A tonoplast-localized half-size ABC transporter is required for internal detoxification of aluminum in rice

Authors


(fax +81 86 434 1209; e-mail maj@rib.okayama-u.ac.jp).

Summary

Toxic aluminum enters the root cells rapidly, therefore internal detoxification is required. However, the molecular mechanisms underlying this process are poorly understood. Here we functionally characterized a rice gene, Os03g0755100 (OsALS1), that is regulated by ART1, a C2H2-type zinc finger transcription factor. OsALS1 encodes a half-size ABC transporter that is a member of the TAP (transporter associated with antigen processing) sub-group. Expression of OsALS1 was rapidly and specifically induced by Al in the roots, but not by other metals or low pH. OsALS1 was localized at all cells of the roots. Furthermore, OsALS1 is localized to the tonoplast. These expression patterns and cell specificity of localization are different from those of the homologous gene AtALS1 in Arabidopsis. Knockout of OsALS1 in three independent lines resulted in significant increased sensitivity to Al, but did not affect the sensitivity to other metals and low pH. Comparison of Al accumulation patterns between wild-type and osals1 mutants showed that there was no difference in Al levels in the cell sap of root tips between wild-type and the mutants, but the mutants accumulated more Al in the cytosol and nucleus than the wild-type. Expression of OsALS1 in yeast resulted in increased Al sensitivity due to mis-localization. These results indicate that OsALS1 localized at the tonoplast is responsible for sequestration of Al into the vacuoles, which is required for internal detoxification of Al in rice.

Introduction

Phytotoxicity of aluminum is characterized by rapid inhibition of root elongation at micromolar concentrations. Although the exact mechanisms of Al toxicity are still not well understood, it seems that Al interacts with multiple sites of root cells (Kochian, 1995; Ma, 2007; Poschenrieder et al., 2008), resulting in functional and structural damage. For example, Al reduces root cell-wall extensibility (Ma et al., 2004b), and has also been reported to block Ca2+ channels on the root-cell plasma membrane in roots of wheat (Triticum aestivum) (Huang et al., 1992). Aluminum causes membrane damage and peroxidation of membrane lipids (Cakmak and Horst, 1991; Wagatsuma et al., 1995). It also affects signal transduction pathways such as the phosphoinositide pathway in the plasma membrane, resulting in disruption of cytoplasmic Ca2+ homeostasis, the phospholipid bilayer of the plasma membrane, and components of the cytoskeleton (Jones and Kochian, 1997). Studies have also shown that Al blocks symplastic transport and communication by inducing callose deposition at plasmodesmata in wheat roots (Sivaguru et al., 2000) and that it affects mitochondrial functions by triggering production of reactive oxygen species in pea (Pisum sativum) roots (Yamamoto et al., 2002). Therefore, in order to survive in a toxic Al environment such as acidic soils, plants have developed strategies for detoxifying Al both externally and internally.

Several mechanisms for external detoxification of Al have been proposed (Kochian et al., 2004; Ma, 2007; Poschenrieder et al., 2008), but the most well-studied strategy is secretion of organic acid anions, including citrate, oxalate and malate, from the roots (Ma et al., 2001; Kochian et al., 2004). These organic acid anions are able to chelate Al externally, preventing Al from binding to root cells, and thereby detoxifying Al (Ma, 2000). A wide variety of plant species, including both dicots and monocots, have been reported to utilize this mechanism (Delhaize et al., 1993; Ma et al., 2001; Hoekenga et al., 2006). Genes encoding transporters for the Al-induced secretion of malate and citrate have been identified (Ryan and Delhaize, 2010). The malate transporter gene ALMT1 was first isolated in wheat (Sasaki et al., 2004), and homologous genes have been isolated in Arabidopsis and oilseed rape (Brassica napus) (Hoekenga et al., 2006; Ligaba et al., 2006). Genes responsible for Al-induced citrate secretion have also been identified in barley (Hordeum vulgare), Sorghum bicolor, Arabidopsis, maize (Zea mays), wheat and rye (Secale cereale) (Furukawa et al., 2007; Magalhaes et al., 2007; Liu et al., 2008; Ryan et al., 2009; Maron et al., 2010; Yokosho et al., 2010). Unlike ALMT1, these genes belong to the multi-drug and toxic compound extrusion (MATE) protein family.

