Distribution and Characterization of Armillaria Complex in Atlantic Forest Ecosystems of Spain

Armillaria root disease is a significant forest health concern in the Atlantic forest ecosystems in Spain. The damage occurs in conifers and hardwoods, causing especially high mortality in young trees in both native forests and plantations. In the present study, the distribution of Armillaria root disease in the forests and plantations of the Basque Country is reported. Armillaria spp. were more frequently isolated from stands with slopes of 20–30% and west orientation, acid soils with high permeability, deciduous hosts, and a rainfall average above 1800 mm. In a large-scale survey, 35% of the stands presented Armillaria structures and showed disease symptoms. Of the isolated Armillaria samples, 60% were identified using molecular methods as A. ostoyae, 24% as A. mellea, 14% as A. gallica, 1% as A. tabescens, and 1% as A. cepistipes. In a small scale sampling, population diversity was defined by somatic compatibility tests and Universally Primed-PCR technique. Finally, the pathogenicity of A. mellea, the species with the broadest host range, was determined on different tree species present in the Atlantic area of Spain in order to determine their resistance levels to Armillaria disease. A significant difference in disease severity was observed among tree species (p < 0.001), with Pinus radiata being the most susceptible tree species and Cryptomeria japonica the most resistant to A. mellea.


Introduction
About 40 species of Armillaria are known worldwide [1].Seven are present in Europe: A. mellea (Vahl) P.Kumm., A. gallica Marxm.& Romagn., A. ostoyae (Romagn.)Herink, A. tabescens (Scop.)Emel, A. cepistipes Velen., A. borealis Marxm.& Korhonen, and A. ectypa (Fr.)Lamoure [2,3].Many of these are pathogens that cause root and butt rot disease on a broad range of trees, shrubs and some herbaceous plants [4].This disease is characterized by some general symptoms such as chlorotic leaves, progressive thinning of the crown, slower leader growth, and excess cone production.Evidence of infection may also include subcortical white mycelial fans, clusters of golden-brownish mushrooms near the tree base, rhizomorphs, rotten stringy-yellow wood with black lines (pseudosclerotia), rapid tree death without the loss of foliage, and/or basal resin or gum exudates [5,6].
In general, conifers seem to be more susceptible to infection than hardwoods [7], although susceptibility depends on the specific Armillaria species involved [2,8].The extent of damage caused by Armillaria is variable and is determined by factors such as the fungal species, the vigor of the host, interaction with other diseases, soil properties, climate, plantation management, and previous land uses, among others [9][10][11][12][13][14].Therefore, forest susceptibility cannot be generalized [15].The composition of the stand can also influence the range of infection; lower density of susceptible tree species and higher species diversity in different forest strata reduce the possibility of disease transmission [16,17].Management procedures such as selective logging, early thinning, and the continued planting of susceptible or moderately susceptible species that are not very well adapted to the location increase the inoculum sources and the potential for spread of infection [14,[18][19][20].Thus, in comparison with natural forests the damage in exotic tree plantations is usually greater [8].
In field conditions, Armillaria can colonize different hosts by direct contact between an infected source and roots by way of hyphae, or by advancing through the ground from an infection point by way of rhizomorphs [21].This is a short distance spreading mechanism but it is considered the most important even if, for some Armillaria species, formation of rhizomorphs in the field is not common [3,15].The capacity to create new infection points by releasing basidiospores varies from one species to another [22] but, in general, this seems to be the least frequent source of infection [23].
The genotypic diversity within a population is an important parameter for the determination of the epidemiology of a disease.Many different small somatic compatibility (SC) groups in a stand imply that spread is predominantly by basidiospores, and a single extensive SC group implies spread by vegetative mycelium.By means of SC tests, Armillaria population structure in the field can be determined [24].Even though SC tests are usually reliable, sometimes they do not differentiate among closely related individuals of Armillaria [2].Molecular techniques such as Universally primed-PCR (UP-PCR) [25] could provide more information in these cases.UP-PCR has been used for the characterization of fungal populations at interspecies and/or intraspecies level [26][27][28].In this technique, the entire genome of an organism is targeted with a single primer or a combination of primers that will anneal to multiple regions resulting in a multiband profile which differs among different genotypes.Universal primers consist of a stable minisatellite-like region, with high GC content, designed using as a template genome regions of different organisms, which allows primer annealing at high temperatures resulting in reproducible PCR product patterns, and a variable, randomly designed region.Highly variable intergenic regions are usually targeted which enables differentiation between closely related isolates [29].
Armillaria spp., including the primary parasites A. ostoyae and A. mellea, have been previously reported in Northern Spain [30][31][32] including the Basque Country [33].The main tree species in Basque Country plantations is Pinus radiata D. Don.(covering an area of 132,084 ha), followed by Eucalyptus spp.(15,197  Liebl.association (16469 ha), and Q. pyrenaica Willd.(13,039 ha) [34].All of these tree species have been reported to be susceptible to different Armillaria spp. in the literature, however, Armillaria spp.distribution, species diversity, and dispersal mechanisms are not known in the Basque Country, where the wood industry is valued at 1 billion euros per year [35].Thus, the objectives of this study were to determine: (i) the Armillaria species distribution and population diversity in this region; (ii) factors affecting Armillaria spp.Distribution; and (iii) host susceptibility.This will contribute to a better understanding of the impact of Armillaria spp.and establishment of management practices.

