The first experiment was set up in a greenhouse under controlled conditions (23±5ºC) with 32 plants equally distributed in four treatments in a 2 × 2 factorial design, consisting of non mycorrhizal plants inoculated or not with A. mellea and mycorrhizal plants inoculated or not with A. mellea. Lavandula angustifolia plants (10±2 cm height) obtained from rooted cuttings were inoculated with Rhizophagus irregularis (former Glomus intraradices BEG 72) [(Blaszk., Wubet, Renker & Buscot) C Walker & A Shüssler comb nov], a native AMF from Spain originally isolated from Citrus (Camprubi & Calvet, 1996) with proven effectiveness in different agricultural systems involving root pathogens infestations (Calvet et al., 2001; Camprubi et al., 2008). Plants were placed in 2-liter containers filled with a mixture of pasteurized sandy soil, sphagnum peat and quartz sand (3:2:1; v/v/v) of pH 7.7. Each container in the mycorrhizal treatments received 10 mL of R. irregularis bulk inoculum obtained from leek (Allium porrum L.) cultures, containing spores (at least 1000 spores in 10 mL) and colonized root fragments. The inoculum was placed below the plant root systems at planting.

The Armillaria isolate was obtained from a heavily infested peach (Prunus persica L.) replant orchard. It was molecularly identified as A. mellea (Mansilla et al., 2000), grown in pure cultures on malt agar, and used to inoculate autoclaved chestnuts (Castanea sativa L) in sterile malt agar medium (Difco®) supplemented with 0.2 g/L streptomycine sulphate and 10 mL/L benomyl. After a 5-month incubation period at 25ºC in the dark, infected chestnuts were used as individual sources of pathogenic inoculum.

One month after the mycorrhizal inoculation, lavender plants were placed in an open-air shade house and each container of the pathogen-inoculated treatments received one infected chestnut, placed at 10-cm depth and 10-cm distance from the plant stem. Fertilization was applied every two weeks (10 mL Hoagland’s nutrient solution per plant). Plants were harvested after ten months, when decay symptoms were apparent on the plants. Plants were cut to measure shoot and root weights, largest plant diameter and number of spikes. The extent of mycorrhizal root colonization was measured in plants inoculated with R. irregularis and A. mellea damage was estimated in the root system of pathogen-inoculated plants. In order to assess R. irregularis internal infection, root samples from each inoculated plant were clarified and stained (Phillips & Hayman, 1980) and the percentage of root colonized by the mycorrhizal fungus was measured with the gridline intersect method (Giovannetti & Mosse, 1980). Disease symptoms caused by A. mellea were recorded in plants that had received the infected chestnuts. Two levels of root necrosis and crown necrosis were determined: “High” (infection ranking from 25% to 100% of the root system length) and “Low” (infected root below 25%) for root necrosis; “High” (necrotic tissue with presence of white mycelium) and “Low” (necrotic tissue with no mycelium) for crown necrosis.

Data obtained for plant growth were analyzed in a multifactorial ANOVA and means were compared by Tukey’s test when there was no interaction between factors: pathogen and mycorrhiza.

