In order to improve understanding of its diversity, 338 isolates of Cryphonectria parasitica, the causal agent of chestnut blight, were sampled from 10 chestnut populations throughout chestnut-growing coastal and continental areas of Croatia. Eighteen vegetative compatibility (VC) types were identified. The VC type EU-1 was the most widespread, comprising 42·9% of the isolates, followed by EU-2 (21%) and EU-12 (14·2%). In respect to the occurrence of the main VC types, the C. parasitica populations in Croatia combined features of both northwestern and southeastern European populations. Perithecia and mating-type ratios of approximately 1 : 1 were found in all populations, suggesting that sexual reproduction of the fungus is common in Croatia. Natural hypovirulence was also evident in all populations, with incidence of hypovirus-infected isolates ranging from 12·7% in Istria-Buje to 66·6% in the continental part of the country. A total of 36 hypovirus-infected isolates sampled throughout Croatia were analysed in ORF-A and ORF-B by RT-PCR/RFLP analysis. All viral isolates belonged to the Italian subtype of Cryphonectria hypovirus 1 (CHV-1) and were closely related to the isolates found in other European countries. The RFLP patterns found were also identical or similar to the patterns of three isolates collected in Croatia 22 years ago, suggesting a slow evolution of the hypovirus.
The ascomycete fungus Cryphonectria parasitica, the causal agent of chestnut blight, was introduced into the USA from Asia at the beginning of the 20th century and devastated nearly all of the American chestnut (Castanea dentata) stands (Anagnostakis, 1987). In Europe, chestnut blight was first observed in northern Italy in 1938, from where it rapidly spread to most of the chestnut (Castanea sativa) growing areas in Europe (Robin & Heiniger, 2001). However, damage caused by this pathogen in Europe was much less severe than in North America and many European chestnut stands recovered from the disease as a result of the natural occurrence of hypovirulent C. parasitica strains (Heiniger & Rigling, 1994).
Hypovirulent isolates are infected with a cytoplasmic double-stranded (ds) RNA hypovirus that reduces virulence (≡ pathogenicity) (hypovirulence) and sporulation of C. parasitica. The hypovirus can be transmitted from infected to uninfected strains via hyphal anastomosis, providing the basis for biological control of chestnut blight (MacDonald & Fulbright, 1991). Horizontal transmission of the hypovirus is limited by a vegetative incompatibility (vic) system of the fungus, involving at least six vic loci (Cortesi & Milgroom, 1998). The hypovirus readily spreads between strains with the same vegetative compatibility (VC) type, but it spreads less frequently between strains of different VC types (Cortesi et al., 2001). The hypovirus is transmitted vertically into asexually produced conidia, but not into sexually produced ascospores (Prospero et al., 2006). Two mating type alleles, MAT1-1 and MAT1-2, at a single locus control sexual compatibility in C. parasitica (Marra & Milgroom, 2001). Sexual reproduction has a negative impact on the spread of hypovirulence because ascospores are hypovirus-free and new VC types can arise through recombination of polymorphic vic genes, thereby increasing VC-type diversity (Cortesi & Milgroom, 1998; Milgroom & Cortesi, 1999). Low VC-type diversity favours dissemination of the hypovirus and is thought to be a critical factor in the success of biological control of chestnut blight in Europe (Anagnostakis et al., 1986; Heiniger & Rigling, 1994; Robin et al., 2000; Cortesi et al., 2001).
Cryphonectria hypovirus 1 (CHV-1) plays a major role in the biological control of chestnut blight. Among the four species of Cryphonectria hypoviruses that have been described, CHV-1 is the best known and the only one that has been found in Europe to date (Allemann et al., 1999; Milgroom & Cortesi, 2004). Its genome is encoded on a large dsRNA which contains two open reading frames, ORF-A and ORF-B. In Europe, five subtypes of CHV-1 have been identified, with the Italian subtype being the most widespread (Gobbin et al., 2003). Although the hypovirus leads to the recovery of chestnut stands in some locations in Europe where chestnut blight is present for a longer time period (Italy, Switzerland south of the Alps, and southwestern France), the disease still causes severe damage in other areas, particularly where it has been recently introduced (Macedonia, Greece, Germany, Spain and Portugal) (Robin & Heiniger, 2001; Sotirovski et al., 2004; Aguin et al., 2005; Bragança et al., 2007).
