Effects of jasmonic acid signalling on the wheat microbiome differ between body sites

Jasmonic acid (JA) signalling helps plants to defend themselves against necrotrophic pathogens and herbivorous insects and has been shown to influence the root microbiome of Arabidopsis thaliana. In this study, we determined whether JA signalling influences the diversity and functioning of the wheat (Triticum aestivum) microbiome and whether these effects are specific to particular parts of the plant. Activation of the JA pathway was achieved via exogenous application of methyl jasmonate and was confirmed by significant increases in the abundance of 10 JA-signalling-related gene transcripts. Phylogenetic marker gene sequencing revealed that JA signalling reduced the diversity and changed the composition of root endophytic but not shoot endophytic or rhizosphere bacterial communities. The total enzymatic activity and substrate utilisation profiles of rhizosphere bacterial communities were not affected by JA signalling. Our findings indicate that the effects of JA signalling on the wheat microbiome are specific to individual plant compartments.

Root and shoot endophytes. Relative to shoots, the diversity of root endophytic communities was richer (Sobs and Chao1) and more even (Simpson's Diversity Index) (R 2 > 83%, P < 0.001) (Figs 2 and S1). This is consistent with the fact that root endophytes typically derive from soil 7 and that shoot endophytes colonise either from root endophytic environments via the vascular tissue or enter via openings on stems and leaves 8,9 . The composition of endophytic communities also differed significantly between roots and shoots (R 2 = 88.9%, P = 0.002; Figs 3 and S2). Shoot endophytes were positively associated with members of the Shewanella (OTU [21][22] and a representative of the Halomonas (OTU 27) (Figs 3 and S2). Root endophytes were positively associated with representatives of the Streptomyces (OTUs [11][12][13][14] and members of the Actinosynnemataeae (OTU 1) and Glycomyces (OTU 4) (Figs 3 and S2). All of these taxa have previously been detected as endophytes in a wide-range of plant species. For example, representatives of the Halomonas have been observed in endophytic root and shoot environments of: Alopecurus aequalis 10 , Typha domingensis 11 and Arthrocnemum macrostachyum 12 . Shewanella spp. have been detected inside potato tubers 13 , rice roots 14 and baby spinach leaves 15 . Actinobacteria, particularly Streptomyces spp., are frequently isolated from endophytic root and shoot environments of maize (Zea mays L.) 16 , rice 17 , tomato 18 and wheat [19][20][21][22] and members of the Streptomycetaceae are key components of endophytic communities in Arabidopsis thaliana roots 23,24 . The influence of JA signalling on the diversity of root and shoot endophytes. Activation of JA signalling led to a significant reduction in the richness (P < 0.001) and evenness (P < 0.001) of root, but not shoot, endophytic communities (Figs 2 and S1). This novel finding may indicate that when under attack plants have evolved a mechanism to generally suppress microbial colonisation. However, absolute rather than relative abundances are needed to test this hypothesis. Previous studies have also reported no effects of JA signalling on the diversity of endophytes associated with aerial parts of plants 25 . Root endophytic communities may be more responsive to JA signalling because, relative to aboveground environments, soils harbour more organisms and, therefore, more potential attackers. Activation of JA signalling also led to a significant change in the composition of root, but not shoot, endophytic communities (P = 0.011; Figs 3 and 4 and S2). Relative to the control, MeJA treatment significantly increased the relative abundances of a Actinosynnemataeae (OTU 1) and a Streptomyces Figure 1. The effect of MeJA application on the transcription of genes associated with the jasmonic acid (JA) signalling pathway in 10-day-old wheat seedlings. Asterisks indicate significant differences between control and MeJA treated plants (*P < 0.05, **P < 0.01, ***P < 0.001, two-tailed student's t test). Error bars represent standard errors of the means (n = 3). (OTU 11) population, and decreased the relative abundances of a Glycomyces (OTU 4) population and several members of the Streptomyces (OTUs 12-14) (Fig. 4). All of these taxa are members of the Actinobacteria, which include many populations that have been shown to promote plant growth, mobilise nutrients and suppress bacterial, fungal or viral phytopathogens [26][27][28][29][30] . For this reason, the observed changes in the relative abundances of actinobacterial populations in our study, may have had functional consequences for the host, which deserve further investigation in future studies.
Rhizosphere and bulk soil microbial communities. Activation of the JA pathway did not significantly influence the richness, evenness or composition of bacterial communities associated with the rhizosphere or bulk soil (P > 0.05) (Figs 2 and 5 and S1). Likewise, activation of the JA pathway did not influence the total enzymatic activity or substrate utilisation profiles of microbial communities associated with rhizosphere or bulk soil (Fig. S3). While all previous studies indicate that JA signalling has no effect on the richness or evenness of rhizosphere bacterial communities 3,31 , the effects on bacterial community composition are inconsistent. When grown in soil collected from areas where A. thaliana grows naturally, stimulation of the A. thaliana JA pathway led to a significant alteration in rhizosphere bacterial community composition 3 . However, when grown in 'non-native' soils, induction of the A. thaliana JA pathway had no effect on the composition of rhizosphere bacterial communities 31 . This suggests that JA pathway-mediated effects on rhizosphere bacterial communities may be influenced by soil type and the length of association between a particular plant genotype and soil. The soil selected in our study had a long cropping history of wheat but we did not detect any effects on rhizosphere bacterial communities within three days of JA signalling. This does not rule out the possibility that effects may become apparent over longer time periods or for plants grown in other soils.
Effects of JA signalling on root and shoot biomass. Relative to the controls, MeJA treatment led to a 14% reduction in root dry weight (P = 0.015) but shoot biomass was not affected (Fig. S7). This is consistent with previous studies in Arabidopsis thaliana 34,35 and sunflower (Helianthus annuus L.) 36 , which reported root inhibition upon activation of JA signalling.

