Influence of different mineral nitrogen sources (NO3−-N vs. NH4+-N) on arbuscular mycorrhiza development and N transfer in a Glomus intraradices–cowpea symbiosis

Labeled nitrogen (15 N) was applied to a soil-based substrate in order to study the uptake of N by Glomus intraradices extraradical mycelium (ERM) from different mineral N (NO3− vs. NH4+) sources and the subsequent transfer to cowpea plants. Fungal compartments (FCs) were placed within the plant growth substrate to simulate soil patches containing root-inaccessible, but mycorrhiza-accessible, N. The fungus was able to take up both N-forms, NO3− and NH4+. However, the amount of N transferred from the FC to the plant was higher when NO3− was applied to the FC. In contrast, analysis of ERM harvested from the FC showed a higher 15 N enrichment when the FC was supplied with 15NH4+ compared with 15NO3−. The 15 N shoot/root ratio of plants supplied with 15NO3− was much higher than that of plants supplied with 15NH4+, indicative of a faster transfer of 15NO3− from the root to the shoot and a higher accumulation of 15NH4+ in the root and/or intraradical mycelium. It is concluded that hyphae of the arbuscular mycorrhizal fungus may absorb NH4+ preferentially over NO3− but that export of N from the hyphae to the root and shoot may be greater following NO3− uptake. The need for NH4+ to be assimilated into organically bound N prior to transport into the plant is discussed.


Introduction
Many studies have reported that the extraradical mycelium (ERM) of arbuscular mycorrhizal (AM) fungi can take up and transfer considerable amounts of N from both organic (Hawkins et al. 2000;Hodge et al. 2001;Leigh et al. 2009; and inorganic (Johansen et al. 1996;Tanaka and Yano 2005;Subramanian and Charest 1999;Govindarajulu et al. 2005;Jin et al. 2005) sources to the host plant. The most common forms of N taken up by AM are nitrate and ammonium. However, under some conditions ammonium (NH 4 + ) supply, compared with nitrate (NO 3 − ) supply, has been shown to decrease mycorrhizal activity (Valentine and Kleinert 2006;Ngwene et al. 2010).
The importance of AM fungi for plant N nutrition is a topic of controversy. The existing experimental evidence has recently been summarized by Smith and Smith (2011). Some authors have argued that the potential N uptake and transport rates of roots are much higher than those of hyphae so that in soils with high mineral N supply, the contribution of AM fungi to shoot N content is likely to be small compared with uptake by the roots (Hawkins et al. 2000;Hodge 2003). However, the case may be different when AM fungal hyphae have access to N sources that are not accessible to roots (George et al. 1992;Frey and Schuepp 1993;Hodge 2003;Leigh et al. 2009). An understanding of N dynamics between AM fungi and host plants may help not only to quantify N uptake processes in individual plants but also to describe plant and AM fungal contribution to global N cycles.
Although studies on N transfer by AM fungi to host plants have shown that N can be transported through AM fungal hyphae, it is still not clear how different forms of N are transferred by the fungi (Smith and Smith 2011). Govindarajulu et al. (2005) proposed that after uptake of mineral N (NO 3 − or NH 4 + ), both forms are assimilated mainly into the amino acid arginine, transported through the hyphae, and then transferred in the form of NH 4 + to the plant. In accordance with this pathway, some studies have indicated a preferred hyphal AM uptake of NH 4 + as opposed to NO 3 − (e.g., Johansen et al. 1996;Govindarajulu et al. 2005;Jin et al. 2005). Despite these insights into plant-fungal N dynamics, inconsistencies exist between published results which may have arisen due to differences in host plant N status, root or hyphal proliferation, localization of N supply in the substrate, and substrate pH or moisture content. Additional inconsistencies could have arisen from a failure to consider that N transfer rates for NO 3 − or NH 4 + from soil to the plant shoot via AM hyphae are controlled by several independent steps: (1) soil adsorption of NO 3 − or NH 4 + , resulting in available and fixed pools in soil; (2) absorption and assimilation by the ERM; (3) transport within the ERM to the fungal tissues within the root; (4) transfer from fungal tissue to root tissue; and (5) root-toshoot transfer. For example, experiments using excised hyphae investigate only step 2, and those on artificial substrates or in nutrient solutions often do not consider step 1.
