Plant material and growth conditions
Cicer arietinum var. Sonali and var. Kyabra, both previously rated as very susceptible to Phytophthora root rot (Bithell et al. 2021), were used in this experiment. Genotypic variation in AM colonisation was previously shown for these two varieties with Kyabra exhibiting higher mycorrhization levels (~21% root length colonised) when compared to var. Sonali (~7.5% root length colonised) when inoculatedin sterilized soil with Funneliformis mosseae (Plett et al., 2016). Surface seed sterilization was performed by soaking the seed in 4% bleach solution for 15 min followed by three 5-minute washes with sterile water.Soil, obtained from a chickpea paddock in Tamworth, NSW, Australia (31.0900° S, 150.9293° E) was mixed with sterile sand in a ratio of 1:1 after which it was sterilized by gamma-irradiation (50 kGy; Steritec NSW). The final plant-available phosphorus concentration was 15 ppm (Colwell). There was no detectable difference in %N by weight (0.1% in sterile and non-sterile conditions). Prior to planting, a microorganism filtrate that contained a living soil microbe community from unsterilised soil of the same location, but which excluded AM fungal spores, was added to the irradiated soil. This soil microbial wash, with the purpose to reinstate communities of other microorganisms into the sterile soil, was obtained by combining and mixing one part of non-irradiated soil with three parts of demineralised distilled water (100 g soil + 300 g water). After 10 minutes of resting, this mixture was washed through a series of sieves with an increasingly smaller pore size (2mm > 125 µm > 20 µm). This mycorrhiza‐free filtrate was collected and added to the soil:sand mix (50 mL kg-1).The seeds were germinated in the soil:sand mix and chickpea plants were cultivated in round pots (9x4 cm) in a blocked arrangement in a plant growth chamber under controlled conditions: 15 h light/ 9 h dark cycle at 18˚C, 70% relative humidity, and 3,500 μmol m-2 s-1 light intensity.Plants were watered daily with distilled waterand fertilized every week starting after week two with 0.5 mL/pot of Long Ashton solution minus nitrogen and phosphorus: 2mM K2SO4, 1.5mM MgSO4 · H2O, 3mM CaCl2 · H2O, 0.1mM FeEDTA and 1 ml Micronutrients (2.86 g l-1 H2BO3, 1.81 g l-1 MnCl2 · 4H2O, 0.22 g l-1 ZnSo4 · 7H2O, 0.08 g l-1 CuSO4 · 5H2O, 0.025 g l-1 NaMoO4 · 2H2O)
AM fungal inoculum
A commercial inoculum (Startup Ultra, Microbesmart, Adelaide, SA, Australia) that contains four isolates of the AM fungus R. irregularis (previously known as Glomus intraradices) (Schüßler and Walker 2010), was used in this experiment. To remove the small particles of Calcined diatomaceous earth, which is an inert carrier used in this commercial inoculum, 10 g of inoculum was mixed in demineralised distilled water and sieved through 2 mm > 125 µm >20 µm sieves and the filtrate on top of the 20 µm sieve, containing the AM fungal spores, was collected. Half of this filtrate was used for the AM fungal treatment (+AMF) and the second half was autoclaved twice and added to the non-AM fungal treatments (-AMF). The +AMF soil mixture was prepared by mixing 300 g of sterilized soil with half of the filtrate (112.5 ml) containing the AM fungal spores (10 g of inoculum contains at least 10,000 Spores). The pot set-up was as follows: 30 g of soil (without AM fungal inoculum) was added to each pot followed by either 7.5 g of +AM fungal soil mixture (for +AMF treatments) or 7.5 g of AM-free soil (for -AMF treatments). On top of this mix, one seed of either chickpea variety Sonali or Kyabra was placed and covered with 10g of sterile soil (Figure1a). The autoclaved AM fungal spore suspension with the same number of spores per plant was equally distributed and added to each pot of the -AMF treatment. Subsequently, the plants were allowed to germinate and grow for 25 days to enable the establishment of the AM fungus based on previous literature (Gutjahr et al. 2015; Renaut et al. 2020), and root colonisation after which pathogen treatments were established as outlined below (Figure1a). We attempted at the end of this experiment to establish the level of root colonization. However, the experiement was run just before COVID-related laboratory shutdown meant that the roots had to be stored for an extensive period of time before observation. The ultimate result was that the fragile root system of the plants used in this experiment disintegrated while performing the root staining method, which made reliable root colonisation assessment unobtainable. Therefore, as we cannot claim that these two genotypes show different colonisation by R. irregularis in the exact experiment, we instead disucss our results (i) by focusing on plant genotype-specific responses, (ii) we use terminology “AM inoculation” instead of “AM colonisation”.
