Pattern of genetic inheritance of grain zinc and iron content, agronomic and biochemical traits in bread wheat using mixed linear models under salinity stress

Wheat is the most important staple food that acts as a primary source of dietary calories, protein and most of the bioavailable micronutrients such as iron (Fe) and zinc (Zn) for the world’s population. Understanding genetic control of micronutrients uptake is necessary for development of good quality wheat genotypes. To study the nature of inheritance of Zn and Fe efficient uptake under none-saline and saline conditions, two Iranian facultative wheat variety; Navid (salt sensitive, Fe and Zn-deficient) and Roshan (salt tolerant, Fe and Zn-efficient); were crossed to generate six basic generations for generation mean analysis. All the genotypes of these six generations were evaluated for grain Zn and Fe content, agronomic and biochemical traits under none-saline and saline conditions. For all the studied traits, the non-additive components were greater than the additive component. Additive effects were negative and significant for all traits under non-saline and saline conditions, except for 100-SW in control, MDA in 100 mM salinity level, EL and LeaNaC in all salinity levels. Additive gene actions were important for grain Zn and Fe content; while for rest of traits both fixable and non-fixable genetic effects were important. Duplicate dominant type of epistasis was involved in inheritance of all the traits. Broad-sense heritability values (>0.6) for most traits under non-saline and saline conditions were high, whereas the narrow-sense heritability values for most of the studied traits were low to moderate. These Zn and Fe efficient uptake indices could be used to select Zn/Fe-efficient wheat genotypes from segregating populations.


Introduction
Among the major abiotic stresses, salinity is one of the most adverse environmental factors limiting the growth, survival and production of wheat in Iran and worldwide. Salinity stress generates hyperionic and hyperosmotic stress in wheat, which cause hormonal imbalance, nutrient disturbances and negative impact on yield (Munns and Tester 2008;Daneshbakhsh et al. 213). Iran is a chief producer of wheat in Asia (FAO 2017), challenging for about 34 million ha of salt-affected land area which includes 4.1 million ha of irrigated land (FAO 2000) leading to an annual economic loss of almost US$ 1 billion (Qadir et al. 2008).
By reducing the micronutrients, the effects of salinity stress can accelerated in cereal crops (Faran et al. 2019). In the other words, high salt concentrations in the soil reduces the root growth leading to poor nutrient uptake, especially Zn and Fe, due to strong competition between Zn and Fe and salt actions which may cause severe deficiencies of these elements in the plants (Rehman et al. 2017). Under salt stress, the production of reactive oxygen species (ROS) has increased which causes structural and functional damage on lipids, proteins, nucleic acids and water relations that consequently suppress the photosynthesis, Zn and Fe uptake in plants resulting the poor yield (Munns and Tester 2008; Rehman et al. 2017).
To surmount the aforementioned problem and ensure higher crop production, different approaches such as agronomic biofortification (foliar spraying, soil application and/or seed priming) and genetic advancement of crops have been suggested to develop genotypes which can maintain sufficient micronutrients under saline soils (Peleg et al. 2009;Pfeiffer and McClafferty 2007). Among the potential solutions, the breeding of wheat varieties especially landraces and wild relatives of common wheat has been proposed as one of the most cost-effective and environmentally safe approaches to alleviate the malnutrition (Peleg et al. 2009). The agronomic solutions are being exhausted in case of minimizing the impact of salinity stress, and consequent Zn and Fe deficiency. Therefore, developing salt tolerant genotypes of crops accompanied with higher Zn and Fe acquisition efficiency would be may be beneficial to minimize micronutrient deficiency gaps in farming communities. Breeding for both tolerance to salt and micronutrients (particularly Zn and Fe) deficiency, in spite of its importance, has even rarely considered in wheat breeding programs.
A primary requirement of such a breeding effort are exploring about genetic variation of component indices of Zn and Fe efficiency, including uptake of Zn and Fe by the roots, translocation, assimilation, and remobilization of them into the grains (Liu et al. 2019). According to Velu et al. (2014), efficient uptake of Zn and Fe from the soil can be achieved by modifying the morphological and structural traits of root.
Generation means analysis (GMA), a biometrical method developed by Mather and Jinks (1982), is a simple but useful tool for designing breeding strategies to take the advantage of gene interaction existing in succession breeding generations (Mather and Jinks 1982). In GMA, joint scaling test is commonly used to study the inheritance types of traits, in which the gene effects (additive and dominance) and digenic epistatic interactions (additive x additive; Additive x dominance; Dominance x dominance) are set up in a linear model and evaluated by the chi-square test (Mather and Jinks 1982). However, this method has several limitations that can restrict its utilization. This method cannot be used directly for models in which the number of effects is equal or greater than to the number of generation means. On the other hand it is possible that additive, dominance, and epistatic effects are over-or underestimated, since these effects are obtained without error via the least square analysis method (Balestre et al. 2012). Additionally, the error terms usually computed based on within-plot variances (Mather and Jinks 1971).
In this case, because between-plot variance for the segregating generations is ignored, results may not be very accurate (Piepho and Mohring 2010). Using Mixed linear model developed by Piepho and Mohering (2010), previous issues and problems have been resolved and estimates of genetic effects and generation means as well as all variance components are executed in a single step, while, the traditional method requires two analyses for means and variances, separately.
Genetic control of different traits under abiotic stress conditions has been studied in wheat. Dashti et al.
(2010) reported significant additive gene effects and high broad-sense heritability for K/Na ratio, K + , and Na + concluding existence of possibilities to improve these traits under salinity stress condition. Moroni et al. (2013) stated that additive effects were larger than dominant effects for manganese tolerance. Abbasi Holasou et al. (2019) showed that additive, dominance and epistatic effects were involved in the inheritance of agronomic traits under water deficit stress condition. Amiri et al. (2020) reported that both additive and non-additive effects are important for Fe and Zn uptake efficiency at both crosses (Marvdasht × Rassoul and Marvdasht × Shahpasand) under normal and drought stress conditions.
To the best of our knowledge, there were no reported records about gene action of zinc and Fe absorption ability under salinity stress conditions in bread wheat. Thus, the main objectives of this study were understanding the types and value of gene effects controlling the inheritance of efficient Zn and Fe uptake and related characters viz. grain yield, 100-seed weight, shoot dry weight, root dry weight, proline, Malondialdehyde (MDA), Peroxidase (POX), Catalase (CAT) by means of six basic generations (P1, P2, F1, F2, BC1, and BC2) of wheat using generation mean analysis according to mixed linear model analysis under salinity stress conditions. Ultimately, this information will be useful for selecting Zn and Fe-efficient wheat genotypes for cultivation in low-Zn and Fe soils under salinity stress, thereby providing health benefits to humans.

