Logo of plntphysLink to Publisher's site
Plant Physiol. Jul 2003; 132(3): 1260–1271.
PMCID: PMC167066

Phosphate Starvation Triggers Distinct Alterations of Genome Expression in Arabidopsis Roots and Leaves1,[w]


Arabidopsis genome expression pattern changes in response to phosphate (Pi) starvation were examined during a 3-d period after removal of Pi from the growth medium. Available Pi concentration was decreased after the first 24 h of Pi starvation in roots by about 22%, followed by a slow recovery during the 2nd and 3rd d after Pi starvation, but no significant change was observed in leaves within the 3 d of Pi starvation. Microarray analysis revealed that more than 1,800 of the 6,172 genes present in the array were regulated by 2-fold or more within 72 h from the onset of Pi starvation. Analysis of these Pi starvation-responsive genes shows that they belong to wide range of functional categories. Many genes for photosynthesis and nitrogen assimilation were down-regulated. A complex set of metabolic adaptations appears to occur during Pi starvation. More than 100 genes each for transcription factors and cell-signaling proteins were regulated in response to Pi starvation, implying major regulatory changes in cellular growth and development. A significant fraction of those regulatory genes exhibited distinct or even contrasting expression in leaves and roots in response to Pi starvation, supporting the idea that distinct Pi starvation response strategies are used for different plant organs in response to a shortage of Pi in the growth medium.

Phosphorus is one of the three essential macronutrients of plants. Phosphorus is not only a constituent of such key cell molecules as ATP, nucleic acids, and phospholipids, but it is also a pivotal regulator in many metabolisms, including energy transfer, protein activation, and carbon and amino acid metabolic processes (Marschner, 1995). Limitation of phosphate (Pi) causes both molecular and developmental adaptation in all organisms. In higher plants, Pi limitation leads to dramatic changes in root growth and architecture, such as an increase root to shoot ratio, an increase of lateral roots, and an increase of root hair number and length, thus Pi use can be enhanced (Bates and Lynch, 1996; Rubio et al., 2001).

Many molecular adaptation responses have been described from various organisms. For example, under Pi limitation, both Escherichia coli and Brewer's yeast (Saccharomyces cerevisiae) activate a multigene-inducible system to scavenge traces of usable Pi from the surrounding medium (Torriani, 1990; Lenburg and O'Shea, 1996). The existence of an analogous multigene Pi starvation-inducible system has also been proposed in higher plants based on studies with tomato (Lycopersicon esculentum) plants (Goldstein et al., 1989). Several individual genes responsive to Pi starvation have been described. These include acid phosphatases with broad substrate specificity (Duff et al., 1994; Trull and Deikman, 1998), phosphoenol-pyruvate (PEP) phosphatase and pyrophosphate-dependent phosphofructokinase (Theodorou and Plaxto, 1993), RNases (Green, 1994), high-affinity Pi transporter (Raghothama, 2000), phosphodiesterases (Abel et al., 2000), β-glucosidase (Malboobi et al., 1998), and others of unknown function (Liu et al., 1997; Burleigh and Harrison, 1999).

Pi starvation-responsive genes appear to be involved in multiple metabolic pathways, implying a complex Pi regulation system in plants. Investigation of Pi starvation-responsive genes at a genomic level is required to establish a more complete inventory of the genes and pathways that are responsive to Pi starvation stress. A whole-genome DNA microarray analysis revealed that the yeast system consists of more than 20 Pi starvation-regulated genes under the same control mechanism (Ogawa et al., 2000). Recently, a MYB transcription factor (TF) responsive to Pi starvation was found based on a mutant screened from an Arabidopsis transgenic line harboring a reporter gene specifically responsive to Pi starvation (Rubio et al., 2001). This TF is directly involved in the expression of the Pi starvation-inducible genes, AtIPS1, AtIPS3, and RNS1. However, a systematic gene expression profile analysis and characterization of metabolic pathways under Pi starvation at the genome level are not available for Arabidopsis or for any higher plants. Such a genome-scale analysis would be very useful for understanding the mechanism and would provide a basis for characterization of specific mutations defective in response to Pi starvation.

Microarray technology has become a useful tool for the analysis of plant genome expression profiles under a variety of developmental or environmental conditions. Recently, several successful studies of gene expression profiles responsive to environmental cues have been reported. They include light (Ma et al., 2001, 2002, 2003; Tepperman et al., 2001; Wang et al., 2002), the circadian clock (Harmer et al., 2000), nitrogen nutrition (Wang et al., 2000), plant defense against pathogens (Maleck et al., 2000), and abiotic stresses (Kawasaki et al., 2001; Seki et al., 2001). In the present study, we applied an Arabidopsis microarray (Ma et al., 2001) to analyze genome expression profiles in response to Pi starvation time points in leaves and roots of Arabidopsis. Our results indicate that approximately 29% of the genes on our microarray were up- or down-regulated 2-fold or more during the Pi starvation in leaves and/or roots. This large number of Pi starvation-responsive genes defines a variety of cellular metabolic pathways and functions.


Experimental Design and Plant Pi Concentration Measurements

Wild-type Arabidopsis (Columbia ecotype) seeds were germinated and grown under normal Pi conditions (Chen et al., 2000) up to 30 d and then transferred to Pi starvation conditions (without Pi in the culture solution) for 6, 24, 48, and 72 h, as well as control containing the normal level of Pi. Both leaves and roots were sampled from two separate experiments at the four indicated time points for free Pi content measurement. For each time point, dried weight of leaf and root and total free Pi were determined for 10 plants from each experiment. The average data from the two experiments was used. These measurements revealed that the Pi concentration in root was decreased by 22% after 24 h of starvation, and the decrease declined to about 11% during the 2nd d and about 6% during the 3rd d of Pi starvation, whereas no obvious changes were detected in leaf tissue (Table I). The results indicate that during the 72 h period of Pi starvation, the starvation triggered a significant drop in free cellular Pi in roots, but not in leaves. This result indicates that removing the Pi from the growth medium results in a clear decline of free available Pi in root early on, which can be compensated to a large extent with longer time. However, this treatment did not affect the level of free available Pi in leaves during the same treatment period.

