Integration of P, S, Fe, and Zn nutrition signals in Arabidopsis thaliana: potential involvement of PHOSPHATE STARVATION RESPONSE 1 (PHR1)

1Among environmental constraints, mineral nutrition plays a key role for plant growth and development. Variations in soil nutrient composition and availability are the rule and plants have evolved mechanisms to cope with conditions ranging from extreme deficiency to toxicity due to excess. Plant breeding was oriented these last 50 years to provide crops to modern agriculture with high intrinsic growth rates and yields, under the condition that mineral nutrition was not limiting. Such condition was obtained by the massive use of fertilizers, in particular considering nitrogen (N), phosphorous (P), and potassium (K). The future of agriculture will undoubtedly require to use so far uncultivated lands, some of them exhibiting unfavorable soil mineral composition, and to reduce the use of fertilizers in order to promote sustainable practices. In such a context of lower input into the environment, new cultivated plant genotypes will need to be selected in a way improving their mineral use efficiency. Reaching such a goal would be facilitated by a knowledge-based approach rooted in the understanding of how plants sense and signal changes in the availability of nutrients.

2A wealth of knowledge was obtained these last 20 years on the physiological and morphological adaptation of plants in response to variations in availability of a given mineral element. Genes encoding proteins involved in uptake, translocation, assimilation, and storage of macro and micro-elements have been characterized and the regulation of their expression in response to mineral status has started to be elucidated. More recently multi-level studies integrating transcriptome to metabolome and to enzyme activities data enabled to begin to understand how plants reprogram various metabolic pathways in response to removal and/or resupply of mineral nutrients. It gives an insight into how plants integrate metabolism adaptation to mineral nutrition deficiency to growth. However, it is well known that interactions between nutrients for uptake can cause imbalances if one of them is deficient or in excess. Multi-level interactions between the various mineral elements need therefore to be studied in order to understand how the different sensing and signaling pathways activated in response to changes in availability of one element are coordinately integrated with the ones of other elements.

3In such a context, the principal aim of this paper is to review interactions between phosphorus (P) and sulfur (S) on one hand, and between zinc (Zn) and iron (Fe) on the other hand. In addition, phosphate (Pi) and sulfate (SO4) not only interact between themselves but also influence Zn and Fe nutrition, and these intricated nutritional cross-talks are presented, pointing out the emerging role of transcription factors (TFs) belonging to the MYB family.

4Due to their sessile nature plants must face and adapt to a variety of biotic (e.g., bacteria, fungi, etc.) and abiotic (e.g., cold, drought, etc.) stresses throughout their life cycle. As a consequence plants have evolved molecular mechanisms allowing a tight control of their growth and development. This process is complex and dynamic, and requires the coordinated expression of several thousands of genes.

5Transcription factors are sequence-specific DNA binding proteins that play a key role in the control of genes expression by acting as transcriptional activators, repressors or both. TFs possess a modular structure that is characterized by two key domains, a DNA-binding domain (DBD) and a transcriptional regulatory domain. TFs have been categorized into various families on the basis of some specific amino acid signatures mostly present in their DBD.

6Among the various classes of TFs found in plants, the MYB family is one of the largest and most diverse. MYB proteins are characterized by their DBD (Figure 1), or MYB domain. It is composed of different numbers (from 1 to 4) of imperfect repeats (R) of approximately 50 amino acids. Each repeat forms a helix-turn-helix (HTH) structure containing three evenly spaced tryptophan residues. These residues form a hydrophobic core playing a key role in the sequence-specific binding to DNA. The MYB gene family is divided into various groups according to the number and the type of repeat(s) found in their DBD.

7To date, no biological role directly related to plant responses to abiotic stresses has been clearly reported for the 3R- and 4R-MYB proteins. 3R-MYBs are found in all eukaryotic cells where they participate to the control of the cell cycle. In contrast, the role of the plant specific 4R-MYBs remains elusive.

