Proteomic analysis of middle and late stages of bread wheat (Triticum aestivum L.) grain development

Proteomic approaches were applied in four grain developmental stages of the Chinese bread wheat Yunong 201 and its ethyl methanesulfonate (EMS) mutant line Yunong 3114. 2-DE and tandem MALDI-TOF/TOF-MS analyzed proteome characteristics during middle and late grain development of the Chinese bread wheat Yunong 201 and its EMS mutant line Yunong 3114 with larger grain sizes. We identified 130 differentially accumulated protein spots representing 88 unique proteins, and four main expression patterns displayed a dynamic description of middle and late grain formation. Those identified protein species participated in eight biochemical processes: stress/defense, carbohydrate metabolism, protein synthesis/assembly/degradation, storage proteins, energy production and transportation, photosynthesis, transcription/translation, signal transduction. Comparative proteomic characterization demonstrated 12 protein spots that co-accumulated in the two wheat cultivars with different expression patterns, and six cultivar-specific protein spots including serpin, small heat shock protein, β-amylase, α-amylase inhibitor, dimeric α-amylase inhibitor precursor, and cold regulated protein. These cultivar-specific protein spots possibly resulted in differential yield-related traits of the two wheat cultivars. Our results provide valuable information for dissection of molecular and genetics basis of yield-related traits in bread wheat and the proteomic characterization in this study could also provide insights in the biology of middle and late grain development.


Introduction
Hexaploid wheat (Triticum aestivum, 2n = 6× = 42, AABBDD) is one of the most important cereals that provides a large proportion of essential nutrients in the human diet. The major constituents of wheat grain are starch (70-80% dry weight) and proteins (10-15% dry weight; Tasleem-Tahir et al., 2012). Of the total wheat grain proteins, the major protein (80%) reserves are the prolamins, which are a mixture of monomeric gliadins and polymeric glutenins located in the starchy endosperm. In contrast to the gliadins and glutenins, the other major protein families of the wheat endosperm, are the non-prolamins, including albumins and globulins (Vensel et al., 2005a).
Wheat grain development is divided into two main stages: (1) grain enlargement, and (2) grain filling and desiccation/maturation. Grain enlargement involves early and rapid division of the zygote and triploid nucleus. Cell division is followed by the influx of water, which drives cell extension. This stage occurs at approximately 3-20 days postanthesis (dpa). During the grain filling stage, cell division slows and then ceases and beginning at around 10 dpa until maturity, storage products are accumulated, at which point the endosperm serves its function as a carbohydrate store (Nadaud et al., 2010). In recent years, different approaches including transcriptomics, proteomics, and metabolomics have been used to understand the diversity and development of grain. However, the expression profiles of accumulated proteins are often poorly correlated with their corresponding mRNAs; e.g., in Arabidopsis (Ruuska et al., 2002), rice (Zhang et al., 2010), and wheat (Dong et al., 2012;Ma et al., 2014). Two-dimensional electrophoresis (2-DE) and mass spectrometry (MS) proteomic approaches have been broadly applied to investigate the dynamic expression profiles of proteins during grain development in different plant species, including Arabidopsis (Ruuska et al., 2002;Li et al., 2007), soybean , maize (Méchin et al., 2007), and rice (Thelen, 2009;Zhang et al., 2012). Further, a significant study on the proteomics of the wheat grain developmental period has been carried out. Proteomic studies on the response to heat stress during grain filling in 10 wheat cultivars indicated that primarily changes in both the amount and activities of enzymes involved in photosynthesis and antioxidant activities contributed to relatively higher heat tolerance (Wang et al., 2015). Identification of proteins in the first 2 weeks of grain development stages showed that a total of 10 clusters of genes were examined in bread wheat (Nadaud et al., 2010). The proteomes of hard and soft near-isogenic wheat lines at four grain developmental stages revealed that kernel hardness is related to the amplification of a stress response during endosperm development (Lesage et al., 2012). Proteome characterization of four grain developmental phases in wheat cultivars Jimai 20 and Zhoumai 16 indicated that differences in seed storage proteins were related to different flour quality performance from these wheat cultivars (Guo et al., 2012).
Ethyl methanesulfonate (EMS) mutants have been widely used as an important method to develop new germplasms in wheat breeding programs due to its high mutant frequency (Henry et al., 2014). The key reasons for this include their highly beneficial mutations, excellent phenotypic characteristics, and novel gene traits. The EMS mutation technique has also reached a mature stage, in which damage in plants is reduced and abundant plant mutations are generated by controlling EMS use. Additionally, EMS mutants have been employed as basic materials in some studies (Botticella et al., 2011;Bonchev et al., 2012;Henry et al., 2014). For example, six EMS-mutagenized lines were validated to improve lodging resistance in Tef (Eragrostis tef ; Zhu et al., 2012). Stay-green and fast-senescing EMS mutated wheat lines with similar anthesis were characterized to investigate the impact on yield and nitrogen partitioning (Derkx et al., 2012).
A combination of 2-DE and EMS mutants in wheat proteomic studies were rarely applied. The Chinese winter wheat cultivar Yunong 201, developed by Agronomy College of Henan Agricultural University, was released as a high-quality noodle wheat cultivar by Henan province in 2006. An elite M 2 line was screened from a large EMS-mutagenized population because of its different plant architecture, larger kernel size, and higher grain weight. This line was self-crossed four times into Yunong 3114. Compared with Yunong 201, Yunong 3114 showed relatively larger kernel size, higher thousand grain weight and higher yield per plot. Therefore, comparison of proteomics of mid and late grain developmental stages of the bread wheat Yunong 201 and Yunong 3114 could provide valuable information for dissection of molecular and genetics basis of yield-related traits in bread wheat, and the proteomic characterization could also provide insights in the biology of middle and late grain development.

