The reference strain of P. sojae, P6497, was provided by Prof. Brett M. Tyler (Oregon State University) and served as the wild-type strain for gene editing to obtain mutants in this study, of which the genome has been sequenced and annotated (Tyler et al., 2006). P6497 and all transgenic strains obtained in this study were routinely cultured on 10% V8 agar medium at 25°C in darkness (Miller, 1955; Erwin and Ribeiro, 1996). Sporangium production and ZO release of P. sojae were performed according to a method described by Ward et al. (1989), with a few modifications.

The susceptible soybean cultivar Japan Blue was grown in plastic pots containing vermiculite at 25°C for 5–7days in a greenhouse (27±2°C, 80% relative humidity, and in the dark) for harvesting etiolated soybean seedlings.

Total DNA was extracted from the mycelia (MY) of P. sojae strains grown on 10% V8 agar medium at 25°C for 4days using the method described by Ristaino et al. (1998), with some modifications (Miao et al., 2018). Total RNA was extracted from the frozen samples using the SV Total RNA Isolation kit (Promega, Beijing, China) and cDNA was synthesized using the PrimeScript RT reagent Kit with gDNA Eraser (Takara, Beijing, China) according to the protocols of the manufacturers. All the DNA and cDNA samples were stored at −20°C until required.

The PsSCPs genes were blasted against the JGI genome database¹ and the FungiDB data set.² The sequences of SCPs from other species were collected from the National Center for Biotechnology Information (NCBI) database.³ The valid accession numbers of protein sequences downloaded from different databases (JGI, FungiDB, or NCBI) have been listed in Supplementary Table S1. Hidden Markov Model was used to predict putative Ptc proteins, based on the homologous Ptc proteins reported in other species (Supplementary Table S1), with the score more than 100 as a threshold (Grewal et al., 2019, 2020). The online open reading frame (ORF) finder⁴ was carried out to predict the ORFs of PsSCPs. The PsSCP gene models were corrected based on the RNA-seq data (FungiDB), and the reverse transcription-PCR (RT-PCR) were carried out to verify the corrected gene models. The putative transmembrane domains of PsSCPs were predicted by using TMHMM Server v.2.0.⁵ Phylogenetic analyses of PsSCPs amino acid sequences with other SCP proteins reported in different species were performed using MEGA6⁶ according to Hall (2013), and conserved domains were detected using SMART.⁷ Multiple sequence alignment of amino acid sequences generated with Clustal W (https://www.ebi.ac.uk/Tools/msa/clustalo/; Thompson et al., 1994).

The PsPTL gene transcription levels were examined by quantitative real-time PCR (qRT-PCR) in different developmental stages and during the infection process. All samples, including MY, hyphae with sporangia (SP), ZO, cystospores (CY), germinated cystospores (GCY), andetiolated soybean seedlings at 0, 1.5, 3, 6, 12, 24, and 48h after inoculation, were prepared as described in the previous study with minor modifications (Ye et al., 2011). The RNA and cDNA were obtained according to the above method and stored at −80°C until required. The primers used for this experiment were shown in Supplementary Table S2, and the P. sojae housekeeping genes RPS and RPL13α were served as endogenous controls (Wang et al., 2011). The qRT-PCR reactions were performed following the manufacturer’s instructions (FastSYBR Mixture, CW Biotech, Beijing, China) by using a qPCRsoft 3.4 system (qTower 2.2, Analytik Jena AG, Jena, Germany) under the following conditions: 95°C for 2min, 40cycles of 95°C for 10s, 60°C for 30s to calculate cycle threshold (Ct) values, followed by 95°C for 15s, 60°C for 1min, and then 95°C for 15s to obtain melt curves. The relative expression levels of each gene were calculated using the 2^−ΔΔCt method (Cheng et al., 2019). Two biological replicates, each containing three technical replicates, were carried out for each sample.

The eGFP gene was fused to the C-terminus of PsPTL by ligation to SpeI/ApaI-digested pYF3 (Fang et al., 2017b). The primers used for sequence amplification are shown in Supplementary Table S2. The pYF3-PsPTL-eGFP construct was introduced into P. sojae P6497 by polyethylene glycol (PEG)-mediated protoplast transformation (Fang and Tyler, 2016), and the empty construct pYF3-eGFP was transformed into P6497 as a negative control. The living hyphae of transformants were picked from the 10% cleared V8 broth containing 50μg/mlG418 after 3days growth, which were observed by using confocal microscopy (Olympus Fluoview FV3000). Images were captured using a 60× oil objective with excitation/emission settings (in nm) 488/510~535 for green fluorescent protein (GFP).

