Endogenous rhythmic growth and ectomycorrhizal fungi modulate priming of antiherbivore defences in subsequently formed new leaves of oak trees

Abstract

en

1. Priming of plant defences provides increased plant protection against herbivores and reduces the allocation costs of defence. Defence priming in woody plants remains obscure, in particular, due to plant development traits such as endogenous rhythmic growth.
2. By using bioassays with oak microcuttings (Quercus robur) and combining transcriptomic and metabolomic analyses, we investigated how leaf herbivory by Lymantria dispar and root inoculation with the ectomycorrhizal fungus Piloderma croceum prime oak defences. We further investigated how defence priming is modulated by the rhythmic growth of the oaks.
3. A first herbivory challenge in oak leaves primed newly grown leaves for enhanced induction of jasmonic acid (JA)-related direct defences, or enhanced emission of volatiles, depending on the specific growth stage at which the plants were first challenged. Root inoculation with Piloderma abolished the enhanced induction of JA-related defences and volatile emission.
4. Synthesis: Our results indicate that a first herbivore attack primes direct and indirect defences of newly formed oak leaves and that the specific display of defence priming is regulated by rhythmic growth, and modulated by the interaction Piloderma. Our results show that the priming memory in oaks can be transmitted to the next growth cycle, to the leaves of the new shoot unit.

Resumen

es

1. El priming de las defensas de las plantas es un mecanismo que les proporciona una mayor protección contra los herbívoros y reduce el coste de la activación de estas defensas. En plantas leñosas, la regulación del priming sigue siendo parcialmente desconocida, en parte, debido a ciertos rasgos del desarrollo de las plantas leñosas como el crecimiento rítmico endógeno.
2. Utilizando bioensayos con microesquejes de roble (Quercus robur) y combinando análisis transcriptómicos y metabolómicos, investigamos cómo la herbivoría foliar por Lymantria dispar y la inoculación de raíces con el hongo ectomicorrícico Piloderma croceum pueden afectar al priming de las defensas del roble. Además, investigamos cómo este priming puede estar modulado por el crecimiento rítmico de los robles.
3. Un primer ataque del herbívoro activó el fenómeno de priming en las nuevas hojas en desarrollo, mediante la potenciación de las defensas directas reguladas por el ácido jasmónico (JA), o mediante una mayor emisión de volátiles, dependiendo de la etapa de crecimiento específica en la que las plantas fueron sometidas al primer ataque. La inoculación de raíces con Piloderma suprimió la inducción de las defensas relacionadas con el JA, así como la mayor emisión de volátiles.
4. Síntesis: Nuestros resultados indican, que, en el roble, un primer ataque foliar por un herbívoro activa las defensas directas e indirectas en las hojas de nueva formación y que este priming de las defensas está regulado por el crecimiento rítmico, y a su vez modulado por la interacción con hongos ectomicorrícicos como Piloderma. Nuestros resultados muestran que la memoria del priming en los robles puede transmitirse a las hojas del siguiente ciclo de crecimiento.

1 INTRODUCTION

Trees are subjected to both constant and seasonal pressure from herbivorous insects and exhibit well-adapted and regulated defence mechanisms (Haukioja, 1991; Nykänen & Koricheva, 2004). Tree responses to herbivore attacks can vary widely, mainly depending on the species identity of both the herbivores and the host tree. Commonly, herbivory results in higher concentrations of defensive compounds in leaves, making them less palatable and/or less nutritious to the herbivores (i.e., direct defences; Bacht et al., 2019; Mageroy et al., 2019). In addition, herbivore attacks also trigger the production of volatile compounds, which may affect the herbivores by attracting their natural enemies (i.e., indirect defences; Ghirardo et al., 2012; McCormick et al., 2019). Plant direct and indirect anti-herbivory defences are regulated by complex signalling networks in which diverse hormones play a major regulatory role (Erb & Reymond, 2019). Among them, jasmonates (JA), a family of oxylipins, emerged as a key signal of plant responses to insect-chewing herbivores (Howe & Jander, 2008). Other hormones, such as salicylic acid, ethylene and abscisic acid may interact with the JA signalling pathway in the orchestration of plant anti-herbivory defences (Erb & Reymond, 2019). Most research on anti-herbivory defence regulation has been conducted on herbaceous annual plants. However, it remains unclear whether the same patterns are also applicable to long-lived woody plant species (Ullah et al., 2019).

Pedunculate oaks (Quercus robur) are large, long-lived deciduous trees, harbouring one of the largest biodiversity of organisms among European forest trees (Brändle & Brandl, 2001). The species is widely distributed in Europe and provides important ecosystem services (Denk & Grimm, 2010) like hardwood for construction, furniture and wine barriques. From a research perspective, common oaks have biological traits that pose challenges for studying them, including their large size, long generation time and longevity (Kremer & Le Corre, 2012). Moreover, the growth of pedunculate oaks is not continuous but consists of successive cycles of endogenous rhythmic growth. The formation of successive growth cycles is not restricted to oak trees. Besides its expression in many tropical trees (Hallé et al., 2012), it is also described in other trees growing under temperate climates, belonging to the genus Betula, Carpinus, Castanea, Fagus and Quercus (Lavarenne et al., 2014). The growth cycles are separated by a rest phase, and each of them consists in an alternating root flush followed by a shoot flush (Harmer, 1990). The growth phases in each flush are distinguishable by morphological traits, like a drastic reduction in internode length at the end of the shoot flush, and have recently been related to transcriptome patterns of genes involved in a multitude of metabolic and physiological processes (Herrmann et al., 2015). These growth cycles are not related to climate seasonality but have an endogenous determination. Successive growth cycles develop on young trees grown under optimal growth conditions with long photoperiod and about 25°C. The period of one growth cycle is strongly determined and is about 40 days for the oak under optimal growth conditions. These shoot flushes occur directly after germination of the acorn and are observed during the whole life of the oaks. The number of developed growth cycles decreases with the age of the tree and varies in the field under various climatic conditions (Herrmann et al., 1998, 2015). The pedunculate oak has become an important model organism due to the possibility of being micropropagated (Herrmann et al., 2016) and its sequenced genome's availability (Plomion et al., 2018). These oaks are therefore good models to explore anti-herbivory defences and their regulations over successive growth units in trees.

