Introduction

Ectomycorrhizal fungi, i.e., which live in symbiosis with tree and shrubs, play important roles in forest functioning and biogeochemical cycles1. In boreal forests, 50–70% of the carbon stored in the soil is derived from roots and root-associated microorganisms such as ectomycorrhizal fungi2. Besides forest ecosystems, ectomycorrhizal trees were also implanted in agroforestry ecosystems and in dedicated orchards for producing non-wood products such as edible fungi. The inoculation of tree seedlings with selected ectomycorrhizal fungi in nurseries (i.e., controlled mycorrhization) started hundred years ago, and this technique has been used extensively to grow truffles since the 1970s3.

True truffles (Tuber spp.) are ectomycorrhizal Ascomycetes producing hypogeous fruiting bodies. The genus Tuber is present globally in temperate areas including over 200 species4,5. Of these, at least 30 species are naturally present in Europe6, but only a few have remarkable organoleptic properties and social value7,8,9. The most valuable truffle species are Tuber melanosporum Vittad, the Périgord black truffle; Tuber magnatum Pico, the white Italian truffle; and Tuber aestivum Vittad, the summer and Burgundy truffle. Since the first commercialization of seedlings inoculated with T. aestivum and T. melanosporum in 1973, considerable progress has been made to improve the quality of the inoculated plants3,10. T. melanosporum plantations are found not only in the Mediterranean area but also outside their native range including Australia, New Zealand, South-Africa, South-America, and West USA10.

T. aestivum is the second most cultivated truffle species worldwide3. Versus T. melanosporum, it has the most extensive natural habitat from Morocco to Sweden and from Ireland to Azerbaijan11,12,13. Its wide natural distribution and long harvest season—make this species very promising for cultivation12,14,15,16.

The entire truffle life cycle occurs in the soil, and therefore soil parameters such as temperature and water availability could influence plantation success. However, information about ecological and agronomic requirements for the establishment and management of T. aestivum orchards (e.g. soil tillage, irrigation, and tree pruning) are largely unknown. To the best of our knowledge, there is no long-term data on T. aestivum ascocarps production in a plantation available in the literature. Molinier et al.17 published a study on the production of T. aestivum followed during 30-years, but this orchard used seedlings inoculated with T. melanosporum. Clear guidelines for orchard management are thus needed to improve cultivation of this species.

The genome of T. aestivum has been recently sequenced18. T. aestivum is a heterothallic species but it is impossible to realize its life cycle in vitro. To better understand its sexual reproduction cycle, an in situ small-scale genetic structure analysis was realized19,20. An important genotype turnover was observed suggesting a ruderal strategy. However, some genotypes spread several hundred meters. Gryndler et al.21 developed a qPCR protocol to better understand T. aestivum soil DNA localisation and dynamics. T. aestivum DNA was found profusely on the surface of roots other than ectomycorrhizae21. Moreover, stimulation of soil T. aestivum DNA with high doses of lime powder and other organic compounds (i.e. gallic acid, cellulose and calcium formate) was observed22. T. aestivum can share the same habitat as other truffle species (e.g. T. melanosporum and T. mesentericum) and thus it is possible that the DNA of all of these truffle species coexist in the soil17,23. This is an important issue because the Gryndler et al.21 protocol also amplified T. mesentericum DNA. Thus, it was necessary to develop another protocol that exclusively amplifies T. aestivum DNA.

Here, we analysed a highly productive T. aestivum orchard composed of 117 trees (hazels, oaks, pines, lime, and hornbeam) belonging to the French truffle grower’s federation experimental network. The aims of this study were: 1) to design a specific qPCR protocol via genomic data to trace and quantify T. aestivum DNA in the soil; and 2) to assess the monthly soil DNA dynamic over 24 months according to soil parameters (i.e. soil hydric potential and temperature).

