XU Li-ming, LIU Chan, CUI Bao-ming, WANG Ning, ZHAO Zhuo, ZHOU Li-na, HUANG Kai-feng,DING Jian-zhou, DU Han-mei, JIANG Wei, ZHANG Su-zhi

1 Key Laboratory of Biology and Genetic Improvement of Maize in Southwest China, Ministry of Agriculture/Maize Research Institute, Sichuan Agriculturaluniversity, Chengdu 611130, P.R.China

2 Crop Research Institute of Binzhou, Binzhou 256600, P.R.China

Abstract Aluminum (Al) toxicity is a major factor limiting crop production and plant growth in acid soils. The complex inheritance of Al toxicity and tolerance mechanisms in maize has uncharacterized yet. In this study, the maize inbred line 178 seedlings were treated with 200 μmol L-1 CaCl2+0 μmol L-1 AlCl3 (control) and 200 μmol L-1 CaCl2+60 μmol L-1 AlCl3 (Al treatment) for 1 and 6 h, respectively. The experiment was repeated three times. Then a detailed temporalanalysis of root gene expression was performed using an Agilent GeneChip with 34 715 genes, only the genes showing more than 2.0-fold difference (P＜0.01)between the controland the Al treatment maize seedlings were analyzed further. Thus, a totalof 832 different expression genes, 689 significantly up-regulated and 143 down-regulated, were identified after the seedlings were treated with Al for 6 h. And 60 genes, 59 up-regulated and one down-regulated, were also detected after the seedlings were treated for 1 h.Replicated transcriptome analyses further showed that about 61% of total significantly genes could be annotated based on plant genome resources. Quantitative real-time PCR (qRT-PCT) of some selected candidate genes was used to demonstrate the microarray data, indicating significant differences between the controland Al-treated seedlings. Exposure to Al for 6 h triggered changes in the transcript levels for several genes, which were primarily related to cell wall structure and metabolism,oxidative stress response, membrane transporters, organic acid metabolism, signaling and hormones, and transcription factors, etc. After Al-treated for 1 h, differentialabundance of transcripts for several transporters, kinase, and transcription factors were specifically induced. In this study, the diversity of the putative functions of these genes indicates that Al stress for a short stage induced a complex transcriptome changes in maize. These results would further help us to understand rapid and early mechanisms of Al toxicity and tolerance in maize regulated at the transcriptional level.

Keywords: aluminum, maize, microarray, mechanism, transcriptome

1. Introduction

Acid soils (pH＜5.5) are the most important limitation to global crops production, which comprise about 40% of the world arable soils and up to 70% of potentially arable land.Aluminum (Al) element is the principal component of mineral soils and is present in a wide range of primary minerals(Kochian et al. 2004). Al become soluble in soils of pH＜5.5.And soluble Al, which exists in Al3+form, is readily absorbed by plant roots even it is at micromolar concentration, then root growth is inhibited, subsequently nutrient and water uptake are decreased, thus crop yield is reduced (Ryan et al. 2001).

Now, the mechanisms of root growth inhibition are not wellunderstood. But, the root apex, where cell division and cell elongation are abundant, is the most sensitive part to Al3+on root (Degenhardt et al. 1998). The rapid response to Al stress indicates that Al can interact with multiple structures in the apoplasm and symplasm of root cells quickly. In the root tissue, as much as 50-90% of the totalabsorbed Al was bound in the apoplasm rapidly and localized in the extracellular compartments (Taylor et al. 2000). Alalso accumulated in the symplasm of root rapidly. Plants exposure to Al would produce many reactions, including disorder of reactive oxygen species (ROS), alterations of the membrane structure, disruption of cytoskeletal dynamics, changes in Ca2+homeostasis and signaling, and mitochondrial dysfunction (Yamamoto et al. 2002). Further,Al3+interacting with the nuclei may inhibit mitotic activity via alterations in DNA composition, chromatin structure or template activity (Frantzios et al. 2001). In summary, these researches demonstrate that Al has deleterious effects on various cellular components, thus root growth is inhibited.