Although most Al is bound to the negatively charged cell wall (Ma et al., 2004b; Jones et al., 2006), a proportion of Al enters the root cells rapidly. Using cryosections of soybean (Glycine max) root tips, secondary ion mass spectrometry analyses showed substantial intracellular Al accumulation after Al exposure of only 30 min (Lazof et al., 1994). Recently, a transporter (Nrat1) facilitating Al uptake into root cells was identified in rice (Oryza sativa) (Xia et al., 2010). Although the concentration of free Al3+ is decreased to <10−10 m at cytosolic pH (pH 7.0), such low concentrations are still potentially phytotoxic because of the strong affinity of Al for oxygen donor compounds such as inorganic phosphate, ATP, RNA, DNA, proteins, carboxylic acids and phospholipids (Martin, 1988). Therefore, internal detoxification for Al entering the cells is required. In some Al-accumulating species such as buckwheat (Fagopyrum esculentum) and hydrangea (Hydrangea macrophylla), internal detoxification of Al is achieved by chelation with oxalate and citrate, respectively, and sequestration (Ma et al., 1997a,b, 2001). In buckwheat, Al is sequestrated into the vacuoles in the form of Al-oxalate complex at a 1:3 molecular ratio (Ma et al., 1998; Shen et al., 2002). However, molecular mechanisms underlying vacuolar sequestration are poorly understood. In Arabidopsis, the half-size ABC transporter ALS1 is implicated in Al sequestration (Larsen et al., 2007). ALS1 is localized at the tonoplast and expressed in all tissues. Knockout of ALS1 resulted in increased sensitivity to Al (Larsen et al., 2007). However, it is not known how this transporter is involved in Al sequestration. Recently, a transcription factor (ART1) for Al tolerance was identified in rice (Yamaji et al., 2009). ART1 regulates 31 genes that are implicated in Al resistance. One of the downstream genes, OsALS1, shows high similarity to Arabidopsis ALS1 (Larsen et al., 2007). In this study, we functionally characterized OsALS1 and found that it plays a crucial rule in Al tolerance in rice. However, OsALS1 shows a distinct expression pattern and cell-specificity of localization that are different from those of AtALS1.

Results

Gene structure of OsALS1

OsALS1 (Os03g0755100) has 17 exons and 16 introns, and encodes a peptide with 641 amino acids (Figure S1a). OsALS1 is a single-copy gene in the rice genome. Phylogenetic analysis identified homologs of OsALS1 in various species, from moss (Physcomitrella patens) and fern (Selaginella moellendorffii) to higher plants including monocots and dicots (Figure S2a). The identity of OsALS1 to AtALS1 is 72% at the amino acid level (Figures S2a and S3). The closest homolog is AAM47580 from Sorghum bicolor, which shares 91% identity with OsALS1 (Figure S2a).

OsALS1 encodes a half-size ABC transporter belonging to the TAP (transporter associated with antigen processing) sub-group, which has typical ABC transporter motifs including the Walker A motif, Q loop, ABC signature, and the Walker B and H motifs (Figure S3). OsALS1 is predicted to have five transmembrane domains (Figures S2b and S3).

Expression pattern of OsALS1

OsALS1 was expressed in the roots as well as the shoots in the absence of Al (Figure 1a). The expression of OsALS1 was up-regulated by Al in the roots, but not in the shoots (Figure 1a). A dose–response experiment showed that expression of OsALS1 in the roots did not greatly increase with increasing Al concentrations (Figure 1b). A time-course experiment showed that expression of OsALS1 was rapidly induced by Al, reaching a maximum at 3 h after Al exposure (Figure 1c). Neither other metals (including Cd and La) nor low pH triggered increased expression of OsALS1 (Figure 1d), indicating that expression of OsALS1 is specifically induced by Al in the roots.

Figure 1.

 mRNA expression pattern of OsALS1 in rice.
(a) Tissue-specific OsALS1 expression in WT and the art1-1 mutant. The seedlings were exposed to 50 μm Al for 6 h. The data were normalized to OsALS1 expression in WT roots without Al treatment.
(b) Dose-responsive expression of OsALS1 in rice roots. The roots were exposed to various concentrations of Al for 6 h. The data were normalized to OsALS1 expression in the absence of Al.
(c) Time-dependent OsALS1 expression. The roots were exposed to 20 μm Al for various times.
(d) Metal- and pH-dependent expression of OsALS1 in rice roots. The seedlings were exposed to various metals or pH for 6 h. The data were normalized to OsALS1 expression in the absence of Al at pH 4.5. The expression was determined by real-time RT-PCR and histone H3 was used as an internal control. All data shown are means ± SD of three biological replicates. Means with different letters are significantly different (< 0.05, Tukey’s test).

Tissue, cellular and subcellular localization of OsALS1

To investigate the tissue-specificity of localization of OsALS1 in rice, we generated a transgenic rice line carrying the 2.3 kb promoter and coding sequence of OsALS1 fused with GFP. Observation of green fluorescence showed that OsALS1 was localized to the primary roots as well as lateral roots (Figure 2a, b). Immunostaining with an anti-GFP antibody revealed that OsALS1 was present in all root cells (Figure 2c). This localization is similar to that of ART1 and other ART1-regulated proteins such as STAR1/2 and Nrat1 (Huang et al., 2009; Yamaji et al., 2009; Xia et al., 2010).

Figure 2.

 Localization of OsALS1.
(a,b) Localization of OsALS1 in a seminal root (a) and a lateral root (b) of transgenic rice carrying pOsALS1:OsALS1-GFP. Scale bars = 100 μm.
(c,d) Cell specificity of localization. Immunostaining was performed using anti-GFP antibody in the root tip region (2 mm from apex) of a pOsALS1:OsALS1-GFP transgenic plant (c) and a non-transgenic wild-type plant (d). Scale bar = 100 μm.
(e,f) Subcellular localization of OsALS1. The 35S:OsALS1-GFP construct was co-transformed with 35S:DsRed into onion epidermal cells by particle bombardment. Scale bars = 100 μm.
(g) Subcellular localization of OsALS1. The 35S:OsALS1-GFP construct was transformed into rice leaf protoplasts by polyethyleneglycol method. Pink color shows chloroplast autofluorescence. Scale bar = 10 μm.

