Leaf processing for the analysis of within-canopy variations in microbial communities
On each branch, four leaves were randomly chosen (Figure 1B), collected using sterilized nitrile gloves and placed on autoclaved filter papers. Each leaf was then wrapped with pliers, placed in a 20mL high-density polyethylene (HDPE) vial (Zinsser Analytic®) and frozen with liquid nitrogen. Vials were then stored in a Nitrogen dry shipper (Voyageur, Air liquide ®). They were cold ground using a GenoGrinder (30s at 1750 RPM), after adding two autoclaved stainless steel balls to the tubes. Aliquots of 20mg of powder were then transferred in individual tubes (Micronic, MP32022, In Vitro A/S, Fredensborg, Denmark) along with two sterile stainless steel balls (TissueLyser II, Retsch, Qiagen®) and stored at -80° until DNA extraction.
Leaf processing for the analysis of within-leaf variations in microbial communities
Moreover, three leaves were collected on the two most distant branches (cut from the lower inner part and higher outer part of the crown, respectively) with pliers and a pair of scissors cleaned with DNaway and 70% alcohol between each use. On each leaf, two 1cm x 3cm pieces of leaf were cut; one in the blade edge and the other in the blade center, avoiding the main rib. The entire rib was also removed and cut into 3 to 4 pieces (Figure 1D). The samples were then stored in Eppendorf tubes. Samples were transported in a dry-shipper and stored at -80°C until further analysis.
Leaf processing for the comparison of epiphytic and endophytic microbial communities
Finally, three leaves per branch were collected to compare variations in epiphytic and endophytic communities (Figure 1E). Each leaf was collected with sterile nitrile gloves and placed in a sterile 50ml Falcon tube in a cooler. Leaves were stored in the lab at -20°C until the epiphytic fractions of microbial communities could be detached by sonication and vortexing (sonication 3min, vortex 5sec, shake 10sec), after adding 20mL of PBS-Tween 20 in each tube under a laminar flow hood. The resulting solution was then centrifuged during 15min at 4,000g. The supernatant was removed using a PipetBoy (Integra Biosciences, Fernwald, Germany), by leaving the pellet and approximately 2ml of solution. For each date, solutions of the first replicate of each tree were re-suspended, pooled and centrifuged a second time. The second and third replicates were processed similarly, resulting in three replicates of each foliar habitat for each date and position in the canopy. The supernatant was removed, by leaving out 800µL of solution. The remaining solution was transferred into a 1.1 mL Micronic tube (NovaZine, Lyon, France, ref: MP32033L) itself placed in a 2ml Eppendorf tube. The latter was centrifuged during 5min at 10 000g and the supernatant was removed. Two stainless steel balls were added to each sample and stored at -80C° until DNA extraction. To recover the endophyte fraction of beech leaves, leaves were removed from each falcon tube after sonication and surface-sterilized by immersion for 3 min in a 70% ethanol solution, immersion for 2 min in a 3% calcium hypochlorite solution and rinsing with DNAway and sterilized water. After drying on sterilized filter papers, leaves were cold-ground in Zinsser vials (as previously described).
Leaf processing for the analysis of decomposer communities
On the last sampling campaign, additional leaves were collected on each tree to characterize the succession of microbial communities throughout leaf decomposition (Figure 1F). Twelve leaves were collected on two branches, cut from the lower inner part and higher outer part of the canopy, respectively. Each leaf was placed in a sterile Petri Dish with 2mL of sterile water and incubated in a climatic chamber at 17.33°C with 85.9% of relative humidity, corresponding to the mean climatic conditions recorded on the sampling site in September. Four leaves were then cold-ground in Zinsser vials (as previously described) 1 day, 15 days, 30 days and 60 days after sampling for each branch and tree sampled. Leaf powders from the same branch on the same tree were pooled resulting in three replicates per branch and tree.
Environmental sampling for the characterization of microorganism habitat range
Seven types of environmental samples were collected to characterize the microbial communities in the environment of each tree: (1) twigs cut from the branches collected in the canopy, (2) common ivy (Hedera helix L.) and (3) bryophytes growing on the tree trunk, (4) dead wood bark, (5) litter and (6) butcher’s-broom leaves and twigs (Ruscus aculeatus) beneath each tree, and (7) river bedding. All samples were collected in triplicates with sterile latex gloves and stored in 50mL Falcons tubes, transported in coolers and stored at -20°C until the epiphytic microorganisms could be detached by sonication and vortexing (as previously described, except that the first solution was filtered with a sterile cell strainer (100µm of porosity, nylon; Dutscher) to remove bark and litter fragments).
Leaf traits measurements
Twenty-three leaf traits representing leaf morphology, chemistry and physiology were measured on each sampled branch (Table 1). All traits were measured on several leaves per branch and then averaged at the branch level.
Table 1. Leaf chemical, physiological and morphological traits measured on each sampled branch. NM indicates the number of independent measurements per branch. Traits highly correlated (R2> 0.7) to each other were removed from the dataset (Fig. S1). Foliar traits selected for further analysis are marked with a cross.
