Tree species richness differentially affects the chemical composition of leaves, roots and root exudates in four subtropical tree species

Forests are important ecosystems that provide many ecosystem services. To humans, they an resource for food, firewood and timber (Jang et al., 2019). represent rich natural pharmacies containing species with known or potential medicinal nutritional value (Karjalainen 2009). In addition, harbour enormous number of different organisms, from microbes to large mammals (Barlow 2018). This is particularly true (sub) tropical forests, which contain 96% all tree on Earth (Poorter 2015). Consequently, forests sustain high levels biodiversity, in particular insects other arthropods. Tree diversity commonly has a positive effect Several studies show relationships between richness biomass production (Guillemot 2020), carbon storage 2015; Steur 2020) insect (Huang 2018; Schuldt Considering the current biodiversity crisis fuelled by anthropogenic changes land-use CO2 (Trumbore 2015), it understand how relates functions. Specialized molecules play role interactions plants their biotic environment. These molecules, often called secondary metabolites, such as alkaloids, phenolic compounds volatile organic compounds, mediate resistance herbivores pathogens, attract pollinators beneficial root microbes, serve allelopathic suppress competing plant (Raguso contrast ubiquitous amino acids sugars, profiles metabolites specific (Endara Gargallo-Garriga 2020; Sedio 2017). Chemical can drive associated trophic levels, therefore mechanism underlying organismal (Richards Recently, untargeted metabolomics emerged powerful approach study chemical plants. Metabolomics principally technology, aims measure possible one extract without prior knowledge its composition (Macel & van Dam, independence also allows application ecology, gave rise field ecometabolomics (Gargallo-Garriga Peters Ecometabolomic approaches enhance our understanding individual well functioning, example functional trait (Sardans Walker editorial this issue). When comes analyses trees, so far, metabolomic have been mostly used quantify related phylogenetic distances genera Salazar 2017), within canopies (Sedio 2017, 2019; Wiggins 2016), herbivore-induced responses variation leaf metabolomes among seasons 2020). Among others, these showed not static, change response abiotic environmental variation. For example, attacks above-ground below-ground pathogens may shift metabolome towards producing more (defensive) (van Dam Heil, 2011; Macel 2014). terms, plastic, genetic boundaries depending type Indeed, ecometabolomic grasses frobs functioning (BEF) level surrounding (Huberty Ristok Scherling 2010). diversity-driven effects could be caused directly plant–plant (Bixenmann 2016) indirectly via plant–soil feedbacks. lead competition, reduce primary acid (Scherling 2010) increase (Fernandez 2016). Soil feedbacks occur because communities facilitate soil microbial (Bardgett der Putten, Individual exposed respond changing Other than grassland species, diversity-related studied at great depth. targeted revealed total concentrations birch Betula pendula leaves increased plot (Poeydebat Different however, very (Whitehead 2021). Untargeted detail identity phenolics, presence compound classes, roles ecological (Peters test whether affected we analysed exudate, four subtropical experimental site BEF-China (Bruelheide 2014; Huang Trogisch Although previous focused metabolomes, explicitly included roots exudates. important, shoots functions, function environments Roots fulfil physiological water uptake, macro- micro-nutrient foraging storage. Niche differentiation, form morphological (Sun 2017) uptake forms N (complementarity), alleviate nutrient limitation maintain (Lang Moreover, produce exudates, secreted into rhizosphere