Introduction

Iron (Fe) is an essential nutrient for plant growth and crop productivity. Although it is abundant in soil, it is not very soluble and not easily available to plants under aerobic conditions in the physiological pH range. Higher plants have evolved two major strategies for Fe acquisition (Marschner et al. 1986). Graminaceous plants secrete mugineic acid family phytosiderophores (MAs) from their roots to chelate and solubilize Fe in the rhizosphere (Takagi 1976); Fe(III)–MAs complexes are then absorbed through Fe(III)–MAs complex transporters in the cell membrane (Curie et al. 2001; Murata et al. 2006). The production and secretion of MAs increases markedly in response to Fe deficiency (Takagi et al. 1984). The biosynthetic pathway to produce MAs from l-methionine has been clarified through intensive physiological and biochemical studies (Mori and Nishizawa 1987; Shojima et al. 1990; Ma et al. 1999; Bashir et al. 2006). The genes encoding enzymes that participate in the MAs biosynthetic pathway are highly induced during Fe deficiency (Higuchi et al. 1999; Takahashi et al. 1999; Nakanishi et al. 2000; Kobayashi et al. 2001). The genes of the methionine cycle are also upregulated in Fe-deficient rice and barley plants (Kobayashi et al. 2005; Fig. 1). Barley secretes a larger amount of MAs, especially mugineic acid and 3-epihydroxymugineic acid, than other graminaceous plants, and is therefore more tolerant of low Fe availability. The secretion of MAs by barley roots follows a distinct diurnal rhythm (Takagi et al. 1984; Marschner et al. 1986). A peak in secretion occurs just after the onset of illumination and ceases within 2–3 h. This diurnal secretion may help prevent the microbial degradation of MAs in the rhizosphere (Römheld and Marschner 1990).

Fig. 1
figure 1

The methionine cycle and the MAs biosynthetic pathway in graminaceous plants. All of the genes except SAMS were induced by Fe deficiency and showed no diurnal expression rhythm. SAMS S-adenosyl-Met synthase; NAS nicotianamine synthase; NAAT nicotianamine aminotransferase; DMAS 2′-deoxymugineic acid synthase; IDS2 and IDS3 deoxygenases catalyzing the hydroxylation of MAs; DMA 2′-deoxymugineic acid; MA mugineic acid; epiHDMA 3-epihydroxy-2′-deoxymugineic acid; epiHMA 3-epihydroxymugineic acid; MTN methylthioadenosine/S-adenosyl homocysteine nucleosidase; MTK methylthioribose kinase; IDI2 eukaryotic initiation factor 2B-like methylthioribose-1-phosphate isomerase; RPI putative ribose-5-phosphate isomerase; DEP methylthioribulose-1-phosphate dehydratase-enolase-phosphatase; IDI1 2-keto-methylthiobutyric-acid-forming enzyme; IDI4 putative aminotransferase catalyzing the synthesis of Met from 2-keto-methylthiobutyric acid; FDH formate dehydrogenase; APT Ade phosphoribosyltransferase; PRPP phosphoribosyl pyrophosphate. Hordeum vulgare L. cv Ehimehadaka no. 1 secretes three types of MAs: DMA, MA, and epiHMA

Full size image

To dissect the molecular mechanisms of MAs secretion, microarray analysis was performed in barley plants grown under Fe-deficient conditions. The first investigation of comprehensive gene expression in Fe-deficient plants was performed in barley roots using a rice cDNA array containing 8,987 (9 K array) rice expressed sequence tags (Negishi et al. 2002). This experiment showed that many genes involved in MAs biosynthesis and the methionine cycle are induced by Fe deficiency in barley roots. The rice 9 K array has also been used to study expression in rice roots, and it has been shown that the genes involved in MAs biosynthesis and the methionine cycle are also upregulated in Fe-deficient rice roots (Kobayashi et al. 2005). Several researchers have used microarray analysis to examine gene expression in non-graminaceous plants subjected to Fe deficiency (Thimm et al. 2001; Wang et al. 2002; Wintz et al. 2003). Several regulatory, signaling, and transporter genes are induced by Fe deficiency in tomato (Lycopersicon esculentum L.) roots (Wang et al. 2002), while in Arabidopsis, several genes involved in respiration and metal transport are induced in the roots and shoots (Thimm et al. 2001; Wintz et al. 2003).

