Analysis of Diversity And Identication of SSR, SCoT, And ITAP Informative Amplicons For Grain Fe And Zn Content In Wheat Genotypes

Biofortication provides a promising method of solving microelement malnutrition in developing countries. For this purpose, a study was conducted to understand the grain Fe and Zn content variation in seventy prevalent Iranian wheat genotypes across three consecutive years, to assess genetic diversity, and to identify informative amplicons for high grain Fe and Zn content using three simple sequence repeat (SSR), start codon targeted (SCoT) polymorphism, and intron targeted amplied polymorphism (ITAP) markers. Grain Fe and Zn content was highly variable each year with high heritability. Despite the highly signicant effect of year-genotype interaction, some stable genotypes were ranked highly all the three years for grain Fe and Zn content. The grain Fe and Zn contents were positively correlated in the second and third years. High genetic diversity was detected among the wheat genotypes using three different marker systems. A number of informative SSR, SCoT, and ITAP amplicons for high grain Fe and Zn were identied overall or in individual years. A few informative amplicons were common and stable for grain Fe and Zn content in the different years. The SSR alleles located on 3A, 4A, 4B, and 6B chromosomes were positively correlated with high Fe and Zn content, indicating that co-location of genes affected Fe and Zn content. Identication of informative alleles and amplicons for high grain Fe and Zn content could contribute to the development of sequence-based markers and improve the selection of genotypes with high micronutrient content. of combined analysis of Fe the Pearson coecient Fe combined among revealed SCoT, and ITAP markers.

Malnutrition of minerals such as Fe and Zn is a signi cant health problem in developing countries that rely on cereals as a vital staple food, leading to illness and even death (Welch and Graham 2005;Srinivasa et al. 2014). Many strategies, including forti cation, supplementation, and use of diverse food sources, have been suggested to compensate for de ciencies. For technical and economic reasons, however, these strategies are not entirely e cient and sustainable, especially for the poor society sector of the world's population (Borrill et Kumar et al. 2018b). Genetic diversity for grain micronutrients is crucial for successful bioforti cation, identi cation, and selection of future donors (Srinivasa et al. 2014). Genetic diversity is fundamental to conventional and molecular plant breeding strategies such as marker-assisted selection, quantitative trait loci (QTL) mapping, and genomic selection (Eltaher et al. 2018). Several DNA markers have been used widely in genetic studies due to simplicity, informativeness, and cost-effectiveness (Xiong et al. 2013;Eltaher et al. 2018; Rufo et al. 2019). However, simple sequence repeat (SSR) and single nucleotide polymorphism (SNP) markers target speci c sequence regions, and function as codominant markers. Therefore, these markers have been used widely for genetic diversity analysis, cultivar identi cation, association, and phylogenic analysis (Yang et al. 2016;Eltaher et al. 2018). Several novel techniques of gene-targeted molecular marker techniques have also been innovated recently, such as start codon targeted (SCoT) and intron targeted ampli ed polymorphism (ITAP), acting as semi-speci c and semi-functional and revealing polymorphism ideally in the plant kingdom (Xiong et al. 2013). The SSR and SCoT markers have been used in several previous studies for investigation of genetic diversity in wheat populations (Naghavi et al. 2009; Senturk Ak rat and Uncuoglu 2013; Nasrollahi et al. 2019; Khodaee et al. 2021). Based on our knowledge, however, there is no report on employment of ITAP markers in genetic studies of wheat. Informative markers (i.e. correlated molecular markers with phenotypic traits) make it possible to select reliable genotypes. Marker-assisted selection (MAS) is a highly e cient, dependable method in plant breeding programs.
The objectives of this research are (i) to investigate variation and genotype-environment interaction in wheat grain Fe and Zn content, (ii) characterize the genetic diversity and population structures of the wheat genotypes prevalent in Iran using SSR, SCoT, and ITAP markers, and (iii) to specify informative SSR, SCoT, and ITAP amplicons for grain Fe and Zn contents.