Previous studies have characterized genes and metabolites (volatile and non-volatile) related to the activation of anti-herbivory defences in pedunculate oaks as thaumatins and pathogenesis-related proteins (Bacht et al., 2019; Ghirardo et al., 2012; Mageroy et al., 2019; Volf et al., 2020, 2021). Remarkably, pedunculated oaks were found to display a strong transcriptomic anti-herbivory defence response specifically during root flush, evidencing the relevance of taking the endogenous rhythmic growth into account (Bacht et al., 2019; Kurth et al., 2015; Maboreke et al., 2016; Tarkka et al., 2021). Moreover, shoot flush and root flush are related to parallel shifts in above- and below-ground carbohydrate allocation and non-structural carbohydrates NSC (Herrmann et al., 2015, 2016), which can in turn strongly affect plant–herbivore interactions.

Several studies demonstrate that plants become more resistant to insect attacks when they already have experienced herbivory or have perceived cues indicating an increased probability of herbivore attack (Bandoly et al., 2015; Bittner et al., 2019; Ton et al., 2007). This phenomenon is known as defence priming and allows plants to respond more strongly and/or faster and/or in a more sustained manner to subsequent attacks (Frost et al., 2008; Hilker et al., 2015; Martinez-Medina et al., 2016; Mauch-Mani et al., 2017; Westman et al., 2019). Strikingly, the durability of the priming can span from hours to months after the priming stimulus, and it can be even transmitted to the next generation (Erbilgin et al., 2006; Mageroy et al., 2019; Pastor et al., 2013). A few studies have found evidence for defence priming in trees (Bittner et al., 2019; Camisón et al., 2019; Mageroy et al., 2019), but little is known about the underlying molecular mechanisms. Moreover, whether and how the endogenous rhythmic growth affects the deployment of defence priming in pedunculate oaks, and whether defence priming can be expanded to newly grown tissues over successive growth cycles is unknown so far.

Root colonization by soil microorganisms can also lead to defence priming against multiple stressors (Gruden et al., 2020). Among them, mycorrhizal fungi can prime plant defences against herbivore and pathogen attacks to promote plant health (Jung et al., 2012; Kempel et al., 2010; Rivero et al., 2021; Sánchez-Bel et al., 2016). Whereas the mechanisms underlying mycorrhizal-related defence priming have been well studied in the interactions between arbuscular mycorrhizal fungi and herbaceous plants, the effect of ectomycorrhiza on tree anti-herbivory defences is much less explored (Gruden et al., 2020). Trees generally harbour a high density and diversity of insect herbivores (Brändle & Brandl, 2001), and they are associated with a vast diversity of ectomycorrhiza fungi (Baar et al., 1999; Kõljalg et al., 2000). Recently, Kaling et al. (2018) found that poplar root colonization by the ectomycorrhiza Laccaria bicolor reduced the performance of the leaf beetle Chrysomela populi, and this effect was associated with a resource shift from constitutive phenol-based to jasmonate-induced specialized defence compounds (as prunasin and phenylalanine). Along the same lines, the ectomycorrhiza Piloderma croceum mediated a shift from direct to indirect defences at the transcriptional level in pedunculate oaks challenged by the caterpillar Lymantria dispar (Bacht et al., 2019). Such results suggest that ectomycorrhiza does not only provide nutritional services to trees (Becquer et al., 2019) but also might affect tree anti-herbivory defence strategies (Dreischhoff et al., 2020). In addition, ectomycorrhiza reinforces the amplitude of the oscillations between above- and below-ground carbohydrate allocation during endogenous rhythmic growth in pedunculate oaks (Herrmann et al., 2015; Tarkka et al., 2021), which can further affect anti-herbivory defences (Bacht et al., 2019). Still, how ectomycorrhiza modulates anti-herbivory defences in pedunculate oaks after successive herbivory events, and how endogenous rhythmic growth modulates the three-way interactions between pedunculate oak, ectomycorrhiza, and herbivore remains poorly understood.

As stated above, in long-lived plants such as trees, the primed state triggered by a first herbivory event may persist across months (Mageroy et al., 2019; Zvereva et al., 1997). Considering the endogenous rhythmic growth of pedunculate oaks, which involves several growth cycles per season, we hypothesized here that (i) the defences primed by a first herbivore event will persist until the next growth cycle, in subsequently formed new leaves. According to the strong influence of the endogenous rhythmic growth on the deployment of plant anti-herbivory defences, we also hypothesized that (ii) the specific growth stage (shoot flush or root flush) at which a first herbivory occurs will modulate the display of defence priming, with a strong priming effect during shoot flush, when plant resources are mostly allocated to leaf tissues. Finally, due to the impact of ectomycorrhiza in anti-herbivory defence strategies, we hypothesized that (iii) the inoculation with an ectomycorrhiza fungus will affect the features of defence priming caused by leaf herbivory. To test these three hypotheses, we performed a controlled laboratory experiment on oak microcuttings from the oak clone DF159 (Herrmann et al., 2016). By combining transcriptomic and metabolomics analyses, with insect performance assessment, we examined the priming phenomenon across successive growth cycles.

2 MATERIALS AND METHODS

2.1 Oak microcuttings

Oak microcuttings were obtained by in vitro propagation and rooting of the Quercus robur L. clone DF159 (www.TrophinOak.de) according to Herrmann et al. (1998), which excludes genetic variability among the treatments and replicates. The oak microcuttings are equivalent to young saplings, where the total root system has several secondary principal roots that show a rhythmic growth like the one described for saplings. The clone DF159 was already micropropagated for 30 years in vitro, having similar physiological traits to old oak trees, and being able to develop acorns only a few years after transfer to the field. Under optimal controlled growth conditions, the rooted microcuttings showed a typical endogenous rhythmic growth where two alternating shoot flush and root flush result in one growth cycle (GC), thus behaving physiologically like mature trees (Herrmann et al., 1998; Herrmann & Buscot, 2008). Oaks microcuttings were grown in a photosynthetic photon flux density of 180 μmol m⁻² s⁻¹, a long-day setting (16 h of light and 8 h of darkness), a constant temperature of 23°C, and a relative air humidity of 75%. Four-week-old oak microcuttings cultivated in the Petri dish system described below were supplied with sterilized tap water every 2 weeks to keep the weight of the Petri dishes approximately constant and to avoid drought stress (Bacht et al., 2019; Herrmann et al., 2015; Kurth et al., 2015; Maboreke et al., 2016; Tarkka et al., 2021).

