Ralstonia solanacearum strains used in this study were listed in Table 1. OE1-1 strain (phylotype I, race 1, biovar 3) is virulent in tomato and tobacco plants (Kanda et al., 2003). RS1002 (Mukaihara et al., 2004) and GMI1000 (Salanoubat et al., 2002) strain (phylotype I, race 1, biovar 4) are virulent in tomato plants but elicit hypersensitive response (HR) in tobacco leaves. R. solanacearum was grown at 28°C in rich B medium or in hrp-inducing medium (hydroponic plant culture medium containing 2% sucrose) (Yoshimochi et al., 2009b). Escherichia coli DH12S and S17-1 (Simon et al., 1983) were grown in Luria-Bertani (LB) medium at 37°C and used for plasmid construction and conjugational transfer, respectively. Antibiotics were added into the media at the following concentrations: ampicillin (Ap), 100 μg ml^-1; gentamicin (Gm), 20 μg ml^-1; kanamycin (Km), 50 μg ml^-1; polymyxin B (PB), 50 μg ml^-1.

In this study, the npdA in-frame deletion mutants were generated by pK18mobsacB- based homolog recombination as previously described (Zhang et al., 2015). Primer pairs 0316A1B (CATGGATCCATCAACCGTTCGAAGGAGA) with 0316B1C (TCAACGGAGACTTGTGCCCCACCCGTTGCATCCGCG) and 0316A2C (CGCGGATGCAACGGGTGGGGCACAAGTCTCCGTTGA) with 0316B2H (CATAAGCTTCCATCTGGACCTTCAGGT) were used for plasmid pK18dnpdA construction with the OE1-1 genomic DNA as PCR template. After validating the sequence, pK18dnpdA was transferred from E. coli S17-1 into the corresponding R. solanacearum strains, and npdA mutants were generated (listed in Table 1).

The complementation was performed by using the pUC18-mini-Tn7T-Gm-mediated site-specific chromosome integration system (Choi et al., 2005) with some modification (Zhang et al., 2015). Primers pairs 0316A1B and 0316B3H (CATAAGCTTCAGGCCGTCGTCGCCGAAG) were used for the pUCnpdA construction with the OE1-1 genomic DNA as the template, which contained an approximately 500 bp-upstream region of npdA and empirically harbored its native promoter. The complemented npdA was specifically integrated into R. solanacearum chromosome at a single attTn7 site (25-bp downstream of the glmS gene) and was confirmed by colony PCR.

The gene expression level was evaluated by a lacZYA fusion-based β-galactosidase assay, and the assay (both in vitro and in planta) was performed as previously described (Zhang et al., 2013). Each assay was repeated for at least three independent experiments, and the averages with their respective standard deviation (SD) were determined. The statistical significance was assessed using the two-tailed Student’s t-test.

Total RNA was isolated from the hrp-inducing medium cultured R. solanacearum cells using the TRIzol reagent according to the manufacturer’s instruction (Life, United States). After validating the RNA quality, cDNA was synthesized using the PrimeScript^TM RT Reagent Kit with gDNA Eraser (Perfect for Real Time, TAKARA, Japan). The One Step SYBR^® PrimeScript^TM PLUS RT-PCR Kit (TAKARA, Japan) was used for the qRT-PCR reactions with an Applied Biosystems 7500 Real-Time PCR System (Life, United States) to quantify the cDNA level. In this study, RipB, RipD, RipE, RipO, RipR, RipTAL, and RipW were selected for mRNA quantification by qRT-PCR and the serC gene was selected for reference (Monteiro et al., 2012). The primers used in this study were used as previously described (Wu et al., 2015). The experiments were repeated from RNA isolation for at least three independent replicates, and each treatment included four replications. The mean values were averaged with the standard deviation (SD), and statistical significance was determined using the two-tailed Student’s t-test.

The virulence assay and HR test were performed as described previously (Yao and Allen, 2007) with some modifications (Zhang et al., 2013). Wilt-susceptible tomato plants (Solanum lycopersicum cv. Moneymaker) and tobacco plants (Nicotiana tabacum CV. Bright Yellow) were used for the virulence test by soil soak (with bacteria at a final concentration of 10⁷ cfu g^-1 of soil) and petiole inoculation (dropped 2 μl of bacteria at 10⁷ cfu ml^-1 onto the fresh-cut surface of petioles). Each experiment included at least 10 plants per treatment, and each assay was repeated at least in triplicate. Mean values were manipulated with standard deviation (SD), and their statistical significance was determined using the two-tailed Student’s t-test. N. tabacum BY was used for the HR test with leaf infiltration, and the symptom development of HR was recorded periodically. Each experiment was repeated for at least three independent replicates, and a representative result was presented.

