Novel mutant alleles of the starch synthesis gene TaSSIVb-D result in the reduction of starch granule number per chloroplast in wheat

Transient starch provides carbon and energy for plant growth, and its synthesis is regulated by the joint action of a series of enzymes. Starch synthesis IV (SSIV) is one of the important starch synthase isoforms, but its impact on wheat starch synthesis has not yet been reported due to the lack of mutant lines. Using the TILLING approach, we identified 54 mutations in the wheat gene TaSSIVb-D, with a mutation density of 1/165 Kb. Among these, three missense mutations and one nonsense mutation were predicted to have severe impacts on protein function. In the mutants, TaSSIVb-D was significantly down-regulated without compensatory increases in the homoeologous genes TaSSIVb-A and TaSSIVb-B. Altered expression of TaSSIVb-D affected granule number per chloroplast; compared with wild type, the number of chloroplasts containing 0–2 granules was significantly increased, while the number containing 3–4 granules was decreased. Photosynthesis was affected accordingly; the maximum quantum yield and yield of PSII were significantly reduced in the nonsense mutant at the heading stage. These results indicate that TaSSIVb-D plays an important role in the formation of transient starch granules in wheat, which in turn impact the efficiency of photosynthesis. The mutagenized population created in this study allows the efficient identification of novel alleles of target genes and could be used as a resource for wheat functional genomics.


Background
Starch granules in plants are classified into two types according to their physiological role. Some starch granules are stored for long periods in reserve tissues as storage starch. This type of starch is an important source of carbohydrates for human beings. Other types of starch are stored temporarily as transient starch to provide carbon and energy for plant growth. Transient starch is synthesized in chloroplasts during the day and consumed during the night. Storage starch granules are synthesized in the plastid stroma by joint action of ADP-glucose pyrophosphorylases (AGPase), starch synthases (SS), starch branching enzymes (SBE), starch debranching enzymes (DBE) and their isoforms; however, the synthesis and regulation of transient granules hold a specific feature [1]. Some synthase isoforms that synthesize storage starch are different from those that synthesize transient starch.
AGPase is the first and rate-limiting enzyme in the biosynthesis of starch, it catalyzes the synthesis of ADPglucose, which is the substrate of amylose starch, and over expression of the AGPase gene enhances the rate of starch biosynthesis [2][3][4]. Amylose, one of the main forms of starch, is synthesized from the substrate, ADPglucose, by granule-bound starch synthase (GBSS). The isoform GBSS2 is responsible for amylose synthesis in leaves and other non-storage tissues [5][6][7][8], whereas amylopectin, the other major form of starch, is synthesized by the coordinated actions of AGPase, SS, and SBE. Five soluble SS isozymes, SSI, SSII, SSIII, SSIV, and SSV are found in plant genomes [9,10]; and among them, SSI, SSII, and SSIII are involved in amylopectin elongation [11]. Isoform SSI is highly expressed in plant leaves [12,13]. Two or three isoforms of SSII and SSIII are found in the genomes of different species, and at least one isoform is expressed in leaves and regulates starch granule biosynthesis [14][15][16][17]. Only one SSIV isoform is found in the Arabidopsis genome; AtSSIV is involved in the initiation of starch granule formation in leaves [18,19]. Overexpression of AtSSIV increases the levels of starch accumulation by 30-40% and results in a higher rate of growth [20]; SSIV mutants have remarkably decreased numbers of granules and abnormal granules [18,19,21].
The functions of SSIV isozymes in the leaves of cereal crops are not as clear as in Arabidopsis. Two SSIV isoforms, SSIVa and SSIVb, are found in rice. SSIVa is mainly expressed in the endosperm and is responsible for starch accumulation in seeds, whereas SSIVb is mainly expressed in leaves at early development stages and is responsible for leaf granule biosynthesis [12,14,22]. It also has overlapping and crucial roles with SSIIIa in rice seeds in determining granule morphology and in maintaining the amyloplast envelope structure [23]. ZmSSIV is highly expressed in the embryo, endosperm and pericarp in maize. In wheat, only one isoform of SSIV has been identified until now. This gene is located on the long chromosome arm of homologous group I and shows high similarity with OsSSIVb, so this isoform is named TaSSIVb [12,24]. TaSSIVb is preferentially expressed in leaves and is not regulated by the circadian clock [24]. TaSSIVb is also highly expressed at the middle stages of seed development [25,26]; however, the impact of TaSSIVb on wheat leaf and seed granule characteristics at different growth stages have not yet been reported due to the lack of mutant lines.
TILLING (Targeting-Induced Local Lesions IN Genomes) is a powerful approach for novel allele identification and has been used to screen mutagenized plant populations for novel alleles of target genes [27][28][29][30][31][32][33][34]. Hundreds of missense and nonsense mutant alleles of key starch synthesis-related wheat genes, such as TaSSII, TaSBEIIa, TaSBEIIb, have been identified using this platform [35,36]. A hexaploid wheat line containing two waxy homozygous mutations created through TILLING and a pre-existing deletion of a third waxy homoeolog displays a near-null phenotype [37]. Novel alleles of TaSBEII have been identified in both durum wheat and bread wheat. Double mutant lines combining a TaS-BEIIa-A mutation with a TaSBEIIa-B mutation in durum wheat and triple mutant lines combining mutations in TaSBEIIa-A, TaSBEIIa-B with TaSBEIIa-D have a significant increase in both amylose and resistant starch content [38][39][40][41][42]. Using this approach to identify novel alleles of TaSSIVb would be helpful in characterizing its role in transient starch granule synthesis in wheat leaves.
In this paper, we used the TILLING approach to identify novel alleles of TaSSIVb-D with mutations in functional regions in an Ethyl methanesulphonate (EMS) mutagenized population derived from an elite hexaploid wheat cultivar. Homozygous mutants were used to investigate the specific role of TaSSIVb-D in starch granule synthesis in leaves at different stages of growth. We found that TaSSIVb-D mutations decrease gene expression and the number of starch granules in leaves.

