Genome-Wide Identification and Expression Profiles of Late Embryogenesis-Abundant (LEA) Genes during Grain Maturation in Wheat (Triticum aestivum L.)

Late embryogenesis-abundant (LEA) genes play important roles in plant growth and development, especially the cellular dehydration tolerance during seed maturation. In order to comprehensively understand the roles of LEA family members in wheat, we carried out a series of analyses based on the latest genome sequence of the bread wheat Chinese Spring. 121 Triticum aestivum L. LEA (TaLEA) genes, classified as 8 groups, were identified and characterized. TaLEA genes are distributed in all chromosomes, most of them with a low number of introns (≤3). Expression profiles showed that most TaLEA genes expressed specifically in grains. By qRT-PCR analysis, we confirmed that 12 genes among them showed high expression levels during late stage grain maturation in two spring wheat cultivars, Yangmai16 and Yangmai15. For most genes, the peak of expression appeared earlier in Yangmai16. Statistical analysis indicated that expression level of 8 genes in Yangmai 16 were significantly higher than Yangmai 15 at 25 days after anthesis. Taken together, our results provide more knowledge for future functional analysis and potential utilization of TaLEA genes in wheat breeding.


Introduction
The dehydration process of embryos during grain maturation is the terminal event of seed development, correlating with the acquisition of desiccation tolerance and with the specific induction of highly abundant hydrophilic proteins described as Late Embryogenesis Abundant (LEA) [1,2]. This process indicates the "switch" of seed transformation from development to germination, acting as an adaptability strategy of angiosperm to ensure the survival of seeds during storage or the environmental stress of long-term evolution [3,4]. Seeds are classified as orthodox and recalcitrant based on their storage behavior and desiccation tolerance in the drying process [5]. Orthodox seeds undergo mature dehydration on the maternal plants. Their water content is usually low and can be further dried to 1 to 5% moisture without damage [4,6].
LEA was first isolated from cotyledons of cotton at the late embryonic development stage [7]. In the following 30 years, LEA proteins have been detected in a variety of higher plant development seeds [8][9][10][11][12][13]. The LEA family is usually composed of multiple members in plants. Previous study identified at least 13 and 34 LEA homologous genes in maize [14] and rice, respectively [15]. With the development of genome-wide sequencing technology, the LEA gene family has been further The nucleotide and protein sequence in the whole-genome of T. aestivum was downloaded from the wheat genome URGI database (https://wheat-urgi.versailles.inra.fr/Seq-Repository). These sequences were first used to construct a local protein database with which to search against known LEA protein sequences collected from A. thaliana (51) [9] and O. sativa (34) [15], through a local protein basic local alignment search (BLASTP) program (https://blast.ncbi.nlm.nih.gov) with an E-value cut-off <10 −10 and an identity of 50% as the threshold. The Hidden Markov Model (HMM) which was constructed using HMMER 3.0 program was used to search all the encoding protein sequences of wheat to find all predicted LEA family members. A self-blast of these sequences was performed by comparing of HMMER and BLAST hits to remove the redundancy, and no alternative splice variants were considered. The aligned sequences were then used as candidate LEA family sequences. The obtained candidate sequences were subjected to domain annotation of the target sequence using the software PfamScan and Pfam A databases, and the sequences containing the complete LEA-1 (Pfam ID PF03760.14), LEA-2 (PF03168.12), LEA-3 (PF03242.12), LEA-4 (PF02987.15), LEA-5 (PF00477. 16), LEA-6 (PF10714.8), Dehydrin (PF00257. 18), and SMP (PF04927.11) domains were identified as the final wheat LEA family sequences. The identified wheat LEA proteins were termed according to the subfamily it belongs and the order of gene ID. The biochemical parameters such as theoretical isoelectric point (PI), molecular weight (MW), amino acid number, instability, and hydrophilicity of the obtained proteins were calculated by the ProtParam (http://web.expasy.org/protparam/). To investigate the evolutionary relationships among these TaLEA genes, multiple sequence alignment was performed using MAFFT (Multiple Alignment using Fast Fourier Transform: http://mafft.cbrc.jp/alignment/software/) with the default parameters. A tree was constructed based on the full-length protein sequences using the Neighbor-Joining (NJ) method with Partial deletion and p-distance model, and a Bootstrap test of 1000 replicates for internal branch reliability. Another two phylogenetic trees for LEA protein family of T. aestivum, O. sativa, and A. thaliana were generated using Maximum-Likehood [ML, by FastTree (an open-source software developed by Morgan N. Price in Adam Arkin's group at Lawrence Berkeley National Lab, Berkeley, USA) with default parameters] and Maximum-Parsimony (MP, by MEGA7.0 software [43] with the Max No. of Trees to Retain as 100) method to better understand the evolution of LEA genes.

