Accumulation of glycolipids in wheat grain and their role in hardness during grain development

Hixi Qin,Dongyun M,b,*,Xin Hung,Jie Zhng,Wn Sun,Gege Hou,Chenyng Wng,b,Tini Guo,**

aNational Engineering Research Center for Wheat,College of Agronomy,Henan Agricultural University,Zhengzhou 450002,Henan,China

bThe National Key Laboratory of Wheat and Maize Crop Science,Henan Agricultural University,Zhengzhou 450002,Henan,China

cCollege of Food Science and Technology,Henan University of Technology,Zhengzhou 450002,Henan,China

ABSTRACT Article history:Received 26 April 2018 Received inrevised form8August2018 Accepted 3 September 2018 Available online 29 November 2018

Keywords:Amyloplast membrane Endosperm microstructure Grain hardness Polar lipid Wheat Grain hardness is an important parameter for wheat quality.To understand the role of glycolipids in the formation of grain hardness,the glycolipid contents in wholegrain wheat flour and the starch granule surfaces of oven-dried and freeze-dried hard and soft wheat grain were analyzed.Changes in endosperm structure and amyloplast membrane integrity during grain development were also examined by electron microscopy.The monogalactosyldigylcerol(MGDG)and digalactosyldigylcerol(DGDG)contents of the starch surface were significantly higher in soft wheat than in hard wheat,regardless of the drying method or developmental stage.Throughout grain development,MGDG content was significantly higher in the starch surface of freeze-dried hard wheat than in the starch surface of oven-dried hard wheat.In contrast,the MGDG content of the starch surface was significantly higher in freeze-dried soft grain at 14 and 35 days after anthesis.No significant difference was observed in puroindoline protein(PIN)accumulation in wholegrain flour from wheat that was dried using the two methods,whereas PIN accumulation on the starch surface of freeze-dried grain was lower than that on the starch surface of oven-dried grain.The gap between the amyloplast membrane and starch granules was larger in hard wheat than in soft wheat,as shown by transmission electron microscopy.For the same wheat cultivar,this gap was larger for oven-dried than for freeze-dried grain.The content of polar lipids in the starch surface was closely related to grain hardness,and the breakdown of the amyloplast membrane may determine the location of polar lipids on the starch surface.

1.Introduction

Grain hardness is important in wheat quality and plays a key role in wheat milling and food production.Hardness is thought to be dependent on the strength of starch-protein interactions[1].Greenwell and Schofield[2]found that friabilin,a 15-kD protein located on the surface of waterwashed starch granules,weakens the starch:protein matrix.N-terminal sequencing showed that friabilin is composed primarily of two major proteins,puroindolines a and b(PINA and PINB,respectively)[3,4].PINA and PINB are encoded by the completely linked puroindoline(PIN)genes Pina-D1 and Pinb-D1,which are associated with the hardness(Ha)locus on chromosome 5D in common wheat[5,6].Variation in Pina and Pinb affect hardness;wild-type Pina-D1a/Pinb-D1a determines a soft endosperm texture,whereas mutations in either Pina or Pinb or both lead to hard endosperm[7-9].Both PINs possess a tryptophane(Trp)-rich domain that has an important role in lipid binding[10].The interaction between PINs and polar lipids,including hydrophobic,electrostatic and other interactions[10-12],may weaken the starch:protein matrix,affecting kernel hardness[3].The interactions between PINB(containing only three Trp residues)and lipids are thought to be weaker than those of PINA,which contains five Trp residues.Pauly et al.[10]postulated that there are three probable interactions between wheat polar lipids and PINs at the starch granule surface:starchlipid-PINs,starch-PINs-lipid,and interconnected starchlipid-PINs.

Polar lipids also play an important role in determining grain hardness.Greenblatt et al.[13]found a close relationship between the level of bound polar lipids on water-washed wheat starch surfaces and friabilin.Like friabilins,these lipids can serve as biochemical markers for endosperm texture[14].Kim et al.[15]reported that soft-textured genotypes contained more phospholipids and glycolipids than hard-textured genotypes,with the differences increasing through out seed development.Feiz et al.[16]also found that soft textured genotypes had the highest digalactosyldigylcerol(DGDG) content,while hard-textured genotypes had the lowest.However,Finnie et al.[17]reported that endosperm hardness was not significantly related to polar lipid contents in wholegrain wheat flour,whereas hardness was slightly related to the polar lipid contents in flour fractions and significantly related to the polar lipid contents located on the water-washed starch surface.

