agronomy Crop Yield, Ferritin and Fe(II) boosted by Azospirillum brasilense (HM053) in Corn

: An increasing global population of over 4.5 billion people drives increasing demand for calories—30% of which are satisﬁed by grain crops, such as maize. High-density farming practices have been implemented but tend to deplete the soil of essential elements resulting in lower nutritional value, notably iron, of cultivated crops. Low iron content in staple crops can contribute over time to severe, even fatal, micronutrient deﬁciencies. Enhancing grain iron content using post-harvest biofortiﬁcation strategies can be costly. However, ﬁeld inoculation using biologics like Azospirillum brasilense (HM053) can be a cost-e ﬀ ective alternative to improving crop nutritional value. Using ion chromatography with chemiluminescence detection, we have shown that maize seeds harvested from outdoor pot-grown plants possessed a four-fold higher iron content as ferrous iron (Fe + 2 ) compared to non-inoculated plants. Seeds from A. brasilense HM053-inoculated plants also contained approximately 13 nmol of ferritin per ground dried weight of kernel compared to 3 nmol from non-inoculated plants. In addition, A. brasilense HM053 inoculation increased crop yield 30–50% relative to non-inoculated plants.


Introduction
Iron (Fe) is a necessary nutrient for life, helping form red blood cells and facilitate oxygen transfer from the lungs to other tissues. The recommended daily intake of iron for adults is at least 8 mg for men and 18 mg for women [1]. Iron deficiency is a common malnutrition disorder, causing approximately 50% of all diagnosed anemia cases and affecting more than two billion people worldwide [2,3]. Even mild and moderate iron deficiency can result in impaired human functioning [4].
The various forms of iron present in food differ in bioavailability to humans [5]. Nutritional iron is available from animal sources as heme iron or plant sources as non-heme iron. Heme iron, prevalent in meat and seafood, comes from the center of hemoglobin and myoglobin proteins and has the greatest Agronomy 2020, 10, 394 3 of 12

Plant Growth
For courtyard studies, 3 maize kernels from Elk Mound Seed Co. (Hybrid 8100) were sown into each of 2.7 gallon pots filled with ProMix. After germination, any excess seedlings were removed from each pot leaving a single plant. A capful of fertilizer (~1.2 g) containing nitrogen, phosphate and potash (14-14-14, Osmocote™ Smart-Release Plant Food Flower & Vegetable™, The Scotts Company, Marysville, OH) was added to the assigned pots at the time of planting. Fertilizer was reapplied to pots 30 days after germination (DAG) and 60 DAG. Plants were placed on elevated tables outside (Figure 1). During the 2018 growing season, four study regimes included the following: (i) non-inoculated control plants; (ii) plants inoculated with A. brasilense HM053 bacterium; (iii) plants inoculated with A. brasilense ipdC bacterium; and (iv) plants treated with 30 µM auxin plant hormone (IAA). Plants were administered liquid inoculants at 21, 42 and 63 DAG. Treatments with auxin hormone followed the same schedule. Treatments were randomized across the planting platforms and pots were rotated weekly during the growing season. The study was repeated in the 2019 growing season with the exception that auxin treatments were not applied.

Plant Growth
For courtyard studies, 3 maize kernels from Elk Mound Seed Co. (Hybrid 8100) were sown into each of 2.7 gallon pots filled with ProMix. After germination, any excess seedlings were removed from each pot leaving a single plant. A capful of fertilizer (~ 1.2 g) containing nitrogen, phosphate and potash (14-14-14, Osmocote™ Smart-Release Plant Food Flower & Vegetable™, The Scotts Company, Marysville, OH) was added to the assigned pots at the time of planting. Fertilizer was reapplied to pots 30 days after germination (DAG) and 60 DAG. Plants were placed on elevated tables outside (Figure 1). During the 2018 growing season, four study regimes included the following: (i) non-inoculated control plants; (ii) plants inoculated with A. brasilense HM053 bacterium; (iii) plants inoculated with A. brasilense ipdC bacterium; and (iv) plants treated with 30 µM auxin plant hormone (IAA). Plants were administered liquid inoculants at 21, 42 and 63 DAG. Treatments with auxin hormone followed the same schedule. Treatments were randomized across the planting platforms and pots were rotated weekly during the growing season. The study was repeated in the 2019 growing season with the exception that auxin treatments were not applied.

