 50 μM phenylmethylsulfonyl fluoride, 0.1% bovine serum albumin, and subsequent filtering through two layers of Miracloth. The filtrate was spun at 2,000 × g for 15 min at 4°C to remove cellular debris and then at 15,000 × g for 20 min at 4°C to pellet mitochondria. After several washes in a buffer containing 400 mM mannitol, 10 mM Tricine (pH 7.2), 1 mM EGTA, 0.2 mM phenylmethylsulfonyl fluoride, mitochondria were layered onto a 14 to 28 to 45% Percoll step gradient and spun at 70,000 × g for 45 min in a Beckman SW-28 rotor at 4°C. The interface band between 28 and 45% was collected and washed several times. Mitochondria were used on the day of isolation for the in organello run-on transcription/RNA editing analysis.

      Preparation of Mitochondrial RNA Editing Extracts

      The RNA editing extract of pea mitochondria was prepared according to a modification of the method used to isolate wheat mitochondrial editing activity(
      • Araya A.
      • Domec C.
      • Begu D.
      • Litvak S.
      ). Briefly, mitochondria were lysed in the presence of 1 M ammonium sulfate and 0.2% Triton X-100. Centrifugation of the lysate at 100,000 × g yielded a membrane-free soluble protein fraction (S100). The S100 fraction was subjected to anion-exchange chromatography using DEAE-cellulose, and proteins were eluted with 50 mM KCl. The fractions were dialyzed against 10 mM Tris-HCl (pH 8.0), 10 mM KCl, 1 mM dithiothreitol, 0.1 mM EDTA, and 10% (w/v) glycerol, concentrated using a Centricon-3 microconcentrator (Amicon), and stored at −80°C.

      Run-on Transcription/in Organello Incubation

      Highly purified mitochondria (200 μg) were resuspended in 1 ml of run-on buffer containing 10 mM Tris-HCl (pH 8.0), 5 mM MgCl2, 1 mM dithiothreitol, 50 mM KCl, 1 mM EDTA, 200 units of RNasin, 0.5 mM each of UTP, ATP, and GTP, and 250 μCi of [α-32P]CTP(
      • Finnegan P.M.
      • Brown G.G.
      ). The reaction mixture was incubated at 37°C, and aliquots were taken after different times up to 1 h. In some experiments, radiolabeled UTP was used instead of CTP to follow the run-on transcription. To investigate the possibility of cytidine to uridine conversion at the nucleotide level high concentrations of cold UTP (5 mM) were used in some run-on transcription assays. After 10, 30, and 60 min of incubation, aliquots were taken and nucleic acids were isolated from mitochondria by two phenol/chloroform extractions. The high molecular weight nucleic acid fraction was purified by a NAP 5 column (Pharmacia Biotech Inc.) and precipitated with ethanol. Nucleic acids were digested to monophosphate nucleotides with 10 units of nuclease P1 (Life Technologies, Inc.) in 10 μl of P1 buffer (50 mM sodium acetate (pH 5.3)) at 37°C for 2-6 h. To verify that transcripts obtained by in organello transcription are real mitochondrial transcripts, and not bacterial contaminations, blots of digested cosmids with mtDNA of Arabidopsis thaliana were hybridized with the transcriptional products of the in organello synthesis (data not shown).

      Thin Layer Chromatography

      Aliquots (5-10 μl) of the nuclease P1 hydrolyzed products were spotted onto cellulose TLC plates (CEL 300-10, Macherey-Nagel, Germany), and one- and two-dimensional chromatography was carried out at 20°C. The solvent systems used for two-dimensional TLC were isobutyric acid/water/25% ammonia (66:33:1 by volume) for the first dimension and 2-propanol/37% HCl/water (70:15:15 by volume) for the second dimension(
      • Nishimura S.
      ). One-dimensional TLC was developed in t-butanol/37% HCl/water (70:15:15 by volume). To identify the positions of the 5ʹ-monophosphates by UV shadowing 2-5 μl each of 100 mM solutions of cytidine, adenosine, uridine, and guanosine 5ʹ-monophosphates (NMPs) were added. The relative positions of the nucleotides varied slightly depending on the time of development and the condition of the solvents, e.g. after repeated use.
      Spots containing hydrolyzed 3H-labeled CMP and UMP were identified after TLC separation by UV shadowing and removed from the TLC plate. After elution of the NMPs from the cellulose material 2 ml of liquid scintillation mixture (Ready Safe, Beckman) were added, and the radioactivity was determined by liquid scintillation counting.

