The effect of Cu-amended medium was evaluated on the growth and sporulation of Trichoderma spp. (Table 1). Petri dishes containing 2% MEA (Malt extract agar, OXOID, Pratteln, Switzerland) and different concentrations of anhydrous CuSO₄ (Sigma-Aldrich, Buchs, Switzerland) (0%, 0.025%, 0.05%, 0.075%, 0.1%, 0.5% and 1%) were inoculated in the centre with 100 μl of Trichoderma spp. conidia suspension adjusted to 10⁶ spores ml^-1. The growth rate (mm day^-1) was determined by colony diameter measurements after 10 days. To evaluate the influence of Cu-toxicity on conidiation rates of Trichoderma spp. the same concentrations of Cu-amended medium were evaluated. After 10 days, the produced conidia were collected, filtered with 5 μm pore filters and counted in a Neubauer improved chamber (Marienfeld, Germany). The yield of conidia was normalised to the surface of the fungal colony according to Gandía et al. [26]. For each experimental treatment (growth and conidiation rates), five Petri dishes for each Trichoderma spp. were performed.

Interaction tests with wood block specimens of Scots pine (Pinus sylvestris L.) sapwood (2 x 2 x 8 cm; radial, tangential, longitudinal) were performed as described by Ribera et al. [27] with following modifications. For evaluation of the preventative effect of Trichoderma spp. against wood destroying basidiomycetes, 45 autoclavable plastic containers (WEZ, Oberentfelden, Switzerland. dimensions; 40L x 60W x 15H cm) with 1000 g of vermiculite and 15 g of Scots pine sawdust were used as a substrate for the cultivation of wood destroying basidiomycetes (Table 1). Each fungus was tested on three autoclavable containers as a biological replicate. Before testing, the moisture content (MC) and water holding capacity (WHC) of vermiculite was determined according to ENV 807 [28]. The amount of water needed to bring the substrate to 75% of its WHC was calculated and added to the vermiculite. After 8 weeks incubation with the five wood destroying basidiomycetes at 22 (±1)°C and 70% relative humidity (RH), the boxes were inoculated with either T. harzianum (T-720) or T. atroviride (T-685) (100 ml of 10⁶ spores ml^-1). Three separate containers (biological replicates) for each wood decay basidiomycete were kept as controls. T. harzianum (T-720) was selected due to its high resistance on Cu-amended medium. T. atroviride (T-685) was selected as a bench mark fungus due its high antagonistic potential as described in previous studies [29, 30]. After 2 weeks incubation with Trichoderma, three wood specimens (3 repetitions) were sterilised with ethylene oxide and placed into each container. Three wood specimens were also placed into each control containers. Determination of the initial wood dry mass was calculated by oven drying (103°C) test specimens during 24h. After 12 weeks incubation, the specimens were removed, carefully cleaned with a brush, oven dried (103°C) and the mass loss recorded. In addition, the lethal effect of Trichoderma spp. was evaluated by re-isolating the wood destroying basidiomycetes from the substrate. Five subsamples of the infected substrate were selected randomly from the surface of each container and then placed into five different Petri dishes containing 20 ml of 2% MEA with 2 ml of thiabendazole dissolved in lactic acid (Merck, Darmstadt, Germany) (T-MEA). Petri dishes were incubated at 22 (±1)°C and 70% RH for 10 days. T-MEA is a selective medium for wood destroying basidiomycetes [31]. If the wood destroying basidiomycete failed to grow on T-MEA, the lethal effect of Trichoderma spp. was considered to be 100%.

Evaluation of the toxicity of Cu-based wood preservatives was performed according to the European Standard EN113 [32]. Wood specimens of Scots pine sapwood with standard dimensions (2.5R x 1.5T x 5L cm) were impregnated using a vacuum-pressure process. Four concentrations of wood preservatives distributed close to the expected toxic value were selected and the retention rates of the wood specimens were calculated (Table 2). Specimens impregnated with distilled water (0% preservative formulation) were used as controls. Wood specimens were kept in a drying vessel during four weeks and inverted twice per week. Culture vessels were inoculated with R. placenta, C. puteana, G. trabeum and F. pinicola from the Empa collection (Table 1). In each culture vessel one Cu-treated wood specimen and one non-impregnated control specimen was placed. For each different amount of retention rates of wood preservative and the respective fungi, four biological replicates were used. After 16 weeks mass loss was recorded (as described above) and toxic values were calculated. The toxic thresholds were calculated for each wood preservative formulation at which the wood was not adequately protected against the wood destroying fungi. The protection provided for the wood by the test preservative was considered to be adequate when the mean loss in mass of the specimens was less than 3.0% of their initial dry mass.