Plant Materials
A Chinese winter wheat cultivar Yunong 201 (released no. Yushenmai 2006006) was treated by 0.8% EMS (ethyl methanesulfonate) in 2007. An elite M 2 line was screened from the EMS mutated population containing 2000 lines due to its differential plant architecture, larger kernel size and higher grain weight, which was self-crossed four times into an M 6 line Yunong 3114. Yunong 201 and Yunong 3114 were planted at the Zhengzhou Scientific Research and Education Center of Henan Agricultural University (longitude 113.6 • E; latitude 34.9 • N) during the 2013-2014 cropping seasons under nonstressed natural soil conditions. Differently developmental seeds of Yunong 201 and Yunong 3114 were collected during the postanthesis period based on thermal times that corresponded to the cumulative average daily temperatures as shown in Table 1, and grain size and weight of each sample were investigated (Figure 1). Sampled grains were stored at −80 • C prior to analysis.

Protein Preparation
For two-dimensional gel electrophoresis (2-DE), protein samples with three biological replicates were prepared according to the  method of Gao et al. (2009). Grain samples of 500 mg were extracted in the mid-ear region of each spike, and were ground into a powder in liquid nitrogen with a mortar and pestle. Ten volumes of cold extraction buffer containing 100 mM Tris-HCl (pH 8.8), 10 mM fresh dithiothreitol (DTT), and 10% sodium dodecyl sulfate (SDS) were added, and further ground for 1 h on ice. After centrifuging at 10,000 g for 10 min at 4 • C, the supernatants were collected to new tubes, and an equal volume of phenol was added, following which samples were shaken gently for 30 min, and centrifuged at 14,000 g for 10 min at 25 • C. Below the top phenol phase, the samples were collected to new tubes, and then the above cold extraction buffer was added again to extract once more. The phase of phenol was acquired again, and samples were precipitated with five-fold volumes of cold ammonium acetate/methanol at −20 • C for 2 h. After centrifugation at 14,000 g for 15 min at 4 • C, the supernatants were discarded and the pellets were washed three times in ice-cold acetone containing 5 mM DTT. The pellets were vacuum-dried and resuspended in lysis buffer containing 8 M urea, 2 M thiourea, 4% 3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulfonate (CHAPS), and 20 mM DTT at 25 • C for 2 h according to method of Li et al. (2013). The suspension was centrifuged at 14,000 g for 40 min at 25 • C to remove insoluble materials. Concentrations of total protein were determined by the Bradford assay (Bio-Rad) based on a bovine serum albumin standard (Li et al., 2013). Detailed standard curves with seven different concentrations of BSA (0-100 µg) resuspended in lysis buffer and water in triplicate were shown in Additional file 2 (Data sheets 11, 12).

