The most drought-tolerant genotypes (N = 10) were selected, and DNA was extracted from two to three leaves of six old seedlings. The DNA extraction was performed in the Biotechnology laboratory in the Genetics Department, Faculty of Agriculture, Assiut University. DNA was extracted from each genotype from two to three leaves using the Thermo Scientific GeneJET Plant Genomic DNA Purification Mini Kit protocol. Nine primer combinations of dehydration-res element binding proteins (nine fixed forward primers in combination with nine reverse primers) developed by Liu et al. (2018) were tested with the ten drought–tolerance genotypes. Primer codes and sequences of the forward and reverse primers are shown in Table 1. A gradient PCR was performed in order to determine the optimal annealing temperature for each primer used. The gradient test was performed using a gradient annealing temperature of 70 < 60 >50. The method of Thermo scientific PCR Master Mix protocol (Scientific, 2012) was followed for PCR reactions; each amplification reaction was carried out in a total volume of 20μL, containing 1x PCR reaction mix buffer (10 μL), 0.2 of each forward and reverse primer and 1 μL of template DNA, 8.6 μL H₂O. Polymerase chain reactions (PCRs) were carried out using the following program: initial denaturation at 94°C for 5 min, 35 cycles, while denaturation at 94°C for 30 s, annealing at 52–69.6°C for 30 s, 72°C for 60 extensions, and a final extension at 72°C for 5 min. The PCR products of each reaction (20 µl) and a 1,000 bp ladder marker (1 µl) were electrophoresed onto submerged agarose gel of 1% concentration containing 0.05 μL Supper Saffa in 50 ml TBE buffer (50X). Electrophoresis was carried out under a constant voltage of around 80 V for approximately 2–2.5 h. The banding patterns were visualized under Transilluminator and photographed using a gel documentation system (Ultra-Violet Product, Upland, CA, USA).

Analysis of variance (ANOVA) results for the morphological (tolerance and recovery) traits scored in this study at the seedling stage are presented in Table 2 and Supplementary Table S2. The results showed a highly significant variation (p < 0.01) among the genotypes for all traits measured under drought stress. For the tolerance traits, the highest and lowest p-values among the genotypes were for LW5 and LW1, respectively. For the recovery traits, RB (11.52**, p > 0.01) had the highest p-value, while DSR (6.24**, p > 0.01) had the lowest. All traits had a wide range of heritability (H²), and the estimates for the tolerance traits ranged from 69.02 (LW1) to 87.76 (LW5), while it ranged from 83.98 (DSR) to 91.32 (RB) in recovery traits. The heritability estimates for the recovery traits were higher than those for the tolerance traits. The drought tolerance index (including TI and RI) had the highest H² among the selection indices at 90.71, while seedling length (SL), which was scored before the drought, had an H² value of 92.27.

The distribution of all genotypes in relation to all traits scored under drought stress is presented in Supplementary Figure S5A,B. Three drought indices were calculated to better describe drought tolerance in wheat. The phenotypic variation among the genotypes for three drought indices (RI, TI, and DTI) is illustrated in Figure 1 and Supplementary Figure S5C. The drought tolerance index (DTI) divided the genotypes into three categories: tolerant (13 genotypes), intermediate (55 genotypes), and susceptible (104 genotypes) (Figure 1). Based on the DTI, the Egyptian cultivar Shandweel-1 (DTI = 2.7) was identified as the most drought tolerant, while the United Kingdom cultivar Little Tich was the most susceptible (DTI = 6.6).

The physiological traits were scored under drought and control conditions, and the ANOVA showed highly significant differences among all genotypes (Table 3 Supplementary Table S3). The ANOVA analysis also showed highly significant differences between treatments (control vs. drought). The differences among the three biological replicates were insignificant except for the protein and proline contents. The interaction between genotypes and treatments was highly significant. The F-values among the genotypes for all of the physiological traits were higher under drought stress when compared to the control (Table 4 Supplementary Table S4), except for those of fructose content. All physiological traits showed very high heritability estimates in both conditions, as the heritability estimates under drought were higher than in the control condition except the fructose trait had an H² of 97.42 and 96.98 under control and drought, respectively. Under control conditions, the heritability varied from 96.41 for Protein to 98.68 for proline, while under drought stress the heritability ranged from 96.98 for Fructose to 99.9 for proline (Table 4).

