Author for correspondence: Huiying Li Tel: +61 (8) 83036787 Fax: +61 (8) 83036511 Email: email@example.com
• We used 32P to quantify the contribution of an arbuscular mycorrhizal (AM) fungus (Glomus intraradices) to phosphorus (P) uptake by wheat (Triticum aestivum), grown in compartmented pots. The soil was from a major cereal-growing area, the Eyre Peninsula, South Australia; it was highly calcareous and P-fixing. Fertilizer P was added to soil at 20 mg kg−1, as solid or liquid. Two extraction methods were used to estimate plant-available P.
• Fungal colonization was well established at harvest (36 d). Application of P decreased both colonization and hyphal length density in soil, with small differences between different P fertilizers.
• Plants showed large positive responses in terms of growth or total P uptake to all P additions, and showed no positive (or even negative) responses to AM colonization, regardless of P application.
• 32P was detected only in AM plants, and we calculated that over 50% of P uptake by plants was absorbed via AM fungi, even when P was added. The results add to the growing body of knowledge that ‘nonresponsive’ AM plants have a functional AM pathway for P transfer to the plant; it should not be ignored in breeding plants for root traits designed to improve P uptake.
Arbuscular mycorrhizal (AM) fungus–plant relationships are usually described as mutually beneficial, because fungi supply mineral nutrients, especially phosphorus (P) to their host plants in return for photosynthate (Smith & Read, 1997). However, functional attributes of AM symbioses in wheat range from mutualism (Khan, 1975) to parasitism (Graham & Abbott, 2000), as also observed in many other plant species (Johnson et al., 1997). The contribution of AM fungi to P uptake by positively responsive plants can be easily identified by comparison of P uptake in AM plants and nonmycorrhizal (NM) plants. However, this approach cannot be used in the absence of positive responses, and it is usually then assumed that the fungi do not contribute P.
Work with compartmented pots combined with 32P or 33P has shown that AM wheat absorbed significantly more isotopic P compared with NM plants, regardless of growth responses (Ravnskov & Jakobsen, 1995; Hetrick et al., 1996). Field-grown wheat can also obtain P via AM fungi (Schweiger & Jakobsen, 1999). Likewise, a tomato variety, which had no response to AM colonization in terms of either vegetative growth or total P uptake, could (depending on the AM fungus) gain nearly all of its P via AM fungal hyphae (Smith et al., 2003). Molecular-genetic studies support the operation of the AM uptake pathway by demonstrating that plant P transporters, operating in the mycorrhizal uptake pathway, are induced by AM fungi (Karandashov & Bucher, 2005). Mycorrhiza-inducible P transporter genes have been reported in wheat (Glassop et al., 2005). However, quantitative estimates of P uptake via hyphae in wheat have not been obtained.
It is widely believed that percentage colonization by AM fungi in wheat is low, but this is not necessarily so. Li et al. (2005) showed that wheat was highly colonized in a soil from the Eyre Peninsula, South Australia. The soil is highly calcareous, P-fixing and used for cereal production. Despite the high total soil P, crops often suffer severe P deficiency and have low responsiveness to the application of conventional granular P fertilizers (Holloway et al., 2001). The P extracted by bicarbonate using the Colwell method (Colwell, 1963; referred to as Colwell P later) is high, and often exceeds published values of the critical concentration (21 mg kg−1) required for wheat grown in the South Australian wheat belt (Reuter et al., 1995). However, Colwell P is a poor predictor of plant-available P determined from plant growth and P uptake, because bicarbonate releases P from this soil that is not plant-available (Bertrand et al., 2003). The P extracted by anion exchange resin strips (McLaughlin et al., 1994; referred to as resin P later) is very low, but may better predict the P availability for plants than Colwell P. Few studies have compared the P extracted by the different methods with plant growth and no information has clearly suggested an appropriate P-extraction method to determine plant-available P for this soil. Research has shown that liquid fertilizers are much more efficient sources of P to crop plants than granular fertilizers and have a large potential to be used for agricultural production (Holloway et al., 2001; McBeath et al., 2005). The mechanism underlying the advantage of using liquid rather than granular fertilizers has been suggested to be further diffusion of P from the point of application, and a larger amount of P remaining potentially available (Lombi et al., 2004).