2-DE and Images Analysis
For 2-DE, 800 µg of protein samples were loaded onto an ReadyStripTM IPG Strip (24 cm, pH 4-7, BIO-RAD, USA) and hydrated passively with 450 µL of protein solution containing 0.5% (v/v) immobilized pH gradient (IPG) buffer (pH 4-7) for 12-18 h at 20 • C using a PROTEAN IEF Cell (BIO-RAD, USA). The first-dimension isoelectric focusing (IEF) was performed with six steps: 250 V for 130 min, 250 V for 90 min, 500 V for 90 min, 1000 V for 2 h, 9000 V for 5 h, and 9000 V for 10 h with a total of 99 kVh and a constant 500 V for the last 12 h. After IEF, the strips were incubated for 15 min in "equilibration buffer I" consisting of 6 M urea, 2% (w/v) SDS, 1.5 M Tris-HCl (pH 8.8), 20% (v/v) glycerol, 0.01% (w/v) bromophenol blue, and 2% (w/v) DTT and then in "buffer II" consisting 6 M urea, 2% (w/v) SDS, 1.5 M Tris-HCl (pH 8.8), 20% (v/v) glycerol, 0.01% (w/v) bromophenol blue, and 2.5% (w/v) iodoacetamide for 15 min. For second-dimension electrophoresis, the strips were transferred to 12% vertical sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) gels. All seed samples were run in triplicate to obtain statistically reliable results. After electrophoresis, gels were fixed in 40% (v/v) methanol and 10% (v/v) acetic acid for 40 min. To visualize the gels, they were stained with staining solution consisting of 0.12% (v/v) Coomassie brilliant blue (CBB) G-250, 20% (v/v) alcohol, 10% (v/v) phosphoric acid, and 10% (w/v) ammonium sulfate, and then destained in double-distilled water . The 2-DE images were scanned at 300 dpi with a UMAX Power Look 2, 100XL scanner (Maximum Tech, Taiwan, China), and quantitative intensity analysis was performed using PDQuest software (version 8.0.1, Bio-Rad, USA). First, the 2-DE gel of 21 dpa seed samples of Yunong 210 and Yunong 3114 was selected as the reference gel. All gels of other stages were matched to the reference gel. Automatic groups formed, and single spots that differed between replicates were manually checked and corrected when necessary. The spots that existed in three independent sample sets were selected. Image quantitative analysis revealed significant differences in protein spot abundance by Student's t-test (abundance variation at least two-fold, P < 0.05).

Two-dimensional Gel Excision, Tryptic Digestion, and Desalting
Protein extracts were separated on preparative gels and 130 proteins of interest were recovered from the gels for identification. Proteins (800 µg) from samples were resolved on separate preparative polyacrylamide gels and were visualized by staining with a modified silver staining method that was compatible with subsequent mass spectrometric analysis (Yan et al., 2000). Protein spots of interest were cut from the preparative gels, destained for 20 min in 30 mM potassium ferricyanide/100 mM sodium thiosulfate (1:1 v/v) and washed with Milli-Q water until the gels were destained. The spots were incubated in 0.2 M NH 4 HCO 3 for 20 min and then lyophilized. Each spot was digested overnight in 12.5 ng/µl trypsin in 25 mM NH 4 HCO 3 . The peptides were extracted three times with 60% acetonitrile (ACN)/0.1% trifluoroacetic acid (TFA). The extracts were pooled and dried completely by a vacuum centrifuge.

