Exploiting Morphophysiological Traits for Yield Improvement in Upland Cotton under Salt Stress

ABSTRACT

The increased land salinization threatens land productivity, food security, and economic losses. The study used a comprehensive set of morpho-physiological, biochemical, and fiber quality parameters to examine the genetic variability of 24 cotton genotypes against 15 dSm-1 salt stress. The general linear model (GLM) effect revealed significant effects of salinity for studied accessions except for lint percentage and fiber strength. The genotype × treatment effects were also significant for all studied traits, while non-significant effects were observed for seed number per boll (SNB), potassium to sodium ratio (K+/Na+), K+, peroxidase (POD), and catalase (CAT). A notable reduction for all traits was observed except for fiber fineness, superoxide (SOD), CAT, POD, carotenoid contents, and hydrogen peroxide (H₂O₂), which were increased under saline conditions. Based on multivariate analyses, hybrids viz. MS-71× KAHKASHAN, followed by MS-71× CRS-2007 and NS-131× CRS-2007, performed well under both normal stressed conditions. Moreover, biochemical and agronomical traits in PCA validate that the MS-71× KAHKASHAN is the most desirable genotype under both conditions. Better hybrid performance under normal and 15 dSm-1 salt stress conditions supports the hybrid adaptability under salinity stress environments. The outcome would assist breeders in developing salt-tolerant cotton varieties under climate change scenarios.

摘要

日益严重的土地盐碱化威胁着土地生产力、粮食安全和经济损失. 本研究使用一套全面的形态生理、生化和纤维品质参数来检测24个棉花基因型对15dSm-1盐胁迫的遗传变异. 一般线性模型（GLM）效应显示，除皮棉百分比和纤维强度外，盐度对所研究材料的影响显著. 基因型×处理对所有研究性状的影响也很显著，而对每铃种子数（SNB）、钾钠比（K+/Na+）、K+、过氧化物酶（POD）和过氧化氢酶（CAT）的影响不显著. 除纤维细度、超氧化物歧化酶（SOD）、CAT、POD、类胡萝卜素含量和过氧化氢（H2O2）在生理盐水条件下增加外，所有性状均显著降低. 基于多变量分析，杂交种MS-71×KAHKASHAN、MS-71×CRS-2007和NS-131×CRS-207在两种正常应力条件下都表现良好. 此外，主成分分析中的生化和农艺性状验证了MS-71×KAHKASHAN是两种条件下最理想的基因型. 在正常和15dSm-1盐胁迫条件下更好的杂交性能支持了杂交在盐胁迫环境下的适应性. 这一结果将有助于育种家在气候变化情景下开发耐盐棉花品种.

KEYWORDS:

• Salt tolerance
• productivity enhancement
• fiber quality
• biochemical traits
• multivariate analysis
• variety development

关键词:

• 耐盐性
• 提高生产力
• 纤维质量
• 生化性状
• 多元分析
• 品种开发

Introduction

Extreme weather events increased the threats of unsustainable agricultural productivity. In order to cater global needs of the fast-growing population, their changing diet patterns and increasing biofuel consumption require doubled crop productivity by the year 2050 (Ray et al. 2013). However, biotic stresses, i.e., disease and insects, and abiotic stresses, i.e., salinity, drought, extreme temperatures, waterlogging, flooding, and many others, are limiting factors to meeting the pace of global food security (Chaudhry UF et al. 2022; Nawaz et al. 2023). Among the stresses mentioned above, salinity is responsible for significant productivity threats to crops as most crop plants cannot grow well across high salt concentrations. It is difficult to reclaim under soil salinity, resulting in the continuous decline in crop productivity over the last decades (Razzaq et al. 2020; Sultana et al. 2023). Different studies have reported that globally 1125 million ha are salt-affected (Hossain 2019). It is proposed that almost half of the global arable land will be declared sodic by 2050 (Ivushkin et al. 2019). Hence, soil salinization is a potentially significantly higher and constant threat to food security and crop productivity (Corwin 2021; Hossain 2019).

Cotton is an important cash crop and a major source of natural fiber, edible oil and biofuel for the sustainability of many people across developing countries (Hassan A et al. 2022; ZAFAR et al. 2020). Generally, cotton is a moderately salt tolerant crop with a threshold salinity level up to 7.7 dS.m⁻¹ (Nadeem M et al. 2023; Zafar et al. 2021). However, it is believed that salinity impedes cotton growth by employing osmotic stress (Sairam and Tyagi 2004), ionic toxicity and nutrient imbalance (Munns 2010), which disturb the normal physiochemical and other processes leading to poor plant growth and yield (Guo et al. 2022). At the reproductive stages of cotton, salt stress results in decreased fruit-bearing positions and boll weight, bringing down the yield with inferior fiber quality (Farooq et al. 2020). Salinity has complicated and detrimental impacts on plants’ growth and physiology. The higher salinity level and enormous soluble Cl⁻ and Na⁺ increased soil solution osmotic potential, which prevents roots from absorbing water and causes deficit stress in cotton (Ma et al. 2021). The stomata of cotton plants closed to reduce the transpiration rate to maintain water balance. Salinity decreases photosynthesis by restricting CO₂ diffusion into the leaves (Abdelraheem et al. 2019).