Chestnut stands in Croatia have also been devastated. In Croatia, the disease was first noticed in the western, coastal region, near Opatija, in 1955 (Halambek, 1988). It probably spread to Croatia from the north, from Slovenia, where it first appeared in 1950 in Nova Gorica, 4 km from the Italian border. In 1961, chestnut blight was first noticed in Cazin and Velika Kladuša in northern Bosnia-Herzegovina, near the Croatian border. In the former Yugoslavia, including the territory of Slovenia, Croatia, Bosnia-Herzegovina and Macedonia, disease progression was at first characterized by the absence of a defined disease front and by outbreak points far from existing diseased areas (Jurc, 2002). All attempts to stop the spread of the chestnut blight during the 1950s and 1960s failed and research ceased until 1978, when callusing cankers were noticed in chestnut stands in continental Croatia (Halambek, 1986). Isolates of C. parasitica having white culture morphology were recovered from those cankers (Halambek, 1988). Disease surveys conducted in the 1980s and 1990s revealed that by then chestnut blight had spread throughout the chestnut-growing regions of Croatia (Halambek 1988; Novak-Agbaba et al., 2000).
In spite of the widespread occurrence of chestnut blight in Croatia, little is known about its VC-type diversity or diversity of the hypovirus. Three hypovirus-infected C. parasitica isolates from Croatia sampled in 1985 were analysed by Allemann et al. (1999). All three isolates were infected by the Italian subtype of CHV-1 and had different VC types: EU-1, EU-12, and EU-13. Knowledge of the distribution and diversity of VC types and mating types is required in order to select appropriate hypovirulent isolates for biological control (Heiniger & Rigling, 1994; Cortesi et al., 1998). Since the genotypes of most VC types in Europe are known, this information can also be used to assess the risk of an increase in VC-type diversity through sexual recombination of polymorphic vic genes (Cortesi & Milgroom, 1998). Finally, knowledge of the occurrence and frequency of hypovirus subtypes will allow the assessment of the state and potential of natural hypovirulence.
The objectives of this study were to (i) determine VC and mating types of C. parasitica isolates sampled from 10 populations sampled throughout Croatia, (ii) estimate VC-type diversity and assess the possibility of an increase in that diversity, (iii) assess the presence of perithecia, and (iv) estimate the incidence of hypovirus-infected isolates in the populations and determine which CHV-1 subtypes are present.
Materials and methods
Cryphonectria parasitica was sampled between 2004 and 2006 from 10 populations, throughout the chestnut growing areas of Croatia (Fig. 1). The exact time of C. parasitica introduction in each sampled population was unknown, but it was known that in Istria C. parasitica was introduced in the 1950s, while in the continental part of the country it appeared in the 1970s (Halambek, 1988). Each population consisted of randomly collected isolates in an area of 1–3 ha. Populations were at least 2 km apart from each other and all of them were sampled in coppice stands. In each population, one canker per tree was sampled, with the exception of 21 trees, in which two cankers were sampled. For isolation, a piece of bark (5 × 7 cm) was taken from the margin of each canker and stored at –20°C until C. parasitica was isolated as described by Cortesi et al. (1996). Before each sampling all tools were sterilized in 96% ethanol. To assess the presence of perithecia, bark samples were examined under a dissecting microscope.
Fungal cultures were maintained on potato dextrose agar (PDA, Biolife) in Petri dishes. All C. parasitica isolates were assessed for the presence of hypovirus using culture morphology. For that assay the isolates were grown on PDA plates at 20–22°C in the dark for 7 days followed by incubation under daylight on the laboratory bench for 7 days. Under these conditions hypovirus-free C. parasitica isolates typically produce yellow-orange mycelium with abundant conidiation, while C. parasitica isolates infected with CHV-1 remain white with little or no conidiation (Bissegger et al., 1997). Isolates which could not be clearly assigned to a white or orange cultural type were classified as intermediate.