MALDI-TOF/TOF Analysis
MS and MS/MS data for protein identification were obtained using a MALDI-TOF-TOF instrument (4800 proteomics analyzer; Applied Biosystems). Instrument parameters were set using the 4000 Series Explorer software (Applied Biosystems). The MS spectra were recorded in reflector mode and a mass range from 800 to 4000 and a focus mass of 2000. MS was used using a CalMix5 standard to calibrate the instrument (ABI 4700 Calibration Mixture).
For one main MS spectrum 25 sub-spectra with 125 shots per sub-spectrum were accumulated using a random search pattern. For MS calibration, autolysis peaks of trypsin ([M+H] + 842.5100 and 2, 211.1046) were used as internal calibrates, and up to 10 of the most intense ion signals were selected as precursors for MS/MS acquisition, excluding the trypsin autolysis peaks and the matrix ion signals. In MS/MS positive ion mode, for one main MS spectrum 50 sub-spectra with 50 shots per sub-spectrum were accumulated using a random search pattern. Collision energy was 2 kV, the collision gas was air, and the default calibration was set using the Glu1-Fibrino-peptide B ([M+H] + 1570.6696) spotted onto Cal 7 positions of the MALDI target. Combined peptide mass fingerprinting (PMF) and MS/MS queries were performed using the MASCOT search engine 2.2 (Matrix Science, Ltd.) that was embedded into the GPS-Explorer Software 3.6 (Applied Biosystems) on the NCBI database with the following parameter settings: 100 ppm mass accuracy, with trypsin cleavage and one missed cleavage allowed, carbamido methylation set as a fixed modification, and oxidation of methionine allowed as a variable modification. Additionally, the MS/MS fragment tolerance was set to 0.4 Da. A GPS explorer protein confidence index ≥95% was used for further manual validation.

Comparison of Grain Size of Yunong 201 and Yunong 3114
Kernel sizes and weights of the grain in both Yunong 201 and Yunong 3114 increased gradually from 21 to 35 dpa, and then decreased from 35 to 42 dpa (Figures 1E,F). Compared with Yunong 201, Yunong 3114 possessed longer kernel length and higher grain weight at the four stages of development (detailed data in Figures 1A-D) but there was no obvious difference on grain width between the two cultivars.

Identification, and Classification of Differentially Accumulated Proteins during Grain Development
Yunong 201 and Yunong 3114 had similar proteomic profiles at the four stages according to the 2-DE protein maps that were extracted from both samples (Figures 2, 3). There were more than 1000 gel spots detected over the gel, and 173 spots were detected that displayed altered abundance, which were then analyzed by mass spectrometry. Finally, 130 spots out of the 173 spots, representing 88 unique proteins, were successfully identified based on BLASTp analyses of NCBI databases (Additional file 1: Data sheets 1-10; Table 2).

Protein Expression Profiles during Grain Development
The expression profiles of the 130 protein spots were investigated by hierarchical cluster analysis ( Figure 5). Four main expression patterns (A-D) were presented and clearly reflected two      distinct grain development phases: grain filling (21, 28), and desiccation/maturation (28-42), as shown in Figure 5. Expression pattern A included 54 protein spots in Yunong 201, and 48 spots in Yunong 3114 that exhibited up-regulation during the four grain developmental stages, which contained many stress/defense-related proteins, such as α-amylase inhibitor 0.19 (protein spots 10 and 19-1), CM 17 protein precursor (protein spots 32 and 33), disease resistance protein RPP13 (protein spot 70), vicilin-like antimicrobial peptides 2-2 (protein spot 71), and 1-Cys peroxiredoxin (protein spot 83), which all accumulated significantly at the desiccation/maturation developmental stages in Yunong 201 and Yunong 3114.
Almost all of the storage proteins including globulin 3 (protein spots 5 and 15), globulin-3A (protein spots 22, 25, and 130), globulin 3B (protein spot 23), and gamma gliadin (protein spot 78) displayed this pattern in both Yunong 201 and Yunong 3114. Besides, the same responses were seen for glutaminyl-tRNA synthetase (i.e., protein spot 101), serpin-Z1C (i.e., protein spot 104), which is involved in protein synthesis/assembly/degradation (Fernando et al., 2015;Kodera et al., 2015) and 27 K protein (Kimoto et al., 2009;i.e., protein spot 76-2), which is involved in transcription/translation. Expression pattern B included the largest proportion of identified proteins whose expression was down-regulated during the mid and late grain developmental stages. Moreover, 71 protein spots in Yunong 201 and Yunong 3114 belonged to this expression group, respectively. All of the proteins associated with glycolysis and most of the proteins involved in starch metabolism, photosynthesis, and energy production and transportation/signal transduction displayed this expression pattern in two samples; for example, ATP synthase CF1 beta subunit (protein spot 107), 2,3-bisphosphoglycerateindependent phosphoglycerate mutase (protein spots 142, 143, and 144), adenosine kinase 2 (protein spots 89 and 90), granule bound starch synthase (protein spot 138), and the 23 kDa oxygen evolving protein of photosystem II (protein spot 166). Expression pattern C showed both down-and up-regulated expression trends, including protein spots 49, 67, and 111, which were seen in the 14-3-3-like protein B, serpin-N3.2, and β-amylase in Yunong 201. There were five protein spots (i.e., 103, 104, 105, 106, and 111) that showed this pattern of expression in Yunong 3114. Unlike expression pattern C, expression pattern D displayed both an up-and down-regulated expression trend. Only protein spot 60 (chitinase 2) belonged to this pattern in Yunong 201, as did protein spots 121 (cerpin-Z2B), 122 (cerpin-Z1C), and protein spot 99 (serpin-N3.2) in Yunong 3114. Protein spots 168 (unnamed) and 167 (unnamed) displayed higher expression levels in Yunong 201 than did Yunong 3114, and protein spot 168 accumulated a single pattern E (not shown in Figure 5) that remained constant during the four grain developmental stage of both cultivars. In addition, protein spot 167 also belonged to this pattern in Yunong 3114, while it displayed expression pattern B in Yunong 201.