Previous studies on interactions of wheat and AM fungi in this soil have shown them to be strongly influenced by P supply. AM fungi increased grain yield and P uptake at maturity when a moderate resin P concentration was achieved by application of P (Li et al., 2005). However, a growth depression was observed at the early growth stage. We previously demonstrated that considerable amounts of 32P were delivered into wheat via AM fungi in the absence of growth responses, but could not quantify P uptake by AM fungi because of the difficulties of determining the soil P pool available to plants (Li, 2005). The aim of this study was therefore to quantify the contribution of AM fungi to P uptake by wheat grown in the calcareous and P-fixing soil from the Eyre Peninsula, South Australia, and to determine if this contribution was influenced by the form of P fertilizer applied.
Materials and Methods
The experiment had a complete factorial design with two AM treatments (inoculated with AM fungi or not) and five P treatments (no additional P or 20 mg kg−1 P added as four P sources). There were five replicate pots in each treatment. Two noninoculated and unplanted pots in each treatment were set up and used to check the specific activity of 32P in the extractable P in the soil at the end of experiment.
Compartmented pots and growth medium
The compartmented pot system, which was the same as used by Smith et al. (2003, 2004), is briefly described here. The main compartment (root + hyphal compartment; RHC) was a plastic pot, containing 1354 g of growth medium (including AM inoculum or not, see the ‘Plant and fungus’ section later). The hyphal compartment (HC) was a small plastic vial, containing 46 g of the same soil (3.4% of the total), labelled with carrier-free PO4 and capped with 30 µm nylon mesh, which allowed fungal hyphae, but not roots, to penetrate from the RHC and absorb P. A layer of unlabelled soil (6 g) was placed on the surface of the 32P-labelled soil as a buffer zone to prevent root hairs from accessing 32P through the mesh and to minimize diffusion of 32P out of the HC. The HC was placed horizontally, 5.5 cm below the soil surface within the RHC, with the mesh towards the centre of the pot.
The soil used in this study was collected from Cungena, Eyre Peninsula, South Australia. It is a highly calcareous grey sandy loam, and classified as a Calcixerollic Xerochrept in the Soil Taxonomy System (Soil Survey Staff, 1994). The soil is P-deficient and highly P-fixing, and has been used in previous studies to examine P-fertilizer efficiency (Holloway et al., 2001) and AM function in wheat (Li et al., 2005). The soil had a pH of 8.0 (0.01 m CaCl2) and contained 320 mg kg−1 total P, 39 mg kg−1 Colwell P, 5 mg kg−1 resin P, 35.5% CaCO3 and 0.86% organic carbon. The soil was air-dried, passed through a 2 mm sieve and diluted with sand to produce a 70 : 30 soil : sand mix as the growth medium. This allowed better extraction of roots that were cleaner than with 100% soil (Li, 2005). Soil and sand were autoclaved separately at 121°C for 1 h twice over a 3 d period before mixing. The soil : sand mix is referred to as soil later. Nutrients were thoroughly incorporated into the soil at the following rates (mg kg−1 dry soil): K2SO4, 75; CaCl2·2H2O, 75; MgSO4·7H2O, 45; CuSO4·5H2O, 2.1; ZnSO4·7H2O, 5.4; MnSO4·7H2O, 10.5; CoSO4·7H2O, 0.39; Na2MoO4·2H2O, 0.18; H3BO3, 0.3; Fe EDTA, 0.4.
Phosphorus was added as a powder of CaHPO4 or as solutions of ammonium polyphosphate (APP; 14 : 21 of N : P; w/v%), H3PO4 or Na3PO4·12H2O, to give 20 mg P per kg. The aim was to simulate application of conventional and fluid fertilizers used in the field. The P was thoroughly mixed into the soil for both RHC and HC. These treatments are referred to as P0, P-Ca, P-APP, P-H and P-Na. Soil for the HCs was well mixed with carrier-free PO4 at 17 kBq g−1 dry soil. This addition of 32P had no effect on soil P concentration. Total and available P were the same in HC and RHC. An extra 100 g of soil with each fertilizer addition was prepared for measuring available P and specific activities of 32P extracted by bicarbonate (Colwell, 1963) or anion exchange resin strips (McLaughlin et al., 1994). This soil was contained in a plastic vial, wetted to the same water contents (12.5%) as the soil in pots, covered with a lid and incubated at room temperature. The water content was maintained by weighing every week. This soil was analysed for extractable P and 32P 2 wk after planting and again at the same time as the soil in HCs of nonplanted pots at harvest. P0 soil contained 30 mg kg−1 Colwell P and 4 mg kg−1 resin P. Additional P resulted in 45–47 mg kg−1 Colwell P and 10 mg kg−1 resin P, except for P-APP, which gave 8 mg kg−1 resin P. The extractable 32P in soil ranged between 13 and 16 kBq g−1 dry soil for the Colwell method and 4 and 5 kBq g−1 for resin, with small differences between different P treatments for each method.
