Arbuscular mycorrhizas contribute significantly to inorganic phosphate (Pi) uptake in plants. Gene networks involved in the regulation and function of the Pht1 family transporters in legume species during AM symbiosis are not fully understood.
In order to characterize the six distinct members of Pht1 transporters in mycorrhizal Astragalus sinicus, we combined cellular localization, heterologous functional expression in yeast with expression/subcellular localization studies and reverse genetics approaches in planta. Pht1;1 and Pht1;4 silenced lines were generated to uncover the role of the newly discovered dependence of the AM symbiosis on another phosphate transporter AsPT1 besides AsPT4.
These Pht1 transporters are triggered in Pi-starved mycorrhizal roots. AsPT1 and AsPT4 were localized in arbuscule-containing cells of the cortex. The analysis of promoter sequences revealed conserved motifs in both AsPT1 and AsPT4. AsPT1 overexpression showed higher mycorrhization levels than controls for parameters analysed, including abundance of arbuscules. By contrast, knockdown of AsPT1 by RNA interference led to degenerating or dead arbuscule phenotypes identical to that of AsPT4 silencing lines. AsPT4 but not AsPT1 is required for symbiotic Pi uptake.
These results suggest that both, AsPT1 and AsPT4, are required for the AM symbiosis, most importantly, AsPT1 may serve as a novel symbiotic transporter for AM development.
Arbuscular mycorrhiza (AM) is a mutualistic endosymbiotic association between fungi of the ancient phylum Glomeromycota and most vascular land plant species (Remy et al., 1994). More than 80% of terrestrial plant species are known to be able to form AM symbiosis (Denison & Kiers, 2011). This mutualism improves the plant's efficiency for uptake of mineral nutrients, particularly phosphates, and in return fungi obtain organic carbon from the plant to complete the fungal life cycle (Smith & Read, 2008). To form AM symbiosis, the two symbionts undergo a series of developmental transitions that enable the fungus to enter the root cortex and establish highly dichotomously branched arbuscules in which phosphate and other nutrients are delivered to the root (Bucher, 2007). During arbuscule development, a plant-derived membrane, the peri-arbuscular membrane (PAM) develops around the branching hypha and separates the fungus from the plant cell cytoplasm. Phosphate transport proteins essential for symbiotic inorganic phosphate (Pi) transfer to the plant cell reside in this membrane (Harrison et al., 2002). Symbiotic phosphate transport in AM plants is mediated by specific phosphate transporters (Karandashov & Bucher, 2005). Phosphorus (P), as an essential mineral nutrient, plays a key role in AM symbiosis. Because of phosphate's low mobility and thus its availability in soil (Ai et al., 2009), plants adopt a series of adaptive strategies to increase the acquisition of poorly available Pi. One of them is the formation of symbiosis with AM fungi.
Plant Pht1 transporters belong to the phosphate: H+ symporter (PHS) family within the major facilitator superfamily (MFS; Pao et al., 1998). So far, Pht1 transporters in the direct uptake pathway occurring through the epidemal cells have been observed in potato (StPT1 and StPT2; Rausch et al., 2001); Medicago truncatula (MtPT1 and MtPT2; Liu et al., 1998b); rice (OsPT1-10; Paszkowski et al., 2002); and barley (HvPT1 and HvPT2; Rae et al., 2003). AtPHR1 and OsPHR2 are key regulators in the Pi signalling pathway (Jia et al., 2011), and overexpression of the two genes results in excessive accumulation of Pi in shoots, and upregulation of the Pht1 genes under Pi-sufficient conditions (Liu et al., 2010). Currently, only a few of these Pi transporters have been functionally characterized. The double pht1;1 and pht1;4 mutant from Arabidopsis thaliana exhibits a significant decrease in Pi uptake and shoot accumulation under both Pi-sufficient and Pi-deficient conditions, indicating compensatory effects between the two transporters (Shin et al., 2004), and AtPht1;5 plays a critical role in mobilizing Pi from source to sink organs in accordance with the P status of the plant (Nagarajan et al., 2011). The fourth Pht1 member functionally characterized so far, AtPht1;9 transporter, mediates root Pi acquisition under Pi-deprived conditions, contributing to the overall plant Pi status (Remy et al., 2012).
Several mycorrhiza-specific Pht1 members have been identified in M. truncatula (MtPT4), rice (OsPT11), potato (StPT4 and StPT5) and tomato (SlPT4) (Harrison et al., 2002; Paszkowski et al., 2002; Nagy et al., 2005; Chen et al., 2007; Xu et al., 2007). Tissue localization studies revealed that the mycorrhiza-induced Pht1 genes are exclusively expressed in arbuscule-containing cells (Harrison et al., 2002). The spatial expression pattern of these genes suggests that a cell autonomous signal involved in arbuscule development, activates expression of these genes (Gomez-Ariza et al., 2009). Lysophosphatidylcholine (LPC) may be one of these signals, as it was shown to trigger mycorrhiza specific Pi transporter gene expression in potato and tomato (Drissner et al., 2007). Functional analysis of these Pi transporter promoters revealed that P1BS and MYCS elements are required to confer the AM-inducible PiT gene expression (Chen et al., 2011). MtPT4 is essential for symbiotic Pi transport and the development of AM symbiosis (Javot et al., 2007b). The mycorrhiza-inducible Pi transporters StPT3, StPT4 and StPT5 from potato, exhibit functional redundancy (Nagy et al., 2005). Knockdown of LjPT3 exhibits a decrease in Glomus mosseae arbuscules and necrotic root nodules (Maeda et al., 2006). However, the functions of other mycorrhiza-specific Pi transporters have not been described.
