Nucleic Acids Research, 1994, Vol. 22, No. 11 2109-2113

RNA editing of transcripts of a chimeric mitochondrialgene associated with cytoplasmic male-sterility in Brassica

Richard Stahl, Sophie Sun, Yvan L'Homme, Troy Ketela and Gregory G.Brown*Department of Biology, McGiII University, 1205 Doctor Penfield Avenue, Montreal, Quebec H3A 1Bi,Canada

Received February 3, 1994; Revised and Accepted April 26, 1994

ABSTRACTThe orf224 gene is a chimeric open reading frameassociated with the Polima or pol cytoplasmic malesterility of Brassica napus. The first 58 codons and 5'upstream region of orf224 are derived from aconventional mitochondrial gene, orfB, while the originof the remaining portion of the gene is unknown.Trancripts of the orf224 gene were found to be editedat a single site in the region of the gene that does notcorrespond to a known sequence. Oligonucleotidescorresponding to the edited and unedited forms wereshown to hybridize specifically to respective in vitroorf224 transcripts. Analysis of floral mtRNA by thismethod indicated that virtually all orf224 transcripts ofboth sterile and fertile, nuclear-restored pol cytoplasmplants are edited. Our results indicate that transcriptsof novel, CMS-associated genes may be edited, butthat, at least in this case, the degree of editing doesnot appear to be directly related to the male-sterilephenotype.

INTRODUCTIONRNA editing refers to the alteration of the sequence of a transcriptso that it differs from its template DNA. RNA editing occursin the mitochondria of a number of different types of organisms,in chloroplasts and in mammalian nuclei, but the types of editingmodifications differ among the various systems. In themitochondria of flowering plants (1-3), RNA editing involvesthe conversion of C to U and rarely, U to C, residues. The editingis highly specific: transcripts of different protein coding genesare edited to varying degrees and the transcripts of a given genemay be edited at only one or a few sites in one species and atmany sites in another. Editing outside of protein coding regionsis infrequent; for example, rRNA transcripts are edited only veryrarely, if at all. Editing tends to alter RNA sequences such thatthe encoded amino acid sequences become more similar to thethose of the corresponding proteins from other organisms.Transcripts of virtually all the standard, functional, protein-codingmitochondrial genes examined have been found to be edited, andtranscripts of non-functional pseudogenes may also be edited(4, 5).

Cytoplasmic male sterility (CMS), the cytoplasmically inheritedinability to produce functional pollen, is a common plant traitthat results from rearrangements of the mitochondrial genome(6, 7). In general, the regions of the mitochondrial genome thatappear to be responsible for the trait contain unusual, chimericgenes that are often co-transcribed with standard mitochondrialgenes. For many instances of CMS, specific nuclear genes,termed restorers of fertility (RJ) have been identified that cansuppress the male sterile phenotype. In some cases it has beenshown that Rf genes alter the expression of CMS-associatedchimeric mitochondrial gene regions.

Several recent studies have examined the role ofRNA editingin CMS. Ward et al. showed that the rearranged gene associatedwith cms-T of maize, T-urf13, is not edited in either CMS orfertility-restored plants (8), although the co-transcribed orJ221(orJ2S), a standard mitochondrial gene, is edited at 12 sites. Bycontrast, Iwabuchi and colleagues (9) have shown that one oftwo atp6 genes on a male-sterile rice mitochondrial genome isassociated with CMS and that the unprocessed transcripts of thisgene found in CMS plants are less highly edited than theprocessed transcripts found in restored plants. They postulatedthat aberrant proteins, translated from the incompletely editedatp6 transcripts of CMS plants, might be responsible for male-sterility. The observation that the targeting of the protein productof an unedited wheat atp9 transgene to Nicotiana mitochondriacan induce male-sterility (10) is consistent with the premise thatCMS may be related to editing defects.The atp6 mitochondrial gene region has also been found to

