A phosphate transporter from Medicago truncatula is expressed in the photosynthetic tissues of the plant and located in the chloroplast envelope


Author for correspondence:Maria J. Harrison Tel: +1 580 224 6670 Fax: +1 580 224 6692 Email: mjharrison@noble.org


  • Phosphate is essential for many cellular processes including the light reactions of photosynthesis. Photosynthesis results in the production of triose phosphates that are transported across the chloroplast envelope to the cytosol in counterexchange for phosphate. Until recently, members of the plastid phosphate transport family, which mediate the exchange of phosphate for phosphorylated compounds, were the only proteins known to transport phosphate into the chloroplast.
  • Here, we characterized a phosphate transporter, MtPHT2;1 of Medicago truncatula. Transient expression of an MtPHT2;1-GFP fusion protein indicates that MtPHT2;1 is located in the chloroplast envelope.
  • The phosphate transport activity of MtPHT2;1 was assayed in yeast where the protein mediates phosphate uptake with a Km for phosphate of 0.6 m m and a pH optimum of 3–4.
  • MtPHT2;1 is expressed in all the photosynthetic tissues of the plant and transcript levels are also influenced by light, development and phosphate status of the plant. The phosphate transport activity and location in the chloroplast envelope membrane suggest a role for MtPHT2;1 in phosphate transport into the chloroplast.


Plants require a range of mineral nutrients for optimal growth and development. Phosphorus is one of the most important and comprises c. 0.2% of the d. wt of each cell (Schachtman et al., 1998). Found within every cellular compartment, phosphorus serves a variety of structural, energy transfer and regulatory roles. Of central importance to plants, the reactions of photosynthesis require phosphate for photophosphorylation and cytosolic phosphate levels regulate the subsequent partitioning of photosynthetically fixed carbon between the sucrose and starch biosynthetic pathways (Walker & Sivak, 1986). Consequently, plants must not only obtain significant quantities of phosphorus to sustain growth, but they must also control the cellular and subcellular concentrations in all tissues of the plant.

Plants acquire phosphorus as phosphate from the soil solution. As a result of its propensity to form complexes that are relatively insoluble, the concentration of phosphate in the soil solution is frequently less than 10 µm (Holford, 1997). Consequently, plants have evolved high-affinity phosphate transport systems in their roots to enable efficient acquisition of this sparingly soluble mineral. In addition, the majority of the vascular flowering plants are also able to obtain phosphate indirectly, via symbiotic interactions with arbuscular mycorrhizal fungi. These associations develop most readily during growth in low phosphate environments and the plant supplies the fungus with carbon while the fungus transports phosphate to the root cortex (Smith & Read, 1997; Harrison, 1999). Those plants that do not form arbuscular mycorrhizal associations have developed other adaptations to enhance their phosphate supply, such as the development of proteoid roots (Neumann et al., 2000).

Over the past few years, the cloning and analysis of phosphate transporters from potato (StPT1, StPT2) (Leggewie et al., 1997), Arabidopsis (AtPT1, AtPT2) (Muchhal et al., 1996; Smith et al., 1997), Medicago truncatula (MtPT1, MtPT2) (Liu et al., 1998; Chiou et al., 2001) and tomato (LePT1, LePT2) (Daram et al., 1998; Liu et al., 1998b) provided the first insights into the molecular nature of the proteins involved in phosphate transport in roots. These studies revealed that plants possess phosphate transporters that share sequence similarity with the high-affinity proton-coupled symporters of yeast (Bun-ya et al., 1991) and Neurospora crassa (Versaw, 1995) and belong to the PHS transporter family of the Major Facilitator Superfamily (Pao et al., 1998). Two closely related phosphate transporter genes were cloned from each of these species and the encoded proteins share between 70 and 95% identity. The ability of these proteins to mediate phosphate transport was verified via expression in yeast and they show Km values in the low- to mid-micromolar range. In a similar manner to the high-affinity phosphate transporters of yeast and N. crassa, most of the plant phosphate transporter genes are derepressed during growth in phosphate limiting conditions; however, their spatial expression patterns vary. The M. truncatula (MtPT1, MtPT2) and Arabidopsis (AtPT1, AtPT2) transporters are expressed only in roots (Muchhal et al., 1996; Smith et al., 1997; Liu et al., 1998b; Chiou et al., 2001), while LePT1 is also expressed in leaves (Daram et al., 1998; Liu et al., 1998a) and StPT1 is expressed in a very wide range of tissues, including roots, leaves, flowers, tubers and stolons (Leggewie et al., 1997). MtPT1 and LePT1 are located in the plasma membrane of root hair and epidermal cells and are probably responsible for phosphate transport at the root–soil interface (Daram et al., 1998; Liu et al., 1998; Chiou et al., 2001). The location and potential roles of the other transporters are not yet known but they may also be involved in transport at the root/soil interface or alternatively, may function to load phosphate into the xylem, a step that is also predicted to require a high-affinity transporter. In addition to these transporters, mycorrhiza-specific phosphate transporters have been identified from potato and M. truncatula. These transporters are expressed exclusively in mycorrhizal roots and their spatial expression patterns are consistent with a role in the acquisition of phosphate released by the arbuscular mycorrhizal fungi (Rausch et al., 2001; Harrison et al., 2002).