The subcellular localization of OsALS1 was investigated by transiently expressing an OsALS1–GFP fusion with the red fluorescence protein gene DsRed as a cytosolic and nuclear marker in epidermal cells of onion (Allium cepa). The green fluorescence of OsALS1–GFP was localized outside of the nucleus and did not overlap with the DsRed signal, which was localized at the nucleus and cytosol (Figure 2e,f). Furthermore, subcellular localization of OsALS1 was also examined by introducing OsALS1–GFP to rice leaf protoplasts. The signal was observed at the tonoplast (Figure 2g). Taken together, these results indicate that OsALS1 is localized to the tonoplast.

Isolation of osals1 mutants

To examine the biological function of OsALS1 in rice, we obtained three independent loss-of-function mutants of OsALS1. osals1-1 was isolated by screening of M3 progeny from gamma-irradiated seeds (cv. Nipponbare). Sequence analysis revealed that this mutant had a 9 bp deletion in the 4th exon of OsALS1 (176–184 bp of the OsALS1 coding sequence) (Figure S1a), causing deletion of three amino acids and one amino acid change (i.e. the sequence TMAL is replaced by a single methionine) in the first transmembrane domain of OsALS1. Allelic tests described below showed that this is a knockout line of OsALS1. osals1-2 and osals1-3 were obtained from the National Institute of Agrobiological Sciences rice Tos17 insertion mutant database and the POSTECH rice T-DNA insertion sequence database, and have an insertion of the retrotransposon Tos17 in the 17th exon and a T-DNA insertion in the 4th intron of OsALS1, respectively (Figure S1a). RT-PCR analysis showed that the full-length transcript of OsALS1 was undetected in osals1-2, and that osals1-3 produced no RNA, indicating that both lines are knockout lines of OsALS1 (Figure S1b).

Phenotypic analysis of OsALS1 knockout lines

When wild-type (WT) rice, osals1-1 and osals1-2 were grown in the absence of Al, their root growth was similar (Figure 3a). However, in the presence of 50 μm Al, the root growth was inhibited more in the mutants than in the WT (Figure 3a). A dose–response experiment showed that all three osals1 mutants were more sensitive to Al than their WTs at all Al concentrations tested (Figure 3b).

Figure 3.

 Phenotypic analysis of osals1 mutants.
(a) Al sensitivity of WT (cv. Nipponbare) and two osals1 mutants. Germinated seeds were exposed to 0.5 mm CaCl2 solution (pH 4.5) containing 0 or 50 μm Al for 7 days. Scale bar = 2 cm.
(b) Al sensitivity of osals1 mutants and their WT. Six-day-old seedlings of Nipponbare (WT1), osals1-1, osals1-2, Dongjin (WT2) and osals1-3 were exposed to 0.5 mm CaCl2 solution containing 0, 10, 25, 50 or 100 μm AlCl3 at pH 4.5 for 24 h. Root length was measured before and after the treatment. Data are means ± SD (= 10). Means with different letters are significantly different (< 0.05, Tukey’s test).

To confirm that osals1-1 isolated from screening of mutated seeds is an allelic mutant of osals1-2, we investigated the Al sensitivity in F1 plants derived from a cross between osals1-1 and osals1-2. The Al sensitivity of the F1 plants was similar to each mutant (Figure S4), indicating that osals1-1 is allelic to osals1-2 (Figure S1). This result also confirmed that the increased sensitivity of the mutants to Al was caused by the loss of function of OsALS1.

To investigate whether osals1 mutants are specifically sensitive to Al stress or not, the sensitivity to other toxic metals including Cd, La, Zn and Cu was compared between WT and the mutants. There was no difference in sensitivity in the presence of 20 μm Cd, 5 μm La, 100 μm Zn or 0.5 μm Cu between WT rice and mutants (Figure 4a). We also compared the sensitivity to various pHs between WT and two osals1 mutants. The results showed that the sensitivity to pH (between 3.5 and 6.5) was similar between WT rice and mutants (Figure 4b). These results indicate that the sensitivity of osals1 mutants to Al is highly specific.

Figure 4.

 Sensitivity to other metals and low pH.
(a) Sensitivity to other metals. Five-day-old seedlings of Nipponbare (WT), osals1-1 and osals1-2 were exposed for 24 h to 0.5 mm CaCl2 solution (pH 4.5) containing 50 μm Al, 20 μm Cd, 5 μm La, 100 μm Zn, 0.5 μm Cu or no metal (NM).
(b) Sensitivity to low pH. Five-day-old seedlings of WT (cv. Nipponbare), osals1-1 and osals1-2 were exposed to a buffer solution containing 0.5 mm CaCl2 at pH 3.5, 4.5, 5.5 or 6.5 for 24 h. Data are means ± SD (= 10). Root length was measured before and after treatments. Means with different letters are significantly different (< 0.05, Tukey’s test).

Altered Al distribution in osals1 mutants

Al accumulation in WT and osals1 mutants was examined. After a short exposure (6 h) to Al, the mutants accumulated similar levels of Al to those in WT for both root tips (0–1 cm) and basal roots (1–2 cm) (Figure 5a). The Al levels in the cell wall and root cell sap were also similar between WT and mutants (Figure 5b, c). These results suggest that knockout of OsALS1 did not affect Al accumulation in the roots during a short exposure to Al.

Figure 5.

