Canker Development and Biocontrol Potential of CHV-1 Infected English Isolates of Cryphonectria parasitica Is Dependent on the Virus Concentration and the Compatibility of the Fungal Inoculums

Biological control of Cryphonectria parasitica fungus, causal agent of chestnut blight, by virus infection (hypovirulence) has been shown to be an effective control strategy against chestnut blight in Europe and some parts of North America. The most studied mycovirus is the Cryphonectria hypovirus 1 (CHV-1) type species of the Hypoviridae family. To efficiently provide biocontrol, the virus must be able to induce hypovirulence in its fungal host in chestnut trees. Here, two different CHV-1 subtype I virus strains (E-5 and L-18), gained by transmissions, were tested for their hypovirulence induction, biocontrol potential, and transmission between vegetatively compatible (VCG) and incompatible fungal isolate groups in sweet chestnut seedlings and branches. Both strains of CHV-1 showed great biocontrol potential and could protect trees by efficiently transmitting CHV-1 by hyphal anastomosis between fungal isolates of the same VCG and converting virulent to hypovirulent cankers. The hypovirulent effect was positively correlated with the virus concentration, tested by four different reverse-transcription PCRs, two end-point and two real-time methods, one of which represents a newly developed real-time PCR for the detection and quantification of CHV-1.


Introduction
Chestnut blight is a disease of the Castanea species caused by the ascomycete Cryphonectria parasitica (Murrill) M. E. Barr. The fungus originates from Eastern Asia [1] but it has caused severe epidemics resulting in death and dieback of both Castanea dentata (Marshall) Borkh. in North America where it was introduced in the late nineteenth century [2], and on sweet chestnut (Castanea sativa Mill.) in most of continental Europe, where it was first introduced in Italy in 1938 [3]. Diseased chestnut trees exhibit crown dieback above girdling cankers on the trunk and/or branches and profuse epicormic growth may be observed below the cankers. Additional signs include orange fruiting bodies that erupt through swollen lenticels and/or whitish mycelial fans that form beneath the bark and spread into the phloem and cambium tissue. Planting stock, timber, bark, and seeds are pathways of introduction and long-distance dispersal.
England was considered free of chestnut blight until 2011, when C. parasitica infections were discovered on 90 young saplings of sweet chestnut planted in a nursery farm in Warwickshire [4]. They originated from the same nursery in Europe; the saplings were imported and planted in 2007 and some trees died and were replaced in 2010. This stimulated surveys between 2011 and 2012 where the fungus was identified on recently On the other hand, four virus-free English fungal isolates were used. LAP731 and FTC687 VIRUS-FREE, both isolated in 2021 from London, WAR706 VIRUS-FREE from Devon, and DIG460 isolated in 2020 from Devon [7], belonging to the VCGs EU-10, EU-10, EU-9, and EU-9, respectively, were used as virulent fungal isolates.

Preservation of Virus-Infected Fungal Strains and Assessment of Their Viral Load
Each virus-infected fungal isolate was preserved after growing on both PDA plates and PDA plates with two sterile filter disks (Whatman 1820-055), square portions of 1 × 2 cm at 25 • C under a 16 h photoperiod of 2500 lux for 14 days. Initially, six mycelial plugs with abundant sporulation were deposited at −80 • C in cryovials with 800 µL glycerol 22%, previously heavily vortexed, and immediately frozen in liquid nitrogen just before entering the ultra-freezer within a mapped cryogenic box.
The filter disk portions were transferred to a plastic box with silica gel at −20 • C in the form of colonized and dried (48 h under laminar flow chamber) filter disk portions included within sterile paper envelopes.
To quantify the virus content of each isolate, before and after two months of the two types of preservation (two replicates), incubation, RT-PCR, and electrophoresis methods were used as previously described in Pérez-Sierra et al. [6].

