Engineering super mycovirus donor strains of chestnut blight fungus by systematic disruption of multilocus vic genes

Significance

Allorecognition, termed “vegetative incompatibility” (vic), in filamentous fungi limits the efficacy of virulence-attenuating mycoviruses for biological control (hypovirulence). Because mycoviruses lead exclusive intracellular lifestyles, horizontal transmission relies on cytoplasmic exchange during hyphal fusion (anastomosis). Fusion between vic-incompatible strains results in localized cell death restricting mycovirus transmission. We report the systematic disruption of multilocus vic genes and excision of exogenic genes to generate strains of the chestnut blight fungus able to transmit hypovirulence to strains with genotypic differences at any or all of the defined vic loci. The results demonstrate the feasibility of modulating fungal allorecognition to promote transmission of virulence-attenuating mycoviruses for enhanced biocontrol potential.

Abstract

Transmission of mycoviruses that attenuate virulence (hypovirulence) of pathogenic fungi is restricted by allorecognition systems operating in their fungal hosts. We report the use of systematic molecular gene disruption and classical genetics for engineering fungal hosts with superior virus transmission capabilities. Four of five diallelic virus-restricting allorecognition [vegetative incompatibility (vic)] loci were disrupted in the chestnut blight fungus Cryphonectria parasitica using an adapted Cre-loxP recombination system that allowed excision and recycling of selectable marker genes (SMGs). SMG-free, quadruple vic mutant strains representing both allelic backgrounds of the remaining vic locus were then produced through mating. In combination, these super donor strains were able to transmit hypoviruses to strains that were heteroallelic at one or all of the virus-restricting vic loci. These results demonstrate the feasibility of modulating allorecognition to engineer pathogenic fungi for more efficient transmission of virulence-attenuating mycoviruses and enhanced biological control potential.

Mycovirus infections have been reported to reduce virulence (hypovirulence) for a wide range of plant pathogenic fungi, providing potential for biological disease control (1–6). For hypovirulence to be effective, the virulence-attenuating viruses must be efficiently transmitted from infected hypovirulent strains to uninfected virulent strains (5, 7). Mycoviruses generally have evolved exclusive intracellular lifestyles (8). With very few exceptions (9), mycovirus infections cannot be initiated by exposure of uninfected hyphae to cell extracts or secretions from infected fungal isolates. Transmission is limited to intracellular mechanisms, vertical transmission through asexual spores, and horizontal transmission through anastomosis (fusion of hyphae).

Horizontal mycovirus transmission to uninfected isolates of the same fungal species is complicated by nonself allorecognition genetic systems, termed “heterokaryon” or “vegetative incompatibility” (vic), which operate widely in filamentous fungi (10). Interactions between genetically distinct individuals of the same species result in an incompatible reaction that triggers localized programmed cell death (PCD), forming a line of demarcation, or barrage, along the zone of contact (10–12) and restricting cytoplasmic transmission of viruses and other cytoplasmic elements (1, 13–15).

A negative correlation between vic diversity and virus transmission has been reported for several fungal hosts (1, 16), but has most extensively been demonstrated for the chestnut blight fungus Cryphonectria parasitica infected with virulence-attenuating hypoviruses (7, 17–20). Genetic analyses have defined six diallelic vic genetic loci for C. parasitica (21). These loci and associated genes were recently identified at the molecular level through a comparative genomics approach (22, 23) (Table 1). Independent gene disruption analysis of 12 genes associated with these loci (22, 23) provided formal confirmation that five of the six loci contribute to restriction of virus transmission; an allelic difference at vic4 was shown previously not to restrict virus transmission (24). Moreover, all but one of these mutations resulted in no observable phenotypic change other than increased virus transmission or loss of the incompatibility reaction (22, 23). We now report the use of sequential vic gene disruption (23, 25) and mating approaches to engineer C. parasitica super hypovirus donor strains. The results demonstrate the feasibility of global modulation of fungal host allorecognition systems to remove restrictions to mycovirus transmission.

Table 1.

