• Open Access

Enhanced phosphorus nutrition in monocots and dicots over-expressing a phosphorus-responsive type I H+-pyrophosphatase

Authors


* Correspondence (fax (860) 486 0534; e-mail roberto.gaxiola@asu.edu)

Summary

Plants challenged by limited phosphorus undergo dramatic morphological and architectural changes in their root systems in order to increase their absorptive surface area. In this paper, it is shown that phosphorus deficiency results in increased expression of the type I H+-pyrophosphatase AVP1 (AVP, Arabidopsis vacuolar pyrophosphatase), subsequent increased P-type adenosine triphosphatase (P-ATPase)-mediated rhizosphere acidification and root proliferation. Molecular genetic manipulation of AVP1 expression in Arabidopsis, tomato and rice results in plants that outperform controls when challenged with limited phosphorus. However, AVP1 over-expression and the resulting rhizosphere acidification do not result in increased sensitivity to AlPO4, apparently because of the enhancement of potassium uptake and the release of organic acids. Thus, the over-expression of type I H+-pyrophosphatases appears to be a generally applicable technology to help alleviate agricultural losses in low-phosphorus tropical/subtropical soils and to reduce phosphorus runoff pollution of aquatic and marine environments resulting from fertilizer application.

Introduction

In worldwide agricultural production, phosphorus is second only to nitrogen as the most limiting macronutrient. In soils, orthophosphate (Pi), the assimilated form of phosphorus, exists primarily as insoluble calcium salts or iron–aluminium oxide complexes that are inaccessible to plants (Holford, 1997). When aggressive fertilization is employed to alleviate available Pi deficiency, runoff from agricultural land represents a serious threat to aquatic and marine environments (Hammond et al., 2004).

In response to limiting Pi availability, plant metabolic and developmental processes are altered to enhance Pi uptake. For example, in Arabidopsis, the coordinated induction of more than 600 genes is seen under conditions of Pi deprivation (Misson et al., 2005). Perhaps the most obvious consequence of altered gene expression in Pi-deprived plants is the expansion of their root architecture and resultant increases in absorptive surface area (Lopez-Bucio et al., 2002; Gahoonia and Nielsen, 2004). Pi-deprived roots exhibit transition of the primary root to determinate growth, greater frequency of lateral root formation and increased recruitment of trichoblasts to form root hairs (Abel et al., 2002; Poirier and Bucher, 2002; Sanchez-Calderon et al., 2006). In some species, Pi-deprived roots form specialized structures to enhance nutrient uptake, as is seen in white lupin (Lupinus albus), which forms clusters of short, hairy lateral roots (proteoid roots) that are specialized for Pi uptake (Yan et al., 2002).

Another adaptation to low soil Pi is rhizosphere acidification, resulting from enhanced plasma membrane H+-adenosine triphosphatase (H+-ATPase) activity in roots (Yan et al., 2002; Zhu et al., 2005; Shen et al., 2006). Increased H+ extrusion results in increased displacement of Pi from insoluble soil complexes (Vance et al., 2003). The advantage of these adaptations to low-Pi conditions is evident in the apparent universality of such responses in plants that prosper in low-Pi soils.

Recently, we have reported that the over-expression of the Arabidopsis vacuolar pyrophosphatase AVP1 results in increased root proliferation and apoplast acidification (Li et al., 2005; Gaxiola et al., 2007), suggesting a mechanism that may be manipulated to produce plants that exhibit increased resilience to Pi deficiency. Moreover, the discovery that Arabidopsis and tomato plants over-expressing AVP1 are resistant to water deficit stress (Gaxiola et al., 2001, 2007; Park et al., 2005) further enhances the potential value of this approach, as low-Pi soils are common in developing nations where water deficits are not easily ameliorated by irrigation (soils.usda.gov). In this article, we describe the expression of AVP1 in response to Pi deprivation, and characterize the improved performance of Arabidopsis, tomato and rice plants over-expressing AVP1 (AVP1OX) under conditions of Pi stress.

Results and discussion

Low Pi increases transcript and protein abundance of AVP1 and P-ATPase in Arabidopsis