Novel alleles of TaSSIVb-D in a mutagenized population
Three primer sets corresponding to TaSSIVb-D were designed for allele mining (Table 1), and their specificities were validated using nullisomic-tetrasomic lines (Fig. 1a). The location of these primers on the TaSSIVb-D sequence is shown in Fig. 1b.
In the mutagenised population, 54 mutations were identified with a mutation density of 1/165 Kb. Among these mutations, 52 are G to A (61.12%) or C to T (35.19%) transition point mutations, and the other 2 are T insertions. Of these mutations, 26 are in the coding region (Table 2), including 1 nonsense mutation (E054-13), 15 missense mutations and 10 silent mutations. The remaining 26 are located in introns and include 2 splice junction mutations.
Using SIFT program, five missense mutations (E2-2-226 and E2-2-437, E2-2-193, E570, E972, E1137) are predicted to have a severe impact on protein function. Based on the PSSM program, three missense mutations (E2-2-226 and E2-2-437, E972, E1137) are predicted to have a severe impact. As the nonsense mutant E054-13 results in the loss of both the starch catalytic domain and glycosyltranferase domain, it might have a severe impact on protein function. Because the E1137 mutation is located in the glycosyltranferase domain which has less similarity with other SS isoforms [24], and might have distinct functions from other starch synthases, The expression of TaSSIVb in mutant leaves at different growth stages Specific primer sets for each sub-genome (Additional file 1: Table S1, Additional file 2: Figure S1) were used to analyze the pattern of TaSSIVb gene expression. In wild type, gene expression profiling revealed non-significant differences in expression between different growth stages (Fig. 2). In the mutants, TaSSIVb-D was down-regulated, and TaSSIVb-A and TaSSIVb-B did not show significant compensatory responses. The expression of TaSSIVb-D in the nonsense mutant E054-13 was reduced by~8-fold compared with the wild type at all three growth stages, whereas in E1137, its expression level gradually increased with plant growth and only showed a significant reduction (36.12%) at the seedling stage (Fig. 2a). In contrast, no significant differences in the expression of TaSSIVb-A (Fig. 2b) and TaS-SIVb-B (Fig. 2c), were detected in the two TaSSIVb-D mutants at all three stages, except for TaSSIVb-A expression in the E054-13 mutant at the seedling stage (9.55% reduction).

