Comparison of Leaf Proteomes of Cassava (Manihot esculenta Crantz) Cultivar NZ199 Diploid and Autotetraploid Genotypes

Cassava polyploid breeding has drastically improved our knowledge on increasing root yield and its significant tolerance to stresses. In polyploid cassava plants, increases in DNA content highly affect cell volumes and anatomical structures. However, the mechanism of this effect is poorly understood. The purpose of the present study was to compare and validate the changes between cassava cultivar NZ199 diploid and autotetraploid at proteomic levels. The results showed that leaf proteome of cassava cultivar NZ199 diploid was clearly differentiated from its autotetraploid genotype using 2-DE combined MS technique. Sixty-five differential protein spots were seen in 2-DE image of autotetraploid genotype in comparison with that of diploid. Fifty-two proteins were identified by MALDI-TOF-MS/MS, of which 47 were up-regulated and 5 were down-regulated in autotetraploid genotype compared with diploid genotype. The classified functions of 32 up-regulated proteins were associated with photosynthesis, defense system, hydrocyanic acid (HCN) metabolism, protein biosynthesis, chaperones, amino acid metabolism and signal transduction. The remarkable variation in photosynthetic activity, HCN content and resistance to salt stress between diploid and autotetraploid genotypes is closely linked with expression levels of proteomic profiles. The analysis of protein interaction networks indicated there are direct interactions between the 15 up-regulation proteins involved in the pathways described above. This work provides an insight into understanding the protein regulation mechanism of cassava polyploid genotype, and gives a clue to improve cassava polyploidy breeding in increasing photosynthesis and resistance efficiencies.


Introduction
Cassava (Manihot esculenta Crantz) is a perennial shrub of the Euphorbiaceae family. It is a major calorie source for sub-Saharan Africans and is ranked the sixth most important source of calorie in human diet worldwide [1]. In south China, cassava mainly provides raw materials to produce starch and biofuel. It is a potential crop to grow in semi-arid lands if a suitable cultivar is bred [2]. However, cassava breeding faces several limitations such as (1) its heterozygous genetic makeup which makes it time consuming to breed efficiently and (2) low tolerance to salinity and cold which makes it difficult to grow in north China and saline soils [3]. Because of its importance, a number of important studies regarding genome, proteome and transcriptome have been performed. A draft genome sequence and limited proteome identification have been generated to provide new clues for cassava breeders to overcome the limitations [3][4][5][6].
One of the potential approaches to increase stress resistance is to produce polyploidy (whole-genome duplication) [2]. The polyploid crops, influenced by nuclear genome size, have much larger cells than the diploid ones [7]. Several studies showed that the significance of the proportional increase in cell volume with increase in DNA content [8,9]. This cell volume increase along with DNA content increase was observed in secondary xylem radial and vasicentric axial parenchyma cells, and radial parenchyma and sieve elements of secondary phloem [2,8,9]. Additionally, DNA content increase in the synthesized polyploids led to the description of structural changes such as chromosomal rearrangements and gains or losses of DNA sequences, demonstrating the occurrence of modifications at the level of gene expression. The studies from Brassica, potato and cotton polyploids have shown that some genes are silenced after polyploidization, while others are derepressed [10][11][12], which suggests that functional and phenotypic evolution may be driven by these genomic changes.
Although these have been significant recent advances on the genomic and transcriptomic consequences of genomic merger and doubling, the fate of translated gene products, i.e., the proteome, remains poorly studied in the context of polyploidization [12,13]. Using two-dimensional electrophoresis (2-DE) combined mass spectrometry (MS), it is often possible to visualize, quantify, and identify hundreds or even thousands of proteins in a given tissue or cell sample. Proteome analysis is increasingly used in functional plant studies. Proteomic analysis has the potential to provide a broad view of plant responses to stress at the level of proteins [14,15]. Some reports have been focused on proteomic analysis of plant polyploids including cabbage [16], wheat [17][18][19], cotton [12], potato [20], banana [21], Arabidopsis [22] and Tragopogon [23]. However, little is known about the effect of polyploidy on the cassava global protein networks that underlie key physiological or developmental processes.
In the present study we investigate and validate the changes of proteome patterns between cassava NZ199 polyploid genotype and its diploid plants using 2-DE combined MALDI-TOF-MS/ MS. All differential proteins were clustered into cohesive groups based on their biochemical functions. The biological network of protein-protein interaction was established to describe the polyploid photosynthetic activity tolerance against stresses. Our data will provide crop breeders a set of protein database involved in cassava polyploidy and would be useful to understand the polyploidy mechanism involved in the key physiological or developmental processes.