8The R2R3-MYB class (two repeats) is the largest group of MYB-proteins exclusively found in plant species. For instance, out of the 196 MYB genes found in the Arabidopsis thaliana genome, 126 encode R2R3-MYB proteins. R2R3-MYBs are specifically involved in the transcriptional control of plant-specific processes, including plant responses to various abiotic stresses, such as cold or drought. This TF sub-family has been extensively studied in Arabidopsis allowing to determine the biological role played by more than half of its members. For example, AtMYB14 and AtMYB15 are involved in the plant response to cold stress. AtMYB60 and AtMYB96 act through the ABA signaling cascade to modulate plant response to drought, by controlling stomatal movement, and root growth and cuticular wax deposition, respectively. In contrast, AtMYB2 and AtMYB44/AtMYBR1 regulate the expression of their target genes in response to drought in an ABA-dependent manner. AtMYB88 and AtMYB124/FLP paralogs are key regulators of stomata differentiation. They have recently been shown to be involved in sensing and/or transducing salt stress (and most probably other abiotic stresses; ). AtMYB20 and AtMYB73 are also involved in salt stress tolerance, whereas AtMYB108/BOS1 displays a less specific role as it is required in the response to both pathogens and abiotic stresses, including drought, salinity, and oxidative stress.

9Last, the single MYB repeat proteins forms a heterogeneous group that gather genes that can be classified into five major subgroups: the CPC-like, the CCA1-like/R-R, the I- box-like, the TRF-like, the TBP-like, and the GARP. To date, only 1R-MYBs belonging to the CPC-like (also called R3-MYB), and CCA1-like/R-R subgroups, have been associated to plant abiotic stress responses. The Arabidopsis AtMYBL2 is a CPC-like MYB protein whose activity is decreased under high light stress, and which acts as a negative regulator of anthocyanin biosynthesis (which provide a natural sunscreen for plants), and. Amongt the CCA1-like/R-R group of MYB proteins, OsMYBS3 has been shown to be involved in cold tolerance in rice (Oryza sativa), whereas AtMYBL plays a role in the Arabidopsis response to salt stress. AtMYBC1 that belongs to the GARP subgroup was found to be a negative regulator of freezing tolerance in Arabidopsis.

10Nutrients availability is also an important environmental factor which modulates plant growth and development, and therefore crop productivity. Consequently, deficiencies in nutrient supplies are abiotic stresses against which plants have evolved signaling cascades aiming to improve nutrient acquisition and homeostasis. Similarly to the above-mentioned abiotic stresses, MYB proteins have also been found to be involved in the plant response to nutrient deficiencies. For example, two homologous R2R3-MYB proteins (namely AtMYB10 and AtMYB72) have been shown to play a key role in improving growth under Fe-limiting conditions. However, most of the MYB proteins identified so far as involved in the regulation of mineral nutrition were associated with Pi starvation. For example AtMYB62 participates in the response to Pi shortage, whereas AtETC1 (ENHANCER OF TRY AND CPC 1) is a CPC-like MYB TF acting to modulate root hair density under Pi limited conditions.

11PHOSPHATE STARVATION RESPONSE 1 (PHR1) and PHR1-Like 1, (PHL1) are two homologous GARP MYB proteins that play a critical role in the adaptation of plant to Pi (inorganic phosphate) deficiency. Originally, PHR1 was first identified as a regulator of Pi nutrition in Chlamydomonas reinhardtii, and named PSR1. It is required for the transcriptional induction of Pi-deficiency responsive genes in this green algae. A genetic screen enabled to identify a PSR1 ortholog in Arabidopsis, AtPHR1. Several years later a redundant gene to AtPHR1 was characterized and named AtPHL1. These two TFs are able to interact, and they recognize the same cis element located in the promoter sequence of their target genes. A consensus sequence of this element, called PHR1 Binding Site (P1BS), has been defined as 5′-GNATATNC-3′ (;; Figure 2). Transcriptome analysis revealed that most of the genes induced in response to Pi deficiency lost this ability in phr1 phl1 double mutant plants. Furthermore, the frequency of occurrence of the P1BS element is much higher in the promoter region of genes induced in response to Pi deficiency than in others. Interestingly, two Arabidopsis PHR1 orthologs, OsPHR1, and OsPHR2, have been characterized in rice, and functional P1BS cis-elements have been reported in barley. These data indicate that the regulatory pathway involving PHR1 to activate the expression of some genes in response to Pi starvation is likely to be conserved between monocotyledonous and dicotyledonous plants. It is noteworthy that a recent study has pointed out that the expression of AtFer1, a key gene involved in Fe homeostasis, was transcriptionally regulated by AtPHR1 and AtPHL1, providing a direct molecular link between Fe and Pi homeostasis (, and see below the “PHR1 involvement in Pi and Fe homeostasis interactions” section).