Bacterial growth in planta was performed as described previously (Zhang et al., 2013). Two to 3-week-old tomato plants were inoculated with bacteria using petiole inoculation, and the stem tissues (5 cm above the soil and 1 cm in length) were removed at 4 dpi (wilt symptoms reached to 1–2) and 7 dpi (wilt symptoms reached to 3–4). The cells were collected from the plant stem tissues and quantified by dilution plating. Each experiment was repeated for at least three independent replicates, and the mean values and SD values were determined. The statistical significance was assessed using the two-tailed Student’s t-test.

The T3SS expression in OE1-1 was monitored with a popA-lacZYA fusion in our study. The popA expression in npdA transposon mutant was reduced to approximately 30% of that in the wild-type strain, suggesting that NpdA might be required for the T3SS expression in R. solanacearum OE1-1 (Zhang et al., 2013). To confirm this phenomenon, we created the npdA in-frame deletion mutant RK5628 (popA-lacZYA,ΔnpdA) in OE1-1 strain and evaluated its popA expression level. First, no differences in the bacterial morphology and growth profiles were observed between the npdA mutants and wild-type strain in rich or hrp-inducing media (data not shown). Similar to that of the wild-type strain in rich medium, RK5628 did not express popA. However, in hrp-inducing medium, the popA expression in RK5628 was substantially reduced (an average of 74 vs. 313 Miller Units) (Figure 1). Additionally, the complemented npdA gene (integrated into the chromosome downstream of the glmS gene) could completely restore its popA expression to that of wild type in hrp-inducing medium (Figure 1), confirming that NpdA is an important factor for popA expression in R. solanacearum OE1-1. The strain RS1002 and GMI1000 exhibit distinctly different phenotype from OE1-1, and hence we generated the npdA deletion mutants from RS1002 and GMI1000. The popA expression was also substantially reduced in RS10013 (npdA deletion from RS1002) (Figure 1) and GF0021 (npdA deletion from GMI1000) (data not shown).

NpdA is highly conserved in R. solanacearum species, and through NCBI blast, the identities of NpdAs amino acids between 28 R. solanacearum strains are greater than 93%. For instance, there is only one residue different from the total 335 amino acids of NpdAs between OE1-1 and GMI1000. Furthermore, the npdA gene of OE1-1 could fully complement the popA expression in RS10013 to that of the wild-type strain (Figure 1), suggesting that the NpdA-dependent expression of popA is highly conserved in R. solanacearum species.

The popA gene exists as part of an operon with popB and popC, and belongs to T3Es. Since NpdA is important for the popA expression, we further investigated whether it was required for the expression of some other T3Es in OE1-1. In this study, RipB, RipD, RipE, RipO, RipR, RipTAL, and RipW were selected for mRNA quantification by qRT-PCR. Among these T3Es, the mRNA levels of RipB, RipD, RipE, and RipR in hrp-inducing medium were significantly reduced in RK5628 (p < 0.01), and the mRNA level of RipW was also distinctly reduced (Figure 2), (p < 0.05). However, the npdA deletion did not change the mRNA levels of RipO and RipTAL, suggesting that the NpdA is also important for the expression of a subset of T3Es in R. solanacearum OE1-1.

The T3SS and some T3Es are important for the pathogenicity of R. solanacearum in host plants. Thus, we evaluated whether NpdA was required for the virulence of R. solanacearum. When tomato plants were challenged with soil-soak and petiole inoculation, RK5628 could wilt and kill most of the test plants, while approximately 25% of the test plants did not die even up to 18 dpi (days post inoculation). For example, 22 of the 96 total soil-soak inoculated tomato plants began to wilt at 8–10 dpi, but the wilt symptoms remained up to 18 dpi; then, the disease index of RK5628 was found to be approximately 3 at 12–18 dpi (Figure 3A). When tomato plants were challenged with petiole-inoculation, the virulence of RK5628 was also significantly reduced (Figure 3B). RK5628 exhibited significantly lower virulence than that of OE1-1 strain in tomato plants. With complementation, the npdA gene could completely restore its less virulence to that of wild type (Figures 3A,B), confirming that NpdA is important for the full virulence of R. solanacearum OE1-1 in tomato plants.