Summary of C. parasitica vic genetic loci and associated genes

vic locus	vic genes	Protein features
vic 1	vic 1a-1, vic1a-2	HET domain
	vic1b-2	DUF domain
	vic1c-1	LTR retrotransposon
	vic1d-1	HET domain
vic2	vic2-1, vic2-2	Patatin-like protein
	vic2a-1, vic2a-2	Sec9-like
vic3	vic3a-1, vic3a-2	Hypothetical protein
	vic3b-1, vic3b-2	Life-guard–like
vic4	vic4-1	Protein kinase c-like
	vic4-2	NACHT/WD40
vic6	vic6-1, vic6-2	HET domain
	pix6-1, pix6-2	DUF 1040 domain
vic7	vic7-1, vic7-2	HET domain
	vic7a-1, vic7a-2	Ankyrin repeats

The six defined diallelic C. parasitica vic genetic loci are designated vic1, vic2, vic3 vic4, vic6, and vic7. Each allele, designated allele 1 or allele 2, contains vic-associated genes that exhibit nucleotide sequence polymorphism or nucleotide sequence idiomorphism (no similarity). For example, allele 1 of the vic3 locus contains the vic3a-1 and the vic3b-1 genes, and allele 2 contains the related polymorphic vic3a-2 and voic3b-2 genes. In contrast, allele 1 of the vic4 locus contains the vic4-1 gene, and allele 2 contains the unrelated idiomorphic vic4-2 gene. The vic1 locus contains both polymorphic and idiomorphic vic-associated genes. Allele 1 contains the polymorphic vic1a-1 gene and the idiomorphic vic1c-1 and vic1d-1 genes, and allele 2 contains the vic1a-2 polymorphic gene and the vic1b-2 idiomorphic gene. Vic locus-specific polymorphic genes are indicated in the same row separated by a comma, and locus-specific idiomorphic genes are indicated in separate rows. The vic-associated genes shown in boldface type were disrupted during construction of the super donor strains described in this report. The vic genes are described in refs. 22 and 23. The vic7a is described in this study.

Results

Construction of C. parasitica Triple vic Mutant Strain.

The convention used to describe the vic genotype of a C. parasitica strain specifies which allele, designated allele 1 or allele 2, is present at the six genetically defined diallelic vic genetic loci (21); for example, for the reference strain EP155, the vic genotype is vic1-2, vic2-2, vic3-1, vic4-1, vic6-2, and vic7-2 (abbreviated 2211–22). Subsequent comparative genomic analyses (22, 23) identified multiple polymorphic and idomorphic genes at the different vic loci. For example, the vic3 locus contains two polymorphic genes, vic3a and vic3b, designated vic3a-1 and vic3b-1 for allele 1 and vic3a-2 and vic3b-2 for allele 2, respectively (see Table 1 for current list of vic-associated genes).

We previously demonstrated that disruption of individual polymorphic vic genes at the different vic loci resulted in both asymmetric and symmetric effects on virus transmission (22, 23). For example, disruption of gene vic6-2 within the vic6 locus resulted in unrestricted virus transmission when paired with a vic6 heteroallelic strain only if the mutant strain was the recipient whereas disruption of the adjacent gene pix6-2 resulted in unrestricted virus transmission only when the mutant strain served as the donor. Disruption of vic3b-1 also promoted the ability of the mutant strain to serve as a donor whereas a vic1a-2 disruption mutant was found to be unrestricted in virus transmission when serving as either a virus donor or recipient in a pairing of vic1 heteroallelic strains. Consequently, vic1a-2, vic3b-1, and pix6-2 were targeted for the first three steps in super donor strain construction.

As indicated in the overview presented in Fig. 1, gene disruptions were performed using strain DK80, a mutant of C. parasitica genomic reference strain EP155, vic genotype 2211–22, in which the nonhomologous end-joining DNA repair pathway ku80 gene homolog was disrupted to promote homologous recombination (26). The vic1a-2 gene was disrupted using a loxP-Pro_gpd-neo-loxP–based disruption cassette as previously described (23, 25) and followed by Pro_gpd-neo excision by anastomosis with Cre-expressing strain EP155-cre (Materials and Methods). The vic3b-1 allele was disrupted in the Δvic1a-2 mutant strain by replacement with a loxP-Pro_tubβ-hph-loxP–based cassette (23, 25) to form the double vic mutant strain Δvic1a-2, Δvic3b-1. Disruption of the pix6-2 allele in the double vic mutant was completed by replacing the corresponding CDS with the loxP-Pro_gpd-neo-loxP cassette to form triple vic mutant Δvic1a-2, Δvic3b-1, Δpix6-2 (Fig. 1). A neo-selectable marker gene (SMG)–free isolate of the triple-mutant strain obtained by excision at the pix6 disruption site was used for subsequent vic gene disruptions.

