Pantoea spp. Associated with Smooth Crabgrass (Digitaria ischaemum) Seed Inhibit Competitor Plant Species

Digitaria ischaemum (Schreb.) Schreb. ex Muhl. and Poa annua L. are competitive, early successional species which are usually considered weeds in agricultural and turfgrass systems. Bacteria and fungi associated with D. ischaemum and P. annua seed may contribute to their competitiveness by antagonizing competitor forbs, and were studied in axenic culture. Pantoea spp. were the most common bacterial isolate of D. ischaemum seed, while Epicoccum and Curvularia spp. were common fungal isolates. A variety of species were collected from non-surface sterilized P. annua. Certain Pantoea spp. isolates were antagonistic to competitor forbs Taraxacum officinale, Trifolium repens. All bacterial isolates that affected T. officinale mortality were isolated from D. ischaemum seed while none of the P. annua isolates affected mortality. Two selected bacterial isolates identified as Pantoea ananatis were evaluated further on D. ischaemum, T. repens (a competitor forb) and P. annua (a competitor grass) alone and in combination with a Curvularia sp. fungus. These bacteria alone caused >65% T. repens seedling mortality but did not affect P. annua seedling mortality. These experiments demonstrate that Pantoea ananatis associated with D. ischaemum seeds is antagonistic to competitor forbs in axenic culture. The weedy character of D. ischaemum could at least in part stem from the possession of bacteria that are antagonistic to competitor species.


Introduction
Digitaria (crabgrass) species are some of the most competitive C 4 weeds of agricultural, horticultural and turfgrass landscapes in tropical and temperate regions. Digitaria ischaemum and D. sanguinalis (smooth and large crabgrass, respectively) are among most problematic weed species in the United States, and are particularly well adapted to turfgrass systems where as a C 4 species they often outcompete C 3 grasses and forbs in the summertime [1]. Digitaria sanguinalis is known to produce allelochemicals that may contribute to its competitiveness, but these have not been studied extensively [2,3].
Poa annua is another prolific early successional C 3 plant that is extremely competitive in turfgrass and horticultural systems and can be found on all seven continents [4]. The success of P. annua as a weed is attributed to genotypic variability, prolific seed production, and a short life cycle [5].
The weed microbiome may also contribute to weed competitiveness. Communities of bacteria and fungi which colonize internal tissues as commensals or mutualists are referred to as endophytes and are ubiquitous in plants (reviewed in [6][7][8]). A wide range of endophytes are commonly found in

Seed Collection
Digitaria ischaemum seed was collected from a single site at the Rutgers Horticultural Research Farm No. 2 in North Brunswick, NJ (40 • 45 25 N 74 • 47 67 W). The site consisted of a mixed stand of D. ischaemum and Lolium perenne maintained as a mowed stand of turfgrass for several years. Mowing was suspended in October to allow inflorescence production and maturity prior to harvest in November 2016. Seed was collected from a 300 m 2 area using a leaf sweeper. Seeds were stored at 13 • C until January 2017 where they were moved to storage at −20 • C for one week prior to preparation to break seed dormancy. Poa annua seed was collected in June 2016 from a single site in University Park, PA (40 • 81 09 N 77 • 86 73 W) that was maintained as a mowed turfgrass stand for several years. Seeds were stored at 13 • C until January 2017 where they were moved to storage at −20 • C for one week prior to preparation to break seed dormancy.

