Differential Transfer and Dissemination of Hypovirus and Nuclear and Mitochondrial Genomes of a Hypovirus-Infected Cryphonectria parasitica Strain after Introduction into a Natural Population

Abstract

Biological control of plant diseases generally requires release of living organisms into the environment. Cryphonectria hypoviruses function as biological control agents for the chestnut blight fungus, Cryphonectria parasitica, and hypovirus-infected C. parasitica strains can be used to treat infected trees. We used naturally occurring molecular marker polymorphisms to examine the persistence and dissemination of the three genomes of a hypovirus-infected C. parasitica strain, namely, the double-stranded RNA genome of Cryphonectria hypovirus 1 (CHV1) and the nuclear and mitochondrial genomes of its fungal host. The hypovirus-infected strain was experimentally introduced into a blight-infested chestnut coppice forest by treating 73 of 246 chestnut blight cankers. Two years after introduction, the hypovirus had disseminated to 36% of the untreated cankers and to 35% of the newly established cankers. Spread of the hypovirus was more frequent within treated sprout clusters than between sprout clusters. Mitochondrial DNA of the introduced fungus also was transferred into the resident C. parasitica population. Concomitant transfer of both the introduced hypovirus and mitochondrial DNA was detected in almost one-half of the treated cankers analyzed. The introduced mitochondrial DNA haplotype also was found in three resident isolates from newly established cankers. The nuclear genome of the introduced strain persisted in the treated cankers but did not spread beyond them.

Hypovirulence in the chestnut blight fungus, Cryphonectria parasitica (Murr.) Barr, has been responsible for the natural recovery of many European chestnut (Castanea sativa Mill.) stands (reviewed in reference 25). Unlike the situation in Europe, hypovirulence has not become well established in North America except in isolated chestnut stands, and the American chestnut [Castanea dentata (Marsh.) Borkh.] has been reduced to an understory shrub in its natural range due to the disease (reviewed in references 2, 4, 19, 32, and 43).

Hypovirulence of C. parasitica is caused by unencapsidated double-stranded RNA (dsRNA) viruses of the family Hypoviridae (13, 17, 18, 26). These viruses are located in the cytoplasm of the fungus and can be transmitted from infected to uninfected C. parasitica strains via hyphal anastomosis (44). Horizontal transmission is restricted by a vegetative incompatibility system (vic) of the fungus involving at least six vic loci (14, 15, 30). The low diversity of vegetative compatibility (vc) types favors dissemination of the hypovirus and probably is a critical factor in the success of hypovirulence in Europe (7, 11, 16, 34). Dissemination of the hypovirus within a fungal population can occur through infected asexual conidia or mycelial fragments but not through sexual ascospores (reviewed in reference 25).

Hypovirulence is being exploited for the biological control of chestnut blight (5, 25, 41). Since the Cryphonectria hypovirus has no extracellular phase, it can be released into a population only by hypovirus-infected C. parasitica strains. Thus, in addition to the hypovirus, the nuclear DNA and mitochondrial DNA (mtDNA) of the fungal host also are introduced into the environment during biological control treatments. The extent of persistence and dissemination of these introduced genomes in the natural environment are important parameters, particularly if genetically engineered C. parasitica strains are to be released (6, 38).

The objective of this study was to examine the persistence and dissemination of the nuclear and mitochondrial genomes and the hypovirus from a hypovirus-infected C. parasitica strain that was experimentally introduced into a chestnut forest affected by chestnut blight. We used molecular markers to differentiate the resident and introduced genomes and to track the hypovirus and the nuclear and mitochondrial genomes of the introduced strain. The data obtained should increase our understanding of the ecology of this model disease system and could be used for development of future biocontrol strategies.

MATERIALS AND METHODS

Study site and resident C. parasitica population.

The experimental plot (the resident population) was located in Choëx (Valais, Switzerland; latitude, 46°14′40"N; longitude, 6°57′50"E) north of the main mountain range of the Alps. Chestnut blight was first observed in this area in 1986 (24). The plot consisted of a C. sativa coppice (approximately 3,000 m²) that had been clear-cut in 1991. At the beginning of the experiment in 1996, this plot contained 36 sprout clusters with an average of 27 sprouts per cluster. The disease incidence was high, and no Cryphonectria hypoviruses had been found at the site or in the region (24, 27).

