This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The Pennsylvania State University
The Graduate School
College of Agricultural Sciences
SOURCES OF INOCULUM, EPIDEMIOLOGY, AND INTEGRATED MANAGEMENT
OF BACTERIAL ROTS OF ONION (ALLIUM CEPA) WITH A FOCUS ON CENTER ROT,
CAUSED BY PANTOEA ANANATIS AND PANTOEA AGGLOMERANS
The dissertation of Emily E. Pfeufer was reviewed and approved* by the following:
Beth K. Gugino Associate Professor of Plant Pathology
Dissertation Advisor Chair of committee
Gary W. Moorman Professor of Plant Pathology
Maria del Mar Jimenez-Gasco Assistant Professor of Plant Pathology
Paul A. Backman Professor Emeritus of Plant Pathology
Shelby J. Fleischer Professor of Entomology
Frederick E. Gildow Professor of Plant Pathology
Head of the Department of Plant Pathology and Environmental Microbiology
*Signatures are on file in the Graduate School
iii
Dissertation abstract
Commercial acreage devoted to onion production has increased exponentially in Pennsylvania
(PA) over the past fifteen years for several reasons, including the development of the PA Simply Sweet®
marketing program, the establishment of grower cooperatives, and renewed consumer interest in fresh,
local produce. Even with increasing acreage of onions in PA, consumer demands for the crop are not met
by current production. Bacterial rots of onion are the most significant diseases reducing harvest and
storage yields of the crop, in some instances diminishing marketable yields by 60%. Growers manage
bacterial rots of onion through combinations of chemical and cultural practices, including copper
fungicides, plastic mulch, and drip irrigation; however, yields remain variable between seasons and farms.
These producers are interested in alternative practices, including plant defense-inducing treatments,
carefully planned applications of fertilizer during drip irrigation (fertigation), and targeted insect
management for more consistent control of bacterial diseases, however, data is lacking on the
effectiveness of these practices in PA. The ultimate goal of the following research is to contribute
additional integrated management tools to the existing grower knowledge base to increase the
profitability of growing onions in PA.
As a relatively new pathosystem in PA, basic and applied research was conducted in order to
better understand the impact of bacterial rots on marketable yield of onion in the state. The principal
bacterial pathogens were identified as Pectobacterium carotovorum subsp. carotovorum, Pantoea
agglomerans, and Pseudomonas marginalis pv. marginalis. Sources of bacterial inoculum, including soil,
transplants, and weeds, were elucidated in addition to investigation of ecological interactions between
these species, their hosts, and the cropping system. Aspects of the production system are suggested to
affect plant disease in both pathogen- and disease-specific ways, such as the association of black plastic
mulch with increased detection of P. agglomerans, and early-season soil nitrate resulting in decreased
detections of leaf pathogens. On-farm management factors as observed in PA and New York indicate that
higher incidences of bacterial rots of onion are associated with low foliar nitrogen and high soil
temperatures near the physiological onset of bulbing. Replicated field trials in which plant defense-
inducing and growth-promoting compounds were compared for their efficacy in managing center rot of
onion (Pantoea ananatis and P. agglomerans) indicated moderate disease incidence among all
treatments, including the copper-based grower standard treatment, which was only effective at low levels
of inoculum pressure. Comparisons of the source and timing of nitrogen fertilizer application were
completed in a replicated field study, and one year of data suggests an association between late-season
fertilizer application and higher incidence of center rot. In addition, data were generated in an effort to
iv
understand the role of thrips in the epidemiology of center rot disease. Taken together, these datasets
have increased the overall knowledge about the bacterial rot – onion pathosystem in Pennsylvania,
elucidated management practices that hold promise for future replicated study, and improved
management of bacterial rots of onion, particularly through dissemination of research results to growers
during extension presentations.
v
Table of Contents
List of Figures .............................................................................................................................................. vi
List of Tables ................................................................................................................................................. x
Acknowledgements ..................................................................................................................................... xii
Chapter 1: A review of current onion production, focused on Pennsylvania .............................................. 1
Chapter 3: Epidemiology and ecology of Gram-negative bacteria potentially pathogenic to onion (Allium
cepa) in Pennsylvania ................................................................................................................................. 47
Chapter 5: Efficacy of plant defense-inducing and growth-promoting products for the management of center rot of onion (Allium cepa), caused by Pantoea ananatis and P. agglomerans ............................. 87
Chapter 7: Revised best practices for onion production in Pennsylvania and future work .................... 126
Appendix: Preliminary work with thrips identification by PCR primers and the impact of onion thrips (Thrips tabaci) on center rot of onion (Allium cepa) in Pennsylvania ..................................................... 129
Figs. 2.1a – b. Results of semi-anaerobic pathogenicity tests using nonpathogenic (a) and pathogenic (b) strains .. 27 Fig. 2.2. Severity of symptom development in pathogenicity tests (duplicate) under aerobic incubation after seven days. 1 = nonpathogenic; 2 = local discoloration or maceration; 3 = local symptom, plus discoloration or maceration on adjacent scale(s); 4 = up to half of bulb discolored or macerated; 5 = more than 50% of bulb discolored or macerated. The right half of each bulb is inoculated; bulbs were rated individually and average was reported if not identical ...................................................................................................................................................................... 28 Fig. 2.3. Number of bacterial species detected in symptomatic onion bulbs from harvest and storage combined, in 2011 and 2012 ........................................................................................................................................................... 39 Fig. 2.4. Pathogenicity of epiphytic and endophytic bacterial isolates originating from transplants, incubated under aerobic and semi-anaerobic conditions in pearl onion bulbs. Pathogenic isolates are indicated by blue-shaded portions of bars, nonpathogenic isolates are indicated by green portions of bars ................................................... 39 Fig. 2.5. Pathogenicity of epiphytic and endophytic bacterial isolates originating from common weeds collected at midseason, incubated under aerobic conditions in pearl onion bulbs. Pathogenic isolates are indicated by blue-shaded portions of bars, nonpathogenic isolates are indicated by green portions of bars. *Endophytic isolates are only from 2012, and were isolated on OEM, which is semi-selective for onion pathogens and onion-associated bacteria ....................................................................................................................................................................... 40 Figs. 2.6a - c. Disease severity induced on onion by P. agglomerans (A), P. carotovorum (B), and P. marginalis (C) isolates, from environmental and transplant sources from 2011 and 2012, in aerobic pathogenicity tests. Nonpathogenic proportions of isolates are indicated by green portions of bars (NP), pathogenic isolates are indicated by pink-red portions of bars. The shade of the bar indicates the severity of induced symptoms as described in Fig. 2.2; severity increases as the bars approach the x-axis. *Weed epiphytes and endophytes were isolated using semi-selective media in 2012; weed epiphytes include isolates from 2011 and 2012, while weed endophytes include isolates only from 2012. N = number of isolates tested from each source ................................................................ 41 Fig. 2.7. Representative gel of rep-PCR genomic fingerprints for tracking strains of Pantoea spp. from two fields’ transplant, weed, and symptomatic onion isolates. These fields were planted with transplants from the same lot, but were located approx. 32 km from each other, on different soil types. Isolates 562 and 563 originated from the same area of the same field, but 562 is an endophyte from purslane while 563 is an epiphyte from crabgrass. Isolate 551 is an endophyte from shepherd’s purse, while isolate 1617 is from an onion that developed symptoms after approx. 4 months in storage. All highlighted isolates are P. ananatis ....................................................................... 42 Figs. 3.1a – b. Symptomatic onions typified by the generic diseases surface rot (a) or inner scale rot (b). The green wire loop was surface-sterilized and indicates roughly where symptomatic tissue was harvested for DNA extraction and bacterial isolation ................................................................................................................................................ 51 Fig. 3.2. Detection of bacterial species in symptomatic bulbs pooled from at harvest (2011 and 2012) and from storage (2012 only) samples from which only one pathogen was detected and separated by the type of rot symptom observed (N=238). This represents approx. 36% of all symptomatic bulbs collected from 2011 and 2012. The ‘Other’ category includes positive detections for Burkholderia cepacia, Pseudomonas viridiflava, and either P. ananatis (surface rot bulbs [SR]) or B. gladioli pv. alliicola (inner scale rot bulbs [ISR]) ........................................................... 61 Figs. 3.3a - b. Pathogenicity of Pantoea agglomerans, Pectobacterium carotovorum subsp. carotovorum, and Pseudomonas marginalis pv. marginalis from onion transplants in aerobic and semi-anaerobic pathogenicity tests, divided by bacterial isolation source. *All 2012 isolates were generated from semi-selective OEM, while 2011 isolates were generated from KB (nonselective). Weed endophytes were only isolated in 2012 ......................................... 62
vii
Fig. 3.4. Bacterial epiphytes of the species Pantoea agglomerans, Pectobacterium carotovorum subsp. carotovorum, and Pseudomonas marginalis pv. marginalis from selected common weed sources collected in PA onion fields in 2011 and tested for pathogenicity under aerobic incubation. Purple-shaded portions of bars indicate pathogenic isolates of each species, while green shaded portions of bars indicate nonpathogenic isolates from each weed source, regardless of species. * Indicates proportionally more pathogenic isolates compared to each of the other weeds by Fisher’s exact test (α = 0.06) ....................................................................................................................................... 63 Fig. 4.1. Simple slopes analysis of projected relationships between average soil temperature three weeks preharvest and the incidence of total bacterial rot of onion, given different levels of foliar N, from PA-2011. Points of each line were calculated based on the covariance matrix of the multiple regression model in Table 1 (Aiken and West, 1991). Foliar N values (2.41% N [low], 2.75% N [avg], 3.09% N [high]; ±1 standard deviation from the sample mean) were chosen, then projected bacterial disease estimates were calculated based on chosen average soil temperatures three weeks preharvest (23.79, 24.29, 24.79°C; ± 0.5 °C from the sample mean). Simple slopes were compared to H0 = 0; for average and high foliar N lines, t was significant at P = 0.03. ............................................................................... 82 Fig. 4.2. Total bacterial disease incidence by foliar C/N ratio, combined data from PA-2011 and PA-2012. Ten leaves per plot were co-mingled, dried, homogenized, and analyzed for total C and N via dry combustion. Total bacterial disease incidence was the sum of the percentages of symptomatic bulbs at harvest and from storage as a total of the bulbs harvested per plot. Three plots were averaged for each field value; each point represents one field ..... 83 Fig. 4.3. Total bacterial rot incidence by cultivar grown, NY-2011 and NY-2012. Data were analyzed using a one-way ANOVA in Minitab 16.2, error bars represent the standard error of the mean, and letters above each bar indicate statistically significant differences by Fisher’s LSD (α = 0.05)..................................................................................... 84 Fig. 5.1. 2012 inoculation diagram with locations of high (red, center front), medium (orange, flanking red), and low (yellow, center rear) inoculum ................................................................................................................................... 99 Fig. 5.2. Foliar disease symptom rating scale. Foliar ratings are as follows: 0 – no lesion, asymptomatic (uninoculated plants only); 1 – local lesion (2.5 cm x 2.5 cm or less); 2 – expanded lesion, but less than ¼ of leaf; 3 – up to ½ of the inoculated leaf is chlorotic or bleached; 4 – more than ½ of the inoculated leaf is chlorotic or bleached, but uninoculated leaves do not show symptoms; 5 – entire inoculated leaf and a portion of an uninoculated leaf are symptomatic; 6 – multiple fully symptomatic leaves; 7 – ≥50% bleached and/or collapsed leaves ........................ 100 Fig. 5.3. Photograph of Landisville field showing topography of blocks ................................................................... 99 Fig. 5.4. Average center rot incidence and percentage of marketable large-size (>7.6 cm diameter) onions by treatment block, Landisville, 2011. Block was analyzed as a random factor in order to control for natural variation within the onion field; the arrow roughly indicates the topography of the field (low-lying, on left, to high-ground, on right). Data were analyzed using PROC GLM in SAS 9.2, with post-hoc comparisons completed using Fisher’s LSD (α = 0.05); statistically significant differences are indicated by different letters above each set of bars (disease incidence [red bars] = a-c; large bulbs [blue bars] = x-z). Bars represent the experiment-wide standard error ...................... 101 Fig. 5.5. Proportion of total marketable yield that was categorized as large (> 7.6 cm diameter) bulbs across all treatments grouped by pathogen pressure based on inoculation status (low, medium, and high) from Rock Springs and Landisville, 2012. Analysis was completed using PROC GLM in Minitab 16 with post-hoc comparisons using Fishers LSD (α = 0.05). Letters above each bar indicate statistical significance ...................................................... 101 Fig. 5.6. Center rot incidence by treatment under varying levels of inoculum pressure, Landisville, 2012. Center rot incidence at harvest was recorded for different areas of plots for each treatment. Inoculum pressure within plots includes low (uninoculated), medium (adjacent to inoculated), and high (inoculated) treatments. Error bars represent experiment-wide standard error. Different letters above each bar of the same inoculum pressure indicate statistically significant differences by Fisher’s LSD (α = 0.05)................................................................................... 102
viii
Fig. 5.7. Weekly foliar disease severity ratings post-inoculation, from Rock Springs in 2012. For each treatment, 20 inoculated plants per plot were rated for disease severity following the scale in Fig. 5.2. * indicates a statistically significant difference between GB03 and the grower standard Cu-EBDC treatment (Fisher’s LSD; α = 0.05). ........ 102 Fig. 5.8. Center rot incidence by treatment under varying levels of inoculum pressure, Rock Springs, 2012. Center rot incidence at harvest was recorded for different areas of plots for each treatment. Inoculum pressure within plots includes low (uninoculated), medium (adjacent to inoculated), and high (inoculated) treatments. Error bars represent experiment-wide standard error. Different letters above each bar of the same inoculum pressure indicate statistically significant differences by Fisher’s LSD (α = 0.05)................................................................................... 103 Fig. 5.9. Center rot incidence on research farms in 2012. Bars indicate the mean of each inoculation level on each farm, regardless of in-season treatment, and error bars indicate standard error of each mean. Statistically significant differences are indicated by different letters above the bars (Fisher’s LSD; α = 0.05) ............................................ 103 Fig. 6.1. Lesion development in differentially-fertilized onion seedlings after foliar inoculation with P. ananatis in a growth chamber assay. Inoculated plants are indicated by INC. Means at each date were compared between treatments using ANOVA with Fisher’s LSD; * indicates a statistically significant difference between the Nitrate and Ammonium-fertilized, inoculated plants, while ** indicates a statistically significant difference between inoculated control and nitrate fertilizer treatments compared to the ammonium-only fertilizer treatment. Error bars represent standard error of the mean ..................................................................................................................................... 119 Fig. 6.2. Center rot progression by fertilizer application timing in inoculated plants in the field. N treatments were combined within the timing variable since no differences were apparent between N types. Means within each type of timing were compared using a one-way ANOVA and Fisher’s LSD (α = 0.05). * indicates a statistically significant difference in disease severity between the full-season and half-season fertilized treatments ............................... 120 Figs. 6.3. Center rot at harvest by inoculation status, N fertilizer source and application timing. Bars indicate the average of four replicate plots, error bars represent standard error of the mean. Bars with different letters indicate statistical significance by Fisher’s LSD (α = 0.05) ..................................................................................................... 121 Fig. 6.4. Center rot incidence at harvest based on inoculation status and fertilizer application timing. Bars indicate the average of the plots within each category (N = 4 [unfertilized plots] or N = 8 [half- or full-season fertilized plots, regardless of N source]). Error bars represent standard error of the mean and different letters above each bar indicate statistical significance by Fisher’s LSD (α = 0.05) ........................................................................................ 121 Fig. 6.5. Sulfur content of asymptomatic bulbs at harvest by N fertilizer type and timing. Bars indicate averages by treatment group; different letters above bars indicate statistical significance by Fisher’s LSD (α = 0.05) .............. 122 Fig. 6.6. Analysis of covariance estimates with treatment (listed under each bar with the timing and form of inorganic N) as a main effect and calcium in foliage after three weekly fertilizer treatments as a covariate, with center rot incidence as the dependent variable. Inoculated treatments are solid color bars, uninoculated treatments are hatched. Dark blue indicates full-season application (N fertility applied weekly throughout the season), medium blue indicates half-season application (N fertility applied weekly prior to midseason), and light blue indicates no additional N fertility. Means shown are estimates with early-season foliar Ca included as a covariate in the model. Bars with different letters are significantly different by Tukey’s test (α = 0.05) ...................................................................... 122 Fig. 6.7. Analysis of covariance estimates with treatment (listed under each bar with the timing and form of inorganic N) as a main effect and nitrogen in foliage after six weekly fertilizer applications as a covariate, with center rot incidence as the dependent variable. Inoculated treatments are solid color bars, uninoculated bars are hatched. Dark blue bars indicate full-season fertilizer application, medium blue bars indicate half-season fertilizer application, and light blue bars indicate no fertilizer application. Means shown are estimates with N included in the model. Bars with different letters are significantly different by Tukey’s test (α = 0.05) ...................................................................... 123
ix
Fig. 6.8. Center rot incidence at harvest compared to amount of N fertilizer of either type applied by midseason. Statistically significant differences only exist between means in different inoculation categories. Error bars indicate standard error of the mean ..................................................................................................................................... 123 Fig. A.1. Frankliniella occidentalis (left) and Thrips tabaci (right) as viewed at 100x magnification ....................... 131 Fig. A.2. Representative electrophoresis gel of PCR reactions using published Thrips tabaci (expected amplicon size, 298 bp; Asokan et al., 2007), Frankliniella occidentalis (expected amplicon size, 249 bp; modified from Huang et al., 2010), and Thrips palmi (expected amplicon size, 390 bp; Asokan et al., 2007) primer sets .................................. 133
x
List of Tables
Table 2.1. Detection of eight bacterial pathogens of onion from environmental and plant samples, collected in 2011 and 2012. The percentage of the total N of each sample type is given in the appropriate column; columns sum to more than 100% because many samples gave positive identifications for multiple targeted bacterial species ...... 38 Table 2.2. Summary of rep-PCR genomic fingerprinting analysis of environmental and pathogenic isolates from symptomatic onion tissue. Isolates of Pantoea spp. collected from a variety of sources in and around 26 onion production fields in 2012 were used as templates in rep-PCR genomic fingerprinting analysis to track bacterial strains through space and time ............................................................................................................................................. 43 Table 2.3. Logistic regression analysis of detections of P. marginalis and P. agglomerans in symptomatic onion bulbs at harvest and from storage predicted by environmental and transplant source species detections, using farm as a factor, from 2011 and 2012. Dependent variable is binary P. marginalis (A; C) or P. agglomerans presence (B; D; presence = 1) in symptomatic onion bulbs from harvest and storage; independents are positive transplant endophyte (A), detections in each of early-season soil (B), transplant epiphyte (C), and weed epiphyte samples (D) .............. 43 Table 3.1. Logistic regression of general types of bacterial rot (surface rot or inner scale rot) modeled by the detection of bacterial species in symptomatic bulbs collected at harvest and from storage in PA in 2011 and 2012. Dependent variable modeled is 1 = inner scale rot (n = 395), while independent variables are species detections (presence = 1) ............................................................................................................................................................ 61 Tables 3.2a – d. Logistic regressions of harvest and storage detections of Pantoea agglomerans (3.2a; N = 225 of 617), Pectobacterium carotovorum subsp. carotovorum (3.2b; n = 366 of 614), Pseudomonas marginalis pv. marginalis (3.2c; n = 171 of 614), and Pantoea ananatis (3.2d; N = 48 of 614) from symptomatic onion bulbs from PA, combined in 2011 and 2012; positive detections rated ‘1.’ Independent variables are detections of other bacterial species in symptomatic bulbs as well as environmental and production factors observed throughout the season .. 64 Tables 3.3a – b. Pantoea agglomerans detections in surface (3a; n = 68 of 174) and inner scale rot (3b; n = 128 of 348) bulbs with other bacterial species detections and environmental and management factors as independent variables ..................................................................................................................................................................... 66 Tables 3.4a – b. Pectobacterium carotovorum subsp. carotovorum detections in surface (3.4a; n = 163) and inner scale rot (3.4b; n = 225 of 360) bulbs with other bacterial species detections and environmental and management factors as independent variables ............................................................................................................................... 67 Table 4.1. Field-averaged results of multiple linear regression analysis of PA-2011 dataset. All variables were placed in a stepwise model selection procedure (α to add = 0.15, α to remove = 0.2), with total bacterial rot incidence as the dependent variable. For this model, R2 = 0.557; adj. R2 = 0.480; P = 0.001 ........................................................ 82 Table 4.2. Field-averaged results of multiple linear regression analysis of PA-2012 dataset. All variables were placed in a stepwise model selection procedure (α to add = 0.25, α to remove = 0.3), with a logistic transformation of total bacterial rot incidence as the dependent variable. For this model, R2 = 0.314; adj. R2 = 0.212; P = 0.052 ............... 83 Table 4.3. Field-averaged results of multiple linear regression analysis of combined NY-2011 and 2012 datasets. Independent variables were observed in 22 and 32 fields, respectively. All variables were placed in a stepwise model selection procedure (α to add = 0.15, α to remove = 0.2), with a square root transformation of total bacterial rot incidence for each field as the dependent variable. For this model, R2 = 0.161; adj. R2 = 0.126; P = 0.019 .............. 83 Table 4.4. Field-averaged results of multiple linear regression analysis of combined PA datasets. All independent variables observed in 54 fields were placed in a stepwise model selection procedure (α to add = 0.05, α to remove = 0.1), with a log transformation of total bacterial rot incidence as the dependent variable. For this model, R2 = 0.179; adj. R2 = 0.147; P < 0.001 ........................................................................................................................................... 84
xi
Table 5.1. Treatments, treatment schedule, and 2011 growth and disease incidence results from Rock Springs, 2011. Similar treatments and application schedules were conducted in the 2011 trial in Landisville as well the 2012 trials in Rock Springs and Landisville. Longest leaf means were separated using Fisher’s LSD (P ≤ 0.05); different letters following the mean indicate statistically significant differences ................................................................................ 98 Table 6.1. Relationships between foliar carbon (C), early-season soil ammonium (NH4), and silt content of soil to the foliar nitrogen (N) content from leaves collected at midseason, on 54 Pennsylvania onion fields over two years .................................................................................................................................................................................. 110 Table 6.2. Midseason growth estimates and thrips pressure by fertilizer treatment prior to inoculation in 2013 … 120 Table A.1. Sequences, expected amplicon sizes, and sources of primers used to identify thrips collected from PA to species ..................................................................................................................................................................... 133 Table A.2. Bacterial isolates from thrips tested for pathogenicity on onion through aerobic and semi-anaerobic pathogenicity tests. Ratios in each column pertain to the pathogenic isolates out of all isolates of that species tested; total columns indicate the total number of strains of each species in the collection ............................................. 134
xii
Acknowledgements
I would like to acknowledge the funding sources that supported parts of this dissertation: the
Northeast IPM Competitive Grants program, the Pennsylvania Vegetable Growers Association, the
Pennsylvania Vegetable Marketing and Research Program, and the Larry J. Jordan Endowment in Plant
Pathology and Environmental Microbiology at Penn State. I would also like to thank the onion growers of
Pennsylvania, who always welcomed me to their farms, even if my being there meant they had a
significant disease issue. Without their cooperation, much of this work would not have been possible, and
I have learned as much from them as they have learned from me. Thank you also to numerous extension
educators who facilitated our grower visits and offered advice along the way.
The Department of Plant Pathology and Environmental Microbiology has been a generous source
of financial, educational, and emotional support over the past six years, during both my Master of Science
and Ph.D. degrees. I would particularly like to thank Dr. Beth Gugino, my dissertation advisor, for her
untiring mentorship, excellent advice, positive attitude, and the multitude of opportunities she opened to
me. I am incredibly grateful for all of our teaching moments, whether in her office, our lab, at Rock Springs,
or on grower farms. Dr. Maria del Mar Jimenez Gasco, Dr. Gary Moorman, Dr. Fred Gildow, and Dr. John
Pecchia have also been wonderful sources of advice and friendship, and this process would not have been
the same without them. Thank you to the committee members I have not mentioned yet, Dr. Paul
Backman and Dr. Shelby Fleischer, who provided thoughtful critiques to help craft this dissertation into
what it has become. Tim Grove and members of the Rock Springs farm crew, and John Stepanchak, Jim
Bollinger, and members of the Landisville farm crew were instrumental in completing the research farm
trials. I am also grateful to all of the staff and current and former students from the department for their
friendship and assistance, including Dr. Michele Mansfield, Dr. Anissa Poleatewich, Ilse Huerta, Dr.
Vasileios Bitas, Sarah Bardsley, Freddy Magdama, Sara May, Roxanne Lease, Steve Conaway, and Anna
Testen.
Many thanks to all of my friends and family, without whose support this would not have been
possible. Thank you especially to my parents, Tony and Jan, as well as my siblings, who sparked my interest
in science at an early age, then fed the fire over the subsequent years! I was lucky to have many talented
teachers and professors early on, including Mrs. Laura Anderson, Mrs. Mary Buerk, Dr. Catharina Coenen,
and Dr. Ann Kleinschmidt. My friends, particularly Kat and groupchat (Liz S., Liz E., KP, Lindsay, and Eva),
saw me through successes and failures, and I am grateful to have their unrelenting encouragement.
Finally, thank you to my best friend and husband, Andy, who was a constant source of support as I became
a ‘plant-killer.’ He shares in this Ph.D., and I could not have done it without him.
1
Chapter 1: A review of current onion production, focused on Pennsylvania
Social and economic significance of onion
Onions (Allium cepa L.) are an important component of countless cuisines around the world,
lending varied flavors from savory richness to striking sweetness to meals, both cooked and raw. An
ancestral vegetable, onion’s medicinal properties were described in a 2000-year-old Indian medical
treatise, it is mentioned in both the Old Testament of the Bible and the Koran, and bulbs have been
recovered from Egyptian tombs, indicating they were cultivated as early as 3200 B.C (Schwartz et al.,
2008). The per capita consumption in the U. S. is approximately 9.5 kg (21 lb) of onions per year. Onions
are rich in vitamin C, fiber, and folic acid (PA-DPN). Approximately 170,000 hectares were planted to onion
in the United States in 2012, and combined among green, summer, and storage onions, onion production
is greater than a two billion-dollar industry nationwide (NASS-USDA, 2013).
Bacterial bulb rots, the most significant diseases in fresh-market onion production in the Mid-
Atlantic region, may cause annual losses of up to 60% on some farms. In 2013, a year with particularly
high disease incidence, a number of sweet onion growers in PA suffered losses averaging $4000 per
hectare ($2500 per acre; J. Stoltzfus, adult educator, Eastern Lancaster County School District). Crop losses
may occur immediately at harvest or when the grower sells the onions after a short period of storage.
Efficient and effective management of bacterial diseases of onion remains a formidable challenge.
Onion production and marketing in the U. S.
Commercial onion production varies widely among regions and states. In perhaps the most well-
known onion producing state, Georgia, short-day sweet onions are produced from transplants of multiple
varieties approved for Vidalia® production, which are set in November or December in mulch-free, raised
beds with overhead irrigation. These onions are grown to maturity in one of twenty counties until March
or April, then harvested by undercutting the roots, field-cured for several days, then manually or
mechanically removed from the field and further cured in front of a fan for several more days. These
onions are marketed April through June as part of the Vidalia® marketing strategy; one estimate is that
1/3 of sweet onions sold are Vidalia®. Bacterial diseases are concerns in Georgia onion production are
primarily associated with late onion varieties or warmer than average temperatures. Fungal diseases
typically receive more attention in disease management programs (Boyhan and Kelley, 2007).
California leads the United States in terms of dry bulb onion production (USDA, 2013). Onions are
direct-seeded from September to May in low and high desert areas as well as in coastal regions. Onion
fields are irrigated in raised beds with overhead, furrow, or drip irrigation. Either short-day or
2
intermediate day varieties are grown depending on the production location, and production soil types are
widely variable. As in all onion production systems, commercial production in California has high nutrition
requirements, particularly for nitrogen (Smith et al., 2011).
Washington, Oregon, and Idaho trail California in dry bulb production (USDA, 2013), but together
account for a large proportion of the bold cooking onions produced in the U. S., with some production of
sweet onions, particularly in Washington. All varieties are intermediate- to long-day onions. In these
states, plants are usually direct seeded into mineral soils in February or March, though some operations
start seeds in fall (particularly for Walla Walla production) or plant farm-grown transplants. Some onion
production on high organic matter, muck soils occurs in western Oregon. Plastic mulch is rarely used in
Pacific Northwestern systems, where growers may use overhead, furrow, or drip irrigation. Onions are
harvested in the region from August to October, depending on variety and crop maturity (Oregon State
Vegetable Production Guides, 2004).
Appreciable amounts of onion are commercially grown in Texas, Nevada, New York, Colorado,
New Mexico, and a number of other states. In New York, onions are primarily grown on high-fertility muck
soils with overhead irrigation. These fields are direct-seeded with some of the same onion varieties grown
in Pennsylvania, however, production is focused on bold cooking onions as opposed to sweet varieties. A
minority of producers in New York use plastic mulch and drip irrigation, much of which is used on mineral
soils.
Onion production and marketing in Pennsylvania
Most onions in PA are grown in small fields, sometimes as little as 0.1 hectare, by growers
producing a variety of other vegetable crops, in addition to other farm enterprises. In 2005, over 40
growers across the state were participating in the PA Simply Sweet® marketing program, on approximately
16 total hectares (38 acres; PVGA, 2010). In 2014, there are approximately 100 growers in Lancaster and
Chester counties commercially growing onions, with most participating in the Simply Sweet® program and
selling their crops through wholesalers, produce auctions, and directly through farm stands (J. Stoltzfus,
pers. comm.) In addition, at least 20 growers who may or may not be participating in the Simply Sweet®
program are also commercially producing onion in central, southwestern and eastern PA, primarily for
direct-market sales like farm markets and community-supported agriculture. Simply Sweet® onions,
Pennsylvania’s only trademarked crop, represent a burgeoning market for growers in the state, with crops
commanding premium prices in vegetable buyers’ markets. In 2012, despite an increasing number of
3
growers and over 100 acres in commercial production with the potential of producing over 5 million
onions, consumer demand continued to exceed supply.
To participate in the Simply Sweet® program, which is managed by the Pennsylvania Vegetable
Growers Association, accepted growers must pay a fee, estimate their expected yield, and plant one of
three approved low-pungency cultivars (‘Candy,’ ‘Expression,’ or ‘Enterprise’). All of these cultivars are
sweet, yellow, intermediate- or long-day summer onions, and intended for use as a fresh-slicing onion,
rather than a dry bulb storage onion. The Simply Sweet® program requires that the onions be larger than
a minimum size (7.62 cm in diameter; 3-in.) and analyzed to contain >6% soluble sugar in composite
samples of harvested bulbs. Growers are also required to grow onions using plastic mulch and drip
irrigation. Although not all PA onion growers are participating in the Simply Sweet® program, the majority
are growing ‘Candy.’
Onion plants are started from seed in fields by commercial producers in Arizona or Texas, or PA
growers produce their own plants in local greenhouses starting in December or January. In general,
transplanting of 10-12 week old plants occurs in late March or early April. Onions are a biennial crop, the
physiology of which is highly influenced by day length and temperature. Transplants only begin to produce
new leaves as the soil warms, typically in mid-April. Prior to transplanting, 20 – 30 cm (8 - 12-in.) tall raised
beds are plowed, two rows of drip irrigation tubes are put into position, and the raised bed is then covered
with plastic mulch. Standard plant spacing is 15 cm (6-in.) between plants within the row and 15 cm
between rows; in most fields, there are four rows of plants running longitudinally down the bed. By
changing plant spacing, bulb size can be manipulated, i.e., wider plant spacing generally results in larger
bulbs; some growers choose to alter plant spacing depending on several factors, such as target market or
field size.
Onion varieties have been developed for various regions based on anticipated air temperature
and daylength during the growing season. As a monocot, onions have unbranched, shallow roots that
extend from an extremely abbreviated ‘stem’ or basal plate, at the bottom of the bulb. Full-size bulbs,
composed of modified leaf tissue of approximately fifteen layers, or scales, evolved as a storage organ for
the biennial plant (Schwartz et al., 2008). Each scale in the bulb corresponds to a leaf blade that extends
from the narrowest point of the foliage, called the onion neck, and this leaf primarily produces
photosynthates to nourish its scale through the season (Mann, 1983). If leaves are removed,
carbohydrates are translocated through the basal plate to the orphaned scales, and the bulb overall will
be smaller than bulbs with all of their leaves (Bartolo et al., 1994; Mann, 1983). While onions may have
fifteen or more layers, plants typically only support ten or eleven leaves at a time, with the oldest leaves
4
progressively senescing, drying, and falling to the soil. After attaining eleven full size leaves in May and
early June, bulbing is induced in long-day varieties (recommended for PA) by fourteen to sixteen hours of
daylight, occurring in mid to late June. Bulbing is characterized by a lack of new leaf formation, rapid
transfer of carbohydrates from the leaves to the bulb, and a rapid build-up of bulb biomass, which is
desirable for harvest (Schwartz et al., 2008).
Onions are generally harvested in early to mid-July in PA, for a total field season length of
approximately three months. Most growers pull onions by hand from the ground, which are laid on the
plastic mulch for 1-2 days, with the leaves of one plant covering the bulbs of neighboring plants to prevent
sun injury before the tops are removed and the bulbs are loaded into bins. Alternatively, some growers
remove onion foliage immediately in-field and bulbs are loaded into bins, which are placed in shade or
cool storage. In either case, binned onions are placed near fan-forced air to expedite drying down of the
necks, which in addition to the outer few scales, dry to papery layers over at least a week. Neck drying
seals moisture inside and helps exclude pathogens; bulbs that are improperly dried typically do not store
well and are more susceptible to postharvest disease, such as Botrytis neck rot (Pfeufer, observation).
Growers who observe foliar bacterial disease symptoms in the field often choose to harvest earlier than
initially planned in order to ensure a larger proportion of asymptomatic bulbs; since bulbs produce a
significant amount of tissue (approx. 0.6 cm diameter / week; D. Zook, grower cooperative board member;
pers. comm.) in the last three weeks of the growing season, timing of harvest may significantly influence
yields. One estimate equated an additional 0.6 cm in diameter to a 20 – 25% increase in yield for the entire
crop (D. Zook, pers. comm.). Harvested onion bulbs are considered ‘colossal’ if they are greater than 10.16
cm (4-in.) in diameter and ‘jumbo’ if between 7.62 (3-in.) and 10.16 cm (4-in.) in diameter. Onions smaller
than 7.62 cm in diameter are sold by growers in one major cooperative for lower prices in netted bags, in
some cases to large grocery stores. These onions may also be kept by the grower for home use.
Primarily marketed as a fresh-slicing onion, the ‘Candy’ cultivar is sold in several different markets,
depending on grower location and management preferences. Most of the PA growers are located in
Lancaster and Chester counties, and participate in onion grower cooperatives. These cooperatives allow
the growers to take advantage of quantity discounts for transplant purchases, reduced capital investment
due to shared equipment, and allow smallholder growers to pool their harvests to sell as larger lots to
wholesalers, typically in 18 kg (40 lb) boxes. In an average year, a grower will plant approximately 125,000
transplants per hectare (50,000 plants per acre), and may expect yields of 1350 kg per hectare (1200 lb /
acre; D. Zook, pers. comm.). Growers in the cooperative may also reserve a portion of their harvest for
sale at produce auctions or at their own roadside stands. Growers not involved in cooperatives produce
5
onions for sale by the bulb at on-site markets, roadside stands, or community farmer’s markets, where
bulbs may sell for $1 each or more. Additionally, a few growers produce onions to include as part of their
own Community-Supported Agriculture (CSA) shares.
Growers’ demand for larger, more valuable bulbs may result in the use of intensive nutrient
management programs, including pre-plant manure applications (particularly by Amish, Anabaptist, and
other sect growers) as well as proprietary programs for applying fertilizer during drip irrigation
(fertigation), which are marketed and distributed by fertilizer industry representatives. While manure
application is often low-cost, fertilizer programs delivered during the season are heavily marketed by
company representatives and subsequently sold at a premium to growers; one figure for such fertility
programs was estimated at $1600 per hectare per season ($1000 per acre; J. Stoltzfus, pers. comm.). Since
the specific components of the fertilizer program are often proprietary, growers rely on company
consultants for weekly fertigation recommendations, which increases the number of farm visits by the
representative and associated costs. While this practice may help ensure more jumbo and colossal size
bulbs, it may also result in higher incidence of bacterial disease and over-fertilization of nitrogen (Diaz-
Perez et al., 2002), both of which reduce grower profits.
Onions are produced across PA, but commercial production is concentrated in the southern half
of the state where soils are more amenable to vegetable production. A majority of the growers included
in the first and second survey years are located in Lancaster and Chester counties, with areas near
Pittsburgh, Philadelphia, and Adams County also represented. The distribution of farm locations in
Chapters 2, 3, and 4 is representative of commercial onion production across the state. While other onion-
producing regions report problems with diseases as white rot (Sclerotium cepivorum), Botrytis leaf spot,
pink root (Fusarium spp.), iris yellow spot (caused by iris yellow spot virus), and purple blotch (Alternaria
porri), bacterial bulb rots are the most significant diseases that reduce marketable yields of onion in PA.
Causal agents of bacterial bulb rots, with a focus on Pantoea spp.
Bulb rots of onion may be caused by a number of different bacterial species, including Pantoea
Though monocots and dicots have been shown to possess many of the same defense-related genes,
differences between plants’ SAR responses have been noted, such as instances of a lack of SA upregulation
in barley following infection and constitutively high SA levels in rice (Balmer et al., 2013). However, SA
treatment in onion was demonstrated to increase callose deposition in response to infection of Botrytis
spp. (Poliakovskiy and Dmitriev, 2011).
Induced systemic resistance (ISR) is similar to SAR in terms of the results it confers to plants,
however, the two defense responses have little else in common. ISR is typically triggered by
nonpathogens, such as rhizospheric associations with plant growth-promoting rhizobacteria (PGPR) or
arbuscular mycorrhizal fungi (AMF; Balmer et al., 2013; Liu et al., 2007). This results in an upregulation of
the signaling molecule jasmonic acid as well as the plant hormone ethylene, which activate defense
cascades against necrotrophic pathogens and insect pests (Spoel et al., 2007; Van Loon, 2007). ISR, like
SAR, may differ between monocots and dicots, however, in one case, Bacillus cereus – induced ISR was
effective at suppressing Botrytis elliptica for up to ten days post application in Lilium formanosum, which
was attributed to heat-labile eliciting factors produced by the bacteria, which remained effective even
after the bacteria were rendered nonviable by autoclaving (Liu et al., 2008).