12P is an essential macronutrient required by living organisms. It is found in essential biological molecules including nucleic acids, ATP (a major energy carrier) and phospholipids. P is intimately linked with energy metabolism, the production of numerous metabolic intermediates, and the post-translational modification of proteins (a key parameter in signal transduction cascades; ). Its deficiency has deleterious effects on plant growth and development, evidenced by a strong decrease in shoot growth, and by the accumulation of anthocyanins. Plants acquire P by their roots as inorganic Pi. Pi concentration is heterogeneous in the soil, and often very low at the root/soil interface. Root architecture responds to Pi deficiency by inhibiting primary root growth and by increasing lateral root density. This is an adaptive strategy to explore more soil volume, resulting in an increased Pi uptake capacity of the plants. After having crossed the plasma membrane of epidermal and cortical root cells, Pi is distributed throughout the plant under the control of a cascade of Pi transporters belonging to the PHT and PHO1 families. For an extensive review of the Pi transporter gene family, readers are referred to, and to. Plant adaptation to Pi deficiency involves wide changes in gene expression. It implicates several TFs. Few of them have been identified including WRK75, ZAT6, bHLH32, MYB62, and PHR1. So far, the major regulations of the expression of numerous Pi deficiency-induced genes was attributed to PHR1, via the PHR1-PHO2–miRNA399 pathway described below (Figure 2). In addition to this pathway, plants respond to Pi deficiency through pathways involving phytohormones and various metabolites. Among phytohormones, cytokinins suppress the up-regulation of several genes in response to Pi deficiency, including Pi uptake transporters. It requires the cytokinin receptor CYTOKININ RESPONSE 1/WOODEN LEG/ARABIDOPSIS HISTIDINE KINASE 4 (CRE1/WOL/AHK4) pathway. Among the metabolites, carbohydrates are involved in the Pi deficiency response. The expression of Pi transporters is sensitive to the carbon status of the plant, either upstream or downstream of the hexokinase (HXK) activity in glycolysis.

13At a molecular level Pi deficiency is regulated both at the transcriptional and post-transcriptional levels. The major actors coordinating these various regulations are PHR1 and PHL1, two TFs likely conserved amongst flowering plants.

14Indeed, sensing of the Pi status of the plant is likely conserved between mono- and dicotyledonous plants, as recently reported. Expression of AtPHR1 and of its rice orthologous gene OsPHR2, are not responsive to Pi, raising the question of how plants sense the intra-cellular variations of Pi concentrations. In this context, recently reported that rice OsSPX1 and 2, which are nuclear proteins whose expression is itself activated by OsPHR2 under low Pi conditions, interact with OsPHR2 by their SPX domain in a Pi-dependent manner. This interaction results in the inhibition of OsPHR2 binding on its cis-acting P1BS sequence (Figure 2). Therefore, this mechanism constitutes a very efficient transcriptional regulatory feedback loop to fine tune the PHR1 dependent expression of Pi responsive genes, according to the intracellular fluctuations of Pi concentrations.

15Among the genes transcriptionally regulated by PHR1/PHL1 in response to Pi deficiency are the PHT1 genes (Figure 2). They encode plasma membrane high-affinity H+/Pi co-transporters, preferentially expressed in the root epidermal or cortical cells, and therefore directly involved in Pi acquisition. Once Pi has entered the root, it is loaded into the xylem sap for translocation to the shoots via PHO1, which is specifically expressed in the pericycle. PHT1 and PHO1 transporters are also post-translationally regulated during their intracellular trafficking to the plasma membrane, and the C terminus phosphorylation of PHT1 proteins retains them at the ER upon Pi refilling. During the post-ER trafficking, the ubiquitin E2 conjugase PHO2 modulates the ubiquitination status of PHT1 and PHO1 transporters in order to control their rate of degradation by the proteasome. PHO2 expression is itself post-transcriptionally repressed by miR399, a small non-coding RNA up-regulated by Pi deficiency at the transcriptional level through PHR1/PHL1 activity (;;;;; Figure 2). Post-translational control of Pi loading into the xylem via the degradation of PHO1 under Pi-sufficient conditions has also been reported to be mediated by PHO2.