Surface Sterilization and Microbe Isolation
P. annua and D. ischaemum seeds were surface sterilized by placing 5 g of seed in a 200 mL container filled with a 160 mL solution of DI water and 4.125% (v/v) NaOCl based on the methods of [20]. This container was placed on an orbital shaker to vigorously agitate the solution for 40 min. The bleach solution was then decanted and seeds were rinsed at least five times with sterile DI water in a laminar flow hood. Seeds were then placed on Petri dishes containing yeast extract sucrose agar (YESA; 1% yeast extract, 1% sucrose, 1.2% agar) and the plates were incubated at room temperature. Four seeds were placed on each plate and there were 10 replicates. The process was repeated with D. ischaemum seed one week later with 8 replicates; a second 5 g sample of seed was sterilized using the same process except that 5 µL of polysorbate 20 (Tween20, Thermo Fisher Scientific, Waltham, MA, USA) was added to the 160 mL of 4.125% (v/v) NaOCl solution. Outgrowing fungi were observed after 48 h of incubation and isolated to pure culture on potato dextrose agar for further study. After 7 to 14 days of incubation, bacteria were observed and isolated to pure culture on Luria-Bertani (LB) agar for further study. No outgrowing fungi or bacteria were observed in surface-sterilized P. annua seed after 14 days of observation. Therefore, using the same methods described above except without NaOCl, P. annua seeds were agitated in sterile water for 5 min and then the water was decanted and seeds were rinsed with sterile water before placing on YESA to isolate bacteria from non-surface-sterilized seeds.

Bacterial Identification and Characterization
Genomic DNA from bacteria was isolated using GenElute Bacterial Genomic DNA Kits (SigmaAldrich, St. Louis, MO, USA). Bacterial identifications were made by obtaining 16S rDNAsequences after methods employed by [35] using universal primers 16SF (5 -AGAGTTTGATCCTGGCTCAG-3 ) and 16SR (5 -CTACGGCTACCTTGTTACGA-3 ). Amplified PCR products were resolved by electrophoresis in 1.5% (w/v) agarose gel stained with SYBR safe for visual examination. The PCR products were purified using a PCR purification kit (Qiagen, Germantown, MD, USA) and sent to Genewiz Inc. (South Plainfield, NJ, USA) for sequencing. For certain isolates determined to be of interest after initial experiments (Pantoea spp. isolates 4 and 5), the tuf gene was also sequenced. Amplification of rDNA was performed in a 25 µL volume comprising: 12.5 µL Taq NEB 2x Master Mix, 9.5 µL water, 1.0 µL forward tufGPR (5 -GATACCAGTTACGTCAGTTGTACGGA-3 ), 1.0 µL reverse tufGPF (5 -ACGTTGACTGCCCAGGACAC-3 ) and 1.0 µL DNA. The reaction was performed in a thermocycler with an initial denaturation at 95 • C for 3 min, followed by 35 cycles at 95 • C for 30 s, 55 • C for 1 min, 72 • C for 30 s, and a final extension at 72 • C for 5 min. Sequences were compared to GenBank accessions using BLASTn (Available online: http://www.ncbi.nlm.nih.gov). Sequences were deposited in GenBank under accession numbers MG100861 to 863 and MK733357 to 370.