The transcription and translation of AVP1 and P-ATPases (Arabidopsis H+-ATPases, AHAs), normally expressed in roots (Arango et al., 2003; Gaxiola et al., 2007) under limiting Pi conditions, were monitored. AVP1 and representative AHA mRNA abundance was assessed in wild-type Arabidopsis plants transferred to limiting Pi conditions by quantitative real-time fluorescence-polymerase chain reaction (RTF-PCR). AVP1 mRNA induction peaked 12 h after the transfer of the seedlings to limiting Pi, AHA1 showed no change, and both AHA2 and AHA6 expression peaked 12 h after AVP1 (Figure 1a). Transcription of the phosphate transporter AtPT1, which is induced under low-phosphate conditions (Muchhal et al., 1996), was up-regulated within 3 h of limiting Pi conditions (Figure 1a). The induction of AVP1 expression by limiting Pi was confirmed with an AVP1 promoter-β-glucuronidase (AVP1::GUS) reporter transformant [Figures 1d, e and S1 (see ‘Supplementary material’)]. This behaviour is consistent with the presence of potential cis-regulatory Pi response elements in the 1.7-kb promoter region used to generate the AVP1::GUS reporter (i.e. one PRH1 element at position –540; two TC elements at positions –79 and –103). These elements are present in genes whose expression has been shown to be up-regulated under limiting Pi conditions (reviewed in Hammond et al., 2004). Western blots of microsomal fractions probed with polyclonal antibodies raised against AVP1 and H+-ATPase, and the relative densities of each confirmed expression, showed that the abundance of both H+ pumps was increased fourfold and twofold, respectively, by Pi starvation (Figure 1b, c). These results suggest that increased AVP1 expression precedes an increase in the abundance of AHA2 and AHA6 H+-ATPases.

Figure 1.

Expression of the Arabidopsis vacuolar pyrophosphatase AVP1 and P-type adenosine triphosphatases (P-ATPases) under low orthophosphate (Pi). (a) Quantitative real-time fluorescence-polymerase chain reaction (RTF-PCR) time points of AHA1, AHA2, AHA6, AVP1 and AtPT1 from wild-type (WT) plants grown under low Pi for 0–48 h. The relative mRNA levels were normalized to ACT2. Values are the means ± standard deviation, n = 3. (b) Immunoblot time points of membrane proteins isolated from WT plants grown under low Pi for 0–6 days and probed with antisera to H+-pyrophosphatase and P-ATPase. (c) The relative densities of H+-pyrophosphatase and P-ATPase in (b) were quantified with Bio-Rad Quantity One software. Values are the means ± standard deviation of three independent experiments. AtAVP1::GUS expression in roots under control (d) and low-Pi (e) conditions. Bar, 50 µm.

AtAVP1OX root systems respond more vigorously than controls when exposed to limiting Pi conditions

To examine whether the root systems of AtAVP1OX plants were capable of responding to low Pi, control and AtAVP1OX seeds were germinated under Pi-deficient (10 µm) conditions and their root development was analysed. AtAVP1OX seedlings developed more robust root systems than wild-type controls under Pi limitation (Figure 2a, b). At 20 days, AtAVP1OX roots were longer and had developed an average of seven more lateral roots than controls (P < 0.01) (Figure 2c, d). Root hairs were also 2.5-fold larger and 1.5-fold denser than those of controls under Pi-deficient conditions (P < 0.01) (Figure S2a–d, see ‘Supplementary material’), increasing the absorptive area of the roots. Primary root apical cell proliferation was monitored in control and AtAVP1OX plants germinated under Pi-deficient and Pi-sufficient conditions using a CycB1::CDBGUS reporter associated with meristem activity/indeterminacy (Li et al., 2005). Cell proliferation, a result of meristem activity, in both AtAVP1OX and wild-type plants was curtailed in Pi-deficient conditions, but the switch to determinate growth, indicated by a loss of CycB1::CDBGUS activity, was delayed for 3–4 days in AtAVP1OX plants (Figure S2f, see ‘Supplementary material’).

Figure 2.

Root development and acidification capacity are enhanced in AtAVP1OX. Seedlings grown for 20 days in synthetic medium with the addition of 1 mm orthophosphate (Pi) (a) or 10 µm Pi (b). Bar, 1 cm. Average primary root length (c) and lateral root number (d) from similar experiments. Values are the means ± standard deviation, n = 15–20, *P < 0.01. Root acidification activity of seedlings 10 days after transfer from control conditions to 10 µm Pi medium (e) or 10 µm Pi + 1 mm vanadate (f) with the pH indicator bromocresol purple.

AtAVP1OX plants exhibit enhanced rhizosphere acidification under Pi deficiency

The more robust root systems developed by AtAVP1OX plants would be expected to increase the acidification of Pi-deficient medium, resulting in more efficient scavenging of Pi. To test this hypothesis, wild-type and AtAVP1OX plants were transferred from Pi-sufficient to Pi-deficient medium. A visual examination of the plates showed that AtAVP1OX plants had a greater capacity than wild-type controls to acidify the medium, as indicated by the intense yellow colour of the pH indicator dye (Figure 2e). Rhizosphere acidification was completely inhibited in wild-type and AtAVP1OX plants by 1 mm vanadate (Figure 2f), consistent with the inhibition of plasma membrane H+-ATPase activity (Yan et al., 2002). Enhanced Pi uptake, measured as Pi depletion from defined hydroponic medium, was visible in both AtAVP1-1 and AtAVP1-2 over-expression transformants within 8 h of incubation, with AtAVP1-1 exhausting the available Pi almost 30 h earlier than AtAVP1-2 (Figure S2e, see ‘Supplementary material’). Total depletion of Pi by control plants was not observed at any time point.