16Phosphate Starvation Response 1 is itself regulated post-translationally in response to Pi deficiency through the action of SIZ1, a sumo E3 ligase (Figure 2). PHR1 can be sumoylated in vitro by SIZ1, and Pi-deficient regulated genes are no more induced in siz1 mutant in response to Pi deficiency. This post-translational regulation of PHR1 could modify its activity, in order to modulate the transcriptional activation of its target genes, among which the IPS1 gene. IPS1 encodes a non-coding RNA whose sequence is in part complementary of the microRNA miR399, enabling post-transcriptional regulation via RNA mimicry. IPS1-miR399 matching would therefore lead to the inhibition of the miR399 RNA activity known to target PHO2 (Figure 2). It could explain the increased expression of PHT1-8 and PHT1-9 in pho2 mutant plants, and the decreased expression of PHO2 and of the PHT1-8 and PHT1-9 Pi transporter genes in miR399 over-expressing transgenic lines. The involvement of PHT1-8 and PHT1-9 in Pi uptake and/or in Pi root-to-shoot translocation has been reported. More recently an iTRAQ based quantitative membrane proteomic approach was used to search for components downstream of PHO2. It enables to show that PHO2 mediates the degradation of PHT1 proteins by interacting in the post-endoplasmic reticulum compartments where ubiquitination of endomembrane-localized PHT1-1 occurs. Finally the systemic effect of miR399 was evidenced by grafting experiments. It demonstrated the transport of this small RNA from leaves to root within the phloem sap, explaining why specific overexpression of the miR399 in leaves led to decrease the expression of PHO2 in roots.

17Sulfur is an essential element for plant growth and development. The major source of S for plants is inorganic sulfate (SO4; ). In the cell, S is involved in different functions and aspects of plant metabolism. It is found both in reduced (amino acids, peptides, and proteins, lipoic acid, iron–sulfur clusters), and in oxidized (polysaccharides, lipids, and sulfonate group modifying proteins) forms. Thus S deficiency has profound effects on plant physiology. When challenged by SO4 deficiency, shoot growth is affected, resulting in a decrease in the total biomass. Modulation of root system architecture has also been observed in response to S limitation. The primary root continues to grow, but the lateral roots form closer to the root tip, with an increased density. During the last two decades, a comprehensive view of SO4 transport in Arabidopsis and other plant species has emerged. Numerous “SULTR” genes encoding SO4 transporters for uptake as well as for inter-organs and subcellular distribution have been functionally characterized. The first key step for SO4 uptake in Arabidopsis is mainly carried out by the two high-affinity SO4 transporters SULTR1;1 and SULTR1;2. The possibility that SULTR1;2 may function as a sensor of S status or as a component of a S sensory mechanism has recently been proposed.

18Some key regulatory molecular mechanisms and components involved in the regulation of SO4 transport have been discovered, among which a regulatory pathway requiring miRNAs, including miR395. This microRNA contributes to the regulation of SO4 root-to-shoot transfer involving the SULTR2;1 gene. The transcript abundance of this gene is strongly decreased under deficient S conditions. It can be explained by the increased level of miR395 under such conditions, leading to a strictly restricted expression of the SULTR2;1 gene in the xylem. miR395 acts downstream of the TF SULFUR LIMITATION 1 (SLIM1), also known as ETHYLENE-INSENSITIVE3-LIKE 3 (EIL3;; ).

19Plants have evolved tightly controlled mechanisms allowing the coordination of the S transport and homeostasis with photosynthesis and the carbon status, in a similar manner to Pi transport system. In addition to these mechanistic similarities in the regulation of the Pi and SO4 transport systems, plants have developed coordinated and tightly controlled mechanisms to maintain intracellular homeostasis of both elements in response to their external availability. It has been reported that plant cells operate a rapid replacement of sulfolipids by phospholipids under S deficiency, and the replacement of phospholipids by sulfolipids during Pi deficiency. Such a metabolic switch attests the P/S nutritional interdependency. Interestingly, two genes necessary for the replacement of phospholipids by sulfolipids in Pi-deficient plants, SQD1 and SQD2, contain a PHR1 binding sequence (P1BS) in their promoter, and are up-regulated by Pi deficiency in a PHR1-dependant manner. The accumulation of SO4 and Pi was affected in Arabidopsis lines characterized by a very low inositol-6-phosphate (phytic acid, PA) content. The expression of genes involved in the SO4 and Pi transport or signaling was altered in these low PA mutants. PA emerged thus as a component of the co-regulation of SO4 and Pi homeostasis. reported that the increased accumulation of miR395, known to be up-regulated by SO4 starvation, is suppressed in Pi-deficient plants. It could be therefore a mean to increase the S translocation from root to shoot by SULTR2;1, enhancing thus the sulfolipid biosynthesis in replacement of phospholipids under Pi deficiency. Evidences for co-regulation of Pi and S signaling pathways are starting to emerge from recent data. reported that the SO4 concentration increases in roots and decreases in shoots of Arabidopsis Pi-deficient plants. These results indicate an adaptive regulation of the SO4 transport process upon Pi deficiency in plants, and particularly the inter-organ SO4 distribution. Interestingly, PHR1 is required for this process to take place, likely through its positive regulatory role on SULTR1;3 expression. PHR1 has also a negative effect on the expression of SULTR2;1 and SULTR3;4 in response to Pi deficiency. Noteworthy, such a function is also preserved in Chlamydomonas reinhardtii via the PHR1 ortholog PSR1. An integrative model for the regulation of the expression of genes involved in intracellular and inter-organ SO4 transport under Pi deficiency in Arabidopsis has been proposed, in which PHR1 plays an integrative role. Considered together, these data reveals an unsuspected level of complexity and interconnection in the regulation of SO4 and Pi homeostasis in plants.

