Genetic Variability of Tocol Content in a Genebank Collection of Temperate Maize Inbred Lines from Southeastern Europe

Abstract

Maize is considered a promising candidate for biofortification through breeding, given its widespread cultivation and significance as a food crop. This cost-effective and sustainable approach could be used to increase the content of different tocol compounds, i.e., vitamin E, in maize grain due to the well-documented genetic variability. In the present study, an evaluation of the content of nine different tocol traits was performed in a genebank collection of 88 inbred lines of temperate maize grown at two locations in Croatia in 2019. A large genotypic variability within the studied material was observed for α-tocopherol, γ-tocopherol, δ-tocopherol, total tocopherols, α-/γ-tocopherol ratio, α-tocotrienol, γ-tocotrienol, total tocotrienols and total tocols with corresponding coefficients of variation of 52, 61, 51, 45, 106, 24, 54, 33 and 38%, respectively. Heritability estimates ranged from 0.66 for α-tocotrienol to 0.95 for γ-tocopherol. The content of α-tocopherol, which has the highest vitamin E activity and is therefore most interesting for selection, was not significantly correlated with either grain color or grain type. Comparison of the effects of simulated selection with an intensity of 20% on increased α-tocopherol content using the two selection criteria, absolute α-tocopherol content and α-/γ-tocopherol ratio, showed that the highest absolute α-tocopherol content was a better selection criterion. Indeed, simulated selection based on the absolute α-tocopherol content resulted in a 64% increase in this compound without significantly affecting the total tocopherols, the total tocotrienols, and the α-/γ-tocopherol ratio.

1. Introduction

Tocopherols and tocotrienols, collectively termed “tocochromanols” or “tocols”, are commonly known as vitamin E, an essential lipid-soluble antioxidant in the human diet [1,2]. Within both tocopherols and tocotrienols, the four different chemical species (α, β, δ, and γ) are distinguished by the number and position of the methyl groups on the chroman-6-ol ring system [2]. The vitamin E activity is usually considered as activity of RRR-α-tocopherol [3], and only natural α-tocopherol has 100% vitamin E activity. Other tocols have lower activity, namely, β-, γ-, and δ-tocopherols and α- and β-tocotrienols have 57%, 37%, 1.4%, 30%, and 5% of α-tocopherol activity, respectively, as determined by the fetal resorption-gestation test in rats [4]. The highest vitamin E activity of α-tocopherol makes its content in food especially important from a nutritional perspective [2]. Research studies have shown that vitamin E in diets provides diverse health benefits [2,5,6,7,8,9]. Although it is mostly known as a potent chain-breaking antioxidant [10], vitamin E also has non-antioxidant functions, such as cellular signaling, the regulation of gene expression, and enzymatic activity [2,11,12]. It has been also shown that adequate levels of vitamin E in animal feed are associated with various benefits in pork, beef, and poultry products [13,14,15]. A major source of vitamin E in the human diet is plant-derived products, dominantly edible oils, derived from seed oil crops. Wen et al. [16] compared the tocol contents in nine different species of first-grade traditional edible oils and found the highest content of total tocols in soybean oil, followed by cottonseed oil, corn oil, sunflower oil, rapeseed oil, rice bran oil, peanut oil, and sesame oil, and the lowest content in camellia oil.

Maize is generally characterized by inadequately low levels of vitamin E in its kernels [17], but its worldwide spread and importance as a crop make it a promising candidate crop for increasing the tocopherol and tocotrienol content by biofortification through breeding [7]. This cost-effective and sustainable approach [8,18,19,20] could provide sufficient amounts of vitamin E to people in developing countries that rely on maize grain as a staple food as well as for developed countries, for which seed oil is a primary source of vitamin E [21].

In maize, considerable variation in grain tocol content has been observed in a range of different genetic materials, ranging from elite to exotic and indigenous inbred lines from the USA, Mexico, India, China, Australia, Africa, and Europe [7,15,19,22,23,24,25,26,27,28,29,30] and experimental/commercial hybrids from India and Europe [8,9]. It has been found that tocol levels in maize kernels vary widely, with the nutritionally most interesting α-tocopherol level accounting for only about 20% and γ-tocopherol constituting the majority of the fraction [6,7,15]. Due to the relative and unique health benefits of tocopherols, Rocheford et al. [6] considered altering the overall levels of tocopherols and the ratios of α- and γ-tocopherol as valid breeding goals in maize breeding for increasing the vitamin E availability.

