Disease development in C. maxima begins with drooping of the half side of the top leaf (Figure 1A) and continues with lower leaf wilt and droop, whole plant wilt (Figure 1B), changes in the color of the vasculature to dark brown (Figure 1D), and death of the plant (Figure 1C). These symptoms are similar to those caused by R. solanacearum in other host plants. Bacterial oozing from the stem crosscut sites of plants infected with R. solanacearum was observed (Figure 1E).

To isolate the putative causative agent from diseased plants, 24 wilted plants of C. maxima were collected from fields in the Baiyun district in Guangzhou City, Guangdong province, China. An approximately 1.5 × 0.5 cm section of the basal stem from each diseased plant was excised, surface-sterilized with 70% (v/v) ethanol for 30 s and 0.5% (v/v) sodium hypochlorite for 30 s, and then rinsed with sterile water 4–6 times. An approximately 0.5 × 0.5 mm section of vascular tissue from the surface-sterilized stem section was mashed and macerated in 50 μl of sterile distilled water and left to stand for 2–3 min. These suspensions were then streaked on 2,3,5-triphenyltetrazolium chloride (TZC) medium plates (Kelman, 1954) and incubated at 30°C for 2 days. Large, irregular, round, fluidal and white colonies with pink center pre-dominated the TZC medium plates. One isolate was obtained from each plate after two successive single-colony isolations. In all, 24 isolates were stored in 15% (w/v) glycerol at −80°C for further use.

Virulence tests were performed on four- to six-leaf stage plants of C. maxima, tomato (Lycopersicon esculentum, cv. Dongqie), pepper (Capsicum annuum, cv. Yuejiao No. 15), eggplant (Solanum melongena, cv. Yuefeng), tobacco (Nicotiana tabacum), banana (Musa nana, cv. Dafeng No. 1) and ginger (Zingiber officinale). Fifteen plants of each host were inoculated with each isolate. Six of the seven hosts (excluding ginger), were inoculated by injuring the roots and soaking them in a bacterial suspension (1 × 10⁸ cfu/ml) for 20 min. The roots of 15 plants of each host were also injured and soaked in fluid nutrient medium as negative controls. Fifteen ginger plants were inoculated by injection with 200 μl of a bacterial suspension (1 × 10⁸ cfu/ml) in their stem bases. Fifteen additional ginger plants were injected with 200 μl of fluid nutrient medium as negative controls. After inoculation, the plants were incubated at 28–30°C and 75–80% relative humidity. Disease incidence (DI) was monitored every week for 5 weeks. Plants with visible symptoms (wilted leaves) were recorded as diseased plants. The disease incidence was calculated as DI (%) = 100 × number of disease plants/15 inoculated plants. The experiment was repeated three times. The bacterial pathogen was re-isolated from inoculated C. maxima plants exhibiting wilt symptoms.

DNA was extracted from each isolate using an EasyPure® Genomic DNA kit (TransBionovo Co., Ltd., China) according to the manufacturer's instructions, and the quality and quantity of the DNA samples were checked by measuring their A260/A280 ratios using a NanoDrop One Microvolume UV-Vis Spectrophotometer (Instrument operating software version 1.2.0, Thermo Fisher Scientific, Inc.). The DNA samples were stored at −20°C until use.

The DNA samples were used as templates for PCR amplification of the 16S rRNA gene using the universal primers 27f/1491r (Woese et al., 1983). Approximately 50 ng of DNA template was added to 12.5 μl of Premix Taq™ (0.625 U of TaKaRa Taq, 0.2 mM dNTP mixture, PCR buffer, 10 mM Tris-HCl, pH 8.3, 50 mM KCl, and 1.5 mM MgCl₂) (TaKaRa Biotechnology Co., Ltd, China), 10 pmoles of the primers 27f and 1491r (Woese et al., 1983) and ddH₂O to a final volume of 25 μl. The PCR program consisted of an initial denaturation at 96°C for 5 min, followed by 35 cycles of 94°C for 1 min, 48°C for 1 min and 72°C for 3 min, and a final extension at 72°C for 10 min in a Mastercycler® Gradient Thermal Cycler (Eppendorf, Germany). The PCR product for each strain was subjected to electrophoresis on a 1% agarose gel at 100 V/cm for 40 min, which was then stained with DNA Green (TIANDZ, China) and visualized using a UV-transilluminator.