Materials And Methods
A set of sixty-seven bread and three durum wheat genotypes obtained from Seed and Plant Improvement Institute (SPII) in Karaj, Iran were planted for three years (2017-2019) in a loam-clay soil at the research eld of the University of Kurdistan. A list of the studied wheat genotypes is presented in Table 1 along with pedigree information. Each genotype was grown in a plot consisting of three rows with 0.2 m distances based on randomized complete block design (RCBD) with three replications. Before sowing, samples were collected randomly from the top 0-30 cm of the soil at the experimental site and subsequently analyzed for micronutrients, organic matter, and pH (Supplementary Table S1). The wheat genotypes were assessed for grain Fe and Zn content. The genetic diversity of the wheat genotypes was also analyzed using SCoT, ITAP, and SSR primers. The Chinese Spring cultivar was used as a reference for examination of SSR primers.

Speci cation of grain Fe and Zn content
The grain samples (approximately 0.5 gr) were digested in a mixture of Chloridric acid (HCL) and Perchloridric acid (HCLO4), according to Singh et al. (1999).
The digested samples were analyzed for iron (Fe) and zinc (Zn) content (expressed as mg Kg − 1 dry weight) using ame atomic absorption spectroscopy (SpectrAA220-Varian Ltd., Mulgrave, Australia) in three replications. Appropriate quality control was carried out for each set of measurements.
Analysis of variance (ANOVA) and combined analysis of variance was conducted for comparison of the grain Fe and Zn content of the wheat genotypes each year and over the three years using UNIANOVA syntax SPSS 18.0, IBM. The bivariate correlation between the grain Fe and Zn contents of the wheat genotypes was analyzed through the Pearson coe cient (SPSS 18.0). Broad-sense heritability (H 2 ) was estimated for grain Fe and Zn content on the basis of a combined three-year analysis using the following formula: where σ 2 g , σ 2 gy , σ 2 e , Y, and R represent the variances of the genotypes, genotype-year interaction, error, number of years, and number of replications per trial, respectively.

DNA isolation and PCR ampli cation
The collected leaves (3-4 young seedlings) were resorted to a -40 freezer before DNA extraction. The DNA was isolated from the leaves using the CTAB  Table S2, S3). The SSR markers were selected and synthesized based on information from previous studies on wheat and the GrainGenes database (https://wheat.pw.usda.gov/GG2). The PCR ampli cation reaction was prepared in a nal volume of 10 µL, consisting of 4.7 µL ddH2O, 2 µL genomic DNA, 2.1 µL PCR master kit, and 1.2 µL of either primer (SCoT or ITAP) or 0.6 µL of the SSR forward and reverse primers. The following program was used for PCR ampli cation: 94 for 3 min followed by 35 cycles of 94 for 1 min, annealing temperature (48-68) for 1 min and 72 for 2 min, and nal extension at 72 for 5 min. The ampli ed SCoT and ITAP products were separated using 1.3% agarose, and the SSR alleles were separated on denaturing polyacrylamide (6% acrylamide) gel and resolved through silver staining. The SSR alleles were scored as codominant markers based on band size. The binary codes 1 and 0 were also used for scoring both SCoT and ITAP amplicons.

Grain Fe and Zn Content
We evaluated Fe and Zn microelements in the grains of each wheat genotype on dry weight bases across the three consecutive years.  Fig. 1.
The error variances for every year were homogenate, according to Bartlett's test of equality of variances. Therefore, a combined analysis of variance was made to verify the interaction between grain Fe and Zn content and their environment. The effects of the years and of the interaction between the years and genotypes were highly signi cant for grain Fe and Zn content (p < 0.001, Table 2). Regardless of the signi cant year-genotype interaction, seven genotypes (including Hirmand, Gaspard, Arya, Shahpasand, CrossShahi, Marvdasht, and Kavir) were ranked highly with respect to grain Fe content in all the three years. Similarly, eleven genotypes (including Gaspard, Sabalan, Shahpasand, Karaj1, Karkheh, Hirmand, Arvand, Rasool, Shoeleh, Azar2, and Yavaros) were ranked highly for grain Zn content every year. High broad-sense heritability was estimated for grain Fe (89.29%, 76.51%, and 95.11%) and Zn (74.19%, 88.73%, and 84.48%) content in different years. The σ 2 g/σ 2 gy ratio was calculated after the combined analysis of data, which was less than one for grain Fe and Zn content. The Pearson correlation test was conducted separately for each year between the grain Fe and Zn contents. The correlation of grain Fe and Zn content for the rst year was not signi cant (r = 0.230, p > 0.05). In contrast, signi cant positive correlations were estimated between grain Fe and Zn contents in the second (r = 0.477, p < 0.001) and third (r = 0.274, p < 0.05) years.