 Al accumulation in WT and osals1 mutants.
(a–c) Five-day-old seedlings of WT, osals1-1 and osals1-2 were exposed to 0.5 mm CaCl2 solution containing 50 μm Al at pH 4.5 for 6 h. Root tips (0–1 cm) and basal roots (1–2 cm) were excised for fractionation and Al determination. Data are means ± SD (= 3). Means with different letters are significantly different (< 0.05, Tukey’s test).
(a) Total Al content in various root segments.
(b) Al content in the cell wall.
(c) Al content in the root cell sap.
(d) Subcellular distribution of Al co-stained with morin (green) and FM4-64 (red). Five-day-old seedlings were pre-stained with 5 μm FM4-64 for 20 min, and then exposed to 0.5 mm CaCl2 solution (pH 4.5) containing 50 μm Al for 6 h. Roots were transversely sectioned at 2 and 15 mm from the apex for morin staining and fluorescence observation. Scale bar = 100 μm. Insets show 4-times magnified images.

To examine the Al distribution in the cells, we performed morin staining with FM4-64 on roots of WT and osals1 mutants. Morin can detect Al in the cytosol, but not cell wall-bound Al (Eticha et al., 2005). FM4-64 is a membrane trafficking dye that initially localizes to the plasma membrane and then moves to tonoplast by endocytosis. After a short exposure to Al, both mutants displayed stronger Al-dependent green fluorescent signal in root cells than WT in both the root tip and basal region (Figure 5d). Further observation revealed that the signal was present in the cytosol and nucleus of the mutants (Figure 5d), localizing outside the red fluorescence of FM4-64. Al in the vacuoles was not stained by morin. There are two possible reasons for this. One is that morin is not permeable to the tonoplast. The other is that Al in the vacuoles is chelated by organic reagents such as malic and citric acids, and morin is not able to detect these complexed Al forms, similar to cell wall-bound Al (Eticha et al., 2005).

Analysis of a double mutant between osals1 and art1

Microarray analysis showed that OsALS1 is regulated by ART1 (Yamaji et al., 2009). To confirm this, quantitative real-time RT-PCR was used to assess the expression of OsALS1 in WT rice and the art1-1 mutant. The results showed that Al-induced expression of OsALS1 was not observed in the art1 mutant (Figure 1a), confirming that OsALS1 is regulated by ART1. Recently, an ART1-targeted cis-element has been identified (Tsutsui et al., 2011). Two copies of this element were present at −490 and −1679 upstream of the translational start codon of OsALS1 (Tsutsui et al., 2011).

To investigate the relationship between ART1 and OsALS1, we generated a double mutant between osals1-1 and art1-2 for Al tolerance evaluation. The art1-2 mutant has recently been isolated from M3 progeny from gamma-irradiated seeds, and shares the same genetic background (cv. Nipponbare) as osals1-1. Sequence comparison showed that the art1-2 mutant has a 11 bp deletion in the coding sequence of ART1 (Figure S5a). Analysis of the Al tolerance of F1 plants from a cross between art1-1 and art1-2 indicated that art1-2 is allelic to art1-1 (cv. Koshihikari) (Figure S5b). An Al tolerance test showed that the art1-2 osals1-1 double mutant is more sensitive to Al than either single mutant at all Al concentrations tested (Figure 6). At 5 μm Al, the root elongation of art1-2 was hardly inhibited, whereas that of osals1-1 was inhibited by 20% (Figure 6). At higher Al concentrations, the Al tolerance of art1-2 was lower than that of osals1-1. These results suggest that basal expression of OsALS1 in the art1-2 mutant plays an important role in detoxification of Al at lower Al concentrations, and that other downstream genes regulated by ART1 in addition to OsALS1 are also required for high Al tolerance in rice.

Figure 6.

 Al sensitivity in the art1-2 osals1-1 double mutant. Six-day-old seedlings of osals1-1, art1-2 and art1-2 osals1-1 were exposed to 0.5 mm CaCl2 solution (pH 4.5) containing 0, 5, 20, 50 or 100 μm Al for 24 h, and root lengths were measured before and after treatment. Data are means ± SD (= 7 or 8). Means with different letters are significantly different (< 0.05, Tukey’s test).

Cooperation between Nrat1 and OsALS1

To investigate the cooperation between Nrat1 and OsALS1, we examined the expression of each gene in two mutants. mRNA expression of OsALS1 in the nrat1 mutant showed a similar pattern to that in the WT rice, being induced by Al in both root tips (0–1 cm) and basal roots (1–2 cm) at similar levels (Figure 7a). These results indicate that knockout of Nrat1 does not affect the expression of OsALS1.

Figure 7.

 Interaction between Nrat1 and OsALS1.
(a) Expression of OsALS1 in the root tips (0–1 cm) and basal roots (1–2 cm) of WT and the Nrat1 mutant treated with or without 50 μm Al for 6 h. The data were normalized to OsALS1 expression in root tips of WT in the absence of Al.
(b) Expression of Nrat1 in the root tips (0–1 cm) and basal roots (1–2 cm) of WT and the osals1-1 mutant treated with or without 50 μm Al for 6 h. The data were normalized to Nrat1 expression in root tips of WT in the absence of Al. Expression was determined by real-time RT-PCR. Data are means ± SD (= 3).
(c) Root elongation inhibition in the nrat1 osals1-1 double mutant. Six-day-old seedlings of osals1-1, nrat1 and nrat1 osals1-1 mutants were exposed to 0.5 mm CaCl2 solution (pH 4.5) containing 0, 30, 50 or 100 μm Al for 24 h. Root lengths were measured before and after treatment. Data are means ± SD (= 6–8). Means with different letters are significantly different (< 0.05, Tukey’s test).