To date most studies of phosphate transporters have focused on the roots. However, a multitude of additional transport steps are required to move phosphate to the cells of the aerial tissues of the plant. Based on their expression patterns, transporters such as LePT1 and StPT1 are candidates for a role in some of these processes (Leggewie et al., 1997; Liu et al., 1998a); however, it is clear that other transport systems will also be involved. Sequence analysis of the Arabidopsis genome predicts the existence of at least seven other phosphate transporters of the PHS family, which may fulfill some of these roles.

In addition to uptake into each cell, transport proteins are also required to mediate the movement of phosphate between subcellular compartments and organelles. Phosphate is stored in the vacuole and measurements of phosphate transport into vacuoles suggest the presence of active transport system and also a channel (Mimura, 1999); however, the nature of the proteins involved is unknown. Phosphate transport into mitochondria is mediated by carriers belonging to the mitochondrial phosphate carrier family and a number of these have been cloned from plants (Takabatake et al., 1999). Likewise, a range of translocators that mediate the counterexchange of phosphate and triose phosphates, phosphoenolpyruvate or glucose-6-phosphate across the chloroplast/plastid envelope have also been cloned (Flügge et al., 1989; Fischer et al., 1997; Kammerer et al., 1998). In addition to these exchangers, it was shown recently that the Arabidopsis chloroplast envelope also contains a phosphate transporter that shares sequence similarity with the high-affinity, Na-coupled transporters from yeast and N. crassa (Versaw & Harrison, 2002). The gene, designated PHT2;1, is expressed in leaves and analysis of the transport properties of PHT2;1 indicated that it mediates low affinity phosphate transport in yeast with a Km of 0.4–0.8 mm (Daram et al., 1999; Versaw & Harrison, 2002). Surprisingly, transport showed a pH optimum consistent with that of a proton-coupled symporter and transport was not stimulated by sodium. PHT2;1 transcripts are present in the mesophyll cells of the leaf (Daram et al., 1999) and transient expression of a PHT2;1-GFP fusion indicated that the protein is targeted to the chloroplast envelope, where it is predicted to play a role in phosphate transport into the chloroplast (Versaw & Harrison, 2002). Analysis of a PHT2;1 knockout mutant revealed that the transporter impacts phosphate allocation throughout the plant and is required for the correct regulation of some phosphate starvation responses (Versaw & Harrison, 2002).

While analyses in Arabidopsis have enabled the identification and dissection of the roles of a number of phosphate transporters, the extension of these studies to include homologs from other plant species has provided further insights (Liu et al., 1998). As part of a research effort centered on the analysis of phosphate transport in M. truncatula with and without symbiotic associations, we cloned and characterized the M. truncatula homolog of PHT2;1. MtPHT2;1 is also located in the chloroplast envelope but in contrast to PHT2;1, it is regulated in response to phosphate and during plant development.

Materials and Methods

Plant materials

Medicago truncatula Gaertn ‘Jemalong’ (Line A17) seeds were submerged in concentrated sulfuric acid for 5 min, surface sterilized in 30% commercial bleach for 5 min, rinsed thoroughly in sterile distilled water and incubated on damp filter paper in Petri plates at 4°C for 14 d. Seedlings were moved to room temperature and then exposed to light. For the materials used in the experiment shown in Fig. 5 , the cotyledons were harvested 48 h after moving to room temperature and 24 h after exposure to the light. The remaining seedlings were planted into pots contained the sterile, acid-washed sand and were watered with deionized water and fertilized twice weekly with modified Hoagland's solution containing 20 µ m or 2 m m phosphate ( Arnon & Hoagland, 1940 ). The plants were maintained in a growth chamber under a 16-h light, 8-h dark cycle at 25°C/23°C. The first trifoliate leaves were harvested 15 d after planting. Leaves, shoots and roots were harvested at 22, 29, 36, 43 and 50 d after planting. Flowers were harvested at 36, 43 and 50 d. All tissue samples were frozen in liquid nitrogen and stored at −80°C. The shoot samples included a mixture of stems and leaves.

Figure 5.

Northern blot analysis with 10 µg of RNA from hypocotyls, cotyledons, leaves, stems, shoots flowers, pods and roots at 15, 22, 29, 36, 43 and 50 d post germination. The numbers in brackets indicate the age of the plant from which the tissues were harvested. The shoot sample is comprised of a mixture of leaves and stems. The blot was probed with MtPHT2;1 (panel 1) and then stripped and reprobed with an 18S RNA probe (panel 2) to enable sample normalisation.

For the light : dark experiment, plants were watered twice weekly with modified Hoagland's solution containing 20 µm, 200 µm or 2 mm phosphate (Arnon & Hoagland, 1940). Samples were harvested at 4 wk post germination during the midpoint of the light or dark periods of the photoperiod. For the experiment in which young and old leaves were harvested, the plants were grown and fertilized as describe above. The young leaf sample consisted of the youngest trifoliate leaf that had just unfolded, while the old leaf sample consisted of the two oldest trifoliate leaves closest to the cotyledons. Harvesting was carried out at 5 and 6 wk post germination at the midpoint of the light period.

