INTRODUCTION
Aluminium (Al) toxicity is a major limiting factor for crop development and production in acidic soils, which cover >0·30 of the world's arable land (Sasaki et al. Reference Sasaki, Yamamoto, Ezaki, Katsuhara, Ahn, Ryan, Delhaize and Matsumoto2004). After oxygen and silicon, Al is the third most abundant element and the most abundant metal, making up 0·07 of the earth's crust (Tesfaye et al. Reference Tesfaye, Temple, Allan, Vance and Samac2001). Under normal conditions, most Al combines efficiently with oxygen and silicon to form insoluble oxides and aluminosilicates, which are harmless to plants (Ma & Furukawa Reference Ma and Furukawa2003). However, in acidic soils, where the pH values are lower than 5·50, Al is released into soils in the form of the aluminium ion Al3+, which is toxic to plants. Al toxicity first inhibits plant root growth, even at micromolar concentrations (Barcelo & Poschenrieder Reference Barcelo and Poschenrieder2002). Consequently, the plant's water and nutrient uptake decreases, which results in poor growth and biomass production (Ma & Furukawa Reference Ma and Furukawa2003). Al has been shown to be capable of affecting the plasma membrane structure, inhibiting the intracellular Ca2+ signalling pathway (Ryan & Kochian Reference Ryan and Kochian1993), eliciting reactive oxygen species (ROS) (Richards et al. Reference Richards, Schott, Sharma, Davis and Gardner1998; Watt Reference Watt2003) and arresting cell division by down-regulation of cell cycle genes (Chandran et al. Reference Chandran, Sharopova, Ivashuta, Gantt, Vandenbosch and Samac2008a).
Fortunately, Al-resistant and Al-tolerant plant species and cultivars have evolved external and internal Al-detoxification mechanisms (Kochian Reference Kochian1995; Furukawa & Ma Reference Ma and Furukawa2003). The Al-resistant plants with external detoxification mechanisms can exude large amounts of organic acids such as citrate, malate and oxalate (Pellet et al. Reference Pellet, Papernik and Kochian1996; Ryan et al. Reference Ryan, Delhaize and Jones2001) to chelate Al and decrease the uptake of free Al3+ by plant roots. Al-tolerant plants that have internal detoxification mechanisms accumulate high concentrations of organic acids that can bind with Al3+ in cells, thus reducing or eliminating Al toxicity (Ma et al. Reference Ma, Hiradate, Nomoto, Iwashita and Matsumoto1997, Reference Ma, Hiradate and Matsumoto1998; Ma & Hiradate Reference Ma and Hiradate2000). The Al-activated release of organic acid anions from root tips has been characterized as the most important Al-resistant mechanism in plants (Sasaki et al. Reference Sasaki, Yamamoto, Ezaki, Katsuhara, Ahn, Ryan, Delhaize and Matsumoto2004; Hoekenga et al. Reference Hoekenga, Maron, Piñerios, Cançado, Shaff, Kobayashi, Ryan, Dong, Delhaize, Sasaki, Matsumoto, Yamamoto, Koyama and Kochian2006). Differential display reverse transcription-polymerase chain reaction (DDRT-PCR), cDNA-amplified fragment length polymorphisms (cDNA-AFLP), suppression subtractive hybridization (SSH), DNA microarray and proteomic approaches have been used widely to analyse plant responses to Al stress. Using these powerful tools, Al-responsive genes have been identified in the roots of Arabidopsis thaliana (Richards et al. Reference Richards, Schott, Sharma, Davis and Gardner1998; Kumari et al. Reference Kumari, Taylor and Deyholos2008; Goodwin & Sutter Reference Goodwin and Sutter2009), sugarcane (Watt Reference Watt2003), rye (Milla et al. Reference Milla, Butler, Huete, Wilson, Anderson and Gustafson2002), soybean (Ermolayev et al. Reference Ermolayev, Weschke and Manteuffel2003), rice (Mao et al. Reference Mao, Yi, Yang, Zheng, Wu, Liu and Wu2004) and Medicago truncatula (Chandran et al. Reference Chandran, Sharopova, Ivashuta, Gantt, Vandenbosch and Samac2008a). Combining these approaches with the isolation of Al-hypersensitive mutants and positional cloning techniques, a number of Al-resistance genes from several Al-resistant or Al-tolerant species, such as wheat (Sasaki et al. Reference Sasaki, Yamamoto, Ezaki, Katsuhara, Ahn, Ryan, Delhaize and Matsumoto2004), rape (Ligaba et al. Reference Ligaba, Katsuhara, Ryan, Shibasaka and Matsumoto2006), Arabidopsis (Kobayashi et al. Reference Kobayashi, Hoekenga, Itoh, Nakashima, Saito, Shaff, Maron, Piñeros, Kochian and Koyama2007) and sorghum (Magalhaes et al. Reference Magalhaes, Liu, Guimarães, Lana, Alves, Wang, Schaffert, Hoekenga, Piñeros, Shaff, Klein, Carneiro, Coelho, Trick and Kochian2007) have been identified.
It is well known that both M. truncatula and M. sativa are sensitive to soil acidity and Al toxicity (Sledge et al. Reference Sledge, Pechter and Payton2005). Recently, a DNA microarray approach was used to analyse Al-responsive genes in the roots of M. truncatula cultivar Jemalong genotype A17 (Chandran et al. Reference Chandran, Sharopova, Ivashuta, Gantt, Vandenbosch and Samac2008a). A multi-drug and toxin efflux protein (MATE) gene showing high homology to the citrate efflux transporter from other plant species was found to be highly up-regulated under Al stress conditions. However, up-regulated MATE expression did not confer Al-resistance on M. truncatula.
