|Home | About | Journals | Submit | Contact Us | Français|
Arbuscular mycorrhizal (AM) fungi benefit their host plants by supplying phosphate obtained from the soil. Polyphosphate is thought to act as the key intermediate in this process, but little is currently understood about how polyphosphate is synthesized or translocated within arbuscular mycorrhizas. Glomus sp. strain HR1 was grown with marigold in a mesh bag compartment system, and extraradical hyphae were harvested and fractionated by density gradient centrifugation. Using this approach, three distinct layers were obtained: layers 1 and 2 were composed of amorphous and membranous materials, together with mitochondria, lipid bodies, and electron-opaque bodies, and layer 3 was composed mainly of partially broken hyphae and fragmented cell walls. The polyphosphate kinase/luciferase system, a highly sensitive polyphosphate detection method, enabled the detection of polyphosphate-synthesizing activity in layer 2 in the presence of ATP. This activity was inhibited by vanadate but not by bafilomycin A1 or a protonophore, suggesting that ATP may not energize the reaction through H+-ATPase but may act as a direct substrate in the reaction. This report represents the first demonstration that AM fungi possess polyphosphate-synthesizing activity that is localized in the organelle fraction and not in the cytosol or at the plasma membrane.
Arbuscular mycorrhizal (AM) fungi are obligate biotrophs that form symbiotic associations with most land plants (29). These fungi promote the growth of host plants via enhanced uptake of phosphate (Pi) and thus play important roles in the terrestrial phosphorus cycle. In the symbiotic phase, AM fungi take up Pi from soil through an extensive network of extraradical hyphae and rapidly accumulate inorganic polyphosphate (polyP). This accumulation was as rapid as that for a polyP-hyperaccumulating bacterium found in activated sludge (6). PolyP is a linear polymer of three to hundreds of molecules of Pi linked by high-energy phosphoanhydride bonds and has been found across all classes of organisms (19). Although polyP is considered to play a central role in long-distance translocation of Pi in AM fungal associations (4, 10, 30, 31), the translocation mechanism, metabolism, and dynamics in the fungi have not been elucidated due to the difficulty in obtaining sufficient fungal material for analysis.
Many enzymes/genes involved in polyP synthesis/metabolism have been identified and characterized in prokaryotes (19). For instance, exopolyphosphatase hydrolyzes the terminal high-energy bonds of polyP, and polyphosphate glucokinase (PPGK) transfers the terminal Pi residue to glucose. Polyphosphate kinase 1 (PPK1) is responsible both for polyP synthesis, using ATP as a phosphoryl donor, and for the reverse ATP-generating reaction. This enzyme is bound to the plasma membrane (18) and has been found in a wide range of bacteria (17). Unlike the case for prokaryotes, knowledge of polyP synthesis/metabolism in eukaryotes remains limited. The first eukaryotic PPK genes, DdPPK1 (32) and DdPPK2 (14), were identified from the social slime mold Dictyostelium discoideum. The products of these genes, as known for bacterial PPK1s, are responsible both for polyP synthesis and for the ATP-generating reaction and have been suggested to be associated with vacuoles or small vesicles (14, 32). Although several homologues of bacterial PPK1 genes have now been found in the genomes of eukaryotic microorganisms (17), yeast Candida humicola is the only organism apart from D. discoideum for which PPK-like activity has been confirmed (22). The model organism Saccharomyces cerevisiae is known to accumulate polyP, to up to 10% of its dry weight (19). A unique polyP synthetic pathway different from those of PPK1 has been proposed for S. cerevisiae based on the observation that vacuolar-type H+-ATPase (V-ATPase)-defective mutants could not accumulate polyP (23). In this hypothetical pathway, Pi would be polymerized by an analogous system (enzyme) of mitochondrial F1-ATPase on the vacuolar membrane, using the proton motive force created by V-ATPase (23). On the other hand, Hothorn et al. (16) demonstrated very recently that vacuolar transporter chaperone 4 (VTC4), a small transmembrane protein associated with the membrane, polymerizes Pi by using the γ-Pi residue of ATP as a phosphoryl donor in S. cerevisiae.
