To study the biocontrol ability of L. enzymogenes OH11 (OH11) against Phytophthora infection, we cultivated OH11 in LB media overnight and transferred the culture to 1/10 TSB (herein refers to as TSB) media for 24 hours to induce HSAF production, as previously described (Qian et al., 2013). The cultures were normalized to OD_{600 =} 1.0 with fresh TSB for further use. Phytophthora zoospores were prepared and normalized to 100 zoospores/μL. Afterwards, 200 zoospores of P. sojae strain P6497 were mixed with the prepared OH11 culture and inoculated on soybean (Glycine max) etiolated seedlings, where both TSB and an irrelevant E. coli strain Top10 were used as negative controls. Two days after inoculation, P. sojae treated by TSB or E. coli caused obvious lesions on soybean seedlings, while OH11 significantly inhibited the virulence of P. sojae ( Figure 1A ). We analysed infected soybean seedlings in more detail at the microscopic level with GFP-labelled P. sojae. At 6 hpi (hours post inoculation), both TSB- and E. coli-treated zoospores had germinated and started to infect soybean, whereas no cyst or hyphae was observed in OH11-treated zoospores. After 24 hpi, the TSB- or E. coli-treated P. sojae invaded into soybean and grew in soybean cells, while OH11-treated P. sojae showed no invasive hyphae ( Figure 1B ), suggesting that OH11 completely inhibited the infection ability of P. sojae to soybean cells. Using a similar method, we found that OH11 also inhibited the virulence of P. capsici (LT263) on N. benthamiana leaves and P. infestans (88069) on potato (Solanum tuberosum) leaves ( Figure 1A ). These results collectively indicated that L. enzymogenes OH11 could effectively prevent infection of its host plants by various Phytophthora pathogens.

To understand how OH11 targets Phytophthora pathogens to suppress their infection, we first tested whether the antifungal metabolite HSAF is a potentially key factor (Yu et al., 2007). For this purpose, we carried out a plate-based anti- Phytophthora assay in which 5 μL cultures of wild-type OH11 or the HSAF-deficient mutant ΔlafB generated in a previous study (Wang et al., 2017) were spotted on the surface edge of V8 or RSA agar plates followed by inoculation of three Phytophthora pathogens (P. sojae P6497, P. capsici LT263 and P. infestans 88069) in the center of the plate. After incubation for 3 days, we found that wild-type OH11 produced a distinct zone of inhibition around the tested Phytophthora colonies, whereas HSAF-deficient mutants almost lost this antagonistic activity ( Figure 2A ).

To confirm the above findings, we carried out similar assays in liquid media in which hyphal plugs of P. sojae P6497, P. capsici LT263 or P. infestans 88069 were mixed with empty V8 broth (negative control) or L. enzymogenes cultures, respectively. After 3-day co-inoculation, we found that wild-type OH11 significantly inhibited mycelial growth of the three Phytophthora pathogens mentioned above compared with the negative control ( Figure 2B ). Under similar testing conditions, we were surprised to find that the HSAF-deficient mutant ΔlafB also showed significant inhibition of mycelial growth compared to the negative control ( Figure 2B ). Via microscopic observation represented by P. sojae, we further observed that wild-type OH11 could completely degrade mycelia, while HSAF-deficient mutants and negative control (TSB) could not, although ΔlafB also effectively inhibited hyphal growth in liquid medium ( Figure 2B ). These results suggested that, while HSAF is important during the antagonistic interaction between L. enzymogenes and Phytophthora pathogens, other HSAF-independent factors were also involved in this process.