Studying the available genetic variation, interaction with environments, and gene action are the primary components of designing an efficient breeding strategy for the effective utilization of inbred lines in the hybrid breeding program [8]. In the previous studies, mentioned above, a large number of inbred lines from the Americas, Asia, and Africa were examined for their variation in grain tocols, but in these studies, only a limited number of inbred lines of European origin have been considered. A collection of temperate maize inbred lines with different proportions of indigenous germplasm from Southeastern Europe in their pedigrees is maintained at the University of Zagreb, Faculty of Agriculture, within the Croatian genebank [31]. These inbred lines have been previously characterized for their carotenoid profile [20]. It was considered worthy to expand the analysis of the same lines for their tocol content as well, in order to identify suitable candidates for the development of biofortified maize hybrids through breeding. The objectives of the present study were (1) to determine the variability of the grain contents of different tocol compounds and their ratios in 88 inbred lines from the Croatian genebank and (2) to evaluate the relationships between different tocols with kernel color and kernel hardness within the studied material.

2. Materials and Methods

2.1. Plant Materials

Maize inbred lines included in the present study are maintained at the Faculty of Agriculture, University of Zagreb, within the Croatian genebank [31]. The names of the 88 inbred lines together with their pedigree information as well as their kernel color and kernel type are shown in Table S1. The inbred lines have different proportions of indigenous germplasm (maize landraces and inbred lines from Southeastern Europe) in their pedigrees, including open-pollinated varieties from the first half of the 20th century [32] and obsolete inbred lines developed in the region. The rest of the material originates either from the U.S. Corn Belt or from Western and Eastern Europe.

2.2. Experimental Design and Cultural Management

A field experiment with the 88 maize inbreds was carried out in 2019 at the locations Osijek and Zagreb, situated in the eastern and western parts of the most important maize-growing region in Croatia. The soil type at Osijek was Eutric Cambisol, with a humus content of 1.93%, a pH (KCl) of 7.40, a phosphorus content of 324.5 mg kg⁻¹, and a potassium content of 270.3 mg kg⁻¹, while the soil type at Zagreb was Gleysol, with a humus content of 2.76%, a pH (KCl) of 7.48, a phosphorus content of 273.7 mg kg⁻¹, and a potassium content of 200.9 mg kg⁻¹ (Table S2). The mean air temperature during the 2019 maize growing season (from May to September) was 0.7 °C higher at the Osijek location than at Zagreb, while the total precipitation during the same period was higher at Zagreb than at Osijek (536.8 vs. 426.9 mm, respectively) (Table S3). The field experiment at both locations was set up as a randomized complete block design with two replicates. Each experimental plot consisted of a 4 m long row. The between-row spacing was 0.70 m. Machine planting was completed on 30 April 2019 at Osijek and on 26 April 2019 at Zagreb at a density of approximately 50,000 plants per hectare.

Standard agronomic practices for intensive maize production were applied at both locations. At the Osijek location, fertilization was carried out with 141 kg N ha⁻¹, 80 kg P ha⁻¹, and 120 kg K ha⁻¹ by applying 400 kg ha⁻¹ NPK (7:20:30) before plowing, 100 kg ha⁻¹ UREA (46% N) before sowing, and 250 kg ha⁻¹ CAN (27% N) as side-dressing together with inter-row cultivation. At the Zagreb location, fertilization was carried out with 151 kg N ha⁻¹, 90 kg P ha⁻¹, and 120 kg K ha⁻¹ by applying 300 kg ha⁻¹ NPK (7:20:30) before plowing, 100 kg ha⁻¹ UREA (46% N), and 200 kg ha⁻¹ NPK (15:15:15) before sowing and 200 kg ha⁻¹ CAN (27% N) as side-dressing together with inter-row cultivation. For weed control, 4 L ha⁻¹ of Lumax (Mesotrione 37.5 g L⁻¹ + Metolachlor 375 g L⁻¹ + Terbuthylazine 125 g L⁻¹) was applied immediately after planting, and 0.6 L ha⁻¹ Dicash (Dicamba 480 g L⁻¹) was applied at the 4- to 5-leaf stage at both locations. No irrigation was applied in this study.

In each plot, six to eight plants were self-pollinated to avoid xenia effects. Individual ears of the self-pollinated plants were harvested by hand on 16 October 2019 at both Osijek and Zagreb, and the grains of the harvested ears were bulked at shelling.