Plant materials and experiment
The experiment was conducted under hydroponic greenhouse condition (temperature between 28:19±2°C day:night on a 16:8-h light:dark photoperiod; humidity between 35-45%; natural light conditions) using a factorial experiment (including two factors; generations and treatments) based on randomized complete block design with three replications. The first factor was the six basic generations (P1, P2, F1, F2, BC1, BC2) derived from a cross of two Iranian facultative wheat variety; Navid (salt sensitive, and Fe and Zn-deficient) and Roshan (salt tolerant, and Fe and Zn-efficient). The second factor was the levels of NaCl salinity, including control (no NaCl), 100 and 200 mM NaCl. Parents were previously screened for salinity stress, Zn and Fe efficiency by Khoshgoftarmanesh et al. (2006a and b), Daneshbakhsh et al. (2013) and Sharifi-Soltani et al. (2016).
The seed of generations were planted in 126 plastic pots (25×25 cm) filled with perlite: vermiculite mixture (4:1 v/v). The 24 plants for each P1, P2, F1, BC1, and BC2, and 48 plants for F2 were used in each replication. Prior to sowing, qualified seeds of each generation were surface sterilized with 5% (W/V) sodium hypochlorite (NaClO) for 8 min and rinsed well with distilled water several times. To provide uniform seed germination, the seeds were sown on filter paper moistened with deionized water for 7 days in the greenhouse conditions (28/20°C day/night max/min). Upon germination, the undamaged seedlings with uniform size and uniform root length were transplanted into pots.
The pots were irrigated using Hoagland solution (Hoagland and Arnon 1950) in 1/4, 1/2 and full ratios on two, four and six days after planting, respectively. The nutrient solutions were changed frequently after every 7 days during the salinity treatment the pH was adjusted to 6.5±0.5 (EC=1.3 dsm-1). Salt treatment was applied at 17 days after sowing (DAS), when the third leaf was appeared (Zodaks score (Z) 13; Zodaks et al. 1974). To avoid osmotic shock, salt treatment was imposed stepwise in aliquots of 50 mM in Hoagland's nutrient solution every alternate daily until the appropriate salt treatments were reached. Pots were variedly irrigated from 600 to 650 ml of tap water and saline solutions every alternate period of days to achieve 100% FC (field capacity) according to weather conditions and growth stage. Avoiding excess salinity due to adding Hoagland solution, perlite substrate within the pots was washed every 14 days, and non-saline and salinity treatments were reapplied. Recorded ECs were 2.1 + 0.18, 9.2 + 0.2 and 17.2+ 0.74 for 0, 100 and 200 mM NaCl, respectively, which were weekly measured.

