A recombination point is conserved in the mitochondrial genome of higher plant species and located downstream from the cox 2 pseudogene in Solanum tuberosum

The potato (Solanum tuberosum L.) mitochondrial cox3/sdh4/pseudo-cox2 gene cluster has previously been identified by heterologous hybridization using a Marchantia polymorpha sdh4 probe. In our present study we used Southern blotting using sdh4 and cox2 probes to show that the sdh4 and cox2 genes are clustered in the mitochondria of potato, soybean and pea. Northern blotting revealed cotranscription of sdh4 and cox2 in potato but not in cauliflower, indicating that these genes are not clustered in cauliflower. A putative recombination point was detected downstream of the cox2 pseudogene (pseudo-cox2) in potato mitochondrial DNA (mtDNA). This sequence corresponds to a 32 bp sequence which appears to be well-conserved and is adjacent to the terminals of some mitochondrial genes in Citrullus lanatus, Beta vulgaris and Arabidopsis thaliana and is probably involved in the genic rearrangements. It is possible the potato mtDNA pseudo-cox2 gene was generated by recombination during evolution in the same way as that of several other mitochondrial genes and remains as an inactive partial copy of the functional cox2 which was also detected in potato mtDNA.


Introduction
Mitochondrial genomes vary greatly in size, being about 5.7 to 76.5 kb in protists (Gray et al., 1998;Gray et al., 1999), 16 to 19 kb in vertebrates, 17 to 176 kb in fungi and 16 to 2400 kb in land plants (including algae) (Ohta et al., 1998).Angiosperm mitochondrial genomes are the biggest and most complex relative to other eukaryotes (Newton, 1998), their large size being due not only to the fact that plant mitochondria encode two to three times more genes than animals and fungi but also to the frequent recombination events, sequence rearrangements producing pseudogenes (Burger et al., 2003) and the integration of DNA from the nucleus or chloroplast (Schuster and Brennicke, 1988).Mitochondrial insertions in the nuclear genome have also been reported (Stupar et al., 2001).There is great variation in genome size and gene organization given that mitochondrial organization is quite variable among species due to the large amount of mitochondrial recombination which produces subgenomic molecules or isomer forms of the master mitochondrial DNA (mtDNA) genome by intra and interspecific recombination mechanisms via repeated sequences in direct or inverted orientation (Lonsdale, 1984;Newton, 1988;Hanson and Folkerts, 1992;Suzuki et al., 1996).Repetitions of small size achieve intragenic recombination, which can result in the formation of new open reading frames (ORFs) or pseudogenes (Mackenzie and McIntosh, 1999).
The high level of recombination occurring in plant mtDNA results in great structural complexity of the mitochondrial genome; the advantage of this to the organelle perhaps being the maintenance of variation which results in different phenotypic subpopulations (Mackenzie and McIntosh, 1999).Recombination and amplification of sequences are responsible for the difference in size among mitochondrial genomes of the Bryophyte Marchantia polymorpha, algae and higher plants (Turmel et al., 2002;Städler and Delph, 2002) because in Marchantia duplications do not recombine whereas in higher plants some sequences are several kilobases in size and participate in homologous recombination which results in a complex arrangement of molecules (Wissinger et al., 1991).The fre-quency of homologous recombination is probably related to the abundance of repeated sequences able to recombine, large repetitions (0.7 to 120 kb) being considered responsible for the appearance of multipartite structures such as subgenomic circles while small repetitions recombine in response to stress and generate possible mutations, deletions and duplications as well as simple genic rearrangements (Nedelcu and Lee, 1998).Recombination may create chimeric genes and a duplicate repetitive region which contributes to genome expansion, as known to be the case in the mitochondrial genomes of cucumbers (Lilly and Havey, 2001).
It is known that in the mitochondria of some plants integrated sequences originated in the nucleus or chloroplast because the insertion point of these sequences are randomly distributed in the mitochondrial genome (Stern and Palmer, 1984;Brennicke et al., 1993).Sequences encoding tRNAs or other genes may or may not conserve their expression ability (Breiman and Galun, 1990).During transfer between different cell compartments a gene sequence duplicates either by RNA synthesis and reverse transcription to produce cDNA which inserts itself in the new compartment or by transposition without intermediate RNA synthesis in which the original gene sequence is replicated and the DNA copy migrates to the new cell compartment where it installs itself and becomes a functional gene.However, these are hypothetical mechanisms of genic transference being that the real process remains unclear (Schuster and Brennicke, 1988;Ayliffe et al., 1998).Moreover, transfer of sequences from mitochondria to the nucleus have been well-studied and it is accepted that not only do the transferred genes produce pseudogenes in the donor genome but they suffer a high proportion of nucleotide substitution in comparison to the normally encoded genes, meaning that the neutral mutations occur in high frequency in these pseudogenes due to its gradual loss of functionality (Laroche et al., 1997).
The study reported in this paper concerned the cyclooxygenase (COX) pseudogene cox2 occurring in the cyclooxygenase/succinate dehydrogenase (SDH) gene cluster cox3/sdh4/pseudo-cox2 of potato mtDNA and other species.We investigated the relationship of the cox3/sdh4/ pseudo-cox2 gene cluster not only to its genomic localization and organization in some angiosperms species but also to its cotranscription profile and the relationship of the cluster to the 32 nt sequence (AAGCAATGCCCAAAGACT CCCATTTCTTTCTT) located downstream from it and which is conserved in the mitochondria of some species and appears to be involved in rearrangement and/or recombination events.This small sequence is presumed to be involved in the formation of pseudogene cox2 and its consequent location downstream to sdh4 in potato.