Inoculation of Sweet Chestnut Seedlings and Branch Segments, and Fungal Re-Isolation
Eighteen-month-old chestnut tree seedlings (C. sativa; provenance, Delamere, United Kingdom) of approximately 1.3 m high and approximately 2.5 cm wide (diameter) at the bole, were purchased from an English nursery free of the disease (Delamere) and grown outside for one year from May 2021 to May 2022. They were moved to a biosafety level 3 greenhouse at Forest Research Holt quarantine laboratory and acclimated to the greenhouse environment for a week (BSL3) which was temperature-controlled (25/20 • C, day/night), with a photoperiod of 8 h of light.
Two different assays (I and II) were done. Both assays were repeated by using sweet chestnut branch segments 25 cm long and about 2 to 2.5 cm in diameter previously kept one week in a cold store. Both plant materials were supplied with tap water twice a week. All inoculations were done using mycelial plugs taken from actively growing cultures after seven days on PDA. The viral content in the hypovirulent isolates was re-checked following Pérez-Sierra et al. [6] at that point. For assay I (individual inoculations to investigate virulence), the six different virusinfected C. parasitica strains were inoculated onto seedlings and branch segments. In the case of the seedlings, four biological replicates were used for each strain and for the virusfree controls plus PDA controls. Thus, a total of 44 seedling lesion areas were measured after approximately two months. In the case of the branch segments, three branch segments were randomly held in buckets distributed across a mapped trial with 11 buckets. After inoculation, the holes were sealed with LacBalsam (Compo, Eggenfelden, Germany) to prevent desiccation. At the time of harvesting, (53 days after the individual inoculations), the length and width of the cankers were measured, and the canker area was calculated using the ellipse formula, A = L/2 × W/2 × pi, where area equals half-length per half-width per pi number.
For assay II (challenge inoculations to estimate biocontrol potential), a total of 64 seedlings (only VCG compatible combinations) or 140 branch segments (all the combinations) were used. Two weeks following primary inoculations with the virulent (virus-free) fungal strains, challenge inoculations were made with the virus-infected strains or PDA. Eight inoculations regularly distributed along the periphery of a virulent canker were carried out [16]. After inoculation, the holes were sealed with sterile water-soaked sterile cotton, parafilm and aluminium foil to prevent desiccation. At the time of challenge inoculation, the length and width of the original cankers were measured, and the canker area was calculated using the ellipse formula. Canker expansion after the biocontrol treatments was also measured after 53 days after the challenges by using the same approach. In the case of the chestnut branch segments, five branch segments were randomly held in buckets distributed across a mapped trial with 28 buckets.
All cankers were sampled at the end of the experiment to verify virus infection. Four bark samples (top, two middles, and bottom of the canker) were taken from each canker using a bone marrow biopsy needle (diameter, 1.6 mm; Microlance 3; BD, Huesca, Spain). Bark plugs were placed on malt agar plates containing streptomycin at a concentration of 40 mg/L and incubated at 20 • C for four days. The outgrowing mycelium was transferred onto PDA (39 g/L, BD Difco) plates, which were then incubated at 25 • C in the dark for 7 days.