Plant and fungus
Spring wheat (Triticum aestivum L.) cv. Brookton was used. Seed of uniform size was selected and it contained c. 98 µg P per seed. Seed was surface-sterilized in 3% sodium hypochlorite solution for 10 min, rinsed with reverse osmosis (RO) water and germinated on moist paper for 3 d. Two seedlings were sown in each pot and thinned to one per pot after emergence.
The AM fungus was Glomus intraradices Schenck and Smith (DAOM 181602), grown on Allium porrum L. Ten per cent of dry inoculum was incorporated into soil in the AM treatments. The inoculum consisted of colonized root fragments and soil containing spores and external hyphae of the AM fungus. Sterile soil/sand mix, as used for the inoculum culture, was added to the NM pots. To establish a similar bacterial community in both AM and NM plants, 100 g of pot culture inoculum was suspended in 500 ml water and shaken for 5 min, then filtered through Whatman no. 1 (11 µm) filter paper. The filtrate (5 ml per pot) was added to all pots.
Growth conditions and harvesting
For treatments other than APP, nitrogen was added as 50 mg N kg−1 (NH4NO3) at the start of the experiment and again after 3 wk (total 100 mg N kg−1). For APP treatments, less N was added to the pots at the beginning to take the N in APP into account. Soil water content was maintained at 12.5% (w/w) by watering with RO water daily, taking into account estimated increases in plant fresh weight during the experiment. The pots were arranged randomly on one bench in the glasshouse and were rearranged when watered. Growth was monitored regularly throughout the experiment and appearance of 32P in shoots of the plants was followed (nonquantitatively) using a hand-held radiation monitor. The experiment was conducted from 25 October to 1 December 2005, with natural light and a diurnal temperature range between 16 and 22°C. Plants were harvested 36 d after sowing.
At harvest the shoots were cut off, washed, weighed and used for determination of dry weight (24 h, 80°C), P content and specific activity of 32P (32P/total P). The HCs were carefully removed from the pot and the surface soil used for the buffer zone cut away. The remaining soil in the HC was well mixed. Small weighed samples of soil were taken from the RHC and HC for determination of hyphal length density (HLD). Fresh subsamples of soil from HC in nonplanted pots were taken for checking extractable P and specific activity of 32P. Roots from the RHC were washed and weighed. A subsample of known weight was taken for measurement of AM colonization. The remaining roots were dried and root dry weight was measured. Total root dry weight was calculated from the fresh weight/dry weight ratio of this sample. Dried ground material was digested in a solution of nitric-perchloric acid (6 : 1, v/v).
The shoot and root P concentrations of plants were determined using the phosphovanado-molybdate method (Hanson, 1950). The 32P in plant digests and soil extracts was measured on a Packard TR 1900 liquid scintillation counter and corrected for isotopic decay. AM colonization was determined by the method of McGonigle et al. (1990), following clearing in 10% KOH and staining in Trypan Blue by a modification of the method of Phillips & Hayman (1970), omitting phenol from the reagents. The HLD in HCs and RHCs was determined on well mixed soil samples (Jakobsen et al., 1992).
Calculations and data analysis
The percentage contribution of the mycorrhizal pathway to the total P uptake was calculated from the specific activity of 32P in the HCs and in the plants and the total extractable P in the RHCs and the HCs measured by the resin or Colwell methods:
(SA, specific activity; P, available P extracted by the resin or Colwell method.) Data (see ‘Results’ section) showed that the assumption in the equation that distribution of HLDs is uniform throughout the soil in both RHCs and HCs was valid. We also assumed that P extracted by either the Colwell or resin methods was equally available to plant and fungus. The calculation ignored the small contribution of seed P to total plant P.
