Plasmodiophora brassicae was isolated from Dayi in Sichuan Province, China. The isolates were identified as pathotype 4 based on the differentials of Williams (1966), which were previously described (Zheng et al., 2019). Resting P. brassicae spores were prepared as previously described (Ji et al., 2014). Root galls were homogenized using 10% (m/v) sucrose in a blender. The slurry was filtered through eight layers of gauze, and the suspension was clarified by centrifugation at 3,000 × g for 10 min. The resting spore precipitate was suspended in sterile water, adjusted to a concentration of 1 × 10⁷ (spores)/ml, and stored at 4°C.

Seeds of a susceptible host cabbage (Brassica rapa) and non-host spinach (Spinacia oleracea) were placed on Petri dishes to germinate for 3–4 days in a plant growth chamber at 25°C under a light-dark cycle of 14:10 h. The germinations were then transferred into tubes containing Hoagland solution and incubated for 3–4 days. When the first euphylla was available, the seedlings were transferred into Hoagland solution containing 1 × 10⁵ (resting spores)/ml in a tube, and it was ensured that the roots were immersed in the resting zoospore suspension. The tube was packaged with adhesive black tape. The inoculated plants were maintained in a growth chamber under 14-h light/10-h dark cycles and a temperature of 25°C. Each of the inoculations was repeated in triplicate (more than 10 seedlings per replicate).

To observe secondary zoospore infection in host and non-host plants, secondary zoospores were produced using a previous method (Feng et al., 2012; McDonald et al., 2014). Spinach seeds were planted in the soil. Seven-day-old spinach seedlings were inoculated with 1 × 10⁷ (resting spores)/ml under 14-h light/10-h dark cycles at a temperature of 25°C. At 5 dpi (days post-inoculation), the spinach (Spinacia oleracea) seedlings were uprooted, and the roots were washed with tap water. The roots were cut off and rinsed three times by shaking at 150 rpm for 20 min in distilled water. Then, the roots were shaken in 50 ml of distilled water at 100 rpm for 24 h to stimulate the release of secondary zoospores. The zoospore suspension was concentrated by centrifugation at 5,000 × g for 5 min, adjusted to 1 × 10⁵ (spores)/ml, and used immediately for inoculation. The zoospore suspension was transferred to a sterilized tube. The tube was packaged with adhesive black tape. Three 7-day-old cabbage or spinach seedlings were inserted into the tube, and it was ensured that the roots were immersed in the secondary zoospore suspension. The inoculated plants were maintained in a growth chamber under 14-h light/10-h dark cycles at a temperature of 25°C. At 4, 7, or 12 dpi, more than 10 seedlings were investigated. The inoculations were repeated three times.

To analyze the pathogenicity of the secondary zoospores, non-host spinach (S. oleracea) seedlings were inoculated with resting spores as described earlier. The infected spinach seedlings were dug out 5 days after the initial date of inoculation with resting spores. The spinach seedling roots were washed with tap water to thoroughly remove the resting spores from the root surfaces and then planted directly into the soil to provide a source of secondary zoospores. Five spinach seedlings infected with cabbage seedlings were planted together. The seedlings were maintained under 14-h light/10-h dark cycles at a temperature of 25°C. At 40 dpi, all of the seedlings were harvested, and the symptoms of 50 cabbage seedlings were determined.

After dyeing with 0.5% fluorescent pink dye for 1 min, the background color of the roots on the slides was washed with ddH₂O. The plasmodia, zoosporangia, and resting spores of P. brassicae were stained red. All infection stages were observed with a Carl Zeiss Microscope (GmbH37081, Göttingen, Germany).

The roots of 50 seedlings inoculated with resting spores were harvested at 4 dpi and 7 dpi, and the samples were named P-Y4 and P-Y7, respectively. The roots of 50 seedlings that were inoculated with secondary zoospores were harvested at 4 dpi, and the sample was named S-Q. The inoculation methods for the resting spores and secondary spores and the growth conditions of seedlings are described earlier. The roots were washed thoroughly with distilled water to remove the spores adsorbed onto the surface. They were then frozen in liquid nitrogen and stored at −70°C.

