Involvement of ASR genes in aluminium tolerance mechanisms in rice


M. Margis-Pinheiro. Fax: +55 51 3308 9823; e-mail:


Among cereal crops, rice is considered the most tolerant to aluminium (Al). However, variability among rice genotypes leads to remarkable differences in the degree of Al tolerance for distinct cultivars. A number of studies have demonstrated that rice plants achieve Al tolerance through an unknown mechanism that is independent of root tip Al exclusion. We have analysed expression changes of the rice ASR gene family as a function of Al treatment. The gene ASR5 was differentially regulated in the Al-tolerant rice ssp. Japonica cv. Nipponbare. However, ASR5 expression did not respond to Al exposure in Indica cv. Taim rice roots, which are highly Al sensitive. Transgenic plants carrying RNAi constructs that targeted the ASR genes were obtained, and increased Al susceptibility was observed in T1 plants. Embryogenic calli of transgenic rice carrying an ASR5-green fluorescent protein fusion revealed that ASR5 was localized in both the nucleus and cytoplasm. Using a proteomic approach to compare non-transformed and ASR-RNAi plants, a total of 41 proteins with contrasting expression patterns were identified. We suggest that the ASR5 protein acts as a transcription factor to regulate the expression of different genes that collectively protect rice cells from Al-induced stress responses.


Aluminium (Al) is the most abundant metal, accounting for approximately 7% of Earth's mass. Regardless of its abundance, Al is not considered an essential nutrient; however, it can occasionally stimulate plant growth or induce other desirable effects when present at low concentrations (Foy 1983). Most Al is chelated by ligands or is present in non-toxic forms, such as aluminosilicates or precipitates. Al-induced toxicity can occur through solubilization of Al in soils under highly acidic conditions (pH below 5.0) (Famoso et al. 2010). It has been estimated that approximately 50% of arable land is negatively impacted by Al toxicity that results from acidic soil (Panda, Baluska & Matsumoto 2009). Al toxicity is considered a primary limiting factor in regard to agricultural productivity (Matsumoto 2000) because it inhibits root growth and mineral absorption (Liu & Luan 2001), leading to a stunted root system that negatively impacts the uptake of water and nutrients. There are many potential cellular locations that could be damaged through interaction with Al, including the cell wall, the surface of the plasma membrane, the cytoskeleton and the nucleus (Panda et al. 2009). For example, it has been demonstrated that Al binds strongly to the cell wall of root epidermal and cortical cells (Delhaize, Ryan & Randall 1993). However, some plants are able to tolerate toxic levels of Al in acidic soils. These plants have evolved mechanisms to detoxify Al that is both present internally and externally (Kochian, Pineros & Hoekenga 2005). To achieve internal detoxification, plants accumulate Al inside vacuoles, where it is chelated with organic acids (OA), such as citrate and oxalate (Ma 2007). In contrast, the majority of Al-tolerant plants exclude Al from the root tip by releasing OA at sites of high Al concentration; examples of these acids include malate, citrate and oxalate (Ma, Ryan & Delhaize 2001; Kochian, Hoekenga & Pineros 2004). In species such as sorghum and wheat, the OA–Al complex prevents Al from entering the cell (Sasaki et al. 2004; Magalhaes et al. 2007), which reduces the concentration and potential toxicity of Al at the growing root tip (Ma et al. 2001).

Rice is considered the most Al-tolerant crop (Fageria 1989; Duncan & Baligar 1990); however, there is variability among different rice genotypes, resulting in widely varied tolerance levels among different cultivars (Ferreira 1995). In two independent studies, Ma et al. (2002) and Yang et al. (2008) observed no OA exudation and increased Al accumulation in the root apex of Al-susceptible rice strains relative to Al-tolerant strains. These results demonstrate that rice plants achieve high levels of Al tolerance through a novel mechanism that does not involve root tip Al exclusion (Famoso et al. 2010).

Although genetic studies in rice have identified more than 10 quantitative trait loci for Al tolerance, the responsible genes have only recently been cloned (Huang et al. 2009). The genes STAR1 and STAR2 were isolated from an Al-tolerant cultivar irradiated with γ-rays (Ma et al. 2005). The disruption of either gene resulted in hypersensitivity to Al toxicity. STAR1 encodes a nucleotide-binding domain protein, and STAR2 encodes a transmembrane domain protein of a bacterial-type ATP-binding cassette (ABC) transporter. Analyses indicated that STAR1 and STAR2 form a complex that functions as an ABC transporter that is required for detoxification of Al in rice. The ABC transporter transports uridine diphosphate (UDP)-glucose, which may be used to modify the cell wall (Huang et al. 2009). Yamaji et al. (2009) isolated the zinc finger transcription factor ART1, which regulates multiple rice genes implicated in Al tolerance. Genes regulated by ART1 include STAR1, STAR2 and Nrat1, the latter of which is a specific transporter that mediates the sequestration of trivalent Al ions into vacuoles to achieve Al detoxification (Xia et al. 2010).

Using a proteomic approach, Yang et al. (2007) identified some proteins responsive to Al in rice roots; ASR5 was more highly expressed in these roots. The ASR (abscisic acid, stress and ripening) gene was first described in tomato (Iusem et al. 1993). Subsequently, ASR genes were found to be widely distributed in the vegetal kingdom, having been identified in potato (Silhavy et al. 1995), pinus (Chang et al. 1996), maize (Riccardi et al. 1998), rice (Vaidyanathan, Kuruvila & Thomas 1999), sugarcane (Sugiharto et al. 2002), grape (Cakir et al. 2003) and others. Nevertheless, ASR genes do not occur in the genome of Arabidopsis thaliana (Maskin et al. 2001). ASR genes are expressed during fruit ripening and are induced in response to ABA and various abiotic stresses, including water and salt stresses (Carrari, Fernie & Iusem 2004). Kalifa et al. (2004a) demonstrated that tomato ASR1 proteins were present as unstructured monomers localized in the cytosol and as structured homodimers in the nucleus, where they can bind DNA. Cytosolic tomato ASR1 performs a chaperone-like activity and can stabilize a number of proteins, protecting them from denaturation induced by repeated freeze/thaw cycles (Konrad & Bar-Zvi 2008). Furthermore, a grape ASR protein binds to the promoter of a hexose transporter gene (Cakir et al. 2003), suggesting that it may be a transcription factor that is involved in sugar metabolism. In silico analyses mapped the locations of six copies of ASR genes in the rice genome in different chromosomes; these loci were confirmed by expressed sequence tags (ESTs) (Frankel et al. 2006).

