Meristem cultures of Atractylodes lancea (collected in Maoshan, Jiangsu Province, China) were established according to Wang et al. [²²]. The explants were surface sterilized and grown in MS medium [²³] supplemented with 0.3 mg/L naphthaleneacetic acid (NAA), 2.0 mg/L 6-benzyladenine, 30 g/L sucrose, and 10% agar in 150 mL Erlenmeyer flasks. Rooting medium (1/2 MS) contained 0.25 mg/L NAA, 30 g/L sucrose, and 10% agar. All media were adjusted to a pH of 6.0 before being autoclaved. Cultures were maintained in a growth chamber (25/18°C day/night, with a light intensity of 3400 lm/m² and a photoperiod of 12 h) and subcultured every four weeks. Thirty-day-old rooting plantlets were used for all treatments.

Reagents used as specific scavengers or inhibitors, including ibuprofen (IBU), nordihydroguaiaretic acid (NDGA), 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline −1-oxyl-3-oxide potassium salt (cPTIO), paclobutrazol (PAC), catalase (CAT), diphenylene iodonium (DPI), and 2-aminoindan-2-phosphonic acid (AIP), were purchased from Sigma-Aldrich (St. Louis, MO, USA). All exogenous signaling molecules and inhibitors were filtered using 0.22 μm diameter microporous membranes before use. Unless stated otherwise, inhibitors were applied 1 d before the application of signaling molecules or fungal inoculation.

The endophytic fungus AL12 (Gilmaniella sp.) was isolated from A. lancea, cultured on potato dextrose agar, and incubated at 28°C for five days [²⁴]. Thirty-day-old plantlets were inoculated using 5-mm AL12 mycelial disks. An equal size of potato dextrose agar was used as a control. All treatments were conducted in a sterile environment and replicated at least three times to examine reproducibility.

Thirty-day-old plants were incubated with fungal mycelia disks with or without inhibitors and were harvested 18 d later for determination of NO or H₂O₂. Inhibitors were 1.25 mmol L^-1 cPTIO, 5.25 mKat L^-1 CAT or 3 mmol L^-1 DPI.

The generation of H₂O₂ by A. lancea plantlets was measured by chemiluminescence in a ferricyanide-catalyzed oxidation of luminol according to Schwacke and Hager [²⁵], with modification. Leaf samples (1 g) were ground with 5 ml double distilled water. The homogenate was centrifuged at 13,000 g for 10 min, then 100 μL supernatant, 50 μL luminol (5-amino-2,3-dihydro-l,4-phthalazinedione), and 800 μL phosphate-buffered saline were mixed in a cuvette. The reaction was initiated with 100 μL K₃[Fe(CN)₆. To compare independent experiments, we used H₂O₂ as an internal standard. Fifty microliters of H₂O₂ (1 μM, freshly prepared) was added to the assay mixture containing 750 μL potassium phosphate buffer. One unit of H₂O₂ was defined as the chemiluminescence caused by the internal standard of 1 μM H₂O₂ per gram fresh weight.

The generation of NO was monitored using a NO detection kit (Nanjing Jiancheng Bio-engineering Inst., Nanjing, China) according to the manufacturer’s instructions. Leaf samples (1 g) were ground with 5 ml of 40 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (pH 7.2) and the homogenate was centrifuged at 14,000 g for 10 min. The supernatant was used for the NO assays. One unit of NO was defined as the absorbance variation caused by the internal standard of 1 μM NO per gram fresh weight.

At least 15 plantlets were assayed for each time point, and all treatments were performed in triplicate.

Thirty-day-old plants were incubated with fungal mycelia disks with or without inhibitors and were harvested 18 d later for determination of SA. Inhibitors were 1 mmol L^-1 PAC or 2.5 mmol L^-1 AIP.

Salicylic acid was extracted followed the method of Verberne et al. [²⁶], with some modifications. Five grams of whole plantlets was ground in liquid nitrogen and extracted in 2 ml methanol by sonication. After centrifugation at 14,000 g for 5 min, the supernatant was rotary evaporated, and the residue was resuspended in 250 μl of 5% trichloroacetic acid. The mixture was re-extracted with 800 μl acetic acid ester: cyclohexane (1:1 v/v). Finally, the organic phase was rotary evaporated until dry, dissolved with 600 μl HPLC mobile phase (methanol: 2% acetic acid: H₂O, 50:40:10, v: v: v), and filtered with a 0.22-μm microporous membrane for determination.