More than 2 decades ago, Capaccio and Callow (3) reported the presence of polyP-hydrolyzing, -metabolizing (PPGK), and -synthesizing (PPK-like) activities in the soluble (cytosolic) fractions of the hyphae of the AM fungus Glomus mosseae. Recently, polyP-hydrolyzing activity was found in both the cytosolic and insoluble (membrane) fractions and then characterized (8). PPGK activity has also been confirmed in the cytosolic fraction, although the activity was quite low and hexokinase (ATP-hexose phosphotransferase) activity appeared to dominate in the glucose phosphorylation process (9). PPK-like activity, however, could not be detected in the same fraction (10), and this seems likely because all other prokaryotic (reviewed in reference 17) and eukaryotic (14, 16, 22, 32) polyP-synthesizing enzymes, so far, are associated with membranes. These observations suggest that AM fungi possess a polyP-synthesizing enzyme that is probably associated with membranes and that ATP may be essential in the synthesis as a phosphoryl donor or via H+-ATPase, as suggested by Ogawa et al. (23). In this study, a cell fractionation technique was applied to demonstrate polyP-synthesizing activity in an AM fungus, and then the role of ATP in the synthesis was investigated.
Glomus sp. strain HR1 (MAFF520076) was isolated from the rhizosphere soil of Lespedeza sp. grown in acidic soil and deposited in the NIAS Genebank (http://www.gene.affrc.go.jp/about_en.php). The small-subunit rRNA gene sequence (GenBank accession number AB220171) showed high similarity to those of Glomus manihotis and Glomus clarum, which belong to Glomus group A in the Glomeromycota (28). Dwarf marigold (Tagetes patula L. cv. Bonanza Orange; Murakami Seed, Ibaraki, Japan) was inoculated with 1,000 spores of Glomus sp. strain HR1 and grown in a mesh bag compartment system in which the root/hyphal (R+H) compartment and the hyphal (H) compartment were separated by a mesh bag (37-μm nylon mesh; 40-ml volume) in a plastic pot (5.5 cm in diameter; 90-ml volume) (8). The R+H compartment was filled with a 1:2 autoclaved washed pumice (4 to 10 mm in diameter)-river sand mixture, while the H compartment was filled with autoclaved river sand. The plants (one batch consisted of 60 to 70 pots) were grown in growth chambers (16-h photoperiod, 25°C, relative humidity of 60%) and received distilled water for the first week, followed by Peters Professional 25-5-20 (N-P2O5-K2O) liquid fertilizer (Scotts-Sierra Horticultural Products, OH) at 50 mg N liter−1 for the second week and then low-P liquid fertilizer (4 mM NH4NO3, 1 mM K2SO4, 0.75 mM MgSO4, 2 mM CaCl2, 0.5 mM Fe-EDTA, and 50 μM KH2PO4) for the third to sixth week every other day in sufficient amounts until the solutions flowed out from the drain holes. At the end of the sixth week, 1 mM KH2PO4 solution was applied to the plants in sufficient amounts 4 to 9 h prior to harvest, and extraradical hyphae were collected from the H compartment for all pots of the batch by wet sieving, combined, cleaned under a dissecting microscope as quickly as possible, and placed on ice.
All of the following experiments were done under ice-cooled conditions. Approximately 0.5- to 1.5-g hyphal samples were homogenized immediately after the harvest by mortar and pestle with a fivefold volume (wt/vol) of buffer A (1 mM Na2ATP and 1.2 M sorbitol in 10 mM HEPES-KOH, pH 7.4) with protease inhibitor cocktail for use with fungal and yeast extracts (Sigma-Aldrich, Tokyo, Japan). The slurry was transferred to a 15-ml plastic tube, the mortar and pestle were washed with the same volume of buffer A, and then the solutions were combined. After centrifugation at 160 × g for 5 min at 4°C, the upper layer was transferred to a new tube, and the pellet was resuspended in a fivefold volume (wt/vol) of buffer A and centrifuged under the same conditions. The upper layers were combined, layered on a continuous density gradient that was prepared by centrifugation of 50% (vol/vol) Percoll (GE Healthcare, Tokyo, Japan) in buffer A at 20,000 × g for 5 h at 4°C, and centrifuged at 2,500 × g for 2 h at 4°C. The resultant fractions, with densities of 1.02 to 1.05, 1.06 to 1.09, and >1.15 g ml−1, were designated layers 1, 2, and 3, respectively (Fig. (Fig.1).1). These fractions were collected separately, mixed with a 20-fold volume of buffer A, and centrifuged at 18,000 × g for 15 min at 4°C. The pellets were washed twice with the same buffer, resuspended in a minimum volume of the buffer, and stored on ice.