Collection of Fungi
Fungal samples and data related to forest characteristics were collected from native forests and plantations of the Basque Country, northern Spain, focusing on trees in pockets of mortality and decayed trees that displayed symptoms resembling those of root rot diseases.For the first sample set (Set 1), the stands were surveyed by systematic random sampling in which samples were collected systematically from randomly chosen focal points [36].A total of 709 foci of tree mortality were examined for the presence of Armillaria spp., and fungal samples and ecosystem characteristics related to infection were collected.The second sampling (Set 2) was carried out to determine the genetic

Description of Armillaria-Infected Ecosystems
To determine environmental factors of Armillaria spp.habitats, a dataset of the environmental variables of the studied ecosystems was constructed based on information supplied by the Environment Department of the Basque Government (http://www.ingurumena.ejgv.euskadi.eus/r49-579/es//publicaciones_c.htm).The variables compiled were stand slope, stand orientation, soil pH, soil permeability, average rainfall, average temperature, tree types, and host optimal conditions.Variables were categorized as shown below.Armillaria spp.presence was coded as a binomial variable being 0 for absence and 1 for presence.
As a preliminary exploratory analysis, multiple correspondence analysis (MCA) was applied to represent the relationships among the categorical variables, including Armillaria spp.presence.MCA projects the variables in a reduced space, facilitating visual interpretation of large datasets.This analysis converts a matrix of data into a graphical display known as factor planes.The rows and columns of the matrix are plotted as points in the factor planes and allow a geometrical representation of the information [39].
This procedure was complemented with contingency tables testing separately each categorical variable including its categories against Armillaria presence or absence.Pearson's chi square test was used to determine the independence between row and column variables, i.e., to determine if Armillaria spp.were more frequently detected than expected by chance in certain categories.For calculating the strength of association between categorical variables Cramer's V was used; Cramer's V ranges between 0 (no relationship) and 1 (perfect relationship).Adjusted standardized residuals were checked in order to determine the significant differences between categories; adjusted residuals Samples of Armillaria spp.basidiocarps, rhizomorphs and mycelia were collected and placed in separate polyethene bags, transported to the laboratory, and stored at 4 • C. Fungi were cultured on benomyl-dichloran-streptomycin agar (BDS) [37] and grown at 20 • C in the dark.Pure cultures were obtained, and routinely grown on malt extract agar (MEA) (PanReac Química, Barcelona, Spain).For preservation of the pure cultures, mycelial fragments were placed in 50% glycerol and, after incubating at 4 • C for 24 h, maintained at −20 • C [38].

Description of Armillaria-Infected Ecosystems
To determine environmental factors of Armillaria spp.habitats, a dataset of the environmental variables of the studied ecosystems was constructed based on information supplied by the Environment Department of the Basque Government (http://www.ingurumena.ejgv.euskadi.eus/r49-579/es//publicaciones_c.htm).The variables compiled were stand slope, stand orientation, soil pH, soil permeability, average rainfall, average temperature, tree types, and host optimal conditions.Variables were categorized as shown below.Armillaria spp.presence was coded as a binomial variable being 0 for absence and 1 for presence.
As a preliminary exploratory analysis, multiple correspondence analysis (MCA) was applied to represent the relationships among the categorical variables, including Armillaria spp.presence.MCA projects the variables in a reduced space, facilitating visual interpretation of large datasets.This analysis converts a matrix of data into a graphical display known as factor planes.The rows and columns of the matrix are plotted as points in the factor planes and allow a geometrical representation of the information [39].
This procedure was complemented with contingency tables testing separately each categorical variable including its categories against Armillaria presence or absence.Pearson's chi square test was used to determine the independence between row and column variables, i.e., to determine if Armillaria spp.were more frequently detected than expected by chance in certain categories.For calculating the strength of association between categorical variables Cramer's V was used; Cramer's V ranges between Forests 2017, 8, 235 4 of 18 0 (no relationship) and 1 (perfect relationship).Adjusted standardized residuals were checked in order to determine the significant differences between categories; adjusted residuals are standardized values allowing comparisons among different cells, and follow a standard normal frequency (with mean zero and standard deviation one) so it can be assumed that if their value lies outside of ±1.96 then it is significant at p < 0.05, if it lies outside ±2.58 then it is significant at p < 0.01 and if it lies outside ±3.29 then it is significant at p < 0.001 [40].All of the analyses were carried out in SPSS version 15.0 (SPSS Inc., Chicago, IL, USA).

