Correspondence: Audrius Menkis, Department of Forest Mycology and Plant Pathology, Uppsala BioCenter, Swedish University of Agricultural Sciences, PO Box 7026, SE-75007 Uppsala, Sweden. Tel.: +4618672729; fax: +4618673599; e-mail: email@example.com
The aim of this study was to assess belowground occurrence, persistence and possible impact of the biocontrol agent Phlebiopsis gigantea (Fr.) Jülich on soil fungi. Sampling of soil and roots of Picea abies (L.) H. Karst. was carried out at 12 P. gigantea-treated and five nontreated control sites representing 1- to 60-month-old clear-cuts and thinned forest sites in Finland and Latvia. The 454-sequencing of ITS rRNA from fine roots, humus and mineral soil resulted in 8626 high-quality fungal sequences. Phlebiopsis gigantea represented 1.3% of all fungal sequences and was found in 14 treated and nontreated sites and in all three substrates. In different substrates, the relative abundance of P. gigantea at stump treatment sites either did not differ significantly or was significantly lower than in nontreated controls. No significant correlation was found between the time elapsed since the tree harvesting and/or application of the biocontrol and abundance of P. gigantea in different substrates. In conclusion, the results demonstrate that P. gigantea occasionally occurs belowground in forest ecosystems but that stump treatment with the biocontrol agent has little or no impact on occurrence and persistence of P. gigantea belowground, and consequently no significant impact on soil fungi.
Stump treatment with the biocontrol agent Phlebiopsis gigantea is currently a common forestry practice aimed to prevent infections by Heterobasidion spp. root-rot fungi – the most economically important pathogens of temperate and boreal coniferous forests (Woodward et al., 1998). In Europe, Heterobasidion species complex consists of three species, that is, Heterobasidion annosum (Fr.) Bref. sensu stricto, Heterobasidion parviporum Niemelä & Korhonen and Heterobasidion abietinum Niemelä & Korhonen, which are mostly found on pine, spruce and fir, respectively (Dalman et al., 2010). Primary infections of Heterobasidion take place when its airborne basidiospores land and germinate on freshly cut stump surfaces. Vegetative mycelia grow into the root systems and infect adjacent trees via root contacts (Redfern & Stenlid, 1998). Phlebiopsis gigantea is a saprotrophic wood-decay basidiomycete that, like Heterobasidion spp., is a primary colonizer of freshly cut conifer wood, able to outcompete the pathogen, simultaneously preventing secondary infection of living trees (Holdenrieder & Greig, 1998). Biological control with P. gigantea is accomplished by applying arthroconidia (oidia) of the fungus to the stump surface soon after felling of a tree, so that the niche suitable for Heterobasidion infections becomes preoccupied (Rishbeth, 1951, 1957; Redfern & Stenlid, 1998). In Europe, mechanized stump treatment with the biocontrol agent P. gigantea exceeds 200 000 ha each year and has been shown to reduce Heterobasidion stump infections by 95–100% compared with untreated stumps (Thor, 2005).
Although biocontrol is commonly considered an environmentally friendly approach to combat pathogens, it may involve potential risks, for example, because of its impact on nontarget organisms. Consequently, risk assessment associated with biocontrol is of considerable importance (Barratt et al., 2010). In the past, studies of persistence and short- and long-term impact of P. gigantea treatment on fungal diversity in stumps have been carried out (Varese et al., 1999; Vainio et al., 2001; Vasiliauskas et al., 2004, 2005), showing that the treatment has no significant impact on communities of wood-inhabiting fungi. However, during the treatment, a larger area than the stump surface alone is affected, as up to 50% of the biocontrol substance (spore suspension of P. gigantea) might be sprayed around stumps onto the neighbouring soil (Thor, 1993). The impact of the treatment on nontarget organisms present in the vicinity of the stump should therefore also be evaluated. However, to date, only a single related publication is available, showing negligible effects of P. gigantea treatment on the surrounding ground vegetation in Picea abies forest (Westlund & Nohrstedt, 2000). It was recently demonstrated that, besides wood, P. gigantea is also able to colonize soil and root tips of living trees, forming mycorrhiza-like associations (Vasiliauskas et al., 2007). A possible consequence of this capacity is that large-scale application of P. gigantea spore suspensions might influence diversity and community structure of resident soil fungi, with concomitant effects on forest ecosystems through competition for the same ecological niches.