Data are presented as means and standard errors of means of five replicates. Data were analysed by analysis of variance (anova) or a two-sample t-test (paired) using GenStat Release 6.1, Lawes Agricultural Trust (Rothamsted Experimental Station, UK). Comparisons between means are based on the least significant differences at the 0.05 probability level.
Arbuscular mycorrhizal colonization and HLD
No colonization occurred in uninoculated plants (Fig. S1). AM colonization in pots with inoculum was well established at harvest, with 57% root length colonized at P0 (Table 1). Application of P fertilizers decreased colonization to between 35 and 43%, with very small differences between different P forms. HLDs in NM pots were approx. 0.18 m g−1 dry soil, which provided a background value of dead hyphae, with no significant differences between RHCs and HCs of each pot and different P treatments. In inoculated pots, HLDs in RHCs or HCs also decreased with P additions, with no differences in HLDs between the two compartments of each pot (t-test) and between the four fertilizers (Table 1).
Table 1. Percentage arbuscular mycorrhizal (AM) colonization of wheat (Triticum aestivum) plants and hyphal length densities (HLDs) in soil in the root + hyphal compartment (RHC) and hyphal compartment (HC) of pots inoculated with Glomus intraradices
P treatments: no additional P (P0) or 20 mg kg−1 P added as CaHPO4 (P-Ca), ammonium polyphosphate (P-APP), H3PO4 (P-H) and Na3PO4·12H2O (P-Na). Values are means of five replicates ± SEM. Values with the same letter in each row are not significantly different (P < 0.05).
57 ± 6c
35 ± 4a
38 ± 6ab
43 ± 4b
36 ± 6ab
HLD (m g−1)
1.4 ± 0.1b
1.0 ± 0.3a
1.0 ± 0.3a
1.0 ± 0.2a
1.0 ± 0.2a
1.5 ± 0.2c
1.2 ± 0.3abc
1.1 ± 0.1ab
1.0 ± 0.2ab
0.9 ± 0.1ab
Plant growth and P uptake
Application of P significantly increased both shoot and root DW, regardless of AM colonization and P forms (Fig. 1a,b). Shoot DW was not significantly affected by G. intraradices inoculation in P0 and P-Ca treatments, but was depressed with P-APP, P-H and P-Na (P < 0.001; Fig. 1a). Root DW was not changed by colonization except with P-Na, where there was a small negative response (P < 0.005; Fig. 1b).
Shoot P concentrations ranged between 2.1 and 2.5 mg g−1 and were only slightly affected by either AM colonization or P additions, although there was a slight but significant increase in AM plants when P-APP was applied (results not shown). Additional P significantly increased root P concentrations (from 1.2 to 1.6 mg g−1) in NM plants only, with small differences between the different P forms. AM colonization increased root P concentrations at P0, but not when P was added (results not shown). It follows from these results that total P uptake of both NM and AM plants was greatly increased by additional P, regardless of form of P added (Fig. 2). The P uptake with P-H or P-Na was slightly higher than with P-APP in NM plants, but the same in AM plants. Colonization by AM fungi did not change total P uptake at P0, but decreased it by approx. 15% when P was applied except for P-Ca, which gave the same P uptake by AM and NM plants.
32P uptake and contribution of the AM pathway to plants
No 32P was detected in any NM plants during growth and after harvest, demonstrating that the mesh and buffer zone effectively prevented the roots accessing the HCs and 32P diffusing to the RHCs. The 32P had appeared in shoots of all AM plants at 14 d after planting, as detected by a hand-held radiation monitor. At harvest, large amounts of 32P were measured in AM plants, with values at P0 higher (170 kBq per plant) than those when P was applied (120–150 kBq per plant), in line with the different specific activities (see below).