Structural Characterisation of Wheat LEA
The exon-intron structures of the TaLEA family genes were determined based on alignments of their coding sequences with the corresponding genomic sequences, and a diagram was obtained using GSDS (Gene structure display server: http://gsds.cbi.pku.edu.cn/). To better understand the similarity and diversity of protein motifs, MEME (Multiple Expectation Maximization for Motif Elicitation: http://meme.nbcr.net/meme) was used to identify the conserved motif structures encoded by each group of TaLEA family genes. The parameters as follow: optimum motif widths of 6-50 residues and a maximum of 15 motifs. The schematic diagram of the amino acid motifs for each TaLEA gene was drawn accordingly.

Distribution of LEA Genes on Wheat Chromosomes
TaLEA were mapped on wheat chromosomes according to the positional information of the TaLEA genes from the wheat genome annotation datebase, and the chromosome physical location map was displayed using MG2C (Map Gene 2 Chromosome V2: http://mg2c.iask.in/mg2c_v2.0/).

Expression Profile Analysis of Wheat LEA Genes
The publicly available wheat RNA-Seq datasets used for generating gene expression levels were downloaded from the Wheat Expression Browser (http://www.wheat-expression.com), then used to analyze the expression profiles of the identified TaLEA genes. These data are from the studies of developmental time-course of common hexaploid wheat variety Azhurnaya and Chinese Spring. The results were summarized according to high expression level tissues as grains, roots, leaves/shoots, and spikes. The original TPM (transcripts per kilobase million) of each gene plus 0.0001 to prevent an expression value of '0'. The heat map was drawn in Log 10 (tpm + 0.0001)-transformed expression values according to the standardization. Genes highly expressed in the grain were selected for qRT-PCR detection.

Plant Materials, Growth Conditions, and Sampling
Based on many years production experience and research, Yangmai16 had a faster grain filling speed, earlier maturity compared with Yangmai15 [35]. Therefore, these two spring wheat varieties Yangmai16 and Yangmai15, which were bred by the author's unit, were planted and cultivated under the same condition. At the flowering stage, two individual florets at the base of the center ears were marked with a marker pen and the plants were tagged with anthesis dates. Wheat grains were sampled at 10,15,20,25,30, and 35 DAA (days after anthesis) for extraction of total RNA and further determination.

Quantitative Real-Time Polymerase Chain Reaction Analysis of TaLEA Proteins
In order to identify the expression profile of TaLEA gene in different wheat cultivars, and to explain the relationship between the LEA protein and maturity, as well as the grain production, high expression level LEA genes were collected for quantitative real-time polymerase chain reaction (qRT-PCR) analysis. Total RNA was extracted from the seeds by Trizol reagent (Biotech, Shanghai), and the first strand of cDNA was synthesized using Transcriptor First Strand cDNA Synthesis Kit (Roche, Penzberg, Germany). The wheat gene Actin was used as an internal reference, and the selected 12 TaLEA genes, including TaDehydrin-1, TaDehydrin-3, TaDehydrin-5, TaDehydrin-7, TaLEA1-1, TaLEA1-3, TaLEA4-13, TaLEA5-1, TaLEA5-3, TaLEA5-4, TaSMP-1 and TaSMP-2, were subjected to qRT-PCR using a fluorescence quantitative reagent Ultra SYBR Mixture (Kang Wei Century, Beijing) on a Roch Light-Cycler 480 fluorescence quantitative PCR machine (Roche, Penzberg, Germany); three independent biological replicates. The primer sequences used are listed in Table S1.