Conclusion
Our study demonstrates that activation of JA signalling in wheat reduces the diversity and changes the composition of bacterial communities in endophytic roots but not in shoots or in the rhizosphere. Most of the root endophytic populations that became more abundant in response to JA signalling were closely related to taxa previously reported to suppress bacterial, fungal or viral phytopathogens, promote plant growth or mobilise nutrients [26][27][28][29][30] . JA signalling also led to a decrease in root biomass, which suggests that plants prioritise defence over growth when under attack. We hypothesis that the change in root endophyte communities in response to JA signalling may reflect a coevolved mechanism by which plants recruit microbial symbionts that enhance host biotic stress tolerance when under attack.

Materials and Methods
Plant growth conditions and experimental design. Wheat (Triticum aestivum) seeds (Crusader variety) were pre-germinated on a moist filter paper in a petri-dish for 36 h and then planted in 30-well punnet trays with three seeds per well (Fig. S5). Plants were grown in soil collected from 0-10 cm depth in a long-term wheat paddock in Condamine, Queensland, Australia (26.90°S, 149.64°E). Key physicochemical characteristics of this soil are summarised in Table S1. The soil was a mesotrophic effervescent Brown Sodosol developed on Cainozoic sand plains and had been under no-till management for 19 years. This paddock has a long cropping history of wheat and the previous crop on this soil was also wheat. The soil contained 25% clay, 14% silt and 61% sand and was homogenised prior to planting using a 2.4 mm sieve. Two additional trays were filled with soil but were not planted (Fig. S5). All trays were transferred to a controlled environment chamber (Percival Scientific, Boone, IA, USA) at 20 °C with a photoperiod of 12 h and light intensity of 150 mmol m −2 s −1 . Throughout the experiment, the plants were watered once per two days with an amount ~10 mL per well, and the positions of the trays within the growth chamber were changed on a daily basis.
After 10 days (two-leaf stage), the JA signalling pathway was activated by exogenously applying methyl jasmonate (MeJA) as previously described 3 . Briefly, 300 μ L, 0.5% (v/v ethanol) of MeJA was applied on a cotton ball attached to the lid of the tray to create an atmosphere containing 0.025 μ L MeJA L −1 . The tray was then immediately sealed with tape and enclosed in two sealed transparent plastic bags. The same procedure was repeated for the control plants but MeJA was omitted and 300 μ L of ethanol which was the solvent used to prepare MeJA solution was applied to the cotton ball. To determine whether MeJA led to any direct effects on soil microorganisms The asterisks indicate significant differences between treatments (*P < 0.05, **P < 0.01, ***P < 0.001, two-tailed student's t test). Each OTU has a unique numeric identifier shown in square brackets that is consistent with those shown in other figures.
Scientific RepoRts | 7:41766 | DOI: 10.1038/srep41766 one of the unplanted trays was treated with 300 μ l MeJA solution and compared to another tray that was treated with 300 μ l ethanol. We included three replicates per treatment. Each plant replicate comprised a pool of 30 plants.
Sample collection. Bulk soil and rhizosphere samples. All samples were collected 72 h post-MeJA treatment (Fig. S5). For bulk soil samples, soil was collected in sterile tubes and then stored at − 80 °C until further processing. For rhizosphere soil samples, roots were carefully removed from each pot, excess soil was removed by shaking and that remaining closely adhered to the roots was considered to be rhizosphere soil 3 . For DNA extraction, rhizosphere soil was recovered by shaking roots in sterile 50 ml tubes each containing 25 ml sterile phosphate buffer (Na 2 HPO 4 7.1 g, NaH 2 PO 4 ·H 2 O 4.4 g, amended to 820 mL, pH 7.0, 0.1 M) for five min at 250 rpm. After shaking, roots were transferred to new tubes and rhizosphere soil was pelleted by centrifugation at 12,000 g for 3 min then transferred to − 80 °C storage until further processing. For MicroResp TM (James Hutton Institute, Invergowrie, Scotland, UK) 37 , rhizosphere soil was physically separated from roots using sterile gloves.