Rhizobia and Pathogen inoculation
As we wished to mirror conditions of chickpea growth as they would experience in an agricultural setting, we inoculated all plants with their nitrogen-fixing symbiotic bacterial(rhizobia)symbiontMesorhizobium ciceri (isolate CC1192; Nour et al. 1994; Laranjo et al. 2014). Based on previous findings, co-inoculation with Rhizobia does not influence AM fungal colonisation levels (Tavasolee et al. 2011). Ten days after planting the seeds, one mL of M. ciceri suspension (10,000 CFU mL-1) was added to each plant by injection into the soil near the stem. Two weeks following inoculation with M. ciceri, and 24 days post-seeding, the plants were inoculated with the pathogen Phytophthora medicaginis isolate 7831. This culture, isolated from a chickpea paddock by the New South Wales Department of Primary Industries in Mungallala, QLD, Australia (26.4466° S, 147.5436° E), was grown on V8 juice agar plates (containing 200 mL L-1 V8TM bottled juice; 3.0 g L-1 CaCO3, 15.0 g L-1 agar) at 25˚C in the dark for a minimum of 5 weeks to allow for optimal oospore production. At the time of inoculation, oospores were collected, and their concentration was counted by a haemocytometer. This solution was diluted to a low number of oospores (700 oospores mL-1) and 1 mL of this solution was added to each pot following flooding with distilled water to a level of 1 cm above the soil level. As both chickpea varieties are susceptible to P. medicaginis, oospore concentration and flooding time were limited to prevent excessive disease pressure. After 48 hours, the water was drained and this was chosen as the starting time point of pathogen inoculation (i.e. T0). For mock inoculation, the same procedure was utilized with the exception that the oospore inoculation had been heat sterilized via autoclaving twice and cooled to room temperature prior to addition to the pots.
Experimental design
A factorial design including both chickpea varieties inoculated (or not) with one or both of AM fungi and pathogen, with one plant per pot (40 pots in total) and five biological replicates (n=5), was used in this experiment (Figure 1b). The treatments comprised non-inoculated control plants (Contr), plants inoculated with the R. irregularis AM fungal mix (AMF), pathogen-inoculated plants (Path), and plants inoculated with a combination of AM fungus and pathogen (AP). The AM fungus and pathogen-free control treatments (Contr) contained the heat-treated non-viable pathogen and autoclaved non-viable AMF inoculum (Figure 1a).
Harvest
The destructive harvest was conducted after 11 weeks of plant/AM fungal growth and 53 days post-pathogen inoculation (dpi), when mild brown lesions on the lower stems started to be visible in Path plants. The shoots were collected for fresh and dry weight measurements. For the untargeted metabolomic analysis, lateral root tissue samples without nodules from the top half of the root (Figure 1a) of each chickpea plant were collected, snap-frozen in liquid nitrogen and stored at -80 °C until further processing.
