Mycorrhizal phosphate uptake pathway in tomato is phosphorus-repressible and transcriptionally regulated


  • Réka Nagy,

    1. Federal Institute of Technology (ETH) Zurich, Institute of Plant Sciences, Experimental Station Eschikon 33, CH-8315 Lindau, Switzerland;
    Search for more papers by this author
  • David Drissner,

    1. Federal Institute of Technology (ETH) Zurich, Institute of Plant Sciences, Experimental Station Eschikon 33, CH-8315 Lindau, Switzerland;
    Search for more papers by this author
  • Nikolaus Amrhein,

    1. Federal Institute of Technology (ETH) Zurich, Institute of Plant Sciences, Universitätstrasse 2, CH-8092 Zurich, Switzerland;
    Search for more papers by this author
  • Iver Jakobsen,

    1. National Laboratory for Sustainable Energy (Risø DTU), Technical University of Denmark, Biosystems Department, Frederiksborgvej 399, DK-4000 Roskilde, Denmark;
    Search for more papers by this author
  • Marcel Bucher

    1. Federal Institute of Technology (ETH) Zurich, Institute of Plant Sciences, Experimental Station Eschikon 33, CH-8315 Lindau, Switzerland;
    2. Present address: University of Cologne, Institute of Botany, Gyrhofstrasse 15, D-50931 Cologne, Germany
    Search for more papers by this author

Author for correspondence:
Marcel Bucher
Tel:+49 221 4702481
Fax:+49 221 4705900


  • • Plants colonized by arbuscular mycorrhizal (AM) fungi take up phosphate (Pi) via the mycorrhizal and the direct Pi uptake pathway. Our understanding of the molecular mechanisms involved in the regulation of these pathways is just emerging.
  • • Here, we have analyzed the molecular physiology of mycorrhizal Pi uptake in the tomato (Solanum lycopersicum) variety Micro-Tom and integrated the data obtained with studies on chemical signaling in mycorrhiza-inducible Pi transporter gene regulation.
  • • At high plant phosphorus (P) status, the mycorrhizal Pi uptake pathway was almost completely repressed and the mycorrhiza-inducible Pi transporter genes were down-regulated. A high plant P status also suppressed the activation of the mycorrhiza-specific StPT3 promoter fragment by phospholipid extracts containing the mycorrhiza signal lysophosphatidylcholine.
  • • Our results suggest that the mycorrhizal Pi uptake pathway is controlled at least partially by the plant host. This control involves components in common with the regulatory mechanism of the Pi starvation response pathway in plants.


Plants take up phosphorus (P) as soluble orthophosphate (Pi). Although the total amount of P in soil may be high, most of it is strongly adsorbed to soil components or chemically fixed, thereby rendering concentrations of free Pi in the soil solution low (c. 1 µm) and often growth-limiting (Marschner, 1995). Plants have developed several adaptive strategies to overcome this Pi deprivation stress, among them the association of their roots with soilborne arbuscular–mycorrhizal fungi (AMF). About 80% of the terrestrial plant species are known to be mycorrhizal (Smith & Read, 2007). Their extraradical fungal mycelium extends the root system, thus giving the host plant access to Pi from the soil volume of the mycorrhizosphere which greatly exceeds that of the rhizosphere proper. AMF are obligate biotrophs, and bidirectional nutrient exchange between the two partners is a key feature for the functionality of the symbiosis. Growth of the AMF depends on photosynthesis-derived carbon, and mineral nutrients, in particular Pi, are transferred in return from the fungus to the plant (Smith & Read, 2007).

As worldwide Pi reserves are finite and, furthermore, as P in fertilizers and manure can have negative environmental effects, Pi sources must be used efficiently and in a sustainable way (Sims & Sharpley, 2005). Therefore, the study of the contribution of AMF to plant P nutrition represents an important area of research in plant mineral nutrition and crop productivity (Zhu et al., 2001; Vance et al., 2003). The beneficial role of the symbiosis is more apparent in plants grown at low P conditions. Ample P supplies reduce the extent of root colonization and delay the development of intraradical fungal structures (Amijee et al., 1989, 1993; Bruce et al., 1994). Delivery of Pi to the plant by the extraradical fungal network (Jakobsen et al., 1992a) and the formation of intracellular fungal structures in the root cortex, which trigger the expression of mycorrhiza-inducible Pi transporters (Harrison et al., 2002; Karandashov & Bucher, 2005; Nagy et al., 2005), are prerequisites for a functional symbiosis.