AtAVP1OX plants also exhibited enhanced growth and Pi uptake when grown on solid Pi-deficient medium (Table S1, see ‘Supplementary material’). AtAVP1OX seedlings germinated and grown in Pi-deficient medium for 20 days exhibited 1.6-fold more root and 1.3-fold more shoot biomass than controls (P < 0.01). The Pi content (per plant) was 1.4-fold higher in AtAVP1OX plants than in controls (P < 0.01), suggesting that AtAVP1OX plants acquire more Pi and grow accordingly (Gilooly et al., 2005; Hermans et al., 2006). Consistent with the Pi limitation of organismal growth, the Pi content (mmol/g dry weight) of AtAVP1OX plants grown either at normal or restrictive Pi conditions was no different from controls (Table S1, see ‘Supplementary material’).

AtAVP1OX plants develop larger shoots when grown in low-Pi soil

To determine whether the enhanced root systems seen in AVP1OX plants confer an advantage to plants grown in natural soils, control and AtAVP1OX lines were germinated and grown in natural low-Pi (~15 µm) soil (Figure 3a–d). The growth of all plants was delayed compared with normal conditions (data not shown). However, as shown previously under Pi-sufficient conditions (Li et al., 2005), AtAVP1OX plants developed more leaves with greater surface areas than their wild-type counterparts at all stages of growth and, when scored at 50 days post-germination, AtAVP1-1 and AtAVP1-2 had six- and twofold greater leaf areas than controls, respectively (P < 0.01) (Figure 3b, d). This suggests that the over-expression of AVP1 results in enhanced Pi scavenging from natural low-Pi soils. Similarly, AtAVP1OX plants outperformed controls in a sandy medium supplemented with insoluble rock phosphate (P2O5) as the only source of Pi, and developed an average of 2.3-fold more shoot biomass (P = 0.01) (Figure 3e, f).

Figure 3.

Performance of wild-type (WT) and AtAVP1OX plants subjected to low-orthophosphate (low-Pi) soil and rock treatment. (a) Representative plants grown in natural low-Pi soil (~15 µm PO43–) for 50 days. (b) Corresponding leaf area data. Values are the means ± standard deviation, n = 7. (c) Representative plants grown in natural low-Pi soil for 90 days. (d) Corresponding leaf area data. Values are the means ± standard deviation, n = 7 plants. (e) Representative plants grown in sand with P2O5 (rock Pi) as the only source of Pi for 30 days. (f) Corresponding shoot fresh weight data. Values are the means ± standard deviation, n = 15–20. *P = 0.01.

Tomato plants over-expressing the Arabidopsis type I H+-pyrophosphatase outperform controls under Pi-limiting conditions

There is a high degree of identity at the amino acid level between the type I H+-pyrophosphatases across the plant kingdom (Drozdowicz and Rea, 2001), suggesting that AVP1 from one species would be functional in another species. Transgenic tomatoes over-expressing the E229D gain-of-function mutant (AVP1D) of the Arabidopsis H+-pyrophosphatase (LeAVP1DOX) develop more robust root systems and are resistant to imposed soil water deficits (Park et al., 2005). As was seen with AtAVP1OX, both LeAVP1D-1 and LeAVP1D-2 over-expression lines developed larger shoots, root systems and fruits than controls when grown under Pi-deficient conditions (Figure 4a–c). Although, at 44 p.p.m. NaH2PO4, neither controls nor LeAVP1DOX lines developed fruits, at 100 p.p.m. NaH2PO4, both LeAVP1D-1 and LeAVP1D-2 lines developed a larger quantity of bigger fruits than controls (Figure 4c, d). The root and shoot dry weights of plants grown in the presence of 100 p.p.m. NaH2PO4 were, on average, 13% and 16% higher (P < 0.01), respectively, in LeAVP1DOX plants than in controls. Furthermore, under the same low-Pi conditions, fruit dry weight data and Pi content per plant were 82% and ~30% higher (P < 0.01), respectively, than in controls (Table S2, see ‘Supplementary material’). Here again, there was no statistically significant difference in the normalized Pi content (mmol/g dry weight) of control and LeAVP1DOX roots or shoots grown under three limiting Pi conditions (Table S2, see ‘Supplementary material’).

Figure 4.

