Exploring the transfer of recent plant photosynthates to soil microbes: mycorrhizal pathway vs direct root exudation

1A significant proportion of plant photosynthates is transported belowground shortly after photoassimilation and subsequently released to soil microbes (Dilkes et al.,; Bahn et al.,; Mencuccini & Hölttä, ). This release can be through direct exudation from the surface of fine roots or by transfer to the extraradical mycelium of mycorrhizal fungi (Jones et al.,,; Drigo et al., ). Both root exudation and transfer to mycorrhizal fungi occur rapidly after photosynthesis, ranging from a few hours in grasses to a few days in trees (Johnson et al.,; Dilkes et al.,; Kuzyakov & Gavrichkova, ). Root exudation stimulates microbial decomposition of soil organic matter, which in turn improves nutrient availability along the rhizosphere (Kuzyakov,; Bird et al.,; Philippot et al., ). Carbon (C) transfer to mycorrhizal fungi benefits the plant through direct nutrient transfer from the fungal hyphal network (Bever et al.,; Fellbaum et al.,; Kiers et al., ). In both cases, the plant's investment in belowground C allocation is rewarded with increased nutrient availability, in particular nitrogen (N) and phosphorus (P) (Hodge & Storer, ).

2Mycorrhizal fungi improve a plant's access to nutrients and water by extending its range into soil areas that are not accessible by roots and to nutrient-rich soil hot-spots via a vast network of extraradical mycelium. Current hypotheses propose that mycorrhizal hyphae also stimulate surrounding soil microbes by exuding labile C, and thus increase local nutrient availability in the hyphosphere (Hodge et al.,; Cheng et al.,; Jansa et al., ). Direct evidence for the transfer of plant-derived C from mycorrhizal hyphae to surrounding soil microbes is however still lacking (Jansa et al., ), although some studies have shown that arbuscular mycorrhizal (AM) hyphae potentially exude plant-derived C (Johansson et al.,; Toljander et al.,; Cheng et al., ). Importantly, mycorrhizal fungi may be able to spatially shift the prolific C transfer to soil microbes via root-inaccessible places, which would be highly beneficial for plant nutrient supply (Cheng et al.,; Herman et al.,; Nottingham et al., ).

3Although there are likely to be different returns for the plant on investment of C released by direct root exudation vs C traded to mycorrhizal fungi, not much is known about the pathways and crossroads of these two C flows in the plant root (Badri & Vivanco, ). It is thought that the largest proportion of root-exudate C is lost passively, driven by the large C concentration gradient between the root cytoplasm and the apoplast (Farrar et al.,; Jones et al., ). However, in most of the root area the endodermal layer (i.e. Casparian strip) would block diffusion of phloem-released C to the root cortex and further to the soil solution. Exudates are thus thought to ‘leak’ primarily from the relatively small immature root zones close to the root tip. Here sugars can diffuse from the protophloem into the apoplast and from there further to the soil solution, without being blocked by endodermal cells (Farrar et al., ). AM fungi, on the other hand, colonise roots predominately behind the root hair zone in mature roots that exhibit a developed phloem and endodermal layer (Smith & Smith, ). Plant C is thought to be transferred to the fungus as sucrose or hexose across mycorrhizal intraradical structures (hyphae or arbuscules in the root cortex), and transported as glycogen or triacylglycerol through the large extraradical hyphae network (Bago et al.,,; Parniske, ). This symbiotic interface comprises plasma membranes from both fungus and plant, separated by an apoplastic compartment (Parniske,; Smith & Smith, ). In exchange, fungal hyphae are thought to take up ammonium (NH4+), nitrate, amino acids and phosphate from the soil solution. These are most likely transported as arginine or polyphosphate granules, before being transferred to the plant as NH4+ or phosphate (Govindarajulu et al.,; Jin et al.,; Parniske, ). Improving knowledge on C and N flow pathways through roots and mycorrhizal fungi is a key step in understanding possible regulations and consequences of plant belowground C allocation. However, this has been hampered by the lack of suitable methods for the in situ tracing and quantification of C and N flow at the subcellular (i.e. micron) scale, especially at the root–mycorrhiza interface.

4Stable isotope labelling experiments using 13C and 15N enable C and N to be traced through plant and soil pools (Olsson & Johnson,; Paterson et al., ). However, traditional mass spectrometry analysis is based on bulk plant or soil samples and cannot provide information on cellular and subcellular C and N dynamics. Yet it is precisely at the subcellular and root–soil-microbial interface that studies need to be conducted if we are to elucidate the transfer pathways involved. NanoSIMS does enable simultaneous imaging and isotopic discrimination of stable isotopes (e.g. 13C vs 12C and 15N vs 14N) at a subcellular spatial scale (c. 100 nm resolution) (Kilburn et al., ). NanoSIMS techniques have been applied to investigate 15N flow through microbes, protozoa, and animal cells (Foster et al.,; Pernice et al.,; Woebken et al., ), as well as rhizosphere bacteria and plant tissue (Clode et al.,; Jones et al., ). Application of this technique to explore 13C flows from photosynthesis through plant tissues is only just emerging, as these C molecules have proved difficult to preserve during the preparation required for NanoSIMS (Clode et al.,; Bougoure et al., ).

