|Home | About | Journals | Submit | Contact Us | Français|
Soybean is one of the most aluminum (Al) sensitive plants. The complex inheritance of Al tolerance trait has so far undermined breeding efforts to develop Al-tolerant soybeans. Discovering the genetic factors underlying the Al tolerance mechanisms would undoubtedly accelerate the pace of such endeavor. As a first step toward this goal, we analyzed the transcriptome profile in roots of Al-tolerant soybean line PI 416937 comparing Al-treated and untreated control plants using DNA microarrays. Many genes involved in transcription activation, stress response, cell metabolism and signaling were differentially expressed. Patterns of gene expression and mechanisms of Al toxicity and tolerance suggest that Cys2His2 and ADR6 transcription activators, cell wall modifying enzymes, and phytosulfokines growth factor play role in soybean Al tolerance. Our data provide insights into the molecular mechanisms of soybean Al tolerance and will have practical value in genetic improvement of Al tolerance trait.
Aluminum (Al) toxicity is a major constraint of crop production on acid soils. In view of the fact that 40% of world's arable land is acidic [1, 2], Al toxicity remains a major hurdle for increasing world food, fiber, and fuel production particularly via expansion of cultivation into acid soils.
Aluminum inflicts a wide range of cellular injuries in plants that ultimately result in reduced root growth, nutrient and water uptake, and productivity [1, 2]. Plants possess some degree of tolerance to Al toxicity that varies among species and genotypes [1, 3–6]. Al tolerance mechanisms include exclusion and internal detoxification. Al exclusion via rhizosphere Al-organic acid anion complex formation is the most widely documented physiological mechanism of Al tolerance in cultivated and wild plants alike [1, 7]. Root-exuded citrate, malate, and oxalate are the key organic acid anions involved in such mechanism. Genes involved in Al-induced root exudation of malate and citrate have been cloned in wheat  and sorghum , and their variants are being discovered in several plant species. Internal detoxification mechanisms involve the formation of Al complexes with organic acids, acidic polypeptides, and/or proteins and subsequent sequestration of Al in organelles away from sensitive sites in the cell [9, 10]. The genetic components of the internal detoxification pathways are yet to be elucidated.
In soybean, Al tolerance is a complex trait perhaps involving several genes and pathways [11, 12]. Quantitative trait loci (QTL) mapping in a population derived from Al tolerant PI 416937 and Al sensitive Young has revealed five DNA markers associated with Al tolerance . Most of the alleles were derived from Al-tolerant PI 416937. Other reported soybean Al tolerance genes include phosphoenolpyruvate carboxylase (PEPC), homolog of translationally controlled tumor proteins (TCTPs), inosine 5′-monophosphate dehydrogenases (IMPDHs) , aluminum-induced 3-2 (Sali3-2), and aluminum-induced 5-4a (Sali 4-5a) . Ermolayev et al.  and Ragland and Soliman  used gene expression as a tool to identify the above genes but the techniques used in these experiments were not sensitive enough to detect large number of genes that might be expected from the quantitative nature of soybean Al tolerance trait. The objective of this study was to discover putative Al tolerance genes in Al-tolerant soybean line PI 416937 using DNA microarrays—a robust genome—wide transcript profiling technology. Such an approach was recently employed in wheat [15, 16], maize , Arabidopsis , and Medicago truncatula [19, 20] to discern the molecular basis of Al tolerance in the respective species.
An Al-tolerant soybean plant introduction (PI 416937) highly characterized for Al response [12, 21] was used in this experiment. Seeds were surface sterilized with 20% household bleach (Clorox) in water for 12min, rinsed with distilled-deionized water several times, and were germinated in deionized water moistened standard germination paper at 25°C in an incubator for 72h. Seedlings uniform in tap root length were transferred to black-painted pots filled with approximately 4L of 800μM CaCl2 background solution with 10μM Al added (treated) or no Al added (control) in a Conviron growth chamber (16/8h light/ dark cycle with respective temp. of 28°C/20°C, photosynthetic photon density of 100μmol m−2s−1). The pH of the culture solution was adjusted to 4.3 and maintained at that level for the entire duration of the experiment. After 2, 12, 48, or 72h of Al treatment 1cm sections of the primary root tips of approximately 15 plants/pot were harvested, immediately flash frozen in liquid nitrogen, and stored at −70°C for RNA extraction. Three independent replicates were used per treatment.