For transmission electron microscopic (TEM) observation of the fractions, the pellets of layers 1 to 3 were fixed with 4% (wt/vol) paraformaldehyde-1% (wt/vol) glutaraldehyde in buffer A for 16 h at 4°C. After the fixation, the pellets were centrifuged and rinsed three times with 10 mM HEPES-KOH at pH 7.4 for 10 min, embedded in a drop of 1% (wt/vol) low-melting-temperature agarose (Sigma-Aldrich), and processed by a standard procedure for TEM. Briefly, the materials were dehydrated with an ethanol series, substituted with propylene oxide, and infiltrated in Supper's resin (Nisshin EM, Tokyo, Japan) that was polymerized at 70°C overnight. Ultrathin sections were cut with glass knives or a diamond knife, put on copper grids, stained with tannic acid followed by lead citrate, and observed using a JEM-1200EX TEM (JEOL, Tokyo, Japan).
Each fraction was mixed with a ninefold volume of polyP extraction buffer (8 M urea in 50 mM Tris-HCl, pH 8.0), vortexed for 30 s, and centrifuged at 18,000 × g for 15 min at 4°C. The supernatant was collected, and urea was eliminated using a Micro Bio-Spin P-6 gel filtration column (Bio-Rad Laboratories, Tokyo, Japan) pretreated with TE buffer (1 mM EDTA in 10 mM Tris-HCl, pH 8.0) according to the manufacturer's instructions. PolyP content was determined by the Escherichia coli PPK/luciferase method (2), with some modifications (6). Total protein in the fractions was precipitated with trichloroacetic acid and reextracted with NaOH (7), and the concentrations were determined by the modified Lowry method, using a DC protein assay kit (Bio-Rad Laboratories, Tokyo, Japan) according to the manufacturer's instructions.
PolyP-synthesizing activity was assessed based on an increase in polyP content after incubation in the presence of ATP. Fifteen microliters of each fraction was mixed with an equal volume of buffer A on ice, and two 10-μl aliquots of the mixture were transferred to a new tube. One was mixed with 90 μl of polyP extraction buffer and left on ice as the time zero control, while the other was incubated for 30 min at 30°C. After incubation, the solutions were mixed with 90 μl of polyP extraction buffer. PolyP concentrations in the mixtures before (time zero control) and after the incubation were determined by the PPK/luciferase method. To examine the involvement of proton motive force in polyP synthesis, the effects of 500 μM vanadate (plasma membrane-type H+-ATPase [P-ATPase] inhibitor), 100 nM bafilomycin A1 (V-ATPase inhibitor), and 50 μM carbonyl cyanide m-chlorophenylhydrazone (CCCP [a protonophore]) on the synthesizing activity were assessed. In this assessment, 0.5% (vol/vol) dimethyl sulfoxide (DMSO) was added to all reaction mixtures, as the stock solutions of bafilomycin A1 (20 μM) and CCCP (10 mM) were prepared with DMSO. PolyP-synthesizing activity was defined based on an increase in Pi residues of polyP per min per unit of protein under the specified conditions.
P- and V-ATPases and cytochrome c oxidase (COX) were employed as marker enzymes for the plasma membrane, tonoplast, and mitochondria, respectively, to characterize the fractions. The fractions used for the assessment of P- and V-ATPase activities were prepared from hyphae to which 1 mM Pi solution was not applied prior to harvest, as fractions prepared from Pi-treated hyphae contained a large amount of polyP that was hydrolyzed during incubation and interfered with the assay. P- and V-ATPase activities were determined as “specific inhibitor-sensitive activities,” as follows: inhibitor-sensitive activity = total ATPase activity − inhibitor-insensitive activity (total ATPase activity = ATP hydrolyzing activity − nonspecific phosphatase activity, which was determined using ADP as a substrate).
