Canna indica L. is a popular landscape ornamental herb of family Cannaceae throughout China. This plant has been extensively cultivated for decoration and as an ornamental in China. C. indica was seriously affected by a disease in the garden in spring of 2021 with an incidence of 21.2 to 45.6%, and it caused economic loss to control plant diseases with chemicals. Both young and older leaves developed necrotic lesions with small water-soaked spots on leaves, which then enlarged and were bordered by chlorosis. Symptomatic tissues were collected from Zhanjiang city (21.2N110.3E) and Wuchuan city (21.4N110.7E) in Guangdong province. Three symptomatic leaves from each city were surface disinfected in a 1% hypochlorite solution for 3 mins followed by being rinsed in sterile distilled water for 3 times. Six bacterial isolates originated from six single colonies were recovered from the samples. Colonies were raised and opaque with smooth margins. The bacteria were gram-negative, rod-shaped with sizes ranged from 0.4-0.7 μm wide and 1.2-2.0 μm long, without endospore, and were facultative anaerobes. For molecular identification, the direct colony PCR method (Lu et al. 2012) was used to amplify the 16S rDNA (Moreno et al., 2002), gyrB, leuS and rpoB of 2 selected strains, EM21ZJ1 and EM21WC1 using the primer pairs 27F/1492R, gyrB3/gyrB4, leuS3 /leuS4 and rpoBjt112/ rpoBjt748 respectively (Deletoile et al., 2009). The resulting sequences were deposited in GenBank (ON600470 and ON600471 for 16S rDNA; ON600472 and ON600473 for gyrB; ON600474 and ON600475 for LeuS; ON600476 and ON600477 for rpoB). BLAST searches with the four gene sequences revealed the greatest similarity with the sequences of Enterobacter mori (Zhu et al. 2011). The available complete genome sequences of Enterobacter species, especially type strains, were downloaded and the sequences of the corresponding 4 genes of each genome was extracted by Bioedit software. The concatenated sequences were aligned by Mega 11.0 with the neighbor-joining method. Multilocus sequence typing analysis showed that the concatenated sequences of the 2 isolates were clustered with E. mori with 100% bootstrap value. The 2 isolates were selected for pathogenicity tests to fulfill Koch's postulates. Six plants at three- to five-leaf stage were inoculated with each isolate separately, 2 sites of each leaf were inoculated, one site was wounded with a sterile needle and the other was not. The 2 sites of each leaf were covered with a piece of cotton drenched with 200 µl bacteria suspension (108 CFU/ml) from 2 isolates separately. Control plants were inoculated identically except Luria-Bertani (LB) medium was used to drench the cotton. Inoculated plants were placed in an incubator at 25°C, and 80% humidity under a 12-h light/dark cycle for 7 days. After 3 days of incubation, water-soaked yellow spots were observed in all the inoculated plants in both wounded and unwounded sites, except the negative control. The water-soaked yellow spots enlarged and became necrotic that matched those seen in garden. The pathogenicity tests were conducted three times with similar results. The bacteria were then reisolated from the lesions and found to have the same colony morphology and 99.9% identity of 16S rDNA sequences as those of the inoculum. According to morphological and sequence analysis, the pathogen was identified as E. mori. To our knowledge, this is the first report of disease of C. indica caused by E. mori.
Keywords: Causal Agent; Crop Type; Ornamentals; Pathogen detection; Prokaryotes; Subject Areas; herbaceous/flowering plants.