MBE Advance Access originally published online on February 13, 2007
Molecular Biology and Evolution 2007 24(5):1101-1112; doi:10.1093/molbev/msm030
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles |
Comparative Analysis of Sequences Preceding Protein-Coding Mitochondrial Genes in Flowering Plants
Biology Department, University of Ottawa, Ottawa, Canada
E-mail: lbonen{at}science.uottawa.ca.
 |
Abstract |
|---|
 TOP Abstract IntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
We examined the nucleotide sequences preceding 23 mitochondrial protein-coding genes held in common by maize, rice, wheat, sugar beet, tobacco, Arabidopsis, and Brassica to look for features related to translation initiation and to assess the degree of conservation in mitochondrial mRNA leaders among these plants. We observed broad variation in sequence similarity as illustrated by dot plot analysis, ranging from a level rivaling that of coding sequences to complete absence of homology due to lineage-specific DNA rearrangements. Genes encoding ATP synthase subunits predominated in the latter category, whereas ones encoding cytochrome c biogenesis proteins and NADH dehydrogenase subunits were primarily of the highly conserved type. Within the region immediately preceding initiation codons, in most cases we did not observe motifs consistent with a bacterial-type Shine–Dalgarno interaction to assist in ribosome binding, nor was any other consensus sequence evident. In fact, indels in the form of tandem repeats were seen among homologues from different plants. We did, however, observe a bias for high adenosine and low cytosine in the proximal
30 nt compared with further upstream. Duplicates of some sequences in our data set were found to be associated with more than one gene within a genome. Indeed, 3 such families of upstream cassettes were identified, and they exhibit a lineage-specific distribution among plants. Moreover, the presence of related sequences at genomic sites distant from known genes raises the possibility of future recruitment as regulatory elements. Our observations point to a dynamic nature in the makeup of the 5' leaders of plant mitochondrial mRNAs and an apparent plasticity in translational control elements.
Key Words: plant mitochondria • 5' UTR • translation initiation • DNA rearrangement
|
Introduction |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
The mitochondrial genetic systems in various eukaryotic lineages have evolved in distinctly different directions, and some features that distinguish those of flowering plants are a low rate of nucleotide substitution, a high rate of genomic rearrangement, and very large genome sizes which range from approximately 0.2–2 Mbp among plants (reviewed in Bullerwell and Gray 2004
). The protein-coding gene content in plant mitochondria typically consists of 35 components which are primarily subunits of the respiratory chain and translational machinery. This set is smaller, however, than that of certain protists such as Reclinomonas americana, which has 67 protein-coding genes in a mitochondrial genome of only 69 kb (Gray et al. 2004). Noncoding sequences account for most of the expanded genome size in plants, and although certain stretches can be identified as acquired chloroplast or nuclear sequences, much is of unknown origin (reviewed in Kubo and Mikami 2007). In addition, most plant mitochondrial genomes exist in complex physical forms due to DNA recombination across repeated sequences, which result in a mixed population of subgenomic forms of the deduced master chromosome. Consequently, gene order often varies (even among closely related plants), and coding regions can become relocated into the context of new flanking sequences and hence acquire new regulatory signals. In exceptional cases, rearrangement events can even impact on coding sequences.
Although progress is being made in our knowledge about transcription and RNA processing events in plant mitochondria, as yet little is known about translation initiation or signals involved in the recognition of the correct start codon. This is despite a longstanding interest in this issue (cf., Dawson et al. 1984; Boer et al. 1985; Schuster et al. 1990). For example, Pring et al. (1992) identified 3 short (10–12 nt) conserved blocks within 100 nt preceding the start codons of respiratory chain genes, such as atp6, atp4, and cox2 in various grasses and eudicots, suggestive of a regulatory role. Based on literature reports as well as our own unpublished data, plant mitochondrial 5' untranslated regions (UTRs) often range from 100 to 400 nt in length and presumably contain expression elements for translational control and perhaps also for mRNA stability. Genes which are cotranscribed sometimes possess RNA cleavage sites, and plant mitochondrial transcripts also typically undergo C-to-U type-RNA editing, as well as splicing of group II introns in some cases. Editing, although predominantly occurring within coding sequences, has occasionally been observed in UTRs (reviewed in Shikanai 2006). For example, an editing site was identified 4 nt upstream of the rps14 initiation codon in Oenothera mitochondria (Schuster et al. 1990) and in rice, there is an editing site within the 3 nt spacer separating rpl2 and rps19 (Kubo et al. 1996). This organization of ribosomal protein genes illustrates traces of an ancestral bacterial-type order of genes occasionally seen in plant mitochondria.
It might be anticipated that translation initiation in plant mitochondria would be similar to that of bacteria given their endosymbiotic ancestry and the highly conserved nature of core regions within their ribosomal RNAs (Gray 1992). In bacteria, initiation codon recognition is assisted by base pairing between the pyrimidine-rich 3' end of the small subunit (SSU) (16S) rRNA and a purine-rich (Shine–Dalgarno) sequence within the mRNA. The latter is typically 4–5 nt in length and located 5–9 nt upstream from the start codon (reviewed in Kozak 1999; Marintchev and Wagner 2005; Nakamoto 2006). This, in conjunction with the fMet-tRNA anticodon–codon interaction is important in correct positioning of the 30S ribosomal initiation complex on the mRNA. The chloroplast translation system in flowering plants also has these features, and Shine–Dalgarno sequences have been experimentally shown to be necessary for the translation of a subset of chloroplast mRNAs (reviewed in Sugiura et al. 1998). In plant mitochondria, however, the extreme 3' terminal region of the SSU (18S) rRNA is slightly shorter based on direct RNA sequencing data (cf., Schnare and Gray 1982), and it lacks a canonical anti-Shine–Dalgarno sequence. Consequently, it has been puzzling as to how the correct initiation site is recognized in plant mitochondria. Interestingly, in Reclinomonas americana mitochondria, which is described as the most bacteria like of any known mitochondrial genome, candidate Shine–Dalgarno-like sequences have been identified upstream of most protein-coding genes (Lang et al. 1997).
