Influence of soil type and vertical zonation on soil fungal communities associated with natural jack pine forests
Data files
May 13, 2024 version files 1.80 MB
-
Chemistry_dataset.xlsx
-
Fungal_ASV_dataset_non_rarefied.xlsx
-
Fungal_sequence_correspondence_dataset.xlsx
-
README.md
Abstract
Natural jack pine forests established on sandy eskers and clayey soils represent a unique habitat in the boreal forest in northeastern Canada and are generally associated with well-differentiated soil horizons, but little is known regarding their soil fungal diversity and how soil type and vertical zonation would shape their communities. To address this question, we characterized soil fungal communities from litter, organic and mineral horizons in 18 natural jack pine forests in esker and clayey soils using ITS2 rDNA metabarcoding. Ectomycorrhizal fungi dominated the organic and mineral horizons in esker soils, while saprotrophic fungi were dominant in clayey soils. The dominance of ectomycorrhizal fungi coincided with higher C accumulation within the organic horizon in esker soils. The nutrient-poor and more acidic esker sites harbored less diverse ectomycorrhizal and saprotrophic communities than clayey soils. Several stress-tolerant and melanin-rich fungal taxa were indicators of esker soils. However, fungal communities were more distinct between horizons than soil types and soil type effects differed along the soil profile. Our findings highlight the need to fill knowledge gaps in soil fungal diversity associated with jack pine forests and pave the way for adapted and multi-resource forest management in esker forest ecosystems considering their underground fungal specificities.
README: Influence of soil type and vertical zonation on soil fungal communities associated with natural jack pine forests
https://doi.org/10.5061/dryad.ffbg79d33
Three different datasets are presented. First, an XLSX file (Fungal_ASV_dataset_non_rarefied.xlsx) containing the occurrence of fungal ASVs across our samples, including extraction and PCR controls, based on an ITS2 Illumina metabarcoding (environmental DNA) technique. This dataset is non rarefied. This dataset also includes the taxonomic classification for each fungal ASV (at the Class, Order, Family, Genus, and Species levels), as well as ecological classification and attributes (Trophic Mode, Guild, Growth Morphology, Trait, etc.). The confidence of these assignations is also indicated, as well as the source if applicable. Second, an XLSX file (Fungal_sequence_correspondence_dataset.xlsx) containing the correspondence between the ASVs presented in the first dataset (column ASV_ID) and the associated sequences (column Sequence). Third, an XLSX file (Chemistry_dataset.xlsx) containing chemical properties for our samples. The enumeration of the samples mirrors the one from the first dataset.
Descriptions
Fungal_ASV_dataset_non_rarefied.xlsx
- ASV_ID: Name (or identity) of the different ASVs.
- ITS2-CTRL-neg-extraction0: Negative control sample for DNA extraction, number 0.
- ITS2-CTRL-neg-extraction1: Negative control sample for DNA extraction, number 1.
- ITS2-CTRL-neg-extraction2: Negative control sample for DNA extraction, number 2.
ITS2-CTRL-neg-PCR: Negative control sample for PCR amplification.
Sample indexation follows this structure: ITS2-(Soil type)(Site number)(Horizon). For soil type, it can be either SE = esker soil, or SA = clayey soil. For esker soil, site number ranges from 1 to 12, and 1 to 6 for clayey soil. For horizons, it can be either L = litter horizon, O = organic layer, or M = mineral soil. The combination of these three parameters made up a unique index for each sample of our study. For exemple, sample ITS2-SA1O corresponds to the organic layer from clayey soil of site 1. Sample are listed from column 6 to 59 (total of 54 samples).
The taxonomic classification fo each ASV, based on the FUNGuild database, is listed according to seven levels: Kingdom, Phylum, Class, Order, Family, Genus, and Species. The taxonomic classification follow this structure: (First letter of the classification level)(Taxonomy). For the first letter of the classification level, it can be either k = Kingdom, c = Class, o = Order, f = Family, g = Genus, and s = Species. If we failed to assign a taxonomic classification at a given level, it is indicated as (First letter of the classification level)unidentified. Since ASVs were filtered for Fungi, the column Kingdom contains only one category (KFungi).