5Here, our specific aim was to elucidate the pathways through which recently photoassimilated C is transferred from plant roots to soil microbes and the possible role of AM fungi in this process. Our research questions were: by what routes is recently photoassimilated C translocated from the vascular bundle to the soil in an AM-colonised wheat root? Are these routes linked to those for NH4+ taken up from the soil solution? What role do AM fungi have in the transfer of recently assimilated plant C to the soil microbial community? And, does N addition to roots or mycorrhizal hyphae affect the amount of recently assimilated plant C released to the soil? To address these questions we (1) used NanoSIMS to visualise the flow of 13C and 15N through wheat roots and their associated intraradical mycelium. For this, plants were exposed to both a 13CO2 enriched atmosphere and a pulse of 15NH4+ applied to the soil. (2) We traced the fate of recently plant-assimilated 13C in both root- and hyphae-associated microbial communities via phospholipid fatty acid (PLFA) and neutral lipid fatty acid (NLFA) stable isotope probing and in dissolved organic carbon (DOC) pools via isotope-ratio mass spectrometry.

6Wheat plants (Triticum aestivum L. var. ‘Wyalkatchem’) inoculated with AM fungi were grown in split-boxes (Supporting Information Fig. S1). Each split-box consisted of two compartments (dimension: 8.5 × 7 × 10 cm) that were separated by an assemblage of two membranes (mesh size: 30 μm) encapsulating a solid 1.5-mm-thick, wide-meshed plastic grid in between. The grid established a small void between compartments to prevent solution flow (Zhang et al., ). Mycorrhizal hyphae, but not wheat roots, were able to grow through these membranes. Each compartment was filled with 500 g of fresh soil. The soil was a coarse-textured agricultural soil routinely used for wheat production in rotation with legumes (88% quartz sand) from Mingenew, Western Australia (29°19′N, 115°44′E; samples collected from the 0 to 10 cm Ap horizon: 10.3 g organic C kg−1 and 0.9 g total N kg−1; sieved < 4 mm). Before use, the soil had been inoculated with 25 g of clipped fine root biomass from bioassay pots growing AM-rich subterranean clover (Trifolium subterraneum). This fostered the establishment of plant AM colonisation. Germinating wheat seeds were planted in one compartment of each box (‘root compartment’ i.e. contains both plant roots and mycorrhizal hyphae). The other compartment only hosted mycorrhizal hyphae and was termed ‘hyphae compartment’. Unplanted control split-boxes did not receive seeds in either compartment.

7In order to simultaneously trace the flow of recent photosynthates and soil inorganic N through the plant root mycorrhizal system, we conducted a double 13CO2 and 15NH4 labelling experiment. Split-boxes with 4-wk-old wheat plants and unplanted controls were subjected to an enriched 13CO2 atmosphere for 8 h. 13CO2-labelling was carried out in three 155-l gas-tight perspex chambers, each containing up to 35 randomly allocated split boxes. Chambers were situated in a temperature-regulated glasshouse at 20°C under natural daylight conditions. 13CO2 labelling began in the morning of a cloudless, sunny day in April. After the initial headspace CO2 concentration in each chamber had decreased to 140 ppm, we added 40 ml 13CO2 (99 atom% 13C; Sigma-Aldrich), equivalent to 340 ppm, into each chamber with a gas-tight syringe. This was followed by regular injections of 30 ml 13CO2, equivalent to 250 ppm, every 30 min. This approach repeatedly restocked photosynthetically driven CO2 depletion in each chamber (average CO2 depletion had been determined in a pre-experiment and was in the range of 8 ppm min−1). This resulted in a quick rise of 13C abundance in the CO2 in the chambers (95 atom% by the 3rd injection) with CO2 concentrations fluctuating between 250 and 500 ppm over 30-min intervals. As split-boxes of all harvests and treatments were distributed randomly across the three chambers, all three chambers needed to be opened after 4 h to take out the split-boxes for the first harvest. This was done quickly to keep air exchange to a minimum. However, chamber 13C concentrations dropped to natural abundance after the first harvest and could only be re-established with a subsequent 13CO2 injection 30 min later. The atom% 13C of CO2 in the chambers was measured with a Delta XL isotope ratio mass spectrometer connected to a GasBench II (Thermo Fisher, Bremen, Germany). Three planted split-boxes were also kept in a separate room (to avoid 13CO2 contamination). These served as unlabelled (natural abundance) controls for calculating 13C excess in plant biomass, microbial biomass, DOC, individual NLFA and PLFAs, and the NanoSIMS-measured compartments within fine roots.