Vegetative compatibility (VC) test
Isolates were assessed for vegetative compatibility according to the merging/barrage response (Bissegger et al., 1997; Cortesi et al., 1998). Strains were paired approximately 3 mm apart and 5 mm from the edge of the Petri plate. European VC type testers from EU-1 to EU-31 were used (Cortesi et al., 1998). Six pairings were made per Petri plate (9 cm in diameter), with each plate containing 25 mL PDA. Petri plates were incubated at 25°C for 7 days in the dark, followed by 7 days in daylight at room temperature. Vegetative compatibility was scored after 7 and 14 days. Merging of the two colonies indicated a compatible pairing, i.e. isolates of the same VC type, whereas a barrage line between the colonies indicated an incompatible pairing (Cortesi et al., 1998). Each VC test for each isolate was performed in triplicate and only when replications showed identical results was the isolate assigned to the particular VC type.
Analysis of VC-type diversity
VC-type diversity was determined using the Shannon diversity index (H′) as described by Anagnostakis et al. (1986) and genotypic diversity index (G). H′ was calculated as –∑pi ln pi, in which pi was the frequency of the ith VC type, and G was calculated as 1/Σ, where pi was the frequency of the ith genotype. To estimate evenness (distribution of genotypes within the sample), the index E5 was used as recommended by Grünwald et al. (2003), where E5 = (G – 1)/(eH′ – 1). Since numbers of isolates in the populations were not constant, richness was compared using rarefaction analysis, which yields the number of VC types expected E(g33) in a sample corresponding to the smallest size n of all populations being compared (Grünwald et al., 2003). To compare VC-type diversity in Croatia with that of other European C. parasitica populations, data from Macedonia (Sotirovski et al., 2004), Italy (Cortesi et al., 1996), Switzerland (Bissegger et al., 1997; Milgroom & Cortesi 1999) and Portugal (Bragança et al., 2007) were used. Allelic diversity at six vic loci was calculated for the Croatian C. parasitica populations as described previously (Milgroom & Cortesi, 1999).
A subsample of 130 C. parasitica isolates was tested for the presence of dsRNA. Each isolate was grown on PDA with a cellophane overlay for 6 days at 25°C in the dark. The culture was stripped from the cellophane overlay, transferred to a 2-mL reaction tube and lyophilized. The dried mycelium was ground to a fine powder with a steel ball using a tissue lyser (Qiagen). DsRNA was extracted from 40 mg of lyophilized mycelium and purified from total nucleic acid by cellulose CF-11 chromatography as previously described (Allemann et al., 1999). Residual DNA was removed by treatment with RQ1 RNase-Free DNase (Promega). After 0·8% agarose gel electrophoresis the dsRNAs were stained with ethidium bromide, visualized under UV light and digitalized by Kodak EDAS 290. A lambda DNA/HindIII marker (Promega) was used as a molecular weight marker.
RT-PCR and RFLP analysis
dsRNAs extracted from 36 randomly chosen C. parasitica isolates sampled from 10 populations were subjected to RT-PCR/RFLP analysis (Požega: PzII19, PzII34, PzII47, PzII56; Samobor: Sm1, Sm14A, Sm16, Sm17; Cres: CR5, CR23, CR26; Istria: IS9, IS14; Ozalj: OS8, OS15, OS44; Markuševac 1: M7, M33, M35, M37/1, M46/1, M54, M56/1; Markuševac 2: MM22; Šamarica: S36, S53, SA18, SA43; Sljeme: SLJ14, SLJ41; Hrvatska Kostajnica: HK9, HK25, HK27, HK33, HK38, HK76). The first-strand complementary DNA (cDNA) was synthesized from 100 ng of dsRNA using random primers (Promega). PCR amplification of two hypoviral open reading frames (ORF-A and ORF-B) was performed using the primer set A (EP713-5 and R2280) and primer set B (EP713-6 and EP713-7) as described by Allemann et al. (1999). Each PCR product was digested separately with the restriction endonucleases BsuRI and HinfI (MBI Fermentas). The digest was carried out in a 20-µL reaction mixture containing 7·5 units of restriction enzyme, 10 µL of PCR product and 2 µL of 10 × enzyme buffer. The restriction fragments were separated in a 2% agarose gel and visualized as described above for dsRNA. Previously characterized CHV-1 isolates (Allemann et al., 1999; Sotirovski et al., 2006) were included in the RFLP analysis as references for the different CHV-1 subtypes. A 100-bp DNA Step Ladder (Promega) was used as a molecular weight marker.