Results

Evolution of the production in the experimental truffle orchard

The production of the truffle orchard (Fig. 1) started in 2009—five years after the plantation (Fig. 2a and Supplementary Table S1). From 2009 to 2014 the annual T. aestivum production increased exponentially. The production then stabilized (about 40 kg corresponding to 266 kg/ha) to reach a maximum of 49 kg in 2017 (320 kg/ha). In 2018, the ascocarp production declined to 11 kg (73 kg/ha; Fig. 2a and Supplementary Table S1). We then compared the production of the five tree species present in the plantation. Considering the production from 2009 to 2018, the mean quantity of truffle harvested under Corylus avellana, Pinus nigra, Quercus pubescens, and Ttilia cordata was similar (between 100 to 200 g per tree), although it was significantly higher for Carpinus betulus (above 400 g per tree) (Figs 2b and S1).

Figure 1
figure 1

Map of the T. aestivum experimental orchards in Nouvelle-Aquitaine (France). Different symbols display the five different species within the site (117 trees in total). Stars, Carpinus betulus (10 plants); circles, Corylus avellana (77 plants); triangles, Pinus nigra (10 plants); squares, Quercus pubescens (10 plants); rhombus, Tilia cordata (10 plants). Seedlings of the five different species were planted randomly. Black stars represent the three highly producing trees (C. betulus) selected in the frame of the national experimental project Culturtruf (trees II-9, III-11 and VIII-5). Areas delimited by points and lines represent the high density (4 × 3 m equivalent to 833 trees per hectares) and low density (4 × 6 m equivalent to 417 trees per hectares) in the orchards, respectively.

Full size image
Figure 2
figure 2

Evolution of the production in the experimental T. aestivum orchard. Annual total production in kg from 2009 to 2018 (a) and average production according to tree species from 2009 to 2018 (b). CB, Carpinus betulus; QP, Quercus pubescens; CA, Corylus avellana; PN, Pinus nigra; and TC, Tilia cordata. Different letters (a,b) indicate significant differences between tree species (Tukey’s HSD test, p < 0.05).

Full size image

Evolution of the rainfall, soil hydric potential, and soil temperature

The daily rainfall, pF (i.e. soil hydric potential), and soil temperature recorded from September 2016 to September 2018 at north and south probes are presented in Supplementary Figs S2–S4 for C. betulus trees II-9, III-11, and VIII-5, respectively. The pF corresponded to the field capacity for all trees from November 2016 to March 2017 and from November 2017 to March 2018. In 2017, the pF started to increase from April to November to reach pF 4 (i.e. temporary wilting point) several days for trees II-9 and III-11. There were few days above pF 4 due to rainfall or watering (Supplementary Figs S2 and S3). In 2018, the pF was higher in May than in 2017; it reached pF 4 over several days for trees II-9 and III-11 during July and August. In both 2017 and 2018, the pF did not reach the temporary wilting point for tree VIII-5 (Supplementary Fig. S4). The daily pF was significantly higher for the probe in the north than the probe in the south of trees II-9 and VIII-5 (Wilcoxon p-value < 0.001). In contrast, it was higher for the probe in the south than the probe in the north for tree III-11 (Wilcoxon p-value < 0.001).

The maximum and minimum soil temperature was recorded in August 2018 with 22.7 °C for south probe of tree III-11 and in February 2018 with 1.7 °C for south probe of tree VIII-5, respectively (Supplementary Figs S3 and S4; Supplementary Tables S4 and S5). The daily temperature was significantly higher for the probe in the south than the probe in the north of trees II-9 and VIII-5 (Wilcoxon p-value < 0.01). In contrast, it was higher for the probe in the north than the probe in the south for tree III-11 (Wilcoxon p-value < 0.01). The daily pF and temperature recorded under tree VIII-5 were significantly lower than under trees II-9 and III-11 (Fig. 3; Wilcoxon p-value < 0.001 and <0.01, respectively).

Figure 3
figure 3

Graphic representations of temperature (a) and pF (b) (i.e. soil hydric potential) registered from September 2016 to September 2018 under tree II-9 (in blue), tree III-11 (in red), and tree VIII-5 (in green). The daily temperature and pF are represented as the average between north and south probes. The temporary wilting point (pF 4.2; black dashed line) is reached by tree II-9 and tree III-11 from June 2017 to October 2017 and from July 2018 to August 2018; this does not occur in tree VIII-5.

