Infectious cDNA Clone of Hypovirus CHV1-Euro7: a Comparative Virology Approach To Investigate Virus-Mediated Hypovirulence of the Chestnut Blight Fungus Cryphonectria parasitica

Abstract

We report the construction of a full-length infectious cDNA clone for hypovirus CHV1-Euro7, which is associated with reduced virulence (hypovirulence) of the chestnut blight fungus Cryphonectria parasitica. Field strains infected with CHV1-Euro7 are more virulent and exhibit less severe phenotypic changes (hypovirulence-associated traits) than strains infected with the prototypic hypovirus CHV1-EP713, for which the first infectious cDNA clone was developed. These differences exist even though the two hypoviruses show extensive sequence identities: 87 to 93% and 90 to 98% at the nucleotide and amino acid levels, respectively. The relative contributions of viral and host genomes to phenotypic traits associated with hypovirus infection were examined by transfecting synthetic transcripts of the two hypovirus cDNAs independently into two different virus-free C. parasitica strains, EP155 and Euro7(−v). Although the contribution of the viral genome was clearly predominant, the final magnitude and constellation of phenotypic changes were a function of contributions by both genomes. The high level of sequence identity between the two hypoviruses also allowed construction of viable chimeras and mapping of the difference in symptom expression observed for the two viruses to the open reading frame B coding domain. Implications of these results for engineering enhanced biological control and elucidating the basis for hypovirus-mediated attenuation of fungal virulence are discussed.

Hypoviruses are a family of cytoplasmically replicating RNA viruses that persistently attenuate virulence (hypovirulence) and alter a number of complex biological processes, e.g., pigment production, asexual sporulation, and mating (hypovirulence-associated traits), of their fungal host, the chestnut blight fungus Cryphonectria parasitica (1, 31, 34, 35). Hypoviruses and accompanying hypovirulence can be transmitted cytoplasmically to vegetatively compatible virulent C. parasitica strains via anastomosis (fusion of hyphae), providing the basis for successful biological control of chestnut blight observed in European forests and orchards (5, 23). Natural or introduced hypovirulence-mediated biological control has been much less successful in North American forest ecosystems (1, 22, 31). Factors contributing to this lower efficacy include but may not be limited to barriers to cytoplasmic spread due to a high degree of diversity in vegetative compatibility in North American C. parasitica populations relative to populations in Europe (2, 4, 5, 30).

Hypovirulent C. parasitica field isolates exhibit a wide range of variability in virulence levels and in the magnitude and constellation of hypovirulence-associated traits (16, 18, 20, 31). Results of cytoplasmic transmission studies with natural hypovirus isolates suggest that this variability is primarily a function of hypovirus genetic diversity (17–19). Early attempts at hypovirulence-mediated biological control in North America involved fungal strains that had severely reduced virulence and debilitated mycelial growth and sporulation. While highly curative, these strains showed limited dissemination and persistence (31). These observations led to the suggestion that more effective control might be achieved by the use of more robust, less debilitated hypovirulent C. parasitica strains (31).

While progress in the successful establishment of hypovirulence-mediated biological control proceeds at the moderate pace necessitated by the completion of lengthy field studies, rapid progress was recently achieved in the molecular characterization of hypoviruses. A general view of hypovirus genome organization has been provided by the cloning and complete sequence determination of two hypoviruses (41, 25) and the partial sequence analysis of several others (27, 28, 39, 43). The prototypic hypovirus CHV1-EP713 is found in infected cells as a 12.7-kbp double-stranded RNA (dsRNA) (41). Two contiguous open reading frames (ORF), designated ORF A and ORF B, are located within the polyadenylated coding strand and specify polyproteins that undergo proteolytic processing (13, 42). The subsequent development of a full-length infectious cDNA clone of CHV1-EP713 RNA unequivocally established hypoviruses as the causal agent of hypovirulence (12).

The full-length CHV1-EP713 cDNA clone has been used to initiate infection in virus-free C. parasitica isolates by either of two protocols, transformation or transfection. Transformation involves integration of the CHV1-EP713 cDNA into the fungal chromosome (12), transcription of cDNA-derived viral coding strand RNA, and initiation of cytoplasmic hypovirus replication. The presence of a nuclear copy of CHV1-EP713 cDNA provides the capacity to transmit virus to ascospore progeny via nuclear inheritance (7). This novel mode of transmission is predicted to circumvent transmission barriers posed by the vegetative incompatibility system, thereby enhancing biocontrol potential. Recent environmental release studies have confirmed hypovirus transmission to ascospore progeny by transgenic hypovirulent C. parasitica strains under actual field conditions (2). Transfection is based on electroporation of a full-length synthetic CHV1-EP713 coding strand transcript into spheroplasts of virus-free C. parasitica strains (8). This versatile method has been used to effectively extend hypovirus infection to fungal species related to C. parasitica (6, 8) and to begin identifying virus-encoded symptom determinants (6, 15).