Twenty-five microliters of each fraction was mixed with an equal volume of reaction mixture consisting of 100 mM KCl, 20 mM MgCl2, 2.4 M sorbitol, 0.2% DMSO, 2 mM substrate (Na2ATP or Na2ADP), and 80 mM HEPES-KOH at pH 7.5 in the presence or absence of inhibitors on ice, and two 20-μl aliquots of the mixture were transferred to new tubes. One was mixed with an equal volume of 10% (wt/vol) sodium dodecyl sulfate and left on ice as a time zero control, and the other was incubated for 30 min at 30°C, and then an equal volume of 10% sodium dodecyl sulfate was mixed with the solution. The levels of released Pi in the solutions before (time zero control) and after the incubation were determined as previously described (24). The activity was expressed as the amount of Pi released per min per unit of protein under the specified conditions. COX activity was measured as previously described (15). Prior to the assessment, 1 ml of 2 mg ml−1 horse heart cytochrome c (Wako Pure Chemicals, Osaka, Japan) in 100 mM potassium phosphate buffer (pH 7.5) was reduced by being mixed with 2 to 3 mg sodium dithionite and passed through a PD-10 Sephadex G-25 M column (GE Healthcare, Tokyo, Japan) equilibrated with the phosphate buffer to remove excess sodium dithionite. Ten microliters of each fraction was mixed with an equal volume of 1% (vol/vol) Triton X-100 in the phosphate buffer and 180 μl of 1 mg ml−1 reduced cytochrome c in the phosphate buffer, and decreases in absorbance at 550 nm were monitored for 3 min at room temperature. The concentration of oxidized cytochrome c was calculated based on an extinction coefficient of 18.5 mM−1 cm−1, and the activity was expressed as an increase in oxidized cytochrome c per min per unit of protein under the specified conditions. PolyP-hydrolyzing activity was measured based on the liberation of Pi at pH 7.5, using 1 mM polyP type 75+ (average chain length, 79; Sigma-Aldrich) as a substrate (8).
Each enzyme assay was performed in triplicate, using the same batch of material (n = 3). The treatments that showed zero or negative values were excluded from subsequent statistical analysis and expressed as “not detected.” Analysis of variance followed by Fisher's protected least significant difference test or Student's t test was performed by StatView software (SAS Institute Inc., NC) for tests of significance.
Cell fractionation was carried out several times using different batches of fungal material (Table (Table1;1; see Table S1 in the supplemental material), and representative results are shown in Table Table1.1. After centrifugation of the hyphal homogenate at 160 × g, 20 to 45% of the polyP was recovered in the upper layer. Layers 1, 2, and 3, obtained from the 160 × g upper layer by Percoll density gradient centrifugation, retained only 0.1 to 0.5% of total polyP, whereas 30 to 70% of the polyP in the 160 × g upper layer was recovered from the supernatant of the density gradient centrifugation. PolyP concentrations per unit of protein in layers 1 to 3 were lower than that of the homogenate.
Ultrastructural observations of the fractions revealed that layers 1 and 2 were composed of membranous and amorphous materials, together with organelles that were identified as mitochondria, lipid bodies, and electron-opaque bodies (Fig. 2a to c). A few bacteria were observed in these layers (data not shown). No conspicuous difference in composition between layers 1 and 2 was found. Layer 3 was composed of mainly partially broken hyphae and fragmented cell walls (Fig. 2d and e). Amorphous material associating with the cell wall fragments was also observed.
PolyP-synthesizing activity was consistently detected in layer 2 of all batches in the presence of ATP (Table (Table2;2; see Table S2 in the supplemental material). Although layer 1 showed polyP-synthesizing activity in some batches, the detection of the activity in layer 1 was poorly reproducible (data not shown). Layer 2 was thus used for subsequent characterization of the polyP-synthesizing activity. P-ATPase activity was enriched in layer 3, while V-ATPase activity could not be detected in any of the layers after density gradient centrifugation (Table (Table3).3). COX activity was enriched in layers 1 and 2. PolyP-hydrolyzing activity was diluted in all layers after density gradient centrifugation.