ASVs were also assigned an "ecological" classification, based on the FUNGuild database, listed as Trophic Mode, Guild and Growth Morphology. The detailed definitions indicated below refers to the information provided by FUNGuild (https://github.com/UMNFuN/FUNGuild
Trophic Mode: One of three trophic categories [Pathotroph = receiving nutrients at the expense of the host cells and causing disease (e.g., biotroph, parasite, pathogen, etc.); Saprotroph = receiving nutrients by breaking down dead host cells (e.g., wood rotters, litter rotters, etc.); Symbiotroph = receiving nutrients by exchanging resources with host cells (e.g., ectomycorrhiza, lichens, etc.)].
Guild: Provide a relevant guild descriptor [Pathotroph: Animal Pathogen (including human pathogens - typically annoated as such), Bryophyte Parasite, Clavicipitaceous Endophyte, Fungal Parasite, Lichen Parasite, Plant Pathogen; Saprotroph: Dung Saprotroph (i.e., coprophilous), Leaf Saprotroph (e.g., leaf litter decomposer), Plant Saprotroph, Soil Saprotroph (e.g., rhizosphere saprobe - typically annoated as a rhizosphere fungus), Undefined Saprotroph (e.g., a general saprobe, or in cases where the ecology is not known but suspected to be a saprobe), Wood Saprotroph (e.g., wood rotting fungi); Symbiotroph: Ectomycorrhizal, Ericoid Mycorrhizal, Endophyte, Epiphyte, Lichenized (i.e., lichen)].
Growth Morphology: Basic morphological categories including "Agaricoid", "Boletoid", "Clathroid", "Clavarioid", "Corticioid", "Facultative Yeast", "Gasteroid", "Hydnoid", "Microfungus", "Phalloid", "Polyporoid", "Rust", "Secotioid", "Smut", "Thallus", "Tremelloid", or "Yeast" should be noted.
Trait: Other functional or morphological traits such as "Brown Rot", "Hypogeous", "Poisonous", "Soft Rot", "White Rot" or "Contact Exploration Type" would be appropriate for this field.
Confidence Ranking: "Highly Probable" (= absolutely certain), "Probable" (= fairly certain), "Possible" (= suspected but not proven, conflicting reports given, etc.).
Notes: Any other relevant information related to the taxon (e.g., "Facultative human pathogen causing coccidioidomycosis" for Coccidioides immitis, "Host - Arecaceae - Casuarinaceae, Palmae, Sapotaceae" for Xylaria obovata, or "Syn. Umbilicaria" for Actinogyra).
Citation/Source: Publication, website, etc. from which the corresponding guild information was derived. Data for taxa that is based on peer-reviewed publications is preferred.
Fungal_sequence_correspondence_dataset.xlsx
- ASV_ID: Same ASV name (or identity) as in Fungal_ASV_dataset_non_rarefied.xlsx
- Sequence: Generated ITS2 sequence for each ASV.
Chemistry_dataset.xlsx
- Site_number: Same site indexation as in Fungal_ASV_dataset_non_rarefied.xlsx. Site number for esker soils ranges from 1 to 12, and from 1 to 6 for clayey soils.
- Soil_type: Esker = esker soil, Clay = clayey soils.
- Horizon: Litter = litter horizon, Organic = organic layer, M = mineral soil.
- Organic_matter: Organic matter content (g/Kg)
- Humidity: Percentage of humidity (%).
- Total_C: Total carbon content (g/Kg)
- Total_N: Total nitrogen content (g/Kg)
- P: Phosphorus content (mg/Kg)
- K: Potassium content (mg/Kg)
- Ca: Calcium content (mg/Kg)
- Mg: Magnesium content (mg/Kg)
- Mn: Manganese content (mg/Kg)
- Zn: Zinc content (mg/Kg)
- Al: Aluminium content (mg/Kg)
- Fe: Iron content (mg/Kg)
- pH: Potential of hydrogen.