Full size image

Development of a specific protocol to quantify DNA of T. aestivum

A total of 47,887 clusters were identified; 647 had one single copy gene model present only in T. aestivum (orphan genes). Clusters 17,691 and 17,699 correspond to gene models GSTUAT00001921001 (proteinId 4119) and GSTUAT00007989001 (proteinId 4108), respectively; these clusters were randomly selected. The gene models were called TuGM4119 and TuGM4108, respectively. TuGM4119 was 1542 nucleotides long, and coded for a protein with 428 amino acids and five exons. The NCBI database showed homology in 5′ with a calpain-like protein from Trichoderma parareesei (70 amino acids, e-value 2.10−04) and cystein protein of T. magnatum (152 amino acid, e-value of 4.10−4). There was no known Pfam motif.

TuGM4108 contained 626 nucleotides, and coded for a protein with 102 amino acid and two exons. It had no homology in the NCBI database or any Pfam motif. A pair of primers was designed in both gene models. PCR on thirteen DNA samples from different Tuber spp. samples showed no amplifications for primers TuGM4119f/TuGM4119r and amplification only for T. aestivum DNA for TuGM4108f/TuGM4108r (Supplementary Fig. S5). The primer pair TuGM4119f/TuGM4119r was not investigated further. In qPCR, TuGM4108f/TuGM4108r primers were amplified for all T. aestivum DNA tested but not in the DNA of other species (Supplementary Fig. S6a). Good amplification was seen in the qPCR of 16T. aestivum isolates harvested from different geographic location (Supplementary Fig. S6b).

Temporal dynamic of T. aestivum DNA in the soil

The T. aestivum qPCR protocol was applied on soil samples harvested each month under trees II-9, III-11, and VII-5. The T. aestivum DNA quantity varied each month with a similar trend for all three trees (Fig. 4). It was significantly lower in March/April of 2017 and 2018, although it was higher in August and December 2016 (Fig. 4). Values were higher from August 2016 to August 2017 than from August 2017 to August 2018 (Fig. 4).

Figure 4
figure 4

Monthly quantification of T. aestivum soil DNA from August 2016 to August 2018 for trees II-9 (a), III-11 (b), and VIII-5 (c). Soil DNA is expressed as milligrams of fresh mycelium per gram of dry soil. Error bars represent standard deviation between three replicates. Different letters (a–j) indicate significant differences between sampling months for each tree (Tukey’s HSD test, p < 0.05).

Full size image

Pearson’s correlation and simple linear regression analyses showed a significant (p values < 0.05 for both) correlation between soil DNA quantities and pF explaining 36% and 12% of the variability, respectively, when all seasons were analysed together (Fig. 5a). A significant correlation was also found between soil DNA quantities and soil temperatures. These correlations explained 30% and 8% of the variability according to the two analyses (Fig. 5b). No significant correlation was found between pF and soil DNA quantities when the analysis was done separately for summer and winter (Fig. 5c,e; Supplementary Table S6). Temperature data showed a high positive correlation in summer (p value < 0.001) although no correlation was observed in winter (Fig. 5d,f; Supplementary Table S6). The combined effect of pF and temperature tested by factorial ANOVA was not significant (p value = 0.59).

Figure 5
figure 5

Pearson’s correlation and simple linear regression between soil DNA quantities, pF, and temperature. Temperature and pF data were analysed together for all years (a,b) and separately for summer (c,d) and winter (e,f). Summer corresponds to the period from April to August. Winter corresponds to the period from October to March. The soil DNA quantities were log and square root transformed. R2: Pearson’s correlation coefficient; r2: simple linear regression’s correlation coefficient.