M. sativa is not only the world's major forage, producing more protein per hectare than grain and oilseed crops (Sledge et al. Reference Sledge, Bouton, Dall'agnoll, Parrott and Kochertc2002), but it is also a model plant for genetic engineering studies (Dragiiska et al. Reference Dragiiska, Djilianov, Denchev and Atanassov1996). However, the sensitivity of M. sativa to Al (Kamp-Glass et al. Reference Kamp-Glass, Powell, Reddy, Baligar and Wright1993) limits its breeding and growth in acidic soils. Recently, a study performed by Tesfaye et al. (Reference Tesfaye, Temple, Allan, Vance and Samac2001) tried to improve its Al-resistance by over-expressing a nodule-enhanced malate dehydrogenase (neMDH). Nevertheless, the results were not sufficient to justify its use on an agricultural scale. To better decipher the mechanism for the sensitivity of M. sativa to Al, the present study investigated the changes in gene expression (using the SSH method) and the physiological characters of an Al-sensitive M. sativa cultivar, Yumu No. 1, in response to Al stress. The results are useful for suggestions of more effective and feasible strategies for genetic engineering to enhance its Al-resistance and support sustainable land use (Humphreys Reference Humphreys2005).
MATERIALS AND METHODS
Plant cultivation and Al treatments
M. sativa L. cv. Yumu No. 1 (henceforth referred to as YM1), developed by Professor Yu at the Southwest University of China, was used as the plant material throughout the experiments. The seeds of YM1 were placed in tap water overnight for germination under dark at 25°C. Seedlings with roots c. 10 mm long were sown on a floating mesh in a polypropylene pot with full nutrient solution (5 litres) as previously described (Lipton et al. Reference Lipton, Blanchar and Blevins1987), and the solution was renewed every other day. The seedlings were grown in a tissue culture room at 25°C under constant light (100 μmol/m2/s) for 5 days. Before Al treatment, the 5-day-old seedlings were pre-grown overnight in a 0·5 mm CaCl2 solution (pH 4·2) at 25°C under constant light as described above.
For the Al sensitivity experiment, the seedlings were then transferred into a 0·5 mm CaCl2 solution containing either 0 (control), 3, 5, 10, 20, 30, 50 or 100 μm AlCl3 (pH 4·2), respectively, and grown at 25°C under constant light as described above. The root length was measured with a ruler before Al treatment (0 h) and after 24 h of Al treatment (24 h). The relative root growth (RRG) was defined as the ratio of the net root growth in plants treated with Al compared with the control. As chrome azurol sulphonate (CAS) can bind with Al efficiently, it has been widely used in the determination of Al concentration (Sombra et al. Reference Sombra, Luconi, Silva, Olsina and Fernandez2001) and the CAS staining method was used as an indicator of the Al bound to the YM1 root tips. After Al treatment, the roots of YM1 were stained with CAS solution for 30 min, washed thoroughly with distilled water and photographed. Spectrophotometric determination of Al content in roots is described later.
For SSH analysis, the plants were grown in a full nutrient solution for 3 weeks under glasshouse conditions at 30/25°C day/night temperatures and 12 h of light (1200 μmol/m2/s). The nutrient solution was renewed every other day. After the roots reached a length of 50–100 mm, the plants were treated with Al as follows: the seedlings were pre-grown overnight in a 0·5 mm CaCl2 solution (pH 4·2) at 25°C under constant light as described above. Then, the seedlings were transferred into a 0·5 mm CaCl2 solution containing 5 μm AlCl3 (pH 4·2) for Al treatment at 25°C under constant light as described above. The roots of the plants treated in the CaCl2 solution without AlCl3 were used as the control; these plants were maintained in the CaCl2 after the others were transferred to AlCl3 and remained there for the same length of time as those in AlCl3. The root tips were sampled at 10 and 30 min, 2, 4, 6, 8, 10 and 24 h after treatment. The samples were immediately frozen in liquid nitrogen and stored at –80°C for RNA extraction.
Total RNA extraction and mRNA isolation
Frozen root samples were ground in liquid nitrogen, and total RNA was extracted with Trizol reagent (Invitrogen, USA) according to the manufacturer's instruction. The total RNA was treated with DNase I (Promega, USA) for 30 min at 37°C to remove the residual genomic DNA (gDNA) and then extracted with phenol/chloroform (1/1). The quality and quantity of the extracted total RNA samples were examined by electrophoresis on a 0·02 g/ml agarose/1×TAE buffer, EtBr-stained gel and by spectrophotometry (Shimadzu, Japan). The total RNA samples showing no gDNA in the EtBr-stained gel and with an A260/280 ratio of 1·9–2·0 (an indication of RNA purity) were used for mRNA isolation. To identify genes differentially expressed over the whole 24-h treatment period but not at a specific time point under the Al stress conditions, the total RNA samples extracted from the roots treated with 5 μm AlCl3 for 10 and 30 min, 2, 4, 6, 8, 10 and 24 h were mixed together and used as the tester. The total RNA samples extracted from the control roots treated with 0·5 mm CaCl2 solution for 10 and 30 min, 2, 4, 6, 8, 10 and 24 h were also mixed together and used as the driver. The mRNA pool of Al-treated roots and control roots was isolated from the total RNA mixtures of tester and driver, respectively, using the Oligotex mRNA Midi Kit (Macherey-Nagel, Germany) according to the protocol described by the manufacturer. The quality and quantity of the isolated mRNA samples were determined by electrophoresis on a 0·02 g/ml agarose/1×TAE buffer, EtBr-stained gel and by spectrophotometry. The mRNA samples with good quality were used for cDNA synthesis.