20Zn is an essential microelement for cell life. It is the only metal represented in all six classes of enzymes: oxidoreductases, transferases, hydrolases, lyases, isomerases, and ligases. It also plays a structural role in regulatory proteins. Despite its essentiality, high concentrations of Zn can be toxic for the cell, and cause oxidative stress, a decrease in accumulation of ATP, disintegration of cell organelles, and biogenesis of vacuoles.

21Zn deficient plants exhibit deformed and chlorotic leaves, and interveinal necrosis, leading to decrease biomass production. Transport across the plasma membrane is achieved by transporters belonging to the ZIP (ZRT, IRT-like protein) family such as AtIRT1. A passive Zn influx can also occur through the depolarization-activated non-selective cation channel (DA-NSCC; ), or by the voltage-independent NSCCs (VI-NSCCs; ). Zn loading into the xylem sap involves plasma membrane transporters which are members of the P1B-ATPases family, namely AtHMA2, and AtHMA4. For further information on the Zn transport in plants, readers are referred to. So far, how plants sense and transmit the signal of Zn deficiency remains poorly understood.

22Fe, as Zn, is essential. It is also potentially toxic when in excess because of its reactivity with oxygen which catalyzes the formation of free radicals able to oxidize organic molecules, ultimately leading to cell death. Fe is a key element to ensure the electron flow through the PSII-b6f/Rieske-PSI complex. It is therefore essential for CO2 fixation by the photosynthesis process. Indeed it has been well documented that Fe is a limiting factor for biomass production, as well as for the quality of plant products. Fe enters the plant via the roots, from where it is distributed within the plant. According to the plant family considered – i.e., graminacea plants versus other plants – two mechanisms prevail for mining Fe from the soil solution. They involve, respectively, chelation of the ferric iron (Fe3+) by small organic molecules, among which methionine derivatives constitute the mugineic acid (MAs) family, or its reduction in its ferrous form (Fe2+) prior to transport across the plasma membrane of root epidermal cells. This transport is achieved by transmembrane proteins exhibiting transporter activities, such as YS1 responsible of the transport of Fe3+-MAs complexes in graminacea, or IRT1 responsible of the root uptake of Fe2+ by non-graminacea plants. Many other transporters, as well as soluble proteins, responsible for Fe long distance distribution, sub-cellular compartmentation, and storage have also been characterized this last decade. The expression of these genes encoding proteins responsible of Fe homeostasis at the whole plant level is precisely regulated through integrated pathways modulated by the sensing of the Fe status of the plant. These controls mainly operate at the transcriptional level via a netwotk of TFs, most of them belonging to the bHLH family, but also at the post-transcriptional level. How plants sense their Fe status is now the new frontier in this field of research. It could be achieved, at least in part, by some TFs and/or by Hemerythrin motif-containing Really interesting new gene and Zinc-finger proteins (HRZs)/BRUTUS (BTS), that were recently identified both in rice and Arabidopsis, respectively.

23Fe deficiency in Arabidopsis leads to the activation of expression of IRT1, the primary transporter responsible of root Fe uptake. IRT1 has a weak substrate specificity and contributes therefore to the accumulation of a broad range of divalent transition metals including Zn. Conversely, excess Zn causes physiological Fe deficiency. Early studies reported an absence of IRT1 protein in Arabidopsis roots from plants grown under Zn excess conditions, suggesting that the known post-translational regulation of IRT1 protein levels through ubiquitin-mediated proteasomal degradation might predominate under this condition. However, it was reported more recently that IRT1 protein in Arabidopsis grown under excess Zn increased in abundance, comparatively to plants grown under standard Zn nutrition conditions. IRT1 could therefore be a major contributor to Zn deficiency.