Usually, an Al-stress environment would evolve plants either to reduce Alaccumulation in the root by excluding Al from the root apex or to neutralize toxic Alabsorbed in the symplasm (Ezaki et al. 2001; Kaczorek et al. 2002). For example, Al can be excluded via rhizosphere Al-organic acid anion complex formation, which is the most popular physiological mechanism of Al tolerance in crops (Delhaize et al. 1993; Pi?eros et al. 2010). And in the plants with the mechanism, root can exude the key organic acid anions citrate, including malate, and oxalate (Emmanuel et al.2007). Other proposed Al exclusion mechanisms also involve secretion of proteins and phenolic compounds (Basu et al. 1994), increased root-mediated of the rhizosphere pH(Degenhardt et al. 1998), changed selective permeability of the plasma membrane, and masking Al-binding sites at the cell wall (Yang et al. 2008). Additionally, tolerance mechanisms are active after Al3+enters into symplasm of root cell, such as Al can be quelled in the cytosolor vacuole,or bound with proteins directly (Delhaize and Ryan 1995;Knaggs and Andrew 2003).

The identification of genes response to Al-stress will improve our understanding of stress mechanisms and provide effective strategies for improving stress adaptation.And many modern technologies, such as differential display reverse transcription-PCR (DDRT-PCR), suppression subtractive hybridization (SSH), cDNA-AFLP, expressed sequence tag (EST) libraries, cDNA libraries, etc., have been used to identify genes or transcripts related to Alstress responses (Vij and Tyagi 2007). And a number of genes were identified in different plant species (Chandran et al. 2008; Kumari et al. 2008). The Al-induced genes were categorized into many functional groups, which included stress and defense response, membrane transporter,organic acid metabolism, polysaccharide and cell wall metabolism, protein metabolism, signaling, hormones,transcription factors (TFs), cell structure and cell growth,vesicular transport, nucleotide processing and modification,transcription regulation, and translation regulation (Chandran et al. 2008; Kumari et al. 2008; Duressa et al. 2011).

Maize (Zea mays L.) is considered as a pioneering crop and widely grown in acid soils of USA, Brazil, and China(Horst et al. 2007). Citrate exudation from the root tips plays an important role for Al stress in maize (Horst et al.2007). Pi?eros et al. (2010) observed that the Al-activated root citrate release was not well correlated with Al tolerance,suggesting that other mechanisms of Al toxicity and resistance may also operate in maize roots. The research indicated that Al tolerance of maize may be a multi-genetic trait. Thus, an integrative molecular study appears to be necessary to further classify the mechanisms of Al toxicity and tolerance in maize. Up to now, the transcriptional expression with microarray techniques has been used to study mechanisms of Al toxicity and tolerance in Arabidopsis(Kumari et al. 2008), wheat (Guo et al. 2007), Medicago truncatula (Chandran et al. 2008), common bean (Eticha et al. 2010), and maize (Lyza et al. 2008), and a series of Al-responsive genes were identified. Further, in the present study, oligonucleotide microarrays was used to study the transcriptional profile of maize and to discover putative candidate genes related to mechanisms of Al toxicity and tolerance in maize.

2. Materials and methods

2.1. Plant materials and treatments

Seedlings cultivation Maize inbred line 178 was provided by Sichuan Agriculturaluniversity, China. A totalof 100 seeds were surface-sterilized in 1.0% (v/v) sodium hypochlorite for 5 min, washed 3-4 times with tap water,embedded in water for 5 h, and germinated at 28°C in dark for 3 d. Then, the seedlings were transferred to a growth chamber and cultured in Hoagland nutrient solution at 28°C/24°C (16 h light/8 h dark) for 3 d (Pi?eros et al. 2010).The 60 uniform seedlings were selected and cultured in 2 lof 200 μmol L-1CaCl2solution (pH=4.0) for a 24-h adaptation period.

AlCl3concentration determination The seedlings were treated with different AlCl3concentrations to decide the proper AlCl3concentrations for the experiment. The Al treatment of the preliminary test was 0, 20, 40, 60, 80, and 100 μmol L-1AlCl3, respectively. Then, 60 μmol L-1AlCl3was selected to treat maize seedlings for the later analysis.At least 20 seedlings were used at every concentration and three replicates were carried out for the experiment.

