Plant growth and development are strongly influenced by the availability of nutrients in the soil solution. Among them, phosphorus (P) is one of the most essential and most limiting macro-elements for plants. In the environment, plants are often confronted with P starvation as a result of extremely low concentrations of soluble inorganic phosphate (Pi) in the soil. To cope with these conditions, plants have developed a wide spectrum of mechanisms aimed at increasing P use efficiency. At the molecular level, recent studies have shown that several proteins carrying the SPX domain are essential for maintaining Pi homeostasis in plants. The SPX domain is found in numerous eukaryotic proteins, including several proteins from the yeast PHO regulon, involved in maintaining Pi homeostasis. In plants, proteins harboring the SPX domain are classified into four families based on the presence of additional domains in their structure, namely the SPX, SPX-EXS, SPX-MFS and SPX-RING families. In this review, we highlight the recent findings regarding the key roles of the proteins containing the SPX domain in phosphate signaling, as well as providing further research directions in order to improve our knowledge on P nutrition in plants, thus enabling the generation of plants with better P use efficiency.
Involvement of the SPX domain in inorganic phosphate (Pi) signaling
It is well established that plant growth and development are highly dependent on the nutrient availability in soil. Inorganic phosphate (Pi), the main source of phosphorus (P) for plants, is present in soluble form at very low concentrations in most soils, as it is often bound to organic and inorganic compounds, thus creating insoluble complexes (Poirier & Bucher, 2002). As a result, Pi deficiency has become a major problem in many agricultural ecosystems, limiting plant growth and yield. For a long time, the application of fertilizer was chosen to overcome these problems. However, the side effects associated with heavy fertilization, such as the eutrophication of lakes, concomitant with the expected phosphate rock shortage in the coming decades, indicate that, in the long term, this solution is neither economically nor ecologically sustainable. Therefore, several scientific programs aimed at improving nutrient use efficiency have been carried out in recent years in order to maximize plant growth on soils with low nutrient availability. To date, our knowledge of the molecular response of plants to Pi starvation has greatly improved with the identification of several key players involved in Pi signaling, and has been well reviewed in recent years (Nilsson et al., 2010; Rouached et al., 2010; Chiou & Lin, 2011; Peret et al., 2011). Among the many and diverse proteins involved in the plant response to Pi starvation, proteins containing the SPX domain are key players controlling a set of processes involved in the maintenance of an internal steady state of phosphate ions at the level of the cell, defined as Pi homeostasis (Hamburger et al., 2002; Duan et al., 2008; C. Wang et al., 2009; Lin et al., 2010; Secco et al., 2010; Kant et al., 2011). The SPX domain (Pfam PF03105) (Fig. 1a) is named after the Suppressor of Yeast gpa1 (Syg1), the yeast Phosphatase 81 (Pho81) and the human Xenotropic and Polytropic Retrovirus receptor 1 (Xpr1). This hydrophilic domain is found at the N-termini of various proteins in all major eukaryotes, from Caenorhabditis elegans and Drosophila to mammals (Stefanovic et al., 2011). Although the organization of the SPX domain is variable, it can be subdivided into three well-conserved sub-domains of 30–40 amino acids each (Fig. 1a–c). In addition, despite having an average length of ∼165 amino acids, the SPX domain is found in stretches of 135–380 amino acids, with the three sub-domains being separated from each other by regions of low similarity. In yeast, the working model for nutrient homeostasis in eukaryotes, several proteins belonging to the PHO regulon, involved in the maintenance of Pi homeostasis, possess the SPX domain. Despite the lack of such a regulon in plants, many plant proteins harboring the SPX domain have been shown to be involved in Pi signaling (Table 1; Figs 2, 3). Recent studies in yeast and Arabidopsis have also suggested that the SPX domain itself could be involved in the fine tuning of Pi transport and signaling through mechanisms such as physical interactions with other proteins (Duan et al., 2008; Hurlimann et al., 2009; Zhou & Ni, 2010). In most plants, proteins harboring the SPX domain can be divided into four groups depending on the presence of extra domains: proteins that exclusively contain the SPX domain (Class 1), and proteins that, in addition, harbor at the C-terminus an EXS domain (Class 2), an MFS domain (Class 3) or a RING domain (Class 4) (Fig. 2; Table 1) (Chiou & Lin, 2011). Here, we summarize and highlight recent advances and future challenges in understanding the important roles of the different SPX domain-containing protein families in the regulation of phosphate homeostasis in yeast, Arabidopsis and rice.
