All bacterial strains used in this study are listed in Table 1. Pta 6605 strains, Escherichia coli strains, and Chromobacterium violaceum CV026 were grown as described previously (Ichinose et al., 2020).

An approximately 5.8 kb DNA fragment encoding mexAB-oprM was amplified with a set of primers, mexA5 and oprM3, using KOD FX DNA polymerase (Toyobo, Osaka, Japan) with the genomic DNA of Pta 6605 as a template, then introduced into a pCR-Blunt II-TOPO plasmid vector (Invitrogen, Carlsbad, CA, USA) to obtain pCR-mexAB-oprM (Fig. 1A). To delete the mexB region from the plasmid, inverse PCR was carried out with a set of primers, mexB-d5NheI and mexB-d3NheI, and self-ligated by a Ligation-Convenience kit (Nippon Gene, Tokyo, Japan) after NheI and DpnI digestions. The mutated genetic region was introduced into pK18mobsacB (Schäfer et al., 1994) via an EcoRI site to generate pK18-∆mexB, then introduced into E. coli S17-1 for conjugation with Pta 6605 strains. Deletion mutants were obtained by conjugation and homologous recombination according to the previously reported methods (Taguchi et al., 2006). The DNA sequence of the mutated region in the bacterium was confirmed by DNA sequencing with a Big Dye terminator cycle sequencing kit and an ABI PRISM 3100 sequencer (Thermo Fisher Scientific, Waltham, MA, USA). Primer sequences were designed based on the sequence of Pta 6605 provided by Dr. Studholme, University of Exeter, UK.

Using a previously generated ∆mexF mutant (Sawada et al., 2018) as a recipient, mexB was deleted by a conjugation and homologous recombination as described above, and the ∆mexB∆mexF double mutant was generated.

An overnight culture was centrifuged at 1,500 ×g for 5 min at 4°C, the bacterial pellet was resuspended in LB medium with 10 mM MgCl₂, and the OD₆₆₀ was adjusted to 0.02. Three ml of the bacterial culture was incubated at 27°C with shaking, and the bacterial OD₆₆₀ was measured with an OD meter at each time point.

Potential antibiotics, trans-cinnamic acid, naringenin, acetovanillone, (+)-catechin, catechol, scopoletin, and coumarin were obtained from Tokyo Chemical Industry Co. Ltd. (Tokyo, Japan), phloretin, chlorogenic acid, and p-coumaric acid were obtained from Sigma-Aldrich (St. Louis, MO, USA), and acetosyringone was obtained from Cayman Chemical (Ann Arbor, MI, USA). Phytoalexins, capsidiol, debneyol, and rishitin were purified in the plant pathology laboratory of Nagoya University.

Potential antibiotics and antibiotics were dissolved in sterilized and distilled water or dimethyl sulfoxide. Bacteria (3 ml) cultured overnight in King’s B (KB) medium were mixed with 50 ml of melted 0.6% agar KB medium, then poured into three Petri dishes.

Compound-absorbed paper discs were put on the Petri dish, and the dishes were incubated at 27°C. After 24 h incubation, the growth inhibition zone was photographed and measured using Adobe Photoshop CC version 19.1.9 (Adobe, San Jose, CA, USA). The photographic data was gray-scaled, then noise was removed with a high-pass filter to a middle value of brightness, the image was converted into two gradations, and the growth inhibition zone was measured.

Overnight cultured bacteria were harvested by centrifugation and further incubated in MMMF medium (50 mM potassium phosphate buffer, 7.6 mM (NH4)2SO4, 1.7 mM MgCl₂, and 1.7 mM NaCl, pH 5.7, supplemented with 10 mM each of mannitol and fructose) for an additional 1 h.