Salinity causes increased oxidative stress in cotton, which is brought on by an increase in reactive oxygen species (ROS) (Czégény et al. 2014). ROS production damages the plasma membrane and the oxidation of carbohydrates, proteins, lipids and even DNA (Farooq et al. 2021). Due to their high reactivity, ROS have the potential to significantly interfere with normal metabolism by oxidizing lipids, proteins, and nucleic acids (Zafar et al. 2022). In addition to controlling ion and water flow, salt-tolerant plants ought to have an improved antioxidant system for effectively removing ROS (Farooq et al. 2021). For mitigation of these increased oxidative stress resulting from ROS, plants produce antioxidant enzymes like peroxidases (POD), catalase (CAT), superoxidase (SOD) and non-enzymatic antioxidant carotenoids, tocopherol, flavonoids, and ascorbate which are responsible for enhancing plant salt tolerance (Manan et al. 2022; Shokat et al. 2020). The higher levels of POD, CAT, SOD, TSP and carotenoids in plants are mainly related to salt tolerance (Zafar et al. 2021).

To sustain crops across such stressful situations, an ample understanding of crop responses against salinity stress and identifying genotypes that can tolerate stress is necessary. Moreover, the adverse impacts of salinity compel the plant breeders to develop cotton genotypes with higher yields under salt stress. Assessment of significant genetic variation in the breeding material is necessary for developing stress-resilient cotton genotypes. Developing salt-resilient germplasm through conventional breeding using morphological, physiological and biochemical markers could be cost-effective (Singh et al. 2018; ZAFAR et al. 2020). The last two decades have seen a stagnant improvement in crop productivity (Figure S1), while climate change is continuously causing significant losses in productivity.

In conventional breeding programs focusing on cotton improvement, univariate statistical techniques have been extensively used based on the analysis of variance (ANOVA) proposed by steel et al. (Steel 1997). The mean comparisons deal with the variability among test accessions and select any desirable trait or genotype with the desired characteristics (Rahman et al. 2013). However, there were limitations in these methods, especially in the complex nature of traits and stresses and consideration of multiple factors simultaneously to select the best suitable material with desirable characteristics (Ali et al. 2011). Hence, the advancement of different biometrical models, including leaner regression models (Least Square, GLM, MLM) and multivariate analysis techniques, including correlation, PCA and cluster analysis, has made a pace for more sophisticated and efficient selection procedures with the highest reliability as compared to previously used methods (Iqbal et al. 2022; Yeater, Duke, and Riedell 2015).

In this perspective, understanding the complexity related to biotic and abiotic stresses can be better accomplished through multivariate analysis, and it allows one to address the difficult questions to be conceived through univariate approaches (Zafar et al. 2021). Multivariate statistics examined how various variables correlate and explained how their association relates to the studied problem. Multivariate analysis is the most important technique to assess the nature and degree of genetic divergence in the germplasm (Myint et al. 2019). Different statistical analyses like cluster and principal component analysis are used to evaluate yield traits in any breeding program (Lupwayi et al. 2011; Zafar et al. 2021).

The development of salt-resilient cotton genotypes can alleviate the yield losses due to salt stress conditions (ZAFAR et al. 2020). In contrast, these genotypes could be cultivated across marginal saline lands to meet the future increased demands of fiber. Numerous studies reported the salinity impacts at germination and vegetative phases, but the complex interaction of the yield, fiber and biochemical traits at maturity stages has remained unexplored. Hybrid breeding was hypothesized to leverage salt tolerance variation in diverse cotton genotypes. Therefore, the objectives of this study were (1) to develop salt-tolerant cotton cultivars using different morphological and physicochemical characteristics and (2) to determine whether hybrid breeding may leverage salt tolerance variation in diverse cotton genotypes. The outcomes of this research will be helpful in future breeding programs for developing high-yielding cotton cultivars under salt stress.

Materials and methods

During the first year, 8 cotton lines were selected from a screening experiment of 32 cotton genotypes. These 8 cotton genotypes were chosen based on SCY under 15 dSm⁻¹ salt stress conditions. The selected parental material was crossed in line × tester mating fashion in the following season to obtain the subsequent F1 hybrids. The four salt-tolerant genotypes (MS-71, FH-114, IUB-65, NS-131) were used as lines and four salt susceptible genotypes (FH-312, CIM-573) as testers. In the next growing season, the 24 genotypes (Table S1), including (8 lines and their 16F1 hybrids), were grown in containers (containing normal soils, 1.6 dSm⁻¹) with three replications under RCBD. The experiment was executed in the research area of the PBG UAF, Pakistan. After emergence, 10 healthy seedlings were maintained for each genotype in each replication and salt stress was applied to all the seedlings in the containers. Salt stress was applied through the irrigation water with known electrical conductivity (EC) following a previously reported procedure (Farooq et al. 2020; Hariadi et al. 2011; Shabala et al. 2012). This method involves gradually increasing the salinity to a target level of ~15 dS m⁻¹ in two distinct phases to avert seedling injury and ensure survival. Sodium chloride (NaCl) was calculated and added to the groundwater using the US Salinity Lab formula (Richards 1954).

$\mathrm{Amount}\mathrm{of}\mathrm{sodium}\mathrm{chloride}\left(\mathrm{g}/\mathrm{kg}\right)\mathit{\hspace{0.17em}}=\frac{\mathit{Equiv}.\mathit{Wt}\mathit{of}\mathit{NaCl}\times \mathit{Saturation}\mathit{Percentage}}{100\times 1000}\mathit{X}\mathit{TSS}$

In this equation, TSS (Total Soluble Salts) represents the difference in EC (desired EC – initial EC) multiplied by 12.88 and the saturation percentage of the soil was established at 43%.