In order to evaluate the interaction between R. irregularis and A. mellea, in medium with different calcium (Ca) levels, L. angustifolia plantlets were used to establish mycorrhizal autotrophic culture systems described by Voets et al. (2005). In such systems, the plant leaves perform photosynthesis while roots grow axenically inside a petri plate filled with modified Strullu-Romand medium (MSR) with no sugar or vitamins (Voets et al., 2005) and 7 g agar/L. Under such conditions, the root can be colonized by an AMF monoxenically grown on Agrobacterium rhizogenes-transformed roots in a root organ culture (ROC) (Declerck et al., 1996). In our experiment, the procedure followed to obtain the tripartite association (lavender plant-AMF-root rot fungus) was basically the same described by Nogales et al. (2010) for grapevine plants. Since the amount of growing medium in a 100 mm petri plate was insufficient to sustain grapevine growth longer than a week, requiring the system to be refilled periodically, which increased the risk of contaminations, we used 350-mL autoclaved plastic containers, in order to avoid unnecessary manipulation (Lovato et al., 2014). Lavandula angustifolia seeds were surface-sterilized in 32% sodium hypochlorite for 30 min and thoroughly washed with sterile distilled water three consecutive times for 10 min. The seeds were germinated on sterilized filter paper in petri plates at 25ºC in the dark, and plantlets with incipient roots were transferred to sterilized glass pots (Le pratique®) containing 150 mL of Murashige Skoog (MS) (Murashige & Skoog, 1962) medium with vitamins (Duchefa Biochemie) and supplemented with 30 g/L saccharose and 7 g/L agar (Difco). Plants were grown for 10 weeks in a growth chamber at 25ºC and 16 h photoperiod (200 µmol/m²/s).

The same four treatments described in the greenhouse experiment were adopted for the in vitro experiment: non mycorrhizal plants inoculated or not with A. mellea and mycorrhizal plants inoculated or not with A. mellea. The containers not inoculated with A. mellea were directly filled with 350 mL of modified MSR medium.

In A. mellea treatments, 80 mL of MSR medium were first poured into the containers, cooled and solidified (Lovato et al., 2014). Two 7-mm diameter plugs of A. mellea mycelium from a pure culture were inserted in the medium after removing agar plugs which were replaced afterwards. Each container received then 270 mL of the same modified MSR medium, previously cooled at 40ºC. An incision was made on the lid of the container, through which a lavender plant was inserted, in order to keep the shoot outside the container and the root inside the growing medium. At the same time, the inoculation with the AMF in mycorrhizal treatments was performed using a six-month old ROC culture in a bi-compartmented dish of R. irregularis (Nogales et al., 2010). Two 1-cm² plugs containing 100 to 150 mycorrhizal spores were excised from the compartment lacking roots and introduced next to the root tips, following the same procedure used for A. mellea plugs. Once both inoculations were completed, the container was carefully closed, the space around the stem in the incision hole was sealed with sterile silicone (Panreac) and parafilm was wrapped around the lid.

In order to evaluate the effects of pH, the growing medium in the container was supplemented with increasing levels of CaCO₃ (Panreac): 0, 100 mg/L and 3000 mg/L, which resulted in different pH values of 5.5, 7.0 and 8.5, respectively. There were eight replicates for each pH treatment, and lavender autotrophic systems were placed in a growth chamber at 25ºC.

After nine weeks plants were harvested, shoot dry weight and root fresh weight were recorded, and morphological symptoms of A. mellea infection were observed. Root samples were excised in order to estimate mycorrhizal colonization after clarifying and staining (see the greenhouse experiment) and to quantify A. mellea by real-time PCR (qPCR).

For quantification of the pathogenic fungus 1.0 g root samples were separated and ground in mortar and pestle with liquid nitrogen. Genomic DNA was extracted from a 70-100-mg (fresh weight) sample of ground roots with phytoplasma grinding buffer (PGB) (Arhens & Seemüller, 1992). Specific primers designed by Baumgartner et al. (2010) to amplify a fungal gene for the nuclear elongation factor subunit 1-alpha, EF1a-F1 (5GGATGGCACGGTGATAACAT) and EF1a-R1 5AGTCTTGCCCTTGACGACAC), were used to amplify a 150-bp section of the EF1a gene.

