Acquisition of dissolved forms of N
In sterile hydroponic culture, nitrate 15N was taken up at the same rate as alanine 15N and five times faster than tetraalanine 15N. In contrast to the equal rate of nitrate and amino acid acquisition in sterile culture, competition from microbes caused a large reduction in acquisition of alanine 15N relative to that of nitrate 15N when the plants were grown in nonsterile soil. Previous investigations have suggested some form of preference of soil microbes for l-peptides over the amino acid monomer (Farrell et al., 2011b, 2012; Hill et al., 2011b, 2012). Consequently, it was surprising that while 15N uptake of alanine was considerably reduced in soil, the rate of acquisition of 15N as tetraalanine relative to nitrate by roots in soil was similar to that in sterile solution. Further, what evidence there is suggests that soil solution concentrations of amino acids bound in short peptides are generally higher than those of free amino acids (Farrell et al., 2011a; Hill et al., 2011a,b,c; Table 1). This may indicate that acquisition of N as short l-peptides by wheat growing in agricultural soil exceeds that as l-amino acid monomers.
Realistic evaluation of the importance of different forms of soil N to plant nutrition presents formidable experimental problems, especially for organic forms. Even when attempting to determine relative fluxes, it is necessary to consider a range of potential caveats in the interpretation of data. We found no statistically significant correlation between recovered 13C and 15N for plants grown in soil. This may principally be a consequence of variation caused by the dilution of the 13C taken up as intact alanine or tetraalanine in a c. 15 000-fold larger pool of plant carbon with a variable 13C content. Some support for this view is provided by the much larger recovery of 13C in roots than in shoots and the fact that rates of recovery of 13C in plants grown in soil did not depart dramatically from those which would be expected from recovery of 15N after losses in respiration (Hill et al., 2011c). Nevertheless, we cannot exclude the possibility that some or even all of the amino acid and/or peptide 15N was acquired by plants following prior mineralization by soil microbes. Taking account of isotopic pool dilution in pre-existing pools of soil N also relies on various assumptions. Even if it is clear what proportion of pre-existing soil soluble N is mixed with, differing flux rates through soil pools for the different forms of N add uncertainty. For instance, the residence time of l-amino acids and short l-peptides in soil is probably only of the order of a few minutes (Hill et al., 2012). Thus, if acquired intact, the flux of N into plants as amino acid or peptide relative to that of forms of N which are less desirable to most soil microbes, such as nitrate, is probably underestimated with a chase period of an hour (Jones et al., 2005). If amino acid and peptide N are only acquired following mineralization, recovery of 15N in plants would be likely to increase with the chase period. In this case, the actual flux of N derived from mineralization of organic forms would be underestimated due to dilution by pre-existing inorganic N and this dilution would probably increase with residence time. Many uncertainties relating to solute production sites and mobility in soil may also be of considerable importance in the design of experiments and interpretation of experimental data. Further, we know little about the composition of individual peptides in soil solution. For instance, 160 000 possible tetrapeptides may be formed from 20 common protein amino acids. To date, the availability to plants of very few peptides has been investigated. Consequently, after correction for pool dilution fluxes of, particularly peptide, N from soil to plant must be interpreted with some caution.
