Introduction

Globally, wheat is the major crop and staple diet of many countries. Water shortage at critical growth stages is the most common problem and resultantly grain yield is significantly compromised. Water stress at flowering stage reduces grain number and grain filling (Shokat et al. 2020a, 2021a). Grain production is seriously hampered if water is not available at flowering and post-flowering stage. Plant genotypes enable to tolerate moisture stress at reproductive stage can contribute to sustainable wheat production in arid and semiarid regions of the globe.

Wheat output under water limited conditions depends upon certain agronomic and physiological traits and few key traits were identified to explain the plant performance (Reynolds et al. 2007; Langridge and Reynolds 2021; Shokat et al. 2023). Previous studies reported that canopy temperature depression (CTD) and normalized difference in vegetation index (NDVI) can be breeder friendly traits to describe the grain yield (Moinuddin et al. 2005; Reynolds and Langridge 2016; Reynolds et al. 2020; Shokat et al. 2023). Likewise, traits like better relative water contents, leaf osmotic potential, osmotic adjustments and water use efficiency at growth stages can explain grain yield (Moinuddin et al. 2005; Shokat et al. 2020a; Arif et al. 2021). Above than all, plants maintaining lower kernel abortion or higher grain number as well better grain weight at flowering stage moisture stress can express better harvest index (Shokat et al. 2021a; Ulfat et al. 2021).

Moreover, grain quality is equally affected under moisture stress as sucrose conversion into starch is reduced and resultantly less starch is accumulated (Lu et al. 2019). Starch is generally comprised of amylose and amylopectin and studies on two contrasting wheat genotypes reported a decrease in amylose contents under water stress conditions (Singh et al. 2008). This study indicating testing of large germplasm for flowering stage moisture stress can help us to validate the previous findings by quantifying the values of amylose and amylopectin in large set of wheat germplasm. Current study was planned to use diverse wheat germplasm obtained from Seeds of Discovery program of International Maize and Wheat Improvement Center (CIMMYT), Mexico. This germplasm has previously been tested for number of biotic and abiotic stresses, and number of QTLs and marker-trait associations were identified (Singh et al. 2018, 2021; Shokat et al. 2020b; Arif et al. 2021; Dababat et al. 2021; Akram et al. 2022; Saleem et al. 2022). Further, a novel gene, i.e., isoflavone reductase like was identified playing role in heat tolerance (Singh et al. 2018; Shokat et al. 2021b). Previous reports indicate that genetic diversity can play a crucial role to sustain grain yield under different biotic and abiotic stresses.

This study was planned to investigate the impact of flowering stage drought stress on starch contents, i.e., amylose and amylopectin of bread wheat germplasm derived from landraces and synthetic derivatives. We constructed a hypothesis that reduction in grain yield is due to less accumulation of amylose and amylopectin in grains.

Materials and methods

Plant material and crossings

Fifty wheat genotypes originated from the three-way top-crosses were evaluated under well-watered and flowering stage drought stress at International Maize and Wheat Improvement Centre (CIMMYT) Obregon, Mexico. Two crossing schemes were followed to make three-way crosses, i.e., the landraces were crossed with high yielding line of CIMMYT to make F1 crosses and F1 plants were crossed again with second best line of CIMMYT. In the same way, synthetic bread wheat derivatives were used in second crossing scheme to make three-way crosses. These elite parents were developed and tested resistant to diseases. Two hundred and forty-four crosses out of 1200 crosses of landraces and synthetic derivatives, performing better for yield and yield contributing traits were advanced. A selected bulk method was used to advance these generations up to TC1F5, and 50 diverse genotypes including three check varieties were selected for further evaluation.

Experimental setup

These 50 genotypes comprised of 47 derivatives of landraces and synthetics and three check varieties, i.e., Reedling, Baj and Kachu were raised during 2017–2018 at the experimental station “Campo Experimental Norman E Borlaug (CENEB)” of CIMMYT located near city of Obregon of Sanora state in Mexico. An alpha lattice design was used with two replications for each of well-watered and drought stress. The sowing date of the trails for irrigated and drought trials was November 7, and seed rate was 120 kg/ha. Further, two rows of each genotype were planted having 2-m length and 40 cm distance between rows. In total, approximately 500 mm of water was applied to irrigated trials. In contrast, drought treatment received only two irrigations (~ 180 mm) and irrigation of anthesis and post-anthesis was skipped. Normally at Obregon, Mexico, there is no rainfall through the wheat crop season, and water is controlled through the irrigation (gravity or drip system) provided during the season.