Generally, the averages for all the physiological traits under drought stress were higher than those under well-watered conditions (Control), except in protein content (Table 4 and Supplementary Figure S6). The physiological changes in the leaves (reduction or increase) due to drought stress are illustrated in Figure 2. On average, all physiological traits increased due to drought stress, ranged from 1.07% (F) to 33.6% (Pro), except the protein content, which had a reduction of 10.12%. Amino acid, total soluble carbohydrate, and glucose increased by 3.8, 12.8, and 13.68 mg/g DW, respectively.

The phenotypic and genotypic correlations among all traits are presented in Table 5. The genotypic correlations among the traits were higher than those of the phenotypic correlations. The phenotypic and genotypic correlations among the recovery traits were higher than the correlations among the tolerance traits. SL, which was scored before exposing the plants to drought, was found to be highly and significantly associated with all traits scored in this study. For the tolerance traits, SL was found be positively correlated with S_LW (r = 0.50**) and negatively correlated with DTW (r = −0.68**). The SL showed a significant correlation with the recovery traits as it was positively correlated with DTR (r = 0.51**) and negatively correlated with both RB (r = −0.55**) and DSR (r = −0.56**). Moreover, SL had a positive and significant correlation with three selection indices.

For tolerance traits, the sum of leaf wilting (S_LW) was negatively phenotypic and genotypic correlated with DTW r _phenotypic (p) = -0.79**, p < 0.01, (r _genotypic (g) = -0.86++). Tolerance traits had a high significant correlation with all three selection indices.

For recovery traits, on the other hand, highly significant phenotypic and genotypic correlations were found among the recovery traits scored after irrigating the drought-stressed plants. Days to regrowth had negative phenotypic (r_p) and genotypic correlations (r_g) with RB (r_p = −0.87**, r_g = −0.92++) and DSR (r_p = −0.95**, r_g = −1.00++). Regrowth biomass had high phenotypic and genotypic correlations with DSR (r_p = 0.90**, r_g = 0.95++). The recovery traits also had highly significant correlations with the three selection indices.

By looking at the phenotypic and genotypic correlations between the tolerance and recovery traits, it was observed that there was a significant correlation between the two groups of traits (Table 5). Sum of leaf wilting (S_LW) was negatively correlated with RB and DSR but positively correlated with DTR. Days to wilting (DTW) had a lower significant phenotypic and genotypic correlation size with the recovery traits compared to SLW. Sum of leaf wilting (S_LW) was found to have the same phenotypic correlation with DTR (r_p = 0.57**) and DSR (r_p = −0.57**), while DTW had the same correlation with DTR (r _p = −0.49**) and RB (r _p = 0.49**).

The tolerance index (including S_LW, DTW, and SL) had a higher significant phenotypic and genotypic correlation with tolerance traits than recovery traits. Likewise, the recovery (including DTR, RB, and DSR) index was found to have highly significant phenotypic and genotypic correlations with the recovery traits compared with the tolerance traits. The recovery index was positively and significantly correlated with the tolerance index, with a correlation value of 0.59**. Interestingly, DTI (including TI and RI) was highly and significantly correlated with all traits (tolerance and recovery traits) assessed in this study. The highest correlation between DTI and the tolerance traits was for S_LW (r = 0.88**), while DTR, among the recovery traits, had the highest correlation with DTI (r = 0.88**). The DTI had the same significant phenotypic correlation with RI and TI (r _phenotypic = 0.89**).

The correlation coefficients between all morphological and physiological traits under drought stress are shown in Table 6. The physiological, protein (P), proline (PRO), and fructose (F) did not have any significant correlations with any of the morphological traits. Amino acid (AM) had positive and significant correlations with S-LW (r = 0.27**), SL (r = 0.24**), and DTR (r = 0.36**). Furthermore, AM had negative correlations with DTW (r = −0.26**), RB (r = −0.38**), and DSR (r = −0.39**). A positive and significant correlation was found between AM and the three selection indices (TI, RI, and DTI). Notably, both G and TSC had similar association trends with the morphological traits, as they had negative and significant correlations with S_LW, SL, and DTR (Table 6) and positive significant correlations with DTW, RB, and DSR. The selection indices had negative and significant correlations with G and TSC.

