Ó Springer-Verlag 2000
Curr Genet (2000) 37: 242±256
ORIGINAL PAPER
Claudia B. Monteiro-Vitorello á Dipnath Baidyaroy Julia A. Bell á Georg Hausner á Dennis W. Fulbright Helmut Bertrand
A circular mitochondrial plasmid incites hypovirulence in some strains of Cryphonectria parasitica Received: 12 August / 9 December 1999
Abstract In the chestnut-blight fungus Cryphonectria parasitica, a plasmid, pCRY1, occurs in the mitochondria of several strains isolated at various locations in the northeastern United States and Canada. The monomer of this plasmid is a 4.2-kb circular double-stranded DNA that has no detectable sequence homology with the 160±kb mitochondrial DNA of Ep155, a standard virulent laboratory strain of C. parasitica. The circular nature and oligomeric characteristics of the plasmid were deduced from the heterogeneous size of plasmid DNA molecules as detected by one- and two-dimensional gel-electrophoresis, the nature and alignment of restriction fragments, and the lack of detectable termini in the nucleotide sequence. The cytoplasmic location of the plasmid was deduced from its co-puri®cation with mitochondria, uniparental (maternal) transmission in sexual crosses, dissociation from the nuclei of the donor strain during its horizontal transfer between vegetatively compatible strains through hyphal anastomoses, and mitochondrial codon usage (UGA Try). The pCRY1 plasmid contains a long open reading frame that is transcribed and potentially encodes a unique 1214 amino-acid, B-family DNA polymerase similar to those
Communicated by B.G. Turgeon C.B. Monteiro-Vitorello1 á G. Hausner2 á H. Bertrand (&) Department of Microbiology, Michigan State University, East Lansing, MI 48824-1101, USA e-mail:
[email protected] Fax +1 517 353 8957 D. Baidyaroy á J.A. Bell á D.W. Fulbright Department of Botany and Plant Pathology, Michigan State University, East Lansing, MI 48824-1312, USA Present addresses: Department of Genetics, Escola Superior de Agricultura Luiz de Queiroz, University of SaÄo Paulo, 13418 Piracicaba, SaÄo Paulo, Brazil 2 Department of Biological Sciences, University of Calgary, Calgary, Alberta, T2 N 1N4, Canada 1
encoded by the LaBelle and Fiji circular mitochondrial plasmids of Neurospora. In this subgroup of proteins, the DTD motif characteristic of B-family DNA polymerases is replaced by TTD. Amino-acid motifs related to those that are characteristic of the 3¢®5¢ exonuclease domains of B-family DNA polymerases have been located in the amino-terminal portion of the proteins. A comparison of isogenic plasmid-free and plasmid-containing cultures indicates that pCRY1 is an infectious agent that eects a reduction in the pathogenicity of some, but not all, strains of C. parasitica. Key words Chestnut blight á Endothia á Hypovirulence á Hypovirus á Mitochondrial plasmid
Introduction Mitochondrial plasmids are relatively frequent in the ®lamentous fungi, appearing in at least 16 genera, and by far the majority are linear genetic elements (Griths 1995). The vast majority of the linear plasmids do not seem to produce obvious physiological or phenotypic eects in their hosts. However, a few, most notably the kalilo and maranhar plasmids of the fungus Neurospora (Bertrand et al. 1986; Chan et al. 1991; Court et al. 1991) and the mF plasmid of the slime mold Physarum polycepahalum (Nakagawa et al. 1998), induce senescence by integrating into the mitochondrial chromosome. The mF plasmid also appears to promote mitochondrial fusions and mtDNA recombination (Kawano et al. 1993). In contrast, the life-span of the fungus Podospora anserina is prolonged substantially by integration of the pAL2±1 linear plasmid into the mitochondrial chromosome (Hermanns et al. 1994). Relatively few circular mitochondrial plasmids have been detected so far; not necessarily because they are scarce, but most likely because they are less-easily detected than linear plasmids (Griths 1995). For Neurospora crassa alone, circular plasmids have been divided into at least seven homology groups, as deter-
243
mined by DNA hybridization (Arganoza et al. 1994). However, the nucleotide sequences of four circular plasmids, named Mauriceville, Varkud, Fiji and LaBelle, reveal that the number of dierent types that actually exist is much smaller than indicated by the homology groups based on DNA-hybridization data. While these four plasmids have been allocated to three dierent homology groups, they represent only two dierent types: those encoding a reverse transcriptase (Mauriceville and Varkud) and those encoding a unique DNA polymerase (Fiji and LaBelle). In addition to the numerous sightings in Neurospora, a circular plasmid that may control mating type has been found in Absidia glauca (Haen¯er et al. 1992), and a circular plasmid that may be related to the Mauriceville and Varkud plasmids of Neurospora may control AK-toxin production in Alternaria alternata (Katsuya et al. 1997). Three circular plasmids have also been observed in Pythium species (Martin 1991), a genus of ®lamentous water molds that is more closely related to heterokont algae than to fungi. Four circular plasmids of Neurospora have been studied at the molecular level. The Varkud and Mauriceville circular plasmids comprise one homology group, as determined by DNA hybridization; they encode reverse transcriptases and are thought to be related to mobile elements, particularly introns (Collins et al. 1981; Nargang et al. 1984; Akins et al. 1988). To facilitate the discussion of optional genetic elements that encode functionally related proteins, the Mauriceville and Varkud retroplasmids are identi®ed in this paper as the prototypes of type-I mitochondrial plasmids. The Mauriceville plasmid can integrate into the mitochondrial DNA (Akins et al. 1986), but there is no indication of integrated copies behaving as introns. Instead, integration of these plasmids elicits fungal senescence. In contrast, the Fiji and LaBelle plasmids of Neurospora intermedia both encode B-family DNA polymerases (Stohl et al. 1982; Li and Nargang 1993), even though they belong to dierent DNA homology groups. Hence, in this paper, these two autonomously replicating elements are considered to be the prototypes of type-II mitochondrial plasmids. While some regions of the DNA polymerase encoded by the LaBelle plasmid are similar to some of the domains characteristic of reverse transcriptases (Pande et al. 1989; Schulte and Lambowitz 1991), these features are not conserved in the polymerase encoded by the Fiji plasmid (Li and Nargang 1993). Furthermore, it has been demonstrated that the proteins encoded by the LaBelle and Fiji plasmids are DNA polymerases and lack reverse-transcriptase activity (Schulte and Lambowitz 1991; Li and Nargang 1993). Although a remnant of a 1.6-kb portion of a plasmid related to LaBelle is present in the mitochondrial chromosome of many Neurospora strains (Nargang et al. 1992), neither this plasmid nor the related Fiji element have been observed to integrate into mtDNA or induce senescence in Neurospora. The Harbin-1 plasmid of Neurospora has DNA homology, as detected by hy-
bridization, to the LaBelle plasmid (Yang and Griths 1993a), but has not been characterized suciently to assert whether or not it encodes a DNA polymerase. In the context of the above classi®cation, type-III mitochondrial plasmids are linear invertrons, as de®ned by Sakaguchi (1990), and encode a DNA polymerase as well as an RNA polymerase (Griths 1995). A comparison of the restriction endonuclease cleavage patterns of the mitochondrial DNAs from isolates of the chestnut-blight fungus, Cryphonectria parasitica, revealed the presence of relatively large amounts of a unique, 4.2-kb restriction fragment in 4 of 11 isolates originating from Michigan and Connecticut (Bell et al. 1996). The DNA that gave rise to the unique, 4.2-kb restriction fragment could be separated from the bulk of the mtDNA when the uncleaved nucleic acids were subjected to agarose-gel electrophoresis, thus revealing that it originated from a mitochondrial plasmid, which has been named pCRY1 (Bell et al. 1996). Since this plasmid appeared to be associated preferentially with isolates of C. parasitica that either had unusual growth patterns or gave rise to hypovirulent sectors, it has been sequenced and characterized genetically to determine if it is related to any of the mitochondrial elements known to induce senescence in other fungi (reviewed by Grif®ths 1992, 1995). The results show that pCRY1 is a typeII circular plasmid and indicate that it can elicit hypovirulence in at least some strains of C. parasitica.