Experiment 1: Effect of Bacterial Isolates on Taraxacum Officinale Seedling Mortality
In turfgrass systems, Taraxacum officinale is a common competitor forb and was selected for use in these experiments. T. officinale seeds (collected from a lawn in South River, NJ, USA) were surface sterilized using the same process described above for D. ischaemum surface sterilization (agitation with 4.125% NaOCl for 40 min). Ten T. officinale seeds were plated onto each experimental unit, which consisted of a Petri dish (85 mm diameter) filled with 0.7% agarose media with no mineral nutrients. Petri dishes were maintained at ambient room temperature (22 • C to 27 • C) 40 cm below a fluorescent light source emitting 65 µmol m −2 s −1 PAR for 16 h per day. Isolates were maintained on trypticase soy agar (TSA) and streaked onto LB agar 12 to 24 h before inoculation. The bacteria were then removed from the agar with an inoculation loop and suspended in 1 mL of sterile water. Using this method, the suspension typically measured 0.8 OD ± 0.1 (600 nm) when diluted 1:3 with sterile water. One 3 µL drop of bacterial suspension was pipetted onto each seed within 1 h of placement on agarose. Isolates listed in Table 1 was evaluated in triplicate on a total of 30 seeds. A non-treated axenic control was included for comparison. A Pseudomonas fluorescens (Sandy LB4; GenBank No. KX665565) and Pantoea sp. isolate (RiLB4; GenBank No. KX752781) isolated from Phragmites australis seed by [20] were included as standards of comparison. The Sandy LB4 isolate has been shown to increase mortality of competitor forbs including T. officinale while RiLB4 did not affect mortality.
T. officinale seedlings that germinated (radicle at least 1 mm in length) were assessed as healthy, injured or dead at 14 days after treatment. Seedlings were considered healthy if cotyledons and leaves were green in color, turgid and not displaying any symptoms of cell membrane leakage (necrosis or greasy, off-color leaf tissues). Seedlings were considered dead if they were completely necrotic. Seedlings were scored as injured if they displayed some injury symptoms (i.e., chlorosis, necrosis) but other sections of the seedling were healthy and the seedling was not completely dead. The number of healthy, injured, and dead seedlings was assessed in each Petri dish and this number was used to determine the percentage of healthy, injured or dead plants in each experimental unit. Isolates were inoculated on different dates. Therefore, separate non-treated axenic controls were used for each experimental run and comparisons were made only to the non-treated control within each experiment. For isolates that increased mortality compared to the non-treated control, this experiment was repeated on T. officinale and T. repens. T. repens is also a common competitor species in turfgrass. T. repens seeds (purchased from a commercial seed source) were surface-sterilized using a similar method as T. officinale except that a 2% (v/v) NaOCl solution was used instead of a 4.125% solution. To ensure this method effectively sterilized clover seed, 100 seeds were placed on LB agar and no bacteria or fungi were observed after 10 d of incubation.
After T. officinale plants were evaluated at 14 days after treatments, the agarose plates containing seedlings were flooded with a 2 mmol/L solution of 3,3-diaminobenzidine tetrahydrochloride dissolved in DI water (DAB; Sigma Aldrich, St. Louis, MO) 12 h prior to observation to aid in visual observation of bacteria [36]. Randomly selected inoculated and non-inoculated plants were removed from the agarose and squash prepared with aniline blue (0.01%) dissolved in water and lactic acid for microscope observation at 1000× magnification.

