Jian Luo,Zhui Li,Fei Mo,Yuncheng Liao,Yang Liu*

College of Agronomy,Northwest A&F University,Yangling 712100,Shaanxi,China

Keywords:

A B S T R A C T Poor filling and low weight of inferior kernels limit the further improvement of wheat yield.Two cultivars,Shuangda 1 and Xinong 538,with different grain weights,were selected to investigate the physiological changes of inferior kernels by removal of superior kernels(RS)at the flowering stage.iTRAQ combined with physiological indexes was used to identify factors limiting the filling of inferior kernels.Removal of superior kernels significantly increased the mean filling rate of inferior kernels and thus increased their weight.A set of 6012 proteins in inferior wheat kernels were differentially expressed between the RS and control.These differentially expressed proteins were involved mainly in carbon metabolism and energy metabolism.The main reason for the promoting effect of RS on the filling of inferior kernels may be that RS downregulated proteins involved in glycolysis and pyruvate metabolism while upregulating proteins involved in carbon fixation and photosynthesis.Consequently,RS greatly increased the ATP content in inferior kernels,supplying energy for them to absorb photosynthetic assimilates.Removal of superior kernels increased the activities of sucrose synthase,soluble starch synthase,adenosine diphosphate glucose pyrophosphorylase,and starch branching enzyme in inferior kernels and promoted starch accumulation in them.Thus,RS promoted the filling of inferior kernels and increased their weight.The promoting effect of RS on starch synthesis in inferior kernels was associated with their endogenous IAA and ABA levels.

1.Introduction

Grain filling determines grain weight and affects wheat yield[1–3].Wheat kernels can be divided into superior kernels(SK)and inferior kernels(IK)according to their position on the spike and their grain filling[4,5].Poor filling and lower weight of IK are among the main factors limiting the improvement of wheat grain weight and yield[6].Identifying the causes of poor filling of IK and adopting appropriate measures to improve the trait could increase wheat grain weight and achieve high yield potential.

Studies of the physiological mechanism of grain filling of cereal crops with SK and IK have considered leaf photosynthesis[7,8],kernel morphology development[9],carbohydrate transport[6,10],hormones[11],enzyme activity[12],and other processes.Liang et al.[6]suggested that poor accumulation of carbohydrates in IK leads to lower grain weight and yield.Lower activity of sucrose–starch synthesis pathway enzymes led to lower starch accumulation and weight of IK[10–13].Hormone balance may regulate cereal grain filling[11].These studies focused mainly on physiological differences between SK and IK.Whether there is an interaction between SK and IK in the filling process and the relationship between the physiological metabolism of SK and IK are still unknown.SK and IK were reported[14]to compete in rice kernel filling.

With advances in molecular biology,research into the development mechanism of IK at the molecular level has been reported[15,16].Proteomics can detect simultaneous changes in multiple proteins,shedding light on complex metabolic processes,and has been used to investigate the mechanism of poor grain filling of IK in rice[17,18]and maize[19].Isobaric tags for relative and absolute quantitation(iTRAQ)study[20]suggested that in the early grain-filling stage,the proteins expressed in wheat kernels are involved mainly in carbohydrate and nitrogen metabolism and cell division,whereas proteins such as W5AIU1,Q8W3V4,D2KFG9,and W5DVL2 play important roles in grain filling in the middle and later grain-filling stages of wheat[21].However,there have been few studies of proteome changes during wheat SK and IK development in wheat.

Removing superior kernels promoted filling of IK in rice[14].However,it is unknown whether the filling of IK is affected by SK in wheat or whether competition between SK and IK accounts for the poor filling of wheat IK.

The purpose of this study was to investigate whether grain filling of wheat IK could be increased by changing their relationship with SK.The experimental approach was to measure changes in protein expression,hormone content,and starch accumulation in IK after removal of SK.

2.Materials and methods

2.1.Experimental design and treatment

Field experiments were conducted at the Crop Specimen Farm(34°22′N,108°26′E)of Northwest A&F University from 2016 to 2018.The altitude of the experimental site is 467 m.The mean annual temperature is 12.9 °C and the mean annual precipitation is 550 mm.The temperature and precipitation of the experimental site during the wheat experiment are shown in Fig.S1.Table S1 shows the status of the soil nutrient content in the 0–20 cm layer at the site.Before sowing,375 kg ha-1urea and 375 kg ha-1diamine phosphate were applied.Standard field management was conducted during the wheat growth period.