At the soil-root interface, Pi can be absorbed through the epidermis, including root hairs, via the direct Pi uptake (DPU) pathway, while the AM fungal mycelium and colonized cortical cells are involved in the mycorrhizal Pi uptake (MPU) pathway (Smith et al., 2003). Both pathways operate through fungal and/or plant plasma membrane-intrinsic transporter proteins which transfer Pi anions from the apoplast to the cytoplasm against a steep concentration gradient (Karandashov & Bucher, 2005). While Pi transporters of the DPU pathway are expressed either constitutively or in response to Pi deprivation (Leggewie et al., 1997; Daram et al., 1998; Liu et al., 1998; Chiou et al., 2001; Mudge et al., 2002; Gordon-Weeks et al, 2003; Zimmermann et al., 2003), the plant transporters of the MPU pathway are exclusively or predominantly expressed at the plant–fungus interface in colonized roots (Rausch et al., 2001; Harrison et al., 2002; Paszkowski et al., 2002; Karandashov et al., 2004; Nagy et al., 2005; Chen et al., 2007; Xu et al., 2007). Mycorrhiza-inducible Pi transporters play a key role in total plant Pi uptake (Maeda et al., 2006; Javot et al., 2007). The fungal transporter genes expressed in the extraradical mycelium are strongly up-regulated under Pi-limited conditions (Maldonado-Mendoza et al., 2001), similar to the plant Pi transporter genes involved in the DPU.

Plant-AMF combinations are characterized by a remarkable functional diversity (Smith et al., 2004). Plants differ in their responsiveness to the symbiosis with respect to growth, P content, and degree of colonization, while AMF differ in hyphal length density, their spatial distribution in the soil, and their capacity to take up and translocate Pi (Jakobsen et al., 1992a,b; Ravnskov & Jakobsen, 1995). Physiological analyses of various plant-AMF combinations have uncovered reduced activity of the DPU pathway in roots colonized by AMF (Pearson & Jakobsen, 1993; Smith et al., 2003; 2004). Detailed analysis of loss-of-function mutants in mycorrhizal plants has allowed an assessment of the distinct roles of two Pi transporters in MPU in legumes (Maeda et al., 2006; Javot et al., 2007).

In a changing root environment, DPU and MPU pathways can be expected to be finely tuned via interactive signal transduction pathways. The function of lysophospholipids as signals in plants, and in particular in the AM symbiosis, was recently highlighted when lysophosphatidylcholine (LPC) was identified as the active component in mycorrhizal phospholipid extracts which elicit the expression of the mycorrhiza-inducible Pi transporter genes StPT3 and StPT4 in potato, and of LePT4 in tomato, respectively (Drissner et al., 2007). It is tempting to speculate that LPC is involved in the regulation of the MPU pathway via Pi transporter gene expression, yet experimental data supporting this hypothesis are presently missing.

Tomato contains three mycorrhiza-inducible Pi transporter genes, LePT3, LePT4, and LePT5, belonging to two subfamilies of the Pht1 family (Nagy et al., 2005). Of these, LePT4 has been investigated in most detail. It is expressed in cortical cells colonized by AMF (Nagy et al., 2005), and initial work using the lept4-1 mutant of tomato (Solanum lycopersicon cv. Micro-Tom; Scott & Harbaugh, 1989) revealed considerable functional similarity among the mycorrhizal Pi transporters during conditions of P deprivation (Nagy et al., 2005). Later, Xu et al. (2007) demonstrated that LePT4 function is not fully dispensable, suggesting that mycorrhizal Pi transporters may have similar, but only partially redundant functions in tomato.

The aim of the present work was to test the hypotheses that high P concentrations repress not only mycorrhiza formation but also the MPU pathway in solanaceous species; and that high P concentrations counteract the induction of the mycorrhizal Pi transporter genes by LPC-containing extracts from mycorrhizal roots.

