2.1 Greenhouse experimental design
We conducted two greenhouse experiments for testing our hypothesis and the design of the two experiments were described as below (Fig. 1). Both greenhouse experiments were conducted at China Agricultural University, Beijing, China (40°1′N, 116°16′E). The soil used in this experiment was collected from Shangzhuang Experimental Station (40°08′N, 116°10′E). The physicochemical properties of the soil were pH 8.2 (soil: water = 1:5), organic carbon 11.5 g kg-1, Olsen-P 2.6 mg kg-1, total N 0.72 g kg-1, and exchangeable K 32.3 mg kg-1. The air-dried soil was sieved through a 2 mm sieve, mixed with river sand (w (soil): w (sand) =2:1) and sterilized by γ-ray (25 KGray) as growth substrate.
2.1.1 Greenhouse experiment Ⅰ:testing the interactions of AMF and soil bacterial suspensions on maize growth
A two-compartment rhizobox with a 30 μm nylon mesh in the middle was used, and each compartment was filled with 0.5 kg of the growth substrate. Maize was sown in two compartments as monoculture or as intercropping with faba bean. In order to test the contribution of the neighbor on the focal crop maize, we inoculated AMF in the neighbor side and defined maize in monoculture or faba bean in intercropping as the donor plant. For the maize as the receiver plant either in monoculture and intercropping, bacterial suspensions made from either LP and HP soils (Method S1) were applied and sterilize soil with equal amount of sterile water was used as control. This design was to explore whether the hyphae initiated from the donor species can reach the receiver maize and form common mycorrhizal networks (CMNs) and work together with PSB to promote P uptake in the receiver maize. In the treatments with AMF, 20 g of inoculum was laid in the middle layer of soil in the compartment. In the non-AMF treatments, the same amount of autoclaved inoculum (120℃, 2 hours) was added. The preparation of AMF inoculum was showed in supplementary file (Method S2). In total, there were 8 treatments with 4 replications in the experiment (Fig. 1a) and the plants were harvested 42 days after sowing.
The seeds of maize and faba bean were surface sterilized in 10% (v/v) hydrogen peroxide for 30 min and rinsed with deionized water at least three times. Two germinated seeds (maize or faba bean) were sown in each compartment and were thinned to one seeding per compartment after emergence. All nutrients that need to be applied in the experiment except P were used before planting and mixed into the soil. The remaining nutrients used were (mg kg-1): 200 N as Ca(NO3)2·4H2O, 113 K as K2SO4, 43 Mg as MgSO4·7H2O, 5.9 Fe as Fe-EDTA, 6.7 Mn as MnSO4·H2O, 10 Zn as ZnSO4·7H2O, 2 Cu as CuSO4·5H2O, 0.67 B as H3BO3, 0.17 Mo as Na2MoO4·5H2O.
2.1.2 Greenhouse experiment Ⅱ: testing the effect of different P sources on maize growth
The experiment was to test which PSB are more important in affecting the focal maize P acquisition and growth. The experimental device was similar to experiment I (Fig. 1b). All donor maize and faba bean were inoculated with 20 g AMF inoculum while the receiver maize was applied with the combination of organic and inorganic PSB. The method to isolate PSB was described in supplementary file (Method S3). In addition, two kinds of exogenous insoluble P sources were put into nylon mesh bags with a pore size of 30 μm and then placed in the middle layer of the soil in the compartment planting with receiver maize. Briefly, Ca(H2PO4)2·H2O (Ca-P, 100 mg P kg-1soil) and phytate-P (100 mg P kg-1soil) were used as the P sources in the inorganic and organic PSB treatments, respectively. When treatments with both PSB were added, phytate-P (50 mg P kg-1soil) and Ca-P (50 mg P kg-1soil) were added together as P sources. Finally, there were 6 treatments with 4 replications in experiment Ⅱ (Fig. 1b). The two-compartment rhizobox used and the amount of basic nutrient solutions added in experiment Ⅱ were the same as that of experiment Ⅰ. The plants were grown for 42 days before harvesting.