Analysis of different ploidy levels in cassava
Cassava cultivar NZ199 autotetraploid genotype was ployploidized artificially by the use of colchicine applied as a solution of 0.001% to lateral buds over a period of 72 h. The emerging shoots were screened for the formation of chimeras or total tetraploids with the chimeras eliminated. To identify autotetraploids, buds were observed for standard chromosome counting and flow cytometric analysis (CellLab QuantaTM SC MPL, Beckman, USA) [24] in addition to observing leaf shape. The shoots were propagated vegetatively and grown at the Cassava Germplasm Pool, Tropical Crops Genetic Resources Institute, Chinese Academy of Tropical Agricultural Sciences. Functional leaves on the top of cassava diploid and autotetraploid plants grown for 180 d at CGB were sampled for protein extraction. One leaf was collected from one cassava plant, and three leaves were collected.

Protein extraction and 2-DE separation
Collected leaves were washed with distilled water and the central nervure was removed. Leaves were cut into pieces and 0.5 g pieces were weighed for protein extraction. Leaf proteins were extracted with phenol according to the procedure of Chen et al. [25]. Protein pellets were dissolved in sample buffer [9.5 M urea, 2 M Thiourea, 4% (w/v) CHAPS, 1% (w/v) DTT, 2.5 mM EDTA, 2.5 mM EGTA], and the protein content was then measured with the protein assay according to the instruction manual (Bicinchoninic acid protein assay kit, product code BCA-1 and B9643, Sigma). For analytical and preparative gels, the 13 cm IPG strips (pH 4-7, linear) (GE Healthcare Bio-Sciences AB) were rehydrated overnight with 312 ml of rehydration stock solution [7 M urea, 2 M thiourea, 3% (w/v) CHAPS, 0.5% (v/v) Triton X-100, 1% (v/v) IPG buffer, and 0.002% (w/v) bromophenol blue], containing 300 mg of protein, at room temperature. Isoelectric focusing (IEF) was conducted at 20uC with a Multiphor II (Amersham Biosciences). The running conditions were as follows: 300 V for 0.05 h, then, increased from 300 V to 3500 V as a gradient over 1.5 h, and finally 3500 V for 4.20 h. The focused strips were equilibrated twice for 15 min each first in 1% (w/v) DTT and then 2.5% (w/v) iodoacetamide prepared in equilibration buffer (50 mM Tris-HCl (pH 8.8), 6 M urea, 30% (v/v) glycerol, 2% (w/v) SDS). The second-dimension electrophoresis was performed by SDS-PAGE in a vertical slab of 12% acrylamide using an SE 600 Series Vertical Slab Gel Unit (Hoefer Scientific Instruments, San Francisco, CA, USA). Preparative gels were stained with Colloidal Coomassie brilliant blue G-250 [26]. Three independent biological replications were carried out.

Image and data analysis
Gel matching for protein quantification was performed by Image Scanner III (GE healthcare) and Delta 2D (DECODON GmbH, Greifswald, Germany) software, and spot pairs were confirmed visually. The differentially expressed spots were determined by using Scheffe's test at P,0.05. The abundance of each protein spot was estimated by the percentage volume (% Vol). Only those with significant and reproducible changes were considered to be differentially accumulated proteins.

Tryptic in-gel digestion
Briefly, the differential spots were excised from 2-DE gels and cut into ,1 mm 3 , and then washed twice in MillQ water for 10 min. The washed gel pieces were subjected to destaining solution (25 mM ammonium biocarbonate, 50% acetonitrile) followed by sonication for 5 min. The gel pieces were again washed twice in MillQ water for 10 min, followed by washing twice in 50% acetonitrile for 10 min. After dehydration with 50 ml acetonitrile for 2-3 min, the gel pieces were digested overnight at 37uC in 15 ml of sequencing grade trypsin (Promega) according to the manufacturer's instructions (1 mg in 100 ml of 25 mM ammonium biocarbonate). The supernatants were transferred to a fresh tube and stored at 4uC until analysis.