By contrast, in the osals1 mutant, expression of Nrat1 was higher in both the root tip and basal region of the roots in the absence of Al (Figure 7b). However, in the presence of Al, expression of Nrat1 was not up-regulated in the mutant as for WT, but was down-regulated. In the basal region of the roots, Nrat1 expression level in the osals1 mutant was less than 1/70th of that of the WT (Figure 7b). These results suggest that loss of function of OsALS1 induces the expression of Nrat1 in the absence of Al, but has a suppressive effect on Nrat1 expression in the presence of Al.

An Al tolerance test showed that root elongation was inhibited more in the osals1 mutant than in the nrat1 mutant at the same Al concentrations (Figure 7c), indicating that OsALS1 is more critical for Al tolerance. A double mutant derived from a cross between nrat1 and osals1-1 showed similar Al tolerance to that of osals1-1 (Figure 7c).

Al transport activity of OsALS1 in yeast cells

The transport activity of OsALS1 for Al was determined using a yeast expression system. In the absence of Al, growth of yeast cells expressing OsALS1 was similar to that of the vector control and Nrat1-expressing cells as a positive control (Figure 8a). However, in the presence of Al, although growth of both the vector control and OsALS1-expressing yeast cells was inhibited, the inhibition was greater in the OsALS1-expressing yeast cells than that in the vector control at all Al concentrations tested (Figure 8b–d). Nrat1-expressing cells showed higher sensitivity than OsALS1-expressing cells. Expression of OsALS1 did not affect the Al content in the yeast (Figure 8e), in contrast with Nrat1, whose expression resulted in enhanced uptake of Al (Figure 8e). Observation of OsALS1–GFP fluorescence revealed that OsALS1 was localized to an endomembrane other than the tonoplast in yeast cells (Figure S6).

Figure 8.

 Al transport of OsALS1 in yeast cells.
(a–d) Al sensitivity of yeast cells carrying the pYES2 control vector, or OsALS1- or Nrat1-containing vectors on agar plates. Ten microliters of cell suspension with an OD of 0.2 and four serial 1:5 dilutions were spotted on plates containing (a) 0, (b) 200, (c) 250 or (d) 350 μm Al, and incubated at 30°C for 3 days.
(e) Al uptake in yeast cells carrying the pYES2 control vector, or OsALS1- or Nrat1-containing vectors. Yeast cells with an OD of 3.0 were pre-treated with galactose for 2 h to induce gene expression, and then exposed to 50 μm Al for 2 h. The Al content of yeast cells was determined by atomic absorption spectroscopy. Data are means ± SD (= 3). Means with different letters are significantly different (< 0.05, Tukey’s test).

Discussion

OsALS1 is required for Al tolerance in rice

We functionally characterized a rice gene (Os03g0755100, OsALS1) that is regulated by the transcription factor ART1 that functions in Al tolerance (Figure 1a) (Yamaji et al., 2009). OsALS1 is expressed in both roots and shoots (Figure 1a), but only expression in the roots is greatly enhanced in response to Al (Figure 1a). Furthermore, the expression in the roots was rapidly and specifically induced by Al (Figure 1d). Knockout of OsALS1 in three independent lines resulted in increased sensitivity to Al (Figure 3), but did not affect the sensitivity to other metals and low pH (Figure 4). These results indicate that the OsALS1 gene is specifically required for Al tolerance in rice.

Knockout of OsALS1 did not alter total root Al accumulation or levels of cell wall-bound Al and Al in the root cell sap (Figure 5a–c), but did alter the subcellular distribution of Al; more Al was observed in the cytosol and nucleus in knockout lines in both root tips and basal zones (Figure 5d). OsALS1 encodes a half-size ABC transporter that is localized to the tonoplast of all root cells (Figure 2). These results suggest that OsALS1 functions to sequester Al into the vacuoles. Failure to sequester Al into the vacuoles causes Al toxicity in the cytosol (Figure 5d), resulting in root elongation inhibition in the knockout lines (Figure 3).

Homologs of OsALS1 are present in numerous species from moss to higher plants (Figure S2a). This suggests that internal detoxification of Al mediated by ALS1 homologs represents a universal mechanism in land plants, although these homologs remain to be functionally characterized.

Distinct expression patterns and cell specificity of localization of OsALS1

Although OsALS1 shares 72% identity with AtALS1 (Figures S2 and S3), they have different expression patterns and cell specificity of localization. The expression of OsALS1 is up-regulated by Al (Figure 1), whereas that of AtALS1 is unaffected by Al (Larsen et al., 2007). Although both OsALS1 and AtALS1 are localized to the tonoplast (Figure 2e–g) (Larsen et al., 2007), OsALS1 is expressed in all cells of the roots, whereas AtALS1 is expressed in the vasculature throughout the plant and in the hydathodes in addition to the root tips (Larsen et al., 2007). These findings suggest that AtALS1 and OsALS1 have different roles in plants. Based on its localization, AtALS1 was proposed to facilitate entry or release of Al from the phloem as part of a mechanism of Al movement from the Al-sensitive root tip to less sensitive tissues (Larsen et al., 2007). However, this is not the case for OsALS1 because OsALS1 is localized in all root cells (Figure 2).