Comparative Proteomic Characterization in Yunong 201 and Yunong 3114 during Grain Development
A total of 12 protein spots with different expression patterns co-accumulated in both samples (Figure 5), including stress/defense, protein synthesis/assembly/degradation, signal transduction, starch metabolism, photosynthesis, and the presence of two unnamed proteins. For example, protein spot 67 was identified as serpin-N3.2 that displayed expression pattern C in Yunong 201, and pattern B in Yunong 3114. Protein spot 7 (Nucleoside diphosphate kinase 1) accumulated steadily at the four developmental stages in Yunong 201, and showed pattern E, although it displayed a down-regulated trend in Yunong 3114. Serpin-Z2B, Serpin-Z1C (protein spots 121 and 122) showed expression pattern A in Yunong 201, but displayed expression pattern D in Yunong 3114. In addition, protein spots 103 (ATP synthase subunit beta, mitochondrial), 104 (UTP-glucose-1-phosphate uridylyltransferase) and 105 (ATP synthase beta subunit) showed expression pattern C in Yunong 201, and showed a down-regulated trend in expression in Jimai 20 during the four developmental stages.
Protein spots with two-fold changes or greater in abundance at particular times between the two cultivars were considered as cultivar-different proteins (Guo et al., 2012). Altogether six protein spots displayed cultivar-different proteins during the four developmental stages, which involved three groups: stress/defense, starch metabolism, protein synthesis/assembly/degradation. Among them, protein spot 99 (serpin-N3.2) was only identified in Yunong 3114; however, this protein spot was absent in Yunong 201. Meanwhile, protein spot 169 (i.e., small heat shock protein Hsp 23.5) was not detected in Yunong 3114, which was only identified in Yunong 201.

Discussion
In this study, a Chinese winter wheat cultivar Yunong 201 and its EMS mutant line Yunong 3114 were selected to study the proteomic expression differences during mid and late stages of grain development. Proteomic expression profiles during four grain development stages and cultivar-variable proteins of the Yunong 201 and Yunong 3114 were investigated by 2-DE and MALDI-TOF/TOF-MS. Proteomic characterization in this study could provide insights in the biology of middle and late grain development.

