The greenhouse and field experiments were executed in the greenhouse of “State Key Laboratory for Conservation and Utilization of Bio-resources in Yunnan,” Yunnan Agricultural University and Xiadabai County, Kunming (25° 2′ N and 102° 42′ E), China. The field is severely infected with clubroot pathogen, and Chinese cabbage has been grown in the field for >8 years. The annual rainfall and annual average temperature at Xiadabai County were recorded at 1,534 mm and 15.1°C, respectively. Chinese cabbage variety Luchunbai No. 1, highly susceptible to P. brassicae, was used as plant material.

The greenhouse experiment was performed from March to May 2021 under controlled conditions: 30/18°C day/night temperature and 10/14 h light/dark photoperiod (Zhang et al., 2022b). The pots (24 × 18 cm diameter) were filled with 3.5 kg/pot of diseased soil heavily infected with clubroot pathogen collected from the field (Xiadabai County, Kunming, China). Lime and ammonium bicarbonate were added to each pot at the ratio 1:1,300 (m/m), mixed thoroughly with the diseased soil, covered with plastic film (0.08 mm), and placed in a greenhouse at >35°C. The experiment was performed under four different conditions: NB; soil fumigation with lime (2.69 g/pot), N; soil fumigation with ammonium bicarbonate (2.69 g/pot), LNB; soil fumigation with combined ammonium bicarbonate (2.69 g/pot) and lime (2.69 g/pot), and GZ; no soil fumigation as control. After 15 days of anaerobic soil fumigation, Chinese cabbage seedlings were transplanted in pots (5 plants/pots). In addition, fertilizer was applied in each pot before the transplantation of seedlings to overcome the nutrient deficiency (Wei et al., 2021). The experiment was repeated three times with 15 plants per replication (in total, 45 plants/treatment) and was accomplished using a completely randomized design.

The control effect (CE), disease index (DI), and disease incidence (Di) were assessed at the end of the experiment, while soil pH was measured after soil fumigation and the end of the experiment. The Di, DI, and CE were calculated according to a five-point disease scoring scale and formulas (Zhang et al., 2022b). The soil pH was calculated in a 2.5:1 water/soil (V/W) suspension using a pH meter (Zhang et al., 2022a).

Field experiment was conducted during the growing season from June to August 2021 under four different situations: NB; soil fumigation with lime (0.15 kg/m²), N; soil fumigation with ammonium bicarbonate (0.15 kg/m²), LNB; soil fumigation with combined lime (0.15 kg/m²) and ammonium bicarbonate (0.15 kg/m²), and GZ; no soil fumigation as control. After soil amendments with ammonium bicarbonate and lime, the field was divided into plots (3 m × 9 m) and covered with 0.08 mm film for 15 days for anaerobic soil fumigation. The experiment was performed using a randomized complete block design and repeated three times with 3 plots/treatment (in total 12 plots), and each plot contained 120 plants of Chinese cabbage. The Di, DI, CE, and soil pH from each treatment were recorded at the completion of the trial as described above using the methodology of Zhang et al. (2022a).

Soil samples were collected in replicates (minimum three biological replicates/treatment) from each treatment from greenhouse and field experiments. Briefly, 5 plants per replication from each treatment were uprooted, and the excess soil was removed by shaking plants (Ahmed et al., 2022). Soil adjacent to the root surface was collected as rhizosphere soil samples for qPCR amplification and rhizosphere microbial diversity analysis. Briefly, five cores of soil samples from each replication were thoroughly mixed to make one composite sample, and three samples were collected from each treatment. Following the manufacturer’s instructions, 0.5 g of soil per sample was used to extract soil DNA using a PowerSoil^® DNA isolation Kit. In order to minimize the error, three cores of DNA were extracted from each sample, mixed thoroughly to make one sample, and kept at −80°C.