Table 1. List of the Arabidopsis and rice SPX domain-containing proteins and their characteristics
Induction by Pi starvation
The column ‘Induction by Pi starvation’ represents transcript regulation under Pi starvation.
‘+’, ‘−’, ‘=’, induction, suppression or unaltered transcript expression, respectively.
OsSPX-MFS4a shows that OsSPX-MFS4 is considered as a pseudogene.
In addition to mediating phosphate uptake, the three low-affinity transporters are involved in the sensing of external Pi and the regulation of Pi signaling (Ghillebert et al., 2011). Although Pho87 and Pho90 are localized to the plasma membrane, Pho91 is localized to the vacuolar membrane and is involved in the export of Pi from the vacuole to the cytosol (Hurlimann et al., 2007). Under low-Pi conditions, Pho87 and Pho90 are repressed by Spl2, a negative regulator, which interacts directly with both transporters via the SPX domain. Therefore, the SPX domain of Pho87 and Pho90 can act as an auto-inhibitory domain involved in the regulation of phosphate accumulation in yeast cells (Hurlimann et al., 2009). In addition, on phosphate, nitrogen and carbon source deficiencies, it has been shown that Pho87 and Pho90 can be inactivated via vacuolar targeting, and that this mechanism requires the SPX domain (Ghillebert et al., 2011).
On Pi starvation, Pho81, in association with inositol heptakisphosphate, inactivates the main regulator of the PHO pathway, thus resulting in increased levels of expression of the phosphate starvation genes (Lee et al., 2007).
The vacuolar localized Vtc proteins have recently been shown to synthesize polyphosphate, using ATP as a substrate, before translocating the phosphate polymers to the vacuolar lumen (Hothorn et al., 2009). A truncated form of Vtc4, devoid of the SPX domain, is still active in polyphosphate synthesis, indicating that the SPX domain is not essential for catalytic activity (Hothorn et al., 2009).
Gde1, encoding the only characterized glycerophosphodiester phosphodiesterase in yeast, is responsible for the hydrolysis of glycerophosphocholine in the cell, and thus the scavenging of phosphate from glycerophosphodiesters under low-Pi conditions (Fisher et al., 2005).
The SPX family in plants
In plants, proteins exclusively harboring the SPX domain are referred to as SPX proteins. In Arabidopsis and rice, the SPX family consists of four and six members, respectively (Fig. 2; Table 1) (Duan et al., 2008; C. Wang et al., 2009; Z. Wang et al., 2009). These relatively small proteins (∼280 amino acids) were named as AtSPX1–AtSPX4 in Arabidopsis and OsSPX1–OsSPX6 in rice. Localization studies revealed a broad range of expression for the members of the SPX family, ranging from roots, leaves, cotyledons, stems and pollen grains. Transcript and histochemical analyses showed that all the SPX genes, with the exception of AtSPX4 and OsSPX4, were highly induced on Pi starvation in roots and/or in shoots (Supporting Information Fig. S1) (Duan et al., 2008; C. Wang et al., 2009; Z. Wang et al., 2009). In addition, studies in Arabidopsis showed that these responses were under the control of AtPHR1 and its closest family member AtPHL1 (Fig. S1).
Individual mutant analysis of AtSPX1, AtSPX2 and AtSPX4 knock-outs did not show any obvious phenotypes, under either Pi-sufficient or Pi-deficient conditions (Duan et al., 2008). However, the overexpression of AtSPX1 increased the expression levels of some of the Phosphate Starvation Inducible (PSI) genes, such as ACP5, PAP2 and RNS1, independent of Pi status, suggesting a potential transcriptional regulation role of AtSPX1 on Pi starvation (Fig. 3). In addition, repression of AtSPX3 by RNAi altered the response to Pi starvation at both the phenotypic and gene expression level, resulting in plants with increased Pi concentration in the shoots, and decreased Pi concentration in the roots, suggesting that AtSPX3 can act as a negative regulator of Pi starvation signaling (Duan et al., 2008).