AlCl3treatment After the proper AlCl3concentration was decided, 5-10 seedlings after adaptation period in CaCl2solution were treated with 200 μmol L-1CaCl2+0 μmol L-1AlCl3(control) and 200 μmol L-1CaCl2+60 μmol L-1AlCl3(Al treatment) for 1 and 6 h, respectively, and the pH was adjusted to 4.0 with HCl. Three replicates were set for the experiment.

Net growth length (NGL) The original growth length of root just before treatment and thefinal growth length after treated for 1 and 6 h were measured respectively, then NGL was calculated as follows: NGL=Final growth length-Original growth length. After treated for 1 and 6 h, root tips (used for microarrays) were collected, and frozen in liquid nitrogen respectively. The treated seedlings of three replicates were all collected and stored at -80°C.

2.2. RNA extraction and cDNA synthesis

Total RNA was extracted using TRIzol Reagent (Invitrogen,Carlsbad, CA, USA) according to the manufacturer’s protocol.Additionalon-column DNase digestion was performed three times during RNA purification using the RNase-free DNase Set (Qiagen, Germany). The concentration and quality of each RNA sample were evaluated using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).Quantified RNA was reverse transcribed into cDNA with the superscript the first-strand synthesis system and the oligo(dT) primers (Invitrogen, Carlsbad, CA).

2.3. Microarray hybridization and data analysis

The cRNA of each sample was obtained by complementation with the first-strand cDNA and labelled with biotin Cy3(Al-treated samples) and Cy5 (the control samples),respectively. A Qiagen RNeasy Mini Kit (Qiagen) was used to purify the fluorescent cRNA probes. An equalamount of cRNA probes were mixed for allof experiment groups and used for hybridisation on an Agilent Maize Whole Genome Microarray (Agilent) following the instructions. And onechannel chip hybridizations were performed by ShanghaiBio Corporation (SBC) (Shanghai, China) with the Agilent Genechip Maize Genome Array (Agilent Technologies). The chip hybridization results were scanned using an Agilent Dual Laser DNA Microarray Scanner (GeneChip Scanner 3000, Affymetrix, USA) and Command Console Software 3.1(Affymetrix) with default settings. The raw data were normalized by the MAS 5.0 algorithm with GeneSpring Software 11.0 (Agilent).

Genes with more than 2-fold difference (P＜0.01) between the controland Al-treated maize seedlings were selected using the SBC Analysis System (http://sas.ebioservice.com/). To avoid false positives, the q-value, the minimum false discovery rate at which the test may be called significant, was selected at P＜0.05. Gene Ontology (GO)terms which could specifically describe genes character and their regulation mechanisms were performed by AgriGO (http://bioinfo.cau.edu.cn/agriGO/). The annotation information was obtained from GenBank (http://www.ncbi.nlm.nih.gov/genbank/).

2.4. Quantitative real-time PCR (qRT-PCR) analysis

qRT-PCR was conducted with TaKaRa ExTaq RT PCR Kit and SYBR green dye (TaKaRa, China) and a DNA Engine Opticon 2 Machine (MJ Research, Waltham, MA). Primers(Appendix A) were designed to amplify 150-250 bp. The efficiency of the primers set was calculated by performing qRT-PCR on several dilutions offirst-strand cDNAs, and they were similar. The specificity of each primer set was checked by sequencing PCR products. Amplification profiles were as follows: 95°C for 10 s, followed by 40 cycles of 95°C for 5 s and 60°C for 31 s, and afinal melting-curve 70-95°C.The melting-curve was used to check the specificity of the amplified fragment. Each sample was performed in triplicate. The maize genes Actin2-like and Ubiquitin, were used as internal controls to normalize all data. Relative expression levels of candidate genes were calculated using the 2?ΔΔCTmethod.

