Effects of continuous and rotational cropping practices on soil fungal communities in pineapple cultivation

Collection of soil samples and measurement of the physicochemical properties

Samples were obtained from inter-row soil (i.e., soil between planting rows) of pineapple (variety: Comte de Paris) crops in Leizhou and Xuwen (Guangdong, China), according to our previous study (Chen, Gong & Xu, 2020). For continuous cropping, the pineapple was planted in succession for at least 18 years. For rotational cropping in Leizhou, three of the five farmlands used pineapple–banana rotation cropping, while the other two used pineapple–capsicum cropping. The time scale of pineapple–banana crop rotation was 16 months for pineapple and 12 months for banana, while that of pineapple–capsicum was 16 months for pineapple and 5 months for capsicum. For rotational cropping in Xuwen, all the five pineapple rotational farmlands involved pineapple–sugarcane cropping. The time scale of pineapple–sugarcane crop rotation was 16 months for pineapple and 12 months for sugarcane. Between each rotation, the stubble of previous plants was destroyed by tractor, and was retted in soil for 3 months. Subsequently, the soil was ploughed by tractor and retted before raking. The next crop was then planted at the farmlands. During cropping, the same fertilization/soil management strategy was applied to both continuous and rotational cropping. Briefly, the foliation fertilizer was firstly applied at about 50 days after the planting of pineapple seedlings, and then applied every other month. Further, the chemical fertilizer (375 kg urea, 300 kg potassium chloride, and 300 kg compound fertilizer per hectare) was applied between rows in every spring. Notably, the selection of rotational cropping systems was in accordance with the prevalence of different rotational systems in these two locations. In total, 80 soil samples were collected in a random zig-zag manner at 0–15 cm depth, categorized into eight groups (continuous cropping soils of Leizhou in summer (LCS) and winter (LCW), rotational cropping soils of Leizhou in summer (LRS) and winter (LRW), continuous cropping soils of Xuwen in summer (XCS) and winter (XCW), and rotational cropping soils of Xuwen in summer (XRS) and winter (XRW)). Each group contained 10 samples from five farmlands (two samples in each farmland) and each sample were a mixed of ten subsamples (Table S1). The physicochemical parameters (pH, total organic matter (TOC), total nitrogen (TN), available nitrogen (N), phosphorus (P), and potassium (K)) of the soil were measured as previously reported (Chen, Gong & Xu, 2020).

DNA extraction and polymerase chain reaction (PCR)

We extracted of DNA from 0.5 g samples of soil according to the manufacturer’s instructions of Power Soil DNA kit (Qiagen, Hilden, Germany). And then we purified the combined duplicate DNA extractions using a DNA gel extraction kit (Axygen, Union City, CA, USA). A BioTek Epoch Microplate Spectrophotometer (Winooski, VT, USA) was used to check the DNA concentration of each sample (Chen, Gong & Xu, 2020).

The extracted DNA samples were used as templates for PCR amplification targeting the internal transcribed spacer 1(ITS1) region of the fungus. The ITS region was targeted using the primers ITS1F (5-CTTGGTCATTTAGAGGAAGTAA-3) and ITS2 (5-GCTGCGTTCTTCATCGATGC-3) (Orgiazzi et al., 2012). The reaction volume for PCR analysis was 25 μL, comprising 10 μL 2×Taq Master Mix, 1 μL template DNA, 0.5 μL forward/reverse primer, respectively. The amplification procedure was as follows: 95 °C for 5 min; 95 °C for 30 s, 58 °C, 30 s and 72 °C, 30 s for 42 cycles. PCR products were stored at −20 °C immediately after the end of the reaction. Using the Illumina HiSeq 2500 platform (Illumina, San Diego, CA, USA) at Biomarker Technologies Co, Ltd. (Beijing, China), the PCR products were used to construct the sequencing libraries according to the standard protocol.

Raw data processing

The sequencing reads of all samples were rarefied to read number of the sample with lowest sequencing reads by a custom script. Then the rarefied paired-end data was merged into a sequence of tags based on the overlap relationship, and the quality of the reads and the merged effect were quality-controlled and filtered. FLASH v1.2.7 software was used to stitch the reads of each sample through overlap and obtain the original tags (Magoč & Salzberg, 2011), and Trimmomatic v0.33 software was used to filter the spliced raw tags to obtain clean tags (Bolger, Marc & Bjoern, 2014). Finally, effective tags were obtained by removing the chimera sequence using UCHIME v4.2 software (Edgar et al., 2011).

