The terminal structure of the linear mitochondrial DNA (mtDNA) from the yeast Candida parapsilosis was investigated. This mtDNA, 30 kb long, has symmetrical ends forming inverted terminal repeats. These repeats are made up of a variable number of tandemly repeating units of 738 bp each; the terminal nucleotide corresponds to a precise position within the last repeat unit sequence. The ends had an open structure accessible to enzymes, with a 5' single-stranded extension of about 110 nucleotides. No circular forms were detected in the DNA preparations. Two other unrelated species, Pichia philodendra and Candida salmanticensis also appear to have a linear mtDNA of similar organization. These linear DNAs (which we name Type 2 linear mtDNAs) are distinct from the previously described linear mtDNAs of yeasts whose termini are formed by a closed hairpin loop (Type 1 linear mtDNA). The terminal structure of C. parapsilosis mtDNA is reminiscent of the linear mitochondrial genomes of the ciliate Tetrahymena although, in the latter, the telomeric tandem repeat unit is considerably shorter.

In most yeast species, the mitochondrial DNA (mtDNA) has been reported to be a circular molecule.However, two cases of linear mtDNA with specific termini have previously been described. We examined the frequency of occurrence of linear forms of mtDNA among yeasts by pulsed-field gel electrophoresis. Among the 58 species from the genera Pichia and Wiliopsis that we examined, linear mtDNA was found with unexpectedly high frequency. Thirteen species contained a linear mtDNA, as confirmed by restriction mapping, end labeling, and electron microscopy. The mtDNAs from Pichia pijperi, WiUliopsis mrakii, and P. jadinii were studied in detail. In each case, the left and right terminal fragments shared homologous sequences. Between the terminal repeats, the order of mitochondrial genes was the same in all of the linear mtDNAs examined, despite a large variation of the genome size. This constancy of gene order is in contrast with the great variation of gene arrangement in circular mitochondrial genomes of yeasts. The coding sequences determined on several genes were highly homologous to those of the circular mtDNAs, suggesting that these two forms of mtDNA are not of distant origins.Mitochondria have their own DNA which specifies several proteins of the respiratory complexes as well as mitochondrion-specific rRNAs and tRNAs. While the majority of these genomes seem to be circular, some organisms have been shown to have a linear mitochondrial DNA (mtDNA). Well-known examples are those of the ciliates of the genera Paramecium (7) and Tetrahymena (17). The first case of a linear mtDNA in yeasts was found in 1981 in Wlliopsis mrakii (formerly called Hansenula mrakii) (19). A second case was reported for a yeast identified as Candida parapsilosis (8; see also reference 1). Occurrence of linear DNAs in yeast mitochondria raises questions about their origin and mode of replication. In a number of other organelle genomes, linear forms of DNA have also been described, although in some cases, the reality of the linearity might be questioned because the terminal structure of these forms has not been defined. The evolutionary and taxonomical positions of these linear genomes with respect to the circular forms need to be clarified. These linear DNAs may replicate as a linear molecule, or they may replicate through circular intermediates. In this study, we examined (i) the frequency of the linear forms of mtDNAs among yeast species and (ii) whether they have any basic structures in common. In the accompanying report (4), we examine their possible modes of replication by analysis of their terminal structure. MATERIALS AND METHODSYeast strains. Yeast strains were obtained from Centraalbureau voor Schimmelcultures (CBS), Delft, The Netherlands. They are independent isolates of known origins. As a rule, the type strain of each species, as indicated in the CBS 1990 catalog, was chosen for analysis unless otherwise stated. The culture medium contained 1% yeast extract * Corresponding author.

The terminal structure of the linear mitochondrial DNA (mtDNA) from three yeast species has been examined. By enzymatic digestion, alkali denaturation, and sequencing of cloned termini, it was shown that in Pichia piperi and P. jadinii, both termini of the linear mtDNA were made of a single-stranded loop covalently joining the two strands, as in the case of vaccinia virus DNA. The left and right loop sequences were in either of two orientations, suggesting the existence of a flip-flop inversion mechanism. Contiguous to the terminal loops, inverted terminal repeats were present. The mtDNA from WlUiopsis mrakii seems to have an analogous structure, although terminal loops could not be directly demonstrated. Electron microscopy revealed the presence, among linear molecules, of a small number of circular DNAs, mostly of monomer length. Linear and circular models of replication are considered, and possible conversion mechanisms between linear and circular forms are discussed. A flip-flop inversion mechanism between the inverted repeat sequences within a circular intermediate may be involved in the generation of the linear form of mtDNA.In the accompanying report (8), we show that linear forms of mitochondrial DNA (mtDNA) are frequently found in yeast species. These genomes have a defined terminal structure, as judged from their constant gene order with respect to the termini and the presence of homologous sequences at both termini. In this study, we analyzed in more detail the structure of the termini, as this information is essential to know the origin and mode of replication of the linear genomes. In a number of other organelle genomes, linear forms of DNA have also been identified. In some cases, the reality of the linearity might be questioned. In the case of Pichia pijperi, which had the smallest linear mtDNA, we could demonstrate, by sequencing, that the termini of the linear molecule possessed a continuous hairpin structure linking the termini of the two strands. The results obtained with two other species also suggested that an analogous structure may form the ends of their linear mtDNAs. MATERIALS AND METHODSStrains, media, and most of the analytical procedures are described in the accompanying report (8). Exonuclease III (ExoIII), nuclease Si, and nuclease BAL 31 were obtained from New England Biolabs (Beverly, Mass.) and used according to the supplier's instruction. The pTZ18/19R vector system (Pharmacia France, Saint Quentin-en-Yvelines, France) was used for cloning and enzymatic DNA sequencing. Electrophoresis of denatured DNA was carried out as described elsewhere (21). For electron microscopy, restriction fragments from termini were treated with the gene 32 protein of Eschenchia coli bacteriophage T4 as described previously (7) and then deposited on a carbon-coated grid that had been activated by glow discharge in the presence of pentylamine as described previously (7). mtDNA denaturation was performed according to the formamide-glyoxal method, and denatured DNA was spread as in the case of native DNA ...