Two cultivars,Shuangda 1 and Xinong 538,were selected and sown on October 12,2016,and October 14,2017,at a seeding density of 150 kg ha-1and a row spacing of 0.25 m.There were two treatments for each cultivar,SK removal(RS)and control(CK).SK and IK were defined following Jiang et al.[4]with minor modification.The basal kernels(the first and second kernels from the bottom of the spikelet)in the middle spikelet were defined as SK and the distal kernels(the first and second kernels from the top of the spikelet)in the middle spikelet were defined as IK.As in our previous study[11],when there were only three kernels on the spikelet,as was common in Xinong 538,the two kernels at the base were classified as SK and the one at the distal end as the IK.For Xinong 538,spikelets 6–17 from the bottom were middle spikelets,and for Shuangda 1,spikelets 5–21 from the bottom were middle spikelets.For the RS treatment,the ovary of SK was removed with tweezers at the flowering stage in 360 spikes in each plot.CK plants were not treated.The experiment was laid out in a randomized block design with each plot area 9 m2(3×3 m),and each treatment repeated 3 times.

2.2.Sampling and measurement

Approximately 360 wheat spikes from the same plot that bloomed on the same day were marked at the flowering stage.Samples were harvested every five days from flowering to maturity.Marked wheat was harvested from each plot:60 spikes at 5 and 10 days after anthesis(DAA)and 40 spikes from each plot during the subsequent sampling period.Half of the sampled spikes were quickly frozen in liquid nitrogen and then stored at-80 °C.The remaining spikes were used for morphological observation and measurement of grain weight,sucrose,and starch.

2.3.Observation of grain filling and morphology

The SK and IK were photographed under a stereoscopic microscope to observe changes in kernel morphology.Kernel weight was recorded.According to the growth equation of Richards[22]and the study of Yang et al.[23],the filling process of grain was simulated.

The derivative of equation(1)was used to calculate the grain filling rate(G)

In the equation,Wis kernel weight(mg);Ais final kernel weight(mg);trepresents the time after anthesis(day);andB,kandNare coefficients determined by regression.The active grain-filling period was defined as the period when the kernel weight was between 5%(t1)and 95%(t2)of the final weight.The mean grain-filling rate during this period was accordingly calculated fromt1tot2[23].

2.4.iTRAQ analysis

The small-grain cultivar Xinong 538 is one of the main wheat cultivars on Guanzhong Plain.In 2016–2017,Xinong 538 was used for iTRAQ analysis.IK from the RS treatment and CK at 10 and 20 DAA(early and middle stages of grain filling)were selected for iTRAQ analysis following Geng et al.[24].Total protein was extracted with cold acetone and its quality was assessed by SDSPAGE.The protein was then digested with sequence-modified trypsin(Promega,Madison,WI,USA)at 37°C.The peptide mixture was labelled with an iTRAQ/TMT tag,separated by strong cation exchange,and then separated by liquid chromatography–tandem mass spectrometry(LC-MS/MS).The MS data were converted into MGF files by ProteomeDiscovery1.2(Thermo,Pittsburgh,PA,USA)and analyzed with the Mascot(2.3.02)search engine(Matrix Science,London,UK).The Mascot search results were averaged using medians and quantified.The relative expression levels of proteins were calculated for the comparison RS vs.CK.When relative expression fold changes were＞1.2 or＜0.83 and the unadjusted significance levelPwas＜0.05,the protein was considered to be differentially expressed.Differentially expressed proteins(DEPs)were annotated with the GO(Gene Ontology)and KEGG(Kyoto Encyclopedia of Genes and Genomes)databases to obtain their functions.Enriched GO functions and metabolic pathways of these DEPs were identified using the conditionP＜0.05.

2.5.Determination of sucrose and starch contents

Sucrose was extracted from 0.1 g of powdered grain with 80%ethanol(v/v)at 80 °C and then centrifuged at 3000×gfor 15 min.The concentration of sucrose in the supernatant was determined by spectrophotometer using the resorcinol–HCl method.After sucrose was extracted,the residue was extracted with 36 mol L-1perchloric acid and then with 18 mol L-1perchloric acid,and the two extract supernatants were mixed and used for starch determination.Starch content was measured by spectrophotometer by the anthrone method[25].