Taxonomic analysis

Operational taxonomic units (OTUs) were clustered using UCLUST in QIIME (version 1.8.0) software at a similarity level of 97% (Caporaso et al., 2010). Taxonomic annotation of OTUs was performed based on the UNITE (fungal) taxonomic database (version 8.0) (Nilsson et al., 2018). The Venn graph shows the number of common and unique OTUs between samples and visually shows the overlap of OTUs between samples. Combined with the species represented by OTUs, it is possible to find common microorganisms in different environments. The Venn diagrams for each classification level were drawn using the R software VennDiagram (Chen & Boutros, 2011).

Fungal diversity analysis

Alpha diversity reflects the richness and diversity of a single sample with four measurement indicators, including the Shannon index, chao 1 index, observed species index, and PD whole tree index. The Chao1 index was used to evaluate species abundance, while the Shannon index was used to evaluate species diversity and was influenced by the abundance of species in the sample community and community evenness. The alpha diversity index of the samples was evaluated using Mothur (version v. 1.30) software (Schloss et al., 2009) and compared among different groups with one-way ANOVA followed by Duncan’s multiple range test. Beta diversity analysis based on OTU relative abundance was carried out to examine the differences in microbial community structure among different samples by calculating the ecological distance of Bray Curtis, Unifrac, etc, and pairwise PERMANOVA was performed to estimate the statistical significance between different groups. The QIIME2 software package (https://docs.qiime2.org/2022.8/) was used to calculate the beta diversity distance, and the module “cmdscale” in R software was used to perform principal coordinate analysis (PCoA).

Statistical analysis

The difference in relative abundance of fungi among different groups was calculated with one-way ANOVA followed by Duncan’s multiple range test. The correlation between the fungal community and environmental parameters in winter was analyzed by redundancy analysis (RDA) using Canoco 5.0 (Ter Braak & Smilauer, 2012). Significant differences between groups were mainly used to find biomarkers among different groups by performing linear discriminant analysis (LDA) effect size (LEfse) based on the normalized abundance table of each species level (Segata et al., 2011). The biomarker screening criteria was LDA score >4.

Overview of the sequencing data

We performed ITS sequencing to investigate the fungal communities in the continuous and rotational cropping soil of pineapple cultivation. The sequencing produced an average of 67,932 (ranging from 49,838 to 72,523) effective tags for each sample (Table S2). All the effective tags were clustered into 12,346 OTUs assigned to the taxonomy of the fungi. The average OTU number for all detected samples was 1,329 (ranging from 677 to 2,913). The number of common fungal OTUs shared between the continuous and rotational cropping soil of pineapple in summer and winter was 679 (Fig. 1). The unique OTUs in LRS, LCS, XRS, XCS, LRW, LCW, XRW, and XCW were 1,433, 203, 405, 98, 495, 454, 361, and 391, respectively. Consistent with the results of fungal community composition in the continuous and rotational cropping soil of pineapple, these results demonstrated that the number of OTUs was significantly increased in the rotational cropping soil of pineapple in summer. Taken together, these results demonstrate that rotational cropping caused structural and diverse changes in the planting soil of pineapple.

Diversity analysis of fungi in the continuous and rotational cropping soil of pineapple

In the present study, the Shannon index, Chao1 index, observed species index, and PD whole tree index were used to evaluate species diversity and richness (Table S3). These indices indicated that the richness and diversity of fungi was significantly higher in the rotational cropping soil of pineapple than in the continuous cropping soil (Duncan’s multiple range test, P < 0.05). However, there was no significant difference between pineapple–banana (LRS01, LRS02, LRS03, LRS06, LRS07, and LRS08) and pineapple–capsicum (LRS04, LRS05, LRS09, and LRS10) rotation cropping farms at Leizhou (Table S3). In addition, rotational cropping and continuous cropping soil in winter showed no significant difference (Figs. 2A–2D). These results demonstrated that both species richness and species diversity were mostly affected by the cropping modes (continuous or rotational cropping).