Full size image

Discussion

The ascocarp production reached in this orchard is especially high

The T. aestivum plantation was not set up to study management techniques; thus, we cannot conclude if the management technique applied here was the best. However, the orchard produced 49 kg per 1500 m2 12 years after planting (about 320 kg/ha equivalent). This production was especially high versus other T. aestivum truffle grounds. In a plantation of seedlings inoculated with T. melanosporum, T. aestivum totally substituted T. melanosporum 16 years after planting17. The maximum of production was about 30 kg/ha. The T. aestivum production was higher in a 41.5 ha forest in Hungary producing about 104 kg/ha24. For T. melanosporum, one plantation in the nineteen-century and another of the beginning of the 21st century had a maximum production of about 70 kg/ha2 25. The production level seen here is thus exceptionally high.

Temporary wilting point was only reached a few days

The temporary wilting point (pF = 4) seemed to be an important threshold for T. melanosporum irrigation26. Indeed, maintaining the pF below this threshold increased the production compared to non-irrigated controls26.

To date the pF was never recorded in a T. aestivum orchard. Three C. betulus (trees II-9, III-11 and VIII-5) were selected for soil hydric potential and temperature monitoring. This species was selected among the five tree species because C. betulus was the most productive according to mean truffle production per plants (Fig. 2). According to these results, the real potential of C. betulus as host for truffle would deserve further study. The pF was significantly lower for tree VIII-5 than for trees II-8 and III-11 (Fig. 3b). Tree VII-5 was present in the 417 trees/ha density area although the other two trees were localized in the higher density area (833 trees/ha) (Fig. 1). The difference in hydric potential could be explained by the rain interception in 833 trees/ha density as well as by higher water consumption of trees. The rain intercept was well illustrated by the north probe under tree II-9 for which the rains of July 24th 2017 did not allow the pF to reach field capacity although this was seen in the southern probe (Supplementary Fig. S2). High-density plantations for T. aestivum are often suggested (about 1000 trees/ha), but this needs to be questioned without a drastic tree-pruning strategy. Beside the water availability, density could also impact the production as it was observed in an experimental truffle orchard in Grand-Est region (France) where four densities were implanted (2000, 1000, 666, and 333 trees/ha)27. The production started 11 years after plantation in the 2000 trees/ha plantation, but it decreased dramatically a few years ago. In the long term, the lowest density (333 tree/ha) was the most favourable because even though the production started later it continuously increased from 14 to 20 years after planting.

Monthly variation of soil T. aestivum DNA are partly explained by soil temperature and hydric potential

The soil mycelium of ectomycorrhizal species plays a key role in soil nutrient uptake and transfer of carbon from plant to fruiting bodies28. Molecular methods have been used to analyse soil DNA of specific ectomycorrhizal fungi to evaluate the ecological and functional impact of a given species on its natural environment29. For example, a targeted qPCR protocol has shown that warmer and drier soil conditions reduced the Lactarius vinosus soil DNA30. A qPCR protocol was designed for quantifying T. melanosporum and T. magnatum DNA in the soil31,32. In the latter species, it has been shown by a model approach that the soil mycelium was favoured at 20 °C with pF of about 233. The genome of T. aestivum was recently sequenced18, this allowed us to design a specific qPCR protocol. Indeed, the existing protocol using ITS region21,34 also amplified T. mesentericum.

Using this qPCR protocol, we detected a monthly variation of the T. aestivum soil DNA. This variation correlated with soil temperature and hydric potential. Soil temperature is an important factor regulating microbial activity and shaping the soil microbial community35. The effect of temperature on truffle growth has been analysed in vitro36,37, but our report showed the impact of temperature on truffles in the soil for the first time by a monthly sampling. In our study, the quantity of DNA was low below 5 °C with the highest quantity detected at 6 °C (Supplementary Table S6). In vitro, T. melanosporum responds to cold temperature with specific gene profiling38. Our results suggest that T. aestivum also has a similar machinery to tolerate cold temperatures. Fungal growth rates are typically optimum around 25–30 °C, but the optimum growth temperature varies according to species and isolates35,39. In our study, the soil temperature was never above 22 °C. The second highest value of DNA was observed at 20 °C, indicating that T. aestivum mycelium in the soil can grow between 6 to 20 °C with more growth at higher temperatures. This range is in accordance with data available for other truffle species. Indeed, Iotti et al.33 determined 20 °C as the optimal temperature for T. magnatum mycelium growth in the soil.