      PCR, in Vitro Transcription, and Incubation

      The templates for in vitro transcription were synthesized by polymerase chain reaction (PCR) using the following primers: T7-5ʹ-primer (cox2): 5ʹ-AGGATC CTAATACGACTCACTATAGGGAGATCTTCCTCATTCTTATTTTGG-3ʹ; 3ʹ-primer (cox2): 5ʹ-AGGATCCAGGTACAGCATCACATTTGAC-3ʹ; T7-5ʹ-primer (orf206): 5ʹ-TAATACGACTCACTATAGGGAGACCACTTCACCTACTTCA-3ʹ; 3ʹ-primer (orf206): 5ʹ-AAATTAACCCTCACTAAAGGGAAAAAGAGATACGAAC-3ʹ. The PCR products obtained represent either an internal part of the cox2 gene (542 nucleotides) from pea(
      • Covello P.S.
      • Gray M.W.
      ), which is linked upstream with a promoter sequence of the T7 RNA polymerase or a highly edited region (114 nucleotides) of the orf206 gene(
      • Schuster W.
      ). For the PCR reaction mtDNA or random primed cDNA from pea mitochondria was used. RT-PCR of orf206 in vitro transcripts was initiated by first strand cDNA synthesis primed by a T3-primer (5ʹ-AATTAACCCTCACTAAAGGG-3ʹ). The PCR reaction was done using the T7-orf206- and the T3-primers, and the PCR amplification products were cloned into the pCR™II vector using a TA cloning kit (Invitrogen). The reaction mixture (100 μl) contained 50 mM KCl, 1.5 mM MgCl2, 10 mM Tris-HCl (pH 8.3), 0.5 μg of each primer, 50 fmol of each dNTP, 10 ng of cDNA, and 2.5 units of Taq polymerase (Boehringer Mannheim). PCR was performed on a Biomed cycler under the following conditions: cycle 1, 1 min at 94°C; cycle 2, 1 min at 45°C; and cycle 3, 3 min at 72°C. All cycles were repeated 30 times, and an extension of 10 min at 72°C was added at the end. In the in vitro transcription reactions 2 μg of PCR product were used. Labeled transcripts were prepared under standard conditions using [α-32P]CTP, [α-32P]UTP, or [5ʹ-3H]CTP and T7 RNA polymerase (Boehringer Mannheim). To obtain RNA of maximal specific activity, 500 μCi of [5ʹ-3H]CTP at 22 Ci/mmol was dried and used per transcription. After transcription, template DNA was removed by DNase I, and the synthesized RNA was precipitated. Aliquots of each reaction step were analyzed on agarose gels to monitor the reactions and to verify the transcription products. The labeled in vitro RNAs were either incubated with mitochondrial lysates or used directly as a control after nuclease P1 digestion in the TLC to analyze the separation conditions. A typical in vitro RNA editing assay contained 50 mM Tris acetate (pH 7.8), 10 mM magnesium acetate, and 25 units of RNasin in a final volume of 50 μl. 40 μg of mitochondrial protein was used per assay.

      RESULTS

      Run-on Transcription in Mitochondria

      To study the biochemistry of the RNA editing process in higher plants, isolated mitochondria were initially chosen as the most intact system. Attempts to transfer in vitro transcripts as editing templates into isolated mitochondria by electroporation had failed to give clear results.
      W. Yu and W. Schuster, unpublished data.
      A run-on transcription system (
      • Finnegan P.M.
      • Brown G.G.
      ,
      • Mulligan R.M.
      • Leon P.
      • Walbot V.
      ) was therefore chosen to approach in vitro the RNA editing process in plant mitochondria. This system allows direct in organello labeling of the editing substrate (mitochondrial RNA) coupled with the possibility to follow the RNA editing process in a time-dependent manner.