Ralstonia solanacearum OE1-1 is also virulent in tobacco plants. Thus, we evaluated whether NpdA was required for the virulence of OE1-1 in tobacco plants. When tobacco plants were challenged with soil-soak inoculation, significantly reduced virulence by RK5628 was also observed (Figure 3C). When the bacteria were directly inoculated into tobacco plants with leaf infiltration, no difference in virulence was observed between the RK5628 and OE1-1 (Figures 3D).

The RS1002 strain is virulent in tomato but elicits HR in tobacco leaves. We also tested the virulence of RS10013 (npdA deletion in RS1002 strain) in tomato plants using both the soil-soak and petiole inoculation methods, respectively. It was very intriguing to find that this npdA mutant exhibited almost identical virulence as that of the wild-type RS1002 strain (Figures 3E,F), suggesting that NpdA was not involved in virulence of RS1002 in host plants. The involvement of NpdA in virulence of R. solanacearum is specific to OE1-1 in host plants.

Some T3Es are responsible for the HR elicitation in resistant plants. Thus, we evaluated whether NpdA is involved in the HR elicitation of RS1002 and GMI1000 in tobacco leaves. When infiltrated into tobacco leaves (N. tabacum BY), RS10013 (npdA deletion in RS1002 strain) caused apparent necrotic lesions at 16 hpi (hours post inoculation), and the area of necrotic lesions reached a maximum at 24 hpi (Figure 4). The npdA mutant in RS1002 caused almost the same development of necrotic lesions as RS1002 in tobacco leaves (Figure 4). Moreover, GF0021 (npdA deletion in GMI1000 strain) caused identical development of necrotic lesions as GMI1000 in tobacco leaves (data not shown), suggesting that NpdA was not involved in the HR elicitation of R. solanacearum strains in tobacco leaves.

As a vascular pathogen, the growth capacity of R. solanacearum in host plants is essential for its virulence. Therefore, we evaluated whether NpdA was required for the growth of R. solanacearum in planta. We inoculated tomato plants with RK5628 using petiole inoculation and quantified the number of cells in stem tissues with dilution plating. RK5628 proliferated in the stem to approximately 10⁸ cfu g^-1 at 4 dpi and approximately 10¹⁰ cfu g^-1 at 7 dpi (Figure 5). RK5628 exhibited almost the same bacterial growth as that of OE1-1 in tomato stem, suggesting that NpdA was not required for the bacterial growth of R. solanacearum in planta.

The expression level of T3SS in planta could be induced to a much-higher level than that in the hrp-inducing medium. Thus, we further investigated whether NpdA was involved in the induction of T3SS expression in host plants. We challenged tomato plants with petiole-inoculation and collected the bacterial cells from the stems for the β-galactosidase assay in planta (normalized by cells number). The popA expression in the RK5050 (derivatives of OE1-1) increased at 3, 4, and 5 dpi and then decreased quickly to a much lower level at 6 dpi (Zhang et al., 2013). We detected their expression daily from 3 dpi (the wilt symptom appeared) to 6 dpi (wilt symptom increased to 3–4). Similar to that of wild type, the popA expression of RK5628 (derivatives of OE1-1) in tomato stems increased from 3 dpi, reached a maximum at 5 dpi, and then quickly decreased at 6 dpi (Figure 6A). It was distinctly less than that of the wild-type strain in tomato stems at 3 dpi (p < 0.05) and significantly less than that of the wild-type strain at 4–5 dpi (p < 0.01), suggesting that NpdA is also important for the popA expression of R. solanacearum OE1-1 in tomato plants. Although RK10013 (derivatives of RS1002) exhibited identical virulence to wild type RS1002, we also evaluated the popA expression in RK10013 in tomato stems. It was intriguing that RK10013 exhibited almost identical popA expression in tomato stems to that of wild type RS1002 (Figure 6B), indicating that the NpdA was not required for the popA expression of R. solanacearum RS1002 in tomato stems. All these suggested that the NpdA-dependent popA expression in planta was specific to OE1-1 strain.

Ralstonia solanacearum uses HrpB to directly control the expression of the entire T3SS and T3Es, and the hrpB expression is regulated by a complex regulation network. We therefore investigated whether NpdA affected the T3SS expression through some known pathway. We generated some npdA deletion mutants from OE1-1 derivatives (Table 1) and evaluated its effect on the expression of these known regulators, including the prhA, prhIR, prhJ, hrpG, prhG, hrpB, phcA, and prhN genes. The deletion of npdA substantially reduced the hrpB expression in hrp-inducing medium (43 vs. 175 miller units), while it did not change the expression of other regulators, either in hrp-inducing medium (Figure 7) or rich medium (data not shown), suggesting that NpdA affected the T3SS expression through HrpB and its influence was independent of these known regulators.