In this study, we analysed changes in gene expression of the rice ASR family in response to Al treatment. We found that all members of the ASR gene family display a variable degree of expression, indicating that ASR genes of the tolerant Japonica rice (cv. Nipponbare) are differentially regulated in response to Al. Conversely, ASR5 did not respond to Al exposure in Indica rice roots (cv. Taim), which unlike the Nipponbare cultivar, is highly sensitive to Al (Freitas et al. 2006). According to Ma et al. (2002), Japonica varieties are often more resistant to Al than the Indica varieties. In addition to gene expression analyses, transgenic plants carrying RNAi constructs targeting the ASR genes were made, and an increased Al susceptibility was observed in T1 plants. In addition, transgenic embryogenic calli of rice carrying an ASR5-green fluorescent protein (GFP) fusion protein revealed that ASR5 is located in both the nucleus and the cytoplasm, suggesting that ASR5 may act as a transcription factor.


Plant material and growth conditions

For Al treatments, rice seeds from Japonica Nipponbare and Indica Taim backgrounds were germinated on filter paper for 4 d in the dark at 28 °C. The seedlings were grown in a hydroponic solution (Baier, Somers & Gusiafson 1995) for 12 d in a growth chamber at 28 °C under 12 h of light. The hydroponic solution was replaced every 4 d. After 12 d, seedlings were treated with 150 µM AlCl3. Root tissue samples were collected at 4 and 8 h after the start of treatment. Identical conditions were used for experiments involving the use of 450 µM AlCl3. Samples were collected after 8 h of treatment; roots and shoots were collected from Nipponbare plants and only roots were collected from Taim. For gene expression analysis at the root base and apex, rice plants were cultivated in the same conditions described earlier and samples of Nipponbare roots were divided into two segments: apex (0.5 cm) and base (4 ± 0.5 cm). Control plants were not treated with AlCl3. All plants were grown in acidic conditions (pH 4.5).

To analyse responses to cold stress, 2-week-old plants (cv. Nipponbare) were transferred to growth chambers and exposed to a temperature of 4 °C for 12 h. Control plants were maintained at 28 °C. For ultraviolet (UV) light treatment, 2-week-old plants (cv. Nipponbare) were transferred to the growth chamber and exposed to two 4 h treatments of continuous UV-B illumination (0.25 kJ m–2 min–1) at a 20 h interval, and subsequently maintained under normal light during the recovery period. Analyses were performed 24 h following initial UV irradiation. For drought stress, cv. Nipponbare seedlings were grown to the four-leaf stage in soil with a normal supply of water. Subsequently, control plants were watered normally and stressed seedlings were not watered for 14 d.

Quantitative real-time PCR (RT-qPCR)

Tissue samples were collected and immediately frozen in liquid nitrogen; total RNA was then extracted with Trizol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. cDNA synthesis was carried out using the Moloney murine leukemia virus (M-MLV) RT reverse transcriptase enzyme (Promega, Madison, WI, USA). The RT reaction was performed in a final volume of 50 µL. A 1:10 solution of total cDNA was prepared (stock solution). For RT-qPCR reactions, the stock solution was diluted at a ratio of 1:100.

The protocol for RT-qPCR is summarized as follows: an initial step of 5 min at 94 °C followed by 40 cycles of 10 s at 94 °C, 15 s at 60 °C and 15 s at 72 °C. Samples were maintained for 2 min at 40 °C to promote re-annealing and were then warmed from 55 to 99 °C to generate relative denaturing curve data for the amplification products. RT-qPCR was carried out with 10 µL of the diluted cDNA (1:100), 2.0 µL of 10× PCR buffer (Tris–HCL at 100 mm at pH 8.0, KCl at 500 mm), 1.2 µL MgCl2 50 mm, 0.1 µL 5 mm deoxy-nucleotide triphosphates (dNTPs), 0.2 µL of each individual 10 µM primer solution, 4.25 µL H2O, 2.0 µL SYBR GREEN (1:100 000) and 0.05 µL Platinum Taq DNA Polymerase (5 U/µL, Invitrogen), in a final volume of 20 µL. All reactions were performed in four technical replications, and calculations were performed using the 2−ΔΔCt method (Livak & Schmittgen 2001). Quantitative PCR was performed using specific primer pairs for ASR1 to ASR6: ASR1 5′-TGGTGGACTACGACAAGGAGA and 5′-GCCACCTCCTCCTTCACC-, ASR2 5′- CATGGCGGCTACGGCTAC and 5′- GGTCCTTCTTCGCCTGGT, ASR3 5′- CACCACAAGAACGACGACAA and 5′-TGTGATGCTCGTGGATGG-, ASR4 5′- CGACTATCGCAAGGAGGAGA and 5′- CGATCCCTTCCTTCATCTTG, ASR5 5′- CCAGGACGAGTACGAGAGGT and 5′- CGATCTCCTCCGTGATCTTG, ASR6 5′- GCCCGGAGAAGTACAGGAAG and 5′- GCCCTCCTCGATCCTGTG, APx1 5′- CTACAAGGAGGCCCACCTCA and 5′- CCGCATTTCATACCAACACA and APx2 5′- ACCTGAGGTCCCCTTCCA and 5′- CTCTCCTTGTGGCATCTTCC. The primers eFα 5′-TTTCACTCTTGGTGTGAAGCAGAT and 5′-GACTTCCTTCACGATTTCATCGTAA, FDH 5-′CAAAATCAGCTGGTGCTTCTC and 5′-TTCCAATGCATTCAAAGCTG and Actin 5′-GACTCTGGTGATGGTGTCAGC and 5′-GGCTGGAAGAGGACCTCAGG were used as reference genes against which the amount of mRNA present in each sample was normalized. The primers STAR1 (Huang et al. 2009) 5′-TCGCATTGGCTCGCACCCT and 5′-TCGTCTTCTTCAGCCGCACGAT were used as positive controls for Al stresses. The efficiency of amplification was calculated for all samples and all different primer pairs using the LinReg program (Ramakers et al. 2003). Only samples with efficiency values higher than 85% were considered for expression analysis, with corrections according to primer efficiencies as previously described (Pfaffl 2001). The quantities of amplified products were compared using an Applied Biosystems (Foster City, CA, USA) 7500 Real-Time PCR System.