To select the most promising genotypes with high drought tolerance from both the tolerance and recovery traits at the seedling stage, the genotypes were sorted for all traits based on the direction of drought tolerance from most tolerant to susceptible. Then, the 20 most tolerant genotypes in each trait were selected. Finally, genotype was selected from the top 20 genotypes if it was tolerance criteria in at least five traits and DTI. A set of 13 genotypes were ultimately identified as drought tolerant (Supplementary Table S1). Six of these 13 tolerant genotypes (MISR1, SAKHA93, Shandweel-1, PI525434, SIDS13, and Hutch) were among the 13 most drought-tolerant genotypes for nine traits, while three (SIDS12, Gimmeiza-12, and Sohag-3) were among the 13 most drought-tolerant genotypes in eight traits. The genotypes Beni Swief-5 and Beni Swief-7 were tolerant to drought for seven and six traits, respectively (Table 7). The ramming two genotypes (Gimmeiza 11 and Gimmeiza-07) were tolerant to drought in five traits. All 13 genotypes had a short height, high days to wiling, regrowth biomass after drought, and high survival rate after drought, and low levels of wilt. Finally, based on the number of traits and the value of DTI, the 10 most drought-tolerant genotypes were selected, and their physiological changes to drought and specific drought genes were investigated (Table 7).

In addition to the 10 most drought tolerant genotypes at the seedling stage, the 10 most susceptible genotypes were also selected to give a reliable image of the main differences between the tolerance and susceptible genotypes (Table 7).

Comparisons between the 10 most drought tolerant and susceptible genotypes (Figures 3A,B), showed that the drought-tolerant genotypes had increased G, TSC, P, and PRO traits, but decreased F and AM traits under drought stress (Figure 3A). For the tolerant genotypes, the soluble proteins, proline content, glucose, and total soluble carbohydrate content increased under drought stress compared to the control. In contrast, F and AM decreased under drought stress compared to the control, whereas PRO increased by 53% compared to control. While F and AM were decreased by drought stress competed with the control conditions (Figure 3A). The susceptible genotypes also had increased PRO, AM, and G under drought stress, while there was a notable decrease in TSC, F, and P compared to the control (Figure 3B). The highest increase in AM was identified in the susceptible group.

The 10 most drought-tolerant genotypes (Table 7) in this study were tested for the presence of nine specific DREB alleles. The genotyping revealed the amplification of only six DREB genes, as three primers did not give any specific bands (Figure 4). For the six amplified primers, the presence/absence of each gene was scored in each genotype (Table 8). The six primers showed clear polymorphism among the 10 most tolerant genotypes. MISR1 and SAKHA93 had six different DREB genes, while the American genotype Huch had one (DREB1-D1). The SIDS12, Gimmeiza-12, Shandweel-1, Beni Swief-5, and SIDS13 genotypes each contained five genes, while the Sohag-3 and PI525434 genotypes contained three.

The genetic distance among the most tolerant genotypes was calculated using 21,450 SNP markers (Figures 5A,B, and Supplementary Table S5). The analysis of the dendrogram divided the tolerant genotypes into three different branches. The branch I included eight Egyptian genotypes and one from Morocco. Branches II and III each included one genotype, Sohag 3 (Egypt) and Hutch (USA), respectively. The genetic distances among the genotypes ranged from 0.189 (Gimmeiza-12 and SIDS13) to 0.488 (MISR1 and Sohag-3). Sohag-3 had a genetic distance of >0.46, and Hutch had a genetic distance of >0.39 when compared with all genotypes.

The high significant genetic variation among the genotypes in all traits was believed to be due to the diversity among these genotypes, as they were from 20 different countries covering all continents (except Antarctica). The germplasm used in this study could be utilized to detect a large amount of genetic variation related to drought tolerance in wheat at the seedling stage and could thus be fruitful for plant breeders to help discriminate between tolerant and susceptible genotypes.

Two types of drought tolerance trait, tolerance, and recovery, were scored in this investigation, and the data provided will help to identify the different mechanisms of drought tolerance in wheat at the seedling stage (Sallam et al., 2018b; Sallam et al., 2022).