2.2 Experimental design

Overall, the design of the study included two treatment groups: oak microcuttings control non-inoculated and oak microcuttings inoculated with the ectomycorrhizal fungus Piloderma croceum. Within every treatment group, we have two further treatment groups according to the plant developmental stage: oak microcuttings during shoot or root flush. In every treatment group we have four treatments with the leaf herbivore Lymantria dispar (no_Ld, Ld_prim, Ld_trig and Ld_prim_Ld_trig). The different leaf herbivory treatments are detailed under the ‘Herbivore challenge’ section and summarized in Figure 1. To summarize, we have 16 treatments with seven replicates (microcuttings), having a total of 112 oak microcuttings for the whole experiment.

FIGURE 1

Experimental design. Oak microcuttings were obtained by in vitro propagation and rooting of the Quercus robur L. clone DF159. The microcuttings showed a typical endogenous rhythmic growth where two alternating shoot flush (SF, leaves in D stage) and root flush (RF, leaves in B stage) result in one growth cycle (GC). The oak microcuttings were inoculated (Pc) or not (Co) with the ectomycorrhizal fungus Piloderma croceum. To test the effect of herbivory, the polyphagous moth Lymantria dispar was used. For each microcutting challenged with L. dispar, leaves from the same GC (which included the 2–4 youngest leaves), were exposed to one larva, and enclosed in a clip cage, to limit the feeding area. After 6 weeks of the P. croceum inoculation, the leaves from GC1 were exposed to third-instar larvae of L. dispar larvae for 6 h; then, plants were kept growing until the next growth cycle (GC2) (another 6 weeks). The leaves from GC2 (leaves in D stage for SF and leaves in B stage for RF) were exposed to second-instar larvae of L. dispar larvae for 48 h. Before removing the larvae from GC2, the volatile organic compounds accumulated during 48 h were captured. The performance of the larvae applied on GC2 and the leaf fresh weight in herbivore-challenged plants were determined. Then the leaves were immediately frozen in liquid N and stored at −80°C for gene expression analysis.

2.3 Ectomycorrhizal fungus inoculation

To elucidate how the inoculation with an ectomycorrhizal fungus might modulate the plant defence responses triggered by leaf herbivory, we inoculated half of the microcuttings with the ectomycorrhiza fungus P. croceum J. Erikss and Hjortst isolate DSMZ 4824; ATCC MYA-4870. P. croceum is a common ectomycorrhizal fungus of both coniferous and hardwood tree species and an established model for both ecological and physiological studies on DF159, as described by Herrmann et al. (1998). Ectomycorrhizal fungus inoculation was performed in Petri dishes filled with ɤ-sterilized soil as described previously by Tarkka et al. (2013). The inoculation was performed by mixing a substrate mixture of vermiculite and sphagnum peat containing a 2-week-old liquid fungal mixture (1:1 [vol/vol]) with the sterilized soil as described by Tarkka et al. (2013). For non-inoculated treatments, the soil was mixed with the same but sterile non-inoculated substrate mixture.

2.4 Herbivore challenge

To challenge the oak microcuttings with herbivory, we used larvae of the polyphagous moth L. dispar (Lepidoptera: Noctuidae). This herbivore can feed on a wide range of plant families but has a strong preference for Quercus species (Alalouni et al., 2013). Egg masses of the New Jersey laboratory strain (reared from eggs of breeding stock of the US Department of Agriculture Forest Service Insect Rearing Facility, Hamden, CT) were raised under the same environmental conditions as the oak microcuttings and reared on artificial gypsy moth diet (based on wheat germ, Stanley-Samuelson et al., 1992). For each microcutting challenged with L. dispar, leaves from the same GC (the 2 youngest leaves) were exposed to one second-instar (as the triggering stimulus, Ld_trig) or third-instar (as priming stimulus, Ld_prim) larva and enclosed in a clip cage (3 cm in diameter), to limit the feeding area (Figure 1). We selected as the priming stimulus specifically third-instar larvae, to ensure enough damage during the priming event (6 h); while second-instar larvae were selected as the triggering stimulus, to ensure that there was remaining material upon the triggering event (48 h). The feeding progress was checked twice an hour, and non-feeding larvae were replaced by new ones. Similarly, we mounted an empty clip cage on similar leaves of the plants not assigned to herbivore treatment.

In summary, we exposed the oak microcuttings to four different L. dispar treatments: (1) no_Ld (not exposed to L. dispar); (2) Ld_prim (exposed to L. dispar in the GC1 as priming stimulus); (3) Ld_trig (exposed to L. dispar in the GC2 as triggering stimulus); and (4) Ld_prim_Ld_trig (exposed to L. dispar in two successive growth cycles, in the first growth cycle as priming stimulus and in the following growth cycle as triggering stimulus).

To analyse the role of endogenous rhythmic growth on the plant responses after the leaf herbivory, during shoot flush, herbivores were applied on sink leaves of plants at stage D (fully expanded leaves and roots at rest stage) as defined by Herrmann et al. (2015) during each of the two successive growth cycles GC1 and GC2. At this stage, the sink leaves are still coloured by anthocyanin and photosynthetic maturation is not achieved. While during root flush herbivores were applied on the source leaves successively during GC1 and GC2 of plants at stage B (roots at maximal elongation and bud swelling) as defined by Herrmann et al. (2015). The source leaves were fully photosynthesizing and provide energy for the growth of the new flush.

Before removing the larvae from GC2, volatile organic compounds (VOCs) were sampled as described below (Figure 1). Larval weight (as a proxy for larval performance) and leaf weight from herbivore-challenged plants were recorded. The leaves on which the larvae were feeding were immediately frozen in liquid N and stored at −80°C for molecular analysis.

2.5 Collection and analysis of oaks VOCs

The collection of VOCs was performed at 23 ± 1°C and 75% relative humidity between 10:00 am and 5:00 pm in a growth chamber. VOC sampling was performed on the oak microcuttings leaves 48 h after the L. dispar challenge. Shoot flushes exposed to L. dispar were enclosed in two plastic cups that were closed around the flush (including all the leaves from the flush) with hair clips. Stainless steel thermal desorption sorbent tubes loaded with 200 mg of Tenax (MARKES, Llantrisant, United Kingdom) were inserted into one of the plastic cups and connected to a vacuum pump system. A charcoal filter was included to avoid contaminations. Charcoal-filtered air was pulled through the thermal desorption sorbent tubes at a flow rate of 0.5 L/min. VOCs were collected for 20 min, resulting in a sampling volume of 10 L per sample. For every treatment, seven individual plants (seven independent biological replicates) were sampled. In addition, we also sampled volatiles from empty cups. Those ‘air blanks’ were used in further data processing to exclude systemic contaminations.