Protein identification by MALDI-TOF-MS/MS
Differential proteins were identified using MALDI-TOF-MS/ MS at Analysis and Testing Center, Jiangsu University. After 1 ml peptide extract produced by the in-gel digestion was placed on an anchor chip and air-dried, it was covered with 1 ml solution of 0.4 mg/ml a-cyano-4-hydroxycinnamic acid in a mixture of acetonitrile and 0.1% trifluoroacetic acid (TFA) (70:30) and then air-dried. The mass spectra were acquired on an Ultraflex-TOF mass spectrometer (Bruker, German). Spectra were internally calibrated with trypsin auto-digestion products. Data captured by MALDI-TOF-MS/MS were matched via Mascot v2.2.03 (http:// www.matrixscience.com) against NCBI [Taxonomy: Viridiplantae (Green Plants)]. Carbamidomethyl (Cys) and oxidation (Met) were considered as variable modifications. A single missed cleavage was permitted. Peptide mass tolerance was set as 3.0 Da and MS/MS ion mass tolerance was set at 1.5 Da. Peptide charge states (+1, +2, +3) were taken into account. Routine protein identification required sequence-confirmed data for a minimum of one peptide with recognition as the top ranking match in the Mascot Standard scoring system [3].

Western blot analyses
Leaves of cassava cultivar NZ199 diploid and autotetraploid genotypes were homogenized. The protein extraction and Western blot were performed according to the method previously reported [3]. Proteins detected by immuno-staining with antiRubiscopolyclonal antibody (AS07218), anti-APX antibody (AS08368) and anti-PrxQ antibody (AS05093) from Agrisera. Western blots were developed according to the method of NBT/BCIP from Roche (11681451001).

Photosynthetic activities measurement by imaging pulse amplitude modulation
The Maxi-version of the Imaging Pulse Amplitude Modulation (Imaging PAM) and the software Imaging WIN version 2.39 (both Heinz Walz GmbH, Effeltrich, Germany) were used to determine the photosynthetic activities according to Behr et al. [27]. The selected leaves attached to cassava cultivar NZ199 diploid and autotetraploid genotypes were adjusted in the dark for 20 min prior to measurement. The leaves were detached at 10:00 a.m (ambient full sunlight). The detached leaf was clamped onto the holder at a distance of 7 cm between the leaf and the CCD camera. The other variables were as follows: light intensity at 2, frequency at 1, gain at 7, damping at 2; saturation pulse with intensity at 10 and width at 8. Each recording started by determining the dark fluorescence parameter F 0 . A saturating light pulse was employed to determine the maximal fluorescence parameter Fm. The maximal quantum yield of PS II was calculated as Fv/Fm = (Fm-F 0 )/Fm. In the presence of light intensity 185 mE m 22 s 21 , the current fluorescence yield (F t ) and the maximum light adapted fluorescence (Fm9) were measured to calculate the effective PS II quantum yield (WPSII = (Fm9-F t )/ Fm9).
Additionally, the nonphotochemical quenching [NPQ = (Fm-Fm9)/Fm] was determined to show the proportion of absorbed light energy that is not used for photosynthetic electron transport. NPQ/4 is NPQ divided by 4 in order to allow for a display in a color code ranging from 0.0 to 100%. For quantitative analyses of the whole leaves, the software used 5 areas of interest of 25 mm 2 that were randomly distributed over each leaf segment measured. For each variant, three individual plants were used and the respective results were averaged [27].

Determination of cyanogenic glucosides (CGs) by HPLC
Five plantlets, grown for 2 months in vitro, were selected from each cassava cultivar NZ199 diploid and autotetraploid genotypes. A leaf disc was sampled from the first unfolded leaf of each plant by snap-closing a 2 ml-Eppendorf lid around one of the leaf fingers. Five replicates were carried out in each diploid and autotetraploid genotypes. To determine the content of CGs directly, the leaves samples were immersed into 300 ml prewarmed 85% (v/v) methanol and boiled at once in a water bath at 100uC for 3 min, then, cooled on ice. The methanol extract was transferred to a new tube and filtered through a 0.45 mm filter [28].
CG content was determined on a Hitachi LC 2130 series coupled with autosampler L-2200, UV detector LC 2030 and chromatography workstation T 2100P (Hitachi, Japan). The HPLC column was a Kromasil 100-5C18 column (25064.6 mm, 5 mm). The mobile phases consisted of water (A) and acetonitrile (B). The flow rate was 0.8 ml/min. The gradient program was as follows: 0 to 30 min, linear gradient 10% to 100% (v/v) B, 30 to 33 min, linear gradient 100% to 10% B, 33 to 38 min. The UV detection wavelength was set at 215 nm. The column temperature was 30uC. The injected volume of samples was 20 ml. The concentration of hydrogen cyanide (HCN) standard sample is 0.5 ppm collected from the National Institute of Metrology of China. The retention time for CG was 28.6 min.