The SA samples were quantified by HPLC using a reverse-phase column (Hedera Packing Material Lichrospher 5-C18, 4.6 × 250 mm, 5 μm, Bonna-Agela Technologies, Wilmington, DE, USA). The mobile phases flow rate was 1 ml min⁻¹. Salicylic acid was detected at 217 nm at 25°C [¹⁴].

Thirty-day-old plantlets of Atractylodes lancea were incubated with 5-mm mycelial disks or PDA disks (control). Inhibitors (0.1 mmol L^-1 IBU or NDGA) were added 1 d before fungal inoculation for JA determination.

Volatile oils were extracted from whole plantlets of A. lancea, including leaves and rhizomes (0.8–1.6% oil content in leaves, 2.2–3.4% in rhizomes), according to Zhang et al. [²⁷]. The volatile oils were dried with anhydrous sodium sulfate and stored in dark glass bottles at 4°C for gas chromatograph (GC) analysis.

Following Juergen et al. [²⁸], JA was extracted by grinding plant material (1 g) frozen in liquid nitrogen and extracting with H₂O: acetone (30:70, v:v). Samples were store in dark glass bottles at −21°C for GC analysis.

GC determination was carried out using an 1890 series GC (Hewlett-Packard, Palo Alto, CA) equipped with a flame ionization detector. A DB-5 ms (30 m × 0.25 mm × 0.25 μm) column (Agilent, Santa Clara, CA, USA) was used with the following temperature program: column held at 60°C for 1 min after injection, increased by 10°C/min to 190°C, held for 2 min, increased by 5°C/min to 210°C, held for 2 min, increased by 10°C/min to 220°C, and held for 8 min. Nitrogen was used as carrier and the flow rate was 4 ml/min. Four main components of the volatile oils, atractylone, hinesol, β-eudesmol, and atractylodin, were quantitatively analyzed according to the method of Fang et al. [²⁹]; their retention times were 14.57, 15.24, 16.21, and 22.18 min, respectively.

Total RNA was extracted from leaves as described by Dong and Beer [³⁰]. First-strand cDNA was synthesized from 1 μg of total RNA (PrimeScript RT Reagent Kit, Takara, Dalian, China). Real-time qPCR was performed using the DNA Engine Opticon 2 Real-time PCR Detection System (Bio-Rad, Hercules, CA, USA) and SYBR green probe (SYBR Premix Ex Taq system, Takara). The constitutively-expressed gene EF1α used as an internal positive control. The gene-specific primers used to amplify EF1α were 5′-CAGGCTGATTGTGCTGTTCTTA-3′ and 5′-TGTGGCATCCATCTTGT-3′ (241 bp product) and for alHMGR were 5′-GGTGAGAAAGGTCCTGAAA-3′ and 5′-CATGGTAACGGAGATATGAA-3′ (154 bp). The GenBank accession numbers of the alHMGR and EF1α genes are EF090602.1 and X97131, respectively.

The thermocycler program was as follows: 90 s at 95°C; 40 cycles of 30 s at 95°C, 30 s at 57°C, and 30 s at 72°C; and 5 min at 72°C. To standardize the data, the ratio of the absolute transcript level of the alHMGR genes to the absolute transcript level of EF1α was calculated for each sample of each treatment.

Data were compiled using Microsoft Excel (Redmond, WA, USA). The values were represented as mean ± SD of three replicates for each treatment. Student’s t-test, one-way ANOVA, and Duncan’s multiple range test were used to identify significant differences (SPSS ver. 13.0, SPSS Inc., Chicago, IL, USA).