We now report the cloning, sequence analysis, and construction of a full-length cDNA clone for hypovirus CHV1-Euro7. C. parasitica strains infected with this hypovirus, while hypovirulent, are more aggressive in colonizing chestnut stem tissue and have higher levels of asexual sporulation than strains infected with the prototypic hypovirus CHV1-EP713. The availability of this second infectious hypovirus cDNA clone provides powerful new comparative approaches for elucidating the mechanisms underlying hypovirus-mediated attenuation of fungal virulence and has implications for continued engineering of hypoviruses for enhanced biological control potential.

MATERIALS AND METHODS

Fungal strains, growth conditions, and phenotypic measurements.

C. parasitica EP155 (ATCC 38755), a virulent hypovirus-free strain, was provided by S. Anagnostakis (Connecticut Agricultural Experiment Station), who obtained the original isolate in 1977 from a canker on Castanea dentata (Marshall) Borkh. in a field plot in Connecticut. Strain EP713 (ATCC 52571) was generated by anastamosis-mediated transfer of hypovirus RNA from the French hypovirulent strain EP113 to strain EP155. Strains EP155 and EP713 are therefore considered isogenic. Strain EP713 was the source of prototypic hypovirus CHV1-EP713 RNA used to construct the corresponding full-length infectious cDNA clone (12). Hypovirulent C. parasitica Euro7 (ATCC 66021) was isolated in 1978 by William MacDonald (West Virginia University) from a superficial canker on a European chestnut coppice sprout in a forested area approximately 30 km north of Florence, Italy. This strain was the source of hypovirus CHV1-Euro7 RNA. Designations used in hypovirus nomenclature include CHV for Cryphonectria hypovirus, a number indicating species relatedness and, following a hyphen, the fungal host from which the virus was isolated (24). Strain Euro7(−v), also supplied by William MacDonald, is a virus-free single conidial isolate derived from Euro7. Stock cultures were maintained on potato dextrose agar (PDA; Difco, Detroit, Mich.) as previously described (26).

Measurements of radial growth and sporulation levels on synthetic media were performed as described by Hillman et al. (26). To ensure consistency for phenotypic measurements, parallel inoculation cultures were initiated by transfer of mycelial plugs directly from transfection regeneration plates to PDA plates. Uniform mycelial plugs were subsequently transferred to replicate test PDA plates for analysis. This protocol was instituted because some fungal strain-virus combinations [primarily Euro7(−v)–CHV1-EP713] were observed to undergo a change in morphology involving reduced growth rate and production of irregular colony margins upon continued passage on PDA. Virulence assays were performed with dormant American chestnut tree stems as previously described (12, 29), with a minimum of six duplicate inoculations per fungal strain. Inoculated stems were kept at room temperature in a plastic bag to maintain moisture.

Isolation of hypovirus dsRNAs.

For routine analyses, dsRNA was extracted from hypovirus-infected fungal cultures grown in liquid EP complete medium (38) for 4 to 5 days at 25°C by the protocol of Hillman et al. (26) through the RQ1 DNase (Promega) digestion step. The quantity and quality of the preparations were examined by agarose (0.8%) gel electrophoresis (26). For purposes of cDNA library construction, contaminating single-stranded RNA and tRNA were minimized by further digestion of the dsRNA preparations with S1 nuclease (400 to 600 μg of partially purified dsRNA digested with 300 U of S1 nuclease [United States Biochemicals] at 37°C for 2 h). The reaction mixtures were then subjected to phenol-chloroform extraction, and the intact dsRNA was recovered following ethanol precipitation and passage through a Spin Column-1000 (Sigma).

Generation of a CHV1-Euro7 cDNA library.

The general protocol used originally by Shapira et al. (41) to generate a cDNA library for prototypic hypovirus CHV1-EP713 dsRNA was used to prepare cDNA from purified CHV1-Euro7 dsRNA. Based on the prediction that CHV1-Euro7 RNA may have nucleotide sequence similarity with CHV1-EP713 RNA, first-strand CHV1-Euro7 cDNA synthesized with oligo(dT) as a primer was used as a template for PCR amplification of the terminal domains with primer pairs specific for the 5′ terminus (primers RSDS10 and BR18 [CHV1-EP713 map positions 1 to 22 and 350 to 369, respectively] [40]) and the 3′ terminus (primers BH23 and RSDS11 [CHV1-EP713 map positions 12131 to 12148 and 12677 to 12697, respectively] [40]). The precise terminal sequences were confirmed with the Rapid Amplification of cDNA Ends (RACE) technique performed on purified CHV1-Euro7 dsRNA as described by Chen et al. (9) for CHV1-EP713 dsRNA. CHV1-Euro7-specific primers used for 5′ RACE of the coding and noncoding strands were oligo-490 (CHV1-Euro7 map positions 490 to 471) and Euro-73 (CHV1-Euro7 map positions 12300 to 12319), respectively.

A cDNA library was synthesized with CHV1-Euro7 primers Euro-71 and Euro-64, corresponding to final map positions 12232 to 12251 and 1231 to 1250 of the noncoding and coding strands, respectively, and a TimeSaver cDNA Synthesis Kit (Pharmacia). The resulting double-stranded cDNA was ligated into plasmid pUC19, followed by transformation of competent Escherichia coli XL1-Blue MRF′ cells (Stratagene). The library was screened for larger cloned cDNA inserts, which were subsequently sequenced and analyzed with Genetics Computer Group alignment programs with reference to the published CHV1-EP713 nucleotide sequence (41). Two gaps in the map (coordinates 5501 to 6310 and 8518 to 9246) were filled by PCR amplification of the region with total cDNA as a template. Multiple independent cDNA clones covering the entire CHV1-Euro7 RNA were sequenced to ensure accuracy.