Development of shoot, root and fruits of control and LeAVP1DOX tomato plants subjected to different orthophosphate (Pi) treatments. (a) Representative plants grown in soil adjusted to contain 44, 100 and 400 p.p.m. NaH2PO4 for 40 days in the glasshouse. (b) Representative roots of plants grown in soil with 44 p.p.m. NaH2PO4 for 40 days. Fruits of plants grown in soil with 100 p.p.m. (c) and 400 p.p.m. (d) NaH2PO4 for 120 days.

Over-expression of Arabidopsis AVP1D results in enhanced Pi nutrition in rice

Rice (Oryza sativa) is the principal staple crop for more than one-half of the world's population. It has been estimated that the world's annual rice production must increase from 618 million tons in 2005 to 771 million tons by 2030 in order to keep pace with population growth (http://www.fao.org/newsroom/en/news/2006/1000387/index.html.). To determine whether increased H+-pyrophosphatase activity improves plant performance under Pi-deficient conditions in monocots, rice (O. sativa var. japonica‘Taipei 309’) was engineered with a 35S::AVP1D cassette (see ‘Experimental procedures’). AVP1 over-expression was confirmed by Western blot and relative density; vacuolar-H+-ATPase (V-ATPase) was not altered (Figure 5e, f). As shown in Figure 5a, b, the OsAVP1DOX line exhibited sustained shoot growth under Pi-deficient (10 µm) conditions, whereas the controls grew poorly. The OsAVP1DOX line tested developed more robust root systems than controls in both Pi-sufficient and Pi-deficient conditions (Figure 5c, d). The dry plant biomass data confirmed that the OsAVP1DOX line grown under limiting Pi conditions developed larger roots (90%, P < 0.01) and shoots (50%, P = 0.01) than controls (Table S3, see ‘Supplementary material’). Therefore, AVP1 over-expression in monocots and dicots results in enhanced root systems under low-Pi conditions. However, in contrast with the results of over-expression of AVP1 in Arabidopsis and tomato, OsAVP1DOX rice seedlings grown under Pi-sufficient conditions accumulated 18% higher Pi content (mmol/g dry weight) than controls (Table S3, see ‘Supplementary material’), suggesting a different mechanism between monocots and eudicots under Pi sufficiency.

Figure 5.

Phenotypes of control and OsAVP1DOX (rice) plants grown under low-orthophosphate (low-Pi) conditions. Representative plants grown in rice medium containing 10 µm Pi (a) and 1 mm Pi (b) for 35 days. Representative roots of plants grown in 10 µm Pi (c) and 1 mm Pi (d) for 35 days. (e) Immunoblots of control and OsAVP1DOX membrane proteins probed with antisera to H+-pyrophosphatase and vacuolar-type adenosine triphosphatase (V-ATPase). (f) The relative densities of H+-pyrophosphatase were quantified with Bio-Rad Quantity One software. Values are the means ± standard deviation of three independent experiments.

Roots from AtAVP1OX, LeAVP1DOX and OsAVP1DOX lines have higher K+ contents and exude greater amounts of organic acids than controls

Pi deficiency is often a problem in tropical soils in which marginal aluminium toxicity limits agricultural production (Kochian et al., 2004). In such soils, although AVP1-dependent rhizosphere acidification can enhance Pi efficiency, it may also be expected to enhance aluminium mobilization and toxicity. Surprisingly, this is not the case. The growth of AtAVP1OX, LeAVP1DOX and OsAVP1DOX was assayed in medium in which AlPO4 functioned as the only source of Pi. AVP1 over-expression did not result in increased sensitivity to AlPO4. Only AtAVP1-1 exhibited aluminium sensitivity similar to controls. LeAVP1DOX, OsAVP1DOX and, to a lesser extent, AtAVP1-2 exhibited greater tolerance to aluminium compared with controls (Figure S3, see ‘Supplementary material’).

Enhanced AVP1-dependent H+ extrusion appears to be charge balanced, as demonstrated by enhanced K+ retention and organic acid extrusion from roots, similar to that seen in copper-challenged Arabidopsis roots (Murphy et al., 1999). Under all conditions tested, root K+ contents were significantly higher (P < 0.05), and approximately twice those of controls, in all AVP1OX crops (Figure S4a, see ‘Supplementary material’). Furthermore, quantification of root organic acid exudates (citrate and malate) in AVP1OX plants grown under AlPO4 stress showed higher levels of organic acid exudation in LeAVP1DOX and OsAVP1DOX than in controls (Figure S4b, see ‘Supplementary material’). Organic acid extrusion has been correlated with enhanced resistance to aluminium toxicity (reviewed in Kochian et al., 2004), suggesting that the enhanced rhizosphere acidification triggered by AVP1 over-expression would not result in increased aluminium toxicity in marginal tropical soils.