Table I.
Plant growth and free phosphate (Pi) concentration in plants within 72-h course of Pi starvation

Two duplicated biological samples for each treatment were used for total RNA preparation. The ratios of gene expression at each time point after Pi starvation compared with the control plants were obtained from four to six replicated hybridizations, including reverse fluorescent dye labeling. The ratios of both inducible and repressible expression were calculated as described (Ma et al., 2001). Only high quality and reproducible data from parallel RNA preparations of independent biological samples, with correlation coefficient constant better than 0.85, are considered as acceptable data for further analysis (see supplementary material at http://www.plantphysiol.org). Nine representative clones induced or repressed from 2-fold to more than 60-fold from the microarray analysis were selected for RNA gel-blot analysis. The RNA gel-blot results are quite consistent with the results from microarray data (Fig. 1), thus validating the microarray result.

Figure 1.
Comparison of the microarray data and RNA gel-blot analysis for nine selected genes. Total RNA (10 μg lane1) from leaves and roots at 6, 24, 48, and 72 h postonset of Pi starvation (–P6h, –P24h, –P48h, and –P72h) ...

General Features of the Pi Starvation-Responsive Expression Profile

A total of 1,835 genes (about 29% of the total genes on the microarray) exhibited alteration in their RNA expression in response to Pi starvation in at least one of the four time points using a median 2-fold ratio cutoff. The total number of genes that exhibited changes in expression pattern is almost double in leaves (1,398) in comparison with roots (730). There are both overlapping and distinct genes regulated in the leaves and roots. There were totals of 680 and 333 up-regulated genes in leaves and roots, respectively, with 2-fold or higher differential expression for at least one time point. About 192 of those induced genes are shared between leaves and roots, and 488 and 141 are specifically induced in leaves and roots, respectively. There were 718 and 397 genes down-regulated in leaves and roots, respectively. Among them, only 101 genes are shared by roots and leaves, whereas 617 and 296 are specifically repressed in leaves and roots, respectively. Thus the majority of the repressed genes are distinct between roots and leaves upon Pi starvation, suggesting distinct strategies used by those two plant organ types in response to Pi starvation in the growth medium.

Across the four time points examined, most of gene expression changes (1,148 and 632 of genes for leaves and roots) showed a similar kinetic pattern, e.g. initiation of induction or repression after 6 h starvation, with maximum induction or repression at 48 h, and a decrease at 72 h (Figs. (Figs.2A2A and and3).3). Several other kinetic patterns for gene expression changes, however, were also observed. For examples, in leaves, the induction of expression under Pi starvation for 84 and 20 genes peaked at 24 and 72 h, respectively. In roots, 10, 30, and 12 genes were maximally induced at 6, 24, and 72 h, respectively. As for down-regulated genes, 6, 115, and 20 genes were maximally repressed at 6, 24, and 72 h, respectively, in leaves, whereas 7, 9, and 24 genes were maximally repressed at 6, 24, and 72 h, respectively, in root tissue (Fig. 4). As shown in Figure 4, the repression pattern of the genes maximally repressed at 72 h in roots was different from that in leaves. There are also some genes that were induced or repressed at one time point, while oppositely regulated at another time point in the same organ (Fig. 2, B and C).

Figure 2.
Clustering analysis of Arabidopsis genes that exhibited 2-fold or more expression regulation in response to Pi starvation at any time point examined in either leaves or roots. A, Overall cluster display. Note only 1,471 genes from the 1,835 genes were ...
Figure 3.
The number of genes whose expression was regulated by Pi starvation. The number of genes induced or repressed over 2-fold after the Pi starvation-regulated genes in leaves and roots at 6, 24, 48, and 72 h was recorded. At each time point, the four ...
Figure 4.
The number of genes whose expression was maximally induced or repressed at different time points of Pi starvation during 72-h course of Pi starvation. The number of genes maximally induced or repressed over 2-fold after the P starvation in leaves and ...

Functional Classifications of the Differentially Expressed Genes

Those genes whose expression is significantly altered represent a large range of functional categories. On the basis of the recent functional classification of Arabidopsis genes (http://mips.gsf.de/proj/thal/db/search/search_frame.html; http://www.ncbi.nlm.nih.gov), Pi starvation-responsive genes include those involved in cell biogenesis, cellular organization, cellular transport and transport mechanisms, cell division, nucleic acid metabolism, amino acid metabolism, protein synthesis, protein destination, carbon metabolism, photosynthesis, respiration, photo-oxidative respiration, nitrogen, phosphorus and sulfur metabolisms, senescence, transporter facilitation, signal transduction, compartments, transcription, second metabolism, developmental regulation, responses to stresses, and others. Some selected genes involved in carbon and nitrogen assimilation, metabolic adaptations, TFs, and signal transduction are described below. All of the specific genes listed in the figures thereafter and presented below were sequenced to confirm their identity in this study.

Down-Regulation of Genes for Photosynthesis and Nitrogen Assimilation

Although there was no detectable change of Pi concentration in leaves within 72 h of Pi starvation, representative genes for photosynthesis and nitrogen assimilation were down-regulated after 24 h of Pi starvation, which could lead to a growth arrest. A total of 29 genes involved in photosystem (PS) I, PSII, and Calvin cycle and chlorophyll A/B-binding proteins were repressed 2- to 7-fold after 48 h of Pi starvation in leaves. The expression patterns of the representative genes are shown in Figure 5A. These genes encode components of PSI and PSII, which harvest light energy and coordinately produce ATP through photophosphorylation and NADPH and reduced ferredoxin (Fdx) catalyzed by Fdx-NADP reductase (FNR) and Fdx-thioredoxin reductase (FTR). There are also genes that encode Rubisco small subunits, glyceraldehyde-3-phosphate dehydrogenase, and sedoheptulose-1,7-bisphosphatase, all key regulatory enzymes in CO2 assimilation, reduction, and the regeneration phase of the Calvin cycle, respectively. The expression of all of those genes were repressed, most of them maximally at 48 h after onset of Pi starvation (Fig. 5A).

Figure 5.
Pi starvation down-regulated genes involved in photosynthesis, nitrogen, and carbon metabolisms. A, The repression of selected photosynthetic genes during Pi starvation in leaves and roots at the four time points. The ratios for each gene listed are ...