Total RNA was extracted from the cells collected by centrifugation using Trizol (Thermo Fisher Scientific). For qRT-PCR, the cDNA was synthesized using 0.5 µg of total RNA and ReverTra Ace qPCR RT Master Mix with gDNA Remover (Toyobo) according to the manufacturer’s instructions. The specific primer sets for mexA (mexA-F1 and mexA-R1), mexE (mexE-F1 and mexE-R1), and pcaG (pcaG-F and pcaG-R) were designed according to the registered sequences of Pta 6605 (Table 2). Real-time qRT-PCR was carried out using a GVP-9600 (Shimadzu, Kyoto, Japan) and KAPA SYBR FAST qPCR Master Mix (2×) Kit (Sigma-Aldrich), with primers at a final concentration of 0.4 µM each and 1 µl of cDNA product as a template in a 20 µl reaction mixture. PCR was performed with one denaturation cycle of 20 s at 95°C and 40 cycles of 15 s at 95°C and 30 s at 62°C. The expression of each gene was normalized by pcaG, which encodes protocatechuate 3,4-dioxygenase subunit alpha, as a stably expressed gene (Taguchi et al., 2015). Melting curve analysis confirmed that each PCR was a single product. The PCRs were performed with a high linearity. The data were analyzed using the ∆∆CT (cycle threshold) method.

An inoculation test of Pta 6605 was conducted with a flood assay (Ishiga et al., 2011) optimized for tobacco plants (Nicotiana tabacum L. cv. Xanthi). Tobacco seeds were sterilized and sown on 0.8% Murashige-Skoog (MS) agar plates containing 1% sucrose and vitamin stock solution (thiamin hydrochloride 3 mg/l, nicotinic acid 5 mg/l, pyridoxine hydrochloride 0.5 mg/l), and grown at 28°C under 16 h light:8 h dark conditions for 2 weeks. Tobacco seedlings were transplanted to 0.8% MS agar plates containing 0.1% sucrose and vitamin stock solution as above and grown for 2 days under the same conditions. Bacteria were grown overnight at 27°C in KB medium with 50 μg/ml of nalidixic acid (Nal). The bacterial inoculum was adjusted to OD₆₀₀ = 0.002 with sterilized water containing 0.025% Silwet L-77 (Biomedical Sciences, Tokyo, Japan). Sterilized water containing 0.025% Silwet L-77 was used as a mock treatment. The inoculation test was carried out on a clean bench. The bacterial suspension (approximately 30 ml) was poured into the plate of tobacco seedlings. After about 10 s incubation, the bacterial suspension was decanted, and the plate was air-dried on a clean bench for 15 min. The plants were incubated under 16 h light:8 h dark conditions at 22°C, and disease symptoms were observed for 4 days post-inoculation (dpi) and 10 dpi. To determine the bacterial population at 0 dpi and 4 dpi, leaf disks were punched out using a disposable biopsy hole punch and then ground with mortar and pestle. The homogenates were serially diluted in sterile distilled water and then spread on KB plates containing Nal. The plates were dried and incubated at 27°C for 2 days, after which the bacterial population

was measured by counting the number of colonies (colony forming unit).

Swimming and swarming assays were performed as described previously (Taguchi and Ichinose, 2011). An overnight culture with 50 µg/ml of Nal was centrifuged at 1,500 ×g for 10 min, the bacterial pellet was resuspended in 10 mM MgSO₄, and the OD₆₆₀ was adjusted to 1.0. For the swimming assay, 3 μl of bacterial suspension was inoculated on the center of 0.25% Bactoagar (Becton Dickinson & Company, Glencoe, MD, USA) (w/v) MMMF plates. For the swarming assay, 3 μl of the bacterial suspension was spotted on the center of a 0.45% Bactoagar (w/v) SWM (0.5% peptone and 0.3% yeast extract) plate and allowed to dry. Plates were incubated overnight at 23°C. Swimming and swarming motilities were observed after 24 h incubations.

Bacterial strains were grown in KB medium for 24 h at 27°C. AHLs were extracted with an equal volume of ethyl acetate and detected using C₁₈ reversed-phase thin layer chromatography (TLC Silica gel 60, Merck, Darmstadt, Germany) and the biosensor C. violaceum CV026 (Taguchi et al., 2006) as described in Ichinose et al. (2020).

Statistical analysis was performed using One-way ANOVA Dunnett’s test and Tukey honestly significant difference test with R software version 3.6.1 (https://www.r-project.org/).

The specific deletion of mexB and/or mexF was confirmed by genomic PCR (data not shown). The growth of each bacterial strain in in vitro culture was compared. There was no significant difference in growth speed between the WT strain and the mutant strains (Fig. 1B).