Besides mycelial growth, cyst germination is another key step preceding Phytophthora infection (Judelson and Blanco, 2005). We therefore investigated whether L. enzymogenes OH11 could inhibit cyst germination of Phytophthora pathogens in an HSAF-dependent and/or HSAF-independent manner. We carried out a corresponding assay to address this issue using GFP-labeled P. sojae as a representative. We mixed P. sojae cysts with cultures of wild-type OH11 or HSAF-deficient mutant ΔlafB in liquid broth, followed by microscopic observation. As shown in Figure 3A , approximately 80% of P. sojae cysts were degraded by OH11 after 6 hours of co-incubation compared to the negative control (TSB), and almost all remaining cysts showed undetectable GFP signals, indicating that L. enzymogenes was able to kill P. sojae cysts instantly. This killing effect appeared to be specific, as another negative control (E. coli) failed to do so. Under similar testing conditions, although ΔlafB failed to directly kill P. sojae cysts, this mutant also significantly reduced germination rate and germ tube length of P. sojae ( Figure 3A ). After 24 hours of co-incubation, wild-type OH11 completely digested cysts, whereas ΔlafB only inhibited cyst germination ( Supplemental Figure S1 ). Similar findings were also observed when P. capsici and P. infestans cysts were applied ( Figures 3B, C ). Together, these results indicated that L. enzymogenes OH11 had the ability to directly degrade Phytophthora cysts, and that HSAF was involved in this process. In the absence of HSAF, L. enzymogenes might also inhibit cyst germination through other uncharacterized factors/pathways.

The interesting findings above prompted us to explore the underlying mechanisms at the molecular level. To this end, we carried out an RNA-seq assay to identify P. sojae cyst genes whose transcription could be significantly altered by L. enzymogenes OH11 treatment. To facilitate this assay and considering that wild-type OH11 can degrade the cysts in liquid medium, we therefore depleted OH11 cells by filter and treated P. sojae cysts with the cell-free supernatant that exhibited resistance to P. sojae activity similar to total OH11 culture ( Supplemental Figure S2 ). The results of RNA-seq assay showed that among the 13014 expressed P. sojae genes, 1741 genes were up-regulated and 2176 genes were down-regulated in the OH11-treated P. sojae compared with the negative control (TSB-treated) ( Figure 4A ). Details of differentially expressed genes (DEGs) were listed in Supplemental Table S1 . The functions of DEGs were enriched in basal metabolism, including translation initiation, RNA binding, and amino acid metabolism ( Supplemental Figure S3 ). Meanwhile, a large number of genes involved in signal transduction (46 in total number), stimulus response (95 in total number), antioxidation (20 in total number), and transporting (210 in total number) were also differentially expressed ( Figure 4B ). Considering that Phytophthora can secrete effectors to subvert plant immunity and promote pathogen infection (Wang and Wang, 2018), and several transcription factors (TFs) of bZIP, MYB and HSF families are key transcriptional regulators of during Phytophthora cyst germination (Zhang et al., 2012; Ye et al., 2013; Sheng et al., 2015), we also investigated whether these genes were included in the RNA-seq data. As shown in Figure 4C , we found that the expression levels of 27 RxLR effector genes and 21 CRN effector genes as well as 15 bZIP TFs, 5 HSF TFs and 20 MYB TF genes were significantly changed when P. sojae interacted with L. enzymogenes OH11. In addition, the expression levels of PsAvh457 (RxLR effector), PsCRN70 (CRN effector), PsBZP4 (bZIP TF), Ps138282 (MYB TF), Ps156741 (homolog of apoptosis-inducing centromere protein) and Ps131084 (cell wall endo-beta-1, 3-glucanase) were verified by quantitative real-time PCR (qRT-PCR) ( Figure 4D ). These results provided molecular evidence supporting the conclusion that L. enzymogenes prevented P. sojae infection by attacking the cysts.