Acquisition of N as microbes
Nitrogen delivered to plant roots as a microbial culture was acquired both by plants with sterile roots and those growing in soil. In both cases the nitrate, amino acid monomer and tetrapeptide forms of N were taken up more than an order of magnitude faster than the microbial suspension. If all of the measured microbial 15N acquired by plants was acquired as intact microbes, our results suggest that sterile roots of plants ingested c. 2 × 106 microbial cells g−1 root DW h−1. This suggests that microbes were ingested at a rate of c. 32 000 cells m−1 root length h−1 or 35 cells mm−2 root surface area h−1, although this does not take account of fine root structure such as root hairs, which may have a role in microbe acquisition (Paungfoo-Lonhienne et al., 2010; Mercado-Blanco & Prieto, 2012). To match the rate of N uptake as nitrate or alanine from sterile solution, plants would need to ingest c. 2 × 108 cells g−1 root DW h−1 and c. 4 × 107 to match N uptake as tetraalanine. When growing in soil, 15N recovery suggests that plants ingested c. 390 000 cells g−1 root DW h−1; c. 6 cells mm−2 root surface area h−1 (assuming a constant 11.6 mg of root in a 7 cm section of rhizotube was accessed by injected microbes). If previous estimates of numbers of bacteria on wheat roots are typical and it is assumed that labelled microbes mixed homogeneously with existing rhizoplane microbes, this suggests that rates of uptake were c. 975 000 cells g−1 root DW h−1: c. 15 cells mm−2 root surface area h−1; c. 4.5% of the standing rhizoplane bacterial biomass d−1 (Liljeroth, 1990). If it is assumed that injected cells mixed with all bacteria on the rhizoplane and in the rhizo-sphere, this value rises to c. 106 cells g−1 root DW h−1 or c. 16 cells mm−2 root surface area h−1. Nevertheless, this flux of N into roots still represents only a maximum of c. 6% of the flux of other forms of N when they are similarly corrected for pool dilution. There are uncertainties in the dilution of amino acid, peptide and nitrate in pre-existing soil pools. However, very poor understanding of the process of direct microbe uptake by roots means that the size of the pool with which added microbes mixed is very difficult to establish. Consequently, our estimates of uptake of N as microbes are probably subject to the greatest uncertainty.
Three percent of the 14C added to soil as microbes was mineralized to 14CO2 within an hour. This was around a fifth of the 14C likely to be mineralized to 14CO2 if added to soil of this type as glucose (Hill et al., 2008). Nevertheless, respired 14CO2 was an almost 40-fold greater proportion of the microbial biomass 14C than the proportion of the microbial 15N which was recovered in plants (0.08%). Similarly, the 0.3% of the microbial 14C respired by the living microbes alone was a six times greater proportion of microbial 14C than the 0.05% of microbial 15N recovered in sterile plants. This may indicate that in both soil and sterile solutions, the microbial 15N recovered in plants was taken up as inorganic 15N after microbial mineralization. However, although there is some uncertainty inherent in the measurement of 14C and 15N in separate plants, the close agreement between values for recovery of microbial 15N and 14C in plants growing in soil strongly suggests that 15N was not acquired only in inorganic forms. We cannot completely exclude the possibility that both N and C were taken up as organic forms of N following prior lysis of microbes. In our opinion this seems unlikely to account for the entire flux of 14C and 15N, as that would necessitate the maintenance of the overall microbial 14C to 15N ratio in the organic forms of N taken up after any losses of C in respiration. Nevertheless, post-lysis, or even post-mineralization, plant uptake probably accounts for part of the flux of microbial N into roots.
Many living endophytes exist in plants, although our knowledge of how widespread the ability to survive within plants is amongst soil microbes is largely restricted to studies on a few species (Hardoim et al., 2008; Ryan et al., 2008; Reinhold-Hurek & Hurek, 2011). Our measurements of incorporation of microbial 15N and 14C cannot distinguish between microbes internalized and metabolized within root cells and those continuing to survive within the plant, that is within the apoplast or root wounds (Gantar, 2000; Hardoim et al., 2008; Paungfoo-Lonhienne et al., 2010). The close agreement between the 15N and 14C recovery from the microbial cells contrasts with the at least 40% of amino acid and peptide 13C which was rapidly metabolized and lost in respiration. This may also indicate that microbes were not metabolized by the plant after uptake. Further, in some cases, endophytes move from root to shoot within the plant without apparent attack or digestion by the host with bacterial movement from the root epidermis to the stele occurring via the apoplast (Reinhold-Hurek et al., 2006; Rosenblueth & Martinez-Romero, 2006; Deering et al., 2011). Consequently, 15N recovery in the shoot cannot be unequivocally attributed to degradation of microbes in the root, with subsequent transport of 15N to the shoot (Paungfoo-Lonhienne et al., 2010). However, even if plant degradation of microbes took place more slowly than other forms of organic N and too slowly to be very obvious over the hour of experiment duration, eventual death and decomposition of some microbes within the plant does seem likely.