Since an approach to better understand AM fungal N dynamics under field conditions requires more studies with soil-grown plants (Smith and Smith 2011), the present study was undertaken using a soil-based substrate for both plant and fungal compartments to follow the transport of different N sources by the ERM of Glomus intraradices to cowpea plants. Small fungal compartments were placed within the plant growth substrate, with an air gap to eliminate mass flow and diffusion, to simulate soil patches containing rootinaccessible, but mycorrhiza-accessible N. The influence of different mineral N sources (NO 3 − vs. NH 4 + ) on AM fungal N transfer to the host plant was investigated using 15 Nnitrogen. The effect of these different N sources on AM fungal development was also studied. Based on previous results from our laboratory , we hypothesized that the fungus may accumulate more N when supplied with NH 4 + as compared with NO 3 − but that NO 3 − supply would lead to better mycorrhiza root colonization, ERM development, and fungal N contribution to host plant shoot N content.

Precultivation of plants and AM fungal inoculation
Seeds of the African cowpea cultivar "IT 18", supplied by the International Institute for Tropical Agriculture, were germinated on a filter paper soaked in saturated CaSO 4 solution. Sixty seedlings with fully established primordial leaves were individually transplanted for precultivation into black, round 220-ml plastic planting pots (Teku-Terra TO 9; Pöppelmann, Germany) containing 260 g of heat pasteurized (85°C for 48 h) dry soil substrate at a bulk density of 1.2 gcm −3 . The soil was a nutrient poor loess subsoil of a luvisol, classified as loamy sand with a pH of 7.2, sieved through a 4-mm sieve before pasteurizing. After pasteurizing, the soil was fertilized with 100 mg N kg −1 dry soil ( (Bécard and Fortin 1988) for approximately 8 weeks before experimentation. The material used for plant inoculation consisted of the AM fungal-colonized roots with the surrounding growth medium, which contained numerous spores and hyphae. Prior to its use, the inoculum was placed in a drying oven at 35°C for 48 h and then cut into small pieces. The remaining 30 plants (non-AM) received the same amount of autoclaved (121°C for 20 min) inoculum. After planting, daily water loss from the pots was estimated gravimetrically and water content in the substrate was maintained at 17 % (w/w) with deionized water. Plants were kept in a greenhouse in Grossbeeren, Germany (long. 13°19′ 60″ E, lat. 51°22′0″ N); inoculation was repeated 3 weeks after planting because no colonization was observed in root samples of AM plants at that time. The inoculum was not dried before use to inoculate plants for the second time and was inserted at three points around each AM plant (20 g in total). Again, non-AM plants received the same amount of autoclaved inoculum. Three weeks after the second inoculation, plants showed AM fungal colonization in roots.

Preparation of fungal compartments and fungal compartment substrate
Fungal compartments (FCs) were constructed from 55-ml plastic net pots with a latticed wall (Teku; Pöppelmann, Germany). The outer surface of the net pots was covered with a nylon membrane with a 30-μm mesh width (Sefar AG; Heiden, Switzerland) that allowed hyphae, but not roots, to grow into the FCs. The nylon membrane was fixed to the net pots using silicon (Sista Silicon Küche; Henkel KGaA, Düsseldorf, Germany). To avoid ion diffusion between the substrate in the FCs and the surrounding pot substrate, an air gap was created by lining the inner wall of the FCs with two layers of nylon net having a 1-mm mesh width (Sefar AG; Heiden, Switzerland). Root + fungal compartments (RFC) construction was similar to that of the FCs but without the 30-μm nylon membrane, thus allowing both roots and hyphae to grow into them.
The compartment substrate (in both FCs and RFCs) was a 1:1 mixture of wet sieved loess soil (particle size <40 μm) and glass beads (Ø 1.7-2.1 mm; Carl Roth GmbH, Karlsruhe, Germany), with 20 % w/w water (Neumann and George 2005). This mixture allows for the extraction of intact fungal ERM after harvest. The wet sieved loess soil (compartment substrate) was the same as the cultivation substrate. All nutritional elements were fertilized at the same rate and using the same salts in both the compartment and cultivation substrates, apart from N which was supplied either in the form of NO 3 − (Ca (NO 3 ) 2 ) or NH 4 + ((NH 4 ) 2 SO 4 ). To minimize the conversion of NH 4 + to NO 3 − , a nitrification inhibitor nitrapyrin (N-Serve; Dow AgroSciences LLC; Indianapolis, USA) was applied (initially 7 mg kg −1 dry soil) to the fertilized compartment substrate.

