Two tobacco species, N. benthamiana and N. tabacum, were grown in 32 cell seedling trays filled with sterile commercial soil (Baroker, Seoul Bio Co., Ltd., Eumseong, Korea) in a growth chamber at 26°C with a photoperiod of 14 h of light and 10 h of darkness. Five- to six-week-old N. benthamiana plants were used for virulence tests via syringe infiltration, stem inoculation, and spray inoculation. N. benthamiana plants at the four- to six-leaf stage (around 3 weeks old) were used for virulence tests via root-dip inoculation. Seven- to eight-week-old N. tabacum plants were used for virulence and bacterial growth tests on the leaves via syringe infiltration.

C. michiganensis type strain LMG7333^T (Hwang et al., 2019), C. capsici type strain PF008^T (Oh et al., 2016), and A. citrulli strain Ac8 (Song et al., 2020) were cultured at 26°C for 48 h in the King’s B medium (20 g/l of protease peptone number 3, 1.5 g/l of K₂HPO₄, 6 ml/l of 1 M MgSO₄, and 16 ml/l of 50% glycerol). Tn::celA (Hwang et al., 2019) and Tn::pat-1 mutants (Hwang et al., 2022) of C. michiganensis LMG7333^T were streaked onto the King’s B medium with 10 mg/l of chloramphenicol at 26°C for 48 h.

For the leaf infiltration, five- to six-week-old N. benthamiana and seven- to eight-week-old N. tabacum plant leaves were infiltrated with either sterilized distilled water (mock) or 5×10⁴ CFU/ml bacterial suspension or approximately 10⁸ CFU/ml (OD₆₀₀ = 0.1) bacterial suspension using a needleless syringe; at least three leaves were infiltrated per treatment. All inoculated plants were placed in a growth chamber at 26°C and 50% humidity. The development of blister-like symptoms and necrosis was observed for 8 days and 48 hours, respectively. This assay was repeated at least three times.

For stem inoculation, five- to six-week-old N. benthamiana plants were injected with 20 ul of 10 mM MgCl₂ (mock) or 10⁸ CFU/ml (OD₆₀₀ = 0.1) bacterial suspension after wounding the stems with a pipette tip. All inoculated plants were placed in a growth chamber at 26°C and 50% humidity. Wilting development was observed for 3 weeks. Wilted leaves were counted and compared with the number of fully grown leaves above the inoculation site. The ratio between wilted leaves and fully grown leaves was calculated by percentage. Bacterial canker symptoms on stem inoculation sites were observed at 0, 5, 10, and 15 days after inoculation (dai), and their sizes were measured as lengths of their longitudinal crack using a ruler. This assay was repeated at least three times.

For the root-dip inoculation, N. benthamiana plants in the four- to six-leaf stage were pulled out and dipped in the bacterial suspension (approximately 10⁹ CFU/ml, OD₆₀₀ = 2.0) for 30 or 40 min, depending on plant size. Then, the inoculated plants were replanted into soil in a new tray. All inoculated plants were placed in a growth chamber at 26°C and 50% humidity. Wilting development was observed for 3 weeks. Root-inoculated N. benthamiana plants were analyzed by the number of plants displaying wilting using five categories defining the severity of wilting symptoms: (1) 0 = no visible wilting; (2) 1 = 1–25% wilting symptoms; (3) 2 = 26–50% wilting symptoms; (4) 3 = 51–75% wilting symptoms; (5) 4 = 76–100% wilting symptoms or dead at around 18 dai. The disease index used in this study was described in previous research (Shinohara et al., 2005). This assay was repeated at least three times.

For the spray inoculation, N. benthamiana plants were sprayed with 50 ml of bacterial suspensions (approximately 5×10⁶ CFU/ml, OD₆₀₀ = 0.005 and 10⁸ CFU/ml, OD₆₀₀ = 0.1) containing 0.02% Silwet L−77 (MOMENTIVE Co., Ltd., Seoul, Korea). Wilting development was observed for more than 3 weeks. All inoculated plants were placed in a growth chamber at 26°C and 50% humidity. This assay was repeated twice.

To measure the bacterial growth inside both N. benthamiana and N. tabacum leaves after syringe infiltration with 5×10⁴ CFU/ml bacterial suspension, three-leaf disks (1 cm in diameter) were collected from the infiltrated leaves of at least three different plants using a cork borer at 0, 2, 4, 6, and 8 dai. The collected leaf disks were washed with 70% ethanol for 30 s for surface sterilization and were rinsed with sterilized distilled water twice. Next, washed leaf disc samples were ground by vortexing them in 2 ml microtubes with two iron beads and 1 ml of sterilized distilled water. The homogenate was serially diluted and put onto a KB agar plate. The numbers of colony-forming units (CFU) were calculated after 48–72 h at 26°C.