Fungal Species Identification by RFLP-PCR
Restriction Fragment Length Polymorphism (RFLP)-PCR was used to confirm the species identity of the Armillaria isolates in sample sets 1 and 2 at the species level.DNA from two-week-old pure cultures was extracted with DNeasy Plant Mini Kit (QIAGEN Gmb, Hilden, Germany) in accordance with the manufacturer's instructions.A region of the intergenic spacer (IGS) of the rDNA was amplified using the primer pair LR12R (5 -CTGAACGCCTCTAAGTCAGAA-3 ) and O-1 (5 -AGTCCTATGGCCGTGGAT-3 ) [41].The PCR mixture contained 1.5 mM MgCl 2 , 200 µM each dNTP, 0.5 µM each primer, 2 U Taq polymerase (Netzyme, Molecular Netline Bioproducts, Valencia, Spain) and 1 µL template DNA in a final volume of 50 µL.The cycling conditions consisted of 90 s at 95 • C, 35 cycles of 30 s at 95 • C, 40 s of annealing at 60 • C, and 2 min at 72 • C, and a final 10 min at 72 • C. The obtained DNA fragment was directly digested with the restriction enzyme Alu I (Invitrogen Life Technologies, Carlsbad, CA, USA), Nde I (Takara Bio Inc., Kusatsu, Japan) or Bsm I (Hoffmann-La Roche Ltd., Basel, Switzerland) [34].The restriction fragments were separated in 3 % agarose gels (agarose D-1 Low EEO, Conda, Madrid, Spain).Species were identified based on the restriction patterns determined by Harrington and Wingfield [41] and Sierra et al. [42].

Fungal Population Analysis
In order to determine mechanisms of dispersal of Armillaria spp. at the stand level, genetic diversity and population structure patterns of the Armillaria isolates from Set 2 were analyzed.Armillaria species were determined by RFLP-PCR (as described above) and intraspecies differentiation was determined by SC tests and UP-PCR.For determination of SC groups (SCGs), diploid isolates from the same sampling area were paired in all possible combinations.Approximately 4 mm 2 of mycelia were placed 0.5 cm apart on MEA plates, and incubated at 20 • C for six weeks.When mycelia of co-cultured isolates fused and grew with a uniform morphology, they were considered to belong to the same species and genet, and the pairings were considered somatic compatible.When a line of demarcation was apparent, isolates were considered somatic incompatible [43].
UP-PCR reactions were carried out following the procedure described by Tyson et al. [44], in 25 µL volumes containing 2 mM MgCl 2 , 0.2 mM each dNTP, 0.8 µM primer (Table 1), 50 ng genomic DNA (extracted as for RFLP analysis), and 1.25 U Taq DNA Polymerase (Invitrogen Life Technologies, Carlsbad, CA, USA).The cycling conditions were 5 min at 94 • C, 5 cycles of 50 s at 94 • C, 2 min at primer specific annealing temperature (Table 1), and then 1 min at 72 • C, followed by 30 cycles of 50 s at 94 • C, 90 s at primer specific annealing temperature, and 1 min at 72 • C, and a final extension at 72 • C for 7 min [44].UP-PCR primers were tested on a representative group of Armillaria isolates consisting of different species and genets of Armillaria (as determined by RFLP-PCR and mating tests), and the primers with the best capacity to distinguish between different SCGs and species were chosen for the analysis of all the isolates.Following gel electrophoresis of the UP-PCR amplicons, the band pattern for each isolate was assessed for the presence (1) or absence (0) of each band and represented in a binomial matrix.Similarities between strains were calculated using a simple matching coefficient [45] and represented on dendrograms constructed by average linkage between groups in Forests 2017, 8, 235 5 of 18 SPSS software (SPSS Inc., Chicago, IL, USA).To determine the consistency between the similarity matrix and dendrograms, the cophenetic correlation coefficient was calculated [46].Table 1.Oligonucleotide sequence of the UP-PCR primers and their respective annealing temperatures [44].