be associated with the Polima or pol CMS system of Brassicanapus (11-13). In this case the atp6 gene is situated downstreamof and co-transcribed with a chimeric mitochondrial genedesignated orJ224 (11) or pol-urf (13). The N-terminal codingand 5' non-coding regions of this chimeric gene are derived froma normal mitochondrial gene, orJB, while the origin of theremaining portion is unknown. The B.napus atp6 gene is editedat a single site in plants with fertile cytoplasm, inpol CMS plantsand in nuclear-restored pol cytoplasm plants (14; Stahl and Brownunpublished results). In contrast to rice atp6, no differences havebeen found in the extent ofatp6 editing between fertile and sterileBrassica plants. In this report we show that, unlike the chimericgene associated with cms-T in maize, transcripts of the chimeric,

*To whom correspondence should be addressed

2110 Nucleic Acids Research, 1994, Vol. 22, No. 11

CMS-associated Brassica orf224 gene are edited. Editing occursat only a single site, and virtually all orJ224 transcripts of CMSand fertility-restored plants are edited.

MATERIALS AND METHODSPlant materialThe Brassica napus strains used in this study were of the Westarcultivar as detailed earlier (11). Normal fertile tissue was obtainedfrom the strain containing nap cytoplasm, whereas thecytoplasmic male sterile (CMS) and fertility restored (RJ) tissueswere of thepol cytoplasmic type. Plants were greenhouse grownto maturity and healthy flowers were harvested for the isolationof mitochondrial nucleic acids.

Nucleic acid isolation and analysisFloral mitochondria were purified by the methods of Perez etal. (15) as modified by Singh and Brown (11). Mitochondria werelysed in a solution containing 1% Sarkosyl, 50mM Tris -HCI,pH 8, lOmM EDTA and the nucleic acids were phenol extractedand precipitated with ethanol. Isolated mtRNA waselectrophoresed on 1% agarose + 5M urea gels (16) andtransferred overnight by capillary blotting onto GeneScreen-Plusmembranes (New England Nuclear) with 1OxSSC (1.5M NaCl,0. 15M sodium-citrate). Hybridizations were performed in5 xSSPE (0.9M NaCl, 50mM sodium phosphate, 5mM EDTA),5 x Denhardt's solution (0.1 % Ficoll, 0.1 % polyvinlylpyrolidone, 0.1% bovine serum albumin) and 0.5% sodiumdodecyl sulfate (SDS) at 51°C for 14-16 hrs in a Hybaid oven.Probes for hybridization were end-labelled with [Ly-32P]ATP(Amersham) using T4 polynucleotide kinase (New EnglandBiolabs). Oligonucleotides for selective probing were designedas follows; 5'GATTCAAGTCGGTTCGAG 3', for edited or1224and 5' GATTCAAGTTGGTTCGAG 3', for unedited orJ224.Membranes were washed twice in 2 x SSC, 0.1% SDS at roomtemperature for 10 min and twice in 1xSSC, 0.1% SDS at 55°Cfor 15 min prior to autoradiography on Kodak X-omat-AR filmfor 4-12 days. RNA for control blots was synthesized usingT7 RNA polymerase (Stratagene) and linearized plasmidcontaining either the edited or unedited orJ224 sequence.

cDNA synthesis and amplificationTotal mitochondrial nucleic acids frompol-CMS andpol-Rfplants(isolated as described above) were treated with 3U (units)/4gRNA RNase-free DNaseI (Pharmacia) at 370C for 30 min.Samples were then extracted twice with phenol and precipitatedin ethanol with l,ug glycogen as carrier. First strand DNAsynthesis was initiated from either the atp6 oligonucleotide primer# 4 (5' GCCCATGTCTTTGAAATCCC 3') or the orJ2241atp6intergenic primer # 3b (5'CGACGGGGAGAGTGATGCAG 3').Approximately 50ng of primer were annealed to 0.5-1.0 ,ugmtRNA in diethlypyrocarbonate-treated H20 by cooling from95°C (5 min) to 500C (ramp - °C/20 sec) followed by 10 minat 50°C. Primer extension was performed at 41 -42°C for 1 hrwith 16U MoMLV reverse transcriptase (Pharmacia) using thesupplied buffer and lmM deoxynucleotide triphosphates.Polymerase chain reaction (PCR) amplifications were performedusing primer # 14 (5'GGATGCTACTTCAGGTAG 3'), whichis positioned upstream of orJ224, to synthesize cDNA's of 1.6and 1.0 kb in length. The locations of amplified sequences inthe pol orJ224/atp6 gene region are indicated in Fig. 1. Reaction