For quantification of necrosis in N. benthamiana leaves, ion conductivity was measured as described in the previous study (Choi et al., 2012). At least 4 leaves selected from different plants were infiltrated with 10⁸ CFU/ml (OD₆₀₀ = 0.1) bacterial suspension using needleless syringe. After drying drenched leaves completely for 60 or 90 min, six-leaf disks (1 cm in diameter) were taken from each dried leaf using a cork borer. The leaf disks were moved to a 50 ml snap tube with 25 ml of sterilized distilled water to remove the initially leaked ion until 3 h after inoculation (hai). Then, water inside a 50 ml snap tube was removed completely, except for leaf disks, and 25 ml of sterilized distilled water was added to the snap tube again. Ion conductivity was measured using a conductivity meter (Acorn CON6 portable conductivity meter, Oakton Instrument, Vernon Hills, IL, USA) at 3, 12, 24, 36, and 48 hai. The snap tubes with leaf disks were incubated on the shaker at 60 rpm. Four of 50 ml snap tubes with six-leaf disks and 25 ml sterilized distilled water were used for each bacterial strain. The experiments were repeated at least twice.

To quantify the HR in N. tabacum leaves, six-leaf disks (1 cm in diameter) were collected from the infiltrated leaves with each bacterial strain at 3, 9, 15, and 21 hai. Leaf disks were moved to a 50 ml snap tube with 25 ml of sterilized distilled water and were washed for an hour. Then, the same volume of water was added to the snap tube. Next, those tubes were incubated for an hour in the shaking incubator at 26°C and 180 rpm, and ion conductivity was measured using a conductivity meter. The experiments were repeated at least twice.

The trypan blue staining procedure was modified from a previous study (Guo et al., 2019). Briefly, the leaf disks (1 cm in diameter) from leaves of at least 3 different N. benthamiana plants infiltrated using a syringe with 5×10⁴ CFU/ml of C. michiganensis LMG7333^T and C. capsici PF008^T were taken using a cork borer at 0, 3, 5, and 7 dai. Sterilized distilled water was used for mock treatment. Leaf disks were treated with trypan blue to stain damaged or dead cells. A 3 ml trypan blue solution (10 ml lactic acid, 10 g phenol, 10 ml glycerol, 10 ml distilled water, and 10 mg trypan blue) was added into a 5 ml tube with leaf disks. One day later, the trypan blue solution was exchanged with 4 ml of absolute ethanol to remove unstained trypan blue. After 1 day of ethanol bleaching, the absolute ethanol in the 5 ml tube was exchanged with 70% ethanol for storage until the time that leaf disks were observed using a microscope. Leaf disks were observed using an optical microscope (ECLIPSE E200LED MV R, Nikon Corporation, Tokyo, Japan) at a magnification x40.

The DAB staining method was followed and modified from a previous study (Shi et al., 2014). Briefly, leaf disks (1 cm in diameter) were collected from at least three different N. benthamiana plants infiltrated with 5×10⁴ CFU/ml of C. michiganensis LMG7333^T and C. capsici PF008^T using a cork borer at 0, 1, 3, 5, and 7 dai. Sterilized distilled water was used for mock treatment. The collected leaf disks were placed in the petri dish and stained with 10 ml of 0.1% DAB solution for 1 day. Next, DAB solution was removed from the petri dish, and 10 ml of absolute ethanol was added to the petri dish to bleach the leaf disk chlorophyll. After 1 day of ethanol bleaching, absolute ethanol in the petri dish was exchanged with 70% ethanol for storage until the time that leaf disks were observed using a microscope. All DAB-stained N. benthamiana leaf disks were observed using an optical microscope at a magnification x40.

PCR analysis was conducted for the detection of Clavibacter bacteria in the plants after inoculation. PCR amplification proceeded according to the manufacturer’s manual for 2x Taqbasic PCR Master Mix 2 (Biofact, Daejeon, Korea). The total volume was 20 uL and consisted of 10 ul of 2X Taqbasic PCR Master Mix 2, 1 ul of template, 1ul of forward and reverse primers (10 pmol/ul), and 7 ul of sterilized distilled water.

Several specific primer pairs were used as follows. Primer pair CMR16S F (5′-gtgatgtcagagcttcctctggcggat-3) and CMR16S R (5′-gtacggctaccttgttacgacttagt-3′) were used for C. michiganensis LMG7333^T and C. capsici PF008^T, because they specifically target the 16S-rRNA gene of Clavibacter species (Yim et al., 2012). For confirmation of C. michiganensis Tn::celA and Tn::pat-1 mutant strains, the H1 (5′-atgacatttcgccaagttcgtgca-3′) and H2 (5′-tcagtgcacagggtagaagcg-3′) primer pair was used for celA gene. The EB15 (5′- actagtagaacgctccctgcggccttcg-3′) and EB16 (5′- aagcttacttgtcgtcatcgtctttgtagtcggaggtcgctaatatgtaatacggt-3′) primer pair was used for pat-1 gene.