24MTP3 (a member of the Cation Diffusion Facilitator family), HMA3 (belonging to the P1B-type ATPase family), and ZIF1 (a member of the Major Facilitator Superfamily transporters) are vacuolar membrane proteins required for Zn tolerance, and encoded by genes which are transcriptionaly activated in response to excess Zn or Fe deficiency. Indeed MTP3 and HMA3 gene expression is decreased in the fit mutant. This mutant lacks a bHLH TF that activates root Fe deficiency responses, in particular the transcriptional up-regulation of the Fe(III)-chelate reductase FRO2 and of the Fe2+ transporter IRT1. Interestingly, transcriptional regulation of the ZIF1 gene under Fe deficiency is independent of Fe deficiency-induced Zn accumulation. The bHLH TF POPEYE, controlling a set of genes expressed in the stele and important for the control of Fe homeostasis was reported to directly interact with the ZIF1 promoter and to repress its transcription, although there is a concurrent net increase in ZIF1 transcript levels, under Fe deficiency.

25These results evidence the complexity of the cross-talks between the pathways at work to regulate Fe deficiency and Zn excess in order to establish an integrated response, and the necessity of additional work in the future to decipher them.

26The two macro-elements Pi and S on one hand, and the two metals Zn and Fe on the other hand, do not only interact between themselves as reported above. Indeed Pi and S status of the plant can also influence Zn and Fe nutrition, and these aspects will now be reviewed below.

27From a biochemical point of view, Fe and S are known to interact for the building of Fe–S clusters, which are a major sink for Fe, and known to be essential for photosynthesis, respiration, and many cellular enzymatic reactions. However, it is only recently that Fe and S interactions have been documented at physiological and molecular levels, although the cross-talks between the pathways regulating the integrated homeostasis of these two elements remain to be deciphered. It is well established that leaf Fe concentration decreases in S-deficient tomato, comparatively to S-sufficient control plants. This is consistent with the observation that the expression level of the LeFRO1 gene encoding a root Fe3+-chelate reductase, the activity of this reductase, and the reduction based of 59Fe uptake, are decreased in response to S starvation. In agreement with these observations a strong repression of expression of the Arabidopsis thaliana IRT1 Fe2+ transporter in response to S deficiency was recently reported. Reciprocally, it has been observed that Fe starvation modifies S uptake and assimilation. At a genome wide level, data mining of transcriptomes from Fe deficient Arabidopsis plants revealed a cluster of S-metabolism related genes (including genes encoding plasma membrane and tonoplast S transporters, and enzymes of the S assimilation pathway such as adenosine phosphosulfate reductase) co-expressed with Fe-deficient genes. Also, the high affinity S transporter SULTR1;1 mRNA abundance is 2.5-fold lower in absence of Fe. However, a contrasted report was recently published concerning this last point. It indicated that the expression of most of the group 2 and 4 S transporters were up-regulated in Fe starved tomato plants. It turned out to be also the case of the SIST1;1 and SIST1;2 tomato high affinity transporters of the group 1, to which belongs the SULTR1;1 transporter from Arabidopsis. Resolution of these discrepancies concerning the impact of Fe deficiency on the regulation of expression of high affinity SO4 transporters clearly requires further work.

28The interaction between Fe and S metabolisms has not only been studied in plants acquiring Fe from the soil through a reduction-based strategy as it occurs in tomato (Solanum lycopersicum) or Arabidopsis. It has also been investigated with graminaceous plants. Synthesis of Fe3+-chelators of the MAs family and their release in the rhizosphere were reduced in barley (Hordeum vulgare L.) plants grown under S starvation. However, transcript abundance of the YS1 gene encoding the Fe3+-MAs transporter increased in response to Fe deficiency at the same level, whatever the S nutrition conditions imposed. The modifications of S metabolism occurring in response to Fe deficiency have also been investigated in durum wheat (Triticum turgidum L.; ). These authors have shown that Fe deficiency under S sufficient nutrition conditions led to a S deficiency response at a molecular level. This response was characterized by an increase in abundance of some of the transcripts encoding enzymes of the S assimilation pathway, including APS reductase, ATP sulfurylase, sulfite reductase, and serine acetyltransferase. Furthermore, the activity of the corresponding enzymes was also found to be increased by Fe shortage. However, changes in mRNA abundance of some other genes of the S assimilation pathway, and the activity of the corresponding enzymes, were observed to be uncoupled in their response to Fe or S deprivation. In addition, the expression of the wheat SULTR1;3 high affinity SO4 transporter was significantly increased in roots and shoots in response to both S or Fe shortage, with the highest expression level observed under Fe deficiency conditions. In contrast, the expression of the SULTR1.1 transporter gene, mainly expressed in roots, was strongly induced in response to S deficiency, but unaffected by Fe deficiency.