Fig. 1.

Flowchart of the sequential vic gene disruption and SMG excision events involved in generation of hph/neo-SMG–free quadruple vic mutant strains. Sequential disruption of vic genes was performed in C. parasitica strain DK80 in the order vic1a-2, vic3b-1, pix6-2, and vic7a-2. Strain DK80 (26) is a mutant of reference strain EP155 in which the nonhomologous end-joining DNA repair pathway ku80 gene homolog was disrupted with a cassette containing a benomyl (ben) SMG (indicated at the top of the chart) to promote homologous recombination. The ben-SMG was not flanked with loxP elements. Cre-mediated loxP-flanked SMG excision via anastomosis events are indicated by “EP155-cre” placed at the right of appropriate event arrows. The final quadruple vic mutant strain, indicated at the bottom of the chart, was subsequently crossed with strain EP146 to generate the super donor strains as illustrated in Fig. 2.

Vic7 Disruption Analysis and Construction of Quadruple vic Mutant Strain.

Previous disruption analysis of the vic7 locus focused on a polymorphic ORF encoding a HET domain protein, designated vic7 (22). Disruption of the vic7-2 allele in strain DK80 did not eliminate barrage formation and only partially reduced resistance to virus transmission when the mutant served as a donor strain in a pairing of vic7 heteroallelic strains (22).

A small predicted ORF (164 codons) adjacent to the vic7 ORF containing an ankyrin motif (Joint Genome Institute predicted putative protein ID 234408; genome.jgi-psf.org/Crypa2/Crypa.home.html) was initially not considered as a candidate vic gene due to a low polymorphism (96% identity) at the amino acid level (Fig. S1). A functional role of the ORF, now designated vic7a, in vegetative incompatibility and restriction of virus transmission was examined by disruption of the vic7a-2 allele in strain DK80. The Δvic7a-2 mutant retained the ability to form a barrage when paired with a vic7 heteroallelic tester strain, genotyped European (EU) vic tester strain EU18 (2211-21) (Materials and Methods), but exhibited an increase in virus transmission from 25 to 30 to 100% when serving as the donor strain (Table 2). Thus, unlike the HET-domain vic7 gene, the ankyrin-motif gene vic7a provided an excellent disruption target within the vic7 locus for inclusion in the super donor strain.

Fig. S1.

Amino acid alignment of vic7a alleles 1 and 2 performed using the CLUSTLW Multiple Sequence Alignment program with default settings (Biology Workbench at workbench.sdsc.edu). Amino acid identity is indicated by an asterisk (*), conservative differences are indicated by a colon (:), and nonconservative differences are indicated by a space. The two alleles show 96% amino acid identity. The line above the aligned sequences indicates the Ankyrin repeat conserved domain.

Table 2.

Hypovirus transmission frequency increase for disruption mutants of the vic7a-2 allele

Donor strain	Recipient strain	Frequency
DK80 (2211-22)	EU18 (2211-21)	6/20
Δvic7aSP1* (2211-22)	EU18 (2211-21)	20/20
Δvic7aSP2 (2211-22)	EU18 (2211-21)	20/20
Δvic7aSP3 (2211-22)	EU18 (2211-21)	20/20

This table reports the frequency of hypovirus transmission from infected donor strains DK80 and Δvic7a-2 disruption strains to recipient EU tester strain EU18, which is heteroallelic with strain DK80 only at the vic 7 locus, when the donor and recipient strains are paired on PDA plates. The vic genotypes are indicated in parentheses with a strikethrough indicating a gene disruption within the vic7-2 allele. The transmission frequency is reported as the number of successful transmissions over the number of pairings.

^*SP indicates isolate derived from a single spore.

The vic7a-2 allele was replaced by the loxP-Pro_gpd-neo-loxP cassette in the triple vic mutant Δvic1a-2, Δvic3b-1, Δpix6-2 to give the quadruple vic mutant strain Δvic1a-2, Δvic3b-1, Δpix6-2, Δvic7a-2 (2211-22) (throughout this article the vic genotypes are indicated in parentheses with a strikethrough indicating a gene disruption within the allele). A hph/neo-SMG–free isolate of the quadruple mutant was successfully obtained by excision of Pro_tubβ-hph and Pro_gpd-neo at disrupted vic7a-2 and vic3b-1, respectively (Fig. 1). Two independent quadruple-mutant single-spored isolates (2211-22) were tested for ability to transmit hypovirus CHV-1/EP713 and found to transmit virus unrestricted to EU tester strains EU37 (1221-11) and EU45 (1222-11) that are heteroallelic relative to wild-type strain EP155 (2211-22) at vic1, vic3, vic6, and vic7. No virus transmission was observed from the negative control CHV-1/EP713–infected strain EP155 to these tester strains.