Experimental setup and growth conditions
Fifty-six precultivated plants (28 AM and 28 non-AM) were individually transplanted into black, round 2-l plastic pots (Teku Container BC 17; Pöppelmann, Germany) containing 1,870 g of dry soil at a bulk density of 1.3 gcm −3 . The pot substrate (PS) was prepared in the same way as the substrate used for precultivation, except that N was fertilized either as NO 3 − (Ca(NO 3 ) 2 ) to 16 AM and 16 non-AM plants (PS NO3 ) or as NH 4 + ((NH 4 ) 2 SO 4 ) to 12 AM and 12 non-AM plants (PS NH4 ). Nitrapyrin was applied to the pot substrate in all treatments at an initial rate of 7 mg kg −1 dry soil. Two FCs and two RFCs were alternately inserted into each pot substrate around the plant (Fig. 1). The compartments were filled (5 mm below the top) with 110 g of fresh compartment substrate (2.0 gcm −3 ). The four compartments in each pot were either all filled with NO 3 − -(C NO3/NO3 ) or NH 4 + -(C NH4/NH4 ) fertilized substrate or two (one FC and one RFC) compartments were filled with NO 3 − and the other two with NH 4 + -fertilized substrate (C NO3/ NH4 ). Eight replicates were prepared for the treatment PS NO3 C NO3/NH4 . This permitted later compartment application of 15 N as 15 NO 3 − to four of the eight replicates and as 15 NH 4 + to the remaining four. All other treatments also had four replicates and the experiment was set up in a completely randomized design.
As in precultivation, daily water loss from the pots was estimated gravimetrically, water content in the substrate was maintained at 17 % (w/w) with deionized water, and plants were grown in a greenhouse (July-September, day/night 24/19°C, rh 68 %). Compartments were watered as required to keep the substrate moist. Two weeks after transplanting, all plants received an additional 50 mg N kg −1 dry soil applied to the pot substrate either in the form of NO 3 − (Ca(NO 3 ) 2 ) or NH 4 + ((NH 4 ) 2 SO 4 ). Nitrapyrin (4 mg kg −1 dry soil) was again applied to all pot and compartment substrates. This procedure was repeated after another 2 weeks when compartments received an additional 100 mg N kg −1 dry soil of their respective N-form. Selected FCs were given nitrogen enriched with the 15 N isotope at a rate of 880 μg 15 N per FC (amounting to 20 % of N applied to the FC). The 15 N-nitrogen was supplied as Ca(NO 3 ) 2 or NH 4 Cl to the respective compartments. When all compartments in a pot were fertilized with the same N-form (C NO3/NO3 or C NH4/NH4 treatments), 15 N was applied to both FCs (2× 880 μg per plant). When the N-form was not the same (C NO3/ NH4 treatments), 15 N was applied only to one FC, either the NO 3 − -or the NH 4 + -fertilized one (1×880 μg per plant).

Harvest
Plants were harvested 2 weeks after application of 15 N to the FCs (6 weeks after transplanting). Shoots were separated into stem, leaf, and pod components, and fresh weights were recorded. Roots were separated from the substrate by washing, and a representative sample (about 1 g) of fresh roots To avoid ion diffusion between the substrate in the FCs and the surrounding pot substrate, an air gap was created by lining the inner wall of the FCs with two layers of a nylon net having 1mm mesh width. Two fungal compartments (FC) and two root + fungal compartment (RFC) were inserted into the substrate of each pot around a cowpea plant. All were either NO 3 − compartments (C NO3/NO3 ) or NH 4 + compartments (C NH4/NH4 ), or one of each pair NO 3 − and the other NH 4 + (C NO3/NH4 ) was collected and stored in 15 % ethanol for staining with 0.05 % trypan blue in lactic acid after cutting into 1-2-cm length (Koske and Gemma 1989). The AM fungalcolonized root length was evaluated by a modified intersection method (Tennant 1975;Kormanik and Mc Graw 1982) using 200 or more intersections per plant. All plant parts were oven dried at 65°C for 3 days, and dry weights were recorded. Roots plus ERM and ERM alone were obtained from the RFCs and FCs, respectively, by washing the contents of the compartment over a 40-μm sieve (Retsch Test Sieve 3310-1; Retsch, Germany) and separating glass beads from roots and hyphae (Neumann and George 2005). Samples were then freeze-dried at −30°C for 4 days in a freeze drier (CHRIST ALPHA; Martin Christ Gefriertrocknungsanlagen GmbH, Osterode am Harz, Germany), and dry weights were recorded.