Quantitative PCR (qPCR) was amplified using the specific primer PB05−F (GAACCGGTCCACACACACACACACTAGAT) and PB05−R (GCCCACTGTGTTAATGATCC) to investigate P. brassicae population dynamics in the rhizosphere of Chinese cabbage plants. For qPCR amplification, each 20 μL reaction mixture contained 10 μL of 2× SuperReal PreMix Plus (SYBR Green I), 1.0 μL of PB05F/PB05R primers, 2 μL of DNA, and 7 μL of sterilized distilled water. The qPCR reaction was carried out in iCycler iQ5 (Bio-Rad, California, United States) thermal cycler under the following conditions: initial denaturation for 3 min at 95°C, with 40 cycles of denaturation for 10 s at 95°C, annealing for 30 s at 56°C, and extension for 20 s at 72°C.

PCR was performed for V3-V4 and ITS2 variable regions of 16S rRNA and ITS genes of bacteria and fungi with primer pairs 341F/806R and ITS5/ITS2, respectively, for analysis of rhizosphere microbial diversity (Ahmed et al., 2022, 2024). The obtained PCR products were purified using the AxyPrepDNAGel Extraction Kit (AXYGEN, United States) and sequenced on an Illumina MiSeq platform at Meggenomics Biotech Co., Ltd. (Guangdong, China) for the construction of paired-end reads.

The paired-end reads collected from the Illumina MiSeq sequencer were initially processed through FLASH v1.2.11 to remove sequences (Magoč and Salzberg, 2011) and trimmed with Trimmomatic v0.36 at a score < 20 for quality control (Bolger et al., 2014). UCHIME v8.1 was used to remove the chimeras and for the production of high-quality reads (Edgar et al., 2011). Using the UPARSE pipeline, the reads were clustered into OTUs (operational taxonomic units) with a 97% similarity score (Edgar, 2013). The taxonomic annotation of fungal and bacterial OTUs was performed using the ribosomal database project (RDP) classifier v2.2 based on UNITE (Kõljalg et al., 2005) and SLIVA (Quast et al., 2012) databases, respectively.

QIIME 2 was used to estimate the alpha and beta diversity metrics of bacterial and fungal communities (Bolyen et al., 2019). The beta diversity of bacterial and fungal communities was visualized using Principal Coordinates Analysis (PCoA) based on Weighted and Unweighted UniFrac distance metrics (Zhang et al., 2020). The differences in the composition of microbial communities were assessed by PERMANOVA using “Adonis” in R v4.2.0 (Anderson and Walsh, 2013). The relative abundance (RA) bar plots at phylum and genus levels and RA abundance heatmap at OTUs level for both bacterial and fungal communities were generated using R scripts in R v4.2.0. A co-occurrence network analysis was performed among the top 20 bacterial and fungal OTUs using the “sparcc package” and visualized by “psych package” in R (v.4.2.0) according to “Spearman correlation coefficient > 0.6 and p < 0.05.” Data were statistically examined using analysis of variance (ANOVA), and the least significant difference (LSD; p < 0.05) was used to assess the significant differences among treatments.

Greenhouse and field experiments were performed to assess the effect of soil fumigation on the occurrence of Chinese cabbage clubroot disease (Figure 1; Supplementary Tables S1, S2). At the end of each experiment, the Di%, DI%, and CE% were calculated from each treatment using a disease rating scale and observation of gall formation on the roots. Results of greenhouse experiment demonstrated that soil fumigation with lime (NB) and ammonium bicarbonate (N) significantly reduced the incidence of clubroot of Chinese cabbage, having Di (62.96 and 51.85%), DI (35.80 and 20.57%), and CE (33.51 and 62.10%), respectively compared to control (GZ) Di (85.19%), DI (53.88%), and CE (0%) (Figures 1A–C). However, minimum Di (33.33%) and DI (13.58%), and highest CE (74.68%) were achieved under the combined application of ammonium bicarbonate and lime (LNB) compared with NB, N, and GZ (Figures 1A–C).

A field experiment was conducted to further examine the potential of soil fumigation to mitigate clubroot incidence in Chinese cabbage under natural conditions (Figures 1D–F). Results of field experiment revealed that soil fumigation with NB and N significantly alleviate clubroot incidence in Chinese cabbage having Di (80.55 and 63.89%), DI (49.69 and 33.64%), and CE (34.48 and 55.45%), respectively than control (GZ) Di (94.45%), DI (75.93%), and CE (0%) (Figures 1D–F). Whereas, combined application of ammonium bicarbonate and lime (LNB) results in minimum values of Di (58.33%) and DI (25.62%) compared with NB, N, and GZ, and the relative control effect of LNB on clubroot of cabbage can be reached up to 66.28% (Figures 1D–F). These results suggested that the application of ammonium bicarbonate (N) and lime (NB) significantly reduced the Chinese cabbage clubroot disease under both greenhouse and field conditions. However, maximum results were obtained under the combined application of ammonium bicarbonate and lime (LNB).