SSR Primers
Of the 20 SSR primer pairs assessed in the wheat genotypes, 15 produced unambiguous, polymorphic amplicons. The total number of generated alleles in the 70 wheat genotypes was 67, with the number of alleles per locus averaging 4.67 and ranging from 2 (edm16) to 9 (wmc617). The effective number (Ne) was 2.628 on average (variable from 1.116 to 6.011). Mean expected heterozygosity (He) was 0.539, with the variable ranging from 0.094 (barc48) to 0.833 (wmc617). Shannon's information index (I) ranged from 0.199 (barc48) to 1.931 (wmc617), with an average of 0.979. The values of PIC varied between 0.086 (barc48) and 0.831 (wmc617), with a mean value of 0.534. The greatest, lowest, and average values of haploid diversity (H) were 0.78, 0.07, and 0.536, respectively. The positive mean value of Fis (0.607) indicated the excess of homozygosity. The calculated average Fst value was 0.1, while Fst ranged from 0.005 (edm96) to 0.240 (gwm18). The barc48 (0.835) and wmc617 (0.04) loci exhibited the highest and lowest PI values, respectively. The frequencies of 34.82% of the alleles were lower than 0.1, and the accumulative PI value was 1.02 × 10 − 10 , indicating the likelihood of the same SSR pro le for two randomly selected individuals (Table 3). Table 3 Genetic diversity characteristics of SSR markers comprising number of observed alleles (N), number of different alleles (Na), number of effective alleles (Ne), polymorphism information content (PIC), haploid diversity (H), Shannon's information index (I), expected heterozygosity (He), probability of identity (PI), inbreeding coe cient (Fis), and xation index (Fst)  Population Structure

SSR Data
The STRUCTURE software was used for speci cation of the population structures of the seventy wheat genotypes in Iran using SSR, ITAP, and SCoT data. Based on the SSR data, the peak was obtained at K = 3 after the number of clusters was plotted against ΔK (Evanno et al. 2005), indicating that the current wheat genotypes could be divided into three subpopulations (Fig. 2). between subpopulations I and III, while the minimum value was calculated between subpopulations II and III. The AMOVA indicated that most of the variation was there within the three subpopulations, 88%, whereas a signi cant variation of 12% could be observed among them (p < 0.001, Table 5). Most of the SSRs were not in Hardy-Weinberg equilibrium in any of the three subpopulations; the exceptions included gwm149 in subpopulation I, gwm60 in subpopulations I and II, and wmc617 in all the subpopulations. Based on the principle PCoA analysis, the wheat genotypes were divided into three groups, approximately in accordance with the STRUCTURE analysis results (Fig. 3).

SCoT Data
The genotypes were assigned to four subpopulations (K = 4) based on the SCoT markers, Structure analysis, and ΔK calculation (Fig. 2). The mean Q value for each individual belonging to one of the four clusters was variable between 0.198 and 0.290. The average distances between the individuals in the four clusters were 0.4773, 0.522, 0.4715, and 0.4633, in that order. The wheat genotypes were partitioned into four subpopulations on the basis of the Nei's distance coe cient and PCoA analysis (Fig. 3). The subdivision groups were almost concordant with the results obtained by STRUCTURE. The parameters for the highest genetic diversity were calculated for subpopulation III (Ne = 1.57, I = 0.481, and He = 0.326), and the lowest variation was estimated for subpopulation I (Ne = 1.444, I = 0.379, and He = 0.256). Private amplicons were found in three of the four subpopulations, and the maximum number (5) was estimated for subpopulation III. Based on the AMOVA results, the highest percentage of variation was estimated within the subpopulations (89%), while the variation among the subpopulations (11%) was signi cant (p < 0.001, Table 5).