Here, we report on the characterization of two novel mycorrhiza-specific Pi transporters, AsPT1 and AsPT4, and four other Pht1 family members from Astragalus sinicus, designated as AsPT2, AsPT3, AsPT5 and AsPT6, respectively. Both AsPT1 and AsPT4 play critical roles during the development of AM symbiosis, but AsPT4 alone is required for symbiotic Pi uptake.
Materials and Methods
Plant materials, AM fungi inoculum and cultivation conditions
Astragalus sinicus L. was used in this study. Seeds were surface sterilized and then spread on 0.8% water agar plates for germination. In pot culture (Expt 1), the plantlets were adapted to glasshouse conditions (see Supporting Information Methods S1) for 7 d first, and then, an equal number of plants were transferred either to pots containing a soil : sand mixture (v/v, 1 : 2), or to pots containing the same mixture supplemented with roots of Astragalus sinicus colonized with Gigaspora margarita (BEG34) or Glomus intraradices (BEG141). All plants were supplied with Hoagland's nutrient solution containing the macro- and micronutrients (see Methods S1). The experiment comprised three replicates for each treatment. Plants were harvested at 4 wk post-inoculation (wpi). The freshly collected roots were washed with deionized water and part of the roots were immediately frozen in liquid nitrogen and stored at −70°C for subsequent RNA isolation. Other parts of the roots were used to evaluate colonization and for visualization of arbuscules.
Monoxenic AM cultures
In monoxenic culture (Expt 2), the AM fungus G. margarita or G. intraradices was grown monoxenically in mycorrhizal association with hairy roots of A. sinicus. The cultures are clones of A. sinicus roots that were initially transformed using Agrobacterium rhizogenes K599 as transformation system with the T-DNA of the root-inducing plasmid transferred to the plant genome (Boisson-Dernier et al., 2001; Li et al., 2008). AM-colonized cultures were maintained at a constant temperature of 26°C on petri dishes with 0.3% (w/v) Phytagel (Sigma-Aldrich) as the gelling agent and with MSR medium containing 3% sucrose as the C source and 30 μM Pi (as 6.41 mg KH2PO4 l−1).
Cloning of the Pi transporter genes
Based on the conserved amino acid sequences of legume plant Pi transporters (Harrison et al., 2002), we designed the following degenerate oligonucleotides: forward (AsPTF), 5′-TTCWGGYT YKGVYTYGGMWTYGG-3′, which encodes FRFWLGF; and reverse (AsPTR), 5′-TTVGGVCCRAAVTTSGCRAAGAAG-3′, which is antisense for FFFANFGP. Subsequent PCRs on the total DNA from A. sinicus resulted in 850–900 bp amplified DNA segments encoding six putative Pi transporters, named AsPT1 through AsPT6.
The 5′ and 3′ regions of these putative Pi transporter genes, were cloned via thermal asymmetric interlaced PCR (TAIL-PCR), inverse PCR (iPCR) and RACE. For TAIL-PCR, four relatively longer AD (LAD) primers of 33 or 34 nucleotides were designed (Liu & Chen, 2007). For iPCR, 3–5 μg genomic DNA isolated from A. sinicus were digested with EcoRI, KpnI, BamHI, NcoI or HindIII, then self-ligated and used for PCR. The 5′ RACE and 3′ RACE were carried out using the classic RACE protocols (Scotto-Lavino et al., 2006a,b). For the primers used, see Table S1.
RNA extraction, cDNA synthesis, RT-PCR and real-time RT-PCR
Total RNA was extracted from A. sinicus by Trizol reagents (Invitrogen) following the manufacturer's instructions. After treatment with DNase I (MBI Fermentus, Canada), the 1st cDNA strand for each RNA preparation was synthesized by using oligo (dT)18 and Reverse Transcriptase M-MLV (RNase H−) (TaKaRa, Dalian, China). RT-PCR was performed as the manufacturer's instructions, using gene-specific primers (Table S1). Real-time PCR was performed using the SYBR Green Master Mix (Toyobo, Japan), and then subjected to qPCR reactions according to the manufacturer's instructions using an Applied Biosystems Cycler System (Applied Biosystems, CA, USA). The gene-specific primers are listed in Table S1. All reaction mixtures were heated at 95°C for 30 s, and then subjected to 40 PCR cycles of 95°C for 20 s, 60°C for 20 s and 72°C for 20 s, with the resulting fluorescence being monitored. All the reactions were performed with three biological replicates, each with three technical replications. The relative quantity (RQ) was determined using the Cycler system with software; the data were analysed using the method.
AsPT1, AsPT4 and AsPT5 promoter-UidA constructs
The isolated fragments of the AsPT1, AsPT4 and AsPT5 promoters upstream of the translation initiator, ATG, were amplified by iPCR using primers (Table S1) which contained HindIII and BamHI restriction sites at the 5′ and 3′ ends. After digestion with the two restriction enzymes, the amplified fragments were purified and cloned into the linearized binary vector, pBI121, to replace the CaMV35S promoter in front of the GUS gene (Chen et al., 2011). The resulting constructs were the AsPT1, AsPT4 and AsPT5 promoter-GUS constructs. To investigate the possible involvement of the 5′UTR intron in the regulation of expression of AsPT5, two expression cassettes one without the intron (pAsPT5-261::GUS) and the other with the 5′UTR intron (pAsPT5-1280::GUS) were generated. The new vectors were transformed into A. sinicus roots by A. rhizogenes-mediated transformation (Li et al., 2008).