Isolation of the MtPHT2;1 cDNA clone

The MtPHT2;1 cDNA was identified from a lambda ZAP II (Stratagene, La Jolla, CA, USA) cDNA library prepared from shoot tissues of 30-d-old M. truncatula plants grown with low phosphate fertilization (J. Liu and M. J. Harrison, unpublished) using the Arabidopsis thaliana PHT2;1 coding region as a probe. The 1.8 kb probe was generated by digestion of PHT2;1/pWV3 with BamHI and XhoI. Screening was carried out according to standard molecular biology procedures (Sambrook et al., 1989). Hybridization was carried out under low stringency conditions in 0.5 m NaHPO4 (pH 7.2), 1 mm EDTA, 7% (w/v) SDS at 50°C. The positive phage clones were converted to pBluescript plasmids according to the manufacturer's instructions (Stratagene, La Jolla, CA, USA). A full length cDNA clone of 2185 bp, called MtPHT2;1/pBluescript was sequenced on both strands. This MtPHT2;1 sequence is available in GenBank (Accession number AF533081).

Sequence analysis

Sequence alignments were prepared by ClustalW (Thompson et al., 1994) at the European Bioinformatics Institute http://www2.ebi.ac.uk/clustalw site. The tree shown in Fig. 2 was created in Treeview (Page, 1996).

Figure 2.

Southern blot of Medicago truncatula DNA hybridized with a probe corresponding to the coding sequence of the MtPHT2; 1 gene. BglII, PstI and EcoRV have more than one restriction site in the MtPHT2;1 coding region. Size markers (bp) are shown to the left of the blot.

ChloroP (Emanuelsson et al., 1999) http://www.cbs.dtu.dk/services/ChloroP/. Target P (Emanuelsson et al., 2000) http://www.cbs.dtu.dk/services/TargetP/.

Subcellular localization of MtPHT2;1

PCR-generated SalI-NcoI fragments containing either the full MtPHT2;1 ORF or the first 508 bp, which corresponds to the N-terminal 170 amino acids, were cloned in-frame to the 5′ end of GFP in the plasmid CaMV35S-sGFP(S65T)-Nos (Chiu et al., 1996) to generate MtPHT2;1-GFP and MtPHT2;1(1–170)-GFP, respectively. These chimeric constructs were introduced into leaves of 5-wk-old M. truncatula plants by particle bombardment using the Biolostic PDS-1000/He particle delivery system (Bio-Rad, Hercules, CA, USA). Bombardment experiments were conducted using 900 psi pressure rupture disks, a vacuum of 28 in of Hg, and 1.0 µm diameter gold particles. After bombardment, leaves were placed in Petri plates on water-saturated filter paper, and incubated at room temperature for 24–48 h before imaging with confocal microscopy. Cells were imaged with a Bio-Rad 1024 ES confocal laser scanning microscope equipped with a × 63 (numerical aperture, 1.2) water-immersion objective. GFP fluorescence was excited using the 488 nm line of the krypton-argon laser and emission detected at 522 nm. Chlorophyll autofluorescence was excited at 488 nm and emission detected at 680 nm. Images shown are single optical sections.

Yeast phosphate transport assays

The MtPHT2;1 cDNA clone was amplified by PCR using primers designed to introduce unique BamHI and XhoI sites at the 5′- and 3′ ends of the gene, respectively. The cDNA fragment was subcloned into the yeast expression vector pWV3 (Versaw & Harrison, 2002) to create MtPHT2;1/pWV3. The plasmid MtPHT2;1(Δ1-66)/pWV3 was constructed similarly but the first 198 bp of the MtPHT2;1 ORF, which corresponds to the encoded 66 amino acid transit peptide has been deleted and a new ATG start codon has been inserted. These constructs were tested for their ability to complement the phosphate uptake defect of the yeast mutant PAM2 (Martinez & Persson, 1998). Transformed yeast were grown to an OD600 of 1.0 in SD medium containing 0.22 mm Pi and 25 mm Na-citrate buffer pH 4.5, washed with Pi-free medium and suspended in same at 200 mg of cells per ml. Ten microliters of the cell suspension was diluted into 190 µl of phosphate-free medium of the desired pH and preincubated at 30°C for 5 min. Transport was initiated by addition of 33Pi (0.17 TBq mol−1) to the desired final concentration. Uptake assays were stopped by the addition of ice-cold 25 mm Na-citrate pH 4.5 followed by filtration on glass fiber filters and an additional wash. Accumulated radioactivity was measured by scintillation spectroscopy. Kinetic data were analyzed by nonlinear regression.

