Collaboration between grass seedlings and rhizobacteria to scavenge organic nitrogen in soils

This article describes a process (termed ‘oxidative nitrogen scavenging’) where grasses scavenge organic nitrogen from microbes on and around roots. The authors propose a diurnal process where during the day roots produce and release hydrogen peroxide that oxidizes microbial exoenzymes around roots; at night hydrogen peroxide production ceases, then roots and symbiotic rhizobacteria secrete proteases that degrade the oxidized proteins to form peptides that are absorbed by roots. The existence of a mechanism for organic nitrogen scavenging in grasses emphasizes the nutritional importance of non-pathogenic microbes that associate with roots. Future applications of this process could result in new methods for the cultivation of crop plants.


Introduction
All organisms require nitrogen (N) throughout development to make proteins, nucleic acids and other biological molecules. For plant growth, N is often the limiting nutrient. It has been believed that N available to plants was exclusively in the inorganic nitrate form. However, it has become clear that plants may obtain N from organic sources (Godlewski and Adamczyk 2007;Paungfoo-Lonhienne et al. 2008Adamczyk et al. 2010). The need to push crop plants to higher yields has provoked heavy reliance on inorganic N fertilizers (Herridge et al. 2008). It has been estimated that only 30 -50 % of applied inorganic N fertilizer is actually taken up by crop plants, with the rest lost in run-off, leaching or volatilization, increasing the probability for unintended environmental pollution (Mulvaney et al. 2009).
Agricultural scientists and plant breeders have targeted improvements in N use efficiency of plants as a way to reduce reliance on N applications (Garnett et al. 2009). Grasses associate with arbuscular mycorrhizae that function to absorb phosphorus and N from soils. Years of investigation on mycorrhizae have yielded volumes of information about the biology of mycorrhizalplant associations (Gutjahr and Parniske 2013). However, comparatively little is known about other types of microbial endophytes or epiphytes that associate with healthy plants (Schulz and Boyle 2005). White et al. (2012) reported the occurrence of seed-vectored bacteria on many cool-season grasses. The bacteria, vectored on seed surfaces attached to adherent lemmas and paleas, rapidly colonize the roots of the seedlings after germination. The frequency of root colonizing bacteria on seeds of grasses suggested that bacteria could be involved in a nutritional symbiosis with plants; however, insufficient information was available to delineate clearly how the rhizobacteria benefited plants. In this paper we report studies conducted to evaluate the role of seedvectored rhizobacteria in the nutrition of cool-season grass seedlings. In addition, we examine the role of grass root-secreted reactive oxygen in the process of protein degradation around roots. Some of the questions we pursued in this research include the following: Is there a survival/nutritional function for the seed-vectored bacteria on grass seeds? Do rhizobacteria facilitate organic N acquisition by grass seedlings? Does reactive oxygen secretion from grass roots relate to the function of seedvectored bacteria?

Plant materials
Seeds used were selections of tall fescue (Festuca arundinacea Schreb.), perennial ryegrass (Lolium perenne L.) and annual bluegrass (Poa annua L.) obtained from the Rutgers University turf grass-breeding programme.

Bacteria isolation and identification
To isolate bacteria, 40 seeds of each grass species were washed for 5 min with constant agitation in sterile water, then plated onto yeast extract-sucrose agar (YES) and 10 % tryptic soy agar (TSA; Difco). YES agar contained 1 % yeast extract (Difco), 1 % sucrose (Sigma-Aldrich, St Louis, MO, USA) and 1.2 % agar (Difco). Twenty seeds were plated on each medium (five seeds per plate) and incubated 1 week at laboratory ambient temperature. The most common colonies appearing were selected for sequencing of the 16S rDNA region. Bacterial DNA was extracted using the UltraClean w Microbial DNA Isolation Kit (MoBio Laboratories, Carlsbad, CA, USA). The 16S rRNA gene was amplified using the universal primers 27F (5 ′ -AGAGTTTGATCMTGGCTCAG) and 1492R (5 ′ -CGGTTACC TTGTTACGACTT). Amplicons were sent to be sequenced by Genewiz (South Plainfield, NJ, USA). High-quality sequences were selected and bacteria were identified by searching for sequence similarity using BLAST (NCBI GenBank).

Removal of bacteria from seeds
To conduct seedling development studies, seeds free of bacteria were prepared by surface disinfection. This was accomplished by agitating seeds on a rotary shaker in a 3 % sodium hypochlorite solution for 35 -40 min, after which seeds were rinsed twice for 5 min in sterile water. We optimized the surface disinfection conditions so that seed germination frequency was not reduced but rhizobacteria no longer could be seen on or isolated from resulting seedlings.