Materials and methods Strains, media, sexual crosses, tests for cytoplasmic transmission The following strains of C. parasitica were used in this study: Ep155 MAT1-1, a standard virulent laboratory strain (ATCC 38755); Ep339 MAT1-2, a virulent strain collected by S. Anagnostakis (Connecticut Agricultural Research Station); and CL25.9, a hypovirulent single-conidial isolate derived from strain CL25 (Mahanti et al. 1993). Strains CL25.9 and Ep339 contain the pCRY1 plasmid. Two nuclear respiratory mutants of Ep155, cyt1 (180.17) and cyt2 (181.3) (Monteiro-Vitorello et al. 1995), were used in crosses made to determine the pattern of sexual transmission of pCRY1. The mitochondria of both mutants are de®cient in cytochrome a and b, and the cells manifest elevated levels of cyanide-insensitive respiration (alternative oxidase activity). A plasmid-free, genetically marked (brown pigmentation) strain, F2.36 br (Huber 1996), was used in studies on the transmission of plasmids by hyphal contact. All cultures of the fungus were maintained on Endothia complete medium (Puhalla and Anagnostakis 1971), and sexual crosses were implemented on chestnut wood (Anagnostakis 1979). Tests for the asexual transmission of genetic elements by hyphal contact between vegetatively compatible strains of C. parasitica were performed as previously described (Monteiro-Vitorello et al. 1995). For the preparation of genomic DNA, standing cultures were grown for 10 days from ®ve to ten mycelial plugs in 250-ml Erlenmeyer ¯asks containing 50 ml of liquid medium. Southern blots of the electrophoretically separated high- and low-molecularweight components of uncut genomic DNAs were used in assays for the sexual and asexual transmission of the pCRY1 plasmid. Genetic nomenclature Whenever possible, the recommendations on genetic nomenclature and symbolism assembled for Neurospora by Perkins (1999) are
244 adopted in this paper. Of particular relevance is the enclosure of mitochondrial genes in brackets. In this paper, this rule has been extended to designate strains that have or lack a mitochondrial plasmid. Respiration, growth-rate and virulence assays Cyanide-sensitive respiration and alternative oxidase activity were determined as described by Monteiro-Vitorello et al. (1995). Growth rate was measured over a period of at least 40 days in race tubes on Endothia complete medium as described for Neurospora by Ryan et al. (1943). The apple-test described by Fulbright (1984) was used to determine the relative virulence of dierent C. parasitica strains. Virulence is expressed as an index de®ned as the area of the lesion in cm2 produced by the fungus on a Golden Delicious apple 15 days after the date of inoculation. Isolation of mitochondrial, genomic and plasmid DNAs, cloning, and hybridization The procedures described by Bell et al. (1996) were used for the preparation of mitochondria and the extraction of DNA from the puri®ed organelles. In experiments requiring the screening of many cultures for the presence and absence of the pCRY1 plasmid by Southern blot hybridization, total (genomic) DNA was prepared from the combined mycelia of 2±4 standing cultures by the method of Yelton et al. (1984), except that diethyloxidoformate was omitted. Genomic DNAs used as templates in polymerase chain reactions (PCRs) were generated from the mycelium in a single 50-ml standing culture by the method described by Kim et al. (1990). Mitochondrial RNA was extracted from puri®ed mitochondria by the procedure described by Solymosy et al. (1968). Standard techniques were used for the routine manipulations of DNA and RNA (Ausubel et al. 1987; Sambrook et al. 1989). For the cloning of the pCRY1 DNA, mtDNA from strain CL25.9 was digested either with EcoRI, which cleaves the plasmid at a single site, or BglII, which cleaves the plasmid at three dierent sites (Bell et al. 1996). The resulting mtDNA and plasmid DNA fragments were separated by electrophoresis through a 0.6% agarose gel and visualized by exposure of the ethidium bromide (EtBr)-stained gels to long-wave UV light. The linearized plasmid DNA was excised from the gel, electroeluted into dialysis tubing, and concentrated by ethanolprecipitation. Competent cells of Escherichia coli were prepared and transformed as recommended by Nishimura et al. (1990). The entire 4.2-kb EcoRI-cleaved pCRY1 DNA was cloned in E. coli DH5a cells as an insert in the EcoRI site of the BluescriptKS+ vector (Stratagene), whereas the BglII fragments were cloned as inserts ligated into BamH1-cleaved pUC19 vector DNA. The clone which contains the entire pCRY1 plasmid as an insert in the EcoRI site of the BluescriptKS+ cloning vector was named pE1. The names of E. coli plasmids containing restriction fragments covering the sub-genomic segments of pCRY1 that were used as probes are described in the Results section of this paper. The protocols provided by Boehringer Mannheim with the digoxigenin-dUTP Genius kit were used for the labeling and detection of DNA probes during the identi®cation of complementary nucleotide sequences on Southern and Northern blots. DNA sequencing and analysis Automated ¯uorescent Taq-cycle sequencing of DNA was performed by the Department of Energy, Plant Biochemistry Facility, at Michigan State University using an ABI Catalyst 800 system with an ABI 373 Sequencer. Manual sequencing was performed by the dideoxy method of Sanger et al. (1977) using 33P-labeled dATP or dCTP for the identi®cation of the reaction products by autoradiography. Initial sequences of inserts in the pUC and BluescriptKS+ were obtained through the extension of vector-speci®c forward and reverse primers. These initial sequences were extended progressively with sequence-speci®c primers synthesized by the
Michigan State University Macromolecular Synthesis Facility. Both strands of the pCRY1 plasmid were completely sequenced. The nucleotide sequences across restriction sites that were used for cloning fragments of the plasmid were established by sequencing PCR products generated by using strain CL25.9 mtDNA as a template for pairs of primers ¯anking the respective cleavage sites. Sequences were aligned and merged initially using the MicroGenie MG-IM-5.0 software (Queen and Korn 1984), and databases were searched using BLAST (Altschul et al. 1990, 1997) computer software provided by the National Center for Biotechnology Information at the National Institutes of Health, Bethesda, Md. The sequence of the pCRY1 plasmid can be found under accession number AF031368 in GenBank and related databases. Two-dimensional gel electrophoresis The mitochondrial DNA used for the analysis of the structure of pCRY1 by two-dimensional (2D) gel electrophoresis was prepared from strain Ep339 as described by Bell et al. (1996). Neutral/neutral 2D gel electrophoreses were performed and interpreted as described by Brewer and Fangman (1987) and Backert et al. (1997), except that the ®rst dimension was developed on a 0.5% agarose gel at 1.0 V/cm for 24 h at room temperature. The second dimension was developed on a 1.1% agarose-gel slab at 5.0 V/cm and 4 °C for 60 h, and in the presence of EtBr (0.3 lg/ml). The dierent molecular forms of pCRY1 were detected by hybridization of a Southern blot of the second-dimension gel with a probe prepared from the cloned, full-length plasmid DNA.