components for PCR included 50ng of each oligo, 2U Taq DNApolymerase (Bio/Can), 1 x supplied buffer, 2mM MgCl2,1mg/ml BSA and 200 itM dNTP's. Amplifications wereconducted with a Hybaid Omnigene thermal cycler, with a typicalprogram consisting of 35 cycles of 94°C (30 sec), 50°C (30 sec)and 72°C (2 min) followed by 72°C (10 min). Amplificationproducts were purified from agarose gels using the GenecleanII kit (Bio 101) and cloned into a pBluescript II KS+ (Stratagene)phagemid vector that had been blunt-cut by EcoRV and T-tailedaccording Marchuk et al. (17).

Mitochondrial DNA analysisThe total mitochondrial nucleic acid preparations were used fordirect amplification by PCR and subsequent cloning. Primers andreaction conditions were identical to those used for cDNAmanipulations.

Sequence analysisThe dideoxy chain termination method was employed using theT7 sequencing kit (Pharmacia). Single-stranded DNA templateswere generated according to the suppliers' recommendations.Helper phage R408 and E.coli strain XL-1 Blue were bothpurchased from Stratagene. Sequencing primers includedStratagene T7 and SK primers, in addition to those that weredesigned based on the genomic mtDNA sequence (11) andpurchased from Sheldon Biotechnology Center (McGill Univ.).

RESULTSSynthesis and cloning of orf224 cDNAsThe organization and expression of the Brassica pol orJ224Iatp6gene region, as deduced from studies conducted in this laboratory(7, 11, 18) are indicated in Fig. 1. In floral tissue of pol CMSplants, transcripts of the region are predominandy dicistronic.Upon nuclear fertility restoration, abundant transcripts withtermini mapping within the orf224 gene appear, and the levelof the smallest atp6 transcript, the 5' terminus of which mapsto a tRNA pseudogene located at the 3' end of orJ224, increases.To examine editing of orJ224 transcripts, cDNAs spanning the

orJ224 coding sequence were first synthesized using DNasetreated RNA preparations from CMS and fertility-restored polcytoplasm plants. These were then amplified by the polymerasechain reaction, cloned and sequenced. Two differentoligonucleotides were used to prime cDNA synthesis. One primermapped to a sequence located downstream of the or1224termination codon (primer # 3b), while the other (primer # 4)mapped downstream of the single editing site in the atp6 gene(indicated by the asterisk in Fig. 1). Both types of cDNAs wereamplified using primer # 14, which maps 264 nt upstream oforJ224, in the region between the 5' termini of the 2.2 and 1.9kb transcripts (18). The amplification products derived fromcDNAs primed with oligonucleotides # 3b and #4 were 1.0 and1.6 kb respectively. The mtDNA sequences from which thecorresponding transcripts were derived are indicated in Fig. 1.To ensure that the amplification products were derived fromcDNAs and not mitochondrial genomic sequences, controlamplification reactions were performed on RNA preparations inwhich the reverse transcriptase had not been added during thecDNA synthesis step. Only the products amplified frompreparations in which no product was generated from the controlwere analyzed further.

