A comparative study of the ori sequences from the mitochondrial genomes of twenty wild-type yeast strains

Decreasing mitochondrial RNA polymerase activity reverses biased inheritance of hypersuppressive mtDNA

Faithful inheritance of mitochondrial DNA (mtDNA) is crucial for cellular respiration/oxidative phosphorylation and mitochondrial membrane potential. However, how mtDNA is transmitted to progeny is not fully understood. We utilized hypersuppressive mtDNA, a class of respiratory deficient Saccharomyces cerevisiae mtDNA that is preferentially inherited over wild-type mtDNA (rho+), to uncover the factors governing mtDNA inheritance. We found that some regions of rho+ mtDNA persisted while others were lost after a specific hypersuppressive takeover indicating that hypersuppressive preferential inheritance may partially be due to active destruction of rho+ mtDNA. From a multicopy suppression screen, we found that overexpression of putative mitochondrial RNA exonuclease PET127 reduced biased inheritance of a subset of hypersuppressive genomes. This suppression required PET127 binding to the mitochondrial RNA polymerase RPO41 but not PET127 exonuclease activity. A temperature-sensitive allele of RPO41 improved rho+ mtDNA inheritance over a specific hypersuppressive mtDNA at semi-permissive temperatures revealing a previously unknown role for rho+ transcription in promoting hypersuppressive mtDNA inheritance.

The mitochondrial RNA landscape of Saccharomyces cerevisiae

Mitochondria are essential organelles that harbor a reduced genome, and expression of that genome requires regulated metabolism of its transcriptome by nuclear-encoded proteins. Despite extensive investigation, a comprehensive map of the yeast mitochondrial transcriptome has not been developed and all of the RNA-metabolizing proteins have not been identified, both of which are prerequisites to elucidating the basic RNA biology of mitochondria. Here, we present a mitochondrial transcriptome map of the yeast S288C reference strain. Using RNAseq and bioinformatics, we show the expression level of all transcripts, revise all promoter, origin of replication, and tRNA annotations, and demonstrate for the first time the existence of alternative splicing, mirror RNAs, and a novel RNA processing site in yeast mitochondria. The transcriptome map has revealed new aspects of mitochondrial RNA biology and we expect it will serve as a valuable resource. As a complement to the map, we present our compilation of all known yeast nuclear-encoded ribonucleases (RNases), and a screen of this dataset for those that are imported into mitochondria. We sought to identify RNases that are refractory to recovery in traditional mitochondrial screens due to an essential function or eclipsed accumulation in another cellular compartment. Using this in silico approach, the essential RNase of the nuclear and cytoplasmic exosome, Dis3p, emerges as a strong candidate. Bioinformatics and in vivo analyses show that Dis3p has a conserved and functional mitochondrial-targeting signal (MTS). A clean and marker-less chromosomal deletion of the Dis3p MTS results in a defect in the decay of intron and mirror RNAs, thus revealing a role for Dis3p in mitochondrial RNA decay.

Helicase Hmi1 stimulates the synthesis of concatemeric mitochondrial DNA molecules in yeast Saccharomyces cerevisiae

Hmi1p is a helicase in the yeast Saccharomyces cerevisiae required for maintenance of the wild-type mitochondrial genome. Disruption of the HMI1 ORF generates rho(-) and rho(0) cells. Here we demonstrate that, in rho(-) yeast strains, Hmi1p stimulates the synthesis of long concatemeric mitochondrial DNA molecules associated with a reduction in the number of nucleoids used for mitochondrial DNA packaging. Surprisingly, the ATPase negative mutants of Hmi1p can also stimulate the synthesis of long concatemeric rho(-) mitochondrial DNA molecules and support the maintenance of the wild-type mitochondrial genome, albeit with reduced efficiency. We show that, in the mutant hmi1-5 background, the wild-type mitochondrial DNA is fragmented; and we propose that, in hmi1Delta yeast cells, the loss of the wild-type mitochondrial genome is caused by this fragmentation of the mitochondrial DNA.