Materials and Methods

Plant and fungal material

The plant material used was tomato (Solanum lycopersicum L., cv. Micro-Tom; Scott & Harbaugh, 1989) and transgenic potato (Solanum tuberosum L., cv. Désirée) carrying the pStPT3-GUS construct (Rausch et al., 2001). The AMF used were two isolates of Glomus intraradices, BEG87 and BEG75, which were propagated in pot cultures with Trifolium subterraneum L. (subterranean clover), and Plantago lanceolata L. (narrowleaf plantain), respectively.

Experimental setup

One experiment (Figs 1, 2 and Supporting Information Fig. S1) was carried out at Risø Technical University of Denmark (DTU), Denmark, using the bi-compartmented system and the same design of root + hyphal compartments (RHC) and hyphal compartments (HCs) as described previously (Smith et al., 2004). The growth substrate was an irradiated (10 kGy, 10 MeV electron beam) 1 : 1 soil : sand mixture containing 9 mg extractable P kg−1 soil (Olsen et al., 1954) and supplied with basal nutrients (Pearson & Jakobsen, 1993) and 30 mg NH4NO3-N kg−1. Three soil P concentrations were established by mixing KH2PO4 into the growth medium at either nil (P0), 20 (P20) or 60 (P60) mg P kg−1. The P treatments were with or without inoculation with G. intraradices BEG87, and each of the six treatments had five replicate pots with 1000 g soil in RHC and 27 g in HC. Preparation of RHC soil for mycorrhizal pots involved the thorough mixing of 920 g irradiated growth substrate and 80 g inoculum consisting of dry soil, containing spores and colonized root fragments from clover. There was no inoculum in the HCs. Soil for the HCs had the same P concentration as in the RHC and was well mixed with carrier-free H333PO4 at 5 kBq kg−1 growth medium to provide 121.6, 82.9 and 39.2 kBq mg−1 bicarbonate-extractable P. Micro-Tom seeds were germinated on moist paper and five seedlings were planted in each pot. These were thinned to three per pot after establishment, and during cultivation NH4NO3-N was supplied to provide an additional total of 100 mg N per pot.

Figure 1.

Root length (a), root colonization (b) and length of hyphae in soil (c) as influenced by inoculation with Glomus intraradices and by supply of 0, 20 or 60 mg P kg−1soil (P0, P20 and P60) to pots with Micro-Tom tomato (Solanum lycopersicum) plants. Light and dark gray bars in (a) and (c) represent data from pots with nonmycorrhizal and mycorrhizal conditions, respectively. Means + SE, and bars with the same letter in (a) and (c) are not significantly different (P < 0.05). In (b), identical comparison of means was made separately for each colonization variable (black bars, total; gray bars, arbuscules; white bars, vesicles).

Figure 2.

Root (a) and shoot (b) concentrations, specific activity of 33P (c), and P uptake via the mycorrhizal phosphate uptake pathway (d) in Micro-Tom tomato (Solanum lycopersicum) plants as influenced by inoculation with Glomus intraradices and by supply of 0, 20 or 60 mg P kg−1 soil (P0, P20 and P60). Light and dark gray bars represent data from pots with nonmycorrhizal and mycorrhizal conditions, respectively. Means + SE, and bars with the same letter within each variable are not significantly different (P < 0.05).

Plants were maintained in a growth room with a 16 : 8 h light : dark cycle at 23 : 18°C, respectively. Osram daylight lamps provided 500 µmol m−2 s−1 PAR (400–700 nm). Plants were watered daily to maintain 65% of the water-holding capacity and were harvested 6 wk after planting. Dry weight (DW), P content and 33P content of shoots and roots as well as root colonization and length of hyphae in soil were measured (see later).

The second experiment (Fig. 3) comprised Micro-Tom plants grown in pots of 1 l volume containing a soil : quartz-sand mixture (1 : 9) supplied (+Myc) or not (−Myc) with dry plantain roots colonized by G. intraradices (BEG75). Micro-Tom seeds were germinated on wet sand, and one seedling was planted in each pot. Fertilization was by drop irrigation with one quarter-strength Hoagland's nutrient solution (Hoagland & Broyer, 1936) containing 5 µm NH4H2PO4. Plants were harvested 4 or 6 wk postinoculation. For high P treatments (+P), the quarter-strength Hoagland's nutrient solution contained 500 µm NH4H2PO4. Randomly chosen root samples were stained with trypan-blue (Phillips & Hayman, 1970) and analyzed for the presence of fungal structures. The remainder of the root system was used for RNA isolation. Pi content in leaves was measured just before harvest as described later.