Southern blot and Northern blot analyses

Genomic DNA was isolated from Medicago truncatula leaves according to Dellaporta et al. (1983). Southern blots were prepared according to standard procedures (Sambrook et al., 1989). Total RNA for northern analysis was isolated from cotyledons, primary roots, leaves, stems, shoots, flowers, young pods and roots using the GibcoBRL TRIZOL Reagent Ultra Plus (GibcoBRL, Life Technologies, Inc. Gaithersburg, MD, USA). RNA samples (5–10 µg) were separated by electrophoresis on formaldehyde agarose gels and then transferred to GeneScreen Plus Membranes (NEN Life Science Products, INC. Boston, MA.) using 10× SSC. Southern and northern blots were hybridized according to standard procedures (Church & Gilbert, 1984; Sambrook et al., 1989). The MtPHT2;1 probe was a 1732-bp fragment comprising the complete coding region and was obtained by digestion of MtPHT2;1/pWV3 with BamHI and XhoI. Northern blots were hybridized with the MtPHT2;1 probe and an Mt4 cDNA probe (Burleigh & Harrison, 1997) simultaneously. The blots were then stripped and rehybridized with an 18S probe. After each hybridization the Northern blots were scanned and quantified using a phosphorimage analysis system (Storm imaging system, Molecular Dynamics, Sunnyvale, CA, USA). The values from the 18S hybridizations enabled normalization of the samples and adjustments for inequalities in loading. The relative levels of MtPHT2;1 transcripts were calculated from normalized values. In some instances, the differences in MtPHT2;1 transcript levels are shown with a standard error. These were calculated from a minimum of three independent samples.

Determination of phosphate content

The inorganic phosphate content of roots and shoots were determined using a phosphomolybdate colorimetric assay as described by Ames (Ames, 1966). Freshly harvested roots and shoots were frozen in the liquid nitrogen, ground and then suspended in 1% acetic acid. Cellular debris was pelleted by centrifugation and the phosphate content of the supernatant was assayed in triplicate.


Identification of Medicago truncatula PHT2;1

A full-length cDNA clone, MtPHT2;1, was isolated from a phosphate-starved shoot cDNA library using a probe corresponding to the A. thaliana PHT2;1 gene (Daram et al., 1999; Versaw & Harrison, 2002). The MtPHT2;1 cDNA is 2185 bp in length and is calculated to encode a protein of 574 amino acids with a molecular mass of 60.55 kDa (AF533081). MtPHT2;1 is predicted to be an integral membrane protein with the 12 membrane-spanning domain structure common to the proteins of the major facilitator superfamily (Pao et al., 1998). Comparison with Arabidopsis PHT2;1 revealed that they share 71% identity. MtPHT2;1 is the second phosphate transporter of this type to be reported from plants; however, a sequence from ice plant (U84890), that shares 81% identity with MtPHT2;1, is present in the NCBI databases. After the plant proteins, MtPHT2;1 shares most similarity (38–41%) with phosphate permeases from a range of bacteria including Chlamydophila pneumoniae, Chlamydia sp. and Pyroccocus abyssi, and limited identity with PHO89 and PHO-4, the sodium-coupled phosphate transporters of yeast and N. crassa (Fig. 1) (Versaw & Metzenberg, 1995; Martinez & Persson, 1998).

Figure 1.

Unrooted tree based on a ClustalW alignment showing MtPHT2;1 (bold) and the most closely related phosphate transporters of the PiT family. Members of the PHS family are within the dotted square.

Hybridization of the MtPHT2;1 cDNA to Southern blots of M. truncatula DNA resulted in one or two hybridizing bands consistent with a single copy gene (Fig. 2). However, under low stringency conditions additional hybridizing bands were visible suggesting that there are other genes with similarity to MtPHT2;1 in the genome (data not shown). Analysis of the M. truncatula gene index (TIGR) also indicated the presence of a second sequence (TC52549) with similarity to MtPHT2;1.

Subcellular localization of MtPHT2;1 in M. truncatula leaves

By contrast to the bacterial and fungal proteins, both PHT2;1 and MtPHT2;1 possess long N-terminal extensions. Recent analysis of PHT2;1 indicated that this region is a chloroplast transit peptide and PHT2;1 is located in the chloroplast envelope (Versaw & Harrison, 2002). The sequence and distribution of charged amino acid residues in this region of MtPHT2;1 are significantly different to PHT2;1. Therefore, to determine whether MtPHT2;1 is also targeted to the chloroplast, the full-length MtPHT2;1 coding region was fused with that of the green fluorescence protein (GFP) and introduced into M. truncatula leaves by particle bombardment. Green fluorescence was not detected in leaves bombarded with this construct although 35S-GFP positive controls yielded numerous cells expressing GFP (data not shown). It is possible that the lack of fluorescence was the result of instability of the large fusion protein, as bombardment with a truncated fusion construct, MtPHT2;1(1–170)-GFP, which encodes the N-terminal 170 amino acids of MtPHT2;1 fused to GFP, yielded green fluorescence signals in the chloroplasts of transformed cells (Fig. 3). Confocal microscopy revealed that the GFP signal is strongest at the chloroplast periphery, whereas the red chlorophyll autofluorescence signal is uniform. All of the MtPHT2;1-GFP signals detected had a corresponding red signal suggesting that MtPHT2;1 is located in chloroplasts but not other plastids.