Construction of an infectious full-length CHV1-Euro7 cDNA.

The general protocol previously used for the construction of a full-length infectious cDNA clone of CHV1-EP713 dsRNA (12, 18) was adapted for the construction of a CHV1-Euro7 infectious cDNA clone. Early in the construction process, cDNA clones of the terminal domains were modified through the use of PCR to incorporate a unique NotI site followed by a T7 polymerase promoter fused to the 5′ terminus of the CHV1-Euro7 coding strand and to add a unique SpeI site following the CHV1-Euro7 3′-terminal poly(A). Several large intermediate clones were generated from overlapping partial cDNA clones by use of common endonuclease restriction sites contained within neighboring clones. The full-length cDNA was obtained by ligating two terminally modified large cDNA clones that spanned CHV1-Euro7 map positions 1 to 5389 and 5220 to the 3′ terminus at a common NarI site (map position 5310) and cloning the ligated clones into plasmid vector pCRScript SK(+) (Stratagene) to form plasmid pTE7. Transcripts corresponding to the CHV1-Euro7 coding strand were synthesized from SpeI-digested pTE7 in a T7 polymerase reaction and used to transfect C. parasitica spheroplasts as described by Chen et al. (6, 8).

Two chimeric CHV1-EP713–CHV1-Euro7 infectious cDNA constructs, A713BE7 and AE7B713, were prepared by precise swapping of the major hypovirus coding domains ORF A and ORF B. PCR and standard cloning procedures were used to construct chimeras from a combination of the full-length CHV1-EP713 cDNA clone pLDST (12, 41), the full-length CHV1-Euro7 cDNA clone pTE7 (this study), and several intermediate CHV1-Euro7 cDNA-containing plasmids generated during the construction of pTE7. The integrity of each chimera was completely verified by sequence analysis of the junctions of interchanged domains. A detailed description of the cloning steps is available from D.L.N. upon request.

Nucleotide sequence accession number.

The GenBank accession number for the nucleotide sequence of the full-length cDNA copy of CHV1-Euro7 genomic RNA is AF082191.

RESULTS

Organizational similarities and regions of sequence conservation observed for the two hypoviruses for which full-length sequence information had previously been published, CHV1-EP713 (41) and CHV2-NB58 (25), suggested the possibility that CHV1-EP713-specific primers corresponding to the terminal portions of the genome might generate reverse transcription-PCR amplicons from purified CHV1-Euro7 dsRNA. This possibility was confirmed by sequence analysis of reverse transcription-PCR amplicons generated with primer pairs that were specific for the 5′ terminus (primers RSDS10 and BR18 [map positions 1 to 22 and 350 to 369, respectively] [40]) and for the 3′ terminus (primers BH23 and RSDS11 [map positions 12131 to 12148 and 12677 to 12697, respectively] [40]) of the CHV1-EP713 coding strand. The terminal CHV1-Euro7 amplicons showed on the order of 93% identity with the published CHV1-EP713 terminal nucleotide sequences. This new sequence information was used to design CHV1-Euro7-specific terminal primers with which to generate a CHV1-Euro7 cDNA library. The high level of identity between the CHV1-EP713 and CHV1-Euro7 nucleotide sequences allowed ordering of randomly sequenced CHV1-Euro7 cDNA clones by comparison to the published CHV1-EP713 sequence (41). Several short gaps in the sequence not covered by cDNA clones were filled by PCR amplification of total cDNA, while the sequences of the terminal ends were confirmed by sequencing of amplicons generated by 5′ RACE performed on purified CHV1-Euro7 dsRNA (9). A comparison of the derived CHV1-Euro7 genome sequence to the genome sequences of CHV1-EP713 and CHV2-NB58 is presented in Fig. 1.

FIG. 1.

Comparison of the CHV1-Euro7 cDNA sequence information determined in this study with that of two previously reported full-length hypovirus cDNA sequences: CHV1-EP713 (41) and CHV2-NB58 (25). (A) Similarities at the nucleotide levels. Previously identified protein coding regions are noted within the open boxes representing the viral genome (34). The lengths in nucleotides for the 5′ and 3′ noncoding (nc) regions and ORFs A and B for CHV1-EP713 are indicated at the top. The numbers of nucleotides for comparable regions of the other hypoviruses were 494, 1,869, 9,494, and 844, respectively, for CHV1-Euro7 and 487, 1,314, 9,873, and 831, respectively, for CHV2-NB58. The percent nucleotide identity for different coding and noncoding regions is indicated between the different viral genome diagrams being compared. (B) Similar information at the deduced amino acid levels. Note that CHV2-NB58 lacks a p29 homolog and contains p50 and p52 as the homologs of p40 and p48 found in CHV1-Euro7 and CHV1-EP713.