Oak volatiles were analysed by a thermal desorption gas chromatograph-mass spectrometer (TD-GC-MS) consisting of a thermos desorption unit (MARKES, Unity 2, Llantrisant, United Kingdom) equipped with an autosampler (MARKES, Ultra 50/50). Tubes were desorbed with helium as carrier gas and a flow path temperature of 150°C using the following conditions: Dry Purge 5 min at 20 mL/min, Pre Purge 2 min at 20 mL/min, Desorption 8 min at 280°C with 20 mL/min, Pre Trap fire purge 1 min at 30 mL/min, Trap heated to 300°C and hold for 4 min. The VOCs were separated on a gas chromatograph (Bruker, GC-456, Bremen, Germany) connected to a triple-quad mass spectrometer (Bruker, SCION). Separation took place on a DB-5MS column (30 m × 0.25 mm × 0.25 μm. Restek, Germany). The conditions of the gas chromatography were as follows: 40°C for 5 min, 5C/min to 185°C, 30C/min to 260 and hold for 0.5 min. The mass spectrometer was operated in full scan mode with the following parameters: transfer line temperature 280°C, ion source temperature 260°C, scan time 250 ms, scan range 40–550 m/z and ionization 70 eV. We only considered peaks with a signal-to-noise ratio >10. Peaks that were also present in air blanks were regarded as systemic contamination and were excluded from further analysis. The peak areas of the remaining m/z signals were calculated using the Bruker Workstation software (v8.0.1). All these peaks were identified by library matching.

2.6 RNA extraction and gene expression analysis

Sink (stage D) and source (stage B) leaves developed during GC2 were used for transcriptomic analysis according to Tarkka et al. (2013). The leaves from each shoot flush per plant were pooled and considered as a biological replicate, and three biological replicates were used per treatment. These leaf pools were used for RNA extraction. Total RNA was extracted using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer's protocol, from 100 mg of leaf material. The total RNA was treated with RNase-free DNase set (Qiagen). The quality and quantity of RNA were checked using a Nanodrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and a Nano Chip and Bioanalyzer 2100 (Agilent Technologies, Palo Alto, CA, United States). First-strand cDNA was synthesized from 1 μg of purified total RNA using the ImProm-II Reverse Transcription System (Promega, Madison, WI, USA) according to the manufacturer's instructions. First-strand cDNA was pre-amplified using MyTaq™ DNA Polymerase (Bioline, according to the manufacturer's instructions) and specific primers (same primers that will be used later for gene expression analysis by qRT-PCR, Table S1). For every reaction added 2 μL of the 5x MyTaq Reaction Buffer, 0.6 μL of MgCl₂ (25 mM), 0.2 μL of dNTPs (10 mM), 0.2 μL of MyTaq DNA Polymerase, 4 μL of water (ddH₂O) and 2 μL of 5× pooled specific primers solution (250 nM). Pre-amplification conditions were 5 min at 95°C followed by 18 cycles at 95°C for 10 s, 60°C for 1 min and 72°C for 15 s. The pre-amplified cDNA was treated with Exonuclease I (New England BioLabs) according to the manufacturer's instructions, and their quality was checked by Real-time qRT-PCR using iQ SYBR Green Supermix (Bio-Rad Laboratories, Hercules, CA, USA) and the iQ5 real-time PCR detection system (Bio-Rad Laboratories). The expression level of the oak 18S ribosomal RNA gene (Bacht et al., 2019; Tarkka et al., 2013, Table S1) was analysed to check the quality of the pre-amplified cDNA. The qRT-PCR conditions were 10 min at 95°C followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. The specificity of each PCR amplification procedure was verified by melt-curve analysis of the PCR product with a heat dissociation protocol (from 60 to 95°C). The qPCR products were diluted at 1/160 with TE buffer. The gene expression of 65 genes selected according to Bacht et al. (2019) (specific primers Table S1) was analysed by qRT-PCR using SsoFast™ EvaGreen® Supermix (Bio-Rad) and the 96.96 Dynamic Array™ IFC chip for Gene Expression (Fluidigm, San Francisco, CA, USA) and the BioMark™ Systems for Genetic Analysis (Fluidigm). This dynamic array is a targeted technique (based on qPCR), where the genes analysed are pre-selected, to produce gene-specific primers (differently from untargeted techniques such as microarray or RNA sequencing). For each assay, a mix composed of 2.5 μL of 25 μM of each forward and reverse primer and 2.5 μL of the 2× Loading reagent was loaded into the assay inlets of the array. Into the sample inlets, 5 μL of the solution containing 2 μL of pre-amplified cDNA and 3 μL of a mix composed of 1× SsoFast EvaGreen® Supermix (Bio-Rad), 1× loading reagent and 1× ROX were loaded. The cycling programme consisted of 10 min at 95°C followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. Relative quantification of specific mRNA levels was performed using the comparative method of Livak and Schmittgen (2001). The gene expression was normalized by the oak 18S ribosomal RNA gene (Bacht et al., 2019; Tarkka et al., 2013).

2.7 Statistical analyses

The gene expression levels of the 65 genes selected from Bacht et al. (2019) were subjected to permutational multivariate analysis of variance (PERMANOVA) using the ‘vegan’ package and visualized by non-metric multidimensional scaling (NMDS) using the ‘vegan’ and ‘ggplot2’ packages (Figure 2). Data for gene expression levels, of each gene, were analysed by four-way analysis of variance (generalized linear model, Proc Mixed in SAS 9.1; Table S4), including the following variables: priming stimulus (Presence/Absence), triggering stimulus (Presence/Absence), flush (Shoot flush/Root flush) and ectomycorrhizal fungus inoculation (Inoculated/Non-inoculated). After this analysis, Duncan's post hoc test was performed (p < 0.05; Figures 3 and 4). Log-transformed data for caterpillar weight (Table S5) and leaves biomass (Table S6) were subjected to analysis of covariance (generalized linear model, Proc Mixed in SAS 9.1) with initial larval weight as a covariate. The differences in the emissions of the selected VOCs were analysed by four-way analysis of variance (ANOVA; Table S7) followed by Duncan's post hoc test (p < 0.05; Figure 6).