13
Plant induced defenses regulated by the induction of SAR may come at the cost of reduced yields
as a result of a potential deficit in the plants’ metabolic resources in order to maintain high levels of
defense proteins (Gent and Schwartz, 2005; Louws et al., 2001; Romero et al., 2001; Walters and
Fountaine, 2009). The reduced-yield effect may be especially pronounced, as product labels warn, during
periods of plant stress. Some mitigation of the yield reduction effect has been reported by the use of PGPR
in amaranthus (Nair et al., 2006) and tomato (Obradovic et al., 2005), so combinations of SAR inducers
and plant-growth promoters may provide consistent disease control while still producing profitable yields.
In addition to direct ISR induction, both PGPR and AMF may alternatively serve to reduce plant nutrient
stress by way of their suggested roles of enhancing nutrient uptake. These additional nutrients could allow
plants to devote more resources to withstand bacterial ingress or mitigate yield reduction effects when
used in combination with plant defense inducers. While AMF colonization was shown to be delayed in SA-
overproducing tobacco in one study, the final level of AMF colonization between wild-type and mutant
plants did not differ (Herrera Medina et al., 2003). These studies suggest complicated interactions
between PGPR, AMF, and defense-signaling pathways, and relatively little research has investigated field
interactions between these organisms and commercial defense-inducing products.
Thrips
In addition to bacterial diseases, onions grown in PA are exposed to herbivory from several
different types of insects, including onion maggot, yellow striped army worm, and black cutworm;
however, the most significant insect pest of onions in the Mid-Atlantic is onion thrips. Thrips are small
(0.5 – 2 mm) insects that are polyphagous pests to many agricultural crops. The thrips lifecycle has two
feeding larval stages followed by the non-feeding pre-pupa and pupa stages, and a feeding adult phase.
Eggs are laid by adults within plant tissues, with larvae emerging a few days later (Alston and Drost, 2008;
Morse and Hoddle, 2006). It is assumed that larvae often feed on the same plant on which they were
hatched, as their wings are not yet developed. Pupal stages occur in the soil beneath the plant, from which
a winged thrips adult emerges. A single thrips individual may complete its entire lifecycle in as little as 14
days, if climatic conditions are warm and humid, but heavy rain does not occur. Thrips feed on plant leaves
and flowers using their ‘punch-and-suck’ mouthparts (Alston and Drost, 2008); the resulting feeding
symptoms are commonly referred to as ‘silvering’ of leaves or flowers, which appears as damage to
superficial layers of tissue. In addition to photosynthetic reduction and plant injury, over 20 species of
thrips are known to vector tospoviruses, including iris yellow spot virus, which are some of the most
economically damaging viruses worldwide (Morse and Hoddle, 2006).
14
Onion thrips, Thrips tabaci Lindeman, have been reported as the primary problematic insect
species for onion production in New York (Shelton et al., 2006), and varying levels of thrips damage were
found on all 32 PA farms surveyed in 2011 and 2012 (Pfeufer and Gugino, unpublished). Western flower
thrips and onion thrips have been demonstrated to vector P. ananatis and P. agglomerans in Georgia
(Wells et al., 2002; Gitaitis et al., 2003; Dutta et al., 2012, 2014), in addition to iris yellow spot virus (Gent
et al., 2006), which illustrates why thrips management for onions may have more significant impacts than
solely yield concerns. Thrips are difficult to identify to species as a result of their small size, plasticity in
morphology, and active lifestyle; in addition, a thorough study of the number and diversity of thrips
species in PA onion fields has never been conducted, and grower awareness of the presence and impact
of thrips on onion yields is lacking.
The goal of this dissertation is to refine integrated management of bacterial rots of onion, the
most locally significant diseases of this crop, to ultimately increase the profitability of producing sweet
onions in Pennsylvania.
Specific objectives are to:
1. Identify sources of bacterial inoculum in and around production fields through the middle of
the onion growing season in Pennsylvania.
2. Elucidate epidemiological relationships in situ among eight potentially pathogenic bacterial
species and environmental or production factors in onion production systems.
3. Identify environmental and management factors associated with high incidence of bacterial
disease in Pennsylvania and New York.
4. Determine the efficacy of plant defense-inducing and growth-promoting products as
alternatives to grower standard, copper-based treatments for the management of center rot
of onion.
5. Determine the influence of the type and timing of nitrogen fertilizer applications on center
rot incidence, severity, and micronutrient content of onions.
Through extension presentations, portions of these results have been disseminated to growers,
and continuing educational opportunities are planned in efforts to expand the grower knowledge base as
well as increase marketable yields of onion in Pennsylvania.
15
References Ahmad, M., Majerczak, D. R., Pike, S., Hoyos, M. E., Novacky, A., Coplin, D. L. 2001. Biological activity of harpin produced by Pantoea stewartii subsp. stewartii. Microbe-Plant Molecular Interactions 14: 1223-1234. Alston, D. G., Drost, D. 2008. Onion thrips (Thrips tabaci). Utah Pests Fact Sheet, Utah State University Extension. Ent-117-08PR. Anderson, P. K., Cunningham, A. A., Patel, N. G., Morales, F. J., Epstein, P. R., Daszak, P. 2004. Emerging infectious diseases of plants: pathogen pollution, climate change, and agrotechnology drivers. Trends in Ecology and Evolution 19: 535 – 544. Araujo, W. L., Marcon, J., Maccheroni, W., van Elsas, J. D., van Vuurde, J. W. L., Azevedo, J. L. 2002. Diversity of endophytic bacterial populations and their interaction with Xylella fastidiosa in citrus plants. Applied and Environmental Microbiology 68: 4906 – 4914. Balmer, D., Planchamp, C., Mauch-Mani, B. 2013. On the move: induced resistance in monocots. Journal of Experimental Botany 64: 1249 – 1261. Barash, I., Manulis-Sasson, S. 2009. Recent evolution of bacterial pathogens: the gall-forming Pantoea agglomerans case. Annual Reviews of Phytopathology 47: 133-152. Bartolo, M. E., Schwartz, H. F., Schweissing, F. C. 1994. Yield and growth response of onion to simulated storm damage. HortScience 29: 1465 – 1467. Bishop, A. L., Davis, R. M. 1990. Internal decay of onions caused by Enterobacter cloacae. Plant Disease 74: 692-694. Bonaterra, A., Marib, M., Casalinib, L., Montesinosa, E. 2003. Biological control of Monilinia laxa and Rhizopus stolonifer in postharvest of stone fruit by Pantoea agglomerans EPS125 and putative mechanisms of antagonism. International Journal of Food Microbiology 84: 93-104. Boyhan, G. E., Kelley, W. T. 2007. Onion Production Guide. UGA Extension. Accessed online June 25, 2014: http://extension.uga.edu/publications/detail.cfm?number=B1198-2 . Bull, C. T., De Boer, S. H., Denny, T. P., Firrao, G., Fischer-Le Saux, M., Saddler, G. S., Scortichini, M., Stead, D. E., Takikawa, Y. 2010. Comprehensive list of names of plant pathogenic bacteria, 1980 – 2007. Journal of Plant Pathology 92: 551 – 592. Butler, S. L., Doherty, C. J., Hughes, J. E., Nelson, J. W., Govan, J. R. W. 1995. Burkholderia cepacia and cystic fibrosis: do natural environments present a potential hazard? Journal of Clinical Microbiology 33: 1001-1004. Costa, E., Teixidó, N., Usall, J., Atarés, E., Viñas, I. 2001. Production of the biocontrol agent Pantoea agglomerans strain CPA-2 using commercial products and by-products. Applied Microbiological Technology 56: 367- 371.
Coutinho, T. A., Venter, S. N. 2009. Pantoea ananatis: an unconventional plant pathogen. Molecular Plant Pathology 10: 325-335. De Baere, T., Verhelst, R., Labit, C., Verschraegen, G., Wauters, G., Claeys, G., Vaneechoutte, M. 2004. Bacteremic infection with Pantoea ananatis. Journal of Clinical Microbiology 42: 4393 – 4395. De Champs, C., Le Seaux, S., Dubost, J. J., Boisgard, S., Sauvezie, B., Sirot, J. 2000. Isolation of Pantoea agglomerans in two cases of septic monoarthritis after plant thorn and wood sliver injuries. Journal of Clinical Microbiology 38: 460-461. De Maayer, P., Chan, W. Y., Venter, S. N., Toth, I. K., Birch, P. R. J., Joubert, F., Coutinho, T. A. 2010. Genome sequence of Pantoea ananatis LMG20103, the causative agent of Eucalyptus blight and dieback. Journal of Bacteriology 192: 2936 – 2937. De Maayer, P., Venter, S. N., Kamber, T., Duffy, B., Coutinho, T. A., Smits, T. H. M. 2011. Comparative genomic of the type VI secretion systems of Pantoea and Erwinia species reveals the presence of putative effector islands that may be translocated by the VgrG and Hcp proteins. BMC Genomics 12:576. Diaz-Perez, J. C., Purvis, A. C., Paulk, J. T. 2002. Bolting, yield, and bulb decay of sweet onion as affected by nitrogen fertilization. Journal of the American Horticultural Society 128:144-149. Dutta, B., Gitaitis, R., Langston, Avci, U., Barman, A., Srinivasan, R. 2012. Acquisition and transmission of Pantoea ananatis and Pantoea agglomerans (causal agents of center rot of onion) by Thrips tabaci (Thysanoptera: Thripidae). Phytopathology S4:88. Dutta, B., Barman, A. K., Srinivasan, R., Avci, U., Ullman, D., Langston, D. B., Gitaitis, R. 2014. Transmission of Pantoea ananatis and Pantoea agglomerans, causal agents of center rot of onion (Allium cepa L.) by onion thrips (Thrips tabaci Lindeman). Phytopathology; doi: 10.1094/PHYTO-07-13-0199-R. Edens, D. G., Gitaitis, R. D., Sanders, F. H., Nischwitz, C. 2006. First report of Pantoea aggomerans causing a leaf blight and bulb rot of onions in Georgia. Plant Disease 90: 1551. Environmental Protection Agency (EPA). Harpin Protein (006477) Fact Sheet. Updated December 8, 2010. Accessed online, January 31, 2011. http://www.epa.gov/pesticides/biopesticides/ingredients/factsheets/factsheet_006477.htm . Enya, J., Shinohara, H., Yoshida, S., Tsukiboshi, T., Negishi, H., Suyama, K., Tsushima, S. 2007. Culturable leaf-associated bacteria on tomato plants and their potential as biological control agents. Microbial Ecology 53: 524 -536. Ferreira, A., Quecine, M. C., Lacava, P. T., Oda, S., Azevedo, J. L., Araujo, W. L. 2008. Diversity of endophytic bacteria from Eucalyptus species seeds and colonization of seedlings by Pantoea agglomerans. Federation of European Microbiological Societies: Microbiology Letters 287: 8 – 14. Frederickson, D. E., Monyo, E. S., King, S. B., Odvody, G. N. 1997. A disease of pearl millet in Zimbabwe caused by Pantoea agglomerans. Plant Disease 81:8.
Gavini, F., Mergaert, J., Beji, A., Mielcarek, C., Izard, D., Kersters, K., De Ley, J. 1989. Transfer of Enterobacter agglomerans (Beijerinck 1888) Ewing and Fife 1972 to Pantoea gen. nov. as Pantoea agglomerans comb. nov. and description of Pantoea dispersa sp. nov. International Journal of Systematic Bacteriology 39: 337-345. Gent, D. H., Schwartz, H. F. 2005. Management of Xanthomonas leaf blight of onion with a plant activator, biological control agents, and copper bactericides. Plant Disease 89:631-639. Gent, D. H., du Toit, L. J., Fichtner, S. F., Mohan, S. K., Pappu, H. R., Schwartz, H. F. 2006. Iris yellow spot virus: an emerging threat to onion bulb and seed production. Plant Disease 90: 1468 – 1480. Gent, D. H., Schwartz, H. F. 2008a. ‘Bacterial Stalk and Leaf Necrosis’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 66 - 67. Gent, D. H., Schwartz, H. F. 2008b. ‘Center rot’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 64 - 66. Giller, K. E., Witter, E., McGrath, S. P. 1998. Toxicity of heavy metals to microorganisms and microbial processes in agricultural soils: a review. Soil Biology and Biochemistry 30: 1389 – 1414. Gitaitis, R., Walcott, R., Culpepper, S., Sanders, H., Zolobowska, L., Langston, D. 2003. Recovery of Pantoea ananatis, causal agent of center rot of onion, from weeds and crops in Georgia, USA. Crop Protection 21: 983-989. Gitaitis, R. D., Walcott, R. R., Sanders, H. F., Zolobowska, L., Diaz-Perez, J. C. 2004. Effects of Mulch and Irrigation System on Sweet Onion: II. The Epidemiology of Center Rot. Journal of the American Society of Horticultural Science 129: 225-230. Goszczynska, T., Moloto, V. M., Venter, S. N., Coutinho, T. A. 2006. Isolation and identification of Pantoea ananatis from onion seed in South Africa. Seed Science and Technology 34: 655-668. Herrera Medina, M. J., Gagnon, H., Piche, Y., Ocampo, J. A., Garcia Garrido, J. M., Vierheilig, H. 2003. Root colonization by arbuscular mycorrhizal fungi is affected by the salicylic acid content of the plant. Plant Science 164: 993-998. Hoepting, C. A. Cornell University extension educator and onion specialist. Personal conversations, April 2012 – present. Kim, H. J., Lee, J. H., Kang, B. R., Rong, X., McSpadden Gardener, B. B., Ji, H. J., Park, C-S., Kim, Y. C. 2012. Draft genome sequence of Pantoea ananatis B1-9, a nonpathogenic plant growth-promoting bacterium. Journal of Bacteriology 194: 729. Lang, J. M., Gent, D. H., and Schwartz, H. F. 2007. Management of Xanthomonas leaf blight of onion with bacteriophages and a plant activator. Plant Disease 91:871-878. Lee, H.B., Hong, J. P., Kim, S. B. 2010. First report of leaf blight caused by Pantoea agglomerans on rice in Korea. Plant Disease 94: 1372.
18
Liu, J., Maldonado-Mendoza, I., Lopez-Meyer, M., Cheung, F., Town, C. D., Harrison, M. J. 2007. Arbuscular mycorrhizal symbiosis is accompanied by local and systemic alterations in gene expression and an increase in disease resistance in the shoots. The Plant Journal 50: 529-544. Liu, Y-H., Huang, C-J., Chen, C-Y. 2008. Evidence of induced systemic resistance against Botrytis elliptica in lily. Phytopathology 98: 830 – 836. Louws, F. J., Wilson, M., Campbell, H. L., Cuppels, D. A., Jones, J. B., Shoemaker, P. B., Sahin, F., Miller, S. A. 2001. Field control of bacterial spot and bacterial speck of tomato using a plant activator. Plant Disease 85:481-488. Ma, B., Hibbing, M. E., Kim, H.-S., Reedy, R. M., Yedidia, I., Breuer, J., Breuer, J., Glasner, J. D., Perna, N. T., Kelman, A., Charkowski, A. O. 2007. Host range and molecular phylogenies of the soft rot enterobacterial genera Pectobacterium and Dickeya. Phytopathology 97: 1150 - 1163. Maher, E. A., Kelman, A. 1983. Oxygen status of potato tuber tissue in relation to maceration by pectic enzymes of Erwinia carotovora. Phytopathology 73: 536 – 539. Mann, J. D. 1983. Translocation of photosynthate in bulbing onions. Australian Journal of Plant Physiology 10: 515-521. Membre, J. M., Burlot, P. M. 1994. Effects of temperature, pH, and NaCl on growth and pectinolytic activity of Pseudomonas marginalis. Applied Environmental Microbiology 60: 2017 – 2022. Mergaert, J., Verdonck, L., Kersters, K. 1993. Transfer of Erwinia ananas (synonym, Erwinia uredovora) and Erwinia sterwartii to the Genus Pantoea emend. as Pantoea ananas (Serrano 1928) comb. nov. and Pantoea stewartii (Smith 1898) comb. nov., respectively, and description of Pantoea stewartii subsp. indologenes subsp. nov. International Journal of Systematic Bacteriology 43: 162-173. Mohan, S. K. 2008a. ‘Other Bacterial Soft Rots’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 62. Morales-Valenzuela, G., Silva-Rojas, H. V., Ochoa-Martinez, D., Valadez-Moctezuma, E., Alarcon-Zuniga, B., Zelaya-Molina, L. X., Cordova-Tellez, L., Mendoza-Onofre, L., Vaquera-Huerta, H., Carballo-Carballo, A., Farfan-Gomez, A., Avila-Quezada, G. 2007. First report of Pantoea agglomerans causing leaf blight and vascular wilt in maize and sorghum in Mexico. Plant Disease 91: 1365. Morohoshi, T., Nakamura, Y., Yamazaki, G., Ishida, A., Kato, N., Ikeda, T. 2007. The plant pathogen Pantoea ananatis produces N-acylhomoserine lactone and causes center rot disease of onion by quorum sensing. Journal of Bacteriology 189: 8333-8338. Morse, J. G., Hoddle, M. S. 2006. Invasion biology of thrips. Annual Review of Entomology 51: 67-89. Nair, C. B., Anith, K. N., Sreekumar, J. 2007. Mitigation of growth retardation effect of plant defense activator, acibenzolar-S-methyl, in amaranthus plants by plant growth-promoting rhizobacteria. World Journal of Microbiological Technology 23: 1183-1187.
19
Nasser, W., Reverchon, S., Vedel, R., Boccara, M. 2005. PecS and PecT coregulate the synthesis of HrpN and pectate lyases, two virulence determinants in Erwinia chrysanthemi 3937. Molecular Plant-Microbe Interactions 18:1205-1214. Nault, B. Cornell University research and extension entomologist. Personal conversation, February 2012. Nischwitz, C., Gitaitis, R., Sanders, H., Langston, D., Mullinix, B., Torrance, R., Boyhan, G., Zolobowska, L. 2007. Use of fatty acid methyl ester profiles to compare copper-tolerant and copper-sensitive strains of Pantoea ananatis. Phytopathology 97: 1298-1304. Obradovic, A., Jones, J. B., Momol, M. T., Olson, S. M., Jackson, L. E., Balogh, B., Guven, K., Iriarte, F. B. 2005. Integration of biological control agents and systemic acquired resistance inducers against bacterial spot on tomato. Plant Disease 89:712-716. Oregon State Vegetable Production Guides. 2004. ‘Dry Bulb Onions.’ Accessed online June 26, 2014: http://nwrec.hort.oregonstate.edu/onionb-w.html#top . Orloff, S., Natwick, E. T., Poole, G. J. 2013. Onion and Garlic Guidelines: Insects and Mites. UC-IPM Online: Statewide Integrated Pest Management Program. http://www.ipm.ucdavis.edu/PMG/pmgauthors.html . Pennsylvania Department of Public Welfare. ‘Onions.’ PA Nutrition Education Network: Volume I, Issue 16. Accessed online May 16, 2014: http://www.panen.org/sites/default/files/SNAC%20Materials/onion_newsletter2.pdf Pennsylvania Vegetable Growers Association (PVGA). Updated December 3, 2010, under ‘Services.’ Accessed online January 25, 2011: http://www.pvga.org/ . Poliakovskiy, S. O., Dmitriev, O. P. 2011. Study of priming for callose accumulation in Allium cepa during treatment with biotic inducers. Cytology and Genetics: 245 - 248. Pusey, P. L., Stockwell, V. O., Rudell, D. R. 2008. Antibiosis and acidification by Pantoea agglomerans strain E325 may contribute to suppression of Erwinia amylovora. Phytopathology 98: 1136 – 1143. Records, A. R. 2011. The type VI secretion system: a multipurpose delivery system with a phage-like machinery. Molecular Plant-Microbe Interactions 24: 751 – 757. Romero, A. M., Kousik, C. S., Ritchie, D. F. 2001. Resistance to bacterial spot in bell pepper induced by acibenzolar-S-methyl. Plant Disease 85:189-194. Schroeder, B. K., Waters, T. D., du Toit, L. J. 2010. Evaluation of onion cultivars for resistance to Enterobacter cloacae in storage. Plant Disease 94:236-243. Schwartz, H. F., Otto, K. L., Gent, D. H. 2003. Relation of temperature and rainfall to development of Xanthomonas and Pantoea leaf blights of onion in Colorado. Plant Disease 87:11-14. Schwartz, H. F., Mohan, S. K., Havey, M. J., Crowe, F. J. 2008. Introduction in: Compendium of Onion and Garlic Diseases. Pp. 1 – 7. APS Press: St. Paul, MN.
Shelton, A. M., Zhao, J. Z., Nault, B. A., Plate, J., Musser, F. R., Larentzaki, E. 2006. Patterns of insecticide resistance in onion thrips (Thysanoptera: Thripidae) in onion fields in New York. Journal of Economic Entomology 99: 1798 - 1804. Smith, R., Biscaro, A., Cahn, M., Daugovish, O., Natwick, E., Nunez, J., Takele E., Turini, T. 2011. Fresh-market bulb onion production in California. UC Vegetable Research and Information Center. Publication 7242. Smith, D. D. N., Kirzinger, M. W. B., Stavrinides, J. 2013. Draft genome sequence of the antibiotic-producing epiphytic isolate Pantoea ananatis BRT175. American Society for Microbiology Genome Announcements 1: 1-2. Spoel, S. H., Johnson, J. S., Dong, X. 2007. Regulation of tradeoffs between plant defenses against pathogens with different lifestyles. Proceedings of the National Academy of Sciences 104: 18842 – 18847. Stoltzfus, J. Eastern Lancaster County School District Adult Educator. Personal communication April 2011 – present. United States Department of Agriculture (USDA) – National Agricultural Statistics Service. Statistics by Subject. Accessed online (Jan. 29, 2013): http://www.nass.usda.gov/Statistics_by_Subject/index.php?sector=CROPS . University of Minnesota. ‘Introduction to Thrips.’ Accessed online November 2011- January 2012: http://www.entomology.umn.edu/cues/inter/inmine/index.html . Van Loon, L. C. 2007. Plant responses to plant growth-promoting rhizobacteria. European Journal of Plant Pathology 119: 243-254. van Wees, S. C. M., de Swart, E. A. M., van Pelt, J. A., van Loon, L. C., Pieterse, C. M. J. 2000. Enhancement of induced disease resistance by simultaneous activation of salicylate- and jasmonate-dependent defense pathways in Arabidopsis thaliana. Proceedings of the National Academy of Sciences 97: 8711-8716. Volksch, B., Thon, S., Jacobsen, I.D., Gube, M. 2009. Polyphasic study of plant and clinic associated Pantoea agglomerans reveals indistinguishable virulence potential. Infection, Genetics, and Evolution 9:1381-1391. Walcott, R. R., Gitaitis, R. D., Castro, A. C., Sanders, F. H., Jr., and Diaz-Perez, J. C. 2002. Natural infestation of onion seed by Pantoea ananatis, causal agent of center rot. Plant Disease 86:106-111. Walters, D. R., Fountaine, J. M. 2009. Practical application of induced resistance to plant diseases: an appraisal of effectiveness under field conditions. Journal of Agricultural Science 147: 523 – 535. Walters, D., Walsh, D., Newton, A., Lyon, G. 2005. Induced resistance for plant disease control: maximizing the efficacy of resistance elicitors. Phytopathology 95:1368-1373. Wei, Z. M., Laby, R. J., Zumoff, C. H., Bauer, D. W., He, S. Y., Collmer, A., Beer, S. V. 1992. Harpin, elicitor of the hypersensitive response produced by the plant pathogen Erwinia amylovora. Science 257: 85-88.
Wells, M. L., Gitaitis, R. D., Sanders, F. H. 2002. Association of tobacco thrips, Frankliniella fusca (Thysanoptera: Thripidae), with two species of bacteria of the genus Pantoea. Annals of the Entomological Society of America 95: 719-723. Wright, P. J., Hale, C. N. 1992. A field and storage rot of onion caused by Pseudomonas marginalis. New Zealand Journal of Crop and Horticultural Science 20: 435 – 438. Wright, S. A. I., Zumoff, C. H., Schneider, L., Beer, S. V. 2001. Pantoea agglomerans strain EH318 produces two antibiotics that inhibit Erwinia amylovora in vitro. Applied Environmental Microbiology 67: 284 – 292. Yamamoto, S., Kasai, H., Arnold, D. L., Jackson, R. W., Vivian, A., Harayama, S. 2000. Phylogeny of the genus Pseudomonas: intrageneric structure reconstructed from the nucleotide sequences of gyrB and rpoD genes. Microbiology 146: 2385 – 2394. Yishay, M., Burdman, S., Valverde, A., Luzzatto, T., Ophir, R., Yedidia, I. 2008. Differential pathogenicity and genetic diversity among Pectobacterium carotovorum ssp. carotovorum isolates from monocot and dicot hosts support early genomic divergence within this taxon. Environmental Microbiology 10: 2746 – 2759. Zook, D. PA onion grower and board member of Lancaster County onion cooperative. Personal communication April 2011 – present.
22
Chapter 2: Sources of bacterial inoculum in and around transplanted onion (Allium cepa) fields in a plastic mulch production system
Abstract
Exclusion of pathogen inoculum from the production system is one of the foremost strategies for
managing bacterial diseases, particular in vegetable production. To determine the prevalent bacterial
pathogens of onion in PA, where growers may suffer yield losses of up to 60% due to several rot diseases,
and help define probable inoculum sources, extensive surveys were undertaken on 28 farms in 2011 and
26 farms in 2012, respectively. Environmental data and plant samples were gathered three times through
the growing season each year, with PCR-facilitated bacterial species detections in extracted samples as
well as isolations of viable bacteria from multiple sources. Two types of pathogenicity tests were
conducted with selected isolates from each potential source. Pathogenic isolates of Pectobacterium
carotovorum subsp. carotovorum, Pantoea agglomerans, and Pseudomonas marginalis pv. marginalis, the
most common pathogens in the PA onion cropping system, occurred in soil, transplant and weed tissues,
in addition to being present on the surfaces of both onion transplants and common weeds. More specific
associations were suggested between P. marginalis and transplants as well as P. carotovorum and P.
agglomerans and weeds. In addition, rep-PCR facilitated strain tracking of Pantoea ananatis indicated
matching strains isolated from surface-disinfested weed tissue collected at mid-season and a latently
infected onion that had been stored for four months. These results suggest research avenues for more
directed disease management strategies for reducing the impact of environmental inoculum sources on
bacterial rots of onion.
Introduction
Exclusion of pathogens from cropping systems is one of the primary methods for the management
of plant diseases. With low efficacy of chemical control methods, reducing primary inoculum is a
cornerstone of effective bacterial disease management in vegetables (Gitaitis et al., 1992; De Leon et al.,
2011). In order to successfully exclude pathogens, sources of bacterial inoculum must first be known.
Examples of exclusionary pathogen control strategies include the surface-sterilization of vegetative
propagation tools in managing TMV in petunia (Lewandoski et al., 2012) and tomato seed-cleaning
techniques in the case of Clavibacter michiganensis subsp. michiganensis (Ivey and Miller, 2004).
Bacterial rots of onion (Allium cepa L.) are the most significant diseases affecting commercial
production in the Mid-Atlantic and Northeast regions. Up to nine different species of bacteria have been
23
shown to induce rot of onions in Pennsylvania, several of which are ubiquitous and inhabit multiple
environmental niches, in addition to causing crop disease. These include the center rot pathogens,
Pantoea ananatis Serrano and Pantoea agglomerans Beijerinck (Pagg), the soft rot pathogens,
Pectobacterium carotovorum subsp. carotovorum Jones (Pcar) and Pseudomonas marginalis pv.
marginalis Stevens (Pmar), and the slippery and sour skin pathogens, Burkholderia gladioli pv. alliicola
Burkholder and Burkholderia cepacia Burkholder, respectively. In addition, the leaf pathogens
Pseudomonas viridiflava Burkholder and Xanthomonas axonopodis pv. allii Hasse, and the storage decay
pathogen Enterobacter cloacae Jordan (Bull et al., 2010), also occur in the pathosystem, albeit
infrequently. It is not unusual to detect and isolate as many as four different pathogenic bacterial species
from a single symptomatic onion, and particular species typically do not induce diagnostic symptoms.
Based on previous work, different sources of bacterial inoculum have been suggested in onion
pathosystems, including B. cepacia from soil (Coenye and Vandamme, 2003), and P. ananatis and P.
viridiflava from the surfaces of perennial weeds (Gitaitis et al., 1998; 2002). In PA, the majority of onions
grown are transplanted into the field; transplants are produced in the southern U.S. or increasingly, locally
in greenhouses as plug plants. At least one of the previously mentioned pathogens, P. ananatis, has been
shown to be naturally seed-transmitted (Walcott et al., 2002), and since shipped transplants are injured
at least twice in the forms of root and leaf trimming, southern-U.S. transplants may potentially serve as
an inoculum source. Transplants have previously been shown to be sources of bacterial inoculum in other
plant pathosystems (Cuppels and Elmhirst, 1999; Gitaitis et al., 1992).
PCR-based detection strategies exist for several of the plant pathogens referenced above (Walcott
et al., 2002; De Paula Lana et al., 2012; Beer, pers. comm.), however, these current strategies only resolve
to the species level. Genomic fingerprinting methods have been employed as subspecies and strain-level
detection strategies in both environmental microbiology (Albert et al., 2003; Ishii and Sadowsky, 2009)
and plant pathology research (Louws et al., 1994; Lange et al., 2006). In particular, Lange et al. used rep-
PCR to track bacterial strains of Xanthomonas campestris pv. campestris, the black rot pathogen of
cabbage, from cruciferous weeds into diseased cabbage crops (2006). The efficacy of using genomic
fingerprinting for bacterial strain tracking, however, is reliant on several factors, including the number of
bands amplified, the analytical methods used, and the inherent genetic diversity in the target species
(Albert et al., 2003).
To identify potential sources of bacterial inoculum in and around sweet onion fields, extensive
surveys were undertaken on 28 and 26 farms in 2011 and 2012, respectively. Using a multiplex PCR
protocol, detections of eight species of bacteria were sought in soil, on or in transplants and weeds, and
24
in symptomatic onions obtained at harvest and during storage. Simultaneously, bacterial isolations from
the samples resulted in a database of over 2500 bacterial isolates from PA from two sample years; select
isolates were then assayed for their ability to induce disease in two different pathogenicity tests as well
as used in a rep-PCR based strain tracking protocol.
Materials and Methods
Sample collection: In 2011 and 2012, 28 and 26 farms, respectively, were visited three times each over
the course of the onion growing season. Farms were selected with the assistance of regional extension
educators; all PA growers grew ‘Candy’ onions (a sweet, yellow cultivar intended for fresh eating) on
raised beds with plastic mulch and two rows of drip irrigation. Fields ranged between 0.25 - 1 hectare in
size, and were actively crop rotated. Some PA growers had been growing onions for as many as ten years,
however, a first-year grower was included in each survey year. There were 24 repeat growers between
both years, but the fields where the onions were grown were not the same, with the exception of two
farms.
Sample visit 1: In April/May of each year, characteristics of each production field were recorded, including
type of plastic mulch used, spacing between plants, approximate field size, general fertilizer information,
and grower approaches to disease and weed management. Three 9.14 m plots were established by
flagging an area of the field visually estimated to be representative of the entire field with respect to field
length, width, and topography. Each plot was one bed wide, with typical bed width of approximately one
meter. Composite soil samples were collected from each plot (six samples equidistantly through the plot,
using a soil probe 2.54 cm in diameter to a depth of 7.6 cm for each sample). Composite samples of
approximately 30 onion plants, cultivar ‘Candy,’ were collected from growers prior to transplanting into
the field.
Sample visit 2: In approximately the middle of the growing season (second and third weeks of June in PA)
each year, five different, prevalent weeds were sampled from within or in close proximity to one of the
plots established in visit 1, placed in individual sealed plastic bags, and stored on ice, transported to the
lab, and then stored at 2⁰C until further processing. The time from collection until further processing was
generally within seven days.
25
Sample visit 3: Visits were timed as closely as possible to when growers began harvesting their full crop.
Soil was sampled according to the protocol described in visit 1, and the inner two rows of the inner 4.6 m
of each plot were harvested, graded by size, and evaluated for bacterial disease incidence (approx. 60
bulbs). Onions were graded into four size categories based on the bulb diameter: < 6.4 cm diameter are
graded small, 6.4-7.6 cm diameter are graded medium, 7.6-10.2 cm diameter are graded jumbo, and >10.2
cm diameter are graded colossal. In 2011 and 2012, 20% and 50% of the symptomatic bulbs at harvest,
respectively, were retained for further processing. Approximately 30 jumbo-size, asymptomatic bulbs
were retained for postharvest storage evaluation. If 30 jumbo-size, asymptomatic bulbs were not
available, 30 representatively-sized bulbs from the plot were sampled.
Postharvest bulb ratings
In both years, approximately 30 asymptomatic, jumbo-size onions per replicate plot were cured
under burlap in a greenhouse with forced air for at least 72 hours, then placed in 4⁰C storage for 75-120
days (depending on harvest date). In mid-late November of the harvest year, approximately four months
postharvest, the bulbs were sliced in half longitudinally, photographed, and evaluated for disease
incidence, denoted as a percentage of the 30 bulbs per replicate plot. In 2011 and 2012, 20% and 50% of
the symptomatic bulbs from storage, respectively, were retained for further processing.
Bacterial isolations from soil
Seven months (2011) or one week (2012) after soil sampling, one gram of soil was shaken for one
hour in phosphate buffered saline (Kphos buffer: 0.05M K2HPO4, 0.5M KH2PO4, 0.85% NaCl, pH 7.0), then
the supernatant was serially diluted onto King’s B medium (KB; King et al., 1954) to obtain single colonies.
After selecting the dilution, 100 µL was spread-plated on KB medium in 2011 or onion extract medium
(OEM; Zaid et al., 2012) in 2012. Plates were incubated for two days in a laboratory at ambient (21 – 23⁰C)
temperature, after which single colonies were streaked onto KB medium and allowed to grow 1 – 2 days.
Each isolate was numbered, photographed, and its source material and location were recorded. Isolates
were stored by individually transferring a single colony from these plates to sterile glass test tubes
containing 4 mL Luria-Bertani broth (LB; Difco Luria-Bertani broth, Miller, Becton, Dickinson and Co.,
Sparks, MD, USA) and incubating at 30⁰C with shaking (150 rpm) overnight. Sterile glycerol was added to
log-phase broth cultures to establish a 15% concentration of glycerol in cryovials (Denville Scientific,
Metuchen, NJ, USA), which were then maintained at -20⁰C. Each isolate was frozen in duplicate, and boxes
26
of frozen cultures were maintained in two separate locations. Leftover soil samples were retained in their
labeled bags at 4⁰C.
Bacterial isolations from transplants and weeds
Within eight days of sampling, roots were removed from a sample of transplants (approximately
25) and agitated in 200 mL sterile Kphos buffer (0.1% Tween-20) in 500 mL or 1 L flasks for 1 hr at 150
rpm. Aliquots of 50 mL were removed for DNA extraction, bacterial isolation, and archiving at -20⁰C. After
identification of the weed to genus (and species, if possible, using Uva et al., 1997), roots were removed
from each weed sample, then samples were individually combined with one mL sterile Kphos buffer / g
fresh weight plant tissue with 0.1% Tween-20, then agitated in sterile 500 mL or 1 L flasks for 1 hr at 150
rpm.
To obtain epiphytic bacteria from onion transplants and weeds, buffer fractions were serially
diluted, plated on KB to select the appropriate dilution to grow 20 - 200 colonies (typically 10-5 – 10-7),
then 100 µL was spread-plated on KB medium (transplants and 2011 weeds) or OEM (2012 weeds).
Bacterial isolations were conducted following the soil isolation protocol described above, to obtain
isolates archived in 15% glycerol and maintained at -20⁰C. Unused portions of epiphytic Kphos buffer
washes were archived at -20⁰C.
On the day of fractionation of the epiphytic sample, endophytic bacteria from transplant and
weed samples were obtained by washing the plants twice in 70% ethanol, then combining with sterile
Kphos buffer to the equivalent of 1 mL / g fresh tissue. Plant samples were then homogenized in a Waring
blender for 1 min. Aliquots of plant slurries were transferred to three sterile 50 mL centrifuge tubes for
DNA extraction, bacterial isolation, and archiving at -20⁰C. To obtain bacterial isolates, one mL of
transplant slurry was serially diluted, plated on KB to select the appropriate dilution, then 100 µL of the
appropriate dilution was spread on KB medium (transplants) or OEM (2012 weeds) using the protocol
described above and maintained at -20⁰C. The Waring blender was rinsed with water, then 70% ethanol
before homogenizing each plant sample.
Bacterial isolations from harvest and postharvest symptomatic onions
After each of the harvest and postharvest ratings, 20% or one bulb (whichever was greater, in
2011), or 50% or one bulb (whichever was greater, in 2012) of the symptomatic onions were
photographed and approximately 0.5 g of tissue was removed from the margin between symptomatic and
asymptomatic areas. Duplicate samples were removed from the same area and each was transferred to
27
sterile 1.8 mL Eppendorf tubes; one sample was used in the DNA extraction protocol and the other sample
was used for bacterial isolations. The isolation sample was ground with a sterile micropestle with an
additional 500 µL of sterile Kphos buffer, then serially diluted onto KB. Then 100 µL of the appropriate
dilution was spread-plated on KB to select single isolates to be maintained in duplicate (as described
above) at -20⁰C.
Semi-anaerobic pathogenicity tests
Isolates generated from samples from 2011 and 2012 were revived from -20⁰C storage by plating on KB
medium and incubating at 30⁰C for 1-3 days, until colonies were visible. A sterile toothpick was used to
transfer a single colony from the KB plate to a sterile tube containing 4 mL LB broth, then tubes were
placed in a 30⁰C incubator with agitation (150 rpm) overnight. The following day, log-phase broth culture
of each isolate was used to inoculate surface-sterilized (10% NaClO) and completely dry, symptomless,
yellow pearl onions purchased from local grocery stores. Sterile needles and syringes were used to
administer approx. 100 µL log-phase LB broth to two pearl onions by inserting the needle approx. halfway
through the bulb midway between basal plate and neck. These inoculated onions were placed in a
sterilized 200 mL screw-top
jar, one atop the other, and
the lid was screwed closed for
semi-anaerobic pathogenicity
test. These jars were incubated
on a lab benchtop at ambient
room temperature (21 – 23⁰C)
for fourteen days, then pearl
onions were halved on the
inoculation point and rated for
symptom development (0 or 1;
absence or presence; Figs. 2.1
a, b).
