Anthropogenically elevated inputs of nitrogen (N), phosphorus (P) and potassium (K) can affect the carbon (C) budget of nutrient-poor peatlands. Fungi are intimately tied to peatland C budgets due to their roles in organic matter decomposition and symbioses with primary producers; however the influence of fertilization on peatland fungal composition and diversity remains unclear. Here we examined the effect of fertilization over 10-yrs on fungal diversity, composition, and functional guilds along an acrotelm (10-20 cm), mesotelm (30-40 cm), and catotelm (60-70 cm) depth gradient at the Mer Bleue bog, Canada. Simultaneous N and PK addition decreased the relative abundance of ericoid mycorrhizal fungi (ErMF) and increased ectomycorrhizal fungi (EcMF) and lignocellulose-degrading fungi. Fertilization effects were not more pronounced in the acrotelm relative to the catotelm, nor was there a shift towards nitrophilic taxa after N addition. The direct effect of fertilization significantly decreased the abundance of Sphagnum-associated fungi, primarily owing to the overarching role of limiting nutrients rather than a decline in Sphagnum cover. Increased nutrient loading may threaten peatland C stocks if lignocellulose-degrading fungi become abundant and accelerate decomposition of recalcitrant organic matter. Additionally, future changes in plant communities, strong water table fluctuations, and peat subsidence after long-term nutrient loading may also influence fungal functional guilds and depth-dependencies of fungal community structure.

Fertilization experiments

Two sets of fertilization experiments were established in 2000-2001 and 2005, respectively. Experiment 1 included six treatments with different rates of N additions with or without P and K. To better understand the effect of different rates of N addition alone, Experiment 2 was established in 2005. In both experiments, triplicate plots (3 m x 3 m) were fertilized every three weeks from early May to late August (7 times per year) by applying ammonium nitrate (for N treatments) and/or mono potassium phosphate (for PK treatments). The two experiments combined represented a total of 27 plots (9 treatments x 3 replicates) and were established on large hummocks which covered 70% of the terrain. There have been no changes to the distribution of hummocks versus hollows with the fertilization treatments.

Vegetation survey

Plant communities were characterized by the point intercept method in July 2014. The number of times (‘hits’) a specific plant species and organ (leaf/woody stem/flower/moss shoot) contacted a metal rod (4 mm in radius) over 61 grid points in a 60 cm x 60 cm frame were recorded. Sphagnum cover (%) was estimated by the number of hits divided by 61 and multiplying by 100. Vascular species abundance (not identical to cover) was estimated by the total number of hits of all organs per m² for each species.

Peat sampling

In mid-July 2014, a single peat core was taken from a location selected at random within each plot, with 3 depth increments saved per core (27 cores x 3 depth increments = 81 samples total). The upper two depth increments of peat (10-20 and 30-40 cm below the peat surface, 10 cm length x 10 cm width x 10 cm height) were collected using a clean bread knife, and an Eijkelkamp auger was used for deeper peat (60-70 cm, 5.2 cm diameter x 10 cm height). At Mer Bleue, the uppermost layer (10-20 cm) represents the typical acrotelm peat which is rarely saturated (i.e., oxic) while the lowermost layer (60-70 cm) is from the catotelm where peat is rarely above the water table (i.e., anoxic). The middle layer (30-40 cm) is from the mesotelm through which the water table has seasonal oscillations and thus it is a biogeochemical ‘hotspot’ where roots never dry out and are rarely exposed to anoxic conditions in the growing season. At the time of collection, water table depth and the location of the water table relative to the mid-point of the depth of the peat sample were determined in each core hole, and peat temperature was measured from each depth increment. Samples were subdivided into two subsamples (for DNA analyses and physicochemical analyses), transported to the laboratory on dry ice, stored at -20°C, and shipped to the USDA Forest Service, Northern Research station (Houghton, MI, USA) where they were stored at -20°C until further processing.

Physicochemical properties of peat

A variety of elemental concentrations, pH, humification index, and organic chemical composition of peat were measured. Using moist peat from each sample, humification level was characterized using the von Post humification index, and peat pH was measured using a 2:1 ratio (volume:volume) of distilled water to peat. A subsample of peat was oven-dried at 60 °C to a constant weight and ground using a Wiley Mini Mill with size 60 mesh. Total C, N, and S concentrations were determined by dry combustion on a VarioMacro CNS Analyzer in the Watmough lab at Trent University. P and K were analyzed using ICP-MS (Varian 810, now part of Agilent Technologies, Santa Clara, CA, USA) in the Spiers lab at Laurentian University after combustion with a muffle furnace and a modified EPA3050A block digestion. Fourier-transform infrared spectroscopy (FTIR) analysis of dried peat and assignment of peaks to carbohydrate and aromatic classes was carried out in the Chanton lab at Florida State University.

DNA extraction, amplicon library preparation, and sequencing

In a 50 mL centrifuge tube, 10 mL of wet peat was placed with twenty 3.2 mm chrome-steel beads. The sample was pulverized for two minutes on a modified mini-beadbeater-96. A subsample of 0.5 g pulverized peat was used for DNA extraction with a PowerSoil® DNA Isolation Kit. The extraction procedure followed the manufacturer’s instructions, with the addition of a 10-min vortex followed by incubation at 65 °C for 30-min, all in the C1 lysis buffer. DNA was cleaned using a MoBio PowerClean® Pro DNA Clean-Up Kit and quantified with a Qubit Fluorometer.

Cleaned DNA extracts were pooled and sequenced at the U.S. Department of Energy Joint Genome Institute. The amplicon library preparation and sequencing were performed following Joint Genome Institute Protocols for Illumina MiSeq community amplicon sequencing. Briefly, the fungal ITS2 region was targeted with the forward primer fITS9 (GAACGCAGCRAANNGYGA) and the reverse primer ITS4 (TCCTCCGCTTATTGATATGC). The full-length primer was composed of an Illumina adapter, an 11-bp barcode unique to each sample on the reverse primer, a 10-bp primer pad, a 3-bp spacer pad and the primer sequence. DNA samples were combined in equimolar aliquots and sequenced on an Illumina Miseq platform with 2×300 bp chemistry. Eight samples were lost due to poor sequencing, which occurred independently and were not treatment- or depth-specific.