After electrophoresis, the banding pattern of each viral isolate was scored for the presence (1) or absence (0) of each restriction fragment (Allemann et al., 1999). Binary data of RFLP analysis were used to construct phylogenetic trees with upgma cluster analysis (unweighted pair-group method with arithmetic averages) in the software program paup 4·0b10 (Swofford, 2003) using mean character difference among isolates as a distance measure. To assess the reliability of the tree obtained bootstrap analysis consisting of 1000 replicates was performed. CHV-1 isolates of known subtypes from other European countries were used as references to determine the genetic relatedness of the CHV-1 isolates from Croatia. These five main CHV-1 subtypes included: subtype I (represented by isolates from Macedonia: Sk32MK, J12MK, Os65MK, Vr42MK, Vr25MK, Vr4MK, Os44MK Vt28MK, J4MK, Po48MK, Po54MK; Italy: 33I, 32I, 38I, 40I, Euro7, 34I, 35I, 39I; Switzerland: 16CH, 5CH, 26CH; Bosnia-Herzegovina: 46BIH, 47BIH; Greece: 53GR; Croatia: 43HR), subtype F1 (France: EP713, 62F, 55F, 56F), subtype F2 (France: 57F), CHV-1 subtype D (Germany: 72D) and subtype E (Spain: 71E) (Allemann et al., 1999; Sotirovski et al., 2006).
Mating type analysis
Mating type was determined only for a subsample of 204 randomly chosen isolates. PCR amplification was conducted with the primers M1-GS1 and M1-GS2-rev for MAT1-1 and the primers M2-GS2 and InvA5n for MAT1-2 (Marra & Milgroom, 1999; McGuire et al., 2001). Amplification reactions were performed in a final volume of 50 µL containing 50–100 ng DNA, 2 mm MgCl2, 0·2 mm dNTPs, 0·2 µm primers, PCR buffer and 1 U Taq DNA polymerase. Cycling conditions were: an initial denaturation step at 95°C for 3 min, then 30 cycles of 0·5 min at 95°C, 1 min at 64°C and 4 min at 72°C, followed by a final extension at 72 C for 10 min. Amplified DNA fragments were separated by electrophoresis in 1% agarose gels at 90 V for 2 h using the 100-bp DNA Step Ladder (Promega) as a molecular size marker.
Diversity of VC types
A total of 18 VC types were identified among the 338 isolates (Table 1) sampled from 10 C. parasitica populations throughout the chestnut-growing area in Croatia (Fig. 1). The dominant VC type was EU-1, which comprised 42·9% of all the isolates and was found in all populations except Sljeme. VC type EU-2 was found in eight populations and represented 21% of all the isolates. VC types EU-12, EU-13 and EU-17 represented 14·2, 6·2 and 3·8% of the isolates, respectively. Nine other VC types occurred, with incidences between 1·8 and 0·6%. Based on vic genotypes, most of these VC types could have been generated by recombination of other VC types present in each population. Four VC types (EU-7, EU-18, EU-21 and EU-22) were found only once. Three of them, EU-18 (present in Istria), EU-21 (Samobor) and EU-22 (Samobor), could have been generated by recombination of EU-13 with EU-1 or EU-2. Two populations (Markuševac 2 and Sljeme) were represented by less than 10 isolates and were not considered further for diversity estimates. The number of VC types in the other populations varied between four and eight (Tables 1 and 2). Only one of these populations was clearly dominated by a single VC type, namely Ozalj, where EU-1 comprised 84% of the isolates. In most other populations different combinations of co-dominating VC types were observed. For example, EU-1 and EU-12 were the main VC types in Markuševac 1, EU-1 and EU-13 in Cres, and EU-1, EU-2, and EU-12 in Hrvatska Kostajnica and Požega.