Construction of the SSH cDNA library
The cDNA was reverse transcribed from the isolated mRNA samples. The quality of the synthesized cDNA was examined by PCR analysis with alfalfa 18S rRNA primers (see Table 1). The cDNA samples showing only one PCR-amplified band with the expected size of the 18S cDNA was used for SSH by using the Clontech PCR Select-cDNA Subtraction Kit (Clontech, USA) according to the protocol described by the manufacturer. The tester cDNA was produced from the mRNA pool of Al-treated roots. The driver cDNA was produced from the mRNA pool of control roots. Briefly, the tester and driver cDNA were digested with Rsa I, respectively. The tester cDNA was split into two pools and then ligated with adaptor 1 or adaptor 2R, respectively. After two rounds of hybridization and PCR amplification with adaptor 1 and adaptor 2R as the primers, the differentially expressed cDNAs were normalized and enriched. Then the second PCR products were purified and concentrated. The cDNA fragments for differentially expressed genes were inserted into the pGEM-T easy vector (Promega, USA) and transformed into Escherichia coli DH5α competent cells.
Table 1. Primer sequences used in the RT-PCR analysis
AAT, cationic amionic acid transporter; ANN, annexin-like protein; GEP, germin-like protein; MCA, type IIB calcium ATPase (MCA2) gene; MsALMT1, Al-activated malate transporter; PPCK, phosphoenolpyruvate carboxykinase; PR10, class 10 PR protein; SQS, farnesyl-diphosphate farnesyltransferase; STOP1, sensitive to proton rhizotoxicity 1.
Amplification of cDNA inserts
cDNA clones, freshly grown overnight at 37°C, were used as the PCR templates. The cDNA inserts were amplified using the nested PCR primers 1 and 2R provided in the PCR Select-cDNA Subtraction Kit. The PCR reaction mixtures (20 μl) contained 16·1 μl of sterile water, 2 μl of 10×Taq reaction buffer (Tiangen, Beijing, China), 0·02 μl of adaptor 1 and adaptor 2R primers (100 μm), 0·2 μl of dNTPs (10 μm each; Tiangen, Beijing, China), 1 unit of Taq DNA polymerase (Tiangen, Beijing, China) and 2 μl of the bacterial culture. The PCR included an initial heating at 94°C for 5 min followed by 30 cycles of 94°C for 30 s, 65°C for 30 s and 72°C for 1·5 min and then a final hold at 4°C. PCR products were run on a 0·01 g/ml agarose/1×TAE buffer, EtBr-stained gel to validate their quality and quantity.
Expression sequence tag (EST) sequence analysis
The cDNA clones with inserts ⩾500 bp were sequenced using T7 promoter primer. The low-quality regions and the vector and adaptor sequences were removed to obtain the EST sequences. The ESTs were screened against the National Center for Biotechnology Information (NCBI) non-redundant nucleotide database or protein database by using the BLASTN or BLASTX program, respectively (available from http://www.ncbi.nlm.nih.gov/guide/data-software/#All_; verified 8 Dec 2010). ESTs with ⩾0·70 homology with the nucleotide database or ⩾0·30 homology with the protein database were considered as known function genes. ESTs showing significant homology (E<0·001) with sequences of unknown function in the M. truncatula genomic database or other databases were defined as unknown functions. The BLAST program identifies similar protein or nucleotide sequences to a given target sequence and provides an S score as a measure of the similarity of the query to the sequence shown. It also provides an E-value as a measure of the reliability of the S score. Thus, the definition of the E-value is: the probability due to chance that there is another alignment with a similarity greater than the given S score. ESTs showing no significant homology with the M. truncatula genomic sequence or other databases were defined as new ESTs. To obtain a higher quality of annotation classification, the ESTs were functionally annotated manually based on comparisons to the Uniprot (http://www.uniprot.org/; verified 8 Dec 2010) and Gene Ontology (GO) databases (http://www.geneontology.org/; verified 8 Dec 2010). All EST sequences were submitted to the EST division of the gene bank to obtain their accession numbers (Tables 2 and 3). Also, see supplementary Tables S1–S3 available at http://journals.cambridge.org/AGS).
Table 2. Selected known Al-responsive genes
* The sequence blast with the NCBI non-redundant nucleotide database using BLASTN.
† The sequence blast with the NCBI non-redundant protein database using BLASTX.
Table 3. Selected novel genes potentially associated with Al stress
* and † are described as in Table 2.
Reverse transcription-polymerase chain reaction (RT-PCR) analysis
The roots of YM1 were treated with 5 μm AlCl3 for 2, 4, 6, 10 and 24 h at 25°C under constant light as described above. The roots treated with 0·5 mm CaCl2 without Al for 6 h were used as the control (CK). The total RNA isolated from root samples with Trizol reagent was treated with DNase I (Promega, USA) for 30 min at 37°C to remove the residual gDNA and then extracted with phenol/chloroform (24/1). First-strand cDNA was synthesized from the total RNA (5 μg) using M-MLV reverse transcriptase (Promega, USA) according to the manual. The primer sequences used in the RT-PCR analysis are shown in Table 1. The PCR reaction mixtures (20 μl) contained 16·1 μl of sterile water, 2 μl of 10×Taq reaction buffer, 0·02 μl of primers (100 μm), 0·2 μl of dNTPs (10 μm each), 1 unit of Taq DNA polymerase (Tiangen, Beijing, China) and 2 μl of cDNA (200 ng/μl). The PCR products were examined by electrophoresis on a 0·01 g/ml agarose/1×TAE buffer, EtBr-stained gel. Alfalfa 18S rRNA was used as an internal control. Each PCR analysis was repeated thrice.