Experiment 2: Effect of Bacterial Isolates on D. ischaemum Germination and Seedling Mortality
Based on the results of Experiment 1, Pantoea spp. bacterial isolates (4, 5, 8, 18, and 22) were evaluated against peeled, surface-sterilized D. ischaemum seedlings. To successfully remove cultivable bacteria from D. ischaemum seeds required several iterations of various sterilization procedures before an effective method was found. Seeds were placed on mesh screen and a wooden block was used to rub the seeds through the plastic screen, which removed the paleas and lemmas. These seeds were then placed on a finer plastic screen and a wooden block wrapped in the same screen was rubbed on the seed to remove the seed coat. These peeled seeds were separated from the debris and put into a 2 mL microcentrifuge tube filled with 1.5 mL of a 1.5% (v/v) NaOCl solution. The tube was vortexed for 5 min, the NaOCl solution was removed and fresh solution was added before vortexing again for 5 min.
Seeds were then rinsed several times with sterile water. To ensure this method effectively sterilized D. ischaemum seed, 100 seeds were placed on LB agar. No bacteria were observed after 10 days of incubation and a Curvularia sp. (determined through morphological characterization) was observed outgrowing from 2/100 seeds.
Surface-sterilized, peeled, D. ischaemum seeds were placed on agarose and inoculated with bacteria in the same manner as described in Experiment 1. T. officinale seeds were prepared using the same methods as in Experiment 1 included in the experiment as a standard of comparison to Experiment 1. This experiment was conducted once, but was repeated with Isolates 4 and 5 within Experiment 3.
2.6. Experiment 3: Effect of Bacterial Isolates Alone and in Combination with Curvularia sp. on D. ischaemum, T. repens and P. annua Seedlings The objective of these experiments was to evaluate the effect of certain bacterial isolates alone and in combination with a Curvularia sp. on germination and mortality of D. ischaemum, P. annua and T. repens. This experiment was conducted twice on separate dates. P. annua selected as it is a C 3 grass and also as a model plant that can be used evaluate the effect of these treatments on root gravitropism and root length based on previous research [20,37].
Surface-sterilized D. ischaemum, T. repens, and P. annua seeds were placed on agarose and inoculated as described previously. Peeled D. ischaemum seeds were used in this experiment. Seeds were inoculated with bacterial isolates 4 and 5, the combination of 4 + 5, RiLB4 (standard), or no bacteria alone and in combination with a Curvularia sp. isolate collected outgrowing from D. ischaemum seed forming a factorial treatment design. Bacterial isolates were prepared and inoculated onto seeds in the same manner as described for Experiment 1; isolates 4 and 5 were combined by taking an aliquot of each bacterial suspension and combining them in a 1:1 ratio. A suspension of Curvularia conidia was prepared by gently washing a potato dextrose agar lawn culture with a solution of sterile water and 0.05% polysorbate 20 and lightly scraping with an inoculation loop. This resulted in a suspension that contained some hyphae but was primarily conidia. Four 10 µL aliquots were sampled and evaluated with a hemocytometer grid to determine the conidia concentration. Within 2 h of spore suspension preparation, a 2 or 4 µL drop (run A and B, respectively) was applied to each seed to inoculate 10 3 ± 200 conidia per seed. The effect of each treatment on germination and seedling mortality was evaluated as previously described in Experiment 1.
For each P. annua seedling, it was determined whether the root penetrated vertically into the agarose or grew horizontally along the surface of the agarose at 14 days after inoculation. Roots penetrating the agarose were determined to have a positive gravitropic response and the percentage of roots demonstrating a positive gravitropic response is presented [38]. P. annua root length was measured by removing the plant from the agarose and measuring the length of the primary root with a ruler.

Statistical Analysis
Experimental units were arranged in a completely randomized design in all experiments. For Experiments 1 and 2, a single-factor ANOVA was conducted using the GLM procedure in SAS (SAS Institute, Cary, NC, USA) v9.4 (α = 0.05). For Experiment 3, treatments were subjected to a factorial analysis with bacteria isolate and Curvularia inoculation serving as main effects. Means were separated using Fisher's Protected LSD test at the 0.05 level.

Bacterial and Fungal Isolates
All isolates identified to the genus level had 16S rRNA sequences with >99% similarity to known sequences except for two (isolates 22 and 23) that had 94 and 95% similarity (Table 1). Pantoea, Staphylococcus and Paenibacillus spp. were isolated from non-surface sterilized P. annua seed.
No bacteria or fungi were observed outgrowing from surface sterilized P. annua seed. A Xanthomonas and Methylobacterium sp. along with eleven Pantoea spp. were selected from surface sterilized D. ischaemum seeds. Two Pantoea isolates of interest (4 and 5) were identified as Pantoea ananatis based on both 16S sequence similarity to Pantoea spp. and tuf gene sequences with >99% similarity to Pantoea ananatis. Two unique fungal species were isolated and determined to be Epicoccum and Curvularia spp. through morphological characterization [37]. A Curvularia sp. was observed on 55% and 28% of the seeds in the first and second experimental run, respectively, and was isolated for further experimentation given its prevalence on surface-sterilized seed in this experiment and previous reports of Curvularia sp. affecting thermotolerance [24][25][26]. An Epicoccum sp. was observed outgrowing from 4 of 40 seeds in the first run and 10 out of 32 seeds in the second run. 38 † Percent mortality is the percentage of emerged seedlings that were determined by visual assessment to be completely dead. Seedlings were scored as injured if they displayed some injury symptoms but were not completely dead. § An asterisk indicates that the level of mortality or injury was significantly different from the non-treated control according to Fisher's Protected LSD test (α = 0.05). Isolates were evaluated in four separate experiments.