Host Resistance
To assess the susceptibility to A. mellea of different tree species present in the Basque Country, 2-year-old trees of different species, including P. radiata, P. nigra subsp.salzmannii var.corsicana, P. sylvestris L., Fagus sylvatica, Prunus avium, Q. petraea, Q. ilex, Cryptomeria japonica (Thunb.ex L.f.) D.Don, Q. robur, Sequoiadendron giganteum (Lindl.)J.Buchholz, and Eucalyptus nitens H.Deane & Maiden (Explotaciones Forestales Jiménez Araba s. l.Nursery, Vitoria, Spain), were infected with the A. mellea strain AMVac isolated from an Acer pseudoplatanus L. (the fungal strain is maintained in a collection at Neiker Tecnalia, Arkaute, Spain).For the preparation of A. mellea inoculum, pieces of fungal mycelia were placed on BDS agar with autoclaved Quercus spp.acorns and incubated for approximately one month at room temperature in the dark [53].Fifty trees of each species were grown in 53 × 53 × 180 mm pots (300 cc volume) using a mix of peat moss (2/3 peat, 1/3 perlite and fertilizer NPK; N = 200-450 mg/L, P 2 O 5 = 200-500 mg/L, K 2 O = 300-550 mg/L) and, after an adaptation period of two weeks, half acorns infected with A. mellea mycelium was placed in contact with tree roots.The seedlings were maintained for 4 months in a biosafety level 2 greenhouse at a mean temperature of 18 ± 5 • C, with a relative humidity of 55-60% and without supplemental light.After this period, roots were cleaned with tap water and lengths of stems, main roots and secondary roots were measured.A. mellea mycelial colonization was determined after removing the bark from the stem and roots.Plants were scored as healthy (without symptoms of infection) or with lesions (when A. mellea mycelium was present under the bark); the length of the lesions was determined by removing the bark of stem and roots and measuring the extent of Armillaria damage with an electronic caliper.
The differences in Armillaria disease severity among different tree species was determined by Pearson's chi-square.The strength of association between categorical variables (tree species and healthy or lesion containing trees) was measured with Cramer's V; adjusted standardized residuals were checked in order to determine the significant differences between categories.Differences in the size of the fungal lesions among tree species was analyzed by Brown-Forsythe and Welch statistics, and Games-Howell test was chosen for the post hoc analysis.The data was not normally distributed and therefore a base10-log transformation was applied.

Determination of Ecosystem Characteristics in Which Armillaria spp. Were Detected
Environmental characteristics were associated with each of the surveyed points and their relationships with the presence of Armillaria spp. was assessed (Table 2).The spread of the category variables for all characteristics is represented in a MCA that reflects the relationships among the variables in each dimension.MCA revealed that the first horizontal dimension explained 23.7% of the total inertia (variance), as the first factor plane represents the largest inertia, while the second vertical dimension explained 22.7%.A measure of the importance of each variable (squared component loading) is computed for each dimension.This measure is also the variance of the quantified variable in that dimension.The variables with higher variance in the first dimension were average rainfall (Rain), optimal conditions for host growth (Hoc), slope, Armillaria spp., soil permeability, and stand orientation (Figure 3).The variables with higher variance in the second dimension were orientation, average temperature (Temperature), average rainfall (Rain), slope, and soil acidity (Soil) (Figure 3).Armillaria spp.detection was related to categories such as west, northwest, and northeast stand orientation, slopes between 20% and 50%, basic soils, high average rainfalls (>1800 mm), and high soil permeability (Figure 4).Armillaria spp.absence was related to This image cannot currently be display ed.Of the total of isolates obtained from the surveyed plots (Set 1), 60% were identified by RFLP-PCR patterns as A. ostoyae, 24% as A. mellea, 14% as A. gallica, 1% as A. tabescens and 1% as A. cepistipes. A. ostoyae was detected mainly in Pinus spp.(P.radiata, P. nigra and P. pinaster).The host range for A. mellea was more varied.A. mellea pattern 1 (PCR fragment sizes: 320 and 155 bp) was found on P. radiata, Quercus spp., F. excelsior, and C. lawsoniana, and corresponded to 53% of the A. mellea isolates, while the remaining 47% were identified as pattern 2 (fragment sizes: 320, 180, and 155 bp) and appeared on Q. pyrenaica and P. radiata. A. gallica was found on A. glutinosa, P. radiata, and Q. robur, A. cepistipes was detected on P. radiata, and A. tabescens on Q. robur.