To examine whether ATP was used as a direct substrate or as an energy source to create the proton gradients for polyP synthesis (Pi would be the substrate in this case), layer 2 was incubated with ATP, Pi, or both, using the fraction prepared in the presence (original method) or absence (ATP was withheld from buffer A used for the washing step after density gradient centrifugation) of ATP. The activity of the fraction prepared with ATP and incubated with ATP was regarded as a positive control. No polyP-synthesizing activity was detected in the fraction prepared without ATP and incubated with no substrate or with 1 mM Pi (Fig. (Fig.3).3). The fraction prepared without ATP and incubated in the presence of ATP showed polyP-synthesizing activity, although the activity was lower than that of the positive control. No activity was observed in the fraction prepared in the presence of ATP and incubated with both ATP and Pi. The effects of P- and V-ATPase inhibitors and a protonophore were examined for further characterization. Vanadate showed an inhibitory effect on polyP-synthesizing activity, whereas bafilomycin A1 did not (Fig. (Fig.4a).4a). CCCP had no inhibitory effect on the activity (Fig. (Fig.4b4b).
Given the fact that AM fungal associations play a key role in phosphorus acquisition of the majority of land plants, it is of significance to clarify the polyP synthetic pathway, the first step of polyP metabolism and translocation, in the biotrophic fungi. PolyP-synthesizing activity in an AM fungus, Glomus sp. strain HR1, was successfully demonstrated in combination with cell fractionation. PolyP-synthesizing activity could not be detected in the soluble (cytosolic) fractions but was associated with the insoluble cellular components (membranous and amorphous material and organelles), whose densities were within the range of 1.06 to 1.09 g ml−1. It has been shown that bacterial PPKs bind peripherally to the inner plasma membrane (19). The association of the polyP-synthesizing activity with the plasma membrane, however, could be excluded in the AM fungus, as the activity of the P-ATPase, a plasma membrane marker, was not concurrent with the polyP-synthesizing activity. The association of the activity with mitochondria may also be unlikely, because polyP-synthesizing activity was consistently detected only in layer 2, but the mitochondrial marker enzyme COX was enriched in both layers 1 and 2 after fractionation, to the same extent. The amorphous material was not specific to layer 2 and was observed in all fractions, and thus the association of polyP-synthesizing activity with the material is unlikely. The membranous material observed in layer 2 is likely to be vacuolar membrane due to the fact that AM fungal cells contain numerous vacuoles, although the vacuolar marker enzyme V-ATPase could not be detected in any fraction. Possibly, V-ATPase of the fungus is susceptible to physical disruption/fractionation or the activity is intrinsically very low in the fungus. It seems likely that polyP-synthesizing activity is associated with the vacuolar membrane due to the following two reasons. First, all other eukaryotic polyP-synthesizing enzymes, DdPPK1 (32) and DdPPK2 (14) in the slime mold and VTC4 in yeast (16), are associated with vacuoles or small vesicles. Second, AM fungi accumulate polyP in vacuoles, as observed in the germ tubes of Gigaspora margarita (20) and in the extra- and intraradical hyphae in Glomus sp. strain HR1 (Y. Kuga and T. Ezawa, unpublished observations). Further characterization is required to localize the activity.
Our results are not consistent with those reported by Capaccio and Callow (3), in which polyP-synthesizing activity was detected in the extract (soluble fraction) of an AM fungus. They incubated radioactive ATP with the extract and identified polyP as a trichloroacetic acid-precipitated radioactive compound. This approach was the most reliable for detection of polyP at that time but may not be specific to polyP. One possibility, therefore, is that the polyP-synthesizing activity reported previously represented an activity of another phosphotransferase-type enzyme that was localized in the cytosol and used ATP as a phosphoryl donor. It is also likely that the extract was contaminated with the membrane fraction that retained polyP-synthesizing activity. The radioactive compound-based method is highly sensitive, such that contamination by trace amounts of membrane may also result in the detection of activity.