Total RNA was extracted from 100mg root tissue samples using Qiagen RNeasy plant RNA isolation kit following the manufacturer's protocol (Qiagen, Inc.). The Affymetrix GeneChip Soybean Genome Array with over 68 000 probe sets, Glycine max L. and wild soybean combined, was used for microarray analysis of the soybean genome for Al tolerance. Three chips were used per treatment. Detailed procedures for RNA labeling and array analysis are described in the Manufacturer's GeneChip Expression Technical Manual (Affymetrix). Briefly, the quality of total RNA was determined using the RNA 6000 Nano Chip on Agilent BioAnalyzer 2100 prior to double-stranded cDNA synthesis. Total RNA in the amount of 2μg was used for double-stranded cDNA generation by linear amplification using oligo dT-T7 primer and reverse transcriptase (RT). Subsequently, biotin-labeled cRNA was synthesized by in vitro transcription (IVT) using the ENZO High Yield IVT kit (ENZO). Quality and quantity of cRNA were assessed using the RNA 6000 Nano chip on Agilent BioAnalyzer 2100. Fifteen-microgram cRNA was used for hybridization. Arrays were hybridized overnight at 45°C for 16h in GeneArray Hybridization Oven 640 (Affymetrix). The next day, arrays were washed and stained in the Fluidics Station 450 (Affymetrix) and scanned by the High Resolution GeneChip Scanner 3000 (Affymetrix).
Gene expression values were determined using theGeneChip Operating Software (GCOS 1.1, Affymetrix). The expression levels were subjected to data query and data mining in Data Mining Tool (DMT). Statistical Analysis of the data was conducted using the software packages ArrayAssist Enterprise together with Pathway Assist (Stratagene/Agilent, Santa Clara, CA). The raw GeneChip files from GeneChip Operating Software (GCOS, Affymetrix, CA) were uploaded, background-subtracted, variance stabilized, and normalized with GC-RMA method . The control group was used as a baseline to calculate the intensity ratio/fold changes of the treatment versus control. The ratio was log2-transformed before further statistical analysis. The P-values were obtained by an unpaired t-test assuming unequal variance. Significantly upregulated and downregulated genes were annotated using protein databases accessed by blastx at National Center for Biotechnology Information (NCBI).
Quantitative real-time PCR quantification of transcript levels for representative genes [Gma. 20326: F-5′-tcactccccaccttatcgag-3′, R-5′-tcatgtggtggagtgtggtt-3′; Gma. 6948: F-5′-ttatctccggcgaaaacctc-3′, R-5′-tcgtggtgcagcagtttaag-3′; Gma.12326: F-5′-agccactcaaatggttcagc-3′, R-5′-tctccttgtccttctccttcc-3′; Gma. 24062: F-5′-tgccgaaggatcatctcaac-3′, R-5′-cgagggataatggttgatgg-3′; Gma.26937: F-5′-tacccaaaaggcaggcatac-3′, R-5′-ggccgaggtacaaacacatc-3′; Gma.4156: F-5′-tccaatgctgacaagtgctc-3′, R-5′-tagggacactccgtccaatc-3′; Gma.2577: F-5′-acgcctatgaacgtgaaacc-3′, R-5′-aacatcagcggagagcattc-3′] from microarray experiments was conducted using the Roche Diagnostics light Cycler 480 System with SYBR green detection (Roche Diagnostic, Corp) using beta-tubulin gene (beta-tubulin: R-5′-CCATCAAACCTCAAGGAAGC-3′, F-5′-TGCTGTCCTCTTGGACAATG-3′) as internal control. mRNA was isolated from plants grown under similar experimental conditions as in the microarray experiments. mRNA extraction and quality test was as described above. RNA samples were treated with Applied Biosystems Turbo DNA-free DNase (Ambion, Inc.) to remove DNA contamination. Briefly, 2μl 10x DNase I buffer and 1μl rDNase I were added to 20μl RNA sample, and the mix was incubated at 37°C for 30 minutes in water bath. Subsequently, 2μl resuspended DNase inactivation reagent was added and the samples mixed well and incubated at room temperature for 3 minutes. Samples were then centrifuged at 10 000g for 1.5min (Eppendorf centrifuge 5415D) in 1.6ml centrifuge tubes and supernatants transferred to fresh tubes.