Construction of plant vector and plant transformation

A chimeric gene for producing RNA with a hairpin structure (hpRNA) was constructed based on the sequence of the ASR5 locus (LOC_Os11g06720.1). The following primers were used to amplify the 417 bp region corresponding to the full ASR5 coding sequence: 5′-CACCATGGCGGAGGAGAAGCAC and 5′-TCAGCCGAAGAGGTGGTG. PCR products were cloned into the Gateway vector pANDA in an inverted repeat configuration, in which the chimeric gene is under the control of the maize ubiquitin promoter with an intron placed upstream of the inverted repeats (Miki & Shimamoto 2004).

To determine the subcellular localization of the ASR5 protein in transgenic rice calli, the complete coding sequence was fused with the GFP coding sequence at the N-terminus into a modified Gateway vector pH 7FWG2 (Karimi, Inze & Depicker 2002). Restriction enzymes were used to replace the 35S promoter with the maize ubiquitin promoter. Agrobacterium-mediated transformation of rice calli was performed as described previously (Upadhyaya et al. 2000) using Nipponbare cultivar. For transient expression of GFP-ASR5 in rice protoplasts, the complete coding sequence of ASR5 was fused to the GFP coding sequence at its N-terminus and cloned into the Gateway vector pART7-YFP (Galvan-Ampudia & Offringa 2007). The amplified cDNA was introduced into appropriate plasmids by Gateway technology. The resulting vector was used to perform protoplast transformation.

Protoplast isolation was performed essentially as described (Chen et al. 2006) and protoplast transformation was performed according to the methodology described in a previous report (Tao, Cheung & Wu 2002). Transformed protoplasts were incubated in the dark for 24–48 h at 27 °C prior to imaging. Fluorescence microscopy was performed with an Olympus FluoView 1000 confocal laser-scanning microscope (Center Valley, PA, USA) equipped with a set of filters capable of distinguishing enhanced green and yellow fluorescent proteins (EGFP and EYFP, respectively) and plastid autofluorescence. Images were captured with a high-sensitivity photomultiplier tube detector.

Characterization of RNAi transgenic plants

T1 or T2 generation and non-transformed (NT) rice seeds were germinated in moistened filter paper at 28 °C in the dark. Four-day-old seedlings were then transferred to plastic pots containing only Baier's solution (Baier et al. 1995) or Baier's solution supplemented with 450 µm AlCl3 or 25 µm of cadmium (CdCl2). Plants were maintained for 12 d, and root length was then measured. The pH (4.5 for both treatments) was monitored daily, and the nutrient solution was replaced every 4 d. Relative root elongation in cadmium or in Al experiments was defined as the percentage of the root elongated by CdCl2 or Al compared to the control (CdCl2-and Al-free).

For drought stresses, T1 generation plants and NT plants were grown in soil for 5 months with a normal supply of water and then subjected to drought for 15 d without watering.

The relative water content (RWC) of leaves was determined by the procedure described by Cairo (1995). A total of 20 leaf discs of 12 mm in diameter were collected randomly, and fresh weight (FW1) was determined with an analytical balance. The discs were then transferred to a Petri dish containing distilled water and placed on the laboratory bench for a period of 12 h under constant illumination (40 µmol m−2 s−1). After this period, leaf discs were removed from the Petri dish, placed on filter paper and subjected to light pressure to eliminate excess water. The discs were immediately weighed again to determine measured mass (FW2). To determine the dry tissue mass [dry weight (DW)], the discs were placed in paper bags and exposed to greenhouse ventilation with an air temperature of 75 °C for a period of 48 h. The RWC calculations were performed according to the following equation:


The humidity percentage of leaf tissues was determined according to the relationship described by Slavik (1974):


Western blotting

Twelve-day-old rice leaves of NT and RNAi plants were macerated and homogenized in 0.5 M Tris–HCl (pH 8.3), 2% Triton X-100, 20 mm MgCl2, 2% β-mercaptoethanol, 1 mm Phenylmethanesulfonylfluoride (PMSF), 2.5% Polyethylene glycol (PEG) and 1 mm ethylenediaminetetraacetic acid (EDTA); samples were incubated at 4 °C for 1 h. Protein extracts were centrifuged and precipitates were discarded. Samples of each lysate were loaded to yield 50 µg of protein, which was separated by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE) in 15% gel. ASR5 protein was detected with a rabbit polyclonal ASR5 antibody (1:500). To prepare the antibody, ASR5 full-length cDNA was cloned into a pGEX (GE, Fairfield, CT, USA) vector; the recombinant clone was introduced into Escherichia coli (BL21). ASR5 expression was induced by isopropyl β-D-1-thiogalactopyranoside (IPTG), and the protein was purified with Glutathione Sepharose 4B (GE). The purified protein was injected into a rabbit, and the serum obtained from that animal contained the antibody. Goat anti-rabbit IgG (1:1000) conjugated to alkaline phosphatase was used as the secondary antibody. The bands were detected with a premixed BCIP/NBT substrate solution (Sigma, St Louis, MO, USA) and recorded on X-ray film.

Two-dimensional gel electrophoresis and mass spectrometry

Leaves of 6-month-old plants, both NT and expressing RNAi, were macerated in liquid nitrogen and homogenized in a buffer containing 0.5 M Tris–HCl, pH 8.3, 2% Triton X-100, 20 mm MgCl2, 2% β-mercaptoethanol, 1 mm PMSF, 2.5% PEG and 1 mm EDTA; samples were then incubated at 4 °C for 2 h. The samples were centrifuged and the precipitate was discarded. An acetone solution containing trichloroacetic acid was added to the supernatant to yield a final concentration of 10%, and the mixture was incubated for 12 h at −20 °C to allow protein precipitation to occur. The resulting precipitate was washed with cold acetone and dissolved in a solution containing 7 M urea and 2 M thiourea (Acquadro et al. 2009). The amount of protein extracted for each sample was quantified via the Bradford (1976) method.