2.3. Data Collection

Inbreds were visually evaluated for their kernel color and kernel type on a bulk sample of 100 kernels per plot using the IBPGR maize descriptor list [33]. To avoid loss of tocols, the grain samples were stored in the dark at 4 °C until the tocols were extracted. Tocols were extracted and quantified using the method adapted to maize grain [34]. Grain samples were ground using a 0.3 mm sieve (Cyclotec 1093, Foss Tocator, Hoganas, Sweden) and 0.6 g of the sample was ultrasonicated (10 min; Sonorex TK 52, Bandelin, Berlin, Germany) and homogenized (1 min per sample; T10 Ultra-Turaxx, IKA, Staufen, Germany) with 6 mL of ethanol containing 0.1% of butylhydroxytoluene (BHT). The samples were incubated in a water bath for 5 min at 85 °C and saponified with 100 μL of 80% KOH for 10 min at 85 °C. After cooling the samples by adding 3 mL of chilled ultrapure water and placing them in an ice bath, tocols were extracted with hexane in aliquots of 3 mL. The upper hexane layer was separated after centrifugation for 10 min at 2200× g (Centric 322A, Tehtnica, Železniki, Slovenia). The extraction procedure was repeated five times. The collected supernatants were evaporated using a rotary vacuum concentrator (RVC 2-25CD plus, Martin Christ, Osterode am Harz, Germany) and dissolved in 200 µL of acetonitrile:dichloromethane:methanol (45:20:35, v/v/v) containing 0.1% BHT. Extractions were carried out under dim light, and the extracts were analyzed further using HPLC on the same day.

Tocols were separated and quantified using a SpectraSystem HPLC instrument (Thermo Separation Products, Inc., Waltham, MA, USA) equipped with a quaternary gradient pump (P4000), an autosampler (AS3000), and an FL detector (FL3000). Compounds were separated on two sequentially connected C18 reversed-phase columns: a Vydac 201TP54 column (5 μm, 4.6 mm × 150 mm; Hichrom, Reading, UK) followed by a Zorbax RX-C18 column (5 μm, 4.6 mm × 150 mm; Agilent Technologies, Santa Clara, CA, USA). The separation columns were protected by a Supelguard Discovery C18 guard column (5 μm, 4 mm × 20 mm; Supelco, Bellefonte, PA, USA). The mobile phase consisted of acetonitrile:methanol:dichloromethane (75:25:5, v/v/v) containing 0.1% BHT and 0.05% triethylamine. An aliquot of 30 μL was injected, and the flow rate was 1.8 mL/min. The separations were performed at room temperature. The tocols were monitored at an extinction of 290 nm and an emission of 330 nm. Separated compounds were identified by comparing their retention times and quantified using external standardization with calibration curves using commercially available standards (r² ≥ 0.99). Tocol standards (α-, γ-, and δ-tocopherol and α- and γ-tocotrienol; purity ≥ 93%) were obtained from Supelco (Sigma-Aldrich, St. Louis, MO, USA). The eight evaluated tocol traits in μg g⁻¹ dry seed were as follows: α-tocopherol (αT), γ-tocopherol (γT), δ-tocopherol (δT), α-tocotrienol (αT3), γ-tocotrienol (γT3), total tocopherols (ΣT, calculated as αT + γT + δT), total tocotrienols (ΣT3, calculated as αT3 + γT3), and total tocols (ΣTT3, calculated as ΣT + ΣT3). In addition, the ratio of α-/γ-tocopherol (αT/γT) was calculated.

2.4. Statistical Analysis

The analysis of variance was conducted using the PLABSTAT software, version 3A [35]. The broad-sense heritability [36] was estimated as follows:

h² = V_G/[V_G + V_G×L/l + V_e/(lr)]

where V_G, V_L×G, and Ve are components of the genotypic variance, the genotype × location interaction variance, and error variance, respectively, and l and r are the number of locations and replicates, respectively.

To compare the effects of selection based on two criteria (the level of α-tocopherol and the α- to γ-tocopherol ratio) in the set of inbred lines, we simulated the selection of the 20% best-ranked inbred lines for each of the two traits (selection intensity of 20%). The predicted response to selection (R) for a given trait was then estimated by multiplying the trait heritability (h²) by the selection differential (SD), according to Falconer and Mackay [37]:

R = h² × SD,

where SD was calculated as the difference between the trait mean in the selected group and the trait mean of the entire group of inbred lines. The same formula was used to estimate R for correlated traits (correlated response to selection). The significance of the SD was tested by one-way ANOVA using the statistical software SAS/STAT, version 9.4 (SAS Institute 2016 [38]), taking the group (entire set or selected set of inbreds) as an independent variable.

3. Results

3.1. The HPLC Profile of the Tocols

The Tocols separated using the HPLC method [34] were identified as α-tocopherol, γ-tocopherol, δ-tocopherol, α-tocotrienol, and γ-tocotrienol, and Figure 1 represents an example of such a chromatogram.