29In conclusion, the interactions between Fe and S metabolisms are attested both in graminaceous and non-graminaceous plants. These interactions have started to be documented at a molecular level, reporting that Fe deficiency modifies the expression of genes involved in S transport and assimilation, and vice-versa. Nevertheless, the characterization of these interactions is still in its infancy, and more work is needed to understand the complexity of the integration of the various pathways involved. Of particular interest would be the study of a possible role of the synthesis of Fe–S cluster, and of their relative abundance in response to various nutritional stress, as driving forces of the Fe–S interactions.

30Clear physiological links have been established between Fe and Pi. The complexation of Fe by Pi in soils leads to the formation of precipitates, decreasing the availability of these two elements for plants. As a consequence, the high affinity root Fe2+ uptake system, which is induced by Fe deficiency, is also activated under Pi excess conditions. Conversely, Pi starvation promotes metal accumulation in plants, mainly aluminum, and Fe. From a developmental point of view, it is also well documented that a decrease of primary root growth under Pi deficiency is partly due to Fe toxicity. Once taken up by the roots, Pi is translocated to the shoots, and Fe can interact with Pi inside roots leading to a reduced Pi translocation to the shoots. The same kind of interaction has been described in leaves in which high Pi content favors chlorosis, even if the leaf Fe level is sufficient. Finally, in seeds Fe is stored in vacuoles under the form of globoids composed in part of phytate (inositol hexakisphosphate = IP6)-Fe complexes. These observations indicate that Pi is a very efficient Fe chelator. As a consequence, changes in Pi homeostasis will strongly influence Fe availability.

31At the molecular level, transcriptome analysis of Pi deficient plants revealed an increase in abundance of transcripts from Fe excess responsive genes, and reciprocally a decrease in abundance of transcripts from Fe deficiency responsive genes. In this context, the induction of expression of the AtFER1 gene, encoding the Fe storage protein ferritin, in response to Pi deficiency was interpreted by some authors as a consequence of an increase in available Fe under such conditions. Indeed this interpretation is likely wrong since the abundance of AtFER1 mRNA in response to Pi starvation was recently shown to be mediated by PHR1 and PHL1, through their binding to a P1BS cis-element found in the AtFER1 promoter, independently of the plant Fe nutrition conditions. Furthermore, the Fe-dependent IDRS cis-acting element present in the AtFER1 proximal promoter is not required for Pi-deficient induction of AtFER1 expression. Moreover, AtFER3 and AtFER4 ferritins genes, lacking P1BS cis-element in their promoter and known to be induced by Fe excess, have their expression unchanged by Pi starvation. Finally, induction of AtFER1 expression in response to Fe excess is not altered in phr1 nor in phr1xphl1 mutant plants.

32At a cellular level, Fe distribution around the vessels was abnormal in phr1x phl1 double mutant, suggesting that PHR1 and PHL1 may link Fe and Pi homeostasis not only by regulating AtFER1 gene expression, but also by having a broader regulatory function of Fe metabolism. Indeed, genome wide analysis of Pi starved wild-type and double phr1xphl1 mutant plants revealed other Fe homeostasis related genes such as NAS3 (NICOTIANAMINE SYNTHASE 3) and YSL8. These genes were induced upon Pi starvation in wild type, and exhibit a decreased induction in the double mutant plants. Moreover, Fe deficiency responsive genes, including FRO3, IRT1, IRT2, NAS1, and FRO6 were repressed upon Pi starvation in wild type and miss-regulated in the phr1xphl1 double mutant plants. These observations strengthen a global role of PHR1 and PHL1 in the overall control of Fe homeostasis that is, by this mean, integrated to the Pi status of the plant.