Bacteria Growth
Two mutant strains of A. brasilense (ipdC and HM053) were grown in liquid NFbHP-lactate medium following published procedures [40]. The medium contained 20 mM ammonium chloride as a nitrogen source and streptomycin antibiotic selection (80 µg mL −1 ). The bacterial cultures were grown in 25 mL flasks in a shaking incubator set to 30 °C and 130 rpm until OD600 = 1.0 (Optical Density at 600 nm, corresponding to 10 8 cells mL −1 ) is reached. Bacterial cultures were washed with sterile water and diluted to approximately 10 6 to 10 8 colony forming units (CFU). Bacterial culture (10 mL) was injected into the soil at the base of the stem at the specified 3 time points during the growing season. Verification of bacteria presence at harvest was not carried out.

Bacteria Growth
Two mutant strains of A. brasilense (ipdC and HM053) were grown in liquid NFbHP-lactate medium following published procedures [40]. The medium contained 20 mM ammonium chloride as a nitrogen source and streptomycin antibiotic selection (80 µg mL −1 ). The bacterial cultures were grown in 25 mL flasks in a shaking incubator set to 30 • C and 130 rpm until OD 600 = 1.0 (Optical Density at 600 nm, corresponding to 10 8 cells mL −1 ) is reached. Bacterial cultures were washed with sterile water and diluted to approximately 10 6 to 10 8 colony forming units (CFU). Bacterial culture (10 mL) was injected into the soil at the base of the stem at the specified 3 time points during the growing season. Verification of bacteria presence at harvest was not carried out.

Plant Growth Measurements
Maize plants were measured with a tape measure from the soil of the pot up to the highest point of the plant in centimeters. They were measured at three time points: 17, 28, and 55 days after germination. Stem diameters were measured at the second internode up from the base of the stem using calibers. Tape was applied at these stem locations after the first time point measurement which served as a visual marker for subsequent measurements.

Chlorophyll and Leaf Thickness Measurements
Vegetative tissues were separated from roots and placed in pre-weighed and pre-chilled Eppendorf™ tubes. They were flash frozen in liquid nitrogen, ground to a fine powder and weighed. A volume of acetone equivalent to four times the milligram mass of the frozen ground tissue (assuming uniform densities) was added to the centrifuge tube. The content was vortexed (VWR analog vortex mixer; Sigma-Aldrich Corp., St. Louis, MO, USA) then sonicated (Branson Bransonic 32; Sigma-Aldrich Corp. St. Louis, MO, USA) at 0 • C for 2 minutes. Additional vortexing was performed during this period to ensure complete mixing. The Eppendorf™ tubes were then centrifuged for 2 minutes at 15,000 rpm to separate the insoluble and soluble portions. The insoluble portion contained mostly cell-wall polymers and starch while the soluble portion contained small soluble compounds such as sugars. The liquid extract was spotted onto a silica TLC plate in a 5 mm band using 2 µL of sample. The silica plate was developed in a 6:1.6:1:0.4 ratio of petroleum ether: hexane: acetone: methanol solution. Once the plate was fully developed, a razor was used to scrape off bands containing chlorophyll a and chlorophyll b. The shavings were placed in two separate Eppendorf™ tubes. Acetone (1 mL) was added to the tube containing chlorophyll a and the tube containing chlorophyll b. The samples were then vortexed and centrifuged for 4 minutes at 15,000 rpm. The absorbance of the resulting supernatant of each sample was analyzed via ultraviolet-visible spectroscopy (UV-VIS). Chlorophyll a was recorded at a wavelength of 663 nm, and Chlorophyll b was measured at a wavelength of 645 nm. Beer's Law (A =ε bc) was then utilized to determine sample concentration of chlorophyll. Precision calipers (Fowler High Precision, Newton, MA, USA) were used to determine leaf thickness.