Biomass and seed yields are enhanced in both Arabidopsis and rice plants when grown under nutrient-sufficient conditions

The over-expression of AVP1 in Arabidopsis results in plants with significantly larger root and shoot biomasses when grown under nutrient-sufficient conditions (Li et al., 2005). Hydroponically grown OsAVP1DOX plants also developed twofold larger shoots and roots, and twofold more tillers and panicles (inflorescences), than control plants (Figure S5 and Table S4, see ‘Supplementary material’). Furthermore, AVP1 over-expression in rice resulted in a 50% increased seed yield (Figure S5 and Table S4, see ‘Supplementary material’). We did not observe similar increases in fruit production in LeAVP1DOX plants grown under nutrient-sufficient conditions (Table S2, see ‘Supplementary material’). The differences in crop performance could be a result of the different shoot branching patterns displayed by tomato and rice (reviewed in McSteen and Leyser, 2005). It should be noted that ionic contents of rice grains of plants grown under nutrient sufficiency showed that phosphorus, iron and zinc contents were enhanced in OsAVP1DOX plants (Figure S6, see ‘Supplementary material’).

Conclusions

Our results are consistent with the model that links root architectural modifications and rhizosphere acidification triggered by Pi deficiency to the stimulation of the activity of the type I H+-pyrophosphatase, AVP1. Manipulation of the type I H+-pyrophosphatase contributes to the regulation of the pH in the apoplast (Li et al., 2005; Gaxiola et al., 2007) and rhizosphere (Figure 2e, f), apparently by mediating the trafficking of the plasma membrane H+-ATPase. Rhizosphere acidification is a central mechanism for plant mineral nutrition. Engineering plants to over-express AVP1 appears to be a generally applicable technology to a range of agriculturally important crops. This technology can help alleviate agricultural losses caused by Pi limitation in acid and calcareous soils (Raghothama, 1999). It can also help to optimize the use of non-renewable phosphorus fertilizers and reduce the pollution of aquatic and marine environments by Pi runoff, as crops over-expressing AVP1 also exhibit enhanced resistance to water deficits (Gaxiola et al., 2001; Park et al., 2005) and are likely to require minimal irrigation.

Experimental procedures

Arabidopsis growing conditions

The control Arabidopsis (Arabidopsis thaliana) plants and AVP1OX plants (Gaxiola et al., 2001) used in this work were of the Col-0 ecotype. Seeds were surface sterilized and imbibed overnight at 4 °C before being sown on agar medium, or soil, or rock wool for hydroponic growth. For plants grown on agar, half-strength Murashige and Skoog (MS) medium (Murashige and Skoog, 1962) and Pi-free medium (Estelle and Somerville, 1987; Hartel et al., 2000) were used with 1% sucrose and 0.8% or 1% agar (Micropropagation/Plant Tissue Culture Grade, PhytoTechnology Laboratories, Shawnee Mission, KS, USA). Pi-free medium contains 20 mm 2-(N-morpholino) ethanesulphonic acid (MES) (pH 5.8), 5.0 mm KNO3, 2.0 mm MgSO4, 2.0 mm Ca(NO3)2, 50 µm iron ethylenediaminetetraacetate (Fe·EDTA), 70 µm H3BO3, 14 µm MnCl2, 0.5 µm CuSO4, 1.0 µm ZnSO4, 0.2 µm NaMoO4, 10 µm NaCl and 0.01 µm CoCl2. The Pi concentration was adjusted with KH2PO4. All experiments were performed with agar (PhytoTechnology Laboratories) that had no detectable trace Pi contamination, as determined by the method of Murphy and Riley (1962). The composition and pH of the natural low-Pi soil used were analysed by the Soil Nutrient Analysis Laboratory of the University of Connecticut (pH 6.1; P, 0.5 p.p.m.; K, 123 p.p.m.; Ca, 467 p.p.m.; Mg, 74 p.p.m.; soil texture classification, sandy loam). Pi-free medium with different KH2PO4 concentrations was used to water plants grown in low-Pi soil. Plants were grown in growth chambers with a 16-h light/8-h dark cycle at 21 °C. For hydroponically grown plants, the conditions described by Gibeaut et al. (1997) were followed. For aluminium treatment, Arabidopsis plants were germinated in half-strength MS medium for 5 days, and then transferred to plates containing 40 µm AlPO3 as the only source of Pi; the medium pH was buffered to pH 4.5 with 1 mmN-2-hydroxyethylpiperazine-N′-2-ethanesulphonic acid (HEPES). For rock phosphate experiments, plants were grown in sand with a 1 : 1000 w/w P2O5/sand ratio as the only source of Pi, and flooded regularly with Pi-free medium.