PsPTL gene deletion mutants were generated using the CRISPR/Cas9 System (Fang and Tyler, 2016). According to the previous protocol (Fang and Tyler, 2016), three single guide RNA (sgRNA) listed in Supplementary Table S2 were designed and cloned into the backbone plasmid pYF2.3G-ribo-sgRNA, respectively. The NPTII gene ligated with two 1,000bp fragments flanking PsPTL gene was used as donor DNA in homology-directed repair (HDR) and infused into the basic donor plasmid pBS-SK+ using the primers listed in Supplementary Table S2 and the In-Fusion HD Cloning Kit (Clontech, Mountain View, CA, United States). All DNA fragments were amplified using TransStart® FastPfu DNA Polymerase (TransGen Biotech, Beijing, China). We transformed P. sojae wild-type P6497 using PEG-mediated protoplast transformation using previously described methods (Fang and Tyler, 2016; Fang et al., 2017a). Putative transformants were transferred to the 10% V8 agar medium containing 50μg/ml G418 and incubated for 3days at 25°C. For verification of the transformants, their genomic DNA was extracted and the homologous recombination events were verified by PCR using primers as shown in Supplementary Table S2. Sanger sequencing and qRT-PCR were also used to confirm that the PsPTL gene was cleanly replaced by NPTII gene and absent in the homozygous deletion transformants.

To determine mycelial growths of P. sojae wild-type (P6497) and transformants, colonies were measured after depositing a mycelial plug (5mm in diameter) on 10% V8 agar medium after 4days (Miao et al., 2018).

To analyze SP production, 10 mycelial plugs (5mm in diameter) from the edge of actively growing colonies on the V8 juice agar medium were placed into a Petri dish, containing 20ml of 10% clarified V8 broth. These plates were incubated at 25°C in the dark for 3days then the hyphae were repeatedly washed with 20ml of sterile distilled water (SDW) for five times, followed by incubation in the dark at 25°C for 4–8h, until sporangia developed. The number of sporangia was counted using the light microscope (magnification ×100).

To gauge the zoospore production of P. sojae, zoospores were produced as described above. After washing the mycelia five times, another 10-ml SDW was added into the Petri dishes, which were incubated in the dark at 25°C for 8–10h. The number of zoospores was counted using a hemocytometer.

To assess cyst germination, the zoospore suspensions were shaken using a vortex for 1min in order to complete cystospore encystment. The cystospore suspension was incubated at 25°C for 6h in the dark. More than 100 cystospores were examined under the microscope, which were considered as germinated if the germ tube was longer than the cystospore diameter.

To evaluate oospore production, the mycelial plug (5mm in diameter) was cultured on 10% V8 agar for 7days. Then, the number of oospores was counted using the light microscope (magnification ×10). Three duplicates of each strain were incubated in darkness at 25°C, and the experiment was repeated three times.

To investigate the pathogenicity of P. sojae, the etiolated seedlings of soybean cultivar Japan Blue were inoculated with mycelium plugs (5mm in diameter) on the hypocotyls, and maintained at 25°C and 80% relative humidity in the dark for 3days. Then, the lesion length was measured. Each strain was tested using at least six seedlings. All assays were repeated three times.

To determine the response of P. sojae to temperature, the wild-type (P6497), control strain (EV), and PsPTL mutants were grown on Plich medium (Qiu et al., 2020) at 10, 18, 25, 30, and 37°C, respectively. Colony diameters were measured after incubation in the dark for 7days, and each treatment consisted of three replicated plates.

To determine the response of P. sojae to pH stress, all strains were grown on Plich medium with a pH value of 5, 6, 7, 8, and 9 at 25°C, respectively, in the dark for 7days. Colony diameters were measured, and each treatment consisted of three replicated plates.

To determine the response of P. sojae to osmotic stress, all strains were grown on Plich medium amended with KCl (0, 0.25, and 0.5M), as well as sorbitol (0, 0.5, and 1M) at 25°C in the dark for 7days, respectively. Colony diameters were measured, and each treatment consisted of three replicated plates.

To determine the response of P. sojae to oxidative stress, all strains were grown on Plich medium amended with 0, 2, and 5mM H₂O₂, respectively, at 25°C in the dark for 7days, respectively. Colony diameters were measured, and each treatment consisted of three replicated plates.