Ion Chromatography Measuring Seed Iron
Ferric, ferrous, and total iron were quantified from corn kernels with ion chromatography coupled with chemiluminescence detection [41] following the collection and drying of the kernels in an oven for 3 weeks at 65 • C. Seeds were pulverized between plastic sheets using a wooden mallet and dissolved in 1 mL of 1M HCl. Samples were subjected to ultrasonication for 5 minutes at 100% amplitude (Branson Bransonic 32; Sigma-Aldrich Corp. St. Louis, MO, USA) then centrifuged for 15 minutes at 3000 rpm. The supernatant was removed for sampling and stored in brown glass vials in a refrigerator (2-8 • C).
Stability studies were completed for the samples, varying time between sample extraction to injection on the ion chromatograph as well as storage conditions of temperature, light, and glass versus plastic storage containers. It was found that plastic Eppendorf tubes were not effective for long term sample storage, and neither was storage at room temperature. Under these circumstances, the oxidation state of iron was observed to change quickly over a short time (< 2 h). Storing samples in brown glass vials placed in the refrigerator allowed sample stability (no change in iron oxidation states) for up to 6 hours.
Iron standards were made in 0.1M HCl. Ferric chloride (FeCl 3 , 1 mg mL −1 ) was used as a ferric iron standard and ferrous sulfate hexahydrate (FeSO 4 •6H 2 O, 1 mg mL −1 ) was used as the ferrous iron standard. The ferric chloride sample was then diluted 50-fold for standard curve generation, while the ferrous sulfate was diluted 10-fold. On average, it was observed that 3% of the ferric standard was reduced to ferrous iron when measured with ion chromatography chemiluminescence likely due to a solvent effect.
Agronomy 2020, 10, 394 5 of 12 The analytical system consisted of a Thermo Scientific Dionex AXP Metal-Free HPLC with a Rheodyne metal-free injector and PEEK tubing 1/20 cm inner diameter. The ion chromatography column was a Thermo Scientific IonPac CS5A (4 × 50 mm guard column and 4 × 250 mm analytical column) designed to separate a broad range of metal complexes by cation and anion chromatography. The outlet of the column was connected to a mixing tee, where the post-column reagents were introduced with the column effluent to elicit the chemilumiescent reaction. Chemiluminescence was measured with the HERM data monitor (HERM LB 500, Berthold Technologies LLC, Oak Ridge, TN, USA) to allow quantification of iron in its oxidation states. Ferric iron had a retention time of approximately 4.3 minutes, while ferrous iron had a retention time of approximately 8.2 minutes.
In the analytical system, sterile water (HyPure™ WFI Quality Water, HyClone Laboratories, Logan, UT, USA) was used in solvent preparation to avoid introduction of additional iron to the samples. The column was rinsed with the WFI Quality Water, followed by flowing 1 M Na 2 SO 3 at 0.5 mL min −1 for 45 minutes to remove any dissolved oxygen on the column. During this time, the pre-column eluent (10 mM 2,4-pyridinedicarboxylic acid [PDCA], 80 mM potassium chloride, 37 mM formic acid, and 33 mM potassium hydroxide pH adjusted to 4) was bubbled with argon gas to remove dissolved oxygen. The pre-and post-column reagents were flowed through the system at rates of 1.00 and 0.75 mL min −1 , respectively. The post-column reagents consisted of 600 mM hydrogen peroxide and 0.025 mM luminol with 50 Mm sodium carbonate pH adjusted to 11.5 [41]. Standard curves were generated daily prior to sample injection and samples were injected in triplicate.

Size-Exclusion Chromatography used in Seed Ferritin Analysis
A stainless steel High-Performance Liquid Chromatography (HPLC) Sonntek Pump was connected to a Rheodyne injector with a 20 µL PEEK injection loop to separate biological metabolites on an Agilent Zorbax GF-250 4 µm size exclusion column (4.6 × 250 mm). Samples were quantified on an Azura UVD 2.15 Knauer UV detector set to 254 nm [42]. Mobile phase was pumped at 0.5 mL min −1 and consisted of 0.2 M ammonium sulfate with 0.05 M tris-HCl buffer at pH of 7. An antimicrobial was added (0.005% sodium azide) to maintain column integrity throughout the study. Maize kernels were acquired from the courtyard grown maize harvest and analyzed for the amount of ferritin across treatment types previously described. Dried kernels (~0.5 g) were pulverized between plastic sheets using a wooden mallet and placed in 1 mL of 50 mM tris-HCl buffer (pH 7) and ultrasonicated for two minutes at 100% amplitude. The sample was then heat treated at 75 • C for 10 minutes followed by immediate cooling on ice for 15 minutes. Centrifugation 15,000 g for 30 minutes pulled large particulates out of the supernatant which was injected (20 µL) for quantification against a ferritin standard (MyBioSource, Inc., San Diego, CA, USA).