Further Characterization of the vic2 Locus and Development of a Mating Strategy to Complete Construction of the Super Donor Strains.

The vic2 locus contains two highly polymorphic ORFs, designated vic2 and vic2a. The vic2 ORF encodes a member of the patatin-like phospholipase family, and vic2a encodes a protein related to a fungal plasma membrane SNARE Sec9 protein (22). These two polymorphic ORFs are separated by a highly conserved ORF encoding a helicase-like protein. Barrage formation with a paired vic2 heteroallelic strain still occurred after disruption of the patatin-like protein ORF, vic2-2, whereas virus transmission became 100%, but only if the mutant strain was the recipient, with no increase if the mutant was the donor (22). Repeated attempts to disrupt the Sec9 vic2a allele failed to give any single-spored disruption mutants, suggesting that this gene may be essential. Disruption of the helicase-like protein gene separating vic2-2 and vic2a-2 was also unsuccessful, and disruption of an ORF encoding a GTPase-like protein adjacent to the vic2a ORF failed to give any change in barrage formation or virus transmission.

The absence of an obvious donor-promoting gene disruption target at the vic2 locus necessitated an alternative strategy to vic gene disruption. An additional consideration in developing this strategy was the presence of the benomyl resistance SMG associated with the ku80 gene disruption in the DK80 strain introduced to promote homologous recombination (26). This SMG lacked the flanking loxP sites and could not be excised by Cre-mediated recombination. The adopted strategy involved mating of the hph/neo-SMG–free quadruple vic mutant strain (2211-22, Δku80^ben, MAT-2) with strain EP146 (2112-11, ku80, MAT-1) and screening for progeny that contained the four disrupted vic alleles and the wild-type ku80 gene in either the vic2-1 or the vic2-2 allelelic backgrounds that, combined, would be able to transmit hypovirus to all six multilocus vic genotypic combinations.

Because the quadruple vic mutant parent produces an orange pigment and the EP146 parent produces a brown pigment (27) (Fig. 2), an equal representation of orange and brown colonies of ascospore progeny collected from single perithecia provided a good indication of outcrossing, i.e., absence of selfing, and appropriate gene segregation. The resulting individual ascospore progeny were independently paired with vic tester strain EU55 containing hypovirus CHV-1/EP713. Because strain EU55 has a vic genotype (1221-22) that is completely heteroallelic with EP146 (2112-11), transmission of virus to ascospore progeny containing the inherited nondisrupted EP146 alleles vic1-2, vic3-1, vic6-1, or vic7-1 should be restricted, allowing rapid screening for candidate progeny that acquired virus due to the presence of disrupted vic alleles. This subset of progeny was then tested by PCR for the presence of all four disrupted vic alleles, the presence of the wild-type ku80 gene, and either the vic2-1 or the vic2-2 allele. This resulted in identification of five candidate strains, designated super donor (SD) followed by the ascospore ID number, with desired genotypes from 840 screened ascospore progeny, two containing the vic2-1 allele and three with the vic2-2 allele (Fig. 2).

Fig. 2.

Mating of quadruple vic mutant and EP146 strains to introduce disrupted vic genes into the vic2-1 allelic background and restore the ku80 wild-type gene in both vic2 allelic backgrounds. For all strains shown in this figure, the top view of the PDA culture is on the Left and the bottom view is on the Right. The quadruple vic mutant strain generated by the sequential vic gene disruption strategy shown in Fig. 1 exhibited growth characteristics and a colony morphology indistinguishable from the wild-type strain EP155, whereas the EP146 parent produced a brown pigment. Colony morphologies representative of the five selected super donor strains and corresponding genotypes are shown at the Bottom. All SD strains contained disruptions of vic1a-2, vic3b-1, pix6-2, and vic7a-2 and the wild-type ku80 gene. Strains SD82 and SD782 contained the vic2-1 allele whereas SD328, SD684, and SD752 contained the vic2-2 allele. All SD strains except SD752 contained the vic4-1 allele, only SD82 contained the MAT-2 mating-type locus, and only SD782 produced orange pigment. The growth characteristics and colony morphologies of the SD progeny strains were similar to that of the parental strains and wild-type strain EP155.

