In organello replication and viral affinity of linear, extrachromosomal DNA of the ascomycete Ascobolus immersus

Horizontal transfer of tRNA genes to mitochondrial plasmids facilitates gene loss from fungal mitochondrial DNA

Fungal and plant mitochondria are known to exchange DNA with retroviral plasmids. Transfer of plasmid DNA to the organellar genome is best known and occurs through wholesale insertion of the plasmid. Less well known is the transfer of organellar DNA to plasmids, in particular tRNA genes. Presently, it is unknown whether fungal plasmids can adopt mitochondrial functions such as tRNA production through horizontal gene transfer. In this paper, we studied the exchange of DNA between fungal linear plasmids and fungal mtDNA, mainly focusing on the basidiomycete family Lyophyllaceae. We report at least six independent transfers of complete tRNA genes to fungal plasmids. Furthermore, we discovered two independent cases of loss of a tRNA gene from a fungal mitochondrial genome following transfer of such a gene to a linear mitochondrial plasmid. We propose that loss of a tRNA gene from mtDNA following its transfer to a plasmid creates a mutualistic dependency of the host mtDNA on the plasmid. We also find that tRNA genes transferred to plasmids encode codons that occur at the lowest frequency in the host mitochondrial genomes, possibly due to a higher number of unused transcripts. We discuss the potential consequences of mtDNA transfer to plasmids for both the host mtDNA and the plasmid.

Anticodon nuclease encoding virus-like elements in yeast

A variety of yeast species are known to host systems of cytoplasmic linear dsDNA molecules that establish replication and transcription independent of the nucleus via self-encoded enzymes that are phylogenetically related to those encoded by true infective viruses. Such yeast virus-like elements (VLE) fall into two categories: autonomous VLEs encode all the essential functions for their inheritance, and additional, dependent VLEs, which may encode a toxin-antitoxin system, generally referred to as killer toxin and immunity. In the two cases studied in depth, killer toxin action relies on chitin binding and hydrophobic domains, together allowing a separate toxic subunit to sneak into the target cell. Mechanistically, the latter sabotages codon-anticodon interaction by endonucleolytic cleavage of specific tRNAs 3' of the wobble nucleotide. This primary action provokes a number of downstream effects, including DNA damage accumulation, which contribute to the cell-killing efficiency and highlight the importance of proper transcript decoding capacity for other cellular processes than translation itself. Since wobble uridine modifications are crucial for efficient anticodon nuclease (ACNase) action of yeast killer toxins, the latter are valuable tools for the characterization of a surprisingly complex network regulating the addition of wobble base modifications in tRNA.

Molecular evolution of a mitochondrial polB gene, encoding a family B DNA polymerase, towards the elimination from Agrocybe mitochondrial genomes

Two genes (Ac-polB O1and Ac-polB O2), each encoding a family B DNA polymerase, were characterized from the mitochondrial genome of the basidiomycete Agrocybe chaxingu. These two polB genes constitute orthologs of the potentially functional Aa-polB gene and its disrupted paralog Aa-polB P1, previously described in the closely related species A. aegerita. Unlike the case in Aa-polB, both gene copies in A. chaxingu are constituted by large but disrupted ORFs, which very probably encode nonfunctional enzymes: Ac-polB O1 has a deletion of 126 bp between the segments encoding the Exo II and Pol I domains and a 78-bp insertion between the Exo II and Exo III domains, whereas Ac-polB O2 has a large deletion of 1208 bp between the Exo II and Pol III domains and a deletion of 54-bp involving the 3' end of the gene. Hence, rearrangements in the Ac-polB ORFs appear to have led to their functional erosion in the mitochondrial genome in this species. Phylogenetic analysis has shown a close relationship between the mitochondrial polB genes and homologous genes carried by fungal linear plasmids, suggesting that they may have been acquired by the integration of linear plasmids into the mitochondrial genome.