3.2. Analysis of Variance

The analysis of variance revealed a significant effect of genotype (G) for all the studied traits, a significant effect of location (L) for αT, γT, ΣT, αT3, and ΣTT3), and a significant G × L interaction for αT, αT3, and αT/γT (

Table 1. The analysis of variance for nine tocol traits in 88 maize inbred lines evaluated at two locations.

Source	df	Mean Square
		αT	γT	δT	ΣT	αT3	γT3	ΣT3	ΣTT3	αT/γT
Location (L)	1	59.7 **	156.6 *	0.0	412.7 **	1.9 *	1.7	7.3	310.6 **	0.033
Genotype (G)	87	46.2 **	545.9 **	1.0 **	587.8 **	1.1 **	11.4 **	11.5 **	598.7 **	0.944 **
G × L	87	6.4 **	29.5	0.1	35.4	0.4 *	2.1	2.6	38.6	0.083 **
Error	198	3.9	28.9	0.1	32.7	0.3	1.7	2.0	35.6	0.037
Heritability		0.86	0.95	0.90	0.95	0.66	0.82	0.77	0.94	0.91

*, ** mean square significant at p < 0.05 and p < 0.01, respectively. The acronyms αT, δT, γT, ΣT, αT3, γT3, ΣT3, ΣTT3, and αT/γT indicate α-tocopherol, δ-tocopherol, γ-tocopherol, total tocopherols, α-tocotrienol, γ-tocotrienol, total tocotrienols, total tocols, and α-/γ-tocopherol ratio, respectively.

). The broad-sense heritability for the three tocopherol compounds varied from 0.86 for α-tocopherol to 0.95 for γ-tochopherol. For the two tocotrienol compounds, the heritabilities were of a lower magnitude than for the tocopherols, reaching 0.66 for α-tocotrienol and 0.82 for γ-tocotrienol. A similar trend was observed for the heritability of the total tocopherols and total tocols versus the total tocotrienols, 0.95 and 0.94 vs. 0.77, respectively.

3.3. Means and Variation in Tocol Contents

The mean values of tocopherol and tocotrienol compounds together with their total values for the two locations are shown in Figure 2. Significant differences between locations were found for α-tocopherol, γ-tocopherol, total tocopherols, α-tocotrienol, and total tocols (8.72 vs. 9.89, 26.28 vs. 28.17, 36.36 vs. 39.42, 3.07 vs. 2.86, and 43.93 vs. 46.59 μg g⁻¹ for Osijek and Zagreb, respectively). The standard deviations in Figure 2 indicate high variability in tocol compound contents among the studied 88 maize accessions.

The levels of tocol compounds in 88 maize inbred lines averaged over the two locations showed considerable and significant variation (

Table 2. Means and variations of nine tocol traits in 88 maize inbred lines across two locations.

	Absolute Value				Relative Value
Trait	Mean	STDEV	Range	CV	% of ΣT	% of ΣT3	% of ΣTT3
		μg g−1		%
αT	9.31	4.8	1.04–26.95	52	30		23.2
δT	1.36	0.7	0.40–3.98	51	4		3.1
γT	27.23	16.5	2.93–73.63	61	66		55
ΣT	37.9	17.1	12.35–79.70	45			81.3
αT3	2.96	0.7	1.60–5.44	24		43.5	8
γT3	4.41	2.4	1.29–12.47	54		56.5	10.7
ΣT3	7.37	2.4	3.60–15.60	33			18.7
ΣTT3	45.26	17.3	18.15–85.20	38
		μg g−1/μg g−1		%
αT/γT	0.65	0.69	0.03–3.86	106

STDEV: standard deviation; CV: coefficient of variation. The acronyms αT, δT, γT, ΣT, αT3, γT3, ΣT3, ΣTT3, and αT/γT stand for α-tocopherol, δ-tocopherol, γ-tocopherol, total tocopherols, α-tocotrienol, γ-tocotrienol, total tocotrienols, total tocols, and α-/γ-tocopherol ratio, respectively. The relative values of tocol traits were expressed as means of the relative values of the individual 88 inbred lines.