The regulatory mechanism of Al tolerance also appears to be different between rice and Arabidopsis. OsALS1 is regulated by the transcription factor ART1 (Figure 1a) (Yamaji et al., 2009; Tsutsui et al., 2011). However, AtALS1 is unlikely to be regulated by STOP1, a homolog of ART1 in Arabidopsis (Sawaki et al., 2009).

OsALS1 is probably a transporter for Al

OsALS1 belongs to the ABC transporter family (Figure S2b), which has 120 members in the rice genome (Rea, 2007; Verrier et al., 2008). Although most of them have not been functionally characterized, previous studies have shown that ABC transporters are involved in the transport of a wide variety of substrates, including lipids, hormones, metals, inorganic acids, glutathione conjugates, peptides, secondary metabolites and xenomolecules (Verrier et al., 2008). We attempted to identify the transport substrate of OsALS1 by expressing it in yeast. Expression of OsALS1 resulted in increased Al sensitivity in yeast (Figure 8a–d), but did not affect the uptake of Al (Figure 8e). If OsALS1 is localized to the tonoplast in yeast, as in rice, its expression is expected to increase the tolerance to Al by sequestering toxic Al into vacuoles. This inconsistency may be attributable to mis-localization of OsALS1 in yeast. In contrast to its localization in rice, OsALS1 is localized to an endomembrane other than the tonoplast in yeast (Figure S6). Nevertheless, the increased Al sensitivity observed in the yeast suggests that OsALS1 is able to transport Al to other organelles from cytosol.

At cytosol pH (7.0–7.5), ionic Al is mostly taken up as Al(OH)4 or chelated by organic anions such as citrate and oxalate (Ma, 2000). OsALS1 belongs to the TAP (transporter associated with antigen processing) sub-group of ABC transporters. Mammalian TAP1 and TAP2 are involved in peptide secretion and translocation through the ER membrane (Rea, 2007). Mdl1, a yeast ABC transporter with 30.2% similarity to OsALS1, is localized to the inner membrane of yeast mitochondria and is required for mitochondrial export of short peptides of 6–21 amino acids (Young et al., 2001). Therefore, OsALS1 might transport Al in complexed form from the cytosol to the vacuole. We tested the transport activity of OsALS1 for ionic Al, Al citrate or Al malate in Xenopus oocytes expressing OsALS1, but failed to detect the activity (unpublished results). Transport activity of AtALS1 for free or chelated AlCl3 also could not be detected in a yeast expression system (Larsen et al., 2007). A possible reason for this is that OsALS1 was not localized to the plasma membrane in the heterologous expression system, making determination of efflux transport activity difficult. Other approaches are necessary to identify the transport substrate of both OsALS1 and AtALS1. As OsALS1 is a half-size ABC transporter, it will also be interesting to examine whether OsALS1 functions as a homodimer or heterodimer.

Relationship between Nrat1 and OsALS1

Nrat1 is a plasma-membrane localized transporter for Al (Xia et al., 2010). Both Nrat1 and OsALS1 expression are induced by Al and regulated by the transcription factor ART1 (Figure 1a) (Xia et al., 2010). Knockout of Nrat1 resulted in increased Al sensitivity and decreased Al uptake (Xia et al., 2010). Furthermore, Nrat1 and OsALS1 show similar cell specificity of localization (Figure 2) (Xia et al., 2010). Therefore, it is possible that Nrat1 and OsALS1 function cooperatively, i.e. that Al transported by Nrat1 is sequestered by OsALS1 into the vacuoles. Expression of OsALS1 was unaffected in the Nrat1 knockout line (Figure 7a); however, the expression of Nrat1 was suppressed in the OsALS1 knockout line in the presence of Al (Figure 7b). These findings suggest that Nrat1 expression is feedback-regulated by OsALS1, although the underlying regulation mechanism is still unknown. This feedback appears biologically reasonable as down-regulation of Nrat1 expression reduces the influx of Al into root cells and thereby attenuates the Al toxicity when the internal detoxification gene OsALS1 is knocked out. This also suggests that functional OsALS1 is a prerequisite for Nrat1 playing a role in Al tolerance. This hypothesis is supported by the observation that over-expression of Nrat1 enhances the sensitivity to Al, as OsALS1 may then become a limiting factor for detoxification of Al (Xia et al., 2011).

Al toxicity in the osals1 mutant is caused by Al accumulation in the cytosol (Figure 5d). Therefore, if Nrat1 is knocked out, it is expected that the Al tolerance in the osals1 mutant will be alleviated. However, analysis of a double mutant between nrat1 and osals1 showed that the Al tolerance was similar between the double mutant and the osals1 single mutant (Figure 7c). These findings suggest that OsALS1 is more critical than Nrat1 for Al tolerance in rice, and that Al is transported into root cells through other unknown transporters. This conclusion is supported by the finding that Al is present in the root cell sap of the nrat1 mutant (Xia et al., 2010).

In conclusion, OsALS1 is a tonoplast-localized half-size ABC transporter. It functions to sequester toxic Al into the vacuoles in rice roots, and is an important component of high Al tolerance in rice.

