Skip to main content
ARS Home » Midwest Area » Ames, Iowa » Plant Introduction Research » Research » Publications at this Location » Publication #291507
Title: Distribution of phytopathogenic bacteria in infested seeds

Author
item DUTTA, BHABESH - University Of Georgia
item Block, Charles
item STEVENSON, KATHERINE - University Of Georgia
item SANDERS, F. HUNT - University Of Georgia
item WALCOTT, RONALD - University Of Georgia
item GITAITIS, RONALD - University Of Georgia

Submitted to: Seed Science and Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/21/2013
Publication Date: 12/1/2013
Citation: Dutta, B., Block, C.C., Stevenson, K.L., Sanders, F., Walcott, R.R., Gitaitis, R.D. 2013. Distribution of phytopathogenic bacteria in infested seeds. Seed Science and Technology. 41:383-397.

Interpretive Summary: Plant disease-causing bacteria were counted from individual infected seeds of five host crops to determine whether a common mathematical relationship existed among amounts of bacteria found on different seeds. Bacterial populations were estimated from cowpea (Vigna unguiculata), pepper (Capsicum annuum), tomato (Lycopersicon esculentum), watermelon (Citrullus lanatus) and corn (Zea mays). Numbers of live bacteria of Xanthomonas axonopodis pv. vignicola (Xav) were estimated from cowpea by soaking individual seeds in a salt buffer and plating samples onto semi-selective agar media. Clavibacter michiganensis subsp. michiganensis (Cmm), Xanthomonas euvesicatoria and Acidovorax citrulli populations were estimated in a similar manner from tomato, pepper, and watermelon seeds, respectively. For Pantoea stewartii, bacterial populations were estimated by measuring ELISA absorbance values (intensity of color change) from individual corn seeds and comparing the values to a standard curve of values made from a pure culture of P. stewartii. Among the four vegetable pathogens (Xav, Cmm, X. euvesicatoria, and A. citrulli), the populations of live bacteria ranged from a low of 10 per seed to a high of 23 million per seed. Bacterial populations of P. stewartii ranged from 160,000 to 8.5 billion per seed. When the bacterial populations were graphed for each crop, none of the populations fit a normal distribution. In a normal distribution the counts would be equally distributed to the left and right of a midpoint. Similarly, none of the populations could be converted to a normal distribution by transforming the original counts to a logarithmic value. Overall, the common feature of the five host-pathogen graphs is that they were highly-skewed and positively-skewed, meaning that most of the seeds had relatively low bacterial counts and fell to the left of the midpoint. Positive skewness means that the distributions had a long tail to the right which represented a relatively small number of heavily-infected seeds. Because seedborne populations of plant pathogens were highly variable and not normally distributed, we concluded that one should not simply assume there is an average bacterial population on an infested seed. Developers of seed health assays should be aware of this fact, as it has potential implications for determining suitable sample sizes for seed health testing.

Technical Abstract: Populations of phytopathogenic bacteria representing five host-pathogen combinations were assessed to determine if there was a mathematical relationship common across seedborne bacterial diseases. Bacterial populations were estimated from naturally-infested seeds of cowpea (Vigna unguiculata), pepper (Capsicum annuum), tomato (Lycopersicon esculentum), watermelon (Citrullus lanatus) and corn (Zea mays). Numbers of viable colony-forming units (CFU) of Xanthomonas axonopodis pv.vignicola (Xav) were estimated from cowpea by soaking individual seeds in phosphate-buffered saline followed by dilution-plating onto MXP semi-selective agar medium. Clavibacter michiganensis subsp. michiganensis (Cmm), Xanthomonas euvesicatoria and Acidovorax citrulli populations were estimated in a similar fashion from individual tomato, pepper and watermelon seeds on appropriate semi-selective media, respectively. For Pantoea stewartii subsp. stewartii, bacterial populations were estimated by comparing ELISA absorbance values from naturally-infested corn seed to standard curves created from a serial dilution of P. stewartii. Among the four vegetable bacterial pathogens (Xav, Cmm, X. euvesicatoria, and A. citrulli), the population ranged from 10 CFU/seed to 2.3 x 10^7 CFU/seed, based on sample subsets of a minimum of 30 infested seeds each. Bacterial populations of P. stewartii ranged from 1.6 x 10^5 to 8.5 x 10^9 CFU/seed, based on a sample size of 212 infested seeds. Based on the Shapiro-Wilk test, none of the bacterial populations fit a normal distribution, and none of the populations were normalized after a log10 transformation. Overall, phytopathogenic bacteria representing the five host-pathogen combinations were best represented by a group of closely related probability distributions known as inverse Gaussian distributions. The common characteristic of these distributions is that they are highly-skewed and positively-skewed. Positive skewness means that the distributions had a long tail to the right which represented a relatively small number of more heavily-infected seeds. Because seedborne populations of plant pathogens were highly variable and not normally distributed, we concluded that one should not simply assume there is an average bacterial population found on infested seeds in any crop. Developers of seed health assays should be aware of this fact, as it has potential implications for determining suitable sample sizes for seed health testing.