Aerobic pathogenicity tests
For aerobic pathogenicity tests, one-half of a surface disinfested, yellow pearl onion was
inoculated by administering approx. 100 µL log-phase LB broth culture through the side of the onion half,
a b
Figs. 2.1a-b. Results of semi-anaerobic pathogenicity tests using nonpathogenic (a) and pathogenic (b) strains.
28
parallel to the full slice. The other half of the onion was left uninoculated. Both halves of the onion were
placed in a sterile glass Petri dish with filter paper that had been moistened by sterile water. Two halved
bulbs per aerobic test were placed in each dish, but bulbs were duplicate pathogenicity tests of the same
isolate. Aerobic tests were incubated on a lab benchtop at ambient room temperature (21 – 23⁰C) for
seven days, then pearl onions were rated (1 – 5) for symptom development; an example of each rating is
shown in Fig. 2.2.
DNA extractions
DNA was extracted directly from soil, transplant epiphyte and endophyte samples, weed epiphyte
and endophyte samples, and symptomatic onion tissue. Soil extractions were completed using a 1 g
sample from the homogenized bag of soil within three weeks of sampling; this was dried for 48 hr in a
forced-air drying oven at 65⁰C. The “Experienced User Protocol” included in the MoBio Ultra Soil DNA kit
(MoBio Laboratories, Inc., Carlsbad, CA) was directly followed, using the vortex attachment option.
Transplant epiphytic and endophytic samples, weed epiphytic samples, and symptomatic harvest
and storage onion samples were processed within two weeks of sampling using the Wizard Genomic DNA
Fig. 2.2. Severity of symptom development in pathogenicity tests (duplicate) under aerobic incubation after seven days. 1 = nonpathogenic; 2 = local discoloration or maceration; 3 = local symptom, plus discoloration or maceration on adjacent scale(s); 4 = up to half of bulb discolored or macerated; 5 = more than 50% of bulb discolored or macerated. The right half of each bulb is inoculated; bulbs were rated individually and average was reported if not identical.
29
Purification kit (Promega, Madison, WI). For epiphytic samples, 50 mL surface wash was centrifuged at
5000 rpm, then the pellet was resuspended in 1 mL sterile Kphos buffer. Transplant endophytic samples
(800 µL) and 0.5 g symptomatic onion tissue were lyophilized for 4 hr at -40⁰C with vacuum (<150 mTorr),
ground, and extracted for DNA using the Promega DNA extraction kit outlined below. Briefly, 600 µL of
Promega Nuclei Lysis Solution was combined with the ground dried tissue, vortexed, and incubated for 30
min at 65⁰C. Next, 200 µL of 10M ammonium acetate (NH4CH3COO) was added, vortexed, and frozen at
-20⁰C for 10 min. Tubes were next centrifuged for 5 min at 13,000 x g, then supernatants were transferred
to sterile tubes with 600 µL isopropanol, which were inverted several times then centrifuged for 5 min at
13,000 x g. Supernatants were decanted, then the pellet was washed with 600 µL 70% ethanol and
centrifuged 1 min at 13,000 x g. The 70% ethanol was decanted and the pellet was dried for 1 hr, followed
by resuspension of the pellet in 100 µL TE buffer (10 mM Tris-HCl containing 1 mM EDTA-Na2, pH 8.0).
Weed endophytic samples were processed within two weeks of sampling using the Qiagen
DNEasy Plant Mini Kit (Qiagen Inc., Valencia, CA). Starting material was 800 µL weed slurry, which was
lyophilized for 4 hr at -40⁰C with vacuum (<150 mTorr), ground, then procedures in the Qiagen kit were
followed without modification.
Duplex PCR for species identification
Bacterial isolates were preliminarily identified to genus based on photographed phenotypic
characteristics. In general, Pseudomonas spp. fluoresce on KB medium, while Pantoea spp. and
Burkholderia spp. appear yellow, and P. carotovorum pv. carotovorum appears more cream-colored. Since
Pantoea spp. are typically more common in PA (Pfeufer, unpublished), phenotypically yellow colonies
were initially screened using the P. ananatis / P. agglomerans primer pairs. Based on these preliminary
identifications, sets of isolates were screened with species-specific primers using an established protocol
(Mansfield and Gugino, 2010). Primers were used in duplex reactions, with the following primer sets
combined and reactions on a thermocycling program with their respective annealing temperature (P.
ananatis and P. agglomerans, 60⁰C; B. gladioli pv. alliicola and B. cepacia, 60⁰C; P. carotovorum pv.
carotovorum and X. axonopodis pv. allii, 62⁰C; P. marginalis pv. marginalis and P. viridiflava, 62⁰C).
Template was approximately 1.5 µL removed directly from the 15% glycerol stock culture from -20⁰C
storage. Other reaction components were 12.5 µL TaqPRO Complete master mix (Denville Scientific,
Metuchen, NJ), 1 µL species 1 forward primer (10 mM), 1 µL species 2 forward primer (10 mM), 1 µL
species 1 reverse primer (10 mM), 1 µL species 2 reverse primer (10 mM), and 7.5 µL PCR water. PCR
conditions were as follows: (1) initial denaturation at 96°C for 10 min, followed by 34 cycles of (2) 99°C
30
for 30 s; (3) 60°C or 62°C for 1 min, (4) 72°C for 1 min; then a final elongation at (5) 72°C for 5 min and
storage at (6) 4°C continuous. Amplification was confirmed with a 1.5% agarose gel electrophoresis.
Specific fragment sizes were as follows: B. gladioli pv. alliicola, 752 bp; X. axonopodis pv. allii, 715 bp; P.
marginalis pv. marginalis, 651 bp; B. cepacia, 475 bp; P. carotovorum pv. carotovorum, 322 bp; P.
agglomerans, 248 bp; P. viridiflava, 181 bp; P. ananatis, 166 bp (Mansfield and Gugino, 2010).
rep-PCR for bacterial strain tracking
Pantoea spp. isolates were revived from -20⁰C storage by plating on KB medium and incubating
24 – 72 hr at 30⁰C. Two loopfuls of bacterial culture from the plate were suspended in 40 µL of sterile PCR
water in a 500 µL Eppendorf tube and vortexed. Tubes were placed in a boiling water bath for 5 min, then
centrifuged for 3 min at 5000 rpm. Repetitive extragenic palindromic-PCR (rep-PCR) protocols were
carried out as described by Versalovic et al. (1994) with the following modifications. Template (1.5 µL)
was pipetted from directly above the pellet from the boiled and centrifuged tubes. Autoclaved, membrane
filtered (0.22 μm) H2O was was reduced by 0.5 µL. Reaction products (12 µL) were mixed with 1.5 µL 6x
EZ-Vision Three dye (AMRESCO LLC, Solon, OH, USA), electrophoresced for 17 hr in a 2% SeaChem agarose
gel (Lonza Corporation, Visp, Switzerland) with two lanes containing 10-kb ladder (Denville Scientific,
Metuchen, NJ, USA).
Data analysis
Nonparametric comparisons
Comparisons of the frequencies of pathogenic and nonpathogenic isolates were conducted using
Fisher’s exact test (for 2x2 tables) and the Mantel-Haenszel-Cochran test (for experiment-wise tables in
which three species were tested). Comparisons of severity distributions were conducted using the Kruskal-
Wallis test (for comparisons between two species) and the Mann-Whitney test (for comparisons between
three species) in Minitab 16 (Minitab, Inc., State College, PA, USA).
Logistic regressions
Logistic regressions were conducted where selected species presence or absence (1, 0) in
symptomatic bulbs from harvest and storage was the dependent variable. Independent variables were
the presence or absence of the same species in soil, transplant epiphytic, transplant endophytic, weed
epiphytic, and weed endophytic samples, with farm of origin included as a factor. Weed epiphytic and
endophytic samples were rated as none (no detections of the selected species), some (1 or 2 detections
31
of the selected species out of 4 or 5 samples processed), or many (3, 4, or 5 detections of the selected
species out of 4 or 5 samples processed), with these species prevalence variables recoded as binary (0 if
false, 1 if true) independent variables in the analysis.
Odds ratios were calculated using SAS 9.2 (SAS, Cary, NC, USA); odds ratios approximate how likely
a binary outcome is to occur by relative risk. To interpret odds ratios, first the difference between the
point estimate and the modeled value, 1, is determined, then this difference indicates the relationship
between the independent variable and the dependent, binary variable. Point estimates below 1 indicate
a negative relationship between independent and dependent variables, while point estimates above 1
indicate positive relationships. The deviation from the modeled value indicates the relative change (for
each independent variable) in odds of detecting the chosen pathogen in each symptomatic onion sample
(Hosmer and Lemeshow, 1989). Variables were chosen using reverse selection in PROC LOGISTIC in SAS
9.2 (α to remove = 0.05). Only odds ratios for statistically significant (P ≤ 0.05) independent variables are
reported in the tables presented here.
rep-PCR fingerprinting
Preliminary analyses using three previously-used genomic fingerprinting primer sets, the
Lending support to these data are logistic regressions of the detection of Pmar in symptomatic onion bulbs
from harvest and storage in 2011 and 2012, in which the odds of detecting Pmar in bulbs substantially
increased if Pmar was detected in either epiphytic (Table 2.3a; 2011) or endophytic (Table 2.3c; 2012)
samples from transplants.
In terms of management, these results suggest disease could be initiated by bacterial epiphytes
or endophytes from plants, and some species may be better suited to survival or transfer via transplants.
As in most agricultural production, growers should avoid planting transplants with noticeable symptoms
of disease. However, as has been suggested with seed-transmitted bacterial blight of onion (caused by X.
axonopodis pv. allii) inoculum may be spread via asymptomatic, latently-infected planting material
(Humeau et al., 2006). Since latent bacterial pathogens may induce symptoms later as a result of host
36
stress or changes in environmental conditions, previously described production practices that minimize
bacterial rots of onion, such as the use of plastic mulches that decrease soil temperatures (Chapter 4;
Gugino et al., 2011), might be encouraged among growers as management options to decrease bacterial
disease due to transplant-associated bacteria. Other options may be surface sterilizing or biological
control-based transplant dips, which may eliminate or compete with pathogenic bacterial populations on
onion foliage, as had been demonstrated on soybean by putative niche-displacing strains of gram-negative
bacteria (May et al., 1997). Finally, regulatory screening of transplants from southern states may be
suggested in order to eliminate the importation of infected plants, though accurate, rapid, and cost
effective sampling protocols need to be designed before regulations are imposed (Gitaitis et al., 1992).
As suggested by pathogenicity tests of Pagg, Pcar, and Pmar isolates from the surfaces and tissue
of weeds, a spectrum of virulence exists weed-derived members of these bacterial species, ranging from
nonpathogenic to highly aggressive (Figs. 2.6a-c). Weeds may play a role in onion diseases by supporting
a wide diversity of strains of these bacteria, and in dilution assays, at least six sampled weeds, including
lambsquarters, redroot pigweed, and purslane, can support endophytic populations of up to 108 CFU / g
fresh weight tissue (data not shown). Perennial weeds have already been demonstrated to harbor
potential bacterial pathogens of onion in Georgia (Gitaitis et al., 2002). PA weeds likely differ in terms of
prevalence, emergence, and biotype, all of which may affect the plants’ abilities to host these bacterial
pathogens, in addition to climatic differences between the two states, which may affect the survival and
colonization of bacteria associated with weeds. Though pathogenic epiphytic Pagg were frequently
isolated from the surfaces of common weeds, logistic regression of the 2012 dataset indicated that when
Pagg was prevalently detected on the surfaces of midseason weeds, the odds of detecting this bacterial
species decreased in symptomatic bulbs from the same farms (Table 2.3d; 62% of 2012 fields were
identified with Pagg as a prevalent weed epiphyte). Together with the logistic regression from 2011 soils
(Table 2.3b), these results reiterate that environmental, potentially nonpathogenic, Pagg strains may
negatively influence potentially onion-pathogenic Pagg strains. These results are based on PCR detection
only, however, and to more fully examine this hypothesis, competition assays between environmental
and onion-pathogenic Pagg strains are necessary, ideally in a field-based experiment. Some Pagg isolates
from multiple sources, including symptomatic onion, have been screened through rep-PCR genomic
fingerprinting, however, no matches have been indicated yet (data not shown).
However, using the same rep-PCR genomic fingerprinting, it was demonstrated that a P. ananatis
strain isolated from surface-disinfested, macerated shepherd’s purse tissue collected at mid-season, had
signficant similarity to a P. ananatis strain isolated from an onion from the same field that developed
37
symptoms in a four-month storage (Fig. 2.7). Matched strains are relatively rare in this type of genomic
assay, even though P. ananatis is fairly well-defined phylogenetically based on multi-locus sequence
analysis (Brady et al., 2008). In similar work with P. ananatis, De Paula Lana et al. (2012) reported high
species diversity among strains from maize, sorghum, and crabgrass, to a level at which rep-PCR using
three different primer sets completely differentiated all 15 strains tested. A prominent limitation of the
current study is that weeds were only sampled at mid-season, so it is difficult to suggest the ultimate
origin of the bacterial strain, whether seedborne (Walcott et al., 2002) on weed seeds; soilborne, then
splashed up to weed tissue; or originating on transplants and using weeds as ‘green bridges’ to move from
onion to onion. Repeated sampling procedures in small plots would help more clearly elucidate the
epidemiology of these pathogens. However these results, coupled with the pathogenicity tests from weed
epiphytes and endophytes, suggest a role for weeds in the in-field movement of bacterial rot pathogens
of onion in PA.
Select isolates originating from early-season soil, transplant surfaces and tissues, and weed
surfaces and tissues may induce discoloration, maceration, or both symptoms on onion in pathogenicity
tests (Figs. 2.4, 2.5, 2.6a-c), suggesting all of these as potential sources of bacterial inoculum. Some
evidence is present, however, for species specificity for some inoculum sources, such as Pmar associated
with transplant epiphytic and endophytic fractions (Tables 2.3a, c; Figs. 2.4, 2.6c), Pcar associated with
endophytic transplant and weed samples (Figs. 2.4, 2.5), and P. ananatis demonstrated to occur as both
an endophyte in a weed sampled at midseason as well as a pathogen in a symptomatic onion bulb from
storage (Fig. 2.7). While pathogenic Pagg isolates were found in or on all environmental and transplant
samples, the relationship between environmental and onion-pathogenic Pagg strains appears to be
complex, with logistic regression analyses suggesting the potential for antagonism between these strains
(Tables 2.3b, d); substantially more evidence than is presented here would be required to more solidly
prove this relationship.
Lastly, results presented here have helped identify several avenues of future research, specifically,
replicated field trials on transplant dips to reduce surface populations of bacteria, as well as potentially
bolstering the transplants’ natural defenses prior to the start of the growing season. In addition, dedicated
weed management should be undertaken both within and between mulched beds; some growers tend to
ignore the latter source of weeds because these plants do not directly compete with the onions growing
mulched beds. Successful implementation of these management strategies will contribute to growers’
integrated disease management programs, in PA and elsewhere, and will potentially increase the
profitability of onion production in the U. S.
38
Tables and Figures
Table 2.1. Detection of eight bacterial pathogens of onion from environmental and plant samples, collected in 2011 and 2012. The percentage of the total N of each sample type is given in the appropriate column; columns sum to more than 100% due to samples testing positive for multiple targeted bacterial species.
Bacterial species
At-planting soilu
Transplant epiphytev
Transplant endophytew
Weed epiphytex
Harvest bulbsy
Postharvest bulbsy
2011
P. agglomerans 8% 78% 61% 33% 41% 36%
P. carotovorum 13% 22% 30% 18% 58% 63%
P. marginalis 4% 70% 70% 21% 31% 32%
P. ananatis 1% 9% 0% 7% 2% 8%
B. gladioli 4% 0% 0% 4% 14% 10%
P. viridiflava 4% 30% 9% 10% 14% 3%
B. cepacia 4% 0% 0% 1% 5% 10%
X. axonopodis 0% 0% 0% 0% 0% 1%
None detectedz 70% 0% 13% 43% 7% 11%
Samples N = 84 N = 23 N = 23 N = 44 N = 171 N = 88
2012
P. agglomerans 26% 64% 100% 40% 40% 32%
P. carotovorum 16% 18% 55% 24% 59% 62%
P. marginalis 22% 36% 45% 27% 18% 33%
P. ananatis 2% 0% 18% 2% 7% 11%
B. gladioli 5% 0% 64% 1% 22% 2%
P. viridiflava 1% 0% 0% 6% 1% 6%
B. cepacia 2% 0% 9% 0% 1% 0%
X. axonopodis 0% 0% 0% 0% 0% 0%
None detectedz 38% 22% 0% 11% 10% 10%
Samples N = 87 N = 11 N = 11 N = 187 N = 188 N = 210 u Soil was collected from six equidistantly spaced points within plots, then homogenized, and 1 g was extracted using the MoBio Ultra Soil extraction kit. v Approximately 20 transplants from growers were shaken in 100 mL phosphate buffer (0.1% Tween-20) for 1 h, then was extracted using the Wizard DNA Extraction kit. w The 20 transplants from the epiphytic sample were shaken in 70% ethanol three times, then were macerated using a Waring blender until homogenous. Homogenate was extracted using the Wizard DNA extraction kit. x Five common weeds were sampled from each field in midseason, then shaken in 1 mL 0.1% Tween-20 phosphate buffer / g fresh wt tissue. Buffer fractions were pelleted and extracted as described in v. y Approximately 1 g of onion tissue was removed from a subset (20%, or one bulb, whichever is greater) of symptomatic bulbs from each farm, placed in a sterile 1.7 mL tube, and macerated with sterile phosphate buffer. Homogenate was extracted using the Wizard DNA extraction kit. z Extractions did not produce a band of the appropriate size following routine PCR.
39
Fig. 2.3. Number of bacterial species detected in symptomatic onion bulbs from harvest and storage combined, in 2011 and 2012.
0
10
20
30
40
50
60
70
80
90
100
Epiphytes(N = 26)
Endophytes(N = 19)
Epiphytes(N = 21)
Endophytes(N = 20)
Iso
late
s te
sted
(%
)
Aerobic Anaerobic
Nonpathogenic
P. marginalis
P. carotovorum
P. agglomerans
Fig. 2.4. Pathogenicity of epiphytic and endophytic bacterial isolates originating from transplants, incubated under aerobic and semi-anaerobic conditions in pearl onion bulbs. Pathogenic isolates are indicated by blue-shaded portions of bars, nonpathogenic isolates are indicated by green portions of bars.
Nonpathogenic
P. marginalis
P. carotovorum
P. agglomerans
40
0
10
20
30
40
50
60
70
80
90
100
Epiphytes(N = 91)
Endophytes*(N = 27)
Epiphytes(N = 57)
Endophytes*(N = 24)
Iso
late
s te
sted
(%
)Nonpathogenic
P. marginalis
P. carotovorum
P. agglomerans
Aerobic Anaerobic
Nonpathogenic
P. marginalis
P. carotovorum
P. agglomerans
Fig. 2.5. Pathogenicity of epiphytic and endophytic bacterial isolates originating from common weeds collected at midseason, incubated under aerobic conditions in pearl onion bulbs. Pathogenic isolates are indicated by blue-shaded portions of bars, nonpathogenic isolates are indicated by green portions of bars. *Endophytic isolates are only from 2012, and were isolated on OEM, which is semi-selective for onion pathogens and onion-associated bacteria.
41
Figs. 2.6 a-c. Disease severity induced on onion by P. agglomerans (A), P. carotovorum (B), and P. marginalis (C) isolates, from environmental and transplant sources from 2011 and 2012, in aerobic pathogenicity tests. Nonpathogenic proportions of isolates are indicated by green portions of bars (NP), pathogenic isolates are indicated by pink-red portions of bars. The shade of the bar indicates the severity of induced symptoms as described in Fig. 2.2; severity increases as the bars approach the x-axis. *Weed epiphytes and endophytes were isolated using semi-selective media in 2012; weed epiphytes include isolates from 2011 and 2012, while weed endophytes include isolates only from 2012. N = number of isolates tested from each source.
a: P. agglomerans
42
Fig. 2.7. Representative gel of rep-PCR genomic fingerprints for tracking strains of Pantoea spp. from two fields’ transplant, weed, and symptomatic onion isolates. These fields were planted with transplants from the same lot, but were located approx. 32 km from each other, on different soil types and under different management. Isolates 562 and 563 originated from the same area of the same field, but 562 is an endophyte from purslane while 563 is an epiphyte from crabgrass. Isolate 551 is an endophyte from shepherd’s purse, while isolate 1617 is from an onion that developed symptoms after approx. 4 months in storage. All highl ighted isolates are P. ananatis.
Transplants Midseason weeds Storage onions
NEG
43
Table 2.2. Summary of rep-PCR genomic fingerprinting analysis of environmental and pathogenic isolates from symptomatic onion tissue. Isolates of Pantoea spp. collected from a variety of sources in and around 26 onion production fields in 2012 were used as templates in rep-PCR genomic fingerprinting analysis to track bacterial strains through space and time.
Species Contributing fields
Transplant epiphytes
Transplant endophytes
Weed epiphytes
Weed endophytes
Thrips endophytes
Harvest isolates
Storage isolates
Total tested
Pairs of matched strains
P. ananatis 10 2 0 0 2 0 2 13 19 1
P. agglomerans 8 2 2 2 4 2 2 2 16 0
Table 2.3. Logistic regression analysis of detections of P. marginalis and P. agglomerans in symptomatic onion bulbs combined from harvest and storage, with species detections from environmental and transplant sources as independent variables and farm as a factor (2011 and 2012). Dependent variable is binary P. marginalis (A; C) or P. agglomerans (B; D; presence = 1) in symptomatic onion bulbs from harvest and storage; independents are detections of P. marginalis as a transplant endophyte (A) or epiphyte (C) and detections of P. agglomerans in early-season soil (B) or as a prevalent weed epiphyte (D).
Parameter df Estimate Standard error Wald chi-square P Odds ratioz 95% conf. interval (OR)y A: P. marginalis in 2011 (N = 187)
Pagg prevalent weed epiphyte 1 -1.464 0.219 4.481 0.034 0.629 0.409 – 0.966 z Briefly, odds ratio point estimates below 1 indicate a negative relationship between independent and dependent variables, while point estimates above 1 indicate positive relationships; the change in odds is determined by the absolute value of the magnitude of the difference between the odds ratio estimate and the modeled value, 1 (see Hosmer and Lemeshow, 1989). y 95% confidence interval of the point estimate of the odds ratio.
44
Acknowledgements Dr. Chris Smart, Holly Lange, and other members of the Smart Lab at Cornell University trained E. Pfeufer in rep-PCR protocols. Jeff Stoltzfus, Lee Stivers, Steve Bogash, and Scott Guiser, regional adult and extension educators in PA, identified grower-collaborator farms. The authors thank Ilse Huerta, Evan Stover, Jill Pollok, Marie Ebner, Khanh Nguyen, Anna Testen, Laura Ramos, Andy Kelly, and Laura del Sol Bautista for their field assistance. References Albert, J. M., Munakata-Marr, J., Tenorio, L., Siegrist, R. L. 2003. Statistical evaluation of bacterial source tracking data obtained by rep-PCR DNA fingerprinting of Escherichia coli. Environmental Science and Technology 37: 4554 – 4560. Brady, C., Cleenwerck, I., Venter, S., Vancanneyt, M., Swings, J., Coutinho, T. 2008. Phylogeny and identification of Pantoea species associated with plants, humans, and the natural environment based on multilocus sequence analysis (MLSA). Systematic and Applied Microbiology 31: 447 – 460. Bull, C. T., De Boer, S. H., Denny, T. P., Firrao, G., Fischer-Le Saux, M., Saddler, G. S., Scortichini, M., Stead, D. E., Takikawa, Y. 2010. Comprehensive list of names of plant pathogenic bacteria, 1980 – 2007. Journal of Plant Pathology 92: 551 – 592. Coenye, T., Vandamme, P. 2003. Diversity and significance of Burkholderia species occupying diverse ecological niches. Environmental Microbiology 5: 719-729. Cuppels, D. A., Elmhirst, J. 1999. Disease development and changes in the natural Pseudomonas syringae pv. tomato populations on field tomato plants. Plant Disease 83:759-764. Gent, D. H., Schwartz, H. F. 2008. ‘Bacterial Stalk and Leaf Necrosis’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 66-67. Gitaitis, R., McCarter, S., Jones, J. 1992. Disease control in tomato transplants produced in Georgia and Florida. Plant Disease 76: 651 – 656. Gitaitis, R., MacDonald, G., Torrance, R., Hartley, R., Sumner, D. R., Gay, J. D., and Johnson, W. C.1998. Bacterial streak and bulb rot of sweet onion: II. Epiphytic survival of Pseudomonas viridiflava in association with multiple weed hosts. Plant Disease 82: 935 - 938. Gitaitis, R., Walcott, R., Culpepper, S., Sanders, H., Zolobowska, L., Langston, D. 2002. Recovery of Pantoea ananatis, causal agent of center rot of onion, from weeds and crops in Georgia, USA. Crop Protection 21: 983-989. Gitaitis, R. D., Gent, D. H., Schwartz, H. S. 2008. ‘Leaf streak and bulb rot’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 59.
45
Hsieh, T. F., Huang, H. C., Erickson, R. S. 2005. Biological control of bacterial wilt of bean using a bacterial endophyte, Pantoea agglomerans. Journal of Phytopathology 10: 608 – 614. Humeau, L., Roumagnac, P., Picard, Y., Robène-Soustrade, I., Chiroleu, F., Gagnevin, L., Pruvost, O. 2006. Quantitative and molecular epidemiology of bacterial blight of onion in seed production fields. Phytopathology 96:1345-1354. Ishii, S., Sadowsky, M. J. 2009. Applications of the rep-PCR DNA fingerprinting technique to study microbial diversity, ecology and evolution. Environmental Microbiology 11: 733–740. Ivey, M. L. L., Miller, S. A. 2004. Evaluation of hot water seed treatment for the control of bacterial leaf spot and bacterial canker on fresh market and processing tomatoes. Acta Horticulturae 695: 197 – 204. King, E. O., Ward, M. K., Raney, D. E. 1954. Two simple media for the demonstration of pyocyanin and fluorescin. Journal of Laboratory and Clinical Medicine 44, 301–307. Lange, H. W., Meeks, G. C., Glover, T. J., Smart, C. D. 2006. Weeds as reservoirs of Xanthomonas campestris pv. campestris in New York 96: S64. Lewandoski, D. J., Hayes, A. J., Adkins, S. 2010. Surprising results from a search for effective disinfectants for Tobacco mosaic virus-contaminated tools. Plant Disease 94: 542 – 550. Louws, F. J., Fulbright, D. W., Stephens C. T., de Bruijn, F. J. 1994. Specific genomic fingerprints of phytopathogenic Xanthomonas and Pseudomonas pathovars and strains generated with repetitive sequences and PCR. Applied and Environmental Microbiology 60: 2286: 2295. Mansfield, M., Gugino, B. 2010. Multiplex PCR for simultaneous detection of eight major onion bacterial pathogens. Phytopathology 100: S77. May, R., Volksch, B., Kampmann, G. 1997. Antagonistic activities of epiphytic bacteria from soybean leaves against Pseudomonas syringae pv. glycinea in vitro and in planta. Microbial Ecology 34: 118 – 124. Mohan, S. K. 2008a. ‘Other Bacterial Soft Rots’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 62. De Paula Lana, U. G., Gomes, E. A., Silva, D. D., Costa, R. V., Cota, L. V., Parreira, D. F., Souza, I. R. P., Guimaraes, C. T. 2012. Detection and molecular diversity of Pantoea ananatis associated with white spot disease in maize, sorghum, and crabgrass in Brazil. Journal of Phytopathology 160: 441 – 448. Pusey, P. L., Stockwell, V. O., Reardon, C. L., Smits, T. H. M., Duffy, B. 2011. Antibiosis activity of Pantoea agglomerans biocontrol strain E325 against Erwinia amylovora on apple flower stigmas. Phytopathology 101: 1234 – 1241. Uva, R. H., Neal, J. C., DiTomaso, J. M. 1997. Weeds of the Northeast. Cornell University Press: Ithaca, NY. 397 pp.
46
Versalovic, J., Schneider, M., de Bruijn, F. J., Lupski, J. R. 1994. Genomic fingerprinting of bacteria using repetitive sequence-based polymerase chain reaction. Methods in Molecular and Cellular Biology 5: 25 – 40. Walcott, R. R., Gitaitis, R. D., Castro, A. C., Sanders, F. H., Diaz-Perez, J. C. 2002. Natural infestation of onion seed by Pantoea ananatis, causal agent of center rot. Plant Disease 86:106-111. Wright, S. A. I., Zumoff, C. H., Schneider, L., Beer, S. V. 2001. Pantoea agglomerans strain EH318 produces two antibiotics that inhibit Erwinia amylovora in vitro. Applied Environmental Microbiology 67: 284 – 292. Zaid, A. M., Bonasera, J. M., Beer, S. V. 2012. OEM – a new medium for rapid isolation of onion-pathogenic and onion-associated bacteria. Journal of Microbiological Methods 91: 520 – 526.
47
Chapter 3: Epidemiology and ecology of Gram-negative bacteria potentially pathogenic to onion, Allium cepa, in Pennsylvania Abstract As many as nine different species of bacteria have the ability to induce rots in onion, a potentially
lucrative crop which is rapidly expanding in Pennsylvania. Intensive surveys were undertaken on 28 and
26 farms in 2011 and 2012 in which bacteria were detected and isolated from symptomatic bulbs at
harvest and from storage, in addition to collection of environmental data from observational plots
throughout the growing season. Pectobacterium carotovorum subsp. carotovorum, Pantoea
agglomerans, and Pseudomonas marginalis pv. marginalis isolates from transplant and weed sources
were tested for pathogenicity under aerobic and semi-anaerobic incubations, and data suggested that
higher proportions of pathogenic strains were isolated as epiphytes from the common weeds
lambsquarters (Chenopodium album) and crabgrasses (Digitaria spp.) than from redroot pigweed
(Amaranthus retroflexus) and purslane (Portulaca oleracea). In addition, management factors influenced
pathogen detection in symptomatic onions in parasite- and disease-specific manners. Examples include
increased odds of detecting P. agglomerans in systems with standard black plastic mulch than in systems
using other plastic mulches, negative relationships between plant tissue carbon content and the detection
of P. carotovorum subsp. carotovorum, and the negative overall influence of early-season levels of soil
nitrate on bacterial disease that initiates in foliage. Elucidation of these factors will help refine integrated
strategies for managing ubiquitous, difficult-to-control bacterial pathogens in the onion cropping system.
Introduction
Onions (Allium cepa L.) are a rapidly-expanding and potentially lucrative crop for diversified
vegetable producers in Pennsylvania, but have only been commercially produced on a large scale for
fifteen years in the state. Bacterial rots, including soft rot (caused by Pectobacterium carotovorum subsp.
carotovorum Jones [Pcar] and Pseudomonas marginalis pv. marginalis Brown [Pmar]) and center rot
(caused by the emerging pathogens Pantoea agglomerans Beijerinck [Pagg] as well as Pantoea ananatis
Serrano; [Gent and Schwartz, 2008a; 2008b; Bull et al., 2010]), are the most significant diseases reducing
marketable yields in PA. In high disease incidence years, commercial growers experience yield reductions
of over 50%. In addition, it is common to detect many as four different species of bacteria from a single
symptomatic bulb (Chapter 2), suggesting primary and secondary plant infection by these bacteria.
48
The most frequently detected pathogens in rotting onion bulbs from PA are Pcar, Pmar, and Pagg,
species of bacteria that are ubiquitous in the environment and have high intraspecific variation ranging
from nonpathogenic to highly aggressive strains (Chapter 2). Pcar has been recognized for years as a
general pectinolytic pathogen of vegetables, fruits, and ornamentals (Ma et al., 2007), and was isolated
from soil, onion transplants, and weeds (Chapter 2; Ma et al., 2007). It is arguably the most prominent
postharvest pathogen as a result of its wide host range and ability to induce disease even at low
temperatures (Maher and Kelman, 1983); effective strategies targeted at managing Pcar have been
elusive as a result of its ubiquitous nature. Pmar is a genetically diverse bacterial species which also
includes pathogens with a wide host range and ubiquity in environments (Yamamoto et al., 2000); in
addition, as a fluorescent pseudomonad, Pmar may also function as a plant growth-promoting
rhizobacterium or niche-displacing biocontrol organism (Kloepper and Schroth, 1984), and may be
influential in diverse carbon cycles and environmental remediation as a result of the species’ adaptability
in substrate usage (Yamamoto et al., 2000). Similarly, Pagg has been found globally (Deletoile et al., 2009),
strains of which have been identified as pathogens on other plants (Barash and Manulis-Sasson, 2009), as
niche-displacing biocontrol agents in the management of plant disease (Bonaterra et al., 2003; Costa et
al., 2001), and even as an opportunistic human pathogen causing disease in immunocompromised
individuals (De Champs et al., 2000). In previous work, isolates of each of these bacterial species have
been found as epiphytes and endophytes from onion transplants and common weeds in and around onion
fields (Chapter 2).
Epiphytic bacteria inhabit the surfaces of aboveground plant parts, particularly the phyllosphere,
while endophytic bacteria asymptomatically reside within host plant tissue; both types of bacterial
lifestyle are characterized by their commensal interactions with their hosts (Hardoim et al., 2008). The
majority of research conducted on endophytic bacterial species has investigated their efficacy as
biocontrol organisms or root mutualists (Hardoim et al., 2008), however, other work has suggested that
soil and endophytic bacterial populations are distinct (Lemanceau et al., 1995). Many species of bacteria
have been isolated as endophytes, including some suggested to have quite stable relationships with their
plant hosts, including Pagg with citrus trees (Araujo et al., 2002) and Eucalyptus (Ferreira et al., 2008).
Frequent detections and isolations of Pcar, Pmar, and Pagg from the surfaces as well as within the tissue
of common weeds were demonstrated in Chapter 2, and pathogenicity tests of these organisms indicated
a range of aggressiveness, from nonpathogenic/asymptomatic isolates (of each species) to highly virulent
isolates capable of rotting a pearl onion bulb in days. While detections and isolations of some onion
pathogens from the surfaces of weeds have been conducted with P. ananatis and Pseudomonas viridiflava
49
Burkholder (Gitaitis et al., 1998; 2002), no investigations have been conducted where bacterial epiphytes
and endophytes have been isolated, tested for pathogenicity, and compared based on weedy host plants.
Phenotypic and genotypic characterizations are available for each of these three bacterial species,
in addition to descriptions of symptoms induced on infected plant hosts. However, some of these
descriptions are limited to type or well-characterized strains, as a result of characterization requiring
substantial time and effort (Holt et al., 1994). The classical plant disease triangle, in which a pathogen, its
susceptible plant host, and a conducive environment coordinate to result in plant disease becomes more
complicated when multiple pathogens are involved. Disease management strategies intended to reduce
the incidence of one pathogen may instead positively affect another pathogen, with the same ultimate
result: reduced crop yields; an example of this is in wheat, where NH4 fertilization and low soil pH have
been documented to reduce take-all (caused by Gaeumannomyces graminis), but in the same system,
NO3 nutrition and alkaline soils are preferable for the management of eyespot (caused by
Pseudocercosporella herpotrichoides; Huber and Thompson, 2007). In addition, while some integrated
management practices are designed to adversely affect pathogens, occasionally host plants are also
affected, perhaps increasing their susceptibility to infection by other pathogens.
With their wide host ranges, ubiquity in the environment, and potential for interspecific
interactions, integrated study of these bacteria was necessary to gain a more comprehensive view of their
lifestyles as epiphytes, endophytes, and pathogens. Observational studies were undertaken on 28 and 26
farms in 2011 and 2012, and approximately 2000 bacterial isolates were generated from soil, transplants,
weeds, thrips, and symptomatic onion at harvest and from storage. Selected bacterial strains were
assayed for their prevalence in symptomatic, single-species detection bulbs, their ability to rot onions
based on their isolation location (epiphyte or endophyte) and type of weed, and the impact of both
management factors and the incidence of other bacterial species in symptomatic onion bulbs.
Materials and Methods
Sample collection for bacterial isolations
In 2011 and 2012, 28 and 26 Pennsylvania farms, respectively, were visited three times each over
the course of the onion growing season. Early, mid-season, and harvest samples were taken from soil,
transplants, weeds, and symptomatic onion bulbs, following protocols outlined in Chapter 2.
50
Environmental and management factors
General production data, soil nitrogen analyses, and pre-plant transplant samples were collected
and processed after visit 1, which occurred in April or early May each year. Midseason ratings of weed
pressure, mulch integrity, foliar thrips damage, and bulb growth were recorded during visit 2, in addition
to the collection of weeds for bacterial isolations and onion foliar tissue samples for total nitrogen
measurements. At visit 3, plots were harvested, rated for disease, graded for size, and additional soil
samples were taken, in addition to retaining approx. 30 asymptomatic onion bulbs for a storage rating,
which took place in November. Three asymptomatic bulbs from each farm were each analyzed for total C
after the postharvest rating. Detailed descriptions of these protocols may be found in Chapter 2.
Bacterial isolations from transplants, weeds, harvest, and postharvest onions
Epiphytic and endophytic bacterial isolates were obtained from transplant and weed samples in
2011 and 2012, after the detailed protocols in Chapter 2. Onion extract medium (OEM; Zaid et al., 2012)
was used in 2012 for isolations from weed samples, while King’s B medium was used in 2011. Before
isolations, types of bacterial rots were grouped into general categories: surface rots, characterized by
discoloration and/or maceration initiating in the outer scales of the bulb and progressing inward (Fig.
