Comparative Proteomic Analysis of Aluminum Tolerance in Tibetan Wild and Cultivated Barleys

Aluminum (Al) toxicity is a major limiting factor for plant production in acid soils. Wild barley germplasm is rich in genetic diversity and may provide elite genes for crop Al tolerance improvement. The hydroponic-experiments were performed to compare proteomic and transcriptional characteristics of two contrasting Tibetan wild barley genotypes Al- resistant/tolerant XZ16 and Al-sensitive XZ61 as well as Al-resistant cv. Dayton. Results showed that XZ16 had less Al uptake and translocation than XZ61 and Dayton under Al stress. Thirty-five Al-tolerance/resistance-associated proteins were identified and categorized mainly in metabolism, energy, cell growth/division, protein biosynthesis, protein destination/storage, transporter, signal transduction, disease/defense, etc. Among them, 30 were mapped on barley genome, with 16 proteins being exclusively up-regulated by Al stress in XZ16, including 4 proteins (S-adenosylmethionine-synthase 3, ATP synthase beta subunit, triosephosphate isomerase, Bp2A) specifically expressed in XZ16 but not Dayton. The findings highlighted the significance of specific-proteins associated with Al tolerance, and verified Tibetan wild barley as a novel genetic resource for Al tolerance.


Introduction
Ionic aluminum (Al 3+ ), highly toxic to plant growth, is a major factor limiting crop productivity on acid soils [1]. The strategies for maintaining production on acid soils include lime application to raise soil pH and use of plants with high tolerance to acid soils. Development and planting of Al tolerant cultivars is a costeffective and practically acceptable approach for full utilization of acid soil [2].
Barley (Hordeum vulgare L.) is one of the most Al-sensitive species among small grain cereals [3]. Al toxicity limits the growth and productivity of barley on acid soils and its expansion as a crop into many agricultural areas in the world [4]. In order to breed barley cultivars tolerant to Al toxicity, it is especially important to identify genetic resources with Al tolerance. Wild barley germplasm is rich in useful genes for crop improvement [5]. Tibetan annual wild barley from Qinghai-Tibet Plateau is regarded as one of the progenitors of cultivated barley and is rich in genetic diversity [6]. We successfully identified Tibetan wild annual barley genotypes with high tolerance to both low pH and Al stress [7]. However, their underlying physiological and molecular mechanisms in Al tolerance remain unclear.
Comparative proteomic analysis and bioinformatics techniques provide powerful tools to identify proteins expressed under abiotic stress [8]. Root proteomic analysis showed that proteins involved in stress defense, metabolisms and signal transduction were important for soybean [9], tomato [10] and Arabidopsis [11] plants survival under Al toxicity. However, only limited information is available on Al accumulation/translocation and Al tolerance mechanisms in barley. Moreover, physiological and proteomic responses to Al stress in Tibetan wild barley genotypes have never been investigated and compared with elite Al-tolerant barley cultivars. Thus, precise knowledge of the proteomic basis is required to dissect the mechanisms underlying acid/Al tolerance in wild barley. In the present study we examined stress-specific proteins for acid/Al tolerance in wild barley by comparing the proteomic responses of the two Tibetan wild barley genotypes XZ16 (high acid/Al tolerant), XZ61 (acid/Al sensitive) and Altolerant cv. Dayton using two-dimensional gel electrophoresis (2-D) and mass spectrometry (MS). These results are useful to better understand the mechanisms of Al tolerance in barley, and provide an effective pathway for the exploration of Al-tolerant genes in plants.