In rice, OsSPX1 has been shown to be specifically induced by Pi starvation and to be preferentially expressed in the roots (C. Wang et al., 2009). Suppression of OsSPX1 by RNAi reduced plant growth and increased Pi accumulation in the shoots, as observed for well-characterized plants overaccumulating Pi, such as Ospho2 mutants or OsPHR2-overexpressing plants (C. Wang et al., 2009). Detailed analysis showed that the increase in shoot Pi concentration in the OsSPX1 RNAi lines correlated with increased expression of some PSI genes, such as the Pi transporters OsPT2 and OsPT8 (C. Wang et al., 2009). By contrast, although growth was still impaired, overexpression of OsSPX1 suppressed the induction of the PSI genes, suggesting that OsSPX1, similar to AtSPX3, is involved in a negative feedback loop to adjust the expression of several PSI genes under Pi-limited conditions (C. Wang et al., 2009). Recently, Liu et al. (2010), using plants simultaneously overexpressing OsSPX1 and OsPHR2, demonstrated that OsSPX1 could counteract the function of OsPHR2 in inducing the expression of OsPT2, which plays a major role in Pi translocation and accumulation, thus demonstrating that OsSPX1 acts as a negative regulator of OsPHR2 (Fig. 3) (Liu et al., 2010).
The SPX proteins have a broad range of subcellular localization. AtSPX1, AtSPX2, OsSPX1 and OsSPX2 are exclusively localized to the nucleus, AtSPX3 and OsSPX4 are localized to some unidentified cytoplasmic speckles, and AtSPX4 and OsSPX4 are localized to the plasma membrane (Duan et al., 2008; Z. Wang et al., 2009). Yet, the subcellular localization of the rice SPX proteins was only monitored in a heterologous system, using onion epidermal cells (Z. Wang et al., 2009). Recent studies have shown that OsSPX1 could regulate the transcription of OsSPX2, 3 and 5 (Z. Wang et al., 2009). However, in this study, the authors did not rule out the possibility that the changes observed in gene expression for OsSPX2, 3 and 5 could be a result of increased Pi concentration in the OsSPX1-overexpressing plants, instead of direct regulation via OsSPX1.
Another feature of the rice SPX family has been demonstrated in the response to cold stress. Constitutive overexpression of OsSPX1 in tobacco plants resulted in decreased total leaf Pi concentration and the accumulation of free proline and sucrose, providing improved cold tolerance compared with the wild-type (WT) (Zhao et al., 2009). It is noteworthy that both cold stress and Pi starvation induce sugar accumulation in plants. A link between low Pi, cold acclimatization and freezing tolerance has also been noted using the Arabidopsis pho1 and pho2 mutants (Hurry et al., 2000). The importance of sugars and, more specifically, sucrose in Pi homeostasis is well documented (Hammond & White, 2011; Lei et al., 2011). Therefore, the deciphering of the cross-talk between phosphate starvation and sugar signaling, as well as the involvement of SPX domain-containing proteins in these pathways, requires further investigation.
The SPX-EXS family in plants
The PHO1 family members are the only proteins in eukaryotes that contain both the SPX and EXS domains (Fig. 2) (Wang et al., 2004). The EXS domain is embedded in a hydrophobic region and has an unknown function. It was named after the yeast ERD1, involved in the localization of endogenous endoplasmic reticulum proteins, the human XPR1 and the yeast SYG1 (Pfam entry PF03124). Unlike other SPX-containing protein families, the SPX domain in the PHO1 gene family is found as a tripartite domain, interspersed with large insertions. Homologs of PHO1 are found in a wide spectrum of eukaryotes, from yeast, C. elegans, Drosophila, mammals and plants (Secco et al., 2010). Interestingly, no PHO1 homologs are found in prokaryotes or in the unicellular alga Chlamydomonas reinhardtii, but PHO1 homologs are found in bryophytes, monocotyledons and dicotyledons (Wang et al., 2008; Secco et al., 2010).