      Run-on Transcription/RNA Editing in Intact Potato Mitochondria

      To analyze RNA editing in intact mitochondria, purified potato mitochondria were supplied with [α-32P]CTP to label the newly transcribed mitochondrial transcripts. Aliquots were taken after 0, 15, 30, 45, and 60 min of incubation, and nucleic acids were extracted. Nucleic acids were treated with DNase I, and the bulk of the unincorporated label was removed by gel filtration chromatography on NAP 5 columns (Pharmacia). After digestion of the labeled high molecular weight RNA with nuclease P1 the resulting 5ʹ-monophosphate nucleotides were separated by one- and two-dimensional TLC (Fig. 2). After 15 min of incubation new discrete spots begin to appear and increase in intensity with the time of incubation. One of the spots comigrates with UMP in both one-dimensional and two-dimensional TLC. Two additional spots (x and y) that are not specific to RNA editing generally appear in relatively constant proportion and are most likely incompletely digested dinucleotides, since they also appear after nuclease P1 digestion of in vitro synthesized transcripts not incubated with mitochondrial lysates. Similar results have been previously reported by other investigators for the human apolipoprotein B and maize mitochondrial editing systems(
      • Hodges P.E.
      • Navaratnam N.
      • Greeve J.C.
      • Scott J.
      ,
      • Rajasekhar V.R.
      • Mulligan R.M.
      ). To investigate whether these results are restricted to potato mitochondria we tested mitochondria of several other plant species. Pea mitochondria showed the highest in organello transcriptional activity of the species investigated (data not shown) and were therefore used for further studies.
      Figure thumbnail gr2
      Figure 2:TLC separation of NMPs from potato mitochondrial in organello transcripts. Panel A shows the one-dimensional TLC separation. Potato mitochondria were incubated with [α-32P]CTP, and aliquots were taken after 15, 30, 45, and 60 min. Nucleic acids were hydrolyzed with nuclease P1, and the resulting NMPs were separated by one-dimensional TLC. Inorganic phosphate (Pi) is the product with the highest mobility, followed by UMP (pU), a product of unknown identity (x), CMP (pC), and another unknown product (y). The mobility of products x and y is consistent with the dinucleotides pCpA and pApC derived by incomplete hydrolysis. Panel B shows the two-dimensional separation of NMPs from potato mitochondria. Hydrolysis products from the 30-min incubation separated on one-dimensional TLC were reisolated from the plate and separated by two-dimensional TLC to verify the mobility of the hydrolysis products. The mobility of unlabeled NMPs identified by UV shadowing is shown schematically above the two-dimensional TLC separation. The spots of CMP (pC) and UMP (pU) are indicated by arrows.

      RNA Editing in a Lysed System of Pea Mitochondria

      A lysed in organello system was used to study run-on transcription (
      • Mulligan R.M.
      • Leon P.
      • Walbot V.
      ) and RNA editing in disrupted mitochondria. Purified pea mitochondria were supplied with [α-32P]CTP, analogous to intact mitochondria, without added osmoticum. After 30 and 60 min aliquots were taken, nucleic acids were extracted, and the purified high molecular weight RNAs were analyzed by TLC after nuclease P1 digestion (Fig. 3). After incubation for 30 min an additional spot appeared comigrating with UMP in the TLC mononucleotide analysis. To exclude that the radiolabeled UMP detected by the TLC analysis is due to a CTP deaminase activity produced at the mononucleotide level rather than modification at the polynucleotide level, we added a 10-fold excess of unlabeled UTP to compete any such reaction. The radiolabeled nucleotide comigrating with UMP retained about the same intensity in both high or low UTP (Fig. 3). These results show that labeled UMP was not integrated into the nascent RNA chain as [α-32P]UTP. Furthermore, this observation indicates that there is no detectable level of CTP deaminase activity in plant mitochondria. RNA editing activity was thus detected in both intact and lysed mitochondria by a coupled transcription/RNA editing system.
      Figure thumbnail gr3
      Figure 3:RNA editing in lysed pea mitochondria. Panel A shows the one-dimensional separation of NMPs from pea mitochondrial run-on transcripts. Mitochondria of pea were supplied with [α-32P]CTP in an osmotically lysed system. Different concentrations of UTP (low UTP, 0.5 mM; high UTP, 5 mM) were used for competition to exclude potential C to U alterations on the NTP level. The mobility of the labeled UMP was determined by separation of a hydrolyzed [α-32P]UTP-labeled in vitro transcript (lane UMP). As control, labeled UMP was added to the nuclease P1-digested products of the 60-min reaction and confirms the mobility of the UMP (pU). Panel B shows the two-dimensional separation of the NMPs from pea mitochondria. The mobility of the products of the one-dimensional TLC separation from the run-on transcription/RNA editing assay was confirmed by a two-dimensional separation. The UMP spots are indicated by arrows. In a control separation the UMP spot was verified by addition of labeled UMP (Control (30 min + UMP*)).