Real-time PCR was carried out using the Step One™ Instrument (Life Technologies) and the SYBR^® Premier Ex Taq^™ kit (Takara, Shiga, Japan). We used a 25 µL reaction volume containing the following components: 12.5 µL SYBR green Super mix, 1 µL of each forward/reverse (F/R) primers (10 µM), 2 µL DNA templates (20 ng/L of DNA), 8 µL sterile HPLC water and 0.5 µL ROX reference dye. Triplicate reactions were routinely used for each sample and each set included controls containing water to check for contamination in the reaction components. The qPCR cycling program was: 10 min at 95°C, followed by 40 cycles at 95°C for 15 s, 60°C for 45 s, and a dissociation stage with 15 s at 95°C, 30 s at 60°C and 15 s at 95°C. The qPCR signal was calibrated by making a standard curve with DNA extracted by the previously described method from known amounts of mycelium from a pure A. mellea culture. The cycle threshold (Ct) was calculated for all the standards and samples to indicate the number of cycles required for the fluorescence signal to cross the threshold above background during the exponential PCR amplification phase. DNA quantities were adjusted for the amount of root material used.

The A. mellea isolate used was highly virulent: five out of eight non-mycorrhizal plants infected by A. mellea were dead at harvest, while all mycorrhizal plants survived and showed less severe disease symptoms (Table 1). None of the mycorrhizal lavender plants showed a high level of necrotic root tissue and only one among them attained a high level of crown necrosis. The presence of A. mellea in R. irregularis-inoculated roots had no significant effect on the root system mycorrhizal colonization extent, that reached mean values of 36±14 % in plants inoculated only with the mutualistic fungus and 29±12% in plants with the dual inoculation. Growth parameters measured at harvest (Fig. 1) indicated that both the AM fungus and the pathogen had significant effects on lavender growth. The effect of mycorrhizal inoculation was highly significant at increasing shoot biomass and plant shoot diameter. Infection by A. mellea lowered shoot biomass and the number of spikes produced at flowering. However, dual inoculated lavender plants were capable to overcome root-rot infection, as there were no significant differences in the number of spikes produced by mycorrhizal plants inoculated with the pathogen (Fig. 1) and by healthy plants. Additionally, they reached the same plant diameter coverage than mycorrhizal plants not infected with the pathogen.

Armillaria mellea was detected by qPCR only in part of the plants inoculated with R. irregularis. For instance, the pathogenic fungus was detected only in one out of eight mycorrhizal plants grown in pH 7.0 medium and in none of the mycorrhizal plants grown in pH 8.5 medium. Consequently, the quantification analysis was discarded and a contingency analysis of chi-square values followed by Pearson’s test (JMP^® 8.0.1 SAS) was used to analyze the effect of mycorrhizal inoculation with R. irregularis on the detection of A. mellea in lavender roots grown in autotrophic systems.

There were significant differences in Armillaria detection due to AM fungal inoculation when substrate pH was increased by CaCO₃ addition (Table 2). At pH 5.5 A. mellea was detected in equal amounts in mycorrhizal and non-mycorrhizal plants, since more than 70% of the plants from both treatments were positive for the pathogen’s presence. When plants grew in higher pH media, detection of A. mellea decreased only in mycorrhizal plants. The percentage of detection in dual-inoculated plants grown in pH 7.0 medium was only 17, and it was null in plants grown in medium with pH 8.5. Differences in A. mellea detection due to pH values were significant according to Pearson’s test (p=0.0036) only for plants inoculated with both fungi, R. irregularis and the pathogen. On the other hand, growth medium pH in autotrophic systems had no significant influence on the detection of A. mellea in non-mycorrhizal plants (Table 2).

Mycorrhizal plants grown in pH 5.5 medium without A. mellea had only 34±3 % of the plant root system colonized by the mycorrhizal fungus (Table 3), while this percentage attained 61±3 % at pH 8.5. The value for dual inoculated plants was similar, as 25±12 % of the root system was mycorrhizal in plants grown in 5.5 pH medium, while in plants grown in pH 8.5 medium this index was 66±23 %. Armillaria mellea infection had no effect on mycorrhizal root colonization after 10 weeks. The percentages of mycorrhizal colonization in plants inoculated with the AMF or with both fungi were similar (34±3 vs 25±12) and (61±3 vs 66±23) at pH 5.5 and pH 8.5 respectively.