In addition to directly targeting pathogens, eliciting host plant immune response is another common biocontrol strategy shared by plant-beneficial Pseudomonas and Bacillus species (Yu et al., 2022). We were therefore also interested in finding out whether L. enzymogenes could use this similar strategy to prevent Phytophthora infection. To address this question, we selected P. capsici-N. benthamiana interaction system as a working model. We first treated leaves of N. benthamiana with TWEEN-20 to promote bacterial adhesion to plant surfaces (Paul et al., 2012), and then sprayed cultures of wild-type OH11 and ΔlafB on the leaves. Meanwhile, empty TSB broth was used as a negative control. After 12 hours and 24 hours of treatment, we examined the expression levels of those well-characterized plant defence-related genes in N. benthamiana that serve as molecular indicators of plant immune responses, as previously described (Yang et al., 2021). The result showed that both wild-type OH11 and ΔlafB induced the expression of NbPR1b (Pathogenesis-related protein 1b), NbPR2, NbPR2b, NbPR4 and NbLOX (Lipoxygenase) at 24h. At 12 h, OH11 induced higher levels of NbPR2, NbPR4 and NbLOX gene expression compared with ΔlafB ( Figure 5A ). Among them, NbPR1b, NbPR2 and NbPR2b are known maker genes of the salicylic acid (SA) signaling pathway, while NbPR4 and NbLOX are well characterized to participate in the jasmonic acid (JA) signaling pathway (Zhang et al., 2015; Yu et al., 2022). These results indicated that OH11 could elicit plant immune response in a HSAF-independent manner by activating the expression of defense genes involved in the SA and JA signaling pathways. To validate this point, we conducted another plant-based biocontrol assay. We inoculated P. capsici zoospores on N. benthamiana leaves at 24 hours after L. enzymogenes treatment and found that both wild-type OH11 and ΔlafB effectively inhibited the infection of N. benthamiana by P. capsici ( Figure 5B ).

The strains used in this study are listed in Supplementary Table S2 . P. sojae strain P6497 and P. capsici strain LT263 were routinely cultured on 10% V8 media at 25°C, while P. infestans strain 88069 was routinely cultured on RSA media at 18°C. Soybean cultivar Hefeng47 was grown at 25°C in the dark for 4 days in environmentally controlled growth chamber, and etiolated soybean seedlings were used for infection assays. N. benthamiana was grown at 25°C in a growth chamber with a 16 h light and 8 h dark photoperiod, and 4-week-old leaves were selected for infection assay. Potato cultivar Desiree was grown at 22°C in a growth chamber with a 16 h light and 8 h dark photoperiod, and 4-week-old leaves were selected for infection assay. Bacteria strains were grown overnight in LB media at 28°C and then transferred to 10% TSB media for 24 h at 28°C to induce HSAF production.

For mycelial growth assays on solid media, 5 × 5-mm hyphal plugs of P. sojae or P. capsici were inoculated onto 10% V8 medium plates at 25°C, while P. infestans hyphal plugs were inoculated onto RSA medium plates at 18°C. When the fresh mycelia grew to a suitable size (3 cm in diameter), 5 μL of bacterial culture (diluted to an OD600 = 1.0) was inoculated to the edge of the plates. Antagonistic activity was measured by an inhibition zone around the colony. For mycelial growth assays in liquid media, 5 × 5-mm hyphal plugs of P. sojae or P. capsici were inoculated onto the plates containing 4 mL 10% V8 liquid medium at 25°C, while P. infestans hyphal plugs were inoculated onto plates containing 4 mL PEA liquid medium at 18°C. 1 mL of bacterial culture (diluted to an OD600 = 1.0) was inoculated onto the plates. Antagonistic activity was measured by the size of Phytophthora colony. Mycelia colonies were observed with a microscope (Zeiss Axio Observer 3).

To induce zoospore production, mycelia of P. sojae or P. capsici were cultured in 10% liquid V8 medium for 3 days, washed with sterile water, and incubated at 25°C until zoospores were produced. To induce zoospore production of P. infestans, mycelia was grown on RSA medium plate at 18°C for 14 days. The sporangia were scraped into sterile water and incubated at 4°C until zoospores were released. Zoospores were diluted to 100 zoospores/μL and mixed with an equal volume of bacterial cultures (diluted to an OD600 = 1.0). A final mixture containing 200 zoospores were inoculated onto the host plants. Infected soybean and N. benthamiana were incubated in the dark at 25°C for 2 days before observation. Infected potatoes were incubated in the dark at 18°C for 3 days before observation. Lesions of plant leaves were observed under UV light. Infection details of GFP-labelled P. sojae zoospores were observed with a microscope (Zeiss Axio Observer 3).