Bacterial growth was expressed as mean and standard deviations. The disease index of stem and root inoculation was calculated as the mean and standard error. Disease severity data were analyzed via non-parametric Kruskal–Wallis test (p < 0.05). Duncan’s multiple range test was conducted for statistical analysis of the parametric comparison between independent groups (p < 0.05). All experiments were repeated at least two or three times.

To investigate whether Gram-positive Clavibacter bacteria can cause any disease symptoms in model plants of Nicotiana species, both C. michiganensis LMG7333^T and C. capsici PF008^T were infiltrated into leaves of two Nicotiana species, N. benthamiana and N. tabacum. N. benthamiana leaves showed white granular particles, called blister-like lesions, when 5×10⁴ CFU/ml of each bacterial species was infiltrated (Figure 1A). These lesions were generated on the leaf surface around 4 dai. The formation of blisters was originally reported in tomato, the host plant of C. michiganensis (Chalupowicz et al., 2017), and blister-like lesions on N. benthamiana appeared phenotypically similar to those in tomato leaves. Conversely, N. tabacum leaves did not display any disease symptoms by infiltration with the same titer of either C. michiganensis LMG7333^T or C. capsici PF008^T suspension (Figure 1B). In N. benthamiana leaves, the infiltrated bacterial pathogens grew significantly up to 10⁶-fold more than the initial inoculum until 8 dai (Figure 1C); an increase was more than that by the known pathogen A. citrulli. However, the bacterial titer of both Clavibacter species increased only 10²- to10³-fold more in N. tabacum leaves from the initial inoculum without any symptoms, and this increase was higher than that by A. citrulli (Figure 1D). Moreover, the bacterial titer of C. capsici was significantly higher than C. michiganensis (Figure 1D). These results indicate that N. benthamiana appears to be a surrogate host plant of two Clavibacter species, whereas N. tabacum appears to be a non-host plant.

To investigate features of blister-like lesions on N. benthamiana leaves after infiltration with C. michiganensis LMG7333^T and C. capsici PF008^T, leaf disks from infiltrated N. benthamiana leaves were collected and treated with trypan blue solution 0, 1, 3, 5, and 7 dai (Figure 2). Dead cells turned to blue after staining, as trypan blue is normally used for observation of such cells. Blue color spots began to appear from 3 dai, and the number and size of those spots were gradually increased (Figure 2). These patterns were exactly matched with blister-like lesions, indicating that plant cells in blister-like lesions are undergoing death.

Generally, ROS generation is accompanied by cell death (Balint-Kurti, 2019). Thus, the 3,3′-diaminobenzidine (DAB) staining was conducted with leaf disks collected from N. benthamiana leaves after infiltration with two Clavibacter species (Supplementary Figure S1). The stained spots by DAB began to appear from 3 dai, and the intensity gradually increased (Supplementary Figure S1). Like trypan blue staining, DAB-stained spots exactly matched with the areas of blister-like lesions, indicating that blister-like lesions are generated by cell death, and accompanied by the generation of ROS, such as hydroxy peroxide.

Previously, it was shown that celA and pat-1 genes of C. michiganensis are critical for disease development in tomatoes (Gartemann et al., 2003; Hwang et al., 2019). Thus, we tried to determine whether these two important virulence genes are required for the development of blister-like lesions on N. benthamiana leaves. For this, Tn::celA and Tn::pat-1 mutant strains of C. michiganensis LMG7333^T were infiltrated into N. benthamiana leaves, and the formation of blister-like lesions was monitored. Intriguingly, when 5×10⁴ CFU/ml of each mutant strain was infiltrated, each caused as many blister-like lesions on leaves as the WT strain (Figure 3A). Moreover, those mutant strains grew as much as the WT strain (Figure 3B). These results indicate that two known important virulence genes are not required for the formation of blister-like lesions in N. benthamiana.

To examine whether different bacterial concentrations cause different symptoms, N. benthamiana leaves were infiltrated with a higher bacterial titer (10⁸ CFU/ml) of C. michiganensis LMG7333^T or C. capsici PF008^T. In this condition, the infiltrated regions by both bacterial pathogens rapidly displayed water-soaking symptoms, followed by typical necrosis by 48 h after inoculation (hai) (Figure 4A). Ion conductivity was measured in infiltrated N. benthamiana leaves to determine the speed of necrosis onset. The ion conductivity began to significantly increase from 12 hai and continuously increased until 36 hai in N. benthamiana leaves infiltrated with both Clavibacter species (Figure 4B). These results indicate that infiltration of N. benthamiana leaves with a high concentration of Clavibacter species causes rapid necrosis without blister-like lesions.

