Development of Caryopsis and Starch Physicochemical Properties in Different Grain Positions of Wheat Spikelets


 Background: Spikelets at different spike positions and the caryopsis at different grain positions grow and develop differently. The caryopsis development and physicochemical properties of starch at different grain positions (the first, second, and third grain positions: G1, G2, and G3) of wheat spikelets were investigated in this study. Results: During the development process, the thickness of both dorsal and abdomen pericarp 8 days after anthesis (DAA) followed the sequence G2 < G1 < G3. However, at 14 DAA the thickness followed the dorsal sequence of G1 < G2 < G3 and the abdomen sequence of G2 < G1 < G3. At 20 and 30 DAA, no difference existed in the pericarp thickness of each grain. The accumulation quantities before 20 DAA varied with starch and protein of endosperm cell in the order G1 > G2 > G3. In mature caryopsis, the caryopsis size and weight indicate that G1 = G2 > G3. The starch content followed the order G1 > G2 > G3, while the essential amino acid, the total amino acid, and the protein content followed the order G2 > G1 > G3. The apparent amylose content followed the sequence G3 > G2 > G1, and A-type starch content followed G3 > G1 > G2. The amorphous ratio followed the order G2 > G1 > G3, whereas the double-helix ratio and the relative crystallinity exhibited the opposite trend. The order of the final degree of hydrolysis through AAG, PPA, and HCL was G2 > G1 > G3. Conclusions: The different material contents were possibly due to the short development time of caryopsis, and the difference in starch physicochemical properties between G2 and the other grain positions might be related to the components and structural characteristics of starch.

more and larger than the central endosperm [22,23]. The above research is mainly focused on the accumulation of substances during the development of wheat caryopsis, while the research on the developmental differences of caryopsis at different grain positions is limited.
In recent years, domestic and foreign scholars have conducted research on different grain positions. Within the central spikelets, the weight at G2 is higher than that at G1, G3, and G4 when moving from the most proximal position to the most distal spikelet position [24,25]. Dai found that the volume and surface area percentage of C-type granules and the protein content in basal grains are higher than those in distal grains, but those of the A, B-type granules in basal grains are lower than those in distal grains [26]. BOZ et al. found signi cant differences between the grain positions within the spike based on hectoliter weights [27]. Yu et al. indicated that grains at upper and basal spikelets can increase their weight by increasing the volume of B-type granules [28]. Within a spike, spikelets are unevenly distributed and largely different due to their unbalanced development [29]. Li et al. found that the effects of spikelet and grain positions on grain weight vary with the grain number of spikelets [30]. Li et al. analyzed the dynamic changes of the weight, protein accumulation, and protein content of superior and inferior grains [31]. The difference between superior and inferior grains lling is due to the ratio of carbon to nitrogen [32,33], sink capacity [34], unbalanced endogenous hormone levels [35,36], as well as the enzyme activities and genes expression involved in the conversion of sucrose to starch [37][38][39], and inconsistencies in assimilation transportation [40]. These studies focus on the protein content, the difference in starch granule size distribution, and the yield of different spike and grain positions, while comparative studies on the physicochemical properties of starch in different grain positions are limited.
Therefore, this study analyzed the accumulation of starch and protein during caryopsis development through microscopic observation techniques, such as slices, and the compositional differences and physicochemical properties of the two in the mature stage to provide a reference for the study of the grouting mechanism of different grain positions in wheat spikes and the regulation mechanism of individual grains.

Experimental Materials
The wheat variety "Yannong 19 (YN19)", purchased from Lixiahe Agricultural Science Institute, Jiangsu Province, was planted in the experimental eld in Yangzhou University from October 2017 to June 2018. The owering wheat spikelets were marked with a black oily marker, and seeds were collected from different locations 8, 14, 20, and 30 DAA, and six grains were xed in 2.5% glutaraldehyde solution for semi-thin sectioning. Starch was extracted from the grain 55 DAA to determine the physicochemical properties.