Total RNA was extracted from two biologically different root samples using TRIzol^® reagent (Life Technologies, Carlsbad, CA, USA) following the manufacturer’s protocol. The RNA quantities and qualities were determined using a 2100 BioAnalyzer (Agilent Technologies, Santa Clara, CA, USA). RNA libraries were constructed with 2 μg of total RNA and subjected to high-throughput sequencing with a HiSeq-4000 sequencer (Illumina, San Diego, CA, USA). The sequencing reads were compared to the reference databases of the P. brassicae genome¹ and B. rapa genome (see text footnote 1), and the length statistics of the reference transcriptome sequences were determined with per scripts (Trapnell et al., 2009). The raw RNA-seq data were uploaded to the NCBI under accession ID SRP392270.

Raw reads from the sequencing machine were generated by Base Calling and saved in FASTQ format. Clean reads were generated by removing reads with adaptors, reads where the number of unknown bases was more than 10%, and low-quality reads (the percentage of the low-quality bases with which value ≤ 5 was more than 50% in one read). Two biological replicates of each sample, showing a correlation coefficient value of ≥ 0.95, were considered for subsequent gene expression analysis. Eight RNA libraries were analyzed by using the RNA-Seq approach and comparative differential expressed gene (DGE) profiling analysis (Wang et al., 2010). The fragments per kilobase million methods were used to calculate the normalized expression data of each gene (Mortazavi et al., 2008).

Differentially expressed genes (DEGs) were selected by the DEseq2 software, with the criteria of absolute Log2 (fold change) ≥ 1 or Log2 (fold change) ≤ −1 and false discovery rate ≤ 0.01 between two samples (Wang et al., 2010). To elucidate the metabolic pathways of these predicted genes, the Kyoto encyclopedia of genes and genomes (KEGG) pathways of DEGs were analyzed as the method reported previously (Wu et al., 2006). The R language software TBtools was used for the generation of the heatmap.

A yeast signal sequence trap assay was adopted to assess the secretory activities directed by signal peptides (Jacobs et al., 1997). DNA extraction of P. brassicae was performed using a previous method (Zheng et al., 2019). The DNA sequences encoding signal peptides were amplified by a high-fidelity DNA polymerase, and the purified DNA was then cloned into the linearized pSUC2T7M13ORI plasmid. The recombinant vectors were transformed into the chemicompetent Saccharomyces cerevisiae yeast strain, YTK12. Positive clones were used to assess the secretory activities using growth assays. Positive clones were confirmed through PCR with the primer pair, pSUC2F/pSUC2R (Supplementary Table 2), and were streaked on CMD-W (0.67% yeast N base without amino acids, 0.075% tryptophan dropout supplement, 2% sucrose, 0.1% glucose, and 2% agar) plates and YPRAA (1% yeast extract, 2% peptone, 2% raffinose, and 2 μg/mL antimycin A) plates and used a 2,3,5-triphenyltetrazolium chloride (TTC) assay according to a previous study (Chen et al., 2019).

Because secondary zoospores of different origins showed different infection abilities (Feng et al., 2012), we first tested the pathogenicity of secondary zoospores from non-host Spinacia. Spinach seedlings inoculated with resting spores were infected, and numerous zoosporangia developed in the root hairs. Healthy cabbage seedlings were planted with the infected spinach, which was used as an infection source of secondary zoospores. The cabbage roots showed obvious clubroot symptoms at 40 dpi; however, inoculation with secondary zoospores resulted in lower disease severity than inoculation with resting spores (Figure 1). The results showed the pathogenicity of secondary zoospores originating from the non-host spinach. No clubs were found on the spinach roots, regardless of whether they were inoculated with resting spores or secondary spores. Therefore, we next compared the infection biology after inoculation with secondary zoospores and primary zoospores.