3.1a), or inner scale rots, characterized by a single or a few discolored and/or macerated scales within the
bulb, which were thought to initiate in the onion foliage (Fig. 3.1b). Bulbs with both types of rots occurred
occasionally, these were not used for bacterial isolations unless no other symptomatic bulbs were
available from that plot, and the disease assignment deferred to inner scale rot. Symptomatic onions were
photo-documented for each isolation, and tissue (approx. 0.5 g and indicated by a surface-sterilized green
wire loop in Figs. 3.1a, b) was removed from the margin between symptomatic and asymptomatic areas.
In 2011 and 2012, respectively, samples were taken from 20% or 50% of symptomatic bulbs per replicate
plot or one bulb (whichever was greater) of each generic type of diseased onions (surface rot or inner
scale rot). Two samples (intended for DNA extraction and bacterial isolation, respectively) were removed
from the symptomatic area and transferred to sterile 1.8 mL Eppendorf tubes. The isolation sample was
ground with a sterile micropestle with an additional 500 µL of sterile Kphos buffer, then was serially
diluted onto KB, then 100 µL of the appropriate dilution was spread-plated on KB to select single isolates
to be maintained in duplicate (described in Chapter 2) at -20⁰C.
51
Figs. 3.1a – b. Symptomatic onions typified by the generic diseases surface rot (a) or inner scale rot (b). The green wire loop was surface-sterilized and indicates roughly where symptomatic tissue was harvested for DNA extraction and bacterial isolation.
Aerobic and semi-anaerobic pathogenicity tests
Aerobic and semi-anaerobic pathogenicity tests were completed using commercial, surface-
sterilized pearl onions that were inoculated with selected isolates of Pcar, Pagg, and Pmar (described in
detail in Chapter 2).
DNA extractions
DNA was extracted directly from soil, transplant, weed, and symptomatic onion tissue using the
MoBio Ultra Soil DNA kit (MoBio Laboratories, Inc., Carlsbad, CA) for soil samples, the Qiagen DNEasy
Plant Mini Kit (Qiagen Inc., Valencia, CA), and the Wizard Genomic DNA Purification kit (Promega,
Madison, WI) for onion samples, as outlined in Chapter 2.
Bacterial species identification
Bacterial isolates were preliminarily identified to genus based on photographed phenotypic
characteristics, then more definitively identified to species using the previously-described duplex PCR
protocol (Chapter 2; Mansfield and Gugino, 2010).
Data analysis
Nonparametric comparisons
Comparisons of the frequencies of pathogenic and nonpathogenic isolates of Pagg, Pcar, and
Pmar were conducted using Fisher’s exact test (for 2x2 tables) and the Mantel-Haenszel-Cochran test (for
52
experiment-wise tables in which three species were tested) using Minitab 16 (Minitab Inc., State College,
PA, USA).
Logistic regressions
Initial logistic regressions where surface (0) or inner scale rot (1) in symptomatic bulbs at harvest
and from storage were conducted with seven (all but X. axonopodis) species presence or absence (1, 0) as
independent variables, to associate generic types of disease with particular pathogens. Logistic
regressions where species presence (1) or absence (0) were dependent variables were conducted with
environmental and production factors as well as the presence or absence of other bacterial pathogen
species were independent variables. In all logistic regressions, the grower, field, and year were included
as class variables, then the stepwise selection procedure (α to add = 0.2; α to remove = 0.25) was used to
narrow the field of potential independent variables. Next, another logistic procedure was conducted
where the selected variables were placed in the model and removed one at a time based on P value
(largest removed), then the analysis was repeated. Removal of variables stopped when all variables had P
values below 0.05. Odds ratios were also calculated (SAS 9.2, SAS Institute, Cary, NC, USA); odds ratios
approximate how likely a binary outcome is to occur by relative risk. To interpret odds ratios, first the
difference between the point estimate and 1 is determined, then this difference indicates the relationship
between the independent variable and the dependent, binary variable. Point estimates below 1 indicate
a negative relationship between independent and dependent variables, while point estimates above 1
indicate positive relationships. In some cases, where a one-unit change in the independent variable
resulted in a very small change in odds, each value was multiplied by a constant in order to attain more
meaningful values (Hosmer and Lemeshow, 1989).
Results
In a logistic regression, where the generic type of bacterial rot symptom was classified either as
surface rot or inner scale rot (Figs. 3.1a, b) and was used as the dependent variable, only two species of
bacteria were consistently associated with each general type of disease. Burkholderia gladioli pv. alliicola
was associated with bulbs with surface rot; the odds of a bulb with inner scale rot testing positive for B.
gladioli were 59% lower than the odds of detecting that pathogen in surface rot bulbs, which was
statistically significant (P < 0.001; Table 3.1). Also statistically significant, the odds of a bulb with inner
scale rot testing positive for Pseudomonas viridiflava were nearly three times the odds of a surface rot
bulb testing positive for this bacterium (P < 0.001; Table 3.1). The incidence of a particular type of rot by
53
other bacterial species could not be statistically predicted (P > 0.05), however, based on the odds ratio
point estimates and their 95% confidence intervals, it may be suggested that Pmar and Pantoea ananatis
trend toward inner scale rots, and Burkholderia cepacia and Pantoea agglomerans (Pagg) trend toward
surface rots; P. carotovorum subsp. carotovorum (Pcar) seems to be equally detected in surface and inner
scale rot – affected bulbs (Table 3.1).
In symptomatic bulbs in which only one bacterial species was detected using the multiplex PCR
protocol, Pcar was the most commonly detected pathogen in surface and inner scale rots (38 – 55%; Figs.
3.1a, b and 3.2). Pantoea agglomerans (Pagg) was also frequently identified as a singular pathogen (22
– 24%), however, for its frequency in the environment, Pseudomonas marginalis (Pmar) was singularly
detected in fewer bulbs, only 12 – 14%. Burkholderia gladioli pv. alliicola and Pantoea ananatis were
detected in 12 and 6% of single-detection surface rot and inner scale rot bulbs, respectively (Fig. 3.2).
In aerobic and semi-anaerobic pathogenicity tests of epiphytic and endophytic Pagg, Pcar, and
Pmar from transplants, over 60% of isolates tested were pathogenic to onion (Fig. 3.3a). No statistically
significant differences between species or isolation source existed in the aerobic test, however,
numerically fewer Pmar isolates from transplants were pathogenic to onion as compared to the
frequencies of pathogenic isolates of Pagg and Pcar. In semi-anaerobic pathogenicity tests,
proportionally more Pmar endophytes were pathogenic than epiphytes, which was nearly statistically
significant (P = 0.059). No other statistically significant differences between epiphytes and endophytes
within each test (and disregarding species) were made apparent by a Mantel-Haenszel-Cochran test (P >
0.1; Fig. 3.3b).
When bacterial epiphytes from weeds collected in 2011 were considered nonspecifically in terms
of the type of weed they were isolated from (Fig. 3.4), proportionally more isolates from lambsquarters
(Chenopodium album L.) were pathogenic in aerobically incubated tests than the proportions of isolates
from each of crabgrasses (Digitaria spp. Haller; P = 0.058), redroot pigweed (Amaranthus retroflexus L.;
P = 0.017), and purslane (Portulaca oleracea L.; P = 0.007) using Fisher’s exact test. There were no
statistically significant differences between each of the other three weeds (Fig. 3.4).
The use of standard black plastic, the partial removal of mulch at midseason by the grower, the
silt content of soil, and the amount of total bulb tissue N at harvest all increased the odds of detecting
Pagg in the symptomatic bulb. Torn mulch increased the odds of detecting Pagg 62% over fields with
mulch left intact, in addition, the use of black plastic mulch doubled the odds of detecting Pagg in
symptomatic bulbs (Table 3.2a). For each 5% increase in silt content of soil, the odds of detecting Pagg
54
increased approx. 20%, and for every 0.25% increase in total N content in bulbs at harvest, the odds of
Detections of Pcar were positively associated with the incidence of purple blotch in fields, while
foliar tissue C at midseason and early-season soil nitrate were negatively associated with the presence of
the bacterial species (Table 3.2b). Purple blotch is a fungal disease of Alliums caused by Alternaria porri
Ellis that commonly occurs on PA farms, particularly in wet weather (Pfeufer, observation). The odds of
detecting Pcar in symptomatic onions increase 59% if greater than 10% of plants have purple blotch at
midseason (Table 3.2b). However, an increase of 0.5% in total foliar C at midseason decreases the odds of
detecting Pcar 12%, as does an increase in early-season soil nitrate of approximately 10 mg nitrate / kg
dry soil (Table 3.2b).
Similarly, the odds of detecting Pmar in symptomatic bulbs decreased 11% for each increase of
early season soil nitrate of 10 mg / kg dry soil (Table 3.2c). Pmar detections were positively associated
with the detection of Pseudomonas viridiflava as well as bulbs that are larger than approx. 6 cm in
diameter at midseason, increasing the odds of detecting Pmar by 119% and 64%, respectively. An
increase in total foliar N of 0.5% increased the odds of detecting Pmar by 83% (Table 3.2c).
In addition to Pcar detections in symptomatic bulbs, the odds of detecting P. ananatis decreased
with both total foliar tissue C and the amount of early-season soil nitrate (Table 3.2d). However, the
opposite relationship with incidence of purple blotch was indicated, suggesting the odds of detecting P.
ananatis actually decreased 45% when more than 10% of plants were symptomatic for purple blotch in
midseason (Table 3.2d). Similar to Pmar, the odds of detecting P. ananatis in symptomatic bulbs more
than doubled if bulbs were at least 6 cm in diameter at midseason. In addition, for each additional ten
hours of soil temperatures at or above 30 ⁰C, the odds of detecting P. ananatis in symptomatic bulbs
increased 4% (Table 3.2d).
When separating diseased bulbs by the type of symptoms expressed as well as those with positive
detections of Pagg, the use of standard black plastic in either case doubled the odds of detecting Pagg in
symptomatic bulbs (Tables 3.3a, b). However, the odds of detecting Pagg in bulbs with surface rots
decreased nearly 65% if the bulb tested positive for Burkholderia gladioli pv. alliicola (Table 3.3a). In bulbs
with inner scale rots only, the odds of detecting Pagg increased with increased silt content of soils, higher
harvest levels of soil nitrate, larger bulb size at midseason, and particularly potentially mineralizable
nitrogen rate, which more than doubled the odds of detecting Pagg in symptomatic, inner scale rot bulbs
(Table 3.3b).
55
Detections of Pcar in surface rot bulbs were positively associated with two soil variables, the total
number of hours of soil temperature ≥ 30⁰C as well as the clay content of soils. An additional 22 hours of
soil temperatures at or above 30⁰C as well as an increase of 1% more clay in field soil increased the odds
of detecting Pcar by approx. 29% (Table 3.4a). A high degree of weed pressure between the raised beds
actually increased the odds of detecting Pcar by 84%, while for every additional 0.33% of total C in
harvested bulb tissue, the odds of detecting Pcar in surface rot bulbs decreased 25% (Table 3.4a).
Similar to surface rot bulbs, for every additional 0.33% of total C in foliar tissue at midseason, the
odds of detecting Pcar in inner scale rot bulbs decreased 8.5% (Table 3.4b). Early season soil nitrate as
well as very high weed pressure between raised beds also decreased the odds of detecting Pcar in inner
rot bulbs. Both the silt content of soil and the incidence of purple blotch symptoms at midseason increased
the odds of detecting Pcar; in particular, purple blotch symptoms on greater than 10% of plants increased
the odds nearly 46% (Table 3.4b).
Discussion
From intensive surveys and laboratory-based studies, the detection, isolation, pathogenicity
status, and influence of environmental variables on several bacterial species was assessed in order to
provide more information about the ecology of these organisms. Pectobacterium carotovorum subsp.
carotovorum (Pcar) occurred most frequently as the single bacterial pathogen identified from
symptomatic bulbs, with Pantoea agglomerans (Pagg) and Pseudomonas marginalis pv. marginalis also
accounting for proportions of single-detection symptomatic bulbs (Fig. 3.2). These results illustrate the
difficulty associated with attributing diagnostic symptoms to particular pathogens – expectations were
that Pcar would be prevalent in surface rot bulbs, but less so in inner scale rot bulbs, which were expected
to have more frequent detections of Pagg and Pmar, bacteria that both begin as leaf pathogens (Mohan,
2008; Gent and Schwartz, 2008). Results from logistic regression analysis with the general type of rot as
the dependent variable indicated two species of bacteria were consistently associated with a single type
of bacterial rot (Table 3.1). Burkholderia gladioli pv. alliicola was typically associated with surface rot
bulbs, and Pseudomonas viridiflava was associated with bulbs with inner scale rot (Table 3.1). These
regressions correspond well with previously published work in which the lifestyle and symptomatology
induced by each pathogen has been described: members of the Burkholderia genus are well-described as
bacterial soil inhabitants (Coenye and Vandamme, 2003), which would facilitate B. gladioli pv. alliicola
surface infections of bulbs, while P. viridiflava is considered the ‘leaf streak’ pathogen, which in advanced
cases, induces a rot of a single scale within symptomatic bulbs (Gitaitis et al., 2008).
56
Since Pcar, Pagg, and Pmar are the three most common pathogens in the onion – bacterial rot
pathosystem in PA (Chapter 2), selected isolates of each species originating from the surfaces and tissue
of transplants were tested for pathogenicity to onion. High frequencies (60% or more) of pathogenic
isolates originated from transplants, and though not statistically significant, Pmar generally tended to
have lower frequencies of pathogenic isolates in either type of pathogenicity test, compared to the other
two species of interest (Fig. 3.3a), which may be expected since Pseudomonas spp. are generally
considered aerobic (Holt et al., 1994). In addition, as might be expected, Pmar endophytes were more
frequently pathogenic to onion in semi-anaerobic tests than epiphytes (Fig. 3.3a), and support from
accompanying work suggested Pmar detections as transplant epiphytes or endophytes actually increase
the odds of detecting Pmar in symptomatic onion tissue (Chapter 2). Taken together, this suggests that
endophytic strains from transplants may actually represent latently infected plants, and may be a source
of bacterial inoculum in the production system. The use of a rep-PCR strain tracking assay will be used in
the future to more conclusively elucidate the hypothesis of pathogenic Pmar arriving with onion
transplants, ideally by generating matching genomic fingerprints between transplant isolates and isolates
inducing bulb symptoms (Chapter 2); this type of assay has been successfully completed with other
Pseudomonas species (Louws et al., 1994).
In a similar analysis, over 60% of Pcar, Pagg, and Pmar isolates from weeds were pathogenic to
onion in aerobic and semi-anaerobic pathogenicity tests (Fig. 3.3b). Though not statistically significant,
Pagg typically had higher proportions of pathogenic isolates than the other two species, with Pmar having
the lowest proportions of pathogenic isolates (Fig. 3.3b). These results may be reinforced by comparisons
between epiphytic isolates from weeds tested for pathogenicity and shown in Fig. 3.3b; for growers, this
may indicate that even if protectant copper bactericides are applied on both cropped onions and adjacent
weeds, bacterial pathogens have the ability to subsist within the weed tissue. Complete removal of weeds
from the field may be necessary in order to remove an inoculum source (weed endophytes) from grower
fields.
In aerobic pathogenicity tests of isolates from the four most common weeds found in PA onion
fields, a larger proportion of isolates originating from lambsquarters (91%) were pathogenic compared to
proportions of pathogenic isolates from crabgrasses (71%), redroot pigweed (55%), and purslane (62%; P
< 0.06), using Fisher’s exact test (Fig. 3.4). These results were unexpected; it was hypothesized that a
larger proportion of bacterial isolates would originate from crabgrasses, which as monocots, are more
closely related to onion than the other weeds and has been shown in other studies (Toussaint et al., 2012;
Yishay et al., 2008). In a study comparing Pcar strains, Yishay et al. found that isolates of Pcar that
57
originated from monocot hosts were more aggressive in monocot pathogenicity tests compared to dicot
pathogenicity tests, and vice-versa (2008). These authors further suggested, through genetic analyses
based on 16S rRNA sequence analysis, that monocot and dicot isolates were clearly differentiated at the
basal level, which may indicate co-evolutionary specialization (Yishay et al., 2008). Similarly, when
Xanthomonas campestris pv. vitians, a bacterial pathogen of lettuce, was inoculated onto foliage of
common weeds, Lactuca spp. supported higher pathogen populations than weeds that were more
distantly related to lettuce (Toussaint et al., 2012).
The present analyses, in which bacterial isolates are grouped nonspecifically into pathogenic and
nonpathogenic proportions, do not support the hypothesis of co-evolutionary specialization; rather,
another aspect influencing interactions between lambsquarters and its bacterial symbionts may impose a
selection pressure on these potential pathogens. For instance, as an early-emerging weed (Myers et al.,
2004), lambsquarters may support larger populations of bacteria, or some aspect of lambsquarters’
biochemical makeup may result in higher populations of bacteria pathogenic to onion. In syringe-
inoculations of wild transplanted lambsquarters, plants inoculated with 100 µL of 108 CFU / mL P.
agglomerans suspension appeared no different than uninoculated plants (Pfeufer, unpublished),
potentially allowing pathogenic bacteria to persist symptomlessly. In addition, in serial dilutions in the
isolation process of this observational study, some endophytic fractions of surface-sterilized weeds were
diluted to 10-8 before bacterial density was such that colonies were countable, indicating the potential for
lambsquarters to support large bacterial populations within their tissues. Hypotheses about relationships
between weed emergence, plant biochemistry, and pathogen populations, however, would need to be
tested further, in addition to conducting pathogenicity tests on isolates from a wider range of weed
sources.
While lab-based biochemical tests and growth assays are well-defined for many of these
pathogens (Holt et al., 1994), those phenotypic traits may not translate directly to pathogen abundance
in field situations, which illustrates why logistic regression analyses of detections of the prevalent species
of bacteria in symptomatic bulbs may contribute additional dimensions to our knowledge about each of
these bacterial pathosystems. Perhaps most strikingly, Pagg was demonstrated repeatedly to be
associated with standard black plastic mulch, which occurred regardless of what type of general bacterial
rot symptoms it was associated with (Tables 3.2a; 3.3a, b). Standard black plastic mulch has been shown
to result in overall higher incidence of bacterial disease (Gugino et al., 2011), perhaps by way of warmer
soil temperatures at and after midseason (Chapter 4). These results corroborate results that indicated the
use of black plastic increased the incidence and severity of center rot of onion, caused by P. ananatis
58
(Gitaitis et al., 2002) in addition to P. agglomerans. It may be hypothesized that warmer soils allow
pathogenic strains of Pagg to overtake nonpathogenic or inhibitory strains of Pagg. Temperature-
dependent virulence gene expression has been widely demonstrated in other plant pathogenic bacteria,
however, most well-studied examples describe downregulation of virulence at temperatures >28°C, with
the most notable exception being the tropical pathogen Ralstonia solanacearum (Smirnova et al., 2001).
Additional experimentation is necessary before concluding this point, but interestingly, if Pagg virulence
is upregulated at higher temperatures, this would more closely resemble a bacterial pathogen of humans
(Smirnova et al., 2001), which Pagg has already been demonstrated to be in immunocompromised people
(De Champs et al., 2001; Volksch et al., 2009).
Other variables influential in the detection of Pagg are the silt content of soil (Tables 3.2a, 3.3b)
and the potential antagonism between Pagg and B. gladioli pv. alliicola in bulbs with surface rots (Table
3.2a). It may be hypothesized that soils with ample silt may not support high populations of Pagg, since
soil texture is established as an important factor in determining rhizobacterial populations (Dey et al.,
2012), however, this would require further evidence. Both Pagg (Wright et al., 2001) and B. gladioli pv.
alliicola (Hu and Young, 1998) have been shown to produce antibiotic-like inhibitory compounds in media,
so further experimentation, such as concurrent culturing in media as well as in onion plants, would be
necessary to determine which bacterial species is the primary inhibitor.
Detection of Pcar in surface and inner scale rot bulbs allowed the further identification of variables
of interest for this difficult-to-manage bacterial pathogen. In particular, the negative relationships
between tissue C content and the incidence of Pcar in both surface rot (Table 3.4a) and inner scale rot
(Table 3.4b), may indicate future strategies to investigate for managing this pathogen: it may be that high
tissue C decreases onion susceptibility to this pathogen. These results correspond well to results
presented in Chapter 4, in which foliar tissue C as well as the C:N ratio of plant foliage appeared to
influence the incidence of bacterial rots at harvest and from storage (Chapter 4). While counterintuitive
when considering bacterial pathogens require plant hosts, in part, as a source of carbon, one suggestion
may be that plants with higher carbon in their tissues are less susceptible to infection as a result of better
overall plant health. Alternatively, higher carbon contained in the plant tissues may be involved in long
carbon chain precursors of defense-related antimicrobials, such as allicin (Kyung, 2012).
Pcar was also positively associated with the incidence of purple blotch in plants at midseason
(Tables 3.2b, 3.4b), while P. ananatis was negatively associated with purple blotch (Table 3.2d). Purple
blotch is a fungal disease of Alliums caused by Alternaria porri Ellis, and less often, Stemphylium vesicarium
Simmons, that commonly occurs on PA farms in wet weather (Pfeufer, observation). Symptoms initially
59
are small water soaked lesions that develop on leaves, turn chlorotic with a pale tan center before
coalescing and developing a purplish-brown color; A. porri-induced symptoms are characterized by
concentric rings (Miller and Lacy, 2008). Under severe purple blotch pressure, growers have informally
attributed rotting bulbs to A. porri (pers. comm., to Pfeufer), even though purple blotch symptoms
typically manifest as purple to brown outer papery scales and rots due to A. porri are thought to occur
infrequently (Miller and Lacy, 2008). Association between these pathogens and Pcar has only been
informally suggested (Paibomesai et al., 2012), and synergistic disease induction by these inter-kingdom
pathogens may be a subject for future research. For growers, this observation reinforces the
recommendation of tank-mixing copper products with ethylene-bis-dithiocarbamate fungicides when
managing onions for disease.
Another interesting point in these datasets is the differential effect of weed pressure between
raised beds in the field on the detection of Pcar in surface and inner scale rot bulbs. While weed pressure
between rows is positively associated with the detection of Pcar in surface rot bulbs, the same variable is
negatively associated with the detection of Pcar in inner scale rot bulbs (Tables 3.4a, b). Since weeds have
been shown to harbor pathogenic strains of Pcar on their surfaces and within their tissues (Chapter 2,
Figs. 3.3, 3.4), the relationship with inner scale rots was unexpected. One potential explanation is that
high weed pressure between the rows shades the raised beds enough to reduce soil temperatures, which
other results suggest may result in lower disease incidence (Gugino et al., 2011; Chapter 4). Replicated
field trials, however, are necessary to confirm this hypothesis.
Considering all the results presented here, it may be concluded that environmental variables
probably exert differential influences on various critical components of the plant disease triangle.
Examples of environmental variables primarily influencing the pathogen include the suggested benefit to
Pagg conferred by the use of standard black plastic mulch (Tables 3.2a; 3.3a, b) and the negative
association between Pcar and local tissue C (Tables 3.2b; 3.4a, b). On the other hand, environmental
variables may serve to increase the susceptibility of the host, or may influence the pathogen-host
interaction such that disease is overall enhanced. One example is the multiple positive relationships
between silt content of soils and the incidence of inner scale rot, which is assumed to initiate in the foliage
and develop as a result of infection by Pagg (Table 3.3b), Pcar (Table 3.4b), Pmar (Table 3.2c), and P.
ananatis (Table 3.2d). Another influential variable in the development of inner scale rot symptoms
appears to be early-season soil nitrate, which is negatively associated with detections of Pcar, Pmar, and
P. ananatis in inner scale rot bulbs, (Tables 3.2b-d, 3.4b), and is particularly interesting given other work
suggesting positive relationships between bacterial rots of onion and preharvest soil nitrate levels
60
(Chapter 4). However, it may be that in the later analysis (Chapter 4), early soil nitrate may positively
influence the surface rot-type diseases while simultaneously reducing the foliar inner scale rot diseases.
As a final word, although Pmar has been well-described in the literature as a primary pathogen of
onion (Wright and Hale, 1992), this species may function as a secondary pathogen in PA, or at least can
be considered less aggressive. This hypothesis is based on the relatively few Pmar single-species
detections in symptomatic bulbs (Fig. 3.2), even though Pmar is frequently detected throughout the
production system (data shown in Chapter 2). In addition, its frequent co-occurrence with P. viridiflava
(Table 3.2c) and other bacterial species, and its consistently lower proportions of isolates found to induce
symptoms in pathogenicity tests (Figs. 3.3a, b) may help reinforce this point. To more conclusively
demonstrate this, future work is necessary using Pmar single and co-inoculations with other bacterial
species isolated from symptomatic onions.
Acknowledgements
The authors thank Dr. Shelby Fleischer and Dr. Mark Otieno (Department of Entomology, Penn State) for several helpful conversations about statistical approaches.
61
Tables and Figures
Table 3.1. Logistic regression of general types of bacterial rot (surface rot or inner scale rot) modeled by the detection of bacterial species in symptomatic bulbs collected at harvest and from storage in PA in 2011 and 2012 (N = 614). Dependent variable modeled is 1 = inner scale rot (n = 395), while independent variables are species detections (presence = 1).
Parameter df Estimate St. Error Wald Chi-square P Odds ratio z 95% C.I. (O.R.) y
P. carotovorum 1 0.335 0.190 0.031 0.860 1.034 0.713 – 1.501 z Briefly, odds ratio point estimates below 1 indicate a negative relationship between independent and dependent variables, while point estimates above 1 indicate positive relationships; the reduced (or increased) relative risk is determined by the magnitude of the difference (see Hosmer and Lemeshow, 1989). y 95% confidence interval of the point estimate of the odds ratio.
0
10
20
30
40
50
60
70
80
90
100
Surface rot (N = 78) Inner scale rot (N = 160)
Sin
gle
spec
ies
det
ecti
on
s (%
)
Other
P. ananatis (ISR)or B. gladioli (SR)P. marginalis
P. agglomerans
P. carotovorum
Other
B. gladioli (SR) or P. ananatis (ISR)
P. marginalis
P. agglomerans
P. carotovorum
Fig. 3.2. Detection of bacterial species in symptomatic bulbs pooled from at harvest (2011 and 2012) and from storage (2012 only) samples from which only one pathogen was detected and separated by the type of rot symptom observed (N=238). This represents approx. 36% of all symptomatic bulbs collected from 2011 and 2012. The ‘Other’ category includes positive detections for Burkholderia cepacia, Pseudomonas viridiflava, and either P. ananatis (surface rot bulbs [SR]) or B. gladioli pv. alliicola (inner scale rot bulbs [ISR]).
62
0
10
20
30
40
50
60
70
80
90
100
Pagg(N = 16)
Pcar(N = 12)
Pmar(N = 18)
Pagg(N = 18)
Pcar(N = 12)
Pmar(N = 18)
Iso
late
s te
ste
d (
%)
Aerobic pathogenicity test Semi-anaerobic pathogenicity test
Nonpathogenicendophytes
Nonpathogenicepiphytes
Pathogenicendophytes
Pathogenicepiphytes
0
10
20
30
40
50
60
70
80
90
100
Pagg(N = 43)
Pcar(N = 28)
Pmar(N = 47)
Pagg(N = 25)
Pcar(N = 25)
Pmar(N = 31)
Iso
late
s te
ste
d (
%)
Aerobic pathogenicity test Semi-anaerobic pathogenicity test
Nonpathogenicendophytes*
Nonpathogenicepiphytes
Pathogenicendophytes*
Pathogenicepiphytes
Figs. 3.3a - b. Pathogenicity of Pantoea agglomerans (Pagg), Pectobacterium carotovorum subsp. carotovorum (Pcar), and Pseudomonas marginalis pv. marginalis (Pmar) from onion transplants (a) and weeds (b) in aerobic and semi-anaerobic pathogenicity tests, divided by bacterial isolation source. *All 2012 isolates were generated from semi-selective OEM, while 2011 isolates were generated from KB (nonselective). Weed endophytes were only isolated in 2012.
a
b
63
Fig. 3.4. Bacterial epiphytes of the species Pantoea agglomerans, Pectobacterium carotovorum subsp. carotovorum, and Pseudomonas marginalis pv. marginalis from selected common weed sources collected in PA onion fields in 2011 and tested for pathogenicity under aerobic incubation. Purple-shaded portions of bars indicate pathogenic isolates of each species, while green shaded portions of bars indicate nonpathogenic isolates from each weed source, regardless of species. * Indicates proportionally more pathogenic isolates from lambsquarters compared to each of the other weeds by Fisher’s exact test (α = 0.06).
64
Tables 3.2a - d: Logistic regressions of harvest and storage detections of Pantoea agglomerans (3.2a; n = 225 of 614), Pectobacterium carotovorum subsp. carotovorum (3.2b; n = 366 of 614), Pseudomonas marginalis pv. marginalis (3.2c; n = 171 of 614), and Pantoea ananatis (3.2d; N = 48 of 614) from symptomatic onion bulbs from PA, combined in 2011 and 2012; positive detections rated ‘1.’ Independent variables are detections of other bacterial species in symptomatic bulbs as well as environmental and production factors observed throughout the season.
Table 3.2a. P. agglomerans
Parameter df Estimate St. Error Wald Chi-square P Odds ratio z 95% C.I. (O.R.) y
Intercept 1 -4.964 0.852 33.943 <0.001
Torn mulch at midseason x 1 0.483 0.140 11.988 <0.001 1.622 1.233 – 2.132
Standard black mulch w 1 0.713 0.198 12.995 <0.001 2.041 1.385 – 3.007
Bulb tissue N at harvest v 1 0.822 1.280 8.594 0.003 2.274 1.313 – 3.939
Silt content of soil u 1 0.038 0.013 8.521 0.004 1.039 1.013 – 1.066
Table 3.2b. P. carotovorum subsp. carotovorum
Parameter df Estimate St. Error Wald Chi-square P Odds ratio z 95% C.I. (O.R.) y
Intercept 1 11.148 2.888 14.905 <0.001
Foliar tissue C at midseason t 1 -0.285 0.075 14.594 <0.001 0.752 0.650 – 0.871
Early season soil nitrate s 1 -0.014 0.005 9.752 0.002 0.986 0.977 – 0.995
Foliar tissue C at midseason t 1 -0.507 0.159 10.153 0.002 0.602 0.441 – 0.823
Early season soil nitrate s 1 -0.021 0.010 4.327 0.036 0.979 0.959 – 0.999
Bulb size at midseason o 1 0.782 0.303 6.660 0.010 2.186 1.207 – 3.960
Soil T >30⁰C (season-long) n 1 0.004 0.002 5.116 0.024 1.004 1.001 – 1.008
65
z Briefly, odds ratio point estimates below 1 indicate a negative relationship between independent and dependent variables, while point estimates above 1 indicate positive relationships; the reduced (or increased) relative risk is determined by the magnitude of the difference (see Hosmer and Lemeshow, 1989). y 95% confidence interval of the point estimate of the odds ratio. x Torn plastic mulch at midseason visit coded ‘1;’ intact mulch coded ‘0.‘ w Standard black plastic mulch coded as ‘1’; other, non-standard mulches (silver-on-black, white, black biodegradable) coded ‘0.’ v Total bulb tissue N as measured from three asymptomatic bulbs per field; continuous variable. u Midpoint of the range of silt content for the field, reported by NRCS web soil survey (websoilsurvey.nrcs.usda.gov); continuous variable. t Total foliar tissue C as measured from thirty fully-expanded leaves, comingled in three samples by plot; value reported is the average of three plots for a field average; continuous variable. s Soil nitrate content early in season, in mg NO3 kg-1; field average derived from three plot averages consisting of six comingled samples each; continuous variable. r Purple blotch (caused by Alternaria porri) pressure on >10% of the plants in the observational plots coded as ‘1.’ q Pseudomonas viridiflava also identified from the same bulb tissue, coded as ‘1.’ p Total foliar tissue N as measured from thirty fully-expanded leaves, comingled in three samples by plot; value reported is the average of three plots for a field average; continuous variable. o Visually estimated bulb size as rated at midseason visit; larger than 6 cm diameter coded ‘2;’ from 4 – 6 cm diameter coded ‘1.’ n Season-long total hours of soil temperature ≥30⁰C by field; continuous variable.
66
Tables 3.3a – b. Pantoea agglomerans detections in surface (3a; n = 68 of 174) and inner scale rot (3b; n = 128 of 348) bulbs with other bacterial species detections and environmental and management factors as independent variables.
Table 3.3a. P. agglomerans in surface rot bulbs
Parameter df Estimate St. Error Wald Chi-square P Odds ratio z 95% C.I. (O.R.) y
Intercept 1 -0.765 0.305 6.299 0.012
Standard black mulch x 1 0.711 0.354 4.041 0.044 2.035 1.018 – 4.069
B. gladioli pv. alliicola w 1 -1.033 0.463 4.987 0.026 0.356 0.144 – 0.881
Table 3.3b. P. agglomerans in inner scale rot bulbs
Parameter df Estimate St. Error Wald Chi-square P Odds ratio z 95% C.I. (O.R.) y
Bulb size at midseason u 1 0.457 0.206 4.924 0.027 1.579 1.055 – 2.362
Standard black mulch x 1 0.979 0.279 12.312 <0.001 2.662 1.541 – 4.600
PMN, early season t 1 1.013 0.388 6.827 0.009 2.754 1.288 – 5.890
Silt content of soil s 1 0.050 0.023 4.683 0.031 1.051 1.005 – 1.100 z Briefly, odds ratio point estimates below 1 indicate a negative relationship between independent and dependent variables, while point estimates above 1 indicate positive relationships; the reduced (or increased) relative risk is determined by the magnitude of the difference (see Hosmer and Lemeshow, 1989). y 95% confidence interval of the point estimate of the odds ratio. x Standard black plastic mulch coded as ‘1’; other, non-standard mulches (silver-on-black, white, black biodegradable) coded ‘0.’ w Positive identification of Burkholderia gladioli pv. alliicola from the same symptomatic bulb coded ‘1.’ v Soil nitrate (in mg NO3 kg-1 dry soil) as measured from each plot at harvest; reported values are the field average of three plots, continuous variable. u Visually estimated bulb size as rated at midseason visit; larger than 6 cm diameter coded ‘2;’ from 4 – 6 cm diameter coded ‘1.’ t Potentially mineralizable nitrogen in early-season soils after a 31-40 day incubation, in mg kg-1 day-1; continuous variable. s Midpoint of the range of silt content for the field, reported by NRCS web soil survey; continuous variable.
67
Tables 3.4a –b. Pectobacterium carotovorum subsp. carotovorum detections in surface (3.4a; n = 163) and inner scale rot (3.4b; n = 225 of 360) bulbs with other bacterial species detections and environmental and management factors as independent variables. Table 3.4a. P. carotovorum subsp. carotovorum in surface rot bulbs
Parameter df Estimate St. Error Wald Chi-square P Odds ratio z 95% C.I. (O.R.) y
Intercept 1 46.814 15.527 9.090 0.003
Soil T > 30⁰C (season-long) x 1 0.013 0.004 9.677 0.002 1.013 1.005 – 1.022
Bulb tissue C at harvest w 1 -1.359 0.431 9.961 0.002 0.257 0.110 – 0.597
Clay content of soil v 1 0.258 0.094 7.559 0.006 1.294 1.077 – 1.555
Weed pressure between bedsq 1 -0.307 0.152 4.097 0.043 0.736 0.546 – 0.990 z Briefly, odds ratio point estimates below 1 indicate a negative relationship between independent and dependent variables, while point estimates above 1 indicate positive relationships; the reduced (or increased) relative risk is determined by the magnitude of the difference (see Hosmer and Lemeshow, 1989). y 95% confidence interval of the point estimate of the odds ratio. x Season-long total hours of soil temperature ≥30⁰C by field; continuous. w Total tissue C in bulbs at harvest as measured by three asymptomatic bulbs; value reported is the average of three bulbs for a field average; continuous. v Midpoint of the range of clay content for the field, reported by NRCS web soil survey; continuous variable. u Midpoint of the range of silt content for the field, reported by NRCS web soil survey; continuous variable. t Total foliar tissue C as measured from thirty fully-expanded leaves, comingled in three samples by plot; value reported is the average of three plots for a field average; continuous. s Soil nitrate content early in season, in mg NO3 kg-1; field average derived from three plot averages consisting of six comingled samples each; continuous. r Purple blotch (caused by Alternaria porri) pressure on >10% of the plants in the observational plots coded as ‘1’ q Visual rating of weed pressure between raised beds; very high weed pressure (shading beds or encroaching on plants) rated ‘1.’
68
References Araujo, W. L., Marcon, J., Maccheroni, W., van Elsas, J. D., van Vuurde, J. W. L., Azevedo, J. L. 2002. Diversity of endophytic bacterial populations and their interaction with Xylella fastidiosa in citrus plants. Applied and Environmental Microbiology 68: 4906 – 4914. Barash, I., Manulis-Sasson, S. 2009. Recent evolution of bacterial pathogens: the gall-forming Pantoea agglomerans case. Annual Reviews of Phytopathology 47: 133-152. Bonaterra, A., Marib, M., Casalinib, L., Montesinosa, E. 2003. Biological control of Monilinia laxa and Rhizopus stolonifer in postharvest of stone fruit by Pantoea agglomerans EPS125 and putative mechanisms of antagonism. International Journal of Food Microbiology 84: 93-104. Coenye, T., Vandamme, P. 2003. Diversity and significance of Burkholderia species occupying diverse ecological niches. Environmental Microbiology 5: 719-729. Costa, E., Teixidó, N., Usall, J., Atarés, E., Viñas, I. 2001. Production of the biocontrol agent Pantoea agglomerans strain CPA-2 using commercial products and by-products. Applied Microbiological Technology 56: 367- 371. De Champs, C., Le Seaux, S., Dubost, J.J., Boisgard, S., Sauvezie, B., Sirot, J. 2000. Isolation of Pantoea agglomerans in two cases of septic monoarthritis after plant thorn and wood sliver injuries. Journal of Clinical Microbiology 38: 460-461. Deletoile, A., Decre, D., Courant, S., Passet, V., Audo, J., Grimont, P., Arlet, G., Brisse, S. 2009. Phylogeny and identification of Pantoea species and typing of Pantoea agglomerans strains by multilocus gene sequencing. Journal of Clinical Microbiology 47: 300-310. Dey, R., Pal, K. K., Tilak, K. V. B. R. 2012. Influence of soil and plant types on diversity of rhizobacteria. Proceedings of the National Academy of Sciences, India 82: 341 – 352. Ferreira, A., Quecine, M. C., Lacava, P. T., Oda, S., Azevedo, J. L., Araujo, W. L. 2008. Diversity of endophytic bacteria from Eucalyptus species seeds and colonization of seedlings by Pantoea agglomerans. Federation of European Microbiological Societies: Microbiology Letters 287: 8 – 14. Gent, D. H., Schwartz, H. F. 2008a. ‘Bacterial Stalk and Leaf Necrosis’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 66 - 67. Gent, D. H., Schwartz, H.F. 2008b. ‘Center rot’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 64 - 66. Gitaitis, R., Walcott, R., Culpepper, S., Sanders, H., Zolobowska, L., Langston, D. 2003. Recovery of Pantoea ananatis, causal agent of center rot of onion, from weeds and crops in Georgia, USA. Crop Protection 21: 983-989. Gitaitis, R. D., Walcott, R. R., Sanders, H. F., Zolobowska, L., Diaz-Perez, J. C. 2004. Effects of mulch and irrigation system on sweet onion: II. The epidemiology of center rot. Journal of the American Society of Horticultural Science 129: 225-230.