We explored the effect of N, NB, and LNB as soil fumigants on soil pH after fumigation (after 15 days) and at the end of experiments (Figure 2; Supplementary Tables S1, S2). In the greenhouse experiment, after 15 days of soil fumigation with NB, N, LNB, and GZ, the soil pH values were recorded as 7.4, 6.3, 7.3, and 5.8, respectively. The soil pH was significantly increased under the application of NB and LNB and was found to be significantly higher than N and GZ, whereas there is no significance among NB and LNB (Figure 2A). At the end of experiment, after harvesting the Chinese cabbage crop, the values of soil pH under different treatments NB, N, LNB, and GZ were recorded as 6.6, 6.0, 6.4, and 5.9, respectively (Figure 2B). In the field experiment, 15 days after soil fumigation with NB, N, LNB, and GZ, the soil pH values were recorded as 6.7, 5.7, 6.8, and 5.1, respectively (Figure 2C). The values of soil pH were significantly increased under the application of NB, N, and LNB compared to GZ, while there was no significant difference between NB and LNB (Figure 2C). At the end of experiment, the values of soil pH under different treatments NB, N, LNB, and GZ were recorded as 5.7, 5.0, 5.6, and 4.9, respectively (Figure 2D). The results demonstrated that soil fumigation with NB and LNB significantly decreased the soil acidity compared with N and GZ.

We determined the effect of soil fumigation with N, NB, LNB, and GZ on the population dynamics of P. brassicae in the rhizosphere of Chinese cabbage in the greenhouse and field experiments. The presence of P. brassicae was confirmed through the amplification of standard qPCR, and population dynamics (copies per gram of soil) were calculated using a standard curve. Results of qPCR analysis showed that soil fumigation with N, NB, and LNB significantly reduced the gene copy number of P. brassicae per gram of soil compared with control (GZ) in the greenhouse and field experiments (Figure 3). The density of P. brassicae decreased from 7 × 10⁵ to 3 × 10⁴ copies per gram of soil and 1.5 × 10⁶ to 9 × 10⁴ copies per gram of soil under greenhouse and field experiments, respectively (Figures 3A,B). In the greenhouse experiment, the density of P. brassicae decreased in the order CK (7 × 10⁵) > NB (4 × 10⁵) > N (5 × 10⁴) ≥ LNB (3 × 10⁴) (Figure 3A). While under field conditions, the density of P. brassicae decreased in the order CK (1.5 × 10⁶) > NB (5 × 10⁵) > N (1.6 × 10⁵) > LNB (9 × 10⁴) (Figure 3B). These results suggest that the combined application of ammonium bicarbonate and lime (LNB) significantly decreased the density of P. brassicae in the rhizosphere of Chinese cabbage compared to that of NB, N, and CK.

The microbiome assembly of Chinese cabbage changed significantly under four treatments (NB, N, LNB, and GZ). In total, 12 rhizosphere soil samples were subjected through the Illumina MiSeq platform for amplification of V3-V4 and ITS2 regions of bacteria and fungi, respectively, which resulted in a total of 1,513,933 (ave; 126,161/sample) and 1,064,123 (ave; 88,677/sample) raw reads, respectively (Supplementary Table S3). After quality control and chimera’s removal 1,396,930 (ave; 116,411/sample) and 1,003,248 (ave; 83,604/sample) high-quality reads were obtained for bacteria and fungi, respectively which were then clustered into 46,743 (ave; 3,895/sample) bacterial operational taxonomic units (OTUs) and fungal 5,627 (ave; 469/sample) OTUs for taxonomic annotation at 3% cutoff level (Supplementary Table S3).