Determination of Ecosystem Characteristics in Which Armillaria spp. Were Detected
Environmental characteristics were associated with each of the surveyed points and their relationships with the presence of Armillaria spp. was assessed (Table 2).The spread of the category variables for all characteristics is represented in a MCA that reflects the relationships among the variables in each dimension.MCA revealed that the first horizontal dimension explained 23.7% of the total inertia (variance), as the first factor plane represents the largest inertia, while the second vertical dimension explained 22.7%.A measure of the importance of each variable (squared component loading) is computed for each dimension.This measure is also the variance of the quantified variable in that dimension.The variables with higher variance in the first dimension were average rainfall (Rain), optimal conditions for host growth (Hoc), slope, Armillaria spp., soil permeability, and stand orientation (Figure 3).The variables with higher variance in the second dimension were orientation, average temperature (Temperature), average rainfall (Rain), slope, and soil acidity (Soil) (Figure 3).Armillaria spp.detection was related to categories such as west, northwest, and northeast stand orientation, slopes between 20% and 50%, basic soils, high average rainfalls (>1800 mm), and high soil permeability (Figure 4).Armillaria spp.absence was related to categories such as south and southeast stand orientation, slopes less than 20% and soils with medium permeability (Figure 4). Figure 3. Measure of the variance of each variable for each dimension.The variables with higher variance in the first dimension were mainly average rainfall (Rain), optimal conditions for host growth (Hoc), slope, Armillaria spp., soil permeability and stand orientation (Figure 3).The variables with higher variance in the second dimension were orientation, average temperature (Temperature), average rainfall (Rain), slope, and soil acidity (Soil).The variables with higher variance in the first dimension were mainly average rainfall (Rain), optimal conditions for host growth (Hoc), slope, Armillaria spp., soil permeability and stand orientation (Figure 3).The variables with higher variance in the second dimension were orientation, average temperature (Temperature), average rainfall (Rain), slope, and soil acidity (Soil).Measure of the variance of each variable for each dimension.The variables with higher variance in the first dimension were mainly average rainfall (Rain), optimal conditions for host growth (Hoc), slope, Armillaria spp., soil permeability and stand orientation (Figure 3).The variables with higher variance in the second dimension were orientation, average temperature (Temperature), average rainfall (Rain), slope, and soil acidity (Soil).The significance of the associations among fungal presence and the environmental variables was determined using Pearson's chi square test and the strength of the association was determined by Cramer's V. A significant association was observed between Armillaria spp.and slope, tree type, soil acidity, soil permeability, and rainfall average (Table 3).When the adjusted standardized residuals were examined, Armillaria spp.were significantly present in stands with slopes of 20-30% (z = 7.2; p < 0.001); stands with western orientation (z = 6.7; p < 0.001); deciduous hosts (z = 3.7; p < 0.001); acid soils (z = 4.7; p < 0.001); high permeability soils (z = 4.3; p < 0.001), and rainfall average (mm) >1800 (z = 4.6; p < 0.001) (Figure 5).Armillaria spp.were significantly absent in stands with slopes <10% (z = −7.1;p < 0.001); stands with southwestern orientation (z = −5.6;p < 0.001); coniferous hosts (z = −3.7;p < 0.001); moderately acid soils (z = −4.1;p < 0.001); medium permeability soils (z = −2.5;p < 0.01), impermeable soils (z = −3.6;p < 0.01), and rainfall average (mm) <1000 (z = −2.5;p < 0.05) (Figure 5).The significance of the associations among fungal presence and the environmental variables was determined using Pearson's chi square test and the strength of the association was determined by Cramer's V. A significant association was observed between Armillaria spp.and slope, tree type, soil acidity, soil permeability, and rainfall average (Table 3).When the adjusted standardized residuals were examined, Armillaria spp.were significantly present in stands with slopes of 20-30% (z = 7.2; p < 0.001); stands with western orientation (z = 6.7; p < 0.001); deciduous hosts (z = 3.7; p < 0.001); acid soils (z = 4.7; p < 0.001); high permeability soils (z = 4.3; p < 0.001), and rainfall average (mm) >1800 (z = 4.6; p < 0.001) (Figure 5).Armillaria spp.were significantly absent in stands with slopes <10% (z = −7.1;p < 0.001); stands with southwestern orientation (z = −5.6;p < 0.001); coniferous hosts (z = −3.7;p < 0.001); moderately acid soils (z = −4.1;p < 0.001); medium permeability soils (z = −2.5;p < 0.01), impermeable soils (z = −3.6;p < 0.01), and rainfall average (mm) <1000 (z = −2.5;p < 0.05) (Figure 5)..96then it is significant at p < 0.05, if y = ±2.58then it is significant at p < 0.01, and if y = ±3.29 then it is significant at p < 0.001 (Field, 2009).