Figure 3.

Expression of phosphate transporter genes and plant growth in response to arbuscular mycorrhizal fungal colonization and phosphate availability. (a) Expression of Pi transporters in roots and mycorrhizas of Micro-Tom. RNA gel blot analysis was performed with 10 µg of total RNA isolated from roots of tomato (Solanum lycopersicum) plants grown at low P (−P/5 µm NH4H2PO4) or high P (+P, 500 µm NH4H2PO4) for 4 or 6 wk, respectively, in the presence (+Myc) or absence (−Myc) of Glomus intraradices as indicated. Fluorescence of ribosomal RNA (rRNA) in the presence of ethidium bromide reflects equal loading of the gel, and the abundance of rRNA in the gel correlated well with that on the blot (data not shown). Hybridization to RNA on the blot was performed with radiolabeled DNA probes corresponding to the respective 3′ regions of the two tomato Pi transporter genes as indicated on the left. (b) Pi concentration in leaves as a function of the degree of colonization and P fertilization. 5 µm NH4H2PO4 represents –P conditions, while 500 µm NH4H2PO4 represents +P conditions. Plants were grown for 4 or 6 wk, respectively, in the presence (+Myc) or absence (−Myc) of G. intraradices. Values represent means (+ SE) of measurements for three plants; bars with different letters indicate significant differences (P < 0.05). (c) Shoot fresh weight of Micro-Tom grown at low P (−P, 5 µm NH4H2PO4) or high P (+P, 500 µm NH4H2PO4) for 4 and 6 wk, respectively, in the presence (+Myc) or absence (−Myc) of G. intraradices.

In the third experiment (Fig. 4), transgenic potato carrying an StPT3:: GUS chimeric gene was grown in a soil : quartz-sand mixture (1 : 9) without mycorrhization and watered with half-strength Hoagland's solution containing 0, 5, or 500 µm NH4H2PO4, respectively. N concentrations were adjusted with NH4Cl. Plants without mycorrhizal colonization were harvested 6 wk after planting. The transgenic root bioassay designed to assess the response of roots to mycorrhiza signals, that is, activation of the StPT3 promoter, was performed as described by Drissner et al. (2007).

Figure 4.

Detection of lysophosphatidylcholine (LPC) in a phospholipid extract from mycorrhizal roots, and GUS activity in transgenic potato (Solanum tuberosum) roots treated with such an extract (PL+myc) as a function of Pi supply. (a) Positive-ion MALDI-TOF mass spectrum of LPC in the PL+myc extract used in (b). Numbers indicate the peak position (m/z) of the protonated form of different LPC molecular species. Species 1 is LPC 16 : 1; 2, LPC 16 : 0; 3, LPC 18 : 3; 4, LPC 18 : 2; 5, LPC 18 : 1; and 6, LPC 18 : 0. (b) Roots were treated with PL+myc for 3 h, and roots with GUS activity were scored (n = c. 100). Means + SD, n = 3. **, not detectable. Different letters indicate significant differences (P ≤ 0.05). P concentration in root (c) and shoot (d) tissue of plants used for the root bioassay shown in (b).

Physiological analysis of the plant material

Total P in the 33P experiment (Fig. 2) was determined by oxidation of dried root and shoot samples, respectively, in a nitric-perchloric acid mixture (4 : 1) and subsequent analysis using a Technicon autoanalyzer II (Technicon Autoanalyzers, Analytical Instrument Recycle, Inc., Golden CO, USA). 33P in shoots and roots was determined in the same extracts in a Packard TR 1900 liquid scintillation counter (Packard Bioscience Company, Meriden, CT, USA). Pi in tomato leaves in the Pi transporter gene expression experiment (Fig. 3) was measured according to Tausky & Shorr (1953), with minor modifications as described by Hurry et al. (2000). Shoot and root tissue P content in the transgenic potato experiment (Fig. 4) was determined colorimetrically using the malachite green phosphate detection method based on the malachite green-molybdate binding reaction (Ohno & Zibilske, 1991).