Principal coordinate analysis (PCoA) based on Weighted UniFrac (abundance of taxa) and Unweighted UniFrac (sensitive to rare taxa) distance metrics was used for the visualization of changes in the structure (beta diversity) of microbial communities (Figure 4). According to PCoA results, the first 2-axis of Weighted UniFrace and Unweighted UniFrace showed a total of 56.20 and 28.20% variation for bacterial communities, respectively (Figures 4A,B) and 68.90 and 40.01% variation for fungal communities, respectively (Figures 4C,D). Further, PERMANOVA of pairwise interaction between treatments for bacterial (R² = 0.666, p = 0.0433) and fungal (R² = 0.732, p = 0.001) communities demonstrated that structure of rhizosphere microbiome significantly changed under different treatments (Supplementary Table S4). However, we cannot observe a significant difference in alpha diversity indices (Chao-1, Shannon, and Simpson) of bacterial and fungal communities under different treatments (Supplementary Table S5).

In addition, an OTUs analysis was performed for the taxonomic annotation of bacterial and fungal recovered OTUs under different treatments (Figure 5). OTUs analysis of different treatments demonstrated that a total of 5,909, 7,971, 6,969, and 6,784 bacterial and 736, 795, 747, and 703 fungal annotated OTUs were obtained from treatments GZ, LNB, N, and NB, respectively (Figures 5A,B). A Venn diagram further showed a variation among the bacterial and fungal annotated OTUs. According to the results of the Venn diagram, a total of 2,189 and 280 common bacterial and fungal OTUs, respectively, and 1991, 3,550, 2,460, and 2,305 bacterial and 155, 161, 124, and 98 fungal unique OTUs were recovered under different treatments GZ, LNB, N, and NB, respectively (Figures 5C,D). These results suggested that the changes in the microbial community structure might be due to the variation in the common and unique OTUs among the treatments.

Relative abundance (RA) analysis at the phylum level showed that bacterial and fungal community composition significantly changed under different treatments. The RA of the top 15 bacterial and fungal phyla in the rhizosphere of Chinese cabbage under different treatments is shown in Figure 6 and Supplementary Tables S6, S7. In all rhizosphere soil samples, Acidobacteria, Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria were the dominant bacterial phyla, accounting for ~89.59% of total RA (Figure 6A; Supplementary Table S6). While Ascomycota, Basidiomycota, Chytridiomycota, Mortierellomycota, Olpidiomycota, and Plasmodiophoromycota were the dominant fungal phyla accounting for ~67.02 of total RA (Figure 6B; Supplementary Table S7). Bacterial phylum Proteobacteria was present in high RA with GZ, and its RA decreased in an order GZ > N ≥ LNB > NB (Figure 6C). The RA of Bacteroidetes was significantly higher in N and NB than LNB and GB (p < 0.05), while Acidobacteria and Chloroflexi were present in high RA in LNB compared with N, NB, and GZ (Figure 6C). Fungal phylum, including Olpidiomycota, Basidiomycota, and Chytridiomycota, were present in higher RA in N and NB than in LNB and GZ (Figure 6D). However, Ascomycota and Plasmodiophoromycota were the dominant fungal phyla in the rhizosphere of GZ compared to N, NB, and LNB. The RA of Ascomycota was decreased in an order GZ > N > NB ≥ LNB. The clubroot pathogen P. brassicae belongs to phylum Plasmodiophoromycota, which was present in 0.69, 0.06, 0.08, and 16.30% average RA in NB, N, LNB, and GZ (Supplementary Table S7), and its RA decreased as GZ > NB > LNB ≥ N (Figure 6D). The results showed that soil fumigation with N, NB, and LNB significantly decreased the RA of Plasmodiophoromycota as compared with non-fumigated soil (GZ).