2.6.Determination of enzyme activities associated with starch metabolism

The activity of sucrose synthase(SS)in kernels was determined following Jiang et al.[4].A sample of 0.6 g of fresh grain was used for enzyme extraction with 1.5 mL mixed solution of 50 mmol L-1Hepes–NaOH(pH 7.5),10 mmol L-1MgCl2,2 mmol L-1EDTA,50 mmol L-12-mercaptoethanol,12.5%(v/v)glycerol,and 5%(w/v)insoluble polyvinylpyrrolidone-40.The activity of SS was calculated from the formation of UDP glucose-dependent sucrose catalyzed by SS.

The enzyme activities of adenosine diphosphate glucose pyrophosphorylase(AGPase),soluble starch synthase(SSS),granule-bound starch synthase(GBSS),and starch branching enzyme(SBE)were determined using the kit produced by Keming Biotechnology Co.,Ltd(Suzhou,China).SSS and GBSS catalyze the reaction of ADPG with dextran to transfer glucose molecules to dextran and ADP is produced at the same time.The activities of SSS and GBSS were calculated from the formation of NADPH in the proportional production of NADPH and ADP.The addition of hexose phosphate mutase and glucose-6-phosphate dehydrogenase to AGP catalyzed the formation of 6-phosphogluconic acid and NADPH.AGP activity was determined by the increasing concentration of NADPH.The activities of GBSS,SSS,and AGPase were determined as the amount of nmol per minute of NADPH produced per gram of fresh grain.Because SBE removes the side branch of amylopectin and reduces the absorption at 660 nm of starch iodine complex,SBE activity was determined as the percentage decrease in absorbance at 660 nm,which was the decrease in the iodine blue value(as a unit of enzyme activity).

2.7.Measurement of ATP content

The ATP content was determined by the phosphomolybdic acid colorimetric method using the kit produced by Gareth Biotechnology Co.,Ltd(Suzhou,China).ATP was extracted from 0.3 g of fresh seeds and then determined according to the kit instructions.Before the determination,a pre-experiment was performed to determine the content of ATP in the sample for adjusting the amount of reagent.

2.8.Hormone determination

Fresh kernel samples of 0.5 g were used to extract endogenous IAA,ABA,and zeatin and zeatin riboside(Z+ZR).The extraction method followed Yang et al.[26].The sample was homogenized with 80%(v/w)methanol containing 1 mmol L-1butylated hydroxytoluene.After centrifugation for 15 min at 4000 r min-1,the supernatant was passed through a ChromoSep C18 column(C18 Sep-Pak Cartridge,Waters Corp.Milford,MA,USA).The fraction was dried under a nitrogen blower and dissolved in 1 mL of phosphate-buffered saline containing 0.1%(v/w)Tween 20 and 0.1%(w/w)gelatine(pH 7.5).The solution was used for the determination of hormones by enzyme-linked immunosorbent assay(ELISA).The ELISA kit was produced by the Institute of Plant Hormones,China Agricultural University(Beijing,China).The quantification of ZR,IAA,and ABA was performed by ELISA as previously described[26].

2.9.Data analysis

Statistic Package for Social Science software(SPSS 16.0)(IBM,New York,USA)was used for statistical analysis.Significance was tested by least significant difference(LSD,P＜0.05).

3.Results

3.1.Grain-filling characteristics

Wheat kernel color changes during ripening from white to green and then to brownish yellow.After removal of SK,the morphological changes of IK with respect to kernel length and width were similar to those of the SK of CK(Fig.S2),indicating that the IK of the RS treatment developed faster and the IK of RS were fuller than those in the CK treatment(Fig.1A,B).RS treatment increased the weight of IK relative to CK(Fig.1 C–F).The weight of IK in the RS treatment was close to that of the SK in the CK treatment at the early filling stage(5–20 DAA)but heavier at the mature stage.The results in the two experimental seasons were similar.

The removal of SK significantly increased the average filling rate of IK(Fig.S3A,B),but there was no significant difference between their active grain filling periods(Fig.S3C,D).The trials showed similar results in the two years.SK removal increased the grainfilling rate,promoting the filling of IK.