Generation of protein interaction networks
All proteins identified in cassava cultivar NZ199 autotetraploid were compared with those in diploid genotype and were used to generate a wider protein interaction map by employing a Pathway Studio software program (www.ariadnegenomics.com) [25].

Protein profiles of cassava autotetraploid and diploid genotypes
A total of approximately 500 protein spots were detected by digital image analysis, and at least 300 spots gave reproducible staining patterns for all samples as judged by eye and by spot intensity ranking ( Fig. 2A and 2B, Figure S1). Using a spot-to-spot comparison and statistical analysis, a total of 65 stained spots (Fig. 2C) from autotetraploid genotype were found to have significant changes (p,0.05) with greater than 2.0-fold altered intensity compared with diploid genotype. The differential spots were determined using Scheffe's test at P,0.05. Of these, the expression of 14 spots was down-regulated and the remainder upregulated (Fig. 2C). As shown in Fig. 3, the detectable spots with differentially abundant expression in Fig. 2C were clearly highlighted (a, diploid genotype; b, autotetraploid genotype). 3-D images of these spots were generated using Delta2D software to detect the significant changes between diploid and autotetraploid genotypes as shown in Figure S2.

Functional grouping of identified proteins
Sixty-five spots with differential expression were isolated from 2-DE gels and identification performed using MALDI-TOF-MS/ MS, of which fifty-two protein spots were identified. Other 13 spots, however, remained unidentified. Forty-seven identified proteins were up-regulated and 5 were down-regulated in autotetraploid genotype compared with diploid genotype. Functions of 52 differentially expressed proteins were annotated via the survey of gene banks (Table 1, Fig. 4). Fifteen proteins are associated with carbohydrate and energy metabolism, of which 4 proteins (spot 35, spots 47-49) were down-regulated. The 32 upregulated proteins were associated with photosynthesis (6 spots, 11%), defense system (12 spots, 23%), HCN metabolism (2 spots, 4%), protein biosynthesis (6 spots, 11%), chaperones (3 spots, 6%), amino acid metabolism (2 spots, 4%) and signal transduction (1 spot, 2%). Other 5 were function-unknown proteins, including 1 down-regulated protein. The characteristics of 4 down-regulated proteins, spot 35 and spots 47-49, relate with energy metabolism. Protein spot 35 was identified as NAD(P)-binding Rossmann-foldcontaining protein and spots 47-49 were putative ATP-binding proteins (Table 1). To ensure reliability of differential proteins on 2-DE gels, the protein expressions of Rubisco, APX and PrxQ in leaves of cassava diploid and autotetraploid genotypes were detected by immunoblotting to validate the proteomic analysis (Fig. 5). For example, the result from western blot showed that the expression of Rubisco small subunit in autotetraploid genotype was more than that in diploid genotype, which is similar with that seen on 2-DE analysis (Figs. 3 and 5).

Photosynthetic activities in cassava diploid and autotetraploid genotypes
Imaging-PAM studies were performed with the leaves from 50 d-old in vitro cassava plants of diploid and autotetraploid genotypes to understand the effects of up-regulated proteins  involved in leave photosynthesis metabolism on the photosynthetic activities of cassava polyploid plants. In the present study, we chose the sensitive photosynthesis parameters Fv/Fm, WPSII and NPQ/4 to determine the changes of photosynthetic activities. Fig. 6 and Table 2 showed that autotetraploid induction could produce significant effects on the efficiency of excitation energy capture by open Fv/Fm, WPSII and NPQ/4, suggesting that an increase in maximal and effective quantum yield and a concomitant increase in NPQ/4 processes are sensitive markers for polyploid genotypes. These data imply up-regulated proteins associated with photosynthesis will result in increase of photosynthetic activities in cassava autotetraploid genotype.