Phenotypic Characterization of, and Enhanced Transmission by, SD Strains.

The quadruple vic mutant strain [Δvic1a-2, Δvic3b-1, Δpix6-2, Δvic7a-2 (2211-22)] exhibited colony growth rates and colony morphologies (Fig. 2) indistinguishable from wild-type strain EP155. As shown in Fig. 2, the growth and morphology of all five mutant SD strains were indistinguishable from the parental Quadruple vic mutant and EP146 strains with mutant strains SD82, SD328, SD684, and SD752 producing the brown pigment similar to the EP146 parent and only mutant SD782 producing the orange pigment produced by the Quadruple vic mutant parent. Mutant strains SD82 and SD328 were further subjected to the dormant chestnut stem virulence assay. Whereas strain SD82 produced larger cankers than did strain SD328 on average, both strains exhibited virulence levels not significantly different from that of wild-type strain EP155 (t test, P > 0.05) (Fig. S2). Note that both SD mutant strains responded to CHV-1/EP713 infection by producing cankers significantly smaller than the wild-type strain, similar to CHV-1/EP713–infected EP155 (P < 0.001) (Fig. S2). Thus, under laboratory conditions, the mutant SD strains containing disruptions of four different vic genes representing four independent vic loci showed no obvious phenotypic changes relative to the parental strains or wild-type strain EP155. However, a final conclusion concerning the phenotypic consequences of these multiple gene disruptions awaits analysis under field conditions.

Fig. S2.

Dormant chestnut stem virulence assays for SD strains. Virulence assay results are reported as relative canker area compared with wild-type EP155-induced canker (error bars: ±SE; sample size: 6). (A) Results are shown for two independent single conidial isolates of strains SD82, SD82-SP2, and SD82-SP3. Columns designated EP155/CHV1 and SD82-SP2/CHV1 indicate the relative canker size produced by the corresponding hypovirus CHV-1/EP713–infected strains. (B) Results are shown for two independent single conidial isolates of strains SD328, SD328-SP1, and SD328-SP2 and for the CHV-1/713–infected strains EP155/CHV1 and SD328-SP1/CHV1 as in A. Canker measurements were taken on day 21 after inoculation for six replicates for each strain, one replicate per stem. Paired, two-tailed t-test suggested no significant differences between wild-type EP155 and virus-free SD strains (P > 0.05), even though strain SD82 produced 25% larger cankers than strain SD328 on average. Significantly smaller cankers were observed for virus-containing strains compared with virus-free wild-type and SD strains (P < 0.001). P values were calculated separately for individual isolates in each case.

The performance of the SD mutant strains as hypovirus donors was examined by pairing infected mutant strains with the 16 EU tester strains with vic genotypes that differed from the wild-type strain at three or more vic loci and by testing for conversion of the recipient EU tester strain to the virus-infected phenotype (Fig. 3). As expected, the transmission rate from the hypovirus-infected wild-type strain EP155 (2211-22) to the 16 EU tester strains was generally in the range of 0–5% (Table 3 and Fig. 3). Mutant strain SD82, containing the vic2-1 allele, was unable to efficiently transmit virus to the eight EU tester strains containing the vic2-2 allele, but transmitted virus at a level of 90–100% to all eight EU tester strains containing the vic2-1 allele (Table 3 and Fig. 3). Similar results were obtained independently for the vic2-1–containing mutant SD782. Conversely, mutant strain SD328 that contained the vic2-2 allele transmitted virus very efficiently to the eight EU tester strains containing the vic2-2 allele, but poorly to those tester strains with the vic2-1 allele. Clearly, heteroallelism at the vic2 locus results in significant resistance to virus transmission. However, as shown in Table 3, virus transmission from the SD328 strain to the vic2 heteroallelic EU tester strains proceeded at a significantly higher rate (22/160) than that from strain SD82 (1/160) (Fisher’s Exact Test, two tails: P = 2.82E⁻⁶).

Fig. 3.