Figure 3.

A MtPHT2;1-GFP fusion protein colocalizes with chloroplasts. A chimeric MtPHT2;1-GFP construct encoding the MtPHT2;1 N-terminal 170 amino acids fused to the N-terminus of GFP was introduced into Medicago truncatula leaves by particle bombardment and the localization of fluorescent signals was examined by confocal microscopy 48 h after transformation. Panels A and B indicate the GFP and chlorophyll fluorescent signals, respectively, and Panel C is their merged image. Bar = 10 µm.

MtPHT2;1 functions as a phosphate transporter in yeast

To determine whether MtPHT2;1 encodes a functional phosphate transporter and to analyze the phosphate transport activity of this protein, the MtPHT2;1 cDNA was cloned into a yeast expression vector pWV3 under the control of the yeast alcohol dehydrogenase promoter and introduced into a yeast phosphate transport mutant strain, PAM2. PAM2 lacks the high affinity proton-coupled and sodium-coupled phosphate transporters, PHO84 and PHO89 (Bun-ya et al., 1991; Martinez & Persson, 1998). Phosphate transport assays revealed that yeast cells expressing MtPHT2;1 show considerably higher phosphate uptake rates than cells expressing the empty expression vector, pWV3. MtPHT2;1 mediates saturable phosphate uptake with a pH optimum of 3–4 and an apparent Km of 0.61 ± 0.03 mm (Fig. 4). As expected, PAM2 cells carrying pWV3 also exhibit saturable phosphate uptake via the remaining low-affinity transport system and show a Km of 1.21 ± 0.09 mm.

Figure 4.

MtPHT2;1 mediates phosphate transport in yeast. (a) Effect of pH on the rate of phosphate uptake. Phosphate uptake by yeast PAM2 cells expressing MtPHT2;1 (cross-hatched bars), MtPHT2;1(Δ1-66) (black bars), or carrying a vector control (open bars) was determined at the indicated pH values. Uptake medium contained 25 m m citrate buffer to maintain the desired pH and 0.5 m m phosphate. Values from one of three independent experiments are shown, mean ± SE for three replicates. (b) Phosphate uptake rate in yeast PAM2 cells expressing MtPHT2;1 (filled squares) or carrying a vector control (open squares) as a function of external phosphate concentration at pH 4.0. Values shown are the mean ± SE for three independent experiments, each with at least duplicate samples. Lines are derived from nonlinear regression analysis.

When expressed in yeast, the full length Arabidopsis phosphate transporter, PHT2;1 was targeted to the mitochondria (Versaw & Harrison, 2002). Deletion of the chloroplast transit peptide resulted in a shift in its intracellular localization from the mitochondrion to the plasma membrane and a concomitant increase in phosphate transport activity of c. 8-fold (Versaw & Harrison, 2002). By contrast, deletion of the first 66 amino acids of MtPHT2;1, predicted to encompass the putative chloroplast transit peptide, had no significant effect on phosphate uptake (Fig. 4).

MtPHT2;1 is expressed in the aerial tissues of the plant and transcript levels decrease in the leaves of older plants

The expression of MtPHT2;1 was examined in primary root, hypocotyls, cotyledons, leaves, stems, roots, flowers and pods at 15, 22, 29, 36, 43 and 50 d postgermination. Northern blot analysis revealed that MtPHT2;1 is expressed in all of the above ground tissues of the plant. Transcripts were present at all stages of plant development from cotyledons to mature leaves and pods with the highest levels in leaves of younger plants (Fig. 5). Transcript levels were lowest in the leaves of older plants, 50 d postgermination. By contrast to the M. truncatula phosphate transporters described previously, MtPHT2;1 transcripts were not detected in roots at any stage of development (Fig. 5).

MtPHT2;1 transcript levels increase during growth under high phosphate conditions but are not affected by development of arbuscular mycorrhizal or nitrogen-fixing symbioses

In Arabidopsis, PHT2;1 transcript levels do not change in response to growth under different phosphate conditions. To determine whether this is also the case for MtPHT2;1, M. truncatula plants were fertilized with 20 µm or 2 mm phosphate twice weekly and leaf and root samples were harvested at 22, 29, 36, 43 and 50 d postgermination. The phosphate content of the leaves and roots was significantly higher in the plants that received the high phosphate fertilization and the phosphate content also increased throughout development (Fig. 6a). The plants that received the low phosphate treatment showed significantly lower levels of phosphate in both leaf and root tissues and this remained constant throughout development. Northern blot analysis demonstrated that MtPHT2;1 transcript levels were slightly higher during growth under phosphate sufficient conditions. This difference was relatively small, 1.4 (± 0.1) fold, but was seen consistently, throughout development (Fig. 6b). As observed previously, MtPHT2;1 was not expressed in roots. As a control, the same blots were hybridized with Mt4, a phosphate-starvation induced gene (Burleigh & Harrison, 1998) that serves as a molecular indicator of the phosphate status of the plant. As expected, Mt4 transcripts were present in both roots and leaves of plants that received the low phosphate treatment, but not in those that received the high phosphate treatment.