variety Bouwmeester, Depending conditions, exudates mineral bioavailability through priming (Dijkstra 2013) attraction mutualistic micro-organisms, mycorrhizal fungi (Ferlian Both exudate deter resist pathogenic invertebrate parasitic plants, (Baetz Martinoia, Zeng, Leaves, hand, responsible assimilation carbohydrates. dense forest canopies, compete light direct neighbours. Intraspecific interspecific competition canopy structure (Guisasola cause differences ecophysiological status trees monocultures versus diverse plots (Grossiord, Lastly, defensive (Agrawal 1999). The stands affect herbivory experienced (Schuldt 2010; but see Wang (Rutten We expect root, directly, is, neighbours, indirectly, interactions, Root selected were sampled diversity, ranging over 2, 4 8 16 24 per plot. recorded local target trees. allowed us (local diversity) leaves, hypothesized roots, species-specific would most similar. Because travel only short (Finzi closest neighbours likely greater influence remote Therefore, expected larger diversity. postulated metabolite physical signals coming increases. Finally, stronger fine might exudation. was carried out Biodiversity–Ecosystem Functioning Experiment China (BEF-China), set up managed since 2009 2014), It located Jiangxi Province, (29°08?–29°11?N, 117°90?–117°93?E) two sites, A B, planted 2010, respectively. former monoculture replanted mixed using ‘broken stick’ design. Using pool 40 extinction scenarios simulated 1, 4, 8, 566 25.8 m × 400 each. rectangular grid 1.29 distance. Thus, every eight distance 1.82 (diagonal). ranges altitude 113 182 slopes 15 43 degrees. climate. From 1971 2000, mean annual temperature 16.7°C precipitation 1,800 mm. 17.9°C 2,076 mm, respectively, during 2013–2017. January coldest month 0.4°C July hottest 34.2°C. (natural) vegetation characterized mixture evergreen deciduous 2011). analyses, Site B (Figure S1). having least three replicates levels. families avoid bias. All evergreens. Cyclobalanopsis glauca (Thunb.) Oerst. (homotypic syn. Quercus Thunb.), Fagaceae, ‘ring-cup oak’, native warmer temperate zones Asia. contains tannins (catechins, procyanidins gallic acids), triterpenoids, flavonoid glycosides steroids (cycloartanols stigmastanes; Kamano 1976; Shen 2012; Sheu 1992; Suga Kondo, 1974; Wakamatsu Schima superba Gardner Champ., Theaceae, dominant broad-leaved distributed spp. triterpenoids saponins flavonoids anthocyanins (Deng Kitagawa 1975; Liang Liu Wu 2015, Yang Yu Daphniphyllum oldhamii (Hemsl.) K. Rosenthal, heterotypic synonym pentandrum Hayata, Daphniphyllaceae, Eastern Species genus group alkaloids unusual ring structures (azaspirodecane derivatives) backbones, probably derived squalene mevalonate (Kobayashi Kubota, Furthermore, lignans described (Chao Gan 2007; Kobayashi 2003; Mu 2006, Shao 2004; Takatsu 2004). Cinnamomum camphora (L.) J. Presl, camphor, Lauraceae, Japan Taiwan cultivated naturalized regions worldwide. monoterpenoids (bicyclic, acyclic, menthane monoterpenoids, e.g. camphene, terpineol, pinene borneol). C. tri- sesquiterpenoids, neolignans, steroids, benzenoids, oxolanes (KNApSAcK database, Afendi 2012). sampling took place 19 October until 28 2019. samples 11 13 experiment species. above-mentioned 1 (4 plots), 2 (two (1 plot), (one plot) (=plot groups D1, D2, D4, D8, D16). As there sufficient found available D16 plot, five collected D24 treated together (D16/24). species), randomly two-species plots, 4–8 each 16/24 1–7 detailed combinations replicate numbers shown Figure S1 data (https://doi.org/10.5281/zenodo.5255811). total, 84 yielding After scrutiny became apparent 9 (exudate) Pre-experiments under excavated collection always originating tree, even when taking care. Eight reassigned based metabolome. One sample reassigned; thus, 83 remained statistical 81 (2 lost preparation) complete 