Representative results of hypovirus transmission assays for wild-type and SD strains corresponding to the results presented in Table 3. For each PDA plate, the virus-infected donor strain is on the left and the virus-free recipient strain is on the right of the paired colonies. Donor strains are labeled at the left and recipient strains are labeled at the top. EU tester strains EU48, EU47, EU33, and EU45 all contain the vic2-2 allele and are heteroallelic with wild-type strain EP155 at vic loci 1,3,4,6; 1,3,4,7; 3,4,6,7; and 1.3.4.6.7, respectively. EU tester strains EU56, EU49, EU10, and EU40 all contain the vic2-1 allele and are heteroallelic with wild-type strain EP155 at vic loci 1,3,4,6; 1,3,4,7; 3,4,6,7; and 1.3.4.6.7, respectively. The mixture of virus-infected SD82 and SD328 was delivered as a 4-μL water suspension of finely ground mycelia produced from equal amounts of mycelia plugs of both strains using a microcentrifuge tube and sterile micropestle. White arrows indicate successful virus transmission and phenotypic conversion of recipient strains.

Table 3.

Hypovirus transmission frequency for super donor strains SD82 and SD328

		Donor
Recipient	Heteroallelic with EP155 at:*	EP155 (2211-22)	SD82 (2111-22)	SD328 (2211-22)	SD82 + SD328
EU52 (1221-12)	1,3,6	0/20	0/20	18/20	18/20
EU48 (1222-12)	1,3,4,6	0/20	0/20	17/20	17/20
EU63 (1221-21)	1,3,7	0/20	0/20	19/20	19/20
EU47 (1222-21)	1,3,4,7	0/20	1/20	20/20	20/20
EU46 (2221-11)	3,6,7	0/20	0/20	20/20	18/20
EU33 (2222-11)	3,4,6,7	0/20	0/20	19/20	20/20
EU37 (1222-11)	1,3,6,7	0/20	0/20	19/20	20/20
EU45 (1222-11)	1,3,4,6,7	1/20	0/20	17/20	18/20
EU50 (1121-12)	1,3,6	0/20	18/20	0/20	17/20
EU56 (1122-12)	1,3,4,6	0/20	20/20	6/20	20/20
EU53 (1121-21)	1,3,7	0/20	19/20	2/20	19/20
EU49 (1122-21)	1,3,4,7	1/20	19/20	2/20	20/20
EU36 (2121-11)	3,6,7	0/20	19/20	0/20	20/20
EU10 (2122-11)	3,4,6,7	0/20	20/20	1/20	18/20
EU34 (1121-11)	1,3,6,7	1/20	19/20	2/20	20/20
EU40 (1122-11)	1,3,4,6,7	0/20	20/20	9/20	20/20

^*The vic loci for which the recipient EU tester strains are heteroallelic relative to wild-type strain EP155. Note also that the top eight recipient EU tester strains contain the vic2-2 allele (same as strain EP155) whereas the bottom eight recipient EU tester strains contain the vic2-1 allele.

Hypovirulent strains with differing vic genotypes have been applied to C. parasitica cankers as mixtures to increase the probability of hypovirus transmission to resident strains of diverse vic genotypes (28, 29). We tested whether a mixture of infected SD82 and SD328 strains would be effective in transmitting virus to all 16 EU tester strains. As indicated in Table 3 and Fig. 3, when combined, strains SD82 and SD328 very effectively donated hypovirus (85–100%) to all 16 EU tester strains, irrespective of the vic2 allele.

Discussion

In a 1994 paper describing suppressor mutants of heterokaryon incompatibility in Neurospora crassa, Arganoza et al. (30) suggested that analogous mutants of pathogenic fungi might serve as universal donors of hypovirulence. The results presented here for C. parasitica and infecting hypoviruses demonstrate that pathogenic fungi can be engineered to transmit virulence-attenuating mycoviruses at a high level by suppressing allorecognition, in this case through disruption of the multilocus vegetative incompatibility loci.

Although gene disruptions that conferred a virus-donor–promoting phenotype had previously been identified individually for vic1, vic3, and vic6, additional analysis was required to identify appropriate disruption targets at vic7. The ankyrin-like gene vic7a identified in this study was an unexpected target candidate because, unlike most vic gene alleles, the allelic forms of this gene showed very little amino acid sequence polymorphism, i.e., 96% amino acid identity. Interestingly, all but one of the allelic polymorphisms resided within the Ankyrin repeat domain (Fig. S1). In this regard, the vic1a gene also showed low polymorphism—91% identity at the amino acid level. Thus, although a high-to-moderate level of polymorphism (37–85% amino acid identity) is the general rule for C. parasitica vic gene alleles, vic1a and vic7a are clear exceptions.