ITAP Data
Based on the ITAP data, the best value of K was calculated through the STRUCTURE software and the STRUCTURE Harvester website. The wheat genotypes were attributed to six subpopulations (Fig. 2). The average Q value of each individual in each subgroup varied from 0.538 to 0.998. The mean distances between the individuals in each subgroup were 0.3126, 0.3388, 0.3397, 0.4074, 0.1793, and 0.2929. Based on the PCoA analysis, the wheat genotypes were subdivided into ve distinct groups closely consistent with the STRUCTURE clusters (Fig. 3). The statistics on maximum genetic diversity, including Ne = 1.35, I = 0.30, and He = 0.20, were measured in subgroup VI, and the minimum values (Ne = 1.21, I = 0.174, and He = 0.12) were obtained for subgroup III. Private amplicons were found in ve out of the six subgroups, of which subgroup I involved the maximum number (6). Using AMOVA, The highest percentage of variation (51%) was estimated within the subgroups. Nonetheless, the percentage of variation between the subgroups (49%) was signi cant (p < 0.001, Table   5).

Informative Markers for the Grain Fe and Zn Content
The relevant informative or eventual SSR, SCoT, and ITAP for grain Fe and Zn content in the three years were examined using GLM, stepwise regression analysis, and Pearson's correlation test, and the results are presented in Tables 6-8. Some SSR amplicons pertaining to the barc48 (4DL, 6BS), gwm219 (6BL), gwm160 (4A), and wmc617 (4A, 4B, 4D) primers were associated with maximization and minimization of grain Fe and Zn content in the different years (Table  6). This possibly demonstrates the importance of their chromosome positions in the control of grain Fe and Zn contents. Approximately speci c SCoT primers, including SCoT1, SCoT12, SCoT22, SCoT33, and the SCoT29/1 combination, were associated with higher and lower values of grain Fe and Zn contents in the three years (Table 7). Commonly, ITAP amplicons comprising ITAP1-EM4, ITAP3-EM10, ITAP1-EM9, ITAP4-EM10, ITAP1-EM9, ITAP3-EM11, and ITAP1-EM3 were correlated with lower and higher grain Fe and Zn contents in the different years, as presented in Table 8.    Shahpasand, Karaj1, Karkheh, Hirmand, Arvand, Rasool, Shoeleh, Azar2, and Yavaros). These genotypes deserve consideration, and could be used as donor sources for increasing grain Fe and Zn content. The σ 2 g /σ 2 gy < 1 ratio indicated that the content is affected by the selection of environment conditions; therefore, successful breeding for enhancement of wheat grain mineral contents relies on the selection environment (Guttieri et al. 2015). In Peterson et al.
(1986), the ratio was less than one for Ca, Zn, and Mn and more than one for Fe. In Guttieri et al. (2015), however, it was estimated to be more than one for Fe and Zn content in wheat grains. The lower proportion of genetic variation (σ 2 g ) than that of genetic-environment interaction could be attributed to the complex nature of the grain Fe and Zn content controlled by polygenes (Khokhar et al. 2018). The genetic diversity of grain Fe and Zn content provides valuable sources for enhancement of the microelement concentration of wheat grains via traditional or molecular breeding approaches (Amiri et al. 2015). In our study, a high broad-sense heritability was estimated for grain Fe and Zn content, which had been reported in previous studies on bread wheat, wheat relative, and other cereal genotypes (Badakhshan et  protein, Zn, and Fe from leaves to the grain, and is also involved in the senescence of earlier ag leaves (Distelfeld et al. 2007;Coco et al. 2019). The positive correlation between micronutrients serves to improve wheat cultivars with enhanced microelements and protein contents (Goudia and Hash 2015) concurrently. A common molecular mechanism for grain micronutrient uptake and metabolism or common transporters could explain the positive correlation between grain Fe and Zn contents (Phuke et al. 2017).