84. position neighbourhood time coordinates (see https://doi.org/10.5281/zenodo.5255811). defined 12 adjacent squares middle 10 determine followed protocol Phillips al. (2008), slight modifications. upper cm soil, starting base identify terminal strands (<2 mm). Adhering particles removed demineralized water. dried paper tissue placed 30-ml glass syringe, then filled new, unused beads (500–750 ?m, washed, ACROS Organics, ThermoFisher, New Jersey, USA/Geel, Belgium). outlet syringe attached plastic tube 50 Ml it. solution (0.50 mM NH4NO3, 0.10 KH2PO4, 0.20 K2SO4, 0.15 MgSO4, 0.30 CaCl2) top. sucked discarded. left equilibrate 20 min, after fresh added procedure repeated. second washing step, closed cotton wool sealed parafilm. Everything wrapped aluminium foil 48 hr. hr, cut-off syringes, contained beads, bags brought laboratory. chosen shaded part crowns. recover 0.22-?m sterile filter vacuum pump manifold 60-mL glass/plastic vial. mounted top 60 solution. Then, started Five millilitres aliquoted non-purgeable (NPOC) analysis below). remaining frozen ?20°C 60-ml HDPE cryo-vials freeze-drying thawed shipment. solution, cleaned remove residual beads. put envelopes, oven-dried 30 min 105°C, hr 60°C. Dried extracted LC-MS according standard protocol: Per mg powdered material, ml extraction buffer (75% v/v methanol, HPLC grade, 25% acetate (2.3 acetic 3.41 g ammonium l 18 M? water, pH 4.8) plus µl 100 IAA-Valin internal standard) shaken ceramic homogenizer (Retch MM400, Retch GmbH, Haan, Germany) 5 Hz. Samples centrifuged RT 15,000g. pellet re-extracted another again both supernatants unified. diluted 1:5 buffer, kept overnight, 15,000g transferred vials. Exudates freeze-dried shipment, re-dissolved bottles washed methanol. new 2-ml reaction tube, completely (vacuum centrifuge 27°C) tried re-dissolve them smaller volume buffer. However, possible, added, dissolution aided vortexing, ultrasonic bath heating 40°C, centrifuged, dissolved that, concentrated, dryness, centrifuge. Non-purgeable determined removal inorganic C (acidification sparging) catalytic thermal oxidation (at 680°C) subsequent detection infrared gas analyser (TOC VCPN) analytical replicates. NPOC values depended heavily morphology differed significantly S2), dilute before analysis. normalizes exudation rates thus better comparison treatments. correct NPOC, concentrated measured pipette. median calculated (5.023 mg/l). minimum 0.84 mg/l, maximum 33.66 mg/l. decided concentration should correspond 300 µl, required follows: volumesample (µl) = NPOCsample (mg/l) µl/median (mg/l). (remaining 50–200 µl) 70% methanol attain (range: 50–2011 µl). Chromatographic separations performed 40°C UltiMate™ 3000 Standard Ultra-High-Performance Liquid Chromatography system (UHPLC, Thermo Scientific) equipped Acclaim® Rapid Separation (RSLC) 120 column (150 mm 2.1 particle size 2.2 ThermoFischer following gradient flow rate 0.4 ml/min:0–1 isocratic 95% [water/formic 99.9/0.1 (v/v %)], 5% [acetonitrile/formic %)]; 1–2 linear 20% B; 3–8 8–16 16–18 18–18.01 18.01–20 B. Data 0 min. injection leaves. Eluted detected m/z 90 1,600 spectra Hz (line only) ESI-UHR-Q-ToF-MS (maXis impact, Bruker Daltonics) ion mode data-dependent collision-induced dissociation (Auto-MSMS mode). instrument settings applied: nebulizer on, 2.5 bar; dry gas, nitrogen, l/min, 220°C; capillary voltage, 4,500 V; end plate offset, 500 funnel radio frequency (RF), 200 Volts peak-to-peak (Vpp); RF, 220 Vpp; in-source (CID) energy, 0.0 eV; hexapole quadrupole low mass, m/z; collision nitrogen; prepulse storage, 7 µs. Stepping: on; basic mode; cell Vpp 1,000 transfer time, 70 µs, timing; 50%/50%, energy MSMS, 80%, timing 50%/50%. Data-dependent CID settings: intensity threshold 600, cycle s, active exclusion spectra, release 0.5 smart exclusion, off, isolation fragmentation