      Incubation of in Vitro Transcripts with Pea Mitochondrial Lysates

      To investigate the editing activity in mitochondrial lysates we incubated [α-32P]CTP-labeled in vitro transcripts of the cox2 and orf206 genes with lysates from pea mitochondria. The cox2 template (Fig. 4) selected for in vitro transcription contains 11 in vivo editing sites (
      • Covello P.S.
      • Gray M.W.
      ) and was amplified by PCR from mtDNA of pea. The labeled in vitro synthesized cox2 transcripts were incubated with S100 supernatant and extracts obtained after DEAE anion-exchange chromatography. After incubation the in vitro transcripts were reisolated, and the monophosphate nucleotide composition was analyzed by TLC (Fig. 4). A small amount of radiolabeled UMP (pU) was visible in the S100 incubation while no pU could be detected in a heat-treated S100 fraction. Heat inactivation thus appears to completely abolish the editing activity. A strong pU signal was observed after incubation of the cox2 transcript with the DEAE fraction.
      Figure thumbnail gr4
      Figure 4:In vitro transcripts are edited in a pea mitochondrial lysate. Panel A, in vitro transcripts of the cox2 gene were incubated with pea mitochondrial lysates. A region of the cox2(128-603) from pea mitochondria, which is edited at 11 positions(
      • Covello P.S.
      • Gray M.W.
      ), was amplified from pea mtDNA. The PCR product was used for in vitro transcription from the introduced T7 promoter to produce the template for the editing assay. Panel B, one-dimensional separation of NMPs from the cox2 transcripts incubated with mitochondrial lysates of pea mitochondria. Cox2 in vitro transcripts labeled with [α-32P]CTP were incubated for 30 min with heat-inactivated S100 (template (CMP*)), S100, and a lysate after DEAE chromatography from pea mitochondria. In the control lane labeled UMP (pU) was added to the DEAE hydrolysis products to confirm the mobility of the UMP spot. The mobility of the pU spot is consistent with the spot of a nuclease P1-digested in vitro transcript that was labeled with [α-32P]UTP. C, two-dimensional separation of the NMPs from the cox2 transcript incubated with mitochondrial lysates. After nuclease P1 digestion of the cox2 transcripts incubated with mitochondrial lysates the NMPs were separated on two-dimensional TLC. The spots comigrating with UMP are indicated by arrows. The mobility of the NMPs (pA, pC, pU, pG) was visualized by UV shadowing, and the relative mobility of the NMP spots is given schematically on the right.