Seedling development experiments
Seeds with and without native bacterial populations were placed onto plates. For each treatment 8 plates per treatment were used with 10 -15 seeds on each plate. Plates were incubated for 10 days at laboratory ambient temperature in a 12-h alternating light/dark cycle. Seedlings were examined and assessed for the percentage of seedlings with roots growing downward into agarose and whether downward growing roots showed development of root hairs (Tables 1 -3). Any inhibition of root growth was also noted.

Microscopic examination of roots
For microscopic examinations (Figs 1 -3), roots were stained for 8-10 h by flooding plates with 2.5 mM diaminobenzidine tetrachloride (DAB) and 5 purpurogallin units mL 21 of horseradish peroxidase (Type VI, Sigma-Aldrich) (White et al. 2012(White et al. , 2014b. Stain was then poured off and plates rinsed several times with water. Seedling roots were examined microscopically using a compound light microscope through the reverse side of plates or by removing seedlings from agarose and placing in a drop of aniline blue/lacto-phenol stain on a slide (White et al. 2012(White et al. , 2014b. For the evaluation of root effects on protein particles, examination was made only through the reverse of plates. DAB/horseradish peroxidase was used to stain H 2 O 2 (red to brown colour); while aniline blue/lactophenol was used to stain proteins and bacterial cytoplasm (light blue to dense blue colour).

Production of 15 N-labelled protein mixture
To label bacterial proteins, Bacillus amyloliquefaciens (strain HF3) was inoculated into 15 N-labelling medium composed of 1 % sucrose, Murashige and Skoogs salt base without nitrogen (Sigma-Aldrich), 0.02 % yeast extract and 0.16 % 15 N-labelled glycine (Sigma-Aldrich) in 200 mL Erlenmeyer flasks. Flasks were then incubated at laboratory ambient temperature with constant agitation on a rotary shaker for 15 days. Cells were harvested from cultures by centrifugation and washed twice with phosphate-buffered saline solution. To extract proteins, cells were suspended in ice-cold acetone for 5 min, washed under nitrogen and then incubated in 1 % sodium dodecyl sulfate for 2 min. Extracted proteins were then freeze-dried to remove water; 0.16 g of labelled proteins was obtained. The labelled proteins were thoroughly mixed with 0.84 g of un-labelled egg albumin (Sigma-Aldrich) prior to use in media.

N-labelled protein absorption experiment
Plastic Petri plates (6 cm diameter) containing either 0.7 % agarose or 0.7 % agarose with 0.05 % of the labelled-protein mixture were prepared. The labelledprotein mixture was added to agarose after autoclaving with the agarose at 70 8C and rapidly cooled to minimize additional protein denaturation. Twelve tall fescue seeds, either with seed-vectored rhizobacteria present or with rhizobacteria removed by surface disinfection as previously described, were placed onto plates of both media.
The following treatments were prepared: (i) no protein in medium and seeds with rhizobacteria in 12-h alternating light/dark cycles; (ii) no protein in medium and no rhizobacteria on seeds in 12-h alternating light/dark cycles; (iii) protein in medium and rhizobacteria on seeds in 12-h alternating light/dark cycles; (iv) protein in medium but seeds without rhizobacteria on seeds in 12-h alternating light/dark cycles; (v) no protein in medium and no rhizobacteria on seeds maintained in total darkness; and (vi) proteins in medium but seeds without rhizobacteria and maintained in total darkness. Three to six replicate plates were made for each treatment (Fig. 4). Plates in dark treatments were wrapped in aluminium foil to exclude light. All plates were placed in a clear plastic sealable canister and placed under fluorescent lights at laboratory ambient temperature for a 10-day period.
After the incubation period, the shoots were excised from seedlings, with all shoots on a plate being combined and dried in an oven at 60 8C overnight. After drying, samples were further processed and analysed for 15 N/ 14 N ratio using mass spectroscopy at the Stable Isotope Laboratory at the Odum School of Ecology at the University of Georgia, Athens, GA, USA. Results were expressed as d 15 N vs. air measurements (Fig. 4). Statistical analysis of data involved application of the Duncan's multiple range test (P , 0.05).