The genes for nitrate reductase, nitrite reductase (NiR), Gln synthase (GS), and Glu synthase (Fdx-GOGAT) were repressed by 3- to 5-fold in both leaves and roots after 24 h of starvation (Fig. 5B). The reduction of nitrate to ammonium requires both nitrate reductase and NiR activities. Nitrite reduction requires Fdx as a reductant. In leaf photosynthetic cells, Fdx is reduced in PSI by FTR. In roots, Fdx is reduced in plastids by FNR, which uses NADPH as a reductant. As shown in Figure 5A, FTR and FNR were repressed in both leaves and roots. Thus it is logical that this reduced supply of reductant is correlated with down-regulation of the nitrate assimilation components.

On the other side, the GS/GOGAT cycle is the principle route of the primary ammonium assimilation and re-assimilation of photorespiratory ammonium and of the synthesis of Glu and Gln in plants. Two classes of GS isoenzymes are found in plants, one in chloroplasts and another in the cytosol. Both genes corresponding to GS were repressed in roots, although marginally or not affected in leaves. Fdx-GOGAT catalyzes a Fdx-dependent reaction to produce Glu using Gln and α-ketoglutarate as substrates, in coordination with GS. In the GS/GOGAT cycle, Asp aminotransferase (Asp), which is involved in producing α-ketoglutarate, was also repressed. This suggests a coordinated down-regulation of primary ammonium assimilation upon Pi starvation.

Interestingly, Glu dehydrogenase (GDH) was induced upon Pi starvation. GDH can catalyze both the synthesis of Glu using NH4+ and α-ketoglutarate as substrates under high NH4+ concentration and its reverse reaction (catabolism) to yield NH4+ and α-ketoglutarate. It has been proposed that the primary role for GDH in vivo is in Glu catabolism, because NH4+ will not be accumulated in higher concentrations in plants and GDH will be induced under carbon limiting conditions. Under Pi starvation, the repression of photosynthesis results in carbon starvation, which can trigger the catabolism of proteins and amino acids and release of free NH4+.

Pi Starvation Alters the Balance of Synthetic and Catabolic Carbon Metabolisms

Together with repression of photosynthetic genes upon Pi starvation, genes for synthetic pathway components for starch, fatty acids, and lipids were also repressed. As shown in Figure 5C, those include genes encoding ADP Glc pyrophosphorylase, ω-3 fatty acid desaturase, and β-ketoacyl-CoA synthase. In contrast, there is a general induction of genes encoding enzymes for the breakdown of fatty acids, lipids, and isoprenoids. Further, Pi starvation also favors carbon transport, because expression of genes for Suc synthesis, Glc transporters, and gluconeogenesis were induced.

There is also a general increase in the expression of genes involved in cell wall synthesis and degradation. Figure 5C lists several representative genes whose expression is induced by Pi starvation. For example, the synthesis of non-cellulose cell wall polysaccharides in the Golgi apparatus uses several nucleotide sugars as substrates. Beginning with formation of UDP-Glc and GDP-Glc, nucleotide sugar inter-conversion catalyzed by specific enzymes produces various nucleotide sugars. UDP-Glc 6-dehydrogenase and uridine diphosphate Glc epimerase are important enzymes in mediating sugar inter-conversion. The intracellular β-xylosidase is an exo-type enzyme that hydrolyzes xylo-oligosaccharides to produce the nucleotide sugar unit that can be used in cell wall biosynthesis. The genes for pectinesterase and pectate lyase, were also induced. The gene for xyloglucan endotransglycosylase (XET) was induced by 14-fold (Fig. 5C).

During glycolysis, the cellular reductant and ATP can be generated by two alternate pathways. In the first case, NADH is generated in a reaction catalyzed by glyceraldehyde 3-phosphate dehydrogenase, whereas ATP is generated in two reactions catalyzed by 3-phosphoglycerate kinase and pyruvate kinase (PK), respectively. In the second pathway, NADPH is produced through a bypass from glyceraldehyde 3-phosphate directly to 3-phosphoglycerate catalyzed by a non-phosphorylating glyceraldehyde 3-phosphate dehydrogenase (Gap), and ATP is produced by PK. Some genes involved in the second pathway were up-regulated under Pi starvation, whereas those in the first pathway were down-regulated (Fig. 5D).

In the cytosol, PEP can be converted to pyruvate in a reaction catalyzed by PK or to malate catalyzed by PEP carboxylase and malate dehydrogenase. Two genes for PEP carboxylase and malate dehydrogenase, respectively, were repressed in both leaves and roots during Pi starvation (Fig. 5D), whereas PK was up-regulated. Thus conversion of PEP to pyruvate, which can be funneled into the citric acid cycle, Ala-related amino acids synthesis, and ethanol and lactate synthesis, becomes the more prominent route during Pi starvation. Consistent with this observation, two genes for Ala glyoxylate aminotransferase, the key enzyme for synthesis of Ala family amino acids, were induced in both leaves and roots upon Pi starvation.

Pi Starvation Down-Regulating Protein Synthesis and Up-Regulating Protein Degradation

During the 72 h of Pi starvation, plant growth was arrested (Table I). As a metabolic adaptation to the growth arrest, down-regulation of protein synthesis and up-regulation of protein degradation could be expected. Actually, there are 42 ribosomal protein genes whose expression was down-regulated by 2- to 5-fold in leaves, and in most cases in roots as well. Those genes encode both chloroplast and cytosol ribosomal proteins (Fig. 6A). This general repression of protein synthesis machinery correlates well with down-regulation of eIF2, a key cytosol translation initiation factor, in both leaves and roots, especially in roots by more than 12-fold (Fig. 6A). Many cellular proteins contain disulfide bonds critical for their function. The proper formation of disulfide bonds is facilitated by an endoplasmic reticulum-localized enzyme named protein disulfide isomerase. Peptidyl prolyl isomerases catalyze the cis-trans isomerization of X-Pro bonds, which helps the protein reach its final folded confirmation more rapidly. All of these genes involved in protein biosynthesis and assembly were repressed after Pi starvation. It is of interest to note that down-regulation of protein synthesis and assembly (Fig. 6A) came somewhat later than those genes involved in carbon and nitrogen assimilation (Fig. 5, A and B), suggesting that the repression of protein biosynthesis may be triggered by the limitation of carbon and nitrogen resources caused by Pi starvation.