Analysis of the structure and inheritance of a linear plasmid from the obligate biotrophic fungus Blumeria graminis f. sp. hordei

A linear plasmid is widespread among isolates of the obligate biotrophic fungus Blumeria graminis f.sp. hordei (synonym Erysiphe graminis) (Bgh), the organism that causes the disease powdery mildew on barley. We cloned and sequenced the entire plasmid of 7965 bp. The plasmid contains two identical terminal inverted repeats (TIR) of 610 bp. Two ORFs are present on opposite strands, one encoding a phage-type DNA polymerase and the other a phage-type RNA polymerase. Two large transcripts of approximately 4.2 and 5.6 kb were identified in conidia, germinating conidia and Bgh -infected barley leaves, indicating that the polymerases are transcribed at most stages of the lifecycle. The transcription start sites were localised within the TIR regions, where a putative 11-bp ARS consensus sequence was also identified. To follow the sexual transmission of the plasmid we screened 27 Bgh isolates for mitochondrial polymorphisms. One polymorphism allowed us to carry out a cross between two isolates that differed in both mitochondrial genotype and presence/absence of the Bgh plasmid. The plasmid was transmitted independently of the origin of the mitochondria. No transfer of the plasmid was observed between two Bgh isolates that were co-cultivated for 1.5 years on a common susceptible barley variety. The plasmid appears to be an autonomous replicon with no phenotypic effect on Bgh.

In vitro replication of mitochondrial plasmid mp1 from the higher plant Chenopodium album (L.): a remnant of bacterial rolling circle and conjugative plasmids?

According to the endosymbiotic theory, mitochondrial genomes evolved from the chromosome of an alpha-proteobacterium-like ancestor and developed during evolution an extraordinary variation in size, structure and replication. We studied in vitro DNA replication of the mitochondrial circular plasmid mp1 (1309 bp) from the higher plant Chenopodium album (L.) as a model system that replicates in a manner reminiscent of bacterial rolling circle plasmids. Several mp1 subclones were tested for their ability to support DNA replication using a newly developed in vitro system. Neutral/neutral two-dimensional gel electrophoresis of the in vitro products revealed typical simple Y patterns of intermediates consistent with a rolling circle type of replication. Replication activity was very high for a BamHI-restricted total plasmid DNA clone, a 464 bp BamHI/KpnI fragment and a 363 bp BamHI/SmaI fragment. Further subcloning of a 148 bp BamHI/EcoRI fragment resulted in the strongest in vitro DNA replication activity, while a 1161 bp-template outside of this region resulted in a substantial loss of activity. Electron microscopic studies of in vitro DNA replication products from the highly active clones also revealed sigma-shaped molecules. These results support our in vivo data for the presence of a predominant replication origin between positions 628 and 776 on the plasmid map. This sequence shares homology with double-stranded rolling circle origin (dso) or transfer origin (oriT) nicking motifs from bacterial plasmids. mp1 is the first described rolling circle plasmid in eukaryotes.

The Cryphonectria parasitica plasmid pUG1 contains a large ORF with motifs characteristic of family B DNA polymerases

The isolation and characterization of the circular mitochondrial plasmid pUG1 from the ascomycete Cryphonectria parasitica is described. The entire sequence (4182 bp) was obtained and high similarities to DNA-dependent DNA polymerases were revealed. Strikingly common features with the DNA polymerases encoded by the Neurospora intermedia plasmids Fiji and LaBelle, such as matches to the conserved motifs A and B and the presence of TTD instead of DTD in motif C, were found, suggesting the existence of a distinct group of members of the B DNA family polymerases. These strong similarities between the plasmids might suggest a common origin of the C.parasitica and the Neurospora plasmids.