8Two hours after the start of 13CO2 labelling, 15 ml of 1 mM (15NH4)2SO4 (99 atom% 15N; Sigma-Aldrich) was added drop wise into either the root- or the hyphae-compartment of the split-boxes. This was achieved with a long needle through pre-installed rubber septa in the upper cover of the perspex chamber (one septum aligned above each split box).

9In order to evaluate the flow of recently photoassimilated plant C through both root and hyphae-associated soil, we measured 13C of DOC, microbial biomass, and NLFAs and PLFAs of soil samples. These were from both root- and hyphae-compartments harvested 4, 8, and 24 h after the start of the 8 h 13CO2 labelling period. This corresponded to 2, 6 and 22 h (respectively) after 15N addition (n = 5 replicate split-boxes for each time point and treatment). Unplanted control split-boxes served as a control to assess the possibility of direct (i.e. photo- or chemo-autotrophic) CO2 fixation by soil microbes. These were harvested 8 h after the start of 13CO2 labelling. In addition, samples of fine roots for NanoSIMS analysis were taken from labelled and control (n = 3) split-boxes 8 h after the start of the 13CO2 labelling period. AM fungal colonisation of roots was assessed by light microscopy after staining with Trypan blue using a gridline intercept method (Newman, ).

10PFLAs and NLFAs were extracted from soil with a mixture of chloroform, methanol and citrate buffer (1: 2: 0.8 v/v/v). PLFAs were separated from NLFAs using silica bonded columns. After fractionation, both PLFAs and NLFAs were converted to methyl esters by alkaline methanolysis (Frostegård et al., ). Dried fatty acid methyl esters (FAME) were re-dissolved in isooctane, and concentrations and C isotope ratios were determined by a Trace Ultra GC (Thermo Fisher) interfaced with an Isotope Ratio Mass Spectrometer (Delta V Advantage; Thermo Fisher) via a combustion unit (GC combustion II/TC; Thermo Fisher). A FAME mixture (Supelco, nrs 47080-U and 47885-U) was used as a qualitative standard. For calculation of FAME concentrations and isotopic ratios an internal standard (19:0) was used. In addition, isotopic ratios were corrected for 13C: 12C ratio of C added during the methanolysis. Fatty acids i15:0, a15:0, i16:0, i17:0, a17:0 were used as indicators of Gram-positive bacteria, cy17:0 as an indicator of Gram-negative bacteria, 14:0, 15:0 and 17:0 as indicators for bacteria in general, and 18:1ω9c and 18:2ω6,9 as general markers for fungi (Zelles,; Leckie,; Joergensen & Wichern,; Kaiser et al., ). Interpretation of 13C enrichment of the fungal biomarkers 18:1ω9 and 18:ω6,9 was undertaken with care as these biomarkers are known to also occur in plants. Earlier work has shown that the contribution of plant-borne 18:2ω6,9 from roots that remain in sieved soil compared to the total 18:2ω6,9 is negligible (Kaiser et al., ). However, as the present approach involves strong labelling of the plant with 13C, the expected contribution of plant-borne 13C-labelled PLFAs to total 13C-labelled PLFAs in sieved soil may be higher. We thus excluded PLFAs 18:1ω6 and 18:2ω6,9 from part of our analysis to get a more reliable pattern of 13C in soil microbial communities. Although PLFAs 16:1ω5 and 18:1ω7 are frequently used as a signature for AM fungi, they are also known to occur in Gram-negative bacteria (Ruess & Chamberlain, ). As our analysis showed that PLFA 18:1ω7 occurred in similar amounts in unplanted controls, whereas 16:1ω5 was absent (Table1), we only used the latter as a PLFA marker for AM fungi. In contrast to their PLFA counterpart, NLFAs 16:1ω5 and 18:1ω7 are known to occur specifically in AM fungi storage compounds and not in other microbial groups, making them excellent biomarkers for AM fungi (Olsson & Johnson,; Drigo et al.,; Ruess & Chamberlain, ). Total PLFA and NLFA concentrations were calculated as mg C g−1 dry soil. The amount of C from recent root exudates incorporated into PLFA/NLFAs was calculated as outlined below.