FIGURE 2

Non-metric multidimensional scaling (NMDS) plot of samples. Expression level of 65 genes described to be induced after leaf herbivory. The gene expression was analysed in control oaks microcuttings (Co) or inoculated with the ectomycorrhiza fungus Piloderma croceum (Pc) during shoot and root flush (SF and RF) (a), only during SF (b) or only during RF (c). During SF and RF, the oak microcuttings were not exposed to Lymantria dispar (no_Ld), exposed only in the first growth cycle (GC1) (Ld_prim), exposed only in the second growth cycle (GC2) (Ld_trig) and exposed in the two successive growth cycles (GC1 and GC2) (Ld_prim_Ld_trig). The gene expression was analysed in leaves of the second growth cycle. Three biological replicates were used (n = 3).

FIGURE 3

Quantification of the jasmonate biosynthesis-related marker genes 13LOX, AOS, AOC and TAN by qPCR. Transcript levels were analysed by qRT-PCR during shoot (SF) and root flush (RF) in leaves from oak microcuttings not inoculated (Co) or inoculated (Pc) with Piloderma croceum. During SF and RF the oak microcuttings were not exposed to Lymantria dispar (no_Ld), exposed only in the first growth cycle (GC1) (Ld_prim), exposed only in the second growth cycle (GC2) (Ld_trig) and exposed in the two successive growth cycles (GC1 and GC2) (Ld_prim_Ld_trig). The gene expression was analysed in the leaves belonging to the second growth cycle (GC2). Figure shows the relative expression of each gene normalized to constitutively expressed oak reference genes 18S. Box plots represent the IQR, the bisecting line represents the median, the whiskers represent 1.5 times the IQR, and the dots represent the data from three independent biological replicates. For every data obtained for one treatment within the same block (delimited by grey lines), a, b and c indicate significant differences according to Duncan's test (p < 0.05).

FIGURE 4

Quantification of the jasmonate responsive-related marker genes CHI, DSP, P450 and MCO. Transcript levels were analysed by qRT-PCR during shoot (SF) and root flush (RF) in leaves from control oak microcuttings (Co) or inoculated with Piloderma croceum (Pc). During SF and RF the oak microcuttings were not exposed to Lymantria dispar (no_Ld), exposed only in the first growth cycle (GC1) (Ld_prim), exposed only in the second growth cycle (GC2) (Ld_trig) and exposed in the two successive growth cycles (GC1 and GC2) (Ld_prim_Ld_trig). The gene expression was analysed in leaves belonging to the second growth cycle (GC2). Figure shows the relative expression of each gene normalized to constitutively expressed oak reference genes 18S. Box plots represent the IQR, the bisecting line represents the median, the whiskers represent 1.5 times the IQR and the dots represent the data from three independent biological replicates. For every data obtained for one treatment within the same block (delimited by grey lines), a, b and c indicate significant differences according to Duncan's test (p < 0.05).

3 RESULTS

3.1 Transcriptomic profile of oak leaves from the second growth cycle after leaf herbivory over two successive growth cycles

We first analysed whether a first herbivory event (priming event: Ld_prim) during the first growth cycle (GC1) affected plant defences triggered by a second herbivory event (triggering event: Ld_trig) in the second growth cycle (GC2). With this aim, we analysed the expression of herbivore-responsive genes in leaves belonging to GC2, in response to L. dispar herbivory. The pre-selected genes for the analysis (65 genes) are known to be regulated by L. dispar herbivory (Bacht et al., 2019; Table S2). We analysed the gene expression in leaves developed during GC2 of plants that had not been exposed (Ld_trig) or had been exposed (Ld_prim Ld_trig) to an initial herbivory event during GC1 (Figure 1). To further understand the impact of the endogenous rhythmic growth on such responses, we analysed the gene expression in response to herbivory treatments either during shoot flush or during root flush (Figure 1). In addition, to analyse the effect of the ectomycorrhiza on such responses, the gene expression was also analysed in oak microcuttings non-inoculated (Co) or inoculated (Pc) with P. croceum.

NMDS (Figure 2a) and PERMANOVA analyses showed that the stage of plant growth (shoot flush or root flush) was the main factor explaining the differences in herbivory-related gene expression between the treatments (R² = 0.25, p < 0.001). Additionally, Ld_prim and Ld_trig were also significant explanatory factors, although to a lesser extent (R² = 0.05, p = 0.004 and R² = 0.04, p = 0.009, respectively). By contrast, mycorrhizal inoculation had no significant effect on gene expression. To better visualize the influence of the different herbivory events, the treatments were then separated between plant growth stages shoot flush and root flush at Lymantria infestation (Figure 2b,c). For plants at shoot flush (Figure 2b), Ld_trig was the only significant factor affecting the gene expression (R² = 0.17, p = 0.003). In contrast, at root flush, Ld_prim was the only significant factor affecting the gene expression (R² = 0.16, p = 0.009; Figure 2c). Again, mycorrhizal inoculation had no significant effect. Our results evidence a clear separation between responses during shoot flush and root flush showing that the endogenous rhythmic growth has a strong influence on this triggering (Figure 2a).

3.2 Leaf herbivory enhances the induction of jasmonate biosynthesis-related genes in the following growth cycle

We next aimed to analyse the transcriptomic profile associated with the different treatments. We focused specifically on the expression level of genes related to the JA-biosynthesis pathway, pre-selected from Bacht et al. (2019). We found that in oak microcuttings that were not inoculated with P. croceum (Co), herbivory during GC2 increased the expression of the JA-biosynthesis-related genes 13LOX (encodes a 13-lipoxygenase enzyme; Wasternack, 2007), AOC (encodes an allene oxide cyclase enzyme; Wasternack, 2007) and TAN (encodes a tannin transcription factor; Wasternack, 2007) at shoot flush (Figure 3, Ld_trig vs. no_Ld). Interestingly, when the Co plants were exposed to herbivory during GC1, subsequent leaf herbivory during GC2 triggered a stronger expression of these genes, compared with plants that were not primed during GC1 (Figure 3, Ld_prim_Ld_trig vs. Ld_trig). Remarkably, this stronger expression was only observed during shoot flush. In addition, when the Co plants were exposed exclusively to priming herbivory event during GC1, either at shoot flush or root flush, the expression level of these genes during GC2 was similar to that observed in control plants not exposed to L. dispar (Figure 3, Ld_prim vs. no_Ld). Interestingly, the above pattern observed in Co plants during shoot flush related to gene expression induction by Ld_trig compared with no_Ld plants, and an even higher induction by Ld_prim_Ld_trig, was not observed in plants inoculated with P. croceum (Pc), neither during shoot flush nor during root flush (Figure 3). In contrast to the JA-biosynthesis genes, we did not find overall significant differences in the expression of genes related to the abscisic acid (NCED1)-, ethylene (ETH)-, gibberellin (GIB2)- or salicylic acid (ICS1)-related pathways (Figure S1). Overall, these results indicate that a first herbivory event during GC1 does not directly affect the expression of JA-related genes in the following growth cycle. However, it primes the plants for a stronger expression in response to a second herbivory event during GC2. This priming effect occurs specifically when the plants are in shoot flush. In addition, our results further indicate that P. croceum inoculation suppresses the priming effect of enhanced expression of JA-biosynthesis-related genes triggered by a first herbivory event.