During the initial phase, the 2-week-old seedlings were subjected to a salinity level of 7.2 dS m⁻¹, which necessitated the addition of approximately 17.32 g of NaCl to each 10 kg pot. Subsequently, in the second phase, the 4-week-old seedlings were subjected to an escalated salinity level of 15 dS m − 1, which required a further addition of about 23.45 g of NaCl per 10 kg pot. Thus, a total of approximately 40.77 g (17.32 g +23.45 g) of NaCl was added per 10 kg pot in the two phases to attain the desired salinity level. After reaching this level, no further salt was added. Following this precise modulation of salinity, standard agricultural procedures were then followed for irrigation. This careful approach enabled us to achieve gradual increments in salinity while preventing seedling injury and ensuring survival throughout the experiment. The characteristics of foil used in this study is given in table S1A. Metrological data has been added as Table S2 and S3. At maturity (140 days of growing), the data were recorded for the following traits like plant height (cm), boll weight (g), number of bolls per plant (NBP), seed cotton yield (g), lint percentage (%), seed index (SI), lint index (LI), lint mass per boll (LMB), seed mass per boll (SMB), fiber strength, fiber length, and fiber fineness from five selected plants from each replication of both controlled and saline treatment.

Fiber quality characters

All the collected seed cotton samples were weighed, and ginning was performed using a single roller ginning machine. Seed and lint samples after ginning were weighed separately and lint percentage was calculated by dividing the sample’s lint weight by the sample’s total weight (seed cotton weight), represented in percentage. A representative lint sample has been prepared and sent to analyze fiber fineness, strength, and length with the help of high-volume instruments (HVI-900, USTER, USA) at the Fiber and Textile Technology Department UAF-Pakistan.

Biochemical analysis

When plants reached vegetative maturity, fresh green leaves were used for the sodium and potassium analysis at noon. These leaf samples were dried in hot air, ground down, and then subjected to a 2:1 molar ratio of nitric acid and sulfuric acid digestion on a hot plate. After digestion, the substance was cooled to room temperature, and a flame photometer was used to record the results (410 Flame Photometer). By dividing the potassium concentration by the sodium concentration, the K+/Na+ ratio was calculated.

A hydrogen peroxide (H₂O₂) measurement was made using Bernt and Bergmeyer’s method (Allen, Farmer, and Sohal 1983). The amount of H₂O₂ was determined by obtaining 0.5 g of leaf samples from each control and treatment group and homogenizing the sample using liquid nitrogen. The powder was then dissolved in 1.5 ml of 100 mM potassium phosphate buffer (pH 6.8). Then, suspensions were centrifuged (Refrigerated SIGMA 2-16Kl centrifuge, UK) at 18,000× g for 20 min at 40°C. The enzymatic reaction was started with 0.25 ml of supernatant and 1.25 ml of peroxidase reagent made up of 83 mM potassium phosphate buffer (pH 7.0), 0.005% (w/v) O-dianizidine, and 40 μg peroxidase/ml at 30°C. After 10 minutes, the reaction was stopped by adding 0.25 ml of 1 N perchloric acid and centrifuging the mixture for 5 minutes at 5,000× g. The absorbance was calculated at 436 nm using a spectrophotometer (NanoDrop^TM 8000 Spectrophotometer from Thermo Fisher Scientific, Sweden), and the amount of H2O2 was determined using an extinction coefficient of 39.4 mM⁻¹ cm⁻¹ (Zhang et al. 2014). SOD was measured in terms of units of the enzyme that inhibited the photochemical reduction of nitro blue tetrazolium (NBT). The reaction mixture contained potassium phosphate buffer (pH 5), 800 μl of distilled water, 100 μl of enzyme extract, 200 μl of methionine, 200 μl of Triton X, and 100 μl of NBT. The resulting solution was then exposed to UV light for 15 minutes and then 100 μl of riboflavin was added. On a spectrophotometer, values at 560 nm absorbance were taken. POD values were computed using the quantity of guaiacol-oxidizing enzyme units. The enzymes extracted for SOD were also used to determine POD activity. In an Eppendorf tube, the reaction mixture contained 800 μl of potassium phosphate buffer (pH 5), 100 μl of H2O2 (40 mM), 100 μl of guaiacol (20 mM), and 100 μl of enzyme extract and the readings were taken using a spectrophotometer at a 470 nm wavelength (Liu et al. 2009). By centrifuging and vortex of leaf tissue extracted in phosphate buffer (pH 4), total soluble proteins (TSP) were calculated. A 40 μl aliquot was blended with 160 μl of Bradford reagent from the same enzyme extract. The mixture was added to an ELISA plate and readings from a spectrophotometer at 595 nm absorbance were taken (Bradford 1976). CAT was observed as the amount of H₂O₂ consumed by catalase and changed into H₂O and O₂. Absorbance reading were taken at 240 nm (Liu et al. 2009).

Chlorophyll contents and carotenoids assay

0.5 g of cotton leaf sample was crushed in 8–10 mL of 80% acetone (v/v) and then homogenization was carried out through filter paper and absorbance of the final solution was taken at 645 and 663 nm (Arnon 1949). The chlorophyll a, chlorophyll b and carotenoids were quantified as under.

$\mathbf{Chlorophyll}\mathbf{a}\left(\frac{\mathit{\mu }\mathbf{g}}{\mathbf{g}}\mathbf{FW}\right)\mathbf{=}\left[\mathbf{12.7 }\left(\mathbf{OD}\mathbf{\text{ 663}}\right)\mathbf{-}\mathbf{2.69 }\left(\mathbf{OD}\mathbf{\text{ 645}}\right)\right]\times \frac{\mathbf{v}}{\mathbf{1000}\times \mathbf{w}}$

$\mathrm{Chlorophyll}\mathrm{b}\left(\frac{\mathit{\mu }\mathrm{g}}{\mathrm{g}}\mathrm{FW}\right)=\left[22.9\mathit{\hspace{0.17em}}\left(\mathrm{OD}665\right)-\mathit{\hspace{0.17em}}4.48\mathit{\hspace{0.17em}}\left(\mathrm{OD}663\right)\right]\mathit{\hspace{0.17em}}\times \frac{\mathrm{V}}{1000\times \mathrm{w}}$