Direct One-Step Reverse Transcription PCR and Comparison with Other Endpoint and Real-Time Virus Detection Methods
The simplified and reliable one-step RT-PCR technique developed by Urayama et al. [17] for virus detection in Magnaporthe oryzae without nucleic acid extraction, more importantly already tested for C. parasitica viruses by the same study and the ones of Aulia et al. [18] and Suzuki et al. [16], was employed to detect the virus in the fungal isolates re-isolated from cankers. This method entails stabbing the central 7 days growing region of mycelial colony on PDA with a toothpick and dipping the toothpick into a 19 µL premixed RT-PCR mixture prepared according to the protocol for PrimeScript One Step RT-PCR version 2 (Dye Plus) (TaKaRa Bio, Saint-Germain-en-Laye, France).
This detection method, therein after called the End-point TaKaRa II Toothpick, was compared with that described in Pérez-Sierra et al. [6], comprising RNA extractions with on-column DNA digestions and named End-point Qiagen Extract, and with a new method, called Real-time TaKaRa III, either from RNA extracts or colony toothpick. This method was used for the first time in the present study and represents, to the best of our knowledge, the first time that a RT-qPCR was designed for CHV-1. It implies the use of a One Step PrimeScript III RT-qPCR kit (TaKaRa Bio, Saint-Germain-en-Laye, France) in combination with the primers and probes indicated in Table 2, designed by using OligoArchitect TM Online (Sigma-Aldrich, Gillingham, UK). Real-time PCRs (qPCRs) were carried out on a LightCycler 480 (Roche, Welwyn, UK). For the assay, 0.4 µM of the primers and probes were used in a 20 µL reaction volume comprising 10 µL 2 × One Step PrimeScript III RT-qPCR Mix, 0.4 µL of each primer and probe, 6.6 µL RNase free water, and 1 µL of RNA sample (or toothpick dipping) per reaction. Each sample was performed in triplicate. Thermal cycling conditions were 50 • C for 30 min and 95 • C for 50 s, followed by 40 cycles of 95 • C for 25 s and 53 • C for 1 min. Fluorescent detection occurred at the end of each 53 • C step. The cycle threshold (Ct) value was calculated automatically using the LightCycler software (Roche, Welwyn, UK) with absolute quantification using a second derivative maximum setting with 465-510 and 533-580 nm channel filters for analysing the specific amplicons and the actin internal control. A ten 1:10 points serial dilution of the 94 bp specific qPCR product synthesized de novo and cloned into a plasmid pUC-GW-Kan (Azenta Life Sciences, Leipzig, Germany), and posterior regression equation analyses, permitted estimating the fragment copy number, actual number of the CHV-1 virus, in real samples depending on the mean threshold cycles number per triplicate wells ( Figure 1). For calculating the copy number in the original plasmid aliquot, the copy number was calculated using the following equation: number of copies/µL = [(6 × 10 23 ) × (DNA concentration, 500 ng/µL)/molecular weight of one plasmid], where 6 × 10 23 is the number of copies per mole, DNA concentration is given in grams per microliter, and the molecular weight of one plasmid is in grams per mole assuming a plasmid size of 2627 bp and a 1 bp molecular weight of 660 g/mole.

Statistical Analyses
Analyses were conducted in R (version 4.1.0), with graphics produced using ggplot2 in R [19][20][21][22]. In all, [23] linear mixed effects models were applied to the data using the square root internal lesion area as the response (to meet normality assumptions) and random effects for isolate. The interaction of row and column position (numeric) was included in all initial models as fixed effects. Akaike Information Criteria (AIC) were used to select the most efficient model. Having selected the most efficient model, an Analysis of Variance (ANOVA) (Type 2 F tests, [24]) was used to determine the significance of fixed effects, with non-significant effects removed from the final model. Estimated marginal means with pairwise contrasts (Tukey's HSD corrections, 22) were used to show significant differences within fixed effects.
Viruses 2022, 14, x FOR PEER REVIEW 6 of 17 is in grams per mole assuming a plasmid size of 2627 bp and a 1 bp molecular weight of 660 g/mole.