Analysis of Cultivar-Different Proteins in Developmental Seeds of Yunong 201 and Yunong 3114
Up to date, a considerable work has been carried out on wheat grain proteomics through different wheat cultivars (Majoul et al., 2004;Kim et al., 2010), for instance, grain storage proteins (Mamone et al., 2009;Dupont et al., 2011), endosperm and endosperm amyloplasts (Vensel et al., 2005b;Dupont, 2008), and kernel peripheral and aleurone layers (Tasleem-Tahir et al., 2011;Nadaud et al., 2015). Those studies provided the important information on biochemical processes of wheat grain development, however, few studies on proteomic of EMSmutagenized cultivars were conducted in bread wheat. Due to the significant differences on kernel size, thousand grain weight and higher yield per plot of Yunong 201 and Yunong 3114, proteomics analysis of Yunong 201 and Yunong 3114 could provide valuable information for further understanding function of candidate cultivar-different proteins (e.g., serpin for spot 99, small heat shock protein for spot 169, β-amylase for spot 111, α-amylase inhibitor for spot 11, dimeric α-amylase inhibitor precursor for spot 12, and cold regulated protein for spot 35) which were possibly associated with yield-related traits in bread wheat.
Heightened stress interrupts normal protein functions. Small heat shock proteins (sHSPs) are produced in seeds during maturation and under various stress conditions, which can form large multimeric structures and display a wide range of cellular functions, as well as being able to act as molecular chaperones. These sHSPs do this by forming stable complexes with folding intermediates of their protein substrates (Omar et al., 2012;Wu et al., 2014). In our study, the abundance of a sHSPs (spot 169) was up-regulated in Yunong 201, but was absent in Yunong 3114. In addition, stress-related cold regulated protein (spot 35) had a higher abundance in Yunong 201 than in Yunong 3114 during the four grain developmental stages. α-Amylase inhibitors are high molecular weight macromolecules that are particularly abundant in certain cereals and leguminosae, which specifically involved in the degradation of α-1,4-linked sugar polymers, such as starch and glycogen, into oligosaccharides (Franco et al., 2000). α-Amylase inhibitors play important roles in protecting starch and protein reserves in the endosperm against degradation, particularly that caused by biotic stresses like insect attack (Franco et al., 2002). In our study, α-amylase inhibitors (spots 11 and 12) accumulated gradually from 21 to 42 dpa in Yunong 201, however, displayed down-regulated trends during β-Amylase is a starch-degrading enzyme that hydrolytically cleaves α-1,4-D-glucosidic bonds to liberate β-maltose from the non-reducing ends of a variety of polyglucans that are synthesized during grain development, and is one of the major proteins in the starchy endosperm (Yin et al., 2002;Vinje et al., 2011). They can only contribute to starch granule hydrolysis by degrading solubilized intermediates that are released from the granules by α-amylase (Sun and Henson, 1991). The identified protein spot 111 (β-amylase), which displayed pattern C, accumulated at a higher level of abundance in Yunong 201 than in Yunong 3114. This is possibly one of the important reasons to result in differences of grain size and weight between Yunong 201 and Yunong 3114.

Analysis of Protein Spots during the Developmental Stage of Yunong 201 and Yunong 3114
A total of 173 identified protein spots showed more than a two-fold difference in abundance in Yunong 201 and Yunong 3114 at the four stages by means of the classic 2-DE method in this study. Of them, 130 were successfully identified by MALDI-TOF/TOF analysis. The identified protein spots had specific functions in stress/defense, carbohydrate metabolism, protein synthesis/assembly/degradation, storage proteins, energy production and transportation, photosynthesis, transcription/translation, signal transduction, and unknown functional groups. Protein spots with unknown functions were those not identified by interrogating three database including NCBI Triticum, NCBI Viridiplantae, and the Universal Protein Resource (UniProt), probably owing to the lack of a genomic sequence of bread wheat until now. A vast majority of the 130 protein spots had similar proteomic profiles at four stages of development in Yunong 201 and Yunong 3114. However, there were still 12 protein spots exhibiting differential expression patterns and six protein spots exhibiting cultivar-differential expression. It suggests that these six protein spots abovementioned possibly contributed to difference of yield-related traits between Yunong 201 and Yunong 3114. Therefore, the identified protein spots showing differential expression could be used for further digging genes related to yield-related traits in bread wheat and characterization of these protein could also provide new insights into the biology of middle and late grain development in bread wheat.