Seedling development experiments
Seedlings that possessed the rhizobacteria showed a majority of the seedling roots to be downward oriented with abundant root hair growth (Tables 1-3). Removal of rhizobacteria from surfaces of seed coats resulted in anomalous root development on agarose without nutrients. Without bacteria, the majority of the seedling roots grew on the surface of the agarose, and any roots that grew downward into the agarose lacked root hair development. Substances that were found to consistently restore normal root development in bacterial-free seedlings included yeast extract and various proteins (Tables 1-3).
In tall fescue the amino acid glycine was seen to restore seedling development, but this was not seen in the other grasses (Tables 1-3). When rhizobacteria were present on seedlings, protein levels that stimulated proper seedling root development in seedlings without rhizobacteria were  Table 2). The addition of the reactive oxygen scavenger sodium ascorbate to agarose-containing proteins was seen to suppress the capacity of the proteins to trigger proper seedling root development in the absence of rhizobacteria (Table 1). The addition of ammonium nitrate to agarose was not sufficient to restore proper seedling root development in the absence of rhizobacteria (Tables 2 and 3).

Visualization of rhizobacteria and proteins on roots
Rhizobacteria could be visualized adhering closely to the surfaces of root hairs (Fig. 1A -D). Proteins could also be visualized in the vicinity of bacteria ( Fig. 1B and D). DAB/ horseradish peroxidase staining showed that H 2 O 2 was abundantly secreted into agarose from roots ( Fig. 2A).
In media that contained protein H 2 O 2 staining appeared to be more intense than in agarose without protein (Fig. 2B). H 2 O 2 appeared to be secreted from the central axis of the seedling roots, rather than from root hairs ( Fig. 2C -E). Visual evidence of root protease activity was seen in the smoothing of protein crystals that were in contact with root hairs (Fig. 3A and B). Seedlings grown on protein media frequently had protein accumulations on root hair surfaces ( Fig. 3C and D).   (Fig. 4). Minimal incorporation of 15 N into shoots was seen in the treatments without rhizobacteria grown on agarose without proteins (Fig. 4). Analysis of shoots of seedlings with rhizobacteria on water agarose free of any proteins showed several times more incorporation of the 15 N label (123.43 + 14.4 d 15 N vs. air) compared with seedlings without bacteria under the same conditions (8.68 + 1.31 d 15 N vs. air) (Fig. 4).

Organic N absorption by roots
Previous research has shown that grasses, and plants in general, secrete proteases from their roots. Adamczyk et al. (2008Adamczyk et al. ( , 2010 conducted studies on plant growth on proteins and presented evidence that plants actively degrade proteins without microbes. Further, Paungfoo-Lonhienne et al. (2008) showed that certain plants were capable of degrading proteins on the surfaces of root cells. Results in our seedling protein absorption experiment may reflect the activities of both plant and rhizobacterial proteases. The lowest 15 N absorption into seedling shoots on media with proteins was in the seedlings without rhizobacteria maintained in darkness (Fig. 4). In darkness, H 2 O 2 secretion is minimized and absorption of organic N may reflect the activity of plant root proteases alone. A 53.4 % increase in incorporation of 15 N into shoots was observed in seedlings maintained under alternating light/dark conditions without microbes where H 2 O 2 was secreted from roots in the light periods. The importance of reactive oxygen secretion from roots to the process of N scavenging may be seen in this 53.4 % increase in N absorption from protein over that seen in the seedlings without bacteria maintained in darkness where reactive oxygen was not secreted from roots. In our experiment, the real benefit of reactive oxygen may be lower than that seen in nature because our proteins were already partially denatured by the use of sodium dodecyl sulfate during the extraction process. The role of reactive oxygen in the process of protein degradation was further emphasized by results of the seedling development experiments (Table 1), where the addition of the antioxidant sodium ascorbate to agarose containing the enzyme lipase was seen to suppress development of root  (5415.46) was seen in shoots of seedlings bearing bacteria grown on agarose containing 15 N-labelled protein in a 12-h alternating light/dark cycle (treatment ¼ protein, bacteria, light/dark). Less 15 N incorporation (3817.49) was seen in shoots of seedlings grown under the same conditions but without bacteria (treatment ¼ protein, no bacteria, light/dark). Slightly less 15 N incorporation (2488.49) was seen in shoots of seedlings grown on labelled proteins but without bacteria in total darkness to suppress reactive oxygen secretion (protein, no bacteria, dark). Minimal 15 N incorporation was seen into shoots of seedlings grown on agarose that did not contain 15 N-labelled protein (treatments ¼ no protein, bacteria, light/dark; no protein, no bacteria, light/dark; no protein, no bacteria, dark hairs. With reduced oxidative denaturing of lipase, the plant proteases alone were insufficient to degrade enough of the enzyme to trigger root hair development. Additional experiments will be needed to get a more accurate assessment of the real value of the pre-oxidation of proteins in the degradation process. The importance of the rhizobacteria to the process of N scavenging from proteins is evident in the 41.86 % increase in incorporation of 15 N into shoots of microbe bearing seedlings incubated in alternating light/dark conditions over that of seedlings without microbes under the same conditions. Secretion of H 2 O 2 from roots and the added degradative capacity of the rhizobacteria increased organic N acquisition by seedlings 119.74 % over what was seen in the seedlings without microbes or secreted H 2 O 2 . Evidence for the direct action of proteases produced by grass roots may be seen not only in accumulation of 15 N in shoots in the bacterial-free seedlings grown on the 15 N-labelled proteins, but also in the smoothing of protein crystals in direct contact with grass seedling root hairs ( Fig. 3A and B). There is evidence that some plants (e.g. rice and wheat) under drought stress enhance their efforts to scavenge organic N from soils, where it has been shown that secreted proteases are up-regulated in plants under water stress (Kohli et al. 2012). The trigger for the shift to organic N use may be restriction of access to mineralized N due to reduced water flow through soils (Kohli et al. 2012).