Figure 6.
Regulation of expression of genes involved in protein synthesis and protein degradation. The ratios for each gene listed are the median values of four to six replicates, and numbers in the parentheses are the sds. The four time points examined are ...

It has been suggested that during nutrient starvation, plants transport a series of cytosolic proteins into the vacuole, where the proteins are degraded into amino acids and carbohydrates that can be exported from senescing tissues. In the vacuole, several proteases such as Cys protease, Asp protease, metalloproteases, and Ser proteases degrade proteins to generate exportable amino acids for cell re-use. All of these protease genes presented in the microarray were induced in leaves and most of them in roots as well (Fig. 6B). It is believed that the protease ClpAP found in chloroplasts plays an important role in the degradation of chloroplast proteins. Two distinct ClpP genes were found to be induced upon Pi starvation. In the ubiquitin-proteasome degradation pathway, proteins destined for degradation are delivered to the proteasome after they are covalently modified by conjugation to a ubiquitin chain. Once ubiquitin tagged, the protein is delivered to the 26S proteasome for degradation. At least 18 genes for ubiquitin and the ubiquitin pathway genes were induced and expression patterns for two of them are present in Figure 6B.

TFs and Signal Transduction Components Involved in Pi Starvation-Regulated Gene Expression

Among the 333 TF genes represented in our microarray, expression for 111 of them (30%) was up- or down-regulated 2-fold or more upon Pi starvation. Among them, 74 were up-regulated, with 14 in both leaves and roots, 52 in leaves only, and eight in roots only. There were 36 TFs whose expression was down-regulated, with four in both leaves and roots, 26 in leaves only, and six in roots only. One gene was induced in roots but repressed in leaves. It is interesting to note that leaves and roots have largely nonoverlapping sets of TFs that are regulated during Pi starvation, implying distinct regulatory changes in leaves and roots. Figure 6A lists those TF genes whose expression is induced or repressed 3-fold or more during Pi starvation.

Among the 111 TFs, eight genes code for MYB proteins. The gene for MYB-related protein CCA1 was induced more than 6-fold in leaves and more than 3-fold in roots at the 24-h point (Fig. 7A). It is interesting to note that at least two reported TFs that have been described as eukaryotic regulators of Pi metabolism are MYB family proteins, including Psr1 from Chlamydomonas reinhardtii (Wykoff et al., 1999) and PHR1 from Arabidopsis (Rubio et al., 2001). Three expression patterns for three genes of the SCARECROW family were found with one induced in leaves, one repressed in leaves, and one induced in both leaves and roots. The gene for SCARECROW has been reported as a TF for root development (Helariutta et al., 2000). Four genes for AP2 domain proteins were induced in leaves, with one induced in roots as well. The AP2 domain is a highly conserved DNA-binding domain in EREBP group TFs, with many of them known to be responsive to environmental stresses (Weigel, 1995). Seven genes for homeobox proteins were regulated in leaves with five induced and two repressed. Five genes for zinc finger proteins were induced in leaves. Several other genes involved in mRNA, tRNA, and rRNA processing and synthesis were regulated in both leaves and roots, including ATP-dependent RNA helicase and RNA-binding proteins (Fig. 7A).

Figure 7.
Pi starvation alters expression of large numbers of genes encoding TFs and signal transduction proteins. A, The expression patterns of selected genes for transcription factors; B, The expression patterns of selected signal transduction genes. Only ...

Among the 435 signaling transduction protein genes in the microarray, the expression of 108 genes (24%) were regulated by 2-fold or more during Pi starvation. Similar to the TFs, there is a largely distinct set of genes that are regulated in leaves and roots. Among those 108 genes, 54 were induced, with 33 in leaves only, 12 in roots only, and nine in both leaves and roots. For the 57 down-regulated genes, 39 were in leaves only, 13 in roots only, and five in both leaves and roots. These genes belong to at least seven distinct classes. Figure 6B lists those genes whose expression is regulated by 3-fold or more, either through induction or repression. It is interesting to note that several induced genes in leaves were already up-regulated at the 24-h point (Fig. 7B), whereas the expression of several genes was already significantly altered at the 6- and 24-h time points. It appears that expression of many signal transduction-related genes responded to Pi starvation signals earlier than the metabolic genes described above.

The signal transduction-related genes belong to essentially all common groups. There are genes encoding an Rho GTPase of plant (ROP) mitogen-activated protein kinase (MAPK) cascade components, and Ca2+- or calmodulin-dependent protein kinase (CDPK) or CDPK-related protein (CDRK) kinases. One CDPK gene on the microarray was induced by 16-fold, the largest among signal transduction-related genes. CDPKs are believed to be involved in the response of plants to environmental stresses such as low temperatures, drought, pathogen defense response, and mechanical wounding (Martín and Busconi, 2001; Romeis et al., 2001; Chico et al., 2002). The gene for phosphatase 2C was induced in leaves but repressed in roots (Fig. 7B). It is also of interest to note that the β-GTP-binding protein was highly repressed, and this repression was evident already at the 6-h point. In Arabidopsis, it has been reported that there is only one gene for β-GTP-binding protein (AGB1; Lease et al., 2001).

Three genes for ethylene response were regulated by Pi starvation with three expression patterns: induced in leaves, induced in both leaves and roots, and repressed in leaves but induced in roots. Four genes for auxin response showed two expression patterns: induced and repressed in leaves. The different regulation patterns indicate that Pi starvation could regulate different hormone-dependent signaling pathways in distinct manners.


Pi Starvation-Triggered Cell Rescue System by Reduction of Carbon and Nitrogen Assimilation

Pi starvation responses in most organisms can be divided into two categories, the specific responses and the general responses (Wykoff et al., 1998). The specific responses promote efficient mobilization and acquisition of Pi from growth medium and intracellular stores. The general responses allow for long-term survival by coordinating the metabolism of cell to nutrient availability and growth potential. One high-affinity Pi transporter AtPT2 (At2g38940) was on the microarray, but no induction of it was detected. It has been reported that the gene could be induced after 5 d of Pi starvation (Muchhal et al., 1996), thus it is still possible that long time of starvation is needed for some of those specific response genes. It has been known that the PHO5 purple acid phosphatase gene in yeast is involved in its Pi starvation-regulated system (Ogawa et al., 2000). In Arabidopsis, one gene for purple acid phosphatase (AtACP5) was found to be Pi starvation induced (del Pozo et al., 1999). Three genes for purple acid phosphatase were on the microarray and one of them (At1g25230) was induced by more than 2-fold (Fig. 1). This is a new purple acid phosphatase gene responsive to Pi starvation. Those examples support a notion that Arabidopsis also has a specific response to promote efficient mobilization of usable Pi.