Next, to examine whether celA and pat-1 genes are required for the development of necrosis, 10⁸ CFU/ml of Tn::celA and Tn::pat-1 mutant strains were infiltrated into N. benthamiana leaves. Results showed that these mutants caused necrosis like the WT strain (Figure 4A) and increased ion conductivity as much as the WT (Figure 4B), indicating that these two virulence genes in tomatoes are not critical for necrosis in N. benthamiana.

C. michiganensis causes not only blister-like lesions, but also wilting and canker symptoms in tomatoes (de León et al., 2011). To examine whether both C. michiganensis LMG7333^T and C. capsici PF008^T cause wilting in N. benthamiana, 10⁸ CFU/ml bacterial suspension of two Clavibacter species was injected into the main stems via the stem inoculation method, and wilting development was monitored. Notably, C. michiganensis LMG7333^T caused severe wilting in N. benthamiana, as in tomatoes, whereas C. capsici PF008^T did not (Figure 5A). Wilting symptoms began to develop on the unilateral side of inoculation sites in N. benthamiana. After 3 weeks, severe wilting and necrosis were observed on the whole N. benthamiana plant only by C. michiganensis LMG7333^T (Figure 5B).

Another main disease symptom of Clavibacter species in host plants is bacterial canker. Bacterial canker development on N. benthamiana stems by C. michiganensis infection has been previously reported (Balaji et al., 2011). Consistently, in this experiment, both C. michiganensis LMG7333^T and C. capsici PF008^T caused significant bacterial canker on inoculated N. benthamiana stems (Supplementary Figure S2).

To confirm that wilting was caused by C. michiganensis bacterium, the polymerase chain reaction (PCR) test was conducted for identification of this bacterium using leaf samples that were taken from locations near and above the inoculation site. PCR results showed that C. michiganensis LMG7333^T could move to the entire plant, likely through its xylem vessels, whereas C. capsici PF008^T could not move to nearby leaves nor in an upper direction in N. benthamiana (Figure 5C).

Clavibacter species normally invade the host plants through wounds and natural openings (Carlton et al., 1998; de León et al., 2011; Tancos et al., 2013). Thus, we sprayed a bacterial suspension of both Clavibacter species onto the N. benthamiana plants to mimic the natural invasion process. However, neither bacterial species caused any visible symptoms (Supplementary Figure S3A). PCR results also showed no evidence of the bacterial presence of both Clavibacter species inside plants (Supplementary Figure S3B). These results indicate that even C. michiganensis LMG7333^T fails to actively infect N. benthamiana through stomata.

To examine whether celA and pat-1 genes are required for the development of wilting in the N. benthamiana plant, 10⁸ CFU/ml of C. michiganensis WT, Tn::celA and Tn::pat-1 mutant strains were inoculated using the stem inoculation method, and the development of wilting was monitored. Both mutants caused delayed wilting in N. benthamiana (Figure 6A), and the wilting severity was approximately 70% of that caused by WT C. michiganensis (Figure 6B). The presence of mutant strains in wilting N. benthamiana plants was confirmed via PCR test (Figure 6C). When N. benthamiana was inoculated using the root-dipping method with 10⁹ CFU/ml of the bacterial suspensions, wilting results were similar to those of the stem inoculation method (Supplementary Figure S4). These results indicate that, like tomatoes, celA and pat-1 genes are partially required for wilting development in N. benthamiana.

We showed that no disease symptoms were observed after infiltration of a low bacterial concentration (5×10⁴ CFU/ml) of two Clavibacter species into N. tabacum leaves (Figure 1). To examine whether these Clavibacter species can induce HR-like cell death in N. tabacum, 10⁸ CFU/ml of C. capsici PF008^T and C. michiganensis LMG7333^T were infiltrated. Results showed that both induced HR-like cell death within 18 hai (Figure 7A), and ion conductivity began to increase from 9 hai until 15 hai (Figure 7B). These results indicate that both Clavibacter species can induce HR-like cell death in N. tabacum, presenting another clue that N. tabacum is a non-host plant of these bacteria.

To determine whether celA and pat-1 genes are required for induction of HR-like cell death, mutant strains were infiltrated. The Tn::celA mutant induced HR-like cell death like the WT strain, whereas the Tn::pat-1 mutant failed (Figure 7A). The ion conductivity was consistent with these phenotypes (Figure 7B). These results indicate that the pat-1 gene, but not celA, is critical for HR induction in N. tabacum.