Results Molecular variants and unit size of pCRY1 As reported previously (Bell et al. 1996), and shown in Fig. 1A, a 4.7-kb DNA species was detected in some strains of C. parasitica when uncleaved mtDNA prepared from ¯otation-gradient-puri®ed mitochondria was subjected to agarose-gel electrophoresis. On the basis of its apparent autonomy from the bulk of the mtDNA, it was surmised that the novel DNA might be a mitochondrial plasmid, which was tentatively named pCRY1 (Bell et al. 1996). When the novel 4.7-kb DNA was extracted from a gel, labeled, and used as a probe on Southern blots of equivalent gels, it annealed not only to molecules of the same size, but also to a series of fastermigrating units as well as many larger molecules that appeared to be multiples of the 4.7-kb species (Fig. 1B). When the mitochondrial DNA from strains that had this novel element were cleaved with EcoRI or KpnI, the 4.7kb species and all the larger variants were converted into a single, quite abundant, 4.2-kb fragment, as shown for strain CL25.9 in Fig. 1C and D. Cleavage of the same DNA with BglII reduced the 4.7-kb element and all its variants to three fragments of approximately 1.6, 1.5 and 1.1 kb, respectively, representing a total length of 4.2 kb (Fig. 1C). These results indicated that: (1) the 4.7kb element that appeared upon electrophoresis of the uncut mtDNA is the relaxed circular form of the 4.2-kb pCRY1 plasmid detected previously in EtBr-stained gels (Bell et al. 1996), (2) the smaller forms that appear on agarose gels as a series beneath the 4.7-kb element are supercoiled forms of the same circular DNA, and (3) the larger units that hybridized with the probe are probably
245
Fig. 1 A ethidium bromide-stained agarose gel showing presence of a circular plasmid DNA (cP) in uncut mtDNA preparations from strains Ep339 and CL25.9 and its absence in strain Ep155. The positions of molecular-size standards of linear DNA are indicated at the left side of the panel. B Southern blot of the gel shown in panel A hybridized with a digoxigenin-labeled probe prepared from DNA obtained by excision of the band marked cP from an agarose gel. The position of circular DNA representing the putative dimer of the plasmid is indicated as 2cP. The slight hybridization of the mtDNAs of strain Ep155 was due to the contamination of the probe with small amounts of mtDNA. No hybridization was observed when Southern blots of mtDNA from the plasmid-free Ep155 strain were hybridized with probes prepared from cloned pCRY1 DNA (data not shown). C agarose gel of electrophoretically separated fragments obtained by cutting the mtDNA of strain CL25.9 with EcoRI. The lane labeled EtBr shows a gel stained with ethidium bromide, the lane labeled Hyb shows a Southern blot of the same gel that was hybridized to the same probe as used in panel B. The position of the linear plasmid DNA is indicated by P. The lane marked M shows linear DNA standards. D Southern blot of uncut, BglII- and KpnI-cut mtDNA from strain Ep339 hybridized to a probe prepared from the cloned EcoRI fragment of pCRY1 shown in panel C. The circular and linear forms of the complete pCRY1 genome are marked as cP and P, respectively, on the right side of the ®gure. Also shown on the right are the sizes in kb of the three BglII fragments of pCRY1
a mixture of linear, relaxed circular and supercoiled circular forms consisting of tandem repeats of the 4.2-kb linear pCRY1 DNA unit. Since uni-dimensional gel electrophoresis of uncut DNA does not distinguish unambiguously between linear and circular DNA molecules of varying sizes and shapes, the mtDNA of the pCRY1-containing Ep339 strain was subjected to two-dimensional gel electrophoresis. The DNA-hybridization pattern of Southern blots of 2D gels are presented in Fig. 2. When the mtDNA was not cleaved, the pCRY1 DNA appeared predominantly as a series of discretely sized units that migrated more slowly than linear DNA and correspond in size to circular monomers, dimers and trimers of the
4.2-kb unit. Circular molecules larger than 13-kb did not appear because they were removed during the processing of the gels. Some of the plasmid DNA appears in a continuous arc of linear molecules, which includes areas of higher concentrations corresponding not only to linear monomers, dimers and trimers, but also of longer chains of the pCRY1 unit. All the linear and circular forms of pCRY1 are reduced to 4.2-kb linear units when cut with EcoRI (Fig. 2B). Exactly the same general patterns of electrophoretic distribution were observed when the Mauriceville, Fiji and LaBelle circular plasmids of Neurospora were subjected to 2D gel electrophoresis (data not shown). The proportion of circular to linear molecules varied in dierent preparations of DNA, circular molecules being predominant when the amount of linear molecules was low and vice versa. Thus, at least some of the linear molecules resulted from random breakage of circular molecules during the preparation of mitochondria and mtDNA, as indicated by the tail of smaller linear pieces of pCRY1 DNA that emanates from the spots of the 4.2-kb linear monomer that were produced by cleavage of the larger units with EcoRI (Fig. 2B). Collectively, the results establish that pCRY1 exists in vivo predominantly as a series of discretely sized circular molecules that are monomers and multiples of the 4.2-kb unit of plasmid DNA. The possibility that an undetermined, but relatively low fraction of the plasmid DNA occurs naturally in a linear form, possibly as products of rolling-circle-replication activity, cannot be ruled out at this time. Inheritance of pCRY1 To determine the mode of inheritance of pCRY1, strain Ep339 MAT1-2 [pCRY1] was crossed reciprocally with
246
Fig. 2A, B Two-dimensional agarose-gel electrophoretic migration patterns of pCRY1. Southern blots of gels containing the 2D electrophoretic separation patterns of the mtDNA of C. parasitica strain Ep339 were hybridized with a probe made from a cloned, complete pCRY1 DNA (EcoRI fragment). A the 2D pattern of the uncut mtDNA. The monomeric, dimeric and trimeric circular forms of pCRY1 are indicated by arrows labeled cP, 2cP and 3cP, respectively, whereas the equivalent linear molecules are marked P, 2P and 3P, respectively, and are positioned on the arc of linear pieces produced by random breakage of large molecules. B the 2D pattern obtained from mtDNA that was cut with EcoRI before electrophoresis. No circular molecules and only linear molecules of unit size, marked P, appear at the high-molecular-weight end of a short arc of randomly sized linear pieces of pCRY1 DNA
two plasmid-free, nuclear respiratory mutants of Ep155 MAT1-1, 180.17 cyt1 and 181.2 cyt2. At least ten random ascospores were collected from several perithecia produced by each cross, and a sample of uncut genomic DNA from each progeny was subjected to electrophoresis on agarose gels to separate the plasmid from the bulk of the DNA. Southern blots of the gels were hybridized with a labeled pCRY1-speci®c probe. As shown in Fig. 3 for a subset of six spores from each unit of a reciprocal cross, the plasmid was inherited by all the
Fig. 3 Maternal inheritance of pCRY1. Undigested genomic DNAs from strains 180.17, Ep339, and six progeny from each of the reciprocal crosses of these two strains (indicated at the top of the gel) were subjected to agarose-gel electrophoresis and Southern blotting. The blot was hybridized with a digoxigenin-labeled probe prepared from a BglII-EcoRI fragment of pCRY1 DNA cloned in E. coli as plasmid pBE3 (see Fig. 7). Circular monomers and dimers of the pCRY1 plasmid are labeled cP and 2cP, respectively
progeny when Ep339 [pCRY1] was the female parent (four crosses comprising a total sample of 81 spores), but appeared in none when the plasmid-bearing strain was the male parent (three crosses comprising a total sample of 87 spores). Similarly, when Ep339 [pCRY1] was crossed as a female with 181.2 cyt2, the plasmid appeared in all of 12 random spores that were sampled, and in none of 12 spores derived from the cross where 181.2 cyt2 was the female, and Ep339 [pCRY1] the male, parent. As expected (Monteiro-Vitorello et al. 1995), the nuclear gene-encoded elevated alternative oxidase phenotype of the cyt-1 and cyt-2 mutants segregated 1:1 from the wild-type respiratory phenotype in all crosses. An apparent exception to the maternal inheritance of pCRY1 was observed in one cross where Ep339 [pCRY1] was the female and 180.17 cyt1 the male parent. In this case, the plasmid appeared in only 11 of 13 ascospores that were collected from four perithecia. This situation could mean that two ascospores that lacked pCRY1 either originated from a perithecium that was formed by the putative male parent, or that the transmission of the plasmid is occasionally suppressed by nuclear genes, as has been documented for some of the mitochondrial plasmids of Neurospora (Yang and Grif®ths 1993b). Because of the uncertainty about the source of the two plasmid-free ascospore progeny, this cross was not included in the calculation of the sample sizes of the progeny given in the previous paragraph. By contrast, transmission of pCRY1 from the paternal parent has not been observed. Since paternal transmission of mitochondrial plasmids happens rarely in Neurospora (Yang and Griths 1993c), it is possible that the number of ascospores sampled to-date is too small to discern whether or not such events also occur occasionally in C. parasitica. Transmission of pCRY1 by hyphal contact At least in Neurospora, mitochondrial plasmids are readily transmitted asexually by hyphal contact between heterokaryon-compatible strains and even between incompatible strains (Collins and Saville 1990; Griths et al. 1990; Griths 1995). To determine if pCRY1 could invade plasmid-free strains, small plugs of mycelium from the Ep339 [pCRY1] strain were inoculated side by side with plugs of mycelium from a brown, heterokaryon-compatible strain, F2.36 br. Conidia were removed from the contact zones between the mycelia and spread on Endothia agar medium for the isolation of orange (donor) and brown (recipient) single-conidial isolates. Genomic DNA was prepared from each singleconidial isolate, cut with EcoRI, and examined for the presence of the plasmid by hybridization of a Southern blot of electrophoretically separated fragments with a labeled probe of pCRY1 DNA (data not shown). In one typical experiment, of ten randomly selected, brown single-conidial isolates, four had acquired the pCRY1 plasmid from the orange-colored donor. Of the ten