The general responses cover a variety of metabolic pathways and most of the 26% of the genes in the array exhibited differential expression in response to Pi starvation. The overall metabolic alteration indicates a Pi starvation-triggered cell rescue system due to the limitation of C and N assimilation. It is generally assumed that plants could not maintain normal metabolisms for more than 1 d under Pi starvation, which is supported by our data that after 1 d Pi starvation, the Pi concentration in root would be decreased. It has been reported that under Pi starvation, plants may alter the rate of the photosynthesis and photosynthetic product partitioning (Duff et al., 1989). The expression profiles of the genes related to the photosynthesis system and to general carbon metabolism in this study lend support for the earlier hypothesis (Fig. 5, C and D).

NADPH and reduced Fdx are required in carbon reduction and primary nitrate assimilation. In leaf cells, Pi starvation resulted in reduction of photosynthetic gene expression and photosynthetic activity. Thus Fdx and NADPH levels are declined. This would certainly affect carbon reduction and nitrate assimilation. Thus it is more energy conserving for the cell to reduce the expression of genes for carbon reduction pathways and nitrate assimilation enzymes. In roots, Fdx is reduced by FNR using NADPH as a reductant, which can be provided by reduced carbon imtermediates that are transported to or stored in the roots. A reduction of photosynthetic activity in leaves will most likely affect the NADPH and thus Fdx levels in roots. Therefore, it is expected that genes for Fdx-NiR and Fdx-GOGAT involved in nitrate assimilation would be repressed after 24 h of starvation in both leaves and roots. GDH will be induced by reduction of carbon and organic nitrogen in plants to provide NH4+ for the cycle of Gln to Glu. The gene for GDH was induced in both leaves and roots with more than 4-fold induction in leaves in this case (Fig. 5B).

New sources of energy and nitrogen re-assimilation are required to meet metabolic and transport demands under Pi starvation. Many genes were coordinately regulated in the catabolism of proteins and carbon compounds after 24 h of Pi starvation. The Pi starvation evidently presents a stress to both roots and leaves. This was not only supported by induced expression of stress response regulatory genes but was also supported by the strong induction of stress-inducible genes. For example, the expression of the SEN1 was most strikingly regulated by Pi starvation. It has reported that SEN1 can be strongly induced in Arabidopsis leaves subjected to senescence by 0.1 mm abscisic acid or 1 mm athephon treatment (Oh et al., 1996). Under our conditions, SEN1 was induced after the 24-h Pi starvation in both leaves and roots, up to 63-fold at the 48-h point in leaves (Fig. 1).

Positive and Negative Regulation of TFs during Pi Starvation May Coordinate Pi Starvation Responses in Plants

At least 22 Pi regulated genes have been identified through a whole-genome DNA microarray analysis in Brewer's yeast (Ogawa et al., 2000). The Pi regulatory system consists of at least five Pi-specific regulatory proteins, including two TFs, Pho2 and Pho4, the cyclin-cyclin-dependent protein kinase (CDK) complex (Pho80-Pho85), and the Pho81 CDK inhibitor (Oshima, 1997). Positive and negative regulators of the PHO5 gene (encoding acid phosphatase) and PHO84 (encoding high-affinity Pi transporter) in yeast have been identified (Lenburg and O'Shea, 1996). Under higher Pi concentrations, Pho80-Pho85 kinase phosphorylates Pho4, and the phosphorylated Pho4 is thus unable to activate target genes. Under low Pi concentrations, Pho81 inhibits the Pho80-Pho85 kinase activity, and the hypophosphorylated form of Pho4 together with Pho2 activates the target genes.

Negative regulation of Pi starvation has been hypothesized in plants (Mukatira et al., 2001), e.g. during Pi starvation TFs may be less abundant and/or incapable of binding with promoters due to post-translational modifications. Both induced and repressed TFs by Pi starvation were found in the present study, including two MYB genes induced and two MYB genes repressed by more than 3-fold (Fig. 7A). Our data support the presence of both positive and negative regulation of distinct TFs during Pi starvation. However, the response to Pi starvation in Arabidopsis is obviously much more extensive than in yeast. Several other types of TFs and translation regulators were regulated by Pi starvation, including one gene for a bHLH DNA-binding protein that was up-regulated by 3-fold (Fig. 7A). A Pi starvation-responsive TF with bHLH domain has also been reported in yeast (Berben et al., 1990).

Pi Starvation and Cell Signal Transduction

Our data revealed that the transition of 30-d-old Arabidopsis plants from normal Pi growth medium to Pi starvation resulted in a more than 20% decrease after 24 h in free available Pi in roots, but it was nearly recovered (to 94% of normal level) by the end of 3rd d. However, no change of free available Pi in leaves was observed during the same period (Table I). The decrease in free available Pi in roots after Pi starvation could potentially serve as the trigger for the alteration in genome expression. However, the maximal 20% reduction in free available Pi in roots would be too small a change to be a sensitive signal for a plant's response to Pi starvation. It is more likely that Arabidopsis roots are able to directly sense the Pi availability of the outside growth medium and are able to relay this perceived information to both roots and leaves for an coordinated response in genome expression.