, Figure 3). Of the five tocol compounds, the most abundant was γ-tocopherol, representing 66% of the total tocopherols and 55% of the total tocols and ranging from 2.93 to 73.63 μg g⁻¹ (Table 2). The second most abundant was α-tocopherol, representing 30% of the total tocopherols and 23.2% of the total tocols and ranging from 1.04 to 26.95 μg g⁻¹. The least abundant was δ-tocopherol, averaging only 4% of the total tocopherols and 3.1% of the total tocols. The two tocotrienol compounds, α-tocotrienol and γ-tocotrienol, represented 8.0% and 10.7% of the total tocols, respectively, with γ-tocotrienol having almost a three-fold range of the α-tocotrienol ratio. Tocopherols were, in their total tocol amount, dominantly represented with 81.3% versus 18.7% of the tocotrienols. The α-/γ-tocopherol ratio ranged from 0.03 to 3.86 with a mean of 0.65. The coefficients of variation for the nine tocol traits ranged from 24% (α-tocotrienol) to 106% (α-/γ-tocopherol ratio).

Maize inbreds 36, 49, 50, 67, and 82 had the highest mean total tocol contents of more than 75 μg g⁻¹ (Figure 3). Three of them (36, 50, and 82) were among the best five ranked inbreds for the content of total tocols at both Osijek and Zagreb, showing high stability for the total tocol content across locations (Table S4). The inbreds 36, 50, and 82 were also characterized with the highest γ-tocopherol levels at both locations with overall means across locations of more than 60 μg g⁻¹, which indicates that high levels of tocopherols dominantly contributed to the total tocol levels in these inbreds. The inbred 67 was on the other hand characterized by a relatively high level of total tocotrienols (13.87 μg g⁻¹), which in its case, also contributed to the high total tocol levels (77.64 μg g⁻¹). In five inbreds, 15, 46, 52, 73, and 88, more than two-fold higher levels of α-tocopherol vs. γ-tocopherol were determined, but the same lines were characterized with below-average total tocopherol and total tocol contents. The inbred 73 had the highest α-tocopherol content in the set, 26.95 μg g⁻¹, which is 7.59 μg g⁻¹ more than the second highest ranking inbred in α-tocopherol content, inbred 9. These inbreds were best ranked at both Osijek and Zagreb, showing a high stability for α-tocopherol content across locations. In the inbreds 11, 39, 41, 46, and 52, more than two-fold higher contents of α- vs. γ-tocotrienol were determined, but the total tocotrienol contents determined in these lines were also below average or average (Figure 3, Table S4).

3.4. Correlations between Grain Tocol Traits

The Pearson correlation coefficients between the studied grain tocol traits are presented in

Table 3. The Pearson correlation coefficients between the tocol traits.

Trait	αT	γT	αT/γT	δT	ΣT	αT3	γT3	ΣT3
γT	−0.11
αT/γT	0.45 **	−0.67 **
δT	−0.22 *	0.71 **	−0.39 **
ΣT	0.17	0.96 **	−0.54 **	0.66 **
αT3	0.44 **	−0.49 **	0.60 **	−0.33 **	−0.37 **
γT3	−0.34 **	0.21	−0.33 **	0.20	0.11	−0.14
ΣT3	−0.21	0.05	−0.14	0.10	−0.01	0.17	0.95 **
ΣTT3	0.14	0.96 **	−0.55 **	0.67 **	0.99 **	−0.34 **	0.24 *	0.14

*, ** indicate the significant correlation coefficients at p < 0.05 and p < 0.01, respectively; The acronyms αT, γT, αT/γT, δT, ΣT, αT3, γT3, ΣT3, and ΣTT3 indicate α-tocopherol, γ-tocopherol, α-/γ-tocopherol ratio, δ-tocopherol, total tocopherols, α-tocotrienol, γ-tocotrienol, total tocotrienols, and total tocols, respectively.

. The α-tocopherol content was negatively correlated with the δ-tocopherol content (r = −0.22) and the γ-tocotrienol content (r = −0.34) but moderately positively correlated with the α-tocotrienol content (r = 0.44). The correlation coefficients between the γ-tocopherol content and the δ-tocopherol content, total tocopherol content, and total tocol content were very strong and positive (0.71, 0.96, and 0.96, respectively), which reflects the dominant γ-tocopherol’s share in the total tocol contents of the studied inbred lines. The correlation coefficient between the α-tocopherol content and γ-tocopherol content was not significant, as were the correlation coefficients between the two tocotrienol compounds, α-tocotrienol and γ-tocotrienol, and between the total tocopherols and total tocotrienols. On the other hand, the correlation coefficients between the total tocol content and all other compounds, except α-tocopherol, were positive and ranged from weak to very strong except with α-tocotrienol, which was negatively moderate (r = −0.34). The α-/γ- tocopherol ratio was in moderate positive correlation with the content of α-tocopherol and α-tocotrienol but in a moderate negative correlation with all other tocol compounds as well as with total tocopherols and tocols.