Nucleic Acids Research, 1994, Vol. 22, No. 11 2111

Transcripts of orf224 are edited at a single siteA total of six cDNA clones from restored plants (four 1.0 kband two 1.6 kb clones) and five 1.6 kb cDNA clones from CMSplants were analyzed. Sequences of all the or1224 cDNAs fromboth sterile and restored plants differed from the genomic or1224sequence at a single site, located 211 nt downstream of the orJ224initiation codon (Figs. 2 and 3). This nucleotide was C inmitochondrial genomic DNA, but T in the cDNA clones. Noother consistent differences were noted between any of the cDNAclones and the genomic sequence nor among the cDNAsthemselves. Although the sequences of most of the cDNA clonesvaried from the genomic sequence at one or more positions inaddition to nt 211, in each case this variation was specific to asingle clone and thus appeared to result from errors introducedduring amplification. A sequence reported for the region upstreamof the pol mtDNA atp6 gene (13) differs somewhat from thatgenerated by this laboratory (11) in that it contains a 105 codonrather than a 224 codon open reading frame. Resequencing ofthe original mtDNA clone and cloned amplification productsderived from the mitochondrial genomic DNA confirmed ourpreviously published sequence and indicated, in particular, thatthe mtDNA nucleotide at position 211 was C (Fig. 3). Theseanalyses showed that the or1224 gene is, indeed, 224 codons inlength and suggested that editing ofor224 transcripts occurs ata single site, just downstream of the region of orfl similarity.Editing results in the conversion of a CGG (Arg) to a UGG (Trp)codon. No evidence for editing in the non-coding regionsupstream or downstream of or1224 was found.

Functional plant mitochondrial genes are almost always editedat more than one site. One of the few exceptions to this rule,is, coincidentally, the Brassica atp6 gene co-transcribed withor224, which is edited at a single site, 103 nt downstream ofits initiation codon. The absence of editing events over an

extended portion of orJ224, together with the finding thattranscripts of another CMS associated chimeric gene, the maizeT-urf13 gene, are not edited, raised the possibility that the singlecommon nucleotide change observed among the cDNA clones

rfpRfp

orf224

orB S'flank Wtny atp6ng I X. I

#14 * #3b #4

14 kb c

1.6 kb cDNA

1.0 kb cDNA-~200 bp

Figure 1. Organization, expression and generated cDNA's of the orf224/atp6gene loci ofBrassicapol mtDNA. The upstream and coding region derived fromocrB is indicated by an open box. The diagonally striped box represents the portionof the orJ224 chimeric gene of unknown origin. The cotranscribed atp6 geneis shown by the large solid box. The region of DNA showing sequence similarityto the radishinitiator-methionine tRNA gene is designated as OtrnfM. The locationof the RNA editing sites are marked by an asterisk (*). Transcripts of male sterileand fertility-restored pol cytoplasm plants are labelled r and Rfp respectively.Small and open arrowheads indicate 5' and 3' termini, respectively. cDNA's ofindicated length represent cloned and sequenced PCR products. The positionsof the PCR primers are marked by the respective primer symbols # 14, # 3band #4.

might have been generated through an experimental artifact. The1.6 kb cDNA clones were expected to span the atp6 editing site.To further rule out the possibility that the or1224 sequencesanalyzed were generated from mitochondrial genomic DNA andnot RNA, the region of the 1.6 kb clones in the vicinity of theatp6 editing site were sequenced. All the clones from both sterileplants and fertile plants were found to contain the sequenceexpected of a transcript edited at both the atp6 and or1224 sites(Fig. 2). Since it is very unlikely that the same nucleotide changeswere artifactually generated co-incidentally at both editing sitesin all 7 clones, this result indicated that the difference in sequenceat position 211 of the orJ224 cDNA clones occurred as a resultof RNA editing.

Editing of orf224 transcripts is virtually complete in sterileand fertile plantsThe finding that all cDNA clones analyzed corresponded to anedited sequence suggested that transcripts of or1224 are editedhighly efficiently. Transcripts of various lengths spanning orJ224are found in both pol CMS and restored plants, and it was ofinterest to determine if a sub-population of these transcripts might

GGATGCTACTTCAGGTAGGGTACGGGCGCTCTATCATTGTCTGATTTTAGGTTT -211

CTGATCGCrTAIGCCCTICCGGGCTGCCCCCGCGATCAAACTATCAATCTCATAAGAIGAAGAA -151

ATCTCTATIGCCCCCTGTTCTTGGTTTTCTCCCATGCTTTTGTTGIGTCAACAACCAACCAC -91

AACTTTCTATAGTTCTTCACTACTCCTAGA GCTTGACQGAGTGAAGCTGTCTG6AGGIIA -31

ATCATTTTGTTGAAATCAATT TTAATCATGCC:TCAAX:TAACTGGATAAATTCC ACTTATTTT 30..: .....