According to Coleman et al.40, three patterns of response to water stress could be observed for ectomycorrhizal mycelial growth: growth only occurred in control treatments without hydric stress (Type I). Growth decreases with increasing stress and the maximum growth is always without stress (Type II). The maximum growth does not occur in the control but rather when the mycelium is under hydric stress (Type III). Here, the two highest values of DNA were observed in non-stressed conditions (pF = 2.2) and in a very high stress condition (pF = 5.9). There was also a linear correlation between pF and DNA quantity (Fig. 5). Thus, we conclude that T. aestivum is a type III species such as Terfezia boudieri and Picoa lefebvrei two desert truffles41. T. aestivum is drought tolerant because it can grow above pF 4.340. On the contrary, T. magnatum seems to belong to type I since the optimal water potential is considered as being at pF around 233. Both truffle species have different ecological requirements. T. magnatum is most of the time harvested in riparian conditions while T. aestivum is harvested in less humid soil conditions. These results have practicable implications for truffle growers: irrigation must be triggered at very different soil water potentials for both species.

Temperature and pF only partially explain the content of soil DNA variability (36% to 30% according the analyses; Fig. 5). Other factors therefore could influence summer truffle soil DNA, but they were not analysed in this study. T. aestivum could enter into competition with other naturally occurring ectomycorrhizal fungi.

According to Stobbe et al.11,12, T. aestivum can fruit all the year, and the only two months without truffles are March and April. Interestingly, we found that months with less soil DNA were also March and April for 2017 and 2018. Unfortunately, because of her personal schedule, the owner of the orchard only harvested ascocarps from May to August, which did not allow us to detect a putative link between ascocarp production and soil DNA.

The quantity of soil T. aestivum DNA decreased from 2016/2017 to 2018. In parallel, the production of ascocarps in 2018 is lower than in 2017 with a decrease of about 88%. For T. magnatum and T. melanosporum, the quantity of DNA is higher in productive versus non-productive plots29,31,42. Thus, it is tempting to link the decrease in ascocarp production to the decrease in soil DNA. This orchard will continue to be analysed to gain more information about the putative link between ascocarp production and quantity of soil DNA.

Material and Methods

Experimental site

The T. aestivum orchard is located at Meux (45.442921, −0.358027) in the Nouvelle-Aquitaine region of France covering 1500 square meters. All the area interested by the plantation is characterized by a calcareous chalky. Wheat occupied the field before the plantation. After wheat harvest, the soil was returned and then tilled with rotary harrow. In 2004, 34 Corylus avellana, 10 Carpinus betulus, 10 Tilia cordata, 10 Quercus pubescens, and 10 Pinus nigra seedlings inoculated with T. aestivum were planted 4 m apart in the line and 6 m between lines (Fig. 1). Seedlings of the different species were planted randomly. In 2005, 43 additional C. avellana seedlings inoculated with T. aestivum were implanted between lines 7 to 17 (Fig. 1). Thus, there are two densities in the orchards: a high density (4 × 3 m equivalent to 833 trees per hectare) and a lower density (4 × 6 m equivalent to 417 trees per hectare). From 2004 to 2006, weeds were controlled by a monthly soil tilling with a rotovator and the proximity of trees was cleaned with a pickaxe. There was no soil tilling from 2007 to 2011; weeds were controlled using a brush-cutter and a few chemical treatments with glufosinate-ammonium (Basta, Bayer). Spores have been added every year since 2008 (from 5 to 100 g of crushed ascocarps). The production started in 2009, and the ascocarps were harvested with a trained dog from May to August. All harvested truffles were assigned to the closest tree. Slight pruning was realized since 2009 to avoid total closure of the orchard. The soil was tilled with a rotovator in September 2012, 2015, and 2017 to avoid soil compaction. In summer, the orchard was irrigated with potable water according to empirical criteria.