In the present study, we have analyzed the 100 nt stretches located immediately upstream of protein-coding genes that are present in all 7 mitochondrial genomes which have been completely sequenced, namely those of maize (Clifton et al. 2004), rice (Notsu et al. 2002), wheat (Ogihara et al. 2005), sugar beet (Kubo et al. 2000), tobacco (Sugiyama et al. 2005), Arabidopsis (Unseld et al. 1997), and Brassica (Handa 2003). Such regions will include mRNA leader sequences that are expected to contain signals involved in translation initiation. Our study has revealed a wide variation in degree of sequence conservation and an apparent plasticity in translational control elements.
|
Methods |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Nucleotide sequences upstream of plant mitochondrial protein-coding genes were obtained from the National Center for Biotechnology Information (NCBI) GenBank entries for the 7 genomes listed in table 1. Genes held in common among these plant genomes are shown in table 2, with the exception that mttB, matR, rps4, and ccmC were omitted from our data set because there is uncertainty as to the position of their start codons in one or more plants. Although ccmC was originally designated as a pseudogene in sugar beet (Satoh et al. 2004
), its status has been reevaluated (Mower and Palmer 2006). Intraspecific comparisons were made using sequences from sugar beet normal-type cytoplasm (BA000009
[GenBank]
) and "Owen cytoplasm" (BA000024
[GenBank]
). In almost all cases, the locations of translation initiation codons were taken from the GenBank annotations. However, there were 7 instances in which there is strong phylogenetic support from comparative sequence analysis for the use of a downstream in-frame AUG, that is the one which would correspond with the initiation codon annotated in other species. Moreover, the presence of frameshifts or in-frame stop codons in homologous sequences from closely related species usually precluded their use of the distal AUG. Those 7 cases for which the proximal AUG was selected (and distance downstream of the annotated initiator) are: Arabidopsis nad2 (33 nt), wheat nad6 (54 nt), wheat nad9 (291 nt), rice cox3 (45 nt), Arabidopsis atp9 (33 nt), maize ccmFN (39 nt), and wheat ccmFN (42 nt). In certain cases, these designations are also supported by experimental analysis (cf., wheat nad6, Haouazine-Takvorian et al [1997]; wheat nad9, Lamattina et al. [1993]).
|
|
We selected a length of 100 nt preceding start codons because, although experimental data are as yet limited, plant mitochondrial 5' UTRs are typically at least this long. Sequence alignments were carried out using ClustalW with the default parameter settings (Chenna et al. 2003
) and then subjected to minor correction by manual inspection. The dot matrix plot was generated using JDotter (http://athena.bioc.uvic.ca/workbench.php?tool=jdotter&db; Brodie et al. 2004) set to default preferences with the stringency level adjusted so that even relatively short sequences of low complexity were visualized, and after identification as such they were omitted from further analysis. This was done to minimize the possibility of overlooking potentially meaningful sequence relationships.
To assess the presence of additional copies within individual plant mitochondrial genomes, we conducted Blast (bl2seq) searches using each 100 nt sequence in our data set as query with default parameters and the filter removed (Tatusova and Madden 1999). A duplicate sequence was designated as an "upstream cassette" when at least 50 nt were found within 100 nt upstream of another gene. This included 3 genes not in our primary data set, namely rps7 in maize, rice, and wheat; rpl5 in sugar beet; and rps13 in tobacco. Copies not closely associated with a known gene were categorized based on length of sequence similarity as well as position within the query sequence. Long copies (i.e., >40 nt) exhibited E values ranging from 10–6 to 10–20, with most being lower than 10–10, whereas short copies (20–40 nt) exceeded 90% sequence identity. We also searched for specific short nucleotide stretches within our data set using FUZZNUC (e.g., allowing for 2 nt degeneracy within an 8 nt stretch) (http://bioweb.pasteur.fr/seqanal/interfaces/fuzznuc.html). To determine whether sequences of chloroplast origin were represented in out data set, we used Blast (bl2seq) to query the rice (NC_001320
[GenBank]
) and Arabidopsis (NC_000932
[GenBank]
) chloroplast genomes.