Experimental procedures

Plant materials

osals1-1 and art1-2 mutants were obtained by screening of M3 progeny from gamma-irradiated seeds (cv. Nipponbare) based on Al-induced root elongation inhibition as described by Ma et al. (2005). The Tos17 insertion line osals1-2 (NC2694) and the T-DNA insertion osals1-3 (3A-07332) were obtained from the National Institute of Agrobiological Sciences rice Tos17 insertion mutant database (http://tos.nias.affrc.go.jp/) and the POSTECH rice T-DNA insertion sequence database (http://www.postech.ac.kr/life/pfg/risd/), respectively. Homozygous lines were identified based on the Al-sensitive phenotype and PCR analysis of the genotype of each line.

Phylogenetic analysis

Phylogenetic analysis of OsALS1 and its homologs was performed using MEGA4.0 software (http://www.megasoftware.net/). The phylogenetic tree was constructed using MEGA4.0 by the neighbor-joining method with 1000 bootstrap trials (Tamura et al., 2007).

RNA isolation and RT-PCR analysis

To examine the expression pattern of OsALS1 in rice, seedlings of WT (cv. Koshihikari or Nipponbare), art1-1 or nrat1 were exposed to various Al concentrations (0–100 μm) for various periods of time, various pH, or Cd and La. Root tips (0–1 cm), basal roots (1–2 cm), or whole roots and shoots were harvested and frozen in liquid nitrogen within 5 min. To compare the expression of Nrat1 in WT and the osals1 mutant, seedlings of WT and the osals1-1 mutant were exposed to 50 μm Al for 6 h, and then the root tips and basal roots were excised for RNA extraction. Total RNA was extracted using an RNeasy plant mini kit (Qiagen, http://www.qiagen.com/). Total RNA (1 μg) was used for first-strand cDNA synthesis using a SuperScript II kit (Invitrogen, http://www.invitrogen.com/), according to the manufacturer’s instructions, using an oligo(dT)20 primer. The primers for RT-PCR analysis of OsALS1 in the osals1-2 and osals1-3 mutants were F2 (5′- CAGTTTTCCAGACCAATACAAGAC-3′) and R2 (5′-CAGTGGATCATCAGGAAACGGTGT-3′), and F1 (5′-GTTCTGCCGCGTCATCAAGT-3′) and R1 (5′-TGTGCCCGTCACATCATCTAG-3′), respectively. The sequences for the internal control ubiquitin1 were 5′-GAGTCCACCCTCCACCTCGTC-3′ and 5′-TGAAGCATCCAGCACAGTAAAA-3′. For real-time RT-PCR analysis, the primer sequences for OsALS1 and Nrat1 were 5′-GGTCGTCAGTCTCTGCCTTGTC-3′/5′-CCTCCCCATCATTTTCATTTGT-3′ and 5′-TCGCATTGGCTCGCACCCT-3′/5′-TCGTCTTCTTCAGCCGCACGAT-3′, respectively. Histone H3 (forward primer 5′- GGTCAACTTGTTGATTCCCCTCT-3′, reverse primer 5′-AACCGCAAAATCCAAAGAACG-3′) was used as an internal control. Experiments were performed using the Mastercycler® ep realplex real-time PCR system (Eppendorf, http://www.eppendorf.com/).

Transient expression in onion epidermal cells and rice protoplasts

For construct of the OsALS1–GFP fusion gene, the coding sequence of OsALS1 except the stop codon was amplified by PCR using primer pair 5′- TAGTCGACATGGGGAAGAACCTGCGTATCAAG-3′ and 5′- TGATATCCCTTGGCCATTGCTTATGGGTTCTATC-3′, and then inserted in-frame between the CaMV 35S promoter and the GFP gene in a pBluescript vector (Agilent, http://www.genomics.agilent.com/). The constructed plasmids 35S:OsALS1-GFP or 35S:GFP mixed with the plasmid 35S:DsRed were coated with 1 μm gold particles and then bombarded into onion epidermal cells. After incubation at 25°C for 18 h, the GFP signal was observed using an LSM700 laser scanning microscope (Zeiss, http://www.zeiss.com/).

The subcellular localization was investigated by introducing 35S:OsALS1-GFP into rice leaf protoplasts. The rice leaf protoplasts were prepared from 2-week-old cv. Nipponbare seedlings grown hydroponically, and used for transformation by the polyethylene glycol method as described by Chen et al. (2006). The GFP signal was observed using an LSM700 laser scanning microscope (Zeiss).

Localization of OsALS1

To investigate the tissue-specific localization of OsALS1 in rice plants, a 2.3 kb promoter sequence of OsALS1 was amplified by PCR using primer pair 5′-TAGGTACCATCTTTATCAGGGCGGAGAACACA-3′ and 5′-AAGTCGACAACGGATGGGTGGGTGGG TG-3′, and then the promoter fragment was fused to the OsALS1–GFP fragment from the 35S:OsALS1-GFP plasmid described above. The fused DNA was then inserted into the pPZP2H-lac vector (Fuse et al., 2001). The derived construct was transformed into calluses (cv. Nipponbare) by Agrobacterium-mediated transformation. T1 transgenic plants were used for observation of GFP signal using an LSM700 laser scanning microscope (Zeiss).

For immunostaining, anti-GFP polyclonal antibody was used. Six-day-old T1 transgenic plants were treated with 10 μm Al for 14 h, and then the roots were used for immunostaining as described previously (Yamaji and Ma, 2007).