Statistical Analysis
Data was subjected to the Shapiro-Wilk Normality Test, which was used to identify outliers so that all data groups represented a normal distribution. Data was then analyzed using the Student t-test comparing individual treatments to non-inoculated plants as the control. A level of statistical significance was set at p < 0.05.

Results and Discussion
An outdoor study using potted plants enabled assessment of crop fitness to maturity and nutritional iron content in the kernel as a function of A. brasilense and chemical auxin treatments. Crop fitness during vegetative growth was measured in the 2019 growth season by changes in plant height and stem thickness (Figure 2). Treatments appeared not to affect rates of plant growth as measured by plant height, but differences were seen in the thickness of the plant stems when measured at the second internode. Measurements taken on days 28 and 55 after germination showed a statistically significant increase in thickness from A. brasilense HM053-inoculated maize relative to non-inoculated control while other treatments had no effect. In support of improved plant growth from A. brasilense HM053 inoculation, we observed that chlorophyll content when measured in 3-week-old lab grown seedlings ( Figure 3A) was significantly elevated from bacterial treatment relative to non-inoculated control plants while plants inoculated with A. brasilense ipdC showed a reduced leaf chlorophyll content. Though chlorophyll content was not measured in the outdoor potted plant studies, we do note that leaf thicknesses measured at the end of the 2019 growing season ( Figure 3B) were significantly greater with A. brasilense HM053 inoculation than non-inoculated controls. Our data suggests that the effects of A. brasilense HM053 inoculation on leaf chlorophyll in the early stages of plant growth must persist throughout the growing season since we observed improved plant growth reflected by increased stem and leaf thickness.
Agronomy 2020, 10, x FOR PEER REVIEW 6 of 12 relative to non-inoculated control plants while plants inoculated with A. brasilense ipdC showed a reduced leaf chlorophyll content. Though chlorophyll content was not measured in the outdoor potted plant studies, we do note that leaf thicknesses measured at the end of the 2019 growing season ( Figure 3B) were significantly greater with A. brasilense HM053 inoculation than non-inoculated controls. Our data suggests that the effects of A. brasilense HM053 inoculation on leaf chlorophyll in the early stages of plant growth must persist throughout the growing season since we observed improved plant growth reflected by increased stem and leaf thickness. based on a comparison of growth rate slopes derived from linear regression analysis. Stem diameters were measured in centimeters at the second internode from the base of the stem. A significant increase in stem thickness was recorded for A. brasilense HM053 relative to non-inoculated control maize. Data (±SE) reflects N = 12 biological replicates. Asterisks indicate significant differences (*p < 0.05) using Student paired t-test comparing treated to non-treated control plant stems when measured at the second internode, in which, at 28 and 55 days after germination, A. brasilense HM053-inoculated maize exhibited a significantly increased stem thickness relative to non-inoculated control, suggesting the biologic improved plant growth. Stem diameters were measured in centimeters at the second internode from the base of the stem. A significant increase in stem thickness was recorded for A. brasilense HM053 relative to non-inoculated control maize. Data (±SE) reflects N = 12 biological replicates. Asterisks indicate significant differences (*p < 0.05) using Student paired t-test comparing treated to non-treated control plant stems when measured at the second internode, in which, at 28 and 55 days after germination, A. brasilense HM053-inoculated maize exhibited a significantly increased stem thickness relative to non-inoculated control, suggesting the biologic improved plant growth.  Finally, crop seed yield was measured as the total kernels of corn per cob (Figure 4). It was found that maize associated with A. brasilense HM053 had significantly higher crop yields compared to non-inoculated control plants (34% and 53% increases for A. brasilense HM053-inoculated plants from the 2018 and 2019 growing seasons). There was also a 26% increase in crop yield from auxin treatment in the 2018 growing season, while A. brasilense ipdC-associated maize exhibited a 26% lower yield than non-treated plants, though not significant, during the 2018 growing season and no difference from non-treated plants in the 2019 growing season (Figure 4). The differences in yield between the 2018 and 2019 growing seasons may be attributed to differences in the outdoor environmental