Stress/Defense
Plants responsiveness to stress entails a complex mechanism and is involved in a large number of enzymes. Chitinase has been proven to play important physiological roles including defense from attack morphological changes, and digestion (Suzuki et al., 2014). Plants with over-expressed Chitinase genes showed stronger disease resistance in different crops (Cletus et al., 2013). In our study, the proteins for spots 60, 61, and 62 were identified as Chitinase 2, which gradually accumulated at four stages of development in Yunong 201 and Yunong 3114, suggesting a vital role of Chitinase in the response to different stress/defense challenges during the mid and late grain developmental periods. The R-proteins recognize pathogenic effectors and activate an efficient defense system that includes a "hypersensitive response (HR)" of programmed cell death or apoptosis at the infection site (Jones and Dangl, 2006). Disease resistance protein RPP13 (i.e., spot 70) is an R-protein, and our proteomic analysis showed that the abundance of RPP13 increased during the grain phase of development, which probably suggested a protective role from pathogens during the grain phase of development during mid and late developmental stages. Vicilin-like antimicrobial peptide 2-3 is a processing product of the 7S globulin precursor that is found in Macadamia integrifolia kernels that display antimicrobial activity (Marcus et al., 1999), the abundance of which (i.e., spots 71 and 100) also gradually increased, which was consistent with RPP13, and might indicate positive roles in antipathogen defense. 1-Cys peroxiredoxin (1-cysPrx) is a novel antioxidant enzyme that reduces phospholipid hydroperoxides, playing an important role in cellular defense mechanisms against oxidant stress (Manevich et al., 2002). The identified protein spots 83 and 84, (1-Cys peroxiredoxin) showed an enhanced trend in expression at the grain phases in Yunong 201 and Yunong 3114. Serpins are likely to participate in a range of biochemical pathways in distinct cell types, tissues and organs in plants to protect cells from oxidative stress, and are highly expressed during seed maturation and occur in tissues during all development stages (Roberts and Hejgaard, 2008). Four types of serpins including serpin-N3.2 (i.e., spots 40, 67, and 99), serpin 1 (i.e., spots 42 and 116), serpin-Z2B (spot 121), and serpin-Z1C (spots 122 and 124) were identified in this study, and spots 121 and 122 showed up-regulated trends in Yunong 201 but down-regulated trends in Yunong 3114. Moreover, spot 67 showed down-regulated trends in Yunong 3114, while it presented a C expression pattern in Yunong 201. In addition, spots 40 and 42 displayed down-regulated patterns of expression during the mid and late grain developmental stages in both cultivars. However, the abundance of spot 124 was increased, which was because of EMS mutagenesis that contributed to four types of serpins displaying differential trends in expression in the two wheat cultivars. Therefore, proteomic studies in this study could also be used in practical applications such as breeding for an enhanced stress tolerance as suggested by Kosova et al. (2014).

Carbohydrate Metabolism
Glycolysis provides energy and intermediates for the synthesis of metabolites, of which, we identified nine protein spots representing six types of proteins. These were referred to fructokinase-2 (spot 87), phosphoglycerate kinase, cytosolic (spot 123), enolase (spot 115), cytosolic 3-phosphoglycerate kinase (spot 127), UTP-glucose-1-phosphate uridylyl transferase (spot 104 and 106), and 2,3-bisphosphoglycerate-independent phosphoglycerate mutase (spot 142, 143, and 144), with the exception of the abundance of spot 104 (UTP-glucose-1phosphate uridylyl transferase), which decreased in Yunong 201, and showed an expression pattern C in Yunong 3114. The remaining spots all showed down-regulated patterns of expression from the 21 to 42 dpa. Moreover, 21 dpa belonged to the late stage of wheat filling, which has enhanced glycolysis that is a significant source of energy for grain filling and accumulation of dry matter. Thus, this coincided with the synthesis stage of starch. During the filling stage, and due to the sharp accumulation of starch, ATPases are activated to provide more energy demands for an organism. Our research suggested that the reduced abundance of energy production from 21 dpa, as indicated by the down-regulated trend of the ATP synthase CF1 beta subunit (spot 107), ATP synthase subunit alpha, mitochondrial (spot 132 and 135), adenosine kinase 2 (spot 89 and 90), and the ATP synthase CF1 alpha subunit (spot 154). However, spot 103 (ATP synthase subunit beta, mitochondrial), and spot 105 (ATP synthase beta subunit) displayed an enhanced trend at the grain mature period in Yunong 3114.