Recently, microarray chips containing barley genes have become commercially available. In this study, we conducted a time-course microarray analysis using a barley DNA chip to elucidate changes in expression that occur over a 24-h period in Fe-deficient barley roots.

Materials and methods

Plant material

Seeds of barley (Hordeum vulgare L. cv Ehimehadaka no. 1) were germinated, and the seedlings were grown hydroponically as described previously (Higuchi et al. 1996) under a daily light regime of 14 h of light (22°C) and 10 h of dark (17°C). To induce Fe deficiency, plants were transferred to culture solutions lacking Fe 2 weeks after germination. After 2 weeks of Fe-deficiency treatment (i.e., when symptoms of severe Fe deficiency appeared), the roots of the Fe-sufficient and -deficient plants were harvested every 3 h from the onset of illumination. Three plants were harvested at each time point and stored at −80°C until RNA and endogenous MAs extraction. We duplicated the plant preparation and harvest, and different plant materials were used for microarray analysis and Northern blotting.

Microarray analysis

Total RNA was extracted from Fe-deficient roots harvested 0, 6, 12, and 18 h after the onset of illumination and from Fe-sufficient roots harvested 0 and 12 h after the onset of illumination. The RNAs were labeled, fragmented using a One Cycle Target Labeling Kit (Affymetrix, Santa Clara, California, USA), and used for microarray analysis (single replicates). The prepared probes were hybridized to Affymetrix GeneChip® Barley Genome Arrays, and the array chips were washed, stained, and scanned on a GeneChip Scanner 3000 (Affymetrix) according to the manufacturer’s recommendations. The six data sets were analyzed using Affymetrix GCOS software. Each data set was linearly normalized using a scaling method to a mean signal intensity of 500 units. Those genes with a signal intensity of <200 were excluded from the subsequent analysis. To select for Fe deficiency-inducible genes, gene expression in the Fe-deficient roots at the start of illumination (0 h) and after 12 h of illumination (12 h) was compared with that in the Fe-sufficient roots at 0 and 12 h, respectively. Those genes with an induction ratio ≥2.0 in the Fe-deficient roots at 0 or 12 h were recognized as Fe deficiency-inducible genes. For the time-course analysis of gene expression in the Fe-deficient roots, the microarray data for the Fe-deficient roots at 6, 12, and 18 h were compared with those for the Fe-deficient roots at 0 h. Those genes showing diurnal changes similar to the pattern of MAs secretion from Fe-deficient barley roots were identified among the Fe deficiency-inducible genes. Changes in gene expression were determined using GCOS software. Genes whose expression was lower at two time points or more than at 0 h were designated as diurnally regulated.

Northern blot analysis of HvYS1 and putative Fe–MAs complex transporter genes

Northern blot analysis was conducted with 32P-labeled probes specific for HvYS1 (AB214183), two putative Fe–MAs complex transporter genes (UniGenes 18532 and 12454; HarvEST ver. 1.50, http://www.harvest.ucr.edu), and genes encoding enzymes that participate in the methionine cycle (IDI1, IDI2, and IDI4; Yamaguchi et al. 2000a, b; Kobayashi et al. 2005). The primers were designed based on the 3′-UTR regions of HvYS1 and two other genes: HvYS1, 5′-AATAGGCATACTTGATTAAAGTTAT-3′ and 5′-GTGTTTTGTTTTACTGGAGATGCTT-3′; UniGene 18532, 5′-TACTTCAAGCCATCTCTTGCCTAGGGCGTAA-3′ and 5′-AACAGAGCTGCTGCTATTTTTGCAA-3′; and UniGene 12454, 5′-ACTAGATGAGTGTCATGCCTGATA-3′ and 5′-ATCTTCACACCCAGTCTTATACTCAT-3′. Using these primers, fragments were amplified by polymerase chain reaction (PCR) with a cDNA library synthesized from the mRNAs of Fe-deficient barley roots. The fragments were ligated into a vector (pCR®-Blunt II; Invitrogen, Carlsbad, California, USA) and sequenced. The vector was then digested with EcoRI and each fragment was purified. Specific probes were synthesized from each purified fragment using a Random Primer DNA Labeling Kit (Takara, Tokyo, Japan). For IDI1, IDI2, and IDI4, excised ORFs were used as templates to synthesize specific probes. A CsCl-gradient ultracentrifugation method (Glisin et al. 1974) was used to extract total RNA from barley roots harvested every 3 h. The materials were prepared at different times and were not the same plant materials used for our microarray analysis. Total RNA (10 μg) was separated, blotted, and hybridized with the probes as described previously (Higuchi et al. 1999). Radioactivity was detected using an FLA-5000 image analyzer (Fuji Film, Tokyo, Japan).