Approximately 500 µg of extracted protein was used for analysis in two-dimensional gel electrophoresis. In the first dimension, proteins were separated according to their isoelectric points under a pH gradient ranging from 4 to 7. In the second dimension, proteins were separated according to their molecular weight in 12.5% polyacrylamide gel containing SDS. The gels were stained with Coomassie brilliant blue R-250 from Thermo Scientific (Rockford, IL, USA) and analysed to obtain images of spots of interest, which were subsequently excised and digested with 10 ng/uL of trypsin. The peptide solution was separated with a multidimensional chromatographic system. The eluted fractions were analysed in a mass spectrometer with electrospray ionization, and two mass analysers – a quadrupole (Q) associated with a tube in which one measures the ion time of flight (TOF) and an ion detector. In conjunction with the online Q-TOF, a capillary chromatography system [Acquity ultra performance liquid chromatography (UPLC)] was used to submit samples in a reverse phase column. To perform MS analysis, the peptides eluted from the column were ionized and then their mass/charge ratio (m/z) was determined. The spectra-generated MS/MS was processed in Mascot Distiller software and compared against the database of the National Center for Biotechnology Information using Mascot software (

Statistical analyses

Data represent mean ±standard error of the mean. Data were analysed by analysis of variance (anova) followed by Duncan's post hoc test, using the statistical program SPSS 15.0 for Windows (SPSS Inc, Chicago, IL, USA) ( The values were considered statically significant when P < 0.05.

In silico analyses

In silico analyses were performed using the Microarray platform (RiceXPro, available at and the BLAST (bl2seq) algorithm for aligning amino acid sequences (available at


Expression profile of ASR genes in rice

The expression patterns of different ASR genes following Al exposure (150 µM) were investigated using quantitative PCR. In Nipponbare plants, a rice cultivar tolerant to excess levels of Al, the expression of ASR genes was up-regulated in response to Al treatment (Fig. 1). Transcript levels corresponding to the genes ASR1, ASR4 and ASR5 were up-regulated after 4 and 8 h of treatment, while ASR2 and ASR6 showed an increase in their transcript levels only after 8 h of treatment; similarly, ASR3 was up-regulated only after 4 h of Al treatment.

Figure 1.

The ASR gene family responds to aluminium stress. Determination of rice ASR family expression levels in response to treatment with Al. Relative expression was plotted using ASR1 gene expression levels in control plants as calibrator. Total RNA was extracted from roots and used for cDNA synthesis. Analyses were conducted using RT-qPCR. Roots of Nipponbare cultivar were collected after 4 and 8 h of treatment with Al (150 µM). Values represent the means ± SD (N = 4). Asterisks indicate significantly different means: *P < 0.05.

When a higher concentration of Al was used to treat Nipponbare plants (450 µM), the expression of ASR5 was increased relative to plants treated with 150 µM of Al. In shoots and roots, ASR5 expression increased after 8 h of treatment with 450 µM Al. In contrast, in the roots of the Al-sensitive cultivar Taim (ssp. Indica), ASR5 expression did not respond to 450 µM Al treatment (Fig. 2a). To confirm the phenotype of the Taim cultivar respective to the tolerance or sensitivity to Al, we have performed experiments comparing the growth of both Nipponbare and Taim plants in the presence of 450 µM Al. Our results show that in response to Al, Taim roots elongated only 18.23%, whereas Nipponbare elongated 59.57% (Fig. 2b,c), indicating that Taim cultivar is significantly more sensitive to Al compared to the Nipponbare cultivar.

Figure 2.

ASR5 does not respond to aluminium stress in an Al-sensitive rice cultivar. (a) Total RNA was extracted from roots and shoots, and used for cDNA synthesis. Analyses were conducted via RT-qPCR. Relative expression was plotted using ASR5 and STAR1 expression levels in control plants as calibrator. Roots and shoots of Nipponbare cultivar and roots of Taim cultivar were collected after 8 h of treatment with Al (450 µM). (b) Relative root elongation (RRE) for Nipponbare and Taim plants. (c) Nipponbare and Taim plants growth for 12 d in a solution containing 450 µm Al. Values represent the means ± SD (N = 4). Asterisks indicate significantly different means: *P < 0.05.

Relative to control roots, ASR5 mRNA expression in Japonica rice was not significantly different in either of the two root segments (apex and base) Al. However, the expression of ASR5 transcripts increased in both segments in response to Al treatment, although stronger expression was observed in the root tip (Fig. 3). To confirm the effect of Al on gene expression, STAR1 was used as a positive control (Huang et al. 2009). As ASR genes are known to be involved in plant responses to several abiotic stresses, the expression patterns of ASR5 in response to cold, drought and UV light were also analysed (Supporting Information Fig. S1). The expression of ASR5 gene did not respond in roots maintained at 4 °C, but increased in leaves when plants were subjected to drought for 14 d. In UV-treated plants, ASR5 transcript levels decreased drastically when relative to control plants.

Figure 3.

ASR5 expression levels increase significantly in root apex after Al exposure. Total RNA was extracted from roots, and used for cDNA synthesis. Analyses were conducted using RT-qPCR. Relative expression was plotted in a logarithmic scale and ASR5 expression levels in root base of control plants was used as calibrator. After 8 h of treatment with Al (450 µM), roots of Nipponbare cultivar were collected and divided into two segments: base and apex (0.5 cm). Bars with different letters are significantly different (anova, P < 0.05).

Using a microarray platform (RiceXPro) to analyse gene expression patterns for Japonica rice grown in natural field conditions (Sato et al. 2010), we observed that ASR1 and ASR5 were expressed in leaves and roots at the vegetative stage and that there was reduced expression at the onset of inflorescence. In later stages of flower development, higher transcript levels were detected in palea, lemma and ovaries (Supporting Information Fig S2).