The terminal protein of the linear DNA plasmid pGKL2 shares an N-terminal domain of the plasmid-encoded DNA polymerase

The 36K protein attached at the 5' end of the linear DNA plasmid pGKL2 from the yeast Kluyveromyces lactis was first purified and characterized. The terminal protein was purified from cells (1 kg wet weight) by ammonium sulphate precipitation and two rounds of centrifugation to equilibrium in CsCl gradients. The pGKL2 was present only in the post-microsomal supernatant. Approximately 10 mg of the purified pGKL2 was recovered and digested with DNase I. The terminal protein (final ca. 0 center dot 8 mg) was homogeneous by electrophoresis and we determined the N-terminal amino acid sequence up to ten residues, showing that it existed in the cryptic N-terminal domain of pGKL2-ORF2 (DNA polymerase) sequence.

Natural plasmids of filamentous fungi

Among eukaryotes, plasmids have been found in fungi and plants but not in animals. Most plasmids are mitochondrial. In filamentous fungi, plasmids are commonly encountered in isolates from natural populations. Individual populations may show a predominance of one type, but some plasmids have a global distribution, often crossing species boundaries. Surveys have shown that strains can contain more than one type of plasmid and that different types appear to be distributed independently. In crosses, plasmids are generally inherited maternally. Horizontal transmission is by cell contact. Circular plasmids are common only in Neurospora spp., but linear plasmids have been found in many fungi. Circular plasmids have one open reading frame (ORF) coding for a DNA polymerase or a reverse transcriptase. Linear plasmids generally have two ORFs, coding for presumptive DNA and RNA polymerases with amino acid motifs showing homology to viral polymerases. Plasmids often attain a high copy number, in excess of that of mitochondrial DNA. Linear plasmids have a protein attached to their 5' end, and this is presumed to act as a replication primer. Most plasmids are neutral passengers, but several linear plasmids integrate into mitochondrial DNA, causing death of the host culture. Inferred amino acid sequences of linear plasmid ORFs have been used to plot phylogenetic trees, which show a fair concordance with conventional trees. The circular Neurospora plasmids have replication systems that seem to be evolutionary intermediates between the RNA and the DNA worlds.

Horizontal transfer of a mitochondrial plasmid

Direct evidence for horizontal transfer of a mitochondrial plasmid from the discomycete Ascobolus immersus to the pyrenomycete Podospora anserina is presented. Southern blot hybridisation analysis, polymerase chain reaction (PCR) amplification, and DNA sequencing demonstrate transmission of a linear plasmid upon hyphal contact. DNA extraction from isolated organelles indicates a mitochondrial localisation for the plasmid in P. anserina. This is the first report of horizontal gene transfer among unrelated fungi. These results have important evolutionary implications for plasmid propagation in fungi.

A novel family of linear plasmids with homology to plasmid pAL2-1 of Podospora anserina

Three recently isolated wild-type strains of the ascomycete Podospora anserina were analyzed for the presence of linear mitochondrial plasmids. In one of these strains, designated Wat, at least 12 distinct plasmid-like elements were identified. From molecular analyses a minimum number of 78 individual linear molecules with proteins bound to their 5' ends was estimated. In addition, the different members of this family of typical linear plasmids were shown to possess a common central region and terminal sequences which differ from one plasmid to another due to the presence of different numbers of a 2.4 kb sequence module. Finally, the pWa6 plasmids share a high degree of sequence similarity with pAL2-1, a linear plasmid previously identified in mitochondria of a long-lived mutant of P.anserina. A mechanism is proposed which explains the generation of these distinct, closely related extra-chromosomal genetic traits.

Evidence for giant linear plasmids in the ascomycete Podospora anserina

In the extrachromosomal mutant AL2 of the ascomycete Podospora anserina longevity is correlated with the presence of the linear mitochondrial plasmid pAL2-1. In addition to this autonomous genetic element, two types of closely related pAL2-1-homologous molecules were detected in the high-molecular-weight mitochondrial DNA (mtDNA). One of these molecules is of linear and the other of circular structure. Both molecules contain pAL2-1 sequences which appear to be integrated at the same site in the mtDNA. Sequence analysis of a DNA fragment cloned from one of these molecules revealed that it contains an almost full-length copy of pAL2-1. At the site of plasmid integration a 15-nucleotide AT-spacer and long inverted mtDNA sequences were identified. Finally, two giant linear plasmid-like DNAs of about 50 kbp and 70 kbp were detected in pulsed-field gels of mutant AL2. These molecules are composed of mtDNA and pAL2-1-specific sequences and may result from the integration of mtDNA sequences into linear plasmid pAL2-1.