The absence of a gene disruption target at the vic2 locus that would promote virus donor activity was addressed by the introduction of a sexual cross between the hph/neo-SMG–free quadruple vic gene mutant (2211-22, Δku80^ben, MAT-2) and strain EP146 (1122-11, ku80, MAT-1) to produce SD strains containing the four disrupted vic alleles in either the vic2-1 or the vic2-2 allelic backgrounds. This approach provided the added benefit of replacing the ku80 disruption cassette with the wild-type ku80 gene to yield completely SMG-free SD strains by eliminating the benomyl-resistance SMG that was refractory to Cre-mediated excision due to the absence of flanking loxP elements. It is anticipated that future field releases of SD strains will benefit from the absence of SMGs as well as the availability of brown- and orange-pigmented SD strains (Fig. 2) that provide phenotypic field markers.

As predicted, SD strains with the vic2-1 allele (e.g., SD82) were able to transmit virus efficiency to multilocus heteroallelic strains as long as they contained the vic2-1 allele, whereas SD strains with the vic2-2 allele (e.g., SD328) efficiently transmitted virus to strains that contained vic2-2. The resistance to transmission by SD82 conferred by the vic2-2 allele in the recipient strain was very strong (Table 3) whereas the transmission by SD328 to recipients with the vic2-1 allele was somewhat leaky, but still restricted. Importantly, a mixture of the two SD strains was able to efficiently transmit virus to all 16 tester strains that were heteroallelic at three or more vic loci irrespective of whether they contained vic2-1 or vic2-2 (Table 3).

The use of mixed inoculum consisting of hypovirulent C. parasitica strains of multiple vegetative compatibility groups to increase hypovirulence transmission to field strains of diverse vic genotypes was reported as early as 1980 (28). Although this approach has been successful in increasing hypovirulence conversion capacity (28, 29, 31), it is limited by the number of different hypovirulent strains needed to convert all strains in a highly diverse vic population as found in most North American forest settings (7, 17) and requires considerable effort for strain propagation and inoculum formulation. In contrast, based on the laboratory transmission results presented here, the application of two SD strains representing the vic2-1 and vic2-2 alleles should convert field strains representing all possible vic genotypic combinations of the defined six diallelic vic genetic loci. The likelihood for successful conversion by SD strains under field conditions is increased when one considers reports that vic-mediated resistance to hypovirus transmission is lower in infected chestnut tissue than in laboratory transmission assays (7, 32–35). It is anticipated that the use of SD strains as vectors to effectively deliver hypoviruses into natural C. parasitica populations will enhance efforts to integrate hypovirulence-mediated biological control with increased disease resistance derived from backcross resistance breeding programs for woodland restoration of the American chestnut tree. Regulatory issues associated with the release of the engineered SD strains should be simplified by the use of the adapted Cre-loxP recombination system to generate SMG-free mutant strains.

Although the currently defined six diallelic vic loci account for the vic diversity of C. parasitica populations in the forest setting in the eastern United States (36), it is clear that additional vic diversity is present in Europe (19, 20, 37) and Asia (38). It is envisioned that the comparative genomic approach (22, 23) for identifying the currently defined vic loci could be used to identify additional C. parasitica vic alleles of existing vic loci or entirely new vic loci. These genes could then be disrupted as described in this report to broaden the effectiveness of the SD strains. Both the comparative genomics approach for identifying vic genes and the multilocus vic gene disruption strategy to promote mycovirus transmission should find general applicability for enhancing the biological control potential for the growing list of pathogenic fungi/mycovirus hypovirulence systems (2, 6, 39).

Materials and Methods

Fungal Strains and Growth Conditions.

The C. parasitica strains used in this study were maintained on potato dextrose agar (PDA; Difco) at 22 °C–24 °C with a 12 h/12 h light/dark cycle. Wild-type and genome reference strain EP155 (ATCC 38755) was originally isolated by Sandra Anagnostakis (Connecticut Agricultural Experiment Station, New Haven, CT) in 1977 near Bethany, CT. Strain EP146 (ATCC64671) was isolated by William MacDonald (West Virginia University) in 1976 near Franklin, WV. Strain DK80, a mutant of strain EP155 that contains a disruption of the nonhomologous end-joining DNA repair pathway ku80 gene homolog to promote homologous recombination (26), was used as the founder strain for vic gene disruptions to construct the super donor strains. The 64 vic genotyped EU tester strains (21) were provided by Michael Milgroom (Cornell University).