Effects of salt stress on cassava diploid and autotetraploid genotypes
Plantlets of cassava diploid and autotetraploid genotypes grown on the MS medium and salt-stressed medium for 30 d and 50 d were examined to observe the effects of up-regulated proteins on salt stress tolerance (Fig. 7). The length and fresh weight of 50 dold cassava plantlets grown on salt-stressed medium showed inhibitory effects of high salt concentrations on the growth of both cassava genotypes. However, length and fresh weight of cassava plantlets from diploid genotype in the presence of salt were significantly lower than those from autotetraploid genotype ( Table 3). These data imply association of the up-regulated proteins with the increase of salt stress tolerance. The spots showing differential expression (.2.0-fold of the normalized volume) were counted after gel analysis and manual editing with Delta2D software. Each value represents the mean ± SE of triplicates. Protein spots whose abundance increased (+) or decreased (2) after polyploidy were shown. The numbers corresponded to the 2-DE gel in Fig. 2 Variation in CG content of cassava autotetraploid and diploid genotypes Variations of CG content in the leaves collected from 50 d-old in vitro cassava plants of diploid and autotetraploid genotypes were studied to validate the effects of up-regulated proteins in relevance with HCN metabolism (spots 6 & 32) in autotetraploid genotype. The CG content of autotetraploid genotypes significantly increased in comparison to diploid genotypes (Fig. 8I-IV). These data imply that an increased conversion of HCN from cyanide containing compounds may be related with up-regulated proteins (linamarase and acetone-cyanhydrin lyase) which are involved in HCN synthesis.

Protein interaction networks
All differential proteins identified in cassava autotetraploid genotypes were used to generate a wider protein interaction map by employing a Pathway Studio software program (Fig. 9). The relationships of binding and regulation were established for 15 differential proteins, responding to plant photosynthesis, yield, adaptation and stresses. CDSP32 is localized in the chloroplast. There are direct interactions between 15 up-regulated proteins, including ATP synthase subunit beta, alcohol dehydrogenase, beta-glucosidase, phosphoglycerate kinase, triose phosphate isomerase (TPI), Rubisco activase (RCA), Rubisco, APX2, CDSP3, peroxiredoxin (PrxQ), thioredoxin, translation elongation factor,  glutamate-ammonia ligase, chaperone and 14-3-3 proteins, whereas RCA can establish relations with other proteins through regulating the processes of photosynthesis and plant yield (Tables S1, S2, S3). Sheffield et al. (2006) compared proteomes between cassava fibrous and tuberous roots and found 292 differentially expressed spots on gels. Of those, 232 proteins were identified [6]. Li et al. (2010) identified 383 proteins from cassava somatic embryos, plantlets and tuberous roots using LC-ESI-MS/MS [3]. These data were helpful to understand proteome patterns between cassava different tissues. To further understand the mechanism of cassava whole-genome duplication induced by colchicine, we first investigated the significant changes in the proteome patterns of cassava diploid and autotetraploid genotypes and assess the potentially cassava-polyploid breeding. Differential protein spots that were found to contribute to this variation included proteins of eight functional categories (Table 1 and Fig. 4), in which there are direct interactions among 15 up-regulated proteins (Fig. 9). Furthermore, the expression levels of the differential proteins in cassava autotetraploid genotype may be directly associated with plant photosynthesis, yield, adaptation and stresses ( Fig. 4 and Tables S1, S2, S3).