TCACAATTCTrCTG&TT7ArGCCTTTCTTC;TTTACT TTTATATTfTCATATGC AATAT 90S0QF F WL.tC F FF TFtYI FlI ON D

::::::: .:.:.: T T Y

GGAGAT1IGAGTACTTGG&GATCAIICAGAATTCTAAACTATGGAACCAACT CTTTCACAC 150GOD G V LIG I S R I L K L W N0Q LL.S H

CCGGGGGAAGAcCtTCcTGAGCAAGGGAAGGCTTGGAAAAAATCGIA&TICA&ATICAAGT 210R G K T LI S K G R L G K N R S SODS S

RTCGAGGTATCAGCGTTGGCCGCCCATTATTTTATCATTTTCGTGGTCCCAAAATTG 270I5 F E VS A LA A H Y F I I F V V P K L

GGACCAGTTTTCTACATTATATATAATTTITITITGTTTGTIGGGGIIGAAAIGGGGGGIA 330G P V F Y11 Y N F F C L L G L KIW G V

IIAGGAAATGAAATTTGTCATTTCGGCGTCGGACCAGAIGGCGICGCGCCCCCAGCGCTG 390OL G N ElI C H F G V G P DIGV AP PA L

GATCTCAACGAGCGCCCGCCITCITGCATCTTTTGTACGCGGATGTTGAGA&TTCCGACTCT 450OL N E R P P L H L L YA DV E S SODS

CAACAAGCGCGAAATAATGACATGIACGCGCATCTTAGGCGCGTACAGGAGATCACCCAA 510OGQA R N N DIM YAH L R RV FEI T 0

AAACIAGAGGGIGAGCGCGATATCGIGCGGC8ICAA&CCCTCCIGG&AIAIAAI&AAATGG 570K L EG FRID I V R ROQA L L D IM K W

GAGGICAGAAGCCTTCAGGAGCACTITCGGATCTTTCGGCACCTTGAICGTCTGCGAGAT 630F V R S LO F H F RI F R H LID R L R D

TCGCAGAGAGCCAAGGIGAACGAGATCCTTGATCICTIICGCIGAAGAAGCTA&GGITCCT 690SO0RA KV NFEI LDL FR-

CCTTGATGTCGTAGGTTCAAATCCTATCTCCGCACTAAGTAAGGGTTTCATTCTGCATCACTCICCCCCGTCG

77507762

Figure 2. Nucleotide consensus sequence of the orJ224 gene loci present in Brassicapol-CMS andpol-Rf plants as derived from11 cDNA clones. The published regionspans the1026bp cDNA clone and includes the predicted 224 amino acid residues.The underlined base contained in a box represents the single C to U edit siteresulting in a codon conversion of an arginine (R) to atuyptophan (W). The shadedregion represents sequence spanning 388 bases and 58 amino acids that has 99%nucleotide sequence homology to Brassica napus orfB. The underlined stretchesof sequence at the 5' and 3' ends indicate the location of the PCR amplificationprimers, #14 and # 3b respectively.

2112 Nucleic Acids Research, 1994, Vol. 22, No. 11

ORF224Genomic cDNAACGT ACGT

;

=

- s

c it'T :._.-G _G 4ST _ * _

A/

ATP6

C

/ T1TGGT *T

A

GenomicA C G T

ACG

AA

cDNAA C G T

A

U- _Am 40 A

_ _

_. \- Anew NW\_

-"

Figure 3. Comparison of the mitochondrial genomic DNA and cDNA sequencesin the edited region of orJ224 and atp6 in Brassica pol. For each cloned gene,the left panel shows part of the mtDNA sequence and the right panel thecorresponding cDNA sequence. The edited nucleotides are labelled by an asterisk(*)