In spring 1996, the plot contained 246 chestnut blight cankers, from which 232 C. parasitica isolates were obtained. None of the 232 isolates had the white culture morphology typically associated with Cryphonectria hypovirus 1 (CHV1) infection (13, 17). Genetic characterization of an arbitrary subsample of 50 isolates showed that there was no diversity or only low diversity for the following markers: vc type (only EU-1 was present), mating type (>90% MAT-1), DNA fingerprint pattern (27), and mtDNA haplotype (only one haplotype was detected).

Fungal strain and hypovirus selection.

Based on characterization of the resident C. parasitica population, strain M1275 from the culture collection of the WSL Swiss Federal Institute for Forest, Snow and Landscape Research (Birmensdorf, Switzerland) was selected as the carrier of the hypovirus. This strain was originally isolated in 1990 in Gnosca (Ticino, Switzerland). It is hypovirus free, belongs to vc type EU-1, is mating type MAT-1, and has a nuclear DNA fingerprint and an mtDNA haplotype that are easily distinguished from the nuclear DNA fingerprints and mtDNA haplotype in the resident population. The hypovirus (CHV1) originated from C. parasitica strain M784, which was isolated in 1976 in Copera (Ticino, Switzerland) and has been stored lyophilized in the WSL culture collection. The virus of strain M784 has a unique restriction fragment length polymorphism (RFLP) pattern compared to the patterns of Swiss and other European viruses (1), which was used to distinguish it from other hypoviruses. Strain M784 has a white culture morphology and is moderately virulent for C. sativa (9). The hypovirus of M784 was transmitted into the hypovirus-free strain M1275 through hyphal anastomosis (40) to obtain strain VSX, which also had a white culture morphology. The growth of strain VSX on potato dextrose agar was comparable to that of hypovirus-free strain M1275 (28).

Release procedure.

In September 1996, C. parasitica strain VSX was introduced into the experimental plot by traditional biocontrol treatments (20, 25). One-half of the sprout clusters were chosen for canker treatment, and every second canker was treated in the selected clusters (Fig. 1). Holes that were about 20 mm apart and to the depth of the cambium (3 to 6 mm) were made at the periphery of a canker lesion with a cork borer (diameter, 5 mm). Mycelium of strain VSX, grown in liquid culture (39) and harvested by filtration, was placed in the holes. The filled holes were covered with masking tape for 2 weeks to prevent desiccation of the mycelium.

FIG. 1.

Map of the chestnut sprout clusters in the experimental plot. Each circle represents a sprout cluster. The circle size is proportional to the number of cankers (initial situation). The numbers below the circles indicate the number of treated cankers (white sectors)/total number of cankers. The cross-hatched circle represents a sprout cluster without cankers.

Sampling and isolation.

We monitored the persistence and dissemination of the introduced strain by annually sampling all cankers. In fall 1997 and 1998 all of the initial cankers (i.e., the cankers present in spring 1996) were resampled. All new cankers (i.e., cankers that developed after spring 1996) were recorded and sampled in fall 1996, fall 1997, spring 1998, and fall 1998. Cankers that merged with other cankers during the surveys were considered a single canker. Three samples were taken from each canker, as follows: one each from the upper and lower margins and one from the center. All instruments (including the cork borer used for treatments and the biopsy needle used for sampling) were sterilized with 96% ethanol and flaming between cankers. Isolation was performed as previously described (27).

C. parasitica and hypovirus characterization.

All C. parasitica isolates recovered were assessed for culture morphology. Isolates were grown as previously described (11) and were considered to be hypovirus infected if they had the white culture morphology typical of infection with CHV1-type hypoviruses (13, 17) and to be hypovirus free if they had an orange culture morphology. We verified the reliability of culture morphology as an indicator of hypovirus infection by testing a sample of 28 orange isolates and 32 white isolates for the presence of dsRNA (1). No dsRNA was recovered from the orange isolates, but all of the white isolates contained viral dsRNA. The identity of the hypovirus was determined by reverse transcription-PCR-RFLP analysis of a fragment from hypovirus open reading frame A (1). A canker was considered to be hypovirus infected if at least one of the three isolates recovered from it had the white culture morphology. Cankers that yielded no C. parasitica isolates or debilitated isolates were omitted from the analysis.