To separately examine the roles of primary and secondary zoospores in the host, inoculation studies using resting spores (source of primary zoospores) and secondary zoospores were conducted. At 7 days in cabbage inoculated with secondary zoospores, limited numbers of zoosporangia were detected, and at 12 dpi, the zoosporangia numbers gradually increased in the root hairs and root epidermal cells (Figure 2). However, at 12 days in cabbage inoculated with resting spores (namely, at 7 days after secondary zoospore inoculation), the developmental asynchrony of P. brassicae was obvious. The polymorphism was characterized by secondary plasmodium, resting sporangial plasmodium, and resting spores (Figure 3). Multinucleate secondary plasmodia that were spherical, compact, and separated from each other were observed, which were more organized with clearly visible membrane boundaries (Figure 3A). Many oblong or amorphous plasmodia were observed in most of the intracellular spaces (Figures 3B,C). A vast number of resting spores had formed (Figure 3D). Given that inoculation with P. brassicae resting spores leads to morphological variations, it is more asynchronous than inoculation with secondary zoospores, and we suggest that the primary zoospores present during root hair or epidermal cell infection can more efficiently affect host resistance.

Next, we examined whether there were differences between the primary zoospore and secondary zoospore infections in non-host spinach. Inoculation studies using resting spores and secondary zoospores were conducted separately. At 7 and 12 days post-inoculation with secondary zoospores (namely, after the time of resting spore inoculation for approximately 12 and 17 days, respectively), zoosporangia were observed within the root hairs and epidermal cells, but we detected no zoosporangial clusters (Figure 4). However, at 12 and 17 days after inoculation with resting zoospores, zoosporangia clusters were observed (Figure 5). In spinach inoculated with resting spores, the proportion of zoosporangia per field of view in the microscope was higher than that with secondary zoospores. The results showed that inoculation with primary zoospores has a higher capability of zoosporangia proliferation than inoculation with secondary zoospores. Empty zoosporangia gradually appeared in seedlings inoculated with resting spores or secondary zoospores, indicating the release of secondary zoospores (Figures 4, 5).

Secretory proteins play critical roles in pathogen recognition and interactions among pathogens and hosts. In a previous study, we predicted some effector proteins based on the P. brassicae genome sequence in the NCBI. Here, to explore the possible reasons for the differences in infection biology, we compared the effector proteins in host plants that were inoculated with resting spores and secondary zoospores.

Differentially expressed genes with fragments per kilobase million (FPKM) values greater than 5 were selected. Among the 19 putative secretory proteins, four genes (e.g., Pb1–Pb4) were upregulated in three samples (namely, two samples at 4 and 7 days after resting spore inoculation and one sample at 4 days after secondary zoospore inoculation). The expression levels of nine effector genes (e.g., Pb5–Pb13) were induced at both 4 and 7 days, and six effector genes (e.g., Pb14–Pb19) had high expressions only at 4 days after inoculation by primary zoospores (Figure 6). No putative secretory proteins were upregulated in the samples that were inoculated only with secondary zoospores at 4 days. The expressions of most effector genes were upregulated after inoculation with primary zoospores. The results indicate that primary zoospores may play a more important role in the host-P. brassicae interaction.

The secretory activities directed by the signal peptides (SPs) of the above 19 putative effectors were tested. The primer sequence was listed in Supplementary Table 1. Eighteen SP sequences showed secretory activities, as demonstrated by growth assays on YPRAA and CMD-W plates in yeast and by the reduction of TTC to form insoluble red-colored 1,3,5-triphenylformazan (TPF) (Figure 7). Only Pb15, a predicted SP, did not demonstrate invertase activity in the chemicompetent Saccharomyces cerevisiae yeast strain, YTK12. Yeast transformed with Pb15 was unable to grow on YPRAA or reduce TCC.

Because the expression levels of many effectors were upregulated by primary zoospores, we next analyzed the host responses using RNA-seq data from root tissues of the resting spore and secondary zoospore inoculations. Approximately 57.84∼61.35 million raw RNA-Seq reads were produced from each sample. After filtering the dirty reads, 55.43∼58.71 million clean sequences were acquired and 70.10∼77.53% of sequences were mapped to the reference genome of Brassica rapa. Clean reads Q30 (%) were larger than 89%, and clean reads ratio (%) was around 95% (Supplementary Table 2), indicating that the accuracy and quality of the sequencing data were sufficient for further analysis.