Tolerance traits reflect the ability of a plant to tolerate prolonged drought stress. Leaf wilting and days to wilting are tolerance traits which directly associated with drought tolerance (Bowne et al., 2012; Muir and Thomas-Huebner, 2015; Drira et al., 2016; Sallam et al., 2019). Leaf wilting indicates a deficiency in soil moisture and subsequent water uptake and transport to the shoots (Sanad et al., 2016). To better reflect the symptoms of water deficiency on the leaves, all five scores (LW1, LW2, LW3, LW4, and LW5) were summed up. SLW is an important accumulative trait that reflects the effects of drought stress on plant leaves. Many previous studies only scored leaf wilting once during the drought treatment (Sayed et al., 2012; Pathan et al., 2014). However, scoring leaf wilting multiple times during the drought treatment enables the precise evaluation of drought tolerance for each respective genotype over time (Sallam et al., 2019). In this investigation, both traits showed large variation among the genotypes, ranging from 14 to 24 and 4–9 days for SLW and DTW, respectively.

In contrast, recovery traits describe a plant’s ability to regrow and recover after irrigation following prolonged drought stress. Bearing in mind that these traits were scored after cutting and reirrigating the drought-stressed plants. Therefore, these traits not only tested the drought tolerance in plants but also reflected their ability to produce new shoots after drought stress. These essential traits should be considered together when scoring plants after irrigation. It was noted that some genotypes started to regrow only a few days after irrigation, such as Gimmeiza-07 and SIDS12, but at the end of the experiment, these genotypes had very little regrowth. Therefore, selection should be made carefully for genotypes that regrow after a few days but also have high regrowth biomass at the end of the drought experiment. Regrowth in wheat after drought conditions has been scored at the seedling stage. Previous studies, such as Pearce (1985), scored the regrowth of wheat seedlings after drought by measuring leaf height before and after rewatering within a few hours and on subsequent days.

The results from this investigation indicate that three selection indices (TI, RI, and DTI) should be utilized to select the most drought-tolerant genotypes. The drought-tolerant index was the main index, as it included both indices. The most important feature of the selection index developed by Falconer (1996) is the possibility of including more than one target trait. In this investigation, high levels of variation were found among all genotypes based on the three selection indices, which could thus be utilized to select promising drought-tolerant genotypes.

The high H² levels found for all traits scored in this study promise effective selection by which to improve drought tolerance in wheat at the seedling stage. SLW and DTW had H² values of 0.84 and 0.83, respectively. Sallam et al. (2018b) reported approximate heritability for SLW with an H² of 0.82, while their DTW value for H² was 0.72, which was lower than that reported in this study. High heritability estimates were also identified for the recovery traits (0.84–0.91), and they were higher than those reported by Sallam et al. (2018b) (0.74–0.88). Likewise, the heritability of the three selection indices reported in this study was also higher than those reported by Sallam et al. (2018b). These results differed due to the nature of the populations used in the two studies. The diverse population of this study possessed higher genetic diversity and variation when compared with the biparental population tested by Sallam et al. (2018b) for drought tolerance. Recovery and tolerance traits were recently reported under drought stress by Sallam et al. (2022) in a diverse winter wheat population at the seedling stage. They reported that the recovery traits had higher H² (0.75–0.89) estimates than the tolerance traits (0.71–0.84). These results further support the H² results obtained in this study on spring wheat.

As described, high levels of genetic variation and heritability estimates allow for selecting the most drought-tolerant genotypes for further improvement through plant breeding programs. Most earlier studies depended on single trait selection to improve tolerance to stress which is not recommendable (Sallam et al., 2019). Instead, multiple trait selection is more fruitful when selecting the target genotypes.

The 13 genotypes selected were from Egypt (11), the USA 1) and one genotypes was from Morocco country. The tolerant genotypes had diversity in the traits correlated to their tolerance. Identifying different genetic resources to investigate drought tolerance will undoubtedly help to improve this trait in spring wheat at the seedling stage and expand the circle of genetic diversity in Egypt. Thus, genotypes could be selected from this set and used as candidate parents for crossing into future breeding programs to improve drought tolerance. Of the 13 genotypes, Shndweel-1, PI525434, Sakha 93, Hutch, Sids 13, Misr 1, Sids 12, Sohag 3, Gimmieza 12, and Beni Swief 5, were selected as the ten most drought tolerant for both the tolerance and recovery traits. They could thus be used in future breeding programs to accelerate the improvement of drought tolerance in wheat. These ten highly drought-tolerant genotypes had the lowest DTI (Table 7). The DTI provides a valuable method by which to truly select the most drought-tolerant genotypes with superior performance in more than one trait. These ten genotypes were used for further physiological and genetic studies.