Extraction and detection of endogenous MAs

Endogenous MAs were purified from the time-course samples of Fe-deficient and -sufficient barley roots described above. Endogenous MAs were extracted as described previously (Higuchi et al. 2001). The extracted MAs were concentrated and detected by HPLC using a method developed for detecting MAs secreted from roots (Mori and Nishizawa 1987). Hordeum vulgare L. cv Ehimehadaka no. 1 secretes three types of MAs: DMA, MA, and epiHMA.

Results

Fe deficiency-inducible genes in barley roots

A barley GeneChip was used to analyze the gene expression profile in barley roots under Fe-deficiency stress. At 0 or 12 h, 82 genes were identified as induced by Fe deficiency (Table 1). Many of these genes had already been reported to be Fe deficiency-inducible, and almost all of them were involved in MAs biosynthesis. Seven genes encoding isozymes of nicotianamine (NA) synthase (NAS), which catalyzes the first step in MAs biosynthesis, were induced under Fe deficiency. Two NA aminotransferase genes (HvNAAT-A and HvNAAT-B), a deoxymugineic acid synthase gene (HvDMAS1), and IDS2 and IDS3 were also found to be inducible by Fe deficiency. In addition to genes involved in MAs biosynthesis, Fe deficiency also upregulated genes encoding all enzymes of the methionine cycle (Fig. 1) with the exception of the SAM synthase gene (SAMS). Because MAs are synthesized from S-adenosyl-l-methionine (SAM), methionine recycling is essential for MAs synthesis. Furthermore, genes encoding sulfate transporters and enzymes involved in the methionine biosynthetic pathway were induced. In total, 21 of the 82 Fe deficiency-inducible genes were related to MAs biosynthesis. In addition to these genes and those involved in sulfur assimilation, the Fe(III)-MAs transporter gene HvYS1 was found to be upregulated in Fe-deficient barley roots. HvYS1 transports Fe–MAs complexes from the rhizosphere to barley roots, and the expression of HvYS1 is increased in Fe-deficient barley roots (Murata et al. 2006). Two genes homologous to HvYS1 were also induced by Fe deficiency. Because the genes expressed in the Fe-sufficient roots changed between 0 and 12 h, the induction ratio of several genes also changed. In particular, the expression of NAS2-7 in the Fe-sufficient plants decreased at 12 h compared to at 0 h; therefore, the induction of these genes by Fe deficiency was significantly increased at 12 h (data not shown).

Table 1 List of genes upregulated in Fe-deficient barley roots 0 and 12 h after the start of illumination
Full size table

Fe deficiency-inducible genes showing diurnal rhythms

To examine the diurnal rhythm in the expression of Fe deficiency-inducible genes under Fe-deficient conditions, additional microarray analyses were performed. Expression in the Fe-deficient roots at sunrise (0 h) was compared with that in the Fe-deficient roots at 6, 12, and 18 h after sunrise. The expression of 14 of the Fe deficiency-inducible genes shown in Table 1 showed diurnal changes in the Fe-deficient roots, similar to the MAs secretion pattern (Table 2). However, the levels of transcripts encoding enzymes involved in MAs biosynthesis and the methionine cycle did not change diurnally. For example, the transcript levels of IDI1, IDI2, and IDI4, which participate in the methionine cycle, were confirmed by Northern blot analysis (Figs. 1, 2a). The expression of these genes showed no diurnal changes, remaining high in the Fe-deficient roots throughout the 24-h period monitored. These expression patterns were in line with our microarray data.