To confirm the differences in the fungal communities between continuous cropping and rotational cropping in winter or summer, principal coordinate analysis (PCoA) based on the Bray–Curtis dissimilarity index was carried out. The results showed that the two main coordinates extracted explained 34% of the variation. Further analysis showed that the distribution of the samples was also significantly different between summer and winter and between continuous and rotational cropping (PERMANOVA, R² = 0.46, P ≤ 0.001; Fig. 3, Table S4), only except for the distance between continuous cropping and rotational cropping samples of Leizhou in the winter. These results demonstrate that a significant difference exists between continuous cropping and rotational cropping.

Community composition of fungi in the continuous and rotational cropping soil of pineapple

The structure and diversity analysis of fungal community showed that fungal diversity significantly increased in continuous cropping soil. At the phylum level, a total of 11 groups were detected at >1% abundance (Table S4). Ascomycota (43.34–90.93%), Basidiomycota (1.98–22.97%), Zygomcota (0.44–6.96%), and Chytridiomycota (0.19–5.17%) were the dominant phyla in the continuous and rotational cropping soil of pineapple both in the summer and winter. Among them, Ascomycota was the most abundant, accounting for more than 50%, which dramatically increased in the continuous cropping soil in summer (one-way ANOVA, P < 0.01). However, in the winter, no obvious changes were observed for Ascomycota in the continuous and rotational cropping soil of pineapple (Fig. 4A). In addition, 26 groups were detected at >1% abundance for fungi at the genus level (Table S5), and the dominant genera of the continuous and rotational cropping soil of pineapple both in the summer and winter were Chaetomium, Penicillium, Fusarium, Trichoderma, and Cryptococcus (Fig. 4B). Further analysis showed that Chaetomium richness was significantly increased in continuous cropping soil in both the summer and winter (one-way ANOVA, P < 0.01). However, it was dramatically lower in the winter than in the summer (Fig. 4B). In addition, our analysis showed that Chytridiomycota at the phylum level and Gibberella at the genus level increased after rotational cropping both in the winter and summer (Fig. 4).

Relationship between physicochemical parameters and fungal community

The correlation between the structure of the fungal community and physicochemical parameters in the continuous and rotational cropping soils of pineapple in winter was investigated using redundancy analysis. In continuous cropping soil, RDA1 and RDA2 accounted for 85.75% and 7.12% of the changes in the structure of fungal community, respectively (Fig. 5A, Table S6), indicating that the detected physicochemical parameters accounted for most of the structural changes in the fungal community. Among the significantly correlated parameters, pH showed the greatest influence on fungal community in continuous cropping soil, followed by total organic matter (TOC), total nitrogen (TN), and available phosphorus (P). Available nitrogen (N) and potassium (K) had no significant correlation with fungal community in continuous cropping soil. In rotational cropping soil, RDA1 and RDA2 contributed to 32.41% and 22.26% of the variations in the structure of fungal community, respectively (Fig. 5B, Table S7). Moreover, TOC and pH significantly influenced the fungal community, while the other parameters did not significantly correlate with the fungal community in rotational cropping soil.

Fungal biomarkers in the continuous and rotational cropping soil of pineapple

We have demonstrated the effects of rotational cropping practices on the structure and diversity of soil fungal community in summer and winter in Xuwen and Leizhou. LEfSe algorithms were used to explore the fungal biomarkers in the planting soil of pineapple. The analysis showed that eight fungal groups had significant differential distribution among the XCS, XCW, XRS, and XRW groups, including Sporobolomyces and Aspergillus at the genus level and Penicillium and fungal sp p1s11 at the species level in rotational cropping soil at Xuwen in winter; Corynascus sp. JHG 2007 at the species level and Corynascus at the genus level in rotational cropping soil at Xuwen in the summer; Sporidiobolales family Incertae sedis in continuous cropping soil at Xuwen in the winter; and Sordariomycetes at the class level in continuous cropping soil at Xuwen in the summer (Figs. 6A and 6B). However, no significantly differentially distributed fungal groups were found in Leizhou groups. The distribution of the eight fungal groups in the two pineapple cropping practices in summer and winter are shown in Figs. 6C–6J. Notably, the biomarkers of XRW (Sporobolomyces, Aspergillus, Penicillium, and fungal sp. p1s11) and XCS (Sordariomycetes) groups also showed similar trend in LRW and LCS groups (one-way ANOVA and Duncan’s multiple range test, P < 0.05) (Figs. 6C–6F and 6J), whereas the biomarkers of XRS group were more abundant in LCS group rather than in LRS group (one-way ANOVA and Duncan’s multiple range test, P < 0.05) (Figs. 6G and 6H).