Seedling development and rhizobacteria
The importance of the role of the rhizobacteria to nutrient acquisition by grass seedlings is emphasized in the results of seedling root development experiments with and without rhizobacteria on various substrates (Tables 1 -3). Growth of seedlings with and without bacteria on nutrient-free agarose showed that bacteria stimulated roots to form root hairs and trigger the root gravitropic response where roots grew downward into the medium (Tables 1 -3). In the absence of rhizobacteria, roots did not develop properly; roots tended to remain on the agarose surface and any roots that did penetrate into media did not produce root hairs. Seedling roots of all three species of grasses responded similarly when germinated on water agarose medium. It seems evident that grass seedlings interact with rhizobacteria in some way that triggers proper seedling root development. In seedlings without rhizobacteria, yeast extract and various proteins could consistently restore proper root development in triggering root hair formation and downward root growth. While we have no data that can clarify the molecular communication between rhizobacteria and grass seedling, the capacity of organic N to restore proper root development suggests that the function of rhizobacteria relates to organic N acquisition. This role is further emphasized because protein levels that were sufficient to stimulate root hair development without the presence of rhizobacteria were inhibitory to seedling growth when bacteria were present (Tables 1 and 3). Rhizobacteria appear to significantly enhance N scavenging capacity of grasses as is evident in results of the 15 N-labelled-protein absorption experiment (Fig. 4). Changes in plant development have also been shown for rice plants containing certain fungal endophytes (Rodriguez et al. 2009). It is entirely possible that improved growth in plants bearing other types of nonpathogenic microbes could also relate to increased supply of organic nutrients, with microbes functioning as collaborators of plants in scavenging organic N from soils or rhizosphere microbes.

Adaptations in grasses
The developmental dependence of grass seedling roots on the rhizobacteria suggests that grasses have evolved to rely on these seed-vectored rhizobacteria in order to maximize scavenging of organic N. Features of development of grasses seem consistent with this idea. Grasses possess lemmas and paleas that adhere tightly to the surface of the seed coat. We suggest that the adherent plant tissues are adaptations for vectoring the rhizobacteria on the seed surface. It is apparent that rhizobacteria are vectored on the seed surface because rigorous surface disinfection of seeds is sufficient to eliminate rhizobacteria from seedlings. Microbes vectored in and on seeds are likely to be essential elements of a plant's microbiome that must be transmitted reliably from parent to offspring. The bacteria that are vectored on seeds of the grasses that we examined are likely essential, since seedlings do not develop properly without them. It seems very likely that the rhizobacteria are important components of the 'functional microbiome' of grasses (Lakshmann et al. 2014;White et al. 2014a). Without competent rhizobacteria seedlings would likely be at a competitive disadvantage in terms of nitrogen supply.

Oxidative nitrogen scavenging
In a previous study we observed degradation of some of the bacteria on surfaces of grass seedling roots (White et al. 2012). We hypothesized that the direct oxidative degradation of bacteria and their proteins on roots could provide a source of N for plants, and we termed the process oxidative N scavenging (ONS). Our results here suggest that ONS involves the rhizobacteria integrally, and rather than being the exclusive target of ONS, they are crucial in its operation in providing additional proteases that efficiently degrade oxidatively denatured proteins. It seems likely that all microbial exoenzymes in the vicinity of roots would be vulnerable to degradation through ONS by grasses.