Wood stakes of Scots pine sapwood (2R x 2T x 8L cm) were impregnated with vacuum-pressure according to EN252 [33] with two different concentrations selected from the results obtained in the “Test method for determining the protective effectiveness of wood preservatives against wood destroying basidiomycetes”. One concentration below the toxic value was used to evaluate the integrated control effect of Trichoderma spp. as a biocontrol agent. The second concentration above the toxic value was used as a positive control for each wood preservative. Wood specimens impregnated with distilled water (0% preservative formulation) were used as controls. After impregnation, stakes were dried for four weeks to allow fixation of the wood preservatives.

Autoclavable containers containing vermiculite/sawdust (as described above) were inoculated with a range wood destroying basidiomycetes. Antrodia serialis was selected as it is the most common fungus isolated from infected wood utility poles in Switzerland and Germany [3]. Serpula himantioides was selected due to its strong decomposition rates observed in 0.1% Cu-treated wood (higher than 5% mass loss after 8 weeks) [3]. R. placenta (EMPA 45) was used as a positive control because of its high Cu-tolerance as demonstrated by Civardi et al. [2] and Ribera et al. [3]. In total, nine containers per wood destroying basidiomycete were prepared (three biological repetitions per fungal treatment). After 8 weeks, three containers (biological replicates) per wood destroying basidiomycete were inoculated with spore suspensions of T. harzianum (T-720) or T. atroviride (T-685) as described above. Three separate containers (biological replicates) for each wood decay basidiomycete were kept as controls. Two weeks after the Trichoderma treatment, three sterilised wood specimens from each Cu-based wood preservative were randomly placed into each container. In total 27 samples per box were incubated.

After 9 months incubation, the wood specimens were removed and the mass loss was recorded as described above. In addition, the lethal effect of Trichoderma was also evaluated.

For light microscopy, the incubated wood was cut into smaller blocks (1 × 0.5 × 0.5 cm), embedded in a methacrylate medium and subsequently polymerized at 50°C. The embedded specimens were sectioned at 3 μm using a rotary microtome (Leica^® 2040 Supercut) fitted with a diamond knife. The selected transverse sections were stained for 12 h in safranine and then counterstained for 3 min in methylene blue and 30 min in auramine [34]. Colour micrographs (Kodak EPY 64T) were taken with a Leitz Orthoplan microscope fitted with a Leitz-Vario-Orthomat camera system. Early stages of brown-rot (seen as loss of birefringence due to cellulolysis) were detected by viewing sections between crossed Nicols [35–38].

Several statistical tests were performed with different objectives applying Analysis of Variance (ANOVA) to the growth, sporulation rates and wood mass losses variables. The normality assumption of the distribution data was checked via the Kolmogorov-Smirnov test. The p-values obtained rejected the normality at a 95% confidence level for growth measurements, sporulation rates and wood mass losses. Therefore, and following previous studies by Civardi et al. [2], a logarithmic transformation for growth and sporulation data and an arcsine transformation for wood weight losses were applied to obtain normality. All data was back-transformed to numerical values for presentation (expressed as mean ± SD). The ANOVA analysis was applied to assess the effect of Cu-amended medium on the growth and sporulation of each Trichoderma spp., considering measurements obtained from each Petri dish as biological replicates. For this purpose, Dunnett’s test was used to evaluate the differences in growth and sporulation between each treatment with CuSO₄ to the control group (0% CuSO₄) (significance levels of p<0.05). To compare the influence of Cu-amended medium on growth and sporulation between the Trichoderma spp., a Tukey’s HSD (Honestly Significant Difference) test was additionally performed for each concentration of CuSO₄ (p<0.05).

To evaluate the preventative effect of Trichoderma spp. (T-720 and T-685) against each wood destroying basidiomycetes another Tukey’s HSD test was also performed. The wood specimens from each container were considered as repetitions. Statistics were then performed on the average of the three repetitions that was considered the value for each biological replicate. All the statistical analysis were performed using the statistical software SPSS^® (Version 22, SPSS Inc., Chicago, IL, USA).