|
Results |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Variation in Conservation and Origin of Upstream Sequences
To assess the degree of conservation in sequences preceding plant mitochondrial genes and to search for potential translational regulatory signals, we compiled sequences extending 100 nt upstream of 23 protein-coding genes, which are held in common among the 7 flowering plant species whose mitochondrial genomes have been completely sequenced (table 1). More specifically, these plants include 3 monocots, maize, rice, and wheat, which diverged from a common ancestor
50–70 MYA (Kellogg 2001) and 4 eudicots, sugar beet, tobacco, Arabidopsis, and Brassica. The latter 2 species are the most closely related, having diverged from a common ancestor 20 MYA (Koch et al. 2000) and the monocot–eudicot split is estimated to have occurred about 150 MYA (Chaw et al. 2004). Among the 23 genes used in our analysis (table 2), 3 are both duplicated and show variation within the 100 nt preceding the start codon, namely atp8 and atp6 in wheat and atp6 in Arabidopsis, so were also included in our data set. Nucleotide alignments are provided in supplementary figure S1 (Supplementary Material online). RNA editing of a genomically encoded ACG in a few cases generates the AUG initiation codon; more specifically, there are 6 cases for nad1, 4 for nad4L, and 1 each for cox1 and atp6. This was based on reported experimental evidence or inferred from comparative analysis. A dot matrix plot was generated for this data set of 164 sequences strung together as a single consecutive sequence, and this is shown as a comparison with itself in figure 1A. Signals along the diagonal (seen as a continuous line) represent identical sequences, whereas signal patterns immediately off the diagonal (squared regions) depict sequence similarity among different plants for a given gene. Signals further off the diagonal represent sequences that are duplicated and located upstream of more than one gene (fig. 1A, illustrated by open arrowheads, discussed below). As can be seen from figure 1A, there is a broad variation in the appearance of the profiles, hence in the sequence similarity preceding different mitochondrial genes. Strikingly, some upstream regions like those associated with nad4, ccmFN, and ccmFC are so similar near the start codon among all the 7 plant species that conservation rivals that of the coding sequence. For yet others, the sequences upstream of genes such as ccmB (fig. 1B) show somewhat greater divergence and reflect the phylogenetic distance between the monocot and eudicot groups. In the case of nad2 (fig. 1C), the monocot–eudicot differences are even more accentuated, with sequences appearing more divergent among the monocots, even though these plants are more closely related than are most of the eudicot species to each other.
|
Among the genes exhibiting conserved upstream sequences, there were several cases in which the sequence from one plant was seen to lack homology with those in all the other species (supplementary fig. S1, Supplementary Material online). One of these, namely nad3, is shown in figure 1D, and in this case mitochondrial DNA rearrangements in the rice lineage have replaced the ancestral-type sequence with one unrelated to anything in the data bank. In the case of maize nad6, the recently acquired upstream sequence was seen to be a paralogue of that preceding atp4 (discussed below). The dot matrix plot in figure 1A also shows that some genes lack prominent signals, and homology among plants is either uncertain or absent. For example, upstream of atp1, clear-cut similarity was detected only between Arabidopsis and Brassica (fig. 1E), although short stretches of similarity could be identified among various plants (supplementary fig. S1, Supplementary Material online) as represented by the fainter signals in figure 1E. Blast searches using sequences in our data set as query revealed numerous matches with short duplicated segments from various genomic sites. For example, the rice nad2 upstream sequence contains a nad7-coding segment of 98 nt beginning approximately 30 nt upstream of the start codon. These observations are consistent with a well-known feature of plant mitochondrial genomes, namely that "bits and pieces" of genic (and flanking sequence) copies are scattered around the genome (cf., Kubo and Mikami 2007
), having arisen through DNA duplication/recombination events or reverse transcriptase-mediated amplification of transcribed sequences with subsequent genomic integration. It is worth noting that none of the sequences in our data set exhibited similarity to sequences of chloroplast origin, which reside in plant mitochondrial genomes, although such sequences can be recruited to serve as mitochondrial expression elements (cf., Nakazono et al. 1996).
From our dot plot analysis, it can also be seen that there are several trends with respect to gene type and degree of variation among plant species. For example, sequences preceding the atp and cox genes are more volatile than those upstream of ccm genes or to a lesser extent nad genes (fig. 1A). The atp6 gene is particularly atypical as it is often preceded by a fused open reading frame, which differs among plants (reviewed in Bonen and Brown 1993) and the resulting amino-terminal region is removed by proteolytic cleavage to generate the mature ATP6 protein (cf., Krishnasamy et al. 1994). In addition, atp6 copies are located on recombinationally active repeats in plants such as wheat (Bonen and Bird 1988) and Arabidopsis (Marienfeld et al. 1996) and consequently differ in their flanking upstream sequences. On the other hand, it should also be noted that some upstream sequences in our data set are constrained because of their close proximity to a cotranscribed gene. In the case of rps12 in all 7 species, the 3' end of nad3 is only 50 nt upstream (supplementary fig. S1, Supplementary Material online). Similarly, rps3 is located 10–20 nt downstream of rps19 in tobacco (or a 
rps19 copy in the other 6 plant species), so that in these 2 cases the query included coding sequences (supplementary fig. S1, bold, Supplementary Material online).
Apparent Absence of a Bacteria-Type Ribosome-Binding Motif
To search for features consistent with a Shine–Dalgarno-type base-pairing interaction, the 20 nt stretch immediately preceding each protein-coding gene in our data set was examined for sequences complementary to at least 4 consecutive nucleotides within 6 nt at the extreme 3' end of the plant mitochondrial 18S rRNA (fig. 2A). It should be noted that the 3' terminus of the mitochondrial SSU rRNA is highly conserved among plants, and this even extends to the bryophyte Marchantia polymorpha. However, the plant mitochondrial ones differ from bacterial and chloroplast in sequence at the expected position of the anti-Shine–Dalgarno element (fig. 2A, boxed) and are slightly shorter in length. This evolutionary shift appears to have occurred after land plants diverged from the green algal lineage, in that the counterpart in Chara vulgaris mitochondria still is bacteria like (Turmel et al. 2003) as seen in figure 2A. Interestingly, within the region shown in figure 2A, only one nucleotide difference is observed between Marchantia and flowering plants, yet it is located within the boxed region and further reduces the pyrimidine content.