Construction of the AsPT4 promoter deletion series
A series of 5′ truncations of the putative promoter sequence of AsPT4 (670 bp) was created by PCR using specific primers (Table S1). The targeted deletion of the predicted MYCS sequence was generated by overlap and extension PCR (Table S1) for the promoter of AsPT4. The deletion and mutation promoter sequences were cloned into the HindIII- and BamHI-digested pBI121 binary vector to drive GUS gene expression. The resulting gene constructs were designated AsPT4P-1 to AsPT4P-5.
Electrophoretic mobility shift assay
In order to examine the putative AM-specific responsive transcription factor binding sequences in promoter regions of AsPT4, the electrophoretic mobility shift assay (EMSA) was performed according to the published report (Frenzel et al., 2006).
In situ hybridization
In situ hybridization was carried out as described previously (Kouchi & Hata, 1993; Liu et al., 1998a,b; Daram et al., 1999) using DIG-labelled antisense and sense probes synthesized from the 3′ end cDNA of AsPT1-4. DIG-labelled probes were generated with the T7 or SP6 RNA polymerase (TaKaRa) from linearized pGM-T (Tiangen, Beijing, China) containing the corresponding cDNA. The probes were labelled with DIG following the procedure described by the manufacturer (Roche; Methods S1).
Subcellular localization of GFP fusion proteins in yeast and onion epidermal cells
The subcellular localization of AsPT1 and AsPT4 in yeast was performed with a C-terminal fusion of AsPT1 or AsPT4 to GFP in strain MB192 as described elsewhere (Reinders et al., 2002). GFP was fused to the 3′ ends of AsPT1 and AsPT4 (Wu et al., 2011), then the fusion constructs CaMV35S-AsPT1 and CaMV35S-AsPT4 and the pBI-121-GFP empty vector were introduced into onion epidermal cells by a particle bombardment system (Biolistic PDS-1000/He System; Bio-Rad, USA) according to the manufacturer's instructions.
Complementation analysis of the Pht1 genes in yeast
Complementation analysis was carried out as described previously (Daram et al., 1999; Ai et al., 2009). The full detailed procedure is given in Methods S1.
In order to determine the kinetic properties of the Pht1 transporters, Pi uptake experiments using 32Pi were performed with the positive yeast transformants. Approximately 1-mg (fresh yeast) cell samples were used following the previously described method (Ai et al., 2009; Sun et al., 2012).
Overexpression of AsPT1 gene in mycorrhizal roots
The full-length ORF of AsPT1 was cloned into the KpnI and BamHI sites of the plant expression vector pU1301. The resulting vector was called pU1301-AsPT1, and the AsPT1 gene was driven by a maize ubiquitin promoter. The subclones were digested and sequenced to confirm their authenticity. Plant transformation was performed as described previously (Li et al., 2008). The empty pU1301 vector was used as the negative control. The transgenic roots were selected using kanamycin (50 mg ml−1) and the constitutive expression of GUS. First, the roots of Plantlets containing pU1301 vector are supposed to be kanamycin resistant. Second, we also did a GUS staining for each root by sampling their tips. We only chose the positively stained roots for further studies. The full detailed procedure is given in Methods S1.
Analysis of transgenic roots for the expression levels of the AsPT1 transcript in G. intraradices symbiosis was performed as described previously (Dermatsev et al., 2010).
AsPT1 and AsPT4 RNAi constructs
Two independent AsPT1 and AsPT4 RNAi constructs were generated. For the AsPT1 RNAi constructs, the regions corresponding to −120 to +99 and 1660 to 1931 nucleotides (relative to the ATG start codon) of AsPT1 were amplified, using the primers AsPT1Ri5F containing SpeI and KpnI sites (5′-TAACTAGTGGTACCGGGAGGGTTCCCCAAGTGC-3′), and AsPT1Ri5R containing SacI and BamHI sites (5′-TAGAGC TCGGATCCAGTGAAGAAACCCATACCAG C-3′); and AsPT1RiF2 containing SpeI and KpnI sites (5′-TAACTAGT GGTACCTGGAACTTCAATAGTTTAG-3′), and AsPT1RiR2 containing SacI and BamHI sites (5′-TAGAGCTCGGA TCCTGATGGTAGACTA CTAAGG-3′). The fragments were introduced into the pDS1301 vector (Chu et al., 2006) at the KpnI–BamHI and SacI–SpeI sites to create the AsPT1 RNAi-5′ and AsPT1 RNAi-3′ constructs, respectively. Similar methods used for generating the AsPT4 RNAi constructs are shown in Methods S1. All the RNAi constructs confirmed by restriction digestion and sequencing were transferred to A. rhizogenes strain K599 by electroporation, and were transformed into A. sinicus roots. The empty pDS1301 vector was used as the negative control.
Histochemical GUS assays and detection of fungal colonization
Histochemical GUS assays were performed as described previously (Chen et al., 2010). Fresh transgenic roots were incubated in 1 mM X-gluc and 0.1% Triton X-100 (v/v) within 0.05 M sodium phosphate buffer (pH 7.0) at 37°C overnight. To visualize fungal structures, the stained material was counterstained at room temperature overnight with 0.01% acid fuchsin (Floss et al., 2008). Stained material was analysed by microscopy (BX51, Olympus, Japan). Roots were stained with Trypan blue, and mycorrhizal colonization was examined and quantified using the program MYCOCALC according to the method described previously (Trouvelot, 1986).