Roots in the 33P experiment were cleared in 10% KOH (20 min at 95°C) and stained with trypan blue (10 min at 95°C) essentially as described by Phillips & Hayman (1970), omitting phenol from the reagents and HCl from the rinse. The percentage of root length colonized, as well as the proportion of root length carrying arbuscules or vesicles were determined according to McGonigle et al. (1990). Hyphal length densities (HLDs) in dried 2 g samples of radiolabeled soil were determined by a grid intersect method as previously described (Jakobsen et al., 1992a). Contribution of the MPU pathway to Pi uptake was calculated using the formula of Smith et al. (2004): (SA 33P plant/SA 33P HC) × (P in pot/P in HC) × (HLD in RHC/HLD in HC) × 100, where SA is specific activity, P is bicarbonate-extractable P, RHC is root-hyphal compartment, and HC is hyphal compartment. Based on previous studies with G. intraradices (Smith et al., 2004), it was assumed that HLD was similar in the root and the hyphal compartments.

Infiltration of the transgenic roots carrying a StPT3:: GUS chimeric gene with a bioactive phospholipid extract from mycorrhizal P. lanceolata roots and subsequent histochemical staining for GUS activity was carried out as described in Drissner et al. (2007). Percentage of root tips with GUS stain after 3 h exposure to the phospholipid extract was scored. LPC species in the extract were identified using mass spectrometry as described later.

Extraction of nucleic acids and RNA gel blot analysis

Total RNA from tomato roots was isolated using the hot-phenol extraction method (Verwoerd et al., 1989). Ten micrograms of total RNA was subjected to electrophoresis on a denaturing formaldehyde gel with 1% agarose (w/v). RNA was blotted onto a Hybond NX nylon membrane (Amersham Pharmacia Biotech Europe GmbH, Dübendorf, Switzerland). Random-primed and 32P-labeled 3′ ends of LePT1, LePT2, StPT3 and LePT4 cDNAs served as probes. Hybridization with the labeled probes was carried out overnight at 68°C using 5× SSC, 5% Denhardt's, and 0.5% SDS (w/v). Membranes were sequentially washed in 2× SSC, 0.1% SDS; 1× SSC, 0.1% SDS; 0.1× SSC, 0.1% SDS at room temperature, with a final wash at 65°C in 0.1× SSC, 0.1% SDS, and subsequently exposed to Kodak BioMax films for autoradiography.

Matrix-assisted laser desorption-ionization time-of-flight/time-of-flight mass spectrometry (MALDI-TOF/TOF MS)

Lipid extraction and fractionation, as well as MALDI-TOF and MALDI-TOF/TOF analysis of the phospholipid fraction of mycorrhizas on a 4700 MALDI-TOF/TOF system (Applied Biosystems, Framingham, MA, USA), were performed as described (Drissner et al., 2007).

Statistical analysis

Data are presented as means and standard errors of the means of five replicates. Data were analyzed by analysis of variance (ANOVA) using STATISTICA (version 6, The significance of differences was determined by the Newman-Keul method for calculation of critical points.


Gradual repression of the MPU pathway with increasing Pi availability

The effect of Pi supply on the functional compatibility between Micro-Tom and G. intraradices was determined by employing the bi-compartmented pot system. Six-week-old Micro-Tom plants colonized or not by G. intraradices (BEG87) were analyzed for growth, the development of intra- and extraradical fungal structures, P content, and specific activity of 33P in shoots and roots as a result of Pi acquisition and transfer by the fungal symbiont. Growth of noninoculated plants correlated directly with increasing P availability (Fig. S1a). Colonization of plants with G. intraradices (BEG87) resulted in a significant growth depression at P0, but there were no differences in growth between colonized and noncolonized plants at the higher concentrations of P (Fig. S1). Similarly, root length was largely unaffected by the various soil P conditions (Fig. 1a). Fungal structures were absent in stained roots of noninoculated plants, whereas plants inoculated with G. intraradices (BEG87) exhibited Arum-type colonization (Gallaud, 1905; Dickson et al., 2007), consisting of intercellular hyphae and intracellular arbuscules (data not shown). Root colonization in inoculated plants was highest at low P and was decreased by 72% in plants supplied with 60 mg P kg−1soil (Fig. 1b). The percentage root length with arbuscules or vesicles also decreased with increasing P supply, but interfaces for nutrient exchange were presumably functional also at the high P supply, at which 14% of the root length contained arbuscules. Increased soil P concentration also reduced the HLD in the HC, but this effect was smaller than the reduction in the percentage of root length colonized (Fig. 1c). The HLD in soil from pots with nonmycorrhizal plants was low, unaffected by soil P concentration and probably represented decaying fungal hyphae (saprotrophic or AMF).