3. Results and discussion

3.1. AlCl3 concentration in the present study

In Fig. 1, NGlof root decreased with AlCl3concentration increasing after treated for 6 h, but NGL was similar under different AlCl3concentration treatments for 1 h. The root growth length was reduced about 40% with 60 μmol L-1AlCl3treatment for 6 h. In addition, NGlunder 100 μmol L-1AlCl3treatment decreased the same as that under 60 μmol L-1AlCl3treatment for 6 h, suggesting 100 μmol L-1AlCl3treatment for 6 h may active a tolerance response in maize. Since the maize transcriptome has no or a little change in adaption to Al environment, 60 μmol L-1other than 100 μmol L-1AlCl3could induce a seriously disorder transcriptome in maize. In this study, 60 μmol L-1AlCl3was the best choice for microarray analysis because it induced acutely Al toxicity symptome. Some similar results have been reported in Medicago truncatula cultivar response to 25 μmol L-1Al for 12 h (Chandran et al. 2008), soybean with 30 μmol L-1Al treatment for 4 h (You et al. 2011), and Al-sensitive maize genotype exposure to 39 μmol L-1Al for 24 h (Lyza et al. 2008). The different Al treatments all induced about 40-50% root length inhibition, respectively.Thus the corresponding concentration was selected to conduct a effective microarray analysis.

Fig. 1 The effect of different aluminum (Al) concentration on the net growth length (NGL). Bars mean SD.

Inhibition of root elongation is the primary and the most dramatic effect of Al toxicity (Sivaguru and Horst 1998;Kollmeier et al. 2000). Most studies showed root growth of plants such as wheat (Guo et al. 2007) and maize (Jorge and Arruda 1997; Pi?eros et al. 2010) is almost entirely abolished after Al-treated for 24-96 h. Further physiologicaland metabolic parameters measured under Al treatment for 24-96 h would probably be distorted. Additionally, previous results in maize (Lyza et al. 2008) and soybean (You et al.2011) showed Al3+stress could induce numbers of genes with differential expression for short stage treatment of 2, 4,or 6 h. Thus, 1 and 6 h were selected as the Al treatment duration in the present study.

3.2. Portrait of the transcriptome response to Al stress

Kumari et al. (2008) has used microarray to investigate the profile of transcriptome changes in plant under Al stress. In the current study, microarrays were also used to measure the transcriptome changes in maize roots. A global gene expression was depicted under Al stress at each time point (Fig. 2-A, Appendix B). The result showed 34 715 distinct genes were represented by the probes spotted on the microarrays. When the log2Ratio value of one gene was approximately equal to 1, its expression would be considered non-significant, or it would be considered expressing significantly (P＜0.01). The majority of gene expression have no significant change under Al stress for 1 and 6 h. The genes with transcripts having more than 2.0-fold difference (P＜0.01) between the controland Altreated maize seedlings were selected. Thus, a totalof 60 genes were identified under Al treatment for 1 h, and abundance transcripts of 832 differently expressed genes were detected under Al treatment for 6 h (Fig. 2-B). Nearly the same number of Al-induced genes were detected in other plant species, such as M. truncatula (Chandran et al. 2008),maize, and soybean (You et al. 2011) using a microarray technology by similar selected criteria at different time points and Al concentration. Among the 60 genes detected at 1 h, 59 genes were up-regulated and only one were down-regulated, while 689 genes were up-regulated and 143 were down-regulated for the 832 genes detected at 6 h. And the majority of up-regulated genes detected in this study were nearly consistent with that in Al-tolerant maize genotype other than that in Al-sensitive maize genotype(Lyza et al. 2008), positively suggesting differential genes underlying different tolerance regulation in early stage of Al treatment. Some genes’ expression had more than 20-fold transcriptional changes, but most of those ranged from 2- to 4-fold (Fig. 2-C). Considering the fact that only one gene was down-regulated after 1 h Al treatment, the number was significantly lower than that of up-regulated genes,suggesting the positive regulation mechanism was most powerful for maize to cope with the early stage Al stress.However, negative regulation was also acted synchronously,which included members missed from the detection due to the weak signalor without the probe, as had been found by Lyza et al. (2008). Some genes were unique for both time points, but some genes could be detected at one time point only. Surprisingly, the 32 genes detected at the both time points for the two groups of differently regulated genes,indicating that genes regulated by the two treatments had a great common feature. Similar result had also been found in Al-sensitive maize genotype (Lyza et al. 2008) following the great overlap of differential genes in short stage of Al stress.

