Amphimeric mitochondrial genomes of petite mutants of yeast. III. Generation by linking two secondary-structure-dependent illegitimate recombination events.

The generation of amphimeric mitochondrial petite genomes of yeast can be explained by a process that links together two illegitimate recombination events, each involving a pair of short inverted repeats. Following "diagonal" double-strand breaks and inter-strand ligations at both possible stem-and-loop structures, a subgenomic single-stranded DNA circle can be excised. This circle comprises four building blocks organized in the so-called datA arrangement where d and t correspond, respectively, to the segments looped out by the upstream and the downstream pair of inverted repeats, a to the sequence separating the two loops, and A to the inverted duplication of segment a. Depending on the different possible "diagonal" recombinations at the inverted repeats, any of four isomeric circles can be excised, representing in its double-stranded form the nascent basic unit of an amphimeric mitochondrial petite genome of yeast. These isomeric basic units differ in the relative orientation of their sequences d and t (called D and T, respectively, when inverted), and are designated datA, DatA, daTA, and DaTA. Any one of these may be replicated to form the previously described regularly arrayed multimeric flip-flop genomes. Our new understanding of the amphimeric mitochondrial petite genomes of yeast emphasizes the role that topological features of DNA can play in mitochondrial genome dynamics. It also permits the re-interpretation of various observations reported in the literature. Some of them, including EtBr-mutagenesis in yeast, are discussed.

Amphimeric mitochondrial genomes of petite mutants of yeast. I. Flip-flop amphimers make up the mitochondrial genomes of "palindromic" petite mutants of yeast.

The mitochondrial (mt) genomes of three spontaneous cytoplasmic "palindromic" petite mutants of yeast were studied by restriction-enzyme analysis. These mt genomes were shown to be made up of an amplified "master basic unit" consisting of two inverted segments (a and A) and of two different unique segments (d and t) separating them. The basic unit was called "amphimeric", this term having been first proposed for certain lambda-phage mutants. We propose that in the mt genomes of the petite mutants studied, the four possible variants of the amphimeric basic unit form two - "flip" and "flop" - tetra-amphimeric repeat units datA-datA-DaTA-DaTA and DatA-DatA-daTA-daTA, respectively. These repeat units make two types of "amphimeric" mt genomes which exist in equal proportions in the cell. In each mt genome, the duplicated segment regularly alternates in its direct and inverted orientation (a...A...a...A...), whereas the unique segments are arranged twice in tandem fashion and twice in inverted fashion (d...d...D...D...d...d...andt...t...T...T...t...t...). The only difference between flip and flop amphimeric mt petite genomes is the different relative orientation of the unique segments in the mono-amphimers. In the mono-amphimers of flip mt genomes, both unique segments are arranged in the same direction (d...t and D...T), whereas in the mono-amphimers of flop mt genomes, both unique segments are arranged in opposite directions (D...t and d...T). Control experiments on one spontaneous petite mutant (which was an ancestor of the mutants studied here) and on three independent, previously investigated, EtBr-induced mutants showed that all of them were, in fact, organized in the same way. Analysing our experimental data and the results published by others, we conclude that amphimeric organization is a general feature of mt petite genomes of yeast previously called "palindromic" or "rearranged".

Amphimeric mitochondrial genomes of petite mutants of yeast. II. A model for the amplification of amphimeric mitochondrial petite DNA.

A model for the recombination-directed replication and amplification of the mtDNA of amphimeric petite mutants of S. cerevisiae is proposed. Replication of an amphimeric master basic unit datA would be initiated in the inverted components a and A. The initiation of replication should be associated with the amphimeric structure of the master basic unit itself, but could be promoted by the presence of ori sequences or of sequences facilitating the initiation of replication in the inverted duplications. The amplification unit of amphimeric genomes is considered to be the double-stranded circular hetero-diamphimer datA-DaTA. Amplification of both diamphimeric strands involves an invasion of the 3' ends of the newly synthesized strands into symmetrical homologous duplex DNA regions promoting the continuation of replication, and leads to the accumulation of two ("flip" and "flop") types of multi-amphimers. We consider that this mode of amplification represents a modified rolling-circle mechanism. By analogy, we propose to call our model of amplification the "rocking-circle model". This model is likely to apply to other genomes organized as amphimeric structures.