247
orange single-conidial isolates that were selected randomly from the same pairing of cultures as a control, all retained the plasmid. Thus, pCRY1 is transmitted asexually by hyphal contact. The observation that the pCRY1 DNA dissociated from the nuclei of the donor strain during the vegetative-transmission process con®rms the extranuclear location of this plasmid. A detailed analysis of the transmission of pCRY1 by hyphal contacts between vegetatively compatible and incompatible strains of C. parasitica and by protoplast fusion will be presented in a separate publication (Baidyaroy et al. 2000). Eect of pCRY1 on the growth and virulence of C. parasitica A preliminary survey indicated that the pCRY1 plasmid is associated frequently, but not exclusively, with virusfree isolates of C. parasitica that either have the cytoplasmically transmissible mitochondrial hypovirulence phenotype or spontaneously develop colony sectors that have this trait (Mahanti et al. 1993; Bell et al. 1996). The tests for the asexual transmission of pCRY1 generated a plasmid-containing line of C. parasitica that is isogenic with F2.36 br. Such pairs of isogenic strains are ideal for testing the eect of plasmids on the growth and virulence of the fungus. In growth tubes and on Endothia solidagar medium, plasmid-free single conidial isolates of F2.36 had a growth rate of 5.84 0.17 mm/day (average standard deviation), whereas the growth rate of the F2.36 br [pCRY1] single-conidial isolates was 5.88 0.02 mm/day. Clearly, the presence of the plasmid per se had no eect on the growth on the F2.36 strain of C. parasitica. Under the same conditions, the plasmid-free Ep155 laboratory strain had a growth rate of 5.61 0.4 mm/day; while, under similar conditions, the growth rates of strains that are hypovirulent because they are infected by hypoviruses (Nuss 1992) or have mtDNA mutations (Monteiro-Vitorello et al. 1995) range from normal to less than 2.0 mm/day. On the basis of apple tests, the virulence index of the plasmid-free single-conidial isolates of the F2.36 br strain of C. parasitica was 19.9 6.4 cm2, whereas the F2.36 br [pCRY1] isolates had an index of only 8.5 1.3 cm2. Under the same conditions, the plasmid-free Ep155 standard-laboratory strain had a virulence index of 28.8 1.1 cm2, whereas the virulence indices of hypovirus-infected and mitochondrially hypovirulent strains are consistently less than 10 cm2 (MonteiroVitorello et al. 1995). Collectively, the results indicate that the cytoplasmically-transmissible pCRY1 mitochondrial plasmid directly depresses the pathogenic potency of strain F2.36 br of C. parasitica. Nucleotide sequence of pCRY1 The complete nucleotide sequence of both strands of pCRY1 was determined and a merging of all the
sequences con®rmed that the plasmid is a 4234-bp circular DNA. Fig. 4 presents the sequence of the coding strand of the plasmid from strain CL25.9 and, together with Fig. 5 provides molecular details of the features that are presented as a map in Fig. 7. The plasmid resembles mtDNA in that it is very AT-rich (A + T 68.6%). A database search revealed that, at the nucleotide sequence level, pCRY1 is very closely related to the 4182-bp pUG1 plasmid found in some European strains of C. parasitica (Gobbi et al. 1997; GenBank accession number Y12637). Very little is known about pUG1 beyond its nucleotide sequence. The plasmids dier from each other in that pUG1 lacks one copy of a 60-bp nucleotide sequence that occurs twice in a tandem repeat in pCRY1, once from positions 4107 to 4166, inclusive, and again from positions 4167 to 4226 (see Figs. 4 and 6). In pCRY1, these two repeats dier from each other by a single base pair, and it is the second copy that is present at the beginning of the nucleotide sequence of pUG1 as published by Gobbi et al. (1997). In addition to this dierence, pCRY1 has only one KpnI site, whereas pUG1 has two, and their nucleotide sequences vary by insertions or deletions of several base pairs in the non-coding segment of their DNAs, as well as by nine base-pair substitutions, involving seven transitions and two transversions (data not shown). Some of these base substitutions are manifested as seven more-or-less conservative dierences in the amino-acid sequences in the homologous proteins encoded by the two plasmids (Fig. 5). Translation of the nucleotide sequences of both strands of the pCRY1 plasmid yielded no open reading frames (ORFs) of signi®cant length when the standard genetic code was applied. However, the application of the mitochondrial genetic code for the ascomycetous ®lamentous fungi (TGA codes for tryptophan rather than being a terminator codon) revealed a very long ORF. If it is assumed that the ®rst in-frame ATG represents an initiator codon, then the predicted protein is 1214 amino-acids long (Figs. 4 and 5). The putative protein encoded by pCRY1 is homologous to the DNA polymerases encoded by the 5268-bp Fiji and 4070-bp LaBelle circular mitochondrial plasmids of Neurospora (Fig. 5; Pande et al. 1989; Li and Nargang 1993). Like the proteins of the Fiji and LaBelle plasmids, the pCRY1 polypeptide has the three motifs (POL I, II and III) that are characteristic of family B DNA polymerases (Ito and Braithwaite 1991). In the POL-I motif, 13 of the 24 amino acids in the pCRY1 protein were identical to those encoded by both LaBelle and Fiji, and amino acids belonging to the same family appeared at an additional eight positions in all three plasmids. The aspartic-acid (D) and tyrosine (Y) residues that are very highly conserved in the POL-I motifs of the B-family polymerases are also present in the pCRY1 protein. In the POL-II motif, 10 of 25 amino acids were identical in the polymerases of pCRY1, Fiji and LaBelle, and nine more were from the same amino-acid family. The most highly conserved amino-acid sequence of the POL-II motif in
248 Fig. 4 Features of the nucleotide-sequence open reading frame of PCRY1. A sequence which resembles the Neurospora mitochondrial promoter consensus sequence (Kubelik et al. 1990), and includes the sequence TTATAAT, is indicated by a box at nucleotides 382±392. The ATG start and TAG stop codons of the open reading frame are given in bold type in the nucleotide sequence. Pertinent restriction sites are presented by white lettering in black ®elds: EcoRI GAATTC, BglII AGATCT, and KpnI GGTACC. The locus corresponding to a KpnI site in the closely related pUG1 plasmid is represented by a shaded stippled box on the nucleotide sequence at position 2359. The 60-bp repeats, one of which is missing in pUG1, are indicated by bracketed bars labeled rep1 and rep2. Amino-acid sequences that contain motifs related to those in the 3¢®5¢ exonuclease domains of type-B DNA polymerases are highlighted in grey, whereas polymerase motifs have thick underlining and are labeled POL I, POL-II, and POL-III, respectively
the B-type polymerases, KÐNS-YG, is also present in the pCRY1 polymerase. In addition, immediately adjacent to the POL-II motif, there appears a block of ®ve amino acids, FDIKT, which is highly conserved in the
proteins of the Fiji, LaBelle, pCRY1 and pUG1mitochondrial plasmids, but does not appear in any of the other members of the B-type polymerases. In the POLIII motif, 8 of 15 amino acids were identical in the
249 Fig. 5 Alignment of the DNA polymerases of the type-II mitochondrial plasmids of C. parasitica and Neurospora. Abbreviations are: Fi for Fiji (accession L08781), pC for pCRY1, pU for pUG1 (accession Y12637), and La for LaBelle (accession X13912). Potential 3¢®5¢ exonuclease motifs are indicated by thin underlining and labeled EXO I, II and III. The polymerase motifs are marked by thick lines labeled POL I, II and III. Amino acids that are conserved in the equivalent motifs of most B-type polymerases, particularly those of linear mitochondrial plasmids, are shown in boxes, whereas amino acids that tentatively are identi®ed as being part of a motif are highlighted by a grey background. Aminoacid substitutions that distinguish pCRY1 from pUG1 are shown in white letters on a black background
250
Fig. 6A, B Identi®cation of transcripts of the pCRY1 plasmid. A ethidium bromide-stained gel showing electrophoretically separated RNAs extracted from the mitochondria of the plasmid-free F2.36 strain (-) and a plasmid-containing derivative of the same strain, F2.36 (pCRY1) (+). B Northern blot of the gel shown in A hybridized with a probe prepared from a cloned DNA covering one entire unit of the pCRY1 genome (pE1). The sizes of the ten distinct species of RNA that could be identi®ed are indicated on the right side of the panel
polymerases of pCRY1, Fiji and LaBelle, and ®ve more were from the same amino-acid family. However, instead of the highly conserved Tyr±AspThrAsp (YBDTD) sequence that is characteristic of the POL-III motif of most of the B-family polymerases, the proteins of the Cryphonectria and Neurospora plasmids all have a unique Ser± Thr-Thr-Asp (SBTTD) motif. While many plasmid-encoded, B-family DNA polymerases have three relatively well-conserved aminoacid motifs associated with the 3¢®5¢ exonuclease activity of the N-terminal ``editing domain'' (Court and Bertrand 1992), these sequences do not appear to be well-conserved in the Fiji and LaBelle plasmids (Li and Nargang 1993) or in pCRY1. Nonetheless, an alignment of the proteins does reveal that the pCRY1, Fiji and LaBelle polymerases share three regions of relatively high sequence similarity (marked EXO I, II and III in Fig. 5) that contain amino-acid sequences related to typical exonuclease motifs. The ®rst of the three contains the sequence DTE (DAE in Fiji), which could be a variant of the consensus DIE motif of the proteinprimed, B-family polymerases (Wang et al. 1989; Ito and Braithwaite1991). The second motif, which has the