Each resulting amplicon was ligated into the vector pMD-20T (TaKaRa Biotechnology Co., Ltd, China). The 16S rRNA fragment-containing recombinant plasmids were transformed into Escherichia coli strain JM109. Three clones were randomly picked from each transformation and sequenced in both directions using primer walking (Invitrogen Life Technologies China Co., Ltd, China). The pathogens were identified by blasting the obtained 16S rRNA gene sequences against the NCBI database.

Biovars of the 24 isolates were determined as described by Hayward (1964). Briefly, 10 μl of a cultured bacterial suspension (1 × 10⁸ cfu/ml) were stabbed into slants of Hayward's medium containing 1% filter-sterilized D-(+)-cellobiose, lactose, maltose, sorbitol, mannitol or dulcitol in glass tubes. All of the inoculated tubes were incubated at 30°C for 3 weeks. Each treatment was repeated three times. Liquid culture media inoculation was used as a negative control. Color changes of the Hayward's medium slants from green to yellow were recorded as positive.

The strain phylotypes was determined utilizing a phylotype-specific multiplex PCR (Pmx-PCR) assay following a previously described method (Fegan and Prior, 2005; Prior and Fegan, 2005). Approximately 50 ng of DNA template was added to 12.5 μl of Premix Taq™ (0.625 U of TaKaRa Taq, 0.2 Mm dNTP Mixture, PCR buffer, 10 mM Tris-HCl, pH 8.3, 50 mM KCl, and 1.5 mM MgCl₂) (Takara Co.); 6 pmoles of the primers Nmult:21:1F, Nmult:21:2F, and Nmult:22:InF; 18 pmoles of the primer Nmult:23:AF; 4 pmoles of the primers 759 and 760 (Opina et al., 1997; Xu et al., 2009); and ddH₂O to a final volume of 25 μl. The PCR program included an initial denaturation at 96°C for 5 min, followed by 30 cycles of 94°C for 15 s, 50°C for 30 s and 72°C for 30 s, and a final extension at 72°C for 10 min in a Mastercycler® Gradient Thermal Cycler (Eppendorf, Germany). Eight microliters of PCR product for each strain was subjected to electrophoresis on a 2% agarose gel at a voltage of 100 V/cm for 40 min, which was then stained with DNA Green (TIANDZ, China) and visualized using a UV-transilluminator.

PCR amplification of the 1,360-bp region of the egl gene was performed using the primers eglF (5′-TCTCCATTTTTCCATTTCGTCATG-3′) and eglR (5′-ATGCCATC CGCCACGGACCCGGC-3′). Approximately 50 ng of DNA template was added to 25 μl of Premix Taq™, 0.5 μM of each primer, 2 μl of dimethyl sulfoxide and ddH₂O to a final volume of 50 μl. The PCR program included an initial denaturation at 96°C for 9 min, followed by 35 cycles of 94°C for 1 min, 62°C for 1 min and 72°C for 2.5 min, and a final extension at 72°C for 10 min in a Mastercycler® Gradient Thermal Cycler. PCR amplification of the 1,434-bp region of the hrpB gene was performed using the primers hrpBF (5′-TGACGCTTCAGGAGCATTGCC-3′) and hrpBR (5′-AGATGGGAAGGGAGAGGACC-3′). Approximately 50 ng of DNA template was added to 25 μl of Premix Taq™, 0.5 μM of each primer, 2 μl of dimethyl sulfoxide and ddH₂O to a final volume of 50 μl. The PCR program included an initial denaturation at 96°C for 5 min, followed by 35 cycles of 94°C for 1 min, 62°C for 1 min and 72°C for 3 min, and a final extension at 72°C for 10 min using a Mastercycler® Gradient Thermal Cycler (Eppendorf, Germany). Six microliters of PCR product from each strain was subjected to electrophoresis on a 1% agarose gel at a voltage of 120 V/cm for 25 min, which was then stained with DNA Green and visualized using a UV-transilluminator. The PCR products were purified and sequenced by Shanghai Sangon Biological Engineering Technology & Service CO., Ltd. The PCR primers were also used as sequencing primers.