Plant Materials and Experimental Design
Hydroponic experiments were performed using two Tibetan annual wild barley XZ16 and XZ61 (H. vulgare L. ssp. spontaneum), acid/Al-tolerant and sensitive genotypes, respectively, and one Altolerant-cultivar Dayton. Seeds were surface sterilized in 1% H 2 O 2 for 30 min, rinsed with distilled water, and then germinated in sterilized moist quartz sand in an incubator at 2061uC. Sevenday-old uniform seedlings were transplanted into 5-L containers filled with 4.5 L basal nutrient solution (BNS). The composition of BNS was described in Wu et al. [12]. The container was covered with a polystyrol plate with 7 evenly spaced holes (2 plants per hole) and placed in a greenhouse. Solution was continuously aerated with pumps and renewed daily after Al addition. The solution pH was adjusted to 5.860.1 with NaOH or HCl, as required.
On the day 7 after transplantation, seedlings were cultured for 1 d in 0.5 mM CaCl 2 at pH 4.3, and then exposed to 0, 50 or 200 mM Al in 0.5 mM CaCl 2 at pH 4.3 for 24 h. The plantlets were kept in a growth room at 25/20uC with a 14/10 h (day/ night) photoperiod and irradiance of 340 mmols m -2 s -1 light intensity. A split-plot design was adopted with treatment as the main plot and genotype as sub-plot with five replicates in each treatment. At 3 days after treatment, plants were harvested from each treatment, and roots were washed with distilled water thoroughly and collected for two 2-DE experiments and qRT-PCR analysis. Dry weights of the plants were determined and used for Al concentration.

Determination of Al Concentration and Root Al Distribution
Al concentrations was determined after digestion in an acid mixture (HNO 3 :HClO 4 = 4:1, v/v) at 150uC for 10 h, using inductively coupled plasma atomic emission spectrometry (ICP/ AES) (Thermo Jarrel Ash, San Jose, CA).
Morin staining for Al in the root tip region was determined according to Zheng et al. [13]. Briefly, 1-cm root tips were stained with 100 mM morin (Sigma-Aldrich, St. Louis, MO, USA) in 10 mM MES buffer (pH 5.5) for 30 min. Images were acquired using a Leica TCS SP2 confocal laser scanning microscope (Leica Microsystems, Heidelberg, Germany) with excitation at 420 nm and emission at 515 nm.