Experiment 1: Effect of Bacterial Isolates on Taraxacum Officinale Seedling Mortality
None of the bacteria isolated from P. annua seed caused T. officinale seedling mortality (Table 1). A Staphylococcus sp. isolate from P. annua caused minor T. officinale seedling injury. This experiment indicates that bacteria isolated from non-surface-sterilized P. annua seed are not pathogenic to T. officinale seedlings.
While bacteria isolated from P. annua had limited or no effects on T. officinale, the most notable isolates from D. ischaemum (4, 5, 21, and 22) were Pantoea spp. that caused seedling mortality. Other isolates from D. ischaemum (8, 9, 10, 12, 21, 22, and 24) were not as antagonistic to T. officinale but still caused seedling injury and these were also Pantoea spp. while one isolate (18) was a Xanthomonas sp. These Pantoea spp. and Xanthomonas sp. were selected for futher study.
When these isolates were evaluated in a second experiment, isolates 5 and 8 caused the greatest T. officinale seedling mortality (Table 2). These isolates also caused injury to T. repens seedlings at 7 days after inoculation, and these seedlings likely would have completely died if observed 14 days after inoculation, as mortality of inoculated seedlings increased between 7 and 14 days after inoculation in subsequent experiments (reported herein). In previous research evaluating bacteria isolated from Phragmites australis, [20] demonstrated that Pseudomonas fluorescens strain Sandy LB4 caused >70% T. officinale seedling mortality while a Pantoea sp. strain RiLB4 did not increase mortality compared to the axenic control. In our experiments, Sandy LB4 nor RiLB4 increased T. officinale mortality compared to the bacteria free control.
Bacteria were observed in squash preparations of T. officinale root tips treated with isolates 4 and 5. Root hairs of seedlings treated with isolates 4 and 5 were malformed and the membrane at the tip of the root hair was often completely destroyed where bacteria were present ( Figure 1). Intact root hairs were observed in the axenic controls. Based on the results of this experiment, we selected isolates 4, 5, 8, 18, and 22 for further study. Percent mortality is the percentage of emerged seedlings that were determined by visual assessment to be completely dead. Seedlings were scored as injured if they displayed some injury symptoms but were not completely dead. § An asterisk indicates that this level of mortality was significantly different from the non-treated control according to Fisher's Protected LSD test (α = 0.05).

Experiment 2: Effect of Bacterial Isolates on D. Ischaemum Germination and Seedling Mortality
Isolate 22 and the combination of isolates 4 and 5 reduced D. ischaemum germination from 57% in the control to 33% and 20%, respectively (Table 3). Other isolates did not reduce germination compared to the bacteria free control. Isolates 8 and 22 increased D. ischaemum seedling mortality at 14 days after inoculation compared to the bacteria free control. Low germination prevented a proper assessment of the effect of the isolate 4 + 5 combination on D. ischaemum seedling mortality; these data were removed from the statistical analysis and are not presented. T. officinale was included to aid comparison to other experiments. All isolates except isolate 18 increased T. officinale mortality; a similar response was observed in Experiment 1. This indicates these bacteria are more pathogenic against competitor forbs than D. ischaemum. Based on the results of Experiments 1 and 2, we selected isolates 4 and 5 for further study. Table 3. Effect of Pantoea spp. isolates on Digitaria ischaemum germination and seedling mortality 14 and 28 days after inoculation in axenic agarose culture. Ten seeds were placed in each Petri dish and each treatment was replicated in three Petri dishes. Taraxacum officinale seeds were also inoculated in separate Petri dishes and included for comparison; mortality was evaluated at 21 days after treatment.  0  b  29  71  a  5  53  ab  6  ab  13  86  a  8  43  ab  18  a  43  78  a  18  43  ab  0  ab  7  33  b  22  33  bc  19  a  25  79  a  4+5  20  c  ---72  a  none  57  a  0  b  0  17 b † Percent mortality is the percentage of emerged seedlings that were determined by visual assessment to be completely dead. § Means followed by the same letter are not significantly different according to Fisher's Protected LSD test (α = 0.05). In columns where no letters are present, the ANOVA determined that the treatment effect was not significant (α = 0.05).