3.2.Basic information of iTRAQ

iTRAQ yielded 397,734 spectra from RS-IK and CK-IK at two stages(10 and 20 DAA).Based on these spectra,2387 DEPs were identified(Fig.S4A).In comparison with CK,removal of SK caused a large change in protein expression in IK at 10 and 20 DAA.At 10 DAA,compared with the IK of CK,269 proteins were downregulated and 230 proteins were upregulated in the IK of the RS treatment.However,the number of DEPs at 20 DAA was markedly lower than that at 10 DAA,with only 40 DEPs upregulated and 73 downregulated in IK of RS compared to the IK of the CK treatment(Fig.S4B).

3.3.GO and KEGG pathway enrichment analysis

In general,GO enrichment analysis suggested more upregulated than downregulated DEPs in cellular components,molecular functions,and biological processes(Fig.2).In the cellular components,these DEPs were distributed mainly in cells,cell parts,and organelles.In the molecular function group,the DEPs were associated mainly with binding,catalytic activity,and membrane and structural molecule activity,of which the protein content of binding and catalytic activity was the highest.In the context of biological processes,DEPs were involved mainly in cellular processes,metabolic processes,localization,single-organism processes,and biological regulation.

Fourteen significantly enriched pathways were identified at 10 DAA(Fig.3A),of which the number of DEPs in the phagosome pathway accounted for up to 8.46%of DEPs,followed by carbon fixation in photosynthetic organisms(8.08%)and RNA transport(7.69%).Some DEPs were involved in oxidative phosphorylation(6.54%).There were 11 significantly enriched pathways at 20 DAA(Fig.3B).Among them,the number of DEPs enriched in the ribosome pathway accounted for 20.3% of DEPs,followed by glyoxylate and dicarboxylate metabolism(15.2%)and starch and sucrose metabolism(10.2%),and some differentially expressed proteins participated in carbon fixation in photosynthetic organisms(8.47%)and nitrogen metabolism(8.47%).From the perspective of KEGG classification,the pathways involved in metabolic processes accounted for the largest proportion in the two periods.Among the metabolic processes,those associated with carbon metabolism and energy metabolism accounted for the largest proportion.At 10 DAA,75 DEPs were involved in carbon metabolism and energy metabolism(Fig.3C),including 46 involved in carbon metabolism(25 upregulated and 21 downregulated)and 29 involved in energy metabolism(15 upregulated and 14 downregulated).At 20 DAA,there were 25 DEPs involved in carbon metabolism and energy metabolism(Fig.3D),including 10 involved in carbon metabolism(6 upregulated and 4 downregulated)and 15 involved in energy metabolism(6 upregulated and 9 downregulated).Thus,the removal of SK strongly influenced carbon and energy metabolism in IK.

Fig.1.Changes in kernel morphology during the grain-filling period and effect of superior-kernel removal on kernel weight change during the grain-filling period.RS,removal of superior kernels;CK,control;superior and inferior kernels are normal.SK,superior kernels.IK,inferior kernels.The vertical bar represents the mean±standard deviation(n=3).

3.4.The relationships of DEPs involved in carbon metabolism and energy metabolism

The relationships in which DEPs participated in the metabolic process were characterized(Fig.4).Most proteins involved in pyruvate metabolism and glycolysis pathways were significantly downregulated,including acetyl-CoA carboxylase(ACACA),pyruvate kinase(PK),phosphoenolpyruvate carboxylase(PPC),and pyruvate dikinase(PPDK).In contrast,most of the DEPs involved in the carbon fixation pathway were significantly upregulated,including ribulose-bisphosphate carboxylase large chain(RBCL),glyceraldehyde 3-phosphate dehydrogenase(GAPDH),triosephosphate isomerase(TPI),fructose-bisphosphate aldolase(ALDO),and malate dehydrogenase(MaD).Sucrose-phosphate synthase(SPS),beta-fructofuranosidase(INV),glucose-1-phosphate adenylate transferase(GLGC),trehalose 6-phosphatesynthase(T6PS),trehalose 6-phosphatephosphatase (T6PP), 1,4-alpha-glucan branching enzyme(GLGB),and glycogen phosphorylase(GLGP)were downregulated(Fig.4).Cytochrome b6-f complex ironsulfur subunit(Cb6-f),F-type H+-transporting ATPase subunit epsilon(ATPase),and photosystem I subunit II(PS)were significantly upregulated.