      Evidence for Cytidine Deamination

      To follow the fate of the cytosine base in order to distinguish between transglycosylation and deamination we used [5ʹ-3H]CTP to label in vitro transcripts. After incubation with pea mitochondrial lysates we determined the amount of radioactivity appearing in the UMP spot after TLC separation (Fig. 5). No label is expected in the UMP spot if the CMP to UMP conversion occurs by a transglycosylation reaction, while [5ʹ-3H]UMP is the expected product of a deamination process. Although [5ʹ-3H]CTP is available at a much lower specific activity than [α-32P]CTP and detection is more difficult, we did observe a significant increase of radioactivity in the UMP spot after incubation of 3H-labeled transcripts with mitochondrial lysates for 60 min at 30°C (Fig. 5). The radioactivity of the CMP spot was determined in independent experiments with 2,360 and 3,800 cpm, respectively. The UMP spot of the incubated transcript was labeled with 140 and 170 cpm. Only background radioactivity (20 ± 5 cpm) could be measured in UMP spots not incubated with mitochondrial lysate. About 10% of the cytidines represent editing sites in the cox2 in vitro transcript, and accordingly only between 236 and 380 cpm are expected for a fully edited transcript. From these data we conclude that only about 30-40% of the editing sites have been changed from C to U in this system.
      Figure thumbnail gr5
      Figure 5:Ring-labeled CMP is modified to labeled UMP during incubation with mitochondrial lysates. T7 in vitro transcripts of the cox2 gene (A) labeled with [5-3H]CTP were incubated with pea mitochondrial extracts for 60 min. After incubation with mitochondrial lysate transcripts were reisolated and hydrolyzed, and the resulting NMPs were separated by TLC. After separation the UMP and CMP spots were identified by UV shadowing and removed from the plate, and the activity of the spots was determined. Only background counts were detectable in the UMP spot that was not incubated with mitochondrial lysate (-lysate), while incubation with the lysate (+lysate) revealed significant radiolabeling of the UMP spot.

      Edited Transcripts Are Not Modified

      To investigate whether the in vitro RNA editing system works at specific cytidines we compared edited with unedited templates of orf206 after incubation with pea mitochondrial lysates. The orf206 gene was selected because it is one of the most highly edited genes detected to date in plant mitochondria(
      • Schuster W.
      ). The unedited in vitro transcript covers 11 editing sites that are replaced by uridines in the edited RNA (Fig. 6). After incubation with mitochondrial pea lysates for 30 min the incubated in vitro transcripts were hydrolyzed and the resulting NMPs were separated by TLC. Radiolabeled UMP was detected only from unedited templates indicating that only specific editing sites have been modified. Incubation of the edited template revealed no label comigrating with UMP although 25 32P-labeled cytidines are still present in the polynucleotide chain.
      Figure thumbnail gr6
      Figure 6:RNA editing modifies only unedited transcripts in the in vitro system. To analyze the specificity of the in vitro RNA editing system transcripts of orf206 were incubated either in the edited or in the unedited version together with pea mitochondrial lysates. After incubation of 30 min the transcripts were reisolated and digested with nuclease P1 to obtain the NMPs for TLC separation. The unedited (orf206 genomic) template contains 36 and the edited transcript (orf206 edited) 25 labeled cytidines, respectively. In the TLC separation of the edited transcript (orf206 edited) two times more radioactivity was loaded to show that no detectable amount of UMP was produced during incubation. An arrow indicates the UMP position, where only the unedited transcript shows a spot comigrating with UMP after incubation.