Next, the Venn diagram of differentially expressed genes (DEGs) was constructed by comparing P-Y4 vs. C-D, S-Q vs. C-D, and P-Y4 vs. S-Q (Figure 8). Among the genes with significant differences, the numbers of special DEGs were 1,145 in sample P-Y4 vs. S-Q, 475 in sample P-Y4 vs. C-D, and 1,155 in sample S-Q vs. C-D, which were involved in a Kyoto encyclopedia of genes and genomes (KEGG) pathway enrichment analysis. The result indicated that the differentially expressed genes related to the infection responses of primary zoospores and secondary zoospores were significantly enriched (Supplementary Table 3). The top one KEGG enrichment pathway was the plant-pathogen interaction (ko04626) by comparing P-Y4 with S-Q, phenylpropanoid biosynthesis (ko00940) by comparing P-Y4 with C-D, and plant hormone signal transduction (ko04075) by comparing S-Q with C-D, respectively.

In the plant-pathogen interaction pathway, the transcription levels of the pathogenesis-related proteins, PR1 and WRKY (e.g., WRKY51, 54, 59, and 62), were upregulated in B. rapa seedlings inoculated with both primary and secondary zoospores. The leucine-rich repeat receptor-like protein kinase BAK1/BKK1, FLS2, GSO2, and PEPR 1/2; calcium-dependent protein kinase CDPK 32; calmodulin-like protein CaM/CML (e.g., 12, 37, 38, 38-like, 39, 39-like, 40, and 43); cyclic nucleotide-gated ion channel CNGC19; RPM1 (At5g66900); RPM1 (disease resistance RPP8-like); disease resistance protein RPS5 and RPM1-interacting protein RIN4; Pti5 (ethylene-responsive transcription factor ERF098); EFR (At2g24130); WRKY (e.g., 23, 30, 33, 40, and 41); and respiratory burst oxidase homology protein Rboh (e.g., F, D, D-like, and G) were not changed or upregulated in samples associated with resting spore inoculation but were downregulated in samples obtained from secondary zoospore inoculation. The expressions of WRKY (e.g., 20, 24, 28, 29, 38, 59, 70, and 71), CML41, and EDS1B-like were upregulated after inoculation with secondary zoospores (Figure 9). The results showed that most genes in the plant-pathogen interaction pathway were induced by inoculation with primary zoospores and suppressed by inoculation with secondary zoospores.

In the plant hormone signal transduction pathway, salicylic acid (SA), jasmonic acid (JA) and ethylene (ET) are important pathogen-response plant hormones. The expression of the SA marker gene, PR1, was upregulated, which indicated that the SA pathway was activated by inoculation with both primary zoospores and secondary zoospores. EIN3 and ERF1/2 in the ET pathway and JAZ and MYC2 in the JA pathway were slightly upregulated or unchanged at 4 and 7 days after inoculation by primary zoospores but were suppressed at 4 days after inoculation by secondary zoospores.

In the cytokinin pathway, B-ARR (PHL5-like) was upregulated after inoculation by both primary zoospores and secondary zoospores. In the auxin pathway, the genes were not changed after inoculation with primary zoospores; however, 2 GH3 and 2 SAUR genes were suppressed, and 9 SAUR, gibberellin pathway GID1 and TF, and ABA pathway PP2C genes were upregulated after inoculation with secondary zoospores alone (Figure 9; Supplementary Table 1). The data showed that the plant growth hormone pathway was mainly induced by secondary zoospore inoculation.

The expressions of 10 target genes were evaluated by qRT-PCR analysis to validate the RNA-seq results. The qRT-PCR results indicated high correlations among the selected genes (Figure 10).

