Dynamics of 15N from a plant perspective
It has been demonstrated under field conditions that wheat has the potential to take up amino acids intact from soil, although its quantitative significance in terms of plant N nutrition remains highly uncertain (Näsholm et al., 2001, 2009). To assess this uncertainty, we took a microcosm approach to critically evaluate the temporal and spatial partitioning of 15N in different plant, microbial and soil pools. Our results showed an almost linear uptake of 15N by the wheat roots over the 1440-min labelling period and that, overall, plants competed better for as a source of N in comparison to glutamate. Further, proportionally more N derived from was allocated to the shoots in comparison to glutamate-derived N. This is in agreement with findings for wheat seedlings grown in sterile hydroponic culture, where a linear rate of N uptake was observed over a 360-min period and was also taken up preferentially in comparison to alanine, trialanine or (Hill et al., 2011). We note that the plant–microbial competition scenario presented here is not reflective of severely N-limited ecosystems where N supply is regulated more by inputs of N from N2 fixation, atmospheric deposition and slow rates of organic matter turnover (e.g. Inselsbacher & Näsholm, 2012).
The type and amount of N taken up by plants are dependent on several factors, including the concentration of N in soil solution, its diffusion rate through soil, the activity and affinity of the root's membrane transporters, the N status of the plant and the competition from microbes inhabiting the rhizosphere (Glass et al., 2002; Forsum et al., 2008; Jones et al., 2009). The soil used here was naturally low in available N and the amount of N added to the soil was chosen to reflect a pulse addition of N that would occur after the addition of an inorganic- or organic-based fertilizer or as a result of the lysis of plant cells in response to physical (e.g. abrasion or osmotic stress), chemical (e.g. metal-induced bursting of epidermal cells) or biological (e.g. toxin) stress (typically 2–30 mM or free amino acids in root cell sap; Jones & Darrah, 1994; Miller et al., 2001). Measurements of soil N availability indicated a rapid depletion of the added glutamate and especially from the soil solution pool. However, the reasons for this disappearance were solute-specific, being mainly attributable to abiotic ion attraction to soil particles for (i.e. cation exchange) and to biological immobilization for 15N-glutamate. This supports the findings of previous studies that have shown amino acids are rapidly mineralized by the soil microbial community, whose growth is more limited by C than N availability (Dempster et al., 2012). These findings also suggest that the timing of plant 15N uptake was significantly different between the two N treatments. Specifically, as was not irreversibly bound to the soil's exchange phase, it was capable of being readily released back into the available pool as the soil solution concentration fell as a result of progressive biotical removal. This indicates that a large reservoir of potentially available existed for the entire duration of our experiment. By contrast, after microbial uptake of almost all of the 15N-glutamate within 30 min, the plant-available 15N amino acid pool was effectively depleted (although roots may have taken up some unlabelled amino acids produced de novo from soil organic matter turnover). Thus, the period for root uptake was actually 50-fold longer than that for 15N-glutamate. Our results further indicate that the microbial community fed 15N-glutmate also effluxed a significant amount of N back into the soil as and that this process occurred within 5 min of glutamate addition in the planted microcosms. This efflux process is induced by the internal catabolism of the added amino acids (in which 40% of the amino acid-C is typically respired) which causes a C : N ratio imbalance between the microbial biomass and the added glutamate, leading to the dumping of excess N (Jones et al., 2005a,b; Manzoni et al., 2008). Thus, our view is that plant 15N uptake in the glutamate treatment was actually a result of a small amount of intact amino acid uptake in the first 60 min followed by the uptake of microbially effluxed over the subsequent 1380-min (23-h) period. This is supported by Roberts et al. (2009) and the lack of reappearance of 15N-glutamate in soil solution after microbial uptake (Fig. 2). This may also explain the apparent lag phase in 15N uptake in the glutamate treatment in comparison to (as revealed by both NanoSIMS and conventional mass spectrometry in addition to the observed slower translocation of 15N to the shoots). This poor root capture of glutamate supports previous work showing that plants could only capture a very small amount of amino acid C supplied to nonsterile wheat roots (Owen & Jones, 2001).