At least two different signaling mechanisms maintaining Pi homeostasis in plants have been proposed, one operating at the cellular level and another involving multiple organs and most probably arising from the shoots (Raghothama, 1999). Although direct evidence for the involvement of Ca2+ in Pi starvation-induced signaling is still lacking, there is evidence of increased Ca2+-ATPase transcript accumulation in the roots of starved tomato plants (Muchhal et al., 1997). In addition, Ca2+-ATPase is involved in the signal transduction chain through alteration of Ca2+ concentration and CDPK. Three clones for CDPK and one clone for CDRK were present in the microarray, and all of them were induced. The expression patterns of one gene for CDPK and one gene for CDRK with a more than 3-fold induction, are shown in Figure 7B. CDPKs are believed to be involved in the response of plants to environmental stresses such as low temperatures, drought, pathogen defense response, and mechanical wounding (Martín and Busconi, 2001; Romeis et al., 2001; Chico et al., 2002), but their involvement in the Pi starvation-signaling pathway was not reported before. It has been reported that a Pi signal is transduced from an unknown Pi sensor to transcriptional activators of PHO5 via a CDK complex in yeast (Lau et al., 1998). Two clones for Arabidopsis CDK were present on the microarray, but their expression was not significantly regulated during Pi starvation.

Many signals are transduced by protein kinase cascades involving small GTP-binding proteins and MAPK cascades (MAPKKK/MAPKK/MAPK). Genes for these proteins involved in the transduction chain were up-regulated in leaves including MAPKKK and RAC-like GTP-binding protein. Down-regulation of the genes for calmodulin and MAPKK was also found. One gene for the heterotrimeric G protein β-subunit was repressed by more than 10-fold at the 48-h point (Fig. 7B). MAP kinases will phosphorylate Ser, Thr, and Tyr residues in target proteins. Four clones for Tyr phosphatase were found on the microarray and two of them were down-regulated in leaves and/or in roots by more than 4-fold.

Up- and down-regulation of genes for ethylene-responsive proteins, auxin-regulated proteins and gibberellin-responsive proteins were found (Fig. 7B). It is likely that ethylene and auxin have roles in altering root architecture and promoting root hair elongation in response to Pi starvation (Lynch and Brown, 1997). It has been known that MAP kinase cascades are involved in the signal transduction cascades triggered by gibberellin, auxin, and ethylene, but the details of the interaction between these hormone-responsive proteins and the MAP kinase is not clear.

Uptake studies in the pho2 Arabidopsis mutant, a hyperaccumulator of Pi in shoots, suggest that the induction of gene expression is initiated in response to changes in internal concentration of Pi in higher plants (Dong et al., 1998). Changes in cellular Pi concentrations are accompanied by the translocation of carbohydrate to the roots. Therefore it is possible that Pi starvation signals are transmitted to the root via certain carbon assimilates or changes in Pi fluxes. The up-regulation of the genes for Glc transporter, carbohydrate catabolism and cell wall biosynthesis, and down-regulation of the genes for a Suc transporter and biosynthesis in leaves were found in this case, which supports the functional relationship of carbon catabolism and translocation with Pi starvation signals.


Plant Growth Conditions

Arabidopsis ecotype Columbia seeds cold-treated in water at 4°C for 3 d were planted in vermiculite for seedling development. After 6 d, the seedlings were transferred to hydroponic growth conditions. Plants were grown in growth chambers (AR-75L, Percival Scientific, Boone, IA) under the light of a photosynthetic photon flux density of 150 μmol photons m2 s1 in a 14-h-light/10-h-dark photoperiod. The day/night temperature and humidity were controlled at 22°C/20°C and 80%. After 30 d, the plants grown hydroponically in nutrient solution containing 5 mm KNO3, 2.5 mm KH2PO4, 2 mm MgSO4, 2 mm Ca(NO3)2, 50 μm Fe-EDTA, 70 μm H3BO4, 14 μm MnCl2, 0.5 μm CuSO4, 1 μm ZnSO4, 0.2 μm Na2MoO4, and 0.01 μm CoCl2, pH 5.7 (modified from Chen et al., 2000) were transferred into a nutrient solution without Pi (KH2PO4 was replaced with K2SO4 [Muchhal et al., 1996]) or a fresh control (normal) medium. The roots and leaves were harvested respectively after 6, 24, 48, 72 h following the growth medium transition and frozen in liquid nitrogen for RNA isolation.

Measurement of Pi Concentration in Plant Tissues

Ten representative plant rosettes and roots were harvested at defined time points after growth medium transition for Pi concentration measurement. After drying in an 80°C oven for 3 d, the Pi concentration of the rosette of each sample was analyzed by the method of phosphomolybdenum blue reaction using a Spectroquant NOVA60 spectrophotometer and Spectro-quant Phosphat-Test Kit (Merck, Darmstadt, Germany). The total free available Pi concentration of each plant was calculated based on the total dried biomass and the Pi content.

RNA Preparation and Blot Analysis

Total RNA was extracted from the roots and leaves of Arabidopsis plants by using Trizol D0410 reagent based on recommended procedure (Invitrogen, Carlsbad, CA). About 50 to 100 mg of tissue was homogenized using a porcelain mortar in 1 mL of Trizol reagent, and the isolation procedure was essentially as previously described (Ma et al., 2001). RNA was resuspended in RNase-free water and stored at –20°C. The total RNA used in microarray analysis was also used for RNA gel-blot hybridization according to a previously used protocol (Deng et al., 1992). The EST clone inserts were used as probes for RNA gel-blot hybridization.

Sequence Analysis

All of the EST clones corresponding to specific genes resented in the figures or text have been sequenced to confirmed the identity using a MegaBACE 1000 sequencer. Sequence homologies were examined with the GenBank/EMBL database using the BLAST program.

Fluorescent Labeling of Probe, Hybridization, Washing, and Scanning

Probe labeling with Cy-3- or Cy-5-conjugated deoxy UTP (Amersham Pharmacia Biotech, Piscataway, NJ), the purification of labeled cDNA, hybridization to the Arabidopsis array, and washing and scanning of hybridized array slides were as described previously (Ma et al., 2001).

Data Analysis

Hybridization signals from the microarray were quantified using Axon GenePix Pro 3 image analysis software. The expression ratios were measured using the GenePix Pro 3 median of ratio method, and they were normalized using the corresponding GenePix default normalization factor. The program GPMERGE was used to merge the replicated GenePix Pro 3 output data files (gpr files; http://bioinformatics.med.yale.edu/software.html). With this program, four to six replicated data sets from each experiment were pooled. A number of quality control procedures were conducted before data points from the replicates of two or three independent biological sample sets were averaged as previously described (Ma et al., 2001, 2002, 2003). A hierarchical clustering analysis was performed as described by Eisen et al. (1998).