69
Gitaitis, R. D., Gent, D. H., Schwartz, H. S. 2008. ‘Leaf streak and bulb rot’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 59. Hardoim, P. R., van Overbeek, L. S., van Elsas, J. D. 2008. Properties of bacterial endophytes and their proposed role in plant growth. Trends in Microbiology 16: 463 – 471. Holt, J. G., Krieg, N. R., Sneath, P. H. A., Staley, J. T., Williams, S. T. 1994. ‘Genus Pseudomonas;’ ‘Genus Pantoea.’ In: Bergey’s Pp. 93, 184. Hosmer, D. W., Lemeshow, S. 1989. Applied Logistic Regression. Wiley-Interscience Publications: New York, NY. 307 pp. Hu, F.-P., Young, J. M. 1998. Biocidal activity in plant pathogenic Acidovorax, Burkholderia, Herbaspirillum, Ralstonia, and Xanthomonas spp. Journal of Applied Microbiology 84: 263 – 271. Huber, D. M., Thompson, I. A. 2007. ‘Nitrogen and Plant Disease’ in: Mineral Nutrition and Plant Disease. APS Press: St. Paul, MN. Pp. 31 – 44. Louws, F. J., Fulbright, D. W., Stephens C. T., de Bruijn, F. J. 1994. Specific genomic fingerprints of phytopathogenic Xanthomonas and Pseudomonas pathovars and strains generated with repetitive sequences and PCR. Applied and Environmental Microbiology 60: 2286: 2295. Kloepper, J. W., Schroth, M. N. 1981. Relationship of in vitro antibiosis of plant growth-promoting rhizobacteria to plant growth and the displacement of root microflora. Phytopathology 71: 1020 – 1024. Kyung, K. H. 2012. Antimicrobial properties of Allium species. Current Opinion in Biotechnology 23: 142 – 147. Ma, B., Hibbing, M. E., Kim, H.-S., Reedy, R. M., Yedidia, I., Breuer, J., Breuer, J., Glasner, J. D., Perna, N. T., Kelman, A., Charkowski, A. O. 2007. Host range and molecular phylogenies of the soft rot enterobacterial genera Pectobacterium and Dickeya. Phytopathology 97: 1150 - 1163. Maher, E. A., Kelman, A. 1983. Oxygen status of potato tuber tissue in relation to maceration by pectic enzymes of Erwinia carotovora. Phytopathology 73: 536 – 539. Myers, M. W., Curran, W. S., VanGessel, M. J., Calvin, D. D., Mortensen, D. A., Majek, B. A., Karsten, H. D.,
Roth, G. W. 2004. Predicting weed emergence for eight annual species in the northeastern United States.
Weed Science 52: 913-919.
Paibomasai, M., Celetti, M., Tesfaendrias, M. 2012. Update on stemphylium leaf blight of onions in
Ontario. OMAFRA Hort Matters. Accessed online July 14, 2014:
Toussaint, V., Benoit, D. L., Carisse, O. 2012. Potential of weed species to serve as a reservoir for Xanthomonas campestris pv. vitians, the causal agent of bacterial leaf spot of lettuce. Crop Protection 41: 64 – 70.
Uva, R. H., Neal, J. C., DiTomaso, J. M. 1997. Weeds of the Northeast. Cornell University Press: Ithaca, NY. 397 pp. Volksch, B., Thon, S., Jacobsen, I.D., Gube, M. 2009. Polyphasic study of plant and clinic associated Pantoea agglomerans reveals indistinguishable virulence potential. Infection, Genetics, and Evolution 9:1381-1391. Wright, P. J., Hale, C. N. 1992. A field and storage rot of onion caused by Pseudomonas marginalis. New Zealand Journal of Crop and Horticultural Science 20: 435 – 438. Wright, S. A. I., Zumoff, C. H., Schneider, L., Beer, S. V. 2001. Pantoea agglomerans strain EH318 produces two antibiotics that inhibit Erwinia amylovora in vitro. Applied Environmental Microbiology 67: 284 – 292. Yamamoto, S., Kasai, H., Arnold, D. L., Jackson, R. W., Vivian, A., Harayama, S. 2000. Phylogeny of the genus Pseudomonas: intrageneric structure reconstructed from the nucleotide sequences of gyrB and rpoD genes. Microbiology 146: 2385 – 2394. Yishay, M., Burdman, S., Valverde, A., Luzzatto, T., Ophir, R., Yedidia, I. 2008. Differential pathogenicity and genetic diversity among Pectobacterium carotovorum ssp. carotovorum isolates from monocot and dicot hosts support early genomic divergence within this taxon. Environmental Microbiology 10: 2746 – 2759.
71
Chapter 4: In-field management factors related to incidence of bacterial rot of onion (Allium cepa) in Pennsylvania and New York
Abstract
Bacterial diseases, including center and soft rots, are significant causes of onion crop loss in
Pennsylvania and New York, resulting in losses of up to 60%, even with conscientious management with
chemical and cultural methods. To identify environmental and production factors associated with high
incidence of bacterial rots of onion, replicated observational studies were undertaken in a total of 108
fields in PA and NY, in 2011 and 2012. Forty environmental and management variables were recorded
through three visits to each field in each season. Prior to data analysis, combined data from NY indicated
higher disease incidence in ‘Red Wing’ than the other eight varieties surveyed. Independent
environmental and management variables were placed in stepwise multiple linear regression models to
determine variables influential to total incidence of bacterial rot. Foliar nitrogen (N) and foliar carbon (C)
were negatively related to total incidence of bacterial rots of onion from the PA datasets, while pre-season
levels of soil nitrate (NO3) were positively related to total incidence of bacterial rots from PA and NY. Foliar
nutrient levels implicate early-season fertility in managing bacterial rots of onion, while associations
between soil NO3 and rot suggest ammonium (NH4) or organic N may be more effective N sources in
managing bacterial rots. In addition, soil temperatures near the physiological onset of onion bulbing were
positively related to total incidence of bacterial rot in PA in 2011 and in the combined NY dataset, which
agrees with previously published results. These results suggest greater complexity is necessary for N
fertility recommendations: not only should a field rate be suggested, but also the timing and type of N
applied may play roles in bacterial disease development. In addition, if possible, growers should also take
steps to lower soil temperatures, particularly near bulbing.
Introduction
Bacterial rots are significant diseases reducing marketable yields in commercial onion production
in Pennsylvania (PA) and New York (NY). Over several recent years, crop losses due to bacterial disease
have exceeded 60% for some PA farms, and up to 40% loss in NY, even with conscientious disease
management by these growers. In PA, leaf blight and bulb rot, caused by Pantoea agglomerans Beijerinck,
and soft rots, caused by Pectobacterium carotovorum pv. carotovorum Jones and Pseudomonas
marginalis pv. marginalis Stevens, are primary diseases, with the three species of bacteria frequently
occurring in complexes as well as being ubiquitous in environments. This contrasts to NY, where growers
typically experience higher losses due to slippery skin (Burkholderia gladioli pv. alliicola Burkholder), sour
72
skin (Burkholderia cepacia Burkholder), and center rot (Pantoea ananatis Serrano); the Burkholderia
species are primarily associated with soil (Coenye and Vandamme, 2003), and P. ananatis is considered
an emerging pathogen (Coutinho and Venter, 2009).
In addition, the two production systems vary in size, predominant soil types, cultivars grown, and
grower inputs. In PA, highly diversified vegetable growers are typically producing ‘Candy’ (a large, sweet,
fresh-market cultivar) onions from transplants in silt loam fields between 0.25 – 1.5 hectares in size, over
a 100-day season. These highly managed systems are typified by raised plastic mulch beds with two rows
of drip irrigation, planned, sometimes proprietary fertigation programs, regular crop rotation, and
directed, integrated management via copper/ethylene-bis-dithiocarbamate (EBDC) spray treatments,
insecticides, hand-weeding, and pre- and post-emergent herbicides. In NY, growers typically direct-seed
bold cooking onions which are suitable for long-term storage, in high organic matter muck soil fields
ranging in size from two to 40 hectares or more. These larger systems usually do not have raised beds or
plastic mulch, most fertilizer is applied pre-plant, and integrated management is focused on copper-EBDC
treatments for disease and rotated insecticides for thrips management over a 130-day season, from
roughly late May to late August or early September.
Chemical strategies used by onion producers are primarily applications of copper hydroxide in
combination with EBDC fungicides (Gent and Schwartz, 2008; Sanchez et al., 2014). Unfortunately,
chemical approaches are only marginally successful, as a result of their lack of systemic activity and
inability to prevent bacterial degradation within the bulb once the plant is infected. Some of the
pathogens of concern in these systems initially infect onion leaves, then move into the bulb over
subsequent weeks (Carr et al., 2013; Pfeufer, unpublished). In addition, copper-tolerant strains of P.
ananatis have been isolated from commercial onion fields in Georgia (Nischwitz et al., 2007), which
suggests resistant strains may also have developed in other areas where copper is relied heavily upon for
in-season management. EBDC fungicides, such as mancozeb, have been suggested to increase the
availability of the copper ion, and therefore increase its efficacy, in bacterial disease management
(Conover and Gerhold, 1981).
Cultural strategies used by onion producers in the management of bacterial rots include practicing
crop rotation, managing soil fertility (Diaz-Perez et al., 2003; Mohan, 2008a), using different types of
mulch (Gitaitis et al., 2004; Gugino et al., 2011), and performing persistent weed management in and
around onion fields to reduce the impact of weeds as pathogen ‘green bridges’ (Gitaitis et al., 2003). In
addition to biotic factors, high rainfall and high humidity may be associated with disease development,
because bacteria proliferate more quickly in moist environments and rain splash may aid in dispersal
73
(Schwartz et al., 2003), in addition to promoting lush susceptible growth of the host plant in non-irrigated
fields. After four seasons of growing onions and observing grower farms, it has been noted that low-lying
areas within onion fields regularly have higher incidences of bacterial disease (Hoepting, unpublished;
Pfeufer and Gugino, unpublished); whether this is implicated in excess soil water or fertilizer accumulation
is unknown. The onion pathogens discussed earlier have varying optimal temperatures for proliferation,
however, warmer air temperatures have been shown to be related to high levels of center rot (P. ananatis)
in Colorado (Schwartz et al., 2003). Since multiple sources of inoculum exist before and during the growing
season, and growers have several opportunities with which to lessen the impact of environmental factors,
it is imperative that they implement integrated disease management strategies at all times before and
during the season.
As noted above, bulb infections are difficult to manage as a result of their inaccessibility: non-
systemic, protectant chemicals will not reach their target pathogens if the bacteria are protected by layers
of host tissue. Several environmental conditions that are beneficial to onion growth, that is, warm
temperatures, high humidity, and transfer of plant nutrients from leaves into the bulb may also favor
pathogen proliferation. To refine disease management strategies for growers in the Northeast and Mid-
Atlantic regions, observational studies were undertaken in PA and NY in 2011 and 2012 to identify
management factors associated with bacterial rots of onion for further investigation in replicated field
trials. It was hypothesized that soil nitrogen and soil temperatures would be positively related to the
incidence of bacterial rots of onion in these systems.
Materials and Methods
Overview of sample sites
In 2011 and 2012, 28 and 26 Pennsylvania fields, and 22 and 32 New York fields, respectively,
were visited three times each over the course of the onion growing season. Farms were selected with the
assistance of regional extension educators; all PA growers grew ‘Candy’ onions on raised beds with plastic
mulch and drip irrigation, fields averaged between 0.5 – 1 hectare in size, and were actively rotated. The
majority of PA growers transplant onions with 15 cm between each plant within the row, four rows across
the bed, and 15 cm between each plant within the row. NY growers grew nine different cultivars (primarily
bold cooking onions, with one red cultivar included), which may or may not have been grown on plastic
mulch, ranged in size from five to 150 acres, with some fields monocropped with onions for more than
ten years. There were 24 repeat growers from PA between both years, however 52 of the 54 fields were
unique. There were five repeat growers from NY from both years, and in some cases the same field was
74
sampled both years and/or multiple onion fields were sampled on the same farm. Soil texture estimates
in PA were determined by selecting the cropped field as a region of interest in the NRCS Web Soil Survey
(http://websoilsurvey.sc.egov.usda.gov/App/WebSoilSurvey.aspx), choosing physical properties under
Soil Data Explorer, and surface layer as the depth under the percent clay (silt, sand) navigation tab.
Grower collaborator visits
Early season (visit 1)
In April/May of each year, in both states, characteristics of each production field were recorded,
including type of plastic mulch, plant spacing, approximate field size, general fertilizer information, and
grower approaches to disease and weed management. Three 9.14 m plots were established in an area of
the field visually determined to be representative of the entire field with respect to field length, width,
and topography. Each plot was one bed wide, with typical bed width of 1.44 m. Soil temperature sensors
(HOBO Pendant Temperature Data Logger; Onset, Pocasset, MA) were placed in two of the three plots at
a 3-in. soil depth beneath the plastic mulch. Composite soil samples were collected from each plot (six
samples equidistantly located in each plot, using a soil probe 2.54 cm (1-in.) in diameter to a depth of 7.5
cm (3-in.) for each sample) and bulk density was calculated. For each plot, soil was homogenized by hand
in plastic zip-bags and approximately 15 g was placed in a screw-top cup containing 100 mL 2M KCl and
shaken well. To measure potentially mineralizable nitrogen, another 15 g was placed in an empty screw-
top soil cup, and all six cups per field were placed on ice in an insulated cooler. On arrival at the lab, PA
samples in KCl were placed in a 4⁰C refrigerator for extraction, and dry cups were placed in a lab drawer,
with the lids loosened, for 31 - 40 days. KCl extracts were filtered through pre-moistened Whatman No. 1
filter paper, then extracts were analyzed for NO3-N and NH4-N using spectrophotometric analysis via the
VCl3 and salicylate-nitroprusside methods, respectively (Doane and Horwath, 2003; Mulvaney et al.,
1996). Soils were also analyzed for gravimetric water content (g H2O g soil-1) and bulk density. NY soil
samples were processed similarly (tests #2820 and 2511) through the Cornell Nutrient Analysis Lab (CNAL;
http://cnal.cals.cornell.edu/index.html).
Mid-season (visit 2)
In mid-season (second and third weeks of June in PA; third and fourth weeks of July in NY) each
year, plots were rated for weed pressure in and between raised beds on a 3-point scale (1 = few weeds, 2
= some weeds, 3 = very weedy), plastic mulch integrity on a three-point scale (1 = intact mulch, 2 = some
tears, 3 = very torn and/or mulch slashed), foliar thrips damage (ten plants per plot; 0-100% damage scale,
modified from Nault and Shelton, 2010), and bulb growth (visual estimate; 1 = < 5.1 cm, 2 = 5.1 to 7.6 cm,
3 = > 7.6 cm). Five different, prevalent weeds were sampled from the field, placed in individual sealed
plastic bags, stored on ice, and then 4⁰C refrigeration until further processing. In PA , the fifth leaf of ten
equidistantly-spaced plants per plot was harvested, comingled in a paper bag, and dried in a 65⁰C drying
oven for ≥ 48 hr, and then was processed according to the tissue N protocol below. In NY, the inner four
leaves of ten plants were harvested, comingled by plot, dried, and processed according to standard CNAL
protocols for tissue analysis (test #6745).
Harvest (visit 3)
Growers were surveyed informally about the growing season in relation to production constraints
due to the weather, pests, etc. Soil was sampled according to the protocol described in visit 1, but only
the 2M KCl cup samples were prepared and processed. In PA, the inner two rows within 4.6 m of each
plot were harvested, graded by size, and evaluated for bacterial disease incidence (approx. 60 bulbs). In
NY, the entire plot was harvested, graded by size, and evaluated for bacterial disease incidence (approx.
130 bulbs). Onions were graded into four size categories based on the bulb diameter: < 6.4 cm diameter
are graded small, 6.4-7.6 cm diameter are graded medium, 7.6-10.2 cm diameter are graded jumbo, and
>10.2 cm diameter are graded colossal. In 2011 and 2012, 20 and 50% of the symptomatic bulbs,
respectively, were retained for further processing. Approximately 30 jumbo-size, asymptomatic bulbs
were retained for postharvest storage evaluation. If 30 jumbo-size, asymptomatic bulbs were not
available, 30 representatively-sized bulbs from the plot were sampled.
Postharvest bulb ratings
In both years, approximately 30 asymptomatic, jumbo-size onions per replicate plot were cured
under burlap in a greenhouse with forced air for at least 72 hr, then placed in 4⁰C storage for 75-120 days
(depending on harvest date). In mid-late November of the harvest year, approximately four months
postharvest, the bulbs were sliced in half longitudinally, photo-documented, and evaluated for disease
incidence, denoted as a percentage of the 30 bulbs per replicate plot.
Tissue N content
In 2011 and 2012, one representative asymptomatic, jumbo-sized bulb from the storage rating
from each grower plot was placed in a labeled paper bag and dried for ≥72 hr at 60⁰C. Thoroughly dried
leaf and bulb tissue samples were each ground to a fine, homogenous powder using a Cyclone Sample
76
Mill grinder (UDY Corporation, Fort Collins, CO, USA) and maintained in tightly capped 50 mL (2011) or 25
mL (2012) centrifuge tubes until analysis. Analysis was performed in the Soils Research Cluster Lab in the
Department of Plant Sciences at Penn State, similar to (Russo, 2008; Westerveld et al., 2003). Briefly, a 4
g sample of each replicate sample was tared in a 502-186 foil cup (LECO Corporation, St. Joseph, MI) then
folded over to completely contain the powdered sample. Samples were analyzed via dry combustion in
EA1110 CHNSO Elemental Analyzer (Thermo Fisher Scientific, Milan, Italy), which produced a report
containing the C and N content of each sample. Elemental contents were calculated based on the weight
of the sample analyzed. NY samples were collected as described above and further processed by the CNAL
(test #6745).
Data analysis
Field values for each variable were represented by the average of three replicate plots and each
field was treated as a sample for each year and state. Total rot incidence was analyzed as the dependent
variable, which was the sum of the incidence of bacterial rot from harvest, plus the incidence of bacterial
rot at storage as a percentage of the remaining total of marketable bulbs. Data were analyzed using
multiple linear regression and the stepwise variable addition procedure (management variables), or the
one-way ANOVA procedure (NY cultivars) in Minitab 16.2 (Minitab Inc., State College, PA, USA). In some
datasets, transformations of total rot were necessary to ensure the datasets satisfied the assumption of
normality in the residuals (square root transformation, Table 3; logarithmic transformation, Tables 2, 4).
In some cases (PA-2011; combined NY), as many as three additional independent variables were not
included in models, since addition of the variables did not explain significantly more variation in the
models.
Interpretation of the interaction in the PA-2011 dataset was by simple slopes analysis (Aiken and
West, 1991), with foliar N values chosen as the sample mean ± 1 standard deviation and average soil
temperature three weeks preharvest chosen as the sample mean ± 0.5 °C (Fig. 1). The sample mean was
inputted as the foliar C value in all equations, since this variable did not interact with other independent
variables (cited in Preacher, 2003). Simple slopes were compared using a t test as described in Aiken and
West (1991), with α = 0.05.
Results
In PA in 2011, bacterial disease incidence at harvest from the set of 28 fields ranged from 0 to
50.2% and bacterial disease incidence from storage ranged from 2.1 to 41.6%. The summed total of
77
bacterial rot incidence for the set of fields, therefore, ranged from 3.3 to 69.5%, which was the dependent
variable for the model in Table 4.1. Independent variables that were significant to the model were foliar
N and C from full-size, asymptomatic leaves collected at midseason, average soil temperature during the
3rd week preharvest (14-21 days prior; 3-wk soil temperature), and the interaction term between foliar N
and 3-wk soil temperature; other interactions in the model were not significant (Table 4.1). Prior to
addition of the interaction term, no multicollinearity was indicated among foliar N, foliar C, and 3-wk soil
temperature (variance inflation factors [VIFs] < 1.5; data not shown). All variables in the final model were
negatively related to total rot except the interaction term, which when investigated more closely by
simple slopes analysis (Fig. 4.1), indicated a positive relationship between average soil temperature and
total bacterial rot except at low levels of foliar N (one standard deviation below the mean), where 3-wk
preharvest soil temperature becomes less influential in the model and total bacterial rot incidence was
always projected to be high (Fig. 4.1).
In PA in 2012, bacterial disease incidence at harvest from the set of 26 fields ranged from 0 to
14.7% and bacterial disease incidence from storage ranged from 6.7 to 64.1%. Total bacterial rot for the
set, therefore, ranged from 7.6 to 68.6%, which was the dependent variable for the model. On initial
stepwise model fitting, no independent variables were selected for the model at α to add = 0.15 and α to
remove = 0.2. When model-fitting parameters were relaxed to α to add = 0.25 and α to remove = 0.3, nine
independent variables were added to the model; early soil NO3, foliar leaf N, and foliar leaf C at midseason
were the first three variables added to the model. When placed in multiple linear regression, these
variables gave an adjusted R2 = 0.212, no interactions were present, and there was no indication of
multicollinearity between the variables (VIFs < 1.7; Table 4.2). A positive relationship between total
bacterial rot and levels of soil NO3 early in the season was indicated (Table 4.2), as well as negative
relationships between each of the foliar tissue variables and the dependent total rot variable (Table 4.2);
the foliar tissue results were consistent with the PA-2011 dataset (Table 4.1). From graphical investigation
of single foliar nutrient variables, foliar N had a much stronger linear relationship with total bacterial
disease incidence than foliar C (data not shown); in some cases, inclusion of foliar N was necessary for
foliar C to be significant to linear regression models (Tables 4.1, 4.2, and 4.4). When foliar nutrients were
represented as the ratio of C to N and years were combined, a positive linear relationship was indicated
between foliar C/N ratio and total bacterial disease incidence (Fig. 4.2), however, inclusion of both foliar
elemental variables in regression analyses explained more total variation in bacterial disease incidence
than stating the variables as a single ratio (data not shown).
78
Total bacterial disease incidence at harvest was very low in NY in both 2011 and 2012, with
averages of 5.2% and 4.6%, respectively. Nine different cultivars were grown and the cv. ‘Red Wing’ had
significantly higher total bacterial disease incidence than seven other cultivars (Fig. 4.3). As a result, the
‘Red Wing’ fields were removed before the stepwise model fitting step for multiple linear regression
analysis. Since total disease incidence did not differ by year, the NY datasets were combined after
removing the five ‘Red Wing’ fields. In the remaining data (N = 49), total bacterial disease incidence at
harvest ranged from 0 to 18.6% and bacterial disease incidence from storage ranged from 0 to 9.9%.
Therefore, total bacterial rot for the dataset ranged from 0 to 22.7%, which was the dependent variable
for the model. On stepwise model fitting with α to add = 0.15 and α to remove = 0.2, midseason soil NO3
and average soil temperature five weeks preharvest were both positively related to total bacterial rot
incidence (Table 4.3). Bulb N, bulb C, and onion thrips damage ratings were also suggested for the model,
but inclusion of all of these variables only explained approximately 2.1% more total variation (data not
shown), thus, the variables were excluded.
In the combined PA dataset, total bacterial rot incidence ranged from 3.3 to 69.5% among 54
fields. In stepwise model fitting, the selection parameters were restricted to α to add = 0.05, α to remove
= 0.1, and both foliar N and C measured at midseason were significant factors in the model (Table 4.4).
These factors were both negatively related to total incidence of bacterial rots of onion and their
interaction term was not statistically significant (Table 4.4).
Discussion
Intensive surveys of farms growing onions in 2011 and 2012 indicated one novel trend among PA
datasets and another that bolsters results from other replicated trials. In ‘Candy’ onion, plant nutrition
and the total incidence of bacterial rots of onion are negatively related. Specifically, foliar nitrogen (N)
and carbon (C) were influential factors in both PA datasets (Tables 4.1 and 4.2), and very strong factors in
the combined PA dataset (Table 4.4); in addition, soil NO3 was influential in the PA-2012 and combined NY
datasets (Table 4.2 and 4.3). Foliar N in the PA-2011 dataset interacted with soil temperatures three weeks
prior to harvest (Fig. 4.1).
The foliar tissue test results are in direct contrast to a-priori hypotheses, wherein it was thought
that excessive N fertility would result in high levels of bacterial disease at harvest, which has been
demonstrated in other systems (Diaz-Perez et al., 2003; Mohan, 2008a; Gitaitis et al., 2008). Instead, the
negative relationship between midseason foliar N and bacterial disease incidence may indicate that
‘Candy’ onion plants that are N-stressed early in the season may be more susceptible to bacterial
79
infection, proliferation, and/or movement into the onion bulb. Other research identified that tissue N
levels differed by onion cultivar (Westerveld et al., 2003), so additional data would be necessary to
determine if the relationship between foliar nutrients and disease incidence is a unique attribute of
‘Candy’ onions or if this relationship may be applied more broadly to all onion cultivars.
In the combined PA datasets, levels of early-season soil NH4 and the silt content of soils were
positively related to levels of foliar N (data not shown), so site selection and the type of N fertilizer applied
may be of interest to growers. Nitrogen cycling is dependent on a number of soil characteristics, including
temperature, moisture, oxygen, pH, organic matter, and microbial diversity (Daroub and Snyder, 2007),
and texturally, more silt in soils may help ensure optimal levels of these characteristics. In the NY
combined dataset, no associations were indicated between foliar nutrient levels and bacterial disease
incidence, but a positive relationship between midseason soil NO3 and total disease incidence was
indicated (Table 4.3). At-harvest levels of soil NO3 and NH4 were not associated with bacterial disease
incidence or the proportion of large-size bulbs, and taken together, these begin to suggest the timing of
N availability may be more influential than a season-long, rate-based recommendation, which aligns well
with anecdotal evidence shared by experienced onion growers included in this study. One onion grower,
who has commercially produced onions for ten years and designs his own fertility program, remarked that
he does not fertigate after the first week of June (roughly two weeks prior to bulbing), and simply irrigates
until harvest the first week of July (pers. communication). In addition, similar timing of nutrient
availability, where N applications were reduced or eliminated after bulbing, indicated no adverse effects
on onion size or maturity (Brewster and Butler, 1989).
Coupled with the midseason foliar N measurements, the PA datasets indicate foliar C at
midseason is another critical predictor of bacterial disease incidence at harvest and from storage (Tables
4.1, 4.2, 4.4). While Pearson’s correlations suggested foliar N and foliar C were correlated (data not
shown), interactions between these variables in linear regression models were not significant in these
analyses. It is generally accepted that phytopathogenic bacteria parasitize plants in an effort to secure
carbon for their own metabolism, and centers of infection may act as photoassimilate sinks in plants
(Kosuge and Kimpel, 1982), however, results pertaining to midseason foliar C indicate a negative
relationship with total bacterial disease incidence. Work completed in Arabidopsis where induction of the
plant defense-related genes ATL31, ATL6, and their knockouts were compared under conditions of C and
N stress indicated increased symptom severity caused by Pseudomonas syringae pv. tomato DC3000
under nutrient-limiting conditions, and ATL31- mutant plants were less sensitive to low nutrient
conditions (Maekawa et al., 2012). ATL31 has been shown to function in both defense responses and
80
coordinate post-germinative growth in accordance with C/N ratio, however, this occurred under high C/N
conditions (Sato et al., 2009). This contrasts to the PA datasets presented here, where C/N ratio was
shown to be positively related to total incidence of bacterial rot of onion (Fig. 4.2). Specific assays would
need to be completed with onion in order to elucidate the effects of high tissue C/N ratios on bacterial
disease severity.
Other nutrient-related variables of interest associated with total bacterial disease incidence in
some of the models include positive relationships with the early-season (Table 4.2) or midseason NO3
levels in soil (Table 4.3). The results pertaining to NO3 levels are similar to results in other pathosystems,
where high levels of soil NO3 were correlated with high bacterial disease incidence (Rotenberg et al.,
2005), and taken together with the previously-mentioned positive relationship between early-season soil
NH4 and midseason leaf N, may indicate that onion, as a crop, prefers NH4 – N to NO3 - N. Associations
between disease incidence and the type of inorganic N fertilizer have been observed for other crops,
however, each pathosystem varies in how these relationships manifest (Elmer, 2000; Elmer and
LaMondia, 1999; Thompson and Huber, 2007). One suggestion from these works is that higher NH4
concentrations promote the acidity of soil and thus allow micronutrients, such as Mn and Zn, to be more
readily taken up by plants (Elmer, 2000; Elmer and LaMondia, 1999); higher levels of these micronutrients
have been implicated in bacterial disease reduction (Thompson and Huber, 2007). Combined with the lack
of reproducibility between years and sample sites, this theme indicates additional replicated trials may be
necessary before changes in fertility recommendations may be made to growers.
Positive relationships between soil temperatures three or five weeks preharvest and the total
incidence of bacterial rots were observed in the PA-2011 (Table 4.1) and the NY datasets (Table 4.3),
respectively. Positive relationships between air temperature as reported by weather stations and
bacterial rot are fairly well-established in the onion pathosystem (Gitaitis et al., 2008; Mohan, 2008a;
Schwartz et al., 2003), and soil temperatures have been suggested to be closely associated with center
rot of onion (Diaz-Perez et al., 2003; Gitaitis et al., 2003), however, explicit relationships between soil
temperatures and other bacterial rots of onion have not been previously reported. In addition, three and
five weeks preharvest in PA and NY, respectively, roughly correspond with the physiological onset of
bulbing, which relates well to results reported by Schwartz et al., where high temperatures at bulbing
were associated with initial onset of Xanthomonas leaf streak in onion (2003). In the PA-2011 dataset, an
interaction between soil temperature three weeks preharvest and levels of foliar N existed, but when
investigated more fully through simple slopes analysis (Aiken and West, 1991), it was predicted that soil
temperature would only play a role in increasing bacterial disease incidence at average and above-average
81
levels of foliar N at midseason (Fig. 4.1). At below-average levels of foliar N, it was predicted that total
bacterial disease incidence would always be high, regardless of soil temperatures three weeks preharvest
(Fig. 4.1). The potential interaction between soil temperatures five weeks preharvest and midseason soil
NO3 was not significant in the combined NY model (Table 4.3). These results relate well to the influence
of soil temperature on bacterial disease incidence in the use of different types of plastic mulches. Through
several replicated trials, it was shown that plots with plastic mulches that kept soils cooler also had lower
incidence of bacterial rots than plots with high-soil temperature mulches (Gugino et al., 2011).
Since the datasets discussed here are largely observational, results may be used to develop more
targeted replicated field trials in the future, with the overall goal of improving management
recommendations for onion growers. The relationships between midseason foliar N and C, soil NO3, and
total incidence of bacterial rots of onion (Tables 4.1, 4.2, 4.3, 4.4) indicate further research into the type
and timing of nitrogen fertility is necessary. The positive relationship between soil temperatures in PA-
2011, NY-2011/2012, and bacterial rot incidence has been indicated in several replicated trials, where
plastic mulches that reduced soil temperatures were shown to result in lower disease incidence (Gugino
et al., 2011). These datasets have resulted in revisions to production guides such that black biodegradable
plastic mulches are now recommended to help moderate soil temperatures, particularly after midseason.
Acknowledgements
Soil nitrate, ammonium, and PMN were processed in Dr. Jason Kaye’s lab in the Department of Ecosystem Resource Management at Penn State, with the assistance of Dr. Kaye, Sara Eckert, and Kaye lab members.
Foliar and bulb tissue N and C content were processed in the Soils Research Cluster Lab at Penn State with the assistance of Dr. Ephraim Govere.
Technical assistance was provided by Dr. Michele Mansfield, Tim Grove, Ilse Huerta, Evan Stover, Jill Pollok, Robert, Andrew Hower, Khanh Nguyen, Lizzie King’ang’i, Rosemary Schwegel, Marie Ebner, Anna Testen, Laura Ramos, Laura del Sol Bautista, and Andy Kelly.
Many thanks to our truck benefactor, Dr. Scott Isard.
82
Tables and Figures
Table 4.1. Field-averaged results of multiple linear regression analysis of PA-2011 dataset. All variables were placed in a stepwise model selection procedure (α to add = 0.15, α to remove = 0.2), with total bacterial rot incidence as the dependent variable. For this model, R2 = 0.557; adj. R2 = 0.480; P = 0.001.
Predictor Coefficient SE of Coefficient T value P value
a The fifth leaf from ten equidistantly spaced plants was collected at the midseason visit, bulked by plot, dried, homogenized, and analyzed for total C and N via dry combustion. Three replicate plots were averaged by field prior to analysis. b Soil temperatures 3-in. beneath the soil surface, beneath the plastic mulch, were recorded hourly throughout the growing season. Datasets were normalized by harvest date, then weekly averages were calculated; this variable represents the time period 14-21 days preharvest.
Fig. 4.1. Simple slopes analysis of projected relationships between average soil temperature three weeks preharvest and the incidence of total bacterial rot of onion, given different levels of foliar N, from PA-2011. Points of each line were calculated based on the covariance matrix of the multiple regression model in Table 1 (Aiken and West, 1991). Foliar N values (2.41% N [low], 2.75% N [avg], 3.09% N [high]; ±1 standard deviation from the sample mean) were chosen, then projected bacterial disease estimates were calculated based on chosen average soil temperatures three weeks preharvest (23.79, 24.29, 24.79°C; ± 0.5 °C from the sample mean). Simple slopes were compared to H0 = 0; for average and high foliar N lines, t was significant at P = 0.03.
83
Table 4.2. Field-averaged results of multiple linear regression analysis of PA-2012 dataset. All variables were placed in a stepwise model selection procedure (α to add = 0.25, α to remove = 0.3), with a logistic transformation of total bacterial rot incidence as the dependent variable. For this model, R2 = 0.314; adj. R2 = 0.212; P = 0.052.
Predictor Coefficient SE of Coefficient T value P value
Constant 7.513 2.971 2.53 0.020 Foliar N, midseason a -0.377 0.147 -2.56 0.018 Foliar C, midseason a -0.136 0.071 -1.91 0.070 Early-season soil nitrate b 0.004 0.002 1.98 0.062
a The fifth leaf from ten equidistantly spaced plants was collected at the midseason visit, bulked by plot, dried, homogenized, and analyzed for total C and N via dry combustion. Three replicate plots were averaged by field prior to analysis. b Soil from the first farm visit was analyzed for nitrate using the VCl3 protocol.
Fig. 4.2. Total bacterial disease incidence by foliar C/N ratio, combined data from PA-2011 and PA-2012. Ten leaves per plot were co-mingled, dried, homogenized, and analyzed for total C and N via dry combustion. Total bacterial disease incidence was the sum of the percentages of symptomatic bulbs at harvest and from storage as a total of the bulbs harvested per plot. Three plots were averaged for each field value; each point represents one field.
Table 4.3. Field-averaged results of multiple linear regression analysis of combined NY-2011 and 2012 datasets. Independent variables were observed in 22 and 32 fields, respectively. All variables were placed in a stepwise model selection procedure (α to add = 0.15, α to remove = 0.2), with a square root transformation of total bacterial rot incidence for each field as the dependent variable. For this model, R2 = 0.161; adj. R2 = 0.126; P = 0.019.
Predictor Coefficient SE of Coefficient T value P value
Constant -7.138 3.884 -1.84 0.073 Midseason soil nitrate a 0.003 0.001 2.51 0.016 Average soil T, 5 wk preharvest b 0.112 0.052 2.17 0.035
a Soil from the second farm visit was analyzed for nitrate using the CNAL protocol. b Soil temperatures 3-in. beneath the soil surface, were recorded hourly throughout the growing season. Datasets were normalized by harvest date, then weekly averages were calculated; this variable represents the time period 28-35 days preharvest.
84
Fig. 4.3. Total bacterial rot incidence by cultivar grown, NY-2011 and NY-2012. Data were analyzed using a one-way ANOVA in Minitab 16.2, error bars represent the standard error of the mean, and letters above each bar indicate statistically significant differences by Fisher’s LSD (α = 0.05).
Table 4.4. Field-averaged results of multiple linear regression analysis of combined PA datasets. All independent variables observed in 54 fields were placed in a stepwise model selection procedure (α to add = 0.05, α to remove = 0.1), with a log transformation of total bacterial rot incidence as the dependent variable. For this model, R2 = 0.179; adj. R2 = 0.147; P < 0.001.
Predictor Coefficient SE of Coefficient T value P value
a The fifth leaf from ten equidistantly spaced plants was collected at the midseason visit, bulked by plot, dried, homogenized, and analyzed for total C and N via dry combustion. Three replicate plots were averaged by field prior to analysis.
References
Aiken, L. S., West, S. G. 1991. ‘Interactions between Continuous Predictors in Multiple Regression’ in: Multiple Regression: Testing and Interpreting Interactions. SAGE Publications, Inc: Thousand Oaks, CA. pp. 12-14. Brewster, J. L., Butler, H. A. 1989. Effects of nitrogen supply on bulb development in onion Allium cepa L. Journal of Experimental Botany 40:1155-1162. Bull, C. T., De Boer, S. H., Denny, T. P., Firrao, G., Fischer-Le Saux, M., Saddler, G. S., Scortichini, M., Stead, D. E., Takikawa, Y. 2010. Comprehensive list of names of plant pathogenic bacteria, 1980 – 2007. Journal of Plant Pathology 92: 551 – 592. Carr, E. A., Zaid, A. M., Bonasera, J. M., Lorbeer, J. W., Beer, S. V. 2013. Infection of onion leaves by Pantoea ananatis leads to bulb infection. Plant Disease 97:1524-1528.