      In Vitro Transcripts of orf206 Are Edited Correctly by the Mitochondrial Lysate

      After incubation of T7-transcribed in vitro transcripts of unedited orf206 first strand cDNA was synthesized from a T3 primer. The PCR product obtained with the T7-orf206- and T3-primers was cloned, and 20 cDNA clones were sequenced (Fig. 7). No PCR product was obtained with either the lysate alone or with pea mtDNA. None of the cDNA clones was fully edited; however, 6 out of 20 clones analyzed showed C to U transitions indicating RNA editing at some of the expected cytidine positions. One clone showed 7 out of the 8 expected C to U changes of the fully edited orf206 transcript. Two clones showed 3 changes, one clone was edited at two sites, and two clones were modified at only a single site. In the 20 cDNA clones sequenced no other modification was observed. The editing pattern of the in vitro edited transcripts indicates that RNA editing occurs at individual sites independently. From this observation it is tempting to speculate that site-specific factors may be involved for individual sites.
      Figure thumbnail gr7
      Figure 7:In vitro transcripts are edited at specific editing sites. A, in vitro transcripts of orf206 were incubated with pea mitochondrial lysate as shown in . After reisolation of the template first strand cDNA was synthesized primed by T3-primer. PCR amplification was done using the T7-orf206 and the T3 primers. B, PCR amplification products are shown from lysate without template(
      • Covello P.S.
      • Gray M.W.
      ), from incubated orf206 template with mitochondrial lysate(
      • Gualberto J.M.
      • Lamattina L.
      • Bonnard G.
      • Weil J.-H.
      • Grienenberger J.-M.
      ), from unincubated orf206 template (
      • Hiesel R.
      • Wissinger B.
      • Schuster W.
      • Brennicke A.
      ), and from mtDNA of pea(
      • Schuster W.
      • Ternes R.
      • Knoop V.
      • Hiesel R.
      • Wissinger B.
      • Brennicke A.
      ). bp, base pairs. Panel C shows the orf206 sequence analyzed in the RT-PCR reaction. Stars indicate the editing sites observed in fully edited orf206 transcripts. PCR products obtained from the unincubated template showed only the genomic sequence of orf206. Sequences from cDNA clones(
      • Covello P.S.
      • Gray M.W.
      ,
      • Gualberto J.M.
      • Lamattina L.
      • Bonnard G.
      • Weil J.-H.
      • Grienenberger J.-M.
      ,
      • Hiesel R.
      • Wissinger B.
      • Schuster W.
      • Brennicke A.
      ,
      • Schuster W.
      • Ternes R.
      • Knoop V.
      • Hiesel R.
      • Wissinger B.
      • Brennicke A.
      ,
      • Covello P.S.
      • Gray M.W.
      ,
      • Schuster W.
      ,
      • Benne R.
      • van den Burg J.
      • Brakenhoff J.P.J.
      • Sloof P.
      • van Boom J.H.
      • Tromp M.C.
      ,
      • Chen S.-H.
      • Habib G.
      • Yang C.Y.
      • Gu Z.W.
      • Lee B.R.
      • Weng S.A.
      • Silberman S.R.
      • Cai S.J.
      • Deslypere J.P.
      • Rosseneu M.
      • Gotto Jr., A.M.
      • Li W.H.
      • Chan L.
      ,
      • Blum B.
      • Bakalara N.
      • Simpson L.
      ,
      • Blum B.
      • Sturm N.
      • Simpson A.M.
      • Simpson L.
      ,
      • Sommer B.
      • Khler M.
      • Sprengel R.
      • Seeburg P.H.
      ,
      • Higuchi M.
      • Single F.N.
      • Khler M.
      • Sommer B.
      • Sprengel R.
      • Seeburg P.H.
      ,
      • Navaratnam N.
      • Morrison J.R.
      • Bhattacharya S.
      • Patel D.
      • Funabashi T.
      • Giannoni F.
      • Teng B.-B.
      • Davidson N.O.
      • Scott J.
      ,
      • Araya A.
      • Domec C.
      • Begu D.
      • Litvak S.
      ,
      • Finnegan P.M.
      • Brown G.G.
      ,
      • Nishimura S.
      ,
      • Covello P.S.
      • Gray M.W.
      ,
      • Hodges P.E.
      • Navaratnam N.
      • Greeve J.C.
      • Scott J.
      ,
      • Mulligan R.M.
      • Leon P.
      • Walbot V.
      ,
      • Rajasekhar V.R.
      • Mulligan R.M.
      ) of incubated templates are given under the listed nucleotide sequence. Identical nucleotides are indicated by dots.