Protein Extraction, Quantification, Visualization and Image Analysis
Total root protein extracts were prepared essentially according to phenol extraction method [14] with minor modification. Root sample (3 g) of control and Al treated plants were grounded in a mortar separately to a fine powder in liquid nitrogen and homogenized in an extraction buffer containing 30 mg PVPP. The homogenate was suspended in 7 ml ice-cold phenol extraction buffer (0.7 M sucrose; 0.1 M KCl; 50 mM ED-TA,0.5 M Tris-HCl, 1% w/v DTT, pH 7.5; complete protease inhibitor cocktail (Roche Applied Science)) and immediately added 7 ml ice-cold Tris buffered phenol and vortexed for 15 s. The sample was then vortexed for 10 s every 5 min and repeated for six times at 4uC. After centrifugation (30 min, 50006g, 4uC) the phenolic phase was collected, and the sample was re-extracted with 14 ml of extraction buffer (added with the same volume of phenol extraction as collected items), and vortexed for 10 s every 5 min and repeated for six times at 4uC. After centrifugation (30 min, 50006g, 4C) the phenolic phase was collected and precipitated overnight with five volumes 100 mM ammonium acetate in methanol at 220uC. After centrifugation at 50006g for 30 min at 4uC, the supernatant was removed and the pellet was rinsed twice in ice-cold acetone/0.2% DTT. Between the two rinsing steps, the sample was incubated for 60 min at 220uC. The pellet was air-dried, resuspended in 200 ml lysis buffer (7 M urea, 2 M thiourea, 4% CHAPS, 20 mM Tris-HCl, pH7.4, containing 1% w/v DTT; Amersham Biosciences), and vortexed for 1 h at room temperature. Protein concentration was determined by standard Bradford assay using bovine serum albumin as standard (Bio-Rad, Hercules, CA, USA). All chemicals used were, if not further specified in the text, p.a. or electrophoresis grade. All electrophoresis units employed were from Amersham Biosciences.
Protein visualization, image analysis and quantification were determined according to Bah et al. [15]. For each sample, at least three independent protein extracts were prepared after each treatment and at least three 2-DE analyses were performed for each protein extract. To analyze the expressed protein patterns, stained gels were scanned and calibrated using a PowerLook1100 scanner (UMAX), followed by analysis of protein spots using GE HealthCare Software (Amersham Biosciences). Spot detection was Figure 2. Representative 2-DE maps of root proteins in XZ16 exposed to Al for 24 h. The proteins were isolated from the root of XZ16 exposed to 50 mM (A, upper) and 200 mM (B, below) Al for 24 h. Total proteins were extracted and separated by 2-DE. In IEF, 100 mg proteins were loaded onto pH 4-7 IPG strips (24 cm, linear). SDS-PAGE was performed with 12.5% gels. The spots were visualized by silver staining. Differentially accumulated protein spots are indicated by green sashes. Arrows indicate the differentially expressed protein spots whose expressions were significantly induced (fold changes $1.5) or unchanged (21.50, folds ,1.5) in XZ16 but down-regulated (folds ,21.50) in XZ61; or up-regulated in XZ16 but unchanged in XZ61, under 50 mM Al (U1-U24) and 200 mM Al (U25-U44) stress. doi:10.1371/journal.pone.0063428.g002 realized without spot editing. The protein spots were quantified using the % volume criterion. Only those with significant and reproducible changes (P,0.05) were considered to be differentially accumulated proteins. The target protein spots were automatically excised from the stained gels and digested with trypsin using a Spot Handling Workstation (Amersham Biosciences). Peptides gel pieces were placed into the EP tube and washed with 1:1 mixture of 50 mL of 30 mM K 3 Fe(CN) 6 and 100 mM NaS 2 O 3 for 10-15 min until completely discolored then washed with 200 mL bidistilled water (two times for 5 min each). The washed solution was drained and washed with 50% ACN (acetonitrile, Fisher A/0626/ 17) and 100% ACN rotationally, and then incubated in 25 mM NH 4 HCO 3 (ammonium bicarbonate, Sigma A6141) for 5 min at 37uC. After leaching out of the incubation solvent, 50% ACN and 100% ACN was rotationally added and dried at 40uC for 5 min respectively. Trypsin digestion was carried out as follows: sequencing-grade porcine trypsin (Promega, Madison, WI, USA) was suspended in 25 mM NH 4 HCO 3 at a concentration of 12.5 ng per ml to rehydrate the dried gel pieces. The trypsin digestion was carried out for 16 h at 37uC. Peptides were extracted from the digest as follows for three times: 10 mL of 50% ACN containing 0.1% TFA (trifluoroacetic acid, GE HealthCare) was added to each tube and incubated for 5 min at 37uC and the supernatants were transferred to new EP tube. The extracts were pooled and then vacuum concentrated for about 2 h. A solution of peptides was filtrated via Millipore (Millipore ZTC18M096) and mixed with the same volume of a matrix solution consisting of saturated a-cyano-4-hydroxycinnamic acid (CHCA) in 50% ACN containing 0.1% TFA. After the peptides were co-crystallized with CHCA by evaporating organic solvents, tryptic-digested peptide masses were measured using a MALDI-TOF-TOF mass spectrometer (ABI4700 System, USA). All mass spectra were recorded in positive reflector mode and generated by accumulating data from 1000 laser shots. The following threshold criteria and settings were used: detected mass range of 700-3200 Da (optimal resolution for the quality of 1500 Da), using a standard peptide mixture

Peptide and Protein Identification by Database Search
Data were processed using the Data Explorer software (Applied Biosystems) and proteins were unambiguously identified by searching against a comprehensive non-redundant sequence database (NCBInr) using the MASCOT software search engine (http://www.matrixscience.com/cgi/search form.pl?FORM-VER = 2&SEARCH = MIS). Folds of increase and decrease in Al exposed vs unexposed roots were calculated as treated/control and -control/treated for up-and down-regulated proteins, respectively. For single-peptide identified proteins, up-and down-regulation were assigned when the regulation factors were above 1.5 (p,0.05). Table 1. Proteins whose expression were significantly induced (+) in XZ16 roots but down-regulated (2)/unchanged in XZ61, or unchanged in XZ16 but down-regulated in XZ61 under 50 mM Al stress.