Linear plasmids, pLm9 and pLm10, can be isolated from the phytopathogenic ascomycete Leptosphaeria maculans by pulsed-field gel electrophoresis

Two linear DNA plasmids (pLm9 and pLm10, sized 9 and 10 kb respectively) were isolated from the phytopathogenic ascomycete Leptosphaeria maculans, using pulsed-field gel electrophoresis. pLm9 and pLm10 are found only in aggressive isolates of L. maculans but, because aggressive and non-aggressive strains appear to be different species, these plasmids are probably not involved in pathogenicity. pLm9 and pLm10 copurify with a mitochondrially-enriched cell fraction, and do not hybridise to chromosomal or mitochondrial DNA, or to each other. Exonuclease digestions suggest that both these molecules contain covalently-bound proteins at their 5' termini. pLm9 hybridises to the RNA polymerase of a linear plasmid from the ascomycete Podospora anserina, and pLm10 hybridises to the DNA polymerase from the same P. anserina plasmid, suggesting that pLm9 and pLm10 encode their own replication and transcription enzymes.

Complex terminal structure of a linear mitochondrial plasmid from Physarum polycephalum: three terminal inverted repeats and an ORF encoding DNA polymerase

The mitochondria of Physarum polycephalum have a linear plasmid (mF) which promotes mitochondrial fusion. To determine the terminal structure of the mF plasmid, restriction fragments derived from its ends were cloned and sequenced. The sequences showed that the mF plasmid has three kinds of terminal inverted repeats (TIRs). The most characteristic feature is a 144-bp repeating unit which exists between a 205-bp TIR at the extreme ends of the plasmid and another 591-bp TIR. All of the clones showed at least one of these 144-bp repeating units. The GC content of the 205-bp TIR (49%) was higher than those of the other TIRs and of another sequenced region (23%). This TIR can form three thermodynamically-stable hairpin structures based on complex internal palindromic components. Moreover, in the right terminal region of the mF plasmid, there is an open reading frame (ORF) which covers the entire 591-bp TIR and most of one of the 144-bp repeating units. This ORF encodes a 547-amino-acid polypeptide, ORF-547, and shows extensive homology with the polymerization domain of the putative DNA polymerases of linear mitochondrial plasmids from other sources.

Three mitochondrial unassigned open reading frames of Podospora anserina represent remnants of a viral-type RNA polymerase gene

The mitochondrial DNA of Podospora anserina is complex, consisting of a characteristic set of genes with a large number of introns and a substantial amount of sequence of unknown function and origin. In addition, as indicated by various types of reorganization, this genome is highly flexible. Here we report the identification of three unassigned mitochondrial open reading frames (ORF P', ORF Q', ORF 11) as remnants of a rearranged viral-type RNA polymerase gene. These ORFs are not transcribed and may be derived from the integration of a linear plasmid of the type recently identified in a mutant of P. anserina.

Genetic identification and nucleotide sequence of the DNA polymerase gene of African swine fever virus

The DNA polymerase gene of African swine fever virus (ASFV) was mapped by marker rescue experiments using a phosphonoacetic acid-resistant mutant and hybridization with an oligonucleotide probe designed from the most conserved motif of family B DNA polymerases. Viral DNA fragments mapping in this region were cloned and sequenced. An open reading frame coding for a 1244 amino acid long peptide with a molecular mass of 142.5 kDa was determined from the sequence. A unique feature of ASFV DNA polymerase is the presence of 13 tandem repeats of the sequence Ala-Gly-Asp-Pro near the carboxyl end of the molecule. Comparison with 30 sequences of alpha-like DNA polymerases of cellular and viral origin showed that ASFV DNA polymerase has all the conserved motifs of family B DNA polymerases. A 3.9 kb transcript was detected by Northern hybridization and the transcription initiation and termination sites were mapped by S1 analysis and primer extension. Late transcription was initiated at a site different from the early transcription initiation site. A 145 kDa protein, consistent with the size of the gene, was identified by an in situ enzyme assay after gel electrophoresis of infected cell extracts.