cDNA was synthesized from 1μg DNase-treated RNA samples using the Roche Diagnostics Transcriptor First Strand cDNA Synthesis Kit (Roche Diagnostics, Corp) according to manufacturer's protocol. cDNA concentration and quality was determined using NanoDrop Spectrophotometer brand ND-1000 (NanoDrop Technologies, Inc.). cDNA samples were diluted with nuclease-free water in varying ratios ranging from 1:4 to 1:10 depending on sample concentration. A total reaction volume of 11μl comprising 2μl cDNA sample, 2μl each of the reverse and forward primers at 0.2μM concentration, and 5μl SYBR mix was prepared in 96-well plates (Roche Diagnostics) in two biological and three technical replicates for each gene. A real-time PCR profile of preincubation at 95°C for 5min, a 45-cycle amplification at 95°C for 10 second, 55°C for 20 second, and 72°C for 20 second, melting at 95°C for 1min, 65°C for 1min, and 95°C continuous, and cooling at 40°C for 30 seconds was used to amplify the samples. Negative controls in which cDNA sample was replaced with PCR grade water for each primer pair were included in each run. Sample wells were individually assessed for data quality by evaluating amplification curves and PCR product specificity was verified by melting curve analysis. The expression level of target genes was normalized using in-run beta-tubulin gene as internal control, and transcript concentration ratios were calculated using the ΔΔCT-Method . The change in gene expression levels (fold change) was calculated as treatment to control ratio and compared with results from microarray.
A total of 38 genes were identified as differentially expressed in the 10μM Al-treated experimental plants compared to no Al added controls at 2h post Al treatment (Figure 1). Thirty-four of them were upregulated and 4 were downregulated with a fold change ranging from 3.08 to 32.55 (Table 1).
The highest number of differentially expressed genes was detected at 48h post Al treatment (Figures (Figures1 and1 and and2).2). A total of 542 genes (97.2% upregulated and 2.8% downregulated) were detected. Those exceeding 13-fold changes are presented in Table 2. The marked fold differences observed in the current research are substantially higher in comparison with results obtained by most authors but are comparable to results of [18, 24]. There were two genes in common between the set of genes detected at 2h and 48h post treatment (Gma.2577, 7-fold downregulated at 2h and 8-fold upregulated at 48h and Gma.26937, 8-fold downregulated at 2h and 115 upregulated at 48h). Similar patterns of gene expression were observed in Arabidopsis roots under Al stress with few overlaps between sets of genes detected at 6h and 48h post Al treatment .
The temporal pattern of Al-induced gene expression changes observed in this study diverges from results of other authors. At 12 and 72h, almost no genes were differentially expressed or detected. The virtually no detection of Al-regulated genes at 12 and 72h post treatment seems a little odd but it is what is expressed in this soybean genotype at detection thresholds of P-value <.01 and 3-fold change in an experiment with 3 replications. Gene expression is species and genotype specific [15–20] making comparison of results across different studies difficult. The most likely explanation for the 72-hour result is that Al toxicity could have already been neutralized by the 72h, making differential gene expression unnecessary. The lack of transcriptional response at 12h, however, is a biological puzzle, and it could represent a very unusual temporal transcriptome response of this soybean genotype to Al stress. Among the few reported Al microarray studies, the results of Kumari et al.  in Arabidopsis is the closest to ours with regard to the number of genes detected at early and late time points. They detected 127 genes at 6 h post treatment and 733 genes at 48h post treatment using a threshold of a 2-fold change whereas we detected 38 genes at 2h and 542 at 48h using a 3-fold change.
All of the differentially expressed genes that were functionally annotated by the Genbank nonredundant protein database were grouped into five functional categories based on their putative cellular function. The functional classification showed that stress- and metabolism-related genes constitute the major fractions of Al-regulated genes (Figure 3).