Due to the potential application of Trichoderma in soils with a high content of leached preservatives around the wood poles [39–42] it was important to identify a competitive strain that was able to colonise Cu-rich substrates. Some authors have described a high Cu-tolerance of Trichoderma spp. to a wide range of Cu concentrations [43–49]. In this study, the toxic effect of CuSO₄ amended medium on the growth of Trichoderma spp., compared by the Dunnett’s test, was statistically significant in concentrations above 0.1% CuSO₄ for all strains (Table 3). The Tukey’s test showed significant differences between the growths of Trichoderma spp. in some concentrations of CuSO₄. T. atroviride (T-722) demonstrated rapid colonization of the 0% Cu-amended medium compared to the other Trichoderma spp. However, T-722 was the most susceptible strain to CuSO₄, being the only strain that ceased to grow at 0.075% (Table 3). T. harzianum (T-720) and T. koningiopsis (T-723) were highly tolerant to CuSO₄-amended medium growing albeit more slowly in concentrations up to 0.75% of CuSO₄.

Results of the Dunnett’s test to measure the effect of CuSO₄ on the production of conidia showed that the effect was statistically significant (p<0.05) in concentrations above 0.075% CuSO_4, compared to the 0% CuSO₄ for all Trichoderma spp. (Table 4). The Tukey’s comparisons between Trichoderma spp. did not show significant differences for concentrations above 0.1% CuSO₄ since no Trichoderma spp. was able to colonise such a high concentrations of CuSO₄ in the substrate. T. harzianum (T-720) produced higher concentrations of conidia in absence of CuSO₄ (0% CuSO₄). Moreover, T-720 was able to sporulate, albeit in reduced yield, in toxic concentrations of CuSO₄ for the rest of the Trichoderma spp. (0% CuSO₄). The toxic effect of CuSO₄ on the production of conidia was more significant on the strains T-721 and T-722 compared to the other strains in 0.05% CuSO₄, producing no conidia.

The diversity of results reported by different authors [22–25] emphasizes the importance to specifically screen a Trichoderma isolate with a high antagonistic potential against the target fungus. The present study indicates that careful selection of a Trichoderma isolate with bioassays will improve the success rate of the antagonist in the field.

Treatment with Trichoderma spp. significantly reduced mass losses of some wood specimens compared to the controls (Fig 1). Trichoderma spp. had a significant impact (p<0.05) on the decay process of four wood destroying basidiomycetes (A. serialis, S. himantioides, G. sepiarium and R. placenta). Even if T. harzianum (T-720) showed a higher preventative effect on the reduction of mass loss against R. placenta, Tukey’s test did not show significant differences between the effects of both Trichoderma spp. treatments against wood destroying basidiomycetes (Fig 1). Additionally, negative re-isolation of wood destroying basidiomycetes after 12 weeks from the wood specimens on selective medium (T-MEA), demonstrated a 100% lethal effect of the wood destroying basidiomycetes by the applied Trichoderma spp. (Table 5).

Different authors have previously reported the successful eradication of wood destroying basidiomycetes with biocontrol agents on wood poles. For instance, Ricard [50] described the use of Scytalidium spp., Ascocoryne sarcoides and Trichoderma to suppress decay by N. lepideus and A. carbonica in the field. By contrast, Bruce and King [51, 52] showed that successfully eradication of L. lepideus only appeared possible during very early stage of wood decay in the field. The latter preventative and remedial studies demonstrated difficulties of Trichoderma to develop and dominate highly infected creosote wood poles in the field. The versatility of Trichoderma spp. to grow on a wide range of acid and basic substrates as well as its tolerance to high concentrations of CO₂ and N increases its potential as integrated control of wood destroying basidiomycetes as described by Score and Palfreyman [53].

All selected concentrations were in the range of the expected toxic values for the fungi used in the EN113 test because all fungi reported mass loss of the specimens less than 3.0% of their initial dry mass for the highest amount of retention of wood preservative (Fig 2). The mass losses reported for the controls (higher than 20%) confirmed the minimum activity of the fungi required for the test. R. placenta (Empa 45) showed the highest Cu-tolerance for all Cu-based preservatives compared to the other fungi. The toxic effect of the Cu-based preservatives on R. placenta was used to select the toxic retention threshold for the integrated wood protection of T. harzianum against the wood destroying basidiomycetes. The non-toxic retentions were defined according to the minimum concentrations at which the wood was not adequately protected for R. placenta and at which other fungi could cause wood decay. The non-toxic value for the CC-treatment was chosen to be 50% of the minimum tested retention to guarantee the non-toxic values.