Grain filling is an important process during grain development.Changes in the hardness index,PIN expression,and polar lipid content of wheat grain during grain filling were investigated previously.The hardness of hard and soft wheat grain can be distinguished at an early stage of grain development.Correspondingly,Turnbull et al.[18]reported that low levels of PINA could be detected in a soft cultivar at 10 days after anthesis(DAA),whereas levels were negligible in a hard cultivar.Chang et al.[19]also found that PINs were significantly higher on starch surfaces of soft wheat than in those of hard wheat at different developmental stages.The polar lipids content of soft wheat grains was higher than that of hard wheat during grain development,but the difference between hard and soft genotypes did not become significant until 28 DAA[15].

According to Pauly et al.,lipids associated with the starch granule surface mainly originate from the amyloplast bilayer lipid membrane[10].During seed desiccation,the amyloplast lipid membrane is degraded[20].Feiz et al.[16]reported that DGDG,monogalactosyldigylcerol (MGDG)and phosphatidylcholine(PC)were the main components of bound polar lipids in wholegrain wheat flour.Finnie et al.[21]also found that these polar lipids were the predominant polar lipids in the starch surface;these lipids resembled those of the amyloplast membranes of potato tubers,which contain high levels of MGDG,DGDG,and PC[22].Kim et al.[15]suggested that PINA and PINB act cooperatively to prevent the breakdown of polar lipids during seed maturation and that this process is central to the development of grain softness.However,no clear evidence was found for the relationship of accumulated polar lipids with grain hardness.In this study,we examined the polar lipid contents of wholegrain wheat flour along with the starch granule surfaces of hard and soft wheat.We also observed the endosperm structure and changes in amyloplasts during grain development to elucidate the mechanism by which polar lipids affect grain hardness.

2.Materials and methods

2.1.Experimental design

Two winter wheat cultivars,hard wheat cv.Zhengmai 366(ZM366)and soft wheat cv.Xuke 316(XK316),were used in this study.The cultivars were grown during the 2014-2015 and 2015-2016 growing seasons at the Henan Agricultural University Experimental Station at Zhengzhou(34°44′N,113°42′E).The soil in this region is loamy fluvo-aquic with an organic matter content of 16.47 mg kg-1(0-30 cm),available phosphorus of 20.83 mg kg-1,available potassium of 212.56 mg kg--1and pH of 7.91.The experiment was a randomized block design with three replicates.Seeds were sown on October 16,2014 and October 15,2015.The plot dimensions were 3 m×7 m,and the sowing density was 160 seeds m-2.Field trials were managed according to local cropping practices.

2.2.Sampling

Spikes undergoing anthesis on the same day were tagged.The grains were harvested at seven-day intervals beginning at 7 DAA and continuing until seed maturation.The sampled seeds were dried using two different drying methods:freezedrying and oven-drying.For freeze-drying,the samples were immediately frozen in liquid nitrogen and then lyophilized(ALPHA1-4LD-PLUS,Germany).For oven-drying,the samples were placed in a forced-convection drying oven at 40°C(DHG-9030A,China).Grain hardness was measured on 300-kernel samples with a Perten SKCS 4100 instrument(Perten Instruments,Springfield,IL,USA)following the manufacturer's instructions.Before hardness testing,all samples were equilibrated to approximately 11%-13%moisture by storing the samples under the same conditions for three days.Grain moisture content was determined using a near-infrared transmittance analyzer(Foss 1241,Foss Tecator AB,H?gan?s,Sweden).Samples harvested at 14 and 21 DAA were divided into three parts:fresh,partially freeze-dried,and partially oven-dried.To obtain the partially freeze-dried samples,the fresh samples were dried to 2/3 of their original moisture content via lyophilization.For the partially oven-dried samples,the fresh samples were dried to 2/3 of their original moisture content using a forced-convection drying oven.