Hyphal length and number of spores were estimated using a modified membrane filtration technique. A subsample of the freeze-dried ERM (about 0.5 mg) was stained with a few drops of 0.05 % trypan blue in lactic acid and homogenized with 300 ml of deionized water in a laboratory blender (Waring Commercial; CT, USA) for 60 s. An aliquot of the hyphal suspension was filtered on a 0.45-μm mesh width membrane filter (MicronSep; GE Water and Process Technologies, USA) using a bottle neck filtration unit (NALGENE Reusable Bottle Top Filter Unit; Nalge Company, NY, USA). The membrane filter was mounted onto a microscope slide, and hyphal length was estimated by a modified gridline intersection method observed at×200 magnification (Newman 1966;Tennant 1975). Spore numbers were counted on the membrane.

Mineral nutrient analysis
For P analysis, 500 mg samples of pulverized plant material were dry-ashed at 500°C, oxidized with 5 ml 21.7 % HNO 3 on a hot plate, and taken up in 25 ml 2.1 % HCl. After filtration (filter circles MN 615, Macherey-Nagel, Germany), P concentration in the filtrate was analyzed colormetrically with a spectrophotometer (EPOS 5060, Eppendorf, Hamburg, Germany) set at a wavelength of 436 nm, after staining with ammonium molybdate-vanadate solution (Gericke and Kurmies 1952). For N analyses, samples of 5 to 15 mg pulverized plant material or freeze-dried fungal material were analyzed after dry oxidation (Elementar vario EL; Hanau, Germany). After total N measurement, the N fraction and helium (carrier gas) were automatically introduced to a coupled emission spectrometer (NOI 7; Fischer Analysen Instrumente, Leipzig, Germany) where 15 N atom percentage labeling was determined. The amount of 15 N ( 15 N content in microgram) in the plant and fungal material was calculated by multiplying the 15 N atom percentage excess with the total N (in millimole per sample) and then converting to microgram 15 N per sample.

Statistical analyses
All data were analyzed for normal distribution and equality of variance before being subjected to analyses of variance (ANOVA). The shoot/root ratio, ERM dry weight, and spore numbers as well as data in the form of percentages were first arcsine square root transformed before being analyzed. Mean values were compared by a one-way ANOVA/Tukey's multiple comparison, or a t test when appropriate. Two-and threeway ANOVAs were used to estimate whether pot substrate N fertilization, compartment N fertilization, or AM fungal inoculation, alone or in interaction, had a significant influence on the mean values. Differences were considered significant when p values were below 0.05. Statistics were performed using the SigmaStat 3.5 program (STATCON, Germany).

Extent of root length colonized by G. intraradices in the pot substrate
No root colonization was observed in non-AM plants. All AM plants showed a high degree of root colonization (Fig. 2). Two-way ANOVA revealed a significant effect of the N-form fertilized to the pot substrate on the extent of AM fungal colonization of roots in the pot substrate (p values <0.001), with a slight reduction in root colonization levels in plants supplied with NH 4 + (PS NH4 ) as opposed to NO 3 − (PS NO3 ). Colonization levels of roots in the pot substrate were not significantly affected by the N-form applied in the compartment substrate.

Plant growth
Four weeks after transplanting, leaves of the NH 4 + -fed plants (PS NH4 ) appeared pale. At harvest, the shoot and root dry weights (DWs) of the cowpea plants were significantly influenced by AM fungal inoculation and pot substrate N fertilization (Tables 1 and 2). Inoculation with G. intraradices increased shoot and root DWs above those of the corresponding non-AM plants. No significant differences in shoot or root DWs were observed between non-AM plants supplied with either NO 3 − or NH 4 + . In contrast, NO 3 − fertilization increased both shoot and root DWs in AM plants, compared to NH 4 + fertilization. This increase was smaller in the presence of NO 3 − in two or all compartments in the NH 4 + -fertilized treatments. When the pot substrate and all compartments were fertilized with NO 3 − (PS NO3 C NO3/NO3 ), the shoot/root ratio was increased by G. intraradices inoculation. No significant differences in the shoot/root ratio were observed among other treatments. No nodules were observed on roots at harvest.