3.5.Sucrose and starch contents

Fig.2.Gene Ontology(GO)classification of differentially expressed proteins(DEPs)in RS-IK and CK-IK by the comparison RS vs.CK at 10 and 20 DAA.Red and blue bars indicate GO enrichment at 10 and 20 DAA,respectively.Down,downregulated DEPs;Up,upregulated DEPs;DAA,days after anthesis.

Based on the iTRAQ results,changes in starch and sucrose contents in the kernels were measured(Figs.5,S5).Removal of SK reduced the sucrose content in the IK of Xinong 538 and Shuangda 1 at 5 and 10 DAA compared to the CK treatment(Figs.5A,B,S5A,B).In contrast,removal of SK reduced the starch content of IK compared to CK from 20 DAA to maturity(Figs.5C,D,S5C,D).Starch accumulation in RS-IK of Xinong 538 and Shuangda 1 was higher than that in CK-SK and CK-IK from 5 DAA to maturity(Figs.5E,F,S5E,F).

3.6.Sucrose-to-starch-associated enzyme activity

The activities of SS,SSS and SBE in kernels of Xinong 538 and Shuangda 1 in the RS-IK at 5 and 10 DAA were significantly higher than those of the CK-IK and close to those of the CK-SK(Fig.6A,D,E).In Xinong 538,GBSS activity in kernels of RS-IK was significantly higher than that of CK-IK at 5 DAA(Fig.6B).With respect to AGP enzyme activity,Xinong 538 and Shuangda 1 showed the same trend,and the AGP activity of RS-IK was significantly higher than that of CK-IK at 10 DAA(Fig.6C).

3.7.ATP content

The RS-IK showed a significantly higher ATP content than CK-SK and CK-IK at 10 DAA,for both Xinnong 538 and Shuangda 1(Fig.S6).At 20 DAA,there was no significant difference among RSIK,CK-SK,and CK-IK in the two cultivars.

Fig.3.KEGG classification of differentially expressed proteins(DEPs)between RS-IK and CK-IK in the comparison RS vs.CK.Metabolism pathways shown in the figure was significantly enriched(P＜0.05).(A)Differential enrichment of proteins at 10 DAA.(B)Differential enrichment of proteins at 20 DAA.The percent labels on bars refer to the proportion of DEPs in the pathway relative to the total of differential proteins.The small plots(C and D)show the up-and down-regulation of DEPs involved in energy metabolism and carbon metabolism in these significantly enriched metabolic pathways.DAA,days after anthesis.

Fig.4.The main differentially expressed proteins(DEPs)involved in carbon metabolism and energy metabolism.(A)Metabolic relationship of DEPs.(B)Statistics of DEPs involved in carbon metabolism and energy metabolism.The labels on arrows represent differentially expressed enzymes involved in each kind of metabolism.The relative expression levels of DEPs were calculated for the comparison RS vs.CK.There were differentially expressed enzymes on the blue(10 days after anthesis)and gray(20 days after anthesis)arrows.The vertical coordinate of the bar chart in(A)represents the relative expression of DEPs.The red arrows indicate proteins that were significantly differentially expressed at both 10 and 20 days after anthesis.Dotted arrows indicate processes with more than one step.Different colors in the network represent different metabolic pathways.T6PP,trehalose 6-phosphate phosphatase.T6PS,trehalose 6-phosphate synthase.SPS,sucrose-phosphate synthase.INV,beta-fructofuranosidase.GLGC,glucose-1-phosphate adenylyltransferase.GLGP,glycogen phosphorylase.GLGB,1,4-alpha-glucan branching enzyme.A1E,aldose 1-epimerase.ALDO,fructose-bisphosphate aldolase.TPI,triosephosphate isomerase.GAPDH,glyceraldehyde 3-phosphate dehydrogenase.PK,pyruvate kinase.AD,aldehyde dehydrogenase.FBA,fructose-bisphosphate aldolase.RBCC,ribulose-bisphosphate carboxylase large chain.G3PD,glyceraldehyde 3-phosphate dehydrogenase.PPC,phosphoenolpyruvate carboxylase.MaD,malate dehydrogenase.PPDK,pyruvate orthophosphate dikinase.ACACA,acetyl-CoA carboxylase/biotin carboxylase 1.ALDH,aldehyde dehydrogenase(NAD+).GLOA,lactoylglutathione lyase.GLDC,glycine dehydrogenase.GLYA,glycine hydroxymethyltransferase.RBCL,ribulose-bisphosphate carboxylase large chain.GLUL,ribulosebisphosphate carboxylase large chain.GLU,glutamine synthetase(ferredoxin).ATPase,F-type H+-transporting ATPase.Cb6-f,cytochrome b6-f complex iron-sulfur subunit.PS,photosystem I subunit II.