The pho1 mutant was first isolated in Arabidopsis as a Pi-deficient mutant with low leaf Pi concentration, reduced shoot growth and increased Pi concentration in the roots (Table 1) (Poirier et al., 1991). Consistent with its function in Pi loading into the xylem, AtPHO1 is expressed in the cells of the root vascular system and the lower part of the hypocotyl (Hamburger et al., 2002). Recent studies have demonstrated that AtPHO1 is capable of mediating phosphate efflux out of the cell, rendering it the first identified protein in plants and animals involved in Pi export (Stefanovic et al., 2011). It has also been shown that AtPHO1 plays a key role in the long-distance Pi-deficiency signaling network. Indeed, reducing the level of AtPHO1 transcripts in Arabidopsis results in a decreased rate of Pi transfer from the root to the shoot, consequently leading to Pi-deficient shoots. Surprisingly, all the usual hallmarks associated with Pi deficiency, such as the induction of PSI genes and reduced growth, are absent in AtPHO1-underexpressing lines, demonstrating a clear role of PHO1 in Pi signaling (Rouached et al., 2011). In Arabidopsis, the PHO1 family consists of 10 additional members (AtPHO1;H1–AtPHO1;H10) with high homology to AtPHO1 (Fig. 2; Table 1). Although the majority of these genes are predominantly expressed in the vascular cylinder of the root and/or the shoot, some genes are more broadly expressed, such as in the trichomes, pollen grains and in response to hormones, implying a broader role for AtPHO1 proteins than simply the long-distance transfer of Pi (Wang et al., 2004; Rouached et al., 2010). To date, four members of the family, AtPHO1, AtPHO1;H1, AtPHO1;H4 and AtPHO1;H10, have been studied and characterized, with only AtPHO1 and AtPHO1;H1 being involved in Pi homeostasis. Genetic complementation studies aimed at rescuing the growth defect phenotype as well as the low shoot Pi concentration of the Atpho1 mutant, with all the members of the AtPHO1 gene family, with the exception of AtPHO1;H10, showed that only AtPHO1 and AtPHO1;H1 were involved in long-distance Pi transfer (Stefanovic et al., 2007). Despite having a similar biological role, and being induced by Pi starvation, AtPHO1 and AtPHO1;H1 are differentially regulated (Fig. 3; Table 1). Although the induction of AtPHO1;H1 on Pi starvation is controlled by AtPHR1, the main transcription factor controlling Pi homeostasis, and can be suppressed by the nonmetabolizable phosphate analog, phosphite, the increase in AtPHO1 expression is independent of AtPHR1 and is not influenced by phosphite (Stefanovic et al., 2007). Analysis of the expression profile of AtPHO1 and AtPHO1;H1 on sucrose and phytohormone treatments revealed further differences in their mode of regulation (Ribot et al., 2008). Recent studies have demonstrated that AtPHO1 is regulated by two WRKY transcription factors, WRKY6 and WRKY 42 (Chen et al., 2009). Both transcription factors negatively regulate the expression of PHO1 by binding to two W-box motifs present in the AtPHO1 promoter, in a Pi-dependent manner (Chen et al., 2009). Taken together, it appears that AtPHO1 and AtPHO1;H1 are regulated by distinct signal transduction pathways.
AtPHO1;H4, also named Short Hypocotyl under Blue Light (SHB1), and AtPHO1;H10 have, to date, no direct links with phosphate homeostasis, being involved in the control of hypocotyl elongation under blue light and in response to numerous biotic and abiotic stresses, respectively (Table 1) (Ribot et al., 2008; Zhou & Ni, 2010).
In rice, the PHO1 family consists of only three genes, namely OsPHO1;1–OsPHO1;3. Interestingly, all the rice PHO1 proteins clustered with AtPHO1 and AtPHO1;H1, the only two Arabidopsis members involved in long-distance Pi transfer (Secco et al., 2010). Mutant analysis revealed that OsPHO1;2, the closest homolog of AtPHO1, was required to transfer Pi from the roots to the shoots (Fig. 3). A key difference between the Arabidopsis and rice PHO1 families is the presence of cis-Natural Antisense Transcripts (NATs) for all three members of the rice PHO1 family (Secco et al., 2010). Despite the unknown function of these NATs, their expression pattern, as well as data from the literature, suggest that they could be implicated in regulating the expression of the sense transcript. The function of the other members of the family is still unknown, but could be involved in the maintenance of Pi homeostasis in other tissues, such as flowers, as suggested by their expression profiles, or be functionally redundant with OsPHO1;2. Moreover, phylogenetic analyses of PHO1 homologs of different mono- and dicotyledonous plants revealed the emergence of a divergent clade of PHO1 proteins in dicotyledons, which include members that have not yet been involved in Pi homeostasis, such as AtPHO1;H4 (SHB1) (Secco et al., 2010). However, the functionality of this PHO1 dicotyledon-specific clade is still unclear.