Plant P and N status
Inoculation with G. intraradices resulted in increased plant P concentrations (Tables 2 and 3). In AM plants, NH 4 + fertilization resulted in an increased shoot P concentration but a decreased total plant P content compared with NO 3 − fertilization.
No effect of compartment N fertilization on plant P status was observed. Non-AM plants had higher shoot N concentrations but lower total plant N contents (   (Table 4). These were, Compartments in pot Mycorrhizal ( however, on average colonized 13 % less than roots outside the compartment. The ERM DWand spore density in FCs obtained from the pots fertilized with NO 3 − were also not significantly affected by compartment N fertilization. However, hyphal length increased in this treatment when the compartment was supplied with NH 4 + , irrespective of whether the other compartment was also supplied with NH 4 + or with NO 3 − (   (Fig. 3). The 15 N shoot/root ratio (shoot 15 N content/root 15 N content) was lower when NH 4 + as opposed to NO 3 − was supplied to the FCs, irrespective of whether the other FC was also supplied with NH 4 + or with NO 3 − (Fig. 4). ) were taken up from the substrate by the ERM of G. intraradices, and N from these forms which was transferred to the cowpea plants reached up to 40 % of the applied 15 N (fraction of applied 15 N detected in plant material). Transport of considerable amounts of N by AM fungi to a host plant has been demonstrated in previous experiments using different soilless media (Johansen et al. 1992(Johansen et al. , 1996Hawkins and George 1999;Subramanian and Charest 1999;Hawkins et al. 2000;Govindarajulu et al. 2005;Tian et al. 2010).
The transfer of 15 N by the ERM of G. intraradices from the FC to the cowpea plants was consistently higher with 15 NO 3 − than with 15 NH 4 + , regardless of whether both FCs in the pot contained the same or different N-forms (NO 3 − -N or NH 4 + -N) and irrespective of the N-form fertilized in the main pot substrate. Bago et al. (2001) proposed that absorbed N (NO 3 − or NH 4 + ) is assimilated into arginine via the glutamine synthetase/glutamate synthase (GS/ GOGAT) cycle in the ERM, then compartmented in the vacuoles, and translocated to the intraradical mycelium (IRM) for storage and cytosolic pool replenishment. In the cytosol, arginine is broken down by arginase and urease to release NH 4 + , which is then transferred to the plant at the AM symbiotic interface, the fungus retaining the fixed carbon component of the arginine. This proposed pathway, which is metabolically efficient for both partners, has Two fungal compartments (FC) and two root + fungal compartments (RFC) were inserted into each pot. All were either NO 3 − (C NO3/NO3 ) or NH 4 + (C NH4/NH4 ) fertilized, or one of each compartment type was NO 3 − while the other one was NH 4 + (C NO3/NH4 ) fertilized. Mean values ± standard deviation are shown. Mean values in rows labeled by the same letter are not significantly different  Fig. 3 The total 15 N content in shoot + root (including internal mycorrhizal structures) in microgram per plant at harvest. Two weeks before harvest, labeled N ( 15 N) was supplied to the ERM in the fungal compartment (FC) in the respective form (NO 3 − or NH 4 + ). When both FC in a pot were fertilized with the same N-form (C NO3/NO3 or C NH4/ NH4 ), 15 N was applied to both (left). When the N-form was not the same (C NO3/NH4 ), 15 N was applied to only one FC (NO 3 − or NH 4 + ) or the other (right). Mean values ± standard deviation are shown. Mean values labeled by the same letter are not significantly different (Tukey's multiple comparison, p<0.05 (left) and t test, p<0.05 (right)) received further support from other studies. Govindarajulu et al. (2005) and Jin et al. (2005) reported that after feeding ERM with 15 N-nitrogen, high levels of 15 N-labeled arginine were detected in the ERM which were then translocated to the IRM. They also observed that when 13 C-substrates were applied to the ERM, 13 C was detected in free fungal amino acids, but not in mycorrhizal roots. This indicates that translocated amino acid is broken down in the IRM and only the N-fraction (NH 4 + ) is transferred to the plant while the carbon is recycled into the fungal C pool. Other studies on spatial and temporal expression of N metabolic enzymes (Cruz et al. 2007) and genes (Tian et al. 2010) have provided molecular evidence consistent with the proposed pathway. Govindarajulu et al. (2005) also postulated that no direct transfer of NO 3 − would