Starch Synthesis and Storage Proteins
Starch is the major energy reserve for a large variety of higher green plants, such as cereals, legumes, and tubers (Miao et al., 2015). The biosynthesis of starch is the major determinant of overall yield in cereal grains (Emes et al., 2003). In all plant tissues capable of starch biosynthesis, adenosine diphosphate glucose (ADPGlc) pyrophosphorylase (AGPase, EC 2.7.7.27) is the enzyme that is responsible for the production of ADPGlc, the soluble precursor and substrate for starch synthases. The AGPase reaction is the first committed step in the biosynthesis of stored starch in amyloplasts (Tetlow et al., 2004). Our results demonstrated that ADP glucose pyrophosphorylase (spot 54), ADP-glucose pyrophosphorylase large subunit (spots 92 and 153), and small subunit ADP glucose pyrophosphorylase (spot 141) were all down-regulated from the filling stage to trace levels at the desiccation phase, which matched the increase in starch content and grain weight during the mid and late grain developmental stages. Glycogen synthase catalyzes the formation and elongation of the α-1,4-glucose backbone using ADP-glucose, the second and key step of glycogen biosynthesis. Elongation and branching of amylopectin is a complex process and it requires an array of enzymes including starch synthases (SS), starch branching enzymes (SBE), and debranching enzymes (DBE). However, synthesis of amylose is brought about solely by the enzyme granule-bound starch synthase I (GBSSI) or waxy protein (Ahuja et al., 2014).
In our study, the abundance of granule bound starch synthase (spot 138) was decreased from the 21 dpa, which belonged to the late stage of grain filling, which was consistent with the filling phase, and the most vital stage of starch accumulation. We identified one β-amylase that was involved in seven spots (i.e., 109, 111, 114, 119, 125, 128, and 133), except for spot 111. The remaining six spots showed a down-regulated trends. There were three types of globulin that were identified at the four grain stages, including the globulin 3 (spot 5 and 15), globulin-3A (spot 22, 24 and 25), and globulin 3B (spot 23). Further, with the exception of gamma gliadin (spot 78) and avenin-like seed protein (spot 74), they accumulated gradually at the mid and late developmental periods in both samples. In general, biosynthesis of seed storage protein is dependent of amino acid synthesis and the transport of nitrogen metabolism (Hernández-Sebastià et al., 2005). We identified one alanine aminotransferase 2 (i.e., spot 151) and one glutaminyl-tRNA synthetase (i.e., spot 101). However, they displayed the opposite expression trend for the two wheat cultivars.
Other Functional Proteins 14-3-3 Proteins function as homodimers or heterodimers and bind a large number of differentially phosphorylated substrates to regulate a wide array of cellular signaling and physiological processes (Lozano-Durán and Robatzek, 2015). We identified one 14-3-3 protein (i.e., spots 46, 47, and 48), and one 14-3-3-like protein B (i.e., spots 49 and 50), among them, 14-3-3 proteins were down-regulated in two samples, while the abundance of protein spot 50 increased in Yunong 201 and Yunong 3114. Nevertheless, the other 14-3-3-like protein B (i.e., spot 49) displayed a steady expression trend in Yunong 201, but was gradually down-regulated at the four stages of development in Yunong 3114. The cytosolic NDPK1 is the main nucleoside diphosphate kinase (NDPK) isoform in plants, which operates in the context of homeostasis of cellular nucleoside triphosphate (NTP) pools that accounts for more than 70% of total NDPK activity in plants (Prabu et al., 2012). In our study, the abundance of nucleoside diphosphate kinase 1 (spot 7) was gradually reduced in Yunong 3114; however, it showed a steady expression trend during the four developmental grain stages in Yunong 201. Thus, nucleoside diphosphate kinase 1 might play a different role in signal regulation in the two wheat cultivars.

Conclusions
In our study, 2-DE and tandem MALDI-TOF/TOF-MS were implemented to characterize protein accumulation at the middle and late stages of grain development in Yunong 201 and Yunong 3114, which differ by grain weight and size. Totals of 130 differentially accumulated protein spots representing 88 unique proteins were identified and they showed four main expression patterns in Yunong 201 and Yunong 3114. Moreover, six cultivardifferent protein spots were examined. These included cultivardifferent protein spot 111 (β-amylase), which accumulated at much higher abundance in Yunong 201 than in Yunong 3114. This difference was possibly related to the difference in grain size and weight between the two wheat cultivars. In addition, the absence or down-regulation of three protein spots identified as 11, 12, and 169 in Yunong 3114 were all related to stress/defense, the results possibly revealed that Yunong 201 and Yunong 3114 possessed differential adaptation to abiotic stress. Our results could provide valuable information for dissection of molecular and genetics basis of yield-related traits in bread wheat as well as new insights into the biology of late grain development.

Author Contributions
FC and DC designed the project. NZ and WH performed the experiments. NZ and FC wrote the paper. NZ performed the analyzed the data. All authors read and approved the manuscript.