Supplementary Material

Supplemental Data:


We thank Jessica Habashi, James Sullivan, Vicente Rubio, and Magnus Holm for reading and commenting on this manuscript. We are grateful to the Yale DNA microarray laboratory of the Keck Biological Resource Center for the production of the microarray used in this study (http://info.med.yale.edu/wmkeck/dna_arrays.htm).


Article, publication date, and citation information can be found at www.plantphysiol.org/cgi/doi/10.1104/pp.103.021022.

1This work was supported by the National Institutes of Health (grant no. GM–47850 to X.W.D.), by the National Natural Science Foundation of China (grant no. 39725002), and by the National Key Basic Research Special Foundation of China (grant no. G199911700). L.M. is a long-term postdoctoral fellow of the Human Frontier Science Program.

[w]The online version of this article contains Web-only data. The supplemental material is available at http://www.plantphysiol.org.


  • Abel S, Nurnberger T, Ahnert V, Krauss GJ, Glund K (2000) Induction of an extracellular cyclic nucleotide phosphodiesterase as an accessory ribonucleolytic activity during phosphate starvation of cultured tomato cells. Plant Physiol 122: 543–552 [PMC free article] [PubMed]
  • Bates TR, Lynch JP (1996) Stimulation of root hair elongation in Arabidopsis thaliana by low phosphorus availability. Plant Cell Environ 21: 529–538
  • Berben G, Legrain M, Gilliquet V, Hilger F (1990) The yeast regulatory gene PHO4 encodes a helix-loop-helix motif. Yeast 6: 451–454 [PubMed]
  • Burleigh SH, Harrison MJ (1999) The down-regulation of Mt4-like genes by phosphate fertilization occurs systemically and involves phosphate translocation to the shoots. Plant Physiol 119: 241–248 [PMC free article] [PubMed]
  • Chen DL, Delatorre CA, Bakker A, Abel S (2000) Conditional identification of phosphate-starvation-response mutants in Arabidopsis thaliana. Planta 211: 13–22 [PubMed]
  • Chico JM, Raíces M, Téllez-Iñón MT, Ulloa RM (2002) A calcium-dependent protein kinase is systemically induced upon wounding in tomato plants. Plant Physiol 128: 256–270 [PMC free article] [PubMed]
  • del Pozo JC, Allona I, Rubio V, Leyva A, de la, Pena A, Aragoncillo C, Paz-Ares J (1999) A type 5 acid phosphatase gene from Arabidopsis thaliana is induced by phosphate starvation and by some other types of phosphate mobilising/oxidative stress conditions. Plant J 19: 579–589 [PubMed]
  • Deng XW, Matsui M, Wei N, Wagner D, Chu AM, Feldmann KA, Quail PH (1992) COP1, an Arabidopsis regulatory gene, encodes a protein with both a zinc-binding motif and a G beta homologous domain. Cell 71: 791–801 [PubMed]
  • Dong B, Rengel Z, Delhaize E (1998) Uptake and translocation of phosphate by pho2 mutant and wild-type seedlings of Arabidopsis thaliana. Planta 205: 251–256 [PubMed]
  • Duff SMG, Moorhead GBG, Lefebvre DD, Plaxton WC (1989) Phosphate starvation inducible: bypasses of adenylate and phosphate dependent glycolytic enzymes in Brassica nigra suspension cells. Plant Physiol 90: 1275–1278 [PMC free article] [PubMed]
  • Duff SMG, Sarath G, Plaxton WC (1994) The role of acid phosphatase in plant phosphorus metabolism. Physiol Planta 90: 791–800
  • Eisen MB, Spellman PT, Brown PO, Botstein D (1998) Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 95: 14863–14868 [PMC free article] [PubMed]
  • Goldstein AH, Baertlein DA, Danon A (1989) Phosphate starvation stress an experimental system for molecular analysis. Plant Mol Biol Rep 7: 7–16
  • Green PJ (1994) The ribonucleases of higher plants. Annu Rev Plant Physiol Plant Mol Biol 45: 421–445
  • Harmer SL, Hogenesch JB, Straume M, Chang HS, Han B, Zhu T, Wang X, Kreps JA, Kay SA (2000) Orchestrated transcription of key pathways in Arabidopsis by the circadian clock. Science 290: 2110–2113 [PubMed]
  • Helariutta Y, Fukaki H, Wysocka-Diller J, Nakajima K, Jung J, Sena G, Hauser MT, Benfey PN (2000) The SHORT-ROOT gene controls radial patterning of the Arabidopsis root through radial signaling. Cell 101: 555–567 [PubMed]
  • Kawasaki S, Borchert C, Deyholos M, Wang H, Brazille S, Kawai K, Galbraith D, Bohnert H (2001) Gene expression profiles during the initial phase of salt stress in rice. Plant Cell 13: 889–905 [PMC free article] [PubMed]
  • Lau WW, Schneider KR, O'Shea EK (1998) A genetic study of signaling processes for repression of PHO5 transcription in Saccharomyces cerevisiae. Genetics 150: 1349–1359 [PMC free article] [PubMed]
  • Lease KA, Wen J, Li J, Doke JT, Liscum E, Walker JCA (2001) A mutant Arabidopsis heterotrimeric g-protein beta subunit affects leaf, flower, and fruit development. Plant Cell 13: 2631–2641 [PMC free article] [PubMed]
  • Lenburg ME, O'Shea EK (1996) Signaling phosphate starvation. Trends Biochem Sci 21: 383–387 [PubMed]
  • Liu C, Muchhal US, Raghothama KG (1997) Differential expression of TPSI1, a phosphate starvation-induced gene in tomato. Plant Mol Biol 33: 867–874 [PubMed]
  • Lynch J, Brown KM (1997) Ethylene and plant responses to nutritional stress. Physiol Plant 100: 613–619
  • Ma L, Gao Y, Qu L, Chen Z, Li J, Zhao H, Deng XW (2002) Genomic evidence for COP1 as repressor of light regulated gene expression and development in Arabidopsis. Plant Cell 14: 2383–2398 [PMC free article] [PubMed]