2.3.Grain milling and starch extraction

Wholegrain wheat flour samples were ground using aCyclotec Sample Mill(Foss Tecator AB,H?gan?s,Sweden).Starch extraction was carried out according to Finnie et al.[21].

2.4.Polar lipid extraction,separation and quantification

The method of polar lipid extraction was adapted from Zhao et al.[23]and Bao et al.[24]with minor modifications.Wholegrain wheat flour(2.0 g),or primary starch samples(6.0 g),were treated with water-saturated n-butanol(1:6 sample-to-solvent)for 48 h.The mixture was then centrifuged at 4000×g for 15 min.The supernatant was concentrated to approximately 5 mL using a rotary evaporator.The concentrate was treated three times with 10 mL chloroform:methanol:water(2:1:0.75,v:v).After each treatment,the bottom yellow layer was collected and dried by N2and then dissolved in 1.6 mL methanol as crude lipid extract.

The crude lipid extracts were separated on thin-layer silica gel G chromatography plates(200 mm×200 mm)with chloroform:methanol:acetic acid:acetone(5:1:1:1,v:v:v:v)as a developing agent.The polar lipids targeted for analysis were DGDG and MGDG.Qualitative identification of the resulting bands was carried out with the use of MGDG and DGDG standards purchased from Sigma.The target band was collected and dissolved in 15 mL chloroform:methanol(2:1)followed by filtration and collection.This process was repeated twice.The filtrate was dried by N2and dissolved in 700 μL methanol.

The lipid contents in wheat extracts were analyzed using high-performance liquid chromatography(HPLC;a Waters 2695)with a DIONEX AD25 absorbance detector and a Symmetry c18(250 mm×4.6 mm)column.The flow rate of the mobile phase,which consisted of 95%ethanol and 5%acetic acid,was 1.0 mL min-1.The injection volume was 10 μL.MGDG and DGDG in samples were identified and quantified based on comparison of the chromatographic retention times and areas with those of external standards.

2.5.PIN extraction and sodium dodecyl sulfate-polyacrylamide gel electrophoresis

The method used to extract whole grain PINs was adapted from Bettge et al.[25]and Chang et al.[19]with minor modifications.Five kernels from the middle of the spike were fully ground and transferred to 1.5-mL tubes.After adding 700 μL of pre-cooled(4 °C)acetone,the mixture was allowed to rest for 5 min at room temperature.The sediment pellet was washed twice with precooled acetone after centrifugation(12,000×g,15 min).The pellet was then dried at room temperature and treated with a 750-μL solution of 50 mmol L--1NaCl in 50%isopropanol(v/v).The mixture was extracted under water bath oscillation at 50°C for 45 min and centrifuged at 12,000 ×g for 10 min.The supernatant(350 μL)was then moved to a new tube,1.5 mL acetone was added,and the mixture was placed at-20°C for 8-10 h.Subsequently,the mixture was centrifuged at 12,000×g for 15 min,and the precipitate was washed with precooled acetone.The precipitate was then dried at room temperature,resuspended in sample buffer(62.5 mmol L-1Tris-HCl,pH 6.8,10%[v/v]glycerol,2%[w/v]sodium dodecyl sulfate[SDS],with the addition of 5%[v/v]2-mercaptoethanol),and heated at 100°C for3 min.The supernatant was centrifuged (12,000×g,15 min),and aliquots(10 μL/lane)were loaded into SDS-polyacrylamide gels.The samples were initially run at 120 V in the stacking gel(T=12.0%,C=2.6%,with T=total concentration of acrylamide and bisacrylamide, and C=bisacrylamide concentration).When the dye indicator had moved into the separation gel,the samples were run for 1.5 h at 80 V(T=5%,C=2.6%).The gels were stained with 0.1%Coomassie brilliant blue R250.

Extraction of PINs from the starch surface was carried out according to the method of Chang et al.[19]and Finnie et al.[22]with minor modifications.Primary starch(80 mg)was then treated with 500 μL of 50 mmol L-1NaClin 50%isopropanol(v/v).The process used to treat the extract was the same as for PIN extraction from the kernel,as described above.