Figure 6.

MtPHT2;1 transcript levels are higher during growth under high phosphate conditions. (a) Phosphate content of roots (open bars) and shoots (shaded bars) of plants fertilized twice weekly with 20 µ m or 2 m m phosphate. Tissues were sampled at 22, 29, 36 and 43 days postgermination. (b) Northern blot analysis with 10 µg of RNA from shoots and roots. The blot was probed with MtPHT2;1 (panel 1) and Mt4 (panel 2) simultaneously and then stripped and reprobed with an 18S RNA probe (panel 3) to enable sample normalisation.

MtPHT2;1 transcript levels were also examined during development of symbiotic associations with an arbuscular mycorrhizal fungus, Glomus versiforme and with a nitrogen-fixing bacterium Sinorhizobium meliloti . Transcript levels were monitored in shoots from nonmycorrhizal and mycorrhizal plants at five timepoints postinoculation with G. versiforme but no significant changes were observed (data not shown). Likewise, MtPHT2;1 transcripts were monitored in shoots of plants during development of nitrogen-fixing nodules but no significant differences were observed (data not shown).

MtPHT2;1 transcript levels are higher during the light phase of the photoperiod

To further elucidate factors regulating MtPHT2;1 expression, plants were fertilized twice weekly with 20 µm, 200 µm or 2 mm phosphate and at 4 wk postgermination, leaf tissue was harvested at the midpoints of the light and dark phases of the photoperiod. As observed previously, MtPHT2;1 transcript levels were slightly higher in plants grown with sufficient phosphate. In addition, these analyses also revealed that MtPHT2;1 transcripts vary through the photoperiod and are 1.4-, 1.5- and 1.1-fold higher in the light relative to the dark in the 20 µm, 200 µm or 2 mm phosphate fertilized samples, respectively (Fig. 7). This difference is greatest in plants grown under phosphate deficient conditions. Consistent with previous studies, the phosphate contents of the leaf tissues are slightly higher in the light than the dark, although this was not seen at the highest level of phosphate fertilization (Table 1) (Morin et al., 1991). The expression of the Mt4 gene also reflects these differences in phosphate content.

Figure 7.

Northern blot analysis with 10 µg of RNA from shoots of plants fertilized twice weekly with 20 µ m , 200 µ m or 2 m m phosphate. Tissues were sampled from plants at 4 wk post germination at the midpoint of the light (L) and dark (D) cycles. The blot was probed with MtPHT2;1 (panel 1) and Mt4 (panel 2) simultaneously and then stripped and reprobed with an 18S RNA probe (panel 3) to enable sample normalisation.

Table 1.  Phosphate content of Medicago truncatula plants grown with three levels of phosphate fertilization and harvested at the mid-point of the light and dark cycles
Phosphate fertilization (µM)TissuePhosphate content (nmol mg−1 f. wt)
20Leaves0.97 (± 0.11)0.42 (± 0.08)
200Leaves1.41 (± 0.04)0.59 (± 0.04)
2000Leaves22.3 (± 1.5)26.8 (± 0.7)

MtPHT2;1 transcript levels are higher in young leaves

Under conditions of phosphate sufficiency, most plants store phosphate in the vacuole and under conditions of phosphate starvation, phosphate from older leaves is mobilized from the vacuole and redistributed to the younger growing tissues of the plant (Mimura, 1995). We examined MtPHT2;1 transcript levels in the youngest and oldest leaves of plants fertilized with 20 µm or 2 mm phosphate. The harvests were made at the midpoint of the light period at 5 and 6 wk postgermination. Northern blot analysis revealed that MtPHT2;1 transcripts were 1.8 (± 0.2) fold higher in the young leaves relative to old leaves (Fig. 8). As expected, the phosphate content of old leaves from plants grown in high phosphate was significantly higher than the young leaves. The phosphate content of leaves from plants grown under low phosphate conditions was much lower and there was no significant difference between the young and old leaves (Table 2).

Figure 8.

Northern blot analysis with 10 µg of RNA from leaves of plants fertilized twice weekly with 20 µ m or 2 m m phosphate. The youngest (Y) and oldest (O) leaves were sampled at 5 (left panel) and 6 (right panel) weeks postgermination. Samples were collected at the midpoint of the light cycle. The blot was probed with MtPHT2;1 (panel 1) and Mt4 (panel 2) simultaneously and then stripped and reprobed with an 18S RNA probe (panel 3) to enable sample normalisation.