Variation in the number of starch granules in the chloroplast
We investigated starch granule characteristics in the chloroplasts of wild type and the two mutants at three growth stages (Fig. 3). Granule number varied plast by plast in each genotype, with numbers ranging from 0 to 8 and more than 90% of chloroplasts contained no more than 4 granules (Additional file 1: Table S2). The difference between the wild type and mutants was mainly in the number of chloroplasts containing no more than 4 granules.    The biggest difference in granule number per chloroplast was observed at the seedling stage, and mutation of TaSSIVb-D led to a reduction in granule number. The percentage of chloroplasts with 0-2 granules in both mutants, E054-13 (96.70%) and E1137 (84.70%), were significantly higher compared with wild type (70.30%) (Fig. 4), whereas at the elongation and heading stages, the difference between E1137 and wild type became smaller; only the E054-13 mutant had a significantly higher percentage of chloroplasts with 0-2 and 3-4 granules (17.74% and 18.72% higher than wild type, respectively). In contrast, the percentage of chloroplasts containing 3-4 granules was lower in the mutants than in wild type. At the seedling stage, the percentage of chlorplasts in E054-13 and E1137 with 3-4 granules was 22.00% and 10.3% lower than wild type, respectively. At the elongation and heading stages, only the E054-13 mutant was significantly different from wild type; the percentage of chlorplasts with 3-4 granules was 14.26% and 15.05% lower than wild type, respectively. The observation of the chloroplast granule number phenotype at all stages for the E054-13 mutant and only at the seedling stage for the E1137 mutant is consistent with the lower levels of TaSSIVb-D gene expression in the mutants at these stages (Fig. 2).

Photosynthesis parameters and the efficiency of PSII
As transient starch in chloroplasts provides carbon and energy for plant growth, we deduced that photosynthesis might be affected by reduced granule numbers. Granule numbers in the mutant E054-13 were more severely reduced than those in E1137, and the impact on photosynthesis parameters are expected to be more severe. Therefore, photosynthesis parameters in the mutant E054-13 and the wild type at the heading stage were compared to determine the impact of reduced granule number. We found that the maximum quantum yield (F v /F m ) and yield of PSII (Y(II)) of E054-13 were significantly lower than in wild type (Table 3), and differences between the mutant and wild type in both Y(II) and electron transport rate (ETR) increased with increasing photosynthetically active radiation (PAR) (Fig. 5). As Y(II) provides the effective quantum yield and could be used to estimate the effective portion of absorbed quanta; while ETR provides information for plant stress reaction, so the results indicates that mutation of TaS-SIVb-D leads to a decrease in relative quantum yield and has a negative effect on PSII efficiency.

Conserved and functional regions of TaSSIVb
As one of the isoforms of the soluble starch synthesis family, wheat SSIV contains the starch catalytic domain (GT-5) and glycosyltranferase domain (GT-1). GT-5 (amino acids 422-661) is more conserved than GT-1 (amino acids 706-886) between SS isoforms [24]. Two mutations that we identified in this study, E972 and E1137, are located in GT-5 and GT-1, respectively, and in the wild type these amino acids are involved in the formation of β-sheets (data not shown). In contrast the N-terminus (amino acids 1-405) is distinct from other SS isoforms, and this region contains two coiled-coil domains and a 14-3-3-protein binding site [24]. The missense mutation E2-2-226 and the truncation mutation E054-13 are located in this region. These novel alleles could be very useful in understanding the specific functions of SSIV in starch granule synthesis in wheat. The gene expression patterns and alteration of granule number in the E1137 and E054-13 mutants demonstrate the important roles of these predicted functional regions. The E054-13 truncation mutation has a more severe effect than the missense mutation; TaSSIVb gene expression is reduced by 5~13-fold compared with E1137 (Fig. 2), and granule number is also significantly decreased compared with E1137 (Additional file 1: Table S2).

Expression of TaSSIVb homoeologs
Wheat has three sub-genomes, and the literature shows that the contribution of each homeologous gene can be different. For example, of the three SSIIa genes, SSIIa on the B genome has the largest contribution to amylopectin structure than those on the other two genomes [43]. Each waxy gene encoding GBSSI has different effects on amylopecitin characteristics [37,44,45]. SS isoform SSIV is encoded by three homeologous genes located on group I chromosomes [24] and the effect of each gene and their interactions is not clear. Here we designed homeologous-specific primers to determine the gene expression patterns of the three SSIV genes. In TaSSIVb-D mutants, decreased expression of TaSSIVb-D did not result in significant changes in the expression of TaSSIVb-A or TaSSIVb-B. This is similar to what was observed in the single null mutant of SBEIIa; a mutation in one subgenome does not result in appreciably different expression in the other two sub-genomes [46]. However, the identification of mutations in TaSSIVb-A and TaSSIVb-B is needed to analyze the interaction between homoeologous copies or their dosage effects.