2.6.Scanning electron microscopy

The internal structure of wheat grain was examined by scanning electron microscopy(SEM).Kernels were fixed by 2.5%glutaraldehyde for 5 h at 0-4°C,and then washed three times in phosphate buffer solution,dehydrated in a graded ethanol series,infiltrated by isoamyl acetate,and subjected to CO2critical-point drying.A small piece of each sample was cut with a knife,and the bottom of the sample was fixed on the sample table with conductive adhesive.The sample surface was coated with gold and viewed using a S-3400 N scanning electron microscope(Hitachi Science Systems,Ltd.,Japan)operating at 3 kV.

2.7.Transmission electron microscopy

The sample preparation method for microstructural observation was adapted from the method of Wang et al.[26].Samples were sliced to 1-3-mm3pieces and fixed by 2.5%glutaraldehyde for ＞3 h at 0-4 °C.The sample was then washed three times with phosphate buffer solution.Samples were then fixed in osmic acid at 4°C for 3 h.After washing three times with phosphate buffer solution,samples were dehydrated in a graded ethanol series,infiltrated by propylene oxide,and embedded in Spurr resin.Serial sections(70 nm)were cut on an ultramicrotome(Leica EM UC6,Germany),stained with uranyl acetate and lead citrate,and examined using a Jeol 1230 transmission electron microscope(JEOL,Japan).

2.8.Statistical analysis

Data were analyzed and evaluated using SPSS 15.0 software.One-way analysis of variance and Duncan's multiple range tests were applied to distinguish differences between the wheat genotypes subjected to different drying methods following sampling on the same day.Duncan's test was also used to test for differences in hardness index between grains of the same wheat cultivar at different times during development.

3.Results

3.1.Grain hardness in developing wheat grains

Hardness values of wheat grains dried using the oven-drying and freeze-drying methods are shown in Fig.1.The hard wheat cultivar ZM366 had the highest hardness index at 21 DAA,regardless of the drying method.Freeze-dried grains of ZM366 had a significantly lower hardness index than ovendried grains.The largest differences in ZM366 hardness between the two drying methods were observed at 14 and 21 DAA(differences of 31.5 and 20.0,respectively),whereas the difference was smallest at maturity(difference of 9).Similarly,oven-dried grains of the soft wheat cultivar XK316 had a higher hardness index than freeze-dried XK316 grains.However,significant differences in hardness between the two drying methods were only observed at 14 and 35 DAA(differences of 12.5 and 15.2,respectively).Additionally,the hardness of the oven-dried grains of soft wheat did not change from 14 to 28 DAA and reached its lowest value at 35 DAA.

3.2.MGDG and DGDG contents in developing wheat grains

Hard and soft wheat grains showed the same trends in glycolipid content during grain development(Fig.2).DGDG and total glycolipid contents of wholegrain wheat flour were least at 35 DAA,with average values of 560.4 and 994.7 μg g-1,respectively.The average contents of MGDG,DGDG,and total glycolipids in oven-dried hard wheat over all time points were 451.58,726.03,and 1177.60 μg g-1,respectively,whereas those of oven-dried soft wheat were479.30,834.73,and1313.53 μg g-1.The corresponding MGDG,DGDG,and total glycolipid contents in freeze-dried hard wheat were 440.24,647.58,and 1087.82 μg g-1,whereas those in freeze-dried soft wheat were 485.61,742.58,and 1241.50 μg g-1.For the same drying method,the glycolipid content in soft wheat was slightly higher than that in hard wheat throughout grain development,although no significant differences were observed.DGDG and total glycolipid contents of freeze-dried soft grain were higher than that of oven-dried soft grain at 28 and 35 DAA.