$\mathrm{Carotenoids}\left(\frac{\mathit{\mu }\mathrm{g}}{\mathrm{g}}\mathrm{FW}\right)=\frac{{\mathit{A}}^{\mathrm{car}}}{\mathrm{Em}}\times 1000$

${\mathbf{A}}^{\mathbf{car}}\mathbf{=}\mathbf{OD}\mathbf{480}\mathbf{+}\mathbf{0.114}\left(\mathbf{OD}\mathbf{663}\right)\mathbf{-}\mathbf{0.638}\left(\mathbf{OD}\mathbf{645}\right)$

where,

W = weight of leaf sample, V = volume of sample, Em = 2500

Statistical analysis

The collected data were subjected to a Generalized Linear Model (GLM) for all the variables under study, considering the replication genotypes and treatments. A generalized linear effect test was carried out to find the possible significant linear effects among replications and estimates of genotypes, treatments, and their complex interactions. The summary of estimated effects has been presented for comparison. After the estimates have been found to have sufficient variation existing through estimate tests and the replications shown non-significant estimates, the data have been pooled for replications. The mean data for genotypes and treatments were used for multivariate analyses, including correlation, PCA, and cluster analyses. Pairwise correlations have been calculated using the person approach using SAS-JMP Pro 16 (SAS Institute Inc., Cary, NC, USA, 1989–2011). The means from the genotype by treatment analysis were used for correlation-based PCA using the default function for standardization/scaling by JMP software: SAS-JMP Pro 16 (SAS Institute Inc., Cary, NC, USA, 1989–2011). Cluster analysis was performed using the Hierarchical clustering approach with two-way clustering to generate a tree diagram based on Euclidean distances by Ward’s method.

Results

From screening experiments, eight cotton genotypes, including four salt-tolerant (MS-71, FH-114, IUB-65, NS-131) as lines and four testers viz. Kahkashan, CRS-2007, FH-312, and CIM-573 (salt susceptible genotypes) were selected based on analysis of mean methods (ANOM)-decision chart. Figure 1 shows that the mean SCY of four salt-tolerant genotypes was lying above the UDL under both saline and normal environment. In contrast, four salt-sensitive genotypes revealed a significant decline in SCY in saline conditions compared to normal conditions.

Figure 1. ANOM-decision chart with decision limits 38.72 to 50.84 for seed cotton yields across normal and salt stress (α > 0.05%). It provides a graphical test for simultaneously comparing the mean performance of these 32 cotton genotypes across normal and salt stress. Red-colored heads represent a significant deviation from the mean, either above the upper-decision level (UDL) or below the lower decision level (LDL).

The impacts of salinity on physio-morphological and fiber quality parameters of parents and their hybrids were observed by GLM ANOVA. The ANOVA was constructed based on generalized linear model (GLM) effect tests. All parents and their hybrids revealed significant differences in morphological, biochemical, and agronomic traits (Table 1).

Table 1. Generalized linear Model effect test ANOVA for genotypes under study across normal and salt stress conditions.

Source		Genotype	Treatment	Genotype × Treatment
Nparm		23	1	23
DFNum		23	1	23
PH	F Ratio Prob > F	33.2517 <.0001*	375.9295 <.0001*	10.3561 <.0001*
NBP	F Ratio Prob > F	7.7883 <.0001*	98.0490 <.0001*	1.7076 0.0382*
BW	F Ratio Prob > F	7.7105 <.0001*	97.0698 <.0001*	1.6906 0.0413*
SCY	F Ratio Prob > F	13.3863 <.0001*	649.6384 <.0001*	10.7088 <.0001*
LP	F Ratio Prob > F	4.2805 <.0001*	0.9911 0.322	2.9322 <.0001*
SI	F Ratio Prob > F	14.1159 <.0001*	51.5221 <.0001*	2.4387 0.0014*
LI	F Ratio Prob > F	2.7871 0.0003*	2.1185 0.1488	2.2401 0.0035*
SNB	F Ratio Prob > F	1.4242 0.1200	3.5785 0.0616	0.7609 0.7616
SMB	F Ratio Prob > F	11.1426 <.0001*	55.8777 <.0001*	1.0563 <.0035*
LMB	F Ratio Prob > F	11.7764 <.0001*	49.9426 <.0001*	1.0876 0.3732*
SV	F Ratio Prob > F	3.3181 <.0001*	7.2662 0.0083*	2.4122 0.0016*
FF	F Ratio Prob > F	9.8892 <.0001*	159.2814 <.0001*	1.7660 0.0299*
FS	F Ratio Prob > F	9.3028 <.0001*	1.2313 0.2700	3.8693 <.0001*
FL	F Ratio Prob > F	6.7657 <.0001*	42.9454 <.0001*	2.6955 0.0004*
K^+/Na^+	F Ratio Prob > F	6.3672 <.0001*	238.0626 <.0001*	1.1056 0.3545
Na^+	F Ratio Prob > F	7.0031 <.0001*	435.0301 <.0001*	1.8358 <.0219*
K^+	F Ratio Prob > F	27.9945 <.0001*	102.8785 <.0001*	1.3522 0.1567
H2O2	F Ratio Prob > F	6.0763 <.0001*	256.3888 <.0001*	2.5443 0.0008*
SOD	F Ratio Prob > F	4.8637 <.0001*	3085.477 <.0001*	4.1453 <.0001*
POD	F Ratio Prob > F	4.9580 <.0001*	1.019 0.031*	0.332 0.998
CAT	F Ratio Prob > F	9.4090 <.0001*	137.8633 <.0001*	0.7787 0.7487
TSP	F Ratio Prob > F	13.869 <.0001*	21.225 <.0001*	1.904 0.0162*
Chla	F Ratio Prob > F	9.9930 <.0001*	170.7653 <.0001*	6.3405 <.0001*
Chlb	F Ratio Prob > F	3.6000 <.0001*	156.3892 <.0001*	2.4787 <.0001*
Car.	F Ratio Prob > F	7.5131 <.0001*	3459.639 <.0001*	4.7220 <.0001*