Table 1. Vegetative compatibility types and mating types of Cryphonectria parasitica isolates from 10 populations in Croatia
The diversity of VC types, calculated with the Shannon diversity index (H′), ranged from 0·63 in Ozalj to 1·68 in Istria-Buje (Table 2). Diversity estimates for Croatian populations were lower than those in populations from Switzerland, but higher than those in populations from southern Italy, Macedonia and Portugal. Populations in Croatia were generally comparable in richness and diversity to the populations in northern Italy (Donnaz, Crevoladossola, Valtellina, Bergamo, Pigna and Corniglio; Table 2). The highest estimate of evenness was for population Požega, where the three most common VC types occurred at frequencies of 21·6, 41·2 and 31·4%, respectively, and the lowest estimate of evenness was for population Ozalj, where the VC type EU-1 occurred at a frequency of 83·7%.
Based on known vic genotypes, allelic diversity at six vic loci was calculated for eight populations composed of at least 14 isolates (Table 3). Seven of these populations had five polymorphic vic loci and one population (Ozalj) had four. One vic locus (vic 3) was not polymorphic in Croatia. Allelic diversity for the other five vic loci was relatively high, i.e. close to 0·5.
Table 3. Allelic diversity at six vic loci in Croatian populations of Cryphonectria parasitica
Allelic diversity was calculated only for populations with more than 10 isolates.
Maximum number of vegetative compatibility (VC) types assuming sexual recombination at all polymorphic vic loci.
Two cankers were sampled from 21 trees. In four cases two different VC types were found on the same tree.
Mating-type diversity and presence of perithecia
Perithecia were observed in all 10 populations assessed. The presence of perithecia was quantified for five populations with more than 25 bark samples available (Table 1). The incidence of cankers with perithecia in these populations ranged from 19·2 to 60·5%, with a weighted mean of 40·1%.
Both mating types of C. parasitica were found in all Croatian chestnut populations (Table 1). Using chi-squared statistics, mating-type ratios were tested for deviation from 1 : 1 in eight populations with large enough number of tested samples. The observed mating-type ratios were not significantly different from 1 : 1 in all these populations (Table 1).
Culture morphology and hypovirulence
Among the 338 C. parasitica isolates sampled throughout Croatia, 147 isolates had white, 99 had intermediate and 92 had orange culture morphology on PDA. The highest incidence of white isolates was found in Šamarica (66·6%) and Požega (60·7%) and the lowest in Istria-Buje (12·7%) (Table 4). In Sljeme, all four isolates collected had white culture morphology. White isolates were found in all common VC types (i.e. EU-1, EU-2, EU-12 and EU-13), with frequencies ranging from 38 to 65% (Table 1). The highest percentage of white isolates in EU-1 was detected in populations Markuševac 1, Ozalj, Hrvatska Kostajnica and Šamarica (56, 47·2, 58·8 and 62·5%, respectively); while a high percentage of white EU-2 isolates was recorded in populations Hrvatska Kostajnica, Šamarica and Požega (50, 77·7 and 80·9%, respectively). Hypovirulent isolates were also found among rare VC types (e.g. EU-3, EU-20 and EU-29) in Croatia.
Table 4. Culture morphology of Cryphonectria parasitica isolates and dsRNA and RT-PCR/RFLP analysis of hypovirus-(un)infected isolates
Out of 130 isolates (82 white, 26 intermediate and 22 orange) that were tested for the presence of dsRNA, 90 isolates (82 white and eight intermediate) contained L-dsRNA. Thirteen of these isolates also contained smaller M-dsRNA. All L-dsRNAs found in C. parasitica were approximately 12·7 kb in length. DsRNAs were not detected in cultures having orange morphology. Hypoviral dsRNAs were found in all 10 populations in Croatia (Table 4). upgma cluster analysis based on the RFLP banding patterns of amplified ORF-A and ORF-B, each digested with two restriction endonucleases, grouped all Croatian hypoviruses to the Italian subtype of CHV-1 (Fig. 2). Within this subtype five groups of Croatian CHV-1 isolates having identical restriction patterns in both ORFs were formed. The largest group contained 18 isolates and all of these isolates shared identical patterns with reference isolates from Macedonia (Sk32 MK), Greece (53GR), Bosnia-Herzegovina (46BIH) and Italy (33I). This group also contained a Croatian isolate sampled 22 years ago (44HR). The other four groups comprised two to five viral isolates. Isolates collected in the same location often had identical RFLP patterns. Six isolates (PzII34, M35, M37 1, S36, OS8 and HK76) showed unique RFLP patterns.