Table 2 Diurnally regulated Fe deficiency-inducible genes in Fe-deficient barley roots
Full size table
Fig. 2
figure 2

Northern blot analysis reveals diurnal changes in the transcript levels of barley genes encoding iron-phytosiderophore transporters and enzymes participating in the methionine cycle. UniGenes 12454 and 18532 encode putative transporter genes. IDI1, IDI2, and IDI4 encode enzymes that participate in the methionine cycle. Time (0) means the start of illumination

Full size image

In contrast, the expression of HvYS1 and a putative metal–MAs complex transporter gene (UniGene 12454, which shows similarity to AtYSL6) showed diurnal rhythms, and both genes were downregulated at 6 and 12 h (Table 2). Northern blot analysis confirmed that the levels of these transcripts increased under Fe deficiency (Fig. 2b). The expression of HvYS1 and UniGene 12454 was high at the start of illumination, gradually decreased during illumination, and increased again at the start of darkness. However, the transcript levels of another gene that encodes a putative metal–MAs complex transporter (UniGene 18532, homologous to AtYSL2) did not change during the day.

Degree of endogenous MAs change during the day in Fe-deficient barley roots

The amount of endogenous MAs in the roots was increased in the Fe-deficient barley, reaching 1,000 times the level in the Fe-sufficient roots at the start of illumination (Fig. 3). Furthermore, the amount of endogenous MAs showed a diurnal rhythm in the Fe-deficient roots, peaking at the start of illumination then decreasing to its lowest level 6 h later. The amount of endogenous MAs then gradually increased again until the start of illumination.

Fig. 3
figure 3

Diurnal changes in the amount of endogenous MAs in Fe-deficient and -sufficient barley roots. MAs include DMA, MA, and epiHMA. Time (0) means the start of illumination. Error bars represent the standard deviation (n = 3)

Full size image

Discussion

We used a DNA microarray designed from barley genes to examine the gene expression profile of Fe-deficient barley roots over 24 h. Among the 25,500 genes on the array, 82 were identified as Fe deficiency-inducible. Twenty of these genes had already been reported to be inducible by Fe deficiency, indicating the reliability of the plant material and the method for expression profiling in Fe-deficient barley plants.

The methionine cycle, which recycles methionine, is required for MAs biosynthesis in Fe-deficient barley roots (Ma et al. 1995; Fig. 1), and a number of genes related to the methionine cycle are Fe deficiency-inducible in barley and rice roots (Negishi et al. 2002; Kobayashi et al. 2005). Our microarray analysis showed that all of the genes involved in the methionine cycle, with the exception of SAMS, were induced by Fe deficiency in barley roots. The intensity of the SAMS signal was high, and its level of expression was among the top 50 on the array. Because SAM plays various roles in plants, SAMS expression is likely to occur at high levels. Therefore, the induction ratio of SAMS during Fe deficiency might be lower than that of other genes involved in the methionine cycle. In addition to the genes involved in this process, genes involved in MAs biosynthesis were highly upregulated in the Fe-deficient barley roots.

The transport of sulfate into the cell is a major regulatory step in sulfur metabolism (Vauclare et al. 2002). There is clear evidence that the levels of transcripts encoding transporters involved in the initial steps of sulfur uptake from the rhizosphere and in the vascular transport of sulfate are controlled by the sulfur supply (Buchner et al. 2004). Genes encoding a high-affinity sulfate transporter and a putative low-affinity sulfate transporter were upregulated in the Fe-deficient barley roots (Table 1). Sulfate transporters have been classified into groups based on their localization, kinetic properties, and sequence similarity (Grossman and Takahashi 2001). High-affinity transporters are expressed primarily in the roots of plants undergoing sulfate starvation; they contribute to sulfate uptake but not to the transport of sulfur to the shoots (Shibagaki et al. 2002; Yoshimoto et al. 2002). Low-affinity transporters are expressed in vascular tissues, and may be involved in the transport of sulfate within the plant. The induction of high- and low-affinity sulfate transporter genes during Fe deficiency suggests that sulfur uptake and translocation increase to meet the increased demand for methionine to support MAs production. In addition, several genes involved in the methionine biosynthetic pathway were upregulated by Fe deficiency. An increase in sulfur assimilation is crucial for the response of barley roots to Fe deficiency. As MAs constitute 1–2% of the root dry weight in Fe-deficient barley, increased methionine synthesis seems to be essential for the proper functioning of the methionine cycle under Fe-deficient conditions.