Experiment 3: Effect of Bacterial Isolates alone and in Combination with Curvularia sp. on D. ischaemum, T. repens and P. annua Seedling Mortality
The effect of experimental run was not significant; therefore, the data from two experimental runs were combined.

Experiment 2: Effect of Bacterial Isolates on D. ischaemum Germination and Seedling Mortality
Isolate 22 and the combination of isolates 4 and 5 reduced D. ischaemum germination from 57% in the control to 33% and 20%, respectively (Table 3). Other isolates did not reduce germination compared to the bacteria free control. Isolates 8 and 22 increased D. ischaemum seedling mortality at 14 days after inoculation compared to the bacteria free control. Low germination prevented a proper assessment of the effect of the isolate 4 + 5 combination on D. ischaemum seedling mortality; these data were removed from the statistical analysis and are not presented. T. officinale was included to aid comparison to other experiments. All isolates except isolate 18 increased T. officinale mortality; a similar response was observed in Experiment 1. This indicates these bacteria are more pathogenic against competitor forbs than D. ischaemum. Based on the results of Experiments 1 and 2, we selected isolates 4 and 5 for further study. Table 3. Effect of Pantoea spp. isolates on Digitaria ischaemum germination and seedling mortality 14 and 28 days after inoculation in axenic agarose culture. Ten seeds were placed in each Petri dish and each treatment was replicated in three Petri dishes. Taraxacum officinale seeds were also inoculated in separate Petri dishes and included for comparison; mortality was evaluated at 21 days after treatment. b † Percent mortality is the percentage of emerged seedlings that were determined by visual assessment to be completely dead. § Means followed by the same letter are not significantly different according to Fisher's Protected LSD test (α = 0.05). In columns where no letters are present, the ANOVA determined that the treatment effect was not significant (α = 0.05).

Experiment 3: Effect of Bacterial Isolates alone and in Combination with Curvularia sp. on D. ischaemum, T. repens and P. annua Seedling Mortality
The effect of experimental run was not significant; therefore, the data from two experimental runs were combined.
Germination of non-inoculated P. annua and T. repens was 93% and 83%, respectively and was not affected by bacteria or Curvularia treatment (data not presented). D. ischaemum germination was significantly reduced from 71% to 34% by Curvularia, but was not affected by bacterial treatment (data not presented).
Bacteria and Curvularia treatment affected T. repens seedling mortality 2 weeks after inoculation (Table 4). Bacterial isolates 4 and 5 alone or in combination with each other caused between 57 and 81% T. repens mortality compared to 0 and 4% for RiLB4 and the bacteria free control (Figure 2). When inoculated with Curvularia, seedling mortality was similar across all treatments (64 to 87%). Curvularia also increased D. ischaemum mortality regardless of bacteria treatment. When averaged across bacterial treatments, Curvularia increased D. ischaemum mortality from 11% to 73%. In the absence of Curvularia, according to the statistical analysis, isolates 4 and 5 alone or in combination did not affect D. ischaemum mortality compared to RiLB4 and the bacteria free control. However, a non-significant trend suggests isolates 4 and 5 inoculated alone increase D. ischaemum seedling mortality, but seedlings are not affected by the mixture of isolates 4 and 5. The bacteria treatments did not affect P. annua mortality (data not presented). Table 4. Effect of Pantoea ananatis isolates 4 and 5 alone and in combination with Curvularia sp. spores outgrowing from surface-sterilized Digitaria ischaemum seeds on Trifolium repens seedling mortality 14 days after inoculation in axenic agarose culture. Ten seeds were placed in each Petri dish and each treatment was replicated in three Petri dishes per experimental run. Combined results of two experimental runs are presented. A Pantoea sp. (RiLB4; Genbank No. KX752781) isolated from Phragmites australis and was included as a standard.