Metabolite extraction and untargeted metabolite profiling
The frozen root samples were ground in 200 μL of cold extraction solution (4:4:2 methanol:acetonitrile:water) using an MP Biomedical bead mill for 30 seconds at a frequency of 6.0 Hz. Following the addition of a further 300 μL of extraction solution, samples were sonicated in a 4°C water bath for 25 minutes followed by 10-minute centrifugation (21,139xg, 4 °C). The supernatant was collected and stored at -80°C until analysis. Prior to analysis, the samples were diluted 4x (based on sample QC dilution series) and the mixtures were centrifuged for 5 minutes at 21,139 x g to remove particulates. Subsequently, 200 μL of supernatant was transferred into 96-well sample collection plates (Waters) in a randomised manner. The final untargeted metabolite profiling, including all biological samples as well as a method blank and a solvent blank, was performed using an ACQUITY UPLC I-Class FTN coupled to a Waters Synapt G2-Si HDMS instrument (Waters, Wilmslow, United Kingdom) with Unispray ionization. 2 µL of sample metabolite extract was injected onto the same column ( ACQUITY UPLC HSS T3 1.8 µm 100 × 2.1 mm Column at 35 °C) from 96-well 1mL sample collection plates kept at 4 °C, alongside pooled biological quality control sample (created by combining small aliquots from several samples across all treatment types). Injection was eluted at a flow rate of 0.5 mL/min with a 9 min gradient, with mobile phase A held at 99% for 1 min, decreased to 85% over 1 min, decreased to 50% over 2 min, decreased to 5% over 2 min and increased to 99% over 2 min. The mobile phases consisted of: Mobile phase A containing (Water + 0.1% Formic Acid) , and mobile phase B containing (Acetonitrile + 0.1% Formic Acid). Mass spectral data from 50-1200 m/z were collected in the positive and negative electrospray ionization mode. Data acquisition was performed using HDMSe with ion mobility separation followed by mass fragmentation and Resolution mode mass analysis. The scan time was 0.2 s and the elevated energy transfer collision voltage was 20–50 eV. For this experiment, the instrument was run in positive ionisation mode with the following settings: Capillary: 0.5 kV, source temperature: 120 °C, sampling cone: 30 V, source offset: 80 V, desolvation temperature: 500 °C, desolvation gas flow: 800 L/h, cone gas flow: 20 L/h. Leucine Enkephalin solution (Waters, 200 pg/mL) was used for lockspray.
Metabolomic data analysis and metabolite annotation
Automated data processed of acquired raw data including peak alignment, peak picking, and deconvolution, was conducted with the program Progenesis QI, version 3.0 (Nonlinear dynamics, Waters Corporation, UK). The data from the positive ion mode (determining the mass-to-charge ratio after positive ion formation) are used throughout the study and information regarding the negative ion (determining the mass-to-charge ratio after negative ion formation) mode can be found in the supplementary material (Supplementary Figure S1, Supplementary Table S3). These processing steps revealed information about peak intensities, which were further used for statistical analysis. Potential metabolite feature identification and annotation of the observed peaks were obtained using the ProgenesisQI support for the web-based Chemspider structure database, including the public databases ChEBI, Phenol-Explorer, PlantCyc, KEGG and Golm Metabolome Database with a precursor tolerance 15 ppm and fragment tolerance 50 ppm. Mass error, isotope similarity and fragmentation score were used to calculate a confidence score for each potential identification. Potential compound identifications were assigned a chemical classification based on their primary structural features.
Statistical analysis
Statistical analysis of peak intensities was performed using MetaboAnalyst 5.0 (Xia et al. 2009; Pang et al. 2021) and R version 4.1.1 (R Core Team 2021). The total metabolite features obtained were filtered by excluding those features with low variability across samples (based on the log10 fold changes between treatments with log10FC<2 removed) and the exclusion of low-intensity peaks (molecular features with a peak intensity of <50 in all conditions). Significant differences between remaining metabolite features by treatment were determined with univariate analysis (ANOVA) followed by post-hoc analysis (Fisher’s LSD, p-value 0.05) using log10 transformed and weight-specific normalised data (Supplementary Table S4). The treatment comparisons for subsequent analysis of significant log2 fold changes in metabolite features are as follow: Pathogen contrast “Pathogen vs Control” (change in expression of metabolite features between the Control and Pathogen treatment), AMF contrast “AMF vs Control” (change in expression of features between the Control and the +AMF treatment), AMF/Pathogen contrast “AMF vs Pathogen” (change in expression of features between the Control and +AMF/Pathogen treatment). The shoot dry weight data were cleaned using the Grubbs method (alpha = 0.2) and after identification of two outliers further analysed with a one-way ANOVA followed by an uncorrected Fisher’s Least Significant Difference (LSD) and significance (P) with standard error of the difference (SED).