Growth inhibition assays of the plant-derived potential antimicrobial compounds in KB soft agar plates showed that 5 µl of 0.9 M trans-cinnamic acid, 0.5 M p-coumaric acid, 1 M acetosyringone, 1 M scopoletin, 1 M chlorogenic acid, 0.3 M phloretin, 1 M naringenin, and 200 mg/ml (+)-catechin did not produce any clear growth inhibition zone in all tested strains (data not shown). Furthermore, 5 µl of 500

µM capsidiol, debneyol, and rishitin, phytoalexins of Nicotiana tabacum, Nicotiana debneyi, and Solanum tuberosum, respectively, also did not inhibit any bacterial growth (Supplementary Fig. 1). When the concentration of capsidiol was increased to 10 mM, no growth inhibition was observed (Supplementary Fig. 1). This concentration of phytoalexins completely inhibits the growth of oomycetes and fungal pathogens (Egea et al., 1996). However, 5 µl of 1 M catechol and 1 M coumarin inhibited the growth of the Pta 6605 WT strain (Fig. 2A and B). Furthermore, the size of the growth inhibition zone by 1 M catechol increased in ∆mexF, and further expanded in ∆mexB and ∆mexB∆mexF. The ∆mexB and ∆mexB∆mexF mutant strains showed a slightly higher sensitivity than the WT strain to 1 M coumarin; however, the ∆mexF mutant produced a smaller inhibition zone than the WT strain. Although 5 µl of 2 M acetovanillone did not inhibit growth of the WT strain, it inhibited those of ∆mexB and ∆mexF, and an additive growth inhibition effect was observed in the ∆mexB∆mexF mutant (Fig. 2C).

The sensitivity of bacterial strains to some bacteria-derived antibiotics was also investigated. The sensitivities to most of the antibiotics used in this study, such as kanamycin, spectinomycin, tetracycline, streptomycin, and hygromycin B, did not differ among the bacterial strains (data not shown). However, only ∆mexB and ∆mexB∆mexF were sensitive to 5 µl of 50 mg/ml ampicillin, while the growths of the WT strain and ∆mexF were not inhibited by it (Fig. 2D). On the contrary, although the WT and ∆mexB mutant strains showed only weak growth inhibition zone to 5 µl of 50 mg/ml chloramphenicol, the growth of ∆mexF and ∆mexB∆mexF mutant strains was remarkably inhibited (Fig. 2E). These results indicate that MexAB-OprM and MexEF-OprN specifically extrude ampicillin and chloramphenicol, respectively.

The effect of the deletion of mexB, mexF, or both mexB/mexF on mexA and mexE gene expression was analyzed by real-time qRT-PCR. As shown in Fig. 3A, the expression of mexA was enhanced in the ∆mexF and ∆mexB∆mexF strains. In addition, the expression of mexE was very weak, and seemed to be slightly increased in the ∆mexB strain (Fig. 3B).

Surface swimming and swarming motilities of the ∆mexB mutant were comparable to WT strain, whereas those of ∆mexF and ∆mexB∆mexF double mutant strains were abolished (Fig. 4A).

Pta 6605 utilizes AHLs as quorum-sensing molecules. AHLs production in ∆mexB and ∆mexB∆mexF mutants was reduced. However, the ∆mexF mutant retained AHL production ability as good as that of the WT strain (Fig. 4B).

To investigate whether MexAB-OprM and/or MexEF-OprN of Pta 6605 is involved in the virulence, a flood inoculation test of the WT, ∆mexB, ∆mexF, ∆mexB∆mexF, and ∆hrcC mutant strains (Marutani et al., 2005) on host tobacco seedlings was performed. The ∆hrcC mutant did not show any disease symptoms, and bacterial growth was limited (Fig. 5). On the other hand, all mutants except ∆hrcC showed disease symptoms similar to those of the WT strain at 4 dpi (Fig. 5A) and 10 dpi (Supplementary Fig. 2). However, the bacterial populations in the leaves of ∆mexB, ∆mexF, and ∆mexB∆mexF mutant strains were significantly lower than that of WT at 4 dpi (Fig. 5B), indicating that MexAB-OprM and MexEF-OprN might partially contribute to virulence at the early stage of infection.