To evaluate the biological functions of Al-responsive genes, the genes identified in maize roots after 6 h of Al treatment were assigned for functional enrichment analysis by GO terms. Nearly 61% of the total differential genes’functions were known, among them approximately 22%of the up-regulated genes and 32% of the down-regulated genes were not classified into any functional classes (Fig. 3).Eleven and nine functional categories were identified for the transcript abundance increased and decreased genes,respectively. Interestingly, the same functional categories between the up- and down-regulated genes indicated that different regulation patterns of certain biological processes might be necessary to allow cells to redirect resources towards adaptation mechanisms or coping with Al toxicity.As “stress and defense response” and “vesicular transport”occurred in the functional classes of the up-regulated genes only, the adaptive reprogramming transcriptome for response to Al stress might require the larger number of up-regulated genes other than down-regulated genes.

Fig. 2 A global gene expression profile under aluminum (Al) stress (A), the number of up- and down-regulated genes (B), and fold change of the up-regulated genes (C) under Al stress.

Fig. 3 The pathway categories of differential expression genes under aluminum (Al) stress.

3.3. Cell wall related genes response to Al stress

Some studies have suggested the cell wall is the first contact site of external stimuli and Al can bind rapidly to the cell wall, then increase cell wall rigidity and reduced cell extension and growth (Sasak et al. 1996; Kochian et al.2003; Kochian et al. 2005; Horst et al. 2010; You et al. 2011).In this study, several genes associated with the cell wall structure and modification were identified with differential expression under Al stress (Fig. 4-A). And the transcription levelof pectin methylesterase (PME) gene (Table 1) was up-regulated by Al stress after 6 h, suggesting PME was activated by Al stress. PME activity correlated with Aladsorption capacity and Al sensitivity due to demethylation of pectin determining the amount of Al binding to the cell wall (Wen et al. 1999; Vercauteren et al. 2002). Thus, the higher levelof PME activity was, the more Al-sensitive the cultivar was (Schmohl et al. 2000; Lyza et al. 2008; Yang et al. 2008).

Cell wallarchitecture can be remodeled for alleviating Al stress (Chandran et al. 2008). And glycosyl transferases(GTs) is necessary for cell wall synthesis (Sasaki et al.1996; Kumari et al. 2008). Typically, the sugar moieties in cell wallare predominantly glucose, galactose, fucose,glucuronic acid, and xylose (Lorenc-Kuku?a et al. 2004).In the present study,five genes encoding different cell wall sugar unit transferases were increased after 6 h profile,and two genes encoding putative glucosyltransferase and GT1 were decreased at the same time point (Fig. 4-A).Recently, several reports showed that the increased expression of GT genes might enhance cell wall synthesis metabolism activity, leading to Al toxicity response in soybean (You et al. 2011) and Arabidopsis (Kumari et al.2008). The similar results in our array indicated that the increased GTs activity with enhanced synthesis of cell wall components may cause cell wall stiffening and root cell expansion inhibition in maize, potentially representing Al toxicity response. Glycosyl transferase family 64 (GT64)(Table 1) might catalyze the formation of a glycosidic bond, that was consistent with a possible role in pectin biosynthesis, which was found in this study, Pedersen et al. (2003), and Singh et al. (2005).

Sasaki et al. (1996) found that Al-sensitive plant would accumulate more lignin in the roots than Al-tolerant plant under Al stress (Sasaki et al. 1996). The shikimate and phenylpropanoid pathway were required for the biosynthesis of lignin, and phenylalanine was thefinal production in the shikimate pathway (Boerjan et al. 2003; Knaggs and Andrew 2003). Herrmann and Weaver (1999) and Dixon et al. (2002)reported that the genes-encoding chalcone synthase and iso flavonoid reductase related to phenylpropanoid pathway were up-regulated in soybean, leading to lignin deposition and Al-induced inhibition of root growth (Dixon et al. 2002).Similarly, one putative chalcone isomerase gene and three genes involved in the shikimate pathway, encoding shikimate kinase, shikimate dehydrogenase, and chorismate synthase,respectively, were up-regulated by Al stress after 6 h in the microarray of this study (Table 1). All these results suggested the up-regulated genes which involved in the phenylpropanoid and shikimate pathways might increase the lignin production, then lead to cell wall rigidification to resistant to Al toxicity.