Fig. 7 Physical map of a circular monomer of the pCRY1 plasmid. The beginning and end of the nucleotide sequence shown in Fig. 4 are indicated by a vertical arrow. In addition, only the most salient features that are characteristic of this plasmid are shown, including the location of restriction sites, the open reading frame (thick semi-circular arrow), the location of the putative EXO and POL motifs (grey diamonds and rectangles, respectively) that are characteristic of B-type DNA polymerases, and the location of the 60-bp repeats (small arrows) that distinguish pCRY1 from pUG1. The segments of pCRY1 DNA that were cloned in E. coli are shown as semicircles on the periphery of the map and are identi®ed by the names of the corresponding recombinant bacterial plasmids
consensus I/Y±HNÐFD, is not obvious in the proteins encoded by pCRY1, Fiji and LaBelle, but variants of this arrangement might exist within the EXO-II region as shown in Fig. 5. The most-conserved feature of the third exonuclease motif, the sequence YÐD, clearly appears in the third conserved region (EXO III in Fig. 5) of all four mitochondrial plasmids. While the similarity between the polymerases of the pCRY1, LaBelle and Fiji plasmids is fairly obvious, the amino-acid sequences of the two Neurospora plasmid proteins are more closely related to each other than either is to proteins encoded by the Cryphonectria plasmids. Nonetheless, the distance between the POL-1 and 2 motifs is much shorter in the polymerases of pCRY1 (132 codons) and LaBelle (139 codons) than in the polymerase of the Fiji plasmid (251 codons). Thus, either an insertion of approximately 400 bp has occurred in an ancestor in the lineage that leads to the Fiji plasmid, or a deletion has taken place in a plasmid which is ancestral to pCRY1 as well as LaBelle. In any case, it appears that the pCRY1, LaBelle and Fiji plasmids have all evolved independently from a common ancestral genetic element. Aside from the uniqueness of the amino-acid sequences encoded by the three plasmids, the features that distinctly separate pCRY1 and pUG1 DNA polymerase from those of the Fiji and LaBelle plasmids are: (1) the length of the relatively unconserved amino-terminal segment that precedes the ®rst motif of
251
the putative editing domain, and (2) a deletion of four or ®ve amino acids between two highly conserved blocks of amino acids in the POL-1 motif of the polymerase domain. In the pCRY1/pUG1 proteins, the amino-terminal segment has insertions of several blocks of amino acids relative to equivalent regions of the Fiji and LaBelle polymerases. Much of the dierence appears to be associated with blocks of perfect and imperfect repeats of three oligonucleotides, TGAAAA, GAA and GTT, at the beginning of the open reading frame in the pCRY1 DNA. Transcription of pCRY1 Potentially, plasmids such as pCRY1 can be maintained by the mitochondrial replication system without expression of their genes. To establish a basis for the assumption that the expression of genetic information encoded by the pCRY1 plasmid might be a requisite for its maintenance or eect on the virulence of the fungus, RNA was extracted from ¯otation-gradient-puri®ed mitochondria derived from plasmid-free and plasmidcontaining isogenic strains. The RNA molecules were sorted by size using electrophoresis and blotted onto modi®ed nylon membranes. The Northern blots were then examined for the presence of speci®c transcripts by hybridization with a probe prepared from a cloned complete unit of pCRY1 DNA (pE1). As expected, no molecules exhibiting sequence-complementarity to either strand of the plasmid DNA were detected in the RNA preparation from the plasmid-free strain (Fig. 6B), even though the amount of this RNA that was loaded onto the gel was higher than that of the plasmid-bearing strain (Fig. 6A). In contrast, in the RNA from the plasmid-bearing strain, the probe hybridized with at least ten dierent classes of discretely sized RNA molecules, ranging from 0.85 kb to more than 10 kb. It is unlikely that the any part of the hybridization pattern was caused by contamination of the RNA by DNA, for the pro®le was not altered when the samples were treated with DNase. The most-abundant class is composed of molecules that are approximately 4.2-kb long and might be full-length transcripts of the basic unit of pCRY1 DNA. The RNAs that are longer than the 4.2-kb class could originate either by the transcription of more than one unit of plasmid DNA in multimeric forms of pCRY1, or by the repeated and continuous movement of RNA polymerase around unit-length circles of plasmid DNA. Consequently, the RNA species that are shorter than 4.2 kb could be partial transcripts or byproducts of the processing of larger molecules into mature molecules of unit length. Regardless of whether or not primary transcripts are processed, the distribution of the smaller than 4.2-kb molecules into ®ve distinct size classes implies that large molecules are either cleaved at very speci®c sites or that transcription is initiated and terminated at a small number of speci®c locations within the plasmid genome.
Geographical distribution of plasmids in the pCRY1 clan As part of a study on the role of mitochondria in the spontaneous appearance of cytoplasmically transmissible hypovirulence phenotypes in natural populations of C. parasitica, strains isolated from chestnut trees that are located at many dierent sites in the Northeastern United States, Southern Ontario in Canada, China and Japan, were examined by DNA hybridization for the presence of plasmids that are related to pCRY1. The results of this survey are summarized in Table 1 and indicate that some local populations of the fungus in Connecticut, Michigan and Ontario may have been colonized by pCRY1 or closely related plasmids, whereas plasmids of this type may be relatively scarce in China and Japan. In a more-limited survey, Gobbi et al. (1997) found that 41% of the C. parasitica isolates from a very restricted area in Italy contained pUG1, but detected no plasmid in an unspeci®ed number of isolates that originated from the same site in Southern Ontario where our study detected plasmids in two out of two isolates. This discrepancy can not be ascribed to dissimilarities in the probes that were used to detect plasmids by Southern-blot hybridization, for the nucleotidesequence identity between pCRY1 and pUG1 is greater than 99%. Interestingly, all the plasmids that were found in the North American isolates of C. parasitica more closely resembled pCRY1 than pUG1 in that their DNAs contain only one KpnI restriction site, rather than two. This feature, together with the number of 60-bp repeats in the non-coding region of the plasmid genome, may prove to be a convenient means for distinguishing members of the North American family from the European family of plasmids in the pCRY1 clan during studies on the origin, distribution and migration of these genetic elements.
Table 1 Occurrence of plasmids of the pCRY1 family in isolates of C. parasitica from North America and Asia Location Connecticut, USA (1)a Michigan, USA (5) Maryland, USA (2) North Carolina, USA (1) Virginia (1) Ohio, USA (1) West Virginia, USA (1) Wisconsin, USA (1) Ontario, Canada (2) Japanb Chinab a
Number of isolates
Number of isolates containing pCRY1
12 24 4 1 2 3 2 2 3 23 23
6 3 0 0 0 0 0 0 2 0 0
The number of dierent sites sampled in the named state (USA) or province (Canada) is provided in parenthesis b A sub-sample of the strains isolated at the sites in China and Japan described by Milgroom et al. (1996) was used in this survey
252
Discussion As pointed out in a relatively recent review by Griths (1995), the mitochondrial plasmids of ®lamentous fungi not only display considerable diversity in their physical structures, abundance, geographical distribution, transmission and host range, but very few have been shown to aect the growth, longevity, host-speci®city or virulence of their respective hosts. In the present paper, we provide the results of experiments that were aimed at establishing the most-fundamental aspects of the structure, genetic properties, transmission and function of a plasmid, pCRY1, that previously was observed to be associated preferentially, but not exclusively, with attenuated strains of C. parasitica under natural ecological conditions (Bell et al. 1996). These results not only impact on some of the current views on the molecular arrangement, infectious characteristics, phylogenetic relationships and distribution of the type-II mitochondrial plasmids in C. parasitica and Neurospora spp., but also show that some members of this group of apparently innocuous genetic elements can in subtle ways alter the biological properties of their host. On the basis of their electrophoretic behavior on unidimensional gels, restriction maps and nucleotide sequences, it was generally accepted that elements like pCRY1, and in particular the type-I (Mauriceville and Varkud) and type-II (Fiji and LaBelle) mitochondrial plasmids of Neurospora, are circular DNA molecules (Nargang 1985; Griths 1995). However, through data obtained by pulse-®eld gel electrophoresis, Maleszka (1992) concluded that over 90% of the Mauriceville and LaBelle DNAs exist in vivo as a group of linear molecules of heterogeneous size, and that circular forms are rare and represent less than 9% of the population of molecules. The resolution of the pCRY1 DNA into circular and linear molecules by 2D gel electrophoresis (Fig. 2) does not support this view, but shows that, in vivo, the prevailing form of this plasmid is a series of circular molecules consisting of multiples of the 4.2-kb unit genome. These arrays of circular tandem repeats are consistent with the notion that these units replicate by a rolling-circle mechanism, as suggested by Maleszka (1992), who reached this conclusion because some of the linear molecules displayed migration inertia in pulsephase electrophoretic ®elds characteristic of the presence of single-stranded regions. However, the mechanisms that are involved in the replication of circular mitochondrial plasmids have not been examined rigorously at this time, except for the demonstration that reverse transcriptases encoded by Mauriceville and Varkud play a signi®cant role in the maintenance of these retroplasmids in Neurospora (Kennell et al. 1994, 1995; Lambowitz and Chiang 1995). Moreover, a rolling-circle mode of replication has been ®rmly established for a circular plasmid-like DNA that exists in the mitochondria of the higher plant Chenopodium album (Backert et al. 1996, 1997, 1998).