qRT-PCR Analysis
Total RNA was isolated from roots with the TRIzol reagent following manufacturers' recommendation (Invitrogen, Karlsruhe, Germany). cDNA samples were assayed by quantitative real time PCR (qRT-PCR) in the iCycler iQTM Real-time PCR Detection System (Bio-Rad, Hercules, CA, USA) using the SYBR Green PCR Master Mix (Applied Biosystems). The PCR conditions consisted of denaturation at 95uC for 3 min, followed by 40 cycles of denaturation at 95uC for 30 s, annealing at 58uC for 45 s and extension at 72uC for 45 s. Gene-specific primers (Table S1) were designed using the Primer Express software (Applied Biosystems). Barley actin gene was used as control (AY145451) fw-59-GACTCTGGTGATGGTGTCAGC-39, rv-59-GGCTGGAA-GAGGACCTCA-39.

Statistical Analysis
Statistical analysis were performed using the Data Processing System (DPS) Software Package [16]. Statistical significance of the data was evaluated by two-way ANOVA using Duncan's multiple range test (SSR).

Tibetan Wild Barley XZ16 Is Highly Tolerant to Al Toxicity
Time of appearance and severity of Al toxicity symptoms differed greatly among the three genotypes ( Figure S1). XZ16 was less affected by 24 h exposure to 50 or 200 mM Al (pH 4.3), whereas XZ61 was obviously affected, as reflected by severe root growth inhibition. No significant difference between control and 50 or 200 mM Al stressed plants was found in root DW (dry weight) of XZ16, and the whole plant DW of XZ16 and Dayton. However, root and the whole plant DW of XZ61 decreased by Table 2. Proteins whose expression were significantly induced (+) in XZ16 roots but down-regulated (2)/unchanged in XZ61, or unchanged in XZ16 but down-regulated in XZ61 under 200 mM Al stress. Al localization in barley roots exposed to different Al levels for 24 h was monitored by morin fluorescence using confocal laser scanning microscopy ( Figure 1A), and fluorescence intensity of image analysis was calculated using Image J software ( Figure 2B). Root Al fluorescence showed that root Al concentration increased with increasing external Al levels. XZ16 exhibited similar fluorescent signal in root tips with Dayton, being significantly (p,0.05) less than that of XZ61 in both Al levels.
As to root Al accumulation, there was no significant difference between XZ16 and Dayton, but both exhibited significantly (p,0.05) less accumulation than that of XZ61 in the two Al levels ( Figure S2A). Shoot Al accumulation, however, was significantly lower (p,0.05) in XZ16 than in Dayton and XZ61 in response to Al treatments ( Figure S2B). The transferring rates from roots to shoots among them were not differed significantly except that Dayton showed 18.2% and 13.6% higher than XZ16 and XZ61 when exposed to 50 and 200 mM Al stress, respectively ( Figure  S2C). The above results indicate that Tibetan wild barley XZ16 is highly tolerant to Al stress.
In addition, expression of three proteins, classified as the energy category ( Figure 5 and Table S2) was significantly higher in XZ16 vs Dayton under control condition at pH 4.3, and simultaneously suppressed at pH 6.0. They are predicted to be 6-phosphogluconate dehydrogenase decarboxylating-like isoform 1 (6-PGDH) (E1), plastid glutamine synthetase isoform GS2b (E2) and iron deficiency specific clone no. 3 (IDS3) (E3). These proteins are potentially responsible for the low pH tolerance in XZ16 compared with Dayton.

XZ16 Shows Higher Expression of Genes Corresponding to Al Up-regulated Proteins
To determine whether the changes in protein abundance detected by 2-DE were correlated with changes at the transcriptome level, quantitative RT-PCR was performed using RNA isolated from the roots of a separate set of plants treated with 0 or 50 mM Al for 24 h (Figure 6 and Table S2). Transcript levels of five Al-regulated genes including SAMS3, MeSe, GS, c-GCS and ATP synthase beta subunit (U1, U2, U4, U5 and U9) were chosen and successfully detected. Among them, SAM3, MeSe, GS and c-GCS were up-regulated in XZ16 and down-regulated in XZ61,  (Table 1). Fifty micromolar of Al significantly up-regulated ATP synthase beta subunit by 3.2 folds in XZ16 compared to the control, but no change was detected in XZ61 and Dayton ( Figure 6).
For genome mapping, BLASTn and CAP3 were used to blast the Brachypodium distachyon protein databases. The top hits were then mapped to barley chromosome based on barley genome zipper and sequence [18], [19]. Finally, 32 (including 30 of the 35 Al-tolerance related proteins, and 2 of the 3 low pH responsive spots) out of the above 38 proteins were linked to their corresponding candidate genes in barley genome (Table S3). The mapping of these candidate genes provides a short cut for the identification and transformation of Al responsible genes in future.