Chemical traits measurement
The four leaves collected on each branch for the analysis of within-canopy variations in microbial communities (Figure 1B) were also used to quantify leaf chlorophyll a and b, amino acids and non-structural carbohydrates (fructose, glucose and sucrose) at the HiTMe platform of the MetaboHUB Bordeaux facility (France). Chemical compounds were extracted from aliquots of 20 to 50 mg leaf powder, by shaking them with 250µL of 80% ethanol then heating at 80°C for 20 minutes and centrifuging at 13,400 rpm for 5 minutes. This step was repeated a second time with 150µL of 80% ethanol and a third time with 250µL of 50% ethanol. At each step, the ethanolic supernatant was collected, representing at the end a total volume of 650 µL of solution. Leaf chlorophyll a and b concentrations were determined as in Arnon (1949) by measuring the absorbance at 645 and 665nm of 35µL of ethanolic extract mixed with 120µL of 98% ethanol. Concentrations of glucose, fructose and sucrose were quantified from the ethanolic extract using a method adapted from Stitt et al. (1989). Aliquots of 5µL of ethanolic extract were added to a 160µL mix containing 0.1 M Hepes-KOH buffer (pH 7), 3 mM MgCl2, 100 mM ATP, 45 mM NADP, 1% polyvinylpyrrolidone and 5 glucose-6-phosphate dehydrogenase units. The glucose, fructose and sucrose contents were quantified from the change in absorbance triggered by the sequential addition of one unit of hexokinase, one unit of phosphoglucose isomerase and 30 invertase units , respectively. The quantification of total free amino acids was based on the reaction of amino acids with fluorescamine (Bantan-Polak et al., 2001). The final 213µL reaction mixture contained 8μL of ethanolic extract mixed with 15µL of sodium borate at 0.1 mM, pH 8.0, 90 µL of 0.1% of fluorescamine and 100 µL of water. The solution was excited at 405 nm and read at 485 nm. Absorbance and fluorescence measurements were performed using SAFAS MP96 and SAFAS Xenius microplate readers, respectively. Flavonol content and Nitrogen Balance Index were measured in the field on 6 leaves per branch, using the Dualex forceps (Dualex®. ForceOne CNRS-LURE. France).
Physiological traits measurement
These six leaves were also used to measure leaf stomatal conductance and water content. Leaf stomatal conductance was measured in the field using a Porometer (SC-1, Decagon Devices Inc., Pullman. WA. USA). Fresh weight was measured on the day of sampling using a balance and dry weight was measured after 48 h of desiccation in an oven at 60°C (Sayer, 2006). Leaf water content was calculated as LWC (%) = 100 x (Fresh weight – Dry Weight) / Fresh weight.
Two additional leaves were taken from each branch to measure leaf osmotic potential. They were immediately immersed in liquid nitrogen using pliers and then placed on filter paper. For each leaf, two 6mm diameter foliar discs were cut with a punch on each side of the midrib, placed in an Eppendorf in a dry-shipper and then stored at -80°C. The osmotic potential (πosm) was measured by placing a frozen foliar disc in a calibrated osmometer chamber (C52, Wescor, USA) connected to a datalogger (Psypro, Wescor, USA). The turgor loss point (πtlp) was then estimated as πtlp = 0.832 πosm – 0.631 (Bartlett et al. 2012). The turgor loss point represents the leaf water potential at wilting and is a more reliable predictor of plant drought tolerance than the osmotic potential (Bartlett et al. 2008; Bartlett et al. 2012).
Leaf water potential was measured in the field at predawn and at the time of sampling using a Scholander-type pressure chamber (DGMeca, France), on two to four leaves per branch.
Morphological trait measurements
Three leaves per branch were also collected to measure the thickness of the upper epidermis, palisade mesophyll, spongy mesophyll, lower epidermis and the total leaf thickness. These leaves were placed in plastic bags, transported in a cooler and stored at 5°C. A transversal, 1cm large strip was cut with scissors in the middle of each leaf. Each strip was then dehydrated by a series of 5 graduated ethanol baths containing 30, 50, 70, 85 and 100% of ethanol for 30, 30, 60, 50 and 30min, respectively (Chen et al., 2016). Ultra-fine cross sections (between 30 and 40 µm) were then finely cut using a microtome (GSL1, Schweingruber inst, Switzerland), avoiding the midrib, by holding the samples between two plastic lamellae. The foliar strips were then stained by adding a solution composed of 300 µL of methylene blue, 100µL of safranin and 1.6mL of distilled water diluted in 9mL of methylene blue. Microscopic pictures (magnification x 200) were taken using a microscopic imaging software (LasX, Leica Microsystems, Germany) and a microscope (DM2500, Leica Microsystems, Germany) equipped with a camera (MC190HD, Leica Microsystems, Germany). For each cut, the thickness of 4 types of leaf tissue and the total thickness were measured using the ImageJ software (http://imagej.nih.gov/ij/). Specific leaf area was estimated by dividing the leaf surface by the dry matter.