At least two completely dierent observations con®rm that pCRY1 is located in the mitochondria of C. parasitica: (1) it appears as a unique class of DNA that co-puri®es with mitochondria but can be separated from the bulk of the mtDNA by agarose-gel electrophoresis, and (2) the element has a nucleotide composition (69% A+T) and codon bias (UGA Trp) that are characteristic of the mitochondrial DNAs of the ®lamentous ascomycetes. Since it is established that the mtDNA of C. parasitica follows a uniparental (maternal) line of inheritance in crosses (Milgroom and Lipari 1993; Mahanti and Fulbright 1995; Monteiro-Vitorello et al. 1995; Polashock et al. 1997), it is not surprising that pCRY1 is also transmitted in this fashion during the sexual process. In contrast to a hypovirulence-associated mitochondrial dsRNA element which is transmitted in crosses from the maternal parent to only 50% of the ascospore progeny (Polashock et al. 1997), pCRY1 is inherited from the maternal parent by almost all, or all, of the progeny. In this respect, the mitochondrial DNA plasmid contrasts sharply from another class of hypovirulence-eliciting cytoplasmic genetic elements, namely the non-mitochondrial dsRNA hypoviruses, which commonly are lost during the sexual process (Anagnostakis 1982; Nuss 1992). Since the sexual elimination of dsRNA viruses diminishes their effectiveness for controlling the virulence of the fungus (Anagnostakis 1990), attempts have been made to circumvent such losses by inserting into nuclear chromosomes heritable DNA copies that are transcribed and elicit the production of cytoplasmic dsRNAs (Nuss 1992, 1996). So far, long-term control of the fungus in a natural ecosystem, however, has not been achieved, possibly because the transgene is rapidly lost from the gene pool of treated populations (Anagnostakis et al. 1998). At least in theory, the use of maternally inherited hypovirulence-eliciting mitochondrial plasmids for the biological attenuation of deleterious ®lamentous fungi that have a sexual cycle circumvents the need to produce and release transgenic pathogens. In the fungi, one of the best-established criteria for the cytoplasmic location of genetic elements is their migration between strains through hyphal anastomoses independently of the transfer of nuclear marker genes (reviewed by Gillham 1978). It appears that the traf®cking of protoplasm through hyphal anastomoses is the primary, and perhaps only, route of natural ``infection'' of mycelia by hypoviruses (MacDonald and Fulbright 1991; Nuss 1992; Heiniger and Rigling 1994 ). In this context, the transmission of pCRY1 by hyphal contact from donor to recipient strains establishes it as being an ``infectious'' agent, even though there is no reason to believe that it is a virus. The processes involved in the colonization of recipient hyphae, however, remain to be elucidated. For example, it is not clear whether mitochondria from donor strains multiply preferentially and displace the organelles of the recipient, or if the plasmid molecules that are ferried by a few mitochondria from the donor across hyphal bridges
253
replicate extensively in the recipient and actively invade the resident, plasmid-free organelles. Since Polashock et al. (1997) detected mtDNA recombination during studies on the vegetative transmission of a mitochondrial dsRNA, it is clear that the mitochondria from donor and recipient strains of C. parasitica fuse with each other at least in the transient heteroplasmons that are formed in the vicinity of points of cytoplasmic exchange. These aspects are further explored in a study on the eects on plasmid transmission of nuclear genes that determine heterokaryon formation and compatibility in C. parasitica (Baidyaroy et al. 2000). It is intriguing to note that pCRY1 diminishes the virulence of at least one strain of C. parasitica, F2.36. That this plasmid might have an eect on pathogenicity was anticipated from the observation that it appeared to be associated preferentially with isolates of the fungus that either manifested or developed hypovirulence syndromes which occasionally were associated with respiratory defects (Bell et al. 1996). At the same time, it is already apparent that this virulence-depressing eect is not universal, for strain Ep339 contains high amounts of pCRY1 DNA but is as virulent on apples as the Ep155 standard wild-type strain (Monteiro-Vitorello et al. 1995). Moreover, as shown elsewhere (Baidyaroy et al. 2000) in a more detailed analysis of the interaction of the plasmid with dierent strains of C. parasitica, pCRY1 has a more severe eect on the virulence of C. parasitica on chestnut trees than anticipated from the apple test. The nucleotide sequence of pCRY1 establishes clearly that it is a type-II mitochondrial plasmid. While plasmids of this type encode type-B DNA polymerases, the unique structural features of the enzymes that are encoded by this particular type of genetic element will become more evident as the nucleotide sequences of new members are added to the existing database. Even though the type-II plasmids are circular, the polymerase domains of pCRY1, like those of the Fiji and LaBelle plasmids (Li and Nargang 1993), more closely resemble those of the protein-primed B-type polymerases of linear mitochondrial plasmids, linear bacteriophages and viruses than the nucleic acid-primed-B-type polymerases of bacteriophages, viruses and bacteria that have circular genomes (Court and Bertrand 1992). The feature that de®nitely distinguishes the polymerases of type-II plasmids from those of type-III plasmids is the appearance of TTD instead of DTD in the highly conserved Pol-III motif. Considering the high degree of dissimilarity between the nucleotide sequences of Fiji, LaBelle and pCRY1/pUG1, it is clear that TTD is one of the signature motifs of the type-II mitochondrial plasmids and the corresponding subgroup of type-B polymerases. Some of the highly conserved regions that appear in the alignment of proteins in Fig. 5, such as the one located immediately downstream from the POL-II motif, suggest that a battery of additional unique amino-acid-sequence signatures might emerge as more type-II plasmids are sequenced. Nonetheless, the three motifs that are characteristic of 3¢®5¢ exonuclease activity as-
sociated with the amino-terminal editing domain of the B-type polymerases (Wang et al. 1989; Ito and Braithwaite 1991) could not be identi®ed by either Li and Nargang (1993) or Gobbi et al. (1997). However, short sequences that are related to the ``classical'' exonuclease motifs of the B-class of polymerases were identi®ed tentatively in all the type-II plasmids in the present study (Figs. 4 and 5). The identi®cation of motifs characteristic of an editing domain is complicated by uncertainty about the association of 3¢®5¢ exonuclease activity or editing function with the proteins encoded by the type-II plasmids. In fact, the appearance of numerous microsatellite-like repeats within the nucleotide sequences of pCRY1 and the loss of one copy of a 60-bp direct repeat from pUG1 can be invoked to argue that at least these plasmids are duplicated by a mechanism that is prone to replication slippage. Moreover, the high degree of nucleotide-sequence dissimilarity among the related pCRY1, Fiji and LaBelle plasmids, and the base substitutions that distinguish pCRY1 from pUG1, could indicate that the polymerases of the type-II plasmids are devoid of an exonuclease activity that removes mismatched base pairs from the growing ends of newly synthesized DNA strands. Potentially, mitochondrial double-stranded DNA plasmids could be replicated by the mitochondrial DNA polymerase complex by a rolling-circle mechanism similar to that seen for the replication of a plasmid-like, circular mitochondrial element in Chenopodium (Backert et al. 1998) and small circular derivatives of the mtDNA in Neurospora (G. Hausner and H. Bertrand, unpublished observation). Since the type-II plasmids, including pCRY1, are transcribed in vivo (Nargang et al. 1992; this communication) and their unique DNA polymerases appear to prefer the corresponding plasmid DNAs as a template for replication (Schulte and Lambowitz 1991; Li and Nargang 1993), it is likely that at least the proteins encoded by these elements are involved in their maintenance and/or biological activity. The expression of the genomes of the type-II plasmids, however, may be more complex than anticipated from sequence data alone, since both pCRY1 (this paper) and the LaBelle plasmid (Pande et al. 1989) are transcribed into several discrete RNAs that are shorter than any transcript that would cover the long ORF that encodes the putative DNA polymerase. The role of these short transcripts remains obscure, and its elucidation awaits the development of techniques for the genetic manipulation of at least one representative type-II plasmid. It is conceivable, however, that the arrogation of the mitochondrial transcription and translation apparatus for the synthesis or translation of the pCRY1 transcripts interferes with the optimal expression of essential mitochondrial genes in C. parasitica. If this eect is modulated by nutritional factors, then it could elicit reduced growth of at least some strains of the fungus on apples and thus result in the manifestation of the hypovirulence phenotype. From an evolutionary point of view, it can be easily accepted that pCRY1 and pUG1 are genetic variants of