Parameters for calculating the contribution of AM fungi to P uptake by plants are presented in Table 2. Results showed that application of P significantly decreased SA of 32P in plants, with no difference between different P fertilizers. Trends of soil 32P SA with different P treatments were similar to those in plants. However, soil 32P SA was lower using Colwell P than resin P as expected from different amounts of P extracted by the different methods (see ‘Materials and Methods’ section). There were no significant differences in soil 32P SA of each treatment between 2 wk after planting and harvest, or between the extra soil incubated and the soil from nonplanted pots of each treatment (results not shown). In other words there was rapid equilibration of added P and 32P in the soil.
Table 2. Parameters for calculating percentage P uptake via hyphae in wheat (Triticum aestivum) plants inoculated with Glomus intraradices: 32P specific activity (SA) in plant and in hyphal compartment (HC) soil all at harvest (36 d)
32P SA (kBq mg−1 P)
The P was extracted by bicarbonate (Colwell) and anion exchange resin strips (resin). Soil P treatments: no additional P (P0) or 20 mg kg−1 P added as CaHPO4 (P-Ca), ammonium polyphosphate (P-APP), H3PO4 (P-H) and Na3PO4·12H2O (P-Na). Values are means of five replicates ± SEM. Values with the same letter in each row are not significantly different (P < 0.05).
28 ± 2b
10 ± 1a
10 ± 2a
10 ± 2a
12 ± 2a
437 ± 51b
310 ± 28a
313 ± 21a
342 ± 34a
316 ± 27a
1106 ± 50c
490 ± 14a
610 ± 56b
534 ± 62a
497 ± 43a
The percentage P uptake via hyphae was calculated from Eqn 1 and results are shown in Fig. 3. Values obtained by the two methods showed the same trends, with those based on Colwell P being much higher than those based on the resin P. The contribution of AM fungi to total P uptake by wheat was high at P0 and decreased with P additions, with small differences between different P forms.
Arbuscular mycorrhizal colonization and HLD
Arbuscular mycorrhizal colonization of wheat by G. intraradices was well established and quickly achieved degrees (∼57%) that are realistic compared with the natural colonization of wheat grown in the field, where it was up to 80% at the end of tillering (Li et al., 2005; H. Li, unpublished). These results show that the potential for colonization in wheat grown in soils from this growing region is high and that the general belief that wheat is usually not highly colonized is an oversimplification. Percentage AM colonization was inversely related to the P supply, as frequently shown in previous experiments with wheat (Baon et al., 1992; Mohammad et al., 2004; Li et al., 2005) and for many other species (Smith & Read, 1997). However, it is increasingly appreciated that percentage AM colonization of roots is not strongly correlated with growth responses and P uptake (Hetrick et al., 1992; Smith et al., 2004; Li et al., 2005). The trend for effects of P additions and forms on the HLD was similar to those for the colonization. Lack of differences in HLD between the RHCs and HCs of each pot indicated that distribution of the hyphae in soil was even, as we assumed in the calculation of the contribution of hyphal P uptake.
Plant growth and P uptake
In this study, plants showed large responses to P additions, regardless of forms of P added, indicating that the soil is P-deficient, as previously observed (Holloway et al., 2001; McBeath et al., 2005). Wheat showed no positive responses to AM fungi, as was also found in our previous experiments (Li et al., 2005; Li, 2005) at the early growth stage (∼6 wk) and in research by others (Graham & Abbott, 2000; Zhu et al., 2001). As with plant growth, total P uptake of plants was significantly increased by P additions, with no differences between the different forms of P in NM plants, but slightly lower with the liquid P (P-APP, P-H and P-Na) than the solid P (P-Ca) in the AM plants. Overall, AM fungi did not show any apparent contribution in terms of net increase in plant growth or total P uptake in the early growth stages investigated here, regardless of P additions and forms. Although vegetative growth is normally used as an index to evaluate the effectiveness of AM fungi on crops, it is a poor predictor of reproductive growth. In our previous study, the growth depression-induced by AM colonization at the early growth stage disappeared at maturity (Li et al., 2005).