Replication and preferential inheritance of hypersuppressive petite mitochondrial DNA

Wild-type yeast mitochondrial DNA (mtDNA) is inherited biparentally, whereas mtDNA of hypersuppressive petite mutants is inherited uniparentally in crosses to strains with wild-type mtDNA. Genomes of hypersuppressive petites contain a conserved ori sequence that includes a promoter, but it is unclear whether the ori confers a segregation or replication advantage. Fluorescent in situ hybridization analysis of wild-type and petite mtDNAs in crosses reveals no preferential segregation of hypersuppressive petite mtDNA to first zygotic buds. We identify single-stranded DNA circles and RNA-primed DNA replication intermediates in hypersuppressive petite mtDNA that are absent from non-hypersuppressive petites. Mutating the promoter blocks hypersuppressiveness in crosses to wild-type strains and eliminates the distinctive replication intermediates. We propose that promoter-dependent RNA-primed replication accounts for the uniparental inheritance of hypersuppressive petite mtDNA.

Nucleo-mitochondrial interactions in mitochondrial gene expression

All proteins encoded by mitochondrial DNA (mtDNA) are dependent on proteins encoded by nuclear genes for their synthesis and function. Recent developments in the identification of these genes and the elucidation of the roles their products play at various stages of mitochondrial gene expression are covered in this review, which focuses mainly on work with the yeast Saccharomyces cerevisiae. The high degree of evolutionary conservation of many cellular processes between this yeast and higher eukaryotes, the ease with which mitochondrial biogenesis can be manipulated both genetically and physiologically, and the fact that it will be the first organism for which a complete genomic sequence will be available within the next 2 to 3 years makes it the organism of choice for drawing up an inventory of all nuclear genes involved in mitochondrial biogenesis and for the identification of their counterparts in other organisms.

Lack of site-specific recombination between mitochondrial genomes of petite mutants of yeast

Previous work from our laboratory showed that mitochondrial (mt) genomes, with tandem repeat units, from spontaneous, cytoplasmic petite mutants of Saccharomyces cerevisiae do not exhibit site-specific recombination in petite x petite crosses [Rayko et al., Gene 63 (1988) 213-226]. Here, we have extended and confirmed these observations by studying other crosses of petites with tandem repeat units, as well as crosses in which one of the parents was, instead, an unstable petite, a-15/4/1, having a palindromic mt genome. In no case was site-specific recombination of the parental mt genomes observed. Progeny cells harbored mt genomes derived from either one or both of the two parents, as shown by analysis of restriction fragments. In the case of biparental inheritance, extensive subcloning of the diploids showed that this was due to a persistent heteroplasmic state and not to intermolecular recombination. The 'new' restriction fragments present in the mt DNA from some diploids were shown to be derived from the unstable parental genome, a-15/4/1, by a secondary excision process. Lack of site-specific recombination is, therefore, not only a feature of crosses involving petite genomes made up of tandem repeat units, but also of crosses in which one parental genome consists of inverted repeats and frequently originates secondary petite genomes formed by tandem repeats. Previous reports of mt recombination in petite mutants are discussed in light of these results.

Genetic approaches to the study of mitochondrial biogenesis in yeast

In contrast to most other organisms, the yeast Saccharomyces cerevisiae can survive without functional mitochondria. This ability has been exploited in genetic approaches to the study of mitochondrial biogenesis. In the last two decades, mitochondrial genetics have made major contributions to the identification of genes on the mitochondrial genome, the mapping of these genes and the establishment of structure-function relationships in the products they encode. In parallel, more than 200 complementation groups, corresponding to as many nuclear genes necessary for mitochondrial function or biogenesis have been described. Many of the latter are required for post-transcriptional events in mitochondrial gene expression, including the processing of mitochondrial pre-RNAs, the translation of mitochondrial mRNAs, or the assembly of mitochondrial translation products into the membrane. The aim of this review is to describe the genetic approaches used to unravel the intricacies of mitochondrial biogenesis and to summarize recent insights gained from their application.

Saccharomyces cerevisiae contains an RNase MRP that cleaves at a conserved mitochondrial RNA sequence implicated in replication priming