A field strain containing hypovirus CHV1-EP713 was originally isolated in 1966 in southern France, while the field strain containing hypovirus CHV1-Euro7, strain Euro7, was isolated later (1978) in northern Italy (see Materials and Methods). The original C. parasitica strain harboring hypovirus CHV2-NB58 was isolated quite recently in North America (27). However, this hypovirus is speculated to be of distant European origin (25, 27). Consistent with the geographical and chronological histories of these virus isolates, CHV1-Euro7 is much more closely related at the nucleotide and amino acid levels to CHV1-EP713 than to CHV2-NB58. Sequence identity is particularly high at the terminus corresponding to the 5′ end of the coding strand, where only four differences occur within the first 100 nucleotides. The level of nucleotide identity for the entire 5′ noncoding domain for these two hypoviruses is 93%, compared to 66% identity over the same region for CHV1-Euro7 and CHV2-NB58. A similar relative level of nucleotide identity was observed for the 3′ noncoding domain (Fig. 1). However, the CHV1-Euro7 genome is 11 nucleotides shorter than the CHV1-EP713 genome (12,701 versus 12,712). Differences relative to the CHV1-EP713 sequence include two single nucleotide deletions and one nucleotide insertion within the 5′ noncoding region, the deletion within ORF B of one codon that corresponds to CHV1-EP713 leucine residue 1400 (nucleotides 6561 to 6563), and seven nucleotide deletions within the 3′ noncoding region upstream from the poly(A) tail. Four of the differences within the 3′ noncoding region occur adjacent to the poly(A) tail: 5′-GAACAACAAAG-poly(A) for CHV1-EP713 versus 5′-GAACAAC-POLY(A) for CHV1-Euro7. Thus, this four-base difference could result from a simple G-to-A transition at CHV1-EP713 map position 12712.

The similarity between CHV1-Euro7 and CHV1-EP713 is even more striking at the amino acid sequence level, ranging from a low of 90% identity for the p40 portion of ORF A to a high of 98% identity in the region between the putative polymerase and helicase domains. Prominent features and their sequence contexts previously identified for CHV1-EP713 (reviewed in reference 34) are conserved in CHV1-Euro7. These include amino acid residues C-162 and H-215, essential for p29 cleavage; the UAAUG pentanucleotide at the junction between ORF A and ORF B; amino acid residues C-341 and H-388, essential for p48 cleavage; and the p48 cleavage site (G-418/A-419). The only notable difference is that the presumptive CHV1-Euro7 p29 cleavage site has the sequence RIG/NQL rather than the sequence RIG/GRL demonstrated for CHV1-EP713 (13). These high levels of identity contrast with the relatively low levels of sequence conservation between CHV1-Euro7 and CHV2-NB58, particularly in ORF A, where CHV2-NB58 lacks a defined p29 domain.

Development of an infectious cDNA clone for CHV1-Euro7.

C. parasitica Euro7, the source of hypovirus CHV1-Euro7, exhibits phenotypic traits that differ significantly from those of strain EP713, the source of hypovirus CHV1-EP713. For example, Euro7 grows more rapidly on solid synthetic medium than corresponding virus-free isogenic strains, while EP713 grows more slowly. EP713 forms small, superficial cankers on chestnut tissue (i.e., it is considered highly hypovirulent) and produces little or no asexual spores either on synthetic medium or on chestnut tissue. In contrast, Euro7 is only moderately hypovirulent (30a). Aggressive canker expansion early after inoculation eventually slows or ceases, concomitant with heavy callus formation at the canker margins. Additionally, Euro7 does produce asexual spores, although at a reduced level relative to corresponding virus-free isogenic strains, on synthetic medium and especially on chestnut tissue. These properties—more effective colonization of bark tissue and higher sporulation levels—are predicted to positively contribute to biological control potential by enhancing persistence and dissemination (31). Both Euro7 and EP713 are deficient in the production of orange pigmentation, a useful laboratory marker of hypovirus infection.

The significant differences in phenotypic traits exhibited by strains Euro7 and EP713, coupled with a high level of sequence identity for hypoviruses CHV1-Euro7 and CHV1-EP713, suggested that the development of a full-length infectious cDNA clone for CHV1-Euro7 would provide unique opportunities to examine the relative contributions of hypovirus and fungal host genomes to hypovirulence and associated traits and would allow the mapping of hypovirulence determinants by the construction of viral chimeras.

With the transfection protocol developed by Chen et al. (8), synthetic transcripts generated from a full-length CHV1-Euro7 cDNA clone (see Methods and Materials) were shown to be infectious, yielding infected colonies with traits typical for the Euro7 field strain (data not shown). To examine the relative phenotypic contributions of viral and fungal host genomes, two virus-free virulent strains, Euro7(−v) and EP155, were transfected independently with CHV1-Euro7 and CHV1-EP713 synthetic transcripts (Fig. 2). Both fungal strains transfected with the CHV1-Euro7 synthetic transcripts resembled the Euro7 field strain in terms of hyphal growth rate and colony morphology, while the two strains transfected with the CHV1-EP713 synthetic transcripts clearly resembled strain EP713. Additionally, the double-stranded form of the corresponding hypovirus RNA accumulated in each of the transfectants (Fig. 3). As has been reported for hypovirulent strain EP713 (40) and passaged transgenic hypovirulent strains transformed with CHV1-EP713 cDNA (7), some of the transfectants examined in this study contained dsRNA species that had internal deletions and that migrated faster than full-length viral dsRNA (Fig. 3, lanes 4 and 7). However, no phenotypic changes have been found to be associated with the appearance of these deletion dsRNA species (7). Transfectants were further analyzed in detail for growth characteristics, canker formation (virulence), and sporulation properties on both synthetic medium and chestnut tissue (Tables 1 and 2).