In this study, the physiological changes at the seedling stage were evaluated by assessing six physiological traits (protein, amino acids, proline, glucose, fructose, and total soluble carbohydrates) to understand the biochemical changes that occur in wheat leaves in response to drought stress and consequently, how to alleviate this stress. All these physiological traits had a direct relation to drought stress, and some previous studies, such as Abid et al. (2018), used these traits to study this trait at the tillering and jointing stages, respectively.

These high levels of genetic variation among genotypes in all physiological parameters could provide an extremely useful resource for both breeders and geneticists to efficiently understand the changes in the physiological parameters that occur in plants to alleviate the effects of drought stress. Moreover, the high levels of genetic variation indicated that there was also variation in the ability of the different genotypes to make substantial changes to their physiological parameters. The presence of significant differences between the two different treatments indicated that drought affects the performance of the genotypes when compared to their performance under the control conditions. Previous studies, such as Nowsherwan et al. (2018), stated that the performance of the genotype varies according to the different conditions and stages. Moreover, it was observed that the genotype responses differed with the different treatments, and this was shown by the presence of significant variation between the G × T interactions in all physiological traits. This effect can be observed by estimating the changes in the physiological parameters due to drought stress in the genotypes (Figure 2). The most significant change was in the proline content, which increased by 33.6% under drought stress, and its accumulation was reported in stressed plants compared to non-stressed plants (Sallam et al., 2018a; Dawood et al., 2020). In the current study, the physiological traits all had high H² estimates under both control and drought conditions, but they were higher under drought conditions when compared with the control. These higher heritability estimates (H²) suggested that using physiological traits to improve drought tolerance would be more successful.

On average, for each population in this investigation, the physiological traits increased under drought stress compared to the control, except for protein content (Table 4 and Supplementary Figure S6). This indicates that the genotypes use different mechanisms by which to deal with and adapt to drought stress fluctuations. This result was expected due to the high levels of genetic diversity in the tested genotypes, which were collected from 20 different countries. In addition, studying the physiological changes that occur because of drought stress was useful as they provided information on the different ways in which plants function to counteract and relieve stress. The study of these changes showed that the content of all physiological traits increased or accumulated in the leaves except protein content. The soluble protein content decreased by 10.12%, indicating that the protein content was the trait most affected by drought stress. It is clear in this study that the protein content was decreased in favor of the proline and amino acid increases (Figure 2). Gilbert et al. (1998) reported that the protein reduction could be associated with an increase in amino acids that may serve as a readily available energy source or as a nitrogen source during limited growth and photosynthesis and the detoxification of excess ammonia under periods of stress.

Physiological and developmental plant responses to drought were shown to occur by reprogramming gene expression and metabolism (Reddy et al., 2004; Chaves et al., 2009; Hayano-Kanashiro et al., 2009; Zhang et al., 2014). Responses to drought stress depend on plant species, the stage of development, the rate of dehydration, and the duration and severity of stress (Reddy et al., 2004; Chaves et al., 2009). To elucidate a plant’s ability to survive under drought conditions, it is important to study the physiological, biochemical, and genetic basis of adaptation and tolerance as well as the mechanisms of recovery under drought stress. The physiological modifications in response to water stress were studied, herein, in terms of soluble proteins, amino acids, proline, glucose, fructose, and total soluble carbohydrate contents, which were analyzed under drought and control conditions to understand the biochemical changes in shoot metabolites in spring wheat plants. The average amount of glucose, fructose, and total soluble sugars (as the average of 172 cultivars or genotypes) increased in response to water stress, but for fructose, the increase was very slight. The accumulation of sugars in response to drought stress was not uniform, indicating that there were categorical tasks in charge of each sugar component. Fructose only slightly responded to the drought stress.

The phenotypic and genetic correlations among all traits scored in this study shed light on the different drought tolerance mechanisms. Overall, the genetic correlations among the traits were higher than the phenotypic correlations, and these results corresponded with those obtained by Sallam et al. (2018b). Genetic correlation is an informative analysis as it provides predictions for the selection response in one or more traits to be made when selection is another trait (Hill, 2013).