Material and Methods
Mitochondrial DNA (mtDNA) was isolated from etiolated seedlings of coix (Coix lacryma-jobi, cv Adlay), maize (Zea mays, cv AGF352), pea (Pisum sativum, cv Mikado), soybean (Glycine max, cv IAC-5), potato tubers (Solanum tuberosum, cv Binje) and cauliflower inflorescence (Brassica oleracea, from a local market).The mitochondria were lysed and the mtDNA separated by cesium chloride-ethidium bromide centrifugation, being after this digested using the BamHI restriction enzyme as described by Siqueira et al. (2001).Standard procedures (Sambrook et al., 1989) were used for restriction enzyme digestions and agarose gel electrophoresis.Potato and cauliflower mitochondrial RNAs (mtRNAs) were prepared as described by Stern and Newton (1986), treated with DNAse I (Sigma, USA) to remove any residual DNA and then phenolextracted and precipitated.
Restriction-digested mtDNAs were transferred to Hybond-N filters (Amersham, UK) by standard procedures (Sambrook et al., 1989).The DNA fragments used as probes were purified from gel-slices by electroelution and labeled by random hexamer priming.Homologous hybridizations were carried out using as probes the potato sdh4 gene, exon 2 of the wheat cox2 gene, and the potato sdh4/pseudo-cox2 cluster against BamHI-digested mtDNA from the angiosperms potato, cauliflower, maize, coix, soybean and pea and the posterior procedures being performed under the conditions described by Sambrook et al. (1989).The mtRNAs isolated from potato and cauliflower were treated with DNase I in order to eliminate DNA traces, fractionated by electrophoresis onto 1.2% (w/v) agarose gel (5 mg mtRNA per lane) and blotted onto Hybond-N filters.Probe labeling, hybridization, solutions to wash the membrane and the full procedures are described in Siqueira et al. (2001).
To detect similar sequences to the potato 32 nt sequence (i.e.downstream from pseudo-cox2) we used the BLAST program (www.ncbi.nlm.nih.gov/BLAST).