*Significance (p < .05); PH = Plant Height (cm), NBP= number of boll per plant, BW= Boll weight (g), SCY= Seed cotton yield (g), LP= lint (%), LI= Lint index (g), SI= seed index (g), SNB= seed number per boll, SMB= seed mass per boll (g), LMB= lint mass per boll (g), SV= seed volume (cm2), FS= Fiber strength (g/tex), FL=Fiber length (mm), FF= Fiber fineness (ug inch − 1), Na+, sodium concentration (mg g−1); K+, potassium concentration (mg g−1); K+/Na+, potassium to sodium ratio, H2O2 = Hydrogen peroxide (μmol/g), CAT = Catalase (U mg-1 protein), POD = Peroxidase (U mg-1 protein), SOD=Super-oxidase dismutase (U mg−1 protein), TSP = Total soluble Proteins (mg g-1 FW), Chl a&b= Chlorophyll contents a & b (mg g−1 FW), Caro. = Carotenoid (mg g−1 FW).

The treatment effects (salinity effects) were highly significant for all characters except lint% (LP) and fiber strength (FS). The genotype × treatment effects were highly significant for all traits, while non-significant effects were observed for SNB, LMB, K+/Na+, POD and CAT. All pairwise comparison using Tukey´s HSD-test is presented in (Figures 2a & 2b and table S4). The figure clearly showed that all traits were significantly affected by salt stress in studied genotypes except SNB. The analysis of mean methods (ANOM)-decision chart was also used to compare the performance of each genotype under normal and stress for studied traits. These charts graphically represent the variation in breeding material for studied characters across both environments (Fig S2A & B and Table S5). The response grid sliders were obtained to assess the variation of each character under normal and saline conditions. The prediction profiler graph allowed us to observe the response of all genotypes for studied characters under saline and normal conditions. It also predicts the desired genotype that performed well based on studied characters under both conditions. The response grid sliders and prediction profiler showed the reduction in plant height (PH), boll weight (g), seed cotton yield (SCY), LMB, SMB, SI, LI, fiber strength (FS), fiber length (FL), K+/Na+, and total soluble proteins (TSP) under salt stress while the fiber fineness (FF), H₂O₂, Na+, catalase (CAT), and POD were increased under saline conditions (Figure S3A & B). Interestingly, the lint% of IUB-65× CIM-573 was increased under salt stress, while most of the genotypes revealed a reduction in lint%. The prediction profiler plot also revealed that the MS-71× KAHKASHAN exhibited higher antioxidant enzymatic activities under salt stress and performed superior for yield and fiber quality parameters in both environments (Figure 3).

Figure 2a. A & b. All pairwise comparisons for biochemical, yield and fiber-related traits under normal and stress conditions, red color bars represent significant comparisons whereas blue color bars indicate non-significant differences in comparisons for traits under study.

Figure 2b. Continued.

Figure 3. Prediction profile plot with highest adjusted desirability and factor values for all studied agronomic, biochemical and fiber quality traits of 24 cotton genotypes under normal and salt stress conditions.

Pairwise correlation

Pairwise correlation revealed a positive relationship between seed cotton yield and NBP with BW, POD, CAT, TSP, K⁺/Na⁺ and chlorophyll a & b under normal and salt stress environments. Interestingly, SCY and NBP under both conditions also revealed a significant negative association with H₂O₂, Na⁺, and carotenoids (Figure 4). Under both conditions, BW exhibited a positive relationship with all within boll yield components and antioxidant traits under both conditions. The trait seed index is negatively associated with LI, SOD, CAT, Na⁺, and carotenoids under normal and saline conditions. Under both environments, the trait SMB revealed a significant positive relationship with SCY, BW, LMB, K⁺/Na⁺ and LI and a negative association with H2O2, Na⁺, and carotenoids. The positive relationship of LP was observed with SCY, BW, NBP, SOD, POD, CAT, Car, TSP, FS, FL, and FF, whereas it was negatively associated with seed index under both environments (Figure 4). The fiber strength and length were positively related to SCY, BW, TSP, K⁺/Na⁺ and chlorophyll contents (a & b) under both conditions. Interestingly, FF revealed a strong positive association with Na+ and a negative relationship with K⁺/Na⁺ and Chlb. Under both conditions, K⁺ showed a significant positive association with SCY, BW, TSP, and chlorophyll contents (a & b). The characters POD, CAT, and TSP were significantly and positively correlated with PH, SCY, BW and NBP under normal and saline environments. The chlorophyll a & b were positively related to all agronomic and fiber quality traits under both conditions (Figure 4).

Figure 4. Scatterplot matrix to visualize several attributes by pairwise dependencies of different traits under study across normal and heat stress conditions. The upper triangle matrix represents correlations among biochemical, yield and fiber-related traits under normal and stress conditions. The lower triangle matrix reveals bivariate density distribution with ellipses between each pair of attributes. The legends at the top right corner of the color gradient (red to blue), and the size of circles show the amount of correlation and log (p) values for the significance threshold, respectively.