Determination of the Dry Weight and Moisture Content of Grains
Sampled grains were used for measurements of grain dry weight and moisture content. The fresh grains then dried at 70 • C to constant weight were weighed for the fresh weight (W F ) and dry weight (W D ). Moisture content of grains was calculated by ( The processes of grain filling were fitted by Logistic' equation as described by [44]: The dehydration progress was fitted by Gompertz equation as described by [45]:

Genome-Wide Identification and Phylogenetic Analysis of Wheat LEA Genes
Based on genome-wide blast, a total of 121 non-redundant LEA homologous containing the complete LEA, Dehydrin, and SMP domains were identified from the hexaploid wheat genome, which were considered as the putative wheat LEA genes. By phylogenetic analysis, these 121 genes were divided into 8 groups (groups LEA1-LEA6, Dehydrin, and SMP) ( Table 1). The largest group Dehydrin contains 47 members, while the smallest groups LEA3 and LEA6 have only two members. Groups LEA1, LEA2, LEA4, LEA5, and SMP contain 19, 15, 15, 4, and 17 genes, respectively. The physical and chemical parameters of 121 TaLEA proteins were shown according to the analysis using the ProtParam online tool. The wheat LEA protein has a maximum length of 457 amino acid residues (TaDehydrin-8) and the shortest of 99 amino acid residues (TaLEA3-2), with isoelectric points (pI values) ranged from 4.18 (TaSMP-15) to 10.74 (TaDehydrin-28). The molecular weights range from 10.64 kDa (TaLEA3-2) to 43.89 kDa (TaDehydrin-8) with an average of 20.98 kDa. Only 16 proteins have a molecular weight >30 kDa. Analysis of the grand average of hydropathicity (GRAVY) index indicated that all TaLEA proteins were hydrophilic, with all of the index <0. The lipid index reflects the thermal stability of the protein in Table 1.
The phylogenetic analysis (Figure 1) revealed two major clades of the TaLEA family. The 47 genes of the TaDehydrin group, the LEA5 group, and part of the LEA1 group were clustered in one clade, while the other six groups were clustered in another one. According to the evolutionary relationship, TaSMP is more closely related to LEA6 and LEA2; LEA1 and LEA4 are more closely related. To classify the TaLEA genes into different subfamilies with higher confidence and to better understand their evolutionary relationship, another two phylogenetic trees were constructed with Arabidopsis and Oryza ( Figures S1 and S2). It was indicated that the subfamilies classification of LEA genes in wheat is credible.

Structural Characterization of Wheat LEA
All Gene structure analysis revealed that ( To elucidate the similarity and diversity of protein motifs, the conserved motifs of wheat LEA proteins in each group were analyzed by MEME software. 15 conserved motifs were identified, namely motif1~motif15 (Figure 3). The results indicated that with the exception of the TaLEA3, TaLEA5, and TaLEA6, there were conservative motifs specific to each group (Figure 3). The composition of structural motifs was diverse among different LEA groups, but similar within the same group. Moreover, the motifs encoding LEA domains were relatively conserved, indicating that the functions of TaLEA proteins are intergroup specific. All of the 19 genes in the LEA1 group have motif15, and 12 of them also contain motif13. Among the LEA2 group, 9 members contained 2 motifs and 6 members contained 5 motifs, while each member possessed motif3 and motif4. The TaSMP group contained motif7 and motif10. All LEA4 members had motif13, with 10 of them also having motif11, and 3 of them having motif5 and motif15. There were 27 genes in the Dehydrin group possessing motif1, motif2, motif5, motif6, motif8, and motif9, and all 47 members contained motif1, motif2, and motif5.