Full-length MsALMT1 cDNA isolation and sequence analysis
For full-length MsALMT1 cDNA isolation, the total RNA was extracted using Trizol reagent (Invitrogen, USA) from YM1 roots treated with 5 μm AlCl3 under constant light as described above for 24 h at 25°C. First-strand cDNA was synthesized from 5 μg of total RNA. The full-length cDNA of MsALTMT1 was amplified by PCR using the forward primer ALMT1F (ATGGTGTCTGAACCAAATTCAAG) and the reverse primer ALMT1R (TAGTTAATTATAATAACATGTTG). The PCR was performed with Taq plus DNA polymerase (Tiangen, Beijing, China). The cDNA fragment of MsALTMT1 was subcloned into a TA cloning vector, pMD-18T (TaKaRa, Dalian, China), to generate pMD-MsALTMT. The pMD-MsALTMT was subjected to sequencing. The amino acid sequence of MsALTMT1 was predicted using the Genetyx program (Software Development, Tokyo, Japan). Phylogenetic relationships between MsALMT1 and ALMT1 family proteins were analysed using the Dnaman program (Version 6.0.40, Lynnon Biosoft, USA). The dense alignment surface (DAS) Program (Cserzo et al. Reference Cserzo, Wallin, Simon, Von Heijne and Elofsson1997) was used for predicting the location and transmembrane domains of MsALMT1.
Hydrogen peroxide (H2O2) and malondialdehyde (MDA) content analysis
The roots of YM1 were treated with 5 μm AlCl3 for 2, 4, 6, 10 and 24 h under constant light as described above at 25°C. For the determination of H2O2 content, the roots (100 mg) were homogenized in 1 ml of tris(hydroxymethyl)aminomethane (Tris-HCl; 50 mm, pH 7·0). The homogenate was centrifuged at 13 000 g (4°C) for 20 min. The H2O2 content was measured as described by Gay & Gebicki (Reference Gay and Gebicki2003). The MDA content was determined with the 2-thiobarbituric acid (TBA) method (Baccouch et al. Reference Baccouch, Chaoui and Ferjani1998). The roots (100 mg) were homogenized in 1 ml of 0·10 mg/ml (w/v) trichloroacetic acid (TCA) and centrifuged at 13 000 g (4°C) for 20 min. Then 0·5 ml aliquots of the supernatants were added to 0·4 ml of 0·01 mg/ml (w/v) TBA in 0·10 mg/ml (w/v) TCA and the tubes were incubated in a water bath at 100°C for 15 min with occasional shaking. The reaction was stopped by placing the reaction tubes into an ice bath. After centrifugation at 13 000 g (4°C) for 20 min, the absorbance of the supernatant was measured at 532 and 450 nm.
Determination of malate content
The roots of YM1 were treated as described for the H2O2 and MDA analyses. Roots and exudates were harvested and collected after Al treatments of 0 (control), 2, 4, 6, 10 and 24 h. The estimation of malate contents in the YM1 roots and exudates was performed using the enzymatic method according to Wang et al. (Reference Wang, Zhao, Yi, Li, Yu and Chen2010).
Measurement of Al content
The roots of YM1 were treated as described for the H2O2 and MDA analyses. The root apices (six tips for each sample) were ground with liquid nitrogen. The powder was suspended in 1·5 ml of 2 m hydrochloric acid (HCl) at room temperature for 24 h with occasional shaking (100 rpm). The Al contents in the YM1 root tips were measured spectrophotometrically with pyrocatechol violet (PCV) as described by Zheng et al. (Reference Zheng, Lin, Yang, Liu and Tang2004) with some modification. Assays were performed by mixing 0·05 ml of the sample solution with 1·35 ml of deionized water and then 0·1 ml of the PCV (37·5 mg/ml, w/v) reagent and 0·5 ml of Tris-HCl (50 mm, pH 6·0) buffer were added to start the reaction. The absorbance was read at 590 nm after 15 min.
Statistical analysis
Experiments were conducted from 3 to 12 replicates and data were expressed as means±s.d. For statistical analysis, one-way ANOVA (SPSS 13.0 program; Norusis Reference Norusis2005) and the t-test were used to determine the significance at P<0·05.
RESULTS
Phenotypic performance of YM1 roots under Al stress and Al content in the root tips
The Al sensitivity and phenotypic performance of YM1 were tested using the RRG method. Increasing Al concentration led to lower RRG (linear regression R 2=0·7651, P<0·001, Fig. 1a). The inhibitions at the Al concentrations of 10, 20 and 30 μm were c. 0·30, 0·40 and complete inhibition was observed at 50 and 100 μm AlCl3. The results were similar to that of Al-sensitive M. truncatula A17 (Chandran et al. Reference Chandran, Sharopova, Vandenbosch, Garvin and Samac2008b). CAS staining is an indicator of the toxic effects of Al on the apices of the roots, and the staining degree of the root tips provides a semi-quantitative measurement of the Al content in the root tips. The YM1 roots were stained with CAS after Al treatments (Fig. 1c). No visible staining was observed in YM1 roots treated with 3, 5 and 10 μm Al, respectively. Weak staining appeared in the root tips after treatment with 20 μm Al, and visible staining was observed in the root tips after treatment with 30 μm Al. The roots were stained deeply after treatment with 50 μm Al. When the concentration was increased up to 100 μm, the root tips of YM1 were fully stained and exhibited severe damage. These results confirm that YM1 is Al sensitive.
Fig. 1. The effect of Al on YM1 root growth and Al content in YM1 roots after Al treatment. (a) RRG of 5-day-old YM1 seedlings. Values are means±s.d. (n=12). (b) The Al content in YM1 roots after treatment with 5 μm Al (pH 4·2) for 2, 4, 6, 10 and 24 h, respectively. Values are means±s.d. (n=3). (c) Chrome azurol S staining of YM1 roots after exposure to 0, 3, 5, 10, 20, 30, 50 and 100 μm AlCl3 (pH 4·2) for 24 h.