Fungal Population Analysis by RFLP, SI and UP-PCR
Analysis of fungal samples (Set 2) for determining the mechanism of Armillaria spp.dispersal in three specific plots in Otxandiano, Amunategi, and Altube (Figure 1) revealed three different Armillaria species, A. mellea, A. ostoyae and A. gallica.The 21 samples collected in the stand located in Otxandiano belonged to the same SCG and were identified as A. ostoyae (data not shown).They were found in Q. robur stumps and trees, C. lawsoniana, Crataegus monogyna Jacq., and grassland.In the stand located in Amunategi, four of the 19 isolates were classified as A. mellea RFLP pattern 2 that belonged to two SCGs, both present in R. pseudoacacia, and 15 isolates as A. gallica in a more complex .96then it is significant at p < 0.05, if y = ±2.58then it is significant at p < 0.01, and if y = ±3.29 then it is significant at p < 0.001 (Field, 2009).

Fungal Population Analysis by RFLP, SI and UP-PCR
Analysis of fungal samples (Set 2) for determining the mechanism of Armillaria spp.dispersal in three specific plots in Otxandiano, Amunategi, and Altube (Figure 1) revealed three different Armillaria species, A. mellea, A. ostoyae and A. gallica.The 21 samples collected in the stand located in Otxandiano belonged to the same SCG and were identified as A. ostoyae (data not shown).They were found in Q. robur stumps and trees, C. lawsoniana, Crataegus monogyna Jacq., and grassland.In the stand located in Amunategi, four of the 19 isolates were classified as A. mellea RFLP pattern 2 that belonged to two SCGs, both present in R. pseudoacacia, and 15 isolates as A. gallica in a more complex population structure located in R. pseudoacacia, Salix alba L. and stumps of deciduous trees (Figure 6).In the stand located in Altube, nine of the 17 samples were identified as A. ostoyae, separated in 3 SCGs, in F. sylvatica and Q. robur, seven as A. mellea pattern 2, separated in 3 SCGs, in F. sylvatica and C. monogyna, and one as A. mellea pattern 1 in F. sylvatica (Figure 6).The larger size of A. ostoyae SCGs indicates dispersal predominantly by vegetative mycelium.In contrast, the smaller SCGs obtained for A. mellea and A. gallica indicate dispersal by basidiospores and vegetative mycelium.The location in the stands from which the samples were collected, the groups obtained by SC tests and their extension are depicted in Figure 7.
Forests 2017, 8, 235 10 of 18 population structure located in R. pseudoacacia, Salix alba L. and stumps of deciduous trees (Figure 6).In the stand located in Altube, nine of the 17 samples were identified as A. ostoyae, separated in 3 SCGs, in F. sylvatica and Q. robur, seven as A. mellea pattern 2, separated in 3 SCGs, in F. sylvatica and C. monogyna, and one as A. mellea pattern 1 in F. sylvatica (Figure 6).The larger size of A. ostoyae SCGs indicates dispersal predominantly by vegetative mycelium.In contrast, the smaller SCGs obtained for A. mellea and A. gallica indicate dispersal by basidiospores and vegetative mycelium.The location in the stands from which the samples were collected, the groups obtained by SC tests and their extension are depicted in Figure 7.   Forests 2017, 8, 235 10 of 18 population structure located in R. pseudoacacia, Salix alba L. and stumps of deciduous trees (Figure 6).In the stand located in Altube, nine of the 17 samples were identified as A. ostoyae, separated in 3 SCGs, in F. sylvatica and Q. robur, seven as A. mellea pattern 2, separated in 3 SCGs, in F. sylvatica and C. monogyna, and one as A. mellea pattern 1 in F. sylvatica (Figure 6).The larger size of A. ostoyae SCGs indicates dispersal predominantly by vegetative mycelium.In contrast, the smaller SCGs obtained for A. mellea and A. gallica indicate dispersal by basidiospores and vegetative mycelium.The location in the stands from which the samples were collected, the groups obtained by SC tests and their extension are depicted in Figure 7.    UP-PCR primer AS4 showed good ability to distinguish the fungal strains at the interspecies and intraspecies levels.Although primer L15/AS19 showed good discrimination in the initial screen, it did not yield specific banding profiles when all the samples where tested (data not shown).The best differentiation patterns were obtained for A. ostoyae strains and in general the clusters were comparable to those generated from mycelia matings (Figure 8).The cophenetic correlation coefficient between the similarity matrix and the dendrogram was 0.886, meaning that the clustering had a good fit.UP-PCR primer AS4 showed good ability to distinguish the fungal strains at the interspecies and intraspecies levels.Although primer L15/AS19 showed good discrimination in the initial screen, it did not yield specific banding profiles when all the samples where tested (data not shown).The best differentiation patterns were obtained for A. ostoyae strains and in general the clusters were comparable to those generated from mycelia matings (Figure 8).The cophenetic correlation coefficient between the similarity matrix and the dendrogram was 0.886, meaning that the clustering had a good fit.