One hypovirus-infected isolate per hypovirus-infected canker or one hypovirus-free isolate per hypovirus-free canker was arbitrarily selected for further analysis. The vc types of the isolates were determined by pairing with an EU-1 tester strain from the resident C. parasitica population (16). vc types were determined only for the isolates obtained from the first sampling of each canker. The fungal DNA isolation, gel electrophoresis, and nonradioactive hybridization procedures used have been described previously (27). Nuclear DNA fingerprints were determined by hybridization with the multilocus probe pMS5.1 (36). Blots were stripped in 0.2 M NaOH-0.1% sodium dodecyl sulfate and reprobed with purified mtDNA of C. parasitica isolate EP67 (= ATCC 38753) (35).

RESULTS

Dissemination of the introduced hypovirus.

Seventy-three (30%) of the 246 initial cankers (i.e., the cankers present in spring 1996) were treated with strain VSX in fall 1996 (Fig. 1). Due to merging of cankers the total number of cankers was reduced to 230 in fall 1997 and to 228 in fall 1998. In the fall 1997 resampling, 174 cankers yielded isolates with an unambiguous orange (hypovirus-free) or white (hypovirus-infected) culture morphology, and in the fall 1998 resampling 177 cankers could be analyzed. The remaining initial cankers yielded either no C. parasitica or isolates with an abnormal, mostly debilitated culture morphology and were excluded from further analysis. The proportions of treated and untreated cankers analyzed were similar in the fall 1997 and fall 1998 resamplings (35% treated, 65% untreated). Eighty-five percent of the treated cankers in 1997 and 78% of the treated cankers in 1998 yielded hypovirus-infected isolates. Twenty-four percent of the untreated cankers analyzed were infected with hypovirus 1 year after the initial treatment, and 36% were infected with hypovirus after 2 years.

Between the initial survey in spring 1996 and the final sampling in fall 1998, 157 new cankers developed. We periodically sampled all new cankers to determine the dissemination of the introduced hypovirus into these cankers (Table 1). Within 1 month after introduction of the hypovirus-infected strain in fall 1996, 1 of the 48 cankers that had developed since spring 1996 became hypovirus infected. Hypovirus infection increased in the new cankers in subsequent years (Table 1). The greatest increase in hypovirus infection occurred in the new cankers detected in fall 1996, where more than one-half of the cankers analyzed became hypovirus infected in the subsequent years.

TABLE 1.

Persistence and dissemination of the introduced hypovirus in the newly established cankers

Appearance period	No. of new cankers	No. of cankers on date of first sampling or resampling in:
		Fall 1996a	Fall 1997	Spring 1998	Fall 1998
Spring 1996-fall 1996	50	1/48b	22/35	NSc	21/36
Fall 1996-fall 1997	58		13/52	NS	17/44
Fall 1997-spring 1998	24			0/20	2/19
Spring 1998-fall 1998	25				3/23

^aSamples were obtained 1 month after introduction of hypovirus-infected strain VSX.

^bNumber of hypovirus-infected cankers/total number of cankers analyzed.

^cNS, no sampling.

Seventeen (12%) of 143 new cankers were hypovirus infected in the first sampling (Table 1), but 75 (33%) of 224 cankers (initial and new) were converted (i.e., they were hypovirus free in the first sampling and hypovirus infected in at least one of the subsequent samplings). The difference between these two proportions is highly significant (P ≪ 0.01, as determined by a χ² test).

The hypovirus spread in a distance-related manner. In both resamplings of the initial and new cankers in 1997 and 1998, the proportion of naturally converted cankers was significantly higher (P < 0.01, as determined by a χ² test) in treated sprout clusters (31 of 63 and 39 of 93 cankers) than in untreated sprout clusters (17 of 84 and 32 of 108 cankers). The proportion of hypovirus-infected new cankers in treated sprout clusters (16 of 100 cankers) also was significantly higher (P < 0.05, as determined by a χ² test) than the proportion of hypovirus-infected new cankers in untreated sprout clusters (1 of 43 cankers).

We identified a sample of 32 hypoviruses with RFLP markers. Twelve of these hypoviruses were from hypovirus-infected C. parasitica isolates from untreated initial cankers, and 20 were from newly established cankers. With one exception, all of the patterns of the 32 viruses were identical to the pattern of the introduced virus. In the exceptional virus, one fragment from the BsuRI digest differed from the RFLP pattern of the introduced virus, possibly as a result of a single point mutation.

Dissemination of the introduced nuclear DNA and mtDNA.