Diurnal model of ONS in grasses
We suggest a diurnal nature to ONS in grasses, with oxidative protein denaturation occurring only during the day and denatured protein degradation and absorption occurring at night (Fig. 5; White et al. 2012). This results in a temporal separation of oxidative and degradative processes. This temporal separation of oxidative and degradative processes is necessary because proteases of the plant and rhizobacteria are also vulnerable to oxidation by H 2 O 2 . In this cyclic process, grass seedlings collaborate with seed-vectored rhizobacteria to scavenge organic N in the vicinity of roots. Both grass and rhizobacteria have active roles in the decomposition of proteins around roots. Grass roots secrete reactive oxygen that denatures proteins, while rhizobacteria and grasses secrete proteases that further degrade proteins into smaller peptides or oligopeptides that may be absorbed by grass roots. Denaturation of proteins by reactive oxygen involves protein unfolding and increased exposure of peptide bonds to proteases (Fligiel et al. 1984;Davies 1987). Denatured proteins become highly susceptible to degradation by proteases (Davies 1987). Bongarzone et al. (1995) found enhanced susceptibility of peroxidized myelin proteins to degradation by Bacillus protease subtilisin. Several of the seed-vectored bacteria we isolated from grasses are known to produce secreted proteases, including B. amyloliquefaciens, Paenibacillus hordei, Pantoea agglomerans and S. epidermidis (Wells et al. 1983;Bott et al. 1988;Rodarte et al. 2011;Kim et al. 2013;Sugimoto et al. 2013). It is interesting to speculate that grasses may manage the rhizosphere to maximize protein production by rhizosphere microbes. The secretion of complex root exudates composed of carbohydrates and proteins into the rhizosphere could be targeted at encouraging secretion of microbial enzymes to degrade the exudate components. In essence, grasses may be cultivating microbes in the rhizosphere and harvesting their proteins to support growth (Walker et al. 2003). Validation of the diurnal oxidative nitrogen scavenging process could provide a key to understanding the mystery of why many plants grow faster during the night (Nusinow et al. 2012). The absorption of protein fragments during the evening hours could provide the rapidly acquired nitrogenous nutrients plants require to build enzymes needed for the growth spurt that occurs in the predawn hours.

Evidence for enhanced absorption of N from air
The results of the 15 N-labelled protein absorption experiment ( Fig. 4) suggests that the seed-vectored rhizobacteria not only enhance absorption of organic N through increased degradation of proteins, but they may also increase N available to plants by scavenging it from air. In this experiment, Petri plates containing agarose media and seedlings were maintained together in a closed plastic canister. We hypothesize that the degradation of proteins by plants and microbes released 15 N-labelled ammonia or other compound into the air of the canister. Analysis of shoots of seedlings with rhizobacteria on water agarose free of any proteins showed a 14-fold increase in N absorption over that of seedlings grown under the same conditions but without rhizobacteria (Fig. 4)

Conclusions
This paper presents observations and experiments that support occurrence of an organic N scavenging process that we hypothesize functions daily in grasses. It is necessary to rigorously evaluate whether ONS, as we have outlined it based on laboratory studies, is actually occurring Figure 5. Proposed cyclic model of the oxidative nitrogen scavenging process in grasses. In daylight grasses secrete hydrogen peroxide from roots in order to denature proteins (microbial exoenzymes) around roots. At night plant and microbial proteases degrade oxidized proteins to form smaller peptides or oligopeptides that may be absorbed by roots. in grasses in soils. This confirmation would likely involve larger scale N tracking experiments. It is necessary to elucidate the culturable and non-culturable rhizobacteria that assist plants to scavenge N. Each bacterium will need to be evaluated for its role in the process of N scavenging. It is possible that the bacteria that were most active in N scavenging were not cultured, or perhaps are not culturable. In most cases we can isolate multiple species of bacteria from grass seeds and seedling roots. It is unknown whether the rhizobacteria function in consortia, or if only a single species is functional. Finally, there are some important applications that can come from this research. If there are rhizobacteria that are more efficient at degrading soil proteins, it may be possible to use them to enhance the N scavenging efficiency of commercial grasses. This could have widespread positive effects, both economically and environmentally in reducing costs to apply fertilizers and reduce inorganic N applications that contaminate bodies of water. Further, Beltran et al.

Sources of Funding
This research was supported by funds from the John E. and Christina C. Craighead Foundation, the Rutgers Turfgrass Science Center, USDA-NIFA Multistate Project W3147 and the New Jersey Agricultural Experiment Station.

Conflicts of Interest Statement
None declared.