Sampling site and soil sampling
In June 2022, soil sampling was conducted in the landward slopes of Ishikari coastal dunes, Hokkaido, Japan (43˚14'N, 141˚20'E), of which vegetation was described in Kawahara and Ezawa (2013) and Cahyaningtyas and Ezawa (2024). Three sampling squares of 1-m-quadrat were designated at 100 m intervals along the coastal line, and six pairs of a root-soil-core sample with or without an aboveground part of M. sinensis plant were collected from each of them by stainless-steel core samplers [5 × 5 cm (diam/ height), 100 mL in vol]. The aboveground part was cut at 5 cm above the ground level before collecting by the sampler, which were designated as NC inocula. Whereas the samples without a living plant were sieved on a 2-mm-stainless mesh to destruct AM fungal hyphal networks before trap culture and termed as SP inocula. Twenty kilograms of sand was also collected from the seaward slope of the dunes where no vegetation was present and autoclaved for 2 h as a base medium for trap culture.
Soil trap cultures
As a main experiment, the NC inoculum was transferred to a cylinder-shaped 37-µm-nylon mesh bag [5 × 5 cm (diam/ height)] inserted in a 400-mL plastic pot, and then the pot was filled with a mixture of the SP inoculum (100 mL) and base medium (200 mL) (Fig. 1). The nylon mesh allows AM fungal hyphae, but not plant roots, to pass through. As a supporting experiment, two treatments were set up to assign either of the two life-history strategies NC and SP to the fungi detected in the main experiment as follows. As a NC treatment, the NC inoculum was transferred to the mesh bag, and the pot was filled with a mixture of an autoclaved SP inoculum and the base medium, whereas as a SP treatment, four-week-old M. sinensis seedling was transplanted in an autoclaved NC inoculum (the living plant was removed prior to autoclaving) in the mesh bag, and the pot was filled with a mixture of the intact SP inoculum and base medium. Then M. sinensis seeds (Snow Brand Seed Co., Ltd., Sapporo) were sown outside the mesh bag as test plants, covered with a thin layer of autoclaved sand, and grown in a greenhouse either under natural light or under a plastic net that reduces natural light by 80% at day/ night temperatures of 26/ 20 ̊C (n = 3). The seedlings were grown with tap water for the first month and with liquid fertilizer made of Peters Professional 25-5-20 (ICL, St. Louis, MO) at 50 µM phosphate once a week thereafter.
After two months, the roots of the test plants were collected, washed with pressured tap water, cut into 1-cm segments, randomized in water, blotted on a paper towel, immersed in RNAlater (Thermo Fisher Scientific, Tokyo) for more than 48 h at room temperature to fix nucleic acid, blotted on a paper towel to remove excess RNAlater, transferred to a 3-mL tube with an O-ring sealed cap (Yasui Kikai, Osaka), and stored at -80°C until DNA extraction.
Molecular identification and strategy assignment
The frozen roots in the 3-mL tube were ground with a metal cone in the presence of liquid nitrogen at 2,500 rpm for 2 × 5 s using Multi-Beads Shocker (Yasui Kikai, Osaka), and DNA was extracted and purified from approx. 100 mg of ground sample by Maxwell RSC Instrument (Promega, Madison, WI) using Maxwell RSC PureFood GMO and Authentication Kit (Promega) according to the manufacturer's instructions, stored at -30˚C, and used as template for PCR amplification. The divergent domain 2 of large-subunit ribosomal RNA gene (LSU rDNA) was amplified in a 25-µL reaction mixture of Expand High-Fidelity PCR System (Roche Diagnostics, Tokyo), 0.5 nmol µL− 1 each of FLd3 (forward) and FLR2 (reverse) primers that were linked to TruSeq-type forward- and reverse-adapter sequences (Illumina, Tokyo), respectively, at the 5'-end (Niwa et al. 2018), and 0.2–2 µL template DNA using C1000 TouchTM Thermal Cycler (BIO-RAD, Tokyo) with the following program: initial denaturation at 94˚C for 2 min, followed by 30 cycles of denaturation at 94˚C for 15 s, annealing at 48˚C for 40 s, polymerization at 72˚C for 1 min, and final elongation at 72˚C for 10 min. The PCR products were sequenced on the Illumina MiSeq platform (2 × 300 bp), and high-quality paired end reads (read 1 and read 2) were merged with a minimum overlap length of 10 nt using FLASH (http://ccb.jhu.edu/software/FLASH/) at Bioengineering Lab (Sagamihara, Kanagawa, Japan). The merged reads were subjected to blastn searches against the fungal LSU rDNA database consisted of 82,769 operational taxonomic units (OTUs) of fungi, including 687 OTUs of AM fungi (http://amfungi.kazusa.or.jp/) and assigned to the OTUs at ≥ 95% similarities over 250-bp alignment with an E-value cut-off of 1e-30. Details of the phylogenetic positions of the 687 OTUs are indicated in the maximum likelihood tree of am03 release (ML-tree of am03) at http://amfungi.kazusa.or.jp/.
The life-history strategies of the AM fungal OTUs detected in the main experiment were assigned with reference to the supporting experiment and previous studies as follows. All the OTUs that occurred in the SP treatment in the supporting experiment, those that were categorized as SP or SP/RD fungi in Cahyaningtyas and Ezawa (2024), and those that occurred in the trap cultures in Kawahara and Ezawa (2013) were assigned to SP strategists, irrespective of the occurrence in the NC treatment. Only the OTUs that occurred "uniquely" in the NC treatment and those that were categorized as SL fungi in Cahyaningtyas and Ezawa (2024) were assigned to NC strategists. The OTUs that occurred only in the main experiment, but neither in the supporting experiment nor in the previous studies were left unassigned. The sequences obtained by Cahyaningtyas and Ezawa (2024) and Kawahara and Ezawa (2013) were subjected to blastn searches and re-assigned to the latest AM fungal OTUs in the recently renewed database (http://amfungi.kazusa.or.jp/) by applying the same criteria prior to the strategy assignment.
Statistical analysis
All read count data were transformed to presence-absence data prior to analysis, and numbers of the sample in which the OTUs occurred were designated as abundance data for abundance-based analyses. Welch's t-test and two-way ANOVA followed by Tukey's post hoc test were performed on the R 4.3.1 platform (R Core Team 2023). Scatter plots were drawn with the dabestr package on the R platform (Ho et al. 2019). Similarity-difference-replacement (SDR) simplex analysis for abundance data was performed to explore OTU distribution patterns between the light and shaded conditions using SDR-abunSimplex program (Podani et al. 2013) (http://podani.web.elte.hu/SYN2000.html). Light-shade preference index of the M. sinensis seedlings (test plants) for the OTUs were defined by the following equation:
Light-shade preference index = OCshade - OClight
where OCshade and OClight represent the occurrence of OTU, that is, the numbers of samples in which the OTU occurred, under the shaded and light conditions, respectively. Positive and negative values of the index imply that the plants associate with the OTU more preferentially under the shaded and light conditions, respectively.