Production of semi-thin resin slices
The preparation of resin semi-thin sections refered to the method of Xiong et al. [11]. The steps were as follows: The fresh caryopsis was cut to a slice of 1-2 mm with a double-sided blade along the middle part, and then soaked in 2.5% glutaraldehyde solution for 48 h. The phosphate buffer (pH7.2) were used to clear the glutaraldehyde solution and did the dehydration step from 20% to 100% absolute ethanol series concentration. After the previous step, we replaced ethanol with propylene oxide, and nally gradually increased the concentration of the low-viscosity embedding medium (Spurr 1969) in a propylene oxide environment, and polymerize it into an embedded block in a pure resin at 70°C. We used a semi-thin microtome (Leica Ultracut R, Germany) to make 1 μm thick sections, evaporated water of the sections stained with 0.5% methyl violet for 30 seconds on a dryer, and observed and photographed under Leica DMLS microscopy.

Determination of protein and amino acid content
The protein content is analyzed using the CHN mode of the elemental analyzer (Vario EL cube, German element manufacturer, Germany) to analyze the total nitrogen content, and the result is multiplied by the coe cient 6.25 to obtain the protein content.
The amino acid content was determined using the principle of ninhydrin reaction (the ninhydrin reagent kit was purchased from Biochrome Company, UK) and measured by the amino acid analyzer of the Agricultural College of Yangzhou University.

Extraction of starch and total starch and amylose content determination
The separation of starch by water extraction method and total starch content was based on the test method of Wang et al. [5]. The apparent amylose content was determined by He et al. [41].

Morphology observation and granule size distribution analysis
The sample processing of starch morphology observation refers to the research method of Yu et al. [13]. We sucked 10 μL of anhydrous ethanol-dispersed starch mixture into a uted sample stage wrapped in tin foil. After ethanol evaporated in the room temperature, gold plating was performed in the etching coater (BAL-TEC SCD 500 Sputter Coater, Leica, Germany). The sample stage was placed under a scanning electron microscope (S4800, Hitachi, Japan) to observe the morphology of starch granules. The granule size distribution of the starch were determined using a MS-2000 laser diffraction particle size analyzer (Malvern Corporation, England). Approximately 50 mg of starch was weighed into 10-ml (Eppendorf) tubes and suspended with 5 ml of double-distilled water. The equivalent volume, the equivalent surface area, and the proportions of starch granules were automatically assessed using a laser diffraction particle size analyzer (Mastersizer 2000), and the diameter of a sphere with the same volume as the starch grain and the diameter of a circle with the same projected area as the real starch grain were measured.

XRD analysis
We use a glass slide to compact a little dried starch sample on the stage, and place the stage on the X-ray diffractometer (D8 Advance, Bruker, Germany) to scan the spectrum, where the scanning range was 3° -40° and the scanning step length was 0.4 s. The calculation method of relative crystallinity was based on Nara and Komiya [42]. The difference was just that the software version was Photoshop CS6 and Image-Pro-Plus image analysis software.

Fourier transform far infrared spectroscopy
The pretreatment of the sample was 30 mg starch mixed with 100 microliters of distilled water to form a paste. The scanning of spectrum background was done with distilled water rst, and then added the paste with a small spoon to the sample stage of the Fourier Transform Far-Infrared Spectrometer (FTIR; 7000, Varian, USA) to scan the sample spectrum, and the range was 800-4000 cm -1 . In the next step, the spectrum processing referred to the method of Wei et al. [17]. The image production was performed with Origin8.0 and the ratio of (1045/1022) and (1022/995) cm -1 was calculated based on the peak intensity at 1045, 1022, 995 cm -1 and repeated three times.

Starch hydrolysis determination
The 13 C Cross Polarization/Magic Angle Spinning Nuclear magnetic resonance ( 13 C CP/MAS NMR) spectrometer (ADVANCE Starch hydrolysis was determined using the method of Wang et al. [45]. After centrifugation (3000 × g) at 4 °C for 10 min, the soluble sugar content [M (mg)] obtained by hydrolysis in the supernatant is determined by the sulfuric acid-anthrone colorimetric method of Wei et al. [17]: Amount of hydrolyzed starch = M × 0.9.

Statistical analysis
The standard error of data was analyzed using the SPSS 19.0 software, and the differences in measured values among the different starch samples were tested at P < 0.05.