85
Coenye, T., Vandamme, P. 2003. Diversity and significance of Burkholderia species occupying diverse ecological niches. Environmental Microbiology 5: 719-729. Conover, R. A., Gerhold, N. R. 1981. Mixtures of copper and maneb or mancozeb for control of bacterial spot of tomato and their compatibility for control of fungus diseases. Proceedings of the Florida State Horticultural Society 94: 154-156.
Coutinho, T. A., Venter, S. N. 2009. Pantoea ananatis: an unconventional plant pathogen. Molecular Plant Pathology 10: 325-335.
Daroub, S. H., Snyder, D. H. 2007. ‘The Chemistry of Plant Nutrients in Soil’ in: Mineral Nutrition and Plant Disease. Editors: L. E. Datnoff, W. H. Elmer, D. M. Huber. APS Press: St. Paul, MN.
Diaz-Perez, J. C., Purvis, A. C., Paulk, J. T. 2002. Bolting, yield, and bulb decay of sweet onion as affected by nitrogen fertilization. Journal of the American Horticultural Society 128:144-149.
Doane, T. A., Horwath, W. R. 2003. Spectrophotometric determination of nitrate with a single reagent. Analytical Letters 36: 2713-2722. Elmer, W. H. 2000. Comparison of plastic mulch and nitrogen form on the incidence of Verticillium wilt of eggplant. Plant Disease 84:1231-1234. Elmer, W. H., LaMondia, J. A. 1999. Influence of ammonium sulfate and rotation crops on strawberry black root rot. Plant Disease 83:119-123. Gent, D. H., Schwartz, H. F. 2008. ‘Bacterial Stalk and Leaf Necrosis’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 66-67.
Gent, D. H., Schwartz, H. F. 2005. Management of Xanthomonas leaf blight of onion with a plant activator, biological control agents, and copper bactericides. Plant Disease 89:631-639.
Gitaitis, R., Walcott, R., Culpepper, S., Sanders, H., Zolobowska, L., Langston, D. 2003. Recovery of Pantoea ananatis, causal agent of center rot of onion, from weeds and crops in Georgia, USA. Crop Protection 21: 983-989.
Gitaitis, R. D., Walcott, R. R., Sanders, H. F., Zolobowska, L., Diaz-Perez, J. C. 2004. Effects of mulch and irrigation system on sweet onion: II. The epidemiology of center rot. Journal of the American Society of Horticultural Science 129: 225-230. Gitaitis, R. D., Gent, D. H., Schwartz, H. S. 2008. ‘Leaf streak and bulb rot’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 59. Gugino, B. K., Hoepting, C. A., Stoltzfus, J., Mansfield, M. A., Klotzbach, K., Reid, J. E. 2011. Management of onion bacterial diseases using alternative mulches and plant spacing. Phytopathology 101: S65. Kosuge, T., Kimpel, J. A. 1982. ‘Altered Metabolism: Response to Infection’ in Phytopathogenic Prokaryotes, Volume 1. Edited by M. Mount and G. H. Lacy. Academic Press, Inc: New York, NY. pp 366-368.
86
Maekawa, S., Sato, T., Asada, Y., Yasuda, S., Yoshida, M., Chiba, Y., Yamaguchi, J. 2012. The Arabidopsis ubiquitin ligases ATL31 and ATL6 control the defense response as well as the carbon/nitrogen response. Plant Molecular Biology 79:217-227.
Mohan, S. K. 2008. ‘Other Bacterial Soft Rots’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 62.
Mulvaney, R., Sparks, D., Page, A., Helmke, P., Loeppert, R., Soltanpour, P., Tabatabai, M., Johnston, C., Sumner, M. 1996. Nitrogen-Inorganic Forms. Methods of Soil Analysis. Part 3-Chemical Methods. pp. 1123–1184.
Nault, B. A., Shelton, A. M. 2010. Impact of insecticide efficacy on developing action thresholds for pest management: a case study of onion thrips (Thysanoptera: Thripidae) on onion. Journal of Economic Entomology 103: 1315 – 1326.
Nischwitz, C., Gitaitis, R., Sanders, H., Langston, D., Mullinix, B., Torrance, R., Boyhan, G., Zolobowska, L. 2007. Use of fatty acid methyl ester profiles to compare copper-tolerant and copper-sensitive strains of Pantoea ananatis. Phytopathology 97: 1298-1304. Preacher, K. J. 2003. ‘A primer on interaction effects in multiple linear regression.’ Vanderbilt University, accessed online July 27, 2014: http://www.quantpsy.org/interact/interactions.htm Rotenberg, D., Cooperband, L., Stone, A. 2005. Dynamic relationships between soil properties and foliar disease as affected by annual additions of organic amendment to a sandy-soil vegetable production system. Soil Biology and Biochemistry 37:1343-1357. Russo, V. M. 2008. Plant density and nitrogen fertilizer rate on yield and nutrient content of onion developed from greenhouse-grown transplants. HortScience 43: 1759-1764. Sanchez, E., Elkner, T. E., Lamont, W. J., Demchak, K., Orzolek, M. D., Gugino, B. K., Halbrendt, J. M., Fleischer, S. J., LaBorde, L., Hoffman, K., San Julian, G. J. 2014. Commercial Vegetable Production Recommendations – Pennsylvania 2014. Penn State Cooperative Extension, Pennsylvania Agricultural Experiment Station, and The Pennsylvania State University, University Park, PA. 329 pp. Sato, T., Maekawa, S., Yasuda, S., Sonoda, Y., Katoh, E., Ichikawa, T., Nakazawa, M., Seki, M., Shinozaki, K., Matsui, M., Goto, D. B., Ikeda, A., Yamaguchi, J. 2009. CNI1/ATL31, a RING-type ubiquitin ligase that functions in the carbon/nitrogen response for growth phase transition in Arabidopsis seedlings. The Plant Journal 60: 852–864.
Schwartz, H. F., Otto, K. L., Gent, D. H. 2003. Relation of temperature and rainfall to development of Xanthomonas and Pantoea leaf blights of onion in Colorado. Plant Disease 87:11-14.
Thompson, I. A., Huber D. M. 2007. ‘Manganese and Plant Disease’ in: Mineral Nutrition and Plant Disease. Edited by L.E. Datnoff, W.H. Elmer, D. M. Huber. APS Press: St. Paul, MN.
Westerveld, S. M., McKeown, A. W., Scott-Dupree, C. D., McDonald, M. R. 2003. How well do critical nitrogen concentrations work for cabbage, carrot, and onion crops? HortScience 38: 1122-1128.
Chapter 5: Efficacy of plant defense-inducing and growth-promoting products for the management of center rot of onion, caused by Pantoea ananatis and P. agglomerans
Abstract
Center rot, caused by the bacteria Pantoea ananatis and Pantoea agglomerans, is an emerging
disease of onion in Pennsylvania. Growers primarily manage onions for bacterial diseases using copper
mixed with ethylene-bis-dithiocarbamate fungicides, however, low efficacy has been reported for this
practice. A number of plant defense-inducing and growth-promoting products are labeled for onion,
however, little data on the efficacy of these products was available. To determine the efficacy of these
products for the management of center rot of onion, replicated field trials were undertaken in two
locations in 2011 and 2012, using two different inoculation techniques. The 2012 inoculation technique
effectively resulted in disease pressure at three different levels. In low inoculum pressure plots, the
grower standard copper-EBDC, hydrogen peroxide, harpin alternated with copper-EBDC, and acibenzolar-
S-methyl treatments resulted in lower disease incidence than the untreated control, but no treatments
were effective in the untreated plots at higher inoculum pressures. Though not statistically significant,
plants treated with Bacillus subtilis GB03 consistently had lower disease incidences than other treatments.
Introduction
Pennsylvania growers may suffer yield losses of 50% or more due to bacterial diseases of onion
(Allium cepa L.), including center (caused by Pantoea ananatis Serrano and Pantoea agglomerans
Beijerinck) and soft rots (caused by Pectobacterium carotovorum subsp. carotovorum Jones and
Pseudomonas marginalis pv. marginalis Stevens). Center rot symptoms typically develop following the
infection of a single leaf, with the bacteria progressively moving proximally into the corresponding
enlarged scale in the onion bulb (Carr et al., 2013). The primary chemical management strategy for
bacterial diseases of onions is the use of preventative applications of copper hydroxide combined with
ethylene-bis-dithiocarbamate (EBDC) fungicides, which are reported to increase the efficacy of copper
(Conover and Gerhold, 1981; Gent and Schwartz, 2005). EBDC fungicides, such as mancozeb, have been
shown to be class B-2 carcinogens, were cancelled for selected uses in 1992, including green onion
(EXTOXNET, 1992), and are not permissible for use in organic production systems, which has led some
growers to call for alternative chemical management strategies for onion production. Most importantly,
even under conscientious management using the copper - EBDC grower standard, extensive losses due to
bacterial disease may still occur.
88
Some interest has been expressed in plant defense-inducing compounds as alternatives to
traditional copper-EBDC treatments for the management of bacterial bulb rots of onion. Several different
types of plant defense-inducing compounds have been suggested as effective management tools for plant
disease, particularly viral and bacterial diseases (Gent and Schwartz, 2005; Louws et al., 2001; Romero et
al., 2001). In addition, commercial plant growth-promoting products have been suggested to activate
induced systemic resistance (ISR) or systemic acquired resistance (SAR), which increase plants’ natural
defenses against pathogens and insects (van Wees et al., 2000; Walters et al., 2005; Siddiqui, 2006).
Salicylic acid (SA), a defense signaling molecule that naturally occurs in plants, activates systemic
acquired resistance (SAR), which provides nonspecific, long-lasting, systemic control against plant
pathogens as part of pathogen associated molecular pattern (PAMP)- and effector- triggered immunity
(Jones and Dangl, 2007; van Wees et al., 2000; Walters et al., 2005; Walters et al., 2009). SA biosynthesis
and SA-induced genes have been shown to be activated in response to plant infection with biotrophic
pathogens (Spoel et al., 2007), and implicated in interactions between Pseudomonas syringae pv. tomato
and tomato (Solanum lycopersicum; Thaler et al., 2002), and a number of other pathosystems.
Acibenzolar-s-methyl (ASM), a synthesized analog of SA, is one of an increasing group of plant defense-
inducing products marketed to conventional vegetable growers, particularly those who suspect resistance
to copper-based treatments (Gugino, pers. comm.) In a study on the onion foliar pathogen X. axonopodis
pv. allii Hasse, Gent and Schwartz (2005) evaluated combinations of plant defense activators and
biological control agents along with traditional copper applications in the management of onion leaf
blight. Reduced-copper applications combined with ASM were comparable or more effective for leaf
blight control compared to a traditional copper - EBDC spray program (Gent and Schwartz, 2005).
Harpin protein, a pathogenicity factor isolated from Erwinia amylovora (Wei et al., 1992), was first
found to elicit a hypersensitive response, and later, to activate systemic acquired resistance (SAR) in
Arabidopsis (Dong et al., 1999). Harpin protein homologues have been shown to be synthesized by several
plant pathogenic strains of bacteria, including causal agents of soft rots (Nasser et al., 2005) and Pantoea
stewartii subsp. stewartii (Ahmad et al., 2001), which is fairly closely related to P. ananatis. By utilizing
transgenic E. coli, harpin protein is now mass-produced and marketed commercially as a plant defense
inducing product (EPA, 2011).
The potential for acibenzolar-S-methyl (ASM) and harpin protein to assist in managing bacterial
disease in onion bulbs suggests these products are worth further investigation. Currently, ASM is
marketed in the US as Actigard 50WG (Syngenta Crop Protection, Greensboro, NC) and onions are
included on the product label. Purified harpin protein (Employ or ProAct; Plant Health Care, Inc.,
89
Pittsburgh, PA) is marketed as a preventative treatment for a variety of diseases on virtually any crop.
While ASM functions as a mimic of salicylic acid (SA), harpin protein triggers plants to naturally produce
SA (EPA, 2011). These types of plant-induced defenses may come at the cost of yield as a result of a
potential deficit in metabolic resources, due to maintenance of high levels of defense proteins (Gent and
Schwartz, 2005; Louws et al., 2001; Romero et al., 2001; Walters and Fountaine, 2009). This ‘yield-drag’
effect may be especially pronounced, as both product labels warn, during periods of plant stress.
Mitigation of this effect has been reported by the use of plant-growth promoting rhizobacteria (PGPR) in
amaranthus (Nair et al., 2006) and tomato (Obradovic et al., 2005), so combinations of SAR inducers and
plant-growth promoters present potential to manage bacterial disease while still producing profitable
yields.
Plant growth-promoting rhizobacteria (PGPR) have been shown to induce positive effects on
plants in a number of ways, including nitrogen fixation, phosphate solubilization (Lugtenberg and
Kamilova, 2009; Van Loon, 2007), soil pollutant remediation (Kuiper et al., 2001), and the stimulation of
root hair growth (Brazelton et al., 2008; Dodd et al., 2010). As biocontrol agents of soilborne pathogens,
PGPR may exclude pathogens from the rhizosphere by resource competition, inhabiting facultative
pathogen niches, or by producing secondary metabolites, such as antibiotic compounds or siderophores
(Kokalis-Burelle et al., 2003; Lugtenberg and Kamilova, 2009; Van Loon, 2007). Others suggest that PGPR
may interfere with pathogen quorum sensing systems (Lin et al., 2003), which help coordinate pathogen
infection of potential hosts. The most notable effect of PGPR, however, in terms of plant disease, may be
the positive regulation of induced systemic resistance (ISR). ISR is similar to SAR in its result (heightened
plant defenses against a variety of plant pathogens), but the two types of resistance differ in terms of
signaling molecules, genetic determinants, and inducing organisms (Siddiqui, 2006).
Arbuscular mycorrhizal fungi (AMF) function in a similar relationship to some PGPR: AMF are
fungal symbionts that form associations with terrestrial plants and have been shown to reduce the
incidence and severity of diseases caused by soilborne fungi, oomycetes, and nematodes (Whipps, 2004).
Some studies have also examined the role of AMF in reducing foliar bacterial diseases in tomato and alfalfa
(Garcia-Garrido and Ocampo, 1988; Liu et al., 2007). Specifically, Liu et al. recorded an overall increase in
stress and defense-associated gene expression in Medicago truncatula shoots and roots following the
development of mycorrhizal associations, which was coordinated with an overall reduction in foliar
symptoms caused by Xanthomonas campestris pv. alfalfae (2007). The authors concluded the M.
truncatula response to the formation of mycorrhizal associations was homologous to induction of ISR (Liu
et al., 2007); this is supported by the results that AMF associations with plants are mediated by the same
90
signaling molecules as PGPR-induced ISR (Pozo and Azcon-Aguilar, 2007). Rather than its effects on plant
defenses, however, the majority of horticultural studies on AMF focus on their ability to improve water
and nutrient uptake, especially phosphorus, in arid or nutrient-poor soils (Borowicz, 2001; Simard and
Durall, 2004). In this way, PGPR and AMF may alternatively serve to reduce plant nutrient stress by way
of their suggested roles of enhancing nutrient uptake. These additional nutrients could mitigate yield
reduction effects when used in combination with plant defense-inducers.
In addition to the previously mentioned plant defense-inducing and plant growth-promoting
products, a number of growers have begun to use surface-sterilizing type treatments to manage bacterial
diseases in their onions. In particular, OxiDate (active ingredient hydrogen dioxide, BioSafe Systems,
Glastonbury, CT) is marketed to growers as a broad-spectrum bactericide/fungicide, and is an OMRI-listed
product approved for organic production systems. The product is labeled for onion for the management
of Botrytis, downy mildew, powdery mildew, and in particular, has some efficacy data available for
management of Xanthomonas leaf blight. In addition, the product label lists two concentrations for
product use on onion, ‘preventative,’ and ‘curative,’ both on a 5-7 day spray interval; curative is simply
the maximum concentration, while preventative is a reduced concentration. Several PA growers have
already begun treating their onion crops using hydrogen peroxide, however, efficacy data for
management of typical PA diseases, such as center rot, are lacking overall.
With expressed grower interest in the aforementioned foliar plant defense-inducing products, soil
treatments for induced disease resistance, and surface sterilizants, replicated studies of selected
commercial products were undertaken on two university research farms in 2011 and 2012. Two
inoculation protocols were explored, and in-season growth and disease estimates were recorded in
addition to harvest yields and disease incidence.
Materials and Methods
Field trial establishment and maintenance
In 2011 and 2012, replicated research trials were established at the Russell E. Larson Agricultural
Research and Education Center in PA Furnace, PA (Rock Springs), and the Southeast Agricultural Research
and Extension Center in Manheim, PA (Landisville); soil in both fields were Hagerstown silt loam. The fields
were plowed and prepared following standard commercial production practices. Raised beds with 1.0 or
1.5-mil standard black plastic mulch and a double row of drip-irrigation tape were formed using a raised-
bed plastic mulch layer (Rain-Flo Irrigation Inc., East Earl, PA). Beds were spaced on 2.4 or 3 m centers
depending on the year and trial location. ‘Candy’ onion transplants (Dixondale Farms, Carrizo Springs, TX)
91
were spaced 15.2 cm apart in the row and 15.2 cm apart between rows, with four rows across an approx.
1 m bed. Plots were one bed wide and 3.7 or 4.3 m long with 0.3 or 1.5 m breaks between plots, depending
on year and location. Insects were managed using Radiant SC (0.44 -0.73 L / ha, Dow AgroSciences,
Indianapolis, IN) and/or Warrior 1EC (0.19 – 0.28 L / ha, Syngenta Crop Protection, Greensboro, NC) as
necessary and weeds were hand-pulled within the row and mowed between rows. Onions were fertilized
through drip irrigation weekly with a water-soluble 30% liquid urea fertilizer (Helena Chemical Co.,
Collierville, TN, USA) following standard commercial production practices, typically application of 16.8 kg
/ ha per week, to a field rate of nitrogen of 168 - 196 kg / ha for the season.
In 2011, onions were transplanted on April 19 and 21 at research farms in Rock Springs and
Landisville, respectively, and arranged in a randomized block design with four replicates in each trial;
untreated controls were included as treatments in this design. In 2012, onions were transplanted on April
17 and 19 at Rock Springs and Landisville, respectively, and treatments were arranged in a randomized
block design, but each treatment plot was split into uninoculated and inoculated subplots in order to
account for yield fluctuations that may occur by the use of the products in the absence of disease.
Treatments
In 2011, the treatments included acibenzolar-S-methyl (ASM; Actigard 50WG, Syngenta Crop
Protection, Greensboro, NC, USA), harpin protein (harpin; Employ, Plant Health Care, Pittsburgh, PA, USA),
Glomus intraradices (GI; Myke Pro WG, Premier Tech Biotechnologies, Rivière-du-Loup, Québec, Canada),
and Bacillus subtilis GB03 (GB03; Companion, Growth Products, Ltd., White Plains, NY, USA). GB03 and GI
were applied as a drench at the base of the plant two days after planting at concentrations of 8.19 mL / L
and 36 spores / 150 mL (159 g / L), respectively, using a backpack sprayer with a concentrated spigot tip.
After sprayer calibration, the application was estimated at 150 mL / plant. At Rock Springs, foliar
treatments were applied using a tractor mounted, CO2-powered side boom sprayer calibrated to deliver
206 L / ha at 24 psi through three TX-18 nozzles, while at Landisville treatments were applied using a Solo
backpack sprayer. Weekly foliar applications of copper hydroxide (Kocide 3000; Dupont, Wilmington, DE,
USA; 1.68 kg / ha) tank-mixed with mancozeb (Penncozeb 75DF; Cerexagri-Nisso LLC, King-of-Prussia, PA,
USA; 1.68 kg / ha) was applied as a grower standard. Harpin was applied at a rate of 146 mL / ha per
treatment, while ASM was applied at 5.5 mL / ha per treatment following the application schedules in
Tables 5.1 and 5.2 for 2011 and 2012, respectively.
In 2012, ASM, harpin, GB03, hydrogen peroxide, and the current grower standard copper
hydroxide + EBDC were applied singly or in combination. ASM and harpin were applied at concentrations
92
consistent with the label and listed above. Hydrogen peroxide (Oxidate, Biosafe Systems Inc., Glastonbury,
CT, USA) was applied weekly at two concentrations, preventative (9.3 L / ha for three weekly applications,
then 2.9 L / ha weekly after) and curative (9.3 L / ha weekly). Treatment dates and rates are outlined in
Table 5.1.
Leaf length
On June 10, 2011, the number of leaves as well as the longest leaf of each of five plants per plot
were measured (cm) as a measure of vigorous growth, which is one of the additional product claims.
Inoculation
In 2011, onion plants were inoculated on June 21 with 108 colony-forming units (CFU) /mL
suspension of six isolates of bacterial species P. ananatis (isolates PA1a, PA1b, and 09-082) and P.
agglomerans (isolates 10-009, 10-022, and 09-063) in a 0.1% Silwet solution. All isolates were initially
isolated from symptomatic onions from PA farms in 2008, 2009, or 2010 and are part of the Gugino Lab
bacterial isolate collection. Plants were sprayed to run-off less than 1 hr before sunset with a backpack
sprayer on June 21. Four leaves from each treatment were sampled the next morning, placed individually
in plastic bags, each shaken in 100 mL buffer, and plated on PA20, a semi-selective medium for Pantoea
spp. (Goszczynska et al., 2006). Epiphytic bacterial populations remained at least 106 CFU / mL (data not
shown).
In 2012, onion plants were inoculated June 13 with a 108 CFU / mL mix of two strains of P. ananatis
(isolates PA1a and PA1b) and two strains of P. agglomerans (isolates 2010-009 and 2009-063). A toothpick
was dipped into the bacterial suspension, and the fourth and fifth or fifth and sixth leaves of the interior
two rows of plants in the bed were punctured with the toothpick less than 15.2 cm from the onion neck
(modified from Gent and Schwartz, 2005; Carr et al., 2012; Fig. 5.1). Based on a laboratory simulation (not
shown), this inoculation protocol results in approx. 10 µL of suspension delivered through each punctured
leaf, for an inoculation of approx. 106 CFU applied per wound. Two leaves were inoculated per plant, and
the two exterior rows of plants in the four-row raised bed were left uninoculated, as were full-size 3.7 m
plots; this represented three levels of inoculum pressure: low (uninoculated), medium (adjacent-to-
inoculated), and high (inoculated).
93
Foliar disease ratings
In 2012, disease ratings were recorded weekly post-inoculation on a 0 to 7 scale (Fig. 5.2). Since
two leaves per plant were inoculated, the leaf expressing higher severity was rated. The scale is as follows:
0 – no lesion, asymptomatic (uninoculated plants only); 1 – local lesion (2.5 cm x 2.5 cm or less); 2 –
expanded lesion, but less than ¼ of leaf; 3 – up to ½ of the inoculated leaf is chlorotic or bleached; 4 –
more than ½ of the inoculated leaf is chlorotic or bleached, but uninoculated leaves do not show
symptoms; 5 – entire inoculated leaf and a portion of an uninoculated leaf are symptomatic; 6 – multiple
fully symptomatic leaves; 7 – ≥50% bleached and/or collapsed leaves (Fig. 5.2). A total of four foliar disease
ratings were recorded at Rock Springs in 2012 on June 26 (one week postinoculation), July 2, July 11, and
July 17. A single foliar disease severity rating was recorded at Landisville using the same scale on July 9.
Notes on the topography of the field in Landisville were also recorded (Fig. 5.3).
Harvest evaluation
In 2011, plots were harvested individually, the number and total weight of bulbs with bacterial
disease symptoms were recorded, and marketable bulbs were graded by size. Briefly, the size categories
are considered the following: small, < 6.4 cm in diameter; medium, 6.4 – 7.6 cm; jumbo, 7.6 – 10.2 cm;
colossal, > 10.2 cm. One-gram samples of tissue were excised from 20% of the symptomatic bulbs for
bacterial pathogen isolation as well as lyophilized and extracted for multiplex PCR detection of all bacterial
species present; both samples were placed in 1.8 mL tubes. In 2012, plots were harvested individually and
separated by inoculation pressure: low (uninoculated), medium (outer rows directly adjacent-to-
inoculated), and high (inoculated). Symptomatic bulbs were culled for bacterial pathogen identification
using the two-sample microbiological and molecular protocols described above. Marketable bulbs were
graded by size, counted, and weighed.
Data analysis
Data were analyzed using one-way ANOVA and Levene’s test in Minitab 16.2 (Minitab, State
College, PA, USA), in addition to post-hoc mean comparisons using Fisher’s LSD (α = 0.05).
Results
Marketable yield and plant growth – 2011
At Landisville in 2011, distributions of bulb sizes were subject to a blocking effect, such that plots
in blocks 1 and 4 had a significantly greater proportion of large-size (combined jumbo and colossal; >7.6
94
cm diameter) bulbs than block 3 (Fig. 5.4; blue bars). At Rock Springs, a blocking effect was again present;
block 1 had significantly fewer large-size bulbs than the other three blocks (data not shown). When block
1 was removed from the Rock Springs analysis, no differences were indicated between any of the
treatment groups, so groups were compared based on their combinations of products. Earlier that season,
growth estimates indicated the harpin, harpin and GI, GB03, and GB03/ASM treatments averaged longer
leaves than untreated controls (Table 5.1).
Disease management - 2011
In 2011, harvest disease incidence in onion bulbs from Landisville averaged 12.9% loss across all
plots. However, these data were subject to a blocking effect such that the lowest-lying area of the field
(block 3; Fig. 5.3) was significantly different from blocks 2 and 1 (Fig. 5.4). In addition, the slightly lower
area of the field (block 4) was significantly different from the highest area, block 1 (Fig. 5.4, red bars; Fig.
5.3). This prevented effective use of ANOVA to compare treatment means since disease incidence was not
uniform across blocks.
At Rock Springs in 2011, disease incidence was overall much lower, at an experiment-wise average
of 2.4% loss across all plots. These data very nearly violated the assumption of equal variances by Levene’s
test (P = 0.06) between treatments; it appeared that very inconsistent disease incidence in the untreated
inoculated control contributed to this, with one block averaging 18% disease incidence and the remaining
three blocks with 0% disease incidence, as indicated by variance within the treatment (Table 5.1). As a
result of the unequal variances, data were not analyzed further.
Marketable yield - 2012
In 2012, a different inoculation technique was used in which plants were subjected to three levels
of inoculum pressure, low (uninoculated), medium (adjacent-to-inoculated), and high (toothpick-
inoculated). In Fig. 5.5, the proportion of marketable bulbs that were large (> 7.6 cm diameter) in the low
and medium inoculum pressure plots was greater than the proportion of large bulbs in the high inoculum
pressure plots, which was apparent at both research farms (P < 0.001; Fig. 5.5).
Bacterial disease management – 2012
In both Landisville and Rock Springs in 2012, only a few statistically significant differences existed
in disease incidence between product treatments within each inoculation group. At Landisville in 2012, in
the medium inoculum pressure plots, the hydrogen peroxide preventative rate treatment had higher
95
bacterial rot incidence than the uninoculated plots, however, no other statistically significant differences
occurred (Fig. 5.6). However, numerically, treatments with the lowest disease incidence included the
curative rate of hydrogen peroxide in low inoculum plants and GB03 in high inoculum pressure plants (Fig.
5.6). For plants under medium inoculum pressure (adjacent-to-inoculated), the lowest average disease
incidences were in the untreated control and GB03-treated plots (Fig. 5.6).
At Rock Springs in 2012, 20 inoculated plants in each inoculated plot were rated for foliar disease
for four weeks after inoculation using the scale described in Fig. 5.2. Three weeks post-inoculation, plots
treated with GB03 had an average severity rating of 4.55, which was significantly less than the grower
standard (copper + mancozeb) severity rating of 4.97 (P = 0.001), with other treatments ranging in severity
from 5.0 to 5.3 (Fig. 5.7). However, when plots were rated for severity immediately prior to harvest the
following week, this difference was no longer observed (Fig. 5.7). At Landisville, a single foliar rating was
recorded 17 days post-inoculation, and the harpin treatment averaged a severity rating of 4.1, which was
significantly higher than the preventative hydrogen peroxide, uninoculated, and ASM treatments with
averages of 3.3, 3.3, and 3.2 disease severity, respectively (P < 0.05; data not shown). In addition, these
foliar disease ratings were correlated with the bacterial disease incidence recorded at harvest in
Landisville (data not shown, P < 0.001, R2 = 0.66).
In terms of disease incidence at harvest, statistically significant differences between treatments
administered at Rock Springs only occurred at the low level of inoculum (Fig. 5.8). The grower standard
copper - EBDC mixture, in addition to the alternative H2O2 (preventative rate), ASM, and Harpin alternated
with copper – EBDC all were lower in disease incidence than the untreated plots (Fig. 5.8; P < 0.05). While
no statistically significant results were indicated at either the medium or high inoculum levels, numerically
the grower standard Cu-EBDC, hydrogen peroxide (curative rate), and GB03 were the treatments with the
lowest disease incidence (Fig. 5.8).
In summary, bacterial disease incidence was low (Rock Springs) or variable (Landisville) in the 2011
research farm trials. With the change of the inoculation protocol, bacterial disease incidence was higher
in 2012, yet few significant differences were present among the treatments included in these trials. Data
from the two locations could not be combined as a result of the variability between sites (Fig. 4.9),
reiterating the general need for product evaluation at multiple sites.
Discussion
Onions have been grown commercially as a trademarked crop in Pennsylvania for approximately
15 years, so the elucidation of a relatively consistent inoculation protocol was a significant result from this
96
work. The 2011 inoculation protocol, in which plants were sprayed to run-off with 108 CFU / mL Pantoea
spp. in 0.1% Silwet, had been shown to induce bacterial disease in growth chamber studies (not shown),
but did not induce consistent disease under field conditions (Table 5.1; Fig. 5.4), even though bacterial
populations were maintained on the plants overnight. The 2012 inoculation protocol, in which plants were
inoculated by puncturing two full-size, inner leaves with a toothpick dipped in a bacterial suspension
containing 108 CFU / mL, was more effective overall in inducing consistent center rot symptoms (Figs. 5.6,
5.8, 5.9). In fact, the 2012 protocol allowed observation of inoculum pressure at three levels, low
(uninoculated), medium (adjacent-to-inoculated) and high (inoculated); which was statistically significant
when inoculum levels at both field locations were grouped (Fig. 5.9; data not shown). When Rock Springs
and Landisville were combined, the incidence of center rot averaged 12.9%, 17.9%, and 27.5% for low,
medium, and high inoculum pressures, respectively.
Few recommendations can be deduced from the 2011 cropping season alone, however, one trend
that became apparent was the association of high disease incidence and low proportions of large (> 7.6
cm diameter) bulbs. This may suggest that even if bulbs are not symptomatic at harvest, high inoculum
pressure may still result in smaller asymptomatic onions (Fig. 5.4), perhaps as a result of latent foliar
infection and reduced movement of photosynthate to onion bulbs. Work with 14C-urea injections into
onion leaves has indicated that each leaf blade primarily nourishes its corresponding bulb scale; when the
leaf blade is removed, portions of photosynthate from both older and younger blades are then shared
with the ‘orphaned’ scale (Mann, 1983). In addition, through leaf removal experiments simulating storm
damage, it was suggested that reduced yield of onions with removed leaf blades may also be attributed
to a later onset of bulbing, due to the reduction in total numbers of light receptors (Bartolo et al., 1995).
Results from both locations in 2012 lend strength to this hypothesis, which indicated more large-size bulbs
in low and medium inoculum plots compared to the high inoculum plots when considering only the
marketable bulbs (Fig. 5.6; P < 0.001). However, other environmental variables, such as field topography,
may also play roles in disease pressure and marketable yields, with Landisville in 2011 as an example; in
this dataset, the lowest areas of the field recorded the highest disease incidence, as indicated in Fig. 5.4.
Some trends may be suggested by combining disease and growth estimates from 2011 and 2012,
which may be pursued in future work. GB03-treated plots were consistently among those with the lowest
disease incidence at harvest (Fig. 5.6 and 5.8), even though these results were not statistically significant.
GB03 - treated plots also had slower foliar disease development (Fig. 5.7) and longer leaves in 2012 (Table
5.1). Similar results have been shown in other bacteria-vegetable pathosystems, such as Pseudomonas
syringae pv. lachrymans on watermelon, muskmelon, and cucumber (Raupach and Kloepper, 1998;
97
Kokalis-Burelle et al., 2003). Bacillus subtilis GB03 is recommended as an at-planting soil drench, and may
be an economically viable disease management treatment as a result of the efficacy of its single
application. In this study, treatment occurred shortly after transplanting, which would reduce overall
management costs compared to weekly foliar applications of copper – EBDC fungicide by tractor or
backpack. Disease incidence in GB03-treated plots was consistently no different from the copper - EBDC
standard (Figs. 5.6 and 5.8), and the product has short re-entry (4 hr) and 0 day pre-harvest intervals.
Conversely, the collective data presented here suggest harpin may not be effective for bacterial
disease management in onion. In 2011, two treatments that included harpin had significantly longer
leaves than control treatments (Table 5.1), which supports a potential growth promotion effect. However,
harpin-treated plots (without copper – EBDC) had the highest disease incidence at both locations in 2012
(Figs. 5.6 and 5.8). Other research has indicated poor bacterial disease management using a harpin
protein-based product (Obradovic et al., 2005), so products with this active ingredient, applied on the
same schedule, may not be the most effective management tools.
The plots receiving plant defense-inducing and growth-promoting products had similar
marketable yields and disease incidences compared to the grower standard in two replicated research
trials in 2012, however, most treatments were also similar to the untreated controls in these studies.
Therefore, it is recommended that selected products be tested for a third year using the 2012 inoculation
protocol and the repetitions of each experimental block are increased which may increase the power of
these experiments and ideally allow more customized recommendations based on perceived inoculum
pressure in grower fields. In summary, plant defense-inducing products, particularly GB03, have some
potential to serve as alternatives to the grower standard copper-EBDC treatments for the management
of bacterial rots of onion.
Acknowledgements
John Stepanchak, Jim Bollinger, Alyssa Collins, and Michele Mansfield offered essential technical
assistance.
98
Tables and Figures
Table 5.1. Treatments, treatment schedule, and 2011 growth and disease incidence results from Rock Springs, 2011. Similar treatments and application schedules were conducted in the 2011 trial in Landisville as well the 2012 trials in Rock Springs and Landisville. Longest leaf means were separated using Fisher’s LSD (P ≤ 0.05); different letters following the means indicate statistically significant differences.
Treatment and rate Applications (days post-transplant)
2011 longest leaf (cm)
2011 Disease incidence (% ± std. deviation)
Untreated 53.1c 2.8 ± 3.7
Untreated, inoculated 53.1c 4.5 ± 9.2
Copper, 1.68 kg / ha Mancozeb, 1.68 kg / ha
55, 62, 69, 76, 83, 90 55, 62, 69, 76, 83, 90
54.9bc
0.6 ± 0.7
Harpin, 146 mL / ha Copper, 1.68 kg / ha Mancozeb, 1.68 kg / ha
55, 76 62, 69, 83, 90 62, 69, 83, 90
57.6abc
4.0 ± 2.8
ASM, 5.5 mL / ha 55, 62, 69, 76, 83, 90 55.9abc 1.4 ± 2.0
GI, 159 g / L 0 (at planting) 58.2ab 3.3 ± 3.4
Harpin, 146 mL / ha 55, 76 60.1ab 1.4 ± 2.0
GB03, 8.19 mL / L 0 (at planting) 60.7a 3.0 ± 3.5
Harpin, 146 mL / ha GI, 159 g / L
55, 76 0 (at planting)
59.2ab
4.1 ± 4.2
Harpin, 146 mL / ha GB03, 8.19 mL / L
55, 76 0 (at planting)
55.1bc
0.6 ± 0.7
ASM, 5.5 mL / ha GI, 159 g / L
55, 62, 69, 76, 83, 90 0 (at planting)
56.1bc
1.9 ± 1.9
ASM, 5.5 mL / ha GB03, 8.19 mL / L
55, 62, 69, 76, 83, 90 0 (at planting)
59.1ab
2.0 ± 1.7
ASM, 5.5 mL / ha Copper, 1.68 kg / ha Mancozeb, 1.68 kg / ha
55, 69, 83 62, 76, 90 62, 76, 90
53.7c
3.5 ± 5.2
Harpin, 146 mL / ha ASM, 5.5 mL / ha
55, 76 62, 69, 83, 90
52.9c
0.7 ± 0.8
99
Rep 1
Rep 2
Rep 3
Rep 4
Fig. 5.1. 2012 inoculation diagram with locations of high (red, center front), medium (orange, flanking red), and low (yellow, center rear) inoculum. pressure sections of each plot.
Fig. 5.3. Photograph of Landisville field showing topography of blocks.
100
Fig. 5.2. Foliar disease symptom rating scale. Foliar ratings are as follows: 0 – no lesion, asymptomatic (uninoculated
plants only); 1 – local lesion (2.5 cm x 2.5 cm or less); 2 – expanded lesion, but less than ¼ of leaf; 3 – up to ½ of the
inoculated leaf is chlorotic or bleached; 4 – more than ½ of the inoculated leaf is chlorotic or bleached, but
uninoculated leaves do not show symptoms; 5 – entire inoculated leaf and a portion of an uninoculated leaf are
Fig. 5.4. Average center rot incidence and percentage of marketable large-size (>7.6 cm diameter) onions by treatment block, Landisville, 2011. Block was analyzed as a random factor in order to control for natural variation within the onion field; the arrow roughly indicates the topography of the field (low-lying, on left, to high-ground, on right). Data were analyzed using PROC GLM in SAS 9.2, with post-hoc comparisons completed using Fisher’s LSD (α = 0.05); statistically significant differences are indicated by different letters above each set of bars (disease incidence [red bars] = a-c; large bulbs [blue bars] = x-z). Bars represent the experiment-wide standard error.