The high level of genetic and phenotypic correlations among all traits promises fruit selection for a group of traits that are directly associated with drought tolerance. It was noted that the tolerance and recovery traits had very high and significant correlations within each group. While there was also a significant correlation between the tolerance and recovery traits, the correlation size was smaller than that within each group. The correlation between RI and TI was higher than the correlation between the traits in the two groups. This indicates that including more than one trait in a selection index could be more beneficial for selecting target traits. Interestingly, DTI had high and significant levels of correlation with all traits in both groups. These results highlighted the importance of using a selection index to improve target traits and optimize the selection to improve drought tolerance. The phenotypic correlations reported by Sallam et al. (2018b)showed that there were no correlations between the tolerance and recovery traits. Even after creating a selection index for each group, RI and TI showed no significant correlation, suggesting that different genetic mechanisms controlled these traits. The absence of correlation between the tolerant and recovery traits was also validated in a highly diverse winter wheat population by Sallam et al. (2022), conducting QTL mapping (winter biparental population) and GWAS (diverse winter wheat population). The study’s results revealed that each group of traits was controlled by different QTLs and concluded that both types of traits were controlled by different genetic mechanism different genetic mechanisms. The gene and SNP networks analysis supported this notion (Sallam et al., 2022). It should also be considered that Sallam et al. (2018b) tested winter genotypes which are completely different from the spring wheat genotypes used in this study. Winter wheat, for example, has a crop coefficient of 0.7 for nonfrozen soil, while the spring wheat crop coefficient is 0.3. The lower the crop coefficient, the lower the water demand and water stress (Allen and Zaplachinski, 1994), consequently, spring wheat is more tolerant than winter wheat. The relationship among these traits expanded our knowledge and understanding of the drought tolerance mechanisms in both spring and winter wheat; both appeared to possess different genetic mechanisms by which to alleviate the effects of drought stress. This information will be valuable for wheat breeders and geneticists as it can be utilized for the selection of drought tolerance in spring and winter wheat.

It was of note that seedling height had a significant correlation with tolerance and recovery traits in this investigation. This indicates that shorter plants were more drought tolerant. This trait could thus be used to predict and select drought tolerance as this trait was scored before exposing plants to drought stress. The same trait did not correlate with recovery and tolerance traits in winter wheat (Sallam et al., 2022).

The correlation between morphological traits (tolerance and recovery) and physiological traits helps elucidate the changes in wheat plants at the seedling stage to alter and alleviate the effects of drought stress. Fructose was previously reported to be associated with secondary metabolite synthesis and not osmoprotectants (Younes et al., 2019). In this investigation, fructose was found to only slightly respond to drought stress conditions. It was thus determined that fructose was not involved in osmotic adjustments, which was also indicated by the correlation data, as it did not correlate with any of the studied traits for wilting or recovery. Younes et al. (2019) demonstrated that fructose might be related to the synthesis of substrates related to phenolic compound synthesis. Glucose, in contrast, which acts as a substrate for cellular respiration or osmolytes to maintain cellular homeostasis, was found to accumulate in the shoots of drought-treated plants, implying that glucose is associated with osmotic adjustments under drought stress conditions, as was reported by Misra and Gupta (2005). It is of note that glucose was found to be negatively associated with witling traits. Thus, a positive relationship was found between days to wilting and glucose content indicating that the tolerant genotypes accumulated high levels of glucose to tolerate a prolonged period of drought. The significant correlations between glucose and the recovery traits indicated that increased glucose accumulation levels improved survival and recovery after reirrigation. This result provides a clear explanation of the importance of glucose as an osmotic against drought stress. The above-described tendencies for glucose were observed for the total soluble carbohydrates, which contributed to the osmotic adjustment of wheat plants exposed to prolonged drought stress. In conformity, carbohydrates can contribute to 30–50% of the osmotic adjustment in glycophyte plants (Al-Thani and Yasseen, 2018).