occur from an AM fungus to the plant and proposed that all N (from NO 3 − and NH 4 + ) is transferred as described above. Although the arginine pathway has been shown to clearly function under specific conditions (Ri T-DNA-trasformed carrot roots colonized by G. intraradices), it is still an open question whether this is the predominant pathway in agricultural (whole plant) systems (see Smith and Smith 2011). If this is the case, then it is unlikely that NO 3 − -N would be transferred faster than NH 4 + -N. However, a larger transfer of 15 N was observed when it was applied in the form of 15 NO 3 − compared with 15 NH 4 + in the present study, similar to that previously reported by Hawkins and George (2001). In addition to this, if arginine is the universal transport form of N in AM fungal hyphae (irrespective of NO 3 − or NH 4 + supply to ERM), similar 15 N shoot/root ratios would be expected in plants where ERM was supplied with 15 NO 3 − and 15 NH 4 + , but this was also not the case in the present study where higher 15 N shoot/root ratios (indicating higher root-to-shoot translocation) were found in the cowpea plants with 15  . Because of the extra energy cost involved in reducing NO 3 − prior to assimilation into amino acids, AM systems may more readily assimilate NH 4 + (e.g., Johansen et al. 1996;Hawkins et al. 2000). The results obtained here with intact hyphae connected to a living plant support this view, in that ERM had higher 15 N enrichment when the FC was supplied with 15 NH 4 + than when supplied with 15 NO 3 − . However, this enrichment (accumulation) of 15 N in the hyphae of G. intraradices exposed to a local 15 N supply is the result of a balance between 15 N absorption and export to other parts of the mycelium or the plant so that high 15 N accumulation in hyphae (after 15 NH 4 + supply) may indicate either high absorption or low export. Thus, larger absorption (assimilation) of 15 NH 4 + over 15 NO 3 − does not necessarily imply greater transfer to other parts of the ERM, to the IRM or to the plant. Considering the potentially high N demand of the AM fungus , rapid assimilation of inorganic N by AM fungi may in part be simply to satisfy the internal fungal demand, for example, during spore formation. Jin et al. (2005) found high levels of 15 N-labeled arginine in mature spores after feeding the ERM with 15 N-nitrogen. It may also be a strategy to prevent the accumulation of toxic amounts of NH 4 + in the ERM (Chalot et al. 2006)  (or its metabolic products) from the root to the shoot or higher accumulation of the metabolic products of 15 NH 4 + in the root and/or intraradical mycelium. These differences in the 15 N shoot/root ratio suggest that the mode of transport and transfer may not always be the same for NO 3 − and NH 4 + taken up by hyphae and also points to a possible additional, alternative transfer pathway for NO 3 − (or its metabolic products). The present results suggest that NO 3 − may either have been assimilated by hyphae in larger amounts than NH 4 + (perhaps due to NH 4 + adsorption in the FC soil substrate) or, more likely, that the metabolic products from NO 3 − are transported faster in the ERM than the metabolic products from NH 4 + . A possible explanation of the present results is also that NO 3 − is transferred directly to the plant after uptake by the ERM. Hildebrandt et al. (2002) found that the transcript levels of a plantspecific NO 3 − transporter increased in AM roots, which may imply an increased NO 3 − acquisition in these roots. From their findings, these authors speculated that AM fungi transfer excess NO 3 − directly to the host plant.
Another factor that may have contributed to the difference in the amount of N transferred from NO 3 − vs. NH 4 + was the availability of the applied N, as mentioned above. In many soilless media, the availability and mobility of applied Nforms (NO 3 − and NH 4 + ) are not distinctly different, whereas NH 4 + is less mobile compared with NO 3 − in most soils. This is particularly true for clay-rich soils, such as the one used in this study; for example, Jensen et al. (1989) observed that clay minerals can adsorb considerable amounts of NH 4 + . This phenomenon could have led to differences in the proportions of applied N available for uptake in the soil-based substrate used in the present study and thus have contributed to the observed differences in the amount of 15 N transferred by the fungus to the plant. It should, however, have no major influence on the 15 N shoot/root ratio.