3.3 Leaf herbivory induces the expression of jasmonate response-related genes in the following growth cycle

In analogy to the JA-biosynthesis-related genes analysed above, a first exposure of Co plants to herbivory during GC1 led to a higher expression of the JA-responsive genes CHI (encodes a Chitinase I enzyme involved in plant defence response against insects; Adrangi & Faramarzi, 2013), DSP (encodes a plant disease resistance protein; Leach et al., 2001), P450 (encodes the cytochrome P450, CYP94 family, involved in plant defence response against herbivores; Jun et al., 2015) and MCO (encodes a multi-copper protein involved in plant defence response; Janusz et al., 2020) in L. dispar-challenged plants during GC2, compared with L. dispar-challenged plants that were not exposed to herbivory during GC1 (Figure 4, Ld_prim_Ld_trig vs. Ld_trig). Furthermore, the enhanced expression of the JA-responsive genes found in L. dispar-challenged Co plants that were exposed to a first herbivory event was observed specifically at the shoot flush stage. Similar also to the JA-biosynthesis-related genes, when the Co plants were exposed exclusively to the first priming herbivory event, the expression level of the JA-responsive genes in the following growth cycle remained similar to that observed in control plants not exposed to L. dispar (Figure 4, Ld_prim vs. no_Ld).

Interestingly, the above expression pattern observed in Co plants during shoot flush, JA-responsive- related genes induction by Ld_prim_Ld_trig compared with no_Ld and Ld_trig plants, was not visible in plants inoculated with P. croceum (Pc) during neither shoot flush nor root flush (Figure 4).

Overall, these results indicate that an herbivory event during GC2 induces the expression of JA-responsive-related genes only when a previous herbivory event has occurred during GC1. Moreover, this pattern is specific to the plants that are herbivory primed at the shoot flush stage, and it seems abolished when plants are co-inoculated with P. croceum.

3.4 A first leaf herbivory event impedes herbivore performance when feeding in the following growth cycle

To assess whether the higher activation of the JA-related pathway in L. dispar-challenged plants triggered by previous herbivory affects herbivore performance, we assessed the weight of L. dispar larvae after 2 days of feeding on oak leaves. We found that when the plants were faced with herbivory at the shoot flush, the weight of larvae fed during GC2 was lower when plants had been exposed to a previous herbivory event during GC1 (Figure 5a, shoot flush, prim_Ld_trig vs. Ld_trig). Interestingly, this decrease in the larval weight was not accompanied by a greater remaining leaf biomass (Figure 5b). Indeed, for the shoot flush treatment, the biomass that remained was similar regardless of whether plants had been exposed to a previous herbivory event (Figure 5b). By contrast, the weight of caterpillars feeding on leaves during root flush was similar regardless of whether the plants had been exposed to a previous herbivory event or not (Figure 5a, Ld_prim_Ld_trig vs. Ld_trig). Remarkably, during root flush, the remaining leaf biomass was higher in L. dispar-challenged plants that had been previously exposed to herbivory compared with treatments without herbivory exposure in GC1 (Figure 5b, Ld_prim_Ld_trig vs. Ld_trig). Noticeably, these effects were regardless of Piloderma root inoculation or not inoculation. Our results indicate that specifically for herbivory during shoot flush, a first herbivory event induces plant resistance to L. dispar feeding in the next growth cycle (Tables S5 and S6).

FIGURE 5

Caterpillar performance and leaves biomass. (a) The corrected end fresh weight of the Lymantria dispar larvae, feeding on the leaves belonging to the (GC2) second growth cycle, was analysed in oak microcuttings during shoot (SF) and root flush (RF). (b) The leaves biomass was analysed in oak microcuttings during shoot (SF) and root flush (RF). The larvae weight and the leaf biomass were analysed in oak microcuttings exposed to L. dispar only in the second growth cycle (GC2) (Ld_trig) or in two successive growth cycles (GC1 and GC2) (Ld_prim_Ld_trig). Ld_prim and Ld_trig treatments include the oak microcuttings inoculated and not inoculated with Piloderma croceum. Values are means ± SE of 14 independent biological replicates. Asterisk indicates statistically significant differences according to Type I and III test of fix effect (p < 0.05; n = 14).

3.5 A first leaf herbivory event enhances the accumulation of several herbivore-induced plant volatiles (HIPVs) upon subsequent herbivore attack in a successive growth cycle

We assessed whether a first herbivory event during GC1 might affect indirect defences associated with L. dispar feeding in the next growth cycle (GC2). Therefore, we analysed the blend of volatile organic compounds (VOCs) released by L. dispar-challenged oak microcuttings. We sampled VOCs from leaves during GC2 of oak microcutting that had been exposed (Ld_prim_Ld_trig) or not (Ld_trig) to a previous herbivory event during GC1. To further understand the impact of the oak endogenous rhythmic growth on such responses, we analysed VOC blends in the case of herbivory during both shoot flush and root flush (Figure 1). A total of 54 VOCs were detected (Table S3). Among them, we focused specifically on those VOCs for which the emissions were increased after L. dispar herbivory (i.e., herbivore-induced plant volatiles, HIPVs) either during shoot flush or in root flush. A total of 15 VOCs were selected and analysed. These analyses revealed that in non-inoculated plants (Co), three of these 15 volatiles triggered by L. dispar herbivory tended to accumulate at a higher level in response to L. dispar-challenged plants that had been exposed to a first herbivory event (Figure 6, Ld_prim_Ld_trig vs. Ld_trig). These three volatiles included 3-hexenyl acetate, 3-hexenyl butyrate and α-farnesene. Remarkably, this enhanced accumulation was specifically observed in oak microcuttings that faced herbivory at the root flush stage, but not at the shoot flush stage (Figure 6). However, the differences between Ld_prim_Ld_trig and Ld_trig were significant only in the case of 3-hexenyl acetate. It is remarkable that when the plants were exposed exclusively to the first herbivory event, the emission of the three volatiles in leaves of the following growth cycle was similar to that observed in control plants (Figure 5, Ld_prim vs. no_Ld). We also noticed that this emission pattern described for the three volatiles was observed only in Co plants non-inoculated with EMF. The oak microcuttings inoculated with P. croceum did not show significant differences in the emission for these three VOCs between no_Ld, Ld_priming, Ld_trig or Ld_prim_Ld_trig plants (Figure 6).