3.4. Oxidative stress response genes

Al stress would elicit the excessive production of reactive oxygen species (ROS), which resulting in membrane damage, chromosomalaberration, andfinally cell death(Mittler et al. 2002; Yamamoto et al. 2002). Thus, many genes for detoxification of ROS components have evolved and up-regulated under Al stress, which include glutathione S-transferases (GSTs), glutaredoxins (GRXs), thioredoxins(TRXs), and lipoxygenase (LOX) genes (Kumari et al. 2008).Particularly, the transcripts abundance of GST generally represented an important protective functions against oxidative damage (Yang et al. 2001). Consistent with these results, three kinds of antioxidant enzymes, GST, GRXs, and TRX, were also identified in this study, and the increased transcripts for GST, tetratricopetide-repeat TRX, and TRX genes were also detected after 6 h of Al exposure (Table 1).Additionally, a unique GRX4 gene with 52% similarity to AtGRX4 sequence was up-regulated by Al stress after 6 h in this study. Tetratricopetide-repeat TRX and TRX are both involved in plant oxidoreduction activities (Sang et al.2010; Meyer et al. 2012). And GRXs are members of the TRX fold protein family (Cheng et al. 2006). The AtGRX4 gene in Arabidopsis was induced by various environment stimuli such as temperature and metal-ion stress, and the seedlings root of atgrx4 mutants were more sensitive to oxidants than wild seedlings (Cheng et al. 2008). Basing on the results, GST, tetratricopetide-repeat TRX, TRX, and GRX4 genes may scavenge ROS activity, thus resulting in a strong maize Al tolerance.

Fig. 4 Differential gene categories under aluminum (Al) stress. A, genes associated with the cell wall structure and modification.B, transporter genes. C, protein kinases and phosphatases genes. D, transcription factor genes.

3.5. Organic acid metabolism and transporter-related genes response to Al stress

Organic acids with Alare able to form strong detoxification complexes which acting in both Al tolerance mechanisms by chelating Al ions in the apoplast or by sequestrating compounds into vacuoles when toxic Al permeate inside the cell (Ma et al. 2000). In the present study, the abundance transcripts offour genes encoding NADP-dependent malic enzyme (NADP-ME), phosphoenolpyruvate carboxylase(PEPC), succinyl-CoA ligase beta-chain, and succinyl-CoA ligase alpha-chain 2 were significantly up-regulated by Al stress after 6 h, respectively (Table 1). Eticha et al. (2010)and You et al. (2011) also reported that some genes with up-regulated expression encoding PEPC and NADP-ME were identified under Al stress.

Al stress could increase citrate content in intracellular matrix and citrate exudation increase (Tesfaye et al. 2001;Eticha et al. 2010). But organic acid synthesis metabolism has no correlation with increased organic acid exudation in triticale (Li and Matsumoto 2000; Hayes and Ma 2003).The relationship between organic acid biosynthesis andAl-induced organic acid secretion is still not well clarified in maize. Hence, whether the increased PEPC activity increased root citrate content and enhanced citrate exudation or not was not sure in maize.

Table 1 The genes discussed in the article

Regulated movement of various substances across biological membranes is an important component of cellular stress responses (Kumari et al. 2008). In this study, 25 and 5 transporter genes with differential expression levels were detected, respectively after 6 and 1 h of Al exposure by using microarray technology (Fig. 4-B). The transcripts abundance for several primary transporters were differentially identified after Al treatment, including the multidrug and toxin extrusion(MATE) proteins, sugar transporters, natural resistanceassociated macrophage protein (Nramp), and ion channel proteins (Table 1). MATE transporters are a large family of membrane transport proteins and can exclude metabolic and xenobiotic organic compounds from the cytosol by exporting them out of the cellor into the vacuole (Debeaujon et al.2001; Diener et al. 2001). Many studies (Furukawa et al.2007; Lyza et al. 2008) have reported that MATE proteins represented the well-characterized function in the regulation of Al-induced citrate efflux from plant root apices in various plant species under Al stress. But the transporters involved in malate and oxalate were not detected in the present study,which might be due to constitutive expression in maize genome and also reported by Lyza et al. (2008).