3.8.ABA,IAA and Z+ZR contents

In comparison with the CK-IK,the contents of ABA and IAA in the RS-IK increased significantly,but there was no significant difference in Z+ZR content(Fig.7).The peak values of the ABA and IAA contents of RS-IK and CK-SK appeared earlier than that of CK-IK.The peak values of RS-IK and CK-SK appeared at 15 DAA and those of CK-IK at 20 DAA(Fig.7A–D).

4.Discussion

4.1.Effect of removing SK on weight of IK

Fig.5.Effect of removal of SK on sucrose and starch content and starch accumulation of IK in 2016–2017.Vertical bars represent standard deviation(n=3).Values within the same day followed by different lowercase letters are significant at P=0.05.

SK were better filled,with higher kernel weight,than IK,and removal of SK increased filling and weight of IK.This finding is consistent with previous results[9,14,15,19].In cereal crops,IK are smaller than SK[15]as well as having fewer endosperm cells,resulting in lower length,width,and weight[19].Following removal of the SK of a large-spike-type rice cultivar,the grain filling of the IK was improved,and the IK reached a higher grain weight than the control at maturity[14].Similar results were achieved in this study.Removal of SK increased filling of IK and increased their weight by increasing their mean filling rate rather than the length of the active grain-filling period(Fig.2).These results support a competitive relationship between the development of SK and IK during grain filling and suggest that the removal of SK increases the filling of IK.

4.2.Changes in protein expression in IK after removal of SK

Based on iTRAQ results,Yang et al.[20]suggested that during the wheat development period,most DEPs participate in metabolic processes,including carbohydrate metabolism,cell division,protein synthesis,and signal transduction.You et al.[27]found that after SK removal in rice,approximately half of the DEPs in IK were involved in carbohydrate metabolism(sucrose-to-starch metabolism and energy metabolism)and protein metabolism(protein synthesis,folding,degradation,and storage),and these DEPs were associated with poor grain filling in IK.The observation that there were more DEPs at 10 than at 20 DAA(Fig.S2)shows that the removal of SK promotes grain filling mainly in the early grain filling stage.The finding from KEGG analysis that these DEPs were involved in carbon metabolism and energy metabolism processes is consistent with the findings of a previous study in rice[27].The DEPs were involved mainly in pyruvate metabolism,carbon fixation,starch sucrose metabolism,and photosynthetic pathways.These metabolic processes support starch accumulation.In rice,the 14-3-3 protein GF14f regulates kernel length and weight to increase grain yield[18].However,there are few studies of the key regulatory proteins in differentially expressed proteins in wheat,and further research may identify key regulators of wheat grain filling.

4.3.Effect of removal of SK on carbon metabolism in IK

Fig.6.Effect of removal of SK on the activities of enzymes involved in sucrose-to-starch metabolism in IK.Vertical bars represent standard deviation(n=3).Values within the same day followed by different lowercase letters are significant at P=0.05.

According to the results of iTRAQ analysis on Xinong 538 in 2016–2017,we analyzed the relevant physiological indicators in the 2016–2017 and 2017–2018 seasons to better explain the mechanism of the change of IK.The growth and development of wheat are influenced by carbon metabolism.Starch is the main component of wheat grain,accounting for 65%–70% of its weight[28].In large-panicle-type high-yielding rice,an insufficient supply of starch assimilates led to poor grain filling[29,30].However,there was no difference in soluble carbohydrate content between IK and SK in the early stage of maize[19].Zhang et al.[31]reported that IK had sufficient soluble carbohydrates at the early grain filling stage.These results suggest that the supply of carbohydrates is not the key limiting factor for the filling of IK.In the present study,the sucrose content in IK was higher than that in SK in the early filling stage,and the removal of SK reduced this sucrose content.This result suggests that the poor filling of wheat IK is not caused by limitation of assimilates in the early stage.The synthesis of wheat starch is influenced by the transport of assimilate to the grain and its conversion to starch.Photoassimilates are first transported to the kernels from stalks via catalysis of cell wall invertase,and then starch synthesis is catalyzed by AGPase,GBSS,SSS,SS,SBE,and other enzymes of starch synthesis[32,33].Starch synthesis is restricted by energy conditions and regulated by hormones[26,32,34].Thus,starch synthase activity,hormone content,and energy status reflect the activity of sink.The slow starch synthesis rate of IK was caused by the poor sink strength of that in the early filling stage[32],such as lower enzyme activity[32],unbalanced hormone content[26],and lower energy state[34].