Observation of the microstructure of pericarp at different developmental stages
According to the structural characteristics of wheat caryopsis, we observed two parts, namely, the abdomen region with a symmetrical structure and the opposite dorsal region ( , G1 was nearly lled with amyloplast and PBs, which were also squeezed to ll the gaps between starches. Meanwhile, regular spherical PBs could still be observed at G2, and many gaps existed between the amyloplast and PBs at G3, indicating that the fullness of G3 was signi cantly less than that of G1 and G2. The above results demonstrate that the development of endosperm cells follows an obvious sequence of grain positions, which ultimately determines the length of development time.
3.3 Analysis of material content and characteristics in mature stage

Observation of Grain Morphology and Measurement of Agronomic Characters
In this study, the large spike wheat YN19 was selected to separate the three grains in the spikelet as the fourth and fth grain positions without grains (Figs. 3A-C), and the grain size was measured and analyzed (Figs. 3D-E). The picture (Figs. 3D-E) shows that the grain length and width of G3 were obviously smaller than those of the other two positions, while the length of G1 was slightly larger than that of G2, and the width of G2 was slightly larger than that of G1. Through precise measurement (Tab. 1), no signi cant difference was observed in the grain size and weight of G1 and G2, while those of G3 are signi cantly smaller than those of G1 and G2.

Table1
Size parameters of mature grains at different grain positions

Determination of Amino Acid Content in Grains
In nutrition, amino acids are classi ed as either essential or non-essential amino acids [46]. Essential amino acids, also known as indispensable amino acids, are a group of amino acids that humans and other vertebrates cannot synthesize from metabolic intermediates. We found evident differences in amino acids content, which are mainly manifested as the highest Aspartice acid, Serine, Glutamic acid, Cryteine, Valine, Methionine, Isoleucine, Leucine, and Histidine content at G2, which ultimately leads to G2 having the highest content of essential and non-essential amino acids (Fig. 4A). The essential amino acids and total amino acid in G1 were lower than those in G3 (Fig. 4B). The result is the parameter difference under unit mass, which is converted to the protein content within a single grain followed the sequence G2 > G1 > G3 (Fig. 4C).
3.6 Determination of starch particle size distribution in different grain positions The granule size distribution in wheat starch is an important factor affecting the end-use quality. According to the granule size, the endosperm starch was divided into B-type granules (diameter < 9.9 μm) and A-type granules (diameter > 9.9 μm) [47]. As shown in Figure 5A, the volume-type diameter parameters of the different grain positions of wheat show a weak difference between G1 and G2, but G3 exhibited a signi cant difference between the two ranges, namely, 2.8-9.9 μm and 22.8-42.8 μm.
The results indicate that the number of small starch granules in G3 was less than those in G1 and G2, but the number of large starch granules was greater than those of the two. Similar results can be seen from the results of the scanning electron microscopy of starch (Figs. 5B, C, D). Interestingly, the data indicates that A-type, Area average granule size and the Volume average granule size were signi cantly different among the three positions and followed the sequence G3 > G1 > G2, whereas the B-type starch granule size followed the order G2 > G1 > G3 (Tab. 2).

Analysis of the structure of starch in different grain positions
According to previous studies, the structural characteristics of starch are mainly re ected in three aspects: the content of components, the order of surface structure, and the degree of crystallinity. The components are mainly amylose and amylopectin. In this study, the apparent amylose content of G1 was the lowest, followed by G2, and that of G3 was the highest (Tab. 3). 13 C CP/MAS NMR spectroscopy is widely used for studying the structure of starch samples, kinetics, and correlation. The single-and double-helix contents formed a crystalline structure, and the amorphous region formed an amorphous structure. Figure 6A indicates that the 13 C CP/MAS NMR spectrum of starch has four main resonance peaks (i.e., 103, 82, 73, and 62 ppm) in the range of 50-120 ppm. Software analysis shows that the amorphous starch proportion of G2 was the highest, followed by G1 and G3; the single-helical starch ratio of G3 was the highest, followed by G2 and G1, and the double helix ratio followed the order G3 > G1 > G2 (Tab. 3). The results of Fourier transform infrared spectroscopy show that G1 > G2 = G3 in the ratio of 1045/1022 and G1 < G2 = G3 of the 1022/995 ratio (Fig. 6C, Tab. 3). The crystallinity of starch was analyzed through the X-ray diffraction spectrum. Obvious characteristic peaks were observed at 15°, 17°, 18°, and 23° of the spectrum, which are typically Atype crystal peaks (Fig. 6D). Data processing results show that G3 exhibited the highest relative crystallinity, followed by G1, and that of G2 was the lowest (Tab. 3). The above results indicate signi cant differences in the order degree of the surface and crystal structures of starch at different grain positions. Table 3 Relative proportions of starch single helix, double helix and amorphous structure