To determine the response of P. sojae to exogenous sterol, the wild-type (P6497), control strain (EV), and PsPTL mutants were grown on basal medium (Jee and Ko, 1997) amended with 0, 5, and 20μg/ml β-sitosterol (Shanghai Yuanye Bio-Technology Co., Ltd., Shanghai, China). All strains were measured after incubation in darkness at 25°C for 7days. Each treatment consisted of three replicate plates, and the experiment was repeated three times.

All mycelial samples of PsPTL (PsSCP3) mutants were collected after incubation in darkness at 25°C for 4days on 10% V8 medium. The RNA and cDNA were obtained according to the above method and stored at −80°C until required. Transcript levels of other PsSCPs genes were measured by qRT-PCR. The primers used for this experiment were shown in Supplementary Table S2, and the P. sojae housekeeping genes RPS and RPL13α were served as endogenous controls (Wang et al., 2011). The qRT-PCR reactions were performed as described above. The relative expression levels of each gene were calculated using the 2^−ΔΔCt method (Cheng et al., 2019). Two biological replicates, each containing three technical replicates, were carried out for each sample.

All statistical analyses were conducted using DPS software version 7.05. Differences between the means were determined by the least significance difference (LSD) multiple range test at p=0.05.

A total of 59 predicted Ptc proteins (Supplementary Table S3) were found in the genome of P. sojae P6497 by searching for the conserved domains in JGI database,⁸ and predicted by Hidden Markov Model. However, it was found that there are positional overlaps among different genes, and only six genes are in different positions, with the protein IDs of 466,676, 318,600, 549,457, 322,952, 389,466, and 564,545, respectively (Supplementary Table S3). We further searched RNA-seq data of these genes in FungiDB, and found that 466,676 and 322,952 are individually expressed in a continuous mode. Although, 318,600 and 549,457 are annotated as two different genes, they are expressed together uninterruptedly, which is similar to 389,466 and 564,545 (Supplementary Figure S1). The prediction of ORF and full-length amplification by RT-PCR were carried out to correct the gene models, and four independent genes were identified, because 318,600 and 549,457 with their sequence interval constituted one independent gene, and 389,466 and 564,545 with their sequence interval constituted another independent gene. We artificially named these four proteins as PsSCP1-4, which referred to 466,676, 318,600 and 549,457, 322,952, and 389,466 & 564,545, and the corrected sequences were shown in the Supplementary Material. Four homologous proteins were also found in the genome of Phytophthora capsici,⁹ and they were denoted as PcSCP1, 2, 3, and 4, correspondingly.

The phylogenic analysis showed that PsSCPs and PcSCPs were organized into a cluster of NPC1 homologs (Figure 1A). The conserved domain prediction of PsSCPs and PcSCPs confirmed that the N-terminal NPC1 domain (NPC1-N), SSD, and Patched domains are present in SCP1, 2, and 4 (Supplementary Figure S2). However, it found that SCP3 did not contain NPC1-N domain (Figures 1B,C). The distribution and composition of conserved domains are also different between PsSCP3 and the other three proteins (Supplementary Figure S2; Figure 1B). In other species, both NPC1 and Patched (Ptc) proteins have SSD and Patched domains, while the N-terminal domain of NPC1 (NPC1-N) is only present in the homologous proteins of NPC1 (Loftus et al., 1997; Malathi et al., 2004). NPC1-N plays a vital role in binding cholesterol (Qian et al., 2020). NPC1 and Ptc proteins are involved in different signaling pathways, and their mutants showed distinct phenotypes, which may be due to the special presence of NPC1-N (Kuwabara and Labouesse, 2002). In Phytophthora, PsSCP1, 2, 4 display similar protein structures to other NPC1 homologs with the specific NPC1-N domain (Loftus et al., 1997; Malathi et al., 2004), and it is found that they are more closely related to NPC1 proteins in the phylogenetic analysis (Figure 1A). Because PsSCP3 does not contain the specific NPC1-N domain and similar protein structure to NPC1 homologs, and has similar domains as Ptc, it may be the PTL protein. We performed a phylogenic analysis of SCP proteins in other oomycetes, such as Phytophthora ramorum, Phytophthora infestans, Peronospora effusa, Phytophthora parasitica, and Phytopythium vexans. We found that PTL (SCP3) proteins were conserved in different species of oomycetes, and evolved independently into a branch that is similar to Patched protein (Supplementary Figure S3). Hence, PsSCP3 was renamed artificially as PsPTL protein, and amplified for a full length of 3,138bp, of which the open reading frame includes one intron and encodes a 1,046-amino-acid protein (Figure 1B), with 11 putative transmembrane domains. The biological function of PsPTL has been further studied in this research.