Yeast mitochondrial DNA contains multiple promoters that sponsor different levels of transcription. Several promoters are individually located immediately adjacent to presumed origins of replication and have been suggested to play a role in priming of DNA replication. Although yeast mitochondrial DNA replication origins have not been extensively characterized at the primary sequence level, a common feature of these putative origins is the occurrence of a short guanosine-rich region in the priming strand downstream of the transcriptional start site. This situation is reminiscent of vertebrate mitochondrial DNA origins and raises the possibility of common features of origin function. In the case of human and mouse cells, there exists an RNA processing activity with the capacity to cleave at a guanosine-rich mitochondrial RNA sequence at an origin; we therefore sought the existence of a yeast endoribonuclease that had such a specificity. Whole cell and mitochondrial extracts of Saccharomyces cerevisiae contain an RNase that cleaves yeast mitochondrial RNA in a site-specific manner similar to that of the human and mouse RNA processing activity RNase MRP. The exact location of cleavage within yeast mitochondrial RNA corresponds to a mapped site of transition from RNA to DNA synthesis. The yeast activity also cleaved mammalian mitochondrial RNA in a fashion similar to that of the mammalian RNase MRPs. The yeast endonuclease is a ribonucleoprotein, as judged by its sensitivity to nucleases and proteinase, and it was present in yeast strains lacking mitochondrial DNA, which demonstrated that all components required for in vitro cleavage are encoded by nuclear genes. We conclude that this RNase is the yeast RNase MRP.

Saccharomyces cerevisiae contains an RNase MRP that cleaves at a conserved mitochondrial RNA sequence implicated in replication priming

Yeast mitochondrial DNA contains multiple promoters that sponsor different levels of transcription. Several promoters are individually located immediately adjacent to presumed origins of replication and have been suggested to play a role in priming of DNA replication. Although yeast mitochondrial DNA replication origins have not been extensively characterized at the primary sequence level, a common feature of these putative origins is the occurrence of a short guanosine-rich region in the priming strand downstream of the transcriptional start site. This situation is reminiscent of vertebrate mitochondrial DNA origins and raises the possibility of common features of origin function. In the case of human and mouse cells, there exists an RNA processing activity with the capacity to cleave at a guanosine-rich mitochondrial RNA sequence at an origin; we therefore sought the existence of a yeast endoribonuclease that had such a specificity. Whole cell and mitochondrial extracts of Saccharomyces cerevisiae contain an RNase that cleaves yeast mitochondrial RNA in a site-specific manner similar to that of the human and mouse RNA processing activity RNase MRP. The exact location of cleavage within yeast mitochondrial RNA corresponds to a mapped site of transition from RNA to DNA synthesis. The yeast activity also cleaved mammalian mitochondrial RNA in a fashion similar to that of the mammalian RNase MRPs. The yeast endonuclease is a ribonucleoprotein, as judged by its sensitivity to nucleases and proteinase, and it was present in yeast strains lacking mitochondrial DNA, which demonstrated that all components required for in vitro cleavage are encoded by nuclear genes. We conclude that this RNase is the yeast RNase MRP.

Sequence rearrangements at the ori 7 region of Saccharomyces cerevisiae mitochondrial DNA

Three ori elements (ori 2, ori 5, and ori 7) have been sequenced in Saccharomyces cerevisiae strain Dip 2 and compared to the equivalent ori elements of a second strain (B). Both ori 2 and ori 5 exhibit 98% base matching between strains Dip 2 and B. In contrast, the third ori element (ori 7) exhibits extensive sequence rearrangements whereby a segment located downstream in the consensus strain occurs within the ori structure in Dip 2. This represents a novel polymorphic form of the yeast mitochondrial genome.

Conversion at large intergenic regions of mitochondrial DNA in Saccharomyces cerevisiae

Saccharomyces cerevisiae mitochondrial DNA deletion mutants have been used to examine whether base-biased intergenic regions of the genome influence mitochondrial biogenesis. One strain (delta 5.0) lacks a 5-kilobase (kb) segment extending from the proline tRNA gene to the small rRNA gene that includes ori1, while a second strain (delta 3.7) is missing a 3.7-kb region between the genes for ATPase subunit 6 and glutamic acid tRNA that encompasses ori7 plus ori2. Growth of these strains on both fermentable and nonfermentable substrates does not differ from growth of the wild-type strain, indicating that the deletable regions of the genome do not play a direct role in the expression of mitochondrial genes. Examination of whether the 5- or 3.7-kb regions influence mitochondrial DNA transmission was undertaken by crossing strains and examining mitochondrial genotypes in zygotic colonies. In a cross between strain delta 5.0, harboring three active ori elements (ori2, ori3, and ori5), and strain delta 3.7, containing only two active ori elements (ori3 and ori5), there is a preferential recovery of the genome containing two active ori elements (37% of progeny) over that containing three active elements (20%). This unexpected result, suggesting that active ori elements do not influence transmission of respiratory-competent genomes, is interpreted to reflect a preferential conversion of the delta 5.0 genome to the wild type (41% of progeny). Supporting evidence for conversion over biased transmission is shown by preferential recovery of a nonparental genome in the progeny of a heterozygous cross in which both parental molecules can be identified by size polymorphisms.