Hydrolysis of starch
The hydrolysis process of starch is divided into two stages: the early rapid and late slow hydrolysis stages. The results of the study indicate that the three hydrolysis modes exhibited two stages, namely, the rapid and slow hydrolysis stages (Fig. 7). In the rapid hydrolysis stage, the hydrolysis times of the three granular starches were 0-4, 0-6, and 0-6 days (Fig. 7A), the hydrolysis times using porcine pancreatic alpha-amylase (PPA) were 0-12, 0-24, and 0-8 h (Fig. 7B), and the hydrolysis times through Aspergillus niger amyloglucosidase (AAG) were 0-8, 0-6, and 0-8 h (Fig. 7C). In the same hydrolysis mode, the order of nal hydrolysis of starch was G2 > G1 > G3 (Fig. 7).

Signi cant sequence of caryopsis development existed in different grain positions
Wheat caryopsis can distinguish the endosperm, the seed coats, and the pericarp [7; 8; 9]. Pericarp is composed of three parts: epicarp, mesocarp, and endocarp [10]. This study mainly observed the pericarp thickness of different grain positions at different developmental stages. The change in pericarp thickness in the development of caryopsis was due to two factors: the degradation of the mesocarp caused by the programmed cell death of pericarps, and the enrichment and extrusion of endosperm cells [12]. Eight DAA (Fig. 1), the thickness of G2 was signi cantly smaller than those of the other grain positions, while the thickness of G3 was signi cantly larger than the other grain positions. Until 20 DAA (Fig. 1), the thickness of each grain position showed no obvious difference. Thirty DAA (Fig. 1), the mesocarp basically disappeared, while the cell walls of the exocarp and the endocarp thickened. Pericarp cells eventually become dead cells that coat the surface of the grain [12]. These change disciplines of pericarp thickness verify that the development cycle of pericarp followed the order G2 > G1 > G3. The delay of PCD in pericarp cells may be due to the su cient photosynthetic assimilates and energy supply [12]. The physiological and anatomical functions of the pericarp are similar to those of leaves and resemble those of storage organs [48]. Similar reports pointed out that wheat pericarp has many functions, such as protection, photosynthesis, mineral accumulation, synthesis, and degradation of starch during the development process [11,12]. The development of pericarp is closely related to the process of transport and accumulation of caryopsis, and these differences in pericarp development largely affect the accumulation of starch and protein.
Endosperm is the most important place for the accumulation of storage materials. The endosperm is full of starch in the form of amyloplast and protein in the form of PB, both of which account for more than 80% of the total dry weight [14]. The accumulation of starch grains starts at 4-5 DAA [15]. In this study, 8 DAA (Fig. 2), all three grain positions accumulated amyloplast with differences in number and size, indicating that the developmental starting points of different grain positions varied. Both were also closely related to the owering time of each grain position. The accumulation of protein was relatively late. A study observed that PB was 7 DAA [23], and another one study found PB was 12 DAA [22]. Protein accumulation was observed in G1 and G2 but not in G3. Fourteen DAA (Fig. 2), the starch granule in each grain position increased in size into A-type starch granules. Studies show that A-type starch granules will continue to increase until 19 DAA [16]. Many small starch granules, that is, B-type starch granules were also observed, which is basically consistent with the previous belief that B-type starch granules originated from 10-16 DAA [17,18]. At 20 DAA (Fig. 2), the amyloplast and PBs in the endosperm were further enlarged and increased. The number of small starch granules in G2 was signi cantly less than those in the other grain positions, while the number of PBs was more than those in the other grain positions, indicating that the B-type starch grains at G2 were enriched later than the other grains. Thirty DAA (Fig. 2), G1 and G2 were basically enriched, while many gaps in G3 were un lled. The protein at G1 was squeezed in the gaps between starches, which is consistent with the observation of Zhou et al. during the maturation period [22]. The above results indicate that the delayed formation of B-type starch grains and the increased accumulation of PBs in the middle stage of development (approximately 20 DAA) of the grains at G2 and the poor lling degree of G3 ultimately lead to the starch content order G1 > G2 > G3 and the protein content order G2 > G1 > G3 in the mature grains. The difference in the development and content of starch and protein was largely due to the nal grain length, width, thickness, and grain weight expressed as G1 = G2 > G3.