The glycolipid content of the starch surface decreased during grain development(Fig.3).The contents of MGDG,DGDG,and total glycolipids in soft wheat cultivar XK316 were significantly higher than those in hard wheat cultivar ZM366,regardless of the drying method or developmental stage.For hard wheat(ZM366),the MGDG,DGDG,and total glycolipid contents were significantly higher when the grains were freeze dried than when they were oven-dried throughout grain development,with the exceptions of DGDG and total glycolipid contents at 28 DAA.However,for soft wheat(XK316),the MGDG content was significantly higher when the grains were freezedried than when the grains were oven-dried at 14 and 35 DAA.In contrast,the DGDG and total glycolipid contents increased at 21 DAA when the grains were freeze-dried.

Fig.1-Changes in hardness during wheat grain development for oven-dried and freeze-dried grains.Different lowercase letters above the column for the same date indicate significant differences(P＜0.05).ZM366-OD and ZM366-FD indicate oven dried and freeze-dried Zhengmai 366,respectively.XK316-OD and XK316-FD indicate oven-dried and freeze-dried Xuke 316,respectively.

Fig.2-Changes in polar lipid contents of oven-dried and freeze-dried wholegrain wheat flour during grain development.Different lowercase letters above the column for the same date indicate significant differences(P＜0.05).Panels A,B,and C indicate MGDG,DGDG,and total glycolipid content,respectively.ZM366-OD and ZM366-FD indicate oven-dried and freezedried Zhengmai 366,respectively.XK316-OD and XK316-FD indicated oven-dried and freeze-dried Xuke 316,respectively.

3.3.PIN accumulation in developing wheat grains

Accumulation of PINs in wholegrain wheat flour did not differ significantly between oven-dried and freeze-dried hard wheat grains during grain filling(Fig.4-A),with one exception:the PIN contents of oven-dried grain were slightly higher than that of freeze-dried grain at 14 DAA.Similar results were found for PIN accumulation in soft wheat during grain development.For the same drying method,accumulation of PINs was slightly higher in soft wheat than in hard wheat.

PIN accumulation on the starch surface(Fig.4-B)was obviously lower than in wholegrain wheat flour.In addition,for both hard and soft wheat,the accumulation of PINs on the starch surface was clearly lower in freeze-dried grain than in oven-dried grain.At 21 DAA,almost no PINs were detected in the freeze-dried starch surface for both hard and soft wheat.PIN accumulation at 14 and 35 DAA was slightly higher than at 21 and 28 DAA,regardless of the drying method or genotype.This finding is partly consistent with the corresponding lower grain hardness at 14 and 35 DAA.

3.4.Grain microstructures of oven-and freeze-dried hard and soft wheat

Oven-dried soft wheat grains exhibited loose internal structures,and the protein and starch granules were loosely bound(Fig.5,S-O-21,-28,and-35).In contrast,the starch granules were tightly embedded in the protein matrix in the oven-dried hard wheat grain(Fig.5,H-O-21,-28,and-35)suggesting a close interaction between starch granules and protein matrix in hard wheat.Similar differences between the hard and soft cultivars were observed for the freeze-dried wheat grains(Fig.5,S-F-21,-28,and-35 and H-F-21,-28,and-35).More pits were observed in the soft wheat endosperm than in the hard wheat endosperm(Fig.5,S-O and S-F).In soft wheat samples,starch and protein were loosely bound,and the starch granules tended to fall off during slicing.Looser bonds between the starch granules and protein matrix were observed in the freeze-dried grains compared to oven-dried grains,especially for the hard cultivar(Fig.5,H-F-21,-28,and-35).

Fig.3-Changes in polar lipid contents on starch surfaces of oven-dried and freeze-dried wheat grains during grain development.Different lowercase letters above the column for the same date indicate significant difference(P＜0.05).Panels A,B,and C indicate MGDG,DGDG,and total glycolipid contents,respectively.ZM366-OD and ZM366-FD indicate oven-dried and freeze-dried Zhengmai 366,respectively.XK316-OD and XK316-FD indicated oven-dried and freeze-dried Xuke 316,respectively.