We found that PsPTL showed higher expression during the mycelial and sporulation stages, compared to the zoospore, cyst, and germinated-cyst stages (Figure 2A), indicating that PsPTL may play an important role in mycelial growth and sporulation. Similar transcriptional expressions of PsPTL were found in the mycelium (0h) and during all indicated time points of infection (1.5–48h after inoculation; Figure 2B), which suggests that PsPTL may not have a prominent function in the pathogenic stage of P. sojae.

The eGFP-tagged PsPTL protein was observed in the hyphae of P. sojae by confocal microscopy (Olympus Fluoview FV3000). The green fluorescence of the transformants expressing solo eGFP distributed throughout the hyphae except some vacuoles, and the strongest fluorescence signal was in the nucleus, which has been observed in the study of Fang et al. (2017b). While, in the transformants expressing the eGFP-tagged PsPTL protein, green fluorescence mainly scattered or accumulated in the cytoplasm with a punctate or spot structure, which was different from the transformants that only expressed eGFP (Figure 3).

To explore the biological function of PsPTL in P. sojae, the wild-type strain P6497 had been genetically modified and four independent homozygous mutants (ΔP-8, ΔP-11, ΔP-14, and ΔP-23) were obtained and confirmed by Sanger sequencing (Supplementary Figure S4) as well as qRT-PCR (Figure 4). The transformant in which the PsPTL deletion was not successful was included as a negative control (EV). The biological characteristics of all PsPTL deletion transformants, the wild-type and EV strains, were analyzed. For mycelial growth, all transformants showed slightly slower growth rates than the wild-type strain P6497 and the negative control (Figure 5). The sporangium production and zoospore production of all mutants displayed no significant difference compared to the wild-type strain P6497 or EV transformant. It was found that the disruption of PsPTL did not significantly affect the cystospore germination of P. sojae, except for mutant ΔP-14, which had a significantly lower cystospore germination rate. In addition, mutants ΔP-8 and ΔP-11 produced a considerable number of oospores relative to P6497, whereas ΔP-14 and ΔP-23 produced fewer oospores (Table 1). In addition, no significant difference was found in virulence among all the strains (Figure 6).

To determine whether PsPTL plays an essential role in coping with external stresses of P. sojae, PsPTL mutants and wild-type strain P6497 were cultured under various environmental stresses, such as temperature, pH, osmotic stress, and oxidative stress. The optimum temperature for mycelial growth of all transformants and P6497 was 25°C. Transformants ΔP-8 and ΔP-11 showed slightly faster mycelial growth rates than the wild-type P6497 and EV transformant, whereas ΔP-14 and ΔP-23 showed slower mycelial growth rates (Figure 7A). Further, all mutants did not have significant changes in response to pH conditions, oxidative stress (H₂O₂), and osmotic stress (KCl and sorbitol), compared to P6497 and EV (Figures 7B–E).

We speculated that the PsPTL deletion may affect the influence of exogenous sterols on P. sojae because of the conserved SSD domain. Therefore, mycelial growths of PsPTL mutants and wild-type strain P6497 were determined with exogenous β-sitosterol. However, the results showed that the growth pattern of the mutants did not change significantly compared with the wild-type or EV strain after adding different concentrations of exogenous β-sitosterol (Figure 7F), indicating that the deletion of PsPTL might not affect the absorption and utilization of exogenous sterols by P. sojae.

Because the other three PsSCPs also have conserved SSD and Patched domains, it is plausible that they may have functional redundancy with PsPTL (PsSCP3). The transcription levels of PsSCPs genes in PsPTL mutants and the wild-type strain were then determined. The transcript levels of PsSCP1 increased insignificantly or even decreased slightly in transformants, compared to the EV strain. Regarding PsSCP2 transcription, although ΔP-11 and ΔP-23 showed a slight increase (relative expression fold change <4) compared to the wild-type P6497 and EV transformant, ΔP-8 and ΔP-14 were not significantly different from the wild-type. Furthermore, transcript levels of PsSCP4 in ΔP-11 and ΔP-14 were slightly increased with the relative expression fold change being less than three times, whereas ΔP-8 and ΔP-23 showed no significant difference compared to the wild-type or EV (Figure 4).