The aim of this study was to assess belowground occurrence, persistence and possible impact of the biocontrol agent P. gigantea on soil fungi. Using 454-pyrosequencing method, we investigated fungi in fine tree roots and soil from P. gigantea-treated and nontreated clear-cuts and thinned forest sites in northern Europe, providing new in situ data on biosecurity and ecology of this fungal species used to control Heterobasidion root-rot.
Materials and methods
Study sites and sampling
This study included 17 sites representing clear-cuts and thinned stands in Finland and Latvia. In 12 of these sites, stumps immediately after tree harvesting were treated with the biocontrol agent P. gigantea (Rotstop®; Verdera Ltd., Finland), and in five sites, stumps were untreated (control) (Table 1). The study areas represented boreal (Finland) and northern temperate (Latvia) forests dominated by P. abies, with an admixture of Pinus sylvestris L. and Betula pendula Roth. Soil at the study sites was a coniferous forest podzol composed of 5- to 10-cm humus soil layer and an underlying sandy mineral soil.
Table 1. Studied Picea abies forest sites and detected diversity of soil fungi
Sampling was carried out in June 2009 and included living fine roots of P. abies trees and soil cores. All sampled trees were 0.3–0.5 m high, grew naturally within a 0.5 m radius of cut P. abies stumps and had presumably already been established on each site prior to tree harvesting and eventual stump treatment with spore suspension of P. gigantea. A total of 53 trees were sampled, two to five per site, except for three sites where such trees were not available (Table 1). The trees were excavated from the ground, and their root systems excised from the stems, packed into the plastic bags, labelled, transported to the laboratory and stored at −20 °C. The roots were washed in running tap water to remove any remaining soil, then 2- to 3-cm-long portions of all fine roots (including root tips) were excised from the end of main roots, placed in 50-mL centrifugation tubes and freeze-dried at −60 °C for 2 days. Resulting root samples (14 in total, representing each site where roots of all plants within the same site were collected) were homogenized, and 15 mg (dry weight) root material per sample was used for DNA isolation.
Additionally, at each site, three soil cores (5 cm in diameter and 70 cm in depth) were taken randomly within a 0.5 m radius of three different stumps (one core per stump). As a result, 51 soil cores were collected, sealed in the extraction tubes, labelled, transported to the laboratory and stored at −20 °C. Humus and mineral soil were analysed individually. Each soil core was separated into humus layer and mineral layer, and samples of each respective layer within the same study plot were mixed together resulting in a total of 17 humus and 17 mineral soil samples, representing each respective site (Table 1). The samples were air-dried at room temperature (ca. 21 °C) for 8–12 h and sieved using a 2-mm-diameter sieve to remove larger particles and roots. Resulting samples of humus and mineral soil were freeze-dried at −60 °C for 5 days. Then, from each sample, 1 g (dry weight) of mineral soil and 100 mg (dry weight) of humus were used for DNA isolation.