Fig.4-SDS-polyacrylamide gel electrophoresis profiles of wheat grain PINs(A)and starch surface PINs(B)for oven-dried and freeze-dried hard wheat(ZM366)and soft wheat(XK316)during grain development.Lanes 1-5 and 6-10,profiles of PINs for whole grains of oven-dried and freeze-dried soft wheat(XK316),respectively.Lanes 11-15 and 16-20,profiles of PINs for whole grains of oven-dried and freeze-dried hard wheat(ZM366),respectively.Lanes 21-24 and 25-28,profiles of PINs for the starch surfaces of oven-dried and freeze-dried soft wheat(XK316),respectively.Lanes 29-32 and 33-36,profiles of PINs for the starch surfaces of oven-dried and freeze-dried hard wheat(ZM366),respectively.M,markers;DAA,days after anthesis;OD,ovendried;FD,freeze-dried.

Fig.5-SEM images showing the internal structures of wheat endosperms of oven-dried and freeze-dried soft and hard wheat.S,soft wheat Xuke 316;H,hard wheat Zhengmai 366;O,oven-dried;F,freeze-dried;21,28,and 35 stands for 21,28,35 days after anthesis,respectively.Bar.50 μm.

3.5.Amyloplast changes in hard and soft wheat grains

The internal structures of fresh hard and soft wheat grains were observed by transmission electron microscopy(TEM;Fig.6).At 14 DAA,soft wheat grain had good cell and amyloplast membrane integrities along with clear cell wall outlines,and the nucleus,amyloplast,protein bodies and mitochondria were clearly visible(Fig.6,S-14-I,-II,-III,and-IV).A similar result was observed for hard wheat grains at 14 DAA(Fig.6,H-14-I,-II,-III,and-IV).The numbers and volume of amyloplasts per cell were higher in hard wheat grain than in soft wheat grain(Fig.6,H-14-I and S-14-I).Small amyloplasts were observed around the large amyloplast.Fresh grain at 21 DAA had numerous amyloplasts(Fig.6,S-21-I,-II,and III and H-21-I,-II,and-III),and the number of small amyloplasts was increased in comparison to 14 DAA(Fig.6,S-21-I and H-21-I).Some small amyloplasts had intact membranes(Fig.6,S-21-II).The long axes of the large amyloplasts reached 15 μm,and amyloplast membranes were clearly visible(Fig.6,S-21-I and H-21-I).

The internal structures of grains after partial dehydration(partial oven-or freeze-dried)were also observed by TEM(Fig.7).In the partially oven-dried soft wheat grain at 14 DAA,most of the large amyloplast membranes were separated from the starch granules,and obvious gaps appeared between the membrane and starch granules(Fig.7,S-14-O-I,-II,and-III).Separation of membranes from starch granules can be seen in small amyloplasts.No obvious separation of the amyloplast membrane was observed in the partially freeze-dried grain;however,many vesicles protruded from the amyloplast membrane edge(Fig.7,S-14-F-I,-II,and-III).Similar results were obtained for partially dried hard wheat.In the partially oven-dried hard wheat grains,the membranes of large and small amyloplasts were separated from the starch granules,and there were clear gaps between membranes and starch granules(Fig.7,H-14-O-I,-II,and-III).Separation of membranes and starch granules was particularly clear in hard wheat,and gaps were larger than in soft wheat.In addition,the cell walls of hard wheat grains were wrinkled,and amyloplasts were closer together.Similar results were observed in dried grain samples at 21 DAA using different methods(Fig.7,S-21-O,S-21-F,H-21-O,and H-21-F).

4.Discussion

4.1.Polar lipid contents,PIN contents and grain hardness

Fig.6-TEM images of wheat endosperms from fresh samples of soft wheat(S)and hard wheat(H)at 14(14)and 21 DAA(21).S,soft wheat Xuke 316;H,hard wheat Zhengmai 366;CW,cell wall;LA,large amyloplast;SA,small amyloplast;P,protein body;AM,amyloplast membrane;N,nucleolus;DAA,days after anthesis.