The compatibility of T. harzianum (T-720) with a range of common wood preservatives demonstrated here and the good performance against wood destroying basidiomycetes was evident (Fig 3). A. serialis did not demonstrate a high resistance to the wood preservative formulations evaluated because it did not cause mass losses higher than 3% of the initial mass (Fig 3a). Nevertheless, the mass loss recorded for the control (absence of Cu-based and Trichoderma treatment) demonstrated a high activity of the fungus (nearly 70% of the initial mass). Wood samples treated without Cu-based preservatives showed a significant reduction of mass loss for both Trichoderma treatments. Besides, the treatment with T. atroviride (T-685) was significantly more effective to prevent wood decay. However, Tukey’s results did not show a positive effect of the Trichoderma spp. treatment for any Cu-impregnated specimens. S. himantioides caused mass losses higher than 5% in the control for all retentions rates of wood preservative formulations (Fig 3b). The application of both Trichoderma spp. to prevent decay against S. himantioides resulted in a reduction of mass loss for most of the Cu-impregnated wood. The preventative effect against wood decay basidiomycetes was only absent for high retentions of Cr-free preservative formulations (27.5 kg m³ Cr-free 1 and 41.89 kg m³ Cr-free 2). T. harzianum (T-720) demonstrated higher significant effect to prevent wood decay for all Cu-based treatments (Fig 3b). R. placenta caused high mass losses in the wood specimens used as controls (higher than 20%), CC preservative (3.85 kg m³ and 37.2 kg m³ CC) and low retentions of Cr-free preservatives (8.27 kg m³ Cr-free 1 and 15.10 kg m³ Cr-free 2) (Fig 3c). The preventative effect of Trichoderma treatment on those wood specimens was significant for all samples. Additionally, Tukey’s results only showed significant differences between both Trichoderma spp. for Cu-untreated wood specimens and 37.2 kg m³ CC (Fig 3c).

Interestingly, the general trend of a treatment with Trichoderma spp. was more evident on wood specimens impregnated with low concentrations of Cu-based wood preservatives that were incubated with S. himantioides (8.09 kg m³ CCB, 3.85 kg m³ CC, 8.27 kg m³ Cr-free 1 and 15.10 kg m³ Cr-free 2) and R. placenta (3.85 kg m³ CC, 8.27 kg m³ Cr-free 1 and 15.10 kg m³ Cr-free 2), respectively (Fig 3b and 3c). T. harzianum (T-720) showed a slightly higher antagonistic potential to protect Cu-impregnated and non-impregnated wood specimens against S. himatioides and R. placenta. Failure to re-isolate the wood destroying basidiomycetes from the wood specimens demonstrated 100% lethal effect by Trichoderma spp. (Table 6).

The microscopic investigations demonstrated typical features of brown-rot i.e. numerous clefts in the secondary walls of the tracheids by R. placenta even when the wood was treated with wood preservatives (Fig 4c, 4e and 4g). Cracks appeared in the secondary wall extending perpendicular within the S2 from the S3 to the S1. The colour of degraded secondary walls changed from blue to red, indicating the preferential decomposition of hemicellulose and cellulose (Fig 4c, 4e and 4g). Under polarized light, secondary walls of early wood tracheids showed a distinct loss of birefringence, as also observed by Schwarze et al. [38]. Higher lignified parts, such as the late wood tracheids and epithelial cells of the resin ducts were more resistant to decay, as also described by Schwarze [54]. In contrast, low retentions of wood preservatives in combination with T. harzianum (T-720) were associated with an absence of features typically associated with brown-rot (Fig 4d, 4f and 4h).

Although previous studies could not demonstrate the protection of Cu-treated wood against a range of wood destroying basidiomycetes [55, 56], the combined use of Trichoderma spp. as a biocontrol agent against two wood destroying basidiomycetes on creosote treated poles has been demonstrated in the past [50]. Thus, years after application of Trichoderma in the field, Bruce et al. [57–59] could demonstrate the viability of the Trichoderma treatment on the same site. The reduction of inoculum of wood destroying basidiomycetes demonstrated in the laboratory by T. harzianum (T-720), shows promising potential as a long term treatment for the field. This method of integrated wood protection could improve the efficacy of Cu-based new generation wood preservatives against wood destroying basidiomycetes on high risk sites in Switzerland and Germany.