Expression of TaSSIV-D in wheat leaves
The function of AtSSIV, which it is highly expressed in leaves in Arabidopsis is very well understood. In wheat, the expression of TaSSIVb is tissue-dependent, and is much more highly expressed in leaves than in the endosperm [24]. Our results are consistent with these reports; all three homeologous genes are highly expressed in leaves during the entire growth period. There is also evidence that TaSSIVb is expressed during seed development [24][25][26]. In ongoing research we are investigating the accumulation of storage starch in both the E054-13 and E1137 mutants to elucidate the role of TaSSIVb in storage starch synthesis.

Starch granule synthesis
Starch granules are initiated early in the development of young leaves in Arabidopsis. Because more starch granules are observed in immature leaves than in mature leaves [47], it is believed that SSIV functions in granule initiation at earlier growth periods [21]. The ssIV mutant in Arabidopsis has reduced granules in most chloroplasts [18,21]. The results of this study demonstrate that in young wheat leaves TaSSIV-D has a function similar to AtSSIV, and it is involved in initiation of transitory starch synthesis. At the seedling stage, the expression patterns of both the E054-13 and E1137 TaSSIV-D mutants are down regulated. In the E1137 mutant, the significant reduction of granule number is only observed when the expression of TaSSIV-D is reduced; while in the nonsense mutant, starch granules are reduced at all the three stages. This indicates that reduction of starch granule number results from the mutation of TaSSIV-D and not other genes.

Effect of homeologous genes
In many cases, homeologous genes from different subgenomes of wheat have dosage effects on phenotypes. Only double or triple mutants of SSIIa have severe  effects on starch properties [48,49], and single-gene mutants do not show significant phenotypes [38]. However in single TaSSIV-D gene mutants, gene expression level, granule number, relative quantum yield and PSII efficiency are significantly reduced with respect to the wild type. It has been reported that single mutants of the genes SBEIIa and gpc (grain protein content) also show significant phenotypic differences compared with the wild type [46,50], so not only double and triple mutants, but also single mutants can lead to phenotypic changes in hexaploid wheat.

Relationship between transient starch content and photosynthesis
It has been reported that the capacity for starch synthesis and the rate of photosynthesis are positively correlated [51]. A starchless tobacco mutant shows about 50% inhibition of maximum photosynthetic capacity [52]. Mutation of SSIV in Arabidopsis also has negative effects on photosynthesis efficiency [53]. Our SSIV mutant in wheat shows the same effects; photosynthesis efficiency is significantly reduced. This is likely because SSIV mutants accumulate high concentrations of the substrate ADPglucose [21], which leads to photooxidative stress and decreased photosynthesis efficiency [53].

Conclusions
We identified novel alleles, including missense and nonsense mutations, of the starch synthesis gene, TaSSIVb-D. Gene expression, granule number and photosynthetic parameters are down-regulated or reduced in the missense mutant E1137 and the nonsense mutant E054-13, and the impact of E054-13 is much more severe than that of E1137. This demonstrates TaSSIVb-D plays an important role in the formation of transient starch granules in wheat, and mutation of TaSSIVb-D results in remarkable reduction in transitory starch granules in leavies. The mutagenized population generated in this study could be used as a resource for gene functional research.

Materials
Wheat (Triticum aestivum L.) cultivar Jing411 was used as the wild type (WT) to create the mutagenized population. Homozygous mutant lines grown in the field with normal management were used for RT-qPCR and phenotyping.

Development of the wheat TILLING population
Seeds from the wild type Jing411 were soaked in tap water for 10 h at 20°C, then incubated in 1.5% EMS in 0.1 M sodium phosphate buffer (pH7.0) for 4 h or 6 h at 20°C. This was followed by washing with running tap water for 4 h at room temperature. The mutagenized seeds (M 1 generation) were planted directly in the field and self-fertilized. The M 2 was developed using the single-seed descent (SSD) method. The mutagenized population consists of 3058 M 2 plants, including 1441 individuals from the 4 h 1.5% EMS treatment and 1617 from the 6 h treatment. A young leaf from each individual was sampled at the seedling stage for genomic DNA extraction using the DNA-quick Plant System kit (Tiangen Biotech, Beijing, China). DNA concentration was measured with the NanoDrop2000 spectrophotometer (Thermo Fisher Scientific, USA) according to the manufacturer's instructions, and the final concentration was adjusted to 50 ng/ μl. Individual DNA samples were pooled 2-fold into 96well plates for TILLING screening, which was done according to Till et al. [54].