The nitrogenous components of the stressed shoots showed that the soluble protein content was reduced in favor of the accumulation of free AM and proline. This increase in AM and proline under stress conditions may result from the degradation of proteins, affect their synthesis, inactivate major enzymes, and destroy membranes (Fahad et al., 2017; Sallam et al., 2018a; Dawood et al., 2020). Increased levels of free amino acids and proline were identified in different plants under abiotic stress (Sadak, 2016b, 2016a; Tawfik et al., 2017; Sallam et al., 2018a). It was concluded that these compounds have an important role in enhancing the tolerance of plant cells to various abiotic stresses by increasing the osmotic pressure in the cytoplasm and increasing the relative water content, which is essential for plant growth and different metabolic processes. The average proline levels identified for the 172 tested cultivars showed no correlation with any of the studied morphological or recovery-r traits. The collection used covered a wide range of cultivars and included variants susceptible to drought stress. Of note is that most of the studied cultivars ranged from moderately tolerant to sensitive, and only 24 cultivars were tolerant. Thus, the categorization of these cultivars based on their drought responses will likely provide a reliable image of their physiological behavior in relation to the recovery and tolerance traits.

The AM was positively correlated with three drought indices (drought susceptibility). In this regard, the high levels of amino acid content may be associated with the degradation of proteins which were found to be reduced in response to drought. Thus, drought stress instigated the solubilization of proteins, which may have negative effects on the main enzymes related to various physiological processes. The production of high levels of amino acids under stress conditions could impact multiple processes. The high synthesis rates of the amino acids can result from proteolysis, or their consumption can be restricted due to a decrease in protein synthesis or secondary metabolite production (Silva et al., 2019). Araújo et al. (2011) reported that amino acids can be utilized as alternative respiratory substrates and provide stressed plants with an additional energy source during an energy deprivation situation. In situations with insufficient carbohydrate supply due to a decrease in photosynthesis rates, which usually occurs during stress conditions, plants can utilize amino acids as alternative substrates for mitochondrial respiration (Braun et al., 2015; Hildebrandt, 2018). The degradation pathways of some amino acids have been identified as essential factors for dehydration tolerance in Arabidopsis (Pires et al., 2016). Overall, the general metabolism of the studied cultivars was found to be shifted toward energy saving and stress defense, which leads to an arrest of growth and development. Plants start to invest their energy resources into the production of protective secondary metabolites and osmolytes to counteract the effects of drought stress.

To understand the physiological changes that occurred in plants to alleviate the effect of drought stress, their physiological characteristics were studied in the most drought-tolerant and susceptible genotypes (Figures 3A,B). The increase or decrease of each physiological trait because of drought stress in both the tolerant and susceptible genotypes was estimated. The increase or decrease in physiological components differed according to the stage in which the stress occurs, the duration and severity of the stress, as well as the performance of the genotypes (Reddy et al., 2004; Chaves et al., 2009). The ability of tolerant plants to respond to drought tolerance depends on the genotype. For example, the drought tolerance mechanisms of some genotypes include the accumulation of soluble sugars, proline content, amino acids, chlorophyll content, and enzymatic and nonenzymatic antioxidant activities (Abid et al., 2016). In this study, clear differences were observed between the drought-tolerant and susceptible genotypes for these physiological components, and there was no clear trend as some traits increased and others decreased. These genetic differences in the physiological components between genotypes can be used to effectively identify the genes controlling the physiological traits and accelerate the genetic improvement of drought tolerance (Bowne et al., 2012). The drought-tolerant genotypes showed apparent increases in their protein, proline, and sugar content and decreased their amino acid levels compared to the control, and the opposite occurred in the drought-susceptible genotypes (Figures 3A,B). The most important characteristic distinguishing the tolerant genotypes from the susceptible genotypes under drought stress is a significant increase in proline content, which increased by 50%, while amino acids were reduced by 65%. The drought-tolerant genotypes accumulated more proline in their leaves under drought stress at the expense of other free amino acids. In contrast, the drought-susceptible genotypes increased the amino acids in their leaves at the expense of their proline content. This indicated that the tolerant genotypes possessed genetic plasticity that allowed them to accumulate increased levels of proline under drought conditions, and this was utilized as a self-protection mechanism by which to counteract drought stress. Previous studies, such as Basudeb et al. (2019), reported that the presence of proline is a common trait in most cereals under drought stress Zali and Ehsanzadeh (2018) previously stated that only a few plant species can produce enough proline to notably reduce abiotic stress effects and thus utilize proline as another tolerance mechanism against drought stress. Moreover, proline is a source for several amino acids and nitrogenous compounds (Britikov et al., 1970). It also contributes to the stabilization of subcellular structures, the scavenging of reactive oxygen species, and the buffering of cellular redox potential under stress (Ashraf and Foolad, 2007).