ASR-silenced plants are sensitive to Al and drought

Transgenic plants were obtained, which expressed RNAi constructs targeting ASR5. Expression analysis of transgenic rice plants demonstrated knockdown of both ASR1 and ASR5 genes. Besides, these genes were not able to respond anymore to Al (Fig. 4a). Western blot analyses showed a decrease in the ASR5 protein levels (Fig. 4b). Under normal conditions, there were no differences between the ASR RNAi plants and NT plants with regard to height, leaf number and root development. However, the onset of flowering was delayed by 15 d in transgenic plants, which also presented an abnormal panicle development and reduced seed number relative to NT plants (Fig. 4c). In addition, transgenic plants showed a reduction in the number of leaf trichomes, palea and lemma (Fig. 4d). T1 generation seeds were germinated and grown for 12 d in a solution containing 450 µm Al. Increased Al sensitivity was observed in T1 plants accompanied by strongly inhibited root elongation (Fig. 5a,b). In shoots, however, there was no difference in fresh weight between NT and RNAi line (Fig. 5c). Furthermore, ASR silencing affects specific Al response since roots of NT plants and ASR RNAi were both strongly inhibited using cadmium (Supporting Information Fig. S3). Transgenic plants subjected to drought for 15 d also showed greater susceptibility to this condition (Fig. 6). These results indicate that ASR5 knockdown affects rice plants by increasing their sensitivity to abiotic stresses.

Figure 4.

Characterization of plants expressing the RNAi construct. (a) Total RNA was extracted from leaves, and used for cDNA synthesis. Analyses were conducted using RT-qPCR. Measurements were performed on 3-week-old leaves of ASR RNAi plants in control condition (red bars), ASR RNAi plants in Al condition (450 µM for 8h – green bars) and non-transformed (NT) plants (blue bars). Values represent the means ± SD (N = 4). Asterisks indicate significantly different means: *P < 0.05. ND means non-detected. (b) Western blot analysis of ASR5 protein expression in RNAi and NT rice plants. (c) Abnormal panicle development in T1 RNAi plants compared to NT plants. (d) Reduction of trichomes in leaves, palea and lemma of T1 RNAi plants compared with NT plants.

Figure 5.

ASR-silenced plants are sensitive to aluminium. (a) T1 generation (RNAi) and non-transformed (NT) growth for 12 d in a solution containing 450 µm Al. (b) Differences in root length for ASR RNAi and NT plants. (c) Differences in shoot fresh weight for ASR RNAi and NT plants. Bars with different letters are significantly different (anova, P < 0.05).

Figure 6.

ASR-silenced plants are sensitive to drought. Analyses of relative water content and percentage of humidity in 5-month-old T1 generation (RNAi) and non-transformed (NT) plants subjected to drought for 15 d. Asterisks indicate significantly different means: *P < 0.05.

Global protein levels were affected in ASR-RNAi plants

Using a proteomic approach to compare NT and ASR-RNAi plants, a total of 41 proteins were identified, 11 of which showed increased expression versus 30 that were down-regulated in ASR-RNAi plants. Of these proteins, 32% are involved in photosynthesis, 20% in carbohydrate metabolism and 17% in response to stress. The other proteins are functionally associated with amino acid metabolism, phosphate metabolism, development, protein degradation, nucleotide binding, cellular composition and electron transport (Table 1) (Supporting Information Fig. S4).

Table 1.  Differentially expressed proteins in leafs of non-transformed and ASR5-RNAi plants identified by mass spectrometry (electrospray ionization-MS/MS)
No spotaΔ ExpressionbProtein nameTheoreticObservedScorecS. P. (%)dAccession (NCBI)e
pIP. M. (kDa)pIP. M. (kDa)
  • a

    Numbers correspond to two dimensional (2-DE) gels.

  • b

    Increase (↑) or reduction (↓) of proteins in response to ASR silencing.

  • c

    Mascot score.

  • d

    % of protein sequencing.

  • e

    Accession number of NCBI.

Protein degradation
 1Heat shock protein 70 (hypothetical protein OsJ_17967)5.12744.777896236gi|222631026
 2Mitochondrial chaperonin (Os12g0277500)5.12614.937369943gi|115488160
 3ATP synthase alpha/beta chain, C-terminal domain (Os01g0711000)5.03545.217023128gi|115439527
 4ATP synthase CF1 alpha subunit5.95566.1770123454gi|11466784
 5Ribulose bisphosphate carboxylase large chain6.22536.516693540gi|11466795
 6ATP synthase gamma chain (Os07g0513000)8.6406.215640534gi|115472339
 7Putative oxygen-evolving enhancer protein 1, chloroplast precursor (Os01g0501800)6.1355.3048128661gi|115436780
 8Putative ferredoxin-NADP(H) oxidoreductase7.98415.965235936gi|41052915
 9Chlorophyll a-b binding protein4.73244.933752960gi|108864186
10Chloroplast 23 kDa polypeptide of photosystem II5.56205.903490067gi|164375543
11Putative photosystem I reaction centre subunit IV9.64165.752343061gi|34394725
12Putative ribulose bisphosphate carboxylase large chain precursor (Os05g0427800)4.88164.61185613gi|115464019
13Ribulose bisphosphate carboxylase, small chain5.89156.111555188gi|56966763
14Ribulose-1,5-bisphosphate carboxylase/oxygenase large subunit7265.502531436gi|290585768
15Ribulose-1,5-bisphosphate carboxylase/oxygenase large subunit7266.192924530gi|290585768
Carbohydrate metabolism
16Putative transketolase6.12815.697869329gi|28190676
17Sedoheptulose-1,7-bisphosphatase precursor5.83434.95766240gi|27804768
18Lactate/malate dehydrogenase, NAD binding domain (Os10g0478200)5.75365.945638849gi|115482534
19Lactate/malate dehydrogenase, alpha/beta C-terminal domain (Os05g0574400)8.22366.685561270gi|115465579
20Glycolate oxidase (Os03g0786100)8.48406.935752944gi|115455773
21Putative uridylyltransferase related5.72304.635120034gi|187608845
22Putative ribose-5-phosphate isomerase4.91274.713826645gi|34393836
Nucleotide binding
24Ribulose 1·5-bisphosphate carboxylase/oxygenase activase small isoform precursor5.59524.866467561gi|62733297
25Chloroplast translational elongation factor Tu5.68565.586387949gi|218191089
Cellular component
26Actin (Os11g0163100)5.31425.436016727gi|115484337
27Fibrillin-like protein5.04344.644815522gi|29367475
28PAP fibrillin family domain containing protein (Os11g0595200)8.92285.034614114gi|115486133
Amino acid metabolism
29Glutamine synthetase, catalytic domain (Os02g0735200)5.51395.565819819gi|115448531
30Putative glycine cleavage system H protein (Os10g0516100)4.92174.451816228gi|115482934
Phosphate metabolism
31Putative inorganic pyrophosphatase5.8325.044812116gi|46805452
Stress response
32Glyoxalase (hypothetical protein OsJ_26316)5.53355.535133041gi|125602450
33Class III peroxidase (OsPrx111)5.77335.735146144gi|20286
34Chloroplastic lipocalin (Os04g0626400)6.23385.01476511gi|115460690
35Cytosolic ascorbate peroxidase – OsAPx1 (Os03g0285700)5.42275.673738154gi|115452337
36Manganese superoxide dismutase6.5256.073142556gi|601869
37Putative superoxide dismutase [Cu-Zn], chloroplast precursor5.79215.711961768gi|42408425
38Putative superoxide dismutase 16.24286.232213936gi|108708142
Transport of electrons
39Oxidoreductase NAD-binding domain (Os02g0328300)5.44314.584536448gi|115445869
402Fe-2S iron-sulphur cluster binding domain (anti-disease protein 1)4.37154.042419157gi|18698985
41Germin-like protein 16.01226.012618515gi|4239821