254
the same plasmid. Even so, they exhibit more dierences in their nucleotide sequences than might be expected in the exons of any two alleles of a mitochondrial gene within the same species. Since the mutation rates for this type of genetic element have not been established, the time of divergence between these two plasmids can not be estimated. However, the lack of a detectable level of similarity of the nucleotide sequences of the three main representatives of the type-II plasmids, LaBelle, Fiji and pCRY1/pUG1, suggests that his may be a class of rapidly evolving genetic elements, related to each other by little more than the functional constraints of their unique DNA polymerases. In spite of their dissimilarity at the nucleotide-sequence level, their kinship can be recognized not only from the relatedness of their proteins, but also from the structure and size of their circular genomes. The Fiji plasmid diers from the other Neurospora member of its type, LaBelle, as well as pCRY1, by an in-frame insert of approximately 400 bp in the spacer region between the POL-I and POL-II motifs. In addition to the apparent relationship of their polymerases to the DNA polymerases of linear mitochondrial plasmids and linear bacteriophages, the molecular characteristics of the type-II elements suggest that they may have played a special role in the evolution of a number of dierent types of mitochondrial plasmids. The source of the closely related pCRY1 and pUG1 plasmids that have colonized at least some local populations of C. parasitica in North America and Europe, respectively, remains a mystery. It is virtually certain that the fungus was introduced into both continents from the Orient during the last 100 years (Anagnostakis 1987; Heiniger and Rigling 1994). Hence, the occurrence of pCRY1, or very closely related plasmids, in 11 of 51 North American isolates of C. parasitica (Table 1) and in 41% of the isolates from one site in Italy (Gobbi et al. 1997), suggests that the plasmid was introduced from Asia with its fungal host. Yet, a survey of 48 isolates collected from a variety of sites in China and Japan has not so far uncovered a plasmid that is even remotely related to pCRY1. Therefore, the possibility that C. parasitica acquired the two plasmids, or an ancestral form of both, from a related species in North America or Europe can not be eliminated, although it must be considered less likely than the introduction of the plasmid(s) with the fungus from eastern Asia. Since pCRY1 can be distinguished from pUG1 on the basis of dierences in their nucleotide sequences, the discovery of their source(s) is a tractable problem that probably can be solved through an extension of existing studies on the intercontinental population structure of C. parasitica (Milgroom et al. 1996; Peever et al. 1997, 1998) and related species. Acknowledgments This work was supported in part by United States Department of Agriculture grants 9200717 and 95±37303± 1785. C.B.M.-V. was supported in part by a fellowship from the CNPq of Brazil and D.B. by a grant from the Michigan Agricultural Research Program. We thank Dr. Michael G. Milgroom for kindly providing strains collected in China and Japan and Brian Shaw for expert technical assistance.
References Akins RA, Kelley RL, Lambowitz AM (1986) Mitochondrial plasmids of Neurospora: integration into mitochondrial DNA and evidence for reverse transcription in mitochondria. Cell 47: 505±516. Akins RA, Grant DM, Stohl LL, Bottdorf DA, Nargang FE, Lambowitz AM (1988) Nucleotide sequence of the Varkud mitochondrial plasmid of Neurospora and synthesis of a hybrid transcript with a 5¢ leader derived from mitochondrial DNA. J Mol Biol 204: 1±25 Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignment tool. J Mol Biol 215: 403±410 Altschul SF, Madden TL, Schaer AA, Zhang J, Zhang Z, Miller W, Lipman, DJ (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 3389±3402 Anagnostakis SL (1979) Sexual reproduction of Endothia parasitica in the laboratory. Mycologia 71: 213±215 Anagnostakis SL (1982) Biological control of chestnut blight. Science 215: 466±471 Anagnostakis SL (1987) Chestnut blight: the classical problem of an introduced pathogen. Mycologia 79: 23±37 Anagnostakis SL (1990) Improved chestnut tree condition maintained in two Connecticut plots after treatments with hypovirulent strains of the chestnut blight fungus. For Sci 36: 113±124 Anagnostakis SL, Chen B, Geletka LM, Nuss DL (1998) Hypovirus transmission to ascospore progeny by ®eld-released transgenic hypovirulent strains of Cryphonectria parasitica. Phytopathology 88: 598±604 Arganoza MT, Min J, Hu Z, Akins RA (1994) Distribution of seven homology groups of mitochondrial plasmids in Neurospora: evidence for widespread mobility between species in nature. Curr Genet 26: 62±73 Ausubel F, Brent R, Kingston RE, Moore, DD, Seidman JG, Smith JA, Struhl K (1987) Current protocols in molecular biology. Greene and Wiley, New York Backert S, DoÈrfel P, Lurz R, BoÈrner T (1996) Rolling-circle replication of mitochondrial DNA in the higher plant Chenopodium album (L.). Mol Cell Biol 16: 6285±6294 Backert S, Meiûner K, BoÈrner T (1997) Unique features of the mitochondrial rolling-circle plasmid mp1 from the higher plant Chenopodium album (L.). Nucleic Acids Res 25: 582±590 Backert S, Kunnimalaiyaan M, BoÈrner T, Nielsen BL (1998) In vitro replication of mitochondrial plasmid mp1 from the higher plant Chenopodium album (L.): a remnant of bacterial rollingcircle and conjugative plasmids? J Mol Biol 284: 1005±1015 Baidyaroy D, Glynn JM, Bertrand H (2000) Dynamics of asexual transmission of a mitochondrial plasmid in Cryphonectria parasitica. Curr Genet 37: 257±267 Bell JA, Monteiro-Vitorello CB, Hausner G, Fulbright DW, Bertrand H (1996) Physical and genetic map of the mitochondrial genome of Cryphonectria parasitica Ep155. Curr Genet 30: 34±43 Bertrand H, Griths AJF, Court DA, Cheng C K (1986) An extrachromosomal plasmid is the etiological precursor of kalDNA insertion sequences in the mitochondrial chromosome of senescent Neurospora. Cell 47: 829±837 Brewer BJ, Fangman WL (1987) The localization of replication origins on ARS plasmids in S. cerevisiae. Cell 51: 463±471 Chan BS-S, Court DA, Vierula PJ, Bertrand H (1991) The kalilo senescence-inducing plasmid of Neurospora is an invertron and encodes DNA and RNA polymerases. Curr Genet 20: 225±237 Collins RA, Saville BJ (1990) Independent transfer of mitochondrial chromosomes and plasmids during unstable vegetative fusion in Neurospora. Nature 345: 177±179 Collins RA, Stohl LL, Cole MD, Lambowitz AM (1981) Characterization of a novel plasmid DNA found in mitochondria of Neurospora crassa. Cell 24: 443±452 Court DA, Bertrand H (1992) Genetic organization and structural features of maranhar, a senescence±inducing linear mito-