For the investigation of arbuscule developmental stages, root fractions were stained at 25°C overnight with 0.01% acid fuchsin. The fungal structures within the roots were visualized by confocal laser scanning microscopy (LSM 510 Meta; Zeiss) using the 543 nm laser line for excitation. Arbuscule populations, arbuscule size and infection unit length were measured using Image J software (http://rsb.info.nih.gov/ij/) according to the method described by Zhang et al. (2010).
For enzymatic GUS assays, root tissues were ground in liquid nitrogen and transferred to microtubes containing the extraction buffer for total protein extraction (Chen et al., 2010). Beta-glucuronidase activity was measured fluorimetrically using 1 μg of total protein extract, as described previously (Andriankaja et al., 2007).
Histochemical vital and polyphosphate staining
Histochemical vital and polyphosphate stainings of AM fungus were performed as described previously (Schaffer & Peterson, 1993; Tatsuhiro & Sally, 2001; Ezawa et al., 2004).
Estimation of total P content in plant tissue
Total P content was estimated using a phosphomolybdate colorimetric assay (Shin et al., 2006).
The data were analysed by ANOVA (SPSS 16.0; SPSS Inc., Chicago, IL, USA), followed by Turkey's HSD test (P <0.05), to test differences between plant genotypes and treatments. The data represent the mean ± SD of three independent replicates.
Characterization of the novel mycorrhiza-specific AsPT1 and PHT1 transporters
Six putative Pht1 genes were isolated from mycorrhizal Astralegus sinicus with degenerate primers, designated as AsPT1 through AsPT6 (Fig. 1). Among them, AsPT1 (JQ956415), AsPT3 (JQ956417) and AsPT6 (JQ956420) contain no introns, whereas AsPT2 (JQ956416) and AsPT4 (JQ956418) have single introns, 196 and 699 bp in size, respectively. Interestingly, 5′ end RACE showed that a large (965 bp) intron exists in the 5′ untranslated region (UTR) of AsPT5 (Fig. S1a). Southern blot analysis indicated that these Pi transporters of A. sinicus were single or low-copy genes (Fig. S1b). The deduced amino acid sequences of AsPT1-5 were aligned with the sequences of the Pht1 family from other plants (Fig. S2). Their multiple sequence alignment shows the five PiTs genes have the consensus sequence of 25 amino residue regions, LCFFRFWWLGFGIGGDYP LSATIMSE, including the Pht1 signature, GGDYPLSATIxSE (Karandashov & Bucher, 2005). However, the Pht1 signature was slightly modified in AsPT1; a Thr (T) was exchanged with a Val (V). All six predicted proteins are highly hydrophobic and contain 12 putative membrane-spanning regions with one big intracellular central loop (Fig. S3).
A phylogenetic tree further demonstrated that all of the Pi transporters cloned in this study are the orthologues of six members of the Pht1 family (Fig. 1). These Pht1 family genes can be roughly divided into four subfamilies with clade I (Pht1;1 and Pht1;4), clade II (Pht1;2), clade III (Pht1;3), and clade IV (Pht1;5 and Pht1;6). Interestingly, A. sinicus possesses at least two phylogenetically distant AM-associated Pht1 proteins.
The cis-elements analysis and phylogenetic footprinting revealed that the cis-regulatory element TAAT motif (TAATATAT) was present exclusively in the AsPT1 and AsPT4. In the AsPT1 and AsPT4 promoters, four motifs (CTTC, TGTT, AAAA, TAAT) are conserved, whereas two of them (AAAA, TAAT) was found in the AsPT3 (Fig. S4a). The consensus sequence of MYCS (CTTC motif: TTTCTTGTTCT; Karandashov et al., 2004) was identified exclusively in the promoter regions of AsPT1 and AsPT4 (Fig. S4b). The conserved motif P1BS (GNATATNC) was also found in the promoters of AsPT2, AsPT4 and AsPT5 but not AsPT1 (Fig. S4b). Therefore, the P1BS-dependent pathway might become merged in the regulation of the Pi transporters (Smith et al., 2011). The AsPT1 promoter also contains two motifs: NODCON1GM and OSE1ROOTNODULE (data not shown), which are conserved elements associated with AM-and nodule-induced leghaemoglobin gene regulation (Fehlberg et al., 2005).
Pht1 expression depends on Pi availability and on AM colonization
The expression level of Pht1 gene was quantified in A. sinicus roots colonized by AM fungi (Glomus intraradices or Gigaspora margarita). In the pot system, AsPT1 and AsPT4 were exclusively expressed in mycorrhizal roots (Figs 2, S5), and AsPT3 expression was enhanced in mycorrhizal roots, whereas AsPT2 and AsPT5 displayed lower transcript levels in symbiotic association, compared to the nonmycorrhizal control (Fig. 2d). In AM roots, AsPT1 transcript levels correlated with those of AsPT4 over time. The expression level increased at 14 dpi, and started to decrease from 35 dpi onwards (Fig. 2).