At high exogenous concentrations (> 1 mM) in axenic soil-less culture, glutamate has been shown to be inhibitory to root growth, particularly when applied to the root apex (Sivaguru et al., 2003; Forde & Walch-Liu, 2009). In this soil-based study, however, the localized addition of glutamate revealed no negative effect of glutamate on either root or shoot growth. We ascribe this lack of negative response to the very rapid lowering of the glutamate concentration in solution by the soil microbial community (and to some extent root consumption), which lowered the glutamate concentration to below the critical response threshold. In addition, inhibitory effects of exogenously applied glutamate on root growth were only observed in hydroponic culture at concentrations > 5 mM (Fig. S5).
Spatial distribution of 15N in plant tissues
Our results showed that the added 15N was quickly taken up and rapidly distributed throughout the root tissues (i.e. within 5 min of addition). Although we did not quantify 15N from root hair or epidermal cells (because of either poor preservation or lack of alignment in the field of view), our results suggest no preferential accumulation of 15N in any root cell type (there was never any accumulation of 15N within xylem tissue; e.g. Fig. S4). This supports the view that 15N is rapidly translocated away from the point of uptake to the active growing regions (e.g. root tips and shoots). In terms of root uptake, NanoSIMS did reveal the presence of 15N super-enriched bacteria in both the ecto-rhizosphere (Fig. 5) and endo-rhizosphere (Clode et al., 2009). It is likely that the typical root-washing procedure will not remove these bacteria and that these may have represented a significant contamination of the root 15N signal in many previous soil-based studies. As we injected 15N into root regions with fully expanded root hairs, this may favour the microbial capture of uptake of amino acids because of the higher microbial population in comparison to the low density of microorganisms surrounding root tips (Marschner et al., 2011). Further, the distribution and expression of root N transporters may be different between the actively and nonactively growing root regions (Yuan et al., 2007).
Dynamics of 15N from a microbial perspective
NanoSIMS data revealed that initial bacterial uptake of was faster than for glutamate but that the internal capacity for quickly became saturated (< 30 min), after which little uptake occurred. By contrast, 15N-glutamate incorporation into the active bacterial community continued to rise throughout the experiment, suggesting that the added C was fuelling the formation of new cell biomass that necessitated the uptake of more N. These highly 15N-enriched bacteria were most often seen close to the root surface or in the endorhizosphere. We hypothesized that these rhizosphere bacteria would already be highly active and primed for glutamate uptake in response to the continual release of acidic amino acids from the roots into the soil via passive root exudation (Ayers & Thornton, 1968). Therefore, the lack of appreciable uptake of 15N derived from glutamate within the bacterial cells within 5 min was surprising. It is possible that wheat rhizodeposition contains low amounts of glutamate, as reported by Phillips et al. (2004), or that the bacterial transporters and internal metabolic capacity were initially saturated by the added glutamate pulse (Jones & Hodge, 1999). The latter is supported by the intrinsically low solution concentrations of total amino acids in this soil (12 ± 1 μM, mean ± SE), suggesting that the microbial community will have been operating high-affinity transport systems (Anraku, 1980). It should also be noted that we only investigated 15N uptake in rhizobacteria and not in the wider microbial community (i.e. fungi and actinomycetes). Although no distinct arbuscular mycorrhizal structures were observed in our roots when viewed by TEM, NanoSIMS or conventional light microscopy of trypan blue-stained roots, we cannot discount the potential for mycorrhizal transfer of 15N from the soil into the inner root tissues. The likelihood that this represents a major uptake N pathway in comparison to direct root uptake, however, is very low (Hawkins et al., 2000).
The C and N status of rhizosphere microorganisms remains an area of hot debate. Depending on the prevailing soil conditions, microbial growth could be C, N or phosphorus (P) limited (Denison et al., 2003; Marschner et al., 2011). Our results obtained with NanoSIMS suggest that they were N limited in the short term and then became C limited. Previous work in sterile hydroponic culture has suggested that passive exudation is dominated by simple low-molecular-weight C compounds that are released in response to the large cytoplasmic–soil solution diffusion gradient (e.g. glucose and sucrose) or by compounds that are released to alleviate stress (e.g. organic acids). In both these cases we would expect this to lead to microbial N limitation in the rhizosphere, which would be exacerbated by root removal of available soil N. The rapid immobilization of observed here would support this view.