Discussion
The current study showed that Al-tolerant Tibetan wild barley XZ16 is characterized by less Al accumulation both in roots and shoots (Figure 1 and Figure S2). Our current data is the first study to identify Al-responsive proteins in Tibetan wild barley (XZ16) using a proteomic approach. Barley is one of the most Al sensitive cereal species and responds to Al toxic ion immediately through releasing organic acid in roots [3,35]. It is important to successfully induce Al-tolerant-specific-proteins before visible morphologicalstress-symptoms for plants to achieve tolerance to Al toxicity. In addition, the micro-and minor-elements, containing in the nutrient solution for the long term Al exposure of 15 days, trends to complicatedly react with Al ion compared to simplified adding Ca 2+ solution. Furthermore, after long term Al treatment, barley roots (especially in sensitive genotype) is severely damaged, out of vigorousness and partly death, thus not suitable for protein extraction. Therefore, in this study, Al-tolerant-specific-proteins in wild genotype XZ16 were verified under 50 mM or 200 mM Al condition for 24 h. The treatment of 200 mM Al on the plants could provide us the special mechanism (proteins) of Al tolerance in XZ16 in comparison to Dayton or XZ61 under high Al concentration condition. Because Tibetan wild barley XZ16 and XZ61 are the two contrasting genotypes with different Al tolerances, the Al-regulated proteins identified using comparative proteomics will provide a good foundation to elucidate the mechanisms involved in Al-tolerance/resistance in Tibetan wild barley. Here we identified 35 proteins associated with Al-tolerance in wild barley XZ16 (Figures 2, 5 and Tables 1, 2). There were 16 proteins, up-regulated in roots of XZ16 but down-regulated or unaltered in both Dayton and XZ61, indicating their specificity and importance for Al tolerance in XZ16. Among them, four proteins (SAMS3, U1; ATP synthase beta subunit, U9; Os09g0535000, U28; Bp2A protein, U29), were markedly induced by Al stress in XZ16 but not in Dayton, while repressed in XZ61. Obviously, XZ16 has different stress response and defense mechanisms against Al stress as Dayton did. Further examination of these proteins may elucidate the mechanism of Al tolerance in XZ16 and provide new genetic materials for developing Altolerant crops. The selected stress-responsive proteins are discussed below according to their function.