Structural Characterization of Wheat LEA
All Gene structure analysis revealed that ( Figure 2) wheat LEA genes generally contained few introns, and 24 had no introns. Only two of the 97 intron-containing LEA proteins contained more than one intron. Among the 47 genes of the TaDehydrin group, 46 genes contained only one intron, while the other one (TaDehydrin-21) contained three introns. In the TaLEA1 group, 16 genes had no introns, and the other 3 genes contained only one intron. Among the 15 genes in the TaLEA2 group, 11 were single-intron genes, and the other 4 were intron-free genes. Neither LEA3 nor LEA6 contained introns. There were 14 single-intron genes and 1 double intron gene among the TaLEA4 group. Both of the TaLEA5 and TaSMP groups contained only one intron.
To elucidate the similarity and diversity of protein motifs, the conserved motifs of wheat LEA proteins in each group were analyzed by MEME software. 15 conserved motifs were identified, namely motif1~motif15 ( Figure 3). The results indicated that with the exception of the TaLEA3, TaLEA5, and TaLEA6, there were conservative motifs specific to each group (Figure 3). The composition of structural motifs was diverse among different LEA groups, but similar within the same group. Moreover, the motifs encoding LEA domains were relatively conserved, indicating that the functions of TaLEA proteins are intergroup specific. All of the 19 genes in the LEA1 group have motif15, and 12 of them also contain motif13. Among the LEA2 group, 9 members contained 2 motifs and 6 members contained 5 motifs, while each member possessed motif3 and motif4. The TaSMP group contained motif7 and motif10. All LEA4 members had motif13, with 10 of them also having motif11, and 3 of them having motif5 and motif15. There were 27 genes in the Dehydrin group possessing motif1, motif2, motif5, motif6, motif8, and motif9, and all 47 members contained motif1, motif2, and motif5. Genes 2019, 10, x FOR PEER REVIEW 8 of 18

Chromosomal Locations of Wheat LEA Genes
Using MapGene2Chrom software to analyze genomic position data, 117 genes were distributed on 21 wheat chromosomes, while the location of the other 4 genes remained unclear (Figure 4). The analysis from the 7 homologous chromosome groups showed that there were 29 LEA genes on the chromosome groups five, representing the most abundant regions, followed by group six with 23, while the minimum was on chromosome groups two, with only 6 genes. Chromosome groups one and three had 18 and 17 genes, respectively. Chromosome groups four and seven had 12 genes. In total, 37, 40, and 40 TaLEA genes were located on the A, B, and D sub-genome, respectively, which were almost evenly distributed.

Chromosomal Locations of Wheat LEA Genes
Using MapGene2Chrom software to analyze genomic position data, 117 genes were distributed on 21 wheat chromosomes, while the location of the other 4 genes remained unclear (Figure 4). The analysis from the 7 homologous chromosome groups showed that there were 29 LEA genes on the chromosome groups five, representing the most abundant regions, followed by group six with 23, while the minimum was on chromosome groups two, with only 6 genes. Chromosome groups one and three had 18 and 17 genes, respectively. Chromosome groups four and seven had 12 genes. In total, 37, 40, and 40 TaLEA genes were located on the A, B, and D sub-genome, respectively, which were almost evenly distributed.
According to the chromosomal distribution of the TaLEA gene, there were 43 of the Dehydrin group distributed on homologous chromosome groups three, four, five, and six, except for the four genes were still unclear. Among them, there were 7, 8, 7 on chromosome 6A, 6B, and 6D, and 3, 4, 5 on 5A, 5B, and 5D, respectively, accounting for 72.34% of all Dehydrin members. Such a concentrated distribution deserves further study in the future.