FIGURE 6

Quantification of HIPV emission in leaves from oak microcuttings. 3–3-Hexenyl acetate, Hexenyl butyrate and α-farnesene emission levels were analysed by thermal desorption gas chromatograph-mass spectrometer in leaves from oak microcuttings not inoculated (Co) or inoculated (Pc) with Piloderma croceum during shoot (SF) and root flush (RF). During SF and RF, the oak microcuttings were not exposed to Lymantria dispar (no_Ld), exposed only in the first growth cycle (GC1) (Ld_prim), exposed only in the second growth cycle (GC2) (Ld_trig) and exposed in the two successive growth cycles (GC1 and GC2) (Ld_prim_Ld_trig). The HIPV emission was analysed in the leaves belonging to the second growth cycle (GC2). Box plots represent the IQR, the bisecting line represents the median, the whiskers represent 1.5 times the IQR and the dots represent the data from seven independent biological replicates. For every data obtained for one treatment within the same block (delimited by grey lines), a and b indicate significant differences according to Duncan's test (p < 0.05).

Overall, these results indicate that a first herbivory event during GC1 does not directly affect the emission of 3-hexenyl acetate, 3-hexenyl butyrate and α-farnesene during the following growth cycle, GC2. However, it could prime the plant for an enhanced emission after a second herbivory event in the following growth cycle. For VOCs, this effect of sensitizing occurs specifically when plants are at the root flush stage, and it seems to be buffered by the inoculation with P. croceum.

4 DISCUSSION

The relevance of defence priming as an adaptive trait for the adjustment of plant defence in unpredictable environments is well established (Conrath et al., 2005; Hilker et al., 2015; Martinez-Medina et al., 2016; Mauch-Mani et al., 2017). However, most of the studies on defence priming focus on herbaceous plants, while studies aiming to understand defence priming and the mechanism underlying this phenomenon in woody plants are scarce (Bittner et al., 2019; Camisón et al., 2019; Mageroy et al., 2019). Besides the mechanistic basis of defence priming in trees, a further key question regarding the defence priming phenomenon is about its resilience, especially in the case of long-lived organisms such as trees (Martinez-Medina et al., 2016). There are some examples of trees that have experienced an insect outbreak in 1 year and are more resistant in the next one (Kaitaniemi et al., 1998; Richards, 1993). Still, it remains unknown how an herbivory event can affect the defence status of newly grown tissues. Here, by using the pedunculate oak clone DF159 as a model system, we demonstrated that a first herbivory attack can result in an enhanced defence response to a second herbivory attack in oaks. Our study further indicates that the priming memory associated with a first herbivory attack in oak leaves can persist across a second growth cycle. This leads to enhanced resistance to herbivory (illustrated by a reduction in larval weight) in newly grown shoot units developed in the following growth cycle.

4.1 Impact of leaf herbivory on jasmonate-related plant defences during the following growth cycle

We found that in plants without ectomycorrhiza fungus treatment, leaf herbivory mainly triggered transcriptional activation of the JA-biosynthesis pathway. Previous studies have evidenced that the JA pathway is a key component in the orchestration of anti-herbivory defences in different woody plant species, including oak trees (Bacht et al., 2019; Kaling et al., 2018). This activation was stronger when the plant was challenged to a leaf herbivory event in the previous growth cycle and it was observed only at the shoot flush stage. The influence of the endogenous rhythmic growth in the defence response displayed by oak trees has recently been described (Bacht et al., 2019). Such differences in the defence patterns between different developmental stages of the endogenous rhythmic growth can be attributed, at least, partially to differences in resource allocation patterns between above and below-ground plant parts during root flush and shoot flush (Herrmann et al., 2015). Noticeably, Bacht et al. (2019) found a specific stronger transcriptional activation of anti-herbivory defences in oak microcuttings of the same clone DF159 at the root flush. An explanation for this difference in observations may rely on the fact that we explicitly investigated plant responses to the triggering stimulus at GC2 after a previous priming stimulus in the GC1, while Bacht et al. (2019) just analysed the plant responses to a single triggering stimulus at the GC1. Thus, these different findings are not contradictory, but rather reflect differences in plant responses depending on the number of herbivory events. Interestingly, we found that the transcriptional activation of the JA-biosynthesis pathway triggered by herbivory was amplified in plants that had suffered from a previous herbivory event during their former shoot growth phase. Moreover, it was associated with enhanced activation of the JA-responsive defence genes analysed here. In accordance with this enhanced immune response, plants that had suffered a first herbivory event at the shoot flush stage displayed enhanced resistance against a following herbivore attack. Indeed, lower biomass was observed in caterpillars fed from plants that had suffered from a previous herbivory event. Noticeably, we did not observe changes in shoot biomass between those plants, and plants that were not exposed to a first herbivory event. We hypothesize that this effect could be related to changes in plant nutritional quality; still, further analysis would be required to decipher the mechanism driving this effect. Overall, these findings indicate that a first herbivore primes the leaves of oak microcuttings for enhanced induction of JA-related defences, leading to increased resistance to a second herbivore attack. It is noticeable that the JA-related defences in plants that had suffered exclusively from the first herbivory event had returned to the basal level at the time of the second herbivory event. This indicates that the information of the first attack (the priming stimulus) serves to respond more efficiently to the second attack (triggering stimulus), further confirming the priming phenomenon.

Along the same lines, several studies have shown the involvement of the JA pathway in defence priming against herbivores in woody plants (Bittner et al., 2019; Mageroy et al., 2019). The ecological relevance of these observations can now be tested in the field, where plant chemistry, herbivory and parasite infections are interestingly modified from the first to the second shoot flush (Gaytan et al., 2022).