The development of DNA-based molecular markers has enabled e cient speci cation of genetic diversity, QTL mapping, e cient gene pyramiding, and precise marker-assisted selection of linked targeted QTLs (Goudia and Hash 2015). We used the three marker systems SSR, ITAP, and SCoT to nd informative amplicons for grain Fe and Zn content. Each marker system was designed and ampli ed for speci c regions of the genome, including sequences anking hypervariable regions of microsatellites (SSR), sequences anking the start codon (ATG) in plant genes (SCoT, Collard and Mackill 2009), and 3' widely distributed conserved intron-exon junction sequences (ITAP, Xiong et al. 2013). The origin of polymorphism detected by ITAP markers could be intron length polymorphism and point mutations of ITAP binding sites. Therefore, variable markers with different genome target sequences enabled us to detect informative markers from different parts of the genome. Informative amplicons could be sequenced and developed as sequence-characterized ampli ed region (SCAR) markers for reliable, precise investigation of grain Fe and Zn content in a wheat genetic background.
Various informative SSR loci were identi ed for grain Fe and Zn content on chromosomes 1A, 3A, 4A, 7A, 1B, 4B, 6B, 7B, and 4D (Table 6), explaining 5.04 to 8.22 percent of the Fe content and 4.29 to 20.56 percent for Zn. It seemed that the homeologous group 4 chromosomes and the 6B chromosome played important roles in the control of grain Fe and Zn content in our wheat germplasm because these loci were detected commonly for grain Fe and Zn content.
Homoeology has been reported previously by (Peleg et al. 2009) for Fe, Zn, protein, and other mineral content in the tetraploid wheat population. Although macro elements such as phosphorous (P) and sulfur (S) were not addressed in the present study, high positive correlation has been reported previously by Peleg et al. (2009) to be there between P, S, Fe, and Zn. Peleg et al. (2009) identi ed QTLs on the homeologous group 4 chromosomes for P and S content in tetraploid germplasm, reinforcing the co-localization of ascribed QTLs for Fe, Zn, P, and S, on similar chromosomes. They concluded that the QTL homoeology might indicate synteny between different genomes. Particular SSR loci, including barc48, wmc617, and gwm160, were commonly correlated with grain Fe and Zn content, demonstrating the co-localization or pleiotropic effect of genes controlling Fe and Zn content. Dissimilar correlated markers with grain Fe content were detected in the three years, presumably because of signi cant genotype-environment interaction (Table 6) Several informative SCoT amplicons were detected for grain Fe and Zn content in the three years ( Table 7). The SCoT amplicons SCoT1 (400, 1300, 1400bp) and SCoT22 (400bp) were positively correlated with grain Fe and Zn content. Combined SCoT29/1 (700bp) was negatively correlated with grain Fe, while SCoT22 (550bp) and SCoT12 (250, 150bp) were negatively correlated, and SCoT35/2 (700, 800bp) was positively correlated with grain Zn content in the three years ( Table 7). The SCoT22 (400bp) amplicon was repeatedly detected in the three years for high Fe and Zn content; thus, it should be considered and sequenced in future studies. Multiple informative ITAP amplicons were identi ed for grain Fe and Zn content in the different years (Table 8) The average values of Ne, He, and I for the SCoT markers were 1.619, 0.350, and 0.515, respectively. On the other hand, the mean values for the ITAP markers were Ne = 1.535, He = 0.419, and I = 0.462. These results indicated that the SCoT markers detected more genetic variability in the wheat population than the ITAP markers. Besides, the discriminating power of the SCoT and ITAP markers was compared based on the EMR and MI values, demonstrating that the SCoT markers could potentially be more e cient in detection of genetic diversity. Previous research has reported that SCoT markers have effectively indicated genetic diversity in the wheat population (Nasrollahi et al. 2019; Khodaee et al. 2021), but there is yet no reported study on application of ITAP markers to wheat genotypes.