The Al content in the YM1 roots was measured after different periods of treatment with 5 μm Al. There was a significant increase in Al content with increasing treatment time (linear regression R 2=0·8233, P<0·005, Fig. 1b). The Al content in the roots was significantly higher after Al treatment as compared to the control (without Al treatment). Longer periods of exposure led to greater Al content in the root tip. This suggests that Al was efficiently absorbed by the YM1 root tips during the treatment time.
Construction of a forward SSH cDNA library for YM1 under Al stress
Based on the above results, YM1 roots were treated with 5 μm AlCl3 for a period of 10 min to 24 h in order to construct the SSH cDNA library. The library constituted 1209 clones. Colony PCR with the nested primers was used to determine the length of cDNA inserts. The results show that 0·90 were positive clones, and the lengths of the cDNA inserts ranged from 200 to 1200 bp, with an average length of c. 500 bp (data not shown). A total of 326 clones with inserts ⩾500 bp were picked for sequencing, and 291 high-quality ESTs were obtained (Tables 2 and 3). Also, see supplementary Tables S1–S3 available at http://journals.cambridge.org/AGS).
Functional classification of the Al-responsive genes in the SSH cDNA library
All 291 ESTs obtained from the SSH cDNA library were screened against the NCBI nucleotide and protein databases. The results indicated that 229 ESTs (0·79) showed similarities to proteins of known function (Tables 2 and 3). Also, see supplementary Tables S1 and S2 available at http://journals.cambridge.org/AGS), 53 ESTs (0·18) showed similarities to genes or genomic sequences of unknown function, and nine ESTs (0·3) were new ESTs (see supplementary Table S3 available at http://journals.cambridge.org/AGS). All 229 ESTs with known functions were classified into eight functional categories (Fig. 2). The genes related to signal transduction and transcription accounted for the largest group, with a corresponding ratio of 0·22. The second categories were genes related to defence, stress and cell death (0·19), and metabolism and energy (0·17), respectively. Protein synthesis and protein destination/storage represented the third categories, with the corresponding ratios of 0·14 and 0·09, respectively. Genes related to transport, cell structure and cell growth and others represented the smallest categories, with the corresponding ratios of 0·80, 0·60 and 0·50, respectively. The genes with known function were carefully compared with the Al-responsive genes that have been reported in other plants. The results showed that 137 ESTs were previously reported to be Al-responsive genes (Table 2). Also, see supplementary Table S2 available at http://journals.cambridge.org/AGS). The other 92 ESTs were not reported and thus were referred to as novel Al-responsive genes (Table 3). Also, see supplementary Table S2 available at http://journals.cambridge.org/AGS).
Fig. 2. The distribution of differentially expressed genes in YM1 roots under Al stress. Two hundred and twenty-nine known ESTs were grouped into eight functional categories according to their putative functions. The proportions of the ESTs representing each category are shown.
Up-regulation of known Al-responsive genes
The SSH data (Table 2). Also, see supplementary Table S2 available at http://journals.cambridge.org/AGS) showed that 42 known Al-responsive genes were involved in signal transduction and transcription. The Al and H+ tolerance-related C2H2-type zinc finger protein (sensitive to proton rhizotoxicity 1 (STOP1)), phosphatase 2C, calcineurin B-like-interacting protein kinase, GTP-binding protein, calcium and calcium/calmodulin-dependent serine/threonine-protein kinase DMI-3 and mitogen-activated protein kinase were up-regulated. However, receptor family proteins and cell wall-associated protein kinases, which have been shown to be related to Al resistance, did not show up-regulation in the YM1 roots. The stress, defence and cell death category contains 30 known Al responsive genes. Typical Al-induced marker genes, such as genes related to pathogen stress (class 10 PR protein), disease-resistance proteins, PR10-1 protein and thaumatin-like protein PR-5b, antioxidants (glutathione S-transferase (GST), peroxidase (POD), superoxide dismutase (SOD), germin-like proteins (GEP) and thioredoxin h), cell death (putative senescence-associated protein, cysteine protease and cysteine proteinase precursor) and metal resistance were present in the library. However, blue copper-binding protein and GDP dissociation inhibitor, which have been determined to be Al resistance-associated genes, were absent. Al up-regulated genes related to metabolism and energy included 20 known Al-responsive genes. Several genes encoding important enzymes involved in the metabolism of flavonoids, such as chalcone synthase, chalcone reductase and chalcone isomerase, were up-regulated. Genes related to lipid metabolism such as fatty acid desaturase 1-like and lipase were also up-regulated. However, genes associated with organic acid metabolism, such as citrate synthase (CS), phosphoenolpyruvate carboxylase (PEPC) and malate dehydrogenase (MDH), were absent from the library. The up-regulation of several known Al-responsive transporter genes, such as Zn transporter, cationic amino acid transporter (AAT) and vacuolar H+-ATPase subunit A, was observed. It is worth noting that a putative Al-activated malate transporter (ALMT1) gene, which had been confirmed to be an Al-resistance gene, was also up-regulated. However, genes including multi-drug and toxin efflux protein (MATE, Al-activated citrate transporter), ATP-binding cassette transporter (ABC transporter), Mg transporter, K transporter and mitochondrial ATPase did not show up-regulation. Known Al-responsive genes associated with cell structure modification and cell growth, including caffeic acid O-methyltransferse, xyloglucan endortransglycosylase, expansion-like protein, annexin-like protein (ANN) and chitinase, were up-regulated. However, genes encoding the cytoskeleton, coumarated: CoA ligase, cellulose microfibril and glycosylphosphatilylinositol-anchored protein were not induced in YM1 roots.