DNA isolation, amplification and sequencing
Genomic DNA was isolated from 14 samples of fine roots, 17 of humus, and 17 of mineral soil using Power Plant or Power Soil DNA Isolation Kits (MoBio Laboratories, Carlsbad, CA), according to the manufacturer's recommendations. In addition, isolated DNA was purified using JETquick DNA Clean-Up System (Genomed, Löhne, Germany). In each sample, concentration of genomic DNA was determined using a ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). Diluted (1–10 ng μL−1) genomic DNA samples were amplified separately using the primer pair ITS9 (5′-GAACGCAGCRAAIIGYGA-3′) (Department of Forest Mycology and Plant Pathology, Swedish University of Agricultural Sciences) and ITS4 (5′-xxxxxxxxTCCTCCGCTTATTGATATGC-3′) (White et al., 1990) containing 8-bp sample identification barcodes denoted by x. Using this primer pair, amplified PCR products were estimated to be between 280–420 bp in size and to include larger part of the 5.8S rRNA gene sequences, complete sequences of noncoding ITS2 rRNA region and partial sequences of the 28S rRNA gene. Besides, using this primer pair, PCR chimeras can be expected to be very occasional or absent because these in ITS region are frequently formed when amplicons encompass both ITS1 and ITS2 regions, with the highly conserved 5.8S region in the middle, where partial fragments derived from different templates combine at the conserved 5.8S region and result in sequences with the two ITS regions originating from different templates. The PCR reactions, 45 μL in volume for each sample, were performed using a Veriti Thermal Cycler (Applied Biosystems, Carlsbad, CA) using DreamTaq Green DNA polymerase (Fermentas, St. Leon-Rot, Germany). The PCR cycle parameters consisted of an initial denaturation at 95 °C for 2 min, 27 cycles of denaturation at 95 °C for 30 s, annealing at 55 °C for 30 s and extension at 72 °C for 45 s, followed by a final extension step at 72 °C for 7 min. The PCR products were analysed on 1.5% agarose gels (Agarose D1, Conda, Madrid, Spain) under UV, and amplicons were purified using QIAquick Gel Extraction Kit (Qiagen, Hilden, Germany). The concentration of purified PCR products was determined using Quant-iT™ dsDNA HS Assay Kit (Life Technologies, Carlsbad, CA), and an equimolar mix of all PCR products was used for pyrosequencing. Construction of FLX library and pyrosequencing in a ¼ run as a part of the larger sample was carried out by Macrogen Inc. (Seoul, Korea), utilizing a GS-FLX Titanium 454 system (454 Life Sciences, Branford, CT). Prior to sequencing, estimated DNA concentration was 74.3 ng μL−1 and the predicted mean product size was 353 bp.
The assessment of sequence quality and sequence sorting according to the barcodes corresponding to the 48 samples of this study and barcode removal was carried out by Macrogen Inc. (Seoul, Korea). To track sample information of individual sequences following their assembly, sorted sequences in individual FASTA files were labelled with a sample identifier using bioedit (Hall, 1999). All labelled sequences were imported into seqman pro v.9.0.4 from dnastar lasergene 9 package (DNASTAR, Inc. Madison, WI) and assembled in tentative consensus sequences using the Pro Assembly option with default parameters. An exception was Minimum Sequence Length, which was set to 200 bp (all shorter sequences were discarded during the assembly). For taxonomic identification, example sequences were compared with the databases at both NCBI and UNITE (Koljalg et al., 2005) using the blastn algorithm. The criteria used for identification were as follows: sequence coverage > 80%; similarity to species level 97–100%; similarity to genus level 94–96%. Sequences not matching those criteria or lacking taxonomic description in the reference sequences were considered unidentified, assigned to a phylum or class and given unique names as in Table 2.
Table 2. List of the 23 most common fungal taxa found in 48 samples (Picea abies fine roots, humus and mineral soil) of this study
Sequence similarity column shows base pairs compared between the query sequence and the reference sequence at UNITE or NCBI databases, and the percentage of sequence similarity in the parenthesis.
Oidiodendron sp. 42_48
Piloderma sp. 277_49
Tomentella sp. 23_52
Meliniomyces sp. 114_45
Ceratocystis sp. 58_121
The rarefaction analysis was performed using analytical rarefaction v.1.3 available at http://www.uga.edu/strata/software/index.html. Chao1 diversity indices (Chao et al., 2005) were calculated using the estimates software package version 7.5.2, R. K. Colwell; http://viceroy.eeb.uconn.edu/EstimateS. The relative abundance of P. gigantea in different datasets (treatments, substrates, study sites, harvesting types and countries) and richness of taxa in different treatments were compared by chi-squared (χ2) tests calculated from the actual number of observations (Mead & Curnow, 1983). In cases where the datasets were subjected to multiple comparisons, confidence limits for P-values of the chi-squared tests were reduced a corresponding number of times, as required by the Bonferroni correction (Sokal & Rohlf, 1995). Pearson correlation coefficients were calculated to examine the relationship between the time elapsed since the application of P. gigantea and abundance of P. gigantea in different sites and substrates. The statistics were computed using Minitab® statistical software (Minitab® Inc., 2003). Possible impact of the treatment on fungal communities in different study sites was analysed using Principal Component Analysis (PCA) in canoco 4.5 (ter Braak & Smilauer, 1998). Fungal communities between different treatments, countries, harvesting types and study sites were compared by calculating qualitative Sorensen similarity indices (SS) (Magurran, 1988).