      DISCUSSION

      To analyze the biochemical nature of the RNA editing reaction in plant mitochondria we followed the fate of the α-phosphate of cytidine nucleotides and the labeled cytosine base to distinguish between the possible reaction mechanisms. The experiments show that intact and lysed mitochondrial systems are useful tools to study organellar run-on transcription (
      • Finnegan P.M.
      • Brown G.G.
      ,
      • Rajasekhar V.R.
      • Mulligan R.M.
      ) and RNA editing in plants (
      • Mulligan R.M.
      • Leon P.
      • Walbot V.
      ). The results obtained by in organello and in vitro systems show that in the RNA editing process the production of uridine from cytidine occurs at the polynucleotide level. In coupled run-on transcription/RNA editing systems of higher plants the product of the RNA editing process in mitochondria was determined as a genuine uridine. Furthermore, incubation of in vitro transcripts with mitochondrial lysates showed that the editing process is not directly linked to transcription in mitochondria.
      During the process of RNA editing the 5ʹ-phosphate of the modified cytidine is maintained, which is consistent with either a deamination/transamination or a transglycosylation process. The replacement of cytidine in the polynucleotide chain by uridine by a deletion and insertion process can now be excluded, because no enzymatic insertion mechanism is known that can introduce a nucleotide into a polynucleotide chain without insertion of the 5ʹ-phosphate of the donor. In a transglycosylation reaction the phosphoribosyl chain is maintained, but the base is exchanged. This mechanism is used to post-transcriptionally modify preformed tRNAs, e.g. by the introduction of hypoxanthine (
      • Elliott M.S.
      • Trewyn R.W.
      ) and queosine(
      • Okada N.
      • Nishimura S.
      ).
      A straightforward hydrolytic deamination appears to be most likely the reaction in plant mitochondrial C to U RNA editing because following a 3H-labeled cytosine base led to labeled UMP. The occasional reverse editings from uridine to cytidine(
      • Schuster W.
      • Hiesel R.
      • Wissinger B.
      • Brennicke A.
      ), however, cannot as easily be explained by a straightforward reverse deamination process. Either the reverse U to C editing is catalyzed by a different mechanism, e.g. by a CTP synthase, or the deamination process involves the transfer of the released amine to another molecule in a transamination reaction.
      A site-specific deaminase catalyzes the conversion from C to U in the apolipoprotein B mRNA editing(
      • Navaratnam N.
      • Morrison J.R.
      • Bhattacharya S.
      • Patel D.
      • Funabashi T.
      • Giannoni F.
      • Teng B.-B.
      • Davidson N.O.
      • Scott J.
      ). The catalytic subunit of this type of RNA editing shares high homology to cytidine deaminases, while the target site in the mRNA is determined by a second additional factor that is proposed to bind downstream of the editing site at a conserved sequence motif. A completely unrelated deamination activity, selective for double-stranded RNAs, converts adenosyl residues to inosines in a process termed double strand RNA unwinding(
      • Bass B.L.
      • Weintraub H.
      ). This enzymatic activity seems to be also involved in the editing process of the mammalian neuronal glutamate-gated ion channel mRNA, where a CIG codon is created from CAG(
      • Higuchi M.
      • Single F.N.
      • Khler M.
      • Sommer B.
      • Sprengel R.
      • Seeburg P.H.
      ).
      In wheat mitochondria an in vitro RNA editing system has been reported (
      • Araya A.
      • Domec C.
      • Begu D.
      • Litvak S.
      ) that edits atp9 transcripts and requires no exogenous nucleotides. This observation supports a deamination reaction for RNA editing in plant mitochondria, although it has to be taken into account that in this lysate endogenous nucleotides have not been depleted. Another recent report, also based on a similar experimental approach, likewise suggests the conversion of cytidine to uridine in intact mitochondria of maize and Petunia by either a deamination or transglycosylation mechanism(
      • Rajasekhar V.R.
      • Mulligan R.M.
      ). To distinguish between deamination and transglycosylation we followed the fate of a 3H-labeled cytosine base, which led to 3H-labeled UMP. These results are consistent with a deamination reaction because in a transglycosylation reaction unlabeled UMP is expected as product.
      To determine the specificity of the in vitro reaction edited and unedited templates were tested in the mitochondrial lysate and analyzed by RT-PCR after incubation. The data show that the in vitro RNA editing system of pea mitochondria described here is specific for cytidines at editing positions while other cytidines in the polynucleotide chain are no targets for C to U RNA editing. How site selection is determined is still enigmatic, while the RNA editing patterns observed in in vitro incubated templates clearly indicate the involvement of site-specific factor(s) for individual editing sites.