One-fourth of the genes upregulated by Fe deficiency were related to MAs biosynthesis, indicating that MAs biosynthesis is a major event in barley roots under Fe deficiency. Our time-course microarray analysis of Fe-deficient roots revealed that the level of transcription of all of the genes involved in the methionine cycle, MAs biosynthesis, and sulfate transport was continuously high over a 24-h period. The expression of IDI1, IDI2, and IDI4, which participate in the methionine cycle, was confirmed by Northern blot analysis. In accordance with our microarray data, the expression of these genes was highly induced by Fe deficiency but did not change diurnally (Fig. 2a). We previously reported that NAS and NAAT activity is induced by Fe deficiency, and that the activity of these proteins is unchanged during the day (Kanazawa et al. 1995). NAS and NAAT transcription in the Fe-deficient barley roots was strongly correlated with NAS and NAAT activity. This suggests that MAs biosynthesis occurs throughout the day, whereas Fe-deficient barley roots secrete MAs diurnally (Takagi et al. 1984; Marschner et al. 1986).

The change in the endogenous amount of MAs in Fe-deficient barley is opposite to the pattern of MAs secretion. MAs secretion peaks 2–3 h after sunrise (Takagi et al. 1984; Marschner et al. 1986). In contrast, the levels of endogenous MAs were highest at the start of illumination and then decreased, reaching their lowest levels 6 h after sunrise. The time points corresponding to the highest and lowest levels of endogenous MAs fell before and after MAs secretion, respectively. Thus, the amount of endogenous MAs increased gradually throughout the day (Fig. 3). Our data indicate that MAs are biosynthesized continuously throughout the day and accumulate in the roots until secretion.

In contrast to the expression of genes involved in MAs biosynthesis, the transcript levels of HvYS1 and UniGene 12454, a putative metal–MAs complex transporter gene, were not only induced by Fe deficiency but also showed diurnal changes (Table 2, Fig. 2). Their expression patterns paralleled the secretion of MAs from barley roots. Peak expression was observed at the start of illumination (0 h), after which it decreased during the day and increased during the night. HvYS1 has been reported to be responsible for the uptake of Fe–MAs complexes by barley roots because of its localization and substrate specificity (Murata et al. 2006). The change in the expression of HvYS1 suggests that the uptake of Fe–MAs complexes occurs mainly during the period of MAs secretion. Diurnal MAs secretion may help avoid MAs degradation by microorganisms (Römheld and Marschner 1990), and the high transport activity of Fe–MAs complexes may be required just after MAs secretion.

Recent studies have shown that plants have many YS1-like genes, each with a different expression pattern and function in plant cells (Curie et al. 2001; DiDonato et al. 2004; Koike et al. 2004; Jean et al. 2005; Murata et al. 2006). Our microarray analysis revealed that genes encoding two candidate metal–MAs complex transporters, UniGenes 12454 and 18532, are upregulated in Fe-deficient barley roots, like HvYS1. Barley roots secrete mainly MA and epiHMA under Fe-deficient conditions. HvYS1 transports Fe–MAs complexes but not other metal–MAs complexes (Murata et al. 2006). However, the transporters that take up Fe-epiHMA or other metal–MAs complexes have not yet been reported. As the expression pattern of UniGene 12454 was similar to that of HvYS1, 12454 is a promising candidate for the uptake of other metal–MAs complexes from the rhizosphere. UniGene 18532 is homologous to Arabidopsis YSL2 (AtYSL2). AtYSL2 transports Fe(II) and copper in complexes with NA (DiDonato et al. 2004), and plays a role in intracellular metal transport (Schaaf et al. 2005). The level of UniGene 18532 transcription was maintained throughout the day in the Fe-deficient barley roots, unlike the expression of HvYS1 and UniGene 12454 (Fig. 2). The protein encoded by this gene may be involved in the intracellular transport of metals and in Fe homeostasis.

