Nicotiana benthamiana seeds were surface sterilized for 5 min in 95% (v/v) ethanol, then for a further 5 min in 5% (w/v) NaClO. Seeds were washed thrice with sterile distilled water, then spread evenly on solid Murashige and Skoog medium (Murashige and Skoog, 1962) to germinate. Seedlings were transplanted to pots containing sterile vermiculite and grown for about 6 weeks in a greenhouse (light intensity 200 μE m^-2 s^-1, 50–60% relative humidity, 25°C) with a 16/8 h light/dark cycle. The bacterial strain B. amyloliquefaciens FZB42 was deposited as strain 10A6 in the culture collection of the Bacillus Genetic Stock Center (BGSC; The Ohio State University, Columbus, OH, United States). P. nicotianae was cultured at 24°C for 2 days on clarified V8-agar medium (von Broembsen and Deacon, 1996).

A hyphal plug of P. nicotianae (7 mm × 7 mm) was fixed on the surface of N. benthamiana leaves (Teng et al., 2014), and the samples were kept in the greenhouse conditions described above. Symptoms of disease were recorded after 48 h. The leaves were then placed in 100% ethanol. The rate of resistance was calculated after measurement of the diameter of P. nicotianae lesions: inhibition rate = [(diameter of control lesions-diameter of treatment lesions)/diameter of control lesions] × 100%.

The rhizobacterial strain B. amyloliquefaciens FZB42 was maintained on Luria-Bertani medium plates. A single colony from a freshly streaked plate was used to grow overnight cultures that were adjusted to a final density of OD₆₀₀ = 0.5 (10⁶ colony forming units/mL). Root inoculation of FZB42 was performed by pipetting 5 ml of the bacterial suspension onto the roots of 6-week-old N. benthamiana plants. Controls were root-inoculated with sterile distilled water. To assess the effect of FZB42 and P. nicotianae on stomata, N. benthamiana plants were co-inoculated concurrently by FZB42 on the roots and P. nicotianae on the leaves.

Stomatal apertures in leaves of N. benthamiana plants inoculated with P. nicotianae, B. amyloliquefaciens FZB42, or co-inoculated with both were measured as described by Kumar et al. (2012). Images of stomatal apertures on epidermal strips were recorded (Olympus BX43 microscope [Tokyo, Japan] and cellSens Standard Software). At least 50 randomly selected stomatal apertures were measured in each treatment, and each assay was repeated three times.

To determine ABA content (Fotopoulos et al., 2008), 1 g of freeze-dried, homogenized leaf tissue was extracted in 80% (v/v) methanol and stirred overnight at 4°C. After centrifuging twice at 4000 ×g for 20 min, the extracts were completely evaporated under vacuum and dissolved in water at pH 3.0. The solutions were partitioned with diethyl ether three times and then passed through anhydrous sodium sulfate. After evaporation of the apolar phase, the dry residue was dissolved in Tris-buffered saline (150 mM NaCl, 1 mM MgCl₂, 50 mM Tris–HCl, pH 7.8), and the ABA was detected immunologically using a Phytodetek Kit (Agdia, Elkhart, IN, United States).

Free and SA-conjugated phytohormones were extracted from flash-frozen leaf tissue (1 g) and quantified according to Schuhegger et al. (2006). SA was detected using a Shimadzu RF 535 fluorescence detector at excitation and emission wavelengths of 305 and 407 nm, respectively.

For JA determination, leaves were flash-frozen in liquid nitrogen and tissue (1 g) was processed as described by Mueller and Brodschelm (1994). JA was quantified by gas chromatography-mass spectrometry (GC-MS; SSQ quadrupole instrument; Finnigan, United States) in negative ion chemical ionization mode with isobutane as the reactant gas. Dihydro JA-pentafluorobenzyl (PFB; m/z = 211) and [molecular anions-PFB]-ions of JA-PFB (m/z = 209) were monitored. JA levels were calculated from the GC peak areas of the selected ions.

The concentration of ET was determined from flash-frozen leaf tissue (1 g) according to the method of De Laat and Van Loon (1983).