Evaluation of sensitivity to Al, low pH and other metals

For evaluation of Al sensitivity, seeds of WT (cv. Nipponbare) and the osals1-1 and osals1-2 mutants were germinated at 30°C for 2 days, and then transferred to a plastic net floating on 0.5 mm CaCl2 solution (pH 4.5) containing 0 or 50 μm AlCl3. After 7 days, the root length was measured using a ruler. In a dose–response experiment, 6-day-old seedlings of Nipponbare (WT1), osals1-1, osals1-2, Dongjin (WT2) and osals1-3 were exposed to 0.5 mm CaCl2 solution containing 0, 10, 25, 50 or 100 μm AlCl3 at pH 4.5 for 24 h. The root length was measured before and after treatment. For evaluation of sensitivity to other metals or pH, seedlings of both WT rice and mutants were exposed for 24 h to 0.5 mm CaCl2 solution (pH 4.5) containing 0 or 50 μm AlCl3, 20 μm CdCl2, 5 μm LaCl3, 100 μm ZnCl2 or 0.5 μm CuCl2, or to 0.5 mm CaCl2 solution at pH 6.5, 5.5, 4.5 or 3.5, buffered with 10 mm homopiperazine-N,N′-bis-2-(ethanesulfonic acid) (homopipes). The root length was measured before and after treatments.

Fractionation and determination of Al content

Five-day-old seedlings of WT (cv. Nipponbare) and the osals1-1 and osals1-2 mutants were exposed to 0.5 mm CaCl2 solution containing 50 μm Al at pH 4.5 for 6 h. Root tips (0–1 cm) and basal roots (1–2 cm) were excised and placed in a plastic tube containing 10 ml of 2 N HCl for total Al determination. For collection of root cell sap, the excised root segments were placed on filter in a tube and frozen at −80°C overnight. After rapid thawing at room temperature, the cell sap was collected by centrifugation at 20 600 g for 10 min (Xia et al., 2010). The pellet was washed three times (5 min for each washing) with 70% ethanol to remove membrane fractions, and then Al was extracted using 2 N HCl. The Al concentration in the extraction solution was determined by ICP-MS using an Agilent 7700 mass spectrometer (http://www.agilent.com).

Morin staining

Five-day-old seedlings of WT (cv. Nipponbare) and the osals1-1 and osals1-2 mutants were exposed to 5 μm FM4-64, a membrane trafficking marker that localizes to the tonoplast, for 20 min, followed by exposure to 0.5 mm CaCl2 solution (pH 4.5) containing 50 μm Al for 6 h. Roots were excised and embedded in 5% agar. Root tips were transversely sectioned at 2 and 15 mm from the apex, and stained in 0.01% morin for 15 min. The green fluorescence signal was observed using an LSM700 laser scanning microscope (Zeiss).

Yeast experiments

OsALS1 was amplified by PCR using primer pair 5′-AAAGCTTAAAATGGGGAAGAACC TGCGTATCAAG-3′ and 5′-ATCTAGACAGTGGATCATCAGGAAACGGTGT-3′, and the fragment was inserted into the pYES2 vector (Invitrogen, http://www.invitrogen.com/), which contains the Gal-inducible GAL1 promoter. The plasmids OsALS1-pYES2, vector control pYES2 and Nrat1-pYES2 constructed previously (Xia et al., 2010) were transformed into yeast strain BY4741. Al sensitivity evaluation on agar plates and Al uptake in liquid culture were performed as described by Xia et al. (2010). To examine the subcellular localization of OsALS1 in yeast, the OsALS1–GFP fragment from the construct described above was inserted into the pYES2 vector. GFP fluorescence was observed using an LSM700 laser scanning microscope (Zeiss).

Evaluation of Al tolerance in double mutants

To generate double mutants, osals1-1 was crossed with art1-2 or nrat1. The F1 plants were used for evaluation of Al tolerance by measuring root elongation in the absence and presence of Al and PCR analysis of genotype and self-pollinated to obtain the F2 generation. Homozygous double mutants were obtained by genotyping of F2 plants and then self-pollinated. For evaluation of Al tolerance, 6-day-old seedlings of osals1-1, art1-2 and osals1-1 art1-2 mutants were exposed to 0.5 mm CaCl2 solution (pH 4.5) containing 0, 5, 20, 50 or 100 μm Al for 24 h, and their relative root elongation was used to assess Al tolerance. The Al tolerance of the osals1-1 nrat1 double mutant and the osals1-1 and nrat1 single mutants was examined by exposing the seedlings to 0.5 mm CaCl2 solution (pH 4.5) containing 0, 30, 50 or 100 μm Al for 24 h.

Accession Numbers

Sequence data for OsALS1 can be found in the GenBank/EMBL databases under accession number AB625451.

Acknowledgements

This research was supported by a Grant-in-Aid for Scientific Research on Innovative Areas from the Ministry of Education, Culture, Sports, Science and Technology of Japan (21248009 and 22119002 to J.F.M.), and a grant from the Ministry of Agriculture, Forestry and Fisheries of Japan (Genomics for Agricultural Innovation, IPG-0006, to J.F.M.). We would like to thank the National Institute of Agrobiological Sciences (Japan) and POSTECH (Pohang University of Science and Technology, Gyungbuk, Korea) for providing the Tos17 and T-DNA insertion lines, respectively.

Ancillary

Advertisement