Integrating specific DNA molecular markers for target traits in the breeding program will help in accelerating the breeding program. In this study, nine specific primers associated with nine different allelic variants of TaDREB1 genes developed by Liu et al. (2018) were used to screen the most drought tolerant genotypes for the presses of DREB genes. DREB genes constitute a large family and belong to transcription factors (TFs) that stimulate the expression of many functional genes (Agarwal et al., 2006; Samarah, 2016). A total of 210 DREB genes associated with abiotic stress tolerance (Niu et al., 2020). Identifying and validating new primers for these genes are very useful for breeding and genetics programs in wheat. Six specific DREB primers were polymorphic among the ten drought-tolerant genotypes (Table 8). The high polymorphism and diversity among the drought-tolerant genotypes indicated that the tolerant genotypes differed in the number of TaDREB1 gene haplotypes. It was reported that the drought-resistant materials showed inconsistent heterogeneity of TaDREB1 gene haplotypes and the nucleic acid polymorphisms of the TaDREB1 gene in wheat (Yousfi et al., 2016). In the present study, it was observed that the largest number of TaDREB1 alleles was present in both the Egyptian genotypes MISR1 and SAKHA93 (six genes), while the lowest number of these genes was present in foreigner genotypes such as Moroccan genotype PI525434 (three genes) and American genotype Hutch (one gene). The result also revealed that the Egyptian genotypes were higher drought tolerance than the Moroccan and American genotypes. By looking at the DTI of these genotypes, the correlation between number of TaDREB1 and DTI was negative (r = -0.66**), indicating that some of high tolerant genotypes may contain other drought genes and TaDREB1 alleles not only the source of drought tolerance in the selected genotypes. PI525434 (Morocco) and Hutch (USA) ranked fifth and sixth among the ten drought-tolerant genotypes, respectively. Therefore, they probably included other drought genes than those used in this study. These two genotypes can be utilized for crossing with the Egyptian genotypes to genetically improve drought tolerance at the seedling stage in spring wheat. Further molecular analyses should be done on these cultivars to discover more genes related to drought stress. The results of genotyping confirmed that primers of TaDREB1 alleles used in this study were effective and valuable for marker-assisted selection to test the presence of TaDREB1 alleles in a large number of genotypes in a short time as an alternative to breeding methods traditional.

The genetic distance analysis among the ten drought-tolerant genotypes provides valuable information on the diversity among these genotypes. Such information can be useful in selecting the candidate’s parents for future breeding programs (Eltaher et al., 2018; Mourad et al., 2020). Here, although the ten drought-tolerant genotypes included eight genotypes from Egypt, the genetic distance among them is still useful for the breeding program. Out of the ten genotypes, Sohag-3 was positioned in a separate cluster and had a genetic distance of >0.649 from all other genotypes. Unexpectedly, the two foreigner genotypes, Hutch (USA) and PI525434 (Morocco), were included in the cluster with the other seven Egyptian genotypes. Hutch had a higher range of genetic distance with the Egyptian genotypes than PI525434. The highest genetic distance was found between MISR one and Sohag-3 (0.664); however, including Hutch and PI525434 in the crosses with the Egyptian genotypes (e.g. Sohag-5) may be more fruitful for two reasons (I) increasing the genetic diversity among the Egyptian wheat gene pool and (II) the two genotypes may have other drought-tolerant genes rather than DREBs and the crossing with the Egyptian genotypes could produce wheat cultivars having more drought tolerance at the seedling stage. Genetic diversity is essential for plant survival in nature against the consequences of climate change and crop improvement (Bhandari et al., 2017). Including new plant genetic resources such as Hutch and PI525434 will undoubtedly provide the opportunity for wheat breeders in Egypt to develop new and improved wheat cultivars not only for drought tolerance but also for other desirable characteristics such as agronomic yield traits, quality traits, tolerance to biotic and abiotic stress tolerance, etc. (Al-khayri et al., 2019; Baenziger et al., 2021; Bhavani et al., 2021)