Grain hardness is an important characteristic for wheat marketing and processing.In this study,oven-dried wheat grains showed distinctive hard and soft characteristics from the early stages of grain filling.Hard wheat grain remained hard,and soft wheat grain was soft throughout grain development,in agreement with previous reports[27].Turnbull et al.[18]found that the grain hardness index decreased as grain filling proceeded,with the lowest value observed at 32 DAA.They also reported that the greatest difference in hardness between hard and soft cultivars occurred at 25 DAA.Similar results were found in this study;the lowest hardness was observed at 35DAA(grainmaturity),and the greatest difference in hardness between the hard and soft cultivars occurred at 25 DAA.Variation in hardness likely resulted from the heterogeneity in grain shape at the filling stage and the technical limitations of the machinery that was designed for measurements on mature seeds[18].Grain chemical composition,starch particle size and starch granule content affect grain hardness during grain filling[28],and the drying method can also affect it.In the present work both the hard and soft wheat cultivars had lower hardness indices when freeze-dried compared to oven-dried.This partly agrees with a previous study indicating that the hardness index of freeze-dried hard wheat grain sampled during grain filling was lower than that of soft wheat(except for the mature grains)[27].Freeze-dried hard wheat grain had lower hardness than oven-dried grain;the hardness index of hard wheat was only lower than that of soft wheat at 14 DAA.This inconsistency might be partly attributed to subtle differences in the freeze-drying process.Bechtel et al.[29]found that freeze-dried wheat grain maintained an almost natural ultra-structure.In this study,freeze-dried grain had a loose internal structure,and the starch granules were loosely bound to the protein matrix.The grain composition and manner of binding of friabilin to starch,lipids and other molecules could cause the different endosperm structures in freeze-dried hard and soft wheat.

PINs,which are abundant on the surface of water-washed starch granules of soft wheat,weaken the protein and starch matrix and thus reduce grain hardness[2].In this study,the PIN contents of wholegrain wheat flours derived from the hard wheat(ZM366)and soft wheat(XK316)did not differ significantly.However,the PIN contents on the starch surface of oven-dried soft wheat were higher than on the starch surface of oven-dried hard wheat,suggesting that the PIN contents at the starch surface were negatively related to grain hardness[2,13,30].Compared to freeze-dried grain,the PIN contents of the starch surface of oven-dried grain were higher.However,the hardness was greater for oven-dried grain than for freeze-dried grain.This may be due to a change in the PIN structure or polarity during freeze-drying,causing inhibition of binding of PINs to the starch surface.The Trp-rich domains of PINs are important regions for lipid binding.The size,shape and partitioning properties of Trp residues affect their membrane-binding properties[10,31-34].Moreover,the configuration of lipids following drying may affect the lipidbinding properties.Husband et al.[35]suggested that PINs bind only with lipids present as micelles,while Dubreil et al.[36]suggested that PINs only interact with highly aggregated lipid structures[36].

Fig.7-TEM images of endosperms of soft(S)and hard(H)wheat samples after partial dehydration(oven-dried or freeze-dried)at 14 and 21 DAA.S,soft wheat Xuke 316;H,hard wheat Zhengmai 366;O,oven-dried;F,freeze-dried;14,14 days after anthesis;21,21 days after anthesis.AM,amyloplast membrane;SG,starch granule;CW,cell wall.

Apart from PINs,lipids are also thought to have a key role in grain hardness in wheat.The predominant polar lipids of the starch surface are DGDG,MGDG and PC.MGDG and DGDG represent 60.4%-78.8%of total bound polar lipids[17,21].Greenblatt et al.[13]found that the content of polar lipids on the surfaces of water-washed starch granules was similar to the content of PINs,and suggested that polar lipids may be involved in the formation of hardness.Kim et al.[15]found that polar lipid contents of wholegrain wheat flour from soft wheat were higher than in wholegrain wheat flour from hard wheat sampled during the grain filling stages;however,the difference became significant only at 28 DAA.In this study,the lipid contents in wholegrain wheat flour from soft wheat were higher than that from hard wheat,but the differences were not significant during the grain filling stages.Soft wheat starch had significantly higher polar lipid contents than hard wheat starch using the same drying method.Similar results were reported by Finnie et al.[17],who found that the polar lipid contents in wholegrain wheat flour had no significant effect on endosperm hardness,whereas polar lipids located on the surface of wheat starch significantly influenced hardness.