11In order to estimate DOC and total microbial biomass in root and hyphal compartments, 20 g soil was extracted with 80 ml of 0.5 M K2SO4 before and after chloroform fumigation (Amato & Ladd, ). Dissolved organic C concentration and 13C of DOC were measured in aliquots of the extracts on an HPLC (Dionex Corp., Sunnyvale, CA, USA) connected to a Finnigan Delta V Advantage Mass Spectrometer (Thermo Electron, Karlsruhe, Germany) linked by a Finnigan LC-IsoLink Interface (Thermo Fisher) by direct injection (without column, direct mode) at a flow of 0.5 ml ultrapure water min−1 (Millipore, Vienna, Austria). Microbial biomass C was calculated from differences in DOC in fumigated and nonfumigated soil samples. The fraction of 13C in the microbial biomass C (13CMB) was calculated by a two-pool mixing model from isotope ratios and DOC concentrations in fumigated and nonfumigated soils:

12(13Cf and 13Cnf, 13C-fractions of C in the fumigated and nonfumigated samples (as atom% 13C), respectively; Cf and Cnf, absolute concentrations of C (μg C g−1 soil) in fumigated and nonfumigated samples, respectively.)

13We calculated the excess of 13C relative to the unlabelled control in the microbial biomass, the DOC pool, or each individual PLFA/NLFA (APE 13CX) as:

14(13CX, measured fraction of 13C in the microbial biomass, DOC or individual PLFA/NLFAs; 13CCONTROL, fraction of 13C of the respective unlabelled control (in atom%).)

15We calculated the total amount of excess 13C (μg C g−1 soil) that has been incorporated in the microbial biomass, DOC or each PLFA/NLFA (13CINC), respectively, as:

16(CX refers to the C concentration in microbial biomass, DOC or PLFAs/NLFAs, respectively (in μg C g−1 soil).)

17Carbon and N concentrations, and isotope ratios of ground plant root and shoot samples were analysed with a PDZ Europa ANCA-GSL elemental analyser interfaced to a PDZ Europa 20-20 IRMS (Sercon Ltd, Cheshire, UK, analysed at UC Davis Stable Isotope facility, Davis, CA, USA).

18Fine roots (diameter < 0.5 mm) from split-boxes (n = 3) that were harvested 8 h after the start of 13CO2-labelling (i.e. 6 h after 15N addition) were dissected and immediately rapid-frozen in N2 slush. Samples were freeze-substituted in 10% acrolein in diethyl ether and embedded under vacuum in araldite, based upon the method of Marshall and described in detail by Kilburn & Clode. All solutions were anhydrous and maintained over a molecular sieve, and embedded samples were stored in a desiccator. Transverse sections from resin-embedded material were cut dry at 1-μm thickness, mounted on glass slides and stained with toluidine blue to identify suitable fields of view for subsequent NanoSIMS analyses (Fig.1). Similar dry-cut sections were prepared for NanoSIMS analysis, but these were mounted directly onto Si wafers, before being coated with 5 nm Au (Notes S1).

19In situ isotopic mapping was performed using a NanoSIMS 50 (Cameca, Gennevilliers, France), with a 16 keV Cs+ primary ion beam. Analyses were performed in multi-collection mode with the trolleys positioned to simultaneously detect the negative secondary ions 12C−, 13C−, 12C14N−, 12C15N− and 31P−. The mass spectrometer was tuned to high mass resolution of c. 9000 (CAMECA definition) to separate the 12C15N− from the 13C14N− peak on mass 27 and the 13C− from the 13C1H− peak on mass 13 using an entrance slit of 30 μm, an aperture slit of 200 μm, and a 10% reduction in the signal at the energy slit.

20For secondary ion imaging, the primary current was set to c. 20 pA to optimise the secondary ion signal using a 750-μm primary aperture (D1), giving a spot size of c. 200 nm. Images were acquired by rastering the beam over an area 30 × 30 μm square, at a resolution of 256 × 256 pixels, giving a pixel size of 117 nm. Optical micrographs of sections were used to navigate around the sample. Image data consist of the total number of counts for a given secondary ion species recorded on each pixel, with count times kept constant at 60 ms per pixel. All areas were implanted to the same ion dose by the primary beam before each acquisition to remove surface contamination and to enhance the generation of secondary ions.

21Images were processed using the OpenMIMS data analysis software (National Resource for Imaging Mass Spectrometry http://nrims.harvard.edu) for the freeware package ImageJ (National Institutes of Health, Bethesda, MD, USA). Images were corrected for detector dead time (44 ns) on individual pixels before any other data processing. Maps representing the 13C: 12C ratio were obtained by dividing the 13C− secondary ion counts by 12C− counts on each pixel. Similarly 15N: 14N ratios were obtained by dividing the 12C15N− counts by 12C14N− counts on each pixel. Numerical 13C: 12C and 15N: 14N ratio data were extracted directly from the images by selecting regions-of-interest (ROI: discrete groups of pixels that define a particular feature), and extracting the total number of counts for the given ROI. Ratios were calibrated by taking daily measurements of the resin surrounding the sections to correct for instrumental mass fractionation (IMF). The resin was also independently analysed by IRMS (δ13C = −30.08, SD = 0.196; δ15N = −3.08, SD = 2.1; n = 11). Quoted uncertainties are the standard error of the mean of the measured ROIs for any given fungal/plant cell component across all fields of views (n = 30) from all root samples (n = 3, and one unlabelled control).