Methods
Study area
The study was conducted in the Abitibi-Témiscamingue region, Quebec (Canada). The study area is located in northern Quebec’s clay plain, within the balsam fir-paper birch (Abies balsamea (Linnaeus) Miller-Betula papyriferaMarshall) bioclimatic domain (Grondin 1996). Climatic conditions are subpolar, subhumid, and continental with an annual average temperature and precipitation of respectively between 1-2 °C and 825-975 mm according to the Canadian climate normal 1981-2010 station data (Environment Canada, 2023). Forest vegetation mainly consisted of jack pine stands with interspersed stems of Picea mariana (Miller) Britton, Sterns & Poggenburgh, Populus tremuloides Michaux, Betula papyrifera Marshall, Abies balsamea (Linnaeus) Miller, and Acer rubrum Linnaeus. Specifically, the Abitibi region is characterized by natural gradients of soil physicochemical properties, from sandy eskers to clayey soil types. Indeed, esker soils account for ~3% of the Abitibi’s surface while the remainder of the region’s flat topography (~57%) is dominated by clayey soils (Robitaille and Saucier 1998; Cloutier et al. 2007). Abitibi eskers have mineral surface deposits with an overall thickness of more than 25 cm, an illuvial horizon with coarse texture, and a high stone content. Conversely, glacio-lacustrine clayey soils, which are associated with Luvisols or Gleysols in the Abitibi region (Laverdière and De Kimpe 1984; Bergeron et al. 2007), are characterized by an illuvial horizon with medium texture and a low stone content (Blouin and Berger 2002).
Site selection
Site selection was based on four main criteria, namely soil type (sandy esker/clayey), dominant vegetation (jack pine, minimum 50-75% of canopy cover), stand age (minimum 50-60 years), stand origin (natural fire-origin, totally or mostly unimpacted by anthropogenic disturbance), and accessibility (maximum 350 m from forest roads). We used satellite images and forest inventory data from the Quebec government’s online platform ‘Forêt Ouverte’ for site selection. We selected 18 natural jack pine stands (12 on sandy eskers, 6 on the clay plain). Only six clayey sites fitting our selection criteria were found within the prospected territory due to the scarcity of unmanaged (natural) jack pine stands on clayey soils. Sites were separated by a minimum distance of 5 km to ensure independence between samples and avoid pseudoreplication. Finally, the study design consisted of a first-level factor (soil type) with two levels (sand and clay), and a second-level factor (soil horizons) with three levels (litter, organic, and mineral).
Soil sampling
Soil samples were collected during the summer of 2022 to characterize soil fungal communities and obtain the physicochemical composition across the soil profile. A circular plot with a 28 m radius and an area of 2,500 m2 was established at each study site, at least 20 m from the forest edge to avoid related variability bias (Dickie and Reich 2005). Within each plot, five jack pine trees were randomly selected (at least 10 m apart), and two soil sample replicates per tree (2-3 m apart) were collected, following a modified protocol by Tedersoo et al. (2014; 2021). Sampling locations around trees were randomly selected with the restriction criteria of being strictly opposite (angle of 180°). In total, 180 soil cores (2 replicates × 5 trees × 18 sites) were collected.
For each sampling spot, forest fallen litter (dead leaves, needles) (hereafter referred to as litter horizon) was collected in plastic bags for DNA and physicochemical analyses, followed by samples from the organic layer (hereafter referred to as organic horizon). The latter represented a dark-colored layer rich in organic matter at various decomposition stages, mainly composed of fallen plant material. These organic samples included both F and H horizons. After the removal of the organic layer, we sampled the mineral soil (hereafter referred to as mineral horizon) using a pedological auger (25 cm in length and 7.5 cm in diameter), capturing both eluvial and illuvial horizons. For esker sites, the mineral soil profile was characterized by the formation and accumulation of organo-metallic assemblages, involving Al and Fe elements, in a leached grey eluvial and rust-brown illuvial horizon, respectively. Due to the variable thickness of the eluvial horizons within both esker and clayey soil profiles, the proportion of this horizon in the mineral soil samples varied among sites. Large roots and coarse woody debris were systematically removed from organic and mineral material while sampling.
Samples were pooled per horizon for each site, resulting in one composite sample for each horizon. This represents 18 composite samples per horizon for a total of 54 composite soil samples (3 horizons × 18 sites). Each composite soil samples were divided into two sub-samples: one for eDNA-based soil fungal community analysis and the other for physicochemical analysis. Composite organic and mineral soil samples for eDNA analysis were sieved with a 6 mm mesh in the field, placed in plastic bags, transported in an ice-filled cooler and stored at – 25°C in the Ecology Research Group of Abitibi RCM (GREMA) laboratory until further processing. Composite litter samples for eDNA analysis were crushed in liquid nitrogen using a mortar and pestle prior to DNA extraction. Composite soil samples for physicochemical analyses were sieved with a 4 mm mesh using an automatic vibrating sieve AS 200 Control (ATS Care Retsch, Haan, Germany) after being forced-air dried at room temperature for 14 days to facilitate the sieving process.