      Acknowledgments

      We are grateful to Dr. Axel Brennicke for advice throughout this study and for helpful discussions. We thank Dr. Charles Andr and Dr. Hugo Sanchez for critical reading of the manuscript. We also thank Waltraut Jekabsons and Iris Gruska for excellent technical assistance.

      REFERENCES

        • Covello P.S.
        • Gray M.W.
        Nature. 1989; 341: 662-666
        • Gualberto J.M.
        • Lamattina L.
        • Bonnard G.
        • Weil J.-H.
        • Grienenberger J.-M.
        Nature. 1989; 341: 660-662
        • Hiesel R.
        • Wissinger B.
        • Schuster W.
        • Brennicke A.
        Science. 1989; 246: 1632-1634
        • Schuster W.
        • Ternes R.
        • Knoop V.
        • Hiesel R.
        • Wissinger B.
        • Brennicke A.
        Curr. Genet. 1991; 20: 397-404
        • Covello P.S.
        • Gray M.W.
        FEBS Lett. 1990; 268: 5-7
        • Schuster W.
        Plant Mol. Biol. 1994; 25: 33-42
        • Benne R.
        • van den Burg J.
        • Brakenhoff J.P.J.
        • Sloof P.
        • van Boom J.H.
        • Tromp M.C.
        Cell. 1986; 46: 819-826
        • Chen S.-H.
        • Habib G.
        • Yang C.Y.
        • Gu Z.W.
        • Lee B.R.
        • Weng S.A.
        • Silberman S.R.
        • Cai S.J.
        • Deslypere J.P.
        • Rosseneu M.
        • Gotto Jr., A.M.
        • Li W.H.
        • Chan L.
        Science. 1987; 238: 363-366
        • Blum B.
        • Bakalara N.
        • Simpson L.
        Cell. 1990; 60: 189-198
        • Blum B.
        • Sturm N.
        • Simpson A.M.
        • Simpson L.
        Cell. 1991; 65: 543-550
        • Sommer B.
        • Khler M.
        • Sprengel R.
        • Seeburg P.H.
        Cell. 1991; 67: 11-19
        • Higuchi M.
        • Single F.N.
        • Khler M.
        • Sommer B.
        • Sprengel R.
        • Seeburg P.H.
        Cell. 1993; 75: 1361-1370
        • Navaratnam N.
        • Morrison J.R.
        • Bhattacharya S.
        • Patel D.
        • Funabashi T.
        • Giannoni F.
        • Teng B.-B.
        • Davidson N.O.
        • Scott J.
        J. Biol. Chem. 1993; 268: 20709-20712
        • Araya A.
        • Domec C.
        • Begu D.
        • Litvak S.
        Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 1040-1044
        • Finnegan P.M.
        • Brown G.G.
        Plant Physiol. 1987; 85: 304-309
        • Nishimura S.
        Prog. Nucleic Acids Res. Mol. Biol. 1972; 12: 49-85
        • Covello P.S.
        • Gray M.W.
        Nucleic Acids Res. 1990; 18: 5189-5196
        • Hodges P.E.
        • Navaratnam N.
        • Greeve J.C.
        • Scott J.
        Nucleic Acids Res. 1991; 19: 1197-1201
        • Mulligan R.M.
        • Leon P.
        • Walbot V.
        Mol. Cell. Biol. 1991; 11: 533-543
        • Rajasekhar V.R.
        • Mulligan R.M.
        Plant Cell. 1993; 5: 1843-1852
        • Elliott M.S.
        • Trewyn R.W.
        J. Biol. Chem. 1984; 259: 2407-2410
        • Okada N.
        • Nishimura S.
        J. Biol. Chem. 1979; 254: 3061-3066
        • Schuster W.
        • Hiesel R.
        • Wissinger B.
        • Brennicke A.
        Mol. Cell. Biol. 1990; 10: 2428-2431
        • Bass B.L.
        • Weintraub H.
        Cell. 1988; 55: 1089-1098