A total of 38800 sequences was generated by 454-pyrosequencing from the 48 samples of this study. However, 29855 of these were shorter than 200 bp and were excluded from further analysis. Among the remaining 8945 high-quality sequences (302 bp per sequence on average, or 3.1 Mbp in total), 8626 (96.4%) represented fungi, 279 (3.1%) protists, 34 (0.4%) plants and 6 (0.1%) animals. Assembly of fungal sequences resulted in 456 contigs (at the ≥ 97% similarity level representing different fungal taxa) of which 194 (42.5%) were singletons and 262 (57.5%) were nonsingletons. All singletons and example sequences for nonsingletons are available at GenBank (accessions JQ312675–JQ313130). Total numbers of fungal sequences and of different taxa obtained from each study site and substrate are shown in Table 1. A plot of fungal taxa from the different substrates (fine roots, humus and mineral soil) vs. the number of sequences resulted in rarefaction curves that did not reach the asymptote (Fig. 1), indicating that a potentially higher diversity of taxa could be detected with increased sequencing effort. In support of this, the nonparametric Chao1 estimator predicted that, depending on the type of substrate, the maximum number of taxa could be higher by ca. 51–72%. In this study, the detected fungi were 47.8% Ascomycetes, 44.1% Basidiomycetes, 7.9% Mucoromycotina, 0.1% Glomeromycetes and 0.1% Chytridiomycetes. The absolute richness of taxa was highest in humus (252 taxa out of 2229 sequences), followed by fine roots (201 out of 4723) and lowest in mineral soil (182 out of 1674) (Table 1), but if the same number of sequences had been taken from each substrate, the highest richness would be in humus, then in mineral soil and lowest in fine roots (Fig. 1). Consequently, the chi-squared test showed that the richness of taxa in fine roots was significantly lower than in humus and mineral soil (P < 0.0001) while the latter two did not differ significantly from each other (P >0.05). Identification at least to genus level was successful for 172 (37.7%) out of 456 of fungal taxa of this study, and those represented 69.1% of all fungal sequences. Information on the 23 most common taxa representing 61.2% of all fungal sequences is shown in Table 2.
In this study, P. gigantea represented 1.3% of all fungal sequences (Table 2) and was found in 14 out of 17 treated and nontreated study sites, encompassing different countries (Finland and Latvia), harvesting types (clear-cuts and thinnings) and substrates (fine roots, humus and mineral soil) (Table 3). Only a single sequence of Heterobasidion sp. (or 0.01% of all) was detected and originated from the mineral soil. Comparison by chi-squared test revealed that the abundance of P. gigantea did not differ significantly either between countries or between the harvesting types (P >0.7 and P >0.08, respectively). In contrast, the abundance of P. gigantea in different substrates differed significantly, being highest in mineral soil (4.7%), followed by humus (1.5%) and lowest in fine roots (0.1%) (chi-squared test, P <0.0001) (Table 3). Moreover, a chi-squared test revealed that the relative abundance of P. gigantea in the biocontrol treatment sites either did not differ significantly or was significantly lower than in nontreated controls (Table 3). Furthermore, the richness of fungal taxa was significantly higher at P. gigantea-treated than nontreated sites (chi-squared test, P <0.044). However, no significant correlation was found between the time elapsed since the tree harvesting and/or biocontrol treatment and the abundance of P. gigantea in different study sites and substrates (P >0.05). Minimum-evolution clustering analysis of all ITS rRNA sequences of P. gigantea of this study and of the biocontrol strain DQ320133 revealed the presence of 65 distinct ITS types, among which the largest was composed of 30 sequences including the biocontrol strain (Supporting Information, Fig. S1). Sorensen indices of similarity were moderate when fungal communities were compared between P. gigantea-treated vs. nontreated sites (0.39), Finland vs. Latvia (0.38) and clear-cuts vs. thinnings (0.43) and were low to moderate when compared in all possible combinations between individual study sites (0.21–0.48). Furthermore, PCA analysis of fungal communities showed no specific distribution pattern among P. gigantea-treated and nontreated study sites that were largely intermingled and in more or less close proximity to each other on Axis 1 (Fig. 2). An exception to this was H5 and G8-control sites, in which the ectomycorrhizal basidiomycete Piloderma sphaerosporum Jülich – the most commonly detected species in this study (Table 2), made up exceptionally high proportion (respectively, 21% and 39%) of all sequences, causing distant placement of those two sites in the plot (Fig. 2).