The JA contents of the plantlets increased significantly after endophytic fungus inoculation (Figure 1A), indicating that the fungus may trigger JA biosynthesis in the plantlets. Concurrently, the total amount of volatile oils increased significantly (Table 1). Both IBU and NDGA are inhibitors of the octadecanoid pathway that synthesizes JA and are usually applied in plant systems as JA-specific inhibitors [¹³]. To investigate whether JA was involved in the fungus-induced volatile oil accumulation, IBU and NDGA were applied; as shown in Figure 1B, both inhibitors suppressed not only the fungus-induced JA generation, but also the fungus-triggered volatile oil production. The results suggested that JA was important for fungus-induced volatile-oil synthesis in A. lancea plantlets. However, volatile oils in the A. lancea plantlets treated with both fungus and JA inhibitors could still accumulate, compared with the control, even though JA generation was lower than control (Figure 1B), implying that fungus-induced volatile oil synthesis is not solely dependent on the JA signaling pathway.

Previous results showed that JA is not the sole signaling pathway involved in fungus-induced volatile oil synthesis; NO, H₂O₂, and SA are also known to mediate this process in A. lancea plantlets [²²]. To investigate a possible relationships between JA and one or more of these other pathways, A. lancea plantlets were treated with the NO-specific scavenger cPTIO, the membrane NADPH oxidase inhibitor DPI/CAT, the SA inhibitor PAC/AIP, IBU, and fungal inoculation. The NO scavenger cPTIO could inhibit JA production in inoculated plantlets, but IBU could not inhibit NO production (Figure 2A), showing that JA may act as a downstream signal of NO. Exogenous H₂O₂ could reverse JA suppression, implying that JA is mediated by NO though H₂O₂ in endophyte-induced volatile-oil accumulation. In addition, the H₂O₂ inhibitor DPI/CAT could inhibit JA production, but IBU could not inhibit H₂O₂ production with inoculation (Figure 2A). The one-way dependence of JA on H₂O₂ confirmed that H₂O₂ was the intermediary factor between JA and NO.

Paclobutrazol is an effective SA biosynthesis-related benzoic acid hydroxylase (BA2H) inhibitor [³¹] that also inhibits gibberellin biosynthesis [³²]. Therefore, we also used AIP, a specific SA biosynthesis-related phenylalanine ammonialyase (PAL) inhibitor [³³,³⁴], to confirm that SA generation was suppressed. Interestingly, PAC and AIP could abolish the suppression of JA by DPI/CAT with fungus inoculation (Figure 2B). This result implied that the SA and JA signaling pathways were closely linked in endophyte-induced volatile-oil accumulation in A. lancea plantlets.

To further investigate the relationship between JA and SA, gradient concentrations of the JA-inhibiter IBU and the SA-inhibitors PAC and AIP were applied. As shown in Figure 3, the fungus-induced JA level of the plantlets decreased gradually as IBU concentration increased, but both SA accumulation and volatile oil content were enhanced as well, although the amounts did not exceed those obtained with fungal inoculation alone. Similarly, SA levels in plantlets were inhibited by 0.1, 1, and 2.5 mmol L^-1 AIP and by 3 mmol L^-1 PAC, whereas JA was enhanced significantly (Figure 3A). Volatile oil accumulation was enhanced by 2.5 mmol L^-1 AIP and 3 mmol L^-1 PAC (Figure 3B). The results suggested that JA may have a complementary interaction with SA to mediate fungal endophyte-induced volatile-oil accumulation. However, combining IBU and paclobutrazol could not completely inhibit volatile oil synthesis. We added the H₂O₂-inhibitor DPI/CAT to IBU and paclobutrazol, which reduced volatile-oil accumulation to the level of the control. The results suggested that H₂O₂, SA, and JA may work simultaneously in fungus-induced volatile-oil synthesis in A. lancea plantlets.

The enzyme 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR) catalyzes the conversion of HMG-CoA to mevalonate, which is the key step in the terpenoid biosynthesis pathway in plants [³⁵,³⁶]. We further investigated the possible mediating role of JA on HMGR gene expression. The results showed that exogenous JA could strongly stimulate HMGR gene expression (Figure 4A). Three sesquiterpenoid components of A. lancea volatile oils, atractylone, β-eudesmol, and hinesol, were all induced by JA and suppressed by IBU with fungal inoculation (Figure 4B).

Thirty-day-old plantlets were incubated with 5-mm mycelia disks or with an equal size of potato dextrose agar (control). Data are mean ± standard deviation (SD) of triplicate samples. Within each row, values followed by different lower-case letters were significantly different (one-way ANOVA, Duncan’s multiple range test, P <0.05).