22The effect of quasi-simultaneous arrivals (QSA) (Slodzian et al., ) on the ratios was tested by applying different beta corrections to individual ROIs from several images. Beta values of 1, 0.75 and 0.5 (cf. Hillion et al., ) were applied to the data directly using the OpenMIMS software, but were found to have a negligible effect on these data (data not shown). As such, the data presented here were not corrected for QSA.

23We used R (R 2.15.0 with package vegan 2.0-4 (http://www.r-project.org/) for multivariate correspondence analysis of C concentrations (μg biomarker-C g−1 soil) and 13C excess (ng 13C excess g−1 soil) of 18 individual PLFA and 2 NLFA biomarkers (Gonzales et al., ). SigmaPlot 12.0 was used for all other statistical analyses (Systat Software GmbH, Erkrath, Germany).

24After 4 wk, wheat plants were c. 15 cm high and 56% of root length of plants in split-boxes were colonised by AM fungi (SE = 4.75, n = 10). Viable mycorrhizal hyphae were abundant in both compartments. Biomarkers for AM fungi (PLFA: 16:1ω5, NLFA: 16:1ω5 and 18:1ω7) were abundant in both root- and hyphae-associated compartments of planted samples, but were absent from unplanted controls (Table1).

25Isotope ratio images enabled visualisation of the in situ flow of recent photosynthates and N through a plant root and its associated AM structures (Figs2–4). We did not encounter mycorrhizal arbuscules in the samples we processed for NanoSIMS analysis, but we did identify mycorrhizal intraradical hyphae. After 8 h of 13CO2 labelling, phloem cells were significantly enriched in 13C compared to other areas of the root stele (Figs2, 3). Apart from the phloem, 13C enrichment also occasionally appeared in apoplastic regions of the stele (i.e. cell walls), but not within the cytoplasm of other cells (Figs2,5). In the root cortex, labelled 13C occurred predominantly in AM intraradical hyphae, with lower amounts also detected in cortex cell walls. This 13C was commonly located at intersections of more than two cells (Figs3, 4), but not in symplastic areas of the cortex. The 15N taken up from the soil solution was detected in both cells and cell walls of the stele (symplast and apoplast), but was only observed in cell walls in the cortex (apoplast only) (Fig.5). AM intraradical mycorrhizal hyphae did not incorporate 15N.

26There was a qualitative difference in the pathways of photosynthates and N through root sections close to the root tip compared to more developed fine root sections. Root sections close to the root tip were characterised by immature sieve tube and xylem elements and a stele diameter < 100 μm (Fig. S2). In these root tip sections 13C and 15N were significantly co-located (Fig.3, R2 = 0.41 and 0.47, P < 0.001, for stele and cortex areas). By contrast, mature root sections, which were characterised by a central metaxylem vessel, clearly visible protoxylem and phloem cells, as well as mycorrhizal intraradical hyphae in the root cortex (Fig.1), 13C and 15N were not correlated in the stele or in the cortex. Additionally, in the cortex of immature roots both 13C and 15N enrichment occurred mainly in widened (apoplastic) intersections between more than two cells (Fig.3k,l), whereas in mature roots 13C-enrichment was located primarily in intraradical mycorrhizal hyphae (Fig.4).

27As expected, intraradical mycorrhizal hyphae were rich in phosphorus (P) (Fig.4b,e,g). Although the 13C signal within the intraradical hyphae cross-section was predominantly linked to the N-rich interface between plant and fungal cells (in between fungal and plant plasma membrane, Fig.4c,f,g), P was not restricted to that area and occurred also in the inner part of the intraradical mycorrhizal hyphae.

28Approximately 75% of the 13C injected during the labelling period was recovered in the plant and microbial biomass of the harvested split-boxes (after 4 h: a total of 94 mg 13C excess was recovered in plant and soil pools of 35 split-boxes, with a total of 125 mg 13C added, Table2).

29Analysis of 13C in plants, soil microbial biomass and extractable DOC showed that photosynthetic C was rapidly transported belowground. After 4 h, 13% of total 13C assimilated by plants was found in root and soil microbial pools (Table2). From this C, 90% was stored in roots whereas 10% was found in the soil microbial biomass (6% and 4% in root- and hyphae-associated microbes, respectively). Twenty-four hours after the start of the 8-h labelling period, a total of 29% of the photoassimilated 13C was found in belowground pools.