FIG. 2.

Colony morphology for virus-free C. parasitica EP55 and Euro7(−v) and related hypovirus transfectants. (Top row) Colonies of virus-free strain EP155 (center) and strain EP155 transfected with CHV1-EP713, CHV1-Euro7, chimeric virus AE7B713, or chimeric virus A713BE7. (Bottom row) Colonies of virus-free strain Euro7(−v) (center) and strain Euro7(−v) transfected with CHV1-EP713, CHV1-Euro7 chimeric virus AE7B713, and chimeric virus A713BE7. Photographs were taken on day 7.

FIG. 3.

Agarose gel electrophoretic analysis of dsRNAs recovered from transfected C. parasitica strains. The migration position of the full-length hypovirus dsRNA is indicated by the arrow on the right. Lane M contains 200 ng of a 1-kb DNA ladder (Gibco BRL) as relative size markers. dsRNA preparations recovered from equal volumes of cultured virus-free and transfected strains were loaded in the following order: lane 1, EP155; lane 2, EP155–CHV1-EP713; lane 3, EP155–CHV1-Euro7; lane 4, EP155–AE7B713; lane 5, EP155–A713BE7; lane 6, Euro7(−v); lane 7, Euro7(−v)–CHV1-EP713; lane 8, Euro7(−v)–CHV1-Euro7; lane 9, Euro7(−v)–AE7B713; and lane 10, Euro7(−v)–A713BE7. The faster-migrating species observed in lanes 4 and 7 correspond to internally deleted defective viral RNAs previously identified in hypovirus-infected strains (7, 40). The presence of these deletion dsRNAs has not been associated with any change in phenotypic traits.

TABLE 1.

Effect of transfection with wild-type and chimeric hypovirus transcripts on colony size and asexual sporulation

Strain	Colony size (cm2) on day:		No. of conidia/mla	Fold reduction
	5	7
EP155	23.2 ± 0.4	42.3 ± 1.3	6.9 × 108 ± 5.0 × 107
Euro7(−v)	22.4 ± 0.4	46.3 ± 0.4	8.7 × 108 ± 5.3 × 107
EP155–CHV1-EP713	13.4 ± 0.4	20.8 ± 1.4	2.8 × 104 ± 3.7 × 103	24,643
Euro7(−v)–CHV1-EP713	15.2 ± 0.6	27.3 ± 0.4	9.0 × 103 ± 4.8 × 103	86,667
EP155–AE7B713	11.0 ± 0.3	14.1 ± 0.9	2.2 × 103 ± 3.1 × 103	300,000
Euro7(−v)–AE7B713	15.9 ± 0.6	30.4 ± 1.8	5.2 × 104 ± 9.2 × 103	16,730
EP155–CHV1-Euro7	24.0 ± 0.3	42.7 ± 0.5	3.5 × 105 ± 6.1 × 104	1,971
Euro7(−v)–CHV1-Euro7	33.3 ± 0.8	55.4 ± 0.0	2.7 × 107 ± 2.2 × 106	40
EP155–A713BE7	19.7 ± 0.6	35.3 ± 1.1	1.9 × 105 ± 3.2 × 104	3,632
Euro7(−v)–A713BE7	29.3 ± 0.4	52.8 ± 0.4	8.6 × 106 ± 2.3 × 106	101

^aFungal cultures were grown in parallel on PDA in 85-mm-diameter petri dishes under standard laboratory bench conditions (6) for 22 days. Conidia were collected in 16 ml of 0.15% Tween 80 per culture from five replicate plates for each strain and quantified as described by Craven et al. (15). The values shown represent 1/16 the mean number of conidia per dish per strain. 

TABLE 2.

Effect of transfection with wild-type and chimeric hypovirus transcripts on canker expansion and production of asexual spores on canker tissue