Edited probe

E U N P Rf

Unedited probe

E U N P Rf

* - 22 kb* 3- -1.9kb

in Fig. 4, the oligonucleotide corresponding to the edited formof orJ224 hybridized to transcripts of the expected sizes in floralmtRNA of both pol CMS and restored plants (lanes P and Rf).No hybridization was observed to mtRNA from fertile napcytoplasm plants (lane N). The same oligonucleotide hybridizedefficiently to an in vitro transcript of an orJ224 cDNA clone onthe same gel-blot (Fig. 4, lane E) but did not hybridize detectablyto an in vitro transcript of the genomic orJ224 sequence (laneU), indicating that the signals arising from the 2.2 and 1.9 kbmtRNAs were specifically generated by hybridization to editedtranscripts. These results suggested that the hybridization to theoligonucleotide was sensitive enough to detect the orj224transcripts and specific enough to distinguish edited from non-edited transcripts.

In contrast to the results obtained using the edited-specificoligonucleotide, the oligonucleotide corresponding to the uneditedform did not hybridize detectably to any transcript of nap male-fertile, pol CMS orpol-restored plants under conditions in whicha strong signal was was generated by hybridization to the in vitrotranscript of the mitochondrial genomic sequence. Even whenthe autoradigraph exposure was lengthened by a factor of three,only heterogenous, non-specific background signals could beobserved in the mtRNA lanes. The absence of hybridization ofthe probe to the in vitro transcript of the cDNA clone againindicated that hybridization was specific to unedited transcripts.Collectively the results showed that, within the limits of detection,orJ224 transcripts are fully edited, and that substantial differencesin editing efficiency do not exist between CMS and fertility-restored plants.

9

Figure 4. Autoradiographs of gel blots containing in vitro RNA and mtRNAtranscripts hybridized by oligonucleotides selective for edited and unedited orf224sequence. In vitro orf224 transcripts corresponding to the edited (E) and unedited(U) forms were equally loaded in the first two lanes. MtRNA loaded in the lastthree lanes correspond to fertile nap (N), pol-CMS (P) and pol fertility restored(Rf) plants. Identically loaded gels were blotted and probed with either the editedor unedited 18-mer oligonucleotides (left and right panel, respectively). Equalloading among mtRNAs and transcripts was judged by ethidium bromide staining.Transcript size detected by hybridization signal is indicated in kilobases (kb).Autoradiograph exposure time was 4 days.

be less efficiently edited. In particular, the finding of Iwabuchiand colleagues that rice atp6 transcnpts specific to restored plantsare more efficiently edited than those ofCMS plants (9) suggestedthe possibility that differences in orf224 editing might existbetween sterile and restored plants. To examine these possibilities,we designed oligonucleotides complementary to both edited andunedited versions of the orf224 sequence, and used these to probegel blots of mtRNA from pol CMS and restored plants, as wellas fertile plants with nap cytoplasm.

Previous studies (11, 18) had indicated that sequences in thevicinity of the orf224 editing site would be expected to occuron two major transcripts of approximately 2.2 and 1.9 kb thatare present in both CMS and, at slightly reduced levels, fertility-restored plants. Detectable levels of these transcripts are notobserved in fertile plants with nap cytoplasm (11). As shown

DISCUSSIONUnusual aspects of orf224 editingSeveral lines of evidence suggest that expression of theorJ2241atp6 gene region of the Brassica pol mitochondrialgenome is related to CMS. It is the only region that is expresseddifferently betweenpol CMS and fertile plants and the only regionwhose expression is affected by nuclear fertility restoration (1 1,M.Singh unpublished observations). The single organizationaldifference between the male-sterile pol and male-fertile cammitochondrial genomes is the occurrence, in pol mtDNA, of anadditional 4.5 kb segment that is located immediately upstreamof atp6 and contains the orJ224 gene (19). There are no openreading frames other than orJ224 in this segment and no otherportion of it gives rise to stable transcripts. As transcripts of theCMS-associated maize chimeric gene, T-urfl3, have been shownto not be edited, the finding that orJ224 transcripts are editedwas unexpected; transcripts of the CMS-associated pcf gene ofpetunia may also be edited (6).Although transcript editing has been reported for other unusual