Total RNA was extracted from leaves of N. benthamiana according to the method of Wu et al. (2017) using a Plant RNA Kit (Omega Bio-Tek, United States). First-strand cDNA was synthesized using reverse transcriptase (TaKaRa Bio Inc., Dalian, China) and oligo(dT) primers. Reverse transcription PCR (RT-PCR) products were examined by agarose gel electrophoresis. Quantitative RT-PCR (qRT-PCR) was performed using SYBR Premix Ex Taq (TaKaRa) and an ABI 7500 Fast Real-Time PCR Detection System (Applied Biosystems, United States). PCR reactions were heated to 95°C for 3 min, followed by 30 cycles of 95°C for 30 s, 60°C for 30 s, and 60°C for 30 s. Gene expression in each sample was normalized to expression of N. benthamiana EF-1α, and relative expression levels calculated by the 2^-ΔΔC_T method (Livak and Schmittgen, 2001). Gene-specific PCR primers shown in Table 1 are same for RT-PCR and qRT-PCR analysis.

At least five replicates were performed for each experiment. Data were evaluated using one-way analysis of variance and Fisher’s least significant difference tests in SPSS software v. 16.0 (Chicago, IL, United States).

Bacillus amyloliquefaciens FZB42 was tested for its capacity to trigger resistance to the oomycete pathogen P. nicotianae. N. benthamiana roots were inoculated with a suspension of B. amyloliquefaciens FZB42 cells. Leaves were inoculated with P. nicotianae. Disease symptoms were assessed 48 h after P. nicotianae infection by comparing the sizes of the lesions. Plants inoculated with P. nicotianae showed typical symptoms of Phytophthora infection; the leaves were water-soaked at 48 h post-inoculation. However, inoculation of roots with B. amyloliquefaciens FZB42 resulted in a significant reduction (P < 0.01, n ≥ 5) in the lesion size compared with control plants treated with P. nicotianae alone; the inhibition rate of FZB42 in controlling the leaf disease caused by P. nicotianae was 60.09% (Figure 1), which suggests that pretreatment with FZB42 can provide enhanced resistance to P. nicotianae.

Previous studies have shown that root-associated B. subtilis restricts the stomate-mediated entry of the foliar pathogen P. syringae pv. tomato DC3000 in A. thaliana (Kumar et al., 2012). To investigate whether stomatal closure contributes to FZB42-mediated defense, roots of N. benthamiana were inoculated with FZB42 and stomatal apertures were subsequently measured by microscopic evaluation of freshly prepared epidermal peels. We observed that root inoculation with B. amyloliquefaciens FZB42 resulted in a decrease in the mean size of the stomatal aperture 3 and 9 h post-inoculation, while stomata in the control leaves remained open over the entire investigation period (Figures 2A,B). P. nicotianae triggered stomatal closure 3 h after inoculation, and that the stomata reopened by 9 h post-inoculation (Figures 2A,B). Because both FZB42 and P. nicotianae can influence the stomata, stomatal apertures were evaluated after co-inoculating N. benthamiana plants with FZB42 on the roots and P. nicotianae on the leaves. At 3 h post-inoculation, FZB42 and P. nicotianae had no significant difference on reducing the stomatal apertures compared with either FZB42 or P. nicotianae inoculation alone. Surprisingly, at 9 h post co-inoculation, B. amyloliquefaciens FZB42 prevented the reopening of stomata by disrupting the effect of P. nicotianae infection (Figures 2A,B).

Since FZB42 treatment can specifically keep stomata closed, we root-inoculated N. benthamiana plants with FZB42 and examined the sizes of stomatal apertures over a 144-h time course. Interestingly, we found that closure of stomata induced by inoculation of roots with B. amyloliquefaciens FZB42 was only a transient response by the plant. The stomatal aperture sizes decreased to ∼1 μm by 3 h after inoculation with strain FZB42, but then there was a gradual increase in the stomatal aperture sizes starting around 24 h (Figure 2C). At 120 h post-inoculation, the sizes of the stomatal apertures were similar to those observed in the control. These data indicated that FZB42-mediated stomatal closure in N. benthamiana was transient.