|
At the location corresponding to the Shine–Dalgarno motif in bacterial mRNAs, only 42 of 164 plant mitochondrial sequences exhibited potential complementary to the SSU rRNA sequence. Such candidate sites for the rice and Arabidopsis subset are shown in figure 2B (black boxes), and they vary in location relative to the start codon. In contrast to plant mitochondria, approximately half of the sequences upstream of the corresponding genes in Reclinomonas mitochondria contain potential complementary sequences (fig. 2C), which are virtually all identical and at a location more similar to that found in bacteria (cf., also Lang et al. 1997
). Short stretches of relatively low complexity (e.g., homopolymeric adenosine as well as uridine) are evident in figure 2B, but we found no obvious consensus sequence among all (or even most) sequences in the data set. Figure 2B also illustrates the degree of sequence variation observed between rice and Arabidopsis. More specifically, among the 23 different genes, only 14 upstream sequences appear to be homologous. In 8 of those cases (fig. 2B, name underlined), differences are limited to a few nucleotide substitutions (or none in the case of nad4) (fig. 2B, vertical ovals) of which about 25% are C-to-U editing candidates. For the other 6 cases (fig. 2B, name broken underlined), insertion/deletions were observed between rice and Arabidopsis (fig. 2B, boxed). Notably, such indels occur almost immediately upstream of 14 different genes in our data set (supplementary fig. S1, Supplementary Material online) and 6 of them are shown in figure 3. Their direct tandem repeat nature suggests that they originated by slippage during DNA replication. Indeed, 3 of the cases shown in figure 3 involve differences between the closely related Arabidopsis and Brassica, presumably reflecting very recent evolutionary events. Such variation would be unexpected if regions near the start codons were under strong functional constraint due to the presence of a ribosome-binding site that has strict requirements for its sequence and distance from AUG. Taken together, our observations suggest that classical bacteria-type ribosome binding is unlikely to be the mode of initiator recognition in plant mitochondria.
|
When we examined the nucleotide composition of sequences immediately upstream of protein-coding genes, we observed a bias for high adenosine (36.2%) and low cytosine (16.2%) in the 30 nt preceding start codons (fig. 4A) compared with further upstream (27.6 and 21.3% respectively). In contrast, both the G and U content are more uniformly represented in the 100 nt region preceding initiators and range from 21 to 28% (fig. 4A). Notably, the nucleotide content in the region extending from
30 to 100 nt upstream reflects the overall A + T content (55%) of flowering plant mitochondrial genomes (reviewed in Sugiyama et al. 2005). The asymmetry in A versus C composition for positions –1 to –30 (fig. 4A) mimics the profile of the A versus U content in the comparable region of yeast cytosolic 5' UTRs (Shabalina et al. 2004), and it will be of interest to learn what parameters control accessibility of the ribosome to the correct start site. We observed at least one additional AUG triplet independent of coding reading frame in about 80% of the sequences in our data set, thus disfavoring a scanning model of initiation codon recognition.
|
Within the coding region examined (namely, the first 100 nt), there is a bias for high uridine, which can at least partially be attributed to the third codon position U-bias observed in plant mitochondria (cf., Maier et al. 1996
). An examination of the nucleotide composition close to the start codon (fig. 4B) revealed only one conspicuous feature, that is, a bias against guanosine at position –2. There was no indication of an "extended" codon–anticodon interaction, as has been suggested for translation in chloroplasts where uridine residues are typically present at position –1 (cf., Esposito et al. 2003) and the A + G content is only approximately 55% for positions –5 to –9 which would correspond to the location of the purine-rich Shine–Dalgarno motif in bacteria (fig. 4B).
Paralogues of Upstream Sequences Preceding Multiple Genes
To explore sequence relationships within a given genome, we investigated dot plot signals that are off the main diagonal in figure 1A (illustrated by open arrowheads) and thus represent sequences upstream of more than one gene in our data set. From previous work (reviewed in Bonen and Brown 1993), it has long been appreciated that stretches preceding genes are sometimes duplicated and copies are present upstream of other genes. They therefore potentially provide common regulatory signals rather than each gene having its own unique cis-elements. In our analysis, which included Blast searches (bl2seq) querying individual genomes, a sequence was considered to be a member of an upstream cassette family if a copy longer than 50 nt preceding the initiation codon was present within 100 nt upstream of another known gene. Three such families, designated as types I–III, were identified (fig. 5A, open, gray, and hatched blocks, respectively) and alignments of sequences in the immediate vicinity of the initiation codon are shown in figure 5B. Full sequence alignments are given in supplementary figure S2 (Supplementary Material online). It can be seen from figure 5A that these upstream cassette families have up to 5 members within a genome (including incomplete copies) and they show lineage-specific distributions among the plants. Using Blast searches, we found that some of these cassette-type sequences are also present in spacer regions (see below) or upstream of genes that had been excluded from our data set because they were not represented in all the 7 plant species. The latter includes rps7 in maize, rice, and wheat (type I); rpl5 in sugar beet (type II); and rps13 in tobacco (type III) (fig. 5A and B).