Expression Profile Analysis of Wheat LEA Genes in Different Tissues
To gain insight about the expression profile of these identified TaLEA genes, the publicly available RNA-Seq data were collected from the Wheat Expression Browser database. The expression analysis was performed on grains, roots, leaves/shoots, and spikes. A heat map was created according to the Log10 (tpm+0.0001) of 121 TaLEA genes ( Figure 5). It was indicated that expression patterns of TaLEA genes in different tissues were quite different. 93 genes highly expressed in grains, and among them there were 35 out of 47 members in the Dehydrin subfamily; 17 out of 19 in LEA1; all members of LEA4, LEA5 and LEA6. Whereas, very low expression levels of these 93 genes were observed in the other three tissues, except for 12 in roots and 8 in spikes were also high. 11 highly expressed genes were only observed in spikes, and among them there were 6 members of LEA2 and two members of the LEA3 subfamily. Almost no gene highly expressed in all four tissues and only several genes had higher expression levels in multiple tissues. According to the chromosomal distribution of the TaLEA gene, there were 43 of the Dehydrin group distributed on homologous chromosome groups three, four, five, and six, except for the four genes were still unclear. Among them, there were 7, 8, 7 on chromosome 6A, 6B, and 6D, and 3, 4, 5 on 5A, 5B, and 5D, respectively, accounting for 72.34% of all Dehydrin members. Such a concentrated distribution deserves further study in the future.

Expression Profile Analysis of Wheat LEA Genes in Different Tissues
To gain insight about the expression profile of these identified TaLEA genes, the publicly available RNA-Seq data were collected from the Wheat Expression Browser database. The expression analysis was performed on grains, roots, leaves/shoots, and spikes. A heat map was created according to the Log 10 (tpm + 0.0001) of 121 TaLEA genes ( Figure 5). It was indicated that expression patterns of TaLEA genes in different tissues were quite different. 93 genes highly expressed in grains, and among them there were 35 out of 47 members in the Dehydrin subfamily; 17 out of 19 in LEA1; all members of LEA4, LEA5 and LEA6. Whereas, very low expression levels of these 93 genes were observed in the other three tissues, except for 12 in roots and 8 in spikes were also high. 11 highly expressed genes were only observed in spikes, and among them there were 6 members of LEA2 and two members of the LEA3 subfamily. Almost no gene highly expressed in all four tissues and only several genes had higher expression levels in multiple tissues.

qRT-PCR Analysis of Wheat LEA Genes in the Grain Maturation of Two Different Cultivers
According to the analysis of gene expression in different tissues, 12 TaLEA genes highly expressed in grains were selected to verify the potential role of TaLEA genes in the maturation of seed. Expression levels at different stages post anthesis of Yangmai16 and Yangmai15 were analyzed by qRT-PCR.
Results showed that the expression of these TaLEA genes have temporal specificity. On the whole, the expression level of TaLEA genes increased from 15 DAA to 35 DAA ( Figure 6) in two genotypes. Interestingly, the expression patterns of TaLEA genes in two genotypes presented different characteristics. In Yangmai16, the maximum expression level of 9 genes except TaDehydrin-3, TaLEA5-1 and TaSMP-1 were observed at 30 DAA, whereas there were 8 genes in Yangmai 15 with maximum expression level at 35 DAA. For most genes, the peak of expression appeared earlier in Yangmai16. Statistical analysis indicated that, at 25 DAA, the expression level of 8 genes in Yangmai 16 was significantly higher than Yangmai 15.

qRT-PCR Analysis of Wheat LEA Genes in the Grain Maturation of Two Different Cultivers
According to the analysis of gene expression in different tissues, 12 TaLEA genes highly expressed in grains were selected to verify the potential role of TaLEA genes in the maturation of seed. Expression levels at different stages post anthesis of Yangmai16 and Yangmai15 were analyzed by qRT-PCR.
Results showed that the expression of these TaLEA genes have temporal specificity. On the whole, the expression level of TaLEA genes increased from 15 DAA to 35 DAA ( Figure 6) in two genotypes. Interestingly, the expression patterns of TaLEA genes in two genotypes presented different characteristics. In Yangmai16, the maximum expression level of 9 genes except TaDehydrin-3, TaLEA5-1 and TaSMP-1 were observed at 30 DAA, whereas there were 8 genes in Yangmai 15 with maximum expression level at 35 DAA. For most genes, the peak of expression appeared earlier in Yangmai16. Statistical analysis indicated that, at 25 DAA, the expression level of 8 genes in Yangmai 16 was significantly higher than Yangmai 15. According to the determination of phenotypic character of maturation, the increase of grain dry weight in Yangmai16 was faster than that in Yangmai15 ( Figure 7A). At 30 DAA, the grain filling of yangmai16 grain was almost completed. Statistical analysis indicated that the average grain filling rate was significantly higher in Yangmai 16 (P= 0.0185). The moisture content curve and the average seed dehydration rate indicated that the dehydration of Yangmai 16 was faster (P = 1.69×10 −4 ) and started earlier than Yangmai15, which could be judged by the appearance and color of the seeds ( Figure 7B and C).  According to the determination of phenotypic character of maturation, the increase of grain dry weight in Yangmai16 was faster than that in Yangmai15 ( Figure 7A). At 30 DAA, the grain filling of yangmai16 grain was almost completed. Statistical analysis indicated that the average grain filling rate was significantly higher in Yangmai 16 (P = 0.0185). The moisture content curve and the average seed dehydration rate indicated that the dehydration of Yangmai 16 was faster (P = 1.69 × 10 −4 ) and started earlier than Yangmai 15, which could be judged by the appearance and color of the seeds ( Figure 7B,C).