Phenotypic measurements

The material was diverse, and we observed variations for heading dates and other agronomic traits. To quantify the plant performance, following parameters were recorded:

  1. (a)

    To identify the cooler canopies an infrared thermometer sixth sense (LT300) was employed, and data of canopy temperature depression (CTD) were recorded. As per instructions, CTD measurements were taken from the middle area of two-meter row by adjusting ½ meter distance and 45° angle.

  2. (b)

    Readings of normalized difference in vegetative index (NDVI) were quantified for five seconds for well-watered and drought treatment using Green-seeker with a portable sensor (Mod: 505) and adjustable rate of application and mapping systems. NDVI measurements were taken during first week of every month starting from stem elongation to physiological maturity of the crop. The average values of 0–1 indicating no green to maximum green were used.

  3. (c)

    Days were counted from sowing to 50% flowering to record the data of heading (HD).

  4. (d)

    Maturity time (MT) was recorded as number of days counted from sowing to until all spikes changed into their specific color at maturity.

  5. (e)

    Five plants from each replication were selected to measure the plant height (PH) and measurements were taken from ground level to the top end excluding awns and data were noted in centimeters (cm)

  6. (f)

    Five spikes from each replication were selected to measure spike length (SL) and measurements were taken from the starting point of spike initiation to the end of spike excluding awns

  7. (g)

    Five spikes from randomly selected plants of each replication were threshed to count grains and average was taken to represent number of grains per spike (NGS)

  8. (h)

    Thousand grains from each replication were counted and their weight was taken in grams to represent thousand kernel weight (TKW)

  9. (i)

    Two rows each with 5-m length were harvested and threshed mechanically to obtain grain yield per plot (GY) of each genotype and this yield was expressed in units of kilograms per hectare.

Quantification of amylose and amylopectin

Flowering stage drought stress affects starch accumulation, i.e., amylose and amylopectin. Grains' amylose and amylopectin was estimated to determine the impact of drought on their ratio. Kits of Megazyme were used (www.megazyme.com) and samples were processed by pre-treating with ethanol to remove the lipids before the analysis (Morrison and Laignelet 1983) and procedure described by Yun and Matheson (1990) was opted with little modifications. Briefly, starch was dispersed by heating and treating with dimethyl sulfoxide and lipids from starch were removed by ethanol through precipitation. Lectin concanavalin A was used to precipitate amylopectin and removed by centrifugation. Likewise, glucose oxidase/peroxidase (GOPOD) reagent was used to hydrolyze amylose into D-glucose. The concentration of amylose in the starch sample is estimated as the ratio of GOPOD absorbance at 510 nm.

Data analysis

A two-way analysis of variance was computed using Rstudio.2022.7.2.576 where genotypes were considered as factor one, and irrigation levels were taken as second factor. Further, bar graphs were also drawn to depict the differences, and correlation was estimated using performance analytics package of Rstudio.

Results

Different yield-related traits were estimated and analyzed to understand the impact of water stress on different genotypes.

Traits related to earliness and physiology

Genotypes were significantly different from each other for PH, HD and MT, and NDVI where highest plant height was recorded for genotype 7,643,835 (120 cm), lowest HD for genotype 5,106,304 (60) and MT for genotype 7,641,562 (105), and highest value for NDVI for genotype 7,642,035 (0.84). Likewise, drought significantly reduced all these parameters in comparison to normal irrigation conditions. Moreover, genotype × irrigation interaction was also significant for NDVI, HD and MT, and pronounced reduction was recorded for these traits under drought stress. It is further indicating that some of the genotypes were more severely affected under drought in comparison others where genotypes 7,643,544, 7,644,626 and 5,106,304 showed highest reduction for NDVI, HD and MT, respectively. These results are indicating the drought sensitivity for these traits. However, non-significant results were recorded for CTD (Table 1, Fig. 1).