Twenty-three sequences of the egl gene and 21 sequences of the hrpB gene were selected for analysis (Table 1). The sequences were edited with BioEdit 7.0.5.1 (Hall, 1999) and aligned using Clustal W in DNASTAR MegAlign software, version 5.01 (DNASTAR Inc., Madison, WI, USA). Phylogenetic trees were constructed using the neighbor-joining (NJ) and maximum-likelihood (ML) methods embedded in MEGA version 6.0 (Tamura et al., 2013). Bootstrapping was performed with 100 replicates for the ML tree and 1,000 replicates for the NJ tree. All of the egl and hrpB sequences from the R. solanacearum strains isolated in this study were deposited into GenBank (Table 1).

Twenty-one pumpkin cultivars, including four cultivars of C. maxima and 17 cultivars of C. moschata, were collected and used for resistance evaluation. Plants with four true leaves were inoculated by injuring the roots and soaking them in a bacterial suspension (1 × 10⁸ cfu/ml of strain RS378) for 20 min. Inoculated plants were transplanted into plastic pots (9 cm in diameter) containing horticultural soil. This experiment included four plot experiments with 30 plants each. The DI was monitored every week for 5 weeks after inoculation. Plants with visible symptoms (leaf wilt) were recorded as diseased plants. The disease incidence was calculated as DI (%) = 100 × number of disease plants/total number of inoculated plants in each plot experiment (30). The DI ranged from 0 (no disease) to 100% (dead). Analysis of variance (ANOVA) was used to analyze the disease incidence data from each assessment using a Data Processing System (DPS) based on Duncan's new multiple range method (Hanzhou RuiFeng Information Technology Co., Ltd).

The resistance of the 21 pumpkin cultivars to strain RS378 was evaluated for 5 weeks after inoculation and categorized as follows: DI value of (i) 0–20.0%, highly resistant (HR); (ii) 20.1–40.0%, resistant (R); (iii) 40.1–60.0%, moderately resistant (MR); (iv) 60.1–80.0%, susceptible (S); and (v) 80.1–100%, highly susceptible (HS).

Pure bacterial cultures were isolated on TZC medium plates from the basal stem tissues of C. maxima plants showing wilt symptoms. A total of 24 isolates were obtained. All of the isolates had irregular, round, fluidal, white colonies with pink center that predominated on TZC medium plates. However, 16 of the 24 isolates secreted a brown pigment after incubation at 30°C for 2 days (Figure 2). The phenotypic characterization for biovar determination revealed a predominance of biovar 4 strains over biovar 3 from C. maxima. Of the 24 tested isolates, eight isolates utilized maltose, lactose, D-(+)-cellobiose, mannitol, sorbitol and dulcitol and thus belonged to biovar 3, and 16 isolates secreted brown pigment on TZC medium and utilized mannitol, sorbitol and dulcitol but not maltose, lactose and D-(+)-cellobiose and thus belonged to biovar 4 (Table 1). No other biovars were found.

Using the universal PCR primers 27f and 1491r, each of the 24 isolates generated a 1,423-bp 16S rRNA gene fragment. BLAST analysis of these 16S rRNA gene sequences revealed that they shared 100% identity with R. solanacearum strain GMI1000 (GenBank accession no. AL646052). The 16S rRNA gene sequences of these isolates were deposited in GenBank under accession numbers KY594769 to KY594788 and KX363800 to KX363803. These results established that the isolates from the diseased C. maxima plants exhibiting wilt symptoms in Guangdong, China, were R. solanacearum.