30Plant-assimilated C was transferred to the soil microbial community from either roots or hyphae without being released to the DOC pool in the soil solution. The microbial biomass in both compartments was significantly enriched in 13C after only 4 h of 13CO2 labelling. However, there was no significant excess of 13C in the extractable DOC pool in either compartment at 4 or 8 h after the start of 13CO2 labelling. Only a small enrichment in DOC of the root compartment was detected after 24 h (Table2).

3113C in PLFA biomarkers showed different temporal dynamics in root- and hyphae- compartments. The 13C in microbial PLFAs of hyphae compartments increased quickly after the start of photosynthetic labelling being similar to, or higher than, 13C in root-associated PLFAs after 4 h (Fig.6b,f). These then decreased close to initial concentrations some 16 h after the end of the labelling period (Fig.6; Table3). By contrast, 13C in root compartments continued to accumulate in PLFAs/NLFAs after the end of the labelling period (Fig.6; Table3). The strongest increase in 13C after the labelling period in root-associated soil microbial communities was found in the saprotrophic fungi biomarker 18:2ω6,9 (Fig.6d). The ratio of excess 13C in the AM fungal PLFA biomarker (16:1ω5) to excess 13C in bacterial PLFA biomarkers was higher in hyphae compartments compared to root compartments (Fig.6e). There was no difference between compartments in the ratio of total C concentrations of PLFA 16:1ω5 to bacterial biomarkers (Table1).

32The addition of 15NH4+ to either the root- or hyphae-compartment generally increased the fraction of 13C in PLFA and NLFA biomarkers after 24 h. 13C atom% excess of the AM fungal biomarker PLFA 16:1ω5 was 0.026 (± 0.010 SE) and 0.010 (± 0.006 SE) with and without N addition, respectively, in the hyphae-associated compartments. In root-associated compartments this excess was 0.039 (± 0.012 SE) and 0.019 (± 0.004 SE) with and without N addition, respectively. Although almost all PLFA and NLFAs showed increased concentrations after N addition (data not shown), most of these trends were statistically insignificant, except for NLFA AM fungal biomarkers in the hyphae-associated compartment (0.117 (± 0.033 SE) and 0.015 (± 0.024 SE) with and without N addition, respectively (P < 0.05, n = 4)).

33A multivariate correspondence analysis based on excess-13C concentrations in 13 PLFA and 2 NLFA biomarkers (excluding potentially plant-born biomarkers 18:1ω6 and 18:2ω6,9) revealed that different microbes were involved in the uptake and turnover of plant C in root- and hyphae-associated microbial communities (Fig.7). Although the microbial community composition per se was similar in root and hyphae compartments, the part of the community utilising recent plant C differed (Fig.7). In hyphae-associated microbial communities the uptake of recent plant 13C was associated to a greater extent with the PLFAs 16:1ω5 (that occur in both AM fungi and Gram-negative bacteria), general bacteria and Gram-negative bacteria biomarkers. By contrast, 13C uptake in root-associated communities were largely dominated by NLFA biomarkers 16:1ω5 and 18:1ω7 (specific for AM fungi), and some Gram-positive biomarkers (Fig.7). When 18:1ω6 and 18:2ω6,9 were included in the analysis it appears that 13C-labelled 18:2ω6,9 is clearly more abundant in root-associated than in hyphae-associated PLFAs, especially after 24 h (Fig. S3b). This, however, is not the case for the abundance of unlabelled 18:2ω6,9 (Fig. S3a). When both potentially plant-borne PLFA biomarkers and the NLFAs 16:1ω5 and 18:1ω7 (that may also occur in mycorrhizal intraradical hyphae inside roots) were excluded from the multivariate analysis, there is a less pronounced, but still notable difference between 13C feeding communities in root and hyphae compartments (Fig. S3c,d).

34A rapid and close coupling between plant photosynthesis and belowground C allocation to roots and subsequent soil respiration has repeatedly been demonstrated (Johnson et al.,; Dilkes et al.,; Bahn et al.,, ). AM fungi are known to act as a major sink for plant photosynthates over time periods of days to weeks (Johnson et al.,; Olsson & Johnson,; Drigo et al., ). The possible role of AM fungi to act as a gateway for direct delivery of recent plant photosynthates to other soil microbes on a rapid timescale (i.e. via hyphal exudates, not by hyphae turnover) has seldom been studied due to practical difficulties in separating fungal hyphae exudates and root exudates (Olsson & Johnson,; Drigo et al., ). AM fungi that provide plant photosynthates to hyphosphere microbes may be rewarded with increased nutrient availability through stimulation of microbial soil organic matter depolymerisation (Hodge et al.,; Jansa et al., ). This seems a reasonable strategy for AM fungi that lack, in contrast to other types of mycorrhiza, an ability to produce extracellular enzymes for degrading complex organic compounds (Smith & Smith, ). In addition, mycorrhizal fungi may have a particularly high demand for nutrients, both for their own use and as an exchange for plant C. To better understand the potential role of AM fungi as a possible ‘primer’ of microbial soil organic matter decomposition (Cheng et al., ), it is important to elucidate the pathways of C release in the plant root via direct root exudations and mycorrhiza.