Results
Probes consisting of cox2 gene (only exon 2) from wheat and sdh4 and sdh4/pseudo-cox2 genes from potato have been used in homologous hybridization in order to detect the conserved sdh4/pseudo-cox2 genic cluster in the BamHI-digested mtDNAs from potato, cauliflower, maize, coix, soybean and pea.Only one of the cox2 exons from wheat was used device to the high conservation degree of this gene sequence in different plant species.Several bands showing the same sizes were obtained in the hybridizations performed with the three probes (cox2, sdh4 and sdh4/pseudo-cox2, Figure 1).The sdh4/pseudo-cox2 probe consisted of these two clustered genes so, the hybridization bands obtained with this probe corresponded to the sum of all bands obtained in hybridization approaches using the cox2 or sdh4 probes separately.
Potato mtDNA hybridized with the cox2 probe resulted in three bands of which the largest produced a weak 84 Tada and Souza autoradiographic signal representing a fragment of about 10 kb.This is the same fragment that appeared in hybridizations using the sdh4 probe and corresponds to the cox2-containing fragment located downstream from the potato sdh4 gene.It thus appears that both the cox2 and sdh4 genes are located in the same BamHI fragment detected by cox2 and sdh4 probes in the performed hybridizations.This same 10 kb fragment has appeared in previous hybridization studies with the cox3 probe, indicating that the cox3 gene is located on the same BamHI fragment that contains the cox2 and sdh4 genes, in a clustered arrangement of genes (Siqueira et al., 2001).The other two bands showing a strong autoradiographic signal may each correspond to two copies of the cox2 gene in the potato mtDNA.There is also the possibility that the BamHI enzyme cut the cox2 gene in two fragments if there is a restriction site for this enzyme inside this gene.In this case, each part of the digested cox2 gene strongly hybridized with the cox2 probe.
Hybridization performed with cauliflower mtDNA resulted in two bands with a strong intensity in the autoradiography, but neither of which corresponded to the ~12 kb fragment appearing in hybridizations with the sdh4 probe.This demonstrates that in cauliflower mtDNA the cox2 gene is located far from the sdh4 gene and does not form a genic cluster in this plant.These two bands may point to the existence of two copies of the cox2 gene in the mitochondrial genome or they may reflect again the existence of a BamHI restriction site inside the cox2 gene.This is supported by the same fact that may occur in potato: each BamHI fragment containing a piece of cox2 gene strongly hybridized with the used probe.
Soybean mtDNA resulted in only single band of ~8 kb size in the hybridizations with the sdh4 and cox2 probes.Probably, this fragment is corresponding to an intact cox2 gene, indicating that in soybean the cox2 gene is located close to the sdh4 gene.
Pea mtDNA also produced one intense band (~6.6 kb) when hybridizing with the cox2 probe and this band is corresponding to the same band that appeared in hybridization with the sdh4 probe, indicating that these genes are located next between them.
Maize and coix mtDNAs produced bands in the hybridization with the cox2 probe that were not appeared in the hybridization with the sdh4 probe, indicating that in these plants the cox2 and sdh4 genes are located in distant positions along the mitochondrial genome.Moreover, in these species, this result indicates that there is only one cox2 sequence without any copies.
Northern blots containing total mtRNA from potato and cauliflower were performed in order to detect sdh4/pseudo-cox2 cotranscription.MtDNAs from these plants were hybridized against the same probes used in the previous hybridizations (Figure 2).The results show that in potato, there is a 4.4 kb transcript common to the sdh4 and cox2 probes; a fact that indicates cotranscription of these genes.It is noteworthy again that the transcripts observed in  the hybridizations with the sdh4/pseudo-cox2 probe correspond to the same obtained in each separate hybridization with the sdh4 or cox2 probes.In the hybridization with cauliflower mtRNA a single 1.6 kb band was obtained with the cox2 gene which did not appear in the hybridizations with the sdh4 probe, indicating that in these plants, cotranscription does not occur between sdh4 and cox2 and that sdh4 and cox2 genes are not clustered.This observation is in agreement with the conclusions obtained from the Figure 1.Downstream from the partial cox2 sequence, itself placed downstream from the sdh4 gene in the potato, there is a small sequence (Figure 3.A) corresponding to 32 nucleotides (AAGCAATGCCCAAAGACTCCCATTTCTT TCTT GenBank: AF280607) which are highly (93-100% similarity) conserved in other mitochondrial genomes.Employing the BLAST program for sequence identity search, it was revealed the presence of this small sequence in other mtDNAs besides potato (Figure 3.B).This sequence was found in watermelon (Citrullus lanatus) (AF288042), sugar beet (Beta vulgaris) (Kubo et al., 2000), Arabidopsis thaliana (Unseld et al., 1997) and Brassica napus (GenBank: AP006444) and seems to be involved in mitochondrial genome rearrangement.In Citrullus, this sequence locates upstream to nad9 gene.In sugar beet the same sequence is located in different regions upstream to the atp6 and atpA genes, to a direct repeat of the mitochondrial genome and downstream to the rps7 gene and cox2 pseudogene.In A. thaliana, this sequence is located upstream to nad9 and downstream to the rpl16 and atp9 genes, while in B. napus the sequence appears upstream to c-type cytochrome synthesis gene (ccmFN2 or cycK) and downstream to the nad9 gene.These results indicate conservation of this sequence between some plant species.In some cases, the 32 nt potato sequence is not totally conserved but appears as a partial sequence in various plants.In sugar beet and some other plants there are several small fragments that share partial homology with the potato fragment (data not shown) which may be involved in recombination events in the mitochondrial genome of these species throughout evolution.