Conversion at large intergenic regions of mitochondrial DNA in Saccharomyces cerevisiae

Saccharomyces cerevisiae mitochondrial DNA deletion mutants have been used to examine whether base-biased intergenic regions of the genome influence mitochondrial biogenesis. One strain (delta 5.0) lacks a 5-kilobase (kb) segment extending from the proline tRNA gene to the small rRNA gene that includes ori1, while a second strain (delta 3.7) is missing a 3.7-kb region between the genes for ATPase subunit 6 and glutamic acid tRNA that encompasses ori7 plus ori2. Growth of these strains on both fermentable and nonfermentable substrates does not differ from growth of the wild-type strain, indicating that the deletable regions of the genome do not play a direct role in the expression of mitochondrial genes. Examination of whether the 5- or 3.7-kb regions influence mitochondrial DNA transmission was undertaken by crossing strains and examining mitochondrial genotypes in zygotic colonies. In a cross between strain delta 5.0, harboring three active ori elements (ori2, ori3, and ori5), and strain delta 3.7, containing only two active ori elements (ori3 and ori5), there is a preferential recovery of the genome containing two active ori elements (37% of progeny) over that containing three active elements (20%). This unexpected result, suggesting that active ori elements do not influence transmission of respiratory-competent genomes, is interpreted to reflect a preferential conversion of the delta 5.0 genome to the wild type (41% of progeny). Supporting evidence for conversion over biased transmission is shown by preferential recovery of a nonparental genome in the progeny of a heterozygous cross in which both parental molecules can be identified by size polymorphisms.

RPO41-independent maintenance of [rho-] mitochondrial DNA in Saccharomyces cerevisiae

A subset of promoters in the mitochondrial DNA (mtDNA) of the yeast Saccharomyces cerevisiae has been proposed to participate in replication initiation, giving rise to a primer through site-specific cleavage of an RNA transcript. To test whether transcription is essential for mtDNA maintenance, we examined two simple mtDNA deletion ([rho-]) genomes in yeast cells. One genome (HS3324) contains a consensus promoter (ATATAAGTA) for the mitochondrial RNA polymerase encoded by the nuclear gene RPO41, and the other genome (4a) does not. As anticipated, in RPO41 cells transcripts from the HS3324 genome were more abundant than were transcripts from the 4a genome. When the RPO41 gene was disrupted, both [rho-] genomes were efficiently maintained. The level of transcripts from HS3324 mtDNA was decreased greater than 400-fold in cells carrying the RPO41 disrupted gene; however, the low-level transcripts from 4a mtDNA were undiminished. These results indicate that replication of [rho-] genomes can be initiated in the absence of wild-type levels of the RPO41-encoded RNA polymerase.

RPO41-independent maintenance of [rho-] mitochondrial DNA in Saccharomyces cerevisiae

A subset of promoters in the mitochondrial DNA (mtDNA) of the yeast Saccharomyces cerevisiae has been proposed to participate in replication initiation, giving rise to a primer through site-specific cleavage of an RNA transcript. To test whether transcription is essential for mtDNA maintenance, we examined two simple mtDNA deletion ([rho-]) genomes in yeast cells. One genome (HS3324) contains a consensus promoter (ATATAAGTA) for the mitochondrial RNA polymerase encoded by the nuclear gene RPO41, and the other genome (4a) does not. As anticipated, in RPO41 cells transcripts from the HS3324 genome were more abundant than were transcripts from the 4a genome. When the RPO41 gene was disrupted, both [rho-] genomes were efficiently maintained. The level of transcripts from HS3324 mtDNA was decreased greater than 400-fold in cells carrying the RPO41 disrupted gene; however, the low-level transcripts from 4a mtDNA were undiminished. These results indicate that replication of [rho-] genomes can be initiated in the absence of wild-type levels of the RPO41-encoded RNA polymerase.