Discussion
In the present study, we report the broad distribution of Armillaria spp. in native forests and plantations of the Basque Country and the ecosystem characteristics that may foster the occurrence of this fungal complex.The diversity of the Armillaria population in three of the infected plots was assessed in order to determine their interspecies and intraspecies diversity and potential dispersal mechanisms.Finally, host susceptibility to A. mellea was determined in a set of native and exotic forest species selected on the basis of their presence in the Atlantic area of Spain.
In the large and small scale surveys, A. ostoyae was the predominant species, mainly detected in conifers, but also in native forests of F. sylvatica and Q. robur.The wide host range displayed by A. mellea in this study has been thoroughly referenced [2,8], and A. gallica was also found in both conifers and deciduous trees.
In the stands infected with Armillaria spp., A. ostoyae was distributed in larger clonal clusters

Discussion
In the present study, we report the broad distribution of Armillaria spp. in native forests and plantations of the Basque Country and the ecosystem characteristics that may foster the occurrence of this fungal complex.The diversity of the Armillaria population in three of the infected plots was assessed in order to determine their interspecies and intraspecies diversity and potential dispersal mechanisms.Finally, host susceptibility to A. mellea was determined in a set of native and exotic forest species selected on the basis of their presence in the Atlantic area of Spain.
While Armillaria was found to be associated with mortality in the forests of the Basque Country, the impact of the fungus can be difficult to evaluate because it is influenced by other biotic and abiotic elements of the ecosystem.Armillaria spp.can cause mortality in susceptible tree species, however, sometimes their effects are less visible and infection does not always have a negative impact on tree development.The outcome of infection is also influenced by forest management strategies and this study is a first attempt to define factors that contribute to the development of Armillaria disease in the area, bearing in mind the complex etiology of this pathosystem.

Conclusions
Although Armillaria was frequently detected in the studied native and plantation forests, control measures have been restricted to urban trees and recreational parks, and implemented to prevent civilian and structural damage that may be caused by instability of affected trees.The control of Armillaria complex in forests is more difficult as it is spread over a wider area, often in areas that are difficult to access.The best way to reduce the vigor of the fungus, which is strongly dependent on the availability of food sources, is by pulling out infected stumps and roots.However, this measure can also disrupt beneficial microbial populations [64], which may act as a natural control of pathogens [65].Fungicidal treatments can kill the fungus in the soil; however, the economic and environmental costs are high.By contrast, early detection of disease and treatment with effective biological antagonists, and planting Armillaria-tolerant tree species are recommended for forests, plantations, urban areas and recreational parks [66].In typical modern plantations, control measures are probably economically justified if mortality from Armillaria spp. is severe early in the previous rotation.It is, therefore, important to keep good stand records that will point out the impact of different factors when a decision may be necessary prior to planting a new species.Even if action is not taken, forest owners should be aware of the presence of the fungi, especially if there is a chance that a change of management practice could inadvertently lead to an increase in disease impact [14,18,20].

Figure 2 .
Figure 2. Distribution of Armillaria spp.and main forest tree species of the Basque Country.

Figure 2 .
Figure 2. Distribution of Armillaria spp.and main forest tree species of the Basque Country.

Figure 4 .Figure 3 .
Figure 4. Location in an Euclidean space of the presence or absence of Armillaria spp.and environmental categories.The first two dimensions of the Euclidean space of the MCA are plotted to

Forests 2017, 8 , 235 8 of 18 Figure 3 .
Figure 3. Measure of the variance of each variable for each dimension.The variables with higher variance in the first dimension were mainly average rainfall (Rain), optimal conditions for host growth (Hoc), slope, Armillaria spp., soil permeability and stand orientation (Figure3).The variables with higher variance in the second dimension were orientation, average temperature (Temperature), average rainfall (Rain), slope, and soil acidity (Soil).