chimeric mitochondrial genes or pseudogenes, these editingevents usually occur in regions corresponding to the codingregions of conventional mitochondrial genes (6, 7). Transcriptsof open reading frames that are not homologous to standardmitochondrial genes are generally found not to be edited. Forexample, the atp6 gene of C male-sterile maize is a triple genefusion: the 5' flanking and N-tenninal coding portions are derivedfrom the atp9 gene, a central region consists of sequence ofunknown origin, and the remaining portion of the gene is similarto the atp6 gene of maize T cytoplasm (20). While the regionsof transcripts corresponding to the atp9 and atp6 portions of the

Nucleic Acids Research, 1994, Vol. 22, No. 11 2113

gene are edited at positions expected from the analysis of othermitochondrial systems, editing does not occur in the central regionof unknown origin (21). Thus, the editing of orJ224 transcriptsin a region that does not correspond to a a known mitochondrialgene is surprising. It is unlikely that this unknown portion ofor1224 represents an essential mitochondrial gene, since thesequence is absent from the closely related, male-fertile cam

mitochondrial genome (19). Moreover, sequences correspondingto the unknown portion of or1224 appear to be absent from bothsunflower or maize mtDNA (results not shown) suggesting thatthis DNA region does not comprise a portion of a gene that isfunctional in the mitochondria of other plant groups.

Thus far, editing of only one other Brassica mitochondrialgene, the atp6 gene, has been reported (13). As with orf224,there is just one editing site on atp6 transcripts. The atp6 codingsequence and the orB portion of ort224 together contain over

950 nucleotides. Only a single editing event is detected intranscripts of these standard mitochondrial gene regions, andhence only about 0.1 % of the nucleotides are edited, a lowerfrequency than that found for any mitochondrial protein codingsequences analyzed thus far (2, 3). It should be of interest todetermine if other Brassica mitochondrial transcripts displaysimilarly low levels of editing. In view of the minimal transcriptediting for these standard mitochondrial gene regions, the findingof an editing event in a region of or1224 that does not correspondto a known sequence seems particularly unusual.

Ediing of orf224 transcripts is not directly related to the CMSphenotypeIt is unclear what role, if any, orJ224 editing plays in the CMStrait. We have recently found that antibodies raised against a

synthetic peptide portion of the predicted translation product ofthe or224-unique region, detect a polypeptide of approximately22 kD that is present in pol, but not nap, cytoplasm plants (22,Y.L'Homme, unpublished results). It remains to be demonstratedif expression of this orJ224-encoded polypeptide is causally relatedto male sterility. Hypothetically, the translation products of eitheredited or unedited transcripts could cause male sterility, but,because the editing of orJ224 transcripts is apparently complete,it is probable that only the edited form occurs in vivo. It seems

unlikely that the degree of editing of ocr224 transcripts is directlyrelated to the male sterile phenotype, since virtually all transcriptsof both sterile and fertility-restored plants are edited. A less directrelationship between or1224 editing and male-sterility cannot beruled out; for example, transcript editing, or the lack of it, couldconceivably signal the processing of orJ224 transcripts that hasbeen proposed to occur specifically in restored plants (18).

In rice it has been found that editing of one form of a normalmitochondrial gene, atp6, differs between CMS and fertility-restored plants (9). Both we (unpublished results) and Handa andNakamura (13), however, have found that atp6 editing also isvirtually complete in both pol CMS and fertile cytoplasm Brassicaplants. This, together with findings reported here and the apparentabsence of an association between RNA editing in the case ofthe maize T-urfl3/or1221 region (8), makes it unlikely that a

direct association between altered editing and CMS is universal.