3.5. Simulation of Selection within 88 Inbred Lines

The results of simulated selection in the set of 88 maize inbred lines based on two selection criteria (an increased α-tocopherol content and higher α-/γ-tocopherol ratio) are shown in

Table 4. The results of the simulated selection of 20% intensity in maize inbred lines based on two selection criteria (SC) showing the population mean before selection (Mean all) and after selection (Mean selected) together with the selection differential (SD) and response to selection (R).

SC	Parameter	αT	γT	δT	ΣT	αT3	γT3	ΣT3	ΣTT3	αT/γT
		μg g−1
	Mean all	9.31	27.23	1.36	37.89	2.96	4.41	7.37	45.26	0.65
αT	Mean selected	16.20	28.96	1.23	46.39	3.22	3.06	6.28	52.67	0.87
	SD	6.89 **	1.73	−0.13	8.50	0.26	−1.35 *	−1.09	7.41	0.22
	R	5.94	1.64	−0.12	7.99	0.17	−1.10	−0.84	6.93	0.20
	R (%)	64	6	−9	21	6	−25	−11	15	31
αT/γT	Mean selected	11.18	7.88	0.98	20.04	3.68	3.38	7.06	27.11	1.75
	SD	1.87	−19.35 **	−0.38 **	−17.85 **	0.72 *	−1.03	−0.31	−18.15 **	1.10 **
	R	1.61	−18.30	−0.34	−16.77	0.47	−0.84	−0.24	−16.98	1.01
	R (%)	17	−67	−25	−44	16	−19	−3	−38	155

*, ** indicate the significant SD at p< 0.05 and p < 0.01, respectively; The acronyms αT, γT, δT, ΣT, αT3, γT3, ΣT3, ΣTT3, and αT/γT indicate α-tocopherol, γ-tocopherol, δ-tocopherol, total tocopherols, α-tocotrienol, γ-tocotrienol, total tocotrienols, total tocols, and α-/γ- tocopherol ratio, respectively; R (%) indicates the response to selection expressed relative to the trait mean in the population before selection.

and Table S5. The mean values of the tocol traits were reported for the population before selection (all 88 inbred lines), while for the selected populations, the mean values of the 18 best-ranked inbred lines (20% selection intensity) were reported together with the two selection parameters selection differential (SD) and response to selection (R). When selection was based on the content of α-tocopherol, the content of this compound increased significantly by 6.89 μg g⁻¹ in the selected group compared to the base population. This was accompanied by a significant decrease in the γ-tocotrienol content in the selected group, while for the other individual tocols, the total tocol traits as well as the α- to γ-tocopherol ratio, the difference between the selected and the entire group was not significant. Therefore, selection for a higher absolute level of α-tocopherol is expected to increase this compound by 64% (relative to the mean in the population before selection) without significantly affecting the levels of the other traits, except the γ-tocotrienol content, which would decrease by 25%. On the other hand, selection for a higher α-/γ-tocopherol ratio would increase this ratio by 155%, and the content of α- tocotrienol by 16%, while the content of γ-tocopherol, δ-tocopherol, total tocopherols, and total tocols, would decrease by 67, 25, 44, and 38%, respectively.

3.6. Correlations of Tocol Content with Kernel Color and Kernel Hardness

The distributions of the content of different tocols in maize inbred lines, grouped by the kernel color (PY = pale-yellow, Y = yellow, O = orange, DO = deep orange) together with the Pearson correlation coefficients between the kernel color intensity and the tocol content are presented in Figure 4. A weak positive correlation was found between the intensity of the kernel color and the content of δ-tocopherol (r = 0.21), γ-tocopherol (r = 0.22), total tocopherols (r = 0.19), and total tocols (r = 0.20). The content of the most abundant γ-tocopherol was 62% higher in the deep-orange color class than in the pale-yellow color class, while the increase in total tocopherols and total tocols from the pale-yellow class to the deep-orange class was 40% and 37%, respectively. However, a large variation within the color classes resulted in relatively weak correlations between the kernel color and tocol content. On the other hand, the correlation between the color class and the content of α-tocopherol, α-tocotrienol, γ-tocotrienol, and total tocotrienols was negligible with correlation coefficients between 0.00 and 0.10/0.10.