|
Upstream cassette type I is present in all 3 cereals, and its longest members extend
200 nt upstream of the initiation codon and include a promoter (cf., Covello and Gray 1991). Interestingly, in the case of wheat atp6-1, atp6-2, and atp4, homology continues 50 nt further downstream (fig. 5A), although it should be noted that these sequences comprise part of an ORF fused to atp6 (cf., Bonen and Brown 1993) and do not contribute to the mature ATP6 protein. Cassette type I has a rather mosaic-like appearance in that certain stretches share greater sequence similarity than do others (fig. 5B, shaded; supplementary fig. S2, Supplementary Material online). This was previously observed by Pring et al. (1992) who noted that one of the conserved blocks is coincident with the 5' end of the shorter members and thus consistent with a recombination site. Moreover, certain members within a genome appear to have undergone gene conversion events in that stretches of greater similarity are seen among subsets (fig.5B; supplementary fig. S2, Supplementary Material online). In the case of rice cox2 and atp6, the cassette extends precisely up to (but not past) the initiation codon, thus illustrating clean displacement of the preexisting regulatory sequences and no apparent requirement for adaptation of downstream coding sequences. For rps7 in all 3 cereals, approximately 90 nt of ancestral-type upstream sequence has been retained (fig. 5A and B; supplementary fig. S2, Supplementary Material online), but interestingly, close to the rps7 start codon there is an 8 nt motif (GGAAATTC) that also precedes the start codons of wheat and rice cox2, rice atp6, and maize nad6 (fig. 5B, lowercase italics). Because of the location, it is a candidate for playing a role in start codon selection. Alternatively, it might be involved in other RNA level processing/stability or regulatory events. A search for this motif using FUZZNUC (and allowing for 2 nt degeneracy) revealed 23 other instances within our data set, which are located within 20 nt of the start codon.
Cassette type II (120 nt in length) is found only in the sugar beet mitochondrial genome (fig.5A; supplementary fig. S2, Supplementary Material online), and the origin of its 4 members can be traced to nad9-type upstream sequences (supplementary fig. S1, Supplementary Material online). This cassette family can be divided into 2 subtypes because the proximal part of the cox2 and rpl5 copies (fig. 5A, black blocks) are nonhomologous to those shared between atp6 and nad9 (fig. 5B). This is again suggestive of either gene conversion or difference in timing of duplication events. The eudicot cassette type III ( 180 nt long) is well-conserved over its length (fig. 5B, supplementary fig. S2, Supplementary Material online) although tobacco cox2 has just a partial copy (fig. 5A, B), and this cassette type appears to have originated from eudicot atp8-type sequences (fig. 5A; supplementary fig. S1 and S2, Supplementary Material online). Notably, all members of cassette III (except the truncated tobacco cox2 one) break in homology precisely at the position of the initiation codon (as was seen above for certain cassette I members). Although no duplicated cassette type III sequences were found upstream of other Brassica genes, 2 atp8-type copies are found in genomic spacer regions (see below). There is a fourth minor upstream repeat family in maize mitochondria (data not shown), in that cox1 and rps2B are preceded by a common 45 nt stretch plus an additional 13 nt extending into the coding region and thus reminiscent of the wheat cassette I case described above.
Additional Copies of Upstream Sequences in Spacer Regions
We extended the Blast (bl2seq) analysis of individual genomes to search for sequences similar to those in our data set that were more distantly located upstream (>100 nt) or downstream from a known gene, and in figure 6 their locations are schematically shown along the linearized master chromosomes for each of the 7 plants. Sequences similar to the upstream cassettes discussed in the previous section are designated as C-I, C-II, and C-III. We found 63 paralogues, which collectively represent almost every gene in our data set, with the exception of ccm genes, which also are distinct in having among the most conservative upstream regions. Sequences were classified based on length and position relative to the start codon of the query sequence: 1) long near (fig. 6, black stars), 2) short near (fig.6, open stars), or 3) long distant (fig. 6, open circles) as described in more detail in the figure 6 legend. Approximately two-thirds are within 1 kb of a known gene and strikingly, all are in the same orientation as that gene. Thus, if actively transcribed, they might contribute regulatory cis-elements, such as for RNA processing or stability. Of the 54 in categories 1 and 2 (fig. 6, black and open stars), 37 extend to or past the start codon and thus potentially include translation initiation signals. For the total set (which excludes any within 100 nt upstream of a known protein-coding gene because they were discussed above), three-quarters lie downstream of a gene, and interestingly, most of those located within 100 nt (that is, 7 out of 9) are linked to genes encoding tRNAs or rRNAs.
|
The genomic density of these intergenic copies varies among the 7 plants, with Arabidopsis being the highest (and representing a diversity of genes) and wheat the lowest. Moreover, all of the copies in Arabidopsis are long (fig.6, black stars), and this might reflect recent duplication without subsequent fragmentation by rearrangement. The master chromosome in Arabidopsis contains only 2 recombinational repeat elements (Unseld et al.1997
), whereas wheat is reported to have 16 repeated regions (Ogihara et al. 2005). It will be of interest to learn more about the interplay between the generation/eradication versus functional recruitment of intergenic copies. Although copies were generally dispersed throughout each genome, there were cases of close linkage. For example, in Arabidopsis, the atp4 and atp6-2 elements are less than 200 bp apart, and they are located on a recombinational repeat element.