The microarray gene expression levels were validated with quantitative real-time PCR for representative genes (Figure 4). In general, the microarray results were in agreement with qRT-PCR but in a few cases quantitative RT-PCR gave higher levels of expression compared to microarray. Such results are obtained by a number of investigators [16, 20, 25]. Detail discussion of factors contributing to the discrepancy between microarray and RT- PCR gene expression levels is covered in . Many authors attribute the phenomenon to the high dynamic range and greater sensitivity of PCR detection. It is worth noting that the gene expression kinetics depicted in Figure 1 shows the efficacy of our experimental design in capturing the full dynamic range of gene expression profiles in the soybean genotype studied. Gene expression peaks at 2 and 48h suggesting that major savings in microarray experimental expenditure could be realized by limiting sampling to these time points in future experiments.
A number of transcription factors including bZIP, WRKY, MYB, ADR6, and NAc were highly upregulated in the present study (Tables (Tables1 and1 and and2).2). Members of these families of transcription factors were previously detected under Al stress in several plant species [16, 18–20, 27]. Cys2His2-type zinc finger (bZIP) and auxin downregulated (ADR6) factors are particularly interesting from Al tolerance perspective. Cys2His2-type zinc finger (bZIP) protein coregulates molecular response to proton and Al toxicities . It controls the expression of AlMT1—a malate transporter protein that acts in Al exclusion mechanism. In this study, Cys2His2 (Gma.4526, Table 2) was upregulated 51-fold at 48h post treatment suggesting that malate plays a major role in Al tolerance mechanism of PI 416937 soybean. Earlier physiological study by Silva et al.  showed that Al stress increases exudation of both malate and citrate during the first 6 h of exposure to Al in both tolerant and sensitive soybean types. But they concluded that the sustained accumulation and exudation of citrate is mainly responsible for the genotypic differences in Al tolerance. In the present work, 48 h after Al exposure the malate transporter regulator protein was highly expressed in contrast with the observation of Silva et al. . We postulate that Cys2His2 might regulate the expression of other Al tolerance genes in addition to malate transporter. It is also possible that malate biosynthesis becomes a limiting step or malate might indeed play a major role in soybean Al-tolerance contrary to earlier conclusions. ADR6 transcription factors were previously reported as Al tolerance genes [14, 18]. In the present study, ADR6 was highly upregulated (14-fold, Table 2). The plant hormone auxin and ADR6 exhibit opposite behavior in plant roots under Al stress. Al has been shown to inhibit auxin biosynthesis and transport genes as one possible mechanism of its toxicity . On the contrary, ADR6—an auxin downregulated transcription factor is induced under Al stress perhaps mimicking auxin's role of promoting root growth. These observations suggest that Cys2His2 and ADR6 transcription factors are important modulators of soybean molecular response to Al stress.
Transporters, specifically malate (ALMTs) and citrate (MATE) transporters are the first Al tolerance genes cloned in plants and represent the well-characterized Al tolerance mechanism in a wide range of plant species [5, 8]. None of the family members of these two genes were detected in the present study which could be due to constitutive expression. In contrast, an ABC transporter, a multidrug resistance glutathione-S-transferase-exporting ATPase (Gma.14080, Table 2), was upregulated 27-fold at 48h post treatment in the present study, which could detoxify xenobiotics by transporting glutathione-S-transferase conjugated toxin to the vacuole from sensitive sites in symplast. The involvement of ABC transporters in Al tolerance mechanism is widely documented [15, 18, 30, 31]. Other Al-induced transporters included heavy metal ion transport proteins (Gma.17184 and Gma.24625), lipid transport proteins (DQ222982 and Gma.17184), carbohydrate transport protein (Gma.11888), and coatomer protein complex subunit 2-protien—a polypeptide complex for membrane trafficking (Gma.1654) (Table 2). Heavy metal transport proteins are either located in plasma membrane or subcellular membranes and detoxify heavy metals by exporting metal-ligand complexes out of the cell or by sequestration or compartmentalization of the complex in the vacuole. The internal detoxification mechanism of Al involves formation of Al-organic acid complexes and subsequent transport of the complex by transport proteins to leaf vacuoles in Al hyperaccumulating plants that are adapted to acid soils [1, 9, 10, 32]. Similar mechanism might operate in cultivated plants, and the heavy metal binding proteins upregulated here might function in such pathway.