Figure 4 .Figure 4 .
Figure 4. Location in an Euclidean space of the presence or absence of Armillaria spp.and environmental categories.The first two dimensions of the Euclidean space of the MCA are plotted toFigure 4. Location in an Euclidean space of the presence or absence of Armillaria spp.and environmental categories.The first two dimensions of the Euclidean space of the MCA are plotted to examine the associations among categories.The values on the axes indicate the coordinates within the Euclidean space in which categories are located.Variable description can be found in Table2.

Figure 5 .
Figure 5. Values of the adjusted residuals (blue bars) for each category within a categorical variable for the presence of Armillaria spp.Red lines represent the value the adjust residuals must have to consider a category significant; if y = ±1.96then it is significant at p < 0.05, if y = ±2.58then it is significant at p < 0.01, and if y = ±3.29 then it is significant at p < 0.001(Field, 2009).

Figure 5 .
Figure 5. Values of the adjusted residuals (blue bars) for each category within a categorical variable for the presence of Armillaria spp.Red lines represent the value the adjust residuals must have to consider a category significant; if y = ±1.96then it is significant at p < 0.05, if y = ±2.58then it is significant at p < 0.01, and if y = ±3.29 then it is significant at p < 0.001(Field, 2009).

Figure 6 .
Figure 6.Fungal population analysis by mycelial pairings.Grey squares indicate strains belonging to the same SCG; white squares correspond to non-compatible strains.Paired sets shown on the left correspond to samples collected in Amunategi, paired sets shown on the right correspond to samples collected in Altube.

Figure 6 .
Figure 6.Fungal population analysis by mycelial pairings.Grey squares indicate strains belonging to the same SCG; white squares correspond to non-compatible strains.Paired sets shown on the left correspond to samples collected in Amunategi, paired sets shown on the right correspond to samples collected in Altube.

Figure 6 .
Figure 6.Fungal population analysis by mycelial pairings.Grey squares indicate strains belonging to the same SCG; white squares correspond to non-compatible strains.Paired sets shown on the left correspond to samples collected in Amunategi, paired sets shown on the right correspond to samples collected in Altube.

Figure 7 .
Figure 7. Spatial distribution of the infected trees and Armillaria spp.genotypes.Numbers correspond to Armillaria spp.isolates.Same genotype is marked with the same color.From top to bottom, stand located in Amunategi, stand located in Altube, and stand located in Otxandiano.

Figure 7 .
Figure 7. Spatial distribution of the infected trees and Armillaria spp.genotypes.Numbers correspond to Armillaria spp.isolates.Same genotype is marked with the same color.From top to bottom, stand located in Amunategi, stand located in Altube, and stand located in Otxandiano.

Figure 9 .
Figure 9. Susceptibility of several tree species found in the Basque Country to A. mellea infection.The relative frequency of healthy young trees and those with fungal lesions was determined four months after infection with A. mellea and growth under greenhouse conditions.Counts are represented as percentage of the total number of plants for each tree species.Black asterisks indicate positive significant z scores (p < 0.05).Red asterisks indicate negative significant z scores (p < 0.05).

Figure 9 .
Figure 9. Susceptibility of several tree species found in the Basque Country to A. mellea infection.The relative frequency of healthy young trees and those with fungal lesions was determined four months after infection with A. mellea and growth under greenhouse conditions.Counts are represented as percentage of the total number of plants for each tree species.Black asterisks indicate positive significant z scores (p < 0.05).Red asterisks indicate negative significant z scores (p < 0.05).

Figure 10 .
Figure 10.Length of lesions (cm) caused by A. mellea in different tree species.Error bars show the standard deviation of the means.Statistically significant differences of p < 0.05 between tree species are presented with different lowercase letters.

Figure 10 .
Figure 10.Length of lesions (cm) caused by A. mellea in different tree species.Error bars show the standard deviation of the means.Statistically significant differences of p < 0.05 between tree species are presented with different lowercase letters.

Table 2 .
Characteristics of the plots included in this study.

Table 3 .
Environmental variables for which a significant association with Armillaria spp. was observed.Pearson's chi square value (χ 2 ), degrees of freedom (df), p-value, Cramer´s V value, and effect size are shown for each environmental variable.theassociationsamong categories.The values on the axes indicate the coordinates within the Euclidean space in which categories are located.Variable description can be found in Table2. examine

Table 3 .
Environmental variables for which a significant association with Armillaria spp. was observed.Pearson's chi square value (χ 2 ), degrees of freedom (df), p-value, Cramer´s V value, and effect size are shown for each environmental variable.