Up-regulation of novel genes associated with Al stress
Of the genes with known functions, 92 novel genes (Table 3). Also. see supplementary Table S2 available at http://journals.cambridge.org/AGS) potentially associated with Al stress were up-regulated. There were 18 genes related to metabolism and energy. Gene encoding enzymes involved in lipid, carbohydrate and serine metabolism, such as acyl desaturase, putative serine decarboxylase, serine acetyltransferase and triosephosphate isomerase were up-regulated. In the category of genes related to defence, stress and cell death, genes encoding ABA 8′-hydroxylase, ABA stress ripening protein, epoxide hydrolase and the peroxisomal short-chain dehydrogenase/reductase family protein were induced. Of the signal transduction and transcription category, genes related to salt- and drought-induced ring finger 1, the WD40-like protein and WD-40 repeat family of proteins were induced. The up-regulation of several novel transporter genes, including inositol transporter 2, organic anion transmembrane transporter and substrate-specific transmembrane transporter, were observed.
Validation of SSH data by RT-PCR analysis
RT-PCR is an economic and a simple method of gene expression profile analysis. Nine genes (Table 1), including two stress-related genes (GEP and PR10), three genes for metabolism and energy (SQS, PPCK and MCA), one gene for signal transduction and transcription (STOP1), one gene for cell structure and cell growth (ANN) and two genes for transporters (AAT and MsALMT1), were selected for RT-PCR analysis. RT-PCR analysis was performed as described in the Materials and Methods. The RT-PCR data (Fig. 3) suggested that the transcription profiles of the selected nine genes were all up-regulated. The results indicated that the data of the SSH library are reproducible and reliable.
Fig. 3. Expression profile analysis of genes selected from the SSH library by RT-PCR. RT-PCR analysis was performed as described in the Materials and Methods. Number of PCR cycles was shown on the right side. CK (control) indicates the sample without Al treatment. Al treatment indicates samples treated with 5 μm Al (pH 4·2) for 2, 4, 6, 10 and 24 h. 18S rRNA was used as an internal control. AAT, cationic amionic acid transporter; ANN, annexin-like protein; GEP, germin-like protein; MCA, type IIB calcium ATPase (MCA2) gene; MsALMT1, Al-activated malate transporter; PPCK, phosphoenolpyruvate carboxykinase; PR10, class 10 PR protein; SQS, farnesyl-diphosphate farnesyltransferase; STOP1, sensitive to proton rhizotoxicity 1.
Cloning and expression analysis of full-length MsALMT1 cDNA
ALMT1s are proteins that contain the UPF0005 motif and possess five to seven transmembrane domains. An ALMT-like gene (MtALMT1, ABD32183) was found in the M. truncatula genomic sequence (Delhaize et al. Reference Delhaize, Gruber and Ryan2007). Using the BLASTX program, it was found that MtALMT1 also contains the UPF0005 motif. Based on the coding region sequence of MtALMT1, the full-length (1347 bp) MsALMT1 ORF (GU550122) was cloned from the YM1 roots by PCR amplification. Sequence analysis indicates that the MsALMT1 ORF encodes a deduced protein with 448 amino acid residues. Based on the sequences of the ALMT1 protein family published by the NCBI, a phylogenetic tree was constructed (Fig. 4). As shown in Fig. 4, all 10 sequences can be classified into three main groups, namely A, B and C at a genetic identity score of 0·40. At a genetic identity score of 0·60, group A was further divided into four subgroups (subgroups A1–A4). The deduced amino acid sequence of MsALMT1 displayed 0·96, 0·46, 0·43 and 0·42 homology with MtALMT1, AtALMT1, GmALMT1 and TaALMT1, respectively. The DAS program analysis predicted that the MsALMT1 protein contains seven transmembrane domains (see supplementary Fig. S1 available at http://journals.cambridge.org/AGS). RT-PCR analysis confirmed that the transcript level of MsALMT1 in the YM1 roots was induced under Al stress (Fig. 3), and no expression of MsALMT1 was detected in the shoots of the plants treated with or without Al (data not shown). The expression pattern of MsALMT1 is similar to that of ALMT1s in other plants.
Fig. 4. The phylogenetic relationship between MsALMT1 and its homologues in other species. AtALMT1, A. thaliana AAF22890; MtALMT1, M. truncatula ABD32183; GmALMT1, Glycine max EU586179; BnALMT1, Brassica napus BAE97280; BoALMT1, Brassica oleracea AAW81734; OsALMT1, O. sativa NP-001042433; ScALMT1, Secale cereale ABA62397; TaALMT1, Triticum aestivum BAD10882; ZmALMT1, Zea mays ABC86748; HvALMT1, Hordeum vulgare EF424084.
Malate content in the YM roots and its exudation rate from the YM1 roots under 5 μm stress conditions
To validate the function of MsALMT1 in the YM1 roots, malate exudation was determined after 0, 2, 4, 6, 10 and 24 h of Al treatment. The malate exudation-rate steeply increased during the period of 2–6 h of Al treatment (Fig. 5a). The data confirm the function of MsALMT1 in the YM1 roots. According to the criteria described by Ma et al. (Reference Ma, Ryan and Delhaize2001), the malate exudation pattern in YM1 was characterized as Pattern II with a 6-h lag phase before the maximum period of malate exudation. However, YM1 could not maintain such a high level of Al-induced malate exudation beyond the 6–10 h and 10–24 h periods of Al treatment (Fig. 5a). Moreover, the SSH data showed that the genes involved in organic acid synthesis, including PEPC, CS and MDH were not up-regulated in YM1 roots. To verify whether the decrease in malate exudation was associated with the internal malate synthesis, the malate content in YM1 roots was determined. As shown in Fig. 5b, YM 1 root-tip tissue had higher malate content (17·73 nmol/root tip) before Al treatment (control). However, the malate content was gradually reduced after Al treatment. At 6, 10 and 24 h, malate contents in the YM1 root tips decreased by approximately 0·50, 0·38 and 0·37, respectively, compared to that in the control. The results suggested that the decrease in Al-induced malate exudation might be due to a reduction of malate synthesis in YM1 roots under Al stress conditions.