In order to gain insights into the expression response of Pht1 gene to Pi supply status and AM colonization, both RT-PCR (data not shown) and qRT-PCR analyses were used to detect the expression patterns of Pht1 in the pot and in vitro systems. Mycorrhizal colonization of the pot and monoxenic cultures was 95% and 44%, respectively. Results showed that the expression of AsPT1 and AsPT4 was upregulated distinctly under Pi deprivation (Figs 2e, S5). Expression of AsPT1 and AsPT4 is correlated with the degree of colonization in A. sinicus roots inoculated with G. intraradices (Fig. S5b). AsPT2, AsPT3 and AsPT5 genes were enhanced by Pi starvation, and depressed by Pi-sufficient solution (data not shown). Notably, the expression of AsPT2 at Pi concentrations of 65 μM was c. 11.5 times higher than that of 1000 μM Pi at 30 d after treatment (Fig. 2f).
In the in vitro system, the expression levels for AsPT1 under Pi-starved conditions was enhanced 45-fold in the mycorrhizal transgenic roots (Fig. 2g). Similar to AsPT1, the expression of AsPT2 and AsPT3 was depressed by resupply of Pi in the transgenic roots (data not shown). Besides, the expression of AsPT2 under Pi-starved conditions was enhanced 3.5-fold in the transgenic roots at 14 d after treatment (Fig. 2h).
Subcellular and tissue localization of Pht1 transporters
Plant Pht1 members were predicted to be localized to the plasma membrane. To investigate the subcellular localization of the AsPT1 and AsPT4 transporters, we constructed C-terminal protein fusions with GFP placed under the control of the strong 35S promoter. As expected, the two fused proteins were targeted to the plasma membrane (Fig. S6), confirming the potential transport activity of AsPT1 and AsPT4.
For histochemical analysis, the promoters and 5′ untranslated regions of AsPT1 and AsPT4 were fused to a GUS reporter gene. Subsequently, the constructs were transformed into A. sinicus. Four weeks after colonization, GUS activity could be detected only in cells containing arbuscules (Fig. 3). To visualize AM fungal structures, the GUS stained material was counterstained with acid fuchsin staining (Fig. 3e,h).
In order to confirm the correlation between the AsPT1 and AsPT4 expression pattern observed by q RT-PCR and GUS staining, we examined the accumulation of AsPT1 and AsPT4 mRNA in mycorrhizal A. sinicus using in situ hybridization. These results revealed that the AsPT1 and AsPT4 transcripts were restricted to the arbusculated inner cortical cells (Fig. 3j,m). The patterns of labelled signals were in agreement with the PCR and reporter gene expression patterns described above. In addition, AsPT2 transcript was detectable in the root tip and epidermis under Pi-starved conditions (Fig. S7). AsPT3 was localized in root sectors where mycorrhizal structures are formed, and in the cortex cells and epidermis (Fig. S7k). In transgenic plants carrying AsPT5 promoter without intron, the GUS activity was present exclusively in the central cylinder of hairy roots (Fig. S7p). However, lines carrying AsPT5 promoter with intron, GUS expression was present in the central cylinder and root tips (Fig. S7q,r).
In order to identify the putative cis-elements located within the putative promoter regions that are required for mycorrhiza-specific AsPT4 gene expression, a series of expression cassettes with truncated promoter fragments and a GUS reporter gene were generated (Fig. 4a), and then introduced into A. sinicus roots. The transgenic plants were subsequently inoculated with G. intraradices and analysed. A significant decrease in the level of GUS expression was observed following the deletion of Pht1;4 promoter up to −226 bp as compared to that of the full-length promoter, even though the mycorrhization rate was not significantly different (Fig. 4b). The two motifs (TAAT and TAAC) located within this region might mediate AM-activated expression of AsPT4. However, further truncation of the promoter to −194, which includes the first P1BS motif, slight GUS activity reduction was observed. Deletion of pAsPT4 down to −161(AsPT4p-4), resulted in a decrease of GUS activity (Fig. 4a,b). This finding suggests that the P1BS motif is necessary for the AM-inducible activation. Further deletion up to −129, which includes the MYCS and TGTT motifs, led to lower GUS activity in the roots.
In orer to determine whether MYCS is essential for the AM-activated expression of AsPT4, we generated a construct (AsPT4p-M) with knockout of this motif (TTTCTTGTTTC) from the truncation AsPT4p-2. As observed, the deletion of MYCS resulted in a strong decrease of GUS activity in the transgenic lines (Fig. 4c). The data highlight that MYCS is a functional cis-element mediating AM-activated AsPT4 gene expression.
In order to specify the binding sites of the AM-specific responsive cis-elements, the putative protein binding sites of the AsPT4 promoter were analysed by EMSA. The fragments for DNA-protein binding assays are the putative motifs: MYCS, TGTT, TAAC and TAAT. These four fragments showed a gel shift after incubation with protein extract from mycorrhizal roots (Fig. 4d). Remarkably, no shifts occurred in the control experiments. It means corresponding promoter fragments are only recognized and bound by protein factors in mycorrhizal roots.
Functional analysis of the AsPT1 and Pht1 family in yeast
In order to obtain the biochemical evidence for the function of the Pht1 family, yeast mutant MB192 was used for complementation analysis. The AsPT1 and AsPT5 functionally complemented the mutant (Fig. 5a,b). By contrast, yeast cells expressing AsPT2, AsPT3 and AsPT4 were not able to grow on low-Pi medium. In addition, GFP-AsPT1 or GFP-AsPT4 fusions were introduced into the mutant MB192. In agreement with the localization observed in plant cells (Fig. S6), the GFP signal was clearly detected at the periphery of the yeast cell (Fig. 5c), indicating that the full-length AsPT1 and AsPT4 proteins are targeted to the yeast plasma membrane, which is a prerequisite for transporting phosphate.