35Our results indicate that recently photoassimilated C is released from phloem (or phloem-adjacent cells) into the apoplast without further involvement of symplastic areas in the stele. In areas of root without a blocking endodermal layer (i.e. in immature regions close to the root tip), this release consequently allows C to further diffuse to the root's surface. This finding supports an earlier proposed hypothesis concerning root exudation being a passive loss of C by plant roots (Farrar et al.,; Jones et al., ). However, before C arrives at the ‘leaky’ immature root area it needs to pass the mature root zone. Here, the Casparian strip would block any diffusional loss from the apoplast of the stele. Movement of C into the root cortex would thus require symplastic (i.e. controlled) transport across the endodermis. At the same time this zone is usually colonised by mycorrhizal intraradical hyphae (Smith & Smith, ). After a few hours of 13CO2 labelling, we found mycorrhizal intraradical hyphae to be clear hot-spots of 13C enrichment within the root cortex. Although we found significant 13C enrichment of both phloem cells and mycorrhizal intraradical hyphae in the same samples, we were not able to identify where exactly C was transferred from the root to the fungus. There was no evidence that the 13C in the intraradical hyphae had been taken up from surrounding root cells, which were not enriched in 13C. Recent plant C in the mycorrhizal intraradical hyphae resided predominantly in the N-rich apoplastic area of the symbiotic interface between plant and fungus, but not in the P-rich fungal compartment in the centre (Fig.4g). This could indicate that the apoplastic part of the interface plays a role for the transport of recent plant C within the mycorrhizal intraradical hyphae. Plant-assimilated 13C may have been transferred to the mycorrhizal interface elsewhere, that is in mycorrhizal arbuscules (which we did not encounter in our samples) or in other places of the intraradical hyphae network. Our results indicate that a significant part of recent photoassimilates may be diverted into mycorrhizal intraradical hyphae even before they reach the passive zone of root exudation at the root tip. It is interesting to note that C transfer to the fungi is most likely under plant control as it requires symplastic transport across the endodermis and transfer to mycorrhizal intraradical structures (whereas root exudation may not be).

36We did not encounter mycorrhizal arbuscules in our root samples and this is likely due to the fact that we were only able to sample and analyse a tiny fraction of the overall root system. Similarly, we may not have detected 15N in mycorrhizal intraradical hyphae because the dropwise addition of 15N to only one of the two compartments may have led to heterogeneous distribution of 15N within the root-mycorrhiza system. Despite these limitations, our results provide the first glimpse into C and N flow across root and mycorrhiza. We present visualisation of the in situ flow of labile C compounds derived from photosynthesis of 13CO2 through plant roots and arbuscular mycorrhizal features. Further studies are now required to explore the potential of NanoSIMS to investigate the exchange of C and N between plant and mycorrhizal fungi, ideally also including mycorrhizal arbuscules.

37Highly enriched intraradical mycorrhizal features in mature root sections indicate that the mycorrhizal hyphae network may be an important gateway for photosynthetic C in the short term. This was further supported by the finding that both 13C-PLFA and 13C-microbial biomass fumigation extraction analyses showed similar total amounts of 13C in the microbial biomass of root- and hyphae-compartments in the short term. For example, after 8 h totals of 1.32 ± 0.49 and 1.76 ± 0.36 ng excess 13C g−1 dry soil were stored in PLFAs of hyphae- and root-associated soil, respectively (Fig. S4, Table S1). Both PLFAs and bacteria-specific biomarkers also showed similar total amounts of 13C in both compartments 4 and 8 h after the start of labelling (Figs6, S4; Table S1). This indicates a rapid transfer of 13C not only from the root, but also from hyphae to soil microbes. Although some studies have shown that AM hyphae exude plant-derived C into the hyphosphere (Johansson et al.,; Toljander et al.,; Cheng et al., ), there is no evidence for the direct transfer of recent photoassimilates from mycorrhizal hyphae to hypersymbiont microbes in the surrounding hyphosphere (Jansa et al., ). Although we cannot rule out that C transfer to soil microbes could have happened via hyphae turnover, our results strongly indicate that transfer was via hyphal exudation, due to the short incubation time.