conditions. While plants were irrigated on an alternating day cycle, there was increased rainfall during the early portions of the 2019 growing season which could have contributed to the yield improvement. Finally, crop seed yield was measured as the total kernels of corn per cob (Figure 4). It was found that maize associated with A. brasilense HM053 had significantly higher crop yields compared to non-inoculated control plants (34% and 53% increases for A. brasilense HM053-inoculated plants from the 2018 and 2019 growing seasons). There was also a 26% increase in crop yield from auxin treatment in the 2018 growing season, while A. brasilense ipdC-associated maize exhibited a 26% lower yield than non-treated plants, though not significant, during the 2018 growing season and no difference from non-treated plants in the 2019 growing season (Figure 4). The differences in yield between the 2018 and 2019 growing seasons may be attributed to differences in the outdoor environmental conditions. While plants were irrigated on an alternating day cycle, there was increased rainfall during the early portions of the 2019 growing season which could have contributed to the yield improvement.
cycle of the plant. Furthermore, Fe 3+ showed no change as a function of inoculation. Both A. brasilense HM053 and ipdC data was loosely clustered at opposite ends of iron values measured in seeds (30-130 nmols iron), whereas non-inoculated control data spanned the entire range of values. Recent studies in tomato indicate early treatments with inoculums can reshape the diversity of the soil microbiome favoring dominance of the inoculum whereas non-inoculated plants can exhibit a broad spectrum of microbiome diversity [43]. Error bars indicate ±SE. Paired comparisons were made using the Student t-tests between single treatments and non-inoculated control plants. Asterisks indicate significant differences (**p < 0.01, ***p < 0.001) in these comparisons.
All kernels associated with A. brasilense HM053 or ipdC, as well as those treated with auxin, showed significant increases in kernel ferritin content ( Figure 6). This corroborates the assertion that PGPR association improves iron integration to seed, ultimately increasing its nutritional value. The auxin treatment increasing ferritin content alludes to the importance of the auxin production capabilities of the bacteria in influencing this effect, which requires further investigation. However, this does not explain the observed increase in seed ferritin content from inoculation with the auxin deficient-mutant, A. brasilense ipdC. Past perceptions have assumed that ferritin biosynthesis is tightly regulated by cellular iron homeostasis [14]. Even so, the present results suggest that its regulation is even more complex.
Finally, while chemical treatments using auxin resulted in similar improvements in crop yield and crop nutritional value as A. brasilense HM053 inoculation relative to untreated control plants, we view chemical treatments as more intrusive to the environment, and likely a more expensive mitigation strategy than application using a liquid inoculant of bacteria. An increase in maize crop yield was noted under A. brasilense HM053 inoculation and auxin treatments; however, the main purpose of leveraging bacterial inoculants in this study was to demonstrate a cost-effective means to increase iron nutrition in the crop product consumed by humans. Thus, it was most relevant to quantify the iron content of seeds as a function of treatment, as well as quantify the bioavailability of that iron bound in the ferritin protein. Ion chromatography was used to quantify total seed iron content as well as distinguish Fe 2+ and Fe 3+ oxidation states ( Figure 5). The results revealed a strong positive correlation between seed Fe 2+ accumulation and total seed iron content, with A. brasilense HM053 generating the highest iron values. This suggests inoculation in the early stage of plant development can result in a lasting effect throughout the life cycle of the plant. Furthermore, Fe 3+ showed no change as a function of inoculation. Both A. brasilense HM053 and ipdC data was loosely clustered at opposite ends of iron values measured in seeds (30-130 nmols iron), whereas non-inoculated control data spanned the entire range of values. Recent studies in tomato indicate early treatments with inoculums can reshape the diversity of the soil microbiome favoring dominance of the inoculum whereas non-inoculated plants can exhibit a broad spectrum of microbiome diversity [43].