Table 1 Ranges, means, standard errors and analysis of variance among 50 genotypes for different phonological, ecophysiological, grain yield and starch contents at two levels of water
Full size table
Fig. 1
figure 1

Box plots depict the minimum, maximum, interquartile range and outliers for the traits related to earliness and physiology under irrigated (IR, blue) and drought (DR, red) conditions at flowering stage a indicates canopy temperature depression (CTD), b shows normalized difference in vegetation index (NDVI), c depicts plant height (PH), d illustrates days to 50% heading (HD) and e explains days to maturity (MT)

Full size image

Grain yield and yield-related traits

Significant differences among genotypes were recorded for yield and related traits (SL, NGS, TKW and GY) per plot. Highest SL was recorded for genotype 7,645,255 (14.4 cm), NGS for genotype 7,642,214 (71.6), TKW for genotype 7,640,871 (71.1 g) and GY for genotype 6,176,013 (71.75 kg/ha). Drought significantly reduced all yield-related traits, and severe reduction was record in grain yield per plot through reduced SL, NGS and grain filling. Further, significant interaction between genotype × irrigation was recorded for the grain yield per plot. It is explaining that although drought reduced the GY significantly yet, some genotypes were affected more severely than the others and highest reduction for GY was recorded for genotype7641502. Moreover, it is also indicating this genotype must not be grown in moisture limited conditions (Table 1, Fig. 2).

Fig. 2
figure 2

Box plots depict the minimum, maximum, interquartile range and outliers for grain yield and yield related traits under irrigated (IR, blue) and drought conditions (DR, red) at flowering stage. a indicates spike length (SL), b shows number of grain spike.−1 (NGS), c depicts thousand kernel weight (TKW) and d illustrates grain yield per plot (GY)

Full size image

Amylose and amylopectin

To see impact of flowering stage drought stress on grains quality, the parameters like amylose and amylopectin were estimated. The genotypes were significantly different for both amylose and amylopectin where genotype 7,645,129 showed highest amylose as well as lowest amylopectin contents (40.94 and 59.6). In contrast, genotype 7,643,587 showed the lowest amylose and highest amylopectin contents, respectively (24.5 and 75.5). It was also a noteworthy that drought significantly enhanced amylose contents and but reduced amylopectin contents. Moreover, interaction between genotype × irrigation was also significant for these two traits, and pronounced effect was recorded under drought stress (Table 1, Fig. 3). It is further indicating that though drought impacted these traits significantly however, some genotypes were affected more severely than the others, and highest change was recorded in genotype 7,641,585.

Fig. 3
figure 3

Box plots depict the minimum, maximum, interquartile range and outliers for starch contents under irrigated (IR, blue) and drought (DR, red) conditions at flowering stage a indicates amylose contents (AML) and b shows amylopectin contents (AMP)

Full size image

Combined correlation analysis

A combined correlation was estimated to identify the association of starch contents with different yield contributing traits under drought as well as irrigated conditions. A strongly significant (***) and positive correlation of GY was recorded with PH, NDVI, MT, SL, NGS and TKW. In contrast, correlation of GY was significantly (*) and negatively correlated with CTD. Likewise, NDVI, HD and MT significantly and positively correlated with number of NGS. Moreover, no significant correlation of starch content was estimated with any of the yield contributing traits (Fig. 4).

Fig. 4
figure 4

Correlation analysis among 50 genotypes for phenology, ecophysiological, starch and yield contributing traits grown with and without irrigation stress. * < 0.05, ** < 0.01, *** < 0.05 where * depicts significance with 95% probability, ** indicates significance with 99% probability and *** illustrates 99.9% probability. HD = days to 50% heading, PH = plant height (cm), NDVI = normalized difference in vegetation index, CTD = canopy temperature depression (°C) MT = days to maturity, SL = spike length (cm), NGS = number of grains spike-1, TKW = thousand kernels weight (g), GY = grain yield (kg ha−1) AML = amylose and AMP = amylopectin

Full size image

Contribution of landraces and synthetic wheat to improve grain yield and amylose contents

Supplementary tables S1 and S2 are presented to understand the contribution of landraces and synthetic bread wheat derivatives to improve GY and amylose contents in comparison to checks under the drought conditions. In total 5 genotypes, three from landraces and two from synthetic bread wheat produced higher GY under drought in comparison to better check (Table S1). Likewise, we identified 7 genotypes from each of the landraces and synthetic bread wheat derivatives gave higher amylose contents than the better check under drought conditions (Table S2).