Stable maintenance of a 35-base-pair yeast mitochondrial genome

Small deletion variants ([rho-] mutants) derived from the wild-type ([ rho+]) Saccharomyces cerevisiae mitochondrial genome were isolated and characterized. The mutant mitochondrial DNAs (mtDNAs) examined retained as little as 35 base pairs of one section of intergenic DNA, were composed entirely of A.T base pairs, and were stably maintained. These simple mtDNAs existed in tandemly repeated arrays at an amplified level that made up approximately 15% of the total cellular DNA and, as judged by fluorescence microscopy, had a nearly normal mitochondrial arrangement throughout the cell cytoplasm. The simple nature of these [rho-] genomes indicates that the sequences required to maintain mtDNA must be extremely simple.

Putative target sites for mobile G + C rich clusters in yeast mitochondrial DNA: single elements and tandem arrays

GC clusters constitute the major repetitive elements in the mitochondrial (mt) genome of the yeast Saccharomyces cerevisiae. Many of these clusters are optional and thus contribute much to the polymorphism of yeast mtDNAs. We have made a systematic search for polymorphic sites by comparing mtDNA sequences of various yeast strains. Most of the 26 di- or polymorphic sites found differ by the presence or absence of a GC cluster of the majority class, here referred to as the M class, which terminate with an AGGAG motif. Comparison of sequences with and without the GC clusters reveal that elements of the subclasses M1 and M2 are inserted 3' to a TAG, flanked by A + T rich sequences. M3 elements, in contrast, only occur in tandem arrays of two to four GC clusters; they are consistently inserted 3' to the AGGAG terminal sequence of a preexisting cluster. The TAG or the terminal AGGAG, therefore, are regarded as being part of the target sites for M1 and M2 or M3 elements, respectively. The dinucleotide AG is in common to both target sites; it also occurs at the 3' terminus (AGGAG). This suggests its duplication during GC cluster insertion. This notion is supported by the observation that GC clusters of the minor classes G and V similarily repeat at their 3' terminus a GT or an AA dinucleotide, respectively, from their putative target sites.

Stable maintenance of a 35-base-pair yeast mitochondrial genome

Small deletion variants ([rho-] mutants) derived from the wild-type ([ rho+]) Saccharomyces cerevisiae mitochondrial genome were isolated and characterized. The mutant mitochondrial DNAs (mtDNAs) examined retained as little as 35 base pairs of one section of intergenic DNA, were composed entirely of A.T base pairs, and were stably maintained. These simple mtDNAs existed in tandemly repeated arrays at an amplified level that made up approximately 15% of the total cellular DNA and, as judged by fluorescence microscopy, had a nearly normal mitochondrial arrangement throughout the cell cytoplasm. The simple nature of these [rho-] genomes indicates that the sequences required to maintain mtDNA must be extremely simple.

The GC clusters of the mitochondrial genome of yeast and their evolutionary origin

We have studied the primary and secondary structures, the location and the orientation of the 196 GC clusters present in the 90% of the mitochondrial genome of Saccharomyces cerevisiae which have already been sequenced. The vast majority of GC clusters is located in intergenic sequences (including ori sequences, intergenic open reading frames and the gene varl which probably arose from an intergenic spacer) and in intronic closed reading frames (CRF's); most of them can be folded into stem-and-loop systems; both orientations are equally frequent. The primary structures of GC clusters permit to group them into eight families, seven of which are related to the family formed by clusters A, B and C of the ori sequences. On the basis of the present work, we propose that the latter derive from a primitive ori sequence (probably made of only a monomeric cluster C and its flanking sequences r* and r) through (i) a series of duplication inversions generating clusters A and B; and (ii) an expansion process producing the AT stretches of ori sequences. Most GC clusters apparently originated from primary clusters also derived from the primitive ori sequence in the course of its evolution towards the present ori sequences. Finally, we propose that the function of GC clusters is predominantly, or entirely, associated with the structure and organization of the mitochondrial genome of yeast and, indirectly, with the regulation of its expression.

The primary structure of the mitochondrial genome of Saccharomyces cerevisiae--a review

We have collated and compiled all the available primary structure data on the mitochondrial genome of Saccharomyces cerevisiae. Data concern 78,500 bp, namely 92% of the 'long' genomes; they are derived from several laboratory strains. Interstrain differences belong to three classes: a small number of large deletions/additions, mainly concerning introns; a large number of small (10-150 bp) deletions/additions located in the intergenic sequences; 1-3 bp deletions/additions and point mutations; the interstrain sequence divergence due to the latter, is of the order of 2% for the strains compared; this low value is, however, an overestimate because of sequence mistakes.