Disruption of vic Genes and SMG Excision.

Disruption of vic genes was performed by homologous recombination in C. parasitica strain DK80 using PCR-generated disruption fragments based on the strategy of Kuwayama et al. (40) with modifications as described by Zhang et al. (25) to include flanking loxP sites as described previously (23) for the single disruptions of the individual vic1a-2 and vic3b-1 genes. The gene disruption fragments containing loxP-flanked SMGs were excised by providing the Cre recombinase in trans via anastomosis with a compatible C. parasitica Cre-producing strain as described by Zhang et al. (25). Multiple rounds of anastomosis may be required to achieve excision of multiple SMGs in a single strain. The benomyl resistance SMG present in strain DK80 lacked loxP flanking elements (26) and was eliminated from the SD strains by replacement with wild-type ku80 gene from parental strain EP146 in a sexual cross between the hph/neo-SMG–free quadruple vic mutant strain (2211-22, Δku80^ben) and strain EP146 (2211-22, ku80) performed as described by Chen et al. (27). Oligonucleotides used in this study are indicated in Table S1.

Table S1.

Oligonucleotide PCR primers used in this study

Primer name	Primer sequence (5′-3′)	Note
AnkF1d	ATC AAC GCC AAC CCG CTA CGG TTC AA	For ∆vic7a-2 analysis
Ank-nestd	CAG CAC GGT GTT CAC CAA TAC CAA GCT	For generating the vic7a-2 disruption fragment with loxP-neo-loxP replacing vic7a-2 gene
AnkF1u	TGC AGC GTA CGA AGC TTC AGC TGG GCA TGG AGC ATT TCC TAG T
AnkneoF2d	ACT AGG AAA TGC TCC ATG CCC AGC TGA AGC TTC GTA CGC TGC A
AnkneoF2u	GTA GTG CCT TCG CTT TCG TAA GGC CAC TAG TGG ATC TGA TAT CA
AnkF3d	TGA TAT CAG ATC CAC TAG TGG CCT TAC GAA AGC GAA GGC ACT AC
AnkF3u	CAG CCA GCG AGG TCT GCT TTC TTG	Also for ∆vic7a-2 analysis
AnkcDNAd	CGT CCC TCG AGA CCA TCA A	For vic7a-2 CDs analysis
AnkcDNAu	GCT CCT TCG CTC TGT CCT
pix6koF	TGA CAA GAT TGA GGG GTG GCC TGA	For ∆pix6-2 analysis
pix6F1d	GGC ATT GAT GCG TAC GTA AGT CAA GTT G	For generating the pix6-2 disruption fragment with loxP-neo-loxP replacing the pix6-2 gene
pix6F1u	AGC GTA CGA AGC TTC AGC TTA CAG GGT CGG ATG CTA ACG
pix6neoF2d	CGT TAG CAT CCG ACC CTG TAA GCT GAA GCT TCG TAC GCT
pix6neoF2u	TAG CCC TGA CTT GTC ATG AAC AAG GCC ACT AGT GGA TCT GA
pix6F3d	TCA GAT CCA CTA GTG GCC TTG TTC ATG ACA AGT CAG GGC TA
pix6F3u	CAA GGA TAG GAA CCC GAC GTA	Also for ∆pix6-2 analysis
vic4F	GCA TGG GCA ATG GTC TAC TT	For vic4 allele analysis
vic4-1R	TCC ATC GCA TAG GTA CGC TC
vic4-2R	GCA ATC AAC GAT CAT CTG TC
M1-GS1F	ATC ACA AGT CGG CTC CCA CGA A	For mating-type analysis
M1-GS2R	TAA TCG CCG CTT GAG TAC TCG TG
M2-GS2F	GGG TCC AAA ATA TGG GTA CAG
M2-GS2R	TGA TGT CGG CAG CCT T

Analysis of Fungal Phenotype and Virus Transmission: Colony Growth, Virulence, and Virus Transmission.

Colony morphology and virulence on dormant chestnut stems were assessed according to standard procedures (41). Virus transmission assays were performed as described in Choi et al. (22). Infections by hypoviruses CHV-1/EP713 were initiated in C. parasitica strain EP155 by transfecting fungal spheroplasts with viral transcripts generated in vitro from full-length viral cDNA as described by Chen et al. (42). Hypoviruses were then introduced into the SD strains by anastomosis.