  • Ma L, Li J, Qu L, Hager J, Chen Z, Zhao H, Deng XW (2001) Light control of Arabidopsis development entails coordinated regulation of genome expression and cellular pathways. Plant Cell 13: 2589–2607 [PMC free article] [PubMed]
  • Ma L, Zhao HY, Deng XW (2003) Analysis of the mutational effects of the COP/DET/FUS loci on genome expression profiles reveals their overlapping yet not identical roles in regulating Arabidopsis seedling development. Development 130: 969–981 [PubMed]
  • Malboobi MA, Hannoufa A, Tremblay L, Lefebvre DD (1998) Towards an understanding of gene regulation during the phosphate starvation response. In JP Lynch, J Deiman, eds, Phosphorous in Plant Biology: Regulatory Roles in Molecular, Cellular, Organismic, and Ecosystem Processes. The American Society of Plant Physiologists, Rockville, MD, pp 215–226
  • Maleck K, Levine A, Eulgem T, Morgan A, Schmid J, Lawton KA, Dangl JL, Dietrich RA (2000) The transcriptome of Arabidopsis thaliana during systemic acquired resistance. Nat Genet 26: 403–410 [PubMed]
  • Marschner A (1995) Mineral Nutrition of Higher Plants. Academic Press, San Diego
  • Martín ML, Busconi L (2001) A rice membrane-bound calcium-dependent protein kinase is activated in response to low temperature. Plant Physiol 125: 1442–1449 [PMC free article] [PubMed]
  • Muchhal JM, Raghothama KG (1996) Phosphate transporters from the higher plant Arabidopsis thaliana (ion uptake/yeast complementation). Plant Biol 93: 10519–10523 [PMC free article] [PubMed]
  • Muchhal US, Liu C, Raghothama KG (1997) Calcium-ATPase is differentially expressed in phosphate starved roots of tomato. Physiol Planta 101: 540–544
  • Muchhal US, Pardo JM, Raghothama KG (1996) Phosphate transporters from the higher plant Arabidopsis thaliana. Proc Natl Acad Sci USA 93: 10519–10523 [PMC free article] [PubMed]
  • Mukatira UT, Liu C, Varadarajan DK, Raghothama KG (2001) Negative regulation of phosphate starvation-induced genes. Plant Physiol 127: 1854–1862 [PMC free article] [PubMed]
  • Ogawa N, DeRisi J, Brown PO (2000) New components of a system for phosphate accumulation and polyphosphate metabolism in Saccharomyces cerevisiae revealed by genomic expression analysis. Mol Biol Cell 11: 4309–4321 [PMC free article] [PubMed]
  • Oh SA, Lee SY, Chung IK, Lee CH, Nam HG (1996) A senescence associated gene of Arabidopsis thaliana is distinctively regulated during natural and artificially induced leaf senescence. Plant Mol Biol 30: 739–754 [PubMed]
  • Oshima Y (1997) The phosphatase system in Saccharomyces cerevisiae. Genes Genet Syst 72: 323–334 [PubMed]
  • Raghothama KG (1999) Phosphate acquisition. Annu Rev Plant Physiol Plant Mol Biol 50: 665–693 [PubMed]
  • Raghothama KG (2000) Phosphate transport and signaling. Curr Opin Plant Biol 3: 182–187 [PubMed]
  • Romeis T, Piedras P, Jonathan DGJ (2001) Resistance gene-dependent activation of a calcium-dependent protein kinase in the plant defense response. Plant Cell 12: 803–816 [PMC free article] [PubMed]
  • Rubio V, Francisco L, Roberto S, Ana C, Martín JI, Antonio L, Javier Paz-Ares J (2001) A conserved MYB transcription factor involved in phosphate starvation signaling both in vascular plants and in unicellular. Gene Dev 15: 2122–2133 [PMC free article] [PubMed]
  • Seki M, Narusaka M, Abe H, Kasuko M, Yamaguchi-Shinozaki K, Carninci P, Hayashizaki Y, Shinozaki K (2001) Monitoring the expression pattern of 1,300 Arabidopsis genes under drought and cold stresses by using a full-length cDNA microarray. Plant Cell 13: 61–72 [PMC free article] [PubMed]
  • Tepperman JM, Zhu T, Chang HS, Wang X, Quail PH (2001) Multiple transcription-factor genes are early targets of phytochrome A signaling. Proc Natl Acad Sci USA 98: 9437–9442 [PMC free article] [PubMed]
  • Theodorou ME, Plaxto WC (1993) Metabolic adaptations of plant respiration to nutritional phosphate deprivation. Plant Physiol 101: 339–344 [PMC free article] [PubMed]
  • Torriani A (1990) From cell membrane to nucleotides: the phosphate regulation in Escherichia coli. BioEssays 12: 371–376 [PubMed]
  • Trull MC, Deikman J (1998) An Arabidopsis mutant missing one acid phosphatase isoform. Planta 206: 544–550 [PubMed]
  • Wang H, Ma LG, Habashi J, Li JM, Zhao HY, Deng XW (2002) Analysis of far-red light regulated genome expression profiles of phytochrome A pathway mutants in Arabidopsis. Plant J 32: 723–733 [PubMed]
  • Wang R, Guegler K, Labrie ST, Crawford NM (2000) Genomic analysis of a nutrient response in Arabidopsis reveals diverse expression patterns and novel metabolic and potential regulatory genes induced by nitrate. Plant Cell 12: 1491–1509 [PMC free article] [PubMed]
  • Weigel D (1995) The APETALATA2 domain is related to a novel type of DNA binding domain. Plant Cell 7: 388–389 [PMC free article] [PubMed]
  • Wykoff DD, Davies JP, Melis A, Grossman AR (1998) The regulation of photosynthetic electron transport during nutrient deprivation in Chlamydomonas reinhardtii. Plant Physiol 117: 129–139 [PMC free article] [PubMed]
  • Wykoff DD, Grossman AR, Weeks DP, Usuda H, Shimogawara K (1999) Psr1, a nuclear localized protein that regulates phosphorus metabolism in Chlamydomonas. Proc Natl Acad Sci USA 96: 15336–15341 [PMC free article] [PubMed]

Articles from Plant Physiology are provided here courtesy of American Society of Plant Biologists
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem Compound links
  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...