The following three technical breakthroughs were indispensable for the detection of polyP-synthesizing activity in the fungus. (i) A small-scale and highly sensitive polyP assay system, the PPK/luciferase-polyP assay system (2), has been applied to AM fungi recently (6, 25). Although the method is relatively less sensitive to polyP species shorter than 20 Pi residues (26), it was essential for the measurement of picomole levels of polyP in the present study. (ii) We selected a fungal species that produces few spores and a large amount of extraradical hyphae. Glomus sp. strain HR1 produced fewer spores and a greater hyphal mass over a 6- to 8-week culture than did other species examined (data not shown), although Glomus sp. strain HR1 produced a large amount of spores after a 4-month culture. This is a quite important characteristic for cell fractionation, because AM fungal spores are, in general, filled with lipids that aggregate with organelles and interfere with cell fractionation. (iii) We enriched the polyP-synthesizing activity by cell fractionation. The activities of exopolyphosphatase (3, 8) and acid (8, 12) and alkaline (1, 7, 11, 13) phosphatases have been detected in AM fungi, and these enzymes are likely to be involved in polyP hydrolysis. In fact, polyP-hydrolyzing activity in the hyphal homogenate was quite high in the present study. Therefore, cell fractionation that could enrich polyP-synthesizing activity and dilute polyP-hydrolyzing activity was necessary for detection. It should be noted, in addition, that the polyP-synthesizing activity estimated in our study might have been underestimated due to the concurrence of polyP-hydrolyzing activity in the fraction.
ATP was essential for polyP synthesis in the fungus, and the results strongly suggest that ATP did not energize the reaction via P- or V-ATPase but acted as a direct substrate in the reaction due to the following reasons. First, the V-ATPase inhibitor and protonophore did not inhibit the activity. Second, P-ATPase activity was enriched in layer 3, which showed no polyP-synthesizing activity. Although vanadate, the P-ATPase inhibitor, inhibited the polyP-synthesizing activity, the involvement of P-ATPase, i.e., the proton gradient across the plasma membrane, could be validated only if the activity was inhibited by both vanadate and the protonophoric reagent, because vanadate has been known to inhibit many phosphoryl transfer enzymes as a structural and chemical mimic of phosphate (5). The fact that all polyP-synthesizing enzymes found in prokaryotes (19) and eukaryotes (14, 16, 22, 32) use ATP as a direct phosphoryl donor also supports our observations. Catalysis of the reverse ATP-regenerating reaction is a typical feature of PPK-type enzymes in both prokaryotes (19) and eukaryotes (14, 32) but is unlikely in the case of VTC4 found in yeast (16). In the present study, it was difficult to examine whether layer 2 catalyzed the reverse reaction due to the high background of ATP that was essential for the protection/stabilization of the polyP-synthesizing activity during fractionation. The polyP-synthesizing activity was inhibited by the coexistence of ATP and Pi. It is postulated that Pi interfered with the activity through binding to the catalytic center or that Pi might act as an allosteric effector that regulates the balance between ATP consumption and polyP synthesis in the cell. However, given the fact that cytoplasmic Pi homeostasis is strictly maintained at millimolar levels, e.g., 5 to 10 mM in the case of plants (27), it is unlikely that cytoplasmic Pi is directly involved in the regulation of polyP synthesis in AM fungi. The underlying mechanism will be clarified if the enzyme is purified and localized at the subcellular level.
It is important to identify enzymes/genes involved in polyP metabolism for a clear understanding of the role of polyP in Pi translocation in AM fungi. The present study suggests that purification/localization of the enzyme will be possible using polyP-synthesizing activity as a marker. Application of forthcoming genomic information (21) in conjunction with biochemical analysis will be one promising approach.
The PPK-overexpressing E. coli strain was a kind gift from A. Kornberg (Stanford University, CA). We thank Derek B. Goto (Hokkaido University) for a critical reading of the manuscript.
This work was supported by a grant-in-aid for scientific research (17380043) from the Japan Society for the Promotion of Science (T.E. and Y.K.).
Published ahead of print on 18 September 2009.
†Supplemental material for this article may be found at http://aem.asm.org/.