Identification and Analysis of Wheat LEA Gene Family
In this study, 121 wheat LEA homologous were identified by genomic analysis. Wheat LEA family is bigger than rice (34), corn (32), soybean (36), potato (74), Arabidopsis (51), and canola (108), but less than upland cotton (242). We speculated that in the process of hybridization and chromosome doubling by three species, the number of genes had also doubled and finally formed an allohexaploid species with a huge genome [46][47][48]. It is also possible that most of the wheat LEA genes were retained during subsequent evolution due to genome replication [49]. Functional differentiation of LEA genes during the evolution process could facilitate the regulation of development and adaptation to the environment.
The largest group of the wheat LEA family Dehydrin accounts for about 40% of the total number of family members. However, the largest group was Dehydrin in rice [15] and LEA3 in soybeans [50]. In plum blossom, the largest groups were Dehydrin and LEA2 [10]. LEA2 group in potato had the most members, accounting for 45 of all 74 LEA genes, while the Dehydrin group had only 5 [13]. The difference above indicated that LEA family genes may have evolved independently after the differentiation of these species.
Gene structure analysis revealed that wheat LEA family genes generally lack introns. For instance, there were 24 genes that don't have introns and 95 genes that have only one intron, while only 2 genes contained more than one intron. The phenomenon was also found in other species [13,[50][51]. Perhaps it could be explained by the early intron model of protein-coding gene origin [52]. In general, genes that respond to stress contain fewer introns. Introns can have a deleterious effect on gene expression by delaying transcript production [18]. Moreover, introns can extend the length of the nascent transcript, resulting in an additional expense for transcription [53]. The small number of introns may be a result of adaptation to the environment, which takes less time from transcription to translation, allowing rapid gene expression and functional protein production in response to stresses [53]. DREB [54], AP2/ERF [55], and HD-Zip [42], which play important roles in response to diverse stresses, showing similar phenomenon.