3.6. Signaling and hormones

In the present study, 12 genes encoding putative serine/threonine protein kinase (STK), two genes encoding calcium-dependent protein kinase (CDPK), and one unique gene possibly related to mitogen-activated protein kinase(MAPK) were identified with up-regulated expression after 6 h of Al exposure, but there were six putative STK genes detected were down-regulated expression at the same time (Fig. 4-C). Additionally, one of 12 STK genes was up-regulated only after 1 h of Al exposure, suggesting STK gene represented early and rapidly response to Al stress.

Among the STK genes detected, two of them were putatively encoded cell wallassociated receptor kinase(WAK1), which had been reported as candidate genes involved in Al signal transduction by Sivaguru et al. (2003)in Arabidopsis. Another STK gene detected in the study was potentially involved in regulating cell division activity under Al stress, which was also detected by Chandran et al.(2008). So STK genes having different expression patterns might regulate diverse biological processes in response to Al stress.

Both Osawa and Matsumoto (2001) and Kobayashi et al.(2007) demonstrated that protein kinase was rapidly and transiently activated following malate efflux in Arabidopsis under Al stress (Osawa and Matsumoto 2001). And one gene encoding MAPK was up-regulated after Al exposure in soybean, which could couple with citrate secretion in response to Al stress (You et al. 2011). In the present study,the up-regulated MAPK-related gene may also correlate with regulation of organic acid efflux.

Previous reports showed that diverse protein phosphatase genes, including (serine/threonine phosphatases) ST and desmoplakin (DSP), were up-regulated by Al stress in Arabidopsis (Kumari et al. 2008) and soybean (You et al.2011). In this study, three protein phosphatase 2C (PP2C)genes, one protein phosphatase 2A (PP2A) gene, one DSP gene, and one unique inositol-1,4,5-trisphosphate 5-phosphatase (5PTase) gene belonging to HP family(Horst et al. 2007) were identified with up-regulated expression after 6 h of Al exposure (Fig. 4-C). PP2A and PP2C, belonging to STs family, have negative effect on the cell division (Wera and Hemmings 1995) and ABA signal transduction (Rodriguez 1998), respectively. The roles of up-regulated ST gene family in this study might be consistent with previous researches, which involved in an arrest of cell division and repressor of hormone signal transduction.

In microarray data of the present study, one ABA1 gene relating with ABA regulation, two indole-3-acetic acid (IAA)oxidase genes acting auxin activity, and one gene encoding ethylene-overproduction protein 1 (ETO1) involved in ethylene biosynthesis acitivity were detected under Al stress for 6 h. The up-regulated ABA1 gene would increase ABA accumulation in root. The same result was also reported in soybean (Shen et al. 2004). And Koshiba et al. (1996)concluded that indole-3-acetaldehyde could be oxidized into IAA by IAA oxidase genes (Koshiba et al. 1996).Further, a high levelof auxin can induce new root formation under Al stress (Kumari et al. 2008; Mattiello et al. 2010).Interestingly, one ETO1 with a dual mechanisms inhibited aminocyclopropane carboxylic acid synthase (ACS) enzyme activity involved in ethylene biosynthesis and degraded protein by interacting with E3 ubiquitin ligase (Wang et al.2004; Koji et al. 2005). From these results, it was proposed that the increased ABA and auxin accumulation and decreased ethylene production may enhance the regulation in organic acid efflux activity and root growth, respectively.

3.7. Transcription factors (TFs)

Various TFs function as terminal points of stress signal transduction and molecular switches for downstream genes expression (Nakashima et al. 2012). A number of TFs including basic helix-loop-helix (bHLH), basic leucine zipper (bZIP), WRKY, zincfinger proteins CCCH, and Cys2-His2 (C2H2) were detected in this study (Fig. 4-D).Zincfinger protein genes play important roles in plants stress response (Jan et al. 2013). And three genes with high similarity to STOP1 gene sequence were increased in transcript abundance of soybean after Al exposure by using microarray analysis (You et al. 2011). Arabidopsis AtDi19-3 could function as a negative regulator in response to salt and drought stresses, while it positively involved in a ROS-mediated process (Qin et al. 2014). So we deduced that the up-regulated Di19 gene might reduce Al-induced ROS damage by increasing ROS scavenging gene expression,then enhancing Al tolerance capacity.