The distributions of the content of different tocols in the kernels of maize inbred lines, grouped by kernel type (D = dent, S-D = semi-dent, S-F = semi-flint, F = flint) together with the Pearson correlation coefficients between the kernel hardness and tocol content are shown in Figure 5. A significant, weak positive correlation was observed between the kernel hardness and δ-tocopherol content (r = 0.26), which was of the same order of magnitude as the observed correlation between the kernel color intensity and tocol content (Figure 4). On the other hand, the correlations of γ-tocopherol, total tocopherols, and total tocols with kernel hardiness were stronger, with correlation coefficients of 0.36, 0.37, and 0.38, respectively (Figure 5). In all cases where there was a significant correlation between the tocol content and kernel hardness, the mean tocol content was generally lower in the dent and semi-dent classes than in the semi-flint and flint classes. For example, the content of the most abundant γ-tocopherol was 40% higher in the dent color class than in the flint color class, while the increase in total tocopherols and total tocols from the dent to the flint class was 38% and 37%, respectively. Similarly to the color classes, the large variation within the kernel-type classes observed for all tocols resulted in relatively weak correlations between the kernel hardness and tocol content.

4. Discussion

In our study, a significant effect of genotype (G) was found for all the tocol traits examined, and location (L) was significant for five of the nine traits examined (α-tocopherol, γ-tocopherol, total tocopherols, α-tocotrienol, and total tocols), and the G × L interaction was significant for only three out of the nine traits examined (α-tocopherol, α-tocotrienol, and α-/γ-tocopherol ratio). Previous studies have also shown that the effect of genotype is much more important than that of environment and the genotype × environment interaction for different tocol compounds in maize inbred lines [29] as well as in hybrids [8,19,39]. The heritabilities of the analyzed tocopherols in the present study varied from 0.86 for α-tocopherol to 0.95 for γ-tocopherol and were higher than the heritabilities for tocopherol compounds found in the study of Shutu et al. [24]. For total tocopherols and total tocols, the observed heritability was very high (0.95 and 0.94, respectively), which is in agreement with Li et al. [23] and Lipka et al. [7], suggesting that the variation for the studied tocols was mainly influenced by genetic rather than environmental effects.

In the present study, a wide genetic variation for different tocol compounds was found in a set of 88 maize accessions, with γ-tocopherol being the most abundant and responsible for more than 50% of the total tocols. The observed ranges of tocol compounds in our study are in general accordance with similar studies in maize [7,8,22,29,39,40,41,42], having in mind different genetic backgrounds, origins, and the types of maize studied but also the different environmental conditions in the studies, including the weather conditions, soil types, and cultural practices applied. The second most abundant tocol compound was α-tocopherol, which in our study showed no or a weak negative correlation with δ-, γ-, total tocopherols, and total tocols, which is in agreement with the findings of Goffman and Böhme [39], Muzhingi et al. [19], and Lipka et al. [7] but in disagreement with Li et al. [23], where α-tocopherol was positively correlated with δ-, γ-, and total tocopherols. In our study, α-tocopherol was moderately positively correlated only with its tocotrienol variant, which is also consistent with the previous studies. These observed low correlations between α-tocopherol and other tocopherol and tocotrienol compounds suggest that independent genetic manipulation of α-tocopherol, being the tocol compound with the highest vitamin E activity, is feasible [2,7]. In our study, the total tocopherols were not correlated with the total tocotrienols, while the correlation between the total tocopherols and the total tocols was almost perfect. On the contrary, the correlation between the total tocotrienols and the total tocols was negligible, which is in agreement with Lipka et al. [7] and Diepenbrock et al. [25], suggesting that the synthesis of these two tocol classes is independently regulated.

We identified several maize inbreds, which produced two- or even three-fold higher levels of α-tocopherol than γ-tocopherol. The same lines were below average or average in the levels of the total tocopherols and total tocols. Rocheford et al. [6] also identified a maize hybrid with more α-tocopherol than γ-tocopherol. The identified hybrid had more than 100 μg g⁻¹ less total tocopherols than the hybrid with the highest level of α-tocopherol (114.5 vs. 231.7 μg g⁻¹) in their study. Therefore, according to the same authors, if increasing the α-tocopherol content is the primary goal, then hybrids should be selected for the highest absolute level of α-tocopherol and not for the best α-/γ-tocopherol ratio. In the present study, we compared the effect of simulated selection for increased levels of α-tocopherol and a higher α-/γ-tocopherol ratio and came to the same conclusion. Namely, selection based on the absolute level of α-tocopherol was expected to increase this compound by 64% without significantly affecting the levels of the total tocopherols, the total tocotrienols, the α-/γ-tocopherol ratio, and the levels of most other tocols. On the other hand, selection for a higher α-/γ-tocopherol ratio would increase this ratio by 155% but with negative effects on the level of the total tocopherols and total tocols, mainly due to a significant decrease in γ-tocopherol and an unchanged α-tocopherol content. The observed decrease in total tocopherols and total tocols after selection for a higher ratio of α- to γ-tocopherol is to be expected, considering the moderate negative correlations between this ratio and the two traits observed in the present study as well as in a previous study by Lipka et al. [7].