Contrary to attacks during shoot flush, a first herbivore attack at the root flush neither boosted the JA-biosynthesis pathway nor activated the JA-responsive defence pathway upon a second herbivore attack in the following root flush. It is remarkable that a first herbivore attack at root flush did not affect plant resistance to a second herbivore attack during the next root flush, but led to an increase in the leaf biomass remaining in the flush. We hypothesize that this observation could be related to changes in the pattern of resource allocation triggered by the first herbivore attack, from roots to shoots. These findings may suggest that a first herbivore at a root flush can enhance oak tolerance to a second herbivore, by altering the patterns of resource allocation (Strauss & Agrawal, 1999). In the same line, Bacht et al. (2019) found that herbivory-induced changes in resource allocation and regulation of genes related to growth and defence in oaks are stronger during the root flush.

According to our results, we hypothesize that a first herbivore attack during shoot flush primes the newly developing leaves of the following growth cycle for enhanced JA-related defences, leading to an enhanced resistance to a second herbivore attack. Differently, when the herbivory events occur at the root flush stages, a first attack enhances plant tolerance to a second attack in the following growth cycle.

4.2 Impact of leaf herbivory on the emission of herbivore-induced plant volatiles in the following growth cycle

In contrast with the JA-related marker genes expression, we found that when plants were at root flush, a first herbivory event led to an enhanced emission of the green leaf volatiles 3-hexenyl acetate, 3-hexenyl butyrate and the sesquiterpene α-farnesene, upon a second herbivory event. Remarkably, these HIPVs are characteristics of the herbivore-damaged leaves of different tree species (Kivimäenpää et al., 2020; McCormick et al., 2014, 2019). These compounds likely play a role as signals in the indirect defence of oaks, by conveying specific information on herbivore location to natural enemies (Volf et al., 2020, 2021). Therefore, our results indicate that the priming memory associated with a first herbivory event in oaks over successive growth cycles is not limited to direct defences. In this line, several studies demonstrate that a first herbivory can prime indirect defences in different plant species (Dicke & Baldwin, 2010; Ghirardo et al., 2012; McCormick et al., 2019). Remarkably, this pattern of priming for enhanced emission of specific HIPVs was observed in plants that were at the root flush stage when facing the first herbivory event. Taken together, our results indicate a relevant role of the endogenous rhythmic growth in the display of defence priming of both direct and indirect defence against subsequent herbivory.

4.3 Effect of Piloderma croceum inoculation on plant defence responses

Interactions with mycorrhizal fungi can strongly influence the plant response to herbivory (Jung et al., 2012; Kaling et al., 2018; Rivero et al., 2021; Song et al., 2013). Mycorrhizal fungi may prime JA-related plant defences against herbivores in different plant species (Jung et al., 2012; Minton et al., 2016; Rivero et al., 2021; Song et al., 2013). Thus, our initial expectation for the third hypothesis was that P. croceum might amplify the defence priming observed in the oak microcuttings. Strikingly, we found that inoculation with P. croceum abolished the priming for enhanced induction of JA-related defences observed in plants at the shoot flush stage, as well as the priming for enhanced emission of HIPVs observed in plants at the root flush stage. Our data on the HIPVs were in accordance to Padmanaban et al. (2022). They observed that beetle feeding stimulated HIPV emission in Populus leaves, with higher levels of green leaf volatile compounds (E)-2-hexenal, (E)-2-heptenal and (Z)-3-hexen-1-ol acetate, but only in trees without mycorrhiza inoculation. Apart from that, Padmanaban et al. (2022) reported the same pattern for the sesquiterpenoids α-amorphene, δ-cadinene and ε-murolene, suggesting that the negative influence on VOC production by mycorrhiza inoculation is not restricted to the green leaf volatiles. Bacht et al. (2019) found that inoculation with P. croceum led to a weaker gene expression response to herbivory compared with non-inoculated plants, suggesting a defence response attenuating role for ectomycorrhiza inoculation. The mechanisms behind the systemic suppression of the display of priming associated with a first herbivory event by P. croceum or other ectomycorrhiza fungi are unknown.

Our results indicate that a treatment with P. croceum has the potential for altering the specific patterns of defence priming across successive growth cycles in oak microcuttings. Inoculation of oak by P. croceum stimulates growth and biomass increase, as well as below and above-ground resource allocation by dropping the expression level of numerous genes in roots and leaves in relation to the endogenous rhythmic growth (Herrmann et al., 2015). Therefore, P. croceum inoculated plants display a completely different functioning and their resistance to herbivory, which we also found here, might rely on such resource enhancement.

Overall, our results show that the priming memory in oaks can be transmitted to the next growth cycle. Our results further demonstrate that the display of defence priming is regulated by endogenous rhythmic growth and modulated by the oak interaction with the ectomycorrhizal fungus Piloderma. Whether this phenomenon takes place in other woody species, and to what extent it is conserved across further ectomycorrhizal-tree interactions await to be investigated.

AUTHOR CONTRIBUTIONS

S. Herrmann, M. T. Tarkka, N. M. van Dam and F. Buscot planned and designed the research. I. Fernández, S. Herrmann and M. Schädler performed experimentation. I. Fernández, M. L. Bouffaud, A. Mártinez-Medina, M. Schädler, M. T. Tarkka and A. Weinhold performed the data analysis. I. Fernández, M. L. Bouffaud, A. Mártinez-Medina, S. Herrmann, M. T. Tarkka, N. M. van Dam and F. Buscot wrote the manuscript. All authors read and approved the final version of the manuscript.

ACKNOWLEDGEMENTS

We acknowledge Dr. Fredd Vergara, Dr. Mario Bauer, Ines Krieg and Barbara Krause for their technical support. Ivan Fernández acknowledges the support from Fundación Salamanca Ciudad de Cultura y Saberes and Ayuntamiento de Salamanca (Grant OTR04036). We thank the Oak platform TrophinOak (www.trophinoak.de) of the Helmholtz-Centre for Environmental Research—UFZ for the realization of the experiments. Open Access funding enabled and organized by Projekt DEAL.

FUNDING INFORMATION

This work was supported by an iDiv Flexpool grant (34600565-02). A.W. and N.M.v.D. acknowledge the support of the German Research Foundation (DFG–FZT 118, 202548816).

CONFLICT OF INTEREST STATEMENT

Not applicable.

Open Research

PEER REVIEW

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/1365-2745.14263.

DATA AVAILABILITY STATEMENT

The raw data of the 96.96 Dynamic Array were deposited in the Gene Expression Omnibus (GEO) Database (https://www.ncbi.nlm.nih.gov/search/all/?term=geo) and are accessible through the ID number GSE212157. The raw data of the volatile organic compounds measured were deposited in the ZENODO Database (https://zenodo.org/) and are accessible through the DOI 10.5281/zenodo.7040241 (Fernandez et al., 2022).