2.5 Plant harvest and sampling
For greenhouse experiments Ⅰ and Ⅱ, shoot of maize and faba bean were harvested 0.5 cm above the soil layer, and dried at 105℃ for 30 min and then at 70℃ till the sample weight reached a constant value. Then the shoot biomass and P content of maize and faba bean were determined in each compartment using dry ashing method (Method S4).
2.6 DNA extraction and metagenomic sequencing
Soil samples from LP and HP treatments (approx. 0.5 g) were used to extract DNA using the E.Z.N.A.® Soil DNA Kit (Omega Bio-Tek, Norcross, GA, U.S.) following the manufacturer's instructions (Method S5). Samples of total DNA extracted from LP and HP soils were used for the preparation of metagenomics libraries. DNA extracts were fragmented to an average size of about 400 bp using Covaris M220 (Gene Company Limited, China) for paired-end library construction. The NEBNext® Ultra™ DNA Library Prep Kit for Illumina (NEB, USA) was used to generate sequencing libraries following the manufacturer's instructions. Paired-end sequencing was performed on Illumina NovaSeq (Illumina Inc., San Diego, CA, USA) at Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China) using NovaSeq Reagent Kits following the manufacturer's instructions (www.illumina.com).
The raw reads from metagenome sequencing were used to generate clean reads by removing adaptor sequences (Lang et al. 2019). Trimming and removing low-quality reads (reads with N bases, a minimum length threshold of 50 bp, and a minimum quality threshold of 20) were done by fastq (https://github.com/OpenGene/fastp, version 0.20.0) in the free online platform of Majorbio Cloud Platform (cloud.majorbio.com). The high-quality reads were then assembled to contigs using MEGAHIT (parameters: kmer_min = 47, kmer_max = 97, step = 10) (https://github.com/voutcn/megahit, version 1.1.2), which makes use of succinct de Bruijn graphs (Li et al., 2015). Protein-encoding genes were queried in Prokka and BLAST against the NCBI non-redundant (NR) protein database using Diamond with default settings (Lang et al. 2019). Genes involved in inorganic P solubilization and organic P mineralization were queried by KEGG database (Lang et al. 2019).
2.8 Statistical analysis
The data were checked for homogeneity of variances with Levene’s test and normality with Shapiro-Wilk test before ANOVAs. One-way ANOVA was conducted to show the difference in shoot biomass and P content of donor plants (maize, faba bean) in greenhouse experiment Ⅰ and Ⅱ. Then two-way ANOVA was conducted to show interactions among cropping systems, AMF and bacterial suspension on shoot biomass and P content of receiver maize in greenhouse experiment Ⅰ. The results showed that AMF inoculation did not affect shoot biomass and P content of receiver maize (Table S5), so the treatments with AMF inoculation were merged and then a second two-way ANOVA was conducted to test the interactions between cropping system and PSB on shoot biomass and P content of receiver maize in greenhouse experiment Ⅰ and Ⅱ.
Paired sample t-tests were conducted to examine the difference between the expression of P-cycling genes in HP and LP soils and were also conducted to examine the effect of adding bacteria on shoot biomass and P content of receiver maize in greenhouse experiment Ⅰ. In addition, Paired sample t-tests were conducted to examine the difference in shoot biomass and P content of receiver maize applied with different PSB in greenhouse experiment Ⅱ. Then Tukey’s test was conducted to show the significant difference among treatments (P < 0.05). Sigmaplot 12.0 was used to draw figures.
All the bioinformation analyses were conducted with Majorbio Cloud platform (https://cloud.majorbio.com/). Non-metric multidimensional scale (NMDS) analysis was performed to visualize the difference between microbial communities and functional gene composition in HP and LP soils using Bray-Curtis distances. Wilcoxon rank sum test was performed in linear discriminant effect size analysis (LDA) to show the distinct microbial communities and genes in LP and HP soils and then genes with significant differences were sorted according to LDA score. In addition, the relative abundance of bacteria taxa contributed to the expression of P cycling-related genes was also examined in LP and HP soils.