255 chondrial plasmid of Neurospora crassa. Curr Genet 22: 385± 397 Court DA, Griths AJF, Kraus SR, Russell PJ, Bertrand H (1991) A new senescence-inducing linear plasmid in ®eld-isolated Neurospora crassa strains from India. Curr Genet 19: 129±137 Fulbright DW (1984) Eect of eliminating dsRNA in hypovirulent Endothia parasitica. Phytopathology 74: 722±724 Gillham NW (1978) Organelle heredity. Raven Press, New York Gobbi E, Carpanelli G, Locci R (1997) The Cryphonectria parasitica plasmid pUG1 contains a large ORF with motifs characteristic of family B DNA polymerases. Nucleic Acids Res 25: 3275±3280 Griths AJF (1992) Fungal senescence. Annu Rev Genet 26: 351± 372 Griths AJF (1995) Natural plasmids of ®lamentous fungi. Microbiol Rev 59: 673±685 Griths AJF, Kraus SR, Barton R, Court DA, Myers CJ, Bertrand H (1990) Heterokaryotic transmission of senescence plasmid DNA in Neurospora. Curr Genet 17: 139±145 Haen¯er J, Teepe H, Weigl C, Kruft V, Lurz R, Woestemeyer J (1992) Circular extrachromosomal DNA codes for a surface protein in the (positive) mating type of the zygomycete Absidia glauca. Curr Genet 22: 319±325 Heiniger U, Rigling D (1994) Biological control of chestnut blight in Europe. Annu Rev Phytopathol 32:581±599 Hermanns J, Asseburg A, Osiewacz HD (1994) Evidence for a lifespan prolonging eect of a linear plasmid in a longevity mutant of Podospora anserina. Mol Gen Genet 243: 297±307 Huber DH (1996) Genetic analysis of vegetative incompatibility polymorphisms and horizontal transmission in the chestnut blight fungus, Cryphonectria parasitica. PhD thesis, Michigan State University Ito J, Braithwaite DK (1991) Compilation and alignment of DNA± polymerase sequences. Nucleic Acids Res 19: 4045±4057 Katsuya S, Kaneko I, Owaki M, Ishikawa K, Tsujimoto T, Tsuge T (1997) Circular DNA plasmid in the phytopathogenic fungus Alternaria alternata: its temperature-dependent curing and association with pathogenicity. Genetics 146: 111±120 Kawano S, Takano H, Imai J, Mori K, Kuroiwa T (1993) A genetic system controlling mitochondrial fusion in the slime mold, Physarum polycephalum. Genetics 133: 213±224 Kennell JC, Wang E, Lambowitz AM (1994) The Mauriceville plasmid of Neurospora spp. uses novel mechanisms for initiating reverse transcription in vivo. Mol Cell Biol 14: 3094±3107 Kennell JC, Saville BJ, Mohr S, Kuiper MT, Sabourin JR, Collins RA, Lambowitz AM (1995) The VS catalytic RNA replicates by reverse transcription as a satellite of a retroplasmid. Genes Dev 9: 294±303 Kim WK, Mauthe W, Hausner G, Klassen GR (1990) Isolation of high molecular weight DNA and double-stranded RNAs from fungi. Can J Bot 68: 1898±1902 Lambowitz AM, Chiang C-C (1995) The Mauriceville and Varkud plasmids: primitive retroelements found in Neurospora mitochondria. Can J Bot 73 (suppl): S173-S179 Li Q, Nargang FE (1993) Two Neurospora mitochondrial plasmids encode DNA polymerases containing motifs characteristic of family B DNA polymerases but lack the sequence asp±thr±asp. Proc Natl Acad Sci USA 90: 4299±4303 MacDonald WL, Fulbright DW (1991) Biological control of chestnut blight: use and limitations of transmissible hypovirulence. Plant Dis 75: 656±661 Mahanti N, Fulbright DW (1995) Detection of mitochondrial DNA transfer between strains after vegetative contact in Cryphonectria parasitica. Mol Plant-Microbe Interact 8: 465±467 Mahanti N, Bertrand H., Monteiro-Vitorello C, and Fulbright D W (1993) Elevated mitochondrial alternative oxidase activity in dsRNA±free, hypovirulent isolates of Cryphonectria parasitica. Physiol Mol Plant Pathol 42: 455±463 Maleszka R (1992) Electrophoretic pro®les of mitochondrial plasmids in Neurospora suggest they replicate by a rollingcircle mechanism. Biochem Biophys Res Commun 186: 1669± 1673
Martin FN (1991) Characterization of circular mitochondrial DNA plasmids in three Pythium species. Curr Genet 20: 91±97 Milgroom MG, Lipari SE (1993) Maternal inheritance and diversity of mitochondrial DNA in the chestnut blight fungus, Cryphonectria parasitica. Phytopathology 83: 563±567 Milgroom MG, Wang KR, Zhou Y, Lipari SE, Kaneko S (1996) Intercontinental population structure of the chestnut blight fungus, Cryphonectria parasitica. Mycologia 88: 179±190 Monteiro-Vitorello CB, Bell JA, Fulbright DW, Bertrand H (1995) A cytoplasmically transmissible hypovirulence phenotype associated with mitochondrial DNA mutations in the chestnut blight fungus Cryphonectria parasitica. Proc Natl Acad Sci US 92: 5935±5939 Nakagawa CC, Jones EP, Miller DL (1998) Mitochondrial DNA rearrangements associated with mF plasmid integration and plasmodial longevity in Physarum polycephalum. Curr Genet 33: 178±187 Nargang FE (1985) Fungal mitochondrial plasmids. Exp Mycol 9: 285±293 Nargang FE, Bell JB, Stohl LL, Lambowitz AM (1984) The DNA sequence and genetic organization of a Neurospora mitochondrial plasmid suggest a relationship to introns and mobile elements. Cell 38: 41±453 Nargang FE, Pande S, Kennell JC, Akins RA, Lambowitz AM (1992) Evidence that a 1.6-kilobase region of Neurospora DNA was derived by insertion of part of the LaBelle mitochondrial plasmid. Nucleic Acids Res 20: 1101±1108 Nishimura A, Morita M, Nishimura Y, Sugino Y (1990) A rapid and highly ecient method for preparation of competent Escherichia coli cells. Nucleic Acids Res 18: 6169 Nuss DL (1992) Biological control of chestnut blight: an example of virus±mediated attenuation of fungal pathogenesis. Microbiol Rev 56: 561±576 Nuss DL (1996) Using hypoviruses to probe and perturb signal transduction processes underlying fungal pathogenesis. Plant Cell 8: 1845±1853 Pande S, Lemire EG, Nargang FE (1989) The mitochondrial plasmid from Neurospora intermedia strain LaBelle-1b contains a long open reading frame with blocks of amino acids characteristic of reverse transcriptases and related proteins. Nucleic Acids Res 17: 2023±2042 Peever TL, Liu YC, Milgroom MG (1997) Diversity of hypoviruses and other double-stranded RNAs in Cryphonectria parasitica in North America. Phytopathology 87: 1026±1033 Peever TL, Liu YC, Wang KR, Hillman BI, Foglia R, Milgroom MG (1998) Incidence and diversity of double±stranded RNAs occurring in the chestnut blight fungus, Cryphonectria parasitica, in China and Japan. Phytopathology 88: 811±817 Perkins D (1999) Neurospora genetic nomenclature. Fungal Genet Newslett 46: 34±41 Polashock JJ, Bedker PJ, Hillman BI (1997) Movement of a small mitochondrial double-stranded RNA element of Cryphonectria parasitica: ascospore inheritance and implications for mitochondrial recombination. Mol Gen Genet 256: 566±571 Puhalla JE, Anagnostakis SL (1971) Genetic and nutritional requirements of Endothia parasitica. Phytopathology 61: 169±173 Queen C, Korn LJ (1984) A comprehensive sequence analysis program for the IBM personal computer. Nucleic Acids Res 12: 581±599 Ryan FJ, Beadle GW, Tatum EL (1943) The tube method for measuring growth rate in Neurospora. Am J Bot 30: 784±799 Sakaguchi K (1990) Invertrons, a class of structurally and functionally related genetic elements that includes linear DNA plasmids, transposable elements, and genomes of adeno±type viruses. Microbiol Rev 54: 66±74 Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a laboratory manual. 2nd edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York Sanger F, Nicklen S, Coulsen AR (1977) DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74: 5463±5467
256 Schulte U, Lambowitz AM (1991) The LaBelle mitochondrial plasmid of Neurospora intermedia encodes a novel DNA polymerase that may be derived from a reverse transcriptase. Mol Cell Biol 11: 1696±1706 Solymosy F, Fedorcsak I, Guylas A, Farkas GL, Ehrenberg L (1968) A new method based upon the use of diethylpyrocarbonate as a nuclease inhibitor for the extraction of undegraded nucleic acids from plant tissues. Eur J Biochem 5: 520±527 Stohl LL, Collins RA, Cole MD, Lambowitz AM (1982) Characterization of two new plasmid DNAs found in mitochondria of wild-type Neurospora intermedia strains. Nucleic Acids Res 10: 1439±1458 Wang TS-F, Wong SW, Korn D (1989) Human DNA polymerasealpha ± predicted functional domains and relationships with viral-DNA polymerases. FASEB J 3: 14±21
Yang X, Griths AJF (1993a) Plasmid diversity in senescent and non-senescent strains of Neurospora. Mol Gen Genet 237: 177±186 Yang X, Griths AJF (1993b) Plasmid suppressors active in the sexual cycle of Neurospora intermedia. Genetics 135: 993±1002 Yang X, Griths AJF (1993c) Male transmission of linear plasmids and mitochondrial DNA in the fungus Neurospora. Genetics 134: 1055±1062 Yelton MM, Hamer JE, Timberlake WE (1984) Transformation of Aspergillus nidulans by using a trpC plasmid. Proc Natl Acad Sci USA 81: 1470±1474