Strain	Canker size (cm2) on day:			Sporulationa
	10	21	31	Stromata/canker	Conidia/canker
EP155	8.6 ± 2.2	17.0 ± 6.7	40 ± 15.7	246.7 ± 115.3	7.4 × 108 ± 8.7 × 108
Euro7(−v)	5.7 ± 2.2	8.3 ± 3.9	21.0 ± 9.3	212.0 ± 71.7	1.1 × 109 ± 8.8 × 108
EP155–CHV1-EP713	2.2 ± 0.8	2.6 ± 1.2	3.2 ± 1.2	0.8 ± 1.1	0
Euro7(−v)–CHV1-EP713	2.2 ± 0.4	2.3 ± 0.5	2.4 ± 0.6	0.2 ± 0.3	0
EP155–AE7B713	2.5 ± 0.7	2.9 ± 0.8	3.3 ± 0.9	26.5 ± 12.8	1.4 × 105 ± 6.0 × 104
Euro7(−v)–AE7B713	2.9 ± 0.4	3.5 ± 0.5	3.7 ± 0.5	33.3 ± 10.0	3.4 × 105 ± 3.5 × 105
EP155–CHV1-Euro7	5.0 ± 1.4	6.4 ± 2.8	12.1 ± 6.4	101.2 ± 23.2	3.1 × 106 ± 2.3 × 106
Euro7(−v)–CHV1-Euro7	4.8 ± 1.0	6.5 ± 1.7	18.2 ± 5.1	172.2 ± 44.5	2.6 × 109 ± 1.1 × 109
EP155–A713BE7	4.5 ± 1.5	8.7 ± 2.6	25.4 ± 8.4	106.0 ± 38.7	7.3 × 106 ± 3.8 × 106
Euro7(−v)–A713BE7	4.5 ± 2.1	6.6 ± 2.2	17.1 ± 6.3	135.8 ± 58.8	1.7 × 109 ± 1.2 × 109

^aThe number of stromata (stromal pustules) per canker was determined at day 30. A subset of 10 stromata was removed from the canker. These stromata were placed between two clean microscope slides in 100 μl of 0.15% Tween 80 and gently crushed. The crushed stroma material was then diluted into a total of 4 ml of 0.15% Tween 80, and the number of conidia was determined with the aid of a hemacytometer (7). 

Relative phenotypic contributions of hypovirus and C. parasitica genomes.

It is clear from an inspection of Table 1 and Fig. 2 that the hypovirus genome contributed significantly to the hyphal growth rates of the transfectants. As has been observed for the Euro7 field strain, CHV1-Euro7 transfectants grew as fast as or faster than the corresponding virus-free strains. Similarly, CHV1-EP713 transfectants, like strain EP713, grew more slowly than the corresponding virus-free strains. However, a subtle contribution of the host genome was also observed. Although strain EP155 grew at a rate similar to that of virus-free strain Euro7(−v), its growth rate was reduced to a greater extent by transfection with a specific hypovirus synthetic transcript; i.e., strain EP155 transfected with either CHV1-EP713 or CHV1-Euro7 grew more slowly than the corresponding Euro7(−v) transfectants (Table 1).

A similar pattern was evident for sporulation profiles (Table 1). That is, transfection of EP155 and Euro7(−v) with CHV1-EP713 RNA resulted in a greater reduction of sporulation than did transfection with CHV1-Euro7 RNA, with relative fold reductions of 24,643 versus 1,971 in EP155 and 86,667 versus 40 in Euro7(−v). A host genome contribution was again seen for CHV1-Euro7 transfectants; e.g., transfection of EP155 with CHV1-Euro7 RNA resulted in a 1,971-fold reduction in sporulation, while sporulation by Euro7(−v) transfected with CHV1-Euro7 RNA was only 40-fold lower than that of Euro7(−v).

More dramatic differences in the effect of the two hypoviruses on host phenotype were observed on dormant chestnut stem tissue (Table 2). Although the growth rates for strains EP155 and Euro7(−v) on solid synthetic medium were similar, strain EP155 was consistently more aggressive in canker production following inoculation of dormant chestnut stem tissue (Table 2 and Fig. 4), producing cankers nearly twice the size of those produced by Euro7(−v). The cankers incited by both virus-free strains produced high levels of stromal pustules (pycnidium-containing stromata that erupt through the bark) and viable conidia by 30 days postinoculation. Transfection with CHV1-EP713 severely reduced the ability of both strains to expand on chestnut tissue, resulting in the production of small, superficial cankers that produced very few stromal pustules, as previously described for hypovirulent strain EP713 (12). CHV1-Euro7 transfectants, in contrast, were much more aggressive in canker formation irrespective of the fungal host background and produced cankers with morphologies very similar to those described for the Euro7 field strain (30a). These cankers had distinctive ridged margins, suggesting the formation of callus tissue. Unlike the cankers produced by the CHV1-EP713 transfectants, these cankers produced prodigious levels of stromal pustules containing viable asexual spores. Thus, canker morphology, canker expansion, and asexual sporulation levels on chestnut stem tissue appear to be controlled to a much greater extent by the hypovirus genome than by the genome of the fungal host.

FIG. 4.

Gallery of representative cankers formed by virus-free and transfected C. parasitica strains. (Top row) Typical cankers formed by virus-free strain EP155 (center) and strain EP155 transfected with CHV1-EP713, CHV1-Euro7, chimeric virus AE7B713, and chimeric virus A713BE7. (Bottom row) Cankers formed by virus-free strain Euro7(−v) and the corresponding set of Euro7(−v) transfectants (as detailed for top row). Cankers were photographed 30 days postinoculation after wetting with ethanol to enhance color contrast of cankers and the surrounding area. Stromal protrusions (stromata that contain asexual spore-forming bodies termed pycnidia) are prominent features of the surface of cankers caused by virus-free strains EP155 and Euro7(−v) as well as the CHV1-Euro7 and A713BE7 transfectants. Spiral structures, termed ceri, composed of conidia are seen extruded from some stromata. These structures are rarely observed on the surface of cankers formed by CHV1-EP713 or AE7B713 transfectants.