Table 3. Occurrence and relative abundance of Phlebiopsis gigantea (shown as a proportion of all fungal sequences) in different study sites and substrates
P. abies fine roots
Within columns of respective substrate, values followed by the same letter in chi-squared test are not significantly different.
n.a., data not available.
The results of this study demonstrated that P. gigantea – a common wood-decay fungus (Ryman & Holmåsen, 1998), occurs belowground in forest ecosystems while inhabiting fine tree roots, humus and mineral soil. Although P. gigantea was only found in P. abies fine roots from Finland (Table 3), in our other study on fungal communities in replanted clear-cuts, it was also sequenced from ectomycorrhizal root tips of P. abies seedlings in Latvia (unpublished data). Our results therefore corroborate previous studies (Menkis et al., 2005; Vasiliauskas et al., 2007), providing new in situ information on ecological niches and the multi-trophic nature of this fungal species. Furthermore, the results suggest that the majority of P. gigantea sequences obtained in this study represented individuals from the natural populations (Fig. S1). However, one must be aware that some of observed diversity within ITS might be due to the 454-sequencing errors (Harismendy et al., 2009), and those results, therefore, should be interpreted with caution. Nevertheless, in agreement with our results, a high genetic diversity in natural P. gigantea populations sampled from stumps and logs was reported previously, suggesting that biocontrol applications of genetically uniform P. gigantea are unlikely to cause any immediate threat to the genetic diversity of this fungal species (Vainio et al., 2001; Samils et al., 2009). In this study, the sites in which the ITS type of P. gigantea biocontrol strain was detected were 1–60 months old following harvesting of the trees and included both treated and nontreated sites (Fig. S1; Table 1). Therefore, the possibility should not be excluded that the biocontrol strain might persist in the soil for up to 60 months and also spread occasionally to the nontreated sites. For example in stumps treated with the biocontrol strain P. gigantea, it was found that this strain dominates the fungal communities for 4–5 years and then its abundance rapidly decreases (Vainio et al., 2001; Vasiliauskas et al., 2005) while its spread to adjacent stands is largely restricted (Samils et al., 2009). However, the results of this study suggest that the relative abundance of P. gigantea belowground is not affected either by the time elapsed since the application of the biocontrol agent or by the treatment itself as in different sites it was generally low over time and did not significantly exceed the levels found in the controls (Table 3).
In contrast to P. gigantea, many ectomycorrhizal and endophytic taxa were dominant (Table 2) and often shared between individual study sites. Consequently, PCA has shown a rather proximal aggregation of many sites both treated and nontreated at the same time showing that stump treatment with the biocontrol agent P. gigantea had little or no impact on fungal community structure (Fig. 2). Estimates of Sorensen similarity indices between P. gigantea-treated and nontreated sites were of the same magnitude as in all other datasets, repeatedly suggesting that fungal communities detected in this study were generally unaffected by the treatment while similarity levels observed indicated the relative importance of rare taxa present at each site. In common with our results, Vasiliauskas et al. (2005) have shown that in stumps treated with the biocontrol agent P. gigantea the fungal community structure remains largely unchanged while species richness is moderately reduced. In this study, in contrast, biocontrol treatment with P. gigantea had a slight but significantly positive effect on richness of fungal taxa. The reason for this is unclear but the possibility should not be excluded that biocontrol treatment, while preventing establishment of Heterobasidion in stumps and eventually in roots, created more favourable conditions for the establishment of diverse fungal communities.
In conclusion, the results of this study demonstrated that P. gigantea occasionally occurs belowground in forest ecosystems but that stump treatment with the biocontrol agent P. gigantea has little or no impact on occurrence and persistence of this species belowground, and consequently no significant impact on soil fungi.
Financial support from Stiftelsen Oscar och Lilli Lamms Minne, Stiftelsen Carl-Fredrik von Horns fond, the SNS project PATHCAR and The Swedish Research Council Formas is gratefully acknowledged.