To examine the immune response of N. benthamiana, leaves of N. benthamiana were quickly soaked into 0.5% (v/v) TWEEN-20 and washed with H₂O to remove TWEEN-20. The TSB, OH11 and ΔlafB cultures were then sprayed on the leaves individually. Treated leaves were incubated at 25°C in the dark before sampling. To detect immune gene expression, leaf samples were collected 12 and 24 hours after treatment. For the infection assay, 200 P. capsici zoospores were inoculated onto N. benthamiana leaves 24 hours after treatment. Infected leaves were incubated at 25°C in the dark for 2 days before observation.

Phytophthora zoospores were induced and harvested into 1.5 mL tubes. Tubes containing 100 zoospores/μL in a 500 μL suspension were vortexed for 90 s to induce cyst formation. 500 μL cyst suspension were mixed with 500 μL 10% V8 medium and 500 μL bacterial culture (diluted to an OD600 = 1.0), and then incubated at 25°C to observe the germination of P. sojae or P. capsici. 500 μL P. infestans cyst suspension were mixed with 500 μL PEA medium and 500 μL bacterial culture (diluted to an OD600 = 1.0), and then incubated at 18°C to observe the germination, which was observed with a microscope (Zeiss Axio Observer 3). GFP-labeled P. sojae cysts were stained with 1 μg/mL chitin-binding CFW (Caleofluor White) for 5 min to make cysts clearer to observe.

For RNA-seq sampling, bacterial culture of L. enzymogenes OH11 were collected and diluted to a density of 1.0 at 600 nm. The bacteria culture was then centrifuged and filtered through a 0.22-μm filter to remove OH11 cells. P. sojae cysts were diluted to 100 cyst/µL with sterile ddH₂O. Bacterial suspension and cysts were mixed in a 1:1 ratio and cultured for 6 hours at 25°C prior to sampling. TSB media-treated cysts were used as negative controls. Three replicates of each sample were collected and sent to BGI Genomics Corporation for RNA-seq. P. sojae RNA was extracted using the TRIzol^® method following the manufacturer’s protocol. RNA-seq was conducted using the DNBseq platform and 100-bp paired-end modules.

Raw reads were filtered for subsequent analysis by removing reads containing adapter, poly-N, and low-quality reads. Clean reads were mapped to the genome of P. sojae (v1.1 for isolate P6497) using Tophat with up to two mismatches. Mapped reads were quantified using the Cufflinks program, and transcript levels for each gene were quantified as RPKM (reads per kilobase transcript length per million reads mapped). Differentially expressed genes were identified using featureCounts software, and log2 fold change (log2FC) values and adjusted P-values were calculated using DESeq2 software. Genes with adjusted P-value < 0.05 and absolute log2FC ≥1 were considered differentially expressed.

For qRT-PCR assays, cDNA was synthesized with PrimeScript First Strand cDNA Synthesis Kit (TaKaRa Bio Inc.) following the manufacturer’s protocol. qRT-PCR was performed in a reaction mixture of 20 μL of SYBR Premix ExTaq (TaKaRa Bio Inc.) following the manufacturer’s protocol. PCR was performed on an ABI Prism 7500 Fast Real-Time PCR System (Applied Biosystems Inc.) The P. sojae actin gene (Psactin, Ps108986) was used as a constitutively expressed endogenous control for P. sojae. The N. benthamiana elongation factor 1-alpha gene (NbEF1α) was used as endogenous control for N. benthamiana. Primers used in this study are listed in Supplementary Table S3 .