Discussion
Effects of polyploid induction on the proteome were considerably pronounced. Statistical analysis showed clear differences in protein patterns between autotetraploid and diploid genotypes. In the present work, the proteins associated with plant defense system including defense proteins [beta-glucosidase (spots 9, 10), chloroplast latex aldolase-like protein (spot 24), metacaspase (spot 29), CDSP3 (spot 33), thioredoxin-like protein (spot 61)], detoxifying and antioxidant proteins were highly accumulated in cassava NZ199 autotetraploid genotype compared with diploid plants. Beta-glucosidase is a key enzyme regulating the abscisic acid (ABA) pool in plants under stresses [29,30]. Chloroplast latex aldolaselike protein plays a role in defense and stress responses [31]. CDSP32 (a chloroplastic drought-induced stress protein) and thioredoxin-like protein, induced by environmental stress conditions are known to participate in the response to oxidative and drought stresses [32][33][34]. It is reported that plant metacaspases, a family of cysteine proteases structurally related to caspases, play important roles in biotic and abiotic stress-induced programmed cell death [35,36]. PrxQ (spot 36) and APX (spots 39,44,45) are reported to participate in the protection against oxidative damage [25,32]. Moreover, the evidences of salt stress on cassava diploid and autotetraploid genotypes showed that up-regulated proteins associated with plant defense could result in the increase of cassava plantlet tolerance to salt stress ( Fig. 7 and Table 3). Interestingly, 14-3-3 protein (spot 37), a signal transformation-related protein,   was up-regulated by more than 4.73-fold in autotetraploid genotype compared to diploid (Table 1). Previous studies revealed that multiple members of the 14-3-3 regulatory protein family act as mediators in ABA signaling through direct interaction with ABA-responsive element binding factors, and then alter the expression of related genes to enhance plant resistance against salinity or drought [37]. In addition, the previous studies described that the anatomical alterations of cassava polyploid plants, showing differences in density, compaction and thickness of parenchyma cells, may confer upon high tolerance to drought in tetraploid plant. The large number of vessel groupings in the tetraploid type may also maintain a larger quantity of water than in case of fewer ones in diploid plants [2,38]. All above described results provide evidences at proteomic and anatomical levels to support polyploid genotype owning high tolerance to stresses. Polyploids consistently exhibit larger mesophyll cells with more chloroplasts and greater photosynthetic capacities per cell than their diploid progenitors [39,40]. The causes of these differences at the level of underlying proteins are unknown. In the present study cassava polyploid genotype could increase the expressed levels of 6 photosynthesis related proteins including nuclear encoded precursor to chloroplast protein (spot 4) [41], Rubisco large subunitbinding protein subunit beta (spot 8), oxygen evolving enhancer protein (spot 22) [25], RCA (spot 31), cytochrome b6-f complex iron-sulfur subunit (spot 55) [42] and rubisco (spot 60) ( Table 1). These proteins are very likely involved in enhancement of photosynthesis, carbohydrate and energy metabolism in cassava leaves [3]. Rubisco serves as the main gateway for inorganic carbon to enter metabolic pathways in most ecosystems and hence is unique in its importance to support life. It catalyzes the key reaction in the photosynthetic assimilation of CO 2 [43]. RCA could regulate the processes of plant photosynthesis and then affect the plant yield (Fig. 9). The evidences of photosynthetic activities in cassava polyploid plantlets showed that up-regulated proteins associated with photosynthesis could result in the increase of photosynthetic capacities in cassava autotetraploid genotype in comparison to diploid genotype ( Fig. 6 and Table 2). The up-regulated proteins involved in carbohydrate and energy metabolism including ATP synthase (spots 13-16, 30, 64, 65), phospoglycerate kinase (spot 26), alcohol dehydrogenase (ADH, spot 28) and TPI (spots 43, 46) ( Table 1). Mitochondrial ATP synthase plays a role in mitochondrial changes through the life cell [44]. Phosphoglycerate kinase catalyzes the phosphorylation of 3phosphoglycerate to 1, 3-diphosphoglycerate using ATP within the Calvin-Benson cycle [45]. ADH plays a significant role in sugar metabolism [46]. TPI catalyzes the reversible conversion of dihydroxyacetone phosphate to glyceraldehyde-3-phosphate and is involved in many pathways including glycolysis, Calvin cycle, and glycerol metabolism [15,47]. However, we also detected 4 down-regulated spots including NAD(P)-binding Rossmann-foldcontaining protein (spot 35) and ATP-binding protein (spots [47][48][49] in autotetraploid genotype. NAD(P)-binding Rossmann-foldcontaining protein plays a role in protein folding [48]. ATP binding proteins, a large and important class, are kinases, which transfer the gamma phosphate from ATP to substrates; however, the functions of both proteins in cassava tetraploid genotype are unclear. The differential proteins regarding photosynthesis, carbohydrate and energy metabolism in autotetraploid genotype indicate that polyploidy would differ phenotypically from their diploid progenitors.
The previous studies showed that glutamate-ammonia ligase (GLUL or GS, previous name: glutamine synthetase) (spot 27), a key enzyme in the GS-glutamate synthase cycle, plays a pivotal role in the recycling of NH4 + that is released during photorespiration by generating glutamate from NH4 + and glutamine [49]. There are two isoforms of GS in higher plants, cytosolic GS1 and chloroplastic GS2, with recycling of photorespiratory NH4 + depending on GS2 [50]. The present study demonstrated that cytosolic GLUL at a higher expression in tetraploid genotype than diploid plants. Likewise, translation factor (spot 2) and elongation factor Tu (spots 3,20), which are associated with the process of protein biosynthesis, showed higher expression levels in polyploid plants.
In the present study, cassava tetraploid genotype increased the expressed levels of HCN metabolism-related proteins including linamarase (spot 6) and acetone-cyanhydrin lyase (spot 32) ( Table 1). Linamarase is an enzyme found in the cell walls of cassava. The major cyanogen in cassava is linamarin stored in the Figure 9. Biological networks generated for combination of twelve differential proteins. Fifteen differentially up-regulated proteins including ATP synthase subunit beta, alcohol dehydrogenase, beta-glucosidase, phosphoglycerate kinase, triose phosphate isomerase, RCA, Rubisco, APX2, CDSP3, peroxiredoxin, thioredoxin translation elongation factor, glutamate-ammonia ligase, chaperone and 14-3-3 in cassava autotetraploid genotypes were used to generate a protein-protein interaction network through Pathway Studio analysis. Regulation is marked as an arrow with R, Chemical Reaction as an arrow with C and Binding as an arrow without any marks . The entity table, relation table and reference table data were  presented in in Tables S1, S2,  vacuole. Upon tissue disruption linamarin is deglycosylated bylinamarase, producing acetone cyanohydrin. Acetone cyanohydrin can spontaneously decompose at pH.5.0 or temperature .35uC, or is enzymatically broken down by acetone cyanohydrin lyase, which is hydroxynitrile lyase involved in the catabolism of cyanogenic glycosides [51], to produce acetone and free cyanide [52]. The evidences of HCN detection by HPLC showed that upregulated proteins associated with HCN metabolism could result in the increase of HCN content in cassava autotetraploid genotype in comparison to diploid genotype (Fig. 8I-IV).
The autopolyploid genotypic variation is obvious as 65 proteins were detected showing significant differences in the present study. The proteomic study based on iTRAQ showed that the levels of protein divergence were relatively high between Arabidopsis autopolyploid and allopolyploid plants [22]. Proteomics is a powerful tool to characterize varieties of autopolyploids and allopolyploids [21]. Allopolyploidy can accelerate evolution through rapid and reproducible genomic changes in the first generation of nascent polyploids, including elimination of DNA sequences, gene silencing, alteration of cytosine methylation and activation of genes [53], and sporadic genomic changes that occur during the lifetime of a polyploidy species, which are not possible at the diploid level, such as diversification of homoeoalleles via mutations [54]. The phenomenon would provide an agronomic point of view; constructing allopolyploids could be more pertinent for crop improvement.