Two Neurospora mitochondrial plasmids encode DNA polymerases containing motifs characteristic of family B DNA polymerases but lack the sequence Asp-Thr-Asp

We have determined the DNA sequence of the mitochondrial plasmid from Neurospora intermedia strain Fiji N6-6. The plasmid contains a 1278-codon open reading frame that is 49% identical to the open reading frame of the mitochondrial plasmid from the LaBelle strain of N. intermedia, which is known to encode a DNA-dependent DNA polymerase. The results of polymerase assays and photolabeling studies, the high degree of identity with the LaBelle plasmid polymerase, and the observation that the Fiji polymerase activity in a reaction utilizing endogenous template is not affected by removal of RNA suggest that the Fiji plasmid also encodes a DNA-dependent DNA polymerase. Comparison of regions of amino acids that are highly conserved in the two plasmid polymerases to family B polymerases reveals good correlates for the three major polymerase motifs and suggests that previously identified motifs characteristic of reverse transcriptase found in the LaBelle sequence are not significant. The polymerases encoded by the Fiji and LaBelle plasmids are unusual in that the amino acid sequence Asp-Thr-Asp, which forms the core of the third motif in family B polymerases, is not present in either Fiji or LaBelle. A version of the motif containing Thr-Thr-Asp exists in both sequences.

Interaction between mitochondrial DNA and mitochondrial plasmids in Claviceps purpurea: analysis of plasmid-homologous sequences upstream of the lrRNA-gene

Homology of two linear, mitochondrial (mt) Claviceps purpurea plasmids, pClK1 and pClT5, to the upstream region of the large ribosomal RNA gene in the mtDNA of three strains (W3, T5 and K) has been investigated in detail to explore the widespread phenomenon of homology between mt plasmids and mtDNA in C. purpurea. Sequence comparison indicates that recombination between free plasmids and mtDNA is the cause of the observed homology. The process is similar to the integration of the structurally related adenoviruses into the mammalian genome. As in other fungi, palindromic sequences seem to be involved in this mitochondrial recombination process.

Organization and nucleotide sequence of the DNA polymerase gene from the archaeon Pyrococcus furiosus

We cloned the gene encoding the thermostable DNA polymerase from the archaeon Pyrococcus furiosus. The DNA fragment of 2785 base pair (bp) containing the structural gene for DNA polymerase was sequenced. DNA polymerase (Pfu polymerase), as deduced from the DNA sequence, consisted of 775 amino acids, had a molecular weight of 90, 109, and was structurally homologous to the alpha-like DNA polymerases (family B) represented by human DNA polymerase alpha and Escherichia coli DNA polymerase II. An unrooted phylogenetic tree of the alpha-like DNA polymerases based on the amino acid sequence alignment was constructed. Pfu polymerase, with two other archaeon polymerases, constitutes a group with some animal viruses. The transcription initiation sites of the pol gene were identified by analysis of in vivo transcripts of both from P. furiosus and E. coli, and the promoters were assigned upstream of the pol coding region. A typical promoter sequence for the archaeon was found at a reasonable distance from the transcription initiation site in P. furiosus.