Metabolism Category
In the root proteomic analyses, seven of the identified proteins are involved in metabolism: SAMS3 (U1), methionine synthase (MeSe, U2 and U3), glutamine synthetase (GS, U4), c-glutamylcysteine synthetase (c-GCS, U5), putative aspartate aminotransferase (AST, U6) and predicted pirin-like protein (U7). SAMS, as an essential enzyme in cellular metabolism, has been long regarded as a 'housekeeping' function and it catalyzes the nucleophilic substitution reaction between methionine and ATP into SAM (S9-adenosyl-L-methionine) (Figure 7). SAM serves as an aminopropyl and methyl donor for ethylene and polyamine (PAs) [20], [21], [22]. It is well documented that ethylene and PAs are involved in response to biotic and abiotic stresses in plants [23], [24]. Thus SAMS plays an important role in Spd and Spm biosynthesis and stress response in plants [25], [26]. Qi et al. [27]   , ATP synthase beta subunit (spot U9), atp1 (F-tyep ATP synthase a subunit, spot U32) and V-ATPase (V-type ATPase AC39, spot 38) were represented by ''3.6.3.14'' in oxidative phosphorylation (B). Glucose 6-P either proceeds through glycolysis (participated by FBA, TPI, PGM and ENO) or the PPP (participated by TK) to produce pyruvate, which is converted to acetyl-CoA, entering the TCA cycle and producing citrate and malate (participated by Aco3). atp1 belongs to mitochondrial membrane ATP synthase generates ATP from ADP in the presence of a proton gradient across the membrane.
found that over-expression of Suadea salsa SAMS2 in transgenic tobacco plants leads to an increase in polyamine (PAs) content and enhancement of salt tolerance. In the present study, SAMS3 was up-regulated in XZ16 but not expressed in Dayton and repressed in XZ61 under 50 mM Al stress, highlighting the function of this protein in Al tolerance of XZ16. Although SAM is the precursor of ethylene, the key enzyme of ethylene synthesis is ACCS [28] (Figure 7). Since ACCS was not identified in this study, further research is required to determine whether up-regulation of SAMS is involved in ethylene-mediated inhibition of root growth and/or the alteration of cell wall and polymer structures in roots under Al stress. On the other hand, SAMS3 was expressed normally in Dayton under Al stress as determined by qRT-PCR ( Figure 6). Therefore, Al-induced expression of SAMS3 is likely to be regulated at the post-transcriptional or translational level in Dayton.
In addition, we also observed that MeSe was depressed in XZ61 (U2 and U3), but not altered in XZ16 under Al stress. Higher MeSe levels would increase the amount of methionine for SAM synthesis (Figure 7). Enhanced accumulation of MeSe contributes to increase PAs biosynthesis, which is critical to plant survival in many environmental stresses [29]. The decreased accumulation of MeSe in XZ61 indicates that Al stress triggered damages in barley roots may be associated with the modification of amino acid metabolism and synthesis of other amino acids derived metabolites. The next important questions is how SAMS3 and MeSe elaborate to cope with Al toxicity and which metabolism factors, e.g., SAM, ethylene, PAs (spermidine or spermine), are responsible for Al detoxification in conjunction with SAMS3 and MeSe. Therefore, it is necessary to investigate the accumulation of related metabolite and enzymatic activities of the SAMS family for a better understanding of their functions in the Al stress response.
Two glutathione (GSH) synthesis proteins c-GCS and GS (U4 and U5), play a pivotal role in various metabolic processes involved in plant growth and development and stress responses including Al toxicity [30], [31]. The pirin-like protein (U7 and U26) plays important roles in a number of different biological processes, however, its physiological role in plants remains unclear [32]. Aspartate aminotransferase (AST, U6 and U25) is a key antioxidative enzyme for detoxifying reactive oxygen species (ROS) under abiotic stresses [33]. Similarly, up-regulation of AST expression was reported in Arabidopsis in response to Al stress [34]. Under Al stress, c-GCS (U5), GS and AST were significantly increased in XZ16, whereas dramatically decreased in XZ61 and Dayton (Table 1), again indicating that there a unique mechanism in Al tolerance for the Tibetan wild barley genotype.