One or two of the youngest leaves of the C. maxima plants inoculated with the R. solanacearum isolates began to successively exhibit wilting at 5 days post inoculation (dpi) (Figure 1F). The initial symptom in diseased plants was apical leave wilting during the day that recovered at night. Subsequently, the wilted leaves failed to recover at night, and the whole plant withered 3 days later. No symptoms were observed in the control plants. Pathogens with the same morphology on TZC medium plates as that of the bacterial isolates mentioned above were also re-isolated from the inoculated plants with wilt symptoms. Based on these results, the pathogen causing wilt in C. maxima was confirmed to be R. solanacearum.

Inoculated plants of tomato, eggplant and pepper began to wilt at 5 dpi. The number of wilted plants stabilized at 35 dpi. Statistical analysis revealed that all of the 24 isolates were pathogenic to tomato and eggplant, with disease incidences of 66.7–100%; to pepper, with disease incidences of 26.6–86.7%; and to tobacco and ginger, with disease incidences of 0–26.67%. These isolates were not pathogenic to banana at 35 dpi (Table 2).

Using the Pmx-PCR protocol, each of the 24 isolates generated the expected 280-bp species-complex-specific fragment and a 144-bp fragment specific to phylotype I of R. solanacearum, suggesting that all of the isolates from C. maxima that cause wilting disease should be phylotype I strains.

Partial sequences of the egl and hrpB genes from these 24 isolates were analyzed. The egl gene sequences from the 24 isolates shared identities of 99–100% with each other. The hrpB gene sequences from the 24 isolates shared identities of 99.1–100% with each other. These sequences were deposited in GenBank under the accession numbers KY594789 to KY594812 for the egl gene and KY594813 to KY594836 for the hrpB gene. Phylogenetic trees were constructed using egl gene sequences of the 24 isolates from C. maxima and 23 reference strains and hrpB gene sequences of the 24 isolates from C. maxima and 21 reference strains of R. solanacearum, which were retrieved from GenBank. The reference strain sequences were added to the phylogenetic tree to position the strains in this study within the known phylogenetic structure (Liu et al., 2017). Two phylogenetic methods (NJ and ML) yielded similar results, but only the NJ tree is displayed for discussion. Based on the egl and hrpB trees, all of the 24 isolates in this study were classified into phylotype I, consistent with results from Pmx-PCR. The branching patterns in the egl and hrpB trees for phylotype I showed a split with 100 and 96% bootstrap values, respectively. In the egl tree, the 24 isolates clustered into three major groups, A, B and C. Group A, which includes 17 isolates, represented sequevar 45. Group B, which contains three isolates, was not identical to any previously designated sequevar and was thus defined as sequevar 56, a new sequevar. Group C, which included four isolates, represented sequevar 17 (Figure 3). Conversely, in the hrpB tree, the 24 isolates clustered into one group (Figure 4).

Inoculated pumpkin plants began to wilt at 5 dpi. The number of diseased plants stabilized at 35 dpi. Statistical analysis revealed that none of the 21 pumpkin cultivars was immune to R. solanacearum strain RS378. The disease incidences for the 21 cultivars ranged from 38.63 to 94.92%. Xiangyu1, a C. moschata cultivar with a DI value of 38.63%, was resistant to strain RS378. Xiangmi, a C. moschata cultivar with a DI value of 57.88%, was moderately resistant to strain RS378. Seven C. moschata cultivars, including Xiangyu101, Pocket-size, C. moschata, Yuemei No. 1, Wushan, Super long pumpkin and Xiangyu2, had disease incidences ranging from 60.46 to 77.50% and were susceptible to strain RS378. Eight C. moschata cultivars, including Butternut squash, Jingou, Zhongnan, Jinling, Xiangmi-Xiangyu, Pumpkin King, Jinniu No. 3 and No. 3 had disease incidences that ranged from 81.50 to 94.92% and were highly susceptible to strain RS378. Four C. maxima cultivars, including Black rose 879, Longxindan, Jinhongxi889 and Red earth, had disease incidences ranging from 83.62 to 93.73% and were highly susceptible to strain RS378 (Table 3).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.