Cytosolic ascorbate peroxidase gene knockdown rendered rice plants more sensitive to high Al concentrations

Among the stress-related genes down-regulated in ASR-RNAi plants, APx1 and APx2 protein levels were markedly reduced. The down-regulation of cytosolic ascorbate peroxidase (APx) in response to ASR gene knockdown (displayed in Table 1) prompted us to examine how plants containing RNAi sequences targeting these peroxidases (APx1/2s plants) could respond to toxic concentrations of Al. These Apx1/2s plants were previously obtained by our group (Rosa et al. 2010). Treatment with 750 µM Al was conducted in hydroponics for 14 d. Seedlings with reduced expression of cytosolic APx genes (Apx1/2s plants) showed greater sensitivity to this concentration of Al. The aerial parts of Apx1/2s plants developed necrotic lesions in high extension of foliar volume. Conversely, unlike RNAi plants, NT plants leaves had expanded after 14 d of stress (Supporting Information Fig. S5a). The plants (both transgenic and NT) did not differ when grown in control conditions (Rosa et al. 2010). In wild-type tolerant cultivar Nipponbare, APx1 and APx2 increase its levels of transcripts in response to Al (Rosa et al. 2010), whereas in the sensitive cultivar Taim, APx1 and APx2 did not respond to Al (Supporting Information Fig. S5b).

ASR5 protein is localized in nucleus, cytoplasm and chloroplasts

To identify the subcellular localization of the ASR5 protein in vivo, the full-length cDNA of ASR5 was fused at its N-terminus to the coding sequence of GFP. A modified pH 7FWG2 vector (see Materials and Methods) was used for stable expression and the pART vector was used to achieve transient expression. In stable expression analyses, ASR5-GFP fusion proteins expressed in transgenic rice embryogenic calli were present in the nucleus and cytoplasm (Fig. 7a). NT embryogenic calli were used as negative control. Conversely, transient expression analyses of rice protoplasts obtained from green leaf showed that ASR5 resides in chloroplasts (Fig. 7b). For this experiment, a vector containing only GFP was used as positive control and an empty vector was used as a negative control. These results indicate that ASR5 is located in multiple cellular compartments.

Figure 7.

ASR5 is localized in nuclei, cytoplasm and chloroplasts. (a) Localization of UBI-ASR5::GFP in rice embryogenic calli. Green fluorescent signal indicates cell wall autofluorescence in non-transformed rice. (b) Localization of ASR5 in the chloroplasts of rice protoplast visualized by transient expression of 35S-ASR5::YFP. Negative and positive control groups with 35S-gus empty and 35S-gus:YFPC, respectively.

Analysis of ASR5 nuclear localization signal

Most ASR proteins have a nuclear localization signal (NLS) in their C-terminal region that is composed of two sets of basic amino acids conserved between different members of monocots and dicots and separated by a segment composed of approximately 10 unconserved amino acids, or a continuous set of basic amino acids (Padmanabhan, Dias & Newton 1997; Cakir et al. 2003; Kalifa et al. 2004a). With the exception of ASR2, all of the rice ASR proteins contain basic lysine (K) residues, which is a characteristic of NLSs that is relatively well conserved in lily ASR – LLA23 (Wang et al. 2005) (Supporting Information Fig. S6). These residues are divided into two sets represented by the letters A and B, the latter of which is found only in the ASR5 protein (Supporting Information Fig. S6).


The ASR gene family is well established as being involved in responses to ABA and abiotic stresses, including drought and salinity; ASR genes are also involved in the process of fruit ripening (Carrari et al. 2004; Kalifa et al. 2004a; Konrad & Bar-Zvi 2008). Here, we have demonstrated that, in addition to ASR5, the whole rice ASR gene family is up-regulated at the transcriptional level when rice plants are subjected to high Al concentrations. Rice ASR5 was previously identified as a protein that responded to Al (Yang et al. 2007). Moreover, global transcriptional analyses comparing the Al responses of two wheat (Triticum aestivum L.) near-isogenic lines of variable Al tolerance revealed an early increase in ASR1 transcript expression (Guo et al. 2007). Therefore, based on their expression profiles, the involvement of ASR genes in response to Al is common to both rice and wheat. Although some literature reports have demonstrated the Al-induced ASR gene response, the direct involvement of this gene family in regard to Al tolerance has not been previously reported. Therefore, our effort to compare ASR mRNA accumulation in response to Al in two varieties of rice with contrasting Al sensitivity represents the first study of the functional role of ASR genes in this plant defence mechanism.