Fig.7.Changes of ABA,IAA and Z+ZR contents during grain filling in wheat kernels.Vertical bars represents standard deviation(n=3).Values within the same day followed by different lowercase letters are significant at P=0.05.ns indicates that there is no significant difference among treatments within the same day.

Starch is the main component of wheat grain.The lower sink strength likely prevented the IK from making full use of carbohydrates transported from the stem.As a result,the IK had lower starch accumulation and kernel weight[32].Wheat grain yield is correlated with the activities of starch synthesis enzymes such as AGPase,GBSS,SSS,and SBE[33].An increase in enzyme activities increases cereal grain weight[33,35].However,study[35]has suggested that the final grain weight of wheat is not limited by the activity of key starch synthases.Our results showed that SK with relatively high weight had higher SSS,AGPase and SS activities at 10 DAA than IK with lower weight.At the same time,the starch accumulation in SK was significantly higher than that in IK,suggesting that the increase in enzyme activity promotes the accumulation of starch in kernels.After the removal of SK,the activities of SS,SSS,and SBE in IK increased,and the activity of AGP also increased at 10 DAA.The starch accumulation and weight of IK also increased markedly after SK removal.The RS treatment increased the activity of enzymes involved in starch synthesis and reduced the content of sucrose,indicating that the synthesis of sucrose to starch was carried out efficiently.Although the starch content of IK did not increase,the starch accumulation and weight in IK also increased,possibly because kernel size increased faster than starch synthesis.These results suggest that the activities of starchsynthesis enzymes may be the key factor limiting the grain filling of IK in wheat.Removal of the SK released the competition between SK and IK and promoted the starch synthesis of IK and increased the starch accumulation and weight of IK.

Plant hormones influence the sink strength and starch synthesis of kernels[26].Cytokinin is considered to be necessary for cell division in the early stage of grain development[36,37].IAA and ABA promote sink strength and carbohydrate transport from rice and wheat stems to kernels,and these two hormones are involved in starch synthesis[26,38].In the present study,removal of SK increased the contents of ABA and IAA in IK.Thus,the promoting effect of RS on the grain filling of IK was associated with IAA and ABA levels in wheat kernels.Our previous study[11]suggested that the high ABA level in grain favors wheat kernel starch accumulation.These results suggest that increasing IAA in the RS-IS may promote endosperm cell division,thus increasing the sink strength of IK and increasing their potential to absorb carbohydrates transported from the stem[14].Removal of SK increased the ABA content in kernels,promoting starch synthesis in IK.This may be one of the main reasons why RS increased starch accumulation and grain weight in IK.

4.4.DEPs and carbon metabolism

The change in protein expression of RG in the early grain-filling stage(10 DAA)was greater than that in the middle grain-filling stage(20 DAA).At 20 DAA,the number of up-regulated proteins in carbon fixation in photosynthetic organizations,glycosylate and dicarboxylate metabolism,and glycolysis and gluconeogenesis pathways were respectively 3.2,6.5,and 1.4 times higher than those at 10 DAA.The photosynthesis pathway was upregulated only at 10 DAA.These pathways are closely associated with carbon metabolism in kernels[17,19,21].Similarly,the difference of hormone content(ABA,IAA and Z+Zr),sucrose content and ATP content of IK and RG was greater in the early stage than in the middle stage of grain filling.The activities of SS,AGP,and SSS in RG were higher than those in IK.Thus,the removal of SK changed mainly the metabolism process of IK in the early grain filling stage to increase kernel weight.Previous study[20]have shown that wheat grain is in a rapid-morphogenesis stage in the early stage of grain filling and that this stage plays a key role in the formation of final kernel weight.Thus,in wheat production,the early stage of grain filling is the key period regulating kernel formation.