Cluster analysis

The hierarchical clustering was executed to classify the 24 cotton genotypes based on their genetic potential for agronomic, fiber and biochemical characters under normal and saline conditions. This study initially considered grouping genotypes in three groups by AHMC clustering, not stating them as susceptible or tolerant genotypes by software but just grouping or clustering them. Another clustering of variables to group traits according to their trend was added along with a heatmap attached further to the two-way clustering to visualize a clear trend of all observations to extract supplementary conclusions from this tree concerning this experiment (Figure 5). The heatmap was constructed for studied traits in which blue and red boxes revealed negative and positive associations, respectively, with increasing color strength reflecting a higher coefficient. In control conditions, 24 genotypes were grouped into 3 clusters. Groups 1, 2 and 3 consist of 6, 6 and 12 genotypes, respectively (Figure 5). The heat map showed that the genotypes MS-71× KAHKASHAN followed by NS-131× FH-312 exhibited higher antioxidant enzymatic activities and performed superior for yield and fiber-related parameters under a normal environment. Under salt stress conditions, 24 genotypes were classified into 2 clusters. Groups 1 and 2 contained 21 and 3 genotypes, respectively (Figure 5). Under both stressed and normal environments, various clusters were depicted in different colors. Based on studied traits, the genotypes of cluster 3 were declared as the most salt-tolerant genotypes. Depending on the pattern of clusters, the genotype MS-71× KAHKASHAN followed by NS-131× CRS-2007 and MS-71× CRS-2007 performed well in yield, physiological and fiber-related parameters under both environments. With these traits mentioned above, three better hybrid genotypes showed an increasing trend of phenotypic values under both environments. The three hybrid genotypes got a separate cluster in the hierarchy due to a similar trend of their studied traits under stress conditions. The hybrid genotype MS-71× KAHKASHAN showed its unchanged behavior regarding superior performance based on studied traits under stress. The two-hybrid genotypes viz; NS-131× CRS-2007 and MS-71× CRS-2007 shuffled their cluster by performing better, just like the hybrid MS-71× KAHKASHAN (Figure 5). The physiological, morphological, and yield-related traits changed their performance under stress conditions, due to which the remaining studied genotypes shuffled their clusters in stress conditions. However, regardless of their tolerance, the genotypes in similar clusters got the same cluster as their fellow genotypes under stress conditions, supporting our selection criteria. Under both environments, the genotypes in obtained clusters are given in (Table S6).

Figure 5. Agglomerative hierarchical clustering (AHC) calculates the Euclidean distance matrix of 24 cotton genotypes for biochemical, yield and fiber-related traits under normal and stress conditions. The plot was constructed in JMP pro. V. 16 (SAS Institute Inc., Cary, NC, USA) using Ward’s minimum variance on standardized data. The horizontal and vertical axes represent clades formation based on the division of traits and genotypes, respectively, following the two-way clustering approach.

Principal component analysis (PCA)

The principal component analysis (PCA) was carried out to extract information regarding genotypic performance based on agronomical, biochemical and fiber quality traits under normal and saline conditions. The genetic variability among studied genotypes was explained based on the association among biochemical and agronomic characters and extracted clusters.

The whole genetic variability was split into 19 principal components (PCs), out of which only the first 5 PCs exhibited >1 eigenvalue (Figure 6). These five PCs contributed 72.99% to the total diversity of cotton genotypes assessed for physiological, agronomic and fiber quality parameters under both environments. While the rest of all PCs contributed 27.01% of the whole variation. The PC1, PC2, PC3, PC4 and PC5 revealed 33.57%, 18.13%, 9.04%, 6.39% and 5.83% variation among the understudied characters’ genotypes. The characters, including PH, SCY, Chlb, Chlb, PH and BW, showed considerable positive factor loadings on PCI. The PC-II was characterized by H₂O₂, FF, CAT, Na+, BW, and POD. The PC-III was explained by variation among genotypes for LMB, SMB, SI, and LI (Table S7).

Figure 6. Summary plots with (leftmost) biplot between PC1 and PC2 displaying the distribution of traits; the biplot (center) of 24 cotton genotypes under normal and salt stress conditions; (right), scree plot showing the number of components to be considered for variability coverage through PCA.

In the biplot, variables were portrayed in vector form and the relative distance covered by variables from the origin regarding PC1 and PC2 represents the contribution of variables to the total variation of the genetic materials. The distance covered from the origin to the tip of the plot provides knowledge regarding the diversity among the genotypes. The biplot of PC1 and PC2 explained 51.7% of the total variation. The PH, SCY, Chlb, Chlb, BW, NBP, K+/Na+, and FL had long vectors and higher correlation, contributing more to this biplot variation, while LI, FS and LP showed the least variability. Interestingly, the PCA biplot revealed that these traits were strongly associated with genotypes of cluster 3. The H₂O₂, FF and CAT were lied toward the direction of PC2 and were highly correlated. The scatterplot matrix of PC1 and PC3 exhibited a lower variation (42.63%) than the biplot of PC2. This biplot was explained by LMB, SMB, SI, and LI and was highly correlated. The biplot of PC1 and PC4 revealed lower diversity (39.96%) compared to PC2 and PC3, respectively (Figure 6 and Table S7). The character LI, SI, FS and Chla are major contributing factors in this biplot. The biplot of PC1 and PC5 showed the least variability than other PCs mentioned above; major elaborating factors were LP, LI and SV (Figure 7 and Table S7).

Figure 7. Scatterplot matrices of PC1, PC2 and PC3 displays different traits and genotypes under normal and salt stress conditions. Different color dots across coordinates of scatter plots show the placement of genotypes under salinity stress, whereas different color shapes in scatterplots represent placement genotypes under normal conditions. Different color schemes represent the grouping of genotypes across stress and normal conditions.