Evolution of linear plasmids

Linear plasmids are genetic elements commonly found in yeast, filamentous fungi, and higher plants. In contrast to all other plasmids they possess terminal inverted repeats and terminal bound proteins and encode their own DNA and RNA polymerases. Here we present alignments of conserved amino acid sequences of both the DNA and RNA polymerases encoded by those linear plasmids for which DNA sequence data are available. Additionally these sequences are compared to a number of polymerases encoded by related viral and cellular entities. Phylogenetic trees have been established for both types of polymerases. These trees appear to exhibit very similar subgroupings, proving the accuracy of the method employed.

The linear mitochondrial plasmid pAL2-1 of a long-lived Podospora anserina mutant is an invertron encoding a DNA and RNA polymerase

The molecular characterization of an additional DNA species (pAL2-1) which was identified previously in a long-lived extrachromosomal mutant (AL2) of Podospora anserina revealed that this element is a mitochondrial linear plasmid. pAL2-1 is absent from the corresponding wild-type strain, has a size of 8395 bp and contains perfect long terminal inverted repeats (TIRs) of 975 bp. Exonuclease digestion experiments indicated that proteins are covalently bound at the 5' termini of the plasmid. Two long, non-overlapping open reading frames, ORF1 (3,594 bp) and ORF2 (2847 bp), have been identified, which are located on opposite strands and potentially encode a DNA and an RNA polymerase, respectively. The ORF1-encoded polypeptide contains three conserved regions which may be responsible for a 3'-5' exonuclease activity and the typical consensus sequences for DNA polymerases of the D type. In addition, an amino-acid sequence motif (YSRLRT), recently shown to be conserved in terminal proteins from various bacteriophages, has been identified in the amino-terminal part of the putative protein. According to these properties, this first linear plasmid identified in P. anserina shares all characteristics with invertrons, a group of linear mobile genetic elements.

Genetic organization and structural features of maranhar, a senescence-inducing linear mitochondrial plasmid of Neurospora crassa

The nucleotide sequence of maranhar, a senescence-inducing linear mitochondrial plasmid of Neurospora crassa, was determined. The termini of the 7-kb plasmid are 349-bp inverted repeats (TIRs). Each DNA strand contains a long open reading frame (ORF) which begins within the TIR and extends toward the centre of the plasmid. ORF-1 codes for a single-subunit RNA polymerase that is not closely related to that encoded by another Neurospora plasmid, kalilo. The ORF-2 product may be a B-type DNA polymerase resembling those encoded by terminal protein-linked linear genetic elements, including linear mitochondrial plasmids and linear bacteriophages. A separate coding sequence for the terminal protein could not be identified; however, the DNA polymerase of maranhar has an amino-terminal extension with features that are also present in the terminal proteins of linear bacteriophages. The N-terminal extensions of the DNA polymerases of other linear mitochondrial plasmids contain similar features, suggesting that the terminal proteins of linear plasmids may be comprised, at least in part, of these cryptic domains. The terminal protein-DNA bond of maranhar is resistant to mild alkaline hydrolysis, indicating that it might involve a tyrosine or a lysine residue. Although maranhar and the senescence-inducing kalilo plasmid of N. intermedia are structurally similar, and integrate into mitochondrial DNA by a mechanism thus far unique to these two plasmids, they are not closely related to each other and they do not have any nucleotide sequence features, or ORFs, that distinguish them clearly from mitochondrial plasmids which are not associated with senescence and most of which are apparently non-integrative.

A deletion derivative of the kalilo senescence plasmid forms hairpin and duplex DNA structures in the mitochondria of Neurospora