Root and shoot P concentrations of nonmycorrhizal plants increased markedly with increasing P supply (Fig. 2a,b). In mycorrhizal plants, P concentrations showed a similar pattern for shoots, but not for roots, where P concentration was high also at the lowest P supply.

Phosphate uptake via the MPU pathway was quantified as the amount of 33P-orthophosphate taken up from soil accessible only to extraradical fungal hyphae. Any 33P in the plant would therefore have been delivered exclusively via the MPU pathway. Accordingly, specific activities of 33P were always negligible in noninoculated plants, but high in mycorrhizal plants, in particular in the absence of additional P (Fig. 2c). Intermediate and high P supplies (P20 and P60) led to a reduction of 71 and 95%, respectively, in the specific activity of 33P in plants. Interestingly, the percentages of both root colonization and HLD were far less sensitive to increasing P supply (Fig. 1b,c) than the specific activity of 33P.

We determined the relative contribution of the MPU pathway to total Pi uptake. A previous study with tomato and G. intraradices (BEG87) had reported no significant differences in HLD between soil compartments containing either both roots and hyphae (RHC) or hyphae only (HC) (Smith et al., 2004). Assuming a similar relationship at any of the soil P concentrations under our experimental conditions, c. 78% of the total Pi was delivered by the MPU pathway at low P conditions (Fig. 2d). The addition of 20 or 60 mg P kg−1soil reduced the MPU contribution to 33 and 12%, respectively, of the total Pi uptake, indicating a shift in the relative contribution from the MPU towards the DPU pathway.

Repression by P of mycorrhiza-inducible phosphate transporter gene expression

We initially failed to detect LePT4 transcripts in roots of plants grown at high soil P concentrations (P60) even though they were colonized by AMF (data not shown). This prompted us to examine the expression of four members of the Pht1 family in more detail, that is, LePT1 through 4, as a function of root colonization by AMF over time in a similar experiment at ETH Zurich. Root colonization of plants grown with G. intraradices (BEG75) increased from c. 35% at 4 wk to c. 57% at 6 wk at low P (−P). RNA was isolated from roots of these plants, and RNA gel blot analysis was subsequently performed. Enhanced colonization coincided with an increase in LePT3 and LePT4 transcript abundance. Abundance of LePT1 and LePT2 transcripts, respectively, barely changed upon root colonization (Fig. 3a). High Pi concentrations in leaves, and a concomitant increase in shoot fresh weight (FW), were only observed in plants supplied with 500 µm NH4H2PO4 in the nutrient solution (+P), indicating that in absence of P fertilization, and independent of the presence or absence of AM fungal colonization, available Pi was growth-limiting during 6 wk of cultivation (Fig. 3b,c). While the Pi concentrations between the +P and –P treatments varied several-fold, the differences in biomass were less pronounced and more variable. Whereas a small decrease in LePT1 and a stronger decrease in LePT2 transcript abundance, respectively, were observed at high Pi concentration in the nutrient medium, expression of LePT3 and LePT4 was strongly suppressed in the same roots (Fig. 3a), despite colonization by AMF to a degree of c. 20%, as estimated by visual inspection. This indicates that mycorrhiza-inducible Pi transporter gene expression in a mycorrhiza is up-regulated during progression of intraradical AM fungal development at low Pi concentration, and is strongly repressed in colonized roots when the plant P status is high.