Metabarcoding of bacterial and fungal communities
DNA extraction and amplification were performed under a hood in a confined laboratory. Total genomic DNA of leaf and environmental samples was extracted using the Qiagen DNeasy plant kit (Qiagen) accordingly to the manufacturer’s protocol, except that samples were incubated with AP1 solution during one hour at 65°C and that DNA extracts were eluted twice, the first time with 50µL of Buffer APE and the second time with the first elution solution.
The V5-V6 region of the bacterial 16S rDNA gene was amplified using 16S primers 799F-1115R (Redford et al., 2010; Chelius et al., 2001) to exclude chloroplast DNA. To avoid a two-stage PCR protocol and reduce sequencing biases, each primer contained the Illumina adaptor sequence, a tag and a heterogeneity spacer, as described in Laforest-Lapointe (2017) (799F: 5’).For 16S primers, HS represented a 0–7-base-pair heterogeneity spacer and for all primers “x” a 12 nucleotides tag. The PCR mixture (20µL of final volume) consisted of 2µL of each of the forward and reverse primers (2µM), 2µL of dNTPs (2mM), 4µL of 5X HotStart Phusion HF Mix, 0.6µL of DMSO, 0.2 µL of Phusion Hot Start II polymerase (ThermoScientific), 1µL of template and water up to 20 µl. PCR cycling reactions were conducted on a Veriti 96-well Thermal Cycler (Applied Biosystems) using the following conditions: initial denaturation at 98 °C for 30 s followed by 35 cycles at 98°C for 15 s, 60°C for 30s, 72°C for 30s with final extension of 72 °C for 10 min.
Each PCR plate had 2 positive controls and 5 to 6 negative controls. Bacterial positive controls were represented by the DNA of two marine bacterial strains (Sulfitobacter pontiacus and Vibrio splendidus); the first positive control included 1 µL of 10 ng. µL-1 DNA of Vibrio splendidus only and the second included 1µL of an equimolar mixture of both strains. Fungal positive controls were represented by the DNA of one fungal strain growing on hyper saline environments (Debaryomyces hansenii) and second strain growing on hyper sweet or salty environments (Wallemia sebi); the first positive control included 3 µL of 10 ng. µL-1 DNA of Debaryomyces hanseni only and the second included 3µL of an equimolar mixture of both strains. Negative controls were represented by one to two controls containing PCR mix without any DNA template and four to five controls containing DNA extraction buffer solution in each PCR plate. Four additional controls were performed each time samples had been opened under non-steel conditions. Potential contamination during leaf weighing and grouding was analysed by opening three empty tubes and rinsing them with 1mL of sterile water. The putative contaminants present in the PBS-Tween solution used to separate the epiphyte and endophyte fraction of the leaves and in the sterile water added to the petri dishes during the decomposition of the leaves were analyzed in 1mL of each solution. Each type of negative control was represented by 1ml of solution of 3 replicates pooled together.
Table of contents :
I. Les micro-organismes, des éléments essentiels au fonctionnement du système Terre
II. La forêt, un écosystème diversifié en habitats microbiens indispensables à son fonctionnement
II.1. Les micro-organismes associés à la formation des sols
1.A. Les micro-organismes décomposeurs des litières
1.B. Les micro-organismes décomposeurs du bois mort
1.C. Les sols représentent le plus important des habitats microbiens en forêt
II.2. La plante et son microbiote forment une entité fonctionnelle : l’holobionte
2.A. Les micro-organismes colonisent les organes reproducteurs des plantes
2.B. Les micro-organismes colonisent les organes végétatifs des plantes
a. La rhizosphère est un habitat riche en nutriments
b. La phyllosphère est un habitat oligotrophe exposé à de multiples contraintes
Caractéristiques anatomiques et chimiques foliaires
Conditions microclimatiques foliaires
Les interactions biotiques à l’intérieur de la phyllosphère
III. Influence des micro-organismes de la canopée des arbres sur le fonctionnement des forêts
III.1. Effets positifs et négatifs des micro-organismes sur la fitness des arbres
1.A. Stimulation et inhibition de la germination des graines et de la croissance des plantes
a. Attaques de pathogènes et d’herbivores sur les graines
b. Régulation hormonale de la croissance des plantes
1.B. Contribution à la nutrition azotée
1.C. Régulation des stress biotiques
a. Effets directs : Interactions biotiques entre micro-organismes
b. Effets indirects : Activation du système immunitaire de la plante hôte
III.2. Contribution aux cycles biogéochimiques
2.A. Cycle des nutriments
2.B. Cycle de l’eau
IV. Métagénomique ciblée : une méthode devenue incontournable pour étudier les communautés microbiennes
IV.1. Préparation des échantillons
IV.2. Amplification des gènes microbiens
IV.3. Séquençage des gènes microbiens
IV.4. Traitements bioinformatiques des séquences
V. Objectifs de la thèse
II Within-canopy variation in phyllosphere microbial communities of european beech (fagus sylvatica): magnitude, drivers and functional consequences
III Maternal effects and environmental filtering shape seed fungal communities in oak trees
IV Quantitative and qualitative assessment of bacterial fluxes among soil, plant phyllosphere and near-surface atmosphere
V Conclusion et perspectives