Bacterial Isolate
Curvularia  4  yes  73  a  77  a  5  yes  87  a  63  a  4 + 5  yes  87  a  68  a  RiLB4  yes  64  a  77  a  none  yes  73  a  83 a † Percent mortality is the percentage of emerged seedlings that were determined by visual assessment to be completely dead. § Means followed by the same letter are not significantly different according to Fisher's Protected LSD test (α = 0.05).

Observation of Bacteria in Seedling Roots
T. repens and T. officinale root hairs stained with DAB after inoculation with isolate 5 were often malformed and bacteria were observed around root hairs but not usually in the cytoplasm; observations of root hairs treated with isolate 4 were similar except that L-forms of bacteria were Curvularia increased P. annua mortality from 0 to 6% when averaged across all bacterial treatments, although this difference may not be biologically significant. Curvularia sp. treatment increased the positive gravitropic response of P. annua roots. When averaged across bacterial treatments, 57% of roots inoculated with Curvularia demonstrated a positive gravitropic response compared to 21% of roots without Curvularia. Bacteria treatments did not affect root gravitropism (data not presented). Curvularia nor bacteria treatments affected P. annua root length (data not presented).

Observation of Bacteria in Seedling Roots
T. repens and T. officinale root hairs stained with DAB after inoculation with isolate 5 were often malformed and bacteria were observed around root hairs but not usually in the cytoplasm; observations of root hairs treated with isolate 4 were similar except that L-forms of bacteria were occasionally observed in root hair cytoplasm (Figure 1). In every observation (n > 20) of DAB stained roots of seeds inoculated with bacteria, large numbers of bacterial rods were present in the rhizosphere, while no bacteria were observed in the rhizosphere of axenic controls. Conclusions from microscopic observations are limited, and the mechanism by which these bacteria kill T. repens and T. officinale should be investigated in more detail.