All kernels associated with A. brasilense HM053 or ipdC, as well as those treated with auxin, showed significant increases in kernel ferritin content ( Figure 6). This corroborates the assertion that PGPR association improves iron integration to seed, ultimately increasing its nutritional value. The auxin treatment increasing ferritin content alludes to the importance of the auxin production capabilities of the bacteria in influencing this effect, which requires further investigation. However, this does not explain the observed increase in seed ferritin content from inoculation with the auxin deficient-mutant, A. brasilense ipdC. Past perceptions have assumed that ferritin biosynthesis is tightly regulated by cellular iron homeostasis [14]. Even so, the present results suggest that its regulation is even more complex.
Agronomy 2020, 10, 394 9 of 12 Agronomy 2020, 10, x FOR PEER REVIEW 9 of 12 Figure 5. Seed iron content as total iron, ferric iron, and ferrous iron as a function of treatment type. Panel A: data on seed ferrous (Fe 2+ ) concentration relative to total iron from the 2018 harvest. There was a positive correlation across all treatment types between ferrous iron and total iron content in maize seed. Panel B: data on seed ferric iron (Fe 3+ ) concentration relative to the total iron. There is no strong correlation between changing total iron content and ferric iron in maize seed kernels. Figure 6. Seed iron-ferritin content as a function of treatment type. All treatment types showed significantly greater concentration of ferritin protein in maize seed relative to control with greatest elevation observed in the A. brasilense HM053-inoculated maize kernels. Data reflects N = 4 biological replicates for each treatment with error bars indicating ±SE. Paired comparisons were made using the Student t-test between single treatments and non-inoculated control plants. Asterisks indicate significant differences (**p < 0.01, ***p < 0.001) in these comparisons. Figure 5. Seed iron content as total iron, ferric iron, and ferrous iron as a function of treatment type. Panel A: data on seed ferrous (Fe 2+ ) concentration relative to total iron from the 2018 harvest. There was a positive correlation across all treatment types between ferrous iron and total iron content in maize seed. Panel B: data on seed ferric iron (Fe 3+ ) concentration relative to the total iron. There is no strong correlation between changing total iron content and ferric iron in maize seed kernels.
Agronomy 2020, 10, x FOR PEER REVIEW 9 of 12 Figure 5. Seed iron content as total iron, ferric iron, and ferrous iron as a function of treatment type. Panel A: data on seed ferrous (Fe 2+ ) concentration relative to total iron from the 2018 harvest. There was a positive correlation across all treatment types between ferrous iron and total iron content in maize seed. Panel B: data on seed ferric iron (Fe 3+ ) concentration relative to the total iron. There is no strong correlation between changing total iron content and ferric iron in maize seed kernels. Figure 6. Seed iron-ferritin content as a function of treatment type. All treatment types showed significantly greater concentration of ferritin protein in maize seed relative to control with greatest elevation observed in the A. brasilense HM053-inoculated maize kernels. Data reflects N = 4 biological replicates for each treatment with error bars indicating ±SE. Paired comparisons were made using the Student t-test between single treatments and non-inoculated control plants. Asterisks indicate significant differences (**p < 0.01, ***p < 0.001) in these comparisons. Figure 6. Seed iron-ferritin content as a function of treatment type. All treatment types showed significantly greater concentration of ferritin protein in maize seed relative to control with greatest elevation observed in the A. brasilense HM053-inoculated maize kernels. Data reflects N = 4 biological replicates for each treatment with error bars indicating ±SE. Paired comparisons were made using the Student t-test between single treatments and non-inoculated control plants. Asterisks indicate significant differences (**p < 0.01, ***p < 0.001) in these comparisons.
Finally, while chemical treatments using auxin resulted in similar improvements in crop yield and crop nutritional value as A. brasilense HM053 inoculation relative to untreated control plants, we view chemical treatments as more intrusive to the environment, and likely a more expensive mitigation strategy than application using a liquid inoculant of bacteria.

Conclusions
The present study has demonstrated that certain PGPR can have beneficial effects in improving crop nutritional value. Specifically, A. brasilense HM053 a hyper N 2 -fixing and high auxin producing strain of A. brasilense provided the greatest benefits in improving seed iron content with improved iron bioavailability when stored in ferritin. It is also noted that this bacteria strain improved crop yields significantly, possibly as a result of host iron assimilation during the growing season impacting carbon input via the photosynthetic machinery. Future studies will seek to translate our findings from pots to the field.