Fig. 5. The malate exudation rate (a) and content (b) of YM1 roots under 5 μm Al stress conditions. YM1 roots were treated with 5 μm Al (pH 4·2) for 0, 2, 4, 6, 10 and 24 h, respectively, as described in the Materials and Methods. Values are means±s.d. (n=3).
Oxidative stress levels in YM1 roots under 5 μm Al stress
A number of genes related to antioxidants were shown to be up-regulated in YM1 roots according to the SSH data (see Table 2). To better understand the molecular mechanisms within YM1 in response to Al stress, the oxidative stress level in the YM1 roots under stress of 5 μm Al was investigated. MDA and H2O2 contents are routinely used as the indicators of oxidative stress level. The contents of MDA and H2O2 in the YM1 roots were determined after treatment with 5 μm Al and both of them showed significant linear relationships with Al treatment time (Fig. 6). As shown in Fig. 6a, an H2O2 burst was triggered within 2 h after Al treatment. The H2O2 content was maintained at a higher level than that in the control throughout the entire treatment period. Similar changes were also observed with regard to the MDA content. The MDA content in YM1 roots was also markedly increased after exposure to 5 μm Al (Fig. 5b). The increase in H2O2 and MDA contents suggested that Al might induce ROS formation and lead to lipid peroxidation. Therefore, oxidation stress was severe in YM1 root cells under the 5 μm Al stress conditions. The up-regulation of antioxidant-related genes might be favoured of protecting root cells from damage by oxidative stress induced by Al toxicity.
Fig. 6. Effects of 5 μm Al stress on H2O2 (a) and MDA (b) contents in YM1 roots. Values are means±s.d. (n=3). YM1 roots were treated as described in Fig. 5.
DISCUSSION
In the present study, the SSH strategy was used to analyse gene expression changes in response to 5 μm Al stress in the roots of an Al-sensitive M. sativa L. cv YM1, which is an important alfalfa cultivar in China. The data obtained from the SSH cDNA library showed the up-regulation of 137 known Al-associated genes that had been found in other Al-tolerant species and 92 novel Al response genes under 5 μm Al stress conditions. The results suggest that the Al-sensitive YM1 shares some Al response mechanisms with Al-tolerant species but still has some mechanisms that are different from those in Al-resistant plants.
It is well known that antioxidant genes are up-regulated in response to Al stress in Arabidopsis (Richards et al. Reference Richards, Schott, Sharma, Davis and Gardner1998), barley (Guo et al. Reference Guo, Zhang and Zhang2007), sugarcane (Watt Reference Watt2003) and wheat (Darkó et al. Reference Darkó, Ambrus, Stefanovits-Banyai, Fodor, Bakos and Barnabas2004) roots. The excessive production of ROS induced by Al causes oxidative damage to cell membrane lipids and DNA strands (Guo et al. Reference Guo, Zhang and Zhang2007). The cellular damages can be protected by antioxidant enzymes, such as GST, POD, catalase (CAT) and SOD. Darkó et al. (Reference Darkó, Ambrus, Stefanovits-Banyai, Fodor, Bakos and Barnabas2004) showed that the activities of CAT and GST enzymes in Al-tolerant plants were greater than those in Al-sensitive plants, resulting in less accumulation of ROS in the roots of Al-tolerant plants. Furthermore, the over-expression of Al response genes, such as tobacco GST gene (parB) and POD gene from Arabidopsis (AtPox), conferred transgenic Arabidopsis with not only oxidative resistance but also Al tolerance (Ezaki et al. Reference Ezaki, Gardner, Ezaki and Matsumoto2000). These results suggest that antioxidant enzymes may play a role in the detoxification of Al-induced ROS in Al-resistant plants. GEPs are ubiquitous plant proteins, and most of them have antioxidant enzyme activities by binding with Mn/Zn/Cu. The expression of GEPs is considered to be associated with biotic and abiotic stresses (Membre et al. Reference Membre, Bernier, Staiger and Berna2000; Gucciardo et al. Reference Gucciardo, Wisniewski, Brewin and Bornemann2007). The analysis of H2O2 and MDA contents confirmed that 5 μm Al stress resulted in pronounced elevation of H2O2 and MDA levels in YM1 roots. MDA formation is used as a general indicator of lipid peroxidation. This suggests that 5 μm Al stress did cause a severe oxidative stress in YM1 roots. The SSH data revealed the up-regulation of a number of genes encoding antioxidant enzymes (GST, POD, SOD and GEP) in YM1 roots by Al stress. This might improve its resistance to the oxidative stress caused by Al toxicity.