The SPX-MFS family in plants
The Major Facilitator Superfamily (MFS) represents the largest group of transport carriers in all organisms, which are often coupled to the movement of another ion. Proteins of this family can function as uniporters, symporters or antiporters, and have a diverse range of substrates, such as ions, sugars, nucleosides, amino acids and peptides. Based on the properties of the SPX and MFS domains, it has been hypothesized that proteins harboring these two domains could be involved in both transport and signaling (Lin et al., 2010).
In rice, although there are four putative genes for this family, namely SPX-MFS1–SPX-MFS4, the latter has no reported full-length cDNA or expressed sequence tag (EST) sequence, suggesting that it may be a pseudogene (Fig. 2; Table 1). Transcript analysis of the rice SPX-MFS genes showed that they were preferentially expressed in the shoots, and that both OsSPX-MFS1 and OsSPX-MFS3 were suppressed by Pi starvation, whereas OsSPX-MFS2 was induced by Pi deficiency (Fig. 3) (Lin et al., 2010). OsSPX-MFS1 and OsSPX-MFS2 have been shown to be specifically regulated by a Pi starvation-induced microRNA, osa-miR827 (Lin et al., 2010). In situ hybridization revealed that OsSPX-MFS1 and OsSPX-MFS2 were preferentially expressed in the leaf mesophyll and parenchyma cells surrounding the xylem, similar to osa-miR827 (Lin et al., 2010). Analysis of the knock-out T-DNA lines of OsSPX-MFS1 and OsSPX-MFS2 and overexpressing plants of osa-miR827 revealed that both OsSPX-MFS1 and OsSPX-MFS2 were negatively regulated by osa-miR827 abundance, despite their different responses to external Pi status. Although Lin et al. (2010) did not find any obvious phenotype in osa-miR827-suppressing or -overexpressing lines, a recent study has shown that overexpression of osa-miR827 or reduced expression of OsSPX-MFS1 increases Pi concentration in the leaves, and reduces Pi re-mobilization from old to young leaves (C. Wang et al., unpublished). Moreover, using OsPHR2-overexpressing plants, it has also been shown that the osa-miR827/OsSPX-MFS1/2 pathway is under the control of OsPHR2 (Lin et al., 2010).
Interestingly, in most of the monocotyledons, the preferential cell type for leaf Pi storage is the mesophyll cell, in contrast with Arabidopsis, where Pi is mainly stored in the epidermal and bundle sheath cells (Conn & Gilliham, 2010; Conn et al., 2011). The compartmentalization of nutrient storage in specific cell types aims at reducing the creation of insoluble complexes. Thus, in rice, specific phosphate transporters should be required to concentrate Pi in mesophyll cells. Hence, it is tempting to hypothesize that some members of the SPX-MFS family, as a result of their localization and function, could perform such a function.
The SPX-RING family in plants
The Really Interesting New Gene (RING) finger domain, a specialized type of zinc finger domain, is involved in the mediation of protein–protein interactions. The presence of a RING finger domain is a characteristic of RING-class E3 ubiquitin protein ligases, which are capable of transferring ubiquitin from an E2 enzyme to a substrate protein. In Arabidopsis, despite the RING domain being present in more than 450 proteins, only two proteins in both rice and Arabidopsis possess the RING and SPX domains (Fig. 2; Table 1).