Identification and Analysis of Wheat LEA Gene Family
In this study, 121 wheat LEA homologous were identified by genomic analysis. Wheat LEA family is bigger than rice (34), corn (32), soybean (36), potato (74), Arabidopsis (51), and canola (108), but less than upland cotton (242). We speculated that in the process of hybridization and chromosome doubling by three species, the number of genes had also doubled and finally formed an allohexaploid species with a huge genome [46][47][48]. It is also possible that most of the wheat LEA genes were retained during subsequent evolution due to genome replication [49]. Functional differentiation of LEA genes during the evolution process could facilitate the regulation of development and adaptation to the environment.
The largest group of the wheat LEA family Dehydrin accounts for about 40% of the total number of family members. However, the largest group was Dehydrin in rice [15] and LEA3 in soybeans [50]. In plum blossom, the largest groups were Dehydrin and LEA2 [10]. LEA2 group in potato had the most members, accounting for 45 of all 74 LEA genes, while the Dehydrin group had only 5 [13]. The difference above indicated that LEA family genes may have evolved independently after the differentiation of these species.
Gene structure analysis revealed that wheat LEA family genes generally lack introns. For instance, there were 24 genes that don't have introns and 95 genes that have only one intron, while only 2 genes contained more than one intron. The phenomenon was also found in other species [13,50,51]. Perhaps it could be explained by the early intron model of protein-coding gene origin [52]. In general, genes that respond to stress contain fewer introns. Introns can have a deleterious effect on gene expression by delaying transcript production [18]. Moreover, introns can extend the length of the nascent transcript, resulting in an additional expense for transcription [53]. The small number of introns may be a result of adaptation to the environment, which takes less time from transcription to translation, allowing rapid gene expression and functional protein production in response to stresses [53]. DREB [54], AP2/ERF [55], and HD-Zip [42], which play important roles in response to diverse stresses, showing similar phenomenon.
The expression profiles of TaLEA family members in different tissues suggested that many TaLEA genes have obvious spatial specificity. However, they were still mainly enriched in the spikes and grains. Our study could be helpful for further investigation of LEA genes with spatiotemporal expression specificity and evaluation of their roles in desiccation tolerance.

Wheat Seed Desiccation and the Roles of TaLEA Genes in Acquirement of Desiccation Tolerance
With the prospect of climate changes that are shaping future agricultural practices, the production of highly vigorous seeds have become a key lever to improve yields and a focus of food security and genetic resources [56]. Because of the long growth duration of common wheat varieties, plants must face various biotic and abiotic stresses in fields, especially during the stage of seed maturity [57]. The stage of seed maturity at harvest, which was also associated with the desiccation tolerance, is a major factor influencing seed vigor and grain quality [58]. Nowadays, faster filling and dehydration of wheat seeds achieved more popularity in China because of the important role for optimizing the new cropping systems.
The wheat seeds undergo three development phases: embryo morphogenesis, material accumulation, and mature dehydration. During the last stage, the metabolism is correspondingly weakened, along with the rapid decrease of water content, which then turn on the transition from morphogenesis to germination [59][60][61]. Before and during dehydration, physiological and biochemical changes contribute to the acquisition of desiccation tolerance.
The correlation between LEA protein and desiccation tolerance had been confirmed in many orthodox seeds [62,63]. In this study, the spatial expression pattern of TaLEA indicated that some genes display high expression levels mainly in spikes and grains. Two spring wheat variety Yangmai16 and Yangmai15 were then used for quantitative verification. The results indicated that Yangmai16 had certain advantages in time and quantity of expression at 25 DAA, in which the maximum expression of 8 TaLEA genes appeared earlier than that of Yangmai15. It could be a hint of the inherent genetic differences between two varieties. We speculated that the phenotype of earlier and faster dehydration, which may also possess higher desiccation tolerance theoretically, were correlated with the different expression levels of LEA genes in grains.
The contribution of LEA to maintain seed desiccation tolerance is possibility due to their thermal stability and high hydrophilicity, acting as a dehydration protectant and substitute of water [64]. During the mature dehydration of seeds, the protein of Dehydrins will be present in the cell like a "space filler", maintaining the dissolved character of cell fluid to avoid damage to cell architecture [21]. Magnetic Resonance Imaging (MRI) has been used for tracking and locating the distribution regular of water during seed maturation. It was shown that under the action of Dehydrins, water moves against potential, and the sufficient water in the surroundings of the seed could maintain the available flowing water content in the cells [65,66]. LEA were also supposed to regulate the expression of other genes by direct binding, which still lacks evidence.
Four questions that need to be considered were raised by Song and Fu in the last century [67]: What are the signals that induce the mature dehydration during seed development? Are these signals from the seed itself (embryo, storage tissue, and seed coat) or the maternal tissues? What are the acceptors in seeds that response to the dehydration signal? And last but not least, what are the subtle effects of LEA proteins in cells and sub-cells leading to desiccation tolerance? Answers to these questions would be valuable for a better understanding of mature dehydration mechanisms and germplasm conservation, as well as the genetic improvement of crops.