Ma et al. (Reference Ma, Ryan and Delhaize2001) described two patterns for Al-stimulated secretion of organic acids from plant roots. For pattern I, Al directly activates an anion channel on the plasma membrane and plants release organic acids immediately after Al treatment, whereas in pattern II, Al interacts with the cell wall and, perhaps via an unknown signal pathway, activates the transcription of genes involved in the metabolism of organic acids or their transports across the plasma membrane. The exudation of organic acids is mediated by membrane-localized organic acids transporters (Delhaize et al. Reference Delhaize, Gruber and Ryan2007), and these proteins belong to two families, ALMT1 and MATE. Recently, genes encoding the ALMT1 and MATE protein families were identified from wheat (TaALMT1), Arabidopsis (AtALMT1 and AtMATE), rye (ScALMT1), rape (BnALMT1 and BnALMT2) and sorghum (SbMATE) (Delhaize et al. Reference Delhaize, Gruber and Ryan2007). The full-length cDNA of the putative ALMT1 gene, MsALMT1, was also cloned from the YM1 roots, and the sequence analysis confirms that it belongs to the ALMT1 transporter family. In Arabidopsis and rice, the organic acid transporter genes are regulated by the STOP1 and ART1 transcription factors, respectively (Liu et al. Reference Liu, Magalhaes, Shaff and Kochian2009; Yamaji et al. Reference Yamaji, Huang, Nagao, Yano, Sato, Nagamura and Ma2009). In YM1 roots, MsALMT1 and STOP1 were up-regulated and their expression levels were increased in a time-dependent manner after Al treatment (see Fig. 3). The induction of malate exudation from YM1 roots was triggered efficiently by Al and occurred within 0–2 h following Al exposure. The malate exudation pattern in YM1 roots was characterized as pattern II. This might suggest that Al-induced malate exudation was achieved through the function of MsALMT1. It is possible that STOP1 is involved in the regulation of MsALMT1 expression in YM1 roots.
The malate exudation pattern of YM1 roots is similar to that of Al-sensitive pattern-II species (Delhaize et al. Reference Delhaize, Ryan and Randall1993; Li et al. Reference Li, Ma and Matsumoto2000; Rangel et al. Reference Rangel, Rao, Braun and Horst2010) but different from that of Al-resistant species in which organic acid exudation is maintained at a constant level or an ascending trend after the induction period (Delhaize et al. Reference Delhaize, Ryan and Randall1993; Yang et al. Reference Yang, Nian, Sivaguru, Tanakamaru and Matsumoto2001; Zheng et al. Reference Zheng, Yang, He, Yu, Zhang, You, Shen and Matsumoto2005; Ryan et al. Reference Ryan, Raman, Gupta, Horst and Delhaize2009; Rangel et al. Reference Rangel, Rao, Braun and Horst2010). YM1 could not maintain high levels of malate exudation after the induction period. The high contents of Al in the YM1 roots during these periods (see Fig. 1b) indicated that the released malate failed to chelate the external Al effectively. Several studies have indicated that the Al-induced secretion of organic acids is poorly associated with the internal organic acid concentration and the activity of organic acid metabolism enzymes (Ryan et al. Reference Ryan, Delhaize and Jones2001; Hayes & Ma Reference Hayes and Ma2003). However, the studies performed by Yang et al. (Reference Yang, Nian, Sivaguru, Tanakamaru and Matsumoto2001) and Rangel et al. (Reference Rangel, Rao, Braun and Horst2010) suggested that an internal accumulation of organic acids conferred the enhanced organic acid exudation in some typical Pattern-II Al-resistant species of soybean and common bean under Al stress. Moreover, the Al-resistant common bean genotype Quimbaya had higher specific activities of CS (4-fold) and PEPC (1·6-fold) than that in Al-sensitive genotype VAX-1 under Al stress (Rangel et al. Reference Rangel, Rao, Braun and Horst2010). In agreement with these results, over-expression of organic acid metabolism enzymes such as CS and MDH in transgenic plants was shown to be able to enhance organic acid exudation and Al-tolerance (Koyama et al. Reference Koyama, Kawamura, Kihara, Hara, Takita and Shibata2000; Tesfaye et al. Reference Tesfaye, Temple, Allan, Vance and Samac2001). The malate content of YM1 roots decreased gradually over the whole period of Al treatment. Genes encoding enzymes associated with organic acid metabolism, such as CS, PEPC and MDH, were absent in the SSH library. As a result, the up-regulation of MsALMT1 did not confer YM1 Al resistance under Al stress conditions. This finding is also supported by the evidence that over-expressing neMDH enhanced the malate exudation and Al resistance of transgenic M. sativa (Tesfaye et al. Reference Tesfaye, Temple, Allan, Vance and Samac2001).
The conclusion from the present study is that very low levels (5 μm) of Al stress induce a severe oxidative stress in the Al-sensitive YM1 roots. This severe oxidative stress induced the expressions of a series of antioxidant genes in YM1 roots. This suggests that YM1 responds to oxidative stress caused by Al toxicity in the same way as that found in Al-tolerant species. The up-regulation of MsALMT1 and STOP1 by 5 μM Al stress might be associated with a steep increase in the malate exudation rate for a short period of Al treatment. However, the gradual reduction of malate synthesis in YM1 roots under Al stress might not allow the malate exudation to sustain a plateau period or an increasing trend as is the case in Al-resistant species. These results suggest that YM1 does not contain a perfect system to cope with Al toxicity. This may be the main reason for the sensitivity of YM1 to Al. Thus, over-expressing genes associated with the enhancement of organic acid synthesis under Al stress might be one of the effective strategies to improve the Al resistance of YM1.
The authors are grateful to Z. B. Song (Kunming University of Science and Technology) for his valuable advice and discussions in writing the manuscript and to Y. N. Zhang (Kunming University of Science and Technology) for her kind assistance in the RT-PCR analysis. This work was supported in part by grants from the National Basic Research Programme of China (No. 2007CB108901) and the Foundation (2004PY01-5) of Yunnan Province and Kunming University of Science and Technology for Training Adult and Young Leaders of Science and Technology.