Assuming that bacteria can be internalized within root cells, it raises questions about the mechanism of cytophagocytosis and to what extent this is under direct control of the plant (Hardoim et al., 2008; Paungfoo-Lonhienne et al., 2010). Due to the small pore size of the cell wall (< 10 nm) relative to the size of bacterial cells (c. 1000 nm), internalization can only occur by loosening/digestion of the cell wall in mature root regions or possibly at weak points in actively growing cells (Reinhold-Hurek et al., 2006; Miralles et al., 2012). If the plant is actively undertaking this process to acquire N, strong selection for nonharmful bacteria is expected, bypassing myriad host-defence processes (Kogel et al., 2006; Rosenblueth & Martinez-Romero, 2006; Hückelhoven, 2007). Perhaps more probable is that many bacterial cells are taken up passively as has been demonstrated for a range of inert micro- and nano-particulates (Solomon & Matthews, 2005; Miralles et al., 2012). However, whilst it is clear from many studies that live bacteria can rapidly enter and survive in the endorhizosphere, little evidence exists for the passive uptake of dead cells into the apoplast (Quadt-Hallmann et al., 1997; Hardoim et al., 2008).
Although considerable uncertainty surrounding mechanisms remains, our results suggest that acquisition of microbes from soil by wheat roots, with subsequent translocation of N, does take place. Thus, if this is actively undertaken, all three plant species investigated to date, Arabidopsis thaliana, Solanum lycopersicum and Triticum aestivum, appear to have this capacity (Paungfoo-Lonhienne et al., 2010). However, if wheat proves to be typical, low rates of uptake of N as intact microbial cells in comparison with uptake of common inorganic and organic forms of soil N suggest that the importance of this process to overall plant N nutrition is minor. Use of other forms of organic N is often considered to be most important in environments where N mineralization is slow (Chapin et al., 1993; Schimel & Bennett, 2004; Näsholm et al., 2009; Hill et al., 2011a). Similarly, wider investigation may establish that uptake and digestion of soil microbes has high functional importance in some ecosystems; perhaps when free-living diazotrophs are abundant in an N-limited rhizosphere. In these respects, the use of microbes may be more significant in some angiosperms than it is in a highly-bred agricultural plant such as wheat. Of course, it may also be that microbes are primarily consumed as a source of some nutrient other than N, or that their consumption is only of transient importance.
Although each form of N differs in its mobility in soil, strong sorption of microbial cells to soil particles results in a slow rate of diffusion to the root surface (Foppen et al., 2005). Consequently, most microbes available to plants live and reproduce very close to roots and derive much of their nutrition from roots, for example, as exudates (Liljeroth, 1990; Brimecombe et al., 2001). This may mean that active consumption of these microbes primarily represents a mechanism by which plants recover lost nutrients, as has been proposed for some organic solutes (Jones et al., 2005).
Uncertainty still surrounds the extent to which plants can acquire different forms of organic N from soil. Previous studies undertaken in sterile hydroponic culture have clearly demonstrated the potential for plant roots to take up intact microbial cells, but the functional significance of this process in soil environments remains unknown. The results presented here for wheat plants grown in both hydroponics and in soil strongly suggest that the rate of uptake of N as intact microbial cells is very low in comparison with uptake of common inorganic and organic forms of soil N. Although wider investigation is needed and other functions cannot be excluded, this relatively low incorporation of microbial N suggests that digestion of soil microbes probably represents only a small component of overall plant N acquisition.