Repression by P of the lysophosphatidylcholine-dependent signaling pathway

Transgenic potato plants carrying an StPT3:: GUS chimeric gene (Rausch et al., 2001) were cultivated under different P conditions and thus differed in their P status. Roots of these plants were then infiltrated with bioactive phospholipids extracted from Plantago lanceolata mycorrhizas (PL+myc). The presence of the mycorrhiza signal LPC in PL+myc was evident from the MALDI-TOF mass spectrum with mass peaks corresponding to different LPC species (Fig. 4a). In young roots, a positive GUS response was related to a low P status of the plant (Fig. 4b). Upon infiltration of PL+myc, GUS staining was most pronounced in roots from transgenic plants that had been cultivated in the absence of Pi in the nutrient solution. Pi as low as 5 µm caused a consistent reduction of the GUS response, concomitant with a slight but consistent increase of the P concentration in roots and shoots. All these plants had P concentrations of ≤ 1 mg P g−1 DW in their roots and shoots (Fig. 4c,d). When Pi was supplied at 500 µm, resulting in a high plant P status (> 3 mg P g−1 DW), GUS expression was fully repressed.


An increasing P supply to the miniature tomato cultivar Micro-Tom suppressed root colonization by the fungal symbiont, as expected, and the observed pronounced effect on root length containing vesicles rather than total colonization or root length containing arbuscules, confirms previous findings, for example, in Trifolium subterraneum (Abbott & Robson, 1979). AMF colonization negatively affected plant growth at P0 but not at higher P. Irrespective of P supply, AMF colonization hardly affected P content, yet the MPU pathway contributed substantially to Pi acquisition by the plant, especially at the lowest Pi availability (P0, Fig. 2b,c). This is in line with previously published data on tomato (Smith et al., 2003, 2004).

However, the almost complete repression of 33P accumulation in shoots of plants grown at high P (P60, Fig. 2b) was unexpected, if one considers the much less pronounced effect on extra- and intraradical fungal mycelia (Fig. 1b,c). Li et al. (2006) found that in wheat the contribution from the MPU pathway was reduced to a lesser extent by higher soil P concentration than we observed here. But such a comparison is complicated by variability among plant species, soil P concentrations and Pi availability, and growth conditions. Furthermore, it is likely that soil P concentration and plant P status interact in controlling the MPU pathway. Shoot P concentration thus increased much less in response to additional P in the study of Li et al. (2006) (from 2.1 to 2.5 mg g−1 DW) as compared with the present study (from 1.6 (P0) to 2.2 (P20) and 3.7 mg g−1 DW (P60)).

Both the highest P supply in the first experiment (P60, Figs 1, 2) and the +P treatment in the second experiment (Fig. 3) markedly raised the shoot P status. While at P60 total shoot P content was 2.3–2.7-fold that at 0P, the shoot Pi concentration at +P was three- to fourfold the value at –P. Similarly, in both experiments the growth differences were roughly about twofold between plants cultivated under the respective conditions. We consider that these responses to added P are sufficiently similar to allow a comparison of the physiological data obtained in two experiments with the expression data of the second experiment. This view is supported by the conversion of the data on Pi concentration in Fig. 2(a) to P content in shoots of the second experiment (Fig. 3b), assuming a ratio of 10 between FW and DW and a P to Pi ratio of 1 and 3 at high and low P conditions, respectively (Clarkson & Scattergood, 1982). Such conversion resulted in 4.5 and 2.5 mg P g−1 DW in +P and –P shoots, respectively (Fig. 3). Thus, shoot P concentrations were in a similar range in all three experiments.

The inactivation of the MPU pathway at high Pi availability was most likely caused by the strong repression of LePT4 and LePT3, and presumably also LePT5 (Nagy et al., 2005; Chen et al., 2007; Xu et al., 2007), in roots colonized by G. intraradices (Figs 2c, 3a,b, and data not shown). Based on our data, combined with those indicating that the MPU and DPU pathways operate simultaneously (Poulsen et al., 2005; Li et al., 2006), we may assume that the relative contribution of the two pathways to total plant P uptake depends on Pi transporter gene regulation, which integrates signals from mycorrhizal colonization and the plant P status. It should be noted that fungal Pi transporters acting at the fungus–soil interface are also constituents of the P transfer in the MPU pathway. Similar to the situation in radix, fungal Pi transporter genes are also repressed at high P conditions (Maldonado-Mendoza et al., 2001; Benedetto et al., 2005).