Mutation screening
Each primer used for TILLING detection was synthesized with and without fluorescent dye label. Unlabeled primers (Sangon Biotech, China), forward primers labeled with the fluorescent dye IRD700 and reverse primers labeled with the fluorescent dye IRD800 (Bioneer, China) were mixed with the ratio of unlabeled forward: labeled forward: unlabeled reverse: labeled reverse = 4: 1: 1: 4. PCR amplification and polyacrylamide gel electrophoresis were carried out as previously described [54]. Identified mutations were sequenced to verify nucleotide variation. Mutation density was calculated by dividing the total number of mutations by the total length of the sampled DNA sequence (length of the amplified fragment × number of individuals sampled). Totally 3058 individual plants were screened for each primer set, and the amplicon size of primer set s4b-D1, s4b-D1 and s4b-D1 was 1191, 1361 and 960 bp respectively. As DNA products at the top and bottom of the gel were difficult to detect, we subtracted 100 bp from the 5' terminus and 100 bp from the 3' terminus of each fragment.

RT-qPCR
The effects of mutations on the expression of TaSSIVb gene homoeologs was evaluated for leaves at the seedling, elongation and heading stages. Total RNA was extracted from three individual plant leaves, and cDNA was synthesized using the FastQuant RT kit (With gDNase, Tiangen Biotech) according to the manufacturer's instructions. RNA extraction and cDNA synthesis was done for three biological replicates per genotype with three technical replicates for each biological replicates, and their concentrations measured with the Nano-Drop2000 spectrophotometer, the same volume and concentration of normalized template was used for the next step experiment. RT-qPCR was performed on a CFX96 system using the SsoFast EvaGreen Supermix kit (Bio-Rad), PCR was performed at 94°C for 3 min, followed by 40 cycles at 94°C for 30 s, annealing for 30 s, and 72°C for 10 s, and then a melt curve stage. Concentration of each primer was 0.3 μmol, amplification efficiency of the three genes ranged from 97% to 100%. The actin gene used as a control was the same as described by Gu et al. [57]. The relative expression level was calculated with ΔΔCt method according to the CFX96 manual, data were analyzed by one-way ANOVA using Microsoft Excel software.

Electron microscopy
At the seedling, elongation, and heading stages, leaves of mutants and WT at the same developmental stage were sampled with 3 replicates for transmission electron microscope analysis. Each sample was processed as described in Guo et al. [58] and photographed with a transmission electron microscope (HT-7700, Hitachi, Japan). Starch granule numbers per chloroplast were calculated, and data were analyzed using the SPSS software [59].

Chlorophyll fluorescence parameters
A pulse amplitude modulation fluorometer (MINI-PAM, Heinz Walz, Effeltrich, Germany) was used to measure chlorophyll fluorescence parameters. At the seedling, elongation, and heading stages, individual plants at the same developmental stage were selected for measurements with three biological and three technical replicates. From 10:00-11:00 am on the same day, after dark adaption by Dark Leaf Clip for 30 min, the value of F v / F m and Y(II) were measured with a photosynthetically active radiation (PAR) of 281 μmol m -2 s -1 according to the manufacturer's instructions. The response of photosynthetic fluorescence parameters to light intensity changes was determined using rapid light curves. Imaging Win software (Heinz Walz, Effeltrich, Germany) was used to collect and analyze fluorescence parameter data. Data were analyzed by one-way ANOVA with Microsoft Excel software.

Additional files
Additional file 1: Table S1. Primers used for RT-qPCR. Additional file 2: Figure S1. Validation of sub-genome-specific RT-qPCR primers using Chinese Spring nullisomic-tetrasomic lines. CS: Chinese Spring; N1DT1B, N1AT1B, N1AT1D, and N1BT1D: Chinese Spring nullisomictetrasomic lines; J411: wild type; s4b-qD: TaSSIVb-D-specific primers; s4b-qA: TaSSIVb-A-specific primers; s4b-qB: TaSSIVb Ethics approval and consent to participate Wheat samples have been used in the article, and seeds of the wild type were deposited in Institute of Crop Science, Chinese Academy of Agricultural Sciences. The mutants were developed by the authors. All plants were grown and conducted in accordance with the local legislation. No voucher specimen was deposited in public herbarium.