During grain filling,plants use photoassimilates synthesized during photosynthesis to accumulate starch and other substances to form kernels.A small amount of carbohydrates provides energy support for plant yield formation by glycolysis[39].Mechin et al.[40]found that proteins involved in glycolysis were highly accumulated during grain filling and believed that these proteins were essential for grain filling.In the present study,the expression of proteins involved in the carbon fixation pathway(RBCL,GAPDH,TPI,ALDO,MaD)in IK was up-regulated after removal of SK.In contrast,proteins involved in glycolysis and pyruvate metabolism pathways(ACACA,PPC,PPDK)were down-regulated.This finding suggests that the removal of SK may promote carbon sequestration and reduce carbohydrate consumption.Glyoxylate and dicarboxylate metabolism also play an important role in plant growth and development and resistance to adversity[41].Succinic acid produced by the glyoxalic acid cycle can be used in the synthesis of plant sugars[42].Removal of SK upregulated the expression of proteins(RBCL,GLDC,GLYA)in the glyoxylate and dicarboxylate metabolism pathways,suggesting that the removal of SK promotes the synthesis of sugars in IK.The up-regulated expression of Cb6-f,ATPase and PS during photosynthesis may provide sufficient ATP(Fig.S6)for the grain-filling process and may compensate for the energy lost from the decrease in glycol.Sufficient ATP could increase the sink strength of IK and provide energy for IK to absorb photosynthetic assimilates.

In summary,we hypothesize that the glycolysis/gluconeogenesis pathway,carbon fixation pathway,and pyruvate metabolic pathways,in conjunction with the glyoxalic acid and dicarboxylic acid cycle pathways,increase sugar synthesis and decrease sugar consumption,thus promoting the sugar accumulation of IK after SK removal.The energy provided by the photosynthetic pathway promoted starch synthesis in the starch and sucrose metabolism pathway.

Removal of SK did not change the starch and sucrose metabolism pathway in IK,but the activity of starch synthesis related enzymes increased,perhaps as a consequence of a change in enzymatic reaction conditions instead of changes in enzyme content.Although GLGP,a protein involved in starch synthesis,was also down-regulated,the total accumulation of starch was increased,suggesting that GLGP is not a key enzyme for starch synthesis in wheat kernels.

Trehalose 6-phosphate synthase(TPS)and trehalose 6-phosphatase(TPP)catalyze the conversion of trehalose 6-phosphate(Tre6P)to trehalose[43,44].Trehalose plays an important role in plant development and carbohydrate metabolism[45].TPS promoted starch accumulation in Arabidopsis[44].Overexpression of TPS and TPP promoted sucrose metabolism and starch accumulation in wheat[46].In the present study,because removal of SK affected the expression of TPS and TPP,we hypothesize that the differential accumulation of TPS and TPP promotes grain filling and starch accumulation in inferior kernels.Whether the synthesis of trehalose catalyzed by TPS and TPP plays a regulatory role in wheat grain filling awaits further study.

In summary,removal of SK increased the grain weight of IK by changing the expression of proteins involved mainly in the processes of carbon metabolism and energy metabolism.Expression of proteins involved in the carbon fixation pathway in IK was upregulated after removal of SK.In contrast,proteins involved in glycolysis and pyruvate metabolism pathways were down-regulated,suggesting that removal of SK promotes carbon sequestration and reduces carbohydrate consumption.The up-regulated expression of Cb6-f,ATPase,and PS during photosynthesis may provide sufficient ATP for the grain-filling process,making starch synthesis more efficient.Removal of superior kernels promoted starch synthesis by increasing ABA and IAA contents and starch-synthesis enzyme activities in the early stage of grain filling.Carbohydrate metabolism and energy supply in the early grain-filling stage may be the key processes regulating the development of inferior kernels in wheat.

CRediT authorship contribution statement

Yang Liu and Yuncheng Liao:conceived and designed the experiments.Jian Luo and Zhui Li:performed the experiments.Jian Luo and Fei Mo:analyzed the data.Yang Liu and Jian Luo:wrote the paper.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was supported by the National Natural Science Foundation of China(31871567),the National Key Research and Development Program of China(2017YFD0300202-2),and Tang Young Scholar(2017).

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2020.12.012.