DDBJ/EMBL/GenBank accession no. Y12637
ABSTRACT
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.
Plasmids are very commonly found in filamentous fungi where they generally have a mitochondrial localization (1 ). Some plasmids represent defective forms of mitochondrial DNA (mtDNA), while others lack any homology to the mtDNA.
The majority of the plasmids so far identified, belonging to the second type, are linear but their functions are still unclear with a few exceptions such as the kalilo and maranhar DNA of Neurospora spp. and pAL2-1 of Podospora anserina (2 ).
Circular plasmids have been reported to occur in fungi, and to date they include only the Mauriceville, Fiji, LaBelle, Java, Mb1, VS, Harbin 2 plasmids in Neurospora spp. (3 ), the plasmids of Absidia glauca (4 ) and of Pythium spp. (5 ). The circular plasmids are not associated with a particular phenotypic trait with the exception of a plasmid of A.glauca which encodes for a surface protein present only in mating type positive strains harbouring the plasmid (4 ).
In the plant pathogenic fungus Cryphonectria parasitica a circular plasmid, pUG1, has been found in virulent strains exhibiting a senescent phenotype of the colony reported as heteroauxesis (6 ). Morphological alterations of plasmid-containing virulent strains of C.parasitica appeared during prolonged vegetative propagation as abnormalities of the colony development. The phenomenon also includes cessation of growth of the colony with limitations of the morphological development and possibly consequences on the dynamics of the development of the pathogen in the environment (7 ).
Cryphonectria parasitica has a great economic and ecological relevance to forestry as the causal agent of Chestnut Blight, the severe disease that destroyed hectars of chestnut stands mostly in USA and Europe (8 ,9 ). Transmissible hypovirulence, i.e. a dramatic decrease in pathogenicity, is associated with an extrachromosomal viral ds-RNA (10 ) and characterizes hypovirulent strains of the fungus, making biological control treatments possible and efficient in Europe (9 ,11 ).
The Neurospora plasmids are some of the best studied and characterized plasmids, nonetheless their origin, functions and the mechanisms of their maintenance are still not well understood. A reverse transcriptase activity has been proposed to explain the formation and replication of the Mauriceville homology group plasmids (12 ), while the LaBelle and Fiji plasmids were found to encode a DNA-dependent DNA polymerase (13 ) like other linear plasmids such as kalilo from Neurospora intermedia (14 ), maranhar from Neurospora crassa (15 ), pEM from Agaricus bitorquis (16 ), pClK1 from Claviceps purpurea (17 ), pAI2 from Ascobolus immersus (18 ) and S1 from maize (19 ).
In the present study, the first reported isolation, the characterization and complete sequence of pUG1 are given. The plasmid shows a high degree of similarity with two N.intermedia plasmids, Fiji (13 ) and LaBelle (20 ), and might encode for a DNA polymerase. The plasmid mitochondrial location is confirmed.
The sequence reported in this paper has been deposited in the EMBL database (accession number Y12637).
Three isolates of C.parasitica harbouring plasmids, Cp5, Cp9 and Cp13, were isolated from evolutive cankers on chestnut trees (21 ), Cp5pl- is a derivative of strain Cp5, which lost its plasmid spontaneously during vegetative propagation. Strain Ep155 (ATCC 38755) is a laboratory standard plasmid-free strain provided by N.K.Van Alfen (Texas A&M University, Texas, USA). Ccp29 and Ccp28 were donated by G.J.Boland (University of Guelph, Ontario, Canada) and are dsRNA-free hypovirulent strains.
The plasmids pECUG1a and pECUG1b contained respectively the 2.4 and 1.7 kb KpnI fragments of the pUG1 plasmid inserted into the KpnI site of pUC19. pECUG1a was cleaved by EcoRI and subcloned in two plasmids, pECUG182 and pECUG322, used in sequencing.
The isolation of mtDNA and nuclear DNA (nDNA) by CsCl-bisbenzimide gradient, was performed according to Gobbi et al. (22 ). The DNAs were digested by KpnI for 2-8 h. MtDNA was treated with Ribonuclease A from bovine pancreas (Sigma, St Louis, MO, USA) at 37oC for 20 min. The digested nDNA and mtDNA were electophoresed in 0.7% agarose gel in TBE buffer with EtBr (0.5 [mu]g/ml) at 50 V/m for 16-18 h at 4oC.
For rapid extraction of total DNA, fungi were grown on PDAmb agar plates covered by a membrane (Bio-rad Laboratories #165-0963, Hercules, CA, USA) for 7 days at 25oC in the dark.
The DNA was extracted from the mycelia as described in Lecellier and Silar (23 ).
The procedures used in plasmid cloning, preparation and transformations were performed according to Ausubel et al. (24 ).