The scatterplot matrix of PC2 and PC3 showed the least variation (27.18%), and LI, SV, SMB, and SI are the most discriminating factors in this biplot (Figure 7). The biplot of PC2 and PC4 was explained by LI, FS and Chla and revealed significant positive associations among themselves. The scatterplot matrix of PC2 and PC5 was characterized by LP, LI, SV and SNPB and have a positive relationship among themselves. The characters Chla, Chlb and FS were more discriminating factors in the biplot of PC3 and PC4. The biplot of PC4 and PC5 is explained by LP and SV and has the least variability of all other PCs (Figure 7 and Table S7). These biplots revealed that the MS-71× Kahkashan was the most superior genotype for studied traits, followed by MS-71× CRS-2007 and NS-131× CRS-2007 under both environments. The genotypes FH-114, IUB-65, and FH-114 × CRS-2007 were not in a desirable direction and were highly susceptible genotypes under saline conditions (Figure 7).

The first three components of PCA contributed 60.75% variability to the total variation. The yield, biochemical and fiber quality parameters are depicted by factor map squared cosines (Figure 8). It is deduced that more values of squared cosines revealed a fair share of the specific variable. The PC1 covered PH, NBP, SCY, FL, K+/Na+, Na+, K+, SOD, Chla, Chlb, and Car. The traits BW, LP, SNPB, FF, FS, H₂O₂, POD, CAT and TSP were covered by PC2. The trait SI, LI, SMB, LMB and SV were covered by PC3 (Figure 8). The biplot of studied traits of cotton revealed three forms of groups of the traits. The on group consists of PH, NBP, SCY, FL, K+/Na+, Na+, K+, SOD, Chla, Chlb, and Car. The 2nd group comprised BW, LP, SNPB, FF, FS, H₂O₂, POD, CAT and TSP. The 3^rd group comprised SI, LI, SMB, LMB and SV. The PC-biplot scattered genotypes to understand the genetic potential of studied genotypes against salt stress with agronomical, fiber and biochemical traits. Grouping genotypes across biplot indicated and validated the clustering results, making the decision more confident regarding classifying genotypes in three groups. The genotypes of cluster 3 revealed the maximum value of agronomical, biochemical and fiber quality traits and were considered salt tolerant.

Figure 8. Squared cosines are associated with the principal components for the studied traits under normal and salt stress conditions.

Discussion

To develop salt-tolerant cultivars, initially, diverse genotypes are evaluated in a breeding program to increase their utility. Salt resistance is a complex character caused by many interrelated mechanisms of morphological and biochemical characters (Shelke et al. 2017; ZAFAR et al. 2020). Different studies have assessed different cotton genotypes using physiological, morphological, and agronomic traits. These traits have proven effective in differentiating between salt-tolerant and sensitive genotypes when exposed to salinity stress conditions (Nawaz et al. 2023). These characteristics are closely associated with coping with the adverse impacts of soil salinity on plant growth and development. In the present study, we examined different agronomical, biochemical and fiber quality characters to investigate the salt resistance ability of 24 cotton genotypes. All genotypes behaved significantly differently for all studied characters under both conditions, revealing the presence of variation for all measured traits. It also suggests that the panel of genotypes used in this study has been selected appropriately (Kumar et al. 2021). Cotton improvement against salt stress depends on the genetic variation in studied germplasm for traits involved in salt tolerance. The studied genotypes exhibited a great magnitude of variation for salinity tolerance and performed differentially under saline conditions. Under salt stress, grid slider plots and prediction profiler showed a reduction in plant height (ZAFAR et al. 2020), boll weight (Ibrahim et al. 2019), seed cotton yield (Hassan et al. 2020), K⁺/Na⁺, K⁺, chlorophyll contents, total soluble proteins, SMB, LMB, SV, fiber length and strength. The reduction in these characteristics is due to ionic imbalance (Zafar et al. 2021), osmotic stress and nutrient deficiency (Munns 2010), which disturb the normal biochemical processes that lead to poor plant growth and ultimately lower the yield (ZAFAR et al. 2020). The reduced boll weight is attributed to higher sodium and chloride ions accumulations inside the cell (Akhtar et al. 2010). It is also possible that the reduced boll weight in cotton genotypes’ was caused by the salt-induced shrinkage of compartments used to store undesirable harmful materials like vacuoles (Ju et al. 2021). Under salt stress, the decline in seed cotton yield and lint% is associated with decreased boll weight (Sharif et al. 2019). Fiber quality parameters significantly declined with increased salt stress (Zhang et al. 2013).

The K⁺/Na⁺ ratio is a fundamental criterion for selecting salt-tolerant cotton genotypes. The salt-resistant genotypes have a higher K⁺/Na⁺ ratio than susceptible plants with increased amounts of Na⁺ inside their cells (Joshi et al. 2022). The lower K⁺/Na⁺ ratio is related to the high mobility of Na+ inside the cell due to cell injury and the ineffective method for excluding Na⁺ outside the cell under saline environments (Amin et al. 2021). The ANOVA results presented significant differences among genotypes regarding CAT, POD, Car, chlorophyll contents and H₂O₂ under both environments at a P > .01 level. The hydrogen peroxide contents, CAT, POD, and fiber fineness increased under saline environments. These outcomes are consistent with the outcomes of (Farooq et al. 2020). Salt stress affects the levels of carotenoid and chlorophyll contents as well as critical photosynthetic enzymes. Measuring a plant’s chlorophyll contents is a reliable way to determine its susceptibility to salt stress (Ibrahim et al. 2019). Salt stress inhibits the photosynthesis process by higher values of ROS, which accelerates oxygen-induced cellular damage (Joseph, Jini, and Sujatha 2011). Therefore, a plant’s ability to regulate its antioxidant defense system, which involves a variety of antioxidant enzymes like like CAT, POD, SOD and TSP, determines its ability to tolerate salt stress (Wang et al. 2017).

