Soil sampling and amplicon analysis of soil fungi
36 soil cores, one from each treatment combination (i.e., one column,
consisting of three 40×40 cm quadrats; see Table 1b), were collected in
June 2018 (near the end of the three-year experiment) for analyzing the
fungal composition. Soil samples were collected from the interstices
between the transplanted seedlings in each column (bulk soils instead of
rhizosphere soils were collected because it was not feasible to trace
the roots of the seedlings that died in the first two years). After
removal of loose surface debris, surface soil samples (0–10 cm in
depth) were collected using a soil corer with a 5-cm inner diameter. The
soil corer was rinsed thoroughly using double distilled water after each
sampling.
The collected soil samples were immediately frozen and shipped on dry
ice to Magigene Technology Ltd. (Guangzhou, China) laboratory and were
stored at -80 °C prior to DNA extraction for Illumina amplicon
sequencing. The sequencing steps are described as follows:
- Fungal DNA was extracted from 1 g of soil sample with the
PowerSoil® DNA isolation kit (MO BIO Laboratories
Incorporation, Carlsbad, USA).
- The internal transcribed spacer (ITS) region 2 (Taylor et al.2016) was selected to determine fungal communities using ITS3 (5′-
GCATCGATGAAGAACGCAGC -3′) and ITS4 (5′- TCCTCCGCTTATTGATATGC -3′)
primers on an Illumina MiSeq/Hiseq 2500 platform, with full
consideration of inaccuracies in reading numbers by Illumina to
estimate fungal abundances, e.g., those caused by variable rDNA copy
number and PCR biases (Taylor et al. 2016); 250 bp paired-end
reads were generated.
- Before amplicon sequencing, concentration and purity of extracted DNA
samples were measured using NanoDrop One (Thermo Fisher Scientific,
Waltham, MA, USA).
- The downstream processes included quality control of paired-end raw
reads according to Trimmomatic (Bolger et al. 2014), assembly
of paired-end clean reads using FLASH (Magoč & Salzberg 2011), and
quality control of raw tags using Mothur software (Schloss et
al. 2009).
- Using USEARCH (Edgar 2010), a 97% threshold of sequence similarity,
which is effective in analyzing microbial community, was selected to
categorize the effective sequences into unique operational taxonomic
units (OTUs) and do taxonomic annotation of OTUs (Koljalg et
al. 2005).
In total, 11407 OTUs (Data S1) were observed. Among these OTUs, 88.73%
(9802) belonged to fungi (Data S2), while the remaining OTUs belonged to
kingdom Animalia (0.05%), Chromista (2.10%), Plantae (1.92%), and
Protista (0.71%), or showed no blast hit (5.51%). According to the
previously published criteria for grouping functional guilds based on
the FUNGuild database (Nguyen et al. 2016; Vetrovsky et
al. 2019; Delgado-Baquerizo et al. 2020), the fungal OTUs
identified to the genus level were further grouped as saprotrophs, plant
pathogens, EcM fungi, etc. An OTU identified to the genus level was
assigned to a guild if it could unequivocally match a guild, otherwise
its trophic strategy was considered unknown (Nguyen et al. 2016;
Hannula et al. 2017; Mommer et al. 2018; Vetrovskyet al. 2019). During analyses of the response of pathogenic fungi
and EcM fungi to experimental treatments, we only chose the fungal OTUs
as plant-pathogenic fungi or EcM fungi with “probable” or “highly
probable” confidence levels (Data S3).
Because the ITS primers used in this study were invalid to classify
oomycete pathogens (belonging to Chromista), Chrmista OTUs were excluded
from the community analysis. Even though oomycete-specific primers are
currently available, they can merely be used to qualitatively analyze
oomycete community and diversity monitoring (Legeay et al. 2019),
infeasible to quantitatively assess relative abundance of oomycete
pathogens.