Population Structures and Relationships
Based on the SSR data, the wheat genotypes were divided into three subgroups, in accordance with the PCoA results. The genotypes with the largest numbers, the highest consistency with the pedigree information, were classi ed as subpopulation III. The lowest-number genotypes were classi ed as subgroup I.
However, the number of common parents of the genotypes in this group was larger than that in subpopulation II. Moreover, two tetraploid durum wheat genotypes (Karkheh and Arya) and ve landraces with similar origins (speci cally, RashaGol, GolSepi, and AarasGolsoor from the Kurdistan region of Iraq), were classi ed altogether into group I. The maximum numbers of private alleles, 0.594 (He) and 1.1080 (I), were calculated for subpopulation III. Private alleles are indicators of private genetic diversity, and loci with different alleles could potentially be useful in breeding programs (Eltaher et al. 2018). The majority of wheat genotypes ranked as possessing high grain Fe and Zn contents were categorized into subgroups I and III, respectively.
Using the SCoT data, the wheat genotypes were divided into four admixture subgroups. The number and quality of genotypes in each subpopulation were in line with those in the PCoA results. Most wheat genotypes were classi ed into subpopulation IV, with the largest number of common parents. Unlike in the SSR subpopulations, none of the above-mentioned landraces or tetraploid genotypes was grouped. Instead, two tetraploids, including Yavaros and Karkheh, were categorized in the same group. The highest genetic diversity (He = 0.326, I = 0.481) and the largest number of private amplicons (5) were estimated for subpopulation III, indicating proprietary genetic diversity within the subpopulation.
Based on the ITAP data, the wheat genotypes were portioned into six subpopulations. Using PCoA analysis, the wheat population was divided into ve subpopulations, closely consistent with the results of the structure analysis. The largest subpopulations were V and VI, comprising 15 wheat genotypes each. However, subpopulation II exhibited the maximum consistency with the pedigree information. As in the SCoT-based subpopulations, two tetraploid wheat genotypes, including Yavaros and Karkheh, were grouped together. The greatest genetic variation (He = 0.2, I = 0.30) was estimated for subpopulation VI. In contrast, subpopulation I involved the maximum number of private alleles (6), indicating unique genetic diversity, presumably due to the presence of wheat genotypes with various, uncommon parents.
Based on the three different marker data, the numbers of subpopulations were different, but the groupings of the genotypes were identical in some cases. The two landraces GolSepi and Sorkhtokhm and cultivars like Golestan and Chamran, Rasool and Shiroodi, and Falat and Karaj2 were paired together due to equal parents, similar geo-ecological conditions, and presumably similar traits as considered by breeders. The SSR markers were the most consistent with the pedigree information, followed by SCoT and ITAP, in that order. This nding was not unexpected because of the speci city of the SSR primers.
The results of AMOVA using the SSR, SCoT, and ITAP data indicated high variation within the subpopulations (88, 89, and 51 percent, respectively).
Nevertheless, the divergence among the subpopulations (12, 11, and 49 percent, respectively) was also signi cant (p < 0.001). The low genetic variation among the subpopulations could be attributed to the gene ow due to the presence of common parents in the subpopulations. It could be assumed that different growth habits, ecological adaptation, breeding goals, and ploidy possibly determined the genetic diversity within the subpopulations. Consistent results have been reported in earlier studies on wheat populations. In accordance with Eltaher et al. (2018), the distinct structure observed in our wheat population could be accounted for broadly in terms of common parents, common breeding goals for selection of genotypes, or common geographic conditions. Most of the studied genotypes were obtained mainly from the International Maize and Wheat Improvement Center (CIMMYT) breeding programs.

Conclusion
Bioforti cation of cereals, as a primary food source in developing countries, is recommended to prevent severe malnutrition and health problems. Substantial genetic variability was estimated for grain Fe and Zn content in the wheat genotypes examined in the current study. Variation in grain Fe and Zn content predisposes improvements in cultivars in terms of nutrient uptake from the soil, bioavailability, and grain lling. By the same token, association of informative molecular markers with responsible genes provides an e cient marker-assisted selection of genotypes and the possibility of pyramiding major genes in the ideal genotypes. A number of informative SSR, SCoT, and ITAP amplicons were identi ed, positively correlated with high Fe and Zn content. They could potentially be used to improve grain Fe and Zn content in wheat breeding programs. The markers considered here have to be further assessed in larger wheat populations using the association mapping approach. Some of the amplicons correlated with grain Fe and Zn content were constant across at least two years and common for grain Fe and Zn, reinforcing the co-segregation of the alleles involving these microelements. Regardless of the signi cant impact of the yeargenotype interaction, a few stable genotypes were ranked highly for Fe, and Zn content could potentially be used in hybridization programs as donor parents.
High genetic diversity among the wheat genotypes was revealed by the SSR, SCoT, and ITAP markers.