Over half of these duplicated upstream sequences also include part of the coding sequence with which it was originally associated. In 15 cases, this has resulted in the seemingly fortuitous creation of mosaic reading frames, which have been annotated as ORFs in the mitochondrial complete genome data bank entries. For example, the 5' end of orf315 in Arabidopsis is composed of a duplicated atp9-coding fragment (of 40 codons), which is preceded by 209 nt of atp9-type upstream sequence. Interestingly, several of these chimeric ORFs are among the ones identified to be expressed at high levels in Arabidopsis plants, which are downregulated for polynucleotide phosphorylase (Holec et al. 2006). This suggests that atp9 upstream sequences may contribute to transcription and/or RNA processing/stability elements. Similarly, in sugar beet with normal-type cytoplasm, orf246 was generated in part by the tandem duplication of a short rps3-coding segment preceded by 154 nt of its associated upstream sequence. This orf246 is actively transcribed, although no translation products were detected (Satoh et al. 2004).
|
Discussion |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Plant mitochondrial genomes exhibit a striking degree of plasticity in organization, even within regions located very close to protein-coding genes, and this is illustrated in our examination of sequences immediately preceding translation initiation codons. Relatively few of the 23 genes in our data set showed strong, uninterrupted conservation in upstream sequences among all 7 plants examined, and in many cases sequences were clearly nonhomologous. The switching of upstream regulatory sequences via DNA rearrangements has occurred frequently during flowering plant evolution, and genes encoding ATP synthase (atp) subunits are particularly dynamic. In addition, certain newly acquired upstream sequences are duplicate copies that also precede another gene (or genes) within a given genome. Our analysis focused on data from the completely sequenced mitochondrial genomes of the 7 plant species, whose divergence times range from
20 MYA to 150 MYA, but the phenomenon of genomic fluidity impacting on upstream regulatory regions is also evident within a given species. For example, between the sugar beet "Owen CMS" and normal-type mitochondrial genomes (Satoh et al. 2004), sequences preceding cox2 and atp1 each show a break in homology within 50 nt of their respective start codons, and in the case of atp8, upstream sequences are completely unrelated in the 2 cytoplasms.
The presence of additional upstream sequence copies within the intergenic spacers of plant mitochondrial genomes provides an opportunity for their future recruitment as regulatory elements as well as the de novo creation of transcripts which might even be translated into novel polypeptides. Indeed, this is a hallmark of certain types of cytoplasmic male sterility (reviewed in Hanson and Bentolila 2004), whereby detrimental proteins are generated from novel chimeric ORFs. In this regard, it is notable that atp sequences (both regulatory and coding) are particularly frequent contributors to the creation of such ORFs. In a recent study of CMS in rice with Boro II cytoplasm, fertility restoration was found to be mediated by pentatricopeptide repeat (PPR) type proteins which act at the mRNA turnover level to downregulate cytotoxic orf79 (Wang et al. 2006). This chimeric orf79 includes a short cox1-coding segment fused behind cox2-type upstream sequences (i.e., cassette type I in fig. 5) and that in turn is preceded by atp6 and its associated type I upstream sequence.
The dynamic nature of sequences preceding protein-coding genes also has implications for translational signals, and at the present time it is unclear how the correct AUG is recognized by the ribosomal machinery. Our analysis has revealed that even though the sequences preceding initiation codons show a bias for adenosine, there do not appear to be conserved purine-rich blocks complementary to the 3' terminal region of the SSU rRNA (and thus candidate elements for a bacterial-type translation initiator recognition) nor was any other consensus sequence evident. These observations in conjunction with the presence of indel sequence variation very close to initiation codons suggest that a classical Shine–Dalgarno-type interaction is unlikely to play a major role in plant mitochondrial translation initiation. Moreover, the function of the S1 ribosomal protein in Escherichia coli to stabilize the mRNA at the initiation codon (Sorensen et al. 1998) is excluded in that the plant mitochondrial S1 counterpart lacks the corresponding carboxy-terminal RNA-binding domains (cf., Mundel and Schuster 1996).
The ability of upstream sequences to be shuffled among protein-coding genes in plant mitochondria, as exemplified by the 3 families of lineage-specific cassettes we have described, is rather reminiscent of behavior seen in yeast mitochondrial pseudorevertants. For example, respiratory function was seen to be restored through DNA duplication/rearrangements, whereby defective cox1 upstream sequences were replaced by a cob leader (Manthey and McEwen 1995). In yeast mitochondria, 5'UTRs are typically long and they possess cis-elements that interact with gene-specific translational activators and localize protein synthesis on the mitochondrial inner membrane (reviewed in Costanzo and Fox 1990). Because of the added complexity of events such as RNA editing in plant mitochondria, it is possible that RNA-binding proteins (such as PPR proteins, cf., Andrés et al. 2007) might perform multiple roles, and this could contribute to an evolutionarily dynamic translation initiation recognition system with independent gene-specific signals. It will be important to have a greater understanding of translational signals in plant mitochondria to better appreciate the potential impact (both beneficial and detrimental) of highly rearranging genomes.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Supplementary figures S1 and S2 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Financial support by the Natural Sciences and Engineering Research Council of Canada is gratefully acknowledged.
|
Footnotes |
|---|
Franz Lang, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Andrés C, Lurin C, Small ID. The multifarious roles of PPR proteins in plant mitochondrial gene expression. Physiol Plant (2007) 129:14–22.[CrossRef]
Boer PH, McIntosh JE, Gray MW, Bonen L. The wheat mitochondrial gene for apocytochrome b: absence of a prokaryotic ribosome binding site. Nucleic Acids Res (1985) 13:2281–2292.
Bonen L, Bird S. Sequence analysis of the wheat mitochondrial atp6 gene reveals a fused upstream reading frame and markedly divergent N-termini among plant ATP6 proteins. Gene (1988) 73:47–56.[CrossRef][ISI][Medline]
Bonen L, Brown GG. Genetic plasticity and its consequences: perspectives on gene organization and expression in plant mitochondria. Can J Bot (1993) 71:645–660.