Discussion
Our hybridization results (Figure 1) show the presence of the cox3/sdh4/pseudo-cox2 cluster (Siqueira et al., 2001) in potato mtDNA but this sequence was not identified in the other plants analyzed in this study (cauliflower, coix, maize, pea and soybean), which is another example of the complexity and diversity of plant mitochondrial genomes due to their inherent dynamism and high level of recombination.The sdh4/cox2 cluster (the pseudo-cox2 cluster) was encountered in potato, pea and soybean but not in cauliflower, coix and maize, where these genes were encountered not as a cluster but separately.Moreover, if in potato the putative single copy of cox2 was not cut by BamHI there is a possibility that the potato genome contains a second copy of this gene.
As mentioned above, the sdh4/cox2 cluster is absent from the mitochondrial genome of cauliflower, coix and maize, where there were single bands for each gene indicating that they were unclustered, but, however, another approach is necessary in order to detect the presence or absence of cox2 in the nucleus.If this gene is encountered in this compartment, this could indicate that genic transfer occurred during evolution and perhaps the mitochondrial copy could have been inactivated and replaced by a pseudogene.Multiple transference processes have been described for structural and respiratory chain genes in plant mitochondria, especially for the respiratory chain genes sdh4 (Adams et al., 2001) and cox2 (Kanazawa et al., 1998;Adams et al., 1999;Subramanian et al., 2001).However, as was observed by us in pea and soybean, the cited studies gave no evidence for the evolution of a recombined sequence or site that could have allowed the separation of cox2 and sdh4 or of fragments of these genes (pseudogenes) dispersed in the rest of the genome.
We found that in maize and coix the cox2 and sdh4 genes are dispersed in the mitochondrial genome and appear as a single copy, indicating the absence of a recombination event that placed these sequences in close together or created pseudogenes.In cauliflower mtDNA, however, the cox2 probe revealed two bands which may corresponding either to cox2 cut by BamHI or two copies of the cox2 gene, although the sdh4/cox2 cluster was absent.Adams et al., (1999) has shown that in some plant species the cox2 gene has been transferred from the mitochondria to the nucleus while in other species it remains in the mitochondria or still it can exists like a functional nuclear copy in addition to a pseudogene that remains located in the mitochondria.Moreover, Lupold et al. (1999) reported that rearrangement of sequences influences the activity of cox2 promoters in maize mitochondria.
Northern Blotting performed with total mtRNA from the potato and cauliflower against sdh4 and cox2 probes indicated that cotranscription of the sdh4/pseudo-cox2 cluster occurred only in potato and not cauliflower (Figure 2).This observation was to be expected because sdh4 and cox2 are separated in cauliflower (Figure 1) which means that cotranscription between these genes is not to be expected.In cauliflower mtDNA there were no transcript bands common to both the probes, while in potato there was a common 4.4 kb transcript that was obtained with both probes.Unseld et al. (1997) stated that cotranscription of mitochondrial genes in plants is a frequent event because many mitochondrial genes are clustered.
The 32 nt sequence we detected alongside pseudo-cox2 sequence is a further example of the small sequences commonly involved in the frequent rearrangements which occur in the mitochondrial genome of plants.The conservation of the this sequence in some plant species may point to 86 Tada and Souza it having a useful function, its association with various genes in diverse plant species reflects its ability to recombine during evolution.In Beta vulgaris, p. ex.these conserved sequence appears to be dispersed along the genome in association with different genes (Kubo et al., 2000) and may be involved in the rearrangement process.In potato this sequence may be responsible for the interruption of the cox2 sequence and could be associated with the relatively high recombination rate of the cox2 genic region, which may result in a partial cox2 sequence located downstream from the sdh4 gene.Ceci et al. (1993) have reported that in sunflower a recombination event occurred downstream A new recombination point in plant mitochondrial DNA 87 from the sdh4 gene and that this event inserted a 417 bp chloroplast DNA fragment which included a ct-tRNA Val sequence.This recombination event took place four nucleotides upstream from the insertion point of the cox2 partial sequence in potato mtDNA, illustrating that this region seems to be a `hot spot' for recombination events in various species.An interesting fact is that BLAST analysis (data not shown) showed that in Arabidopsis and sugar beet mtDNA the cox2 pseudogene is conserved adjacent to the 32 nt sequence just as in potato mtDNA.
Our analysis of the cox2 gene and the associated recombinatory sequence is another example of the very complex and dynamic structure of plant mitochondrial genomes.Further analysis should be carried out with the plant species investigated by us to establish whether or not transcription initiation and processing sites exist and to determine the transcript promoter region and also to verify the presence or absence of the cox2 sequence in the nucleus.

Figure 2 -
Figure 2 -Hybridization between mtRNA from potato and cauliflower with the following probes (indicated at the top of the pictures): A = sdh4, B = sdh4/cox2 and C = cox2.The numbers on the left of the pictures indicate the transcripts size in kb based on an Invitrogen 0.24 to 9.5 kb RNA Ladder molecular size marker (M).(•) indicates the common transcripts for cox2 and sdh4 probes.The observed transcripts in the hybridization with the cox2/sdh4 probe correspond to the same transcripts observed in the separate hybridizations with sdh4 and cox2.

Figure 3 -
Figure 3 -Diagrams (not to scale) showing the position of the conserved potato mitochondrial sequence.(A) Linear nucleotide potato sequence.The atga sequence at site 795 indicates the atg start codon for sdh4 and the tga stop codon for cox3.Nucleotide numbers are shown on the right and square brackets indicate the start/stop for each gene.The 32 nt recombinatory sequence conserved in some genomes is underlined and in bold.(B) Representative diagram showing the total sequence analyzed in this research.Numbers above the bar indicate the start and stop positions for the cox3 and sdh4 genes, pseudo-cox2 and the 32 nt conserved sequence (in evidence below the diagram).Each cm in the bar corresponds to 50 nt.