38The lack of 13C in the DOC of the soil solution of both plant root- and hyphal-compartments (Table2) indicates rapid uptake of root- or hyphae-exuded C by surrounding soil microbes. The surface of both plant roots (Watt et al., ) and mycorrhizal extraradical hyphae (Scheublin et al., ) are colonised by specific assemblages of bacterial species that differ from the bulk soil bacterial community. This may explain why, overall, microbial communities were similar in root- and hyphae-associated soil, but the fraction of the community utilising root- or hyphae-released C was different (Fig.7). It is therefore reasonable to assume that different species assemblages are associated with fine plant roots and AM hyphae.

39In contrast to PLFA biomarkers, NLFA biomarkers for AM fungi showed consistently higher 13C enrichment in root compartments than in hyphae compartments after 4 and 8 h of labelling (Fig.6). A possible explanation is that the NLFA biomarker 16:1ω5 occurs in high numbers in mycorrhizal intraradical features and in spores (Olsson & Johnson, ). A short-term increase of 13C in NLFAs of root-associated soil samples may reflect contamination of the soil with fractions of fine roots containing mycorrhizal intraradical hyphae, whereas the incorporation of 13C in spores may only happen at longer timescales.

40Unlike in root compartments, 13C in bacterial PLFAs of hyphae-compartments almost returned to initial concentrations 16 h after the end of the labelling period. This indicates that plant-derived C exhibits a faster turnover in microbial PLFA biomarkers of hyphae-associated soil than in root-associated soil. Mycorrhizal hyphae seem to preferably transport recently assimilated (i.e. only a few hours old) labile plant C, which were rapidly replaced by nonlabelled components after the labelling period. Plant roots, by contrast, seem to have continued to deliver C that was photosynthesised several hours before, most likely in the form of more complex compounds such as mucilage or root border cap cells, which are naturally not transported through mycorrhizal hyphae. A significant fraction of the C photoassimilated by plants during daytime is stored as starch, from which it is mobilised during the subsequent night as a substrate for aboveground and belowground growth and respiration (Bahn et al.,; Graf & Smith, ). The finding that less of the 13C signal was found in microbial communities of hyphae compartments compared to root compartments after 24 h (i.e. after the subsequent night after 13CO2-labelling) may thus also indicate that the hyphal pathway is less important for night-time belowground C transport. More research is needed to explore the possible different dynamics of C being transported via the root and the hyphal pathway to soil microbes.

41The particularly strong increase of 13C in the PLFA 18:2ω6,9 in root-associated soil may indicate the feeding of saprotrophic fungi on this different C source. However, as this PLFA is also known to occur in plants this could be due to an increased contribution of 13C-enriched plant remains to soil samples.

42NH4+ added to either the root or the hyphae compartment increased the fraction of 13C in AM fungi biomarkers after 24 h. Although the overall amount of NH4+ added to the soil was small (c. 3% of the total NH4+ in the soil solution), it may have caused localised increased NH4+ concentrations around plant roots and mycorrhizal hyphae. We speculate that his could have acted as a first signal for the encounter of an N-rich organic patch in the soil. This may have then triggered an increase of belowground plant C allocation towards that spot in order to further stimulate microbial decomposition of the encountered substrate. If root exudation was mainly a passive diffusion process driven by a lack of a Casparian strip in immature root areas as proposed (Farrar et al.,; Badri & Vivanco, ), well-directed priming effects induced by plants at substrate-rich spots in the soil would be impossible. Our finding that NH4+ addition to the hyphal compartment increased the amount of plant C delivered to this compartment, indicates that mycorrhizal fungi may play a key role in a controlled response of plants to varying soil nutrient content. As these results were only weakly significant, however, further studies are needed to clarify the role of mycorrhizal fungal as a potential partner for well-directed soil priming.

43Our findings indicate that a significant fraction of recently photoassimilated C delivered to soil microbes leaves plant roots via mycorrhizal hyphae already upstream of the place of passive root exudations. AM extraradical hyphae, which enlarge root areas, may thus not only improve nutrient uptake by plants, but also act as an effective distributor of recently assimilated plant C to the soil microbial community. Moreover, the amount of plant C channelled through the AM pathway seems to respond quickly to changing soil nutrient availability. Based on our results we propose qualitative differences between the two pathways of recent plant C release to soil microbes: hyphosphere C release is more intimately coupled to plant photosynthesis than rhizosphere C release, and processed by a different set of microbes. Furthermore, the involvement of the mycorrhizal partner may allow for a finer regulation of plant responses to soil nutrient conditions as C is not necessarily lost passively, as likely happens through root exudation.