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.

Temperature can reversibly modify the structure and the functional efficiency of ori sequences of the yeast mitochondrial genome.

We have compared the suppressibility of three isonuclear spontaneous, cytoplasmic petite mutants of Saccharomyces cerevisiae, as measured at three temperatures, 23 degrees C, 28 degrees C and 33 degrees C. The three petites have mitochondrial genomes made up of repeat units which are about 400 bp in size, and carry an origin of replication, ori1. This ori sequence is intact in petite Z1, whereas it lacks GC cluster A in petite 26 and cluster A plus some contiguous nucleotides in petite 14. These deletions lead to the impossibility to form a stem-and-loop structure of the ori sequence, the 'A-B fold', which involves two GC clusters, A and B, and the nucleotides in between. Instead, a 'replacement fold', only involving AT base pairs, is feasible. In petites 14 and 26, suppressivity decreases when the temperature is raised from 28 degrees C to 33 degrees C, and increases when the temperature is lowered from 28 degrees C to 23 degrees C. In contrast, no changes are seen in petite Z1. These temperature effects correlate with the stability of the 'A-B fold' and the instability of the 'replacement folds'. Since suppressibility measures the replicative competitiveness of the petite genome relative to the wild-type genome, these results indicate that an environmental parameter, temperature, can reversibly affect the structure and the functional efficiency of ori sequences in vivo. The evolutionary implications of these findings are discussed.

Regions flanking ori sequences affect the replication efficiency of the mitochondrial genome of ori+ petite mutants from yeast.

The mitochondrial genomes of progenies from 26 crosses between 17 cytoplasmic, spontaneous, suppressive, ori+ petite mutants of Saccharomyces cerevisiae have been studied by electrophoresis of restriction fragments. Only parental genomes (or occasionally, genomes derived from them by secondary excisions) were found in the progenies of the almost 500 diploids investigated; no evidence for illegitimate, site-specific mitochondrial recombination was detected. One of the parental genomes was always found to be predominate over the other one, although to different extents in different crosses. This predominance appears to be due to a higher replication efficiency, which is correlated with a greater density of ori sequences on the mitochondrial genome (and with a shorter repeat unit size of the latter). Exceptions to the 'repeat-unit-size rule' were found, however, even when the parental mitochondrial genomes carried the same ori sequence. This indicates that noncoding, intergenic sequences outside ori sequences also play a role in modulating replication efficiency. Since in different petites such sequences differ in primary structure, size, and position relative to ori sequences, this modulation is likely to take place through an indirect effect on DNA and nucleoid structure.

Saccharomyces cerevisiae ARS on a plasmid from Staphylococcus aureus.

Staphylococcus aureus plasmid pC194 carries three sequences closely related to a consensus sequence defined previously by analysis of different genetic elements which replicate autonomously in yeast Saccharomyces cerevisiae. Two of these enable the plasmid to replicate in yeast, the third does not. A new consensus sequence A/T T T T A T R T T T, 1 bp shorter than the previous one, can be deduced from our results. Replacement of the T with G at the position 9 of the sequence abolishes its activity. The presence of the two active sequences on pC194 genome can be explained by the A + T-rich base composition of the plasmid.

Nuclear mutations affecting the stability of the mitochondrial genome in S. cerevisiae.

We have studied a pleiotropic mutation petD in S. cerevisiae which both confers the inability to grow on glycerol (Gly-) and greatly increases the frequency of cytoplasmic petites (Het). The first phenotype, Gly-, is recessive, whereas the second, Het, is dominant. Genetic and biochemical analysis showed that the majority of the petites in petD strains are not of the rho degree type (completely lacking mit-DNA), but of the rho- type (containing partially deleted mit-DNA). This finding and the fact that the phenotype Het is dominant argue in favour of the involvement of the petD product in the excision process of the mit-DNA. Another nuclear mutation, mod, was shown to exhibit a dominant epistasy with respect to the Het phenotype of the mutation petD. Two types of Gly+ revertants from petD mutants were isolated: rpa revertants, which restore completely the wild-type phenotype, and rpb revertants, which restore only the growth on glycerol, but still allow the production of high frequencies of cytoplasmic petites. Thus the mutations mod and rpb permit the genetic uncoupling of two phenotypes induced by the mutation petD.