Statistical Analyses
Analyses were conducted in R (version 4.1.0), with graphics produced using ggplot2 in R [19,20,21,22]. In all, [23] linear mixed effects models were applied to the data using the square root internal lesion area as the response (to meet normality assumptions) and random effects for isolate. The interaction of row and column position (numeric) was included in all initial models as fixed effects. Akaike Information Criteria (AIC) were used to select the most efficient model. Having selected the most efficient model, an Analysis of Variance (ANOVA) (Type 2 F tests, [24]) was used to determine the significance of fixed effects, with non-significant effects removed from the final model. Estimated marginal means with pairwise contrasts (Tukey's HSD corrections, 22) were used to show significant differences within fixed effects.
Considering assay I with seedlings, three initial models were fitted to the data using different virus fixed effects: (1) virus strain, (2) virus absence/presence, and (3) original virus concentration.
Regarding assay I, using branches, the same three initial models were fitted to the data.
Considering assay II with seedlings, three separate models were applied to the data, one including the interaction of primary and challenge inoculation, one with the interaction between the primary inoculation and challenge inoculation strain, and one with the interaction between the primary inoculation and control/virus, with all models including row and column position (numeric) and initial lesion area (all challenge inoculations were compatible, so this effect was excluded) as additional predictors.
Regarding assay II using branches, the data indicated that most samples had reached the maximum achievable lesion size by the end of the experiment, therefore data were analysed in two ways: (1) using the binary response of not achieved/achieved maximum lesion size, and (2) the analysis of actual lesion size for the subset of samples that had not achieved maximum lesion size. The number of samples achieving maximum lesion size versus not was used as a two-column matrix response in a generalized linear model with binomial errors and a logit link function. Two separate models were applied to the data, one including the interaction of primary and challenge inoculation and one with the interaction between primary inoculation and challenge inoculation strain, with both Considering assay I with seedlings, three initial models were fitted to the data using different virus fixed effects: (1) virus strain, (2) virus absence/presence, and (3) original virus concentration.
Regarding assay I, using branches, the same three initial models were fitted to the data. Considering assay II with seedlings, three separate models were applied to the data, one including the interaction of primary and challenge inoculation, one with the interaction between the primary inoculation and challenge inoculation strain, and one with the interaction between the primary inoculation and control/virus, with all models including row and column position (numeric) and initial lesion area (all challenge inoculations were compatible, so this effect was excluded) as additional predictors.
Regarding assay II using branches, the data indicated that most samples had reached the maximum achievable lesion size by the end of the experiment, therefore data were analysed in two ways: (1) using the binary response of not achieved/achieved maximum lesion size, and (2) the analysis of actual lesion size for the subset of samples that had not achieved maximum lesion size. The number of samples achieving maximum lesion size versus not was used as a two-column matrix response in a generalized linear model with binomial errors and a logit link function. Two separate models were applied to the data, one including the interaction of primary and challenge inoculation and one with the interaction between primary inoculation and challenge inoculation strain, with both models including challenge inoculation compatibility as an additional predictor. AIC values were used to determine the most efficient model, with the analysis of deviance (likelihood ratio chi-square tests, [25]) used to determine significant effects.

Transmissions
All the co-culture hypovirus transmissions were successful, with gained virus concentrations ranging between 234.37 and 612.83 ng/µL of RNA extract after the virus-specific RT-PCR [6], equivalent to about 536.7-1403.3 ng/mg of mycelium. This represents a mean increment of 912.8 ng of amplicon per mg of mycelium with respect to the determined concentrations in the wider environment.

Preservations
Both glycerol and the paper disk methods worked nicely in general to preserve the CHV-1 strains in the infected fungal isolates, and although there were no significant differences, the glycerol method results were on average slightly better than the paper disks preservation method (Table 3). Table 3. Viral concentrations (ng/µL) before and after two months of two types of preservations (see methods second epigraph).

Assay I, Pathogenicity Test
Using seedlings, initial correlations suggested strong (>0.7) positive correlation between all pre-and post-experimental viral concentrations. The most efficient model included the absence/presence of the virus only (not strain or viral concentration; (F 2,8 = 40.1, p < 0.0001). Lesions caused by virus-free isolates were significantly larger than those caused by virus-infected isolates (Figures 2A and 3A). Lesion areas of virus-infected isolates were not significantly larger than the control inoculations with PDA ( Figure 2A). There was a notable variation across fungal isolates, with the LAP731 (virus-free) isolate having markedly larger lesion areas than the other virus-free isolates ( Figure 3A). The results showed a negative correlation between the lesion area and the original or post-harvesting viral concentration (Supplementary Table S1). This is consistent with that as the viral concentration is higher, the lesion area is smaller. Besides, the original viral concentration was positively correlated with all post-harvesting viral concentrations, except virus copy number (Supplementary Table S1). Strain L-18 was correlated with higher virus concentrations, even virus copy number, and thus with smaller lesion areas (Supplementary Table S1).
Similarly, the most efficient model using branch segments also included the absence/presence of the virus only (not strain or viral concentration; (F 2,8 = 11.8, p = 0.004). Lesions caused by virus-free isolates were significantly larger than those with the virus or negative controls ( Figure 2B). Lesion areas of those fungal isolates with the virus were not significantly larger than the controls (Figures 2B and 3B). There was notable variation across fungal isolates, with the FTC687 (virus-free) isolate having markedly smaller lesion areas than the other virus-free isolates ( Figure 3B). Similar correlations were found than those using seedlings (Supplementary Table S2). areas than the other virus-free isolates ( Figure 3B). Similar correlations were found than those using seedlings (Supplementary Table S2).    (Figure 4). The results showed a negative correlation between the lesion area and the original or post-harvesting viral concentration (Table 4). The original viral concentration was positively correlated with all post-harvesting viral concentrations, including virus copy number. Strain L-18  areas than the other virus-free isolates ( Figure 3B). Similar correlations were found than those using seedlings (Supplementary Table S2).    (Figure 4). The results showed a negative correlation between the lesion area and the original or post-harvesting viral concentration (Table 4). The original viral concentration was positively correlated with all post-harvesting viral concentrations, including virus copy number. Strain L-18