The pH dependence of Pi transport by AsPT1 was measured over a range of pH values. Growth of the MB192 yeast strain expressing AsPT1 was optimal on acidic pH medium, with a sharp pH optimum at 4.5 (Fig. 5e).
In order to determine the kinetic properties of AsPht1 family, Pi uptake experiments using 32Pi were performed in the transformed yeast mutant. Uptake rates of 32Pi at different Pi concentrations showed that Pi uptake mediated by AsPT1 followed Michaelis–Menten kinetics, exhibiting an apparent mean Km of 31.12 ± 5.87 μM Pi (Fig. 5f). These results indicate that AsPT1 is a high-affinity Pi transporter. We next monitored direct Pi uptake of other Pht1 family transporters, the apparent Km values ranged from 46.75 ± 19.54 μM for AsPT5 (Fig. S8) to 128.7 ± 60.17 μM for AsPT3 (data not shown), showing that the AsPT5 transporter is able to mediate high-affinity Pi acquisition as well.
AM symbiosis in A. sinicus showing development promotion by overexpression of AsPT1
In order to characterize the role of AsPT1 in the development of AM symbiosis in planta, an AsPT1 overexpression assay was performed. Overexpression efficiency of AsPT1 in roots was confirmed by qRT-PCR analyses. Under the Pi-deficient condition, the AsPT1-OE roots showed enhanced amounts of the transcript compared with the controls (Fig. 6f). A hastening growth of 42-d-old AsPT1-OE AM roots under LP (30 μM) supply was observed (Fig. 6b). The total amount of Pi accumulated in the transgenic roots was increased by AsPT1-OE (Fig. 6c). Relative to the controls, AsPT1-OE roots showed an increased density of mycorrhizal structures and abundance of arbuscules (Fig. 6d). Moreover, the majority of arbuscules in AsPT1-OE roots were significantly larger and more mature (Fig. 6a). However, no significant difference in the total colonization was found between control and AsPT1-OE roots. Hence, AsPT1 is able to maintain development of the AM symbiosis in A. sinicus.
In addition, overexpression of AsPT1 also affected expression of other Pht1 members in A. sinicus. The expression levels of AsPT2 and AsPT5 were elevated in mycorrhizal AsPT1-OE transgenic roots, whereas the AsPT6 gene was downregulated upon overexpression of AsPT1 (Fig. 6g).
AsPT1 and AsPT4 are indispensable for AM development in A. sinicus
In order to further examine the function of AsPT1 and AsPT4 in AM symbiosis, AsPT1 and AsPT4 RNAi constructs were generated, to investigate their effects on the development of the AM symbiosis. Knockdown efficiency of AsPT1 and AsPT4 was confirmed by qRT-PCR analyses. In AM roots, the transcript levels of AsPT1 in the RNAi lines were c. 6% and 23% of the controls, respectively (Fig. 7a). Similarly, AsPT4 was effectively silenced in AM roots (Fig. 7b). Relative to the controls, AsPT1 RNAi roots showed a reduced extent of arbuscule formation and overall fungal colonization (Fig. 8b). Similarly, total colonization and arbuscules rate were significantly reduced also in AsPT4 RNAi roots (Fig. 8c). Hence, AsPT1 and AsPT4 are indispensable but not functionally redundant for normal development of the AM symbiosis in A. sinicus.
In order to examine the symbiotic phenotype at the molecular level, we analysed the expression of other Pht1 members in AsPT1 or AsPT4 RNAi lines. Interestingly, the AsPT4 mRNA levels in the AsPT1 RNAi lines remained as high as in the controls (Fig. 7c), despite the reduced total colonization and arbuscule formation, indicating compensatory effects between the two transporters. Transcript levels of AsPT1 in AsPT4 RNAi lines were significantly reduced relative to the controls (Fig. 7d). AsPT2 was also downregulated upon suppression of AsPT1 or AsPT4; however, AsPT3 showed unaltered transcript levels in all the RNAi lines. In AsPT1 RNAi lines, no reduction of AsPT5 mRNA was observed, by contrast, AsPT5 mRNA in AsPT4 RNAi lines was reduced relative to the controls. Transcript levels of AsPT6 were upregulated in the AsPT1 and AsPT4 RNAi roots (Fig. 7c,d).
Knockdown of AsPT1 or AsPT4 by RNAi results in a premature death of the arbuscule phenotype
AsPT1 and AsPT4 RNAi lines were analysed to determine the morphology of arbuscule. Transgenic roots expressing RNAi construct designed to target AsPT4 showed a mycorrhizal phenotype identical to that of AsPT1 RNAi lines. The fungus was able to penetrate the roots, but arbuscule development was impaired, degenerating and premature death arbuscules were visible in the cortical cells (Fig. 8a). To evaluate the proportion of the various stages of arbuscule development, three stages of arbuscule development and decay were defined. As deduced from a total of c. 900 evaluated arbuscules, a shift towards degenerating and dead arbuscules at the expense of mature ones occurred in AsPT1 or AsPT4 RNAi roots relative to the controls (Fig. 8d). Furthermore, lengths of arbuscules and infection units in AsPT1 RNAi lines were significantly shorter than those in the controls (Fig. 8e). Septate hyphae were observed in small arbuscules (Fig. 8a), indicating arbuscule collapse and death in AsPT1 or AsPT4 RNAi roots. The vital staining revealed that the incidences of live arbuscules were low in AsPT1 or AsPT4 RNAi roots (Fig. 8f). Thus, AsPT1 and AsPT4 are not only indispensable for intraradical fungal development, but also essential for arbuscule development in AM symbiosis.