The activation of Pi transporter gene expression in the Pi starvation response appears to be a universal phenomenon in the biosphere. The genes encoding LePT1 of tomato and MtPT1 of M. truncatula, two members of the plant Pht1 protein family involved in the DPU pathway (Bucher, 2007), are under transcriptional control (Muchhal & Raghothama, 1999; Chiou et al., 2001). Mukatira et al. (2001) suggested a negative regulation of a Pi starvation-inducible Pi transporter gene (in Arabidopsis), which is part of the DPU pathway (Misson et al., 2004; Shin et al., 2004). Activation of this gene upon onset of Pi deprivation was rapid and reversed at high P conditions. Transcription factors involved in the Pi starvation response of Arabidopsis have been identified (Rubio et al., 2001; Devaiah et al., 2007), and both miRNA- and E2 conjugase-mediated Pi starvation signaling have also been demonstrated (Bari et al., 2006; Chiou et al., 2006). Present knowledge on Pi starvation response regulation, combined with our data on P repression of the MPU pathway activity (Fig. 2), of mycorrhiza-inducible Pi transporter gene expression (Fig. 3), and of StPT3 promoter activation through mycorrhizal signals (Fig. 4), supports our model whereby the plant P status mediates regulation of both DPU and MPU pathway activity through Pi transporter expression predominantly at the transcriptional level. It will be interesting to elucidate whether the various genes involved in the DPU and MPU pathway, including their regulatory units, share common targets of the P repression mechanism.

Our previous work indicated that the signal recognition and transduction pathways in AM symbiotic Pi transport are evolutionarily conserved among dicotyledonous plant species (Karandashov et al., 2004). Moreover, evidence was provided that the regulatory mechanism(s) controlling mycorrhiza-enhanced Pi transport in potato are operative in a cell-autonomous way in cortical cells harboring different types of AM fungal structures (Rausch et al., 2001; Karandashov et al., 2004). This conclusion was supported by detailed studies on MtPT4 expression in M. truncatula where reporter gene expression and protein abundance were localized exclusively to cortical cells colonized by AMF (Harrison et al., 2002). The mycorrhizal signal LPC, a lysophospholipid of fungal or plant origin contained in phospholipid extracts from mycorrhizal roots, triggered mycorrhiza-inducible StPT3 promoter activation in Pi-starved potato roots (Drissner et al., 2007). The bioassay developed by Drissner et al. (2007) was used in the present work to study the responsiveness of roots to the mycorrhizal signal(s). Induction of GUS expression in transgenic potato roots by a mycorrhizal extract (Fig. 4a) was a function of the plant P status (Fig. 4b), and the roots kept under high P conditions were insensitive to the mycorrhizal signal(s). In our model, at least in solanaceous species, Pi starvation signaling is epistatic to the induction of the MPU pathway by mycorrhizal signals, and repression of the MPU pathway by Pi occurs at the level of cis-acting transcriptional regulation of Pi transporter genes. The cellular and molecular bases of the regulatory cross-talk between the mycorrhizal and the Pi repression signaling, respectively, in the symbiosis are presently unknown.

Overall, our work suggests that the plant P status is a major regulator controlling induction/derepression of Pht1 genes at both the soil–root interface during conditions of Pi deprivation, and the intraradical symbiotic interface via modulation of LPC signaling. As a result, enhanced Pht1 gene expression determines total Pi uptake in a mycorrhiza through increased Pi transporter abundance at these interfaces.


Special thanks are due to Anne Olsen and Anette Olsen for their excellent technical assistance. We are grateful to Peter Gehrig (Functional Genomics Center Zurich) for his support with MALDI analysis. Many thanks also to Sarah Wegmüller for carefully reading the manuscript and helpful suggestions. We also wish to thank unknown colleagues for suggestions to improve the manuscript. This work was supported by ETH Zurich (grant no. TH/3-00/3 to MB), the National Laboratory for Sustainable Energy (Risø DTU) and the Marie Curie Foundation (contract no. HPMT-CT-2000-00194).