In the present study, the color intensity of the kernel was weakly positively correlated with the content of δ-tocopherol, γ-tocopherol, total tocopherols, and total tocols, with correlation coefficients ranging from 0.19 to 0.21. The mean values of tocols were generally higher in the deep-orange and orange color classes than in the yellow and pale-yellow color classes. Previously, Hossain and Jayadeep [43] reported a higher level of the total tocols in red maize than in yellow maize, and Suriano et al. [44] found the highest total tocol content in red maize, followed by yellow and purple, and the lowest in blue maize. However, in these studies, the color class was represented by only one or two genotypes [43] or by a single genotype [44]. Consistent with our results, Mladenović Drinić et al. [28] reported that 37 maize inbred lines with orange kernels had a higher mean value of β + γ tocopherols compared to 29 lines with yellow and 4 with white kernels. Similarly, in the study by Das et al. [29], 44 maize inbreds with yellow/orange-colored endosperm had significantly higher γ-tocopherol and total tocopherol contents than 10 maize inbreds with white endosperm types. The weak correlation found in the present study between the content of tocols and the kernel color is to be expected knowing that tocols are predominantly present in the maize germ [45], while kernel color is related to the carotenoids found in the aleurone, phlobaphenes that accumulate in the pericarp, and anthocyanins found in the pericarp and in small amounts in the aleurone [46].

In the present study, a significant weak positive correlation (r = 0.26) was found between the kernel hardness and δ-tocopherol content and was of the same order of magnitude as that between the content of different tocols and the intensity of the kernel color (r from 0.19 to 0.22). The correlation between the kernel hardness and γ-tocopherol, the total tocopherols and total tocols, was slightly stronger with correlation coefficients of 0.36, 0.37, and 0.38, respectively. In fact, since tocols are predominantly present in the maize germ [45], while kernel hardness in maize reflects the proportion of vitreous hard endosperm in the kernel [47,48], a relatively weak correlation was found in the present study between the kernel hardness and the content of different tocols was expected. Some previous studies in maize have investigated the relationship between the kernel hardness and kernel carotenoid content [20,47,48], but studies reporting the association between the kernel hardness and tocol content are lacking. In these studies, kernel hardness was generally associated with increased levels of total carotenoids, but the carotenoid profile was also influenced by the kernel type [47,48], with flints having higher levels of β-branch carotenoids, while dents had higher levels of α-branch carotenoids. The correlations between the kernel hardness and the content of different tocols in the present study are slightly stronger than the correlations between the kernel hardness and the content of β-branch carotenoids (zeaxanthin, β-criptoxanthin, and β-carotene) in the study of Šimić et al. [20], who analyzed the same set of inbred lines used in the present study. Nevertheless, the magnitude of the correlations between the tocol content and kernel color and hardness found in the present study was too small to have significant effects on selection. Even considering the highest correlation found in the present study, the one between the kernel hardness and total tocol content (r = 0.38), the corresponding coefficient of determination (r²) of 0.14, indicates that only 14% of the total variability between inbreds for tocols can be explained by kernel hardness. In addition, the content of α-tocopherol, which has the highest vitamin E activity and is, therefore, most interesting for selection, was not significantly correlated with either kernel color or kernel type.

5. Conclusions

Our results revealed a wide genotypic variability for different tocol compounds within the 88 studied maize genebank accessions, with γ-tocopherol being the most abundant compound responsible for more than 50% of the total tocols. Three inbred lines (36, 50, and 82) were among the five best ranked inbred lines in the γ-tocopherol and total tocol content at both locations of the study, while inbred lines 73 and 9 were the two best ranked in terms of the α-tocopherol content, also at both locations. The biofortification of individual and total tocopherols and tocotrienols as well as total tocols through breeding in the studied set of inbred lines appears to be feasible due to the high levels of heritability observed for all studied tocol compounds. The content of α-tocopherol, which has the highest vitamin E activity and is, therefore, the most interesting for selection, was not significantly correlated with either kernel color or kernel type in the studied genetic material. Using the absolute α-tocopherol level as a selection criterion for increasing α-tocopherol, total tocopherol, and total tocols seems to be the most promising strategy for breeding.