Plasmids pECUG1a and pECUG1b DNAs were labelled by random priming according to the manufacturer (Boehringer Mannheim Biochemicals, Mannheim, GE) with 7dUTP digoxigenin (BMB). The visualization of membranes was performed following the manufacturer's instructions (BMB). After hybridization, the membranes were stripped and rehybridized with different probes.
Fresh mycelium was obtained from 8 day-old biomass grown in 250 ml complete medium (25 ) inoculated with 5 * 105 conidia and incubated at 25oC with agitation (100 r.p.m.). The mycelium was ground, with glass beads, on ice in a mortar containing cold extraction buffer (10 mM TES-NaOH pH 7.5, 1 mM EDTA pH 8, 0.33 M saccharose) as described (26 ). The homogenate was centrifuged at 1100 g for 10 min to remove cell debris and the supernatant was centrifuged at 15 000 g for 20 min to give a pellet enriched in mitochondria. The pellet was suspended in extraction buffer and purified by a discontinuous sucrose gradient. Mitochondria were recovered from the 1.6-1.2 M interface and washed with extraction buffer prior to the DNAse digestion at 37oC for 1 h. Mitochondrial DNA was isolated by proteinase K treatment (100 [mu]g/ml) in the presence of sarkosyl (10%), followed by phenol extractions and ethanol precipitation.
Both strands were sequenced using the Thermosequenase flourescent labelled primer cycle sequencing kit (Amersham International plc, Buckinghamshire, UK) and the automated sequencing system A.L.F. DNA Sequencer (Pharmacia Biotech Europe, Brussels, Belgium). M13 universal and reverse primers were used to sequence inwards and the internal regions of clones pECUG1b and pECUG182 were sequenced by primer walking. Each region was sequenced at least twice. The analyses of the sequence data were performed using the sequence analysis software package from the Genetics Computer Group of the University of Winsconsin (27 ) and by the BLAST (28 ) programs for database searches.
Low molecular weight plasmid-like DNA bands were detected during a screening of mtDNA of C.parasitica strains isolated from nature. All these bands disappeared when treated with DNAse but were resistant to the RNAse treatment (data not shown). In the KpnI-digested mtDNA of strain Cp9, two bands, which fluoresced more intensely than the larger fragments derived from the mtDNA digestion (Fig. 1 A, lane 1), were supposed to be of plasmid origin and cloned. When used as probes in Southern analysis, the two resulting recombinant plasmids pECUG1a and pECUG1b can only detect the two corresponding bands of 2.4 and 1.7 kb in the KpnI-digested mtDNA (Fig. 1 B and C, lanes 1). In Figure 1 , the additional fainter hybridization signal, corresponding to a DNA fragment 4.1 kb in size and not belonging to the standard pattern of KpnI-digested mtDNA, was an incomplete cleavage product, as proved by longer digestion (data not shown).
The sequence homology of the plasmid with nDNA and mtDNA was evaluated by hybridization studies.
The plasmid showed no homology with the nDNA or with the mtDNA of plasmid-containing and plasmid-free strains (Figs 1 , 2 and 3 ), the probes pECUG1a and pECUG1b hybridized to their homologous targets, 2.4 and 1.7 kb respectively, when present but not to other genomic DNAs. Moreover, when the two strains Cp5 and its isogenic Cp5pl- were examined, signals appeared only in strain Cp5 (Fig. 4 , lanes 3 and 4) corresponding to the plasmid fragments, this result confirms the absence of homology between pUG1 and mtDNA. In addition, the mtDNA patterns of these two strains were indistinguishable, as shown in detail in the EcoRI and KpnI digestions, suggesting that no gross mutation has taken place as a consequence of the plasmid existence. Southern analysis (Fig. 3 ) showed a significantly high homology of pUG1 with the plasmids of similar sizes present in the strains Cp5 and Cp13; these plasmids also show an identical undigested and EcoRI- or KpnI-digested pattern, apparently very similar to that of pUG1.
The plasmid pUG1 has always been detected only in the mitochondrial fraction from CsCl-bisbenzimide gradient. The conclusive proof of the mitochondrial localization of pUG1 came from the isolation of the plasmid after the purified mitochondria of Cp9 were treated with DNAse. The digested DNA extracted from DNAse treated mitochondria contained the plasmid (Fig. 5 ).
The plasmid pUG1 is 4182 bp long and is a circular molecule as indicated by electrophoresis analysis. In Figure 1 , the undigested mtDNA sample contains many bands in addition to the high molecular weight mtDNA band. The fastest migrating band hybridizes with both pECUG1a and pECUG1b. The apparent size of this band, if considered linear, would be 2.4 kb (Fig. 3 B and C lanes 3), while, if compared to a supercoiled molecular weight marker, the resulting size of 4.0 kb (data not shown) would be consistent with the hybridization results, suggesting the supercoiled conformation of pUG1.