Phenotypic evaluation
Sixteen morpho-physiological traits including grain Fe and Zn concentration (GFeC and GZnC), root length (cm), root fresh weight, root dry weight, leaf Na + concentration (LeaNa + C), leaf K + concentration (LeaK + C), Na + /K + concentration ratio (LeaK + /Na + C), Relative water content (RWC), Electric leakage (EL), weight of 100 grains (100-SW), grain yield per plant (g/plant), proline content, Catalase (CAT), Peroxidase (POX) and Malondialdehyde (MDA) were measured after 10 weeks from sowing. The shoots were cut off at the perlite-vermiculite mixtures surface, and the roots were separated gently from the perlite-vermiculite mixtures via soaking them in deionized water for 10 min. The remaining perlite-vermiculite mixture adhering to the roots was then washed away.
Electric leakage (EL) was determined by the method of Lutts et al. (1996) using the portable electrical conductivity meter on the third leaf. Relative water content (RWC) was measured according to Barrs and Weatherly (1962). Proline concentration was assessed according to Bates et al. (1973). MDA, POX and CAT were assayed by Heath and Packer (1968), Gueta-Dahan et al. (1997) and Singh et al. (2010), respectively. Concentration of Na + and K + was measured on the second leaf after nine weeks using a flame photometer according to Poustini and Siosemardeh (2004) with minor modification, and Fe and Zn concentration was measured according to Chapman and Pratt (1961) using Atomic Absorption Spectrometer (SpectrAA-220, VARIAN, Australia).

Statistical analysis
The analysis of variance was done using the PROC GLM procedure of SAS System ver. 9.2 (SAS Institute 2007).  Broad-sense and narrow-sense heritabilities (Wright 1968) as well as degree of dominance (Halluer and Miranda 1988) were estimated using following formula: Where VG, VA, VD, and VE are genotypic variance, additive variance, dominant variance and environmental variance, respectively.

ANOVA
Analysis of variance revealed significant (p ≤ 0.01) effects of salinity and generations for all the traits indicating the presence of sufficient genetic variability for carrying out generation mean analysis and estimating heritability. The salinity × generations interaction was also significant (p ≤ 0.05) for all the traits except 100-SW, proline content, GFeC and GZnC (Table 1). The CV varied from 2.05% for RWC to 15.35% for Na + /K + ratio.
The means and standard errors of the six main generations at three salinity levels (0, 100 and 200 mM) are represented in Table 2. It is worth noting that the mean values for GFeC, GZnC, RWC, LeaNa + C, LeaK + C, LeaK + /Na + C, CAT, POX, MDA and proline in F1 was close to the average of two parents at all three levels of salinity stress, while the means for 100-SW, GY, RL, RFW and RDW in F1 was greater than that of the two parents (Table 2). These results indicated that heterosis relative to mean parents which may be used to help develop hybrid varieties. Heterosis (over best parent) for grain yield was 43.56%, 28.13% and 10.34% at control, 100 and 200 mM salinity levels, respectively. Salinity stress reduced GFeC, GZnC, 100-SW, GY, RL, RFW, RDW, RWC, LeaK + C and LeaK + /Na + C although the decrease at 200 mM was more than 100 mM. The value of EL, LeaNa + C, CAT, POX, MDA and proline were increased by imposing the salinity treatments in all generations (Table 2).

Gene effects and genetic parameters
Wald-F test revealed significant digenic epistatic interactions for majority of the traits at different salinity conditions. However, the additive-dominance model explained the genetic basis of GFeC and RL at the two higher salinity regimes (100 and 200 mM), RFW, RDW, MDA and proline under non-saline condition, RWC, EL and CAT at 100 mM NaCl and GZnC and LeaK + C at 200 mM NaCl.
Generally, the non-additive effects were greater than the additive effect. Additive effects were negative and significant for all traits at all salinity levels (P1 had smaller values than P2), except for 100-SW in control, and EL and LeaNa + C in all salinity levels. The negative additive effect indicates the inheritance of favorable alleles from P2 (Roshan) to the progenies. The same trend was observed for the dominance effect for RWC and LeaK + /Na + C at all salinity levels, for 100-SW and LeaK + C under non-saline condition and 100 mM level, for GY, MDA and RFW at 100 mM salinity level, for POX at 200 mM salinity level, and for RDW and proline at 100 and 200 mM salinity levels.
Duplicate epistatic (dominance and dominance × dominance components were in opposite directions) digenic effects were ascertained critical in the inheritance of LeaK + /Na + C (at all salinity levels), RDW, LeaNa + C, proline (under 100 and 200 mM salinity levels), 100-SW, LeaK + C (at non-saline and 100 mM salinity levels), RWC, POX (at non-saline and 200 mM salinity levels), GY, RFW, MDA (at 100 mM salinity level) and EL (at 200 mM salinity level). It was suggested that these traits were governed by complex genetic architecture. Delayed selection until the high level of gene fixation is suggested for improvement in these traits (  Table 4). Most of the studied traits exhibited high broad-sense heritabilities (more than 80%) in all three salinity levels, indicating the environmental effects constitute a minor portion of the total phenotypic variation for these traits. The narrow sense heritability ranged from 0.03 for LeaK + /Na + C (at 200 mM salinity level) to 0.69 for MDA (at 100 mM salinity level). The average dominance ratio for all studied traits was greater than unity in the three salinity levels except GFeC, RFW, LeaNa + C, POX and proline under 200 mM salinity level, GY, RWC and MDA under 100 mM salinity level and RDW under control condition, indicating the predominance of the dominant gene action for these traits.