Our results reveal that the ASR5 expression levels were not affected by Al treatment in the Al-sensitive cultivar (Taim), but were significantly increased in the Al-tolerant Nipponbare rice cultivar (Fig. 2). ASR5 gene expression was induced in the salt-resistant rice cultivar ‘Pokkali’ in response to NaCl treatment, but not in the sensitive rice cultivar ‘IR29’ (Salakdeh et al. 2002). In our experiments, ASR5 mRNA levels increased significantly in the root apex following exposure to Al (Fig. 3). According to Panda et al. (2009), the root is the most Al-sensitive plant region. Al phytotoxicity blocks cell division, and, as a result, root hair development is inhibited and the root apex becomes swollen and damaged (Clarkson 1965). The root apex and elongation zone are highly sensitive to Al and easily accumulate Al. As a result, more severe physical damage happens in these zones relative to the mature root regions (Panda et al. 2009). In contrast to other stresses, cold treatment did not affect ASR5 expression. Nevertheless, it is possible that ASR5 plays a protective role in leaf tissues rather than roots, as Kim et al. (2009) has reported that ASR5 transcripts are up-regulated in rice leaves after 3 and 6 h of stress. Furthermore, the overexpression of ASR5 increased the cold tolerance of transgenic rice. In other plants, ASR overexpression indicates the potential use of ASR as a gene for plant breeding. Stable expression of tomato ASR1 in tobacco and potato has been shown to increase salt tolerance and influence glucose metabolism, respectively (Kalifa et al. 2004b; Frankel et al. 2007). In Arabidopsis, overexpression of a lily ASR ortholog resulted in increased salt tolerance (Yang et al. 2005). Although many studies involving ASR genes have been performed previously, there are presently no reports characterizing the responses of the entire rice ASR family to abiotic stresses and, more specifically, Al treatment. These genes respond to different abiotic stresses and Al exposure, suggesting that they play an important role in protecting rice plants from environmental stresses.

To further evaluate the functional role of rice ASR genes, we generated transgenic plants carrying an RNAi construct that targeted the ASR5 gene. ASR5 knockdown was confirmed at the mRNA and protein levels (Fig. 4a,b), and ASR1 was also knocked down at mRNA level (Fig. 4a). Both genes were not able to respond anymore in ASR RNAi plants stressed with Al. Due to sequence homology between all members of the ASR family, it is reasonable to expect that the ASR5-RNAi construct could also silence other ASR genes. However, as ASR2, ASR3, ASR4 and ASR6 showed 100-fold less transcripts than ASR1 and ASR5 (Fig 1), we were not able to detect expression of these transcripts in RNAi plants. These results suggested that the RNAi construct was able to reduce expression of the whole ASR gene family. ASR knockdown plants showed a delay in flowering, underwent abnormal panicle development and presented reductions in the seed number relative to NT plants (Fig. 4c). Transgenic plants also showed a reduction in the number of trichomes in leaf, as well as in the palea and lemma (Fig. 4d). Microarray analysis revealed that the ASR gene expression in reproductive tissues is developmentally regulated (Supporting Information Fig S2).

Leaf trichomes may be important to numerous biological functions, and trichome production has been shown to be a variable character with simple inheritance in several species (Karkkainen, Løe & Agren 2004). Furthermore, elevated trichome density has been shown to decrease rates of water loss in several species (Ehleringer & Mooney 1978; Ehleringer 1981; Perez-Estrada, Cano-Santana & Oyama 2000). This may partially explain the susceptibility of the RNAi rice plants to drought stress (Fig. 6).

ASR knockdown plants were more sensitive to Al exposure (Fig. 5a). The primary phenotype of Al exposure was dramatic root inhibition, which is the primary symptom of Al phytotoxicity (Fig. 5b). This result is consistent with the expression pattern observed for the ASR RNAi plants, suggesting that ASR proteins play a functional role in the mechanism of Al tolerance. Using cadmium as a heavy metal stress, both NT and ASR RNAi plants showed strong inhibition of root elongation which suggests that ASR proteins are specific for Al response in rice (Supporting Information Fig. S3). In addition, rice plants were increasingly sensitive to drought, containing reduced water content and percentage humidity relative to control plants (Fig. 6). The involvement of ASR genes in the response to drought stress has been already proposed by other groups (Padmanabhan et al. 1997; Riccardi et al. 1998; Sugiharto et al. 2002); however, our results are the first confirmation of the involvement for these proteins in the mechanism of Al tolerance.

Most ASR proteins have a NLS at their C-terminus (Padmanabhan et al. 1997; Cakir et al. 2003; Kalifa et al. 2004a). Recently, Takasaki et al. (2008) identified ASR5 protein from rice in nuclear and cytosolic compartments of rice leaves, suggesting that ASR5 might localize in these two subcellular compartments. In our stable expression analyses with transgenic rice embryogenic calli expressing ASR5-GFP fusion proteins, we confirmed that ASR5 is present in both the nucleus and cytoplasm (Fig. 7a). These results are consistent with previous reports that have demonstrated that ASR proteins are located mainly in the nucleus where they could regulate specific promoters (Cakir et al. 2003; Wang et al. 2005; Yang et al. 2005).

Surprisingly, in transient expression analysis, ASR5 localized in chloroplasts of rice protoplasts obtained from green leaves (Fig. 7b). This is also the first report of a chloroplastic localization pattern for ASR proteins. The nuclear localization suggests that ASR5 might act as a transcription factor in rice plants as reported in grape (Cakir et al. 2003). ASR5 may do so specifically in the root apex of rice plants, where it could possibly regulate the expression of other genes involved in Al tolerance. However, the precise physiological explanation for the localization of ASR5 in chloroplasts is not clear. It is possible that ASR could regulate the expression of chloroplastic genes. Alternatively, ASR proteins may shuttle to the chloroplast as a chaperone for other proteins typically located there. A disturbance in the architecture of the chloroplast may occur in response to Al toxicity. Moreover, there is a decrease in the photosynthesis ratio as a result of reduced electron transport in photosystem II (Ali et al. 2008 and Zhang, Ryan & Tyerman 2001). It is not surprising that ASR might perform some function in the chloroplast, but it does seem unlikely that it is efficiently shuttled there, as the N-terminal regions of ASR proteins lack the peptide required to target proteins to the chloroplast. Therefore, further studies must be conducted to confirm this surprising subcellular localization biological significance.