Lipid and sugar transport proteins are among other transporters detected. Lipid transport proteins transport lipids to cell wall for biosynthesis of cutin layers and surface waxes as a defense mechanism against pathogen attack . They are also induced by abiotic stresses including aluminum [15, 33]. Lipid transport proteins loosen cell wall in a nonhydrolytic mode and enhance cell elongation, a role traditionally attributed to expansins . Aluminum stress inhibits root growth by restricting cell wall extension ; hence, there should be a significance to the upregulation of lipid transport proteins under Al stress. Plant sugar transporters have been reported to be induced by pathogen attack and Al stress [18, 35], as is the case in the present study (Gma.11888, Table 2).
Aluminum toxicity has been shown to elicit a wide range of stress-related proteins [19, 20, 36, 37]. In this study, genes known to be responsive to pathogens, oxidative stress, toxins, or Al were classified under this category. Several pathogenesis-related proteins including syringolide-induced protein, acidic endochitinase, PR-5, basic secretory protein, pathogenesis related protein STH-2, and proteinase inhibitors were upregulated at 48h post Al treatment (Table 2). The confluence between plant molecular response to aluminum toxicity and pathogen infection likely arises from the fact that both cause oxidative stress. However, the role of pathogenesis-related proteins in Al tolerance is equivocal. Overexpression of peroxidase and proteinase inhibitor genes in Arabidopsis did not improve Al tolerance for the transformed plants relative to controls . On the other hand, overexpressing pepper basic pathogenesis-related protein 1 gene in tobacco resulted in enhanced tolerance to heavy metal cadmium and pathogen infection .
Other Al-upregulated stress-related genes included carbohydrate oxidase, glutathione-S-transferase, and glutathione-based reductase (Tables (Tables1 and1 and and2).2). Carbohydrate oxidase and cell wall peroxidases have been reported to provide protection against pathogens by generating hydrogen peroxide from carbohydrate substrates in the apoplast . Hydrogen peroxide has antimicrobial property and also acts as signal molecule for defense genes expression. In the case of aluminum, the activity of these enzymes is correlated with plant Al sensitivity [40, 41]. Glutathione-S-transferase and glutathione-based reductase are the key enzymes of cellular detoxification and antioxidation system . Glutathione reductase catalyses the conversion of oxidized glutathione to reduced form. Glutathione-S-transferase conjugates toxins and electrophilic compounds to reduced glutathione. The glutathione-conjugated toxin is then exported out of the cell or into the vacuole by the ABC transporter proteins discussed above. The concurrent upregulation of glutathione-based reductase, glutathione-S-transferase, and ABC transporter protein suggests that PI 416937 soybean may guard itself against Al by extruding Al out of the cell or by compartmentalization of Al to the vacuole. Yet there are conflicting evidences with respect to the role of the glutathione defense system in plant Al tolerance. Overexpression of glutathione-S-transferase in Arabidopsis thaliana has been shown to enhance plant Al tolerance . On the other hand, Maron et al.  found more oxidative stress genes upregulation in Al sensitive cultivar of maize than in Al tolerant cultivar and argue that oxidative stress genes upregulation is a symptom of Al toxicity rather than a tolerance mechanism, an assertion that is supported by findings of . In addition, these genes are responsive to several biotic and abiotic stress factors and, therefore, should not be regarded as major Al tolerance genes while partial role is certainly possible.