A novel deletion derivative, delta kal, of the kalilo senescence plasmid from Neurospora intermedia, was recovered from a culture treated with chloramphenicol. The deletion derivative exists in mitochondria as two different, equally abundant forms: a 2.8 kb duplex DNA molecule (delta kal-2.8) and a 1.4 kb hairpin form (delta kal-1.4). The delta kal-2.8 plasmid contains the 1366 bp terminal inverted repeats and a partially duplicated 102 bp segment of the unique sequence of the 8.6 kb kalilo plasmid. In contrast, the delta kal-1.4 hairpin plasmid appears to result from the folding of single strands that are generated during the replication of delta kal-2.8. Both forms of delta kal have covalently linked terminal proteins. Sequence analysis suggests that delta kal was generated either by slippage of the tip of a growing strand during the replication of kalilo, or by illegitimate recombination between two copies of the plasmid at non-homologous palindromic sequences that might form cruciform structures. In either case, the deletion process was mediated at least in part by an inverted repeat of 5 bp in the unique region of kalilo. Since the terminal segments of kalilo DNA that are implicated in plasmid integration might also form cruciform structures, it is possible, but improbable, that the process that generated the first delta kal molecule is related to that which mediates integration of the plasmid into mitochondrial DNA.

Identification of Acanthamoeba at the generic and specific levels using the polymerase chain reaction

We have adapted the polymerase chain reaction to identify strains of Acanthamoeba. Using computer-assisted analysis, primers were designed from an anonymous repetitive sequence and from published sequences of 18S and 5S ribosomal RNA genes of A. castellanii. Amplification of a short ribosomal DNA target (272 base pairs) at restrictive annealing conditions (greater than 50 degrees C) resulted in a single band that was unique for the genus and distinguished Acanthamoeba from Naegleria. This assay functioned with fresh and formalin-fixed cells as starting material. Amplification of longer targets (400-700 base pairs) at less restrictive annealing conditions (less than 47 degrees C) led to more than one band. This multiple banding pattern could reproducibly classify Acanthamoeba at the strain level and was, in certain cases, diagnostic for known pathogenic strains. However, these assays need to be further refined to make them relevant for clinical purposes.

The kalilo linear senescence-inducing plasmid of Neurospora is an invertron and encodes DNA and RNA polymerases

The nucleotide sequence of kalilo, a linear plasmid that induces senescence in Neurospora by integrating into the mitochondrial chromosome, reveals structural and genetic features germane to the unique properties of this element. Prominent features include: (1) very long perfect terminal inverted repeats of nucleotide sequences which are devoid of obvious genetic functions, but are unusually GC-rich near both ends of the linear DNA; (2) small imperfect palindromes that are situated at the termini of the plasmid and are cognate with the active sites for plasmid integration into mtDNA; (3) two large, non-overlapping open-reading frames, ORF-1 and ORF-2, which are located on opposite strands of the plasmid and potentially encode RNA and DNA polymerases, respectively, and (4) a set of imperfect palindromes that coincide with similar structures that have been detected at more or less identical locations in the nucleotide sequences of other linear mitochondrial plasmids. The nucleotide sequence does not reveal a distinct gene that codes for the protein that is attached to the ends of the plasmid. However, a 335-amino acid, cryptic, N-terminal domain of the putative DNA polymerase might function as the terminal protein. Although the plasmid has been co-purified with nuclei and mitochondria, its nucleotide composition and codon usage indicate that it is a mitochondrial genetic element.

High-frequency transfer of linear DNA containing 5'-covalently linked terminal proteins: electroporation of bacteriophage PRD1 genome into Escherichia coli

Using electroporation with the phage PRD1 genome, we set up a high-frequency DNA transfer system for a linear dsDNA molecule with 5'-covalently linked terminal proteins. The transfer was saturated when more than 100 ng of PRD1 genome was used. Electroporation efficiency was about four orders of magnitude higher than that obtained with transfection. Removal of the terminal protein abolished plaque formation, which could not be rescued by supplying the terminal protein or phage DNA polymerase or both in trans.