Table 2.  Phosphate content of young and old leaves of Medicago truncatula grown at 2 levels of phosphate fertilization
Phosphate fertilization (µM)Age of plant (wk post germination)Phosphate content (nmol mg−1 f. wt)
oldest leafyoungest leaf


Until recently, the only transport systems known to mediate the movement of phosphate across the chloroplast envelope were members of the phosphate translocator (PT) family. These proteins mediate the counter-exchange of inorganic phosphate and phosphorylated compounds and include the triose phosphate/phosphate translocator (TPT) (Flügge et al., 1989; Schulz et al., 1993), the phosphoenolpyruvate/phosphate (PPT) translocator (Fischer et al., 1994) and the glucose-6-phosphate/phosphate (GPT) translocator (Kammerer et al., 1998; Neuhaus & Wagner, 2000). However, in addition to these proteins, recent analyses revealed that an Arabidopsis transporter (PHT2;1) belonging to the inorganic phosphate transporter (PiT) family is also located in the chloroplast inner envelope membrane (Versaw & Harrison, 2002).

Here we have identified and characterized a PHT2;1 homolog, MtPHT2;1, from M. truncatula. MtPHT2;1 is a single copy gene but in contrast to Arabidopsis, low-stringency Southern blots indicated the presence of at least one other similar sequence in the M. truncatula genome and a candidate EST that might represent this gene, has been identified from the TIGR gene index. Overall, MtPHT2;1 shares 71% identity with PHT2;1 and 81% identity over the mature protein. At the sequence level, MtPHT2;1 and PHT2;1 are distinct from the proton-coupled transporters of the PHS family, such as the Arabidopsis AtPT1/AtPT2 (Muchhal et al., 1996) and M. truncatula MtPT1/MtPT2 (Liu et al., 1998b) and they share most similarity with proteins from a range of bacteria and archea. These include obligate, intracellular pathogens such as Chlamydophila pneumoniae, opportunistic pathogens such as Psuedomonas aeruginosa and extremophiles, such as Thermatoga maritima and cyanobacteria such as Nostoc sp. That plants share a more recent ancestor of PHT2;1 with prokaryotes rather than other eukaryotes may be a reflection of the prokaryotic origins of the chloroplast.

PHT2;1, MtPHT2;1 and the ice plant protein (U84890) possess long N-terminal extensions that are not present in the related bacterial, archae or fungal phosphate transporters. ChloroP, a program designed to identify chloroplast transit peptides (Emanuelsson et al., 1999), predicted that the N-terminal region of PHT2;1 was a chloroplast transit peptide and experimental evidence demonstrated that PHT2;1 is located in the chloroplast envelope (Versaw & Harrison, 2002). ChloroP predicts that MtPHT2;1 is located in the chloroplast and suggests two possible cleavage sites for the transit peptide. The first, between residues 16 and 17 and the second, between residues 66 and 67. The second cleavage site results in a transit peptide of 66 amino acids and most closely matches the cleavage site of PHT2;1. Transient expression of a MtPHT2;1-GFP fusion protein revealed that like PHT2;1, MtPHT2;1 is located in the chloroplast envelope. Although not proven directly, it is likely that MtPHT2;1 is located in the inner envelope membrane. The inner membrane is the permeability barrier between the chloroplast stroma and the cytoplasm and is the location of TPT and other transport proteins (Heldt & Sauer, 1971). In addition, the structure of the MtPHT2;1 transit peptide sequence resembles those of other inner envelope membrane proteins, and the majority of proteins destined for the outer envelope membrane lack transit peptides (Li & Chen, 1996; Lee et al., 2001). The ice plant phosphate transporter is also predicted to be located in the chloroplast envelope with a transit peptide sequence of 64 amino acids. It has been shown experimentally that the transit peptides of proteins destined for the chloroplast inner envelope membrane actually provide the information for targeting to the chloroplast stroma and the sequences important for integration of the protein into the inner envelope membrane reside within the mature protein (Brink et al., 1995). With this in mind, it is interesting to note that a motif identification program, MEME (Bailey & Elkan, 1994), suggests that the three plant phosphate transporters share a unique sequence motif (ENDDDFPGMA QAFHISSNTASAISICIAFAALTFPFFMTSLGQGLALKTK) just downstream of the transit peptide cleavage sites. While there is currently no experimental evidence to indicate a function for this region of the protein, it is conspicuous by its absence from the bacterial, archae and fungal sequences and might comprise a region important for location of these proteins in the chloroplast inner envelope membrane.