Brodie R, Roper RL, Upton C. JDotter: a Java interface to multiple dotplots generated by dotter. Bioinformatics (2004) 20:279–281.
Bullerwell CE, Gray MW. Evolution of the mitochondrial genome: protist connections to animals, fungi and plants. Curr Opin Microbiol (2004) 7:528–534.[CrossRef][ISI][Medline]
Chaw SM, Chang CC, Chen HL, Li WH. Dating the monocot-dicot divergence and the origin of core eudicots using whole chloroplast genomes. J Mol Evol (2004) 58:424–441.[CrossRef][ISI][Medline]
Chenna R, Sugawara H, Koike T, Lopez R, Gibson TJ, Higgins DG, Thompson JD. Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res (2003) 31:3497–3500.
Clifton SW, Minx P, Fauron CMR, et al, (13 co-authors). Sequence and comparative analysis of the maize NB mitochondrial genome. Plant Physiol (2004) 136:3486–3503.
Costanzo MC, Fox TD. Control of mitochondrial gene expression in Saccharomyces cerevisiae. Annu Rev Genet (1990) 24:91–113.[CrossRef][ISI][Medline]
Covello PS, Gray MW. Sequence analysis of wheat mitochondrial transcripts capped in vitro: definitive identification of transcription initiation sites. Curr Genet (1991) 20:245–251.[CrossRef][ISI][Medline]
Dawson AJ, Jones VP, Leaver CJ. The apocytochrome b gene in maize mitochondria does not contain introns and is preceded by a potential ribosome binding site. EMBO J (1984) 3:2107–2113.[ISI][Medline]
Esposito D, Fey JP, Eberhard S, Hicks AJ, Stern DB. In vivo evidence for the prokaryotic model of extended codon-anticodon interaction in translation initiation. EMBO J (2003) 22:651–656.[CrossRef][ISI][Medline]
Gray MW. The endosymbiont hypothesis revisited. Int Rev Cytol (1992) 141:233–357.[ISI][Medline]
Gray MW, Lang BF, Burger G. Mitochondria of protists. Annu Rev Genet (2004) 38:477–524.[CrossRef][ISI][Medline]
Handa H. The complete nucleotide sequence and RNA editing content of the mitochondrial genome of rapeseed (Brassica napus L.): comparative analysis of the mitochondrial genomes of rapeseed and Arabidopsis. Nucleic Acids Res (2003) 31:5907–5916.
Hanson MR, Bentolila S. Interactions of mitochondrial and nuclear genes that affect male gametophyte development. Plant Cell (2004) 16:S154–S169.
Haouazine-Takvorian N, Takvorian A, Jubier MF, Lejeune B. Genes encoding subunit 6 of NADH dehydrogenase and subunit 6 of ATP synthase are co-transcribed in maize mitochondria. Curr Genet (1997) 31:63–69.[CrossRef][ISI][Medline]
Holec S, Lange H, Kuhn K, Alioua M, Borner T, Gagliardi D. Relaxed transcription in Arabidopsis mitochondria is counterbalanced by RNA stability control mediated by polyadenylation and polynucleotide phosphorylase. Mol Cell Biol (2006) 26:2869–2876.
Kellogg EA. Evolutionary history of the grasses. Plant Physiol (2001) 125:1198–1205.
Koch MA, Haubold B, Mitchell-Olds T. Comparative evolutionary analysis of chalcone synthase and alcohol dehydrogenase loci in Arabidopsis, Arabis, and related genera (Brassicaceae). Mol Biol Evol (2000) 17:1483–1498.
Kozak M. Initiation of translation in prokaryotes and eukaryotes. Gene (1999) 234:187–208.[CrossRef][ISI][Medline]
Krishnasamy S, Grant RA, Makaroff CA. Subunit 6 of the Fo-ATP synthase complex from cytoplasmic male-sterile radish: RNA editing and NH2-terminal protein sequencing. Plant Mol Biol (1994) 24:129–141.[CrossRef][ISI][Medline]
Kubo T, Mikami T. Organization and variation of angiosperm mitochondrial genome. Physiol Plant (2007) 129:6–13.[CrossRef]
Kubo T, Nishizawa S, Sugawara A, Itchoda N, Estiati A, Mikami T. The complete nucleotide sequence of the mitochondrial genome of sugar beet (Beta vulgaris L.) reveals a novel gene for tRNA-Cys (GCA). Nucleic Acids Res (2000) 28:2571–2576.
Kubo N, Ozawa K, Hino T, Kadowaki K. A ribosomal protein L2 gene is transcribed, spliced and edited at one site in rice mitochondria. Plant Mol Biol (1996) 31:853–862.[CrossRef][ISI][Medline]
Lamattina L, Gonzalez D, Gualberto J, Grienenberger JM. Higher plant mitochondria encode an homologue of the nuclear-encoded 30-kDa subunit of bovine mitochondrial complex I. Eur J Biochem (1993) 217:831–838.[ISI][Medline]
Lang BF, Burger G, O'Kelly CJ, Cedergren R, Golding GB, Lemieux C, Sankoff D, Turmel M, Gray MW. An ancestral mitochondrial DNA resembling a eubacterial genome in miniature. Nature (1997) 387:493–497.[CrossRef][Medline]
Maier RM, Zeltz P, Kossel H, Bonnard G, Gualberto JM, Grienenberger JM. RNA editing in plant mitochondria and chloroplasts. Plant Mol Biol (1996) 32:343–365.[CrossRef][ISI][Medline]
Manthey GM, McEwen JE. The product of the nuclear gene PET309 is required for translation of mature mRNA and stability or production of intron-containing RNAs derived from the mitochondrial COX1 locus of Saccharomyces cerevisiae. EMBO J (1995) 14:4031–4043.[ISI][Medline]
Marienfeld J, Unseld M, Brandt P, Brennicke A. Genomic recombination of the mitochondrial atp6 gene in Arabidopsis thaliana at the protein processing site creates two different presequences. DNA Res (1996) 3:287–290.[Abstract]
Marintchev A, Wagner G. Translation initiation: structures, mechanisms and evolution. Q Rev Biophys (2005) 3:197–284.