The overall C. parasitica population in Croatia is characterized by a relatively high VC type diversity. The dominant VC type in the Croatian populations was EU-1 (42·9%), followed by EU-2 (21%) and EU-12 (14·2%). The same three VC types were also dominant in northwestern Bosnia-Herzegovina, which borders Croatia in the south (Trestićet al., 2001). EU-1 and EU-2 are dominant in northern Italy, southern France, Switzerland and eastern Spain (Robin & Heiniger, 2001), whereas EU-12 is the dominant VC type in southern Italy, Greece, Macedonia, and other Eastern European countries (Cortesi et al., 1996; Radócz, 2001; Robin & Heiniger, 2001; Sotirovski et al., 2004, Perlerou & Diamandis, 2006). The fourth most frequent VC type in Croatia, EU-13, was also commonly found in northwestern Bosnia-Herzegovina (Trestićet al., 2001), Slovakia (Adamčíkováet al., 2006), Hungary (Radócz, 2001) and in some populations in northern Italy, Switzerland and the Czech Republic (Cortesi et al., 1998; Haltofová, 2006). This VC type is relatively rare in other European countries. The other VC types found in Croatia occur with much lower frequencies and were generally also rare in other regions in Europe. In respect to VC types, the C. parasitica populations in Croatia combine features of both northwestern and southeastern European populations. It is likely that EU-1 and EU-2 migrated into Croatia from northern Italy, where chestnut blight was first detected in Europe. Likewise, EU-12 might have migrated from the southern Balkans via Bosnia-Herzegovina into Croatia.
The number of VC types and the Shannon diversity index (H′) was higher in the Croatian populations than in the populations in southern Italy (H′ = 0·4–0·8) (Cortesi et al., 1996) and Macedonia (H′ = 0·1–0·7) (Sotirovski et al., 2004). In contrast, H′ values of Croatian populations were lower than those of populations in southern Switzerland that comprised 14 and 16 VC types with H′ values from 1·9 to 2·1 (Bissegger et al., 1997). The diversity of VC types in Croatia could be compared with VC type diversity observed in populations in northern Italy where 4–16 VC types were found and H′ diversity ranged from 0·9 to 1·7 (Cortesi et al., 1996). High VC-type diversity has been reported from areas where the disease has been present for longer periods, such as southwestern France (Dordogne), Italy, and Switzerland south of the Alps, and also where sexual reproduction of C. parasitica is frequent (Cortesi et al., 1996; Bissegger et al., 1997; Robin et al., 2000). On the other hand, low VC-type diversity was reported in Hungary (Radócz, 2001), Slovakia (Juhásováet al., 2005), Switzerland north of the Alps (Hoegger et al., 2000), Germany (Seemann et al., 2001), Turkey (Çeliker & Onogur, 2001; Gurer et al., 2001), Macedonia (Sotirovski et al., 2004), Portugal (Bragança et al., 2007) and northwestern Spain (Aguin et al., 2005), where C. parasitica has been present for relatively short period. In Bosnia-Herzegovina, low VC-type diversity was found in the eastern, central and southwestern regions, but a high diversity was found in the northwestern region, with a Shannon index of 2·26 (Trestićet al., 2001). In Croatia, there is no obvious geographic pattern of VC-type diversity. Populations with comparable VC type diversity were found in different chestnut-growing regions. The VC type diversity in populations of C. parasitica in Croatia could increase over time, since the presence of perithecia and of both mating types in similar proportions indicate potential for sexual reproduction.