Tomato growing conditions

Tomato (Lycopersicon esculentum Mill. cultivar Money Maker) control and AVP1D over-expressing plants have been described elsewhere (Park et al., 2005). Seeds were surface sterilized and imbibed overnight at 4 °C before being sown on half-strength MS medium. Ten-day-old seedlings were transferred to pots containing 1 kg of natural low-Pi soil (see above) mixed with 44, 100 or 400 mg NaH2PO4. Tomato pots were kept in plastic bags to prevent loss of Pi. Plants were randomly placed in the glasshouse. Pi-free medium, described above, was used to water the plants every 2 weeks. For aluminium treatment, tomato plants were germinated in half-strength MS medium for 10 days, and then transferred to sand supplemented with 20 µm AlPO3 as the only source of Pi; the pH was buffered to pH 5.0 with 1 mm HEPES.

Rice growing conditions

Rice control Oryza sativa var. japonica‘Taipei 309’ and AVP1D over-expressing plants (see ‘Rice transformation’) were grown in sand under normal and low-Pi conditions. Rice seeds were surface sterilized and germinated in agar plates containing medium supplemented with either 1 mm or 10 µm Pi, as described above (Estelle and Somerville, 1987; Hartel et al., 2000). Ten-day-old seedlings were then transferred to pots with sand. These pots were placed in a plastic tray filled with 2 L of rice Pi-free liquid medium (modified from Yoshida et al., 1976) supplemented with either 1 mm or 10 µm Pi. Rice Pi-free liquid medium contains 1.43 mm NH4NO3, 0.51 mm K2SO4, 1 mm CaCl2, 1.64 mm MgSO4·7H2O, 9.5 µm MnCl2·4H2O, 0.048 µm (NH4)6Mo7O24·4H2O, 19 µm H3BO3, 0.15 µm ZnSO4·7H2O, 0.155 µm CuSO4·5H2O, 35.6 µm FeCl3·6H2O, 71 µm citric acid monohydrate and 1.1 mm sulphuric acid, pH 5.0. The Pi concentration was adjusted with KH2PO4. Plants grown under Pi-sufficient conditions were watered with 1 mm Pi medium, and plants grown under Pi-deficient conditions were watered with rice Pi-free medium, every 2 weeks. Plants were incubated in growth chambers with a 16-h light/8-h dark cycle at 25 °C. For aluminium treatment, rice plants were germinated in half-strength MS medium for 10 days, transferred to pots with sand and flooded with Pi-free medium + 10 µm AlPO3 as the only source of Pi. The pH of the medium was buffered to pH 5.0 with 1 mm HEPES.

Hydroponic conditions

Rice plants were germinated and grown on agar plates supplemented with half-strength MS medium for 10 days, transferred into pots with sand and watered with one-eighth-strength MS medium. After 15 days of growth in sand, plants were transferred to the hydroponic system (General Hydroponics, Sebastopol, CA, USA). The nutrients added were from General Hydroponics following their directions for ‘general purpose’ growth conditions. The medium was supplemented once with 10 mL of Iron Max Ac 6% (15-0-0) (Growth Products, White Plains NY, USA) per 120 L of hydroponic medium solution.

Quantitative RTF-PCR

Arabidopsis thaliana plants were germinated in half-strength MS medium for 2 weeks, and then transferred to 10 µm Pi plates for 0–48 h. Total RNA was isolated from seedlings (10–15 seedlings per sample) with TRI-reagent (Molecular Research Center, Cincinnati, OH, USA), according to the manufacturer's manual. After being treated with DNase I (DNA-free Kit, Ambion, Austin, TX, USA), 1 µg of each RNA sample was used to synthesize cDNA with an iSript cDNA Synthesis Kit (Bio-Rad Laboratories, Hercules, CA, USA) in a total volume of 20 µL at 41 °C for 1 h.

RTF-PCR was performed in a LightCycler 2.0 (Roche Applied Science, Mannheim, Germany), in a total volume of 25 µL containing 0.1 µL reverse transcriptase reaction (diluted into 5 µL) as template, using a LightCycler FastStart DNA MasterPLUS SYBR Green I Kit, according to the manufacturer's manual (Roche Applied Science). The transcription levels of AtPT1, AVP1 and AHAs genes were normalized to ACT2. The following are the specific primer pairs for the different genes designed with LightCycler Probe Design2 software (Roche Applied Science): ACT2, 5′CCCGCTATGTATGTCGC3′ and 5′TCCAGCAAGGTCAAGACG3′; AtPT1, 5′CCTCCTCAAGTTGACTACATT3′ and 5′CTCGATATCTGT TTGTAAGACCT3′; AVP1, 5′GTTTCGTCACTGAGTACTACAC3′ and 5′TCATGATAGCAATAGCAAAGATTGGA3′; AHA1, 5′TCCATCCCTGT TGAGGAGT3′ and 5′ATATCTGCTTTCTTCAAAGCGG3′; AHA2, 5′ATTGACGGCAGTGGTAAC3′ and 5′CGAGCAACAGCCAACGA3′; AHA6, 5′AGATGAGATAATTGACAAGTTTGCT3′ and 5′TCTGCAC TGTCATGTCTTGGA3′. Their specificity was confirmed by blastn in the National Center for Biotechnology Information (NCBI).