Development of a T. aestivum qPCR specific protocol

The T. aestivum genome has recently been sequenced18. In the U.S. Department of Energy Joint Genome Institute (DOE-JGI) genome browser (https://genome.jgi.doe.gov/Tubae1), T. aestivum orphan genes (i.e. genes present only in the T. aestivum genome and absent in the other genomes) were identified using the cluster analysis considering eight Pezizomycete species: Ascobolus immersus, Choiromyces venosus, Morchella importuna, Pyronema confluens, Terfezia boudieri, T. aestivum, T. magnatum, and T. melanosporum. Only clusters with one single gene copy in T. aestivum were considered. Two gene models (clusters 17,691 and 17,699) with numbers GSTUAT00001921001 (proteinId 4119) and GSTUAT00007989001 (proteinId 4108), respectively, were randomly selected. The primers TuGM4119F/TuGM4119R and TuGM4108F/TuGM4108R were then designed with Amplify4 v1.0 software in these gene models (Table 1). Primers were designed for use with qPCR via the SYBR® Green dye.

Table 1 Primers used in this study.
Full size table

The specificity of the two gene model primers was tested via PCR using 13 DNA samples from a first set of Tuber species ascocarps (Supplementary Table S2a): T. aestivum, T. mesentericum, T. magnatum, T. borchii, T. melanosporum, T. brumale, and T. indicum. The DNA of each sample was extracted using the DNeasy™ Plant Mini Kit (Qiagen SA, Courtaboeuf, France). A second set of DNA was extracted from 16 DNA samples of dried T. aestivum ascocarps of different geographical locations (Supplementary Table S2b). The universal fungal primers ITS1F and ITS243,44 (Table 1) were used to check that the DNA was suitable for PCR amplification (Supplementary Figs S7a and S8a). The PCR amplification was performed in a 10-µL mixture composed of 2X REDExtract-N-Amp™ PCR ReadyMix™ (5 µL), 10 µM primer forward (0.4 µL), 10 µM primer reverse (0.4 µL), 16 mg/mL bovine serum albumin (0.35 µL), water (1.85 µL), and 1/10 diluted DNA (2 µL). PCR reactions were carried out on a C1000 Touch™ Thermal Cycler (Bio-Rad, Schiltigheim, France) with denaturation at 94 °C for 4 min followed by 35 cycles of denaturation at 94 °C for 30 s, annealing at 55 °C for 30 s, extension at 72 °C for 45 s, and a final extension at 72 °C for 5 min.

To confirm that the ascocarps belonged to T. aestivum, PCR amplifications were performed with the T. aestivum specific primers TuITS1 and TuITS445 (Table 1; Supplementary Figs S7b and S8b). The PCR mixture was the same as those used for the ITS universal primers with the following thermal profile: denaturation at 94 °C for 4 min, 30 cycles of denaturation at 94 °C for 30 s, annealing at 65 °C for 30 s, extension at 72 °C for 45 s, and a final extension at 72 °C for 5 min.

The first set of 13 DNA samples were amplified with primers TuGM4119f/TuGM4119r and TuGM4108f/TuGM4108r (Table 1) by PCR with the same PCR mixture described above and according to the following conditions: denaturation at 94 °C for 4 min, 30 cycles of denaturation at 94 °C for 30 s, annealing at 65 °C for 30 s, extension at 72 °C for 45 s, and a final extension at 72 °C for 5 min. All PCR products were run on 2% agarose gel and visualized with a UV trans illuminator (Geldoc™ Universal Hood II, Biorad) after ethidium bromide staining.

According to the direct PCR results, real-time PCR amplification with SYBR® Green dye was then carried out on both sets of DNA with primers TuGM4108f/TuGM4108r only. The 20 µL of qPCR mixture was composed of Takyon™ No Rox SYBR® Core Kit (16.4 µL), 10 µM primer forward (0.8 µL), 10 µM primer reverse (0.8 µL), and template DNA (2 µL). The qPCR reactions were performed with a StepOnePlus™ Real-Time PCR System machine provided with the StepOne software v. 2.3 (Life Technologies, Applied Biosystem) on 96-well plates with carry over prevention at 50 °C for 2 min, Takyon™ activation at 95 °C for 3 min, 40 cycles of denaturation at 95 °C for 5 s, annealing at 60 °C for 15 s, and extension at 72 °C for 15 s.