33Pi and Zn homeostasis in plants are known to strongly interact. Nevertheless, the molecular bases of the Pi–Zn interactions remain so far poorly understood. Long-term Pi deprivation leads to Zn over-accumulation, and inversely Zn starvation appears to cause an over-accumulation of Pi. Transcriptome data from roots of Pi- or Zn-deficient Arabidopsis plants, could induce the expression of many Pi-related genes, and that Pi starvation modifies the expression of genes involved in the maintenance of metal homeostasis such as Zn and Fe. Considered together, this transcriptome data reinforce the evidence of a cross-talk between Pi and Zn signaling pathways in planta. Cross-analysis of these expression data showed that Pi and Zn deficiencies often act in anopposite manner on these sets of genes. Many genes induce by Zn deficiency are repressed by Pi deficiency. Interestingly, it was observed that most of these genes were further repressed in the phr1 mutant genetic background. This observation further supports the implication of PHR1 in the regulation of expression of Pi-related genes under Zn limitation in Arabidopsis. It agrees with a report by showing the involvement of AtPHR1 in the cross-talk between Pi and Zn homeostasis. As aforementioned, PHR1 was already known as a major regulator of Pi deficiency signaling through its involvement in the PHR1-miR399-PHO2 regulatory pathway. However, this regulatory pathway is not involved in the over-accumulation of Pi in the shoot in response to Zn deficiency. The Zn-responsive signaling pathway in which PHR1 is involved thus remains to be elucidated. also identified two additional genes that are involved in the control of Pi accumulation in response to Zn deficiency, namely PHO1 and its homolog PHO1;H3. PHO1 is a well-characterized root-to-shoot Pi transporter. It is likely one of the final targets of the Zn-deficiency signaling pathway. Since its level of expression is unchanged in response to Zn deficiency, one favorite hypothesis is that its activity would be regulated through a protein–protein interaction, considering that a similar mechanism involving PHO1 and PHO2 has already been reported. The PHO1;H3 gene appeared to be specifically and strongly induced by Zn deficiency. PHO1;H3 constitutes thus an interesting entry point to study Pi–Zn crosstalk and the regulation of Pi transport under Zn limitation. A first model illustrating Pi–Zn crosstalk in plant has been proposed. In the future, this model would benefit from an in depth characterization of the molecular mechanisms underlying the effect of Pi starvation on Zn nutrition, together with the effects of Zn deficiency on Pi nutrition. The identified and characterized key genes and mechanisms acting in the coordination of Pi and Zn transport and signaling in plants could help to point to specific mutants or genetic variants that could be used in breeding programs. Taken together, this knowledge should have important consequences for both basic and applied research in agronomy and should be valuable to plant biologists, agronomists and breeders.

34PHR1 was initially described as a major transcriptional regulator of Pi homeostasis. It activates the transcription of Pi deficiency responsive genes encoding Pi transporters, as well as regulatory RNA and proteins involved in the post-transcriptional and post-translational regulation of these transporters, respectively (;;; Figure 2). More recently, few reports indicated that PHR1 regulated also the expression of genes required for S, Fe, and Zn transport and homoeostasis, linking the metabolism of these nutriments to Pi nutrition (Figure 3). PHR1 is therefore the first molecular link common to various pathways controlling mineral nutrition of both macro- and micro-elements. It can be extended to some aspects of the control of plant response to water stress, for which AtPHL1 activity is required. As mentioned above, in response to Pi starvation PHR1 controls the transcription of post-transcriptional and post-translational regulators such as miR399, IPS1, and PHO2 (Figure 2). Whether or not these PHR1-dependent regulators play a role in the regulation of S, Fe, or Zn metabolisms is unknown, and would deserve to be investigated. Furthermore, the requirement of PHR1 has been documented in the case of two by two interactions of only some of the elements considered: Pi and S, Pi and Fe, Pi and Zn but to our knowledge no reports mentioned a direct role of PHR1 in the Fe–S or Fe–Zn interactions.

35PHR1 has been the most studied regulator of Pi deficiency response, but it is known that other regulators are involved. Among them TFs including WRKY75, ZAT6, MYB62, PTF1, bHLH32, and WRKY45 have been reported. Furthermore, not only additional TFs have to be considered. In addition to the well-characterized role of miR399 in the regulation of Pi homeostasis in Arabidopsis, other miRNAs (miRNA778, miRNA827, and miRNA2111) were reported to be specifically and strongly induced in response to Pi starvation. The role of these miRNA in the potential cross-talks between pathways regulating homeostasis of various mineral nutriments has already been suggested for miRNA82, involved in the cross-talk between Pi-limitation and nitrate-limitation signaling pathways affecting anthocyanin synthesis. Finally, chromatin modification could also be concerned through H2A.Z histone deposition via the nuclear actin-related protein ARP6 in order to activate the expression of many genes related to the Pi deficiency response.

36Integration of pathways controlling two by two nutriment homeostasis has started to be documented. However a survey of transcriptome data reveals that the role played by PHR1 and PHL1 in these interactions could be wider (Figure 4). Indeed PHR1 and PHL1 control transcript accumulation of key genes of Fe homeostasis as well as genes whose expression is directly dependent on S or Pi availability. In consequence, a major challenge in the future will be to consider mineral nutrition as a system, and to develop tools enabling to model integrative gene networks that will take into account the availability of a maximum of macro-and micro-elements, and their interactions, at a given time.

37The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.