We identified nuclear DNA fingerprints and mtDNA haplotypes (Fig. 2) in a subset of C. parasitica isolates recovered from the initial cankers and in all of the isolates recovered from the new cankers. Of the 52 hypovirus-infected isolates obtained in the 1997 resampling from the treated cankers, 33 were analyzed. Eight isolates had the nuclear and mitochondrial genotypes of the resident C. parasitica population, indicating that these strains were residents that had been infected with the introduced virus. Ten isolates had the same nuclear and mitochondrial genotypes as the introduced strain, strain VSX. The remaining 15 isolates had the nuclear DNA fingerprint of the resident strain but a mitochondrial haplotype identical to that of the introduced VSX strain (Fig. 2). All 15 isolates recovered from these cankers in spring 1996 (i.e., prior to the biocontrol treatment) had the resident mitochondrial genotype, indicating that transfer of both the hypovirus and mtDNA occurred. Hybridization analysis of undigested total DNA extracts from strain VSX and a selection of isolates in which mtDNA transfer was observed ruled out the possibility that plasmids (cytoplasmic or organellar) were involved (data not shown).

FIG. 2.

Nuclear DNA fingerprints and mtDNA haplotypes of the introduced C. parasitica strain VSX and representative isolates from untreated and treated cankers, sampled in fall 1997: Southern blot of PstI-digested total DNA hybridized with digoxigenin-labeled probe and detected by the chemiluminescence method. The numbers on the left indicate molecular weights (in kilobases). All isolates were hypovirus infected. The asterisks indicate isolates in which mtDNA transfer occurred. (a) Nuclear DNA fingerprints resulting from hybridization with probe pMS5.1. (b) mtDNA haplotypes resulting from hybridization with purified mtDNA.

Furthermore, we analyzed 23 hypovirus-free and 27 hypovirus-infected isolates obtained from untreated cankers in the 1997 resampling. All 50 strains had the nuclear and mitochondrial genotypes of the resident C. parasitica population. Therefore, only the hypovirus was disseminated into these cankers. For the newly established cankers, none of the 143 isolates obtained from the first sampling had the introduced nuclear genotype, and 140 of the 143 isolates had the resident mtDNA haplotype. One isolate with the introduced mitochondrial haplotype was hypovirus free and was recovered from a new canker detected in fall 1996. The two other isolates, one of which was hypovirus free and one of which was hypovirus infected, were obtained from new cankers in 1997.

DISCUSSION

We found different patterns for dissemination of the viral, nuclear, and mitochondrial genomes of the introduced hypovirus-infected C. parasitica strain. The hypovirus spread most efficiently and infected a considerable portion of the resident C. parasitica population. The incidence of hypovirus infection 2 years after the introduction of the hypovirus was 78% for the treated cankers and 36% for all untreated (initial and new) cankers. All of the members of the resident fungal population had the same vc type, so no vegetative incompatibility barriers were present and each hypovirus-infected strain could, potentially, transfer its hypovirus to every other strain in the plot (15, 30). In North America, where vc type diversity is normally high (7, 34), hypovirus CHV1 persisted and disseminated only to a limited extent after release for biological control (21, 31).

The extent of hypovirus dissemination which we observed may have underestimated the amount of dissemination that was actually occurring. Hypovirus infection often is unevenly distributed throughout a canker (21, 42), and hypovirus may not be detected when only three samples are taken per canker. This sampling problem also could explain why only some of the treated cankers yielded hypovirus-infected isolates after resampling. Other explanations for this problem include (i) the possibility that the inoculum failed to become established in the plant tissue and (ii) the possibility that the hypovirus was eliminated due to unfavorable conditions in the fungal host.

Our results demonstrate that the hypovirus spread predominantly within existing cankers rather than by initiating new hypovirus-infected cankers. The proportion of converted cankers was significantly higher than the proportion of new hypovirus-infected cankers. Furthermore, the new cankers might have been initiated by hypovirus-free strains that were converted to hypovirus-infected strains afterwards, especially since there was at least a 6-month gap between the times that newly established cankers were sampled. In genetically heterogeneous populations, in which conversion is restricted by vegetative incompatibility, however, canker initiation by hypovirus-infected strains may be more important.

