Introduction
Bacterial fruit blotch (BFB) caused by Acidovorax citrulli (Schaad et al., 1978), is a devastating disease of many cucurbitaceae hosts (Burdman and Walcott, 2012). Since a highly aggressive strain was first reported from commercial watermelons at USA in 1989 (Somodi et al., 1991), BFB outbreaks have occurred throughout the world. In Korea, the pathogen was first recovered from commercial watermelons at Cheonbuk province in 1991 (Song et al., 1991) and has since been reported from other cucurbit plants (Song et al., 2015). Like other bacterial diseases, use of chemical control against BFB is a little effective. The most effective strategy was known to use resistance cultivars (Wechter et al., 2011), the great genetic diversity of A. citrulli causes resistant cultivars to act restrictedly. Thus, the analysis information on the genetic diversity of A. citrulli containing the relationship within population and host’s geographical origin can provide a very useful data base for developing resistant cultivars and for effective control strategy of the disease. Recently, many researchers have studied the genetic diversity of A. citrulli population by physiological, phenotypic characteristics, serology, biochemical reaction and pathogenicity tests (Walcott et al., 2004; Melo et al., 2014). Additionally, Song et al. (2015) also reported the existence of two groups in Korean A. citrulli population. However, these studies couldn’t give enough information to understand genetic characteristics of Korean population. Therefore, the current study was tried to clarify the intraspecific genetic diversity within Korean A. citrulli population through multi-locus, repetitive-sequence-polymerase chain reaction (Rep-PCR) and universal rice primers (URP)-PCR analysis in order to obtain reliable data for developing resistance cultivars and the integrated control of bacterial fruit blotch in Korea.
Materials and Methods
Results and Discussion
Bacteria strains and DNA extraction
A total of 29 A. citrulli strains were collected from seeds, fruits, leaves and stems of melon, pumpkin, cucumber and watermelon plants in the cultivation fields of Korea from 2011 to 2014 (Table 1). All of these bacteria were stored in 15% sterile glycerol at - 80. The identification of all strains were performed through pathogenicity test by inoculation on watermelon seedlings and 16s rDNA sequencing (Song et al., 2015). Bacterial DNA was extracted using the DNeasy Tissue Kit (Qiagen Inc., USA).
Multi-locus analysis
For the multi-locus analysis, DNA sequences of six housekeeping genes; adk (437 bp), glyA (563 bp), gltA (487 bp), pilT (405 bp) obtained from our previous work (Song et al., 2015) and ugpB (452 bp) and phaC (479 bp) were selected and PCR products were performed using Primers of ugpB (ugpB1 & ugpB2) and phaC (phaCF & phaCR). Each PCR reaction was performed by mixture (Solgent Co., Ltd., Korea) under the PCR conditions described by Kang et al. (2002). All of the PCR amplicons were sequenced (Macrogen Ltd., Korea). The sequences were visualized and aligned with MEGA5 (Tamura et al., 2011) and also compared with foreign A. citrulli strains (ATCC29625; group I and 30002; group II). A maximum likelihood tree was constructed with the concatenated dataset of the four loci (gltA, pilT, ugpB, and phaC) as the DNA sequences for five loci (adk, gltA, glyA, pilT, ugpB) of all Korean clonal complex (KCC)1 and KCC2 strains of A. citrulli were identical according to the group. A. avenae subsp. cattleyae were served as outgroups.
Rep-PCR and URP-PCR
DNA fingerprinting was conducted using Rep-PCR primers; ERIC1R/ERIC2 and BOXA1R (Louws et al., 1994) and URP-PCR primers; URP2, URP5, URP6, and URP8 from Kang et al. (2002). Conditions for PCR amplification reactions were performed as described by Kang et al. (2002). RAPD bands were scored from the agarose gel and recorded as present (1) or absent (0) and assembled into a data matrix. The combined six results were analyzed with NTSYS (Exter Biological Software), and dendrograms were generated using the unweighted pair group method with average (UPGMA).
Multi-locus relationship analysis
Upon multi-locus relationship analysis using 1,823 bp of the total base sequence of the genes which included gltA, pilT, ugpB, and phaC, A. citrulli were classified into two groups; KCC1 and KCC2 within the same clade apart from A. avenae subsp. avenae and A. subsp. cattaleyae (Fig. 1).
MFig. 1. aximum likelihood phylogeny of 29 Korean strains of Acidovorax citrulli with two foreign strains; A. citrulli CC1 and CC2 types (Feng et al., 2009) and two Acidovorax spp. as the outgroup. The tree was constructed from 1,823 bp using four genes: pilT, gltA, ugpB and phaC. Bootstrap values with 1,000 replicates for the major divisions of A. citrulli from the outgroups. C, cucumber; M, melon; W, watermelon; P, pumpkin.