Fig. 5.5. Proportion of total marketable bulbs that were categorized as large (> 7.6 cm diameter) bulbs across all treatments grouped by pathogen pressure based on inoculation status (low, medium, and high) from Rock Springs and Landisville, 2012. Analysis was completed using PROC GLM in Minitab 16 with post-hoc comparisons using Fishers LSD (α = 0.05). Letters above each bar indicate statistical significance.
a
ab
z yz
xy x
c
bc
Lowest Highest
a a
b b
a a
Inoculum pressure
102
Fig. 5.6. Center rot incidence by treatment under varying levels of inoculum pressure, Landisville, 2012. Center rot incidence at harvest was recorded for different areas of plots for each treatment. Inoculum pressure within plots includes low (uninoculated), medium (adjacent to inoculated), and high (inoculated) treatments. Error bars represent experiment-wide standard error. Different letters above each bar of the same inoculum pressure indicate statistically significant differences by Fisher’s LSD (α = 0.05).
Fig. 5.7. Weekly foliar disease severity ratings post-inoculation, from Rock Springs in 2012. For each treatment, 20 inoculated plants per plot were rated for disease severity following the scale in Fig. 5.2. * indicates a statistically significant difference between GB03 and the grower standard Cu-EBDC treatment (Fisher’s LSD; α = 0.05).
*
103
Fig. 5.8. Center rot incidence by treatment under varying levels of inoculum pressure, Rock Springs, 2012. Bacterial disease incidence at harvest was recorded for different areas of plots for each treatment. Inoculum pressure within plots includes low (uninoculated), medium (adjacent to inoculated), and high (inoculated) treatments. Error bars represent experiment-wide standard error. Different letters above each bar of the same inoculum pressure indicate statistically significant differences by Fisher’s LSD (α = 0.05).
Fig. 5.9. Center rot incidence on research farms in 2012. Bars indicate the mean of each inoculation level on each farm, regardless of in-season treatment, and error bars indicate standard error of each mean. Statistically significant differences are indicated by different letters above the bars (Fisher’s LSD; α = 0.05).
a
b b
c
bc
d
104
References Ahmad, M., Majerczak, D. R., Pike, S., Hoyos, M. E., Novacky, A., Coplin, D. L. 2001. Biological activity of Harpin produced by Pantoea stewartii subsp. stewartii. Microbe-Plant Molecular Interactions 14: 1223-1234. Bartolo, M. E., Schwartz, H. F., Schweissing, F. C. 1994. Yield and growth response of onion to simulated storm damage. HortScience 29:1465-1467. Bonaterra, A., Marib, M., Casalinib, L., Montesinosa, E. 2003. Biological control of Monilinia laxa and Rhizopus stolonifer in postharvest of stone fruit by Pantoea agglomerans EPS125 and putative mechanisms of antagonism. International Journal of Food Microbiology 84: 93-104. Borowicz, V. 2001. Do arbuscular mycorrhizal fungi alter plant-pathogen relations? Ecology 82: 3057-3068. Brazelton, J. N., Pfeufer, E. E., Sweat, T. A., McSpadden Gardener, B. B., Coenen, C. 2008. 2,4-diacetylphloroglucinol alters plant root development. Microbe-Plant Molecular Interactions 21: 1349-1358. Carr, E. A., Zaid, A. M., Bonasera, J. M., Lorbeer, J. W., Beer, S. V. 2013. Infection of onion leaves by Pantoea ananatis leads to bulb infection. Plant Disease 97:1524-1528. Conover, R. A., Gerhold, N. R. 1981. Mixtures of copper and maneb or mancozeb for control of bacterial spot of tomato and their compatibility for control of fungus diseases. Proceedings of the Florida State Horticultural Society 94: 154-156. Dodd, I. C., Zinovkina, N. Y., Safronova, V. I., Belimov, A. A. 2010. Rhizobacterial mediation of plant hormone status. Annals of Applied Biology 157: 361-379. Dong, H., Delaney, T. P., Bauer, D. W., Beer, S. V. 1999. Harpin induces disease resistance in Arabidopsis through the systemic acquired resistance pathway mediated by salicylic acid and the NIM1 gene. The Plant Journal 20: 207 – 215. Environmental Protection Agency (EPA). Harpin Protein (006477) Fact Sheet. Updated December 8, 2010. Accessed online, January 31, 2011: http://www.epa.gov/pesticides/biopesticides/ingredients/factsheets/factsheet_006477.htm . EXTOXNET Pesticide Information Notebook. EBDCs (General Information) Mancozeb Research Report. Accessed online 18 March 2014: http://pmep.cce.cornell.edu/profiles/fung-nemat/aceticacid-etridiazole/edbc/index.html . Garcia-Garrido, J. M., Ocampo, J. A. 1988. Interaction between Glomus mosseae and Erwinia carotovora and its effects on the growth of tomato plants. New Phytologist 110: 551-555. Gent, D. H., Schwartz, H. F. 2005. Management of Xanthomonas leaf blight of onion with a plant activator, biological control agents, and copper bactericides. Plant Disease 89: 631 - 639.
Goszczynska, T., Venter, S. N., Coutinho, T. A. 2006. PA20, a semi-selective medium for isolation and enumeration of Pantoea ananatis. Journal of Microbiological Methods 64: 225 – 231. Herrera Medina, M. J., Gagnon, H., Piche, Y., Ocampo, J. A., Garcia Garrido, J. M., Vierheilig, H. 2003. Root colonization by arbuscular mycorrhizal fungi is affected by the salicylic acid content of the plant. Plant Science 164: 993-998. Kokalis-Burelle, N., Vavrina, C.S., Reddy, M.S., Kloepper, J.W. 2003. Amendment of muskmelon and watermelon transplant media with plant growth-promoting rhizobacteria: effects on seedling quality, disease, and nematode resistance. HortTechnology 13:476-482. Kuiper, I., Lagendijk, E. L., Bloemberg, G. V., Lugtenberg, B. J. J. 2004. Rhizoremediation: A Beneficial Plant-Microbe Interaction. Microbe-Plant Molecular Interactions 17:6-15. Lang, J. M., Gent, D. H., and Schwartz, H. F. 2007. Management of Xanthomonas leaf blight of onion with bacteriophages and a plant activator. Plant Disease 91:871-878. Lin, Y. H., Xu, J. L., Hu, J., Wang, L. H., Ong, S. L., Leadbetter, J. R., Zhang, L. H. 2003. Acyl-homoserine lactone acylase from Ralstonia strain XJ12B represents a novel and potent class of quorum-quenching enzymes. Molecular Microbiology 47: 849-860. Liu, J., Maldonado-Mendoza, I., Lopez-Meyer, M., Cheung, F., Town, C. D., Harrison, M. J. 2007. Arbuscular mycorrhizal symbiosis is accompanied by local and systemic alterations in gene expression and an increase in disease resistance in the shoots. The Plant Journal 50: 529-544. Louws, F. J., Wilson, M., Campbell, H. L., Cuppels, D. A., Jones, J. B., Shoemaker, P. B., Sahin, F., Miller, S. A. 2001. Field control of bacterial spot and bacterial speck of tomato using a plant activator. Plant Disease 85:481-488. Lugtenberg, B., Kamilova, F. 2009. Plant growth-promoting rhizobacteria. Annual Reviews of Microbiology 63: 541-556. Mann, J. D. 1983. Translocation of photosynthate in bulbing onions. Australian Journal of Plant Physiology 10: 515-521. Nair, C. B., Anith, K. N., Sreekumar, J. 2007. Mitigation of growth retardation effect of plant defense activator, acibenzolar-S-methyl, in amaranthus plants by plant growth-promoting rhizobacteria. World Journal of Microbiological Technology 23: 1183-1187. Nasser, W., Reverchon, S., Vedel, R., Boccara, M. 2005. PecS and PecT coregulate the synthesis of HrpN and pectate lyases, two virulence determinants in Erwinia chrysanthemi 3937. Molecular Plant-Microbe Interactions 18:1205-1214. Obradovic, A., Jones, J. B., Momol, M. T., Olson, S. M., Jackson, L. E., Balogh, B., Guven, K., Iriarte, F. B. 2005. Integration of biological control agents and systemic acquired resistance inducers against bacterial spot on tomato. Plant Disease 89:712-716.
106
Pozo, M. J., Azcon-Aguilar, C. 2007. Unraveling mycorrhiza-induced resistance. Current Opinion in Plant Biology 10: 393-398. Romero, A. M., Kousik, C. S., Ritchie, D. F. 2001. Resistance to bacterial spot in bell pepper induced by acibenzolar-S-methyl. Plant Disease 85:189-194. Siddiqui, Z. 2006. PGPR: Prospective biocontrol agents of plant pathogens. PGPR: Biocontrol and Biofertilization. Springer Netherlands: Pp. 111-142. Simard, S. W., Durall, D. M. 2004. Mycorrhizal networks: a review of their extent, function, and importance. Canadian Journal of Botany 82: 1140-1165. Raupach, G. S., Kloepper, J. W. 1998. Mixtures of plant growth-promoting rhizobacteria enhance biological control of multiple cucumber pathogens. Phytopathology 88: 1158-1164. Thaler, J. S., Fidantsef, A. L., Bostock, R. M. 2002. Antagonism between jasmonate- and salicylate-mediated induced plant resistance: Effects of concentration and timing of elicitation on defense-related proteins, herbivore, and pathogen performance in tomato. Journal of Chemical Ecology 28:1131 – 1159. van Wees, S. C. M., de Swart, E. A. M., van Pelt, J. A., van Loon, L. C., Pieterse, C. M. J. 2000. Enhancement of induced disease resistance by simultaneous activation of salicylate- and jasmonate-dependent defense pathways in Arabidopsis thaliana. Proceedings of the National Academy of Science 97: 8711-8716. Walters, D. R., Fountaine, J. M. 2009. Practical application of induced resistance to plant diseases: an appraisal of effectiveness under field conditions. Journal of Agricultural Science 147: 523 – 535. Walters, D., Walsh, D., Newton, A., Lyon, G. 2005. Induced resistance for plant disease control: Maximizing the efficacy of resistance elicitors. Phytopathology 95:1368-1373. Wei, Z. M., Laby, R. J., Zumoff, C. H., Bauer, D. W., He, S. Y., Collmer, A., Beer, S. V. 1992. Harpin, elicitor of the hypersensitive response produced by the plant pathogen Erwinia amylovora. Science 257: 85-88.
107
Chapter 6: Effects of nitrogen fertilizer on growth characteristics, center rot incidence and severity, and other nutrient status in ‘Candy’ onion
Abstract
Center rot of onion, caused by Pantoea ananatis and P. agglomerans, is an emerging disease in
onion production regions across the United States, but economic losses can be especially severe in
Pennsylvania, where large amounts of inputs are integrated into these highly managed systems. As a
result of poor efficacy of chemical controls, commercial growers are interested in the impact of fertility
programs on bacterial disease management. From 2011 and 2012, intensive surveys suggested a weak
relationship between early-season levels of soil NH4 and the incidence of bacterial disease at harvest, and
in 2013, preliminary growth chamber studies suggested delayed center rot progression in plants fertilized
primarily with (NH4)2SO4 compared to Ca(NO3)2 or a 1:1 mixture of the two fertilizers. A small-plot
randomized block trial was conducted during the 2013 growing season in which ‘Candy’ onion plants were
fertilized at the same concentrations with either Ca(NO3)2 or (NH4)2SO4 on two different application
schedules, half-season (recommended N fertility applied weekly prior to bulbing) or full-season
(recommended N fertility applied weekly through the growing season). Consistent differences were
indicated between inoculated and uninoculated plants, with a possible interaction with the timing of N
application, and some pairwise differences in inoculated plants. Additionally, the inclusion of early-season
foliar Ca and mid-season foliar N were significant factors in ANCOVAs with center rot incidence as the
dependent variable when treatment was included as a factor; in particular, early-season foliar Ca was
negatively related to center rot incidence. Fertilization with (NH4)2SO4 resulted in onion bulbs with
significantly higher levels of sulfur, a major component in onion pungency and a restrictive factor in
marketing onion bulbs in PA. Taken together, these results suggest the timing of N fertility applications in
onion fields may have implications for bacterial disease management, foliar Ca may be influential in
slowing the spread of bacterial pathogens from infected foliage to onion bulbs, and using (NH4)2SO4 as the
sole N fertility source should be avoided by growers involved in the PA Simply Sweet® marketing program.
Introduction
Bacterial rots of onion (Allium cepa L.), including center rot, are the most significant diseases
affecting commercial production of the crop in Pennsylvania. Center rot is caused by Pantoea ananatis
Serrano and P. agglomerans Beijerinck, which are considered recently emerging bacterial pathogens
(Gitaitis and Gay, 1997; Edens et al., 2006). Growers attempt to manage center rot using copper-based
108
fungicides tank-mixed with ethylene-bis-dithiocarbamate, hydrogen dioxide, and cultural methods like
alternative mulches, some of which have efficacy against bacterial diseases (Sanchez et al., 2014; Gugino
et al., 2011). The majority of onions in PA are grown in small, frequently rotated, highly managed fields
on raised beds with plastic mulch and drip irrigation. Some growers contract with commercial fertilizer
companies or consultants for their fertigation programs, spending as much as $1600 per hectare on crop
fertility (J. Stoltzfus, pers. comm.). Data on the efficacy of these fertility programs in ensuring high yields,
however, is lacking, and high N fertility has been suggested to increase bacterial disease in onion (Diaz-
Perez et al., 2003; Mohan, 2008; Gitaitis et al., 2008).
Nitrogen (N) is the most frequently studied elemental nutrient in horticultural production and
plant disease management (Huber and Thompson, 2007). A number of studies have been conducted
investigating relationships between N and plant disease; in particular, different plant-available forms of
N, specifically NO3 and NH4, have been implicated in complex relationships with disease management and
plant uptake of other macro- and micronutrients. In work with eggplant, strawberry, rice, and wheat,
fertilizing with NH4-N was demonstrated to increase uptake of manganese, which in turn is implicated in
pathogen inhibition and plant physical defenses (Elmer, 2000; Elmer and LaMondia, 1999; Huber and
McCay-Buis, 1995). However, relationships between macro- and micronutrients, the environment,
pathogens, and their host plants are generally pathosystem-specific, as management of other diseases
appears to favor NO3 nutrition (Huber and Thompson, 2007). Research completed in the onion-white rot
(Sclerotium rolfsii) pathosystem indicated decreases in disease following NH4 nutrition (Huber and
Graham, 1999), in addition to similar relationships shown in many soilborne pathosystems (Huber and
Thompson, 2007), however, the pathogens primarily investigated here are foliar (Pfeufer and Gugino,
unpublished; Carr et al., 2013).
Results from an observational study of 54 farms in PA suggested that foliar N at midseason had a
strong negative relationship to the total incidence of bacterial rots of onion at harvest and from storage
(Chapter 3), and a weak negative relationship was suggested between soil NH4 shortly after transplanting
and the incidence of bacterial rot at harvest (Pfeufer and Gugino, unpublished). In analyses on the same
dataset, early-season soil NH4 was positively associated with onion foliar N at midseason, as was the silt
content of soil (Chapter 3; Table 5.1). In contrast, horticultural studies on NO3 and NH4 nutrition have been
completed for onion in a modified hydroponic system, which indicated reduced plant canopy, early onset
of bulbing, reduced water usage, and low bulb weight when onions were fertilized solely with NH4
(Gamiely et al., 1991). While different N sources were included in another horticultural study concerning
critical N concentrations in Canada, the combination of N source data from this research indicates no
109
significant differences between the type of N used and the overall yield in those plots (Westerveld et al.,
2003a; 2003b).
Investigations into the timing of N fertility in onion concluded reduced yield and delayed bulb
ripening when onions were underfertilized early in the season, but sufficient early-season N combined
with N reduction later in the season had no effect on horticultural yields (Brewster and Butler, 1989).
While no statistical significance was indicated, Westerveld et al. (2003a) reported highest onion yields on
mineral soils in Ontario when the N fertility concentration was either equally split between pre- and in-
season applications (2000) or when no additional N fertility was added besides the pre-plant application
(2001). These yield studies complement the result that higher incidence of storage losses due to several
bacterial onion pathogens were recorded when plants were fertilized late in the season (Wright, 1993).
Literature suggests N deficiency is only indicated when plants respond to increased levels of N fertilizer
(Lorenz and Tyler, 1976; cited by Rumpel et al., 2004), so the lack of response at the later tissue sampling
dates suggests N fertilizer did not need to be applied on these dates. More locally, one experienced PA
grower has anecdotally shared that he never fertilizes through drip irrigation after the first week of June,
and simply waters his onions until harvest in early- to mid-July (pers. comm., to Pfeufer). The effects of
the timing of N fertility on the incidence of center rot of onion, therefore, have not been explicitly
addressed and are relevant to PA growers.
In addition, N levels in plant foliage have been suggested to be influential in insect pest preference
for plants, and this relationship has previously been demonstrated with onion thrips (Thrips tabaci
Lindeman; Malik et al., 2009), which are considered the major insect pest of onions in PA. Onion thrips
have been shown to vector Pantoea spp. to previously uninfected onion plants (Dutta et al., 2014), so the
role of N in thrips pressure was also briefly investigated. In one study on the impact of the role of N
fertility in onion thrips damage completed in Ontario, the authors observed no differences in thrips counts
between differentially-fertilized onions (Westerveld et al., 2003b). However, in another study in Utah,
authors demonstrated that 1/3 the recommended season concentration of fertilizer N reduced onion
thrips by 23 – 31% (Buckland et al., 2013).
Intensive survey support
Results from a two-year observational study in which different management factors were
investigated for associations to bacterial rot disease incidence suggested strong relationships between
foliar N and C at midseason and high disease incidence (see Chapter 3). These data were further mined
with midseason foliar N as a dependent variable, using variables previously investigated or unchanging
110
over the observation time as independent variables. When farm was treated as a sample, these results
indicated the level of early-season soil NH4 was positively associated with foliar N content at midseason,
along with the silt content of each field’s soil (Table 6.1).
Table 6.1. Multiple linear regression between foliar carbon (C), early-season soil ammonium (NH4), and silt content of soil to the foliar nitrogen (N) content from leaves collected at midseason, from 54 Pennsylvania onion fields over two yearsa.
Predictora Coefficient SE of Coefficient T value P value
The present study was planned based on the results from preliminary observational studies and
local interest by growers coupled with concerns about over-fertilization and the expense of proprietary
fertility programs. As an emerging disease of onion, there is a general absence of data in the present
published literature on the effects of N fertility on center rot, caused by P. ananatis and P. agglomerans.
A randomized, replicated field trial was conducted at the Russell E. Larson Agricultural Research and
Education Center, Pennsylvania Furnace, PA in 2013 in order to determine the effects of N source and N
fertility timing on onion growth, onion thrips pressure, micronutrient content, and center rot incidence at
harvest.
Materials and Methods
Preliminary growth chamber assays
In two growth chamber assays, ‘Candy’ seeds were sown in a flat eight weeks prior to
transplanting, which was watered 2-3 times per week and placed under growth lamps with a 16-hr light
period. Plants had 2-4 leaves when transplanted into 7.62 cm diameter round clay pots with pasteurized
Metro-Mix 360 (Sun Gro Horticulture, Agawam, MA), then were allowed to grow for two more weeks
prior to fertilizer applications. Applications were composed of Ca(NO3)2, NH4(SO4)2, or a 1:1 mixture of
the two nitrogen sources. Final concentrations of N applied were the equivalent of 84 kg N / ha per
application, which occurred weekly. Two experiments were performed; one in which plants were
fertilized at and after inoculation, and another where plants were fertilized for two weeks prior to
inoculation, then weekly thereafter. Plants were bottom-watered to prevent nutrient leaching.
Inoculations were performed in the growth chamber where a toothpick was used to puncture
the longest leaf of the growing plant approx. midway up the leaf. This toothpick was dipped in 108 colony-
111
forming units (CFU) / mL mixture of P. ananatis (strain PA1a) and P. agglomerans (strain 10-009).
Uninoculated controls were punctured with a sterile, dry toothpick. Lesions were measured on two axes
each week, with the axes multiplied to calculate symptomatic area. The total number and length of the
longest leaf of each plant was also recorded. Replicate plants (twelve or fifteen, depending on
experiment) of each treatment were blocked within the growth chamber and treatments were
randomized within the block.
Field trial setup and maintenance
Based on a soil test, field nitrogen levels were brought to 112.1 kg N / ha using urea prior to field
preparation, while levels of other nutrients met or exceeded recommendations for onion (soil test on
file, from Penn State Agricultural Analytical Services Lab). Small plots were established on a Hagerstown
silt loam in four rows with standard black plastic mulch and two lines of drip irrigation. Each plot was
1.83 m long and approximately 1 m wide across the top of the raised bed. ‘Candy’ onion transplants
(Dixondale Farms, Carrizo Springs, TX) were planted at 15.24 cm spacing on April 25, 2013. In-season
fertilizer treatments were arranged in a randomized block configuration with unfertilized, half-season
NO3, full-season NO3, half-season NH4, and full-season NH4 treatments; half-season treatments were
twice as concentrated as full-season treatments. Uninoculated and inoculated plots were maintained for
each of the previous treatments. Fertility was applied as a sidedress approx. weekly for four or eight
weeks, depending on the treatment, to a total, season-long field concentration of 179.4 kg N / ha, except
unfertilized plots which received only 112.1 kg N / ha broadcast prior to bed preparation, which is similar
to some recommended fertility rates for onion for the total season (transplant supplier recommendation;
Rumpel et al., 2008; Diaz-Perez et al., 2003). Nitrogen sources were Ca(NO3)2 or (NH4)2SO4, which were
special ordered from Peters Fertilizer (Allentown, PA). A calibrated backpack sprayer was used to
administer approx. 100 mL of concentrated liquid fertilizer directly to the side of each onion plant. If
necessary, plastic mulch was loosened around the necks of each onion plant in order to apply fertilizer
treatments. Weeds were managed by tilling between raised beds and hand weeding within the bed;
there was no insect management in these plots.
Plants in the appropriate treatments were inoculated by puncturing the fifth non-senescing leaf
approx. 15 cm from the leaf whorl with a sterile toothpick dipped in a mixture of three P. ananatis (12-
766, 12-980, and 12-1764) and three P. agglomerans strains (12-823, 12-835, and 12-2058), all
originating from symptomatic onions sampled from PA farms in 2012, and similar to Carr et al. (2013).
Inoculum was prepared by streaking two large (15 x 100 mm) King’s B (King et al., 1954) plates with a
112
single strain, then flooding each plate with 2 mL phosphate buffer the following day, using a sterile plastic
scraper to dislodge the cells, and removing this liquid to six 50-mL tubes containing 6 mL sterile
phosphate buffer, one for each strain. One milliliter was removed from each of these tubes, combined
with 10 mL phosphate buffer, vortexed, and this mixture was aliquoted into sterile 1 mL Eppendorf tubes
for field inoculations. The final concentration of inoculum was 7.6 x 1013 CFU / mL, and estimates suggest
approx. 10 µL inoculum is delivered per puncture. Tubes were gently agitated in the field by inverting
several times prior to inoculations, which occurred on June 22, 2013. Plants were observed over several
days, and since symptoms did not progress quickly, the same leaf per plant was inoculated again on June
29, 2013 using 9.2 x 1013 CFU / mL mixture of P. ananatis strains PA1a and 10-082, and P. agglomerans
strains 09-009 and 10-063. Second inoculation strains were collected from symptomatic bulbs from PA
in 2008, 2009, and 2010 and are part of the Gugino lab bacterial isolate collection. Center rot symptoms
progressed quickly following the second inoculation.
Data collection
At three points in the season (May 13, May 31, and June 12; 18, 35, and 47 days after
transplanting [dap]), the longest leaf, bulb diameter, and number of thrips were recorded for ten plants
per replicate plot. Onion foliage, which was the total green tissue above the onion neck, from three
random plants in the plot was comingled in a paper bag on May 29 (33 dap) and June 22 (57 dap), then
dried for at least 72 hr in a 60⁰C forced-air drying oven. Bulbing estimates, as a ratio of the bulb diameter
at its widest point to the neck diameter from the narrowest point of the plant to (B:N ratio), were
recorded from the same three plants from which foliar tissue was sampled. In addition, at harvest, two
asymptomatic onion bulbs were split, comingled by plot, and dried for at least 72 hr in a 60⁰C drying
oven. Foliar and bulb tissue samples were submitted to Penn State Agricultural Analytical Services Lab,
University Park, PA for tissue grinding and complete analysis for N, P, K, and micronutrients via total acid
digest.
Plants were rated weekly for center rot symptoms after the second inoculation using the 0-7
point scale described in Chapter 4. Plots were harvested on July 16, 2013. Bulbs were initially rated for
bacterial disease by removing those with discolored scales in the neck or macerated area on their surface,
then marketable bulbs were graded by size into the categories small (< 6.35 cm diameter), medium (6.35
– 7.62 cm diameter), jumbo (7.62 – 10.2 cm diameter), and colossal (> 10.2 cm diameter).
113
Data analysis
Data were analyzed using the one-way ANOVA and Correlation procedures in Minitab 16.2
(Minitab Inc., State College, PA) with disease severity, harvest incidence of center rot, sulfur content, leaf
length, thrips density, and neck diameter as dependent variables. Using the General Linear Model
procedure, treatment, fertilizer timing, and fertilizer type were input as class variables, while macro- and
micronutrient data were included as covariates in some analyses. Post-hoc comparisons were completed
using Fisher’s LSD (ANOVA) or Tukey’s HSD (ANCOVA), α = 0.05.
Results
In preliminary growth chamber assays with onion seedlings, plants fertilized with (NH4)2SO4 had
smaller (P < 0.05) lesions after bacterial inoculation than inoculated plants fertilized with Ca(NO3)2 (12
days post-inoculation [dpi] and 19 dpi; Fig. 6.1) or a 1:1 mix of Ca(NO3)2:(NH4)2SO4 (19 dpi). Likely due to
high variation, the statistically significant differences were not present at a final disease severity
measurement at 26 dpi (Fig. 6.1), and no significant differences in either leaf length or in number of leaves
were indicated between any of the treatments (data not shown). In a replicate experiment where plants
were fertilized with the same concentrations of N fertilizer prior to inoculation, the average lesion size of
(NH4)2SO4- fertilized plants was smaller at 7 dpi compared to both the Ca(NO3)2 and Ca(NO3)2 - (NH4)2SO4
– fertilized plants. Seven days later, however, lesion sizes among the treatments did not differ (data not
shown).
In a replicated field trial in 2013, uninoculated and inoculated plants were rated weekly for
bacterial symptom development on two dates following the second inoculation. Significant differences
were indicated by inoculation status, that is, uninoculated plants developed symptoms more slowly than
inoculated treatments (data not shown). No differences were suggested based on N source and
application timing. As a result, treatments were grouped according to inoculation status and fertility
timing; at 9 dpi, inoculated, full-season N fertilized plants had a higher average foliar disease severity
rating than inoculated, half-season N fertilized plants (Fig. 6.2).
Midseason growth and thrips pressure estimates were recorded two (B:N ratio) or three (all other
estimates) times during the 2013 onion growing season. No differences in B:N ratio were indicated
between treatments on either of the data collection dates (Table 6.2). The number of thrips per leaf 35
dap, was numerically largest in the unfertilized treatments (A and B), which was statistically significant
when compared to the half-season fertilized plants, but not the full-season fertilized plants, which each
averaged fewer thrips per leaf at that timepoint (Table 6.2). This difference was not apparent in the data
114
from the other two data collection dates, which flank the date described. When neck diameter of ten
plants was recorded at the narrowest point on ten onion plants’ necks from each replicate plot 47 dap,
the full-season fertilized plants had significantly larger neck diameters compared to the unfertilized and
half-season fertilized plants (Table 6.2).
Center rot incidence at harvest, the percent of bulbs with center rot symptoms out of total bulbs
per plot, was numerically highest in the full-season Ca(NO3)2 fertilized plots, and was numerically lowest
in the uninoculated, unfertilized plots (Fig. 6.3). Disease incidence was generally higher in inoculated
plants than in uninoculated plants. Within only inoculated treatments, the full-season Ca(NO3)2 fertilized
plants had significantly higher center rot incidence than both the half-season Ca(NO3)2 fertilized plants
and the full-season (NH4)2SO4 fertilized plants, but there was no significant difference compared to the
half-season Ca(NO3)2 fertilized plants (Fig. 6.3). When treatments were combined by fertilizer application
timing and inoculation status (Fig. 6.4), the uninoculated, unfertilized plots as well as the uninoculated,
full-season fertilized plots had lower center rot incidence at harvest than the inoculated, full-season
fertilized plots, which was statistically significant (Fig. 6.4). Analysis indicated inoculation status
(uninoculated vs. inoculated) was a highly significant factor determining center rot incidence, and
suggested a nearly-significant interaction existed between inoculation status and timing of fertilizer
application (P = 0.15; data not shown).
A series of two foliar tissue tests 35 dap and 57 dap, in addition to a bulb nutrient test on
asymptomatic harvested onions, were conducted to determine potential roles of micro- and
macronutrient status on center rot incidence in onion at harvest. Asymptomatic, jumbo-size bulbs from
plots fertilized with (NH4)2SO4 had higher levels of sulfur, the primary component in onion pungency, as
compared to sulfur levels in bulbs from unfertilized and Ca(NO3)2 treated plots. This occurred regardless
of the timing of (NH4)2SO4 fertilization (Fig. 6.5).
Calcium levels in foliar tissue sampled 35 dap (after three weekly fertilizer treatments) showed a
weak negative relationship to the incidence of center rot at harvest, and when Ca was included as a
covariate in an ANCOVA, both treatment (P = 0.012) and 35 dap foliar Ca (P = 0.020) were significant
factors influencing center rot at harvest. Estimated means based on the inclusion of this late-May foliar
Ca as a covariate are shown in Fig. 6.6, where inoculated treatments trend toward higher disease
incidence. The model including treatment as a factor and 35 dap foliar Ca as a covariate with a negative
relationship with center rot incidence had an adjusted R2 = 0.403 pertaining to Fig. 6.6). In the 57 dap
tissue test, foliar N levels were weakly positively correlated to the incidence of center rot at harvest, and
when N was included as a covariate in ANCOVA, treatment (P = 0.02) and 57 dap foliar N (P = 0.09) were
115
significant or nearly-significant factors influencing center rot at harvest (data not shown). Estimated
means based on the inclusion of N as a covariate are shown in Fig. 6.7, where again, inoculated treatments
trend toward higher disease incidence. The model including treatment as a factor and 57 dap foliar N as
a covariate with a positive relationship with center rot incidence gave a model with an adjusted R2 = 0.353
(pertaining to Fig. 6.7). Combining treatment as a factor and both the 35 dap foliar Ca and 57 dap foliar N
as covariates indicated a model with an adjusted R2 = 0.417 (data not shown). Interestingly, a multiple
linear regression model not including treatment as a factor indicated the same, strong (P = 0.032) positive
relationship between center rot incidence and 57 dap foliar N, but the negative relationship with 35 dap
foliar Ca was diminished (P = 0.37; data not shown).
When harvest center rot incidence was averaged by inoculation status and N fertility timing, and
displayed by the amount of N fertilizer applied prior to midseason, there appear to be linear relationships
between N fertilizer applied prior to bulbing and disease incidence, which differ according to inoculum
pressure (Fig. 6.8). Analysis of N status in each of the tissue tests indicated higher levels of N in foliage in
plants fertilized at the 179.4 kg N / ha rate than the 112.1 kg N / ha concentration, but this difference was
only apparent at 35 dap tissue sample, where plants had at that point received the equivalents of 50.4
and 25.2 kg N / ha additional N over unfertilized and full-season fertilized plots, respectively (data not
shown).
Discussion
Grower interest in fertility programs as a component of integrated bacterial disease management
in onion, in addition to anecdotal evidence and preliminary data from observational studies, prompted
more thorough investigation of the roles of N fertility in onion growth, micronutrient status, disease
severity and incidence and thrips pressure in ‘Candy’ onion produced in PA. In two growth chamber
studies, it was indicated that plants fertilized only with (NH4)2SO4 [AmmINC] had smaller lesions than
plants fertilized with Ca(NO3)2 [NitINC] (at two timepoints, Fig. 6.1) or plants fertilized with an (NH4)2SO4
– Ca(NO3)2 mixture [ConINC] (at one timepoint) following inoculation with P. ananatis and P. agglomerans
(Fig. 6.1). This result, however, was temporary when compared later in the season, where there were no
differences in disease severity between differentially fertilized plants. Since bulb symptoms typically take
two to three weeks to develop following foliar inoculation (Pfeufer and Gugino, unpublished; Carr et al.,
2013), and in PA, an onion forms a bulb over a period of three to four weeks, even a temporary delay in
disease progression may produce higher marketable yields with a carefully timed harvest.
116
A small-plot, replicated field trial was set up to determine the effects of different N sources and
application timings on onion growth, micronutrient status, center rot incidence at harvest and thrips
pressure. No differences in disease progression were indicated by the type of N source used, however,
ten days after inoculation, plants fertilized on the half-season schedule (that is, no N applications after
mid-season) had lower foliar disease severity than plants that were fertilized on the full-season application
schedule. This difference was not present at the later severity rating, immediately prior to harvest (Fig.
6.2). Growth estimates and thrips counts were completed at several points prior to inoculation, and in
contrast to other published work (Gamiely et al., 1991), there were no significant differences indicated in
B:N ratio when treatments were grouped according to N source (Table 6.1). In that previous study, on
hydroponically-fed onions, plants fertilized with mostly NH4-N bulbed precociously compared to plants
fertilized with ratios higher in NO3-N, even though the authors concluded that precocious bulbing
ultimately had no impact on final bulb fresh weight (Gamiely et al., 1991).
At the May 31 rating, 35 dap, thrips densities were higher on unfertilized plants than on plants
fertilized on the half-season schedule (Table 6.1), which conflicts with other published results which
suggested that high N fertility promotes thrips pressure in onions (Buckland et al., 2013; Malik et al.,
2009), but is in agreement with another published report where onion thrips pressure was highest on
unfertilized plants (Westerveld et al., 2003a). However, no differences were observed 18 or 47 dap, so it
may be that high densities temporarily developed for a different reason, such as a lack of pressure on
thrips populations by natural predators. Neck diameter at 47-dap (June 12) was significantly larger in
plants fertilized on the full-season schedule compared to the unfertilized and half-season fertilized plants
(Table 6.1). This result was expected based on previously published work with onion (Wright, 1993) and
is significant because as the center rot pathogens infect onion leaves and subsequently bulbs (Carr et al.,
2013), logic follows that bulbs with smaller necks will dry down faster and halt this bacterial advance
(Wright, 1993), potentially resulting in reduced incidence of center rot in storage.
Inoculated plants as a group had significantly higher disease incidence than uninoculated plants,
however, the only statistically significant differences within inoculation type were between full-season
Ca(NO3)2 fertilized plants and each of half-season Ca(NO3)2 and full-season (NH4)2SO4 plants (Fig. 6.3).
Combining treatments by fertilizer application timing suggested a nearly-significant interaction between
inoculation status and timing of fertilizer application (P = 0.15), and when compared in a different format,
where total kg N / ha applied at roughly midseason is the only independent variable considered, well-fit
(R2 > 0.8) linear relationships with opposing slopes are indicated by this same dataset (Fig. 6.8). Taken
together, these suggest that the optimal timing of fertilizer application depends on the inoculum pressure
117
within a field, and it may be advantageous for growers to err on the side of half-season N fertility
applications if they are concerned about bacterial diseases. (Figs. 6.4 and 6.8). These results have not been
previously reported for management of center rot of onion, although sources recommend restricting N
fertilization late in the season (Dixondale Farms website), or note high levels of bacterial decay when
onions are fertilized at above-optimal levels (Diaz-Perez et al., 2002).
Foliar nutrients were analyzed in Pearson’s correlations with the incidence of harvested onions
with center rot symptoms, and correlated nutrients were included initially as covariates in ANCOVA
analyses. From the 35 dap foliar tissue test following three weeks of fertility applications, Ca was
suggested to be negatively correlated with the incidence of center rot at harvest (P = 0.02; data not
shown). This is intriguing because Ca plays primary roles in cell membrane integrity, plant defense
signaling, and the maintenance of plant tissue structure through the middle lamella (Rahman and Punja,
2007). In research completed in another bacteria-vegetable pathosystem, a negative relationship was
demonstrated between tuber Ca levels and severity of soft rot due to Erwinia carotovora pv. atroseptica
(now Pectobacterium carotovorum pv. atrosepticum) in potato (McGuire and Kelman, 1984). Including Ca
as a covariate in an ANCOVA with treatment as a factor indicated significant differences between adjusted
means of several inoculated and uninoculated treatments (Fig. 6.6). While NH4+ in soil solution has been
suggested to depress uptake of Ca ions (Rahman and Punja, 2007), no differences between fertilizer
treatments and Ca levels in foliage was indicated in the present study (data not shown). Since Ca has been
shown to accumulate in foliage as compared to other potato tissues (McGuire and Kelman, 1984; Rahman
and Punja, 2007) and Ca nutrition has been implicated in firmer onion bulbs and decreased pungency
(Coolong and Randle, 2009), one suggestion for the center rot – onion pathosystem is that Ca’s role in
strengthening foliar tissue structure may also slow the movement of the Pantoea spp. from inoculation
point to onion bulbs; however, additional replicated trials would be necessary to definitively demonstrate
this relationship and define critical Ca levels that confer its effect.
From the 57 dap foliar tissue test, following six weeks of fertility applications and immediately
prior to the first inoculation, N was suggested to be positively correlated to the incidence of center rot at
harvest (data not shown). This observation is not consistent with the survey results reported in Chapter
4, however, it is consistent with reports that over-fertilization promotes bacterial disease in onion (Diaz-
Perez et al., 2003; Mohan, 2008; Gitaitis et al., 2008). Including 57 dap foliar N as a covariate resulted in
only one significant difference between tested treatments (Fig. 6.7), and including the same variable in an
ANCOVA already containing treatment as a factor and 35 dap foliar Ca as a covariate only marginally
improved the model (R2 value increased 1.2%; data not shown). The interaction term between covariates
118
was not significant to the model (data not shown), which suggests that 35 dap foliar Ca and 57 dap foliar
N are either explaining the same variation in the model, or that each covariate’s explanatory power is
diminished by inclusion of the other variable in the model. When Ca and N were used as dependent
variables with treatments grouped by N source as the independent variable, the type of fertilizer was not
a significant factor in determining the content of these nutrients in onion foliar tissue (data not shown).
This suggests that while these foliar nutrients may influence disease development, the treatments applied
here were not the only factors dictating foliar nutrient content. Overall, relationships between
micronutrients, macronutrients, and the incidence of center rot of onion are complicated, and repeated
experiments are necessary to identify if these relationships are consistent before detailed interpretation
and conclusions are to be drawn.