Soil sampling, treatment, and DNA extraction

Three highly producing C. betulus were selected and equipped with two tensiometers (TEROS-21, Meter Group). These trees were randomly selected from the best producing trees (Supplementary Table S1). Two trees (II-9 and III-11) were selected from the 833 trees/ha area and tree VIII-5 was sampled from the 417 trees/ha area (Fig. 1). Tensiometers were installed 1 m from tree trunks: one at the north and another at the south—both at 12 cm depth. The TEROS-21 records soil water potential and soil temperature. Soil water potential is expressed as pF corresponding to the logarithm of the suction force expressed as centimetre height in a water column. The probes were installed on August 30, 2016. Soil sampling was performed every month 1 m from the trunk at the four cardinal points (North, West, East and South) under all the three trees at a depth of 5–10 cm. After sampling, the soil samples were immediately sent to the laboratory. They were initially dried separately about 2 hours at 40 °C and sieved using a 2-mm mesh and then stored at −20 °C until DNA extraction. The DNA of each sample was extracted using the DNeasy PowerSoil™ DNA Isolation Kit (Qiagen SA, Courtaboeuf, France) from 0.25 g of sieved soil. The extracted DNA was stored at −20 °C until analysis. For the molecular analysis described below, the four DNA samples around one tree were pooled to obtain one sample per tree and per sampling time.

Quantification of soil DNA

A standard curve was created as described in Parladé et al.32 by mixing a known amount of soil harvested in a field close to the truffle orchard with a known amount of fresh T. aestivum ascocarp. The absence of T. aestivum in the soil was first controlled by qPCR and the ascocarp was then ground in a mortar with liquid nitrogen. The DNA was extracted from the mixture in the same way as other soils; serial 10-fold dilutions were done to obtain a standard curve.

The soil T. aestivum DNA was measured with real-time PCR with SYBR® Green dye. The TuGM4108 primers (Table 1) were used to amplify the T. aestivum model gene in soil samples as described above. qPCR was performed on 96-well plates with three replicates for each sample, standard, and negative control. One sample corresponded to the mix of the four soil DNA extractions around one tree at one sampling date. The qPCR mixture and the thermal conditions were the same as those described above for the specificity assay. The software calculated the absolute quantity of T. aestivum in soil samples by interpolating the threshold cycle (Ct) values on the standard curve (R2 = 0.99; Eff = 96.02%).

Statistical analysis

The difference in daily pF and temperature between the north and south probes within the three productive trees and difference of daily pF and temperature (average between north and south probes) between the three productive trees were analysed by pairwise comparison using a Wilcoxon rank sum test (p < 0.05). The differences between qPCR replicates were analysed with multifactor analysis of variance (ANOVA), differences of T. aestivum DNA quantifications between sampling months were detected with Tukey’s HSD test (p < 0.05).

In order to compare the mean production per tree species, a subsampling was realized for C. avellana (Supplementary Fig. S1). Indeed, 34C. avellana were implanted in 2004 and 43 in 2005 although for all other species 10 plants were implanted in 2004. A random subsampling of 10C. avellana was realized 10 times among the 34 plants implanted in 2004 using sample function in R. Boxplot function in R was used to visualized the mean value obtained for the 10 subsampling. The differences were statistically tested with Tukey’s HSD test.

The relationships between T. aestivum soil DNA quantities and soil parameters (pF and temperature) were tested using two methods: (1) linear correlation using Pearson’s method with correlation coefficient and p-value calculation using cor.test function in stats R package; and (2) simple linear regression with lm function of the stats R package. The combined effect of pF and temperature was tested by a factorial ANOVA using aov function in R. Soil DNA quantities were log and square root transformed to fit normality and improve variance equality.