To date, the only characterized member of the SPX-RING family is the Arabidopsis Nitrogen Limitation Adaptation (NLA) gene (Peng et al., 2007), also called benzoic acid hypersensitive 1 (BAH1), for its role in the immune response (Yaeno & Iba, 2008). The Atnla mutant was first identified for its altered growth response on nitrogen (N) starvation, being unable to accumulate anthocyanins, resulting in an early senescence phenotype (Peng et al., 2007, 2008). A recent study has demonstrated the involvement of AtNLA in phosphate homeostasis (Kant et al., 2011). Phosphate analysis revealed that the Atnla mutant showed increased Pi uptake and Pi content, especially under low-nitrate and high-phosphate availability, relative to WT plants. The phosphate uptake capacity and Pi content of the Atnla mutant were similar to those of the well-characterized Pi overaccumulator Atpho2 mutant. The early senescence phenotype detected in the Atnla mutant under low nitrate appeared to be a consequence of shoot Pi toxicity, as observed in Atpho2 mutants. Phosphate overaccumulation was drastically increased under low-nitrate conditions, for both Atnla and Atpho2 mutants, suggesting that nitrate and phosphate levels have an antagonistic interaction (Kant et al., 2011). The Arabidopsis and rice PHO2, encoding an E2 conjugase, are modulated by the Pi starvation-induced miR399 (Rouached et al., 2010; Chiou & Lin, 2011). PHO2 is a key player in Pi homeostasis, regulating a subset of the PSI genes, such as some Pi transporters. Surprisingly, AtNLA has also been shown to be regulated by an miRNA, miR827. Under Pi starvation, miR827, similar to miR399, is specifically up-regulated and targets the degradation of NLA mRNA, thus activating Pi uptake and root to shoot translocation (Fig. 3). Consequently, both AtPHO2 and AtNLA act as negative regulators of Pi uptake, and are regulated by miRNAs, in order to avoid Pi overaccumulation and thus leaf Pi toxicity. Moreover, AtPHO2, an E2 conjugase, and AtNLA, an E3 ligase, are both part of the ubiquitination pathway, targeting proteins for degradation via the ubiquitin-26S proteasome, and could thus interact together. A yeast two-hybrid screen has already demonstrated that AtNLA could interact with the ubiquitin E2 conjugase UBC8 via the RING domain (Peng et al., 2007).
Additional evidence of the involvement of the Arabidopsis AtNLA in Pi homeostasis was provided by the identification of two suppressors of the Atnla mutation recovering the WT phenotype, namely the Phosphate Transporter Traffic Facilitator1 (AtPHF1) and Phosphate Transporter 1.1 (AtPHT1;1) (Kant et al., 2011). In addition, AtPHF1 and AtPHT1;1 are probably direct or indirect targets of both AtNLA and AtPHO2, as mutation of Atphf1 and Atpht1;1 in the Atnla and Atpho2 mutant backgrounds restored the Pi concentration to WT levels. Thus, in the future, it will be interesting to gain further details of the cross-talk between AtNLA and AtPHO2, and their role in controlling nitrate-dependent phosphate homeostasis.
Conclusion and perspectives
In recent years, the role and importance of the SPX domain-containing proteins in the regulation of Pi homeostasis have become increasingly clear. The recent discovery of the involvement of NLA and some members of the SPX-MSF family in the control of Pi homeostasis shows that all families of proteins harboring the SPX domain have some members involved in Pi signaling and/or transport in plants. The function of these families is well conserved between Arabidopsis and rice, with the exception of the SPX-RING and SPX-MFS families that ‘overlap’. In the future, a key challenge to improve our knowledge on the function of the plant SPX domain-containing protein, and thus Pi homeostasis, will be to determine whether the SPX domain is involved in protein interaction, as observed in yeast, and, if so, to identify the different molecular players. Indeed, several SPX domain-containing proteins have been shown to localize to the nucleus, and thus may interact with transcription factors. It is also interesting to note that several yeast and plant proteins possessing the SPX domain have been shown to negatively regulate Pi signaling and/or transport. Whether such a negative feedback regulatory function can be generalized to other proteins harboring the SPX domain has yet to be determined. Gaining further information on the localization of the SPX domain-containing proteins will be a key process in the identification of their function. Ultimately, the identification of the role of the SPX domain, and SPX domain-containing proteins, will greatly improve our knowledge of Pi signaling and transport in plants, assisting efforts to breed plants with better P use efficiency.
This work was supported by the Australian Research Council Super Science Fellowship (FS100100022), the Key Basic Research Special Foundation of China (2011CB100303), the Science and Technology Cooperation Project of China and Australia (20080242) and the Swiss National Fund (3100A0-122493 and 31003A-138339).