Root and shoot endophytic samples. After removal of rhizosphere soil, root tissues were washed with distilled water and 0.1% Silwet L-77 in phosphate buffer three times 38 , sonicated at 20 kHz for five min to remove rhizoplane microorganisms 24 , washed in sterile phosphate buffer, air dried, ground in liquid nitrogen and then stored at − 80 °C for DNA extraction. For shoots, half of the tissues were immediately submerged in liquid nitrogen and stored at − 80 °C for RNA extraction (Fig. S5). The other half were washed with 0.1% Silwet L-77 in phosphate buffer three times, surface sterilised using 0.5% (v/v) hypochlorite for two min, air dried, ground in liquid nitrogen and then stored at − 80 °C for DNA extraction. Determination of plant growth. The MeJA treated and non-treated wheat seedlings were collected 72 h post-treatment and root attached soils were thoroughly removed by washing under distilled water. Thirty plants were pooled in each bioreplicate, and three bioreplicates were included for each treatment. Shoots and roots samples were cut to separate and oven dried (65 °C) for three days, and then the weight of wheat roots and shoots were recorded.
Quantification of JA signalling pathway-related transcripts. Total RNA was extracted from wheat shoots using the SV Total RNA Isolation Kit (Promega) according to the manufacturer's recommendations. The cDNA was synthesised by reverse transcription of 1.5 μ g of total RNA using the Superscript III kit (Life Technologies) and both random hexamers and oligo dT primers. Quantitative real-time PCR (qRT-PCR) assays were performed on a ViiA ™ 7 sequence detection system (Applied Biosystems, USA). Ten JA defence-related genes in wheat, namely PR1.1, PR2, PR4a, PR5, PR9, WCI2, WCI3, CHI3, TaAOS and LIPASE were examined for gene expression in shoots. Primer sequences are shown in Table S2. The wheat 18S rRNA gene was used as an internal reference gene for normalisation. PCR conditions and the relative expression of each target gene was investigated as previously described 6  Processing of sequence data. Data were processed as described previously 40 . Briefly, sequences were quality filtered and dereplicated using the QIIME script split_libraries.py with the homopolymer filter deactivated 41 , checked for chimeras against the GreenGenes database (October 2013 release) using UCHIME ver. 3.0.617 42 , homopolymer error corrected using Acacia 43 and then subjected to the following procedures using QIIME: (1) OTUs were picked at 97% similarity, (2) OTU representative sequences were assigned GreenGenes (October 2013) taxonomy using BLAST, and then (3) tables with the abundance of different operational taxonomic units (OTUs) and their taxonomic assignments in each sample were generated. The number of reads was rarefied to 1,250 per sample to allow comparisons of diversity without the bias of uneven sampling effort. The mean number of OTUs (observed richness) and Simpson's Diversity Index values corresponding to 1,250 sequences per sample were calculated using QIIME.
Microbial community activity. Community-level physiology profiles (CLPPs) were generated by characterising the induced respiratory responses of microorganisms associated with 0.4 g of each soil sample to 20 substrates using MicroResp TM 37 as described in Liu et al. 44 . The substrates included carboxylic acids (citric acid, methyl pyruvate, oxalic acid, D+ galacturonic acid and succinic acid), carbohydrates (beta-d-fructose, D-(+ )trehalose, D-glucose, L-malic acid, D-xylose, mannitol, L-(+ ) Arabinose, cellulose), amino acids (L-alanine, gamma-aminobutyric acid, L-arginine, L-Asparagine), urea, uric acid and tween 40. Milli-Q water was added to controls. Fluorescein diacetate (FDA) hydrolysis assays were used to provide a measure of total microbial enzyme activity and were performed as described by Green et al. 45 .

Statistical analyses.
The effect of MeJA treatment on enzyme activities and the richness and equitability of bacterial communities was investigated using ANOVA. Differences in transcript abundances and wheat dry weights were assessed using two tailed t-tests. The effects of MeJA treatment on the composition of bacterial communities and on substrate utilisation patterns were investigated using Permutational Multivariate Analysis of Variance (PERMANOVA). PERMANOVA was performed using Hellinger transformed OTU abundances. Differences in the abundances of individual OTUs between treatments were identified using ANOVA with posthoc Tukey's HSD tests. All analyses were implemented using R (version 2.12.0). Differences in the composition of microbial communities or the utilisation of substrates between samples were visualised using principal component analysis (PCA) and/or heatmaps.