Discussion
A good strategy to improve and sustain bread wheat production in Zn and Fe-deficient soils with minimum or no Zn and Fe application and knowledge of genetic components of tolerance have become a research priority.
Understanding the inheritance and genetic basis of Zn and Fe efficient uptake-related traits may assisted to achieve high seed yield and salinity tolerance of bread wheat.
In this study, significant differences among generations for the studied traits indicted the presence of For all the studied traits, mean effects (m) were highly significant and had more value than additive gene effects (d), which implicated on existence of common genes between the two parents and the sufficient genetic variation in these traits. Data presented here indicate that dominance effects were generally greater than additives, explaining that the studied traits were generally controlled by major genes under all salinity levels. The negative or positive sign for dominance effect is a function of F1 generation mean value in relation to the mid parent value and indicates which parent is more contributing to the dominance effects. In the current study, the negative sign of dominance effect indicates that the F1 generation was more similar to the high Zn and Fe efficient parent "Roshan", because dominance is originated from the parent containing alleles responsible for high value of the given trait.
Although this study identified the important gene effects controlling Zn and Fe efficient uptake-related traits, it should be noted that these traits are highly influenced by environmental conditions ( Our results according with some other studies (e.g. Abu et al. 2017;Ravari et al. 2017) were shown that Na + , K + and K + /Na + are important traits in relation to salinity tolerance in wheat. James et al. (2011) shown that by introgression of Nax1 and Nax2 genes from T. monococcum into hexaploid wheat, the Na + concentration is reduced in leaves and increased in leaf sheats, resulted in improving grain yield consequently. The results of the genomewide association by Genc et al. (2019) revealed Nax genes associated with high Na + accumulation, which may be involved in osmotic stress/tissue tolerance. They stated that the most modern bread wheat, especially MW#293 genotype, are efficient in excluding Na + and introduced as new paradigm in breeding for salinity tolerance. Utilizing of breeding programs in modern genotypes with high intrinsic Zn and Fe leaded to facilitation of uptaking and transporting of these mineral nutrients from roots to the seeds under salinity stress.
The derived results showed the importance of the main and epistatic effects in LeaNa + C, LeaK + C and LeaK + /Na + C inheritance and demonstrated their concern in designing of breeding programs to improve salt tolerance in wheat cultivars and populations. Wheras additive gene effects did not contribute to Na + , K + uptake and K + /Na + ratio under saline and non-saline conditions, it seems that progress in selecting for low Na + uptake is very slow. additive and dominance genetic effects involved in controlling root length, biomass, grains per spike, 100-grain weight, Na + , K + and K + /Na + ratio under both saline and non-saline conditions in bread wheat. Halward and Wynne (1991) claimed that with increasing of difference between parents, dominance and epistatic effects may play more significant roles in the inheritance of quantitative traits. In this study the parent named Navid variety with pedigree (Kirkpinar 79) 63-112/66-2×7C improved from Iran, while Roshan is an Iranian local cultivar. The pattern of gene effects for quantitative traits (especially yield and 100-SW) were similar under the non-saline and saline conditions explaining fixable and non-fixable gene actions are the basis for yield components.
In contrast, under field conditions, Ravari et al. (2017) showed that in both crosses (Kavir × Arta and Kavir × Moghan3) under acidic soil conditions, the dominance effects were more important than additive and epistatic effects in the genetic control of grain yield. The differences between those results and our study may be due to differences in the level of salt stress, different genetic backgrounds, and/or screening environment used. Concerning According to Table 4, all the traits under non-saline and saline conditions showed high broad sense heritability and low to moderate narrow sense heritability, suggests that dominance gene action is more important than additives in controlling the majority of studied traits. Selection in the early segregation generations for such traits could be misleading, therefore, required further progeny testing. Such traits could be improved by crossing potential genotypes of segregating population by means of recombinant breeding approach (Samadia 2005     6.65 2.05 7.57 6.77 6.87 ns, *, **: non-significant and significant at 0.05 and 0.01 probability levels, respectively.