Mower JP, Palmer JD. Patterns of partial RNA editing in mitochondrial genes of Beta vulgaris. Mol Gen Genomics (2006) 276:285–293.[CrossRef][ISI][Medline]
Mundel C, Schuster W. Loss of RNA editing of rps1 sequences in Oenothera mitochondria. Curr Genet (1996) 30:455–460.[CrossRef][ISI][Medline]
Nakamoto T. A unified view of the initiation of protein synthesis. Biochem Biophys Res Commun (2006) 341:675–678.[CrossRef][ISI][Medline]
Nakazono M, Nishiwaki S, Tsutsumi N, Hirai A. A chloroplast-derived sequence is utilized as a source of promoter sequences for the gene for subunit 9 of NADH dehydrogenase (nad9) in rice mitochondria. Mol Gen Genet (1996) 252:371–378.[ISI][Medline]
Notsu Y, Masood S, Nishikawa T, Kubo N, Akiduki G, Nakazono M, Hirai A, Kadowaki K. The complete sequence of the rice (Oryza sativa L.) mitochondrial genome: frequent DNA sequence acquisition and loss during the evolution of flowering plants. Mol Genet Genomics (2002) 268:434–445.[CrossRef][ISI][Medline]
Ogihara Y, Yamazaki Y, Murai K, et al, (14 co-authors). Structural dynamics of cereal mitochondrial genomes as revealed by complete nucleotide sequencing of the wheat mitochondrial genome. Nucleic Acids Res (2005) 33:6235–6250.
Pring DR, Mullen JA, Kempken F. Conserved sequence blocks 5' to start codons of plant mitochondrial genes. Plant Mol Biol (1992) 19:313–317.[CrossRef][ISI][Medline]
Satoh M, Kubo T, Nishizawa S, Estiati A, Itchoda N, Mikami T. The cytoplasmic male-sterile type and normal type mitochondrial genomes of sugar beet share the same complement of genes of known function but differ in the content of expressed ORFs. Mol Genet Genomics (2004) 272:247–256.[CrossRef][ISI][Medline]
Schnare MN, Gray MW. 3'-terminal sequence of wheat mitochondrial 18S ribosomal RNA: further evidence of a eubacterial evolutionary origin. Nucleic Acids Res (1982) 10:3921–3932.
Schuster W, Unseld M, Wissinger B, Brennicke A. Ribosomal protein S14 transcripts are edited in Oenothera mitochondria. Nucleic Acids Res (1990) 18:229–233.
Shabalina SA, Ogurtsov AY, Rogozin IB, Koonin EV, Lipman DJ. Comparative analysis of orthologous eukaryotic mRNAs: potential hidden functional signals. Nucleic Acids Res (2004) 32:1774–1782.
Shikanai T. RNA editing in plant organelles: machinery, physiological function and evolution. Cell Mol Life Sci (2006) 63:698–708.[CrossRef][ISI][Medline]
Sorensen MA, Fricke J, Pedersen S. Ribosomal protein S1 is required for translation of most, if not all, natural mRNAs in Escherichia coli in vivo. J Mol Biol (1998) 280:561–569.[CrossRef][ISI][Medline]
Sugiura M, Hirose T, Sugita M. Evolution and mechanism of translation in chloroplasts. Annu Rev Genet (1998) 32:437–459.[CrossRef][ISI][Medline]
Sugiyama Y, Watase Y, Nagase M, Makita N, Yagura S, Hirai A, Sugiura M. The complete nucleotide sequence and multipartite organization of the tobacco mitochondrial genome: comparative analysis of mitochondrial genomes in higher plants. Mol Genet Genomics (2005) 272:603–615.[CrossRef][ISI][Medline]
Tatusova TA, Madden TL. Blast 2 sequences, a new tool for comparing protein and nucleotide sequences. FEMS Microbiol Lett (1999) 174:247–250.[CrossRef][ISI][Medline]
Turmel M, Otis C, Lemieux C. The mitochondrial genome of Chara vulgaris: insights into the mitochondrial DNA architecture of the last common ancestor of green algae and land plants. Plant Cell (2003) 15:1888–1903.
Unseld M, Marienfeld JR, Brandt P, Brennicke A. The mitochondrial genome of Arabidopsis thaliana contains 57 genes in 366,924 nucleotides. Nat Genet (1997) 15:57–61.[CrossRef][ISI][Medline]
Wang Z, Zou Y, Li X, et al, (13 co-authors). Cytoplasmic male sterility of rice with boro II cytoplasm is caused by a cytotoxic peptide and is restored by two related PPR motif genes via distinct modes of mRNA silencing. Plant Cell (2006) 18:676–687.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A.-V. Bohne, S. Ruf, T. Borner, and R. Bock Faithful transcription initiation from a mitochondrial promoter in transgenic plastids Nucleic Acids Res., December 18, 2007; 35(21): 7256 - 7266. [Abstract] [Full Text] [PDF]
|