The patterns of taxonomic distribution of bacterial and fungal communities at the genus level become more evident. Results demonstrated that bacterial and fungal community composition significantly changed at the genus level under different treatments. The RA abundance of the top 15 bacterial and fungal genera is shown in Supplementary Tables S8, S9. Flavobacterium was the most dominant bacterial genus in the rhizosphere of Chinese cabbage under different treatments, having an average RA of ~8.23%, and its RA decreased in an order NB > N > GZ ≥ LNB (Figure 7A; Supplementary Table S8). The RA of Sphingobacterium and Pseudomonas significantly increased in the rhizosphere of GZ compared to that of N, NB, and LNB. In contrast, Sphingomonas was found in higher RA in N, NB, and LNB than in GZ, and Bacillus was found in high RA with GZ, NB, and LNB compared with N (Figure 7A). Olpidium was the most dominant fungal genera with an average RA of about ~32.74% in the rhizosphere of Chinese cabbage under different treatments, and its RA decreased in an order NB ≥ N > LNB ≥ GZ (Figure 7B; Supplementary Table S9). Fusarium was dominant in NB and GZ more than N and LNB, while Mortierella was dominant in NB more than N, LNB, and GZ (Figure 7B). The RA of plant pathogenic genera such as Cylindrocarpon and Plasmodiophora significantly decreased after soil fumigation with N, NB, and LNB as compared to non-fumigation soil (GZ). The RA of genus Cylindrocarpon lowered in an order GZ > N, NB, and LNB, while the RA of Plasmodiophora was reduced in an order GZ > NB > N and LNB (Figure 7B). These results suggested that soil fumigation with lime and ammonium bicarbonate significantly reduced the population dynamics of the genus Plasmodiophora, the causative agent of clubroot of Chinese cabbage, and also reduced the RA of other harmful fungal genera pathogenic to plants. The results of microbial diversity analysis at the genus level are similar to the results of qPCR gene copy number, which showed that the population of P. brassicae significantly decreased under soil fumigation (N, NB, and LNB) than non-soil fumigation (GZ).

An OTU enrichment analysis was carried out to determine which bacterial and fungal taxa are sensitive to specific treatments. A total of 20 bacterial and 20 fungal-specific OTUs were recovered (RA > 0.01%) in the rhizosphere of Chinese cabbage under different treatments (Figure 8; Supplementary Tables S10, S11). A number of bacterial OTUs including OTU_7, OTU_37, OTU_17, OTU_6, OTU_40, OTU_18, OTU_200, and 32 were highly abundant to N followed by OTU_14, OTU_5, OTU_4, and OTU_9 dominant to GZ, OTU_3, OTU_27, OTU_2, and OTU_11 enriched to GZ and OTU_1 and OTU_73 dominant to LNB (Figure 8A; Supplementary Table S10). In the fungal community, OTUs such as OTU_14, OTU_15, OTU_18, OTU_37, OTU_23, OTU_3, and OTU_5 was predominant to GZ, while OTU_9, OTU_17, OTU_24, and OTU_13 was present in high RA with LNB. OTU_21, OTU_6, and OTU_51 were highly abundant in the rhizosphere soil of NB, and OTU_4 and OTU_7 were most dominant to LNB (Figure 8B; Supplementary Table S11).

The highly abundant bacterial OTUs were identified as Providencia, Sphingomonas, Pseudomonas, Bacillus, Pedobacter, Flavobacterium, Acidovorax, Stenotrophomonas, Sphingobacterium, Bradyrhizobium, Pseudarthrobacter, Devosia, and Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium. The most abundant fungal OTUs have belonged to Olpidium, Plasmodiophora, Sordariomycetes, Plectosphaerellaceae, Fusarium, and Mortierella. It is worth noticing that OTU_3 was present in high RA in GZ and was identified as P. brassicae, while its RA significantly decreased in N, NB, and LNB (Figure 8B; Supplementary Table S11). These results demonstrated that soil fumigation with lime and ammonium bicarbonate significantly reduced the population of P. brassicae in the rhizosphere of Chinese cabbage, which mitigate the incidence of clubroot disease.

A microbial co-occurrence network analysis was conducted among the top 20 bacterial and fungal OTUs according to “Spearman correlation coefficient” to investigate the overall correlations among the microbial communities (Figure 9). In network analysis, nodes represent the core bacterial and fungal OTUs, while edges show the positive and negative correlation among the microbial OTUs. The highest positive correlation was observed among the clubroot pathogen and rhizosphere microbes in GZ (70.21%) than that of NB (66.85%), N (55.84%), and LNB (58.28%). This suggested that rhizosphere microbes enhance the population of P. brassicae and competition for resources among microbes, which results in the highest disease incidence. In contrast, soil fumigation with N and LNB reduced the population of P. brassicae and competition for resources, which mitigates the incidence of clubroot by making the rhizosphere microbiome healthier and more complex.