Western blot analysis

Seeds of A. thaliana plants were germinated in half-strength MS medium for 2 weeks, and then transferred to 10 µm Pi plates. Samples of seedlings (1–3 g) were taken at different time points (0, 2, 4 and 6 days), and membrane protein was extracted as described elsewhere (Schumaker and Sze, 1985). For rice plants, seeds of control and T2 OsAVP1DOX plants were germinated in half-strength MS medium. Microsomal fractions were isolated from 10-day-old seedlings, as described elsewhere (Schumaker and Sze, 1985). The protein concentration was determined using the bicinchoninic acid (BCA) protein assay reagent (Pierce, Rockford, IL, USA); 15 mg per sample was electrophoretically resolved in 12% Tris-HCl–sodium dodecylsulphate (SDS) gel (Bio-Rad Laboratories) and transferred to Immobilon-P transfer membrane (Millipore, Bedford, MA, USA). Membranes were incubated for 1.5 h with an antiserum raised against a synthetic peptide corresponding to the putative hydrophilic loop IV of the AVP1 protein (Rea et al., 1992) or the 2E7 monoclonal antibody against V-ATPase (Ward et al., 1992), or with a polyclonal antiserum raised against Arabidopsis P-ATPase (Bouche-Pillon et al., 1994). After 1.5 h of incubation with a secondary antibody conjugated with alkaline phosphatase, the membranes were treated with a nitroblue tetrazolium chloride/5-bromo-4-chloroindol-3-yl phosphate p-toluidine salt (NBT/BCIP) substrate solution (Roche, Indianapolis, IN, USA) for staining.

Constructs

pAVP1::GUS

DNA from Bac (F7H2) containing the AVP1 promoter region was employed to amplify a 1.7-kb fragment upstream to the ATG codon using the following primers: sense, 5′GCTCTAGACGTTTACCACACC AGTCACCAC3′ with an XbaI restriction site at the 5′ end; antisense, 5′CGGGATCCCTTCTCTCCTCCGTATAAGAGA3′ with a BamHI restriction site at the 5′ end. The amplified ~1.7-kb AVP1 promoter was ligated to the pGEM-T vector (Promega, Madison, WI, USA), sequenced, and then subcloned into the XbaI/BamHI site of the pBC308 vector in front of the GUS open reading frame (Xiang et al., 1999). The vector pBC308 contains the Bar gene (phosphinothricin acetyltransferase) for selection with the herbicide phosphinothricin (BASTA).

Transformation and selection

The construct was transformed into Agrobacterium tumefaciens strain GV3101, and then introduced into A. thaliana Col-0 ecotype via the floral dip method (Clough and Bent, 1998). Plants transformed with the pAVP1::GUS cassette were seeded in soil and selected by spraying with BASTA (T1). Seeds obtained from self-pollinated transformants (T2) were scored again for herbicide resistance on soil. Complete BASTA resistance identified homozygous pAVP1::GUS plants of the T3 progeny.

Leaf area

The rosette leaves were carefully excised with a scalpel blade, and the leaf areas were measured with a Li-Cor 4100 area meter (Lincoln, NE, USA).

Pi determination

Plant samples were placed in glass scintillation vials and dried at 70 °C for 72 h. The fresh and dry weights were determined on an analytical balance. The samples were ashed at 500 °C for 6 h. The ash samples were dissolved in 1 m HCl, and the Pi contents were determined using a colorimetric method (Murphy and Riley, 1962).

Lateral root, root length and root hair measurements

Root lengths were measured directly with a ruler. The lateral root number and the root hair number were counted under an Olympus SZ40 stereomicroscope (Tokyo, Japan). Root hair photographs were taken and printed, and the root hair lengths on the photographs were measured with a ruler. The final values were converted to the actual size of the root hair.

Root acidification

Plants were germinated in half-strength MS medium for 7 days, transferred to low-Pi medium as described above with 1 mm MES, pH 6.8 and 0.04 g/L bromocresol purple, and incubated for 10 days. The pH change was visualized via changes in medium colour. Comparisons were made with a colour bar generated by documenting the colour change of bromocresol purple in the same medium at specific pH values.

Pi uptake determination

Pi uptake experiments were performed in 125-mL flasks wrapped with aluminium foil. Plants grown hydroponically were used 2 weeks after bolting. After 2 days of incubation in distilled water, the plants were transferred to the flasks filled with 120 mL of medium supplemented with 50 µm Pi. The solution volume was maintained by adding distilled water every 4 h. The Pi concentrations in the medium were determined at 8-h increments for 96 h using the method of Murphy and Riley (1962). This method can determine Pi concentrations as low as 1 µm in seawater, and the salt error is less than 1%.