In this study, the allelic profile analysis using the base sequence of the genes including adk, gltA, glyA, pilT, ugpB except phaC presented the existence of species-specific and group-specific base mutation in 7 sites of KCC1 and KCC2 groups in 29 Korean A. citrulli strains with foreign isolates (Table 2). The 30th bases were cytosine (C), thymine (T) in adk; 439th, 442th, and 451th bases were C, guanine (G), adenine (A) and G, A, C in gltA; 452th were A and G in glyA; 94th were C and T in pilT; and 280th were G and C in ugpB in the strains of KCC1 and KCC2 groups, respectively (Table 2). DNA sequences of gltA, pilT, and ugpB genes for CC1–1 type shared 86% of CC1 group (group I) and CC2-5 type shared 82% of the CC2 group (group II); foreign A. citrulli strains used in studies of Feng et al. (2009) that showed the consistency with Korean strains of KCC1 and KCC2 group, respectively. On the other hand, no group-specific mutation occurred in the base sequence of phaC gene, while the 84th base appeared randomly with 20 : 9 ratio in the individual strains of G to A, demonstrating another mutation within each group in Korean A. citrulli population (Table 1). DNA sequences for CC1-1 type (ATCC29625) and CC2-5 type (30002), foreign A. citrulli strains that showed the consistency with G types of KCC1 and KCC2 group having only G at the 84th base.
DNA fingerprinting Rep-PCR and URP-PCR
According to the cluster analysis using DNA fingerprinting data with these results, we have confirmed two clusters forming 29 unique haplotypes from 29 strains (Fig. 2). 65% of similarity was shown between the clusters. Cluster 1 included 11 recovered strains from watermelon and melon, and cluster 2 included 18 recovered strains from watermelon, melon, cucumber, and pumpkin, which were consistent with the classification results of KCC1 and KCC2 groups by multi-locus analysis. When dividing by two clades of isolated strains from watermelon and melon, 72% of similarity was shown within the group of KCC1 cluster. KCC2 group showed 81% similarity, which was higher than KCC1 group; isolated hosts did develop the cluster but not clade.
Consequently, KCC1 group showed higher isolation rate in the melon than in the watermelon unlike the strains in KCC2 group. 103 bands derived from comprehensive results of both Rep-PCR and URP-PCR, 6 were monomorphic and 97 were polymorphic (Supplementary Fig. 1). Therefore, we could secure the fingerprinting data with more reliable genetic diversity from the results of Rep-PCR and URP-PCR than those from multi-locus phylogenetic analysis. Meanwhile, the strains in KCC2 group which were predominant in the watermelon showed higher occurrence rate within the group of A. citrulli species than the KCC1 group. They were widely recovered from various cultivation regions of the cucurbits in Gyeongbook, Gyeonggi, Gyeongnam, Jeonnam, Jeonbuk, and Chungnam, etc.
Accordingly, using both multi-locus analysis and Rep-PCR and URP-PCR analysis, we have revealed that two groups of genetically differentiated A. citrulli existed in Korea, very similar to two foreign groups widely distributed worldwide, with a variety of genotypes in each group resulting in the genetic diversity and this is very meaningful. Upon further comparison analysis on the pathogenicity differences using various hosts of these strains and on the features of the responses against various control agents, it is considered to be the very useful data in the development of effective resistance cultivars and for the establishment of the disease control system against BFB due to A. citrulli.
Fig. 2. Cluster analysis of Acidovorax citrulli strains and A. valerianella 12066 based on DNA fingerprint profiles generated by polymerase chain reaction (PCR) amplification of repetitive elements and universal rice primers (URP) primers. Distance matrix data was generated using Dice’s (1945) coefficient of similarity and the dendrogram was constructed using the unweighted pairwise group method with arithmetic mean (UPGMA) algorithm. C, cucumber; M, melon; W, watermelon; P, pumpkin.
Supplementary Fig 1. Representative DNA fingerprinting by four universal rice primers-polymerase chain reaction (URP-PCR) and two repetitive elements amplification of Acidovorax citrulli strains. M; 100 bp plus DNA ladder, Lane 1; 11246, 2; 11247, 3; 11248, 4; 14194, 5; 14201, 6; 14202, 7; 12091, 8; 12158, 9; 14222, 10; 11147, 11; 11162, 12; 11163, 13; 11164, 14; 11165, 15; 11201, 16; 11251, 17; 11259, 18; 12089, 19; 12090, 20; 13034, 21; 13211, 22; 17005, 23; 17912,24; 13255, 25; 11171, 26; 12170, 27; 13217, 28; 14236, 29; 17913 of A. citrulli and 30; A. valerianella 12066.