Obvious developmental asynchrony of P. brassicae occurred at 12 days in B. rapa inoculated with resting spores. A previous report indicated that cytoplasmic cleavages yielded massive amounts of uninucleate resting spores in cortical cells at more than 20 dpi (Liu et al., 2020a). Here, we first found that resting spores can be produced at 12 dpi. Meanwhile, limited numbers of zoosporangia were detected in root hairs and root epidermal cells after inoculation with secondary zoospores. The results demonstrate that primary infection influences the subsequent development of P. brassicae. A limited developmental process may be the reason that inoculation with secondary zoospores alone resulted in lower disease severity than inoculation with resting spores alone in canola (Brassica napus) (McDonald et al., 2014). The infectious difference between primary and secondary zoospores provides more evidence that the interaction of P. brassicae and the host mainly occurs during primary infection.

In spinach inoculated with resting zoospores alone, some zoosporangia groups formed, but when inoculated with secondary zoospores alone, we did not detect any zoosporangial clusters within the epidermal cells and root hairs. The infection frequencies of root hairs and epidermal cells with primary zoospore inoculation were higher than those with secondary zoospore inoculation. However, a previous report proved that in resistant hosts, inoculation with secondary zoospores resulted in much higher levels of cortical infection than inoculation with resting spores (McDonald et al., 2014), indicating that the response associated with non-host resistance was different from that of host resistance. The secondary zoospores from infected non-host spinach successfully infected cabbage and spinach, which indicated that P. brassicae was able to complete the primary infection in the non-host. This result was different from the previous observation that P. brassicae could initiate but not complete the primary infection in non-hosts (Liu et al., 2020b). Although root hair infection with P. brassicae appears to be not host-specific, repeating the zoosporangial stage was more obvious in cabbage seedlings with primary zoospore inoculation than those with secondary zoospore inoculation, which indicated that primary zoospores undertook an important interaction with plants. The results also indicated that intercropping Brassica crops with non-host crops was not suitable for managing P. brassicae.

To establish a successful infection, P. brassicae needs to either suppress the host resistance or fail to trigger host resistance. In either case, the response is likely triggered by effectors. Because the early infection stage is a key phase for the recognition and interaction between the host and P. brassicae (Zhao et al., 2017), we detected the gene expressions of P. brassicae after inoculations with primary and secondary zoospores. Although P. brassicae effectors were presented (Chen et al., 2019 and Pérez-López et al., 2020), the differential expressions and specific effectors released by primary zoospores and secondary zoospores were not compared. Here, among the 18 effectors identified, it was interesting to note that 14 effectors were induced after inoculation with primary zoospores. The initial recognition of P. brassicae may occur during the primary infection phase.

Only a few predicted effectors had conserved domains. The effectors Pb13 and Pb17 (chitin recognition protein), Pb18 (cystatin domain), Pb5 (thioredoxin), and Pb12 (ankyrin repeats) were induced by inoculation with primary zoospores. Considering the widespread nature of chitin perception in plants, pathogens could evolve strategies to overcome detection (Sánchez-Vallet et al., 2015), and the secretion of effectors Pb13 and Pb17 might provide cell wall protection or target host immune responses. Pb18 was identified as a cysteine protease inhibitor. Many previously identified plant pathogen effectors, namely, EPIC1 and EPIC2B, from the oomycete Phytophthora infestans (Song et al., 2009) and Pit2 from Ustilago maydis (Mueller et al., 2013), inhibit plant apoplastic cysteine proteases, an interaction that appears as a key for establishing compatibility (Lo Presti et al., 2015). P. brassicae might avoid host recognition or combat host defenses by secreting these effectors.

The expressions of effectors Pb1 (alginate lyase) and Pb4 (thaumatin family protein) were induced after inoculation with primary and secondary zoospores. Alginate lyase is effective in disrupting recalcitrant cell walls in Laminaria digitata (Costa et al., 2021), and plant cell wall-degrading enzymes secreted by biotrophic plant-pathogenic fungi facilitate their penetration of host plant cells (Gibson et al., 2011). When host cell walls are broken down, the effector may have a crucial role in zoospore infection of root hairs and epidermal cells.