Introduced mtDNA also was transferred into the resident C. parasitica population. About one-half of the analyzed hypovirus-infected isolates from treated cankers had a resident nuclear DNA fingerprint and the introduced mtDNA haplotype. The introduced mtDNA haplotype also was found in three resident isolates from newly established cankers, indicating that both the mtDNA and the hypovirus were transmissible. It is possible that complete mitochondria were transferred from VSX to the resident strains. However, we found no mitochondrial heteroplasmons (i.e., strains with a mixture of the two mtDNA haplotypes). Instead, the distinctive bands in the hybridization patterns were either completely absent or present at equimolar levels, indicating that only one mtDNA haplotype was present. These patterns could have resulted from cytoplasmic segregation after heteroplasmon formation (reviewed in reference 22) or if only a portion of the mitochondrial genome was transferred. Mitochondrial group I and group II introns can act as mobile genetic elements and have been reported to be present in C. parasitica (23, 29). These introns can spread efficiently and result in a homoplasmon. Although these introns can transpose to other locations, they normally transfer to intronless alleles of the same gene, which could explain why we always observed the same change in the band pattern. Although the horizontal transfer mechanisms remain unknown, our results provide rare direct evidence that this transfer phenomenon occurs under field conditions.

The mtDNA transmission rate in the treated cankers was high enough that biological control with mitochondrial hypovirulence (10, 37) could be considered. C. parasitica strains with virulence-attenuating mtDNA mutations have been found in field populations and have been induced under laboratory conditions (8, 33, 37). In contrast to hypoviruses, mitochondrial hypovirulence also is transmissible sexually (37). Therefore, a combined treatment strategy with both hypovirus and mitochondrial hypovirulence could, under certain circumstances, enhance biological control (e.g., in populations with frequent sexual reproduction).

The nuclear genome of the introduced strain persisted in the treated cankers but did not spread beyond them. In fall 1997, about one-half of the analyzed hypovirus-infected isolates from the treated cankers contained either resident or introduced nuclear and mitochondrial genomes. Strain VSX was applied to the periphery of the treated cankers, so we expected to find both resident and introduced strains in these cankers. The nuclear DNA fingerprint of the introduced strain was detected neither in the naturally converted initial cankers (resampled in 1997) nor in the newly established cankers. The most likely explanation for the observed lack of dissemination of the introduced nuclear genome is strong suppression of conidiation in biocontrol strain VSX. When inoculated into healthy chestnut stems, VSX produced only small cankers, which were rapidly callused by plant tissue and rarely produced any stromata or conidiospores.

Lack of dissemination of the introduced nuclear DNA is desirable in traditional biological control of chestnut blight because it reduces the risk of increasing the diversity of the resident fungal population. However, transgenic hypovirulent strains require dissemination of nuclear DNA for success. These strains contain a chromosomally integrated cDNA copy of the hypovirus genome and cytoplasmic viral dsRNA derived from the cDNA and can be used to disseminate the hypovirus through sexual reproduction (6, 38). Although both natural and transgenic hypovirulent strains are female sterile, they can still serve as male parents in sexual crosses (3, 12). The role of sexual reproduction in dissemination of the nuclear DNA of an introduced hypovirus-infected strain requires further study. The population in Choëx is asexual, with one mating type dominant (27). The VSX strain introduced in Choëx had the same mating type as the resident population, so its nuclear DNA could not be readily disseminated through sexual reproduction, and no strains with recombinant fingerprints were detected. The observed lack of dissemination of the introduced nuclear DNA in this population, however, should not be extrapolated to populations in which sexual reproduction is frequent and crosses between resident and introduced strains could take place.

To explain the dissemination patterns that we observed, we hypothesized that shortly after the treatment, introduced strain VSX and the resident canker strain came in contact and hyphal anastomosis occurred. The hypovirus was transmitted from VSX to the resident strain in most of the treated cankers, and in many cases mtDNA was transferred as well. This transmission resulted in cankers containing converted resident C. parasitica strains with or without the introduced mtDNA. In some of these cankers, the hypovirus and, less frequently, the mtDNA moved into already established stromatal tissue in which conidia were produced by the resident fungus, which may have contained the hypovirus and the introduced mtDNA. The suppressive effects of the hypovirus on sporulation may not have played a role, as the structures for sporulation were formed by the hypovirus-free resident fungus before hypovirus infection. The conidia could move, mostly over short distances, and convert existing cankers or initiate new cankers. The VSX biocontrol strain with its nuclear DNA persisted at the edges of the treated cankers but did not spread due to reduced asexual sporulation. Further analysis in this experimental plot in a few years should show whether the introduced viral, nuclear, and mitochondrial genomes still persist, whether the nuclear genome has been disseminated, and whether recombination between VSX and resident strains has occurred.