In nutrition, amino acids classify as either essential or non-essential amino acids. These classi cations were a result of early studies on human nutrition [46]. Essential amino acids refer to amino acids that the human body cannot synthesize by itself or whose synthesis speed cannot meet the needs of the human body and must be ingested from food [49]. In this study, the content of most of the amino acids at G2 was greater than those of the other grain positions. Meanwhile, the difference between G1 and G3 was mainly manifested in the content of essential amino acids, and the content of G3 was higher than that of G1. However, the protein content in a single grain followed the order G2 > G1 > G3 because the grain weight of G3 was signi cantly smaller than those of the other positions (Fig. 3). The above comparison of nutrient content can also provide reference for eating and processing wheat.
The structure and physicochemical properties of starch in G2 were quite different from those in the other grain positions Starch is the substance with the largest proportion of caryopsis dry weight, whose characteristics will directly affect quality and use. Starch is composed of two major glucose polymers, namely, amylose and amylopectin [50]. This study found that the amylose content increased with the grain size. The seed endosperms exhibited a unique bimodal starch granule size distribution: B-type granules (diameter < 9.9 μm) and A-type granules (diameter > 9.9 μm) in mature wheat grains [51,52]. Given the separation and quanti cation limitations, C-type starch granules (diameter < 5 μm) are usually classi ed as B-type starch granules [53][54][55]. The analysis of starch granule size indicates that the proportion of A-type starch in each grain position, the area average particle size [3,2], and the volume average particle size [4,3] all followed the position sequence G3 > G1 > G2, while the ratio of B-type starch granules exhibited an opposite trend.  [59]. The results show that the relative crystallinity and the double helix ratio follow the sequence G3 > G1 > G2, while the amorphous ratio shows the opposite trend.
The 1045 cm −1 /1022 cm −1 of G1 showed a larger position than the other grain positions, while 1022 cm −1 /995 cm −1 showed a smaller G1 position than the others. These results indicate that G3 had higher crystallinity and order of surface structure, while G2 had the lowest.
Starch hydrolysis is complex because it is closely related to the physical and chemical properties of starch grains [60]. The susceptibility of starch to acid and enzyme hydrolysis was in uenced by factors, such as amylase content, crystalline structure, granule size, and relative surface area of granules [60, 61]. G2 had the largest ratio of B-type starch with a high surface area to volume ratio [62], which exhibited a large contact area with enzymes and acids and accelerated starch hydrolysis. The highest amorphous starch ratio and lowest relative crystallinity of G2 made it less resistant to enzymatic and acid hydrolysis, resulting in the following degree of hydrolysis: G2 > G1 > G3.

Conclusion
The differences of the grain positions in wheat spikelets were mainly manifested in two aspects: structure and properties. The pericarp development on the dorsal and abdomen at G2 was early and long, and G3 developed later within a short duration. The order of endosperm cell enrichment and the total starch content in the mature stage followed G1 > G2 > G3, whereas the amino acid and protein contents were G2 > G3 > G1. Compared with other grain positions, G2 has a high B-type starch granule content and an amorphous structure, resulting in the highest degree of hydrolysis at G2 and the lowest at G3. Availability of data and material All data generated or analyzed during this study are included in this manuscript.

Competing intends
The authors declare that they have no competing interests.