The TFs of bZIP play critical roles in abiotic stresses such as drought, low temperature, and high salinity (Zg et al.2014). In this study, two genes encoding unknown function bZIP and one putative bZIP60 TF gene were identified with up-regulated expression after 6 h of Al exposure(Fig. 4-D), and an unknown function bZIP TF gene and a low temperature-induced protein 15 (LIP15) gene belonging to bZIP Tffamily were up-regulated after 1 h of Al treatment,these suggesting bZIP TF genes responsed Al stress early.Three differential bZIP TF genes were also detected in Arabidopsis by using microarray technology after 6 h of Al exposure (Kumari et al. 2008). And LIP15 gene was upregulated by different stresses (Kusano et al. 1995). This result might also be demonstrated by present study that the increased LIP15 gene might increase the Al-induced gene expression by specially binding to these gene promoters.In Arabidopsis, bZIP60 was response to endoplasmic reticulum stress (Iwata et al. 2008). And overexpression of OsbZIP60 TF gene was shown to enhance heat and drought tolerance capacities of transgenic rice lines when compared with that of wild-type rice lines (Oono et al. 2010; Xu et al.2011). Therefore, we deduced from these results that the increased bZIP60 TF gene of present study might activate endoplasmic reticulum (ER)-resident chaperones genes to correctly execute protein folding which alleviated the Alinduced ER stress, thereby conferring Al tolerance response.

Table 2 Quantitative real-time PCR (qRT-PCR) primer sequences

Fig. 5 The correlation between the microarray data and quantitative real-time PCR (qRT-PCR) results.

3.8. Validation of microarray results by qRT-PCR

Fifteen genes with various functions and differential expression levels were selected from microarray result of Al stress for 6 h (Appendix C), and their transcript abundances were monitored by qRT-PCR (Table 2). The qRT-PCR detection showed a similar tendency with the data assigned by microarray, both of which exhibited an up-regulated expression pattern (Fig. 5). And a significant correlation was observed between the microarray and qRT-PCR results (R2=0.7362) (Fig. 5). The fact was also found in soybean (You et al. 2011) and maize (Chandran et al. 2008), representing a low R2of 0.8044 and 0.7174,respectively. In general, qRT-PCR results appeared to be a great discrepancy with microarray results, thereby causing a low R2value. Yoshida et al. (2005) attributed the phenomenon to the much wider dynamic range and greater sensitivity of qRT-PCR detection than microarray analysis.

4. Conclusion

The exposure of maize to toxic Al levels resulted in global changes in gene expression. Microarray technique was used to analyze transcriptomic responses to Al stress in maize. And many differential genes contacting with either Al resistance or Al toxicity responses activity were detected under Al stress. These genes were involved in cell wall structure and modification, oxidative stress, organic acid metabolism and transport, signal transduction, hormone metobolism, and TFs. Further, some genes belonging to cell wall structure and polysaccharides metabolism proteins potentially were induced by Al stress in the study,and primarily involved in Al toxicity response such as pectin metabolism, lignin biosynthesis, and GT-related genes. Also, genes encoding GST, TRX, and GRX were detected in the study; they were up-regulated expression to maintain normal ROS homostasis and to alleviate Al-induced oxidative stress. The different signal molecules and TFs involved in the regulation of Al stress were detected in this study, and significantly more signal molecules and TFs were found to regulate Al tolerance response rather than Al toxicity response. And only after 1 h Al stress, some genes were differential expression, suggesting these genes represented a early response to Al stress. Furthermore, many genes with unknown biological function were also identified and differentially expressed at a high levelunder Al stress.These unknown genes need to be further studied.

Acknowledgements

This work was supported by the National Basic Research Program of China (973 Program, 2014CB138705) and the National Natural Science Foundation of China (30800687,31071434), and the Ph D Programs Foundation of Ministry of Education of China (1.20125103110011).

Appendices associated with this paper can be available on http://www.ChinaAgriSci.com/V2/En/appendix.htm