Conclusions
The results suggest that 2-DE combined MS and bioinformatics techniques are a valid method to detect the changes of cassava polyploidy genotypes in proteome levels. The present study differs from the previous genomic-level studies of polyploidy in which investigate the differentially expressed proteins in relation with cassava autotetraploid genotype. The functional classification and expression levels of differential proteins showed that polyploid formation would be a complicated process of polyploidization, in which polyploidy provides a reservoir of duplicate genes as substrates for crop potential innovation. Following cassava polyploidization, new gene copies may undergo modifications allowing functional diversification in plants through the biological network of protein-protein interaction and the functional analyses of differential proteins. A useful proteome data set we provided would be helpful to predict the tolerance mechanism at a protein level of cassava polyploidy plants against environmental stresses. Further analysis of the diversity levels and the proteome patterns of duplicate gene not only traces variation of polyploids, but also provide insights into improvement of cassava polyploidy breeding in increasing photosynthesis and resistance efficiencies. Figure S1 Scatter plot showed the ratios of the relative volumes. A, Uses spots for normalization, detected spots,500, reproducible spots,300; B, Scatter plots of reproducible protein spots on two 2-DE images from diploid and autotetraploid genotypes, respectively. (DOC) Figure S2 3-D maps of 65 differential proteins generated by Delta2D software based on the abundance of spots in 2-DE maps. a, differential spots from diploid genotype, b; differential spots from autotetraploid genotype. Blue circles indicated the location of protein spots from diploid genotype; yellow circles indicated spots from autotetraploid genotype. White letters indicated downregulated; black letters indicated up-regulated. The numbering corresponded to the 2-DE gel in Fig. 2