Discussion
Our experiments demonstrate that D. ischaemum seed contains cultivable bacteria of the Pantoea genus and at least two cultivable fungi in genera Epicoccum and Curvularia. Certain Pantoea ananatis isolates caused more seedling mortality of competitor forbs T. repens and T. officinale than of D. ischaemum in axenic culture, indicating these isolates selectively antagonize forbs. This suggests that D. ischaemum seed harbors microbes that may antagonize competitor forbs. However, a natural environment in consortia with other microbes would likely affect the antagonistic characteristic of these microbes.
Pantoea spp. have been isolated as epiphytes from the phyllosphere of other plant species including Oryza sativa and Phragmites australis and can function as growth promoters, plant pathogens, and disease suppressors [20,39,40]. P. ananatis is often studied as a plant pathogen, but endophyte isolates can function as growth promoters in pepper and switchgrass (Panicum virgatum L.) [41][42][43][44]. Non-pathogenic P. ananatis endophyte isolates have also been found in sugarcane and Miscanthus x giganteus [45,46]. Most of these P. ananatis endophytes were found in root and shoot tissue. However, strains have been isolated from seedlings of surface-sterilized maize kernals [28]. Others investigated the function of three P. ananatis endophytes of maize seed and found one isolate was weakly pathogenic and other growth-promotional [29]. Despite several investigations into growth promotion and pathogenicity of host plants, we are not aware of any research demonstrating that P. ananatis can kill competitor plants.
Research evaluating grass seed-associated bacteria as antagonists of competitor forbs is also limited. Others found that mixtures of certain bacteria associated with Phragmites australis seed increased mortality of T. officinale seedlings in axenic culture to a similar degree as bacteria evaluated in these experiments [20]. Previous research investigating deleterious rhizobacteria (for commercialization) evaluated bacterial strains further if they reduced root length of the target and non-target species by more than 30% and 10%, respectively, in Petri dish culture [47][48][49]. We did not measure forb root length in our experiments, but forb: D. ischaemum mortality ratios ranged from 2:1 to 5:1.
This research demonstrates that Pantoea spp. isolated from D. ischaemum seed have the capacity to antagonize seedlings of competitor forbs in axenic culture. Pantoea ananatis isolates consistently caused necrosis indicative of membrane leakage in T. repens and T. officinale cotyledons beginning about 7 days after inoculation and progressing until complete death was observed 14 to 21 days after inoculation. In our research, RiLB4 was included as a standard and did not increase mortality of T. repens and T. officinale while several Pantoea spp. isolates from D. ischaemum caused 50 to 80% mortality of T. repens and T. officinale seedlings. The effect of these isolates on D. ischaemum seedlings was inconsistent, but results indicate that they may increase seedling mortality. However, in some experiments, mortality was reduced when isolates 4 and 5 were combined, thus, this combination demonstrated the greatest amount of selectivity for the competitor forbs. It is also possible that removing the D. ischaemum seed coat (required to surface sterilize the seed) increased D. ischaemum susceptibility to these bacteria and reduced selectivity for the competitor forbs. Future research should evaluate whether these bacteria are less pathogenic when in consortium with one another or other uncultivable bacteria.
In our experiments, Curvularia functioned primarily as a pathogen of T. repens and D. ischaemum, although with a different combination of bacterial cohorts it may function as a symbiont [10]. Curvularia was much less pathogenic to P. annua than D. ischaemum and increased the positive gravitropic response of P. annua roots. To potentially avert the pathogenic effect against D. ischaemum, future research could use more mature seedlings, perhaps allowing plants to develop at least one true leaf before inoculating with this Curvularia sp. isolate to evaluate stress responses similar to [25] and [26]. Extensive observations of seeds inoculated with Curvularia sp. were not conducted in these experiments, but in every observation on multiple dates, segmented hyphae was observed colonizing intercellular spaces of inoculated P. annua and D. ischaemum root cortex tissue, which suggests it has endophytic capabilities ( Figure 3).    These experiments demonstrate that Digitaria ischaemum seeds contain bacteria that antagonize competitor species in Petri dish experiments. Pantoea ananatis may provide D. ischaemum a competitive advantage against competitor species. The effect of these isolates alone and in consortia on D. ischaemum as compared to competitor forbs should be investigated further. Future research could inoculate these bacteria onto D. ischaemum seeds to determine they affect competitiveness.
One limitation of this research is that seeds were collected from only one location for each plant species. Plant microbes likely vary across geographic sites and genotypes. Future research could investigate bacteria associated with P. annua and D. ischaemum seeds collected from different geographic and climatic regions using culture-dependent and culture-independent methods.
The mechanism of selective pathogenicity should also be elucidated. Cyanide production has been implicated as a mechanism by which rhizobacteria suppress plant growth [50], but preliminary experiments suggest that our Pantoea isolates do not produce cyanide. Interestingly, we consistently observed cotyledons with a water-soaked appearance prior to necrosis and death. Similar symptoms can be observed in corn plants affected by Pantoea stewartii subsp. stewartii, (the causal agent of Stewart's wilt) as a result of biofilms encased in an exopolycaccharide [51]. However, we did not observe exopolycaccharide blockages in the xylem tissue. The role of Large Pantoea Plasmids in the phenotype and selective forb control we observed should be investigated. These plasmids are common to many Pantoea spp. including P. ananatis are thought to be responsible for certain strains to occupy specific ecological niches with various phenotypes such as toxin production and viruluence factors that enable pathogenicity or biocontrol [52][53][54][55]. The proteins encoded by Large Pantoea Plasmids may shed light on the bacterial mechanism of action.