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.

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

The ori sequences of the mitochondrial genomes of 20 wild-type strains of Saccharomyces cerevisiae were compared with those of the previously studied strain A (de Zamaroczy et al., 1984). The seven canonical ori sequences of this strain appear to be present in all strains tested, but in most strains ori1 is replaced by an extensively rearranged ori1 * sequence, and an additional ori sequence, ori8, is present between the oxi3 and the 15S RNA genes; one strain, B, lacks ori4. The location and orientation of ori sequences of three strains, B, C and K, were found to be the same as in strain A. The primary structures of four ori sequences from three different strains (ori1 of strain J69-1B, ori3 and ori5 of strain K, ori6 of strain D273-10B) were found to be identical with the corresponding ori sequences previously investigated. Hybridization experiments with different ori probes indicated a conservation of ori2-ori7 sequences in all strains tested. The primary structure of a petite genome derived from strain B and carrying ori1 * is reported and discussed.

Experimental hypothalamic or genetic obesity in the non-insulin-dependent diabetic rat.

Non-insulin-dependent diabetes was obtained in adult rats by neonatal administration of streptozotocin (100 mg/kg). Obesity was obtained in the same animals either by a ventromedial hypothalamic lesion in adult non-insulin-dependent diabetic Wistar rats, or by using genetically obese Zucker rats. In diabetic rats, weight gain was similar to that in non-diabetic rats, whether hyperphagia was due to a ventromedial hypothalamic lesion or to a genetic factor. Glucose-induced insulin release in vivo was increased in obese diabetic rats compared with non-diabetic rats. Despite this enhanced insulin secretion, both diabetic 'fatty' Zucker rats and diabetic rats with hypothalamic obesity showed a deterioration of glucose tolerance. Moreover, about one-third developed overt diabetes with permanent or transient glycosuria. We conclude that when insulin-deficient rats are made hyperphagic, they are able to increase their insulin secretion and become obese. In some of these animals the occurrence of obesity aggravates the diabetes. The obese diabetic rat appears to be a suitable laboratory model for the study of the relationship between obesity and diabetes.

Plasmids from Staphylococcus aureus replicate in yeast Saccharomyces cerevisiae.

It is known that some plasmids, such as RP4, can replicate in many Gram-negative bacteria. Certain small Staphylococcus aureus plasmids have an even broader host range, being able to replicate in not only phylogenetically distant Gram-positive bacteria such as Bacillus subtilis or Streptococcus pneumoniae, but also in the Gram-negative bacterium Escherichia coli. Here we have examined whether these plasmids can also replicate in a lower eukaryote, the yeast Saccharomyces cerevisiae. For this purpose we constructed hybrids between a S. aureus plasmid pC194 and an E. coli plasmid YIp5, which carries a ura-3 gene easy to select for in yeast but cannot replicate in this host. We found that the hybrids transformed yeast with high efficiency (as did hybrids between YIp5 and three other S. aureus plasmids); were maintained extrachromosomally in yeast; and were not modified during residence in yeast. We conclude from this evidence that S. aureus plasmids can replicate in yeast, which raises the questions of whether the replication signals used by prokaryotes and eukaryotes are similar, and how far up the phylogenetic tree the organisms still able to be hosts to S. aureus plasmids may be.

A region of extreme instability in the mitochondrial genome of yeast.