The P. brassicae secondary infection effectors identified by Perez-Lopez did not include the chitin recognition protein, alginate lyase, thioredoxin, or thaumatin family protein. Our results indicated that primary zoospore infection in root hairs or epidermal cells acted as a major “battleground” with the host. Cysteine protease inhibitors and ANK-containing effectors were also identified during the late infection period (Pérez-López et al., 2020). This result indicated that these effectors played an important role in P. brassicae pathogenicity during the whole infection stage.

How does the host deal with a large number of effectors released by P. brassicae during the primary infection stage? Although the transcription data indicated the differentially expressed genes during the early or late stage of P. brassicae infection (Siemens et al., 2006; Päsold et al., 2010), no study has compared the host gene expression differences between primary and secondary zoospore infections. In the plant-pathogen interaction pathway, most genes were induced at 4 days after inoculation with resting spores (which corresponds to primary infection). Pathogen perception, the key PRR genes triggered by PAMPs, such as brassinosteroid insensitive 1-associated kinase 1 (BAK1), flagellin sensing 2 (FLS2), and EFR, which transduce signals that trigger PTI, showed slightly upregulated expressions after inoculation by primary zoospores but downregulated expressions after inoculation by secondary zoospores. Because inoculation with primary zoospores does not affect further development, PTI does not play a role in the interaction between B. rapa and P. brassicae, as previous reports have shown (Chen et al., 2016). After inoculation with primary zoospores, the CDPK signaling pathway and Rboh were also induced. Some identified effectors, such as Pb5, a thioredoxin protein that mediates redox signaling in plant immunity (Mata-Pérez and Spoel, 2019), may influence the host oxidation-reduction reaction.

Most of the defense pathway induction by primary zoospores was associated with a hypersensitive response (HR) or programmed cell death. Previous reports discovered that the damage was characterized by the degradation of cell walls in susceptible hosts but not in resistant hosts (Donald et al., 2008). The HR activated by primary zoospores might lead to cell wall breaks, which favor P. brassicae penetration and movement.

Otherwise, WRKY exhibited different expression patterns. Previous reports pointed out WRKY 29, 38, 59, and 62 upregulation and WRKY 23, 33, and 40 downregulation in resistant lines (Chen et al., 2016). Therefore, we suggest that WRKY 29, 38, 59, and 62 acted negatively and WRKY 33 acted positively in response to the presence of P. brassicae. The effectors released by P. brassicae primary zoospores might tend to avoid host defenses and establish compatible interactions.

The expressions of the SA signal marker, PR1, showed significant increases in three samples, which was consistent with a previous study showing that PR1 was upregulated in a susceptible line (Wei et al., 2021). The SA accumulations in the susceptible line were higher than those in the resistant line (Wei et al., 2021), which indicated that SA did not play a key role against P. brassicae. Inoculation with secondary zoospores inhibited EIN3, JAZ, and MYC2 expressions in the JA signaling pathway. JA is believed to promote P. brassicae development and lead to host susceptibility (Xu et al., 2016; Li et al., 2018).

The expression levels of several genes in plant growth hormone pathways, such as BR pathway BRI1 and BES1/BZR1, cytokinins B-ARR (PHL5-like), gibberellin pathway GID1 and TF, ABA pathway PP2C, were changed, which indicates that these plant hormone pathways were regulated at an early stage of infection. Most IAA pathway-related genes were upregulated after inoculation with secondary zoospores (e.g., auxin-induced protein 15A, SAUR22, 32, 36, and 41). By considering that SAUR32 expressions were upregulated in a resistant line (Wei et al., 2021), we suggested that SAUR32 and several other SAURs might negatively regulate IAA levels. Previous studies have shown that auxin conjugate synthetases (GH3) are upregulated when auxin levels are higher (Jahn et al., 2013). Here, the downregulation of GH3.1 and GH3.2 might be related to low auxin levels after inoculation with secondary zoospores alone. These genes suppressed by inoculation with secondary zoospores might contribute to a susceptible reaction in the host at the subsequent infection stage. This might be one of the reasons why inoculation with secondary zoospores resulted in lower disease severity than inoculation with resting spores.