32P uptake and contribution of the AM pathway to plants
As in the preliminary studies (Li, 2005), no 32P appeared in any NM plants, but considerable amounts of 32P were measured in all AM plants, indicating that hyphae were delivering P to plants in the absence of growth responses. The results show that lack of growth or P responses of plants does not necessarily mean that AM fungi have made no contribution to P uptake. The findings are supported by previous work on wheat (Ravnskov & Jakobsen, 1995; Hetrick et al., 1996; Schweiger & Jakobsen, 1999; Li, 2005) and other nonresponsive species, including cucumber (Pearson & Jakobsen, 1993). However, these studies could not quantify hyphal contribution as done here, because there was no measurement of specific activity of 32P in the soils. The findings are relevant to agricultural practices since, based on the total growth and P uptake of NM and AM plants, as most commonly used before, function of AM fungi in P uptake by nonresponsive plants such as wheat has often been underestimated or even dismissed as unimportant. Compartmented pots with 32P or 33P are valuable approaches for directly measuring the relative contribution of hyphae and roots to the total uptake of P by the mycorrhizal plants. Smith et al. (2003, 2004) quantified the hyphal P uptake from different AM fungi with both responsive (medic and flax) and nonresponsive (tomato) plants, as done here. The finding that nonresponsive plants such as wheat obtain a large proportion of P via AM fungi highlights the fact that direct root uptake must be reduced, because total P uptake is not increased.
The approach is now being combined with studies of plant P transporter gene expression (Poulsen et al., 2005). The measurement of the P availability in soil is a key factor in the accurate assessment of the hyphal contribution, particularly in this highly calcareous and P-fixing soil from the Eyre Peninsula, South Australia. The Colwell P in this soil unamended with P was high, but clearly unreliable to predict P available for plants, because all plants responded to P additions. The resin P was much lower and better reflected the poor availability of P to plants. Application of P to the soil significantly increased both Colwell and resin P. However, approx. 80% of applied P or 32P in the soil could be extracted by the Colwell method, whereas only approx. 30% was extracted by the resin method. These results confirm that the Colwell reagent dissolved the calcium-bound P from the soil and resulted in over-estimation of the available P as proposed by Bertrand et al. (2003). Because the different methods of P extraction gave different values of available P and specific activities of 32P, they greatly affected the final calculation of the hyphal P uptake. The calculation based on the Colwell P gave a high value of hyphal contribution of over 100% for P0 and P-Na. The unrealistic values obtained indicate that, as for roots, Colwell P is a poor measure of P available for uptake by hyphae. The contribution of AM fungi to P uptake was between 50 and 80% using the much more conservative specific activity of 32P in resin P. Interestingly, there were small or no decreases in P transferred by the AM pathway with P addition (Fig. 3; resin P method), although there were decreases in percentage colonization and HLDs (Table 1), showing that additional P had little or no effects on the contribution of the mycorrhizal pathway for P transport. Solid P fertilizer resulted in a slightly higher plant growth and P uptake than liquid P fertilizer in AM treatments, indicating there may be effects of different P forms on interactions between AM fungi and wheat. Liquid fertilizer did not show advantages for plant growth and P uptake compared with the solid. This contrasts with increased growth and P uptake when fluid fertilizers were compared with granular P fertilizers (Holloway et al., 2001; McBeath et al., 2005), probably because in our experiment all forms of P applied had been well mixed with soil for the purpose of obtaining uniform 32P labelling and available P. This procedure was essential for calculation of AM contribution of P uptake, but differences in diffusion between different forms of P (which has been shown to underlie positive effects of fluid fertilizers in the field) were likely have been minimized. This highlights one difficulty of using our method to make realistic comparisons of fertilizers in the field, where application is not uniform.
In conclusion, this work shows that wheat grown in a highly calcareous soil from a major cereal-growing area, the Eyre Peninsula, was well colonized by AM fungi. The mycorrhizal pathway of P delivery to roots was highly active, contributing 50–80% of plant P, depending on the P treatment. In consequence, the pathway should not be ignored in research aimed to improve the efficiency of P utilization by crops, for example, by breeding for changed root traits. It is also likely to be significant in plant competitive interactions in natural ecosystems that involve ‘nonresponsive’ AM plants (Smith & Smith, 2005).
We thank the South Australia Grains Industry Trust (project no. 05SAGIT_UA1/05) for providing funding to support this research programme. We also thank Dot Brace (Minnipa Station, South Australia Research and Development Institute) for providing soil, Debbie Miller and Colin Rivers (The University of Adelaide), and Gill Cozens (CSIRO Land and Water Division, Adelaide) for technical support.