Asymptomatic onion bulb tissue was dried and analyzed for total sulfur (S), soil levels of which
have been shown to influence S accumulation in sweet onions (Randle et al., 1999). In the present study,
plants fertilized with (NH4)2SO4 had significantly higher levels of S in their bulbs, regardless of fertilizer
application timing, than unfertilized plants or plants fertilized with Ca(NO3)2 (Fig. 6.5), which is consistent
with other publications (Randle et al., 1999). While total tissue S does not directly correlate with
enzymatically-developed pyruvic acid (the determining factor in onion pungency), and onion varieties vary
in their partitioning of S as a nutrient (Randle et al., 1999), the sensitivity of sweet onions to soil S levels
and the result of higher pungency when onions are fertilized solely with (NH4)2SO4 (Gamiely et al., 1991)
may suggest growers participating in the PA Simply Sweet® program should avoid (NH4)2SO4 as their
primary N source. Indeed, crops intended for marketing through the Simply Sweet® program have been
turned away if their pungency levels are too high (J. Stoltzfus, pers. comm.). This may result in an economic
loss to growers who paid membership fees to participate in the program prior to the season, and
additional effort is required to find an alternative market for the crop.
Several conclusions are suggested by the results of the study presented here; among them, a
trend emerges in which full-season fertilizer applications may result in larger onion necks in late-season
(Table 6.1), as well as higher center rot incidence (Figs. 6.3, 6.4, 6.8) as compared to half-season fertilizer
applications, particularly in plants under high inoculum pressure. Since uninoculated plants in unfertilized
plots, which received only 63% of the full season concentration, had the highest marketable yields (data
not shown), and generally, plants’ failure to respond to additional nutrients may suggest adequate N
availability (Lorenz and Tyler, 1976; cited by Rumpel et al., 2004), this suggests that N fertility
concentrations lower than current recommendations may be feasible for PA growers while maintaining
current yields. While the full-season Ca(NO3)2 treatment on inoculated plants had significantly higher
119
disease incidence than both the half-season Ca(NO3)2 and the full-season (NH4)2SO4 treatments, the lack
of relationship between this treatment and the half-season (NH4)2SO4 treatment (Fig. 6.3) necessitates at
least another year of replicated trials, preferably in multiple locations with different soil types. In these
further trials, an alternative NH4 source would be advantageous to include, since fertilizing onions with
(NH4)2SO4 increases bulb S and likely increases onion pungency (Fig. 6.5). In addition, the roles of foliar
nutrient levels in center rot progression may be manipulated through plant nutrition applications as an
additional disease management approach, given that foliar Ca and N were significant variables in the
models presented here (Figs. 6.6 and 6.7).
Acknowledgements
The authors thank Tim Grove and Michele Mansfield for technical assistance in field maintenance and inoculations.
Tables and Figures
Fig. 6.1. Lesion development in differentially-fertilized onion seedlings after foliar inoculation with P. ananatis in a growth chamber assay. Inoculated plants are indicated by INC. Means at each date were compared between treatments using ANOVA with Fisher’s LSD; * indicates a statistically significant difference between the Nitrate and Ammonium-fertilized, inoculated plants, while ** indicates a statistically significant difference between inoculated control and nitrate fertilizer treatments compared to the ammonium-only fertilizer treatment. Error bars represent standard error of the mean.
*
**
120
Fig. 6.2. Disease progression over time by fertilizer application timing in inoculated plants in the field. N treatments were combined within the timing variable since no differences were apparent between N types. Means within each type of timing were compared using a one-way ANOVA and Fisher’s LSD (α = 0.05). * indicates a statistically significant difference in disease severity between the full-season and half-season fertilized treatments.
Table 6.2. Midseason growth estimates and thrips pressure by fertilizer treatment prior to inoculation in 2013.
Treatment by fertilizer type and timinga Treatment by application timingb
a Treatments were as follows: unfertilized (none); half-season fertilized with Ca(NO3)2; full-season fertilized with Ca(NO3)2; half-season fertilized with (NH4)2SO4; full-season fertilized with (NH4)2SO4.
b Groups in application timing are combined regardless of N source. Different letters following measurements within the row indicate statistical significance via Fisher’s LSD (α = 0.05).
c Ratio of the bulb diameter, at its widest point, to the neck diameter, at its narrowest point, for three plants per replicate plot. (N = 8 reps for unfertilized plots; N = 16 for fertilized plots).
d Neck diameter in mm.
*
121
Fig. 6.3. Center rot at harvest by inoculation status, N fertilizer source and application timing. Bars indicate the average of four replicate plots, error bars represent standard error of the mean. Bars with different letters indicate statistical significance by Fisher’s LSD (α = 0.05).
Fig. 6.4. Center rot incidence at harvest based on inoculation status and fertilizer application timing. Bars indicate the average of the plots within each category (N = 4 [unfertilized plots] or N = 8 [half- or full-season fertilized plots, regardless of N source]). Error bars represent standard error of the mean and different letters above each bar indicate statistical significance by Fisher’s LSD (α = 0.05).
Fig. 6.5. Sulfur content of asymptomatic bulbs at harvest by N fertilizer type and timing. Bars indicate averages by treatment group; different letters above bars indicate statistical significance by Fisher’s LSD (α = 0.05).
Fig. 6.6. Analysis of covariance estimates with treatment (listed under each bar with the timing and form of inorganic N) as a main effect and calcium in foliage after three weekly fertilizer treatments as a covariate. Inoculated treatments are solid color bars, uninoculated treatments are hatched. Dark blue indicates full-season application (N fertility applied weekly throughout the season), medium blue indicates half-season application (N fertility applied weekly prior to midseason), and light blue indicates no additional N fertility. Means shown are estimates with early-season foliar Ca included as a covariate in the model. Bars with different letters are significantly different by Tukey’s test (α = 0.05).
123
Fig. 6.7. Analysis of covariance estimates with treatment (listed under each bar with the timing and form of inorganic N) as a main effect and nitrogen in foliage after six weekly fertilizer applications as a covariate. Inoculated treatments are solid color bars, uninoculated bars are hatched. Dark blue bars indicate full-season fertilizer application, medium blue bars indicate half-season fertilizer application, and light blue bars indicate no fertilizer application. Means shown are estimates with N included in the model. Bars with different letters are significantly different by Tukey’s test (α = 0.05).
Fig. 6.8. Center rot incidence at harvest compared to amount of N fertilizer of either type applied by midseason. Statistically significant differences only exist between means in different inoculation categories. Error bars indicate standard error of the mean.
Full None Half Full Half Half Full Half Full None NO
3
NH
4 NH
4 NO
3 NH
4 NO
3 NO
3
NH4
124
References Buckland, K., Reeve, J. R., Alston, D., Nischwitz, C., Drost, D. 2013. Effects of nitrogen fertility and crop rotation on onion growth and yield, thrips densities, Iris yellow spot virus, and soil properties. Agriculture, Ecosystems and Environment 177: 63 – 74. Carr, E. A., Zaid, A. M., Bonasera, J. M., Lorbeer, J. W., Beer, S. V. 2013. Infection of onion leaves by Pantoea ananatis leads to bulb infection. Plant Disease 97:1524 - 1528. Coolong, T. W., Randle, W. M. 2008. The effects of calcium chloride and ammonium sulfate on onion bulb quality at harvest and during storage. HortScience 43: 465 – 471.
Diaz-Perez, J. C., Purvis, A. C., Paulk, J. T. 2002. Bolting, yield, and bulb decay of sweet onion as affected by nitrogen fertilization. Journal of the American Horticultural Society 128: 144 - 149. Edens, D. G., Gitaitis, R. D., Sanders, F. H., Nischwitz, C. 2006. First report of Pantoea agglomerans causing a leaf blight and bulb rot of onions in Georgia. Plant Disease 90: 1551. Elmer, W. H. 2000. Comparison of plastic mulch and nitrogen form on the incidence of Verticillium wilt of eggplant. Plant Disease 84: 1231-1234. Elmer, W. H., LaMondia, J. 1999. Influence of ammonium sulfate and rotation crops on strawberry black root rot. Plant Disease 83: 119-123. Gamiely, S., Randle, W. M., Mills, H. A., Smittle, D. A., Banna, G. I. 1991. Onion plant growth, bulb quality, and water uptake following ammonium and nitrate nutrition. HortScience 26: 1061-1063. Gitaitis, R. D., Gay, J. D. 1997. First report of a leaf blight, seed stalk rot, and bulb decay of onion by Pantoea ananas in Georgia. Plant Disease 81: 1096. Gitaitis, R. D., Gent, D. H., Schwartz, H. S. 2008. ‘Leaf streak and bulb rot’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 59. Huber, D. M., Graham, R. D. 1999. ‘The role of nutrition in crop resistance and tolerance to diseases’ in: Mineral Nutrition of Crops: Fundamental Mechanisms and Implications. Food Products Press, New York. Pp. 169 – 204. Huber, D. M., McCay-Buis, T. S. 1995. A multiple component analysis of the take-all disease of cereals. Plant Disease 77: 437 – 446. Huber, D. M., Thompson, I. A. 2007. ‘Nitrogen and Plant Disease’ in: Mineral Nutrition and Plant Disease. APS Press: St. Paul, MN. Pp. 31 – 44. King, E. O., Ward, M. K., Raney, D. E. 1954. Two simple media for the demonstration of pyocyanin and fluorescin. Journal of Laboratory and Clinical Medicine 44, 301–307.
125
Malik, M. F., Nawaz, M., Ellington, J., Sanderson, R., El-Heneidy, A. H. 2009. Effect of different nitrogen regimes on onion thrips, Thrips tabaci Lindemann, on onions, Allium cepa L. Southwestern Entomologist 34: 219 – 225. McGuire, R. G., Kelman, A. 1984. Reduced severity of Erwinia soft rot in potato tubers with increased calcium content. Phytopathology 74: 1250-1256.
Mohan, S. K. 2008a. ‘Other Bacterial Soft Rots’ in: Compendium of Onion and Garlic Diseases and Pests, 2nd edition. APS Press: St. Paul, MN. pp. 62.
Rahman, M., Punja, Z. K. 2007. ‘Calcium and Plant Disease’ in: Mineral Nutrition and Plant Disease. APS Press: St. Paul, MN. pp. 79 – 89.
Randle, W. M., Kopsell, D. E., Kopsell, D. A., Snyder, R. L. 1999. Total sulfur and sulfate accumulation in onion is affected by sulfur fertility. Journal of Plant Nutrition 22: 45 – 51.
Rumpel, J., Kaniszewski, S., Dysko, J. 2004. Effect of drip irrigation and fertilization timing and rate on yield of onion. Journal of Vegetable Crop Production 9: 65 – 73.
Westerveld, S. M., McDonald, M. R., Scott-Dupree, C. D., McKeown, A. W. 2003a. The effect of nitrogen on insect and disease pests of onions, carrots, and cabbage. Journal of Vegetable Crop Production 8 (2): DOI: 10.1300/J068v08n02_09.
Westerveld, S. M., McKeown, A. W., Scott-Dupree, C. D., McDonald, M. R. 2003b. How well do critical nitrogen concentrations work for cabbage, carrot, and onion crops? HortScience 38: 1122-1128.
Wright, P. J. 1993. Effects of nitrogen fertilizer, plant maturity at lifting, and water during field-curing on the incidence of bacterial soft rot of onions in store. New Zealand Journal of Crop and Horticultural Science 21: 377 – 381.
126
Chapter 7: Revised best practices for onion production in Pennsylvania and future work
Pennsylvania onion growers participating in the Simply Sweet® program are faced with the
combined challenges of bacterial inoculum pressure, use of highly susceptible onion varieties, and
environmental conditions that favor development of bacterial disease. The advantages of participation in
the program include the fact that the supply of Simply Sweet® onions, the state’s only trademarked crop,
currently does not meet consumer demand, the timing of the onion growing season is convenient in
relation to other vegetable crops, and no specialized equipment is necessary. Thus, onions are an excellent
crop for vegetable growers to diversify their farm enterprises. Based on four years’ worth of intensive
sampling and observation, field experimentation, and data analysis, revisions to the current best practices
for onion production may be made to enhance the profitability and expansion of the crop in Pennsylvania.
While most commercial growers in Pennsylvania use plastic mulch and drip irrigation for weed
and disease suppression, respectively, individuals have additional production choices to consider as they
plan their onion fields. One recommendation, based on several years of replicated trials (Gugino,
unpublished) in addition to soil temperature data (Chapter 4), is to use biodegradable mulch to reduce
bacterial disease through moderation of soil temperatures. Early season soil warmth is important for
onion transplant establishment, however, soil temperatures, especially at and after bulbing, are positively
associated with bacterial disease (Chapter 4). Another early-season aspect of effective bacterial disease
management is preventing the introduction of bacterial inoculum into the onion crop. Onion transplants
should be carefully inspected for symptoms of bacterial disease, with only healthy plants planted into the
field. The majority of Pantoea agglomerans, Pectobacterium carotovorum, and Pseudomonas marginalis
isolates from transplants induce rotting symptoms in onion (Chapter 2), and analyses from 54 different
locations suggest this may particularly be the case with P. marginalis (Chapter 3). In the future, at-plant
treatments should be investigated for their potential to reduce bacterial inoculum from transplants;
suggested products to evaluate are surface-sterilizing treatments like hydrogen peroxide and biorational
products with Bacillus spp. as active ingredients, such as those used in Chapter 5. Pathogenic bacteria
have also been isolated from early-season soil (Chapter 2), so growers should choose a field with well-
drained soil that has not been planted to Alliums within the last two seasons, and low-lying areas should
be avoided (Chapter 5). There is evidence for antagonism between environmental (soil and weed-derived)
strains of Pantoea agglomerans and detections of the same bacterium in symptomatic onions (Chapter
3), which could be further investigated in laboratory and field experiments for potential biological control.
Some surface sterilants and plant defense-inducing products, such as those containing Bacillus
subtilis, were shown to be as effective as grower standard copper-mancozeb treatments in managing
127
center rot of onion under low bacterial inoculum pressure (Chapter 5), so these products may play a role
in integrated disease management in the future. Additional experimentation on how best to integrate
these products for disease as well as potential copper resistance management is necessary before making
full recommendations, however. Weed control is also a potential in-season disease management strategy.
Bacterial isolates from weeds were shown to induce symptoms in pathogenicity tests, and one isolate of
Pantoea ananatis was obtained from both a midseason weed and a symptomatic onion bulb from storage,
from the same farm (Chapter 2). Additional work into the role of weeds as sources of bacterial inoculum
may be undertaken, such as the potential for pathogens of interest to be seedborne on weed seeds, the
timing of bacterial epiphytic and endophytic colonization of weeds, and the roles of weeds as potential
green bridges for inoculum movement from onion to onion. There may also be a role for onion thrips, the
most common insect pest of onions, in the epidemiology of bacterial rots (Appendix A), so thrips should
be managed on farms with high pest pressure. Additional research into the practical significance of these
insects is necessary to determine thresholds for management.
Onion growers participating in the Simply Sweet® program as well as other onion growers benefit
from producing a greater proportion of jumbo- and colossal-size bulbs compared to medium and small
bulbs, so fertigation programs are of significant interest to Pennsylvania growers. From one year of data,
it was suggested that while the primary form of supplied nitrogen (nitrate or ammonium) did not affect
center rot incidence, the timing of nitrogen application may play a role in center rot development. Using
the same fertilization rates, treatments where total nitrogen for the season was applied prior to bulbing
were similar between uninoculated and inoculated plots, but this was not the case for unfertilized and
full-season fertilized onion plots (Chapter 6). This suggests that in-season nitrogen application should be
completed in the early part of the season in addition to avoiding sulfur-containing products for the sole
fertilizer source, however, additional replicated experimentation is necessary to verify these preliminary
results. In addition, associations between foliar calcium and nitrogen levels suggest further investigation
of the role of these nutrients in center rot disease management (Chapter 6).
As the onion season draws to a close, growers are perennially faced with the question of when to
harvest their crops. Growers who observe high bacterial disease pressure often harvest early, trading bulb
size for the assurance of a larger proportion of asymptomatic onions. Other growers favor postponing
harvest to maximize bulb size, but at the cost of higher disease incidence. Current and ongoing work using
the disease rating scale (Chapter 5) in combination with regular scouting will help to identify a threshold
foliar disease severity value growers may use to determine when to harvest individual fields to maximize
yield with an economically tolerable level of bulbs with bacterial rot. Following harvest, rapid drying-down
128
of onion necks can prevent the movement of leaf-infecting bacteria into onion bulbs, reducing postharvest
disease. Other future work includes an onion variety trial to identify less-susceptible cultivars and
repeated experiments into plant defense-inducing compounds and fertilizer programs for managing
bacterial disease. Combinations of these cultural and disease management practices will ideally increase
the profitability of producing onions in Pennsylvania, with some production practices broadly applicable
to reduce the incidence of bacterial disease in a variety of pathosystems.
129
Appendix: Preliminary work with thrips identification by PCR primers and the impact of onion thrips
(Thrips tabaci) on center rot of onion (Allium cepa) in Pennsylvania
Introduction
Onion thrips, Thrips tabaci Lindeman, have been reported as the primary insect pest in onion
production in New York (Shelton et al., 2006), and varying levels of thrips damage were found on all 32
Pennsylvania farms surveyed in 2011 and 2012 (Pfeufer and Gugino, unpublished), but the prevalent
species in PA are unknown. Other thrips with potential to be found in Pennsylvania include western flower
thrips (Frankliniella occidentalis Pergande), tobacco thrips (F. fusca Hind), flower thrips (F. tritici Fitch),
and tomato thrips (F. schultzei Trybom; Felland et al., 1995; B. Nault, pers. comm.). Specifically, western
flower thrips and onion thrips have been previously demonstrated to vector P. ananatis and P.
agglomerans (Wells et al., 2002; Gitaitis et al., 2003; Dutta et al., 2012, 2014), in addition to serving as
vectors for iris yellow spot virus, which illustrate why thrips management for onions may have more
significant impacts than solely yield concerns in PA.
Western flower thrips, F. occidentalis, has been analyzed in terms of its gut bacteria, which
includes P. agglomerans, in several studies (de Vries et al., 2001; de Vries et al., 2004). These studies
suggested that western flower thrips became infested during larval stages, the bacteria reside in the
hindgut, and bacterial relationships with the insect change based on the food source (de Vries et al., 2001;
de Vries et al., 2004). The relationship between the insect and its gut symbionts was shown to be parasitic
when the insect food source was nutrient-rich, and beneficial when the food source was poor (de Vries et
al., 2004). In another publication, complicated interactions between F. occidentalis and chrysanthemum
nitrogen content demonstrated that the rate of increase in insect number was highest on plants grown in
high N, but that plant phenology was more influential overall in thrips population development (Chau et
al., 2005). Fertility management has been suggested to be an important factor affecting onion thrips
populations in multiple locations (Buckland et al., 2013; B. Nault and C. Hoepting, pers. comm.).
A comparison of F. occidentalis (western flower thrips) gut bacteria and T. tabaci (onion thrips)
gut bacteria based on 16S rDNA sequence and biochemical characteristics identified the type species of
four out of five T. tabaci gut bacterial populations as Erwinia herbicola (de Vries et al., 2008). In 1989,
however, it was suggested that E. herbicola isolates be placed into the P. agglomerans species, along with
the former Enterobacter agglomerans and Erwinia milletiae (Gavini et al., 1989). A different phylogenetic
analysis of several different F. occidentalis populations identified two consistent, distinct bacterial
symbionts associated with the insects. Both isolates were identified as members of the family
Enterobacteriaceae, with one grouping with Erwinia species, and the other into an unknown genus most
130
closely related to E. coli and Enterobacter sakazakii (Chanbusarakum and Ullman, 2008). It is important to
note, however, that neither of the author groups used any Pantoea species in their analyses. In 2010 and
2011, P. agglomerans was successfully amplified from macerated F. occidentalis found feeding on onions
in the Penn State greenhouses (Pfeufer, unpublished).
Very recently, it has been shown that onion thrips have the ability to acquire as well as transmit
P. ananatis and P. agglomerans (Dutta et al., 2012, 2014). This research group showed that at least 92%
of onion thrips had acquired either bacterium after feeding for 48 hours on inoculated plants, and
transmission efficiency was approx. 60% for P. ananatis and 75% for P. agglomerans (Dutta et al., 2014).
While some PA growers manage thrips populations in their fields through the use of insecticides, it is
common for growers to discount the role of thrips in yield losses, and certainly in terms of disease
incidence in their crops (Pfeufer, unpublished). Little data exists pertaining to the impact of these insects
on yield losses due to bacterial rots of onion in PA. These lines of evidence suggest undertaking a more
thorough investigation into the potential role of thrips as field vectors of the Pantoea pathogens of onion.
Thrips may be identified to species using morphological keys, PCR, DNA fingerprinting,
sequencing, or combinations of methods. The use of morphological keys requires training, and can be
ambiguous because environmental conditions, such as temperature and moisture, can alter phenotypic
characteristics of thrips within the same species (Asokan et al., 2006). In addition, thrips larvae are
morphologically indistinguishable to the species level. RFLP-DNA fingerprinting methods generally
restriction-digest amplified copies of the cytochrome oxidase gene (CO1), generating banding patterns
that are species-specific (Brunner et al., 2002), however, restriction digest procedures are costly in terms
of time and materials. PCR identification methods are based on amplifying DNA of a known sequence
within the insect genome and generating DNA fragments of a specific size. PCR species identification
methods have been published for T. tabaci, T. palmi, and F. occidentalis (Asokan et al., 2006; Huang et al.,
2010; Toda and Komazaki, 2002), however, the use of some published primer sets and protocols has
produced inconsistent results (Pfeufer, unpublished).
To more clearly elucidate the impact of thrips on onion production in PA and determine the
potential for these insects to vector Pantoea ananatis and P. agglomerans, thrips were collected from
grower and research farms, visually identified to species, then bacterial isolates were generated from
these insects. Preliminarily, published PCR primers were tested for their specificity to Thrips tabaci, then
novel primers were designed in order to facilitate more rapid identification of onion thrips. These activities
were not continued past February 2013.
131
Materials and Methods
Sample collection and identification
Thrips were collected from fourteen of the 26 farms visited for the observational study in 2012 by
removing an onion leaf very close to the plant neck and placing the tissue, along with the insect, into a 50
mL screwtop tube. Tubes were placed on ice in a cooler until return to the lab, where they were placed in
4⁰C storage. Within four weeks, thrips were removed from the tubes and tentatively identified to species
by inspecting them under at least 100x magnification for morphological characteristics. Both T. tabaci and
F. occidentalis are golden to gray-tan in color, adults 1-2 mm long, with two pairs of fringed-wings typical
of insects in the order Thysanoptera (UM Extension). Both species of thrips have lateral brown bands
alternated with yellow bands on the thorax. T. tabaci may vary more toward the grayish coloration and
are slightly smaller than F. occidentalis. Both species’ coloration can vary with environment, but the
following descriptive characters are consistent: T. tabaci have 7 antennal segments, 3 gray ocelli posterior
to the eyes, and shorter post-ocellar setae, while F. occidentalis have 8 antennal segments, 3 red ocelli
posterior to the eyes, and longer post-ocellar setae (UM Extension; Hoddle et al., 2012; Fig. A.1). T. tabaci
are more prevalent in the NY production system and can overwinter there (Nault, pers. comm), however,
F. occidentalis are a common greenhouse pest in PA, but seem to be found less frequently in the field
(Pfeufer, unpublished), perhaps as a result of difficulty in overwintering or poor competitorship.
Fig. A.1. Frankliniella occidentalis (left) and Thrips tabaci (right) as viewed at
100x magnification (E. Pfeufer).
132
Sample preparation and bacterial isolation
After putative identification to species, insects were washed once in buffer (for epiphytic
bacteria), then washed twice in 70% ethanol, rinsed in sterile H20, then each was placed in a sterile tube
containing 20 µL of sterile buffer. Insects were macerated in this buffer using a sterile micropestle, then
five µL of buffer was used to inoculate culture tubes containing five mL of sterile nutrient broth (NB).
Tubes were incubated at 30⁰C with agitation (150 rpm) overnight, then broth was spread on King’s B
media, from which bacterial isolates were chosen. Isolates were generated by choosing a single colony
from KB with a sterile toothpick, then inoculating glass tubes containing five mL sterile Luria-Bertani broth
(Becton, Dickinson Co., Sparks, MD, USA), which was allowed to grow overnight before freezing at -20⁰C
in 15% glycerol.
Bacterial species identification
Bacterial isolates with phenotypic characteristics typical of Pantoea spp. (yellow on KB media)
were subjected to direct-colony multiplex PCR using the Pantoea ananatis and P. agglomerans primers
described in Chapter 2. PCR template was a toothpick touched to a single bacterial colony grown on KB.
In addition, a subset of unknown bacterial isolates were amplified for 16S rDNA sequencing, using the
primer sequences 530F and 1492R (Borneman et al., 1996). PCR reactions were cleaned using ExoSAP-it
(USB®, Cleveland, OH, USA), then 5 µL of reaction mixed with 1 µL of 590F primer was submitted to the
Penn State Nucleic Acid Core Facility for sequencing. Sequence data was edited, then NCBI BLAST was
used to identify the bacterial isolates to species.
Pathogenicity tests
A subset of isolates was tested for pathogenicity using aerobic and semi-anaerobic incubations
described in Chapter 2, where surface-sterilized pearl onion bulbs were inoculated with approx. 100 µL
overnight broth culture, then incubated one or two weeks, respectively. Symptom development was rated
according to the scale in Chapter 2.
Thrips species identifications
Previously published primers for T. tabaci (Asokan et al., 2007), F. occidentalis (Huang et al., 2010),
and Thrips palmi (Asokan et al., 2007; Table A.1) were tested for specificity using thrips collected from
onion fields as well as the Penn State greenhouses. Insects of the order Thysanoptera were putatively
identified to species based on morphological characteristics; no T. palmi were identified. Thrips were
133
macerated in 20 µL sterile H20, then macerates were used as templates for PCR reactions composed of
12.5 µL TaqPRO Complete master mix (Denville Scientific, Metuchen, NJ), 1 µL species 1 forward primer
(10 mM), 1 µL species 2 forward primer (10 mM), 1 µL species 1 reverse primer (10 mM), 1 µL species 2
reverse primer (10 mM), and 7.5 µL PCR water. PCR conditions were as follows: (1) initial denaturation at
94°C for 3 min, followed by 40 cycles of (2) 94°C for 30 s; (3) 58°C (OT / MT) or 60°C (WFT) for 35 s, (4)
72°C for 1 min; then a final elongation at (5) 72°C for 20 min and storage at (6) 4°C continuous. PCR
amplicons were visualized by combining 10 µL reaction with 1.5 µL EZ-Vision Three dye (AMRESCO Inc.,
Solon, OH, USA) and electrophorescing in a 1.5% 1X TAE agarose gel for 45 min at 90V with either a 1kb
or 500 bp ladder (Denville, Metuchen, NJ, US).
Table A.1. Sequences, expected amplicon sizes, and sources of primers used to identify thrips collected from PA to species.
Primer Sequence Expected amplicon size
Source
T. tabaci F CGTTTTATCATTCAGGACC 298 bp Asokan et al., 2007
T. tabaci R AAGGTGTTGATATAAAACAGGGTCC
F. occidentalis F GTTTCCGTAGGTGAACCTGC 249 bp Modified from Huang et al., 2010 F. occidentalis R TGTTTTGGGCCATCTCCC
T. palmi F TTGACTTCTTCCACCCTCTTTAACTCTT 390 bp Asokan et al., 2007
T. palmi R TAGATGTTGATAAAGTACAGGATCT
Results and Discussion
Fig. A.2. Representative electrophoresis gel of PCR reactions using published Thrips tabaci (OT; expected amplicon size, 298 bp; Asokan et al., 2007), Frankliniella occidentalis (WFT; expected amplicon size, 249 bp; modified from Huang et al., 2010), and Thrips palmi (MT; expected amplicon size, 390 bp; Asokan et al., 2007) primer sets.
Several PCR reactions were attempted using published T. tabaci (Asokan et al., 2007), F.
occidentalis (Huang et al., 2010), and T. palmi (Asokan et al., 2007) primers to determine if these primer
sets would reliably differentiate the three species of thrips, forgoing morphological identification. F.
occidentalis was accurately identified in approx. 80% of successful reactions using visually identified,
134
macerated F. occidentalis as the template, and failed reactions may have occurred as a result of
inappropriate cycling parameters. F. occidentalis primers did amplify some putatively identified T. tabaci,
though the fragment size was not the appropriate size (approx. 175 bp; Fig. A.2). Nevertheless, these
primers hold promise for the simple, rapid detection of F. occidentalis. Combining all three primer sets
and amplifying using the OT/MT PCR parameters still resulted in amplification of F. occidentalis at the
expected amplicon size, however, T. tabaci was no longer identified in the triplex reaction (Fig. A.2).
Visually identified T. tabaci were identified using the published T. tabaci primers in 50% of
reactions attempted (12 of 24; Fig. A.2). However, the amplicon generated was quite faint and was the
incorrect size; approximately 500 bp, when the expected amplicon size was 298 bp (Asokan et al., 2007).
Since the published primers were developed based solely on onion thrips from India, it may be that PA
populations of onion thrips have insertion(s) in the CO1 gene, resulting in larger than expected amplicon
sizes. However, additional PCR protocols would need to be completed in order to determine if this is the
case. The published T. palmi primers did not amplify any of the extracts used, which was expected, since
none of the extracted insects were visually identified as T. palmi (Fig. A.2).
Table A.2. Bacterial isolates from thrips tested for pathogenicity on onion through aerobic and semi-anaerobic pathogenicity tests. Ratios in each column pertain to the pathogenic isolates out of all isolates of that species tested; total columns indicate the total number of strains of each species in the collection.
Aerobic test Anaerobic test Total Aerobic test Anaerobic test Total
P. ananatis 3 / 5 10 1 / 2 1 / 2 4
P. agglomerans 0 / 2 1 / 2 3 3 / 4 3 / 4 4
Enterobacter sp. 3 1
Rahnella sp. 1 4
Other 5
Not yet identified 20 50
Total isolates 3 / 7 1 / 2 37 4 / 6 4 / 6 68
Pathogenicity tests of bacterial isolates from the surfaces and within the tissues of thrips sampled
from Rock Springs and selected grower farms indicated bacteria pathogenic to onion were isolated from
these insects (Table A.2). In particular, pathogenic P. ananatis and P. agglomerans were isolated in low
numbers, which supports previous reports of these insects vectoring the Pantoea spp. bacterial pathogens
(Dutta et al., 2014). However, significantly more evidence is necessary to determine the true impact of
thrips on bacterial rots of onion in PA. While onion thrips have recently been demonstrated to acquire
and transmit the Pantoea spp. in laboratory assays (Dutta et al., 2014), additional data is necessary
135
pertaining to the frequency of thrips vectoring these pathogens in PA, the relative losses incurred by high
thrips pressure, both in terms of yield and disease, in addition to more basic information about the
prevalent species of thrips in and around onion fields in PA.
Acknowledgement
Caroline Black assisted E. Pfeufer initially in visually identifying thrips to species based on morphology.
References Alston, D. G., Drost, D. 2008. Onion thrips (Thrips tabaci). Utah Pests Fact Sheet, Utah State University Extension. Ent-117-08PR. Asokan, R., Krishna Kumar, N. K., Kumar, V., Ranganath, H. R. 2007. Molecular differences in the mitochondrial cytochrome oxidase I (mtCOI) gene and development of a species-specific marker for onion thrips, Thrips tabaci Lindeman, and melon thrips, T. palmi Karny (Thysanoptera: Thripidae), vectors of tospoviruses (Bunyaviridae). Bulletin of Entomological Research 97:461-470. Brunner, P. C., Fleming, C., Frey, J. E. 2002. A molecular identification key for economically important thrips species (Thysanoptera: Thripidae) using direct sequencing and a PCR-RFLP-based approach. Agricultural and Forest Entomology 4: 127-136. Chau, A., Heinz, K. M., Davies, F. T. 2005. Influences of fertilization on population abundance, distribution, and control of Frankliniella occidentalis on chrysanthemum. Entomologia Experimentalis et Applicata. 117:27-39. Chanbusarakum, L., Ullman, D. 2008. Characterization of bacterial symbionts in Frankliniella occidentalis (Pergande), western flower thrips. Journal of Invertebrate Pathology 99:318-325. de Vries, E. J., Jacobs, G., Breeuwer, J. A. J. 2001. Growth and transmission of gut bacteria in the western flower thrips, Frankliniella occidentalis. Journal of Invertebrate Pathology 77:129-137. de Vries, E. J., Jacobs, G., Sabelis, M. W., Menken, S. B. J., Breeuwer, J. A. J. 2004. Diet–dependent effects of gut bacteria on their insect host: the symbiosis of Erwinia sp. and western flower thrips. Proceedings of the Royal Society of London 271: 2171-2178. de Vries, E. J., van der Wurff, A. W. G, Jacobs, G., Breeuwer, J. A. J. 2008. Onion thrips, Thrips tabaci, have gut bacteria that are closely related to the symbionts of the western flower thrips, Frankliniella occidentalis. Journal of Insect Science 8:23 (11 pp., available online). Doda, S., Komazaki, S. 2002. Identification of thrips species (Thysanoptera: Thripidae) on Japanese fruit trees by polymerase chain reaction and restriction fragment length polymorphism of the ribosomal ITS2 region. Bulletin of Entomological Research 92: 359-363.
136
Dutta, B., Gitaitis, R., Langston, Avci, U., Barman, A., Srinivasan, R. 2012. Acquisition and transmission of Pantoea ananatis and Pantoea agglomerans (causal agents of center rot of onion) by Thrips tabaci (Thysanoptera: Thripidae). Phytopathology S4:88. Dutta, B.., Barman, A. K., Srinivasan, R., Avci, U., Ullman, D., Langston, D. B., Gitaitis, R. 2014. Transmission of Pantoea ananatis and Pantoea agglomerans, causal agents of center rot of onion (Allium cepa L.) by onion thrips (Thrips tabaci Lindeman) through feces. Phytopathology 104: 812 – 819. Felland, C. M., Teulon, D. A. J., Hull, L. A., Polk, D. F. 1995. Distribution and management of thrips (Thysanoptera: Thripidae) on nectarine in the mid-Atlantic region. Journal of Economic Entomology 88(4):1004-1011. Hoddle, M.S., Mound, L.A., Paris, D.L. 2012. Thrips of California. CBIT Publishing, Queensland. Hoepting, C. A. Cornell University extension educator and onion specialist. Personal conversations, April 2012 – present. Huang, K. S., Lee, S. E., Yeh, Y., Shen, G. S., Mei, E., Chang, C. M. 2010. Taqman real-time quantitative PCR for identification of western flower thrip (Frankliniella occidentalis) for plant quarantine. Biology Letters 6:555-557. Morse, J. G., M. S. Hoddle. 2006. Invasion biology of thrips. Annual Review of Entomology 51: 67-89. Nault, B. Cornell University research and extension entomologist. Personal conversation, February 2012. Shelton, A. M., Zhao, J. Z., Nault, B. A., Plate, J., Musser, F. R., Larentzaki, E. 2006. Patterns of insecticide resistance in onion thrips (Thysanoptera: Thripidae) in onion fields in New York. Journal of Economic Entomology 99: 1798 - 1804. Toda, S., Kamazaki, S. 2002. Identification of thrips species (Thysanoptera:Thripidae) on Japanese fruit trees by polymerase chain reaction and restriction fragment-length polymorphism of the ribosomal ITS2 region. Bulletin of Entomological Research 92:359-363. Wells, M. L., Gitaitis, R. D., Sanders, F. H. 2002. Association of tobacco thrips, Frankliniella fusca (Thysanoptera: Thripidae), with two species of bacteria of the genus Pantoea. Annals of the Entomological Society of America 95: 719-723.
Emily E. Pfeufer VITA
Education Doctor of Philosophy in Plant Pathology .................................................................. expected August, 2014 Master of Science in Plant Pathology ....................................................................................... August, 2010 Bachelor of Science in Biology, cum laude ................................................................................... May, 2008
Awards Paul Hand Award for Graduate Student Teaching Achievement, Penn State .......................... March, 2014 2nd place, Life Sciences Division, 2013 Penn State Graduate Exhibition ..................................... April, 2013 1st place, Biological Sciences Division, 2013 PSU - GSD Graduate Research Competition .......... April, 2013 I.E. Melhus Graduate Student Symposium, American Phytopathological Society ............... February, 2013 Henry W. Popp Award, Penn State Department of Plant Pathology and Env. Micro. .............. August, 2012 Larry Jordan Endowment, Penn State Department of Plant Pathology and Env. Micro. .............. May, 2012 Henry W. Popp Award, Penn State Department of Plant Pathology and Env. Micro. ............... August, 2010 4th place, Biological Sciences Division, 2010 PSU – GSD Graduate Research Competition .......... April, 2010 Crouch Graduate Fellowship, Penn State College of Agricultural Sciences ............................. August, 2008 Alden Scholar, Allegheny College...................................................................... Academic terms, 2005-2008 Trustee Scholarship, Allegheny College ........................................................... Academic terms, 2004-2008 Peer-reviewed publications Pfeufer, E. E., Hoepting, C. A., Gugino, B. K. 2014. Environmental and management factors related to bacterial rots of onion in Pennsylvania and New York. Plant Disease; in preparation.
Mansfield, M. A., Pfeufer, E. E., Gugino, B. K. 2014. Diagnostic multiplex PCR method to detect onion bacterial pathogens in plant tissue and environmental samples. Plant Disease; in preparation. Ramos, L. S., Lehman, B. L., Sinn, J. P., Pfeufer, E. E., Halbrendt, N. O., McNellis, T. W. 2013. The fire blight pathogen Erwinia amylovora requires the rpoN gene for pathogenicity in apple. Molecular Plant Pathology 14: 838 – 843. Pfeufer, E. E., Ngugi, H. K. 2012. Orchard factors associated with resistance and cross-resistance to sterol demethylation inhibitor fungicides in populations of Venturia inaequalis from Pennsylvania. Phytopathology 102: 272-282. Brazelton, J., Pfeufer, E. E., Sweat, T., McSpadden Gardener, B., Coenen, C. 2008. 2,4-diacetylphloroglucinol alters plant root development. Molecular Plant Microbe Interactions 21: 1349-1358.