The plasmid sequence was determined from the two KpnI clones pECUG1a and pECUG1b. To confirm that pECUG1a and pECUG1b actually represent the entire pUG1, we isolated PCR fragments obtained by using primer pairs derived from the ends of both pECUG1a and pECUG1b and total DNA as template. The sequence data of these PCR fragments are consistent with the notion that the fragments are oriented as shown in the physical map in Figure 6 (data not shown).
The entire sequence of pUG1 was determined and the A+T content was estimated to be ~68.6%.
A plasmid-like DNA was reported in strains of C.parasitica in 1990 (34 ,35 ). It was described as being associated with a degenerative phenomenon called heteroauxesis. More recently Bell et al. (36 ), during the construction of the physical map of mtDNA of C.parasitica, found a plasmid of a similar size to pUG1.
The aims of this study were to characterize and localize the extrachromosomal plasmid, pUG1, and to assess the extent of sequence similarity with other DNAs to gain more insight into its biological role.
We report here the characterization and complete sequence of the plasmid pUG1 which exists as a mixture of multimeric molecules in mitochondria.
The mitochondrial plasmid pUG1 appears to constitute a distinctive genetic element: it shows virtually no homology with the standard mtDNA in DNA-DNA hybridizations, it achieves a high copy number without suppressive behaviour towards wild type mtDNA, and it also shows no homology with the nDNA. Since the detection of pUG1 from strain Cp9, all preparations of mtDNA of C.parasitica,from cultures isolated directly from natural populations, were assayed for the presence of plasmid DNA. The strains Cp5 and Cp13 harbour a plasmid very similar to pUG1 as confirmed by the hybridization experiments. Using pUG1 as a probe, plasmids were detected from 41% of the strains which belong to a population from a very restricted geographic area (E.G., unpublished), hence, confirming the notion that the plasmid-containing strains can be the rule rather than the exception in fungal populations (1 ).
The impact of pUG1 on the host phenotype is still unknown. It should not be associated with reduction or variation in virulence as all the plasmid-containing and virus-free strains, except one, were virulent on chestnut trees (E.G., unpublished). As far as the senescent phenotype is concerned, no definitive association has been determined between heteroauxesis and the presence of pUG1. There is no correlation between a specific mitochondrial RFLP complex and the existence of mitochondrial plasmids. No plasmid was detected in the Canadian strains which are hypovirulent with a defective mitochondrial function.
The possible existence of a genetic mobile element was postulated to explain the strikingly high level of RFLP shown by the mtDNA of the strains of C.parasitica (34 ,36 ). However there is no evidence yet to confirm such a role for pUG1 because of the absence of integration of the plasmid sequence into the mtDNA, as shown by hybridization. Therefore the plasmid biological implications have not yet been elucidated.
Based on sequence similarity between the mitochondrial plasmid ORF of C.parasitica and the two of Neurospora, it seems likely that the putative DNA polymerase coded by pUG1 represents a particular group B DNA polymerase. The characteristics of this group of enzymes are that they contain TTD instead of the highly conserved DTD in the motif C region and that they are closely related to the linear plasmid polymerases because of their protein primed initiation. pUG1 as a new member of this group enforces the suggestion that TTD is a possible alternative motif C as noted in Fiji and LaBelle polymerases (13 ).
The origin of pUG1 is unknown, neither nDNA nor mtDNA share any homology with the plasmid DNA which apparently is not derived from the nDNA because of the distinct codon usage of pUG1 from that of nDNA of C.parasitica (data not shown). Although the genetic code of the mtDNA of C.parasitica has not been formally demonstrated, two reported mitochondrial elements, pUG1 and a mt ds-RNA (30 ) of C.parasitica, which use the genetic code of mitochondrial DNA of molds (37 ), might provide some hints.
To our knowledge only one case of natural intergeneric distribution of plasmids in fungi has been reported, the kalilo plasmid of Neurospora and Gelasinospora spp. (38 ); recently one case of artificial transfer of a mitochondrial plasmid, pAI2 from A.immersus to P.anserina has been described (39 ). This accumulating evidence for horizontal gene transfer contributes to the hypothesis of a common origin of pUG1 and the Fiji and LaBelle plasmids, considering the strong similarities of the nucleotidic and amino acid sequences and the fact that Cryphonectria and Neurospora are possibly taxonomically close genera.
We thank Dr Lei Zhang (Texas A&M University, Texas, USA) for comments on the manuscript and Prof. Neal Van Alfen (Texas A&M University, Texas, USA) for encouragement. We also thank Igor Tomada for the excellent technical assistance and Jaqueline Rogers for revising the English text. This work was supported by a grant from the National Research Council (C.N.R.) of Italy.
REFERENCES