Rice transformation

To generate AVP1DOX rice plants, seeds from O. sativa var. japonica‘Taipei 309’ were surface sterilized and germinated on LS 2.5 medium [MS medium supplemented with 2.5 mg/L 2,4-dichlorophenoxyacetic acid (2,4-D), 2 g/L casein hydrolysate and 3% sucrose, solidified with 7 g/L agar] (Abdullah et al., 1986) to initiate calli. After 2 weeks of incubation in the dark at 25 °C, scutella were isolated from the starchy endosperm, transferred to fresh LS 2.5 medium and returned to the dark. Calli that formed yellowish, globular clusters were proliferated on the same medium to be used for microprojectile bombardment experiments on a Particle Delivery System (Model PDS-1000/He Biolistic, Bio-Rad Laboratories).

DNA of the following plasmids [pDM302 (Cao et al., 1992) + pRG389 = pRT103 (Topfer et al., 1987) with AVP1D cDNA (Zhen et al., 1997)] was precipitated onto 1-µm gold particles, as described previously (Kausch et al., 1995). The DNA-coated particles were pelleted by centrifugation at 28 g for 5 min, washed with 100% ethanol and resuspended in a final volume of 55 µL of 100% ethanol (anhydrous). The suspension was sonicated for 10 s immediately before dispensing. Ten microlitres of the DNA suspension were aliquoted onto each macrocarrier and allowed to air dry. The bombardment parameters were as follows: 7584-kPa rupture discs (Bio-Rad); rupture disc to macrocarrier gap distance, 5 mm; macrocarrier fly distance, 10 mm; stopping screen to target distance, 5 cm; partial vacuum, 94.82 kPa.

Three-week-old rice scutella bombarded with pDM302 + pRG389, or pDM302, were transferred to LS 2.5 medium supplemented with 3 mg/L glufosinate at 3 days post-bombardment, and incubated in the dark at 25 °C. Calli were subcultured to fresh LS 2.5 medium supplemented with 3 mg/L glufosinate at 4-week intervals and incubation was continued as described. Glufosinate-resistant calli were transferred to rice shoot regeneration medium (4.4 g/L MS salts, 2.0 mg/L 6-benzylaminopurine, 30 g/L sucrose, 7 g/L TC agar), and regenerated shoots were transferred to full-strength MS medium with 2% sucrose and 0.7% agar. Positive candidates were confirmed by Southern blot analysis.

Quantification of organic acids

The determination of organic acids was performed as described previously (Murphy et al., 1999). Plants were grown axenically in one-fifth Hoaglands medium, pH 4.85 with a 16-h day (140 µE/m/s light). Rice and tomato plants were grown at 23 °C and Arabidopsis seedlings were grown at 20 °C. For assays of tomato and rice plants, 20-day-old seedlings were used. Seedlings were transferred to treatment medium (control, 10 µm Pi and 10 µm Pi + 20 µm Al) for 12 h. Roots were washed once in a sterile hood for 5 min with deionized water, followed by washing with the medium to which they were to be transferred. Twenty seedlings were measured for each experiment. The experiments were repeated three times. The organic acid content was normalized to the fresh weight, and then expressed as a percentage of control levels with the percentage sum standard deviations.

K+ content determination

Plants were treated in the same way as in the organic acid exudation experiment. Root K+ contents were quantified using flame atomic absorption spectroscopy, as described by Murphy et al. (1999), and confirmed by inductively coupled plasma mass spectroscopy.

Statistical analysis

Student's t-test was used to analyse the significance of the growth differences under hydroponic conditions between OsAVP1D-2 and wild-type plants. One-way analysis of variance (anova), followed by pair-wise Holm–Sidak post-hoc analysis, was used to compare the development of root hairs, roots, shoots and fruits, as well as Pi contents, in AVP1OX Arabidopsis, tomato and rice plants vs. those in control plants under the different growing conditions tested. All statistical calculations were made using Sigma Stat (Systat Software, Point Richmond, CA, USA).

Acknowledgements

We thank H. Sze and R. Serrano for the V-ATPase and P-ATPase antisera, respectively. We are grateful to E. Lander and A. Rakin for help in data gathering and evaluation. We thank C. Morse for help with the rice growth experiments. We are grateful to B. Lahner for help with the ionic studies. This work was supported by grants from the National Research Initiative, US Department of Agriculture, Cooperative State Research, Education, and Extension Service no. 2006-35304-17339 to R.A.G. and A.M., and Storrs Agricultural Experimental Station Hatch to R.A.G. USDA 58-1940-0-007 to L.K.S.

The authors have no competing financial interests.

Ancillary