Escherichia coli DNA polymerase II is homologous to alpha-like DNA polymerases

The Escherichia coli polB gene encodes DNA polymerase II and is regulated by the SOS system. We sequenced a 4081 nucleotide segment of the E. coli chromosome that contains the polB gene and its flanking regions. DNA polymerase II, as deduced from the DNA sequence, consists of 782 amino acids, has a molecular weight of 89,917, and is structurally homologous to alpha-like DNA polymerases, which include eukaryotic replicative DNA polymerases. Comparison of the sequences of the alpha-like DNA polymerases including E. coli DNA polymerase II showed that there were nine highly conserved regions, and we constructed an unrooted phylogenetic tree of the DNA polymerases based on the differences in these conserved regions. The DNA polymerases of herpes groups viruses and the DNA polymerases that use protein priming for the initiation of replication form two separate subfamilies that occupy opposite locations in the tree. Other DNA polymerases, including E. coli DNA polymerase II, human DNA polymerase alpha, and yeast DNA polymerase I, occupy the central regions between the two subfamilies and they are rather distantly related to each other. The transcription initiation site of polB was identified by analysis of in vivo transcripts, and the promoter was assigned upstream of the polB coding region. The recognition sequence of the LexA repressor (SOS box) was identified by a footprinting experiment. It overlaps the -35 sequence of the polB promoter.

A general structure for DNA-dependent DNA polymerases

In addition to the general 3'-5' exonuclease domain described by Bernad et al. [Cell 59 (1989) 219-228] significant amino acid (aa) sequence similarity has been found in the C-terminal portion of 27 DNA-dependent DNA polymerases belonging to the two main superfamilies: (i) Escherichia coli DNA polymerase I (PolI)-like prokaryotic DNA polymerases, and (ii) DNA polymerase alpha-like prokaryotic and eukaryotic (viral and cellular) DNA polymerases. The six most conserved C-terminal regions, spanning approx. 340 aa, are located in the same linear arrangement and contain highly conserved motifs and critical residues involved in the polymerization function. According to the three-dimensional model of PolIk (Klenow fragment), these six conserved regions are located in the proposed polymerization domain, forming the metal and dNTP binding sites and the cleft for holding the DNA template. Site-directed mutagenesis in the phi 29 DNA polymerase supports some of these structural predictions. Therefore, it is likely that a 'Klenow-like core', containing the DNA polymerase and 3'-5' exonuclease activities, has evolved from a common ancestor, giving rise to the present-day prokaryotic and eukaryotic DNA polymerases.

Site-directed mutagenesis of the YCDTDS amino acid motif of the phi 29 DNA polymerase

The Bacillus subtilis phage phi 29 DNA polymerase, involved in protein-primed viral DNA replication, contains amino acid consensus sequences common to other alpha-like DNA polymerases. Using site-directed mutagenesis we have studied the functional significance of the most conserved C-terminal segment mainly represented by the YCDTDS motif. A series of single point mutants has been constructed and the corresponding proteins have been overproduced and characterized. Measurements, on crude fractions, of the activity of the mutant proteins in the formation of the protein p3-dAMP initiation complex and in an in situ DNA polymerase assay, indicate that the YCDTDS domain is involved both in initiation and in elongation reactions.

The highly conserved amino acid sequence motif Tyr-Gly-Asp-Thr-Asp-Ser in alpha-like DNA polymerases is required by phage phi 29 DNA polymerase for protein-primed initiation and polymerization

The alpha-like DNA polymerases from bacteriophage phi 29 and other viruses, prokaryotes and eukaryotes contain an amino acid consensus sequence that has been proposed to form part of the dNTP binding site. We have used site-directed mutants to study five of the six highly conserved consecutive amino acids corresponding to the most conserved C-terminal segment (Tyr-Gly-Asp-Thr-Asp-Ser). Our results indicate that in phi 29 DNA polymerase this consensus sequence, although irrelevant for the 3'----5' exonuclease activity, is essential for initiation and elongation. Based on these results and on its homology with known or putative metal-binding amino acid sequences, we propose that in phi 29 DNA polymerase the Tyr-Gly-Asp-Thr-Asp-Ser consensus motif is part of the dNTP binding site, involved in the synthetic activities of the polymerase (i.e., initiation and polymerization), and that it is involved particularly in the metal binding associated with the dNTP site.