Genes involved in catabolic or biosynthesis of various metabolites were differentially expressed. The most interesting ones from Al tolerance perspective are genes for biosynthesis of ascorbic acid and genes encoding cytochrome P450 and endo-xyloglucan transferases/hydrolases. All were upregulated in the present study, and the last two were previously reported to be upregulated in Arabidopsis [19, 20] and wheat [15, 16] roots under Al stress. Ascorbic acid is an important component of cellular antioxidation system. Oxidative stress is one aspect of Al toxicity, and maintenance of cellular ascrobate homeostasis has been reported to be an essential component of plant Al tolerance . Cytochrome P450 may serve as monooxygenase in the biosynthetic pathways for lignin, defense compounds, hormones, pigments, fatty acids, and signaling molecules or in the detoxification pathway to catalyze the breakdown of numerous endogenous and exogenous toxic compounds . We detected two genes (Gma.28852 upregulated 43-fold and Gma.29655- upregulated 15-fold) which code for cytochrome P450 (Table 2). Gma.28852 encodes protein involved in pathways of ascorbate metabolism, coumarine and phenylpropanoid biosynthesis, and gamma hexachlorohexane degradation. Endoxyloglucan hydrolases are cell wall metabolism enzymes. Members of this family of enzymes have been implicated in Al tolerance [16, 18–20]. There is a causal relationship among endoxyloglucan hydrolases, cell wall composition, and Al tolerance. Al induced increases in cell wall pectin and hemicellulose increases plant Al sensitivity . Pectin and hemicellulose form complexes with Al resulting in increased cell wall rigidity and reduced cell extension and growth [27, 43, 45]. Endoxyloglucan hydrolases appear to relax the Al-rigidified cell wall presumably by hydrolyzing the Al-sugar complexes.
Perception of stress signal by the cell is the starting point for cascade of events leading to gene expression and change in cell metabolism in response to a stress factor. Aluminum perception and signaling is currently poorly understood. Cell wall-associated receptor kinase (WAK1) was the first Al signaling gene discovered , but there is no evidence that demonstrate, WAK1's major role in Al tolerance . Microarray analyses have shown kinases, phosphates, and EF hand Ca2+ binding proteins as possible components of Al signaling pathway [16, 18]. In the present work, a Ca2+ sensor protein (Gma.35830), calcium-binding EF hand family protein (Gma.7726), oxidative signal kinase (Gma.8262), and a gene for growth factor phytosulfokines precursor (BK0001191) were upregulated 48h post Al treatment (Table 2). The phytosulfokines growth factor is a novel Al-induced gene, and it is involved in cell proliferation and growth, characteristics that confer Al tolerance.
We conducted a transcriptome analysis in Al-tolerant soybean line PI 416937 to identify potential genetic factors underlying Al tolerance trait. Our results uncovered several genes which might potentially have influence on soybean Al tolerance. Among these, two transcription factors, cell wall metabolism enzymes and a cell proliferation gene are particularly interesting from perspective of the physiological and molecular mechanisms of plant Al tolerance. The first transcription factor, Cys2His2 zinc finger protein, coregulates molecular response to proton and aluminum toxicities, the major acid soil stress factors . The second transcription activator, ADR6 is an auxin downregulated gene. Al suppresses auxin biosynthesis and transport in root system which might be one possible mechanism of Al induced root growth inhibition . Conversely, ADR6 is triggered under Al stress probably acting in a parallel pathway to auxin to restore root growth under Al stress. Root cell wall rigidification by Al binding is one principal mechanism of Al toxicity. Cell wall metabolism enzymes and proteins are induced under Al stress and may counteract Al effects on root cell walls. It is increasingly evident that these proteins as well as cell wall pectin and hemicellulose content are important determinants of Al tolerance in cereals [3, 4, 43]. Evidence from this study also implies that cell wall remodeling enzymes and proteins may play role in soybean Al tolerance. Inhibition of cell division and proliferation is another major mechanism of Al toxicity. We identified a novel cell proliferation stimulating gene phytosulfokines growth factor which might reverse this effect of Al. Taken together; our findings provide important insights into the molecular mechanisms of aluminum tolerance in soybean. The genes we identified may guide efforts to improve plant Al tolerance trait.
The authors are grateful to USDA Plant Genetic Resources and Dr. Thomas Carter of North Carolina State University for generous gifts of the seed of soybean genotype used in this study. They thank Dr. Ernst Cebert of Alabama A&M University for his kind assistance in seed multiplication. They would like to express their appreciation to Dr. Michael Crowley, Genomic core facility at University of Alabama at Birmingham for his assistance and training on microarray hybridization in this project. This research was supported in part by NSF/EPSCOR Grant No. 05026. They are also grateful to Alabama Graduate Research Scholars Program for its financial support. Contributed by Agricultural Experiment Station, Alabama A&M University.