In contrast to the barley genes, the expression of genes involved in MAs biosynthesis in rice, OsNAS1, OsNAS2, OsNAAT1, and OsDMAS1, is induced by Fe deficiency and shows diurnal changes in Fe-deficient roots (Nozoye et al. 2004). The Fe deficiency-responsive cis-acting elements IDE1 and IDE2 were identified in the promoter region of HvIDS2 (Kobayashi et al. 2003). Sequences homologous to IDE1 are also present in Fe deficiency-inducible genes in Arabidopsis, barley, and rice, including the above genes, and these elements act in rice and barley roots under conditions of Fe deficiency (Kobayashi et al. 2003). Recently, the rice transcription factor IDEF1, which specifically binds IDE1, was identified (Kobayashi et al. 2007). IDEF1 regulates the response to Fe deficiency in plants. In addition, promoter elements conferring diurnal expression have been identified in the promoter regions of the above four rice genes, and these elements are thought to contribute to their diurnal expression in Fe-deficient rice roots (Nozoye et al. 2004). The evening element (EE; AAATATCT) is a motif for evening-specific expression in Arabidopsis (Harmer et al. 2000), and a mutation in this element could cause a shift in rhythm from evening- to dawn-specific expression (Harmer and Kay 2005). The CBS element (CIRCADIAN CLOCK ASSOCIATED 1 binding site; AAAAATCT) is another motif for the diurnal regulation of gene expression, and this element differs from the EE element by only one nucleotide (Michael and McClung 2002). In addition, some clock-controlled genes predicted by enhancer trapping analysis do not have an exact EE sequence or CBS elements but contain a single mismatched sequence (Michael and McClung 2003). The expression of rice genes, which contain EE- or CBS-like elements, is regulated diurnally (Nozoye et al. 2004). We previously reported the genomic fragments of HvNAS1, HvNAAT-A, HvNAAT-B, IDS2, and IDS3, which participate in MAs biosynthesis (Higuchi et al. 1999; Takahashi et al. 1999; Nakanishi et al. 2000; Kobayashi et al. 2001). To examine the differences in the diurnal expression of these genes in rice and barley roots under Fe-deficient conditions, we searched for CBS and EE elements in the promoter regions of the barley genes (Table 3). The promoter regions of these genes contained EE-like sequences, a CBS element, and a CBS-like element, similar to the rice genes. Interestingly, more of these elements were present in the promoter region of the barley genes than the rice genes, although the expression of these genes in the Fe-deficient barley roots showed no diurnal pattern. These results suggest that Fe deficiency is dominant over diurnal changes in the expression of MAs biosynthesis-related genes in barley roots. However, these results do not mean that there is no diurnal regulation of Fe deficiency-inducible genes in barley roots. For example, the level of transcription of HvYS1 and UniGene 12454, a putative metal–MAs complex transporter gene, showed diurnal changes in addition to induction by Fe deficiency. Because the sequences of the promoter regions of these transporter genes have not been reported, it is unknown whether they contain the cis elements described above. The transcript levels of these transporter genes are regulated by two factors, Fe deficiency and the diurnal cycle, which is different from the expression of genes involved in MAs biosynthesis. The biosynthesis of large amounts of MAs probably requires the expression of genes involved in MAs biosynthesis to remain high throughout the day. In contrast, MAs are secreted diurnally from Fe-deficient barley roots (Takagi et al. 1984; Marschner et al. 1986), and the expression of genes encoding metal–MAs complex transporters may be synchronized with MAs secretion to effectively take up Fe- and metal–MAs complexes from the rhizosphere. This regulation of the expression of genes involved in MAs biosynthesis and Fe–MAs metal complex transporters probably plays a key role in the ability of barley to tolerate low Fe availability.

Table 3 Number of CBS and EE elements in the promoter regions of barley genes involved in MAs synthesis
Full size table