Discussion

Deficient irrigation at critical growth stages of plant especially at flowering and post-flowering hampers the grain yield by reducing grain number and weight (Shokat et al. 2020b; Ulfat et al. 2021). This reduced grain yield is due to disruption of optimum water available from soil, higher ABA contents (Shokat et al. 2021a) and lower WUE (Ulfat et al. 2021). Further, reduced activity of carbohydrate metabolic and antioxidants enzymes can contribute to less grain number and filling at this stage of drought stress. Moreover, diverse germplasm derived from landraces or synthetic and bread wheat derivatives has been reported to contributed flowering stage drought or heat stress (Singh et al. 2018, 2021; Shokat et al. 2020b), however, little is known about the impact of flowering stage drought stress on starch (especially the amylose and amylopectin) of the germplasm derived landraces and synthetic bread wheat.

Previous studies reported that NDVI can be a good indicator of drought stress and can be selected at pre-breeding traits (Bushra et al. 2019). Moreover, canopy temperature can indirectly indicate the performance of better root system under drought and heat stress conditions (Pinto and Reynolds 2015). In current study, we recorded reduction in NDVI under drought however no significant impact of drought on canopy temperature was recorded. Further, drought stress at flowering stage also reduced grain yield significantly through reduction in grain number and less grain filling. Previous studies also reported a significant reduction in grain number and weight (Farooq et al. 2014; Shokat et al. 2020b; Ulfat et al. 2021). Further, we also identified a strong positive correlation of grain yield with NDVI but significant negative correlation with CTD validating the previous findings (Pinto and Reynolds 2015; Bushra et al. 2019). Further, to identify the role of flowering stage drought stress on starch contents, we investigated the amylose and amylopectin contents of grains after harvesting.

Plant grown under flowering stage drought stress can up-regulate the drought tolerant proteins but down regulate the grain filling cycle (Li et al. 2023). We also recorded a significant reduction in life cycle where plants matured early under drought stress indicating a reduction in grain filling period too (Table 1 Figs. 1, 2 and 4). Starch work conducted for two years in two wheat genotypes explained a significant reduction in amylose contents under severe drought stress indicating drought can impact grain quality as well (Zhang et al. 2017). Likewise, studies conducted on two wheat genotypes indicated that plant genotypes tolerant to flowering stage drought stress can maintain better amylose to amylopectin ratio (Li et al. 2023). Moreover, another study reported that drought stress shortened the grain filling process and total accumulation of starch (Lu et al. 2019). Likewise, our result of 50 genotypes showed reduction in both amylose and grain filling under drought conditions (Table 1 Fig. 3). Further, the ratio of amylose/amylopectin remained same in our studies, and similar results were reported in the previous studies (Lu et al. 2019; Li et al. 2023); however, we did not find any association of grain yield or grain filling with amylose contents indicating our hypothesis is non-conclusive with our findings. Moreover, we found that genotype with highest grain yield under drought was derived from landraces, and genotype with highest amylose contents was also derived from landraces (Table S1 & S2). Moreover, 14 lines derived from landraces and synthetic bread wheat derivatives were having higher amylose contents than the better check Baj (Table S2). To add to it, we suggest that simultaneous improvement of both traits is complicated.

Conclusion

This study highlights that drought at flowering stage not only has impact on grain yield but also impacted the amylose and amylopectin. There was no clear association of amylose and amylopectin with yield-related traits, but amylose contents were enhanced under drought conditions. Further, we also identified that genetic diversity derived from landraces, and synthetic derivatives can improve the grain yield or starch contents. Five lines were having better grain yield than the better check, and 14 lines were having better amylose contents then the better check were identified. However, better grain yield and starch contents were not present in the same genotype indicating that simultaneous improvement of both traits is elusive and our hypothesis of simultaneous improvement of grain yield as well as starch contents together was no proved. Moreover, these genotypes can be used in future breeding program to improve grain yield or starch contents of wheat genotypes grown under drought conditions.