To identify potential target genes of the ASR5 mechanism, we used a proteomic approach to compare the proteomic profiles of NT and ASR-RNAi plants. Several chloroplastic proteins were differentially expressed in transformed rice plants. At least three chloroplastic proteins with reduced expression were identified: Chloroplastic lipocalin (gi | 115436780), chloroplast translational elongation factor Tu (gi | 218191089) and heat shock protein 70 (gi | 222631026). Levesque-Tremblay, Havaux & Ouellet (2009) showed through mutational analyses that chloroplastic lipocalin protects Arabidopsis from oxidative stress. In contrast, Latijnhouwers, Xu & Muller (2010) showed an essential role for 70 kDa heat shock proteins in the development of Arabidopsis chloroplasts. Several up-regulated proteins identified in our analysis were also chloroplastic proteins. In total, five proteins exhibited increased expression: 23 kDa polypeptide of chloroplast photosystem II (gi | 164375543), putative photosystem I reaction centre subunit IV (gi | 34394725), putative oxygen-evolving enhancer protein 1, chloroplast precursor (gi | 115436780), chlorophyll a-b binding protein (gi | 108864186), putative superoxide dismutase (SOD) [Cu-Zn] and chloroplast precursor (gi | 42408425). Interestingly, the response of chlorophyll a-b binding protein to Al has already been reported in Arabidopsis: Its expression decreased in response to Al exposure (Richards et al. 1998).

At least two classes of enzymes of the antioxidant system were found to be reduced in ASR-RNAi plants: a putative SOD (gi | 108708142) and the cytosolic APx – APx1 and APx2 (LOC_Os03 g17690) (gi | 115452337). Al cannot catalyse redox reactions on its own, but the involvement of oxidative stress in Al toxicity has been suggested (Jones et al. 2006). According to Jones et al. (2006), Al triggers the accumulation of reactive oxygen species (ROS), such as hydrogen peroxide (H2O2) and superoxide anion (O2-), which correlates positively with Al stress and ROS production. Furthermore, to alleviate oxidative damage during Al stress, several antioxidant enzymes are up-regulated, including SOD, APx and catalase (CAT) (Chen, Qi & Liu 2005; Kumari, Taylor & Deyholos 2008; Sharma & Dubey, 2007 and Panda & Matsumoto 2010).

In canola, the overexpression of manganese SOD conferred Al tolerance (Basu, Good & Taylor 2001). Yin et al. (2010) described that the overexpression of the dehydroascorbate reductase conferred tolerance to Al in tobacco. An Al-tolerant cultivar of Melaleuca trees showed higher antioxidant enzyme activity than their Al-sensitive counterparts during Al stress (Tahara et al. 2008); this demonstrates that an enhanced antioxidant capacity can enhance Al tolerance.

Under most conditions, plants can efficiently scavenge H2O2 by CAT or ascorbate glutathione cycles, wherein APX reduces it to H2O. Ma, Jianmin & Shen (2007) showed that H2O2 production in response to Al stress was more pronounced in an Al-sensitive rice cultivar than in an Al-tolerant cultivar. The tolerant cultivar had significantly higher activities of CAT, APx, dehydroascorbate reductase, glutathione peroxidase, glutathione reductase and concentrations of reduced glutathione higher than those in the Al-sensitive cultivar. In a previous study, our group determined that the knockdown of both cytosolic APx (APx1/2s plants) resulted in a plant that was more tolerant in low concentration of Al (150 µM) (Rosa et al. 2010). Because ASR-RNAi plants present lower level of both cytosolic APxs, we analysed the responses of APx1/2s plants to high concentration of Al. As expected, APx1/2s plants were more sensible to Al stress than the NT plants (Supporting Information Fig. S5a). Besides, APx1 and APx2 transcript levels responded to Al in the tolerant cultivar Nipponbare (Rosa et al. 2010) but did not responded to Al in the sensitive cultivar Taim (Supporting Information Fig. S5b). These data strongly suggest that ASR proteins may directly or indirectly regulate the expression of APx, helping to maintain the redox homeostasis during the onset of stress, and subsequently contributing to the tolerance of toxic Al levels. The increased sensitivity to high Al concentrations exhibited by APx1 and 2 knockdown plants contrasts with a previous report by Rosa et al. (2010), in which APx1/2s conferred elevated Al tolerance. However, in the study by Rosa et al., a lower Al concentration was used (150 µM). Because these plants have higher H2O2 content, they assume a state of constitutive acclimation that permits them to tolerate mild stresses, such as 150 µM Al. However, the low levels of cytosolic APx were not sufficient to compensate for more stressful conditions, resulting in a more Al-sensitive phenotype.

Although ASR proteins from lily present a NLS, they can be found in nuclei and in the cytoplasm (Wang et al. 2005). Among rice ASR proteins, the putative NLS of ASR5 is the only one that contains two lysine residues that comprise the set B (Supporting Information Fig. S6) and is more directly related to the lily ASR. With the exception of ASR2, all other rice proteins present the first two lysine residues conserved in set A (Supporting Information Fig. S6). Wang et al. (2005) showed that the set A is more important in determining the nuclear localization of lily ASR proteins and that replacement of the first two lysine residues with alanine accounts for the mutation that most severely affects nuclear translocation. The fourth lysine residue of the set A in lily ASR protein is replaced by a glutamine residue in the rice ASR6 protein. Nevertheless, ASR6 has a lysine residue at position 202 that regulates subcellular localization, raising the possibility of nuclear localization of this protein. ASR1, ASR3 and ASR4 have all the lysine residues which are similar to those of the set A and identical to lily ASR protein. Due to the importance of this set for the nuclear localization of lily ASR proteins, it is likely that these proteins behave similarly. ASR2 proteins from rice present only the second lysine residue in the conserved set, and lack three of the lysine residues within that grouping in addition to the two residues of the set B, raising the possibility that rice ASR2 localizes in cellular compartments other than the nucleus.

The potential for the ASR5 protein to be a transcriptional factor indicates that this protein may act by regulating the expression of different genes that collectively contribute to the protection of the cell. Future challenges include the identification of the promoters of genes that are targeted by the ASR protein. This would help explain the involvement of ASR in determining the Al tolerance of rice plants. Our results highlight the potential of the ASR gene family as promising subjects for plant biotechnology studies that seek to obtain plants with greater resistance to Al, particularly in acidic soils.


This work was supported by the International Centre for Genetic Engineering and Biotechnology (CRP/06/003), UNESCO, (CAPES:, Fundação de apoio a Pesquisa do Rio Grande do Sul and the Brazilian National Council of Technological and Scientific Development (CNPq). M. Margis-Pinheiro and R. Margis were supported by grants from Conselho Nacional de Desenvolvimento Científico e Tecnológico, CNPq, Brazil (472174/2007-0 and 303967/2008-0).