Since the genetic basis for vegetative incompatibility is known for nearly all VC types in Europe (Cortesi & Milgroom, 1998; Robin et al., 2000), it is possible to determine the potential increase in number of VC types according to the current situation. Through sexual recombination at all five polymorphic vic loci, the number of VC types could increase to 25 = 32 VC types in most Croatian C. parasitica populations. An increase of VC-type diversity could be theoretically expected in those populations where EU-1 co-dominantly occurs with EU-12 or EU-13 (e.g. in Markuševac 1, Hrvatska Kostajnica and Cres). EU-1 differs by four vic loci from EU-12 and EU-13; consequently, sexual recombination would generate 16 VC types. Recombinant VC types were only found at low frequency, suggesting that to date, sexual recombination between the main VC types has rarely occurred. Low levels of recombination between two main VC types of C. parasitica was also found in Teano, Italy, and was attributed to a possible spatial isolation of the two VC types (Liu et al., 1996).
Although a relatively high diversity of VC types was observed in Croatia, the hypovirus was widespread in the C. parasitica populations. White isolates were found in all main VC types and in all populations sampled, with percentages varying from 12·7 (Istria-Buje) to 66·6% (Šamarica). This variation was not obviously associated with VC-type diversity. The percentage of white isolates, however, did not always represent the actual incidence of CHV-1, because hypovirus-infected isolates were also found in some isolates with intermediate culture morphology (e.g. in Cres). Other factors, such as differences among populations in dispersal pattern of VC types or temporal introduction of CHV-1, might also explain the lack of association between VC-type diversity and the incidence of white isolates. Nevertheless, the VC-type diversity in all Croatian populations was lower than the theoretical threshold of approx. 2·0 (Shannon index), above which hypoviruses are not expected to become established (Milgroom & Cortesi, 2004).
A low incidence of callusing (i. e. hypovirus-infected) cankers in the Istria region was previously observed by Novak-Agbaba et al. (2000). The virus seemed to spread quite slowly through the C. parasitica populations in this area and it contributed only a little to biological control by natural hypovirulence. Therefore, it could be useful to treat cankers with hypovirulent isolates to promote hypovirulence. In the survey of Novak-Agbaba et al. (2000), the highest percentage of callusing cankers (20%) was found in the Sisak region, where the Šamarica and Hrvatska Kostajnica populations are situated. In accordance with this finding, high incidences of white C. parasitica isolates (50 and 66%) were found in these two populations in the present study, indicating good natural spread of the virus. Comparable incidences of hypovirus-infected isolates were previously observed in areas where hypovirulence was naturally established (Bissegger et al., 1997) and in areas with low VC-type diversity (Sotirovski et al., 2006). All identified Croatian hypovirus isolates belonged to the Italian subtype of CHV-1. This subtype is widespread in southern and southeastern Europe (Allemann et al., 1999). A large number of Croatian CHV-1 isolates had RFLPs identical to isolates from Bosnia-Herzegovina, Macedonia, Greece and Italy. This finding supports the hypothesis that this virus haplotype has migrated throughout southern Europe (Sotirovski et al., 2006). One hypovirus sampled in Croatia in the 1980s also had this haplotype, suggesting that the evolution of the hypovirus is quite slow. Nevertheless, the occurrence of other viral RFLP types in Croatia might be the result of mutations in resident hypoviruses. Alternatively, new virus haplotypes may have migrated into Croatia from other areas. Chestnut forests where hypovirulence has established are characterized by high disease incidence, but low disease severity (Bissegger et al., 1997). The high incidence of hypovirus-infected isolates of C. parasitica in Croatia could have a positive impact on biological control of the disease and recovery of chestnut stands. Sexual reproduction by the fungus, however, and the potential for an increase of VC-type diversity by recombination could create an obstacle for biological control. Therefore, periodical monitoring of VC-type diversity and hypovirus spread is required.
We would like to thank Paolo Cortesi and Michael Milgroom for providing EU testers (EU reference strains) from EU-1 to EU-64. This research was supported by the Swiss National Science Foundation (SCOPES project IB73A0-111089), the company Tanin Sisak and the Ministry of Science, Education and Sport of the Republic of Croatia (project no. 119-1191192-1215).