Assay II, Biocontrol Potential
Using seedlings, the most efficient model included the challenge inoculation virus absence/presence only. There was a significant interaction between virus absence/presence and primary inoculation isolate; (F [3,26] = 8.21, p = 0.0001) along with a main effect of virus absence/presence (F [1,56] = 165, p <0.0001). Challenge inoculation virus presence was associated with significantly smaller lesions (Figure 4). The results showed a negative correlation between the lesion area and the original or post-harvesting viral concentration ( Table 4). The original viral concentration was positively correlated with all post-harvesting viral concentrations, including virus copy number. Strain L-18 was correlated with bigger virus concentrations and thus with smaller lesion areas (Table 4). inoculation compatibility only; (LR chi-sq = 34.1, p < 0.0001). Those treatments where challenge inoculations were compatible were significantly less likely to achieve maximum lesion area versus incompatible and control challenge inoculations. The second and third most efficient models included challenge inoculation viral strain; (F1,13 = 7.7, p = 0.016) along with primary inoculation isolate (F3,13 = 4.6, p = 0.021). Challenge inoculation strain 2 (L-18) was associated with smaller lesions. Smaller lesion areas were related with higher post-harvest virus concentrations (Table 5).   Using branches, all successful challenge inoculations were compatible, so this effect was excluded in the analyses. Thus, the most efficient model included challenge inoculation compatibility only; (LR chi-sq = 34.1, p < 0.0001). Those treatments where challenge inoculations were compatible were significantly less likely to achieve maximum lesion area versus incompatible and control challenge inoculations. The second and third most efficient models included challenge inoculation viral strain; (F 1,13 = 7.7, p = 0.016) along with primary inoculation isolate (F 3,13 = 4.6, p = 0.021). Challenge inoculation strain 2 (L-18) was associated with smaller lesions. Smaller lesion areas were related with higher post-harvest virus concentrations (Table 5).

New Real-Time PCR
Real-time TakaRa III Toothpick method data revealed that this new PCR will be very useful and faster for detecting the CHV-1 virus in more and even less concentrated (common in England) cultures because those results indicate around four-fold times more amplification efficiency than the protocol normally used (End-point Qiagen Extract) and without the need to perform either RNA extractions or gel electrophoresis.
The triplicate tests' standard deviation was very low, giving very similar results per sample. Furthermore, the use of the ten 1:10 serial dilution of the cloned qPCR product permitted determining the detection limit of this PCR protocol in two viruses. This realtime PCR was tested against five different subtypes of the mycovirus (I, F1, E, D, and G) with positive results. However, the unique weak point of this method is that it could be more difficult to discriminate statistically significant differences among the different infected isolates in comparison with the other three methods, especially the end-point methods ( Figure 5).  Bars with an asterisk indicate significant differences against the same isolate treatment analysed using the Qiagen Extract End-point PCR method for the CHV-1 mycovirus.