To study the phosphate transport activity of MtPHT2;1, we expressed the protein in a yeast phosphate transport mutant, PAM2. These analyses demonstrated that MtPHT2;1 functions as a phosphate transporter with an apparent Km for phosphate of 0.61 mm and a pH optimum between 3 and 4. The transport characteristics of MtPHT2;1 are very similar to those of PHT2;1 and are consistent with operation via proton-coupled symport (Daram et al., 1999; Versaw & Harrison, 2002). The location of the protein coupled with the phosphate uptake activity observed in yeast leads us to suggest that MtPHT2;1 is involved in phosphate uptake into the chloroplast. Furthermore, the relatively high Km for phosphate is consistent with concentration of phosphate in the plant cytosol, which is in the order of 10 mm (Mimura, 1995). Heterologous expression of chloroplast inner envelope proteins in yeast is not always straightforward as some of the inner envelope proteins have transit peptides that form a positively charged amphiphilic α-helix (Flügge et al., 1989; Dreses-Werrlingloer et al., 1991; Willey et al., 1991), which is a feature of mitochondrial presequences. When expressed in yeast, these proteins are targeted to the mitochondria (Brink et al., 1994). This occurred for PHT2;1 and removal of the transit peptide was required before PHT2;1 could be detected in the plasma membrane. This shift in location was associated with a large increase in the phosphate transport activity of the yeast cells (Versaw & Harrison, 2002). By contrast, deletion of the MtPHT2;1 transit peptide had no significant effect on the phosphate transport activity in yeast. The most likely explanation is that the MtPHT2;1 chloroplast transit peptide is not recognized by the yeast mitochondrial trafficking systems and that both the full length and truncated forms of MtPHT2;1 are located in the plasma membrane. In further support of this hypothesis, the PHT2;1 and MtPHT2;1 transit peptides do not show a high degree of sequence conservation and Target P, a program designed to predict the location of chloroplast and mitochondial proteins (Emanuelsson et al., 2000), gives PHT2;1 a mitochondrial location but suggests that MtPHT2;1 is located in the chloroplast. Thus, the transit peptide of MtPHT2;1 may share fewer of the structural features of the mitochondrial targeting sequences with the result that the protein is not targeted to the mitochondria in yeast.

Characterization of MtPHT2;1 expression patterns provided clues about its regulation and revealed significant differences to PHT2;1. MtPHT2;1 is expressed in all of the above ground tissues of the plant including cotyledons, leaves, stems, flowers and pods, but not in roots. By contrast to PHT2;1, MTPHT2;1 transcript levels varied with development and were highest in the leaves of younger plants and declined as the plant grew older. In addition, expression of MtPHT2;1 is affected by the phosphate status of the plant and transcript levels were consistently lower in plants grown under low phosphate conditions. This differential expression was observed at all stages of development. The phosphate content of the tissues and the expression of a phosphate starvation-inducible gene, Mt4, indicated that the plants were experiencing phosphate deprivation under the low and medium phosphate fertilization regimes. From these initial expression analyses, we might have expected that MtPHT2;1 transcripts would increase in the leaves of mycorrhizal plants, which generally have a higher phosphate content than the nonmycorrhizal controls (Smith & Read, 1997). The finding that MtPHT2;1 transcript levels did not change suggests that the regulation of its expression is not simply a response to internal phosphate levels. The results of subsequent experiments, in which we monitored MtPHT2;1 expression in the youngest and oldest leaves of plants grown under a range of phosphate regimes, also support this conclusion. MtPHT2;1 transcript levels were highest in the youngest leaves; however, as expected, the oldest leaves of plants grown in high phosphate had a significantly higher phosphate content than that of the youngest leaves. Overall, MtPHT2;1 transcripts were highest in the youngest leaves of plants grown in high phosphate conditions and lowest in the oldest leaves of plants grown in low phosphate conditions. In these experiments, it is likely that the young leaves from plants grown with sufficient phosphate were the most photosynthetically active. A number of annual plants show a progressive, age-related loss of photosynthetic function that has been referred to as functional senescence (Wittenbach, 1982) and the old leaves sampled in this experiment probably had a lower photosynthetic rate, particularly those sampled from plants growing under phosphate-starvation conditions (Wittenbach, 1982). From these expression analyses and the finding that MtPHT2;1 transcript levels were consistently higher at the mid-point of the light cycle relative to the midpoint of the dark cycle, we propose that MtPHT2;1 transcript levels are highest in tissues predicted to be the most photosynthetically active. Based on the location of the protein in the chloroplast envelope and its role in the transport of phosphate, an essential substrate of photosynthesis, it is possible that MtPHT2;1 expression is regulated in response to photosynthetic cues.

Phosphate plays central roles in photosynthesis (Walker & Sivak, 1986). It is required in the chloroplast for photophosphorylation but also plays a regulatory role controlling the flow of carbon between the sucrose and starch biosynthetic pathways. While previous discussions of phosphate transport into the chloroplast centered on the triose phosphate/phosphate exchange reaction mediated via TPT, the PHT2;1 and MtPHT2;1 data indicate that chloroplasts have a second mechanism of accumulating phosphate, which is not linked to the export of triose phosphates. If the phosphate transport activity measured in yeast is a true reflection of the transport activity in the chloroplast, then this transporter provides a mechanism via which the chloroplast could accumulate significant quantities of phosphate at the expense of a proton gradient. This phosphate transport activity enables the chloroplast to obtain phosphate without a requirement for the end products of photosynthesis and may play a role in fine tuning the phosphate concentrations within the chloroplast. It will be interesting to determine the amount of MtPHT2;1 relative to TPT, the most abundant translocator of the inner envelope membrane (Flügge & Heldt, 1993; Schulz et al., 1993) and to determine whether plastids of heterotrophic tissues possess this type of phosphate transport system also.


The authors thank Cuc Ly for assistance with photoshop layout, Gary Dewbre for critical reading of the manuscript and members of the Harrison lab for general discussion. The work was supported by The Samuel Roberts Noble Foundation.