Use of infectious chimeric hypovirus cDNA transcripts to map differences in hypovirus-mediated alterations of fungal host phenotype.

The high level of sequence identity shared by CHV1-Euro7 and CHV1-EP713 suggested the possibility that viral coding domains responsible for differences in host phenotypic changes caused by these two hypoviruses could be mapped through the construction of chimeras of infectious viral cDNAs. The feasibility of this approach was tested to a first approximation by interchanging the two viral polyprotein coding domains, ORF A and ORF B (see Methods and Materials). Transfection with the synthetic transcripts derived from the chimeras, designated AE7B713 and A713BE7, resulted in productive infections, as judged by the accumulation of hypovirus dsRNA (Fig. 3). In the EP155 genetic background, transfection with the AE7B713 chimera resulted in a colony morphology similar to that of CHV1-EP713 transfectants, while transfection with the A713BE7 chimera resulted in a colony morphology similar to that of CHV1-Euro7 transfectants (Fig. 2). Comparable results were observed for these same chimeras in the Euro7(−v) genetic background (Fig. 2), indicating that the ORF B portion of the chimera generally determined colony morphology.

Differences in growth rates and sporulation levels on synthetic medium conferred by transfection with CHV1-EP713 and CHV1-Euro7 RNAs also mapped to ORF B (Table 1). Although AE7B713 transfectants clearly grew much more slowly than A713BE7 transfectants in both fungal strain backgrounds, very minor differences relative to the corresponding wild-type virus transfectants were observed. For example, the EP155-AE7B713 and EP155-A713BE7 transfectants grew slightly more slowly than the EP155–CHV1-EP713 and EP155-CHV1-Euro7 transfectants, respectively. Additionally, the Euro7(−v)–AE7B713 transfectants consistently grew slightly faster than the Euro7(−v)–CHV1-EP713 transfectants. As was observed for CHV1-EP713, the AE7B713 chimera also reduced asexual sporulation to a much greater extent than CHV1-Euro7 or the A713BE7 chimera in both fungal strains (Table 1).

As noted above, differences in the phenotypic consequences of transfection with the two infectious hypovirus cDNA transcripts were most pronounced when the resulting transfectants were inoculated onto chestnut stem tissue (Fig. 4). Similarly, the contribution of ORF B to these differences was most clearly observed within this context (Fig. 4 and Table 2). Transfectants containing the chimeric viruses that have CHV1-Euro7 ORF B produced cankers strikingly similar in appearance to those caused by the corresponding wild-type CHV1-Euro7 transfectants. These similarities extended to the formation of raised, apparently callus-forming canker margins and the production of stromal pustules. As indicated in Table 2, EP155-A713BE7 transfectants produced cankers even larger than those formed by EP155–CHV1-Euro7 transfectants (mean ± standard deviation, 25.4 ± 8.4 versus 12.1 ± 6.4 cm² by day 31), while cankers caused by the Euro7(−v)–A713BE7 and Euro7(−v)–CHV1-Euro7 transfectants were of essentially the same size (18.2 ± 5.1 versus 17.1 ± 6.3 cm² by day 31). Like the CHV1-Euro7 transfectants, the A713BE7 transfectants of both fungal species produced significant levels of stromal pustules and conidia within the canker area (Fig. 4 and Table 2). Consistent with the results observed for the A713BE7 transfectants, EP155–AE7B713 and Euro7(−v)–AE7B713 transfectants produced cankers of a size and a morphology very similar to those caused by the corresponding CHV1-EP713 transfectants (all on the order of 3 cm² by day 31). Again, very minor differences relative to the wild-type virus transfectants were observed. Cankers incited by CHV1-EP713 transfectants generally failed to yield conidia, while cankers produced by AE7B713 transfectants did yield a very low level of stromal pustules and recoverable viable conidia. Nevertheless, it is clear that the contribution of ORF B to differences in host phenotypic changes caused by CHV1-EP713 and CHV1-Euro7 extended to canker morphology, size, and spore production.

DISCUSSION

Numerous surveys of European and North American C. parasitica field isolates have revealed considerable variability in virulence and morphological traits (16, 18, 20, 31). Contributors to diversity include hypovirulence and associated symptoms linked to mitochondrial dysfunction (32, 33) or infection by a variety of virus-like dsRNAs, including those of hypoviruses (14, 20, 36, 37). Recent detailed analyses have revealed considerable differences in the spectrum and severity of hypovirulence-associated symptoms even for C. parasitica strains infected with hypoviruses that are related at the nucleotide level (25, 27; this study). The construction and manipulation of an infectious CHV1-Euro7 cDNA, as described in this report, illustrate how this diversity in phenotypic traits and conservation of nucleotide sequence can be exploited to examine issues such as the relative contributions of viral and host genomes to hypovirulence-associated symptom expression and to map viral hypovirulence determinants.