AsPT4 but not AsPT1 is required for Pi transfer in AM symbiosis
In order to evaluate the contribution of AsPT1 and AsPT4 to symbiotic Pi uptake, Pi uptake measurements were performed. Pi concentration was determined in shoot tissue of AsPT1 and AsPT4 RNAi lines when grown under low Pi supply in the presence or absence of G. intraradices. The shoot Pi content and shoot mass of NM plants were similar in all transgenic lines, indicating that silence of AsPT1 or AsPT4 function had no effect on Pi nutrition in the absence of symbiosis (Fig. 9a,b). However, a significant increase in shoot Pi concentrations and shoot mass associated with G. intraradices treatment was apparent in the controls (Fig. 9). As observed in M. truncatula PT4 mutants (Javot et al., 2007a), AsPT4 RNAi plants inoculated with G. intraradices did not differ significantly from NM plants, suggesting that AsPT4 is essential for the symbiotic Pi uptake. By contrast, the AsPT1 RNAi plants displayed Pi concentration in colonized plants as high as in the controls (Fig. 9a).
As an alternative, we used polyP staining to confirm whether loss of AsPT1 or AsPT4 function affected the symbiotic Pi transfer in the AM roots. The polyP granules accumulated in intraradical fungal vesicles and arbuscules within the AsPT4 RNAi roots but not in wild-type, vector controls or AsPT1 RNAi lines (Figs 9c, S9). The polyP staining patterns are consistent with the shoot P data (Fig. 9a,b) and suggest that loss of AsPT4 function results in a block in symbiotic Pi transfer from the fungus to the plant. In summary, these results showed that AsPT4 is required for symbiotic Pi transfer in AM symbiosis and AsPT1 do not compensate for the loss of AsPT4.
AsPT4 paralogue AsPT1 is a novel AM-specific Pht1 phosphate transporter
MtPT4, OsPT11 and StPT4 probably derived from the same ancient ancestor protein that existed long before the divergence of monocot and dicot lineages. MtPT4 orthologue AsPT4 is an AM-specific Pi transporter with expression in the presence of symbiosis, supporting the evolutionary conservation of proteins involved in symbiotic Pi transport. Evolutionarily more recent, the legume species obtained an additional mycorrhiza-specific Pi transporter, AsPT1, which is phylogenetically distant from AsPT4. This finding indicates that AsPT1 should play a different role in AM symbiosis. In addition, the AsPT1 homologues in di- and monocotyledons currently form a novel gene subfamily attribute to gene duplication. GmPT7 from Glycine max, for instance, is induced in the later stages of symbiosis (Tamura et al., 2012); and the homologous PtPT8 is phylogenetically associated to the AM-inducible Pht1 subfamily (Loth-Pereda et al., 2011). Indeed, BdPT12 and BdPT13 are induced during AM symbiosis (Hong et al., 2012), similar to OsPT13 (Yang et al., 2012).
Functional diversity of the Pht1 family in arbuscular mycorrhizal Astralegus sinicus
Spatial expression patterns for the Pht1 family were performed by cellular localization analysis. The AsPT1 and AsPT4 were detected exclusively in arbuscule-containing cells of the cortex (Fig. 3). The AsPT2 transcript is localized to the root tip and epidermis during Pi starvation (Fig. S7). The expression pattern is similar to LePT1 and LePT2 (Liu et al., 1998b; Bucher et al., 2001), suggesting that the AsPT2 transporter plays a significant role in phosphate acquisition at the root–soil interface from soil. AsPT3 occurred in mycorrhizal roots and root cortex cells. AsPT5 sharing similarity with OsPT2 (Ai et al., 2009), is expressed in the central cylinder and root tip (Fig. S7). It is reasonable to conclude that AsPT5 probably functions in the transport of Pi from roots to shoots. Thus, the distinct expression patterns of the Pht1 members of A. sinicus in response to the arbuscular mycorrhizas, Pi starvation, and tissues indicated functional diversification.
Pht1 proteins share similarity with the proton-coupled Pi symporters from yeast and the mycorrhizal fungi (Fig. 1). Apparent Km values of Pht1 transporters for Pi have been calculated from many such proteins via their expression in yeast mutants (Rausch et al., 2001; Harrison et al., 2002; Ai et al., 2009; Jia et al., 2011; Remy et al., 2012). Complementation of the yeast Δpho84 mutant's growth defect at micromolar Pi concentrations, supported by a Pi uptake assay, demonstrates that A. sinicus Pht1;5 protein is indeed capable of mediating high-affinity Pi transport. The apparent Km value for Pi uptake of 46.75 μM determined here for Pht1;5. The Km values calculated for mycorrhiza-specific Pi transporters were 64 μM of StPT3 (Rausch et al., 2001) and 493/668 μM for MtPT4 using two different yeast mutants (Harrison et al., 2002), respectively. Importantly, our detailed funtional analysis of AsPT1 and AsPT4 in yeast, indicated that AsPT1 encoded a functional high-affinity transporter, whereas AsPT4 as its orthologue MtPT4 was s low-affinity transporter. According to a previous hypothesis (Nagy et al., 2005), the perifungal Pi concentration in roots can vary substantially, probably depending on the physiology of the symbiotic fungus. The coordinated activity of the symbiotic Pi transporters would ensure efficient uptake of fungal Pi (Smith et al., 2003). In addition, the differences between complementation results and uptake measurements are due to the Δpho84 mutant that possesses other high- or low- affinity Pi transporters.