Discussion
Cryphonectria hypovirus 1 is known to induce hypovirulence in C. parasitica by reducing pathogenic growth and sporulation, hence the virus is used in Europe for disease control [12]. The virus has been introduced into continental Europe on multiple occasions in association with C. parasitica from countries such as Japan, China [1,26] and Korea [27], which are known to be the geographical origin of the fungus. Since those introductions, both C. parasitica and CHV-1 have spread widely [27]. Originally, such introductions are most likely to have occurred through the plant trade and/or the importation of Asiatic planting stock often intended for use in breeding resistance programs to chestnut ink disease caused by Phytophthora cinnamomi and P. cambivora. Six genetically distinct CHV-1 subtypes have been identified in Europe (I, D, E, F1, F2 and G) [14,28]. Subtype I (also known as the Italian subtype), and which is the only subtype that has been detected in England [6,7,13], is the most widespread, because it is commonly associated with mild hypovirulence. It is dominant in Italy, Switzerland, south-eastern France, Greece [29], Bosnia [30], Croatia [31], Slovenia [32], Macedonia [33], Turkey [34], and now also England [6,7,13], where the haplotype exactly matches haplotype E-5. The European distribution of the E-5 haplotype has not been well studied. It was described as a rare haplotype by Gobbin et al. [14], and it has been introduced in an experimental site near Monthey (Switzerland), where it became widely established [35]. The other subtypes, especially F1 with the type member (CHV1-EP713), are usually less abundant because the increased aggressiveness of the virus causes a very relevant reduction in fungal growth and therefore makes infected individuals less likely to persist. F1 and other subtypes have been found in some parts of France [36], Spain [37], eastern Turkey [34], Georgia [28], and Germany [38]. The CHV-1 virus does not occur naturally in the USA but is present in a few locations such as in Virginia [39,40], Wisconsin [41] or Maryland [42] where it has been released for the biological control of American chestnut blight. In Europe, several genetic re-combinations have contributed to the evolution of CHV-1 [28,43].
There have been relatively few studies on the impact of CHV-1 induced hypovirulence using plant material [39,44,45]. To the best of our knowledge, this is one of the first times that the pathogenicity and biocontrol potential of virulent and hypovirulent isolates of C. parasitica have been evaluated on both chestnut seedlings and cut branches [13,16], although other studies have used only cut branches [16,45], naturally infected trees [39,42], or both [44].
Our findings showed that there is a considerable difference in mean lesion area between virus-free isolates and those harbouring a high concentration of the CHV-1 virus, both when using seedlings and branches. The consistency of the plant material and controlled conditions used in our four assays gave high confidence about our observations and the conclusions that could be drawn from the data in relation to the efficacy of those artificially (by co-culture) infected English isolates.
Following findings of chestnut blight in the wider environment in England from late 2016 onwards [6], CHV-1 was first detected in C. parasitica populations at the end of 2017 but does not appear to be widespread based on the very low frequency of findings and the low concentration of detectable virus in affected isolates [6,7,13]. The results from this study suggest that the isolates FTC687 and WAR706 infected with the virus strain E-5 have the most potential for development as biocontrol agents of C. parasitica in the UK. These fungal isolates, which are now virus infected, initially originate from the east London area and Devon, where 77 and 60% of the total isolates belong to single VC groups (EU-10 in London and EU-9 in Devon) [6,7]. These outbreak clusters are possibly longstanding. The prevalence of a single VCG in each geographic area is likely to facilitate the spread of the virus and consequently hypovirulence [46,47]. It also opens the possibility of further evaluation of the field effectiveness of these two hypovirulent isolates in London and Devon FTC and WAR sites as pilot sites. However, further field experiments are needed before escalating the field inoculations by obtaining other VCGs transmissions for other locations where outbreaks have been detected with a low VCG diversity. Those will be comprised of periodic assessments of the treated and untreated cankers in the field every 6 months for a period of two years to determine their efficacy to control chestnut blight in the wider environment.

Conclusions
(1). Two CHV1 strains originating from Switzerland were successfully transferred by hyphal anastomosis to British isolates of Cryphonectria parasitica belonging to vc type EU10 and EU9, the predominant vc type in London and Devon, respectively.
(2). The biocontrol potential of the CHV1-infected British isolates was experimentally verified and found to be dependent on the inoculum compatibility and on the virus concentration, on both seedlings and cut branches.
(3). Real-time TakaRa III Toothpick method data revealed that this new PCR will be faster (no need for RNA extractions or electrophoresis), and more sensitive (around four times) for detection and quantification of the CHV-1 virus.
(4). Following a completed Pest Risk Assessment (PRA) and regulatory approval, the CHV1-infected isolates characterized in this study will be available for further biocontrol testing under field conditions. Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/v14122678/s1, Table S1. Correlation results among all the tested parameters in assay I, using seedlings. Table S2 Correlation results among all the tested parametersin assay I, using branches.