The question of the relative contributions of hypovirus and C. parasitica host genomes to hypovirulence-associated symptom expression has been difficult to approach. The introduction of different hypoviruses into specific virus-free C. parasitica strains by anastomosis complicated interpretation due to the potential transmission of organelles or nuclear genetic information along with hypovirus RNA. The availability of infectious cDNA clones of related hypoviruses derived from infected C. parasitica strains that exhibited very distinct phenotypes provided the opportunity for a more rigorous examination of this issue. The observation that the transfectants obtained resembled the hypovirulent strain from which the infectious cDNA clone had been derived [e.g., strains EP155 and Euro7(−v) transfected with CHV1-EP713 transcripts both resembled hypovirulent strain EP713 (Fig. 2 and 4 and Tables 1 and 2)] clearly indicates that the viral genome is the primary contributor to the morphological differences observed between hypovirulent C. parasitica strains EP713 and Euro7. However, subtle contributions by the host genomes were also observed. For example, strain Euro7(−v) infected with the CHV1-Euro7 virus produced considerably more spores on chestnut stem tissue than did strain EP155 transfected with the same infectious transcripts (Table 2). The combined results present several implications for the field release of natural and transgenic hypovirulent strains.

While hypoviruses are transmitted via anastomosis to vegetatively compatible virulent C. parasitica strains, transgenic hypovirulent strains which contain a chromosomally integrated hypovirus cDNA can also transmit virus to ascospore progeny (7). For effective biological control, hypoviruses must be effectively transmitted from the field-released hypovirulent strain throughout the virulent strain population. This process, for both natural and transgenic hypovirulent strains, results in the generation of a spectrum of hypovirus-fungal host genomic combinations. The results observed in this study concerning the relative contributions of C. parasitica and hypovirus genomes to the level of hypovirulence and the spectrum of associated traits agree with conclusions reached earlier by Elliston based on an extensive series of anastamosis transmission studies (17–19). Combined, these results predict that as natural or cDNA-derived hypovirus RNA disseminates throughout the fungal population, the characteristics of the hypovirulent strains generated will primarily reflect the contributions of the viral genome. Of course, extensive testing of infectious hypovirus cDNAs in a larger number of fungal hosts, now in progress, is required to fully confirm this prediction.

The ability to produce viable chimeras from the infectious CHV1-EP713 and CHV1-Euro7 cDNAs provides a potentially powerful tool for mapping viral determinants responsible for the differences in hypovirulence-associated symptoms conferred by these two viruses. Use of the AE7B713 and A713BE7 chimeras to map these differences almost exclusively to ORF B (Fig. 2 and 4 and Table 1 and 2) firmly established the utility of this approach. This result was somewhat surprising in light of previous reports that the p29 portion of CHV1-EP713 contributes to virus-mediated reductions in orange pigmentation and asexual sporulation (15, 25). The fact that the differences in symptom expression observed for the two viruses mapped almost exclusively to ORF B indicates that the ORF A portions of the two viruses make similar contributions to the overall level of symptom expression. It is anticipated that the mapping of determinants responsible for differences in symptom expression will ultimately lead to the identification of the domains that are directly responsible for the underlying symptoms. Preliminary detailed mapping studies suggest that different portions of ORF B may influence virulence to a greater extent than they influence associated traits, such as reduced sporulation or mycelial growth. Thus, it may be possible to further uncouple hypovirulence from associated traits by appropriate swapping of hypovirus coding domains.

Several lines of evidence suggest that one of the primary ways in which hypovirus infection alters host phenotype is by an alteration of cellular G protein-linked, cyclic AMP-mediated signal transduction (10, 11, 21). It is anticipated that an extension of the chimeric mapping approach will provide the most efficacious route to the identification of hypovirus determinants responsible for an alteration of cellular signaling pathways. Such studies will be aided by the availability of promoter regions from C. parasitica genes previously identified as being transcriptionally regulated in response to hypovirus infection and/or perturbation of G protein-linked signal transduction (10).

MacDonald and Fulbright (31) have discussed the view that successful hypovirulence-mediated biological control requires a continual reservoir of hypovirulent inoculum. They further noted that hypovirulent strains that have been used in most North American biological control efforts, while highly curative, were quite debilitated in their ability to colonize and produce spores on chestnut bark, resulting in limited persistence. In this regard, limited persistence (2 years) of transgenic hypovirulent strains containing an integrated cDNA copy of CHV1-EP713 was recently reported after a single-season release (2). As noted elsewhere (30a) and as partially confirmed in this report (Fig. 4 and Table 2), CHV1-Euro7-infected C. parasitica strains differed from strains infected with CHV1-EP713 in precisely those properties that are predicted to have a direct impact on persistence: colonization of and spore production on bark tissue. By use of the infectious CHV1-Euro7 cDNA to construct transgenic hypovirulent C. parasitica strains, it will be possible to combine properties of enhanced colonization and spore production with a novel mode of virus transmission to ascospore progeny. Future potential enhancements for biological control may also be derived by the construction of transgenic hypovirulent strains with chimeras of CHV1-EP713 and CHV1-Euro7 or additional infectious hypovirus cDNAs as they become available.