About half of the spontaneous petite mutants produced by wild-type Saccharomyces cerevisiae strain B (as well as by several other strains) have the same defective mitochondrial genome. Its repeat unit is a segment, 2200 base pairs (bp) long, which derives from an excision between the origins of replication ori 2 and ori 7 of the wild-type genome, and contains a hybrid ori 2-ori 7 sequence. The spontaneous petites carrying this defective ori.h genome are supersuppressive , i.e., they very rapidly compete out the wild-type genome in crosses. The main reasons for the exceptional frequency of ori.h petites are an extremely high excision frequency, due to the extended homology between the two tandemly oriented ori sequences 265 bp long and the short distance separating them. Such an excision frequency is very strongly increased in petite genomes encompassing the ori 2-ori 7 region, because of their higher concentration in these ori sequences.

Surrogate origins of replication in the mitochondrial genomes of ori-zero petite mutants of yeast.

We have investigated the mitochondrial genome of eight ori-zero spontaneous petite mutants of Saccharomyces cerevisiae. The tandem repeat units of these genomes do not contain any of the seven canonical ori sequences of the wild-type genome. Instead, they contain one, or more, ori-S sequences. These 44-nucleotide long surrogate origins of replication are a subset of GC clusters characterized by a potential secondary fold with two sequences ATAG and GGAG , inserted in AT spacers, two AT base pairs just following them, a GC stem (broken in the middle, and, in most cases also near the base, by non-paired nucleotides), and a terminal loop. This structure is reminiscent of that of GC clusters A and B from canonical ori sequences and supports the view (Bernardi, 1982a ) that the GC clusters of the mitochondrial genome arose, by an expansion process, from the canonical ori sequences. Like the latter, ori-S sequences are present in both orientations, are located in intergenic regions, and can be used as excision sequences when tandemly oriented. Again as in the case of canonical ori sequences, the density of ori-S sequences on the repeat units of petite genomes are correlated with the replication efficiency of the latter, as assessed by the outcome of crosses with wild-type or petite tester strains.

Effect of early and chronic hypoinsulinism on adipose tissue cellularity in the rat.

The effect of chronic hypoinsulinism on the development of retroperitoneal adipose tissue was studied in rats injected with streptozotocin at birth. The streptozotocin injection induced an acute neonatal diabetes which regressed spontaneously after one week and led to a chronic state of chemical diabetes in the young and in the adult rat. Growth of chemically diabetic rats was normal although the retroperitoneal adipose tissue showed a relative hypoplasia which appeared at two months and evolved with age so that at 10 months the number of adipose cells in the retroperitoneal adipose tissue was largely decreased with respect to control animals (1.34 +/- 0.12 x 10(6) versus 2.23 +/- 0.11 x 10(6)). This relative hypoplasia was still present at 20 months. Whereas the hypoplasia associated with the chemical diabetes was highly reproducible, the mean adipocyte size was modified in a variable manner but was never significantly decreased in chemically diabetic rats. These findings indicate that insulin is involved in the control of retroperitoneal adipose tissue cellularity and suggest that the effect of hypoinsulinism on adipocyte number does not depend on a decrease of the mean adipocyte volume.

Physical localisation and cloning of the structural gene for E. coli initiation factor IF3 from a group of genes concerned with translation.

The structural genes for translational initiation factor IF3, threonyl-tRNA synthetase (TRS), the two subunits of phenylalanyl-tRNA synthetase (PRS), and a 12 000 mol. wt. protein of unidentified function are carried by the lambda p2 transducing phage. The localization of these genes on a restriction map of the Escherichia coli DNA insert was achieved by deletion mapping. In addition a set of plasmids carrying fragments of the original phage was constructed and helped to confirm these assignments. One plasmid, containing a 3.3 kb PstI fragment inserted into pBR322, does not code for any of the synthetase genes but causes strains carrying it to overproduce IF3.

Supersuppressive "petite" mutants of yeast.

A class of suppressive "petite" mutants of S. cerevisiae, called here supersuppressive, is characterized by a) the fact that their unmodified mitochondrial genomes are the only ones found in the progeny of crosses with wild-type cells; b) very short repeat units (400-900 base pairs) in their mitochondrial genomes. The repeat units of the three supersuppressive "petites" investigated here share a common 83 nucleotide sequence, which seems to correspond to an initiation site of DNA replication; the multiple copies of this site in the mitochondrial genomes of supersuppressive "petites" might explain why these genomes can compete out those of wild-type cells.