The non-AM plants in the present experiment had only minimal levels of 15 N in their tissues (concentration and total content), and these were not influenced by the form of 15 N (NO 3 − -N vs. NH 4 + -N) applied to the FC. This suggests that there was no significant leakage of N into the soil outside of the compartment and that the transferred 15 N from the compartments to the AM plants was mainly through the ERM. Although N transfer in surface water on hyphae cannot be completely excluded (as suggested for small amounts of, for example, thiophenes by Barto et al. 2011), the high level of 15 N detected in ERM samples (after washing the ERM from the substrate) is an indication that the main transfer route was within the ERM.
It can thus be concluded from the present data, together with those from earlier studies, that (a) at equal concentrations in the supply solution, AM fungal hyphae may take up more N from NH 4 + than from NO 3 − ("preference" for NH 4 + ), (b) at equal application rates to soil of both mineral N-forms, more NO 3 − than NH 4 + is available for take up by AM hyphae due to soil absorption of NH 4 + , and (c) after uptake of NO 3 − , some of the NO 3 − (or a specific metabolic product of NO 3 − ) may be directly translocated via the ERM to the root and then to the shoot. More experimental evidence, including experiments with other AM fungal isolates under realistic supply conditions close to those in soil, is required to test the general validity of these conclusions.
It is clear that AM fungal hyphae can transport nutrients such as P and N, but in an ecological perspective, it is also important to consider that P or N supply affect mycorrhizal colonization and the formation of the ERM. It is assumed that plants regulate the extent of AM fungal root colonization depending on their P-nutritional status (Smith and Read 2008). However, there appear to be also distinct effects of NH 4 + and NO 3 − in this respect. Although the poor P status in the present study (seen from the deficient P-nutritional status of non-AM cowpea plants) resulted in a high degree of AM fungal root colonization in all AM plants, the twoway ANOVA results indicate an apparent negative effect of NH 4 + supply in the pot substrate on root colonization rate. Such negative effects of increasing levels of NH 4 + on AM fungal root colonization, as compared with a NO 3 − supply, have been previously reported by Azcon et al. (1992), Valentine and Kleinert (2006), and Ngwene et al. (2010). Proposed explanations for this effect include a reduced rhizosphere pH following plant NH 4 + assimilation (Habte 1999;Rohyadi et al. 2004), reduced carbohydrate allocation to fungal development due to competition between root development, NH 4 + assimilation in the root, and fungal growth (Raven and Smith 1976), or improved plant Pstatus that may result from NH 4 + nutrition (Johnson et al. 1984). In the present study, a higher shoot P concentration was observed in NH 4 + -fed cowpea plants, indicating a better P-status compared with NO 3 − -fed plants. It could therefore be that the better P-status of NH 4 + -fed plants contributed to this slight reduction in root colonization rate. The contribution of pH changes to this effect cannot be confirmed since AM colonization levels in root samples from compartments (RHC) in the pots supplied with NO 3 − did not show any significant difference between NO 3 − -fertilized and NH 4 +fertilized compartments. If substantiated in further studies, this would mean that decreased AM colonization of NH 4 + fed roots is not principally linked to local changes in rhizosphere pH but to overall changes in root physiology.
A considerable amount of ERM was harvested from the FCs of nitrate-fed plants. Hyphal length (5-11 mcm −3 substrate) was similar to values observed by Neumann et al. (2009) in a comparable substrate (3.5-8.5 mcm −3 substrate) and was also within the range obtained by Drew et al. (2006) in a sand substrate. Spore density was up to five times higher than values reported for pot cultures using a 1:1 sand/vermiculite substrate (up to 200 spores cm −3 substrate) (Silva et al. 2005) and ten times the values reported from field samples (up to 100 spores cm −3 substrate) (Oehl et al. 2005), but within the range reported by Neumann et al. (2009). In the present study, ERM growth in the FC was clearly decreased by the supply of NH 4 + to the pot substrate. While this could also be a result from the improved P-status in the cowpea plants supplied with NH 4 + , as described above, below-ground carbon allocation may also be a contributing factor. Since NH 4 + is generally assimilated in the roots (Marschner 1995), there is a higher demand for carbon in NH 4 + -fed roots leading to reduced carbon allocation for fungal growth. However, this effect probably did not play a major role in the present study because there was no significant difference in the shoot/root ratio between cowpea plants supplied with NH 4 + and plants supplied with NO 3 − .