Sequence organization of the mitochondrial genome of yeast--a review

We have compiled the available primary structural data for the mitochondrial genome of Saccharomyces cerevisiae and have estimated the size of the remaining gaps, which represent 12-13% of the genome. The lengths of sequenced regions and of gaps lead to a new assessment of genome sizes; these range (in round figures) from 85 000 bp for the long genomes, to 78 000 bp for the short genomes, to 74 000 bp for the supershort genome of Saccharomyces carlsbergensis. These values are 8-11% higher than those previously estimated from restriction fragments. Interstrain differences concern not only facultative intervening sequences (introns) and mini-inserts, but also insertions/deletions in intergenic sequences. The primary structure appears to be extremely conserved in genes and ori sequences, and highly conserved in intergenic sequences. Since coding sequences represent at most 33-35% of the genome, at least two thirds of the genome are formed by noncoding and yet highly conserved sequences. The G + C level of genes or exon is 25%, and that of intronic open reading frames (ORFs) 22%; increasingly lower values are shown by intronic closed reading frames (CRFs), 20%, ori sequences, 19%, intergenic ORFs, 17.5% and intergenic sequences, 15%.

A new putative gene in the mitochondrial genome of Saccharomyces cerevisiae

The 2200-bp ori2-ori7 region of the mitochondrial (mt) genome of Saccharomyces cerevisiae has been sequenced on the genome of a petite, b7, excised at those ori sequences from wild-type strain B. The region contains an open reading frame, ORF5, which is transcribed into a 900-nucleotide (nt) RNA in both the parental wild-type strain and its derived petite, b7. This RNA uses as a template the strand used by most mt transcripts. Its start point is located 337 nt upstream of ORF5; and a messenger termination site has been found 900 nt downstream of the initiation site. These data suggest that ORF5 is a new mitochondrial gene. The G + C content of ORF5 is only 15.7%; 90% of the G + C base pairs of ORF5 are comprised in a palindromic G + C cluster similar to that present in the varl gene. The coding capacity of ORF5 is 46 amino acids (aa), mainly represented by methionine, phenylalanine, arginine, valine, asparagine, isoleucine and tyrosine. The aa composition and the codon usage of ORF5 are reminiscent of those of varl and of other intergenic ORFs.

Polymorphic variations in the ori sequences from the mitochondrial genomes of different wild-type yeast strains

We determined the restriction maps and primary structures of two as yet poorly characterized regions of the mitochondrial genomes of different wild-type strains of Saccharomyces cerevisiae. These regions respectively comprised the ori1 sequence and the newly identified ori8 sequence. Ori1 and ori8, together with their flanking sequences, exhibit a large polymorphism, resulting from specific variations due to insertions or deletions of optional GC clusters at different locations. The mechanisms underlying such sequence rearrangements are discussed.

The ori sequences of the mitochondrial genome of a wild-type yeast strain: number, location, orientation and structure

We have investigated the number, the location, the orientation and the structure of the seven ori sequences present in the mitochondrial genome of a wild-type strain, A, of Saccharomyces cerevisiae. These homologous sequences are formed by three G + C-rich clusters, A, B and C, and by four A + T-rich stretches. Two of the latter, p and s, are located between clusters A and B; one, l, between clusters B and C; and one r, either immediately follows cluster C (in ori 3-7), or is separated from it by an additional A + T-rich stretch, r', (in ori 1 and ori 2). The most remarkable differences among ori sequences concern the presence of two additional G + C-rich clusters, beta and gamma, which are inserted in sequence l of ori 4 and 6 and in the middle of sequence r of ori 4, 6 and 7, respectively. Neglecting clusters beta and gamma and stretch r', the length of ori sequences is 280 +/- 1 bp, and that of the l stretch 200 +/- 1 bp. Hairpin structures can be formed by the whole A-B region, by clusters beta and gamma, and (in ori 2-6) by a short AT sequence, lp, immediately preceding cluster beta. An overall tertiary folding of ori sequences can be obtained. Some structural features of ori sequences are shared by the origins of replication of the heavy strands of the mitochondrial genomes of mammalian cells.