settings, size- charge-dependent, width 3–15 m/z, 20–30 eV, charge states included: 1z, 2z 3z. Calibration scale raw files sodium formate cluster ions obtained automatic infusion 1.66 µl/min NaOH 50/50 (v/v) isopropanol 0.2% formic (HPC Three QCs prepared (QC QC exudates) run 8th sample. mix commercial standards (MM8; Böttcher 2007). blanks: blanks (ACN); (empty tubes buffer; exudates: empty methanol), furthermore campaign stored cryo vials cryo-vials, passed whole preparation process. Roots, separate batches cross-contamination. Within tissues, randomized throughout process processing measurement. LC-qToF-MS processed Compass MetaboScape Mass Spectrometry Software, Version 5.0.0 (Build 683; Daltonik GmbH). recalibration, peak picking, alignment, region feature extraction, grouping isotopes, adduct T-ReX algorithm Metaboscape. Settings: Peak detection: threshold, counts, length, signal, intensity. Minimum length recursive spectra. Retention range, 0–18 90–1,600 m/z. MSMS import method, average, grouped energy. Ion deconvolution: EIC correlation, 0.8, ion, [M+H]+, seed ions, [M+Na]+, [M+K]+, [M+NH4]+, common [M+H-H2O]+ T-ReX-Positive Recalibration Auto-Detect. Feature filters: samples: present 3 302, extraction: filter: 80% tissue). Features Blanks excluded signal divided ?3. Quality checks stability retention intensity, check carry-over identity. With tables created: classified around (‘plot diversity’) (‘local diversity’). processing, dataset 39,077 features 22,520 fragment assigned; 3,883 removed, leaving 35,194. diversity: 38,105, assigned 22,104; 3,649 excluded, 34,456. further statistics, exported Metaboanalyst 4.0 (Chong general principal component (PCA) PCA), tissues together. IQR filtered Pareto scaled. Chow–Ruskey diagrams produced ‘intervene’ (https://asntech.shinyapps.io/intervene/; Khan Mathelier, counting ‘present’ >/=1,000. partial squares-discriminant (PLS-DA), created including respective datasets too zeros. PLS-DA, either maximized S1; Table S1) varied due dieback misplantings, some had fewer n level, (monoculture), mixtures. case Cy. glauca, latter (3, 5) combined obtain replicates, no 6 7. D. oldhamii, combined, invaded monocultures. compare matrices, multiple permutation (MRPP) R-package vegan (v.2.5-6; https://CRAN.R-project.org/package=vegan ) R (version 3.6.0; RCoreTeam, sum >=1,000 Shannon vegan. Principal coordinate (PCoA) version 4.0.4 ‘pco’ package ecodist: (v. 2.0.7), matrix ‘Bray–Curtis’ dissimilarity measure. 2.5-7) ‘vegdist’. PERMANOVA ‘adonis2’ ANOVA ‘aov’ model: richness/Chemical Diversity~Tissue*Species*Species Diversity (Local Plot). Spearman's rank-order correlation index ‘stat_cor’ r ggpubr 0.4.0) along nonlinear ‘loess’ fit. First, matched parallel search against databases: NIST17 (The NIST Center, U.S. Department Commerce), MoNa (https://mona.fiehnlab.ucdavis.edu/), ReSpect (Sawada 2012), Riken public databases http://prime.psc.riken.jp/compms/msdial/main.html#MSP), Bank EU (https://massbank.eu/MassBank/Index), GNPS (https://gnps.ucsd.edu/) database (Döll, unpublished), Spectral Library Search [Parameters: Filter: exact match entry precursor mass; tolerances (narrow wide): 10–30 mDa, mSigma 20–100, MS/MS score 900–800]. Afterwards, manual annotation literature KNApSAcK ‘Target species’) formula generation (Smart Formula algorithm, MetaboScape), patterns spectral similarities mentioned databases. annotations, small library (197 compounds). mass find similar already annotated ones. classification approach, searched mass. Tolerances above, limited 17 computational limitations. Relevant PCA (features highest loading separating components) PLS-DA (20 VIPs—variable importance projection first component) manually above. Annotated workflow (Supporting Information, Script). brief, table in.csv format CAS numbers, if available, read (v 4.0.3; employed ‘webchem’