Abscisic acid plays a major role in closure of stomata in response to water stress and pathogen challenge (Acharya and Assmann, 2009; Sawinski et al., 2013; Su et al., 2017). Therefore, to investigate the involvement of ABA in B. amyloliquefaciens FZB42-mediated closure of stomata, we measured the stomatal aperture sizes in N. benthamiana abi1 plants that have a dominant negative mutation in a phosphatase, which impairs ABA transduction during stomatal regulation (Leyman et al., 2000). In the abi1 mutants, FZB42-mediated closure of stomata was disrupted 3, 9, and 24 h after the addition of B. amyloliquefaciens (Figure 3A); thus, ABA is required for FZB42-induced stomatal closure. Simultaneously, we measured the mRNA levels of the ABA biosynthetic gene nced1, which encodes a 9-cis-epoxycarotenoid dioxygenase that catalyzes the key step in ABA biosynthesis (Tan et al., 1997), and the ABA contents in the leaves of wild type plants at 3, 9, and 24 h after root inoculation with FZB42. There was a significant increase in ABA content (Figure 3B), and an increase in transcription of the nced1 gene (Supplementary Figure S1) compared to non-inoculated control. Our data also showed a marked reduction in ABA content in N. benthamiana infected by P. nicotianae, while the ABA content remained high during co-inoculation with P. nicotianae and FZB42 (Figure 3B).

Salicylic acid signaling acts upstream of ABA signaling in B. subtilis-triggered stomatal closure in A. thaliana (Zeng and He, 2010; Kumar et al., 2012). Therefore, we determined if SA is also involved in FZB42-mediated closure of stomata in N. benthamiana. To achieve this, after root inoculation with FZB42 we measured the stomatal apertures in transgenic plants expressing the bacterial nahG gene; these plants do not accumulate SA. NahG plants showed disrupted FZB42-mediated stomatal closure 3, 9, and 24 h after root inoculation (Figure 3A). We also measured the levels of free and conjugated SA in wild type leaves, and the transcriptional levels of the ICS1 gene (which is involved in biosynthesis of SA) (Zhu et al., 2014). The SA content was significantly elevated (Figure 3C and Supplementary Figure S2), and ICS1 expression was upregulated after FZB42 treatment compared to control or P. nicotianae treatment (Supplementary Figure S3). Collectively, these findings indicate that SA and ABA are both involved in FZB42-mediated stomatal closure in response to P. nicotianae.

To investigate the potential signal transduction pathways involved in FZB42-mediated resistance, we analyzed the expression of the SA-responsive gene PR-1a, the JA synthesis-related gene LOX, and the ET-responsive gene ERF1 in the leaves of wild type N. benthamiana plants in response to FZB42 treatment alone, to P. nicotianae inoculation alone, and to FZB42 treatment combined with P. nicotianae inoculation. At 24 h following treatment with FZB42, expression of the PR-1a, LOX and ERF1 genes was evident in wild type plants (Figure 4A). We observed similar gene expression patterns in N. benthamiana plants treated with B. amyloliquefaciens FZB42 then challenged with P. nicotianae. In plants inoculated with P. nicotianae alone, expression of the three marker genes was either very low or not detectable (Figure 4A). Results from transgenic NahG, JA signaling-related gene COI1-silenced and ET signaling-related gene EIN2-silenced N. benthamiana plants (Shibata et al., 2010) also have stated this (Figure 4A and Supplementary Figure S4). qRT-RCR analysis also showed that transcription of the PR-1a, LOX, and ERF1 genes was significantly upregulated after treatment with FZB42 and P. nicotianae (Supplementary Figure S5). Meanwhile, after treatment with B. amyloliquefaciens FZB42, the free and conjugated SA contents were elevated between four- and sixfold (Figure 3C and Supplementary Figure S2), and the JA and ET contents increased five and three-fold, respectively (Figure 4B). On the other hand, reduced resistance of transgenic NahG, COI1- and EIN2-silenced plants was scored by visible development of disease symptoms and inhibition rate (Supplementary Table S1). These data show that SA-and JA/ET- dependent signaling pathways were important in plant defenses against P. nicotianae.