Energy Category
Similar to the proteomic results of Zhou et al. [10], ATP synthase beta subunit (U9) was only detected in XZ16 after 24 h of 50 mM Al stress ( Table 1), indicating that the energy provided to ATPase for active Al efflux and detoxification in XZ16 was increased when exposed to Al stress. qRT-PCR of ATP synthase beta subunit confirmed a remarkable increase in XZ16 but not in X61 and Dayton under 50 mM Al ( Figure 6). Therefore, we conclude that ATP synthase beta subunit significantly contributes to the ATPase-mediated active Al efflux and detoxification, and is regulated at both transcriptional and translational level in XZ16 in response to Al.
In addition to ATP synthase beta subunit, we also identified another type of ATPase, atp1 (U32), which was markedly downregulated in XZ61 after Al stress, but remained unchanged in XZ16 and Dayton ( Table 2). The efflux of Al-induced citrate and malate is usually mediated by anion channels and ATPase-driven active co-transporters [35], [36]. Interestingly, a significant increase in ATP synthase beta subunit and unchanged levels of atp1 in XZ16 were consistent with higher root citrate and malate efflux, whereas the lack of ATP synthase beta subunit expression and the Al-induced severe inhibition of atp1 in XZ61 were associated with lower citrate and malate secretion. These results firmly suggest that high level of ATP synthase beta subunit and atp1 in XZ16, unlike that in Dayton, may partly contribute to the active OA transport and secretion during protection from Altoxicity. Thus, we may speculate that a novel protein synthesis of ATP synthase beta subunit may be involved only in Al inducing organic acid secretion in Tibetan wild barley XZ16. Obviously, it deserves a more detailed investigation in the future.
A BLAST search revealed that Os09g0535000 (U28) is a homologue of triosephosphate isomerase (TPI). TPI was reported to be induced in rice [37] and maize [38] under drought stress, indicating the importance of cellular homeostasis maintenance and emphasizing the role of this protein in energy production. Enolase (ENO, U31) is responsive to salt, low and high temperature and anaerobic stresses [8], [39]. Spot U10, identified as aconitate hydratase 3 (Aco3), plays a role in regulating resistance to oxidative stress and cell death in Arabidopsis and Nicotiana benthamiana [40]. Spot U29, the Bp2A protein, was first identified in the wheat genome [41]. The present study is the first to examine Bp2A (U29) expression under Al stress. However, the function of this protein and its direct involvement in Al tolerance are poorly understood. Therefore, the mechanisms underlying the differential expression of this protein in different barley genotypes should be further explored.
In the glycolytic pathway, the levels of other Al stress responsive proteins fructose-bisphosphate aldolase (FBA, U11 and U27) and phosphoglycerate mutase (PGM, U30) were increased in XZ16 under Al treatments, consistent with the findings of Fukuda et al. [42] in rice under Al stress and Yan et al. [43] in the rice response to chilling. Down-regulation of Os06g0133800 (U8), a homologue to transketolase (TK), in XZ61 may reversibly inhibit ribose 5phosphate, which is a substrate for nucleic acid synthesis associated with the synthesis of RNA under Al stress [42].

Cell Growth/Division and Protein Biosyntheses Category
Proliferating cell nuclear antigen (PCNA) is an essential component in eukaryotic DNA synthesis [44]. PCNA interacts with many proteins and participates in a variety of metabolic processes, such as cell cycle control, nucleotide excision repair and post-replication mismatch repair [45]. Translation elongation factor 1b (EF1B, U13) is a highly conserved protein that catalyzes the exchange of bound GDP for GTP on EF-1a, a required step to ensure continued protein synthesis. In our study, PCNA (U12 and U33) and EF1B (U13) were up-regulated in in roots of XZ16, but were unchanged in XZ61 and Dayton. It suggests that XZ16 might possess a high actitity of DNA replication machinery in response to Al stress, and EF1B may play an important role in maintaining the root elongation rate of XZ16 under Al stress.
In addition, the other specific protein up-regulated in XZ16 is the RNA-binding Ras-GAP SH3 binding protein (G3BP, signal transduction category, U20). Zhou et al. [46] reported that the expression of the G3BP gene was associated with fertility conversion in male-sterile wheat. However, the exact function of G3BP in RasGAP-dependent signaling remains to be defined. These results provide a starting point for further investigation into the functions of these proteins using genetic and other approaches.
In conclusion, the response and defense mechanisms of Al stress in XZ16 appear different from those of Dayton, as reflected by the different expressions of these specific proteins associated with Al tolerance under Al stress between XZ16 and Dayton or XZ61. There are four proteins (i.e. SAMS3, ATP synthase beta subunit, TPI, Bp2A protein), which are exclusively expressed in XZ16 not in Dayton and XZ61 under Al stress, indicating their crucial role in development of Al stress tolerance in XZ16, and novelty of genetic resource for Al-tolerance. In addition, as the functions of some differentially expressed proteins and their direct involvement in stress tolerance are poorly understood, further studies are warranted to elucidate the underlying molecular and metabolic pathways for better understanding the mechanisms involved in Altolerance of wild barley and providing new genetic resources in Altolerant crop breeding.