Soil physicochemical analyses
Composite soil samples (> 50g) were sent to the organic and inorganic chemistry laboratory of the Forest Research Direction (Quebec, QC, Canada) for physicochemical analysis. Total nitrogen (N) and carbon (C) contents were determined by combustion using a CN 928 elemental analyzer (LECO Corp., St Joseph, MI, USA) with thermal conductivity detection for nitrogen and non-dispersive infra-red (NDIR) cell detection for carbon. Organic matter content (hereafter OM) and percentage of humidity (hereafter humidity) were determined by incineration, a method commonly referred to as the loss on ignition (LOI) method (Davies 1974). Soil pH was measured using 10 g of soil mixed with 20 mL of distilled water with an Orion VersaStar Pro pH meter (Thermo Fisher Scientific Inc., Pittsburgh, PA, USA). Elements (P, K, Ca, Mg, Mn, Zn, Al, Fe) were extracted using the Mehlich-III method (Mehlich 1984) and measured by plasma atomic emission spectroscopy (Optima 8300 model, ICP-OES, Perkin Elmer, Waltham, MA, USA). Particle size distribution and textural class determination were performed on soil samples consisting of fine earth (<2 mm) containing 5% or less carbon using the Bouyoucos method (Bouyoucos 1962).
DNA extraction, amplification and library preparation
DNA extraction was carried out at the Environmental Genomic Laboratory of the Laurentian Forestry Centre (EGL-CFL) on 150 mg of composite litter and organic samples, and 250 mg of composite mineral samples using a DNeasy powersoil pro kit (Qiagen, Valencia, CA, US) and the QIAcube automated instrument (Qiagen, Hilden, Germany) in accordance with the manufacturer’s instructions. DNA was quantified with the Qubit™ dsDNA BR Assay Kit (Thermo Fisher Scientific Inc, Wilmington, USA). Fungal DNA amplicon libraries were prepared at the EGL-CFL following a previously described procedure (Samad et al. 2023) and sequenced on an Illumina MiSeq platform at the Next Generation Sequencing Platform of the CHU de Québec-Université Laval Research Centre using a MiSeq v3 600-cycle Reagent Kit. Even though one ITS region contains less genetic information than the entire ITS (Tedersoo et al. 2022), the ITS2 region of the fungal ribosomal DNA was amplified using the primer set ITS9F (5’-GAACGCAGCRAAIIGYGA-3’) and ITS4R (5’-TCCTCCGCTTATTGATATGC-3’ (White et al. 1990; Ihrmark et al. 2012; Rivers 2016) because of limitations to sequence length that can be obtained with Illumina Sequencing. Negative controls in the DNA extraction and PCR amplification were used to control potential contaminants during the process.
Bioinformatic and sequence analyses
All bioinformatics analyses were performed in QIIME2 v2023.2.0 (Bolyen et al. 2019). Demultiplexed FASTQ (R1 and R2) files were first imported in QIIME2 using the ‘qiime import’ command which converts the data into a QZA archive file so it can be processed by the rest of the QIIME2 workflow. First, primers were removed using the QIIME2 implementation of CutAdapt (Martin 2011). Resulting forward and reverse reads went through the DADA2 (Callahan et al. 2016) pipeline for sequence quality control and feature table construction using the ‘’qiime dada2 denoise-paired’’ command. During this process, low-quality regions of the sequences were trimmed, paired reads assembled, chimeric sequences filtered, remaining high-quality sequences dereplicated, and singletons and very low-frequency abundance ASVs removed using the “qiime feature-table filter-features” command. The output of this step was a feature table (i.e., ASV table) which contains read counts for each unique sequence in each sample of the dataset, and feature data which contains sequences corresponding to each ASV.
A taxonomy was assigned to each ASV based on the UNITE reference database (version 9.0) (Abarenkov et al. 2010, 2023; Kõljalg et al. 2005, 2013; Nilsson et al. 2019b) using the ‘‘classify-sklearn’’ with a Naive Bayes classifier. All ASVs not assigned to the ‘Fungi’ kingdom were removed from further analyses using the “qiime taxa filter-table”. The resulting fungal ASV table was used to perform downstream statistical analyses. Functional annotation of ASVs was performed for most fungal guilds using FUNGuild version 1.1 (Nguyen et al. 2016).