AsPT1 and AsPT4 are required for AM symbiosis
It is interesting that dicot species have at least two AMF-inducible phosphate transporter genes clustering separately with two Pht1 subfamilies. Knock out of LePT4 in tomato indicated a high degree of functional redundancy between the two mycorrhiza-specific phosphate transporters, LePT4 and LePT5 (Nagy et al., 2005). It would be useful to know whether the AsPT1 and AsPT4 also have a similar genetic redundancy of a Pi transporter as observed in tomato. If not, the distinct functions between AsPT1 and AsPT4 represent a novel subfamily of AM-specific PHT1 proteins in plants. A detailed phenotypic analysis revealed that AsPT1 or AsPT4 is indispensable for the formation of AM symbioses. AsPT1 or AsPT4 suppression showed decreased levels of total colonization and arbuscule development (Fig. 8), and led to premature death arbuscules. Silencing of AsPT4 by RNAi resulted in a phenotype identical to the MtPT4 mutants, suggesting that Pi delivery to the cortical cell by AsPT4 might serve as a signal to permit continued development of the arbuscule and consequently to sustain proliferation of the fungus (Javot et al., 2007a,b).
Symbiotic phosphate transport in AM A. sinicus was mediated by the specific phosphate transporter AsPT4 but not AsPT1. Loss of AsPT4 function resulted in a block in symbiotic Pi uptake, and knockdown of AsPT1 also affected arbuscule morphology and growth of arbuscules of G. intraradices, surprisely, did not alter Pi transfer in AM symbiosis, indicating compensatory effects between the two transporters. The AsPT1 do not compensate for the loss of the AsPT4, thereby showing that the functional establishment of the AM symbiosis enhanced the Pi status of AsPT1 RNAi plants, but not of AsPT4 RNAi genotypes. As observed in A. sinicus, in rice, mutations in the Pht1;11 gene caused a decrease in symbiotic Pi uptake capacity of the mycorrhizal root that resulted in lower Pi content of shoot tissues, whereas pht1;13 mutant plants did not affect this process (Yang et al., 2012). The combined phenotypic and cellular localization analysis indicated that AsPT1 and AsPT4 play an indispensable role during the development of intraradical fungal structures, most prominently the arbuscules. It became apparent that AsPT1 and AsPT4 are not functionally redundant and no additive effect in AM symbiosis of A. sinicus.
AsPT1, probably encoding a transceptor, mediates Pi concentrations at the periabuscular interface
The most intriguing study on the function of AM-specific AsPT1 originates from its distinctions from AsPT4. Although AsPT1 encodes a functional plasma membrane-localized transporter that mediates high-affinity Pi activity in yeast, and AsPT1 is required for arbuscule development in A. sinicus, AsPT1 is not necessary and sufficient to mediate the symbiotic Pi uptake. A higher level of AsPT1 expression promoted the intraradical fungal development of G. intraradices and, interestingly, altered Pi transfer in mycorrhizal roots (Fig. 6). The results suggest that AsPT1 may moderately contribute to symbiotic Pi uptake in A. sinicus. So far, the detail function of AsPT1 protein at the periabuscular interface is still unclear. It is probable that AsPT1 plays an important role in symbiotic signalling. Currently, the discovery of nutrient transceptors, transporter-like proteins with a receptor function, suggests that receptors for environmental signals may have been derived in evolution from nutrient transporters (Thevelein & Voordeckers, 2009). In rice, a symbiosis-specific OsPT13 protein was suggested to have a nontransporting transceptor role in symbiotic signalling affecting arbuscule development (Yang et al., 2012). Therefore, AsPT1, a probable nutrient sensor protein with high transport activity, might have been discovered in A. sinicus. The synergism of AsPT1 and AsPT4 might permit continued development of the arbuscule and effective fungal symbiont. Altogether, the phenotypic analysis and evidence led us to assume that AsPT1 probably encodes a functional transcepter involving Pi concentrations at the periabuscular interface.
Taken together, there also exists a second Pi-transporter AsPT1 indispensable for the development of AM symbiosis in dicots. AsPT4 but not AsPT1 is necessary and sufficient to mediate symbiotic Pi transfer. The conserved cis-acting elements mediate the mycorrhiza- activated AsPT1 and AsPT4. Future studies should further address the isolation and subsequent characterization of the trans-acting factors binding to cis-acting elements, and explore their potential interaction with other transcription factors. This will help to strengthen our understanding of the regulation during the establishment of the AM symbiosis. The data provide new insights into plant proteins required for AM symbiosis, phosphate transport and pave the way for the identification of novel regulation factors essential for AM symbiosis.
This work was supported by the National Natural Science Foundation of China (30870082, 30800001). We are very grateful to Dr Guohua Xu for kindly providing the yeast mutant stain MB192 and the expression vector 112A1NE, and for his support in the experiments with the complementation analysis. We would like to thank Professor Shiping Wang for providing the pDS1301 vector for the RNAi experiment, Professor Zhongming Zhang for providing strain K599 for the plant transformations, and Professor Hong Long and student Xiaoning Fan for supporting microscopic studies. We would also like to thank Dr Christina Baker and Dr Ton for revising the manuscript.