Analysis of transcript editing using oligonucleotides specificfor edited and unedited formsThis is, to our knowledge, the first time blot-hybridization witholigonucleotides specific to edited and unedited forms has been

successfully used to examine transcript editing. We show that18-mer oligonucleotides can hybridize with sufficient specificityto the corresponding transcript to allow for an approximateassessment of the extent of transcript editing. This approach isadvantageous for analysis of editing at a particular site on atranscript because it is less laborious than sequencing multiplecDNA clones and because it allows the degree of editing ofdifferent sized transcripts of a particular genomic sequence tobe compared. However, the hybridization and washing conditionsrequired to achieve the desired specificity may need to be workedout empirically for each case that is investigated, and analysisof editing at some other sites may be less amenable to thisapproach. Nevertheless, the technique should prove useful foranalyzing the role played by RNA editing in other CMS systemsand possibly in other aspects of plant mitochondrial geneexpression.

Editing of transcripts of a sunflower CMS-associated geneSince our original submission, Moneger et al. (23) have reportedthat transcripts of the chimeric, CMS-associated ORF522 geneof sunflower mtDNA are edited. As with orj224, these transcriptsappear to be completely edited in both CMS and restored plants,providing additional evidence for the lack of general associationbetween RNA editing and CMS.

ACKNOWLEDGEMENTSThis work was supported by grants from the Natural Sciencesand Engineering Research Council of Canada (NSERC). SophieSun was the recepient of an NSERC Undergraduate StudentSummer Award.

REFERENCES

1. Gray, M.W., Hanic-Joyce, P.J., and Covello (1992) Ann. Rev. Plant Physiol.Plant Mol. Biol. 43, 145- 175.

2. Bonnard, G., Gualaberto, J.M., Lamattina, L. amd Grienenberger, J.-M.(1992) CRC Critical Reviews in Plant Science 10, 503-524.

3. Wissinger, B., Brennicke, A. and Schuster, W. (1992) Trends Genet. 8,322-328.

4. Schuster, W. and Brennicke, A. (1991) Nucleic Acids Res. 19, 6923-6928.5. Gualaberto, J.M., Bonnard, G., Lamatfina, L., and Grieneberger, J.M. (1991)

Plant Cell 3, 1109- 1120.6. Hanson, M. (1991) Ann. Rev. Genet. 25, 451-486.7. Bonen, L. and Brown, G.G. (1993) Can. J. Bot. 71, 645-660.8. Ward, G.L. and Levings, C.S.III (1991) Plant Mol. Biol. 17, 1083-1088.9. Iwabuchi, M., Kyozuka, J. and Shimamoto, K. (1993) EMBO J. 12,

1437-1446.10. Hemould, M., Suharsono, S., Litvak, S., Araya, A. and Mourmas, A. (1993)

Proc. Natl. Acad. Sci. U.S.A. 90, 2370-2374.11. Singh, M. and Brown, G.G. (1991) Plant Cell 3, 1349-1362.12. Witt, U., Hansen, M., Albaum, M. and Abel W.O. (1991) Curr. Genet.

19, 323-327.13. Handa, M. and Nakajima, K. (1992) Curr. Genet. 21, 153-159.14. Handa, M. and Nakajima, K. (1992) FEBS Lett. 310, 111-114.15. Perez, M., Bonavent, J.-F. and Berville, A. (1990) Plant Mol. Biol. Rep.

8, 104-113.16. Finnegan, P.M. and Brown, G.G. (1986) Proc. Natl. Acad. Sci. U.S.A.

83, 5175-5179.17. Marchuk, D., Drumm, M., Saulino, A. and Collins, F.S. (1991) Nucleic

Acids Res. 19, 1154.18. Singh, M. and Brown, G.G. (1993) Curr. Genet. 24, 316-322.19. L'Homme, Y. and Brown, G.G. (1993) Nucleic Acids Res. 21, 1903-1909.20. Dewey, R.E., Timothy, D.H. and Levings, C.S.ml (1991) Curr. Genet. 20,

475-482.21. Kumar, R. and Levings, C.C.II (1993) Curr. Genet. 23, 154-159.22. Singh, M. (1992) Ph.D Thesis, McGill University.23. Moneger, F., Smart, C. and Leaver, C.J. (1994) EMBO J. 13, 8-17.