SOD is a key antioxidant enzyme and regulates O²⁻ and H₂O₂ concentration. The increased SOD activities were observed under salt stress in tolerant cotton genotypes. During the ROS scavenging process, POD and CAT played a vital protective role in the presence of SOD. The existence of high peroxidase (POD) enzymes detoxifies H₂O₂ inside the cytosol and chloroplasts of the plant cells (Kusvuran et al. 2016). Catalase is also involved in converting harmful H₂O₂ into water and oxygen. It was observed that the activity of POD and CAT was increased at 15 dS m⁻¹ NaCl stress (Wu et al. 2014). This increase in H₂O₂ permeates the paroxysmal membrane, disrupting the balance of total soluble proteins by further increasing the cytosolic H₂O₂ concentration. Salinity in cotton plants increased the accumulation of H₂O₂ through enzymatic and non-enzymatic pathways. H₂O₂ induces the generation of other antioxidants inside the plant cell, e.g., APX. Thus, H₂O₂ is considered a signal for the plant to prepare itself against the onset of stress (Locato et al. 2008). The activities of POD and CAT are effective indicators for assessing the salt-tolerant ability of various plant species because these play a major role in the detoxification of H₂O₂ (Zhang et al. 2014). Increased POD activity enhanced photosynthetic activity, revealing the role of antioxidants’ defense mechanism in mitigating salt stress. The POD and CAT were positively associated with SCY, BW, NBP, and chlorophyll contents, confirming their role in increasing photosynthetic activities under salt stress and ultimately increasing the yield under stress conditions.

Correlation analysis revealed a positive relationship of SCY with chlorophyll contents, BW, NBP, POD, CAT, TSP and K⁺/Na⁺. So, selection for these characters will increase the lint yield under both environments. ANOVA findings, genotype performance across normal and saline environments, and correlation between quantitative and biochemical traits provided confidence in moving forward with the current panel of genotypes to develop salt-tolerant cotton cultivars. Cluster analysis was done to investigate the genetic diversity among 24 cotton genotypes for studied traits, which gave directions to breeding programs (Chunthaburee et al. 2016). Hierarchical cluster analysis showed that cultivars of cluster 2 under salt stress revealed a higher ability to tolerate salinity stress as determined by all under-studied characters. The hierarchy of parents regarding their tolerance ability remained intact under stress conditions. Almost all the designated tolerant and susceptible genotypes in a specific cluster got assigned a similar cluster with their fellow genotypes (Zafar et al. 2021). The PCA results of the current study affirmed variations for ionic and agronomic characters in studied genotypes that could be further used in the cotton breeding program against salt stress. This approach divides the genetic variation of the breeding material into various components and exploits a particular character’s efficiency in a breeding program (Mohi-Ud-Din et al. 2021). In PC-biplot, K⁺/Na⁺, TSP, CAT, chlorophyll contents and POD showed a positive relation with PH, NBP, BW, SCY, LP, and fiber quality traits. The CAT, TSP, Chla, Chlb and POD exhibited a positive relationship with LP, FF, FL and FS. It means these biochemical characteristics have a direct impact on agronomic characteristics. These biochemical traits will be helpful for the selection of improved cotton genotypes for yield and fiber quality traits. The PC-biplot of studied traits of cotton revealed that the genotypes of cluster 3 revealed the maximum value of agronomical, biochemical and fiber quality traits and were considered salt tolerant. Our PCA results were the following (Krishnamurthy et al. 2016; Munawar, Hameed, and Khan 2021). In our study first two PCs covered the maximum diversity present in the studied material. It was in line with previous reports (Krishnamurthy et al. 2016; Munawar, Hameed, and Khan 2021).

Our findings revealed that the clustering of cotton genotypes was well interpreted and validated by the results derived from PCA. Altogether, it is evident that 24 cotton genotypes have notable differences under both environments for studied characters and the multivariate analyses could be practical in identifying salt-tolerant cotton genotypes. On clustering and PCA, under normal and salt-stress conditions, MS-71× KAHKASHAN followed by NS-131× CRS-2007 and MS-71× CRS-2007 were superior genotypes for most of the studied traits. These genotypes also revealed an increased TSP, CAT and POD level, positively associated with yield and fiber quality traits. In conclusion, analyzing a suite of physio-morphological characters provides a valuable paradigm evaluation of salt tolerance among studied cotton genotypes. Our findings follow the previously reported studies having similar results (Malik et al. 2020; Munawar, Hameed, and Khan 2021). We have identified key characteristics of salt resistance and potentially useful experimental materials for future work.

Conclusion

The research hypothesis was based on the concept that cotton germplasm could be screened out for its potential against salt stress, and genotypes can be brought forward for breeding programs aiming to improve cotton crops against salt stress. Salinity disturbed the normal physiological, biochemical, and molecular processes that reduced the seed cotton yield and fiber quality. The genotype MS-71× KAHKASHAN performed well under normal and saline conditions and can be effectively utilized in future breeding programs focusing on cotton improvement for salinity tolerance to enhance productivity and quality across climate change widows.

Highlights

• Global climate change disturbing cotton ecosystem

• Rise in temperature exposes cotton roots to high salinity

• Selection of salt tolerant genotypes on the basis of physiochemical attributes is amenable

• Selection of salt tolerant genotypes pave the path of future breeding and marker assisted selection

Statement of permission

The genotypes used in this study were obtained with the permission of Department of Plant Breeding and Genetics, University of Agriculture Faisalabad.

Availability of data and materials

The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request

Acknowledgments

The authors appreciate the Researchers Supporting Project number (RSP2023R403), King Saud University, Riyadh, Saudi Arabia for funding this research. Authors acknowledge the Department of Plant Breeding, University of Agriculture Faisalabad, Pakistan, for providing cotton genotypes used in the current study.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15440478.2023.2282048

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article. The authors appreciate the Researchers Supporting Project number (RSP2023R403), King Saud University, Riyadh, Saudi Arabia for funding this research.