Grain hardness is affected by the drying method.In this study,the water-washed starch granules of freeze-dried hard wheat cultivar ZM366 showed significantly higher MGDG,DGDG,and total glycolipid contents than corresponding oven dried samples during the grain filling stage,with the exceptions of DGDG and total glycolipid contents at 28 DAA.The high polar lipid contents of the freeze-dried starch granule surface resulted in decreased grain hardness.However,for soft wheat cultivar XK316,significant differences were observed only for MGDG content at 14 and 35 DAA.The differences may be due to different lipid types,lipid content or polarity and configuration affecting grain hardness.MGDG may play a greater role in grain hardness than DGDG.Of course,other polar lipids such as PC may also affect grain hardness.

4.2.Amyloplast membrane and grain hardness

Many studies have shown that grain lipids,which are mainly derived from the residues of the amyloplast membrane and other membranes, play a role in grain hardness[13,15,16,37,38].Furthermore,the residues of lipid membranes adhere to the starch surface during seed desiccation,affecting grain hardness[20].Feiz et al.[16]showed that the bound polar lipid composition of wheat was similar to that of amyloplast membranes from potato tubers,which contain higher levels of MGDG,DGDG,and PC than other lipid membranes[22].The dehydration process may play an important role in the effect of lipids on hardness.In this study,fresh samples of hard and soft wheat had both large and small amyloplasts,and the amyloplast membranes were clearly visible.However,the number of amyloplasts per cell and the amyloplast volume were slightly larger in soft wheat than in hard wheat.When the grains were partly dehydrated,the amyloplast membrane was more separated from the starch granules in the hard wheat cultivar than in the soft wheat cultivar.As mentioned above,hard wheat starch granules have fewer membrane lipids than soft wheat starch granules.The greater separation of the membrane from the starch granules during drying may decrease the likelihood of membrane lipids adhering to the starch surface.This may lead to a decrease in bound lipids and an increase in grain hardness.

Grain drying is known to affect hardness.Air-drying and oven-drying had similar effects on grain hardness,but freezedried samples had a very low hardness index[27].In this study,large and small amyloplast membranes of partially oven-dried samples were separated from the starch granules.In contrast,no obvious separation was observed in partially freeze-dried samples,and only the detectable difference was the more protruding vesicles adhering to the amyloplast membranes.The separation of the amyloplast membrane from starch granules during oven-drying would reduce the content of membrane lipids on the starch granules,thereby increasing grain hardness.

It is not clear how grain dehydration affects hardness.Feiz et al.[16]postulated that active PINs stabilize bound lipids on the surfaces of starch granule membranes,preventing breakdown during seed desiccation and maturation.However,lipids have also been suggested to adhere directly to the surfaces of starch granules without PINs,whereas PINs cannot adhere directly without lipids[39,40].PINs could be detected in the protein matrix and starch granule surface,but not in the amyloplast membrane[10,41].As mentioned above,a high lipid content was detected on the starch surface of freeze-dried grain in this study,but no PINs were detected.We postulate that the breakdown of the amyloplast membranes allows the membrane lipids to adhere to the starch granules during seed drying.PINs can then bind to the starch granules via lipids.Lipids may play a more important role in determining grain hardness than currently thought.Therefore,further investigation is required into the mechanism by which PIN proteins bind to the starch surface during grain dehydration.A better understanding the role of lipids in PIN binding may help elucidate the mechanism of hardness formation.

5.Conclusions

The MGDG and DGDG contents of the starch surface were significantly higher in soft wheat than in hard wheat.Freezedried hard wheat had a lower hardness index and a significantly higher glycolipid content than oven-dried hard wheat.Compared to soft wheat grain,hard wheat grain had large gaps between the amyloplast membrane and starch granules.The amount of polar lipids on the starch surface is closely related to grain hardness,and breakdown of the amyloplast membrane may determine the location of polar lipids on the starch surface.The findings of this study provide new insights into the